Automation Solutions for Analytical Measurement: Theory, Concepts, and Applications 9783527342174, 9783527805396, 9783527805327, 9783527805389, 9783527805297, 3527342176

Table of contents :
fmatter......Page 2
ch1......Page 12
ch2......Page 35
ch3......Page 82
ch4......Page 101
ch5......Page 175
ch6......Page 212
ch7......Page 237
index......Page 260

Citation preview

Automation Solutions for Analytical Measurements

Automation Solutions for Analytical Measurements Concepts and Applications

Heidi Fleischer Kerstin Thurow

Authors Priv.-Doz. Dr.-Ing. habil. Heidi Fleischer University of Rostock Institute of Automation Richard-Wagner-Straße 31 18119 Rostock Germany Prof. Dr.-Ing. habil. Kerstin Thurow University of Rostock Center for Life Science Automation Friedrich-Barnewitz-Straße 8 18119 Rostock Germany Cover The material was kindly provided by the authors. © celisca 2016 Background image: © Sasa Prudkov

All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at . © 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Boschstr. 12, 69469 Weinheim, Germany All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Print ISBN: 978-3-527-34217-4 ePDF ISBN: 978-3-527-80539-6 ePub ISBN: 978-3-527-80532-7 Mobi ISBN: 978-3-527-80538-9 oBook ISBN: 978-3-527-80529-7 Cover Design Bluesea Design, McLeese Lake, Canada Typesetting SPi Global, Chennai, India Printing and Binding Printed on acid-free paper

v

Contents Preface ix 1 1.1 1.1.1 1.1.2 1.2 1.3 1.3.1 1.3.2 1.3.3 1.4 1.4.1 1.4.2 1.4.3

Introduction 1

Life Sciences – A Definition 1 A Short Definition of Life 1 What Is Life Sciences? 2 Automation – A Definition 4 History of Automation 5 Automation from the Beginnings to the Nineteenth Century 5 Automation Since the Nineteenth Century 10 History of Laboratory Automation 12 Impact of Automation 15 Advantages and Disadvantages of Automation 15 Social Impact of Automation 16 Limitation of Automation 17 References 18

2

Automation in Life Sciences – A Critical Review 25

2.1 2.2 2.3 2.3.1 2.3.2 2.3.2.1 2.3.2.2 2.3.2.3 2.3.2.4 2.3.3 2.4 2.4.1 2.4.2 2.4.2.1 2.4.2.2 2.4.2.3 2.4.3

Overview 25 Definitions and Basics 26 Automation in Bioscreening 28 Overview 28 Automation Devices in Biological Screening 31 Standardization of Sample Formats 31 Robots in Bioautomation 33 Liquid-Handling Systems 34 Additional Components 37 Application Examples 40 Automation in Chemical Sciences 43 Overview 43 Automation Devices for Combinatorial Chemistry 45 Vessel Based Systems 46 Microplate-Based Systems 48 Robot-Based Synthesis Systems 48 Application Examples 49

vi

Contents

2.5 2.5.1 2.5.2 2.5.3 2.6 2.6.1 2.6.1.1 2.6.1.2 2.6.1.3 2.6.2

Automation in Analytical Measurement Applications 51 Overview 51 Process Analytical Technology 52 Automation Systems for Analytical Measurement Applications 54 Requirements for Automating Analytical Processes 56 Bioscreening vs. Analytical Measurement 56 Vessels and Vials in Analytical Processes 56 Liquids and Reagents in Analytical Measurement 58 Process Structure 58 Automation Requirements 58 References 61

3

Automation Concepts for Life Sciences 73

3.1 3.2 3.3 3.3.1 3.3.2 3.3.2.1 3.3.2.2 3.3.3 3.4 3.5

Classification of Automation Systems 73 Classification Concept for Life Science Processes 75 Robot Based Automation Systems 78 Robot Based Systems in Industrial Automation 78 Robot-Based Automation Systems in Life Sciences 79 Concept of the Central Robot as System Integrator 79 Concept of the Flexible Robot 80 Summary and Application of Concepts 81 Degree of Automation 83 Statistical Evaluations 86 References 89

4

Automation Systems with Central System Integrator 93

4.1 4.1.1 4.1.2 4.1.3 4.1.4 4.1.5 4.1.6 4.2 4.2.1 4.2.1.1 4.2.1.2 4.2.2 4.2.2.1 4.2.2.2 4.2.3 4.2.4 4.2.4.1

Centralized Closed Automation System 93 Background and Applicative Scope 93 Automation Goals 98 System Design 99 Process Description 102 Control of the Automation Process 103 Evaluation of the Automation System 104 Centralized Open Automation System 109 Background and Applicative Scope 109 Determination of Mercury in Waste Wood 109 Determination of Methacrylates in Dental Materials 111 Automation Goals 114 Determination of Mercury in Waste Wood 114 Determination of Methacrylates in Dental Materials 115 System Design 116 Process Description 121 Process Description for Determination of Mercury in Waste Wood 121 Process Description for the Determination of Methacrylates in Dental Materials 122

4.2.4.2

Contents

4.2.5 4.2.6 4.3 4.3.1 4.3.2 4.3.3 4.3.4 4.3.5 4.3.6 4.4 4.4.1 4.4.2 4.4.3

Control of the Automation Process 124 Evaluation of the Automation System 126 Decentralized Closed Automation System 130 Background and Applicative Scope 131 Automation Goals 132 System Design 134 Process Description 135 Control of the Automation Process 136 Evaluation of the Automation System 136 Decentralized Open Automation System 143 System Design 144 Process Description 144 Control of the Automation System 145 References 148

5

5.1 5.1.1 5.1.2 5.1.3 5.1.4 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.2.5 5.2.6 5.3 5.3.1 5.3.2 5.3.3 5.4 5.4.1 5.4.2 5.4.3

167 Centralized Closed Automation System 167 System Design 167 Process Description 174 Control of the Automation System 174 Results 179 Centralized Open Automation System 180 Background and Applicative Scope 180 Automation Goals 183 System Design 184 Process Description 186 Control of the Automation System 187 Results 189 Decentralized Automation System 191 System Design 192 Process Description 193 Control of the Automation System 193 Automation Systems with Integrated Robotics 194 System Design 196 Process Description 198 Process Control 198 References 200

6

Automated Data Evaluation in Life Sciences 205

6.1 6.2 6.3 6.4 6.4.1 6.4.2 6.5 6.6

Specific Tasks in Data Evaluation in Analytical Measurements 205 Automation Goals 207 System Design 208 System Realization 211 Software Structure 211 Software Operation 214 Process Description 220 Application Examples 222

Automation Systems with Flexible Robots

vii

viii

Contents

6.6.1 6.6.2 6.6.3

Automated Data Analysis in the Elemental Analysis 222 Automated Data Analysis in the Structural Analysis 224 Automated Data Analysis in Special Applications 225 References 226

7

Management of Automated Processes 231

7.1 7.2 7.3 7.3.1 7.3.2 7.3.3 7.4 7.4.1 7.4.2 7.4.3 7.4.4 7.4.5 7.4.6 7.4.7

Laboratory Information Systems 231 Laboratory Execution Systems 231 Process and Workflow Management Systems 232 Overview 232 Intelligent Scheduling 234 Human Machine Interaction 236 Business Process Management Systems 239 Initial BPM Activities 239 Relationship to Scientific Workflow Management 241 Life Science Automation Industry Application of BPM 241 Status of Life Science Automation 242 Laboratory IT Integration Status 245 Innovation in End-to-End Process Automation 245 Workflow Automation as a New Top-Level Process Automation Approach 246 Outstanding Position of LIMS as an Established Process Documentation System 248 References 249

7.4.8

Index 255

ix

Preface Automation systems with applied robotics have already been established in industrial applications for many years. In the field of life sciences, a comparable high level of automation can be found in the areas of bioscreening as well as high-throughput screening. Strong deficits still exist in the development of flexible and universal fully automated systems in the field of analytical measurements. Reasons are the heterogenous processes with complex structures, which include sample preparation and transport, analytical measurements using complex sensor systems as well as suitable data analysis and evaluation. Furthermore, the use of non-standard sample vessels with various shapes and volumes results in an increased complexity. The state of the art includes automated workstations, semi-automated systems, or proprietary fully automated systems, which have been developed for specific applications. In general, a flexible use of automation systems for different applications is a challenging scientific task. The development of appropriate automation systems in the field of analytical measurements using analytical instruments and complex sensor systems initially requires a systematic analysis of the processes to be automated with the aim to develop suitable structures and allocate them to these processes. In industrial applications, eight different structures can be distinguished according to their centralized or decentralized process structure, local, and functional structure. In analytical measurement technology, there are limitations regarding a general applicability of these structures, thus, an adequate adaption is required. Analytical processes are always characterized by a decentralized process structure. This enables a distinction according to their local and applicative structure. Depending on the robot technology used, two basic automation concepts can be applied to processes in analytical measurements: central system integrators and flexible robots. For a maximum versatility of the processes to be automated an extension to a third concept – integrated robotics – is possible. Due to their high flexibility, robots can be used as transport systems. This enables a connection of the individual subprocesses and workstations, whereby the robot has the function of a central system integrator. A higher flexibility of an automation system can be achieved when, besides transportation tasks, the robot additionally performs active manipulation tasks, whereby the robot has

x

Preface

the function of a flexible robot. A further increase in flexibility can be achieved using mobile robots, which perform both, transportation tasks between various subsystems and manipulation tasks. For an efficient workload of such robots, some of these tasks can be performed even during the transport. This book will provide a substantial contribution to the development and systematization of appropriate automation systems in the life sciences, in particular, in the field of analytical measurement technique. The first chapter gives a widespread overview about the history and the impact of automation systems in the field of life sciences. The second chapter involves a critical review of existing automation systems in bioscreening, chemical sciences, and analytical measurement applications. The chapter begins with general definitions and basics and concludes with the requirements for automating analytical measurement processes. The third chapter is particularly dedicated to the theoretical view on automation structures and presents general automation concepts for analytical measurement processes. The theoretical considerations are completed with delineations regarding the degree of automation and statistical evaluations. The fourth and fifth chapters present realized automation concepts with a central system integrator and a flexible robot. Therefore, special applications from various areas are introduced. This includes applications in environmental measuring technology, medicine, drug development, and drug discovery as well as quality assurance. The goal is to achieve a high degree of automation with maximum sample throughput, short processing, and measurement times with a special focus on the applicative flexibility of the automated systems. The systems are described in detail and the evaluation is done on both, the process performance and the measurement results achieved. The sixth chapter is related to the software development for automated data evaluation. The challenge was developing a flexible solution, which enables the integration of several analytical measurement instruments from different manufacturers to ensure a fully automated process, including the sample preparation, the measurement, and the final data evaluation. The last chapter is dedicated to the high-level management of automated processes and discusses several management systems used in the field of laboratory automation. The authors would like to express their personal thanks to Prof. Dr.-Ing. Norbert Stoll for his support and valuable discussions. Our special thanks go to the company Yaskawa, especially Dr.-Ing. Michael Klos and B.Eng. Wolfgang Schuberthan for providing the dual-arm robot SDA10F and for the support in generating the robot jobs. We would like to acknowledge our thanks to the Federal Ministry for Education and Research (BMBF) for partially supporting several projects. For the realization of the automation systems in detail, we thank the members of the following research groups at CELISCA (Center for Life Science Automation) at the University of Rostock: research group “Life Science Automation – Systems” under the guidance of Dr.-Ing. Steffen Junginger, research group “Life Science Automation – Mobile Robotics” under the guidance of Priv.-Doz. Dr.-Ing. habil. Hui Liu, research group “Life Science

Preface

Automation – Process IT” under the guidance of Dr.-Ing. Sebastian Neubert, and research group “Life Science Automation – Processes” under the guidance of Priv.-Doz. Dr.-Ing. habil. Heidi Fleischer. Finally, we wish to thank all the students for their contributions within the scope of their bachelor and master theses. We wish all users of this book an interesting and informative read. December 2016 Rostock, Germany

Heidi Fleischer Kerstin Thurow

xi

1

1 Introduction 1.1 Life Sciences – A Definition 1.1.1 A Short Definition of Life

The term life sciences is ubiquitously integrated into our everyday life. It has become a standard expression. To understand the content, challenges, and tendencies of life sciences, it is necessary to define the term life. Today, there exist more than 50 different definitions depending on the scientific field and the strategic focus. In general, life can be defined as a characteristic property, which separates living matter from inorganic matter. The main characteristics include the exchange of matter and energy, reproduction, and growth. The definition of the term life in philosophy also follows these criteria [1]. Aristotle differentiated three levels of life in a hierarchical order. The lowest level included plants, whose life is characterized only by nutrition and reproduction. The next level included animals, which have the additional features of sensory perception and movement. The human, whose life is, besides the fundamental functions, characterized by thinking processes, is the highest level in Aristotle’s hierarchy. In the western philosophy of the modern era, two contrary general opinions developed: mechanism [2] and vitalism [3]. Promoters of mechanism explain life processes from the concept of physical laws of movement. The living organism is considered a machine. Main supporters of this idea were William Harvey (1578–1657), Rene Descartes (1596–1650), and Wilhelm Roux (1850–1924). In contrast to this idea, vitalism proposed a significant difference between organic and inorganic matter, whereby life is connected to organic compounds. A targeted living power (vis vitalis) characterizes all living matter. Main supporters of vitalism include Jan Baptist van Helmont (1577–1644), Georg Ernst Stahl (1660–1734), Albrecht von Haller (1708–1777), and Johann Friedrich Blumenbach (1752–1840). Since the synthesis of urea by Friedrich Wöhler (1800–1882), this approach was deprecated, since it could be shown that no special living power is required for the synthesis of organic compounds. A combination of mechanism and vitalism is the organicism [4]. This approach explains processes of life using science principles from physics and chemistry. Living organisms are supposed to have properties that cannot be found in inorganic matter.

Automation Solutions for Analytical Measurements: Concepts and Applications, First Edition. Heidi Fleischer and Kerstin Thurow. © 2018 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2018 by Wiley-VCH Verlag GmbH & Co. KGaA.

2

1 Introduction

Supporters of this idea introduced the hypothesis, that biological questions can only be answered considering the whole organism. Main supporters of the organicism were William Emerson Ritter (1856–1944) [5], Karl Friederichs (1831–1871) [6–8], and August Thienemann (1882–1960) [9]. In natural sciences, the term life describes characteristic properties that define living organisms. This includes the exchange of energy, matter and information, growth, reproduction as well as reactions toward changing environmental conditions. RNA and DNA are supposed to be the main building blocks of life. This general definition of life was already defined by natural scientists of earlier centuries. Carl-Friedrich Kielmeyer (1765–1844) defined sensibility, irritability, reproduction force, secretion force, and propulsion force as important criteria of life [10]. As an early evolution scientist, he described long before Charles Darwin (1809–1882) own ideas regarding the evolution of living organisms [11]. Perception, reproduction, movement, nutrition, and growth were the main characteristics of life according to the German zoologist and philosopher Ernst Haeckel (1834–1919) [12]. For Francis Crick (1916–2004), who was awarded the Nobel Prize for the discovery of the DNA in 1962 together with James Watson (born 1928), self-reproduction, genetics, evolution, and metabolism were the main characteristics of life. In 1944, the well-known physicist and philosopher Erwin Schrödinger (1887–1861), who is considered the father of quantum physics, introduced in his work “What is Life?” [13] the term negentropy, which had great influence on the further development of molecular biology. Supporters of this idea tried to explain biological topics with physical sciences and started to focus on the mechanism of inheritance, which was still unknown at this time. John Maynard Smith (1920–2004) introduced the “Concept of Evolutionary Stabile Strategy” and identified all “populations of units, which are capable of proliferation, inheritance, and variation” as life [14]. The definition of life generally accepted today is based on the eight columns genetic program, reproduction, adaptation, compartmentation, metabolism, catalysis, regulation, and growth. 1.1.2 What Is Life Sciences?

The term life sciences originated in the 1990s when it was established as a marketing instrument of the pharmaceutical and the chemical industry. With the definition of the term, attention has been directed to the large number of pharmaceutical products and pesticides that are required for sufficient nutrition as well as for health maintenance of the world’s population. Meanwhile, the term has gained individuality and has a much greater meaning. All processes, products, and conditions that are connected to “life” itself, are summarized in the term life sciences. Today, life sciences include not only biological sciences, but also parts of medicine. The development of new drugs (e.g., using bio-catalytic methods) or the environmental friendly optimization of processes are included as well as the sequencing of the genome of plants, animals, and humans or the development of new therapeutic forms for different diseases. Life sciences in general are research fields dealing with processes and structures of living organisms. Besides classical biology, this also includes related areas

1.1 Life Sciences – A Definition

such as medicine, biomedicine, biochemistry, molecular biology, biophysics, bioinformatics, human biology, as well as agricultural technology, food, and nutrition sciences up to the use of biogenic natural resources and biodiversity research. The spectrum of methods comprises the complete device and analysis inventory and enters into the fields of humanities and social sciences. The methodical work and the theoretical background are highly interdisciplinary, but always have a clear relation to living organisms, especially to humans. Thus, life sciences comprise a similar huge scientific group as do the humanities. According to the generally accepted classification of biotechnology, a distinction of “green” (agriculture, plants), “red” (medicine, pharmaceutics), “blue” (products from the ocean), “white” (industrial products), “gray” (environmental), and “yellow” (production of food) life sciences is possible. Due to the still large number of diseases that cannot be treated with suitable medication (such as viral diseases), the occurrence of new diseases (such as SARS or bird flu), the knowledge about the side effects of commercially available drugs [15–17], or the increasing resistance against microorganisms (e.g., against antibiotics) [18, 19], a high demand for the development of new drugs exists today. In addition, the expiration of numerous patents and thus the synthesis of cost-efficient generics forces the development of new blockbusters for the pharmaceutical industry. Therefore, numerous new potential drugs have to be synthesized and tested; known drugs and drug candidates with side effects have to be modified; or completely new compounds (such as natural compounds) have to be discovered. The red life sciences are dedicated to the development of new drugs and therapeutic methods including modern developments in the fields of genomics and stem cell research. Plant research (healing plants, nutritional plants, plants as raw materials, plants as energy suppliers), farming methods, cultivation, and resistance against vermin are typical topics of focus of the green life sciences. In addition, plant cells can be used for the production of industrial compounds or drugs [20]. Their use for the remediation of soils (phytoremediation) or as environmental sensors has also been reported. The blue life sciences deal with the enormous potential of still undiscovered natural drugs from marine organisms. Extraction, isolation, testing, and identification of new compounds as well as the development of suitable synthesis concepts for industrial-scale production of these compounds are the focus of the current research. Especially, bacteria living in deep-sea regions under extreme conditions are seen as a potential source for biological substances, which can be used for technical processes. Usual enzymes denaturate at higher temperatures, whereas biocatalysts of deep-sea bacteria still work in extremely hot surroundings. Gray life sciences (also called environmental life sciences) are dedicated to the processes for the preservation and regeneration of the environment. This includes all life science methods of drinking water purification, purification of wastewater, treatment of waste, and remediation of contaminated soil and air [21]. Methods of gray life sciences for environmental protection are mainly used at the end of the process chain (wastewater purification, bio filters, bio washers, etc.). Innovative approaches not only include the disposal of occurring environmental

3

4

1 Introduction

pollutions, but also focus on the avoidance of pollutants during the production process. Genetically modified enzymes can reduce energy and raw material consumption in the production of textiles, food, detergents, and drugs and reduce the occurrence of unwanted side products. White life sciences (also called industrial life sciences) refer to the use of life science methods in industrial manufacturing processes. Industrial life science transfers biological and biochemical knowledge and processes into technical applications. Bacteria such as Escherichia coli and Corynebacterium glutamicum, yeasts, and enzymes are used in these processes. The term industrial life sciences is quite young, but some methods have been used for over 1000 years by mankind. Many cultures used yeasts for alcoholic fermentation, lactobacillus strains for malic acid fermentation, or acetobacter species for acetic acid production. Due to the advances in the development of life science methods and applications, the field of industrial life sciences gained high importance in the last years. Further potential is expected based on the optimization of production processes (e.g., basic and fine chemicals), the reduction of raw material dependency (e.g., through the use of renewable resources), the reduction of energy and disposal costs (e.g., by the substitution of chemical processes), and the development of new products and system solutions with high value-added potential (e.g., utilization of biological metabolic pathways using genetic methods) [22]. The yellow life sciences (also called food life sciences) are used in food industry, for example, for the production of beer using yeasts or yoghurt and sauerkraut using lactic acid bacteria. Life science engineering is a scientific field at the interface of engineering and life sciences. It deals with the technical use and engineering realization of knowledge from the life sciences. The understanding of the functional mechanisms of living organisms is of great importance to ensure the application of this knowledge in modern technology. On the other hand, engineering knowledge is required to integrate biological systems into technical processes. A typical example is the production of pharmaceutical compounds in sufficient quantity and quality. A life science engineer can thus be considered an engineer who understands the scientific aspects and coherences in the life sciences and can integrate this knowledge into technical solutions. Aspects of sustainability as well as the use of suitable software (bioinformatics, simulation) are necessary.

1.2 Automation – A Definition Automation is the transfer of tasks to automatic machines, usually realized by technological progress. Automation is a multidisciplinary field of technology and an engineering science containing all methods for automating machines and facilities to work independently without the involvement of humans. The term automation can be traced back to ancient times. In ancient Greece, people admired the goddess αυτoματια [automatia] – “who manages things according to her own will” – and dedicated chapels to her [23]. Aristotle

1.3 History of Automation

formulated in his work “Politics,” that self-working machines performing proper tasks would lead to a situation, where no assistance for the supervisor or a slave for the master would be required [24–27]. The definition of automation has changed during the ages depending on the actual state of the art of technology. Lothar Litz mentions a very early definition of the word automat in Meyers Neuem Konversationslexikon in 1862, which still has validity. In this definition, an automat is any self-moving mechanical tool, which for a certain time can be operated without interaction from the outside by hidden forces inside the machine; sensu stricto any mechanical art, which due to an inner mechanism can simulate the activity of living organisms, humans, or animals, and which form is analog to them [28]. About 100 years later Kienitz and Kaiser defined automation as a high degree of the substitution of human work by machines including control, decision, and adaptation functions [29]. The current industrial standard DIN IEC 60050-351, in contrast to previous versions, does not include a direct definition of automation, but is limited to the terms automatic and automat [30]. An actual definition of automation was given by Lothar Litz in 2013, whereby consciously a general form was used, to ensure the validity of the definition in the future: by automation, dynamic processes are captured over time and specifically modified, that they can independently execute predefined tasks and functions [28]. According to DIN IEC 60050-351, a process is defined as the unison of different interacting events in a system, which enables the conversion, transportation or saving of matter, energy or information [30]. Applied to technical processes, automation in this area is called process automation [31].

1.3 History of Automation 1.3.1 Automation from the Beginnings to the Nineteenth Century

The history of automatic machines originates in ancient Greece. Besides numerous myths and legends, the first historical evidences for automats can be found. The developers of such automats tried to investigate physics and copy nature using technical tools. A number of artificial birds, moving and speaking statues, and artificial servants and attendants can be found in Greek mythology. The aspect of usefulness was not the priority; the first useful automatic machines have been reported for waterworks and military applications. Homer reported in his “Ilias,” that Hephaistos, the god of craft, developed self-driving vehicles and artificially intelligent handmaidens that could learn different crafts. Many reports of historians from ancient Greece and ancient Rome with detailed explanations about self-driving mechanisms and androids are known. Similar reports are also known from other early cultures, especially from China. The main problem of all reports is that it is difficult to distinguish between myth and reality. The first real automatic devices are known from the era of the Alexandrian school [32]. Excellent natural philosophers such as Heron of Alexandria (died later than 62 after Christ), Pythagoras (570–495 before Christ), Euklid

5

6

1 Introduction

(approximately third century before Christ), and Archimedes (287–212 before Christ)1 were researching and teaching in Alexandria. The Alexandrian developers combined simple tools such as screws, wedges, and levers for the execution of complicated movements and used water, vacuum, or air pressure as driving forces for the designed machines. In his work Automata, Heron von Alexandria described self-opening temple doors. He also developed music machines and automated theaters with remarkable effects. Different researchers in this era proposed numerous designs of artificial birds that could move their wings and chirp, or automated machines that alternatingly delivered wine and water. An automat was developed that provided a defined amount of holy water after insertion of a specific amount of money. The researchers of the Alexandrian school developed many “programmed simulators and automates as well as tools for a feedback,” which are still in use today such as the water flushing in toilets [33]. A vast amount of antique literature was lost in Central Europe with the collapse of the Roman empire, but could be preserved in the Arabian world [33]. In the ninth century, the Khalif of Bagdad Abdallah-al-Manun instructed the sons of his astrologer to systematically search all scientific literature. They translated the work of Heron and revised it. The work of Kitab al-Haiyal “Book of artful tools” can be affiliated with the Alexandrian school and became a standard book during this time. The aspect of practical applications was very important for the authors: one of them was manufacturing clocks. Most of these historical devices have been destroyed by the Mongolians in their attack on the Arabian domain in 1258. Scholars in monasteries in Central Europe started the investigation of old antique literature and the Arabian revisions and advancements at the end of the middle age. A legend reports that Albertus Magnus (1193–1280) designed a speaking statue, which was destroyed by his student Thomas von Aquin (1225–1274) [33, 34]. The scholars during the second half of the thirteenth century relied on the doctrine of soul according to the work “De anima” by Aristotle (384–322 before Christ) [35]. The existence of a soul is the requirement for self-movement; thus, a statue should be given a soul to make it move. For Thomas von Aquin this was only possible with magic, he named people who were able to “make statues speak and move” as necromantici (supporters of black magic) [36]. During this period, the first pure mechanical clocks were developed (complex clocks so far were driven by water) and the craftsmen started to combine the measuring of time with moving figures and chiming mechanisms (early fourteenth century). One example is the clock of the Straßburger Münster with its mechanical cock (around 1350), which was able to move the wings and crow at noon. Following the tradition of the Alexandrian school, complete scenes with religious and secular backgrounds were designed, now driven by mechanical watches. In addition to the mainly humanoid automats, music automats in the form of self-playing instruments were developed. The oldest mechanical instruments that are still preserved are the glockenspiel in monumental clocks from the late middle ages. During the renaissance, craftsmen in Augsburg 1 Archimedes lived in Syrakus, which belonged to the Alexandrian cultural environment.

1.3 History of Automation

developed precious music automats and self-playing spinets that were controlled using pinned barrels. The renaissance was an important era for the history of technology as well, which is called as the “technology revolution of the renaissance” [33]. The Alexandrian school mainly developed models, whereas the technological progress during the renaissance enabled the construction of life-size automatic devices. Engineers such as Leonardo da Vinci (1452–1519) studied books about mechanics and designed many new machines. Around 1550, they started to write technical literature. In the 1950s, a draft of a robot from Leonardo da Vinci was found [37, 38]. This robot could move its arms, stand up, and move the head. The French engineer Salomon de Caus (1576–1626) was the author of a comprehensive work “Les raisons des forces mouvantes” (about moving forces. Description of some artificial and enjoyable devices). He described many automats from Heron and revised them. He constructed the Hortus Palatinus (Pfälzischer Garten) in Heidelberg and later a series of moving figures driven by water mills and cam rollers in the palace of the duke of Burgund in Saint-Germain near Paris. Similar constructions were set up in 1613 in the Castle Hellbrunn; between 1748 and 1752 the facility was extended with at least 256 figures [39]. A hydraulic organ was used to eliminate the noise from the driving mechanism. De Caus can be acknowledged as a pioneer in the construction of life-size automats. At many courts engineers were employed to develop androids and other automats “which worked rough-and-ready” [33]. Another development was the construction of automated toys as in ancient times. One of the founders of this technology field was Juanelo Turriana (1500–1585), who worked as an engineer for Karl V and revolutionized the water supply of Toledo. His reputation was so great, that it is reported that he designed an android, which could go shopping for him [33]. During the seventeenth and eighteenth century, the natural sciences turned toward engineering. Engineering was understood as applied natural sciences. Different engineering schools, such as the École polytechnique, were founded in France. In Germany, some goldsmiths and precision engineers, who were leading in the field of automat construction, were located in Nuremberg and Augsburg. Around 1585, Hans Schlottheim (1545–1625) constructed the most important ship for Karl V. The ship had wheels and was moving on a wriggling track. An organ and different other instruments were playing and cannons crumped periodically. In addition, “people” were performing different tasks on board the ship such as hoisting the sails and making an inspection round. The imperator himself was sitting on a baldachin throne and moving his head [40]. Investigating the control mechanisms of this time shows that highly developed technologies were used, which “since that time have been rediscovered several times independently from one another” [33]. Automatic devices were widely known in the seventeenth century and gained high interest under the philosophers of the burgeoning enlightenment. Cartesianism [41] was characterized by the great advances of scientific and rational dealing with reality. The mechanism postulated parallels between the laws of mechanics and machines and natural bodies. Rene Descartes (1596–1650) [42] explained in his work “Discours de la method” (1637) the difference between

7

8

1 Introduction

humans and animals [43]. Both are God created, but only the human has an immortal soul. Animals can be seen as very complicated machines (la bête machine), as automats. He was convinced that humans would be able to construct an animal-like machine in the future, which would also behave like an animal. He compared the heart with a hydraulic pump and described other organs such as tendons and muscles analogously to automated tools that were available in this period. The German Jesuit Athanasius Kircher (1602–1680) realized the ideas of Descartes and constructed a speaking head, singing birds, and figures playing instruments. He followed the tradition of Salomon de Caus (1576–1626) and developed automated theaters for gardens [33]. Public interest in automats was very high during the eighteenth century, especially since different cases of fraud were uncovered. In the seventeenth and eighteenth century, many reports were published about autonomous driving vehicles and other automats; however, a majority of these were unmasked as fakes. The technological development was far behind the wishes and expectations of the people at this time. A highlight in the development of real automatic devices was the achievement in the constructions of Jaques de Vaucanson (1709–1782). He studied anatomy since he wanted to design three-dimensional anatomic models (anatomis mouvante – moving anatomy). This would have led to a realization of the philosophical basics from Rene Descartes. He designed a life-size flute-playing shepherd. This was not a realistic anatomic model, but an automat driven by clockwork and bellows. Nevertheless, this automat was highly acknowledged. Vaucanson continued following the idea of a moving anatomy with the development of a mechanical duck, which was able to waddle, and could also eat, digest, and excrete [44]. Vaucanson was also a highly valued member of the Académie des Sciences. Later, Vaucanson stopped building automatic devices and, in 1741, became the director of the state silk mill factory. With his further inventions and constructions, he provided many suggestions for its mechanization and automation. He thus initiated the change from pure invention due to scientific technological interest to the introduction of automation devices into industry. “For thousands of years, the construction of automats was more an enjoyable then a useful dissipation. Thanks to the contributions of Vaucanson, it was possible to overcome this level and use automats in industry. Only now the ideas of the Alexandrian school could mature so that an automatically controlled system could become reality” [33]. Vaucanson developed a mechanical weaving loom for patterned fabrics, which control used the same principle as his flute player; however, the system was not used. In 1804, Jacquard revised this weaving loom and invented his weaving automat. Vaucanson can also be seen as the father of modern fabric. In 1756, he built a silk mill near Lyon and designed every detail of the building and the driving. This can be considered as the first industrial facility in a modern sense. He recognized that manufacturing has to be done in a facility, where devices are driven by one central force. Although his inventions are in great contrast to the simple constructions of cotton spinning machines of the Englishman, Vaucanson’s inventions, just like many other ideas, could not gain access to the market in the catholic Ancient Régime (France) in the middle of the eighteenth century [45].

1.3 History of Automation

The industrialization started then in England, where complete different sociological conditions existed. Cotton was used instead of silk, which enabled a mass sale. In addition, the promoters of industrialization were mainly social climbers instead of aristocrats or established citizens. They built their factories in relatively new cities that were not bound by old guild regulations [45]. After Vaucanson, many, and sometimes very complex, androids performing real tasks have been reported. The automats from father and son Jaquet–Droz (1721–1790) are the most well known. Along with a mechanical engineer Jean-Frederic Leschot (1746–1824), they developed the three nicest automats that can be seen today in the museum in Neuchatel (Switzerland). The “writer” is a small human-like automat with moving head and eyes. It is able to write a text with up to 40 letters. The text is coded on a wheel and the letters will be written one by one. The writer can write in different lines and considers blanks. This machine can be considered as a precursor of a computer, since the machine has a program and a memory and can be programmed with different texts. At the end of the eighteenth, and the beginning of the nineteenth century, many automats were developed. One of the master constructors was Johann Nepomuk Mälzel (1772–1838), who constructed numerous music automats including the highly applauded trumpet player, which played in Vienna [46, 47]. Famous composers such as Jan Ladislav Dusík and Ignaz Pleyel composed concert pieces for this trumpeter. The organ clock was invented in the eighteenth century, for which Haydn, Mozart, and Beethoven composed original compositions. The introduction of pneumatics enabled the development of self-playing pianos with a satisfying dynamical gradation. Additional masters of automats were Johann Gottfried Kaufmann (1751–1818) and his son Johann Friedrich Kaufmann (1785–1866), who developed a trumpet player in Leipzig and revised Mälzels development. In Paris, organ developer Beaudon developed in 1810 a mechanical elephant, which was made of 4800 parts and could eat and drink [48]. Preserving the spirit of the ancient times as well as the knowledge transfer from the Arabian era, especially in mathematics, enabled new academic progress in physics during the renaissance. In 1745, the English blacksmith Edmund Lee developed an early automation device for a self-dependent movement of wind mills [49]. According to records from the ancient times, there were already machines during this time, which were able to actuate a windmill. They thus executed work, which was previously done by humans or animals. In the middle ages, windmills were constructed with a vertical axis. The windmills were moved in the direction of the wind to enable continuous working. With the invention of Lee, who integrated an additional windmill with a gear, the machine was able to react autonomously to changes in its environment as required to fulfill its tasks. With the advances in the field of mechanics and the new drive technologies such as the steam engine, the age of industrialization started. Mass fabrication in factories became possible. Animal and human power could be substituted by motors. In 1785, Edmund Cartwright (1743–1823) patented a mechanical weaving machine, which was still hand driven. Only one year later in 1786, he introduced a revised version of the machine, which used a new mechanical drive for the moving parts of the weaving machine [50].

9

10

1 Introduction

These were the first automated machines for industrial production. Although Cartwright was not successful with his weaving mill, his inventions gained acceptance and had extensive social impact. The automated weaving machines destroyed many work places. In 1811 in England, weavers started to revolt against the machines. The machine breakers destroyed the machines and attacked the supporters. The revolts were finished with the help of the military; the participants were executed or banished. Similarly motivated revolts also occurred in Switzerland, whereas the revolts in Germany known as weavers’ uprising were focused against foreign workers and suppliers. The beginning of the industrial revolution is connected with the invention of steam engines. James Watt (1736–1819) did not invent the steam engine, but brought numerous technological advances to it in 1788, including the use of a centrifugal governor to ensure a constant number of revolutions [51]. With the operation of the weaving machines using steam, the industrial revolution started. Tasks and functions, which had been executed before by humans, could now be realized with machines. Similar processes could thus be standardized and the productivity was significantly increased. This general endeavor to execute rigid and repeating procedures by machines has been transferred to many other technical fields. A great number of automat developers in the first half of the nineteenth century were magicians or engineers inspired by illusions. Jean Eugène Robert-Houdin (1805–1871), the father of modern magic, constructed numerous real automats, which he showed in a special theater. In addition, he also developed trick automats that were controlled by wire rope hoists or pedal systems driven by humans inside the objects [52]. The French magician Stèvenard was the most talented precision engineer of this time, since he developed very small but complex automats, which he introduced to the public in 1850 in Paris. The described automats were single devices and thus quite expensive. A small automat-producing industry developed in Paris in the nineteenth century. Families such as Vichy, Lambert, Decamps, and Roullet produced the automats in limited editions [40]. This development was stopped at the beginning of World War I. Many small automats were produced in Germany, such as singing birds, which have been produced up to the 1970s in Black Forest. 1.3.2 Automation Since the Nineteenth Century

The discovery of electricity and the developments in electrical engineering (nineteenth century) enabled the decentralization of production. It became possible to send energy over long distances. First attempts had been made to utilize electricity for tasks such as measuring, regulating, and controlling. In 1833, Samuel Morse (1791–1872) constructed the first workable electromagnetic telegraph. The signals were coded and were serially transmitted. This led to the development of the telex machine and standardized serial interfaces. These were the basis for our current bus systems [53]. The development of the relay by Joseph Henry (1797–1878) in 1835 was another milestone in the development of modern automation technology [54, 55].

1.3 History of Automation

In 1939, Hermann Schmidt (1894–1968) founded in Berlin a Section for Control within the Society of German Engineers (VDI). He defined the term general control technology, which includes technical and biological systems and thus conforms to the definition of cybernetics by Norbert Wiener (1894–1964) [56]. The scientific field includes, for example, the feedback mechanisms on technical and biological systems. Computer technology started a technological development, which led to an increase in the degree of automation in the production with industrial robots, complete automated production lines, and technologies such as pattern recognition in artificial intelligence. The Z3 calculator, which Konrad Zuse (1910–1995) introduced in 1941 was the first workable digital calculator working with binary floating-point numbers [57]. According to our current linguistic usage, this was the first computer. The century of the digital revolution started. In 1948, William B. Shockley (1910–1989) introduced the term transistor for already developed semiconductor devices [58]. The development of the microprocessor followed in 1970/1971. Innovations in the field of electronics, especially the development of transistors, led to a radical decrease in the size of electrical circuitries. With the decreased dimensions, the effort for switching algebra applications was reduced as well. The development of integrated circuitries enabled the equipment of devices with logic circuits without great efforts. Digital technology became the main driver of automation. Innovative field devices such as sensors and actuators communicate with the control and guarantee a constant quality of the products, even in the case of process variations. In 1953, John W. Backus (1924–2007) proposed the advanced programming language Fortran (formula translator) [59]. The production of the first numerical control with tubes can be dated back to 1954. In 1958 followed the market introduction of the first electronic control SIMATIC. The invention of the first solid-state sequential logic solver by Richard E. Morley in 1969 was the basis for the further development of the programmable logic controller. Odo J. Struger (1931–1998) significantly contributed to the formulation of the necessary institute standards in the United States. The PLC control replaced pneumatic and relay-based control systems. A further important step in the automation of industrial processes was the establishment of the Universal Product Coding (UPC) and the introduction of the barcode in the United States in 1970. The EAN code (European Article Number) followed in 1977. These identification methods built the basis for automated logistics. Currently, the barcode technology, which reads the barcodes with optoelectronics, is increasingly substituted by RFID technologies (radio frequency identification). In 1981, IBM introduced the first personal computers, which could be used in offices and schools. Before this time, several MS-DOS based computers have been on the market. As smaller computers became more powerful, they could be linked together, or networked, to share memory space, software, and information, and communicate with each other. The first satellite-based global positioning system (GPS) was established in 1995. Besides the development of navigational systems, this led to the possibility of automated guidance of agricultural engines.

11

12

1 Introduction

One of the most important people influencing the modern industrial automation was Henry Ford (1863–1947). His concept of a modern manufacturing of vehicles revolutionized not only the industrial production (basis for line production), but also had significant influence on modern art (Fordism) [60]. Today’s automation technology is based on a variety of knowledge. Realizations have entered industry as well as our everyday life, and are taken for granted. Today, design, implementation, and commissioning are mainly method oriented and dedicated to specific processes. Automation is a solid part of our society. Communication technology and the underlying networking are used by everybody. Field bus systems such as profibus, interbus, real time Ethernet, and wireless communication systems are used for networking. Machine-oriented networking is in most cases part of the closed functional chain and has thus to fulfill real time requirements. The Internet of Things, where machines are communicating with each other and with their workpieces can be seen as the temporary highlight of automation. 1.3.3 History of Laboratory Automation

Laboratory automation is a special field of automation dealing with the automation of laboratory processes in chemical, biological, pharmaceutical, and food technologies as well as in medicine. It is a discipline combining the knowledge of laboratory science with process engineering. Reports of automated devices for scientific investigations have existed since 1875. These first automated devices were special solutions built by scientists for specific applications in their laboratories [61]. The first time automation was mentioned in the chemical literature of the United States, was an unattended device for washing filtrates [62]. In 1894, Edward Robinson Squibb (1819–1900) created an automatic zero burette [63]. An automated pipette for use in the Babcock milk test was described by Greiner [61, 64]. During the 1920s, different devices were developed for solvent extraction in botanical research. The first continuous liquid–liquid extractors with internal diffusers were reported by Palkin et al. [61, 65]. Besides laboratory use, automation also developed in the coal and power generation industry. The first device specifically manufactured as a piece of laboratory automation was a grinder for preparing coal samples [61]. At the beginning of the twentieth century, different systems were used for measuring carbon dioxide in flue gases in order to optimize combustion control. A continuous system has been sold as the Autolysator by Stache, Johoda, and Genzen since 1912 [61, 66]. The first system for measuring carbon monoxide was reported by Guy B. Taylor and Hugh S. Taylor, who used an automatic volumetric approach [67]. A conductivity-based measuring system was developed by Edelmann in 1921 [61]. Since the 1920s, a new era started in laboratory automation with the increasing development of electronics. The need for rapid gas analysis during the First World War led to the development of new principles such as the use of thermal conductivity as the basis of gas analyzers [68]. The next step in the development of laboratory devices was the introduction of pH electrodes; one of the main drivers was the sugar industry. Early tungsten–calomel electrodes showed only

1.3 History of Automation

insufficient reliability and it thus took some time to install automated pH control in this industrial field [69]. In 1929, Partridge and Muller of the Department of Chemistry at New York University introduced their first automated titrator. A photocell was used to detect the color change in the titration process; signal amplification was realized with a radio tube [61]. A more sophisticated titrator was built for acid-based titrations at Eastman Kodak in Rochester. It included a set of valves that enabled emptying of the previous sample from the titration vessel [70]. In 1942, Bassett Ferguson introduced a semi-automated still for petroleum analysis. This development is a typical example of the type of automation for conserving manpower [71]. An example of devices specifically designed for alleviating skilled labor shortages is the mercaptan automatic titrator. The potentiometric titration procedure was developed by Shell Oil Company in 1941 [72] and automated in 1943 [73]. Different companies started to provide automated equipment after the Second World War since the use of automatic control devices had become routine in chemical laboratories. A new type of titrator was designed in 1948, which used a motor-controlled syringe instead of a dripping burette for adding the titrant. A recording potentiometer was used for plotting the titration curve [74]. The automated Karl Fischer titration was announced in 1952. A polarization end-point was used, where an increase in the current between two platinum electrodes corresponded to a depolarization of the electrodes found at the end point of the titration [75]. Another important step was the introduction of the coulometric Karl Fischer titration in combination with a regeneration of the Karl Fischer reagent from the iodine present in a solution of depleted reagent [76]. Due to an increasing use of automation technologies, a new journal devoted to automation and instrumentation started in 1952, the Instrument Engineer. The first computer used in connection with laboratory automation was an analog computer that, for the first time, allowed chemical researchers to create electronic simulations of their processes [61]. The first use of a digital computer was reported in 1952, when the Atlantic Refining Company introduced a mass spectrometer and a digital computer for the determination of hydrocarbon mixtures [77]. The development of the transistor also revolutionized laboratory automation since it offered the possibility of collecting thousands of data points. The development of laboratory automation was heavily influenced by medical applications. The first truly automated systems appeared in medical laboratories in the mid-1950s. In 1956, a blood analyzer for the determination of urea, sugar, and calcium was introduced (AutoAnalyzer, manufactured by Technicon) [78]. Later designs offered the possibility of simultaneous determination of over 20 analytes with 150 samples per hour. Many other batch analyzers had been developed, which could test up to 100 samples in continuous mode. The introduction of the photodiode array for spectrometers with grating monochromators in the early 1980s allowed the simultaneous detection of multiple analytes using various wavelengths [79]. A different approach to clinical automation appeared in 1959 with the production of the Robot Chemist (Research Specialty Company) [80].

13

14

1 Introduction

The Robot Chemist automated all manual steps performed by lab technicians using conventional cuvettes and automatic pipetting and mixing but was too complex to be practical [81]. The introduction of robotics and informatics to clinical laboratory automation in the 1970s led to the development of total laboratory automation (TLA) [82]. In the early 1980s, Masahide Sasaki opened the first fully automated laboratory [83–85]. Driven by the requirement of increasing the laboratory services without increasing the costs, he and his team assembled robotic manipulators, conveyor systems, and a software system into a complete automation system first installed in 1982 [86]. Many other researchers were inspired by the work of Masahide Sasaki and developed laboratory automation systems for different applications. These automation systems usually consisted of various automated work stations and a connecting conveyor belt based transport system. In the following years, more than 800 laboratories invested tremendous amounts of money in laboratory automation to reduce costs, speed up turnaround, and to develop high-quality sample handling in clinical laboratories [87]. In 1993, Rod Markin designed the worldwide first clinical automated laboratory management system at the University of Nebraska Medical Center. In the mid-1990s, he was the chair of the Clinical Testing Automation Steering Committee of the American Association for Clinical Chemistry [88], which later became an area committee of the Clinical and Laboratory Standards Institute. In 2004, the National Institutes of Health (NIH) and more than 300 leading partners from science, industry, government, and the general public completed the NIH roadmap to increase medical discovery to improve health. Despite the success of Sasaki’s laboratory, the high cost of automated clinical laboratories prevented the spreading of this technology [89]. Another limiting factor was the proprietary interfaces and protocols of different vendors, which did not allow for a communication of devices from different manufacturers. The recent development of scripting languages like AutoIt enabled the integration of equipment from different manufacturers [90]. Automation in the field of life sciences developed rapidly over the last 20–30 years. Main drivers besides medical applications were bioscreening and the development of high-throughput screening technologies (HTS) according to the requirements of the pharmaceutical industry. HTS means the use of highly developed and completely automated laboratories, which enable 24/7 operation. HTS also means the interdisciplinary collaboration of life sciences, natural sciences, and engineering sciences [91], whereby robotics, electronics, information technology, analytical chemistry, chemical synthesis, optics, imaging technologies, cell biology, molecular biology, and biochemistry play important roles [92]. The main goal of biotechnological HTS is the efficient utilization of new active compounds by testing high numbers of samples with special focus on time and cost reduction and increasing information content [91]. Using the example of the pharmaceutical company Pfitzer Global Research and Development, Pereira and Williams described the origin and development of HTS [93]. The early development stage can be dated back to 1984–1995. Between 1995 and 2000, conceptual use occurred for the investigation of drug metabolism and toxicity. This included three consecutive phases that are generally valid for the development of HTS processes. The first phase is concept development and

1.4 Impact of Automation

planning followed by implementation. The second phase includes the technical development and practical realization. The final third phase includes the logical extension with the integration of additional specialist disciplines [93]. It can be seen that the target, which means the screening target compounds, developed over time. Around 1984, mainly natural compounds were the matter of interest for screenings and, until 1990, target molecules for therapeutic applications. Since then and up to now, ADME targets (absorption, distribution, metabolism, excretion, toxicology) are the focal point [93]. In other areas, an increasing demand for suitable high throughput automation solutions exists as well. This includes agricultural and environmental laboratories, quality control, and the academic sector [93–95]. The developments in the automation of life science processes have mainly been driven by the requirements of the pharmaceutical industry. Within the last 25 years, the number of samples to be screened increased significantly. Until the end of the 1980s, approximately 10 000 compounds have been tested per year and target. In the early 1990s, the sample throughput increased up to 10 000 samples per month and target and increased only 5 years later to 10 000 samples per week [96]. Today, high throughput screening is defined as the investigation of different thousand samples per day. An enhancement is the ultra-high throughput screening, which enables the processing of more than 100 000 samples per day [92, 97–100]. With the help of these high throughputs, huge compound libraries can be established within a short time. Today, the big pharmaceutical companies use compound libraries with synthetic compounds for the area of drug discovery (including lead discovery). The development of HTS was shown using the example of the continuous extension of the Bayer HTS library [101]. In 1996, Bayer decided to extend the low molecular drug development with a significant extension of the proprietary compound library. Today, more than 1.5 million single compounds are routinely tested in screening programs. A further increase of the expansion rate of such company libraries is limited. Reasons include the logistic effort to allocate the compounds on microplates, the required storage, as well as the relocating of the compound samples in big compound storage places for subsequent medical studies [101].

1.4 Impact of Automation 1.4.1 Advantages and Disadvantages of Automation

Automation technology is an ancillary discipline for all parts of engineering sciences, including all methods for the automation of machines and facilities. While the original approach was to use automation for mass production, the focus today is on releasing humans from dangerous, exhausting, and routine tasks. Automation has many advantages including increased throughput or productivity, improved quality or increased predictability of quality, an improved robustness or consistency of processes or products, an increased consistency of the output, and reduced direct human labor costs and expenses. Improving the productivity, quality, or robustness can be achieved with different methods.

15

16

1 Introduction

One main point is the installation of automation in operations to reduce the cycle time or in processes, where a high degree of accuracy is required. Another point is to replace human operators in tasks that involve monotonous or hard physical work [102] or those that have to be done in dangerous environments such as fire, underwater, volcanoes, nuclear facilities, or chemical facilities. In addition, tasks that are beyond human capabilities in terms of size, weight, speed, endurance, and so on, can be performed by automation systems. Automation can significantly reduce operation times and work handling. Today, most factories in industrial countries manufacture their products with the help of machines. In principle, it is possible to automate the production of small batch series up to single devices. A higher automation degree in industry and other economic sectors results in increased productivity, simplification, and changes of work processes. Thus, rationalization is always a result of automation. Due to higher productivity in the manufacturing industry, other industrial fields had to follow the automation trend as well. This resulted in a significant increase in the economic output of these industrial fields and companies. For example, this can be seen in Germany or Japan, which increased their incomes due to automation in the twentieth century. For a long time, automation was only used in industry to reduce the amount of monotonous work for humans. In the last years, automation also gained more influence on other fields. This includes, for example, the service industry, where electricity billing has been automated and online banking is possible. In addition, many other things became possible due to automation, such as safety technologies in automobile industry, including electronic stability control/electronic stability program (ESP) or airbags. In this connection, the question of safety plays an important role as well. The observance of regulations is a basic requirement for the development of efficient and reliable working machines and facilities. Automation contributes significantly to this development. Besides the positive effects of automating industrial processes, there are also some disadvantages. One of them is security threats because automated systems may have a limited level of intelligence. Therefore, they are more susceptible to errors since they are typically unable to apply the rules of simple logic to general problems. A second great disadvantage is the high initial cost, since the automation of a new product line or a plant usually requires a large initial investment. This applies also to unpredictable and excessive development costs. The research and development costs especially for complex processes can exceed the cost, which might be saved due to the automation itself. 1.4.2 Social Impact of Automation

A social result of the automation is often the loss of workplaces. Taylorism tried very successfully to establish a rational and efficient mode of production (assembly line production) and thus changed the work environment and the role of work. The efficiency of work was increasing constantly, but was repeatedly connected

1.4 Impact of Automation

with physical and psychological burdens for the employees. Repetitious work led to exhaustion and alienation of the employees from their work. It also produced conflicts between employees and employers since the increase of productivity was not correlated to the wages. In the 1980s, automation was linked to the loss of workplaces. Many simple but also dangerous, monotonous or very precise and fast tasks can be realized with automation technology using machines. This can be much more productive compared to manual operations. Automation frees up workers to take on other roles and provides higher-level jobs in the development, deployment, maintenance, and running of the automated processes. The role of a human in the production process is changing from production to administration, planning, control, maintenance, and services. It also has to be mentioned that the high degree of automation contributed to the further existence of a high amount of industrial manufacturing in Germany: one good example is the automobile industry. In 2013, a group of scientists calculated that computers would be able to take over every second job [103]. Frank Rieger (Chaos Computer Club) warned that increased automation will lead to a significant loss of classical workplaces (e.g., truck drivers due to self-driving cars). Rieger argues for a “socialization of the automation dividend,” meaning a taxing of non-human work, so that economic growth also affects general prosperity and its fair distribution [104]. 1.4.3 Limitation of Automation

Automation has many advantages, but also limitations. The currently available technology is unable to automate all the desired tasks. One main limiting factor for the practical realization is the economic effectivity. The automation of complex processes is very expensive; thus, it is often more economical to purchase expensive robots for simple often-used steps in the production process. In principle, the use of robots is possible for all complex processes; however, operation and programming of these robots is very costly. Thus, only companies producing high quantities can afford such automation facilities. For many small companies, the use of human labor is more reasonable. Partial automation in combination with human labor is more profitable for these companies. Another limitation is the assembly of fragile, very delicate, and complex technology. Highly complex machines are required to automate these production steps. These often cost more compared to the economic savings due to the decreased number of employees. In addition, machines currently do not have creativity or the possibility for flexibility, since they can only execute preprogrammed process steps. If a product requires this creativity, such as in the case of single pieces, the machine is reaching its limitations. Many operations using automation bind large amounts of invested capital and produce high product volumes. This makes malfunctions extremely costly and potentially hazardous. Therefore, trained personnel are required to ensure that the entire system functions properly and that safety and product quality are maintained.

17

18

1 Introduction

Sometimes, all optimizations that could result from automation are limited. This can be the case if manual tasks are more economic compared to the complex automation solution or if human creativity has priority. Activities where humans still have advantages compared to machines include, in general, higher qualification than automated tasks. At the same time, humans must obtain this qualification using simulators, as production lines and facilities should not be interrupted and learning by doing is not possible or connected with high risks (e.g., flight simulator).

References 1 Edwards, P. (1967) Life, meaning and value of art, in The Encyclopedia of

Philosophy, 2nd edn, vol. 5, Macmillan, New York, pp. 345–359. 2 Herbermann, C.G., Pace, E., and Pallen, C.B. (1913) The Catholic Encyclo-

3 4

5 6

7

8 9 10

11 12 13 14 15

pedia. An International Work of Reference on the Constitution, Doctrine, Discipline, and History of the Catholic Church, vol. 1–17 Index, Supplement,, The Universal Knowledge Foundation, New York. Bechtel, W. and Richardson, R.C. (1998) Vitalism, in Routledge Encyclopedia of Philosophy (ed. E. Craig), Routledge, London. Mayr, E. (1997) What is the meaning of life? in This is Biology – The Science of the Living World, Belknap Press of Harvard University Press, Cambridge, MA. Ritter, W.E. (1919) The Unity of the Organism; or, The Organismal Conception of Life, by William Emerson Ritter, R.G. Badger, Boston, MA. Friederichs, K. (1927) Grundsätzliches über die Lebenseinheiten höherer Ordnung und den ökologischen Einheitsfaktor. Die Naturwiss., 15 (7), 153–157. Friederichs, K. (1927) Grundsätzliches über die Lebenseinheiten höherer Ordnung und den ökologischen Einheitsfaktor. Die Naturwiss., 15 (8), 182–186. Friederichs, K. (1934) Vom Wesen der Ökologie. Sudhoffs Arch. Gesch. Med. Naturwiss., 27 (3/4), 277–285. Thienemann, A. (1941) Vom Wesen der Ökologie. Biol. Gen., 15, 312–331. Schumacher, I. (1979) Karl Friedrich Kielmeyer, ein Wegbereiter neuer Ideen: der Einfluss seiner Methode des Vergleichens auf die Biologie der Zeit. Medizinhistor. J., 14 (1/2), 81–99. Bach, T. (2001) Biologie und Philosophie bei C.F. Kielmeyer und F.W.J. Schelling, Frommann-Holzboog, Stuttgart-Bad Cannstatt. Haeckel, E. (1899) Die Welträtsel: Gemeinverständliche Studien über monistische Philosophie, Strauß, Bonn. Schrödinger, E. (1951) Was ist Leben? Die lebende Zelle mit den Augen des Physikers betrachtet, Lehnen, München. Smith, J.M. (1998) Evolutionary Genetics, 2nd edn, Oxford University Press, Oxford. Mutschler, E., Geisslinger, G., Kroemer, H.K., Ruth, P., and Schäfer-Korting, M. (2013) Arzneimittelwirkungen. Lehrbuch der Pharmakologie, der

References

16

17

18

19 20

21 22

23 24

25

26 27 28 29 30

31

Klinischen Pharmakologie und Toxikologie, 10th edn, WVG Wissenschaftliche Verlagsgesellschaft, Stuttgard. Lüllmann, H., Mohr, K., Hein, L., and Wehling, M. (2016) Pharmakologie und Toxikologie. Arzneimittelwirkungen verstehen – Medikamente gezielt einsetzen: ein Lehrbuch für Studierende der Medizin, der Pharmazie und der Biowissenschaften, eine Informationsquelle für Ärzte, Apotheker und Gesundheitspolitiker, 18th edn, Georg Thieme Verlag, Stuttgart. Langbein, K., Martin, H.-P., and Weiss, H. (2014) Bittere Pillen. Nutzen und Risiken der Arzneimittel: ein kritischer Ratgeber, 82nd edn, Kiepenheuer & Witsch, Köln. White, D.G., Alekshun, M.N., and McDermott, P.F. (2005) Frontiers in Antimicrobial Resistance: A Tribute to Stuart B. Levy, ASM Press, Washington, DC. Mahajan, R.C. and Therwath, A. (2000) Multi-Drug Resistance in Emerging and Re-Emerging Diseases, Indian National Science Academy, New Delhi. Ma, J.K.C., Barros, E., Bock, R., Christou, P., Dale, P.J., Dix, P.J. et al (2005) Molecular farming for new drugs and vaccines. Current perspectives on the production of pharmaceuticals in transgenic plants. EMBO Rep., 6 (7), 593–599. Reineke, W. and Schlömann, M. (2015) Umweltmikrobiologie, Springer-Verlag, Berlin. Bundesministerium für Bildung und Forschung (BMBF) (2007) Weiße Biotechnologie – Chancen für neue Produkte und umweltschonende Prozesse, Bundesministerium für Bildung und Forschung (BMBF), Berlin. Nilsson, M.P. (1950) Geschichte der griechischen Religion: 2. Band. Die hellenistische und römische Zeit, Beck, München. Schütrumpf, E., Grumach, E., and Flashar, H. (1991) Aristoteles. Buch I: Über die Hausverwaltung und die Herrschaft des Herrn über Sklaven, Akad.-Verl, Berlin. Grumach, E. and Flashar, H. (1991) Aristoteles. Buch 2: Über Verfassungen, die in einigen Staaten in Kraft sind, und andere Verfassungen, die von gewissen Männern entworfen wurden und als vorbildlich gelten, Akad.-Verl., Berlin, p. 590 S. p. Schütrumpf, E., Gehrke, H.-J., Grumach, E., and Flashar, H. (1996) Aristoteles. Buch IV–VI, Akad.-Verl, Berlin. Schütrumpf, E., Grumach, E., and Flashar, H. (2005) Aristoteles. Buch VII–VIII. Über die beste Verfassung, Akad.-Verl, Berlin. Litz, L. (2012) Grundlagen der Automatisierungstechnik: Regelungssysteme – Steuerungssysteme – Hybride Systeme, 2nd edn, Oldenbourg, München. Kienitz, H. and Kaiser, R. (1966) Automation und Analytik in der chemischen Industrie. Fresenius’ Z. Anal. Chem., 222 (2), 119–127. DIN IEC 60050–351 (2014) Deutsches Institut für Normung e.V. Internationales Elektrotechnisches Wörterbuch – Teil 351: Leittechnik. Berlin: Beuth Verlag GmbH. Lauber, R. and Göhner, P. (1999) Prozessautomatisierung 1: Automatisierungssysteme und -strukturen, Computer- und Bussysteme für die Anlagen- und Produktautomatisierung, Echtzeitprogrammierung und

19

20

1 Introduction

32 33 34 35 36

37 38 39

40 41 42 43 44

45

46 47 48 49 50

51 52

Echtzeitbetriebssysteme, Zuverlässigkeits- und Sicherheitstechnik, 3rd edn, Springer-Verlag, Berlin. Watts, E.J. (2006) City and School in Late Antique Athens and Alexandria, University of California Press, Berkeley, CA. Strandh, S. (1992) Die Maschine: Geschichte, Elemente, Funktion. Ein enzyklopädisches Sachbuch, Weltbild-Verlag, Augsburg. Scheeben, H.C. (1932) Albertus Magnus, Verlag der Buchgemeinde. Shields, C.J. (2016) Aristoteles: De Anima, Clarendon Press, Oxford. von Aquin, T., Schönberger, R., and Spaemann, R. (2001) Über sittliches Handeln. Summa theologiae I–II. Lateinisch/Deutsch, Reclam, Stuttgart, pp. 18–21. Müntz, E. (1898) Leonardo Da Vinci: Artist, Thinker and Man of Science, vol. 1, Heinemann. Müntz, E. (1898) Leonardo Da Vinci: Artist, Thinker and Man of Science, vol. 2, Heinemann. Museum für Europäische Gartenkunst, Stiftung Schloss und Park Benrath (2008) Wunder und Wissenschaft: Salomon de Caus und die Automatenkunst in Gärten um 1600. Katalogbuch zur Ausstellung im Museum für Europäische Gartenkunst der Stiftung Schloss und Park Benrath 17. August bis 5. Oktober 2008, Grupello Verlag. Soriano, A., Battaïni, A., and Bordeau, A. (1985) Mechanische Spielfiguren aus vergangenen Zeiten, Monaco, Sauret, Weber, Genf. Schmaltz, T.M. (2005) Receptions of Descartes: Cartesianism and Anti-Cartesianism in Early Modern Europe, Routledge, London. Descartes, R. and Gaukroger, S. (1998) The World and Other Writings, Cambridge University Press, Cambridge. Descartes, R. (1637) Discours de la méthode, Maire, Leyde. Priebe, C. (2008) Eine Reise durch die Aufklärung: Maschinen, Manufakturen und Mätressen. Die Abenteuer von Vaucansons Ente oder Die Suche nach künstlichem Leben, Books on Demand, Norderstedt. Giedion, S., von Moos, S., and Ritter, H. (1987) Die Herrschaft der Mechanisierung. Ein Beitrag zur anonymen Geschichte. Special ed, Athenäum-Verlag, Frankfurt am Main. Wilson, J.G. (1968) Appletons’ Cyclopaedia of American Biography, Reprint of 1886–1901 edn, Gale Research, Detroit. Augsburgische Ordinari Postzeitung. 1809; Nro. 299 (Freytag, den 15. Dezember):1. Augsburgische Ordinari Postzeitung. 1810;Nro. 151(Montag, den 25. Juni):1. Bennett, S. (1979) A History of Control Engineering, 1800–1930, Reprint 1986 edn, Peter Peregrinus Ltd., London. Strickland, M. (1843) A Memoir of the Life, Writings and Mechanical Inventions of Edmund Cartwright. Inventor of the Power Loom, etc., Saunders and Otley, London. Dickinson, H.W. and Vowles, H.P. (1949) James Watt and the Industrial Revolution, Reprint edn, Longmans, Green & Co., London. Steinmeyer, J. (2003) Hiding the Elephant: How Magicians Invented the Impossible and Learned to Disappear, Carroll & Graf Publishers, New York.

References

53 Jerome, H. (1934) Mechanization in Industry, National Bureau of Economic

Research, New York. 54 Henry, J. (1886) Scientific Writings, vol. 1, Smithsonian Institution, Washing-

ton, DC. 55 Henry, J. (1886) Scientific Writings, vol. 2, Smithsonian Institution, Washing-

ton, DC. 56 Wiener, N. (2007) Cybernetics or Control and Communication in the Animal

and the Machine, 2nd, 14th print edn edn, MIT Press, Cambridge, MA. 57 Rojas, R., Bauer, F.L., and Zuse, K. (1998) Die Rechenmaschinen von Konrad

Zuse, Springer-Verlag, Berlin. 58 Constable, G. and Somerville, B. (2003) A Century of Innovation: Twenty

59 60 61 62 63 64 65 66 67 68

69 70 71 72

73 74 75

Engineering Achievements that Transformed Our Lives, Joseph Henry Press, Washington, DC. Allen, F.E. (1981) The History of Language Processor Technology in IBM. IBM J. Res. Dev., 25 (5), 535–548. Batchelor, R. (1994) Henry Ford, Mass Production, Modernism, and Design, Manchester University Press, Manchester. Olsen, K. (2012) The first 110 years of laboratory automation: technologies, applications, and the creative scientist. J. Lab. Autom., 17 (6), 469–480. Stevens, T.M. (1875) Rapid and automatic filtration. Am. Chem., 6 (3), 102. Squibb, E.R. (1894) Automatic zero burette. J. Am. Chem. Soc., 16 (3), 145–148. Greiner, E.A. (1894) New automatic pipette. J. Am. Chem. Soc., 16 (9), 643–644. Palkin, S., Murray, A.G., and Watkins, H.R. (1925) Automatic devices for extracting alkaloidal solutions. Ind. Eng. Chem., 17 (6), 612–614. Dennis, L.M. (1913) Gas Analysis, Macmillan, New York. Taylor, G.B. and Taylor, H.S. (1922) Automatic volumetric analysis carbon monoxide recorder. J. Ind. Eng. Chem., 14 (11), 1008–1009. Shakespear, G. and Weaver, E. (1920) Notes and correspondence: automatic methods of gas analysis depending upon thermal conductivity. J. Ind. Eng. Chem., 12 (10), 1027–1028. Holven, A.L. (1929) Experimental application of automatic pH recorders to sugar-refinery alkalinity control. Ind. Eng. Chem., 21 (10), 965–970. Hickman, K. and Sanford, C.R. (1933) Automatic titrating devices. Ind. Eng. Chem. Anal. Ed., 5 (1), 65–68. Ferguson, B. Jr. (1942) Semiautomatic fractionation. A rapid analytical method. Ind. Eng. Chem. Anal. Ed., 14 (6), 493–496. Tamele, M.W., Ryland, L.B., and Irvine, V.C. (1941) Potentiometric determination of mercaptans in aqueous alkaline solutions. Ind. Eng. Chem., 13 (9), 618–622. Olsen, K.K. (1997) Rosie the robot, laboratory automation and the Second World War, 1941 to 1945. Lab. Rob. Autom., 9 (3), 105–112. Lingane, J.J. (1948) Automatic potentiometric titrations. Anal. Chem., 20 (4), 285–292. Frediani, H.A. (1952) Automatic Karl Fischer titration: apparatus using dead-stop principle. Anal. Chem., 24 (7), 1126–1128.

21

22

1 Introduction

76 Kelley, M.T. (1959) Automatic coulometric titrator for the Karl Fischer

77 78

79 80 81

82

83

84 85 86 87

88 89 90 91 92

93

94

determination of water application of the leeds & northrup alternating current-operated pH indicator to the control of titrations. Anal. Chem., 31 (2), 220–221. Olsen, K. (1952) C&EN reports: pittsburgh conference on analytical chemistry and applied spectroscopy. Chem. Eng. News Arch., 30 (11), 1092–1094. Plumb, R.K. (1956) Science in Review: Clinical Chemists Report on New Methods and Devices in the Field of Medicine. New York Times (Nov. 16), p. E13. Durner, J. (2010) Clinical chemistry: challenges for analytical chemistry and the nanosciences from medicine. Angew. Chem. Int. Ed., 49 (6), 1026–1051. Rosenfeld, L. (2000) A golden age of clinical chemistry: 1948–1960. Clin. Chem., 46 (10), 1705–1714. Armbruster, D.A., Overcash, D.R., and Reyes, J. (2014) Clinical chemistry laboratory automation in the 21st century – Amat Victoria curam (Victory loves careful preparation). Clin. Biochem. Rev., 35 (3), 143–153. Streitberg, G.S., Angel, L., Sikaris, K.A., and Bwititi, P.T. (2012) Automation in clinical biochemistry: core, peripheral, STAT, and specialist laboratories in Australia. J. Lab. Autom., 17 (5), 387–394. Sasaki, M., Kageoka, T., Ogura, K., Kataoka, H., Ueta, T., and Sugihara, S. (1998) Total laboratory automation in Japan: past, present and the future. Clin. Chim. Acta, 278 (2), 217–227. Felder, R.A. (2006) The clinical chemist: Masahide Sasaki, MD, PhD (August 27, 1933–September 23, 2005). Clin. Chem., 52 (4), 791–792. Boyd, J. (2002) Robotic laboratory automation. Science, 295 (5554), 517–518. Sasaki, M. (1987) The robotic system of the clinical laboratories. Rinsho Byori Jpn. J. Clin. Pathol., 35 (10), 1072–1078. Felder, R.A. (2007) The Lab Automation Survey Demystified. CAP Today, http://www.captodayonline.com/Archives/surveys/0307LASsurvey.html (accessed 01 December 2016). Hawker, C.D. and Schlank, M.R. (2000) Development of standards for laboratory automation. Clin. Chem., 46 (5), 746–750. Felder, R.A. (1998) Modular workcells: modern methods for laboratory automation. Clin. Chim. Acta, 278 (2), 257–267. Carvalho, M.C. (2013) Integration of analytical instruments with computer scripting. J. Lab. Autom., 18 (4), 328–333. Burbaum, J.J. (1998) Miniaturization technologies in HTS: how fast, how small, how soon? Drug Discovery Today, 3 (7), 313–322. Fox, S., Farr-Jones, S., and Yund, M.A. (1999) High throughput screening for drug discovery: continually transitioning into new technology. J. Biomol. Screening, 4 (4), 183–186. Pereira, D.A. and Williams, J.A. (2007) Review: historical perspectives in pharmacology. Origin and evolution of high throughput screening. Br. J. Pharmacol., 152 (1), 53–61. Smith, A. (2002) Screening for drug discovery: the leading question. Nature, 418 (6896), 453–459.

References

95 Chapman, T. (2003) Lab automation and robotics: automation on the move.

Nature, 421 (6923), 661–666. 96 Major, J. (1998) Challenges and opportunities in high throughput screening:

implications for new technologies. J. Biomol. Screening, 3 (1), 13–17. 97 Mayr, L.M. and Fuerst, P. (2008) The future of high-throughput screening.

J. Biomol. Screening, 13 (6), 443–448. 98 Hertzberg, R.P. and Pope, A.J. (2000) High-throughput screening: new

technology for the 21st century. Curr. Opin. Chem. Biol., 4 (4), 445–451. 99 Sundberg, S.A. (2000) High-throughput and ultra-high-throughput screening:

solution- and cell-based approaches. Curr. Opin. Biotechnol., 11 (1), 47–53. 100 HighTech Business Decisions (2002) High-Throughput Screening 2002: New

Strategies and Technologies – Report, Market Study, Moraga, CA. 101 Hüser, J. (2006) High-Throughput Screening in Drug Discovery, Wiley-VCH

Verlag GmbH, Weinheim. 102 Lamb, F. (2013) Industrial automation hands-on. A practical guide to indus-

trial automation concepts, terminology, and applications, McGraw-Hill Education, New York. 103 Frey, C.B. and Osborne, M.A. (2013) The Future of Employment: How Susceptible are Jobs to Computerisation? Oxford Martin School, University of Oxford. 104 Rieger, F. (2012) Roboter müssen unsere Rente sichern. Frankfurter Allgemeine.18.05.2012.

23

25

2 Automation in Life Sciences – A Critical Review 2.1 Overview Besides the classical goals of industrial automation (such as increasing the productivity, reducing costs, improving of product quality), the goals of the automation of life science processes include the increase of the degree of automation, the combination of island solutions to highly flexible complete automation systems, and the integration of manual subprocesses [1]. In the field of laboratory automation, two principal directions have been established: total laboratory automation (TLA) and modular laboratory automation. The following general objectives can be summarized for the laboratory automation [2–4]: • • • • • • • • • • • • •

Reduction of operating and test costs to provide services at a competitive level Summarizing and reduction of repetitive manual process steps Increase of sample throughput without additional personnel costs Reduction of the laboratory times Decreasing sample processing times of analytical measurement systems including a decrease of time delays Continuous operation of automation systems and avoidance of standstill times Ensuring the traceability of samples by monitoring single process steps Minimizing the risk for human operators and increasing occupational safety Minimizing the number of human errors Reduction of the sample volume Minimizing time and standardization in data evaluation and report generation Securing an optimal sample archiving for later repeating and extension of experiments/investigations Optimization of kind and number of available analytical measurement systems.

In the field of laboratory automation, this means the automation of processes for chemical, biological, biotechnological, as well as food and medical laboratories; manifold applications such as automated syntheses [5], production of formulations [6], clean up, sample preparation, and analytical measurement technologies have to be considered. Additional applications include sample logistics [7], cell cultivations and bioprocessing [8], toxicity assays [9], diagnostics and medicine as well as workflow management. Typically, laboratory Automation Solutions for Analytical Measurements: Concepts and Applications, First Edition. Heidi Fleischer and Kerstin Thurow. © 2018 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2018 by Wiley-VCH Verlag GmbH & Co. KGaA.

26

2 Automation in Life Sciences – A Critical Review

automation comprises the integration of liquid handlers and robots for the handling of conventional labware (e.g., microplates) in automated islands, which are often configured for a specific protocol [10]. Actual research is thus focusing on flexible hardware solutions, which means functional and structural flexibility regarding the sample throughput [11]. Sample processing includes the most labor-intensive processes in a laboratory, especially pre-analytical processes. Thus, a high demand for automation solutions exists in this field to increase productivity, decrease test times and laboratory cost, and to reduce human errors [3]. These automation tasks include the sorting of samples, loading and unloading of centrifuges, opening and closing of sample tubes, and the assignment of samples to different analytical measurement devices.

2.2 Definitions and Basics Automation solutions are used in the fields of biotechnology, chemistry, pharmacy as well as in food and beverages industries. Differences can be found in the tasks to be automated, the used devices, and the software components. A broad spectrum of devices can be integrated in laboratory automation systems depending on the degree of automation. This includes single devices such as shakers or centrifuges or complex fully automated systems. The components of automation systems have to be exactly defined and integrated into the complete automation process. A number of definitions were published in 1998 [12]. Actual definitions with international relevance include the concept of Laboratory Unit Operations (LUOs). These LUOs are the basic units for all laboratory processes (a sequence of steps or functions), which can be summarized to a unit [13]. LUOs can be divided into three main categories. Sample transport (manual, robot-based, linear transport devices such as conveyer belts, fluidic currents, positioning tables such as electric powered wheels and Cartesian platforms), sample processing (weighing, manipulation, separation, conditioning, milling), and data acquisition and evaluation (direct measurement, data acquisition with complex sensor systems, data processing and archiving, documentation). All laboratory processes include one or more LUOs whereby some or all LUOs are candidates for automation. The technical challenge is the development of a suitable architecture to enable the automation of the required LUOs [13]. Using these definitions, it can be distinguished between devices, workstations, and integrated systems. A device is defined as a laboratory instrument or tool, which is capable of performing one specific LUO. This includes shakers, centrifuges, oscillating mills, or microwave digestion devices (see Table 2.1). A laboratory instrument, which is capable of performing a limited number of LUOs (at least two LUOs) in an automated way, is defined as an automated work station [14]. Automated work stations are widely used; the transport between different stations is done manually. Examples include multifunctional readers, automated liquid handling work stations, or specialized multi-operational work stations, for sample preparation. In general, also microfluidic systems are included in this group. The work stations are controlled by a PC.

2.2 Definitions and Basics

Table 2.1 Examples for laboratory devices and their LUO. Device

Laboratory unit operation

Shaker

Homogenization

Laboratory centrifuge

Separation

Oscillating mill

Surface enlargement

Microwave digestion system

Digestion of solids and solid containing materials

Automated work stations can be part of integrated systems. This group includes laboratory automation systems, in which different single devices or work stations are connected via transportation elements [14]. These systems can perform many laboratory operations. Usually these are customer-specific configurations, which can be reconfigured or extended [15]. A complete plug-and-play (as in computer

Table 2.2 Selected technologies and methods in physico-chemical analysis. Methods on the basis of chemical reactions [30]

• Chemical equilibrium, electrolytes • Acid-based reaction, titration • Precipitation reaction for gravimetric analysis, titrimetric analysis, masking reactions • Complex formation reactions, complexometric analysis • Reduction–oxidation reactions, redox titration • Extraction, ion exchange • Kinetic methods • Thermal methods (thermo-gravimetric analysis, differential thermal analysis, differential scanning calorimetry)

Methods on the basis of chemical and physical principles [31]

• • • •

Methods and technologies on the basis of instrumental analysis [32]

• Chromatography (e.g., gas and liquid chromatography) • Electrophoresis (gel and capillary electrophoresis) • Molecular spectroscopy (UV/vis spectroscopy, fluorescence and infrared spectroscopy, nuclear magnetic resonance spectroscopy, mass spectrometry) • Atomic spectroscopy (atomic emission spectroscopy, atomic absorption spectroscopy, ICP-MS) • Coupling technologies (sample preparation–separation method, separation method–eparation method, separation methods–, separation methods–pectroscopy, MS/MS-coupling)

Conductometric analysis Potentiometric analysis Amperometric analysis Photometry

27

28

2 Automation in Life Sciences – A Critical Review

technology) is usually not possible due to the still missing standardization. Thus, the reconfiguration requires detailed system knowledge and experience. The control is realized using a global system control (usually a PC), whereby the involved work stations can be controlled by a global system control or their own control software. Integrated systems include table top versions up to space-filling complex systems. Very large systems combine one or more transport elements. A preconfigured and commercially available off-the-shelf system is the automated work cell, a special case of an integrated system. These systems can be purchased as standard systems for a specific application [16], whereby typically a liquid handling station is the central component of a work cell [14, 17]. The highest demand for high-throughput solutions (HTS) can be found in the pharmaceutical industry, especially in drug development and design, as well as in the field of biotechnology. Both areas represent the main drivers for the development and optimization of processes [18–20]. Besides the classical high-throughput applications, an increasing demand exists in the quality control of food technological processes [21], chemical analysis and environmental analytics [2, 22], chemical syntheses [23–25], medicinal chemistry in clinical laboratories [3, 26, 27], and forensics [28, 29]. Physico-chemical metrology is applied, which uses technology and methods based on chemical reactions, chemical and physical principles as well as instrumental analytics (see Table 2.2). The following sections will give an introductory overview about the principles and automation solutions in bioscreening, automated synthesis, and automated analytical measurements.

2.3 Automation in Bioscreening 2.3.1 Overview

Drug development and drug discovery currently are characterized by an unprecedented productivity due to the high degree of automation. The financial pressure, which arises from increasing costs for the market introduction of new drugs, resulted in an escalation of fusion and takeover activities, which on one hand guarantees profits in the short run, but on the other hand has a negative influence on the scientific dynamics [33]. Together with increasing safety requirements from the regulating authorities this is seen as the driving force for further research and development activities in pharmaceutical and biotechnological enterprises that develop new drugs in close cooperation [34]. In [35], a summary is given of the main categories and parts of an HTS process, which enable these high sample throughputs in HTS or uHTS and a process optimization: assay methods and detection, liquid handling and robotics, and process control and information management. The most important factors for the optimization of HTS processes can be divided into three main groups: time, costs, and quality. Together they build the “magical triangle of the HTS” [36, 37]. Goals of the HTS are the increase of screened samples per day and the number of screening experiments per year with minimal processing time per sample and short project duration. Another important factor is the minimizing of

2.3 Automation in Bioscreening

costs for reagents, consumables, instrumentation, and operating personnel [36]. The successful efficiency increase due to HTS can be demonstrated using the example of Novartis. The number of single compounds per well could be increased between 1998 and 2006, whereby the costs per substance significantly decreased (see Figure 2.1). Miniaturization of the screening processes is an essential tool to achieve this increase in efficiency. HTS and uHTS processes are characterized by a constantly increasing density of the wells (96, 384, 1536, or 3456 wells) due to the handling of microplates [18, 38]. Table 2.3 shows a comparison of parameters in classical screening and HTS. Automation in bioscreening is dominated by methods of HTS. HTS is a method, which is mainly used in pharmaceutical research to execute 120 100 80 60 40 20 0 1998

1999

2000

2001

Entities [% of 2007]

2002

2003

2004

Cost/Entity [% of 1998]

2005

2006

2007

FTEs [% von 2007]

Figure 2.1 Development of throughput, costs, and resources invested in HTS projects of Novartis from 1998 to 2007; dark gray: single compounds per well (entities); light gray: costs per single compound; dotted line: full-time employees (FTEs) working on these HTS projects. (Redrawn from Ref. [36]). Table 2.3 Comparison between classical screening and high throughput screening [38]. Classical screening

High-throughput screening

Single vials

Vessels in array-format (e.g., 96 well plate)

Large volumes (ca. 1 ml)

Small volumes (50–100 μl)

Large sample amounts (ca. 5–10 mg)

Small sample amounts (ca. 1 μg)

Sequential addition of components

Simultaneous addition of components

Mechanical impact 1:1

Mechanical impact 1 : 96

Dried components, customer-specific solutions

Components in solution, unique solvent (e.g., DMSO)

Slow and labor intensive

Fast and efficient assays (ca. 1 min per step and 96 well plate)

29

30

2 Automation in Life Sciences – A Critical Review

biochemical, genetic, or pharmacological tests for ten thousand to millions of compounds. If more than 100 000 substances are screened per day, this is called ultra-high throughput screening (uHTS). Using HTS enables the investigation of new biological active compounds, from which lead structures are derived for the development of new drugs [39]. Two principal test procedures can be used for the development of pharmacological active compounds: target-based assays and phenotype-based assays. In target-based screenings the interaction of test compounds with defined goal structures (targets) is investigated [40]. Targets include, for example, proteins, which are connected to a disease or a physiological process. Target-based screenings represent the most common form of the screening of low molecular compounds for the determination of their biological activity in pharmaceutical industry. They are usually performed in microplates with purified or nonpurified proteins or indirect with cells, which represent the target protein. The interaction of a test compound can directly be determined in binding assays (usually elimination of a labeled reference ligand from the target) or indirectly over the influence of signal paths activated by the target protein (e.g., activation of second messenger, protein–protein interaction, protein-phosphorylation, and gene activation), and enzymatic reactions. Thereto especially biochemical methods are used, which are based on the change of the color intensity, fluorescence, or luminescence. Luminescence-based methods produce a high signal change and are thus better qualified than photometric or fluorimetric methods. The signal-to-noise ratio in photometric or fluorimetric methods is decreased for substances with inherent color or inherent fluorescence. Another type of highly sensitive methods is scintimetric test methods such as radio ligand binding studies. In addition, compound properties such as solubility and stability play an important role and have to be taken into consideration [41]. In the phenotype-based screening, effects of test compounds on living cells or tissue, which means the influence of the compound addition on the cell or tissue phenotype, are investigated [42]. The effect of a test compound is determined using a phenotypic change, such as the change of the cell form, the cell growth, or the cell function. The knowledge of the molecular target prior to the screening is not required: often, the screening method is used for the identification of a molecular target. Numerous parameters have to be controlled to avoid a falsification of the screening. Phenotype-based screenings are mainly used for the screening of molecule libraries containing higher molecular compounds such as proteins, DNA, and siRNA. Besides cells and tissue, complete organisms such as fish embryos can also be used as model systems. Automated microscopy (high content screening, HCS) is the main tool for phenotype-based screenings. HCS usually enables a lower throughput compared to a target-based screening. The borders between HTS and HCS are flowing, since often similar processes are used [43]. HTS enables the rapid determination of thousands to hundreds of thousands of small molecules in one measuring series using in vitro and cell-based assays, whereas HCS methods change the high throughput capability against a great biological and phenotypical complexity in the assays [44].

2.3 Automation in Bioscreening

Automated systems are used in all of these areas. Besides industrial applications, an increasing demand exists for HTS methods in academic research [45]. HTS laboratories put in a remarkable pioneering effort and realized strict quality assurance methods such as Z factor monitoring, algorithm for pattern recognition, the regular use of pharmacological standards or performance monitoring for liquid handlers and plate readers. In addition, the latest innovations in precision of low volume liquid handling (e.g., with acoustic dispensers) and the performance of measuring devices (e.g., multimodal plate readers) could be achieved due to the high demands of HTS laboratories [33]. Table 2.4 gives an overview about drugs that have been developed using methods of HTS, their indication, target class, and the year of HTS and FDA approval. 2.3.2 Automation Devices in Biological Screening 2.3.2.1 Standardization of Sample Formats

HTSs are time and cost intensive and thus mainly realized in full or at least partially automated laboratory automation systems. Robots or automats for liquid handling, data acquisition (reader, cameras), and also for cell culturing are used. To ensure the simultaneous processing of high numbers of samples, microplates with 96, 384, 1536, or 3456 wells are used. This is also connected with a significant reduction of the test volumes. The mostly rectangular microplates are made of plastics (usually polystyrene, rarely polyvinyl chloride) or glass for very special applications. They contain different isolated wells in rows and columns. The American National Standards Institute (ANSI) and the Society for Laboratory Automation and Screening (SLAS) defined standards for the dimensions of 96 well microplates [55–59]. These standards are used by manufacturers of microplates but also vendors of devices using microplates. In the early 1990s, 96 well microplates were mainly used in HTS processes in the pharmaceutical industry and biotechnology. Within the last 20 years, microplates with an increasing number of wells and decreasing volumes per well have been introduced to the market [36, 37] (Table 2.5). The wells are available in different forms: F bottom (flat bottom), C bottom (flat bottom with minimal rounded corners), V bottom (conical bottom), and U bottom (U shaped deepening). The most common format in pharmaceutical and clinical diagnostic laboratories is the 96 well format with an 8 by 12 matrix. Typical volumes are in the range of 100–400 μl per well. Higher density microplates with 384, 1536, or more wells are typically used for screening applications, where high throughputs and low assay cost per sample are required. The volumes are decreased down to 5 μl. An even smaller working volumes of only 25 nl can be used on 20 000 well plates [60]. The development of microplates with high well numbers and reduced volumes per well follows the general trend of an increasing miniaturization in the field of HTS [61]. The goals of miniaturization are closely connected with the goals of the pharmaceutical industry and biotechnology. This includes a reduction in costs, faster turnaround, and low storage requirements [18]. Thus, a cost reduction from

31

Cancer HIV Diabetes Cancer HIV Cancer Pulmonary hypertension HIV Hyponatremia Thrombocytopenia

Tipranavir (Aptivus; Boehringer Ingelheim)

Sitagliptin (Januvia; Merck & Co)

Dasatinib (Sprycel; Bristol-Myers Squibb)

Maraviroc (Selzentry; Pfizer)

Lapatinib (Tykerb; GlaxoSmithKline)

Ambrisentan (Letairis; Gilead)

Etravirine (Intelence; Tibotec Pharmaceuticals)

Tolvaptan (Samsca; Otsuka Pharmaceutical)

Eltrombopag (Promacta; GlaxoSmithKline)

Erlotinib (Tarceva; Roche)

Sorafenib (Nexavar; Bayer/Onyx Pharmaceuticals)

Cancer Cancer

Gefitinib (Iressa; AstraZeneca)

Indication

Drug (US brand name, vendor)

Table 2.4 Overview about drugs developed using HTS methods [33].

Cytokine receptor

G protein coupled receptor

Reverse Transcriptase

G protein coupled receptor

Tyrosine kinase

G protein coupled receptor

Tyrosine kinase

Protease

Protease

Tyrosine kinase

Tyrosine kinase

Tyrosine kinase

Target class

1997

app. 1990

app. 1992

app. 1995

app. 1993

1997

1997

app. 2000

app. 1993

1994

app. 1993

app. 1993

Year HTS

2008

2009

2008

2007

2007

2007

2006

2006

2005

2005

2004

2003

Year FDA approval

[54]

[53]

[52]

[51]

[46]

[33]

[50]

[49]

[48]

[47]

[46]

[46]

References

2.3 Automation in Bioscreening

Table 2.5 Microplate formats [36, 37]. Number of wells

Volume range

Standard volume

96

100–1.000 μl

200–500 μl

384

30–100 μl

50 μl

1 536

2,5–10 μl

5 μl

3 456

1–2 μl

1–2 μl

20 000

25 nl

25 nl

35 million dollars for assays performed using 96 well microplates to 1.1 million dollars for assays performed using 1536 well microplates was reported [18]. 2.3.2.2 Robots in Bioautomation

Robots have been used in laboratory automation since the early 1980s. In 1982, Zymark Corp. (Hopkinton, USA) developed the worldwide first system for laboratory applications, which was named the “one-armed chemist” by Forbes business magazine [13, 22]. The requirement of the pharmaceutical industry was the main driver for using robot technology in the field of life sciences. Today, robots play an important role for many types of laboratory methods including drug development, drug testing, and DNA fingerprinting. Different reasons can force the use of robots in laboratory processes [22]. This includes the execution of repeating tasks, the reduction of manpower requirements, and the guarantee of equality, and also the minimization of health risks for human operators working with hazardous materials. In addition, the avoidance of sample contaminations and faster and more reliable execution of tasks compared with manual processing are important points to be considered. The definition of the term robot has changed over time. The International Organization for Standardization (ISO) defines a robot in ISO 8373 (2012) as an “actuated mechanism programmable in two or more axes with a degree of autonomy, moving within its environment, to perform intended tasks” [62]. This definition is used by the International Federation of Robotics, the European Robotics Research Network (EURON), and many national standards committees. The robot comprises the manipulator including the actors and control (hardware, software). According to ISO 8373 (2012), a manipulator is a “machine in which the mechanism usually consists of a series of segments, jointed or sliding relative one to another, for the purpose of grasping and/or moving objects (pieces or tools) usually in several degrees of freedom.” A manipulator can be controlled by operating personnel, a programmed electronic control, or any logical system (e.g., cam control, wired logic). An industrial robot is defined as an “automatically controlled, reprogrammable multipurpose manipulator, programmable in three or more axes, which can be either fixed in place or mobile for use in industrial automation applications.” A critical point in the design of robots is robot kinematics. Robot kinematics comprises a freely selectable combination of linear and rotational axes. This results in a large number of possible combinations, whereby not all combinations

33

34

2 Automation in Life Sciences – A Critical Review

are technically meaningful. The selection and arrangement of the axes directly influence the properties and performance features, such as positioning precision, load capacity, efficiency, kinematic mobility, and footprint [63]. Different mechanical structures have been established for robots. The most important structures including the main axes and the resulting work spaces are summarized in Table 2.6. Cartesian robots have one arm with three prismatic joints, whereby the axes are arranged in a Cartesian coordinate system. The cylindrical robot has at least one prismatic and one revolute joint; the axes form a cylindrical coordinate system. The axes of a polar robot form a polar coordinate system. The arm of this robot type has two revolute joints and one prismatic joint. A horizontal articulated robot has two parallelly arranged revolute joints and enables a passive or active compliance in a special selected level. This structure is also called a SCARA robot (selective compliance arm for robot assembly). The arm of a vertical articulated robot has three revolute joints. Parallel robots are characterized by a robot arm, whose limbs form a closed circular structure [63–65]. In addition, there exist a number of other mechanical structures of stationary robots [64, 65] as well as a combination of different synchronized robots such as dual arm robots [66, 67]. Common mechanical structures of mobile robots include wheel- and chain-driven robots, legged robots with two or more legs, robots with omnidirectional moving mechanisms or mobile platforms and driverless transportation vehicles [65]. The selection of a robot kinematic for the tasks to be fulfilled depends on different factors. This includes the requested performance features, investment costs, and the operability. Some robot kinematics are suitable for specific task areas. Parallel robots have a high dynamic and are thus often used for pick-and-place tasks. SCARA robots are characterized by high rigidity in a vertical direction, which makes them ideal for use in assembly tasks [63]. Industrial robots are used for automation solutions in handling and manufacturing. In 2006, the majority of robots was used in the field of mechanical engineering (38%), followed by the automobile industry (28%). Food technology as well as chemical industry had a market share of only 3% each [68]. In bioautomation and HTS, robots are mainly used as central system integrators that transport the samples and labware between the different stations of an automation system (see Chapter 3). Robots can also be used for other tasks such as pipetting and filtration [13]. Dual arm robots enable extended functions, since they can perform motions and tasks analogous to the human operator [69]. They thus open new interesting possibilities in laboratory automation (see Chapter 3) [67]. Due to the combination of two articulated robots that can move either autonomously or synchronized, many transportation and manipulation tasks can be realized. 2.3.2.3 Liquid-Handling Systems

Laboratory robots are of central importance in liquid handling tasks and procedures [70]. In scientific literature, the terms automated liquid handler (ALH), liquid handler, and liquid handling robot can be found. The simplest versions dispense the liquid with motorized syringes or pipettes. Today, liquid handlers

2.3 Automation in Bioscreening

Table 2.6 Typical configuration of robot main axes and resulting work space. Robot type Cartesian robot

Cylindrical robot

Polar robot

Horizontal articulated robot

Vertical articulated robot

Parallel robot

Redrawn from Ref. [63].

Axis configuration Main axis

Work space

35

36

2 Automation in Life Sciences – A Critical Review

are usually robots with a Cartesian structure that dispense a selected quantity of reagents, samples, or other liquids to a designated container. They can have one or two arms. The operator can optimize the liquid handling process with the help of suitable software. This includes, for example, the aspirating and dispensing speed, which has great importance in relation to the viscosity of the liquids being handled. Another important feature is the dispense contact. The pipetting can take place in empty air (air displacement) or the tip can be lowered into the liquid that is already present in the target location (positive displacement). In addition, features such as tip touch (tip can be touched against the side wall of the target container to wipe off any drop that might be hanging on), blowout (final burst of pressure to blow out the last drop of liquid that tends to stay behind in the tip), or mixing (repeatedly aspirating and dispensing of the liquid while the tip remains in the liquid) can improve the pipetting results [71]. Liquid handling systems can perform the dispensing of liquids in different ways. Some systems use single pipettes/cannulas (single channel pipetting system) or eight parallel pipettes (eight channel pipetting system), which in some cases can be individually programmed. This allows a high flexibility in the system, but lowers the possible throughput. Higher throughputs and thus higher efficiency can be achieved with so-called pipetting heads that enable parallel pipetting of 96, 384, or even 1536 liquids (multichannel pipetting system). In general, fixed tips or disposable tips are used in liquid handling systems. Disposable tips are made of various types of plastic and are designed to be replaced after each pipetting action. They are available in different volumes. The tips are usually provided in holders in the microplate format; thus, an easy integration into the automation system is possible. Fixed tips are thin stainless steel tubes that are used repeatedly and thus have to be washed between the pipetting steps. They can be Teflon coated to improve the washing. Due to the required washing steps, these systems are slower compared to systems with disposable tips. Liquid handling systems can also be equipped with ultrasonic or capacitive sensors for liquid level and phase detection or for the detection of leaks and clogs in the tips. A fast detection of the pipetting success can also be realized with optical methods [72]. More complex liquid handling systems can perform different tasks including sample transport, sample homogenization (e.g., shaking or stirring), manipulation [73], incubation, or the transport of sample vials to or from the included devices (e.g., barcode reader, measuring instruments, sample storage, centrifuges, or incubators). Some liquid handlers can fulfill the normal pipetting tasks, but can also realize the sample injection into measuring instruments [23, 70]. They can be customized using different add-on modules, such as centrifuges, PCR (Polymerase Chain Reaction) machines, colony pickers, shaking modules, heating modules, and others. A partial or complete housing is possible for handling toxic compounds. Liquid handling systems are not limited to Cartesian robot kinematics. One example is the Andrew pipetting robot, a horizontal articulated robot, which uses manual pipettes for the pipetting [74]. Automated pipetting systems are most frequently used for liquid handling applications, such as plate replication and plate reformatting, reagent and compound addition, or serial dilutions. Plate replication and plate reformatting are basic operations for a multichannel pipetting system. The goal in plate

2.3 Automation in Bioscreening

replication is to replicate the contents of the source plate to the destination plate(s). Source and destination plates are identical with respect to the number of wells and the well content. Plate reformatting involves moving samples between the 96 and 384 well microplate formats. Plate expansion is a reformatting operation that spreads the wells of a 384 well plate to four 96 well plates, whereas the reverse process of moving the contents of four 96 well plates to a single 384 well plate is called plate consolidation (or compression) [71]. Serial dilutions are used to set up a series of different sample concentrations for an experiment. A high concentration of sample is used as a starting point. A lower concentration results from the transfer of a small sample amount to another well and mixing with dilution solvent. Meanwhile automated liquid handling robots have matured and can be used for a variety of different laboratory processes, including PCR preparation, IC50 plate creation, determination of metabolic stability and solubility, compound purification, P450 inhibition, ELISAs (Enzyme-Linked Immunosorbent Assay), cherry-picking, titration studies as well as screening assays [75]. Classical liquid handling systems are limited to dispensing volumes between 1 and 1000 μl. When handling liquids in volumes below 1 μl, common pipetting systems are generally insufficiently accurate and need to be supplemented by specific nano-dispensing technologies. In general, it is possible to distinguish between dispensing with direct contact to the liquid and contactless dispensing. Contactless dispensing can be realized with piezoelectric, magnetic, or other physical principles [76]. The so-called drop-on-demand technologies enable the direct contactless dispensing from microplates onto any desired substrate. This includes microplates with up to 1536 wells or any other plate or slide format. Depending on the type of solvent, drops of approximately 1 nl can be dispensed from every nozzle at frequencies of up to 600 Hz. This enables ultra-high throughputs in the volume range between 1 and 1000 nl. A direct determination of the drop volume is possible using camera-based approaches [77]. 2.3.2.4 Additional Components

Besides liquid handling, different other tasks have to be implemented for fully automated bioscreening. All devices that provide the system with functionalities that cannot be delivered by the robot or enable a more cost efficient completion of the tasks, are called robot periphery. This comprises all devices and appliances that directly cooperate with the robot [78]. In the field of laboratory automation, process periphery includes providing stations for labware (e.g., hotels or racks for microplates, vials, sample containers, pipetting tips, etc.), and additional components, such as thermomixers, centrifuges, incubators, sealers, piercers, or devices for opening and closing labware. For screening applications, measuring devices such as plate readers on the basis of spectroscopy have to be included in the systems. Manipulation is the physical handling or treatment of laboratory materials. This includes the opening and closing of containers (e.g., screwing, crimping), the application of labels and barcodes, the handling of consumables as well as the movement of objects, such as doors of laboratory balances or covers of thermomixers and centrifuges [13].

37

38

2 Automation in Life Sciences – A Critical Review

Sealing Systems [79]: To avoid evaporation from the microplate wells or contamination of the samples in the microplate wells, the plate must be sealed suitably. In the easiest case, this is realized with simple lids that are placed over the plates. These lids ensure contamination-free processing and transport of the plate. However, due to the loose fitting, they often show significant evaporation effects from the outer wells compared to the wells inside the plate. To overcome this problem, different sealing materials have been developed, which can be automatically applied to the plates. This includes adhesive backed films, cap mats, and heat applied foil and film sheets. Adhesive backed films adhere to the surface of the microplate. They are intended for the use with aqueous solutions to reduce the evaporation of the liquid from the wells. Adhesive films are usually used at room temperature and have only a limited chemical stability. Thus, organic solutions in the wells may dissolve the adhesive materials, which leads to a contamination of the samples. In addition, the adhesive material can clog the needles of pipetting and auto sampler systems and the material can be detected in chromatographic system. To overcome these problems, new materials such as pressure-activated adhesives have been developed, which are firmly pressed onto the microplate surface to form a strong seal. The main applications are in HTS, genomics, and clinical applications. Another possibility is the use of cap mats, which are single semi-rigid covers that fit into each well of the microplate. Theses mats are used to provide protection from the atmosphere at room temperature; they can be used in freezing applications but are not intended for heating applications. Cap mats are mainly produced of plastics such as ethyl vinyl acetate (EVA, not chemically resistant) or silicone elastomers (where the underside of the sheet can be treated with polytetrafluoroethylene (PTFE) for improved solvent resistance). They are mainly used in pharmaceutical analysis for sealing microplates prior to the injection into analytical systems. A main disadvantage is the lack of a tight fit. Heat applied foils are placed onto the surface of the microplate and heat is applied evenly for several seconds using a semi-automated benchtop sealer unit. This technology offers many advantages compared to other plate sealing methods, which include the secure and uniform sealing of wells, no contamination from the adhesives, solvent resistance, a quick application process, and a tight sealing even with vigorous mixing and shaking processes. In addition, they are very cost-effective. The films are composed of a thin layer of polypropylene on one side and aluminum on the other side. The seals are optimized for peelable and pierceable uses. Thermosealers are available for manual and automated processing. Mixing [79]: Mixing of liquids in sealed microplates is commonly performed in many bioscreening applications. Vigorous mixing is required when reagents, buffers, or standards are added to the wells of the microplate. Also liquid–liquid extractions require a good mixing to ensure an efficient transition of the compounds from the aqueous to the organic phase. Liquids in microplates are usually vortexed in flat-bed shakers. These shakers allow the continuous shaking with adjustable speeds or pulsed mixing; heat can also be applied using thermoelectric coolers or conventional recirculating chillers. Automated shakers are available for single plates or parallel shaking of up to eight plates. Centrifuges: Centrifugation is one of the most important and widely applied research techniques in biochemistry, cellular and molecular biology, and in

2.3 Automation in Bioscreening

medicine. Current research and clinical applications rely on the isolation of cells, subcellular organelles, and macromolecules, often in high yields. Centrifuges use centrifugal forces (g-force) to isolate suspended particles from their surrounding medium on either a batch or a continuous-flow basis. Applications for centrifugation are many and may include sedimentation of cells and viruses, separation of subcellular organelles, and isolation of macromolecules such as DNA, RNA, proteins, or lipids [80, 81]. The centrifugation of microplates puts high requirements on the centrifugation systems. Special microplate rotors are required to accommodate the use of microplates; usually different rotors are required for conventional and deep-well plates depending on the height or z-axis clearance. The typical rotors can hold two or four microplates. Centrifugation of microplates is used, for example, for protein precipitation processes of plasma or to remove particles and clots from plasma prior to taking an aliquot for further investigations. It also plays an important role in cell cultivation and cell handling. In combination with heat and vacuum, centrifugation can also be used for the evaporation of solvents [79]. Barcode readers: A barcode reader or scanner is an electronic device that can read printed barcodes. It typically consists of a light source, a lens, and a light sensor translating optical impulses into electrical ones. Originally, barcodes were developed to systematically represent data by varying the widths and spacing of parallel lines. These are called linear or one-dimensional (1D) barcodes. Later, two-dimensional (2D) codes were developed, using dots, hexagons, rectangles, dots, and other geometric patterns in two dimensions. Barcode readers can be differentiated by the technology used. Common technologies include pen-type scanners, laser scanners, LED (Light-Emitting Diode) scanners (also called charge coupled device readers, CCD), camera-based scanners (video camera readers and large-field-of-view readers) or omnidirectional barcode scanners. Also, cell phones and smart phones can be used as barcode readers. Based on the housing design, barcode readers can be differentiated into handheld scanners, pen scanners, stationary scanners, fixed position scanners, automatic readers, or wireless scanners [82]. Plate readers: Plate readers are used to detect biological, biochemical, or physical events of samples in microplates. They are widely used in drug discovery, bioassay validation, or quality control in pharmaceutical and biotechnological industries or in academic institutions. Common detection modes used in readers include absorbance [83], luminescence, fluorescence intensity, fluorescence polarization, time-resolved fluorescence [84] or light scattering, and nephelometry. Plate readers enable the simultaneous readout of complete plates and can thus ensure a high throughput also for the measuring phase. Many of the detection modes are available stand-alone in dedicated plate readers, but can today also be combined into one instrument (multi-mode plate reader). The range of applications for multi-mode plate readers is extremely large and includes, besides the HTS of compounds and targets in drug discovery, ELISAs, protein and cell growth assays, protein–protein interactions, reporter assays, nucleic acid quantitation, molecular interactions, enzyme activity, cell toxicity, proliferation, and viability, ATP (Adenosine Triphosphate) quantification or immunoassays [85].

39

40

2 Automation in Life Sciences – A Critical Review

Solid phase extraction: Solid phase extraction (SPE) is a versatile and selective method of sample preparation. The analytes are bound onto a solid support material and a wash solution is passed through the sorbent bed to ideally remove adsorbed contaminations from the sample matrix, while the analytes retain in the solid phase. They are selectively eluted for further workup and analysis with a specific solvent [86]. SPE is usually performed in cartridges with different sorbents, but there exist also a variety of 96 well SPE microplates. This includes single piece molded plates and modular plates with removable wells. The different vendors offer the plates with a varying number of sorbents including polymer sorbents; the sorbent bed mass varies between 10 and 300 mg. Different automated SPE systems in microplate format are available on the market, which use either vacuum based or positive pressure extraction of the samples [87]. 2.3.3 Application Examples

There exist many applications of automation systems in bioscreening investigations. This includes protein precipitations, the testing of enzymatic or cellular assays, or cell culturing processes. The identification and development of antiviral drugs is one example for bioanalytics in automated HTS processes. A method for the independent screening of two compound libraries using a cell-based phenotypical HTS assay was described in [88]. A fully automated system was used for the screening process including different substations such as liquid handling, incubation, storage of the 384 well plates and a fluorescence plate reader for the detection. The central transport element for combining the different stations is a laboratory robot on a linear rail. Twenty-five out of 45 identified hits could be validated. The results formed the basis for further investigations for the determination of the chemical structure focusing on their biological activity. Another automation system for industrial and academic HCS is composed of a completely integrated cell analyzer on a platform with a 6 m long linear rail, two stand-alone microscopes, and an additional cell analyzer [89]. The complete system thus has three automated islands that are connected via manual sample transport. The automation system was designed for HCSs of living and fixed cells. Four RNAi screening campaigns were successfully applied and run fully automated. The formulation of protein-based drugs in general is a complicated and time consuming process due to the complex protein structure and the specific physical and chemical properties. Different formulation techniques for these special compounds and their applicability in high-throughput processes were summarized in [6]. This can also be called high-throughput formulation. The automated processing of the design of experiment, sample preparation, analysis, and data evaluation in high-throughput processes enables a fast and comprehensive insight into the physical and chemical properties of protein-based drugs. The sample preparation is realized using robot supported liquid handling systems, which can be directly coupled to suitable analytical systems. Typical compounds used are water, buffers, salts, sugars, preservatives, proteins, dyes, and reagents. Depending on the project goal and the required throughput,

2.3 Automation in Bioscreening

these systems can be extended with handling systems for powders, robot arms, incubators, cooling stations, pH sensors, barcode readers, stackers, or shakers. The analysis includes the determination of the protein concentration using UV absorption, which is also used for quality control. The turbidity can be measured with wavelengths above 320 nm. The formation of particles is determined by the measurement of light scattering. Additional technologies include fluorescence spectroscopy or microscopy. Such systems can be enhanced on demand with additional separation technologies such as chromatography or electrophoresis [6]. The investigation of the toxicity of compounds in bioscreening processes is another important task; an overview about actual and future approaches is given in [9]. The challenges for toxicity investigations include the testing of high numbers of existing chemicals without information regarding their toxicity, the testing of high numbers of new chemicals and new materials, the assessment of potential adverse effects in all phases of life, the assessment of the potential toxicity for risk groups in the human population, the minimization of animal experiments, the reduction of time and costs, data acquisition for the assessment of the human risk as well as providing a help for regulatory decisions [9]. In 2008, the Tox21 initiative was started with the goal of toxicity determination of about 10 000 environmental compounds and storage of the acquired data in compound libraries [90, 91]. Toxicity testing is usually performed in vivo using animal experiments. Thus, information regarding the oral, dermal, and ocular toxicity, immunotoxicity, genotoxicity, reproduction, and development toxicity as well as cancerogenity can be generated. In vitro tests have the advantage of significantly decreased test times and the avoidance of a high percentage of animal testing. The use of human cell cultures enables a higher correlation of the results, which is not given in animal experiments due to the different species. To ensure a high throughput in the testing of high numbers of compounds, a fully automated HTS system was established [90]. The screening system Tox21 enables the screening and profiling for 10 000 compounds with triplicates within a week [91]. The core of the system is an articulated robot with six axes, which is used as a central transport element and is equipped with a special gripper and a barcode reader. Different stations are positioned around the robot. This includes two incubators, a pin tool station, and two contactless acoustic dispensers for the liquid handling in nanoscale. Two additional devices are used for dispensing in microscale. One of them enables the simultaneous addition of up to four reagents in 1536 well plates and the second a high-speed dispensing for one reagent with eight tips. Two plate readers on the basis of fluorescence and luminescence are used for the detection. Different technologies from different scientific disciplines are available for the development of new lead structures for agricultural products. The classical way to find suitable fungicides, herbicides, insecticides, or antiparasitic lead structures is the random screening of chemical and natural compounds in a test organism. This approach can be very time consuming and cost intensive and does not give any evidence regarding the involved mechanism of action. The application of HTS methods is also possible for the development of agricultural chemicals [92]. These methods are characterized by a combination of elements from the rational design

41

42

2 Automation in Life Sciences – A Critical Review

with elements from random screening. They use high-throughput assays on the basis of the agar diffusion method. To use these assays in an automated operation, different processes are required in the assay preparation. This includes also special high-density agar spotting technologies to distribute the media in high density on 9 × 9 in. bioassay plates. For the validation of these assays, the 96- or 384-well format was used, where the latter reduced the requirements for sample storage by a factor 4. Synthetic compounds or extracts are prepared robot assisted in the required concentration according to a chemical library and transferred to 96 well microplates. A laboratory robot system and a liquid handler are used. The media for the agar plates is prepared in flasks (vol. 2 l) and applied to 9 × 9 in. bio assay plates with a dispenser. After autoclaving and cooling off, the specific recombinant test organisms are applied to the plates. The compounds to be tested are applied to the agar plates using a 96-pin replicator system, whereby up to six 96 well arrays (576 different samples) can be positioned at one bioassay plate. The natural compounds are tested analogously whereby other formats of bioassay plates are used. The described assay and automation methods are not limited to the field of agro chemistry and can also be used in pharmaceutical processes [92]. Another example is an automated procedure for the investigation of pathogens in potatoes using immunofluorescence (IF) for use in large-scale routine testing. The bacterium Ralstonia solanacearum has a pathological activity against potatoes and is responsible for brown rot, which leads to enormous economic damage in agriculture. In a routine test, a high-throughput system is used for staining and washing of antibodies on IF slides. The preparation of the slides, the fixation of the samples on the slides, and the microscopic evaluation remain manual steps in the process. Nevertheless, one person can test 288 samples with two different antiserums or 576 samples with one antiserum per day. Fully manual processing would require approximately 2 days [93]. Efficient sample logistics using a database in the field of drug development is a crucial point. Automated production processes in medicinal chemistry including an innovative combination of compounds to be screened (screening collections) and the execution of the assays were described [7]. The samples are distributed into vials with volumes of 4 or 20 ml, where the latter are sent to the Central Compound Management Group (Rahway, USA). The samples are then transferred from the 4 ml vials into 1.4 ml 2D barcode labeled tubes (matrix tubes), marked with a barcode, stored in an automated sample storage system and the sample data are stored in a database. Processes of sample generation, data acquisition, sample delivery, and sample reformatting were realized using a specific sample handling module [7]. The automation of cell culturing processes is a relatively young field in the automation of bioprocesses. The cell cultivation is a general process in the research and laboratory work to expand cells for subsequent bioscreening. Traditionally, these are repetitive and long lasting manual work steps with a risk of contamination and human errors associated with high costs. Regarding this, the automation of cell cultivation is largely required to simplify the processes in combination with increased quality, repeatability, precision, and stability of the individual batches under consistent sterile environmental conditions with

2.4 Automation in Chemical Sciences

efficient costs and time [61, 94, 95]. The automated methods are independent of skill levels and daily constitution of laboratory staff in combination with consistent quality and performance of the methods. Currently, there exist a variety of solutions for automated cell cultivation to improve the laboratory work [96]; these systems are mainly oriented to the culturing of 2D adherent cell lines. A suitable system enabling the cultivation of 2D suspension cells as well as 3D cells including pellet cultures, alginate beads, or hanging drops has been reported [97, 98]. The system consists of a liquid handler for all liquid handling steps, an incubator for the incubation of the different cell lines, a centrifuge, and a cell counter. The central element of the system is the liquid handler for the aspiration, dispensing, and transport of solutions. It is equipped with a span-8 pipetting head and has an integrated gripper at one rail. The one to eight pipetting channels are independent. The liquid handler deck is equipped accordingly to the requirements of automated cell cultivation. Automated laboratory positioners (ALP) and devices are available for specific handling of cell culture flasks or as placeholders. Furthermore, a microplate shaking ALP is integrated on the liquid handler deck for mixing solutions and cells in microplates or flasks with an adjustable shaking motion. For short incubation periods of the cell culture flasks at 37 ∘ C, two on deck incubators are available. These can directly be accessed from the liquid handler deck. For simulating the manual handling of the cell flasks, two 3D tilt racks have been integrated. The racks enable the tilting of the flasks in x, y, or x/y axis, pivoting, and knocking the flask with an adjustable angle and speed.

2.4 Automation in Chemical Sciences 2.4.1 Overview

Automation in chemical sciences was mainly driven by the development of methods of combinatorial and parallel chemistry. Combinatorial chemistry is a method of medicinal chemistry, which was developed in the 1980s. This method combines simultaneously several hundred thousands of organic compounds to bigger molecules, which are the basis for finding new biologically active compounds. Combinatorial chemistry tries to synthesize a multitude of new molecules with an identical scaffold by the combination and variation of different substituents, from which the ones with the wanted or optimized properties can be selected. In classical synthesis, few educts (A and B) react with each other and it is expected that one single or a main product (AB) is formed. In combinatorial chemistry, in one synthesis approach, complete groups of educts (A1 to An ) reacted with a second group (B1 to Bn′ ). It is anticipated that a broad range of final products (A1 B1 to Ai Bi ) are formed. While starting with n + n′ educts a number of n × n′ products can be formed. In most cases, combinatorial synthesis produces even more products. A greater variability can be achieved using polymer compounds that are designed of a limited number of units, but differ in their order (sequence).

43

44

2 Automation in Life Sciences – A Critical Review

Important examples include nucleic acids and proteins. Here, an exponential relation exists between the number of units (n) and the sequence length (k). A nucleotide sequence with 3500 base pairs can thus be assembled in 1.62 × 1021 different ways. Combinatorial drug design uses the principle that a high number of compounds to be tested increase the number of possible new lead structures. A second advantage can be seen after new lead structures have been found in a trial and error approach. Since it is relatively easy to systematically modify structures, the necessary time for the optimization of lead structures is dramatically reduced with methods of combinatorial chemistry. The majority of combinatorial compound libraries is developed using solid phase synthesis [99]. The advantages of this method are a nearly complete conversion in the reactions (often more than 99.8% per reaction step) and a simple compartmentation and automation of the synthesis in a combinatorial reaction approach. After completion of the reaction, the products are tested in a complex, where the polymers or the polymers and linkers are separated from the reaction products using suitable methods such as low pH or UV radiation. An important method of solid phase synthesis uses small polymer beads as support materials (one bead-one compound) and is based on a sophisticated distribution system of the educts. This method is also called split-and-combine or mix-split technology. Thus, it can be ensured that each bead contains only one reaction product. Another approach is solution phase synthesis [100, 101], since solid phase syntheses has also some disadvantages. It is, for example, complicated or even impossible to monitor the progress of solid phase synthesis. Another deficiency is the limited repertoire of known reaction mechanisms. Three stages are required for the development of a compound library, namely, the combinatorial synthesis, the screening of the library components, and the determination of the chemical structures of the bioactive compounds. The main problem in the workflow is the decision, which theoretical producible compounds (virtual library) should be produced in reality. Due to the purpose of the compound library, very similar or very diverse compounds are planned to increase the hit rate in biological tests. For the synthesis, standardized processes and robots are mainly used. Main users are medicinal chemistry and pharmacology. In these areas, new drugs can be found or optimized. Whereas in the 1980s, scientists tried to synthesize as many compounds as possible (so called compound libraries), the trend changed later to the synthesis of purified and well characterized single compounds. Thus, the boundaries to parallel synthesis are flowing. Parallel synthesis is a method for simultaneous and automated synthesis and purification of a high number of structurally similar compounds for the pharmaceutical chemical research. Thus, compared to classical synthesis, multiples of test compounds can be synthesized within a short timeframe. Parallel synthesis is a widely used technology, which enables the development of new compounds and the screening for optimal process conditions. In pharmaceutical industry, parallel synthesis is used for faster drug discovery and the development of potential medicinal compounds. In addition, it enables the optimization of

2.4 Automation in Chemical Sciences

processes due to a better understanding of synthesis ways, solvent systems, optimal temperatures and concentrations, correct reagents, reaction times, and the selection of catalysts. Since in parallel synthesis single reactions are not monitored individually, the hit ratio of desired molecules is expectably lower compared to the classical single synthesis of the compounds. This also applies to the purity of the resulting products. Thus, automated purification methods such as automated HPLC (High Performance Liquid Chromatography) fractioning, the so-called parallel purification, plays an important role. The combination of parallel synthesis and parallel purification results in a significant increase of efficiency. The parallel synthesis can be realized in solution or as a solid phase reaction. In the solid phase reaction, a substance is produced in a chemical reaction, whereby the base molecule is chemically linked to a polymer carrier. On this carrier, different steps including the reaction, washing steps with solvents for the purification of the bonded molecule, reaction, washing steps, reaction are executed, until the final target molecule has been produced. This molecule is then separated from the carrier and selected as a pure substance. Modern data evaluation yields the databases for the selection of a reagent, the design of virtual compound libraries, and the analysis of different batches [102]. Deconvolution methods summarize substances of a compound library mathematically in groups (pooling) and calculate the most probable candidates for a biological activity. For the final isolation of the active compounds more or less syntheses are required. The tagging and labeling methods use chemical or physical properties of inert markers, which are added to the polymer carriers. The chemical or molecular tags create a chemical binary code, which clearly labels the compound. This code can be read and the substance can be identified with capillary chromatography (GC). Unfortunately, it is not easy to create really inert chemical markers [103]. For this reason, electronic tags have been used since 1995. In this method, a group or even each individual polymer bead has a label (ID) in the form of a microchip. During the synthesis, the ID of each reaction partner is added to a microchip with a radio signal. Once all reaction blocks are added, the microchips contain the complete synthesis path of each compound in the library. Since each microchip is sensitive to a different wave length of the radio signal, all products can be identified immediately. These new technologies in chemical pharmaceutical drug discovery have been established in many companies. Nearly all research oriented companies use them for time-optimal, economical, and efficient drug discovery. 2.4.2 Automation Devices for Combinatorial Chemistry

Combinatorial synthesis is currently characterized by full automation. In analogy to automated test procedures, it is called high-throughput synthesis. Different automation stages can be realized. The first group includes semi-automated systems that can perform a specific task, while other tasks still have to be done manually. Systems that perform washing procedures with up to six online solvents are one typical example. The washing steps are performed

45

46

2 Automation in Life Sciences – A Critical Review

automatically, whereas the addition of reagents and other process steps have to be done manually [101]. In contrast, fully automated synthesizers can perform both: the addition of the reagents and the purification steps. In both cases, benchtop as well as modular systems are available. The requirements for automated synthesis include the preparation of multimilligram amounts of compounds as well as microscale synthesis; cheap, adaptable, easily transportable and cleanable reaction vessels; modularity of equipment for a range of options; option for diverse chemical reaction conditions (e.g., inert atmosphere, temperature control, corrosive materials); processing and monitoring for flexible work-up procedures including extraction and evaporation methods; avoidance of cross-contaminations during the reagent transfer and synthesis as well as a rapid delivery of reagents and rapid wash cycles [104]. 2.4.2.1 Vessel Based Systems

The early peptide synthesizers had one reactor vessel only and were based on pressure and valve-controlled distribution systems for reagents and reactants. Newer systems still use this concept, but enable parallel processing of several dozen syntheses with different temperatures and a variation of the solvents, reagents, and reaction times. Flow synthesizers are characterized by the storage of all solvents, reagents, and building blocks in a closed system under inert gas atmosphere. The chemical components are transferred to the reaction vessels and can also be removed from the vessels by applying positive pressure. This enables reactions under complete inert conditions, which is especially important for air- or water-sensitive compounds. Many automation devices ensure an independent temperature control of the different vessels. Thus, defined reaction conditions are adjustable for each experiment. Current systems for solidphase-based automated synthesis enable a completely automated workflow of, for example, peptide synthesis. This includes the preparation and distribution of the resins, the dilution and addition of the amino acids and coupling reagents, the different synthesis steps, the washing process between the coupling steps, the protection and removal of protecting groups as well as the cleavage of the final products. The synthesis can be realized in reaction blocks with 96, 48, 24, or 8 reactors with volumes from 1 to 40 ml in closed vials. Special patented filtration needles have been developed for the filtration steps [105]. Especially, traditional organic synthesis is still performed in a conventional way in classical glass apparatus. These syntheses can also be automated with suitable systems in order to save valuable time during product development. Due to different requirements depending on the respective chemistry, customer-specific, modular system solutions are in demand. The requirements range from systems with low volumes for the screening of reaction conditions or solvents over parallel syntheses systems for process optimization to classical automated laboratory reactors with volumes of 1–2 l. After the first steps in the process of drug development or combinatorial chemistry, additional screening steps are necessary, which are performed with a smaller number of substances. Different reaction blocks have been developed as a central part of automated synthesis systems. They are mainly constructed as solid blocks or simultaneously controlled reactor arrays. They contain different numbers of reactors that can be stirred and tempered simultaneously

2.4 Automation in Chemical Sciences

or independently depending on the specific system. Thus, all vessels can be operated at individual temperatures between −90 and +200 ∘ C (depending on the thermostats used) and variable stirring speeds between 100 and 1200 rpm. Especially in solid phase synthesis, gentle stirring is required. Different possibilities exist for the filtration of the reaction solutions and the removal of the filtrates into waste. Tempering of the reaction vessels is mainly realized with thermostats or cryostats with silicone oils. If lower temperatures below −40 ∘ C are required, liquid nitrogen evaporators are used as cooling devices. In combination with electrical heating, they can enable individual tempering of the reaction vessels [106]. The reaction volumes vary between 5 and 10 ml; up to several hundred reactors can be used in parallel. Different adapters are available to ensure the use of a great variety of reactors. The reactors are mainly made of polypropylene, Teflon, or glass. With the help of these systems, synthesis paths, reaction temperatures, or different solvents can be screened. In addition, these reaction blocks are also used in the synthesis of building blocks. In order to save reagents and solvents, smaller reaction devices have been developed as well. An eightfold parallel reactor was introduced for the screening of catalysts [107]. The reactor system consists of up to three arrays with eight modular single reactor blocks. The concept includes independent stirring using magnetic stirring and independent pressurization of every single reactor with pressures of up to 150 bar. Tempering of the system can be realized with a conventional thermostat that provides identical temperatures for each reactor array. The single reactor blocks are made of chemically inert V4A steel 1 4571. They are equipped in the upper part with gas connections and injection modules. The exchangeable reactors with up to 3 ml reaction volume are fixed in the reactor block. Other systems can process up to 3840 reactions in parallel using 10 reactions blocks with 348 (0.5 ml), 96 (3.5 ml), or 40 (8 ml) vessels. Once possible synthesis paths have been identified, detailed process development and optimization is required. Therefore, the reaction conditions have to be precisely adjustable to ensure good reproducibility. An automated and parallel approach is used for efficient and economic processing. Available compact work stations enable fully automated parallel processing of up to eight reactions with 100–450 ml reaction volume. In general, up to 1000 building blocks can be synthesized per year with these machines. Due to the system dimensions, installation in a classical chemical hood is possible. The included software not only enables the control of the system, but can also be used offline for synthesis preparation or the evaluation of protocols. Simple and complicated synthesis procedures, such as the automated addition of solvents and reagents, temperature control (−50 up to maximum +170 ∘ C), control of the temporal operational synthesis sequences in homogenous phase or in solid phase can be realized with a few steps. The automation concept is aligned to changing synthesis requirements; only glass reaction vessels are used. Besides the general synthesis, a semior completely automated liquid–liquid extraction unit can be integrated to enable the production of an extracted raw product for further evaporation of the solvent. Semi-automated synthesis stations exist for less complex tasks. When the reaction conditions are known, automated laboratory reactors are used for the scale-up processes. These systems can be customer-specifically

47

48

2 Automation in Life Sciences – A Critical Review

equipped with automated solid, gas, or liquid delivery devices. Tempering is realized with a recirculating chiller directly coupled to the reactor or in combination with electrical heating. The exact reaction regime requires the use of different sensors. Besides the classical measurement of temperature, pressure, and stirring speed, different other parameters can be captured with the control software. The coupling of online-Fourier-transform infrared spectroscopy (FTIR) enables direct monitoring of the reaction. The control software enables a direct intervention into active experiments to shorten reaction steps, change temperatures, or stop dosage processes. The integration of special measuring and sensor systems is also possible in short times due to the modular principle of the control software. 2.4.2.2 Microplate-Based Systems

To enable the use of classical equipment known from biological screening automation (e.g., pipetting systems), many reaction blocks are constructed in the microplate format containing 96 or 384 wells. Especially in combinatorial chemistry, it is possible to perform the reactions in standard microplates or filter plates. If filter plates are used, simultaneous collection of the resulting products can be achieved in combination with suitable collection plates [101]. Many available reaction blocks designed for combinatorial synthesis use the microplate format. They can contain 48 or 96 reaction wells to ensure the pipetting steps and the collection of the resulting components in standard microplates. The reaction blocks are made of chemically inert materials and can be solid blocks or can contain single exchangeable reaction vessels. Mixing is usually realized by placing the reaction blocks on an orbital shaker. Also, heat/cooling shakers or ovens are available for performing reactions at different temperatures. Many vendors provide reaction blocks for primary use in solution phase synthesis. Systems for parallel processing of up to 96 or 384 wells at high temperatures and high pressures have been described [25, 108]. These systems have a pressure tank made of a chemically inert nickel-based stainless steel alloy. The lower part of the tank houses the reaction module. For pressurization, two gas supplies are integrated. On the upper part of the pressure tank (the lid), the actuators for opening and closing of the reaction module are mounted. The system can be operated with up to 50 bar pressure. The homogenization of the reaction compounds is realized through a magnetic stirring system [109]. A tempering system based on thermoelectrical coolers is used for tempering the pressure tank. Glass plates with 96 or 384 wells are used as reaction modules. The system was used for the screening of the enantioselective hydrogenation of methyl-2-acetamido acrylate or the screening of a laccase catalyzed coupling reactions [110]. 2.4.2.3 Robot-Based Synthesis Systems

Another approach uses classical laboratory robot-based systems for performing automated synthesis procedures. An early example was a robotic platform, ARCoSyn, based on a central robot with high accuracy and weight capacity. The

2.4 Automation in Chemical Sciences

system was used for fully automated synthesis and purification of compound arrays. The used industrial robot was equipped with a gripper changing system [104]. A fully automated system for synthesis optimization was described, which was applied for yield and purity optimization of a multicomponent reaction for the synthesis of intermediates of lycoricidine analogs [111]. A laboratory robot mounted on a linear track was used to integrate the different system components. This includes a storage rack for 2 ml reaction vials, a modified reaction system for heating, cooling, and shaking of the reaction mixtures, a crimper/decrimper station, a centrifuge, filtration units, and liquid delivery stations. A many-to-many solid dispensing system is included for handling very small amounts of solids. To ensure a fully automated process including automated analysis, analytical measurement systems such as a GC/MS system and an HPLC are integrated into the system as well. Complete reaction procedures including sample preparation, reaction, preanalytical sample treatment, and analysis can be programmed [112, 113]. 2.4.3 Application Examples

Automated chemistry has many applications [104]. The classical application is the automated combinatorial synthesis of peptides and proteins. The history of automated peptide synthesis can be dated back to 1965, when Merrifield and Steward introduced their approach [114]. A simple and general methodology for solid phase synthesis of peptide 𝛼-thioesters for the convergent synthesis of proteins via native chemical ligation (NCL) was reported. The synthesis of this new compound can be fully automated using inexpensive commercially available materials and does not require any post-synthetic steps prior to NCL [115]. While the solid phase synthesis of DNA and peptides has become routine for decades, access to glycans has been technically difficult, time consuming, and confined to a few expert laboratories. Based on the central glycosidic bond forming reaction, a general concept for the protecting groups and leaving groups has been developed, which can automatically produce high numbers of diverse glycans [116]. The automated synthesis of several oligosaccharides on a solid phase synthesizer was first described in 2001 [117]. Today, methods are available for the solid-phase syntheses of oligosaccharides containing sialic acids [118], arabinoxylan-oligosaccharides [119], dermatan sulfate oligosaccharides [120], and chondroitin sulfate oligosaccharides [121] or the synthesis of carbohydrates and functionalization of polyanhydride nanoparticles [122]. The generation of small-molecule libraries is necessary for drug discovery processes. High-throughput synthesis is a major source for compound libraries utilized in academia and industry, seeking new chemical modulators of pharmacological targets. An overview about crucial factors of library design strategies from the perspective of synthetic chemistry and a summary of current approaches and case studies can be found in [123]. The design of organic

49

50

2 Automation in Life Sciences – A Critical Review

libraries containing 2-arylquinolines, benzamides/urea phenols, pyrrolidines, or β-lactams has been reported [124]. High-throughput experimentation can also be used in catalysis research. The discovery of new catalysts is of essential importance, since a variety of processes still exist in the chemical industry, for which no useful catalysts are available at all. As in drug discovery, it must also be ensured in the field of catalyst development that the libraries are sufficiently diverse in order to scan a broad range of parameters. The use of combined catalysis and optical screening for the high throughput discovery of solar fuel catalysts was also reported [125]. A total of 21 amino acid based ligands was combined with [Ru(p-cymene)Cl2 ]2 and [RhCp*Cl2 ]2 , and used as catalysts for the asymmetric transfer hydrogenation of four different heteroaromatic ketones in 2-propanol. The reactions were performed on an automated high-throughput screening robotic platform [126]. A screening system, which was originally designed for homogeneous enzymatic reactions, could be used without further modifications for continuous catalysis with polymer-bound chemical catalysts or for quasi-homogeneous systems like reverse micelles [127]. Biocatalysis became a widely used method in the synthesis of new molecules and libraries. A majority of the used enzymes are commercially available and inexpensive; these enzymes can perform a wide range of chemical transformations under very mild conditions. Enzymatic reactions are further characterized by the absence of reaction byproducts, their broad specificity in terms of substrates, and their complete region and stereo selectivity. The biotransformation strategy for the synthesis of solution phase libraries of BOD products (bicyclo[2.2.2]oct-5-ene-2,3-trans-dimethanol) was reported. Other libraries were reported using adenosine, 2,3-(methylene dioxy)benzaldehyde or taxol as substrates [128]. A procedure has further been described for the automated screening and lead optimization of a supramolecular-ligand library for the rhodium-catalyzed asymmetric hydrogenation of five challenging substrates relevant to the industry. The automated optimization of the resulting two leads showed that an increase of catalyst loading, dihydrogen pressure, and temperature had a positive effect on catalyst activity without affecting catalyst selectivity [129]. Since heterogeneous catalysts have a profound impact on the chemical industry, high-throughput combinatorial methods for heterogeneous catalyst design and development are of great interest. A key problem of synthesis and screening of heterogeneous catalysts is the reproducibility of their catalytic properties. Thus, the adjustment of the reaction conditions such as pH, liquid volume, temperature, pressure, and so on, is important. The automated synthesis of mesoporous silica sieves (MSS) as well as testing of the catalytic properties was reported [130]. Even biological libraries can be established. These libraries consist of pools of microorganisms expressing different polypeptides on the surface. Such libraries are available for phages, bacteria, or yeasts. They are used for affinity ligand identification, pharmaceutical application, or the display of cDNA, genomic DNA, proteins, and antibodies [131].

2.5 Automation in Analytical Measurement Applications

2.5 Automation in Analytical Measurement Applications 2.5.1 Overview

Analytical measurement studies and uses instruments and methods that are used to separate, identify, and quantify matter. Qualitative analysis identifies analytes, while quantitative analysis determines the numerical amount or concentration. Instrumental methods may be used to separate samples using chromatography, electrophoresis, or field flow fractionation. Then, qualitative and quantitative analyses can be performed, often with the same instrument and may use light interaction, heat interaction, electric fields, or magnetic fields. The processes of separation, identification, and quantification can be combined in one instrument. Analytical measurement has broad applications in forensics, medicine, science, and engineering [132]. It can be distinguished as elemental analysis and structural analysis. Atomic or emission spectroscopy as well as X-ray fluorescence is a classical method for the determination of the elemental composition of samples. UV/vis and IR-spectroscopy instead belong to the structural methods giving information regarding the structural coupling of functional groups in molecules. Mass spectrometry is a combined method providing structural and elemental information regarding samples. The goal of analytical measurement is the unambiguous qualitative and quantitative determination of compounds in pure solutions, complex mixtures, and complicated matrices. To ensure the required selectivity of elements and compounds, the concept of pre-, intra-, or post-sensoric selectivity can be used (see Figure 2.2) [133]. Main part is the intra-sensoric selectivity, which can be

Pre-sensoric selectivity

Intra-sensoric selectivity

Measurement object in complex matrix

Single sensor or complex sensor system

• Separation of the measurement object from the complex matrix

• Sensors with suitable selectivity related to the measurand

• Application of mechanical, thermal, chemical, electrophoretic, or chromatographic techniques

• Chemical, physical, or biological single sensors

• Selectivity before the sensor

• Complex sensor systems • Selectivity inside the sensor

Post-sensoric selectivity

High-level information • Interpretation of measurement values • Qualitative evaluation • Quantitative evaluation • Selectivity after the sensor Low-level information • Data visualization

Figure 2.2 Concept of pre-, intra-, and post-sensoric selectivity.

51

52

2 Automation in Life Sciences – A Critical Review

achieved using suitable physical properties and phenomena. This includes single chemical, physical, or biological sensors or complex sensor systems providing a suitable selectivity with respect to the compounds to be measured. Multidimensional information directly correlating to the concentration and the structural composition of the compounds to be analyzed is generated in this step. To increase the selectivity an additional step can be integrated, which realizes the separation of the compounds of interest from the complex matrix (pre-sensoric selectivity). Mechanical, electromagnetic, thermal, or physico-chemical methods can be used for this separation step (see Figure 2.2). The measuring data resulting from the pre- and intra-sensoric selectivity can be interpreted, evaluated, and visualized in a third step, the post-sensoric selectivity. Besides a manual, partly or fully automated commercial or application-specific knowledge-based software systems can be used [133]. A degree of automation as known from automated bioscreening is still not available in analytical measurement technology. General ideas and concepts can be dated back to the 1960s. An automated system for the detection of the influence of contaminants on the determination of creatinine was described in 1965 [134]. Small amounts of liquids have been dispensed in microscale; in addition, a filtration process and photometric measurements were included. Kienitz and Kaiser stated that automation in chemistry requires an automation of the measurement technology [135]. While they focused mainly on the automation of the measuring process, the processes connected to the analytical measurement must be integrated as well [136]. Thus, the devices (automated subprocesses) do not fulfill the purpose of automation, as long as they are limited to sampling and analysis [136]. In the 1960s, Malissa introduced and extended a symbolism for the description of different processes from valid DIN standards of chemical technology and electrical engineering to automated processes in chemical and analytical laboratories [137–140]. Starting from 1980, automation in the field of analytical measurement has been an increasingly expanding business [141]. 2.5.2 Process Analytical Technology

The possible and necessary combination of automation and chemical analysis in industrial processes was already described around 1966 [135]. Today, this is called Process Analytical Technology (PAT) [142]. PAT was described as a combination of methods and techniques for process tracking in chemical, biotechnological, pharmaceutical, or physical transformations. This includes also modeling and simulation, knowledge of compound and process data, data evaluation, and process control [143]. The PAT methods are of increasing importance in process industries and pharmaceutical industries regarding regulatory (proof of process quality) and economic requirements (optimization of production processes due to increasing energy costs and environmental requirements) [143]. Sensors for the determination of pH, redox potentials, and conductivity values followed by gas sensors and photometers are mostly used in the field of chemical industry [144]. Optical technologies, such as spectroscopy (fluorescence, Raman, transmission, reflection), particle measuring technologies, and laser-based methods as well as chromatographic technologies (e.g., process

2.5 Automation in Analytical Measurement Applications

chromatography) also belong to the typical analytical procedures of PAT [145]. Highly automated systems can be found in PAT. The actual focus here is on customer-specific solutions so that individual products with a specific profile of properties can be manufactured [143]. One of the first commercially available analytical devices for online operating procedures was the so-called Autolysator from 1912 [146–148]. This device was used for the determination of carbon dioxide in smoke and combustion gases, which enabled the analysis and optimization of combustion reactions. Today, different operation procedures can be differentiated in PATs. If the sample is taken automatically and the measurement and data evaluation are executed close to the process, an online or inline operation can be defined depending on the sampling (a sample is taken in online mode, no sample is taken in inline mode). In the case of offline or atline mode, sampling is done manually and the measurement is not performed close to the process [143]. To derive product-specific information from a procedural process, suitable process analytical devices as well as an adaptation of the processes are required. Therefore, different functional groups are necessary, which together form the process analytical measuring system [142]. It is not necessarily required that each functional group must be available and can be dropped especially for inline analytical devices (e.g., sampling, sample preparation, sample disposal) [142, 149]. Optical methods such as fluorescence measuring technology, Raman spectroscopy, particle measuring technology, or transmission and reflection spectroscopy are used as analyzers in PAT (see Table 2.7). Besides the numerous advantages of these classical optical measuring techniques (e.g., contactless and nondestructive measuring, high measuring speed), their use is limited in some areas of process measuring technology [144]. Reasons include the often difficult access to the analytes (e.g., windows, cuvettes, etc.), a possible contamination of the optical access, and a safety risk in case of defects of the optical access. Some classical methods can only be used in atline or offline mode due to extreme process conditions (e.g., temperature, pressure, vibrations) [145]. Laser technology developed increasingly in the field of online and inline PAT to overcome the restrictions of common optical methods [145]. Quantum cascade laser and super continuum radiators provide new application possibilities. Other laser light sources, which can be used in the field of process analytical technologies, include surface emitting lasers, compact optical parametric Table 2.7 Typical optical measuring methods in process analytical technology and their applications [145]. Measuring method

Application example

Fluorescence spectroscopy

Quality control in surface cleaning

Raman spectroscopy

Determination of the composition of solid materials

Particle measuring technology

Determination of particle size distribution

Transmission and reflexion spectroscopy

Determination of the composition of gas mixtures, liquids and solids

53

54

2 Automation in Life Sciences – A Critical Review

oscillators, and laser-based plasma light sources [145]. Proprietary automation solutions are used for many applications in PAT to ensure consistent quality. One example is the automated method for the investigation of magnetically soft alloys (e.g., FeSiB) using optical emission spectroscopy and flow injection analysis [150]. Here, microwave digestion of the solids is completed using an automated microwave digestion system Microdigest A 301 (Prolabo). Due to the automated sample preparation, sample delivery into the analytical measuring system, and simultaneous determination of elements, the method enables time and material savings compared to classical chemical analytical methods (e.g., gravimetric analysis, titration, differential spectrophotometry) [150]. On summarizing it can be mentioned that PAT is characterized by a high degree of automation. This is caused by the requirements of industrial production for economic processes and the observance of the branch-specific quality criteria. The automation systems have been developed for specific applications and provide only little flexibility regarding any changes in the processes. In addition, the used measuring principles have only a limited selectivity. 2.5.3 Automation Systems for Analytical Measurement Applications

In the field of chemical analytics outside process analytical applications in the industry, only a few automation solutions exist, which are additionally characterized by massive specialization. Automation mainly applies to the direct transfer of samples into the measuring systems. Auto samplers close an important gap in analytical measurement. They automate sample introduction into the measuring systems and, depending on their equipment, also some parts of sample preparation. Dilution, mixing, and derivatization can be realized with auto samplers without manual intervention [151]. In some cases, a sophisticated auto sampler can be used, which also enables easy operations, such as automated desorption, heating, shaking, or SPE [152]. An example for automated measuring methods is the determination of mercury in environmental applications. In the last years, special fully automated analytical devices have been developed on the basis of atomic absorption (AAS) and atomic fluorescence spectroscopy (AFS) for direct analytical measurements and mercury cold vapor analysis. These devices enable an automatic introduction of solid samples into the system. They are used for the investigation of soil and plant samples near metallurgical factories [153]. Other available systems enable direct as well as cold vapor analytical measurements and use two detection methods (AAS and AFS) [154]. These mercury analyzers are characterized by a high sensitivity with corresponding low limits of detection. Another advantage is the handling of solid and liquid samples. Unfortunately, the measurement is limited to the detection of mercury or a few elements, which have to be detected sequentially. With respect to flexible automation, ICP-OES and ICP-MS are suitable methods for parallel detection of elements, which enable (except for some elements) the detection of nearly all elements of the periodic system. These systems require the use of liquid samples. Therefore, suitable sample preparation is required for the investigation of solid samples. Automated sample

2.5 Automation in Analytical Measurement Applications

preparation systems have been introduced, but they mainly present single, separate work stations, which, for example, do not include any digestions steps [155]. Another example in instrumental analytics is a system for the automated addition of internal standards for the following analysis using ICP-OES [156]. A relatively simple connection system using flow injection to enable the automated addition of internal standards to sample solutions during measurement was developed. Although the automation solution is in this case limited to the sample delivery, it results in a significant increase in performance by reducing manual efforts and increasing precision compared to the manual addition of the internal standard to the different samples. This sample delivery principle is currently used in numerous ICP-MS and ICP-OES devices. One example for automated high-throughput sample preparation in structural analysis was described for the global quantification of the biochemical condition of organisms. Indol-3-acetic acid and additional auxins (naturally occurring phytohormones in plants) are quantitatively analyzed in plant tissue of Arabidopsis thaliana [157]. After the homogenizations, the samples are purified with a positive pressure SPE using an automated liquid handler (ALH). The samples are evaporated in a complex process with simultaneous production of diazomethane, sample derivatization (methylation), and concentration. The evaporated samples are dissolved in ethyl acetate, transferred to the GC vials, and closed. This process requires a processing time of approximately 25 min. Despite the GC-SIM-MS measurement, 96 samples can be simultaneously processed, which enables a high sample throughput. Besides bioanalytical methods, physico-chemical methods are also used in the field of drug development, drug discovery, and pharmaceutical industry. A fully automated robot-based system for the sample preparation and analyses has been introduced in 2008 [158]. This system was developed to reduce costs and processing time for routine measurements in quality control of active compounds and pharmaceutical end-products. The system includes two units, the basis and the HPLC unit. Four different elements are used for handling (transport and manipulation): a multifunctional handling arm, an extension arm, a robotic manipulator arm, and a shuttle device for the sample transport between basis and HPLC unit. The work area is equipped with different stations and laboratory devices. This includes racks for samples, reagents, and solvents, stations for the lid handling (screwing, crimping), homogenization (shaker, magnetic stirrer, tumbler), pH measurement, labeling, tempering, dispensing, and dosage. The analytical measurement is realized using HPLC. The automation system is controlled by two software solutions. The first software was implemented in VisualBasic.NET 2003 and controls the different devices in the basic unit. The second software controls the auto sampler of the HPLC system. While the before described automation systems are used for structural analytics, the following automation system is an example from the field of elemental analytics. An automated system for the sample preparation of educts and intermediate products followed by a determination of the concentration of heavy metals (Pd, Pt, Pb, Hg, Bi, Sb, As, Sn, Cd, Ag, Cu, Mo) using ICP-AES was described in 2011 [159]. The samples are dissolved in a mixture of hydrochloric acid (HCl) and dimethyl sulfoxide (DMSO) followed by an ultrasonic treatment.

55

56

2 Automation in Life Sciences – A Critical Review

The automation system includes an acid-resistant work station with an ultrasonic probe. An ICP-AES system equipped with an auto sampler for automated sample delivery was integrated as well. Experiment design and data evaluation are realized using a specific software. The system is an example for partial automation; sample preparation and sample delivery into the ICP-AES are automated. The sample transport between the work station and AES auto sampler and data evaluation are executed manually. Another design for an automation system was described for an automated HPLC method for the detection of mycophenolic acid and its glucuronide conjugate in human plasma [160]. A laboratory robot with cylindrical kinematics is used as the central transportation element. The different stations for the sample preparation are positioned concentrically around the robot. Besides stations for weighing, diluting, dispensing, pipetting, and sample storage, the system also includes two high-performance SPE units, two online HPLC systems with optical detectors, and a PC for the control software [160].

2.6 Requirements for Automating Analytical Processes The above described applications demonstrate the increasing demand for flexible and universally applicable automation solutions in the field of physico-chemical analytics for the detection of elemental composition as well as the generation of structural relevant information. Currently, this area is dominated by automated work stations, partially automated systems, or proprietary fully automated systems for specific applications. In comparison to current bioscreening systems, strong deficits still exist in the development of flexible applicable automation solutions for analytical measurement. In contrast to process analytics in industrial applications [135, 143], automation of clinical or environmental laboratories is still limited. This is due to the significant differences between automated and manual processes [161]. Even big international laboratories prefer manual procedures, since automated systems compared to manual methods often require relatively expensive consumables and a higher amount of solvents [161]. Additional reasons in favor of manual procedures include maintenance susceptibility, required laboratory space as well as special equipment such as laboratory hoods. Mainly, subprocesses are automated in these fields; thus there exists a high demand for flexible automation solutions with a preferably high degree of automation [161]. 2.6.1 Bioscreening vs. Analytical Measurement 2.6.1.1 Vessels and Vials in Analytical Processes

In contrast to bioscreening where the samples are usually provided on microplates with different numbers of cavities (wells), storage, sample preparation, and measurement in the field of analytical measurement are executed in a variety of vials, vessels, and tubes with different formats and volumes. A main characteristic of chemical analytical processes is the different volumes of the sample vials used. Until now no standardized sample vessels exist in contrast

2.6 Requirements for Automating Analytical Processes

to the establishment of 96, 384, or 1536 well plates in the field of biological screening. Table 2.8 summarizes typical vessels used in analytical measurement. Standards have not been defined either in instrumental chemical analytics. Different analytical systems use different sample volumes with volumes of 100 μl to 50 ml depending on the technology used. Modern auto samplers in liquid chromatography enable the sample introduction via vials or microplates. But not all physico-chemical measuring techniques allow the use of microplates. In contrast,

Table 2.8 Examples for sample vessels in analytical measurement. Sample container

Analytical method

Microwave vessels

• • • •

Bottles

• ICP-MS • ICP-OES • F-AAS

• 125 ml • Sample container • Container for rinsing solution

Falcon tubes

• ICP-MS • ICP-OES • F-AAS

• 8–50 ml • Sample container • Container for rinsing solution

AAS vials

• GF-AAS

• 2.000 μl • Sample Container

GC vials

• LC-MS • GC-MS

• Volume: 2000 μl, 100/500 μl with micro vial insert • Reaction container, sample container • Container for rinsing solution

96-well MTP

• LC-MS • Bioscreening

• Volume: 200–1.000 μl • Reaction container • Sample container

384-well MTP

• LC-MS • Bioscreening

• Volume: 30–100 μl • Reaction container • Sample container

ICP-MS ICP-OES F-AAS GF-AAS

Maximum volume/operating area

• 10, 25, 75, ml • Microwave-assisted acid digestion

Examples

57

58

2 Automation in Life Sciences – A Critical Review

samples for gas chromatography, ICP-MS, ICP-OES, and AAS are provided in single vials with different volumes. The necessity for higher volumes results from the higher inhomogeneity of the samples compared to biological applications; thus, processing of higher sample amounts is required for ensuring representative results. Miniaturization is a central topic in analytical measurement, especially in the field of sample preparation, to reduce costs and time. Thus, an automated clean-up system for the preparation of environmental samples for the subsequent determination of different dioxins using GC-MS was developed [162]. The required sample volumes could be reduced by 95% from 550 to 30 ml. In addition, the required processing time could also be reduced significantly by 75% from 8 to 2 h. 2.6.1.2 Liquids and Reagents in Analytical Measurement

While preferably aqueous media are used in bioscreening, different media must be used in compound oriented measuring technology. This includes organic and inorganic solvents, which differ with respect to their viscosity, surface tension, and vapor pressure. Thus, they have to be handled differently in pipetting processes. In addition, chemical aggressive media (e.g., acidic and alkaline solutions, highly inflammable compounds) are used, which require special safety facilities to avoid damage to the automation equipment. The environmental conditions of, for example, derivatizations of samples (high temperatures and pressures) differ significantly from the conditions in bioscreening, where mild temperatures up to 37 ∘ C are mainly used (e.g., incubation). In addition, the buffers usually handled in biological application can cause damage to analytical measurement systems and therefore have to be avoided or separated from the samples prior to the introduction into the measuring system. 2.6.1.3 Process Structure

The processes in bioscreening are relatively unique and simply structured. They mainly include liquid delivery processes, incubation steps, and readout with simple optical methods. In contrast, the processes in analytical measurement are very different depending on the specific application and often have a very complex structure, since especially in sample preparation a high number of subprocesses (derivatization, acid digestion, extraction, evaporation, etc.) have to be integrated. In addition, sampling and sample delivery are more complex and sophisticated, especially when using single vials. Figure 2.3 shows a comparison of the general process workflows for biological processes and analytical measurements. 2.6.2 Automation Requirements

The general goals for automated systems in analytical measurements can be derived from the previous chapters. The main point is the increase in sample throughput, which can be achieved either with a decrease in processing times in sample preparation and sample transport or with a reduction in measuring

2.6 Requirements for Automating Analytical Processes Start

Start

Pipetting

Pipetting, dosing, re-formatting

Incubation, cell disruption, concentration, dilution, etc.

Derivatization, microwave digestion, filtration, solid phase extraction, concentration, dilution, etc.

Pipetting

Sample introduction (samples on microplates)

Bioscreening Plate reader

Pipetting, dosing, re-formatting

Sampling, sample introduction (samples in single vessels)

Sampling, sample introduction (samples on microplates)

Analytical measurement Instrumental analysis

Analytical measurement Instrumental analysis

Data evaluation Data evaluation

End End

Figure 2.3 General process workflow in bioscreening and analytical measurement.

times and required times for data evaluation. On the other hand, 24/7 operation of automated processes also contributes significantly to the increase in sample throughput. Manual effort should be reduced to a minimum; thus, automation of these subprocesses is necessary. The reduction of errors produced by the operating personnel, including operator errors, unintentional contamination of samples, or variations in the precision, in the case of personnel changes, is another positive effect of automation. Errors and downtimes can occur in automated systems due to technical problems. This must already be considered during the design and development stages to ensure highest possible robustness against external influences (e.g., robustness against aggressive chemicals) and maximum reliability. Another aspect is the reduction in sample volume and material consumption (solvents, reagents, preservative, additives, etc.) and the resulting costs, which can be achieved with a miniaturization of the processes [18, 163]. Furthermore, the use of flexible measuring methods after a sample preparation procedure can contribute to a reduction in material consumption. The same applies to the preferred use of multi element analysis (e.g., ICP-MS and ICP-OES) compared to single element analysis (e.g., F-AAS and GF-AAS). The linkage of automated

59

60

2 Automation in Life Sciences – A Critical Review

processing and evaluation of the measurement data is an additional important task in order to achieve the goal of complete automation. This includes the application of identical data formats, suitable sample logistics as well as efficient process and data management systems. Besides the economic goals, environmentally relevant aspects gain importance as well. The idea of Green Analytical Chemistry (GAC) has its origins in Sustainable Development [164]. The first activities for sustainability in chemistry were focused on industrial methods and products, which can also be defined as Green Chemistry [165]. The beginnings of Green Chemistry were dominated by the Green Organic Synthesis in different branches of the chemical industry, especially the pharmaceutical industry. In 2000, the sector of GAC has been developed with the main tasks of reduction the use of chemical compounds, minimization of energy consumption, and proper power waste management. The acronym SIGNIFICANCE summarizes the 12 principles of GAC [164], which are also valid for modern automation systems. Based on these principles (see Table 2.9), economic and environmentally compatible automation systems can be developed. The challenge for the development of automated systems in the field of life sciences is the integration of different hard and software components (heterogeneous environment with normally proprietary interfaces). The subprocesses to be realized such as transport, manipulation, analytical measurement, data evaluation, and sample storage depend on their type and the number from the relevant application. In general, these tasks include pipetting of liquids, dosage of solid material, and reformatting into different sample containers. This requires often the application and removal of lids with screwing, crimping, or pressing. Derivatization, incubation, or decomposition is often necessary. Additional technologies, such as filtration, liquid/liquid or liquid/solid extraction, SPE, ultrasonic treatment or microwave treatment, the avoidance of static electricity charge with ion beams, or solvent evaporation can also be used as options. Suitable sampling and Table 2.9 SIGNIFICANCE – the 12 principles of green analytical chemistry [164]. S

Select direct analytical technique

I

Integrate analytical processes and operations

G

Generate as little waste as possible and treat it properly

N

Never waste energy

I

Implement automation and miniaturization of methods

F

Favor reagents obtained from renewable sources

I

Increase safety for operator

C

Carry out in situ measurements

A

Avoid derivatization

N

Note that the sample number/size should be minimal

C

Choose multi-analyte or multi-parameter method

E

Eliminate or replace toxic reagents

References

Hierarchical workflow management

Process control

Pipetting Dosing Dilution Concentration Re-formatting Opening Closing

Derivatization Filtration Extraction Centrifugation Microwave digestion Solid phase extraction

Pipetting Dosing Dilution Concentration Re-formatting Opening Closing

Sampling Sample introduction

Material flow

Instrumental analysis

Data evaluation

Information flow

Figure 2.4 General subprocesses in analytical measurement.

sample delivery, analysis, and data evaluation are required as well in all analytical processes. A connection of these subprocesses requires manual or automated material transport. Figure 2.4 demonstrates the general subprocesses in analytical measurement.

References 1 Hamilton, S.D. (2007) 2006 ALA survey on industrial laboratory automation.

J. Assoc. Lab. Autom., 12 (4), 239–246. 2 Frost & Sullivan (2008) European Markets for Laboratory Automation Sys-

tems: Market Study, Frost & Sullivan, London. 3 Kalorama Information (2008) The Worldwide Market for lab Automation:

Market Study, Kalorama Information, New York. 4 Lago, P., Vallone, I., and Fenili, A.M. (2014) Total laboratory automation and

clinical engineering. IFMBE Proc., 41, 1083–1087. 5 Ackermann, U., Plougastel, L., Wichmann, C., Goh, Y.W., Yeoh, S.D.,

Poniger, S.S. et al. (2014) Fully automated synthesis and coupling of [18 F]FBEM to glutathione using the iPHASE FlexLab module. J. Labelled Compd. Radiopharm., 57, 115–120. 6 Capelle, M.A.H. and Arvinte, T. (2008) High-throughput formulation screening of therapeutic proteins. Drug Discovery Today Technol., 5 (2–3), e71–e79. 7 Moore, K.W., Chandler, G., Whalley, P., Gannon, D., and Simpson, P.B. (2006) Efficient sample logistics: from the chemist to the assay plate and beyond. J. Assoc. Lab. Autom., 11 (2), 92–99. 8 Bareither, R., Bargh, N., Oakeshott, R., Watts, K., and Pollard, D. (2013) Automated disposable small scale reactor for high throughput bioprocess development: a proof of concept study. Biotechnol. Bioeng., 110 (12), 3126–3138.

61

62

2 Automation in Life Sciences – A Critical Review

9 Krewski, D., Acosta, D., Andersen, M., Anderson, H., Bailar, J.C.,

10

11

12 13 14

15

16 17 18 19 20 21

22 23 24

25

26

Boekelheide, K. et al. (2010) Toxicity testing in the 21st century: a vision and a strategy. J. Toxicol. Environ. Health Part B, 13 (2–4), 51–138. Katz, A.H. (2009) Lab automation protocols and virtual workcells flexible screening system can be created with the varyscreen using softlinx software. Genet. Eng. Biotechnol. News, 29 (9), 40–41. Najmabadi, P., Goldenberg, A.A., and Emili, A. (2006) Hardware flexibility of laboratory automation systems: analysis and new flexible automation architectures. J. Assoc. Lab. Autom., 11 (4), 203–216. Felder, R.A. (1998) Modular workcells: modern methods for laboratory automation. Clin. Chim. Acta, 278 (2), 257–267. Hurst, W.J. and Mortimer, J.W. (1987) Laboratory Robotics: A Guide to Planning, Programming, and Applications, VCH Publishers, New York. LabAutopedia – SLAS Knowledge Network Society for Laboratory Automation and Screening, http://www.labautopedia.org/ (accessed 01 December 2016). Michael, S., Auld, D., Klumpp, C., Jadhav, A., Zheng, W., Thorne, N. et al. (2008) A robotic platform for quantitative high-throughput screening. Assay Drug Dev. Technol., 6 (5), 637–657. Valid, J.-M. (2010) Progressive automation – the solution of choice for improving lab efficiency. J. Med. Biochem., 29 (4), 325–329. Ray, D. (2011) LEAN labs use work cells, auto data capture for more efficient workflow. Med. Lab. Obs., 43 (7), 24. Burbaum, J.J. (1998) Miniaturization technologies in HTS: how fast, how small, how soon? Drug Discovery Today, 3 (7), 313–322. Hüser, J. (2006) High-Throughput Screening in Drug Discovery, Wiley-VCH Verlag GmbH, Weinheim. Kyranos, J.N. (2004) High Throughput Analysis for Early Drug Discovery, vol. XV, Elsevier Academic Press, San Diego, CA, 186 p. Situ, C., Mooney, M.H., Elliott, C.T., and Buijs, J. (2010) Advances in surface plasmon resonance biosensor technology towards high-throughput, food-safety analysis. TrAC, Trends Anal. Chem., 29 (11), 1305–1315. Bogue, R. (2012) Robots in the laboratory: a review of applications. Ind. Robot, 39 (2), 113–119. Thurow, K., Decker, S., and Allwardt, A. (2005) Automatisierte Reaktionssysteme. Nachr. Chem., 53 (10), 1046–1048. Allwardt, A., Stoll, N., Wendler, C., and Thurow, K. (2004) New developments in reaction engineering – the HPMR 50–96 multiparallel reactor. Chem. Ing. Tech., 76 (11), 1679–1682. Allwardt, A., Stoll, N., Wendler, C., and Thurow, K. (2005) New developments in the field of reaction technology: the multiparallel reactor HPMR 50–96. J. Autom. Methods Manage. Chem., 2005, 26–30. Horowitz, G.L., Zaman, Z., Blanckaert, N.J.C., Chan, D.W., DuBois, J.A., Golaz, O. et al. (2005) Modular analytics: a new approach to automation in the clinical laboratory. J. Autom. Methods Manage. Chem., 2005, 8–25.

References

27 Thurow, K. and Fleischer, H. (2011) Automatisierung für das Klinische

28 29

30 31 32 33

34 35 36 37 38

39

40 41

42

43 44

Labor – Aktuelle Herausforderungen und Tendenzen. GIT Labor Fachz., 55 (7), 466–467. Middleberg, R.A. (2014) Forensic toxicology: meeting the demands of two slaves – science and the law. Am. Lab., 45 (1), 41–43. Usenko, S., Subedi, B., Aguilar, L., and Robinson, E. (2013) High-throughput analysis of PPCPs, PCDD/Fs, and PCBs in biological matrices using GC-MS/MS. Compr. Anal. Chem., 61, 143–158. Otto, M. (2011) Analytische Chemie, 4th edn, Wiley-VCH Verlag GmbH, Weinheim. Hering, E. (2012) Sensoren in Wissenschaft und Technik: Funktionsweise und Einsatzgebiete, Vieweg + Teubner Verlag, Wiesbaden. Gey, M.H. (2008) Instrumentelle Analytik und Bioanalytik: Biosubstanzen, Trennmethoden, Strukturanalytik, Applikationen, Springer, Berlin. Macarron, R., Banks, M.N., Bojanic, D., Burns, D.J., Cirovic, D.A., Garyantes, T. et al. (2011) Impact of high-throughput screening in biomedical research. Nat. Rev. Drug Discovery, 10 (3), 188–195. Swinney, D.C. and Anthony, J. (2011) How were new medicines discovered? Nat. Rev. Drug Discovery, 10 (7), 507–519. Hertzberg, R.P. and Pope, A.J. (2000) High-throughput screening: new technology for the 21st century. Curr. Opin. Chem. Biol., 4 (4), 445–451. Mayr, L.M. and Fuerst, P. (2008) The future of high-throughput screening. J. Biomol. Screening, 13 (6), 443–448. Mayr, L.M. and Bojanic, D. (2009) Novel trends in high-throughput screening. Curr. Opin. Pharmacol., 9 (5), 580–588. Pereira, D.A. and Williams, J.A. (2007) Review: historical perspectives in pharmacology. Origin and evolution of high throughput screening. Br. J. Pharmacol., 152 (1), 53–61. Chen, T. (2010) A Practical Guide to Assay Development and High-Throughput Screening in Drug Discovery, Taylor & Francis, Boca Raton, FL. Sams-Dodd, F. (2005) Target-based drug discovery: is something wrong? Drug Discovery Today, 10 (2), 139–147. Hellerstein, M.K. (2008) A critique of the molecular target-based drug discovery paradigm based on principles of metabolic control: advantages of pathway-based discovery. Metab. Eng., 10 (1), 1–9. Lee, J.A., Uhlik, M.T., Moxham, C.M., Tomandl, D., and Sall, D.J. (2012) Modern phenotypic drug discovery is a viable, neoclassic pharma strategy. J. Med. Chem., 55 (10), 4527–4539. Downey, W. and Hartigan, J. (2013) Adoption of high content screening at HTS laboratories. Drug Discovery World, 14 (2), 51–57. Varma, H., Lo, D.C., and Stockwell, B.R. (2011) High-throughput and high-content screening for huntington’s disease therapeutics, in Neurobiology of Huntington’s Disease: Applications to Drug Discovery, Chapter 5, CRC Press, Boca Raton, FL.

63

64

2 Automation in Life Sciences – A Critical Review

45 Frearson, J.A. and Collie, I.T. (2009) HTS and hit finding in academia – from

46

47

48

49

50

51

52

53

54

55

56

57

chemical genomics to drug discovery. Drug Discovery Today, 14 (23–24), 1150–1158. Fry, D.W., Kraker, A.J., McMichael, A., Ambroso, L.A., Nelson, J.M., Leopold, W.R. et al. (1994) A specific inhibitor of the epidermal growth factor receptor tyrosine kinase. Science, 265 (5175), 1093–1095. Wilhelm, S., Carter, C., Lynch, M., Lowinger, T., Dumas, J., Smith, R.A. et al. (2006) Discovery and development of sorafenib: a multikinase inhibitor for treating cancer. Nat. Rev. Drug Discovery, 5 (10), 835–844. Thaisrivongs, S., Tomich, P.K., Watenpaugh, K.D., Chong, K.-T., Howe, W.J., Yang, C.-P. et al. (1994) Structure-based design of HIV protease inhibitors: 4-Hydroxycoumarins and 4-hydroxy-2-pyrones as non-peptidic inhibitors. J. Med. Chem., 37 (20), 3200–3204. Thornberry, N.A. and Weber, A.E. (2007) Discovery of JANUVIATM (sitagliptin), a selective dipeptidyl peptidase IV inhibitor for the treatment of type 2 diabetes. Curr. Top. Med. Chem., 7 (6), 557–568. Das, J., Chen, P., Norris, D., Padmanabha, R., Lin, J., Moquin, R.V. et al. (2006) 2-Aminothiazole as a novel kinase inhibitor template. Structure-activity relationship studies toward the discovery of N-(2-chloro-6-methylphenyl)-2-[[6-[4-(2-hydroxyethyl)-1-piperazinyl]-2methyl-4-pyrimidinyl]amino]-1, 3-thiazole-5-carboxamide (Dasatinib, BMS-354825) as a potent pan-Src kinase inhibitor. J. Med. Chem., 49 (23), 6819–6832. Riechers, H., Albrecht, H.-P., Amberg, W., Baumann, E., Bernard, H., Bohm, H.J. et al. (1996) Discovery and optimization of a novel class of orally active nonpeptidic endoathelin-a receptor antagonists. J. Med. Chem., 39 (11), 2123–2128. De Corte, B.L. (2005) From 4,5,6,7-tetrahydro-5-methylimidazo[4,5,1-jk](1,4) benzodiazepin-2(1H)- one (TIBO) to etravirine (TMC125): fifteen years of research on non-nucleoside inhibitors of HIV-1 reverse transcriptase. J. Med. Chem., 48 (6), 1689–1696. Yamamura, Y., Ogawa, H., Chihara, T., Kondo, K., Onogawa, T., Nakamura, S. et al. (1991) OPC-21268, an orally effective, nonpeptide vasopressin V1 receptor antagonist. Science, 252 (5005), 572–574. Duffy, K.J., Darcy, M.G., Delorme, E., Dillon, S.B., Eppley, D.F., Erickson-Miller, C. et al. (2001) Hydrazinonaphthalene and azonaphthalene thrombopoietin mimics are nonpeptidyl promoters of megakaryocytopoiesis. J. Med. Chem., 44 (22), 3730–3745. ANSI/SLAS 1-2004. (2011) Microplates – Footprint Dimensions, American National Standards Institute Society for Laboratory Automation and Screening. ANSI/SLAS 2-2004. (2011) Microplates – Height Dimensions, American National Standards Institute Society for Laboratory Automation and Screening. ANSI/SLAS 3-2004. (2011) Microplates – Bottom Outside Flange Dimensions, American National Standards Institute Society for Laboratory Automation and Screening.

References

58 ANSI/SLAS 4-2004. (2011) Microplates – Well Positions, American National

Standards Institute Society for Laboratory Automation and Screening. 59 ANSI/SLAS 6. (2011) Microplates – Well Bottom Elevation, American

60 61 62 63

64 65 66

67

68 69

70 71

72

73

74

75

National Standards Institute Society for Laboratory Automation and Screening. Smith, A. (2002) Screening for drug discovery: the leading question. Nature, 418 (6896), 453–459. Chapman, T. (2003) Lab automation and robotics: automation on the move. Nature, 421 (6923), 661, 663, 665–661, 663, 666. ISO 8373. (2012) Robots and Robotic Devices – Vocabulary, International Organization for Standardization. Gevatter H-J, Grünhaupt U. Handbuch der Mess- und Automatisierungstechnik in der Produktion, 2., vollst. bearb. Aufl ed., Berlin [u.a.]: Springer; 2006. vol. XIX, 854 S. p. DIN EN ISO 8373. (1996) Deutsches Institut für Normung e.V. Industrieroboter – Wörterbuch, Beuth Verlag GmbH, Berlin. DIN EN ISO 8373 Entwurf. (2010) Deutsches Institut für Normung e.V. Roboter und Robotikgeräte – Wörterbuch, Beuth Verlag GmbH, Berlin. Kock, S., Vittor, T., Matthias, B., Jerregard, H., Källman, M., Lundberg, I. et al. (2011) Robot concept for scalable, flexible assembly automation: a technology study on a harmless dual-armed robot. IEEE International Symposium on Assembly and Manufacturing, ISAM 2011, pp. 1-5. Do, H.M., Park, C., Choi, T.Y., and Kyung, J.H. (2013) Design and control of dual-arm robot for cell manufacturing process. IEEE International Conference on Mechatronics and Automation, ICMA 2013, pp. 1419-1423. VDI Nachrichten (2006) Anschlussfertige systeme statt einzelner Komponenten, (19), 27. Cunningham, A., Keddy-Hector, W., Sinha, U., Kruse, D.W.D., Braasch, J., and Wen, J.T. (2014) Jamster: a mobile dual-arm assistive robot with Jamboxx control. Conference Record – IEEE International Conference on Automation Science and Engineering CASE, pp. 509–514. Thurow, K. (2004) Roboter in der Analytik. Nachr. Chem., 52 (5), 618–620. Apricot Designs (2004) Basic Liquid Handling Concepts – Visualized, Apricot Designs, Covina, CA, http://www.apricotdesigns.com/liquid_handling_ concepts.htm (accessed 01 December 2016). Thurow, K., Stoll, N., Ritterbusch, K., and Hitzmann, B. (2011) A fast optical method for the determination of liquid levels in microplates. J. Autom. Methods Manage. Chem., 2011, 10–16. Drexler, D.M., Edinger, K.J., and Mongillo, J.J. (2007) Improvements to the sample manipulation design of a LEAP CTC HTS PAL autosampler used for high-throughput qualitative and quantitative liquid chromatography-mass spectrometry assays. J. Assoc. Lab. Autom., 12 (3), 152–156. Ottleben, I. (2013) Pipettierroboter – Automatisiert Manuell Pipettieren, http://www.laborpraxis.vogel.de/labortechnik/articles/427855/ (accessed 01 December 2016). Gurevitch, D. (2010) Liquid Handling: Theory and Practice, LabAutopedia, http://www.labautopedia.org/ (accessed 01 December 2016).

65

66

2 Automation in Life Sciences – A Critical Review

76 Kong, F., Yuan, L., Zheng, Y.F., and Chen, W. (2012) Automatic liquid han-

77 78 79 80 81 82 83

84 85

86 87 88

89

90

91

92 93

94

dling for life science: a critical review of the current state of the art. J. Lab. Autom., 17 (3), 169–185. Krüger, T., Thurow, K., and Stoll, N. (2009) Optical measurement of minimal volumes in dispensing. Tech. Mess., 76 (1), 8–15. Hesse, S. and Malisa, V. (2010) Taschenbuch Robotik – Montage – Handhabung, Fachbuchverlag Leipzig im Carl-Hanser-Verlag, München. Wells, D.A. (2003) High Throughput Bioanalytical Sample Preparation. Methods and Automation Strategies, Elsevier Science, Amsterdam. Rickwood, D. (1989) Centrifugation: A Practical Approach, 2nd edn, IRL Press, Oxford. Ford, T.C. and Graham, J.M. (1991) An Introduction to Centrifugation, BIOS Scientific Publishers, Oxford. Burke, H.E. (1984) Handbook of Bar Coding Systems, Van Nostrand Reinhold, New York. Ashour, M.-B.A., Gee, S.J., and Hammock, B.D. (1987) Use of a 96-well microplate reader for measuring routine enzyme activities. Anal. Biochem., 166 (2), 353–360. Marktübersicht microplate reader. (2014) Biospectrum, 20 (6), 658–665. Sittampalam, G.S., Coussens, N.P., Nelson, H. et al. (2004) Assay Guidance Manual, Eli Lilly & Company and the National Center for Advancing Translational Sciences, Bethesda, MD, http://www.ncbi.nlm.nih.gov/books/ NBK53196/ (accessed 01 December 2016). Simpson, N.J.K. (2000) Solid-Phase Extraction: Principles, Techniques, and Applications, Dekker, New York. Helle, N., Baden, M., and Petersen, K. (2011) Automated solid phase extraction. Methods Mol. Biol., 747, 93–129. Patel, D.A., Patel, A.C., Nolan, W.C., Huang, G., Romero, A.G., Charlton, N. et al. (2014) High-throughput screening normalized to biological response: application to antiviral drug discovery. J. Biomol. Screening, 19 (1), 119–130. Radu, C., Adrar, H.S., Alamir, A., Hatherley, I., Trinh, T., and Djaballah, H. (2012) Designs and concept reliance of a fully automated high-content screening platform. J. Lab. Autom., 17 (5), 359–369. Shukla, S.J., Huang, R., Austin, C.P., and Xia, M. (2010) The future of toxicity testing: a focus on in vitro methods using a quantitative high-throughput screening platform. Drug Discovery Today, 15 (23–24), 997–1007. Attene-Ramos, M.S., Miller, N., Huang, R., Michael, S., Itkin, M., Kavlock, R.J. et al. (2013) The Tox21 robotic platform for the assessment of environmental chemicals – from vision to reality. Drug Discovery Today, 18 (15–16), 716–723. Walsh, J.C. (1998) High throughput, mechanism-based screening techniques for discovering novel agrochemicals. J. Biomol. Screening, 3 (3), 175–181. van de Bilt, J.L.J., Derks, J.H.J., and Janse, J.D. (2008) Automated, high-throughput immunofluorescence staining: a new approach. Eur. J. Plant Pathol., 120 (1), 91–96. Markin, R.S. and Whalen, S.A. (2000) Laboratory automation: trajectory, technology, and tactics. Clin. Chem., 46 (5), 764–771.

References

95 Carere, A., Stammati, A., and Zucco, F. (2001) In vitro toxicology methods:

96 97

98

99 100 101 102

103

104

105

106

107

108

109

110

111

impact on regulation from technical and scientific advancements. Toxicol. Lett., 127 (1–3), 153–160. Kempner, M.E. and Felder, R.A. (2002) A review of cell culture automation. J. Assoc. Lab. Autom., 7 (2), 56–62. Lehmann, R., Severitt, J.C., Roddelkopf, T., Junginger, S., and Thurow, K. (2015) Biomek cell workstation: a variable system for automated cell cultivation. J. Lab. Autom., 21 (3), 439–450. Lehmann, R., Gallert, C., Roddelkopf, T., Junginger, S., Jonitz-Heincke, A., Wree, A. et al. (2016) Manually and automatically produced pellet cultures of human primary chondrocytes: a comparative analysis. Eng. Life Sci., 16 (3), 272–282. Kates, S.A. and Albericio, F. (2000) Solid-Phase Synthesis. A Practical Guide, Dekker, New York. Jung, G. (2001) Combinatorial Chemistry. Synthesis, Analysis, Screening, Reprint ed,, Wiley-VCH Verlag GmbH, Weinheim. Bannwarth, W. (2000) Combinatorial Chemistry. A Practical Approach, Wiley-VCH Verlag GmbH, Weinheim. DeWitt, S.H., Kiely, J.S., Stankovic, C.J., Schroeder, M.C., Cody, D.M.R., and Pavia, M.R. (1993) “Diversomers”: an approach to nonpeptide, nonoligomeric chemical diversity. Proc. Natl. Acad. Sci. U.S.A., 90 (15), 6909–6913. Ohlmeyer, M.H.J., Swanson, R.N., Dillard, L.W., Reader, J.C., Asouline, G., Kobayashi, R. et al. (1993) Complex synthetic chemical libraries indexed with molecular tags. Proc. Natl. Acad. Sci. U.S.A., 90 (23), 10922–10926. Winter, M., Wiesmüller, K.-H., and Jung, G. (1999) Automated combinatorial chemistry, in Combinatorial Chemistry Synthesis, Analysis, Screening (ed. G. Jung), Wiley-VCH Verlag GmbH, Weinheim, pp. 543–559. Zinsser Analytic GmbH SOPHAS – An Innovative Solution to Automation of Organic Synthesis, http://www.zinsserna.com/files/ab101a_sophas.pdf (accessed 01 December 2016). Winter, M. (2000) Robotik und Automationskonzepte in der Kombinatorischen Chemie – Synthese- und Pipettierroboter. Laborwelt, 2000 (1), 25–29. Stoll, N., Allwardt, A., Dingerdissen, U., and Thurow, K. (2006) An 8-fold parallel reactor system for combinatorial catalysis research. J. Autom. Methods Manage. Chem., Article number 56329, 1–9. Allwardt, A., Holzmüller-Laue, S., Wendler, C., and Stoll, N. (2007) Multiparallel high-pressure reaction in a 384 microplate. J. Assoc. Lab. Autom., 12 (6), 368–373. Cleveland, P.H. and Markle, J.R., inventors (2002) Google Patents, assignee. Magnetic levitation stirring devices and machines for mixing in vessels patent US Patent 6357907 B1. Wendler, C., Allwardt, A., Thurow, K., and Schauer, F. (2005) Screening of a Laccase Catalyzed Coupling Reaction, LabAutomation, San Jose, CA, p. 164. Zhu, J., Wang, Q., and Wang, M.-X. (eds) (2015) Multicomponent Reactions in Organic Synthesis, Wiley-VCH Verlag GmbH, Weinheim.

67

68

2 Automation in Life Sciences – A Critical Review

112 Wendler, C., Krüger-Sundhaus, T., Beller, M., Gördes, D., and Thurow, K.

113

114 115

116 117 118

119

120

121 122

123

124

125

126

(2005) Single-pot solution phase synthesis optimization using fully automated systems for combinatorial screening. J. Assoc. Lab. Autom., 10 (4), 237–241. Von Wangelin, A.J., Neumann, H., Gördes, D., Klaus, S., Jiao, H., Spannenberg, A. et al. (2003) Unusual coupling reactions of aldehydes and alkynes: a novel preparation of substituted phthalic acid derivatives by automated synthesis. Chem. Eur. J., 9 (10), 2273–2281. Merrifield, R.B. and Stewart, J.M. (1965) Automated peptide synthesis. Nature, 207 (4996), 522–523. Terrier, V.P., Adihou, H., Arnould, M., Delmas, A.F., and Aucagne, V. (2016) A straightforward method for automated Fmoc-based synthesis of bio-inspired peptide crypto-thioesters. Chem. Sci., 7 (1), 339–345. Seeberger, P.H. (2015) The logic of automated glycan assembly. Acc. Chem. Res., 48 (5), 1450–1463. Plante, O.J., Palmacci, E.R., and Seeberger, P.H. (2001) Automated solid-phase synthesis of oligosaccharides. Science, 291 (5508), 1523–1527. Lai, C.H., Hahm, H.S., Liang, C.F., and Seeberger, P.H. (2015) Automated solid-phase synthesis of oligosaccharides containing sialic acids. Beilstein J. Org. Chem., 11, 617–621. Schmidt, D., Schuhmacher, F., Geissner, A., Seeberger, P.H., and Pfrengle, F. (2015) Automated synthesis of arabinoxylan-oligosaccharides enables characterization of antibodies that recognize plant cell wall glycans. Chem. Eur. J., 21 (15), 5709–5713. Kandasamy, J., Schuhmacher, F., Hahm, H.S., Klein, J.C., and Seeberger, P.H. (2014) Modular automated solid phase synthesis of dermatan sulfate oligosaccharides. Chem. Commun., 50 (15), 1875–1877. Liang, C.F., Hahm, H.S., and Seeberger, P.H. (2014) Automated synthesis of chondroitin sulfate oligosaccharides. Methods Mol. Biol., 1229, 3–10. Carrillo-Conde, B.R., Roychoudhury, R., Chavez-Santoscoy, A.V., Narasimhan, B., and Pohl, N.L.B. (2012) High-throughput synthesis of carbohydrates and functionalization of polyanhydride nanoparticles. J. Vis. Exp., 65, e3967, doi: 10.3791/3967. Rademacher, C. and Seeberger, P.H. (2016) High-throughput synthesis of diverse compound collections for lead discovery and optimization, in Handbook of Experimental Pharmacology, vol. 232, Springer, New York, pp. 73–89. Felder, E.R. and Berta, D.G. (2005) Solid-phase synthesis of organic libraries, in Combinatorial Chemistry and Technology, 2nd edn (eds G. Fassina and S. Miertus), Taylor & Francis, Boca Raton, FL, pp. 75–88. Gregoire, J.M., Xiang, C., Mitrovic, S., Liu, X., Marcin, M., Cornell, E.W. et al. (2013) Combined catalysis and optical screening for high throughput discovery of solar fuels catalysts. J. Electrochem. Soc., 160 (4), F337–F342. Buitrago, E., Lundberg, H., Andersson, H., Ryberg, P., and Adolfsson, H. (2012) High throughput screening of a catalyst library for the asymmetric transfer hydrogenation of heteroaromatic ketones: formal syntheses of (R)-fluoxetine and (S)-duloxetine. ChemCatChem, 4 (12), 2082–2089.

References

127 Lyagin, E., Drews, A., and Kraume, M. (2015) Fully automated reactor sys-

128

129

130

131

132

133

134 135 136 137

138

139

140 141 142

tem for continuous characterization of (Bio)catalysts. Chem. Eng. Technol., 38 (1), 15–22. Ferritto, R., de Magistirs, E., Missio, A., Paio, A., and Seneci, P. (2005) Solution phase combinatorial libraries of small organic molecules, in Combinatorial Chemistry and Technology, 2nd edn (eds G. Fassina and S. Miertus), Taylor & Francis, Boca Raton, FL, pp. 107–152. Meeuwissen, J., Kuil, M., Van Der Burg, A.M., Sandee, A.J., and Reek, J.N.H. (2009) Application of a supramolecular-ligand library for the automated search for catalysts for the asymmetric hydrogenation of industrially relevant substrates. Chem. Eur. J., 15 (39), 10272–10279. Dominguez, J.M. (2005) High throughput combinatorial methods for heterogeneous catalysts design and development, in Combinatorial Chemistry and Technology, 2nd edn (eds G. Fassina and S. Miertus), Taylor & Francis, Boca Raton, FL, pp. 369–388. Dani, M. (2005) Biological libraries, in Combinatorial Chemistry and Technology, 2nd edn (eds G. Fassina and S. Miertus), Taylor & Francis, Boca Raton, FL, pp. 389–409. Skoog, D.A., West, D.M., Holler, J.F., and Crouch, S.R. (2014) Fundamentals of Analytical Chemistry, 9nd edn, Brooks/Cole, Cengage Learning, Melbourne. Thurow, K. (1999) Ein Methodenspektrum zur selektiven messtechnischen Bestimmung stofflicher Spezies durch spektroskopische Messmethoden am Beispiel ausgewählter Arsenverbindungen. Habilitation thesis. Universität Rostock. Beyermann, K. (1965) Zur Automatisation mikrochemischer Untersuchungen. Fresenius’ Z. Anal. Chem., 210 (1), 1–9. Kienitz, H. and Kaiser, R. (1966) Automation und Analytik in der chemischen Industrie. Fresenius’ Z. Anal. Chem., 222 (2), 119–127. Malissa, H. (1966) Automation in der analytischen Chemie und mit der analytischen Chemie. Fresenius’ Z. Anal. Chem., 222 (2), 100–119. Malissa, H. and Jellinek, G. (1968) Automation in der analytischen Chemie und mit der analytischen Chemie. II. Fresenius’ Z. Anal. Chem., 238 (2), 81–89. Malissa, H. and Jellinek, G. (1969) Automation in und mit der analytischen Chemie. III: Zur Frage der Programmiersprachen und Symbolik. Fresenius’ Z. Anal. Chem., 247 (1–2), 1–8. Malissa, H. (1971) Automation in und mit der Analytischen Chemie. IV: Bedeutung der Semeiotik und des Informationsgehaltes in der automatischen Analyse. Fresenius’ Z. Anal. Chem., 256 (1), 7–14. Malissa, H. and Simeonov, V. (1978) Symbolik in der Analytischen Chemie. Fresenius’ Z. Anal. Chem., 289 (4), 257–263. Malissa, H. (1980) Die Zukunft der Analytischen Chemie. Fresenius’ Z. Anal. Chem., 300 (1), 9–11. Oesterle, G. (1995) Prozessanalytik: Grundlagen und Praxis, Oldenbourg, München, Wien.

69

70

2 Automation in Life Sciences – A Critical Review

143 Kessler, R., Küppers, S., Stieler, S., and Mannhardt, J. (2007) Trends in der

144 145

146

147 148 149

150

151

152

153

154 155

156

157

158

Prozessanalytik – Aktueller Stand und Ausblick. GIT Labor Fachz., 51 (4), 260–262. Wühr, M. (2012) PAT auf der Datenautobahn – Richtige PAT-Analysenmethode für die Überholspur. Laborpraxis, 36 (12), 18–19. Lambrecht, A. (2013) Trends in der laserbasierten Prozessanalytik: Grundlagen, Komponenten und Applikationen im Überblick. GIT Labor Fachz., 57 (6), 363–365. Keane, C.A. and Burrows, H. (1908) The autolysator an apparatus for the automatic determination of carbon dioxide. J. Soc. Chem. Ind., 27 (12), 608–612. Dennis, L.M. (1913) Gas Analysis, Macmillan, New York. Olsen, K. (2012) The first 110 years of laboratory automation: technologies, applications, and the creative scientist. J. Lab. Autom., 17 (6), 469–480. DIN EN 61115. (1994) Deutsches Institut für Normung e.V. Angabe zum Betriebsverhalten von Probenhandhabungssystemen für Prozessanalysengeräte, Beuth Verlag GmbH, Berlin. Wüstkamp, D., Kucharkowski, R., and Broekaert, J.A.C. (1996) Highly accurate analysis of soft magnetic alloys by automated sample digestion and simultaneous ICP-OES. Fresenius J. Anal. Chem., 355 (3–4), 281–283. Spörkel, O. and Platthaus, M. (2006) Autosampler in der Chromatographie, Laborpraxis, http://www.laborpraxis.vogel.de/labortechnik/articles/106031/ (accessed 01 December 2016). Foster, F.D., Stuff, J.R., and Pfannkoch, E.A. Automated Desorption, SPE Extraction, and LC/MS/MS Analysis of Dried Blood Spots – Gerstel Application Note 4/2015, http://www.gerstel.com/pdf/p-lc-an-2015-04.pdf (accessed 01 December 2016). Millan, R., Schmid, T., Sierra, M.J., Carrasco-Gil, S., Villadóniga, M., Rico, C. et al. (2011) Spatial variation of biological and pedological properties in an area affected by a metallurgical mercury plant: Almadenejos (Spain). Appl. Geochem., 26 (2), 174–181. Kopp, G. (2012) Quecksilberanalytik – Für jede Applikation die optimale Methode. LABO – Instrument. Anal., 2012 (6), 20–23. Rohasys Robotic and Handling Systems, Rouwette Consultancy Probenvorbereitung für Schwermetallanalysen, http://issuu.com/rouwetteconsultancy/ docs/probenvorbereitung_f__r_schwermetal (accessed 01 December 2016). Milburn, J.W. (1996) Automated addition of internal standards for axial-view plasma ICP spectrometry using the Optima 3000 XL. At. Spectrosc., 17 (1), 9–14. Barkawi, L.S., Tam, Y.-Y., Tillman, J.A., Pederson, B., Calio, J., Al-Amier, H. et al. (2008) A high-throughput method for the quantitative analysis of indole-3-acetic acid and other auxins from plant tissue. Anal. Biochem., 372 (2), 177–188. Saitoh, S. and Yoshimori, T. (2008) Fully automated laboratory robotic system for automating sample preparation and analysis to reduce cost and time in drug development process. J. Assoc. Lab. Autom., 13 (5), 265–274.

References

159 Sims, J., Smith, A., Patel, D., Batchelor, R., and Carreira, J. (2011) Automated

160

161

162

163 164

165

sample preparation for icp analysis of active pharmaceutical ingredients and intermediates. J. Lab. Autom., 16 (5), 377–380. Tsina, I., Chu, F., Hama, K., Kaloostian, M., Tam, Y.L., Tarnowski, T. et al. (1996) Manual and automated (robotic) high-performance liquid chromatography methods for the determination of mycophenolic acid and its glucuronide conjugate in human plasma. J. Chromatogr. B, 675 (1), 119–129. Bernsmann, T. and Baumann, M. (2014) Probenvorbereitung für die Dioxin Analytik: Die Automation eines manuellen Prozesses. GIT Labor Fachz., 58 (11), 41–43. Hong, J., Miki, Y., Honda, K., and Toita, H. (2012) Development of the automated cleanup system for the analysis of PCDDs, PCDFs and DL-PCBs. Chemosphere, 88 (11), 1287–1291. Sundberg, S.A. (2000) High-throughput and ultra-high-throughput screening: solution- and cell-based approaches. Curr. Opin. Biotechnol., 11 (1), 47–53. Gałuszka, A., Migaszewski, Z., and Namie´snik, J. (2013) The 12 principles of green analytical chemistry and the SIGNIFICANCE mnemonic of green analytical practices. Trends Anal. Chem., 50, 78–84. Anastas, P.T. (1999) Green Chemistry and the role of analytical methodology development. Crit. Rev. Anal. Chem., 29 (3), 167–175.

71

73

3 Automation Concepts for Life Sciences 3.1 Classification of Automation Systems According to Lauber and Göhner, there exist different possibilities for an adequate classification of technical processes [1]. Based on the type of the converted and transported medium, these processes can be categorized as material, energy, and information processes. Additional processes can be differentiated according to the type of their impact. These include generation, distribution, and storage processes. A classification of procedural, productional, and material handling processes is useful considering the type of the material transformation. Besides these classifications, which are relevant to specific industrial branches, the consideration of process parameters collected and analyzed in technical processes is also of great interest. Continuous, sequential, and object-related processes can be derived from these process parameters. Using the dominant processes, a differentiation of steady flow processes (continuous), follow-up processes (sequential), and object-related processes is possible. According to Lauber and Göhner, process automation systems consist of three subsystems connected together: the technical system, in which the technical process is realized; the computer and communication system; and the operator personnel [1]. Früh and Maier defined a similar terminology: (i) the system to be automated includes the manufacturing plant without the automation devices; (ii) the automation system consists of the computer system, sensors, and actuators. Both subsystems together form the complete automated system [2]. The literature offers a variety of possibilities for the classification of automation systems from different aspects. Gevatter distinguishes between automated control systems with open and closed procedures as well as combinations of both [3]. Früh and Maier differentiate centralized and decentralized automation systems [2]. In a centralized system, all close-to-process functions are realized on one computer, whereby the visualization and operating functions can be localized on the same computer or on one or more separate computers. In contrast, in decentralized or distributed automation systems all close-to-process functions are distributed on different computers whereby the visualization and operating functions can be realized on one or more different computers. Decentralized automation systems also enable a hierarchical structure of the components for the execution of the close-to-process functions. Automation Solutions for Analytical Measurements: Concepts and Applications, First Edition. Heidi Fleischer and Kerstin Thurow. © 2018 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2018 by Wiley-VCH Verlag GmbH & Co. KGaA.

74

3 Automation Concepts for Life Sciences

Lauber and Göhner applied the differentiation criteria centralized/ decentralized systems from different perspectives, which led to three different types of automation structures [1]: • Process structure (meaning the structure of the technical processes) • Local structure (meaning the structure/distribution of the several automation devices) • Functional structure of the automation system (meaning a serial, centralized or parallel, decentralized execution of the automation functions). Table 3.1 shows a schematic overview of the mentioned automation concepts that consider the typical structures of industrial processes. According to Lauber and Göhner five elemental criteria can be used for the evaluation of automation structures [1]. The first criterion includes business aspects such as the financial expenditures for the hardware and software and the wiring systems as well as maintenance. The second criterion is the reliability of the automation system under operating conditions with respect to hard and software failures and errors. The third criterion is the flexibility of the system related to changes and updates. The coordination of the different subprocesses and the optimization of the overall process can be considered as a fourth criterion. The fifth and last criterion is the operability and usability of the automation system. With respect to these criteria, the consideration of the above mentioned automation structures leads to the following general statement [1]: The automation structure should be as decentralized as possible and as centralized as necessary. The listed automation structures are usually applied to classical industrial production and manufacturing processes [1–3]. Currently, many processes from the fields of the production and manufacturing industry as well as the transport industry are already automated. Automation is used in industrial processes to increase the profitability, productivity, and efficiency as well as the availability and flexibility. Additional goals include increase of process safety and security, product quality, environmental compatibility, and optimal operability. A special case is the processes from the field of life sciences, which have complex, flexible, knowledge-intensive, distributed, and parallel characteristics. Table 3.1 Automation structures according to Lauber and Göhner [1] (C: centralized, D: decentralized, left character: structure of the technical processes, center: local structure of the automation devices, right character: functional automation structure). Process centralized (Techn. process = unity)

Process decentralized (Techn. process = subprocesses)

Locally centralized

Locally decentralized

Locally centralized

Locally decentralized

Functionally centralized

CCC

CDC

DCC

DDC

Functionally decentralized

CCD

CDD

DCD

DDD

3.2 Classification Concept for Life Science Processes

In contrast to classical automation processes, these processes use mainly heterogeneous resources. Automated, partially automated, and manual tasks are combined in highly variable process chains using a large number of different control structures. The automation in life science laboratories is currently dominated by the automation of process methods using partial or fully automated islands and numerous IT systems [4]. An important criterion that determines the difference between classical industrial automation and automation of life science processes is the required flexibility of the automation systems. This applies especially to analytical measurement processes and was already described in 1965 by Beyermann: Since many new methods have to be investigated, the flexibility of automation systems used is of great importance [5].

3.2 Classification Concept for Life Science Processes The general classification of automation systems according to Lauber and Göhner can also be transferred to the processes of life sciences. As mentioned above, three general automation structures can be categorized, which include the process structure, the local structure, and the functional structure of the automation system. The structure of a technical process can be defined as a centralized process, if the process can be considered a unit [1]. In contrast, processes that consist of different subprocesses form decentralized process structures [1]. Processes of analytical measurement technologies have such a decentralized structure, since, due to the concept of pre-, intra-, and post-sensoric selectivity every process is divided into subprocesses. A classification of functionally centralized and decentralized structures is not useful, since, due to the automation hierarchies, a combination of both structures is realized [1]. In laboratory automation, the focus must be placed on the flexibility of the automation systems instead. The modular expandability and adaptability of the automated systems to new applications in the field of analytical measurements is another important criterion of quality [6]. An example from physics can be used to describe the terms flexibility and expandability. In thermodynamics, a classification of open, closed, and isolated systems is common. Open systems enable an exchange of material as well as of energy. In closed systems, only energy exchange is possible, whereas in isolated systems no exchange with the environment occurs. These definitions can be analogously transferred to automation systems in life science applications, where the case of the isolated system is practically irrelevant. Adaptive systems can thus be defined as open systems, and non-adaptive systems as closed systems. Based on the already described subprocesses, sample preparation, analytical measurement, and transport, which occur in all analytical measurement processes, the following system concepts can be classified (Table 3.2). If all processes of sample preparation, sample transport, and analytical measurement of the samples are executed in one complex facility; a centralized closed automation system can be defined (Figure 3.1). Defined processes can be performed in such self-contained systems; the flexibility depends on the range of

75

76

3 Automation Concepts for Life Sciences

Table 3.2 Automation structures adapted to life science processes: (C: centralized, D: decentralized, O: open, Cl: closed, left character: local distribution of the automation devices; right character: flexibility of the automation structure). Process decentralized (Techn. process = subprocesses) Locally centralized

Locally decentralized

Closed

CCl

DCl

Open

CO

DO

Centralized/closed SP

Figure 3.1 Schematic visualization of a centralized closed automation system. (SP: sample preparation for one application, A: analytical measurement for one application, CE: connection element).

Stationary CE A

system functions. The process is completely determined; the automation system is proprietarily adapted to this process. A fast and effective sample processing with very short transport distances is possible, since all required components (sample preparation, analytical measurement) are installed in one compact facility. Low flexibility is the disadvantage since the process has to be defined in advance and changes are only possible with a reprogramming of the system. The integration of additional components is not possible either. A more flexible design of the systems is possible, if the available system components can be combined depending on the required process operation. This can be defined as a centralized open automation system, since all components are still integrated in one central facility and the process is thus executed locally centralized (Figure 3.2). The combination of the different components is realized via a connection element. Limitations regarding the processes, which can be performed, occur if the range of functions required by the process is not available on the system platform. Additional components can be integrated into the systems. Thus, the range of functions can be extended. Different life science processes have different requirements; thus, no automation systems are known, which can be used for all process classes. In the field of analytical measurement this can mean that subprocesses have to be executed in closed environments or closed subsystems due to safety requirements. For example, a decomposition of samples using microwave technology has to be performed under a safety hood. The development of a complex automation system is in this case associated with very high costs (e.g., due to the requirements of explosion protection for the whole system). In addition, the development of one

3.2 Classification Concept for Life Science Processes

Figure 3.2 Schematic visualization of a centralized open automation system. (SP 1,2,…, X: sample preparation for different applications, A 1,2,…, X: analytical measurement for different applications, CE: connection element).

Centralized/open SP 2 SP X

SP 1 Stationary CE

AX

A1 A2

Figure 3.3 Schematic visualization of a decentralized closed automation system. (SP: sample preparation for one application, A: analytics for one application, CE: connection element).

Decentralized/closed SP Mobile CE A

closed system is not possible due to the constraint of available space resources. In other cases, high-performance analytical devices have to be accessible for different applications and users so that a direct integration of all possible analytical devices is not meaningful. Here, the space resources also play a vital role. Analytical measurement systems require specific environmental conditions with respect to air quality (contamination-free air), which cannot be guaranteed in chemical applications, for example, due to the evaporation of solvents. The fact that, in analytical processes, some manual subprocesses can be included is another complication, since the automation of these subprocesses is connected with rather high costs. This requires suitable concepts for these processes. A complete automation can require the combination of different automation islands, and also the implementation of manual sample preparation steps. The transport of the samples between different automation islands can be realized using conveyers or robots. Mobile robots can be used to guarantee maximum flexibility. According to the general classification, these systems can be defined as decentralized closed automation systems (Figure 3.3). Sample preparation, sample transport, and the analytical investigation of the samples are distributed to different facilities with one or more connection elements. The process to be executed is predetermined and can only be modified by reprogramming the whole system. Thus, the system is considered a closed system. The integration of additional components is only possible with a considerable amount of time and money. Much better flexibility can be guaranteed in decentralized open automation systems (Figure 3.4). An open system consists of different autarkic automation islands with a limited range of functions for each system. Single islands can be

77

78

3 Automation Concepts for Life Sciences

Decentralized/open SP 1

SP 2

SP X

Figure 3.4 Schematic visualization of a decentralized open automation system. (SP 1,2,…, X: sample preparation for different applications, A 1,2,…, X: analytics for different applications, CE: connection element).

Mobile CE A1

A2

AX

combined in any desired way by mobile transport systems. Thus, a flexible use of the available systems and a flexible design of the processes are enabled. The number of automation islands to be integrated is theoretically not limited; the integration of new components and stations is always possible. Due to the completely open structure, complex and long transport distances might be necessary. The hardware configuration is essential for the classification of the automation systems. Besides this, the control software of the systems contributes to the flexibility as well and thus to the degree of openness, which can be achieved in the automation system. The higher the degrees of flexibility of an automation system, the higher are the requirements for the design of the process control software. Especially in the case of decentralized open systems, a shift from process control systems to process management systems can be seen.

3.3 Robot Based Automation Systems 3.3.1 Robot Based Systems in Industrial Automation

Two principal possibilities exist for the use of robots in automation. On the one hand, they can be used as pure transport systems; on the other hand, the possibility of robotic component implementation exists for a direct execution of different tasks. The latter possibility has been integrated in industrial automation for many years. The goal is to perform the manual tasks with a robot; the human hand, specifically the human arm should be replaced and the robot should emulate the human skills. Examples can be found in the automotive industry, precision mechanics, food industry, heavy machinery construction, electrical engineering, and medicine [7]. Single arm or dual arm robots can be used relevant to the tasks that have to be performed. In automotive industry one or more single arm robots are classically used for painting big workpieces such as vehicle bodies [7]. The use of dual arm robots is currently gaining more importance. Examples can be found in industrial manipulation tasks [8], the support of humans with health limitations [9, 10], or in the field of laboratory robotics [11–13]. Here, two tasks can be performed within a limited workspace without the need of a transport step. Three different modes can be defined: in the uncoordinated operation both arms perform separate tasks, independent of what the other arm performs (e.g., one arm is stacking parts, while

3.3 Robot Based Automation Systems

the other arm is working on a weld seam). If both arms are involved in performing a task, the operation mode is goal oriented (e.g., both arms packing parts into the same package). In the bimanual operation, both arms are required to perform one task (e.g., lifting and transporting of a package with both arms) [8]. The use of robots for transportation tasks has been reported in the field of industrial warehouse management. In this case, the task of the robotic components is not limited to the pure transport between different stations. Instead, they also perform grasping and placing of transport objects. Grippers or other suitable end effectors grasp the relevant objects from the storage rack. Subsequently, they are transported directly or after placing on a transportation tray to the place of destination. Here, handover of the transported goods takes place. Classically, robots are used with rails, which guarantee the mobility of the robot(s). Furthermore, mobile robots exist, which show a greater operating range and higher flexibility due to the lack of rails. 3.3.2 Robot-Based Automation Systems in Life Sciences 3.3.2.1 Concept of the Central Robot as System Integrator

With the beginning of the automation of life science processes, suitable robotic concepts had to be developed to support a complete automation. In contrast, at this time of already highly developed industrial automation, the concept of using the robot as a central system integrator was preferred for life science processes. This was mainly due to the high flexibility of life science processes, which has also to be guaranteed by the automation system. Industrial solutions are mainly designed and realized as proprietary systems for a specific application and will be used for a couple of years without bigger changes to the system. In these systems, robots are used for detailed tasks such as welding, painting, screwing, polishing, and so on [7]. Compared to this, the main goal of automating life science processes is the design and realization of systems, which ensure the typical range of functions required for biological, chemical, biochemical, and analytical processes. According to the included functionalities, different specific applications can be performed on these automation systems. This is enabled due to the strict hierarchical design of the systems. The robot is used as a system integrator and acts only as a transportation element. It transports samples between different subsystems of a complex automation system. Autonomous intelligent devices realize the essential tasks such as dispensing, aliquoting, heating, shaking, measuring, and so on. The main advantages of this concept are the simple design and easy programming of the robots, since it only executes transportation tasks but no manipulation of the sample material. The task performing devices have to be adapted to the robot handling to enable an automated delivery and pick up of the samples by the robot. Such completely automated systems are widely used in the field of bioscreening in the pharmaceutical industry. Conventional devices can usually not be used, since they are optimized for the manual operation by a human operator. Thus, the adaptation of existing devices or the development of completely new devices is required to enable a completely automated process. Both versions are connected with high costs. Additionally, these systems have the disadvantage that real 1 : 1 automation is not possible without changes in the

79

80

3 Automation Concepts for Life Sciences

process chain. This is a big problem for highly regulated areas such as pharmaceutical or medical processes and is one of the reasons why these fields, for example, analytical processes, still show a very low degree of automation. Any change in the process requires a completely new validation and approval of the processes, which is connected with considerable costs in addition to the costs of the automation system itself. 3.3.2.2 Concept of the Flexible Robot

Although robots are classically used for the execution of different tasks in the industry, this type of automation has not yet been applied to the automation of life science processes. Main reasons include the variety and diversity of the processes, the frequent change of the tasks to be performed, and the absence of a constant number of samples, which have to be processed by the system. The current development of modern dual arm robots brings into focus the concept of new strategies for the automation of life science processes. In such systems, the robot again is used as a system integrator. Analogous to the classical version of laboratory automation, the robot transports the samples between different subsystems, stations, and devices. In the easiest case analog to the concept of a central robot, the devices that must be loaded by the robot require adaptation. This is connected with considerable costs. Alternatively, the end effectors of the robot arms can be designed in a way that enables copying the functionality of the human hand. This also includes the pressing of buttons or pulling out drawers. Beside transportation tasks, such dual arm robots can perform typical laboratory tasks. Thereto classical devices and tools can be used, which are known from routine manual laboratories. For the reliable picking of the devices and tools, only the design of suitable brackets or holders is necessary. The operation of the devices is realized analogous to the manual process steps by programming the motion sequences. Systems with dual arm robots can be designed in two general configurations: • Robots with a fixed position can be surrounded by a deck, on which all devices, tools, labware, and consumables are positioned, which are required for the execution of the processes. With intelligent programming, a flexible design of the deck according to the requirements of the actual processes is possible. • If mobile dual arm robots are used, they can approach different laboratories and laboratory stations. Sample manipulation at the different stations is possible besides the pure transport function between the stations. This concept completely matches the terminology of an open automation system; the new term integrated robotics can be introduced. Both versions have the advantage that the automation of life science processes is possible with classical laboratory equipment and tools. Thus, the cost for specialized automated stations as used in combination with central robots can considerably be reduced. In addition, real 1 : 1 automation, this means complete automation of the processes analogous to the manual processes without changes of the process steps, can be realized. The only difference is the transfer of the tasks from the human operator to the robot. Thus, such automation concepts

3.3 Robot Based Automation Systems

can also be used in fields with a high degree of standardization and regulation, since due to the exact transfer of the manual processes to the robot, a new validation and approval of the processes is not required. 3.3.3 Summary and Application of Concepts

The combination of the concepts of a general classification of automation systems and the use of robots in life science applications leads to a complex classification as illustrated in Figure 3.5. The introduced concept can generally be applied to all life science processes. Centralized systems, which can either follow a closed or open concept, are Central robot as system integrator Centralized/open

Centralized/closed SP

SP 1

CSI

SP 2

SP X

Decentralized/closed

Decentralized/open

SP

CSI

SP 1 SP 2 SP X CSI CSI CSI

Mobile CE

Mobile CE

CSI A1

A

Single application

AX

A

A2

Flexible applications

A1

Single application

A2

AX

Flexible applications

Flexible robot Centralized/closed

Centralized/open

SP

SP 1

SP 2

SP X

Decentralized/closed

A1

A

Single application

FR

SP 1 SP 2 SP X FR FR FR

Mobile CE

Mobile CE

SP

FR

FR

AX

A

A2

Flexible applications

Decentralized/open

A1

Single application

A2

AX

Flexible applications

Integrated robotics Decentralized/closed

Decentralized/open

SP 1 SP 2 SP X

SP 1 SP 2 SP X

CSI/FR

CSI/FR

CSI/FR

CSI/FR

IR

A1

A2

CSI/FR

CSI/FR

IR

AX

A1

Single application

A2

AX

Flexible applications Mandatory element Optional element

Figure 3.5 Automation concepts in life sciences (SP 1,2,…, X: sample preparation, A 1,2,…, X: analytical measurement, CSI: central system integrator, FR: flexible robot, CE: connection element, IR: integrated robotics).

81

82

3 Automation Concepts for Life Sciences

currently the state of the art in the automation of life science processes. An example for a centralized closed system has been reported by van de Bilt et al. [14] for the investigation of potatoes regarding the Ralstonia solanacearum bacterium, which causes high agricultural damages. This automation system has been designed for a specific application; the use of the system for other applications is not intended. The high-throughput screening system Tox21 is an example for a centralized open system. This robot-based bioscreening system has been designed for the determination of the toxicity of approximately 10 000 environmental substances and result storage in a substance library. The open structure enables the execution of different cell-based assays [15, 16]. An extension to decentralized systems has already been described by Chambers [17] and Scypinski et al. [18]. The authors especially highlighted the flexibility of such systems. An example for such a decentralized open system is the automation system reported by Radu et al., which includes three automation islands combined by a manual sample transport [19]. Due to the manual transport processes, the automation degree is less than 100% and the system is thus only a partially automated system. The high focus on biological screening processes is remarkable, whereas the automation of analytical processes has not yet been addressed in research and development. This book thus focuses on the automation of analytical processes. The automation of the determination of the enantiomeric excess of chiral compounds will be described as an example for a centralized closed system with a robot as system integrator. Sample preparation, analytical measurement of the samples, and data evaluation are realized on one closed automation system. An ORCA laboratory robot is used as a system integrator. The system has a proprietary configuration for a specific application; the use for other applications requires extensive changes to the system. In a centralized open system with a system integrator, different applications can be performed according to the range of functions of the system. An example is an automated sample preparation platform, which can be used for element- or structure-specific investigation of samples. The hardware configuration matches the concept of a central system, where all system components are combined into one complex system. Exemplarily applications include the sample preparation for the determination of mercury in wood (elemental composition) and the investigation of dental materials (structural composition). The sample preparation platform can be extended to a decentralized system. In this case, the components for the sample preparation and the analytical systems are isolated in different locations. Transport between the different resulting substations is realized by mobile robots, which only perform transportation tasks, but do not manipulate the samples. If the system is configured for a specific, exactly defined process, the result is a decentralized closed system. An application here is the analytical determination of calcium and phosphorus in bones including sample preparation and analytical measurement. The result is a decentralized open system, if different automation systems with robots as system integrators on which different applications can be executed are

3.4 Degree of Automation

combined with each other and/or additionally with partially automated or not automated stations by mobile robots. Applications include the flexible sample preparation for the analytical determination of mercury in wood and calcium and phosphorous in bones. The open system is also capable of performing sample preparation and analysis of dental materials. The system is distributed locally and functionally. As an example of a centralized closed system with a dual arm robot (flexible robot), the automation of the determination of the enantiomeric excess of chiral compounds will be introduced. A fixed central dual arm robot SDA 10 (Yaskawa, Kitaky¯ush¯u) is surrounded by a deck, on which all laboratory devices and tools required for the process realization are positioned. Classical laboratory devices that are used in the manual laboratory can be integrated. The robot manipulates the samples and transfers them to the integrated analytical devices. The system has a proprietary configuration and programming for a specific application. The system can flexibly be extended for the processing of additional applications to a centralized open automation system. Examples are the determination of chiral compounds and the determination of cholesterol in biliary stents. The result is a decentralized system, if mobile robots realize a sample transport to and from the SDA-platform. Depending on the fact, whether a proprietary process is performed in the system or the system can be flexibly used for different applications, the concept matches the terminology of a decentralized closed or a decentralized open system. As an example for a closed system, the sample preparation and analytical measurement of biliary stents should be described. In this system, the elemental calcium content and the concentration of cholesterol are investigated. The calcium determination requires microwave digestion, which due to safety reasons was not integrated on the SDA platform. The ICP-MS (Inductively Coupled Plasma Mass spectrometry) system was not integrated either, thus sample transport is necessary. The closed system can be extended to an open system with the integration of additional applications. This was realized with the integration of the determination of calcium and phosphorus in bones and of steroid hormones. A complete decentralized system structure (integrated robotics) can be realized if a mobile dual arm robot is used, which can approach different stations and manipulates the samples at these stations. In this case, closed as well as open systems can be configured. The sample preparation of cell cultures using ultrasonic digestion and solid phase extraction followed by an analytical determination of cyclophosphamide will be used as an example for this system concept.

3.4 Degree of Automation According to DIN IEC 60050-351, the degree of automation can be defined as the amount of automated functions relative to the entirety of all functions of a system or a technical facility [20]. The degree of automation can only be defined for a precisely described and defined system with a weighting of its functions.

83

84

3 Automation Concepts for Life Sciences

According to Hesse, the degree of automation can be calculated according to Equation 3.1 [21]: ∑ (Fautomated ⋅ P) ⋅ 100 (3.1) A= ∑ ∑ (Fautomated ⋅ P) + (Fnot automated ⋅ P) A:

Degree of automation

F automated :

Automated functions

F not_automated : Not automated functions P:

Weighting factor (weighting of the specific function)

If all functions of a system are automated (except start and shut down procedures, emergency intervention), a complete automated operation is the result, and the degree of automation reaches 100%. If no tasks are automated in the system, the degree of automation is 0%. All values between 0% and 100% indicate a partially automated operation [1, 20]. Endsley and Kaber developed a very detailed, hierarchical system of degrees of automation, which includes 10 different steps and allows a precise classification of automation systems [22]. The first level is the manual control (MC), where the human operator is performing all tasks. This includes the supervision/monitoring of the system status, the generation and selection of processes (decision-making), as well as the starting and performing of the processes. At the second level, the system supports the user with the execution of a selected process (action support, AS), whereby some control actions have to be realized manually. A typical example is a remote control system with a human input (master) and a manipulator (slave). The third level is the batch processing (BP). The user designs and selects the processes that have to be automatically executed by the automation system. The automation in this case includes only the execution of the processes. Examples include batch processing systems in industrial production or a cruise control system in vehicles. The next level is the shared control (SHC), whereby the user as well as the computer can design possible decision options. The user still has the complete control over the selection of processes that are to be executed. The execution of the processes is realized by the cooperation between the human and the automation system. Level five includes decision support (DS). The computer generates a list of decision options, from which the user can choose. Own processes can still be designed. This degree of automation is representative for a variety of expert and decision systems. In contrast to the SHC (level 4), the system can execute tasks with this degree of automation. Blended decision-making (BDM) is the next higher degree of automation (level 6). At this level, the computer generates a list of decision options, selects a process, and executes the process after confirmation by the user. The user can accept the processes selected by the system, select other processes, or design a new process. The selected action is then executed by the system. This degree of automation represents decision systems of higher degrees, which are able to select between different alternatives and execute the selected task. A rigid system (RS) is defined at level 7. This includes systems that only have a limited amount of

3.4 Degree of Automation

actions for the operator. The task of the operator in this concept is the selection of such a set of actions. It is not possible to design other processes. The system is relatively rigid and does not give the user discretionary powers. Complete realization of the selected tasks is done. In the automated decision-making (ADM) at level 8, the system autonomously selects the best options and executes them. The decision is made based on a list of alternatives provided by the system. This list can be extended with additional alternatives provided by the user. Such systems automate decision-making in addition to the development of options (such as in decision help systems). At the level of supervisory control (SC, level 9), the system generates processes, selects a suitable process, and executes the process. The main task of the user is the supervisory control and intervention if necessary. The intervention enables the user to select alternative processes (which might be designed by the computer or the user), which then means a shift to level 5 (decision support). This level represents a typical control system with supervisory control by the user and intervention on demand in connection with a highly automated system. The highest degree of automation (level 10) is the complete or full automation (FA). At this level, all actions are executed by the system. The human is completely outside the control loop, and therefore an intervention is not possible. This level is characteristic of fully automated systems, in which human and manual processing is not necessary. An overview about the mentioned degrees of automation is given in Figure 3.6 [22]. According to an actual market report from Kalorama Information, the actual trend of automation systems is changing from full automation of the complete laboratory to modular approaches [6, 23, 24]. Hardware-controlled automation systems are increasingly replaced by software-controlled process control and process management systems. Additionally, it is intended to replace the different proprietary single solutions with standardized processes and labware [24]. The goals of automation, miniaturization, and high throughput thus focus on the development of uniform assay formats in the field of bioscreening and uniform synthesis, sample manipulation, and analytical methods in the field of chemical analytics, which enable the processing of high sample numbers within short times. Table 3.3 summarizes the actual trends in the field of laboratory automation. Figure 3.6 Overview about the degrees of automation.

(10) Full automation (FA)

Level of automation

(9) Supervisory control (SC) (8) Automated decision making (ADM) (7) Rigid system (RS) (6) Blended decision making (BDM) (5) Decision support (DS) (4) Shared control (SHC) (3) Batch processing (BP) (2) Action support (AS) (1) Manual control (MC)

85

86

3 Automation Concepts for Life Sciences

Table 3.3 Trends in laboratory automation [24]. Full automation

→ Modular systems

Hardware-controlled automation systems → Software-driven process control Single solutions

→ Standardization

“Cool technology” (useful “toys”)

→ Laboratory tool

Another market report from Frost & Sullivan shows that a combination of individual workstations and integrated systems with single components is the preferred version of automation. About 23% respondents prefer individual workstations, 7% prefer integrated systems with several components and 70% prefer both types of approaches [25]. This again demonstrates the trend toward flexible automation structures.

3.5 Statistical Evaluations The methods of physico-chemical measurement have a higher complexity compared to bioanalytical methods. This leads to more complex evaluation and validation parameter. General terms of measuring technology are described in DIN 1319 Part 1, which can also be applied to analytical measurement [26]. Validation parameters such as the limit of detection (LOD) and the limit of quantification (LOQ) also are widely discussed currently by different authors [27–34]. The separate norm DIN 32645 defines these parameters especially for analytical measurement [35]. The basics, terms, and methods for the determination of accuracy and precision of measuring methods are defined in the DIN ISO 5725 Parts 1 and 2 [36, 37]. Especially there exist specific requirements for the competence of test laboratories and calibration laboratories, which are defined in the DIN EN ISO/IEC 17025 [38]. The Horwitz criterion can be used as an index for the performance of measurement methods with respect to their precision [39, 40]. The fundamental work of William Horwitz provides an empirical equation for the calculation of the common coefficient of variation CV (also known as relative standard deviation) in correlation to the analyte concentration. Equation 3.7 defines the typical value of the coefficient of variation for the between-laboratory precision. The within-laboratory precision and the repeatability respectively are defined at the half up to two-thirds of the between-laboratory precision Equation 3.8. Additional validation parameters such as the selectivity, specificity, and sensitivity of analytical methods are discussed in [41, 42]. The David test for the determination of normal distribution of the measured data [31], the t-test for the evaluation of two or more measurement series [31, 43], calibration and regression [44], variance analysis [45], and the determination and evaluation of outliers [46] are the main statistical parameters and calculations in data evaluation. In the following chapters practice-oriented definitions according to the Equations 3.2–3.10 are used, which correspond to the

3.5 Statistical Evaluations

Table 3.4 Validation parameter. 1 ∑ x n i=1 i √ √ n √ 1 ∑ 𝜎=√ (x − x)2 n − 1 i=1 i n

Average

x=

Standard deviation

𝜎 ∗ 100 x

(3.2)

(3.3)

Coefficient of variation

CV =

Limit of detection

LOD = xBlank + 3 ∗ 𝜎Blank

(3.5)

Limit of quantification

LOQ = xBlank + 10 ∗ 𝜎Blank

(3.6)

Horwitz criterion (coefficient of variation) for betweenlaboratory precision [39]

CV = 2(1−0.5⋅log C) (3.7) C: Analyte concentration (1 ppm = 10−6 )

Horwitz criterion (coefficient of variation) for within-laboratory precision

2 1 (1−0.5⋅log C) ≤ CV ≤ ⋅ 2(1−0.5⋅log C) ⋅2 2 3

(3.4)

(3.8)

DIN 1319 Part 1 and DIN 32645 and scientific literature [31, 32, 39, 40, 47] (see Table 3.4). Although the validation parameters usually depend on the specific application and the related regulations, standard procedures can be found for the determination of the parameter working area, linearity, repeatability, recover rate, within-laboratory precision, method stability, LOD, and LOQ [31]. For all applications described the repeatability (also called intraday precision) was determined with at least 25 replicates that have been prepared and measured on one day. Furthermore, a test on normal distribution by David test was performed. According to Equation 3.9 the test score can be calculated. The borders for a defined statistical certainty (usually 99%) are defined due to the statistic table of David. In normal distributed measurement values, the test score must be inside these borders [31]. Test score by David: TS =

max. value − min. value R = standard deviation 𝜎

TS

Test score

R

Range

𝜎

Standard deviation

(3.9)

For the determination of the within-laboratory precision, measurements were performed on five consecutive days with identical numbers of samples (3 up to 10 samples depending on the application). The mean values, standard deviations, and coefficients of variations have been considered for the evaluation of the single data points.

87

88

3 Automation Concepts for Life Sciences

The robustness of a measuring method is evaluated by the failure proneness of the measurement values due to parameter changes or changes of the conditions. This includes the method robustness, method stability, and the applicability (or transferability) [31]. For the investigation of the method robustness the measurement method is tested for its failure proneness due to different method parameters. This includes, for example, the reaction temperature and the reaction time. The method stability describes the stability of the measurement values during one measurement series. A sample set with 3–10 replicates were prepared. The sample solutions were divided into five equal parts resulting in five sample sets. One sample set was measured immediately, while the other sets were stored at −18 ∘ C. Each following day one sample set was defrosted at room temperature and measured. Other criteria for the validation of analytical methods are the applicability and transferability, which means the failure proneness of the method due to different operators, devices, and/or different laboratories. The measurement precision, which is also called device or system precision describes the amount of variations caused by the measurement system [31]. For the determination of the measurement precision one sample was prepared and measured 10 times. In general, coefficients of variation of less than 2% are acceptable in analytical measurement [31]. The discrimination threshold of a measurement system according to DIN 1319 Part 1 is the smallest change of the input variable that can cause a recognizable change of the output variable [26]. In analytical measurement, a distinction between LOD, limit of decision, and LOQ is common, whereas different definitions of these terms can be found in literature. According to DIN 32645 the LOD is defined as the smallest signal significantly different from the blank value; this is the smallest detectable value of a compound [35]. According to this reference, the limit of decision is the minimum amount, which can be determined with high probability. Often this value is defined as the double LOD and should be smaller than the lowest calibration point. The limit of quantification defines the smallest amount of a compound, which can be quantified (e.g., the mass percentage or the mass concentration). For all described applications in the following chapters the LOD and the LOQ were calculated according to the Equations 3.5 and 3.6. Ten blank samples containing only solvents and typical reagents were prepared and measured. Some analytical methods are connected with minimal sample losses. One example is microwave digestion, where nitrous gases, solvents, or high volatile analytes can evaporate from the opened vessels. An evaporation of solvents would cause the detection of higher concentration and a loss of highly volatile compounds would cause the detection of lower concentrations. Variations in the analyte concentration can also be caused in the sample delivery process, for example, due to the peristaltic pump in an ICP-MS or ICP-OES (Inductively Coupled Plasma Optical Emission Spectroscopy). To adjust these errors, an internal standard with a known concentration is added to the samples. The adjustment of the measurement values is then realized according to Equation 3.10, which is usually done in the instrument software. For the correction of two systematic errors (e.g., evaporation and sample injection) the use of two internal standards is preferred. Since the instrument software can only process one internal standard the calculations have to be done manually or with the help of spread sheet programs.

References

Measurement correction Canalyte, measured − Cinternal standard, theoretical Canalyte, corrected = Cinternal standard, measured

(3.10)

C: concentration (μg/l or mg/l).

References 1 Lauber, R. and Göhner, P. (1999) Prozessautomatisierung 1: Automa-

2

3 4

5 6

7 8

9

10

11

12

13

tisierungssysteme und – strukturen, Computer- und Bussysteme für die Anlagen- und Produktautomatisierung, Echtzeitprogrammierung und Echtzeitbetriebssysteme, Zuverlässigkeits- und Sicherheitstechnik, 3rd edn, Springer, Berlin. Früh, K.F. and Maier, U. (2004) Handbuch der Prozessautomatisierung: Prozessleittechnik für verfahrenstechnische Anlagen, 3rd edn, München, Oldenbourg. Gevatter, H.-J. (2000) Automatisierungstechnik, Springer, Berlin, Heidelberg. Thurow, K., Junginger, S., Stoll, N. (2005) System integration for full automation – From single components to total system (Systemintegration für die Vollautomation – Von Einzelkomponenten zum Gesamtsystem). BioSpektrum, 11 (5), 666–670. Beyermann, K. (1965) Zur Automatisation mikrochemischer Untersuchungen. Fresen. Z. Anal. Chem., 210 (1), 1–9. Horowitz, G.L., Zaman, Z., Blanckaert, N.J.C., Chan, D.W., DuBois, J.A., Golaz, O. et al. (2005) Modular analytics: a new approach to automation in the clinical laboratory. J. Autom. Methods Manage. Chem., 2005, 8–25. Haun, M. (2007) Handbuch Robotik: Programmieren und Einsatz intelligenter Roboter, 1st edn, Springer, Berlin. Smith, C., Karayiannidis, Y., Nalpantidis, L., Gratal, X., Qi, P., Dimarogonas, D.V. et al. (2012) Dual arm manipulation – a survey. Rob. Autom. Syst., 60 (10), 1340–1353. Cunningham A., Keddy-Hector W., Sinha U., Kruse D.W.D., Braasch J., Wen J.T. 2014 Jamster: a mobile dual-arm assistive robot with Jamboxx control. Conference Record – IEEE International Conference on Automation Science and Engineering CASE, pp. 509–514. Oh, K.W., Lee, K., Ahn, B., and Furlani, E.P. (2012) Design of pressure-driven microfluidic networks using electric circuit analogy. Lab Chip, 12 (3), 515–545. Moore, K.W., Newman, R., Chan, G.K.Y., Leech, C., Allison, K., Coulson, J. et al (2007) Implementation of a high specification dual-arm robotic platform to meet flexible screening needs. J. Assoc. Lab. Autom., 12 (2), 115–123. Chu X., Fleischer H., Klos M., Stoll N., Thurow K. (2015) Application of dual-arm robot in biomedical analysis: sample preparation and transport. IEEE International Instrumentation and Measurement Technology Conference, I2MTC, 2015, pp. 500–504. Fleischer, H., Drews, R.R., Janson, J., Chinna Patlolla, B.R., Chu, X., Klos, M. et al. (2016) Application of a dual-Arm robot in complex sample preparation and measurement processes. J. Lab. Autom., 21 (5), 671–681.

89

90

3 Automation Concepts for Life Sciences

14 van de Bilt, J.L.J., Derks, J.H.J., and Janse, J.D. (2008) Automated,

15

16

17 18

19

20

21 22

23 24 25 26 27 28 29

30 31 32

high-throughput immunofluorescence staining: a new approach. Eur. J. Plant Pathol., 120 (1), 91–96. Shukla, S.J., Huang, R., Austin, C.P., and Xia, M. (2010) The future of toxicity testing: a focus on in vitro methods using a quantitative high-throughput screening platform. Drug Discovery Today, 15 (23–24), 997–1007. Attene-Ramos, M.S., Miller, N., Huang, R., Michael, S., Itkin, M., Kavlock, R.J. et al. (2013) The Tox21 robotic platform for the assessment of environmental chemicals – from vision to reality. Drug Discovery Today, 18 (15–16), 716–723. Chambers, D. (1994) Decentralized management of laboratory automation. J. Autom. Chem., 16 (4), 135–137. Scypinski, S., Nelson, L., and Sadlowski, T. (1995) Automation in the pharmaceutical analysis laboratory: a centralized/decentralized approach. J. Autom. Chem., 17 (2), 47–49. Radu, C., Adrar, H.S., Alamir, A., Hatherley, I., Trinh, T., and Djaballah, H. (2012) Designs and concept reliance of a fully automated high-content screening platform. J. Lab. Autom., 17 (5), 359–369. Deutsches Institut für Normung e.V (2014) Internationales Elektrotechnisches Wörterbuch – Teil 351: Leittechnik. DIN IEC 60050-351, Beuth Verlag GmbH, Berlin. Hesse, S. and Malisa, V. (2010) Taschenbuch Robotik – Montage – Handhabung, München, Fachbuchverlag Leipzig im Carl-Hanser-Verlag. Endsley, M.R. and Kaber, D.B. (1999) Level of automation effects on performance, situation awareness and workload in a dynamic control task. Ergonomics, 42 (3), 462–492. Information, K. (2008) The Worldwide Market for lab Automation: Market Study, Kalorama Information, New York. Markin, R.S. and Whalen, S.A. (2000) Laboratory automation: trajectory, technology, and tactics. Clin. Chem., 46 (5), 764–771. Frost & Sullivan (2008) European Markets for Laboratory Automation Systems: Market Study, Frost & Sullivan, London. Deutsches Institut für Normung e.V (1995) Grundlagen der Meßtechnik – Teil 1: Grundbegriffe. DIN 1319-1, Beuth Verlag GmbH, Berlin. Kaiser, H. (1965) Zum Problem der Nachweisgrenze. Fresen. Z. Anal. Chem., 209 (1), 1–18. Kaiser, H. (1966) Zur Definition der Nachweisgrenze, der Garantiegrenze und der dabei benutzten Begriffe. Fresen. Z. Anal. Chem., 216 (1), 80–94. Luthardt, M., Than, E., and Heckendorff, H. (1987) Nachweis-, Erfassungsund Bestimmungsgrenze analytischer Verfahren. Fresen. Z. Anal. Chem., 326 (4), 331–339. Liteanu, C. and Rica, I. (1973) Über die Definition der Nachweisgrenze. Mikrochim. Acta, 61 (5), 745–757. Kromidas, S. (2011) Validierung in der Analytik, Weinheim, Wiley-VCH Verlag GmbH. Kromidas, S. (2000) Handbuch der Validierung in der Analytik, Weinheim, Wiley-VCH Verlag GmbH.

References

33 Moreda-Piñeiro, J., Alonso-Rodríguez, E., López-Mahía, P.,

34

35

36

37

38

39 40

41

42 43 44 45 46 47

Muniategui-Lorenzo, S., Prada-Rodríguez, D., Moreda-Piñeiro, A. et al. (2007) Determination of major and trace elements in human scalp hair by pressurized-liquid extraction with acetic acid and inductively coupled plasma-optical-emission spectrometry. Anal. Bioanal. Chem., 388 (2), 441–449. Belter, M., Sajnóg, A., and Barałkiewicz, D. (2014) Over a century of detection and quantification capabilities in analytical chemistry – historical overview and trends. Talanta, 129, 606–616. Deutsches Institut für Normung e.V (2008) Chemische Analytik – Nachweis-, Erfassungs- und Bestimmungsgrenze unter Wiederholbedingungen – Begriffe, Verfahren, Auswertung. DIN 32645, Beuth Verlag GmbH, Berlin. Deutsches Institut für Normung e.V (1997) Genauigkeit (Richtigkeit und Präzision) von Messverfahren und Messergebnissen – Teil 1: Allgemeine Grundlagen und Begriffe. DIN ISO 5725-1, Beuth Verlag GmbH, Berlin. Deutsches Institut für Normung e.V (2002) Genauigkeit (Richtigkeit und Präzision) von Messverfahren und Messergebnissen – Teil 2: Grundlegende Methode für die Ermittlung der Wiederhol- und Vergleichspräzision eines vereinheitlichten Messverfahrens. DIN ISO 5725-2, Beuth Verlag GmbH, Berlin. Deutsches Institut für Normung e.V (2005) Allgemeine Anforderungen an die Kompetenz von Prüf- und Kalibrierlaboratorien. DIN EN ISO/IEC 17025, Beuth Verlag GmbH, Berlin. Horwitz, W. (1982) Evaluation of analytical methods used for regulation of foods and drugs. Anal. Chem., 54 (1), 67A–76A. Horwitz, W. and Albert, R. (2006) The Horwitz ratio (HorRat): A useful index of method performance with respect to precision. J. AOAC Int., 89 (4), 1095–1109. Thurow K. 1999 Ein Methodenspektrum zur selektiven messtechnischen Bestimmung stofflicher Spezies durch spektroskopische Messmethoden am Beispiel ausgewählter Arsenverbindungen. Habilitation thesis. Universität Rostock. Kaiser, H. (1972) Zur Definition von Selektivität, Spezifität und Empfindlichkeit von Analysenverfahren. Fresen. Z. Anal. Chem., 260 (3), 252–260. Burke, S. (2001) Understanding the structure of scientific data. LC GC Europe Online Supplement, pp. 3–7. Burke S. 2001 Regression and calibration. LC GC Europe Online Supplement, pp. 13–8. Burke S. 2001 Analysis of variance. LC GC Europe Online Supplement, pp. 9–12. Burke S. 2001 Missing values, outliers, robust statistics & non-parametric methods. LC GC Europe Online Supplement, pp. 19–24. Kromidas, S. and Kuss, H.-J. (2008) Chromatogramme richtig integrieren und bewerten: ein Praxishandbuch für die HPLC und GC, Weinheim, Wiley-VCH Verlag GmbH.

91

93

4 Automation Systems with Central System Integrator In this chapter, realizations for the different automation concepts are described and validated for specific applications. This includes centralized and decentralized systems, which have been configured for a specific dedicated application (closed systems) or which can be used for different applications with suitable adaptations of the system design (open systems). For each of the four automation concepts, a detailed system description including the robot type, functional subunits, and peripheral devices is given. The system requirements are described considering the process steps. The data evaluation is only briefly described, since this part is independent from the general automation concept. A detailed description regarding data evaluation concepts is given in Chapter 6.

4.1 Centralized Closed Automation System A classic example for a centralized closed automation system has been realized for the determination of the enantiomeric excess of chiral compounds. The chiral samples can involve the compound classes of proteinogenic amino acids, carbonic acids, alcohols, amino alcohols, amino esters, and chiral natural compounds. The process operation is identical for all compound classes; only the used standard and auxiliary compounds have to be changed with respect to the compound class. Thus, the concept of a centralized closed automation system is suitable for this application (see Figure 4.1). 4.1.1 Background and Applicative Scope

The phenomenon of handiness is called chirality. Chiral molecules have characteristic properties since they exist in two mirror-imaged forms and do not have a symmetrical plane. A single molecule is called enantiomer. Enantiomers have identical scalar (this means direction independent) physical properties (e.g., mass, boiling point, melting point, density, refraction, IR spectrum, UV spectrum, etc.), but differ regarding their optical activity [1]. In addition, enantiomers can cause different interactions in combination with other chiral reagents or in a chiral environment, such as an organism [2, 3]. This results in different biological effects or stimuli of these compounds in organisms [4–10]. Chiral compounds are of special importance in pharmacy and medicine, since Automation Solutions for Analytical Measurements: Concepts and Applications, First Edition. Heidi Fleischer and Kerstin Thurow. © 2018 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2018 by Wiley-VCH Verlag GmbH & Co. KGaA.

94

4 Automation Systems with Central System Integrator

Figure 4.1 Schematic visualization of a centralized closed automation system with a central system integrator. (SP: sample preparation, A: analytical measurement, CSI: central system integrator)

Centralized/closed

SP

CSI A

O N

(a)

O

O

H N

H N O

O

(b)

O

O N O

Figure 4.2 Thalidomide and its enantiomers – (a) (R)-(+)-thalidomide, (b) (S)-(−)-thalidomide.

the organism handles enantiomers in different ways. About 80% of all currently available pharmacological important drugs are enantiomers [11]. Due to the increasing sales of enantiopure compounds [12] it is expected that this number will increase to 95% in 2020 [13]. The most well-known compound is Thalidomid (see Figure 4.2), which was used between 1957 and 1961 under the name Contergan as a soporific and tranquilizing agent in Germany and about 46 other countries worldwide [14]. The R enantiomer has a sedative and soporific effect, whereby the S enantiomer has a teratogenic effect and can cause significant malformation of the limbs of a fetus [15–17]. Amino acids are ubiquitous in biological systems as they are the main constituents of proteins. Nineteen out of the 20 known 𝛼-amino acids are chiral compounds. The investigation of the effects of enantiomers on plants [18, 19], animals [20, 21], and humans [22, 23] has far-reaching importance in the field of drug development and drug discovery [24–27]. Natural proteins usually contain the L-enantiomers of the amino acids. Some organisms also contain D amino acids and their derivatives. They are also included in foods such as dairy goods, sour dough (produced by fermentation with lactic acid bacteria), fruits and vegetables, and in a variety of processed food. The effect of these D-amino acids is not fully known today [28]. In the field of neurosciences it has been observed that L-proline can cause amnesia in chicks [29, 30]. D-Proline in contrast causes dissociated convulsions and lethality, although a significant effect regarding amnesia could not be found [31]. Experiments with chicks in the field of stress and brain research showed effects of L-proline on the stress-induced metabolism of dopamine and serotonin [32]. Both enantiomers can cause sedative and hypnotic effects [33]. L-Tryptophan can be used for the treatment of insomnia and depression [34, 35].

4.1 Centralized Closed Automation System

The syntheses and measurement of chiral compounds requires a precise and uniform definition regarding the description of the quantitative composition [36–38]. The ratio of two enantiomers (the amount of the enantiomeric purity) can be defined in different ways depending on the type of analysis. One possibility is the optical purity, which is based on the different optical activity of the enantiomers (Table 4.1, Equation 4.1). This form is preferably used in optical chiral analytical methods. Additional variants for the description of the composition of an enantiomeric mixture are the enantiomeric ratio (Equation 4.2) and the enantiomeric fraction (Equation 4.3). Both values deliver simple information, which has to be interpreted according to the specific application. An often used definition is enantiomeric excess, which can be defined as dimensionless (Table 4.1, Equation 4.4) or in percentage value (Equation 4.5) [39]. In contrast to the enantiomeric excess, the enantiomeric ratio can be directly calculated from the peak areas from chromatograms. Equations 4.6 and 4.7 visualize the correlation between the two values. In this application, the composition of the chiral compounds is defined using the enantiomeric excess ee%.

Table 4.1 Definitions for the description of the chiral composition. Optical purity

p=

[𝛼] × 100 [𝛼0 ]

(4.1)

[𝛼]: Specific rotation of the sample [∘ ] [𝛼 0 ]: Specific rotation of the pure enantiomer [∘ ] Enantiomeric ratio

A B A: Quantity enantiomer 1 (in excess)

er =

(4.2)

B: Quantity enantiomer 2 Enantiomeric fraction

ef =

A A+B

(4.3)

Enantiomeric excess

ee =

(A − B) (A + B)

(4.4)

ee% = Correlation between enantiomeric ratio and enantiomeric excess

Definition of the enantiomeric excess ee% used in this book

(A − B) •100 (A + B)

(4.5)

er =

(1 + ee) (1 − ee)

(4.6)

ee =

(er − 1) (er + 1)

(4.7)

ee% =

(VA − VB ) •100, V + V = const. A B (VA + VB )

VA : Volume of solution with enantiomer 1 VB : Volume of solution with enantiomer 2

(4.8)

95

96

4 Automation Systems with Central System Integrator

The calibration samples for the mass spectrometric determination are prepared according to Equation 4.8 with different volumes of the enantiomers in identical concentrations. Table 4.1 gives a general overview about the definitions and their calculation. In general, no single sensors are available for the determination of chiral compounds. Single sensors have only been developed for specific applications. A microbiosensor on the basis of D-amino acid oxidase (DAAO) has been reported for the in vitro and in vivo determination of serine in rat brains. The goal was the investigation of the D-serine metabolism in the central nervous system, which seems to be connected to neurological and psychiatric disorders (e.g., schizophrenia) as well as the testing of possible active compounds [40]. Classical measuring technologies for the determination of chiral compounds use complex sensor systems such as HPLC, GC, or capillary electrophoresis (CE). HPLC-based methods are widely used for the analyses of food [41, 42], biological samples [43], and pharmaceutical compounds [44]. They usually have quite long analysis times of up to 85 min or more. A method for the simultaneous detection of the enantiomers of 16 proteinogenic amino acids in plants (Arabidopsis thaliana) was reported, which uses a liquid chromatography mass spectrometry (LC-MS) in combination with a triple-quadrupole mass spectrometer with analysis times of 25 min [45, 46]. A GC-MS-based method has been described for the detection of liquid and syrupy dietary saps and juices of plant origin with analyses times about 45 min [47]. Amino acid enantiomers can be detected with CE-, HPLC-, or GC-based methods with analysis times between 15 and 30 min [48]. An overview about different methods for the analysis of D-amino acids in biological samples can be found in [49]. For the fast determination of D-serine within seconds, an in vivo and in vitro enzyme based assay has been described. Hydrogen peroxide is produced using DAAO and oxidized at a platinum microelectrode; the resulting current correlates with the concentration of D-serine [49]. This method can only be applied for the determination of D-amino acids and is not suitable for the determination of the enantiomeric excess. In addition, different methods and the resulting analysis times have been described [49]. Some of these methods are characterized by long analysis times, while some are cost-intensive due to the use of chiral columns, a high consumption of solvents, and/or chiral additives. Often, these methods can only be used for single or a limited number of applications. In contrast, mass spectrometry enables short analysis times and a high degree of automation. Due to the identical masses of both enantiomers, direct determination is not possible. Different methods can be used for the determination of the chiral composition including ion-molecule reactions [50] or the kinetic method based on the collision-induced degradation of diastereomeric complexes [51–53]. The principle of the parallel kinetic resolution was developed by Vedejs and Chen [54] and Finn et al. [55] on the basis of earlier work from Horeau [56]. The principle is based on the different reaction kinetics between different enantiomers and other chiral compounds [54]. In a pre-analytical derivatization step, the enantiomers of the chiral analyte react with two pseudo-enantiomeric mass labeled auxiliaries. The latter have opposite chiral configurations, but differ in their molecular mass. Analytes with a chiral center in the molecule thus form four

4.1 Centralized Closed Automation System

Auxiliary 1 Substrate D + Auxiliary 1

Auxiliary 2 Substrate D + Auxiliary 2

Substrate D

Fast

Slow

+

Molecule mass 1

Molecule mass 2

Auxiliary 1 Substrate L + Auxiliary 1

Auxiliary 2 Substrate L + Auxiliary 2

Substrate L

Slow

Fast

Figure 4.3 Principle of the parallel kinetic resolution. (Redrawn from [55, 57].)

reaction products with different molecular masses (see Figure 4.3). A correlation between the chiral composition of the analyte and the ratio of the two characteristic masses of the four reaction products exists due to the different reaction speeds between the chiral analyte and the pseudo-enantiomeric auxiliary. The Marfey’s reagent is used as a derivatization agent in classical HPLC [58]. Diastereomers are formed, which enable the use of achiral columns. Two variants of the Marfey’s reagent, N 𝛼 -(2,4-dinitro-5-fluorophenyl)-L-valinamide (L-FDVA) and N 𝛼 -(5-fluoro-2,4-dinitrophenyl)-D-leucinamide (D-FDLA), have an opposite chiral configuration and a mass difference of m/z = 14. The combination of these reagents as auxiliaries with the principle of kinetic resolution enables the determination of the chiral composition of proteinogenic amino acids [59–63]. Figure 4.4 visualizes the principle of parallel kinetic resolution for the derivatization of L- and D-proline with the two auxiliaries. M = 395.14

M = 300.09 NO2 F H

O OH

O2N HN

H

HN NH2

NH

H O

HO HN

NH2

H

NH2

O

NaHCO3

+

+ NaF

O2N H O

N

N

O2 N HN

NO2

NO2 F

O OH

H

O2N

O

NO2 H

N H O

HO

O M = 115.06

NO2 N

O2N

NH

M = 409.16

NO2

HO HN

NH2

H O

H O

O2N

HO HN

NH2

H

H O NH2

O

M = 314.10

Figure 4.4 Derivatization of L- and D-proline with two different forms of the Marfey’s reagent.

97

98

4 Automation Systems with Central System Integrator

Due to the derivatization of standard and sample solutions with an equimolar mixture of the two auxiliaries, a new calibration has to be done to increase the precision by using identical auxiliary solutions for samples and the corresponding calibration. Thus, the first sample preparation step is the dilution of a stock solution of each enantiomer and the preparation of five calibration standards (enantiomeric excess: −100ee%, −50ee%, 0ee%, +50ee%, and +100ee%). The stock solutions are prepared manually from the solid materials and can be stored over a longer time at −18 ∘ C. The following step includes the preparation of an equimolar solution of the two pseudo-enantiomeric mass labeled auxiliaries (the solids are weighted manually). The auxiliary solution is pipetted to a 96 well microplate together with sodium bicarbonate for the derivatization. The closed microplate is derivatized on a thermomixer for 1 h. A suitable dilution has to be realized prior to the analytical measurement using TOF-MS [60–62]. 4.1.2 Automation Goals

All processes except the preparation of the stock solution are candidates for automation. The proteinogenic amino acids are dissolved in hydrochloric acid and neutralized, whereby the pH is controlled manually. Since the stock solutions can be stored for a long time, the automation of this subprocess does not have a high priority. Another manual process is the weighing of the pseudo-enantiomeric mass labeled auxiliaries. Exact weighing is required since the ratio of both compounds has a direct influence on the calibration curve (sensitivity). The following steps can be automated very well and include the dilution of the solution and the pipetting of the calibration standards. The preparation of the auxiliary solution has four substeps for the mixing of the auxiliaries and the rinsing of the used vials and pipettes to ensure a complete transfer of the compounds (equimolar mixture). The following subprocess includes pipetting of the reagents (chiral substrates, auxiliaries, and sodium bicarbonate) onto a 96 well microplate. This process can be done in parallel using multi-channel pipettes. The derivatization is done on a thermomixer at constant time and temperature. The reaction is quenched due to the addition of hydrochloric acid and again homogenized. A final dilution is the last step before the analytical measurement. These steps can also be realized in parallel mode. The analytical measurement is followed by data evaluation and a final storage or disposal of the samples. Figure 4.5 shows an overview of the manual process workflow [64]. The transportation tasks include the initial transport for providing the amino acid stock solutions, the weighted auxiliaries, reagents, solvents, and labware. In addition, the transport of the sample solutions from the automation system to the thermomixer, the transport of the measuring solutions to the TOF-MS, and the transport of the samples to the sample storage or sample disposal have to be realized. Automation would enable a reduction of human errors such as pipetting or dilution errors as well as mistakes due to accidental change of the sample vials and microplate wells. In addition, the precision during the preparation of the equimolar auxiliary mixture can be increased since a robot-based and always identical

4.1 Centralized Closed Automation System Proteinogenic amino acids

Preparation of stock solutions

Dilution of standard and sample solutions

Pipetting of calibration solutions

Preparation of auxiliary solutions

Pipetting of reagents

Sample storage or disposal Derivatization

Reaction quenching

Homogenization

Material flow Information flow

Dilution

Measurement using mass spectrometry Data evaluation

Figure 4.5 Process workflow for the determination of chiral amino acids using TOF-MS.

movement ensures a higher consistency of the mixture composition. If the initial transport can be ensured, 24/7 operation is possible. 4.1.3 System Design

The designed automation system includes the functional subunits “sample storage,” “liquid handling,” “sample treatment,” and “analytics” in addition to the central system integrator. Figure 4.6 shows the realized automation system. System integrator: The connection between the different functional units as well as the transport of the materials and samples is realized using a central system integrator. An ORCA laboratory robot (Optimized Robot for Chemical Analysis, Beckman Coulter, Krefeld) is integrated for this task. The ORCA is an articulated robot with six degrees of freedom (DOF) mounted on a rail with 3 m length. The robot has a range of 58.4 cm and a precision of ±0.25 mm. The payload is 0.5 kg in continuous operation, whereby temporary loads of up to 2.5 kg are possible. The gripper is designed for handling microplates. In the delivery stations, the microplates are oriented with the small side toward the robot, whereas on the destination positions of the liquid handler (automated labware positioner, ALP), an orientation with the long side is required. Thus, a regrip station is used, which allows the robot a regripping to enable handling of the microplates in both orientations (small and long side of the plate). The laboratory robot is controlled with the device software ORCA NT, which was integrated into the SAMI scheduling software (Beckman Coulter, Krefeld) to enable a maximum of automation (see Section 4.1.5). Sample storage: All standard and sample solutions, reagents, and solvents required for the measurement are provided in glass vials or microplates. The use of microplates is preferred since they show a high degree of standardization and can be used in different automation devices from various vendors. The chemicals and the required labware are stored in dedicated storage systems (hotels) in the form of racks with 2 × 8 or 3 × 18 shelf spaces in the microplate format, which are positioned on the automation system. Liquid handling: The preparation of the calibration standards, the dosage of the reagents and additives as well as the final dilution are realized on an integrated liquid handler Biomek 2000 (Beckman Coulter, Krefeld), which enables the parallel filling of 96 well microplates using an eightfold pipetting tool in the volume range between 5 and 200 μl (MP200). The handling of single vials is also possible; a single pipetting tool in the volume range between 50 and 1000 μl is

99

100

4 Automation Systems with Central System Integrator

5

3 2 8

1

6 7

(a) 1 2

3

2

4 9 5 6 7 8

(b)

Figure 4.6 Fully automated system for the determination of chiral compounds ((a) front view, (b) top view, 1: liquid handler Biomek 2000, 2: storage system for labware, reagents, and samples, 3: multi-parallel reaction system HPMR 50-96, 4: thermomixer, 5: ORCA laboratory robot (central system integrator), 6: HPLC system with autosampler, 7: ESI-TOF-MS, 8: computer for control and data evaluation, 9: regrip station).

used here. Changeable tips are used to avoid cross-contaminations. The use of closed mats for covering the microplates requires high pressure while applying and high force for removing them. Thus, this process step cannot be realized by the used robot. In addition, the injection needles of the HPLC autosampler cannot pierce the mats; thus the analytical measurement has to be done with open microplates. Depending on the type of solvent, evaporation and concentration of the samples can occur. This is not a critical point for the realized process, since no changes in the chiral composition of the samples have been discovered due to

4.1 Centralized Closed Automation System

evaporation. Different possibilities exist for avoiding evaporation. In the easiest way, the microplates can be processed with covers. A suitable station for handling the covers (lifting and attaching the covers) has been integrated for this case. One disadvantage is the incomplete sealing of the plates, when covers are used. Evaporation effects can thus be reduced, but not eliminated. Alternatively, the integration of sealers and peelers is possible. These automated devices enable the application and a sealing film (sealer) or the removing of the film (peeler). Their integration into the automation system is possible, but involves considerable costs for the both devices. Using organic solvents may cause residues of the adhesive films at the injection needle of the autosampler, which may damage the needle or the autosampler robotic drive. A third way for minimizing the evaporation effects is the integration of a reusable cover made from alumina foil. This cover can be handled by the liquid handler and can also be pierced by the injection needle of the HPLC autosampler [65]. The cover consists of a two-piece frame in microplate format; the alumina foil is clamped between the two parts of the frame (see Figure 4.7). After use, the alumina foil can be changed easily without additional tools. A suitable number of frames have to be available in the system with respect to the number of samples to be processed. Sample treatment: The treatment of the samples includes a chemical derivatization reaction. This reaction is realized in a multi-parallel reaction system HPMR 50–96 [66, 67]. The system enables tempering in the range between 0 and 100 ∘ C. The reaction solutions in the single wells of the microplates are electromagnetically stirred to ensure optimal mixing. This requires the delivery of microplates with one stirring plate per well. The stirrers are supplied manually via a filling station; an automation of this procedure is possible with the integration of an additional specialized device. The removal of the stirring plates is done at the end of the complete process during the cleaning steps.

(a)

(c)

(b)

(d)

Figure 4.7 Reusable cover frames for microplates (a) CAD design, (b) CAD explosion view, (c) realized cover, and (d) cover after use with HPLC system.

101

102

4 Automation Systems with Central System Integrator

Pressurization with hydrogen, carbon monoxide, or other gases with pressures up to 50 bar is possible on demand [66–68]. Alternatively, the derivatization can also be done using the thermomixer Comfort (Eppendorf, Hamburg). This device enables tempering up to 100 ∘ C. Cooling with the integrated Peltier element is possible up to 13 ∘ C (reaction vials) or 10 ∘ C (microplates) under room temperature. Sample homogenization is realized by shaking at frequencies between 300 and 1500 rpm. Analytics: A LC-TOF mass spectrometer G1969A (Agilent Technologies, Waldbronn) is used for the analytical measurements of the enantiomeric excess. The samples are injected with the autosampler after they have been transferred by the robot. The overlapping injection mode is used to achieve maximum throughput. In this mode, the following sample is already prepared for injection during the run of the forgoing sample. The data is automatically evaluated after each single measurement in online mode. 4.1.4 Process Description

All necessary solutions and labware are provided manually to the automation system. During automated sample processing, standard solutions, reagents, solvents, and the labware are transported by the central robot from the storage positions to the final positions on the liquid handler deck. This includes the five calibration solutions with −100ee%, −50ee%, 0ee%, +50ee%, and +100ee%, the equimolar auxiliary solution (L-FDVA and D-FDLA), as well as the sample solutions. Subsequently, the reaction solutions are prepared in a 96 well microplate by sequential addition of the chiral substances (amino acids), pseudo-enantiomeric mass tagged auxiliaries (equimolar mixture with L-FDVA and D-FDLA), and sodium hydrogen carbonate (NaHCO3 ). Figure 4.8 gives an overview about the 1

2

3

4

5

6

7

8

9

10

11

12

A

A1

A2

A3

A4

A5

A6

A7

A8

A9

A10

A11

A12

B

B1

B2

B3

B4

B5

B6

B7

B8

B9

B10

B11

B12

C

C1

C2

C3

C4

C5

C6

C7

C8

C9

C10

C11

C12

D

D1

D2

D3

D4

D5

D6

D7

D8

D9

D10

D11

D12

E

E1

E2

E3

E4

E5

E6

E7

E8

E9

E10

E11

E12

F

F1

F2

F3

F4

F5

F6

F7

F8

F9

F10

F11

F12

G

G1

G2

G3

G4

G5

G6

G7

G8

G9

G10

G11

G12

H

H1

H2

H3

H4

H5

H6

H7

H8

H9

H10

H11

H12

Figure 4.8 Microplate layout with 15 calibration standards (dark gray), one blank (white), and maximum 80 samples (light gray).

4.1 Centralized Closed Automation System

plate layout of a typical 96 well plate. The prepared solutions are transferred to the multi-parallel reaction system HPMR 50-96, where they are stirred for a defined time of 1 h at 20 ∘ C. After the derivatization is completed, the solution is quenched with concentrated hydrochloric acid (HCl) on the liquid handler deck. For optimal mixing, the microplate is transported back to the HPMR and the plate wells and the reaction mixture are homogenized with electromagnetically stirring. Alternatively, the derivatization and homogenization can be performed on the integrated thermomixer. Since the homogenization in this case is realized by shaking, stirring plates are not required. The plates with the final solutions are transported by the system integrator into the HPLC autosampler. Data acquisition of the mass spectrometric measurements is realized with the device software “MassHunter Data Acquisition” followed by the extraction of the required masses by the software module “MassHunter Qualification.” The results are stored in an Excel report file. Once the measurement is completed, the file is automatically uploaded to a webserver. The web application “Analytical Data Evaluation” (ADE) realizes the final data evaluation including the calculation of the calibration curve and the enantiomeric excesses of the samples. The data is finally visualized in 2D and 3D diagrams (see Chapter 6). Figure 4.9 shows the process flow chart with all required process steps. 4.1.5 Control of the Automation Process

A process control software or a process control system is required to ensure 24/7 operation with high throughput. The used software “SAMI Version 3.6” (SAGIAN Automated Method Development Interface) (Beckman Coulter, Krefeld) has an open software structure enabling the integration of peripheral devices. With a driver program (middle ware), the commands from the supervisory control software (SAMI) are translated for the peripheral devices and reverse. The SAMI software enables the execution and supervision of the methods using the “SAMI Run Time” module. Using the “SAMI Method Editor,” the operator can program methods and optimize specific single tasks using a graphical interface (see Figure 4.10). Access to the control software of the integrated devices ensures the greatest possible flexibility. The liquid handler is integrated into the configuration as labware, but can also be used independently. All movements are included in the scheduling and optimized. The chronological procedure of all different steps is calculated and set by the software. All process steps on the liquid handler are controlled by the “BioWorks Method Editor” (BioWorks Edit Version 3.2) and stored in a method (see Figure 4.11). This editor enables, for example, the individual adjustment of the aspirating and dispensing velocities in pipetting steps or the immersion depth of the pipette tips. Functions such as Prewet, TipTouch, and Blowout are also available. This provides the possibility to work with different solvents and matrices on the system. Besides the basic configurations of the platform, all other object movements and dosage steps are stored in a BioWorks method. The pipetting instructions have to be separately created for each pipetting procedure (see Figure 4.12). The execution

103

104

4 Automation Systems with Central System Integrator

Start Initial transport Supply of samples, stock solutions, reagents, solvents, and labware

Transport Storage system → Liquid handler (samples, stock solutions, reagents, solvents, and labware)

Pipetting Calibration standards, sample solutions, auxiliary solution, and sodium bicarbonate

Transport Liquid handler → Reactor (microtiter plate)

Derivatization 1 h, 20 °C / 68 °F, 250 rpm

Transport Reactor → Liquid handler (microtiter plate)

Opening Remove cover from microtiter plate

Transport Liquid handler → Reactor (microtiter plate)

Homogenization 1 min, 250 rpm

Transport Reactor → Liquid handler (microtiter plate)

Opening Remove cover from microtiter plate

Transport Liquid handler → Autosampler (microtiter plate)

Closing Put cover on microtiter plate

Pipetting Reaction quenching using hydrochloric acid

Closing Put cover on microtiter plate

Pipetting Sample dilution

Closing Put cover on microtiter plate

Analysis Measurements using ESI-TOF-MS

Transport Autosampler → Storage system (microtiter plate) Liquid handler → Storage system (chemicals and labware used)

Final transport Storage/disposal/cleaning of samples, stock solutions, reagents, solvents, and labware

End

Figure 4.9 Process flow chart for the automated determination of chiral amino acids.

of the instructions is realized column by column with an eightfold pipetting head. The finally created method can be integrated into the “SAMI Method Editor” and started from here. 4.1.6 Evaluation of the Automation System

The described automation system performs all subprocesses of pre-, intra-, and post-sensoric selectivity. Sample preparation including the pipetting steps requires 4 min; 60 min are necessary for the derivatization. The mass

4.1 Centralized Closed Automation System

Figure 4.10 Example for an automated method in “SAMI Method Editor.”

Figure 4.11 Example for an automated liquid handling method in “BioWorks Method Editor.”

spectrometric acquisition time is set to 30 s per sample. Using the mode of overlapping injection for the HPLC autosampler, a processing time of 132 min is required for a complete 96 well plate (including injection times). In online mode, extraction of the data files containing the processed measuring data is started after measurement completion. Thus, the processing time (several seconds) is not relevant, as the next sample can already be measured during this time. The total processing time for one 96 well microplate is about 200 min. The time limiting factor is the mass spectrometric analysis; up to 11 microplates (1056 samples) can be processed per day (see Figure 4.13) [61]. The number of samples with an unknown enantiomeric excess can be calculated from the difference of the maximum possible samples and the number of calibration samples. Up to 7392 samples can be processed per week, which is in accordance to the requirements of high throughput screening [69].

105

106

4 Automation Systems with Central System Integrator

Figure 4.12 Example for an automated pipetting procedure.

Liquid handler Approx. 4 min

multi-parallel reactor 60 min

Liquid handler Approx. 4 min

TOF-MS

Software (web application)

132 min

Sample Derivatization Sample post Analytical preparation processing measurement Pre-sensory selectivity

Intra-sensory selectivity

Online

Approx. 200 min / MTP 11 MTP / day 1.056 samples / day 7.392 samples / week

Data evaluation visualization Post-sensory selectivity

Figure 4.13 Time requirements for an automated screening of chiral compounds.

4.1 Centralized Closed Automation System

× 104

2

408.15

394.13

1.5 1 0.5 0 375

380

385

390 395 400 405 410 415 420 Counts versus mass-to-charge (m/z)

425

430

Figure 4.14 Mass spectrum of the derivatives of proline.

Intensity ratio Int(394.13) / Int(408.15)

Extensive validation experiments have been performed to demonstrate the performance of the automation system. Figure 4.14 shows exemplarily a mass spectrum with the two characteristic masses of the derivatives of proline (ion mass M1 = 394.13 and M2 = 408.15). For the model compound proline (M = 115.13 g/mol), which was used in a concentration of 1 mmol/l for the derivatization, the coefficient of variation (CV) for an acceptable and expected between-laboratory precision is calculated with V K = 7.83%. The CV for the acceptable and expected repeatability and the within-laboratory precision is thus between 3.92% and 5.22%. To determine repeatability, 25 replicate samples were prepared with five defined values of enantiomeric excess (between −100ee% and +100ee%). The CV for the measurements with different enantiomeric excess varies between V K = 1.92% for −100ee% and V K = 2.91% for +100ee%, and is thus under the calculated and acceptable values according to Horwitz (3.92% < V K < 5.22%). The results are visualized in Figure 4.15 [61]. y = 0.64e0.005x R2 = 0.9996

1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

−100

−50

0

50

Enantiomeric excess L-proline [ee%]

Figure 4.15 Repeatability results for the amino acid proline.

100

107

108

4 Automation Systems with Central System Integrator

For the determination of within-laboratory precision, repeated measurements were done on five consecutive days. Every day, 15 samples with 5 defined enantiomeric excess and 3 replicate samples were prepared and measured. The method stability was determined with a sample set with five enantiomeric excess and three replicate samples. The sample solutions were distributed to five microvials after the derivatization with 100 μl each, thus five complete sets of samples were available for the measurements. One set of samples were measured immediately; the four remaining sets were stored at −18 ∘ C. Every day, one set was defrosted at room temperature and measured. The coefficients of variation were below or slightly above the expected values on all days, according to Horwitz (3.92% < V K < 5.22%) [61]. The measurement precision was determined with 5 defined enantiomeric excess samples of the chiral compounds and measured 10 times. The coefficients of variation of the directly measured intensities of the m/z ratio of the diastereomers fulfill maximum V K = 1.94% of the requirement (V K ≤ 2%). The detection limit was determined at LODPro = 0.01 mmol/l and the limit of quantification LOQPro = 0.02 mmol/l. This measuring method is characterized by short processing times per sample using 96 well microplates. Time advantages of up to 82–96% can be achieved compared to classical analytical methods (e.g., achiral or chiral HPLC-MS or chiral GC-FID). Table 4.2 compares the required processing times of the introduced method and classical analytical methods [60, 61, 64].

Table 4.2 Processing time of different methods for the determination of chiral compounds [60, 61, 64]. Method

Compound

Derivatization (h)

Chiral MS

All

1.00

1.38

3.33

—

Achiral HPLC-TOF/MS

Proline

2.00

10.00

18.00

82

Chiral GC-FID

Serine

3.50

48.60

81.26

96

Glutamic acid

3.50

48.60

81.26

96

Borneol

—

15.00

24.00

87

Menthol

—

15.00

24.00

87

Tryptophan

—

25.00

40.00

92

Histidin

—

10.00

16.00

80

Phenylglycinol

—

10.00

16.00

80

N-Boc-HPME

—

15.00

24.00

87

Chiral HPLC-DAD

Achiral HPLC-MSD

Measuring time per sample (min)

Total processing time per 96 well plate (h)

Time reduction Chiral MS (%)

4.2 Centralized Open Automation System

4.2 Centralized Open Automation System A sample preparation system is described as an example for a centralized open automation system, which optionally allows the automated sample preparation of wood samples for the analytical determination of mercury as well as the sample preparation and analytics of dental materials or their extracts for the determination of the material composition or the eluted rest monomers. The methods used in the system have different process sequences and use different labware. Thus, the concept of a centralized open automation system is preferred for the automated execution of the application (Figure 4.16). 4.2.1 Background and Applicative Scope 4.2.1.1 Determination of Mercury in Waste Wood

The transition metal mercury is the only known existing metal, which has a liquid form under normal conditions. In daily life, mercury is used in thermometers or in amalgam-based tooth filling materials. In the past, mercury was involved in an important commercial application: impregnation of wood materials with the sublimate mercury(II) chloride [70]. This technique is called kyanization and was patented in 1832. Since 1838, railroad ties made of wood were impregnated in Germany and other European countries [71]. Furthermore, kyanization was used for the impregnation of telegraph poles [72], construction timbers, and for dry rod clean up in buildings [71]. Mercury(II) chloride provides a potent fungicidal action and a low leachability from wood materials [73]; however, the use has always been controversial due to the potential health impacts [71, 74, 75]. Besides inorganic mercury compounds, organically bound mercury was of major importance [76, 77], which was also used for wood impregnation. The main application fields are fungicides in seed mordants [78, 79] and fruit cultivation [80, 81], additives in paints to prevent mold [82–85], and conservation of textiles [77, 86]. The use of mercury compounds as disinfectants and preserving agents in pharmaceutical and ointment production has declined over time [87–90]. Due to the direct environmental impact of mercury, the use of organic mercury compounds in agriculture was terminated in 1970 [78]. Figure 4.16 Schematic visualization of a centralized open automation system with a central system integrator (SP 1,2,…,X: sample preparation, A 1,2,…,X: analytical measurement, CSI: central system integrator).

Centralized/open SP 2 SP X

SP 1 CSI

AX

A1 A2

109

110

4 Automation Systems with Central System Integrator

Furthermore, gold and silver mining [91–95], metal extraction and processing [96, 97] as well as increasing municipal waste [98, 99] result in an enrichment of mercury in the environment. Additionally, burning of fossil fuels also may cause mercury enrichment in the surrounding environment [100]. The use of mercury compounds in the field of wood protection was decreased, terminated, and prohibited in Germany in 1993 [101, 102]. Nevertheless, pollutants from construction and demolition waste may result in undesirable environmental effects [103, 104]. The accumulation of mercury can be monitored to a high degree in soil, wetlands, and water [105–108]. From there, mercury may reach the human food chain [109, 110] and cause intoxications in humans and animals with damage to liver, kidneys, and brain as well as the central nervous system [90, 111–120]. Mercury shows biological activity already in very low concentrations (approximately 10−5 %), whereby the toxicity depends on the binding form of mercury and the kind of resorption through the organism [113]. Alfred Stock was one of the first scientists who determined small amounts of mercury in various matrices [121–128]. He determined metallic mercury, which was electrolytically generated at a copper wire and subsequently distilled. The mercury droplet produced was measured with a microscope. This technique enabled the determination of 0.1 up to 1000 μg mercury [122, 129, 130]. In 1952, Theden and Engelbrecht investigated the depth of penetration of mercury(II) chloride in wood. Their technique using an ammonium sulfide solution provides a limit of detection of 0.14% mercury compound [131]. Another technique of the same authors is based on copper iodide containing paper for visualization of the active compounds. The detection limit here was 0.006% mercury compound in wood [131]. The measurement methods for mercury determination were further developed and the limit of detections significantly decreased. Widespread and critical overviews related to a wide variety of techniques are given by Reimers et al. [132] and Chilov [130]. The most important techniques include thin-layer, paper, ion-exchange, and gas chromatography as well as spectroscopic methods such as atomic fluorescence, atomic absorption, and atomic emission spectroscopy. The kinds of matrices involve fish, seafood, honey, soil, water, plants, and coal [133–141]. Hamilton et al. determined heavy metal contamination of fish and seafood, water, plants, and sediments in a protected nature reserve (Rockefeller Wildlife Refuge) as well as in industrially polluted river water. The detection limits were determined at 0.02 mg/l using inductively coupled plasma optical emission spectroscopy (ICP-OES) [142]. Ure reviews the mercury determination using graphite furnace atomic absorption spectroscopy (GF-AAS) as well as atomic fluorescence spectroscopy (AFS) for various kinds of samples [143]. For mercury enrichment in the AAS and AFS, the mercury cold vapor method (CV) is often used (also CV-AAS, CV-AFS), which is regulated in the United States in the EPA standard 1631 [144] and in Germany in the DIN EN 1483 standard [145]. Mercury cold vapor is also commonly used in the ICP-OES [146, 147]. A further technique for mercury enrichment is amalgamation [103, 148, 149]. In the last years, a new plasma activation source based on microwaves has been established (microwave induced plasma, MIP), which enables economical gas consumption [150]. The coupling with classical chromatographic techniques such as gas and liquid chromatography is used in speciation analysis. Janzen et al. described the

4.2 Centralized Open Automation System

use of coupling of a MIP-OES (Microwave Induced Plasma Optical Emission Spectroscopy) system with a gas chromatograph for the speciation of mercury, tin, and lead compounds. A limit of detection of 1.9 pg mercury (absolute) could be reached using this technique [151]. Besides spectroscopic techniques, mass spectrometry with inductively coupled plasma has been established in the mercury analysis [152–158]. Broekaert compared in [159] the limit of detections for various elements in water achieved by ICP-OES and inductively coupled plasma mass spectrometry (ICP-MS). Therefore, the limit of detection for mercury determination using ICP-OES was 25 μg/l, using ICP-MS with mercury single-element optimization was 0.02 μg/l, and with multi-element optimization was 0.1 μg/l [159]. A multi-element method for the investigation of food using a double focusing sector field ICP-MS (DF-SF-ICP-MS) was described by Khouzam et al. A detection limit of 8.5 μg/kg could be achieved using this technique [160]. In speciation analysis using ICP-MS, coupling with chromatographic or electrophoretic separation techniques is possible [161]. The determination of mercury in wood materials is of general importance in environmental monitoring, materials research, and waste disposal [162–164]. The German waste wood regulation prescribes the proper disposal of wood waste, whereby the kind of disposal depends on the grade of contamination. The permitted limit value of mercury in dry wood was set to 0.4 mg/kg [104]. Concentration limits for various matrices can also be found in the reports from the US Environmental Protection Agency [165, 166]. The EPA mercury toxicity limit for solid waste is 0.2 mg/kg [167, 168]. The sample preparation generally depends on the kind of sample, physical state, and the composition of the matrix. If necessary, liquid or gaseous samples may be stabilized, cleaned up by filtration or centrifugation, enriched or diluted. The determination of total mercury content in solid samples or samples with a high content of solids requires a complete digestion and transition into the liquid or gaseous state. Microwave-assisted acid digestion is an established sample preparation technique, which is suitable for their use in trace analysis [169, 170]. In the literature, acid digestion procedures are described for various sample materials including dust, polymers, crude oil, and plants [171–174]. For volatile elements such as mercury, microwave-assisted acid digestion is also suitable, since the procedure is performed in pressure tight vessels [113, 175–182]. Methods for digestion of waste and waste wood materials are described in the international regulations ISO 12846, ISO 11466, and ISO 16772 [183–185], the US regulation EPA 7471B [186], and the German regulations DIN EN 13657 and DIN 52161-1 [187, 188]. Sample weights in the range of 0.1–0.8 g and volumes of the digestion solution in the range of 6–10 ml are used [187, 188]. Further methods published in the literature use similar weights and volumes [154, 155, 189]. Subsequent to the digestion, a dilution of the acidic samples to a concentration suitable for the following analysis is required. 4.2.1.2 Determination of Methacrylates in Dental Materials

The challenge of producing a natural-looking tooth reconstruction with optimal mechanical, biological, and optical properties results in the development of a wide variety of synthetic resins in modern chemistry. As early as 1937

111

112

4 Automation Systems with Central System Integrator

methacrylate-based resins were introduced in dental medicine [190]. These dental materials did not completely provide the required properties. There were significant differences in the coefficients of thermal expansion between tooth and resin. Furthermore, the adhesive properties and the color stability were not optimal [190, 191]. In the following years, new techniques in the pretreatment of the enamel surface [192] and new filling materials were developed. These included silica-reinforced polymers [193] and composite resins of different compositions [194–196]. Today, patients prefer dental methacrylate-based materials in natural colors [197], which are often used in general dental practices. Dental materials can be classified into four main groups: (conventional) glass-ionomer cement (GIC), resin-based composites (RBC), resin-modified glass-ionomer cement (RMGIC), and polyacid-modified resinous composites (also called compomer) [198]. In a multitude of studies, differences in physical properties and in shelf-life between classical and modern dental materials were presented [199–202]. In contrast to classical materials such as gold or amalgam, composite filling materials wear off much faster and age in an oral environment. Furthermore, oxidation and hydrolytic degradation result in a discoloration of the tooth filling [203]. These processes are supported by leaching of material compounds by acidic components of saliva, acidic food, exposure to hot food, or drinks as well as by alcoholic drinks. As a result, residual monomers [204–207] and inorganic additives [208] of the composites are eluted. This decreases the biocompatibility of the material and results in undesirable side effects [209, 210]. Commercially available methacrylate-based dental materials (also called composite fillings) consist of an organic matrix and embedded inorganic filling materials [211]. In general, the organic matrix contains the di- or tri-ester of methacrylic acid such as methyl acrylate (MA) and methyl methacrylate (MMA). Further components are monomers, which can be classified into two main groups: base monomers and co-monomers with a lower molecular weight [194, 196]. Base monomers such as bisphenol A dimethyl acrylate (BisDMA), bisphenol A glycidyl methyl acrylate (BisGMA), and urethane dimethacrylate (UDMA) are hardenable binders for reinforcing filling materials. These compounds have a higher molecular mass and a higher viscosity. To reduce the viscosity, co-monomers such as 2-hydroxyethyl methacrylate (HEMA) and triethylene glycol dimethacrylate (TEGDMA) with lower molecular masses and viscosities are added [194, 196, 211]. The base monomer BisGMA and the co-monomer TEGDMA are included in the majority of commercially available composite materials [194, 196]. An overview about basic components of dental composite materials and their properties is given in Table 4.3. The principle of manufacturing composite materials is based on photo polymerization using ultraviolet (UV) light and is performed in three steps: initiation, chain propagation, and chain termination [211, 213]. Further components of dental filling materials are inhibitors for photo polymerization, UV stabilizers, dyes, and pigments as well as inorganic filling materials such as glass, glass ceramics, silicates, and silicon dioxide [194, 207, 210]. Figure 4.17 shows selected components of dental composite materials. Besides the advantages of composites over classical dental materials such as esthetic colors and good processing characteristics, the biological compatibility

4.2 Centralized Open Automation System

Table 4.3 Basic components of dental composite materials [212]. Compound

Abbreviation

Molecular formula

Molecular weight

Methyl acrylate

MA

C4 H6 O2

86.09

Methyl methacrylate

MMA

C5 H8 O2

100.12

Bisphenol A dimethyl acrylate

BisDMA

C23 H24 O4

364.43

Bisphenol A glycidyl methyl acrylate

BisGMA

C29 H36 O8

512.59

Urethane dimethacrylate

UDMA

C23 H38 N2 O8

470.56

2-Hydroxyethyl methacrylate

HEMA

C5 H8 O2

130.14

Triethylen glycol dimethacrylate

TEGDMA

C14 H22 O6

286.32

H3C CH3 O H2C

O

O O

CH3

H2C

CH2

O

CH3

BisDMA

OH

O

H3C CH3

CH3

O

(a)

H N

CH3 CH3 R

O

R

UDMA

N H

O

O

O

O

O

CH2 O

CH3

TEGDMA O

O

CH3

O

O

(b)

CH3

OH

O

CH2

O

BisGMA

O CH3

O

OH

O H2C

H2C

O O

CH3

O

O H2C

HEMA

CH3

H2C

OCH3

CH2

H2C

CH3 CH3

O

(c)

MA

MMA

Figure 4.17 Selected components of dental composite materials, (a) base monomers, (b) co-monomers, and (c) organic matrix. (Redrawn from [212].)

of the materials components and their effects on the human organism are often controversially discussed. It has been reported that HEMA is one of the most common allergens of patients and dental personnel [206]. Some composite-based dental restorative materials may cause an inflammation of the pulp (pulpitis) [214–217], irritations in oral tissues, stimulation of bacteria growth, or trigger allergies [210]. Furthermore, cytotoxic, genotoxic, mutagenic, and estrogenic effects were observed [204, 209, 218–222]. The leaching effect of components from the hardened composites depends on several factors. This includes the conversion rate of the monomers in the polymerization, which is usually about 35–77% [223]. The chemical composition of the leaching solvent as well as the size and the chemical properties of the leachates and extractables influences the leaching effect. Smaller molecules show a higher mobility. For this reason, a higher elution rate is expected than for larger and bulkier molecules [204]. For identification and quantification of leachable residues of dental composite materials after polymerization various analytical methods have been developed. Spahl et al. provided methods using GC-MS and HPLC particle beam mass spectrometry (HPLC-PB-MS) for the determination of water and methanol extracts of light-curing hybrid-type composite resins [223]. Behrend et al. described a

113

114

4 Automation Systems with Central System Integrator

method for the determination of residual monomers using thermal desorption and GC-MS [224]. Rogalewicz et al. presented two methods using HPLC-MS and HPLC diode array detection (DAD) for the determination of RMGIC [225]. The analysis of base monomers using GC-MS is difficult due to their low volatility and decomposition processes in the injector at high temperatures [198]. It is only possible to detect the decomposition products using GC [218, 225]. For testing the automation concept using a central system integrator, two methods for the automated determination of the composition of modern dental materials were used: a HPLC-MS based method for the determination of the base monomers TEGDMA, UDMA, BisGMA, and BisDMA and a GC-MS based method for the determination of MA, MMA, HEMA, TEGDMA [212, 226]. 4.2.2 Automation Goals 4.2.2.1 Determination of Mercury in Waste Wood

In general, all subprocesses of this application can be fully automated. Due to additional costs for special automated devices, some subprocesses such as sample drying, crushing, milling, and weighting are currently performed manually. The microwave digestion procedure can also be automated, but the device was not integrated into the automation system due to safety issues. Highly toxic and corrosive vapors are generated, which require an additional exhausting system to prevent damage to the sensitive mechanical robot parts. The analytical method using CV-ICP-OES enables very low detection limits for mercury of 3.41 μg/kg [227, 228]. The determination of mercury using ICP-MS also provides a very low detection limit of 2.29 μg/kg for mercury [229]. The simultaneous determination of several elements is an advantage compared to mercury cold vapor methods, which are single-element methods. Therefore, this method is well suited for the integration into a system with automated sample preparation [230], sample transport, and analysis. A miniaturized version of the original ICP-MS method was developed for economic and efficient operation with the automation system presented [231]. This increases the throughput and reduces the cost for reagents and for the disposal of sample residues. Figure 4.18 gives an overview of the general process flow with the included subprocesses [64]. Wood sample

Drying, roughly crashing, fine milling, weighing

Pipetting of reagents and internal standard

Data evaluation Measurement using ICP-MS Microwave digestion

Sample storage or disposal

Dilution reformatting Measurement using ICP-OES

Material flow Information flow

Data evaluation

Figure 4.18 Process workflow for the determination of mercury in wood samples using ICP-MS or ICP-OES.

4.2 Centralized Open Automation System

The subprocess “liquid handling” includes pipetting of the standard solutions, digestion acid addition to the solid sample, and the internal standard addition to the sample solutions. The subprocesses “dilution” and “reformatting” combine the final dilution of the digestion solution to an adequate acid concentration (maximum 8%, v/v) and the transfer of the sample solutions from the microwave digestion vessel in vials, which are suitable for the autosampler of the measurement instrument (vol. 8, 15, 30, or 50 ml). The subprocess “analysis” using ICP-MS is characterized by short analysis times with the use of multi-element methods. The subprocess “transport” involves the initial transport to supply the ground and weighed wood samples, reagents, solvents, and labware. Further transport tasks are the transfer of samples, chemicals, and labware between the various stations of the automation system and are realized by the central system integrator. The final transport realizes sample storage and/or disposal. The automated process reduces pipetting errors, unintentional contamination of the samples or digestion solutions, dilution errors using measuring flasks (e.g., reading error) or errors due to accidental transposition of the sample vessels. The miniaturization of the process leads to a reduction of the required sample amount, a reduction in the volumes of digestion acid, internal standard, and diluting solution. Furthermore, sample containers with smaller sample volumes can be used, which reduces space requirements in the automation system. In addition, higher operating safety is achieved by protecting the human operator from potential hazards (e.g., formation of nitrous gases). 4.2.2.2 Determination of Methacrylates in Dental Materials

Depending on the consistency of the starting material, the samples are pipetted or weighted. This step is carried out manually due to the low viscosity of the starting materials. The following subprocess “dilution” is automated and the samples are dissolved in an organic solvent. The subprocess “vaporization” is realized for methacrylates eluted in extraction solvents. Last step of the sample preparation is the final dilution for the subsequent analysis. The following subprocesses “analysis” and “sample storage/sample disposal” are identical for both types of samples and have also been automated. Figure 4.19 illustrates the overall process and its subprocesses [64].

Non-polymerized material

Data evaluation Pipetting, weighing

Measurement using HPLC-MS

Dissolve sample

Sample storage or disposal

Dilution

Extraction solution Add solvent

Measurement using GC-MS

Vaporization using nitrogen

Material flow Information flow

Data evaluation

Figure 4.19 Process workflow for the determination of dental materials using HPLC-MS and GC-MS.

115

116

4 Automation Systems with Central System Integrator

In addition to the initial transport (supply of sample material or extraction solutions, solvents, and labware), the transportation tasks involve the sample transport between the different stations on the automation system. The final transport realizes the sample storage and/or disposal. Automating this process prevents and reduces human errors as described above. 4.2.3 System Design

The designed automation system includes the functional subunits “sample storage,” “liquid handling,” “sample treatment,” “sample handling,” “analytics,” and the central system integrator [230, 232–234]. Figure 4.20 shows the realized system. System integrator: A central system integrator is used for the connection of the different functional subunits and the transport of the materials. The integration of a high number of subunits requires the extension of the workspace. Thus, two ORCA laboratory robots (Beckman Coulter, Krefeld) are used, which are moving on two orthogonal rails. Both robots are connected over a regrip station, which enables the bilateral exchange of labware as well as regripping for different orientations of the microplates. Another advantage is the significant increase of the sample throughput due to the use of two ORCA robots. Sample storage: The standard and sample solutions, reagents, and solvent are provided in plastic or glass containers or microplates depending on the type of application. Because of standardization, microplates are preferred. Some reagents and sample solutions cannot be stored in microplates due to the required volumes in the sample preparation or analytical measurement. They are processed in single vials with volumes between 14 and 125 ml. To use the microplate format also for single-vial handling, special racks have been developed. Figure 4.21 shows a variety of containers and adapters used on the system. The racks are equipped with PTFE (polytetrafluoroethylene) covers to ensure tight sealing of the different single containers. This avoids additional tasks for unscrewing the different sample containers for elemental analysis (see Figure 4.22). The presented concept comprises the idea of capping the whole MTP-footprint including the implemented vessels with just one lid. Irrespective of the port diameter of the vessels, the Biomek gripper tool is capable of gripping this lid in one single step allowing for simultaneous opening and covering of up to 24 vessels. Regarding the specifications of the liquid handler software, the conceived trays (including the lid) have to be smaller than 120 mm (parameter: lid on height) to provide reliable lid gripping. For gripping the whole tray, a gripper-lip is necessary. The highest feasible level of this lip amounts to 60 mm (parameter: lifter height). However, these levels must be ensured to avoid collisions between the vessels and the Biomek gripper tool (in vertical direction). Moreover, a second lip has been attached providing tight fit of the vessels. A storage system with 196 positions for sample containers in microplate format with different heights is integrated on the deck, which can be handled automatically or manually. Liquid handling: The preparation of calibration standards, dosage of digestion acid, solvent, and internal standard as well as the final dilution are realized on

4.2 Centralized Open Automation System

5

3

1

7

4 6

2

2

(a) 8 1 3 2 4

2

5

6

7

(b)

Figure 4.20 Complete automated system for sample preparation of different sample types; (a) front view, (b) top view, 1: liquid handler Biomek 2000 (Beckman Coulter, Krefeld) with housing and exhaust ventilation, 2: central system integrator – 2 ORCA laboratory robots on orthogonal robot rails (Beckman Coulter, Krefeld), 3: storage system for labware, reagents, and samples, 4: thermomixer MKR23 (HLC BioTech, Bovenden), thermomixer Comfort (Eppendorf, Hamburg), antistat (CEM, Kamp-Lintfort), 5: SCARA robot TS60 (Stäubli, Bayreuth) with balance and crimp station, 6: GC-MS system, 7: ICP-MS system with optional coupling to HPLC, 8: single vial liquid handler.

117

118

4 Automation Systems with Central System Integrator

(a)

(c)

(e)

(b)

(d)

(f)

(A)

(B)

Figure 4.21 Racks in microplate format for the handling of different sample and chemical containers; (A) CAD construction, (B) realized racks with containers; (a, b) rack for two narrow bore bottles (LDPE, vol. 125 ml); (c, d) rack for six microwave digestion vessels CEM Xpress (PFA, vol. 25 ml); and (e, f ) rack for 24 sample containers (PP, vol. 14 ml). (a)

(c)

(e)

(b)

(d)

(f)

(A)

(B)

Figure 4.22 Racks in microplate format for the handling of different sample and chemical containers; (A) CAD construction, (B) realized racks with containers; (a, b) rack for two beakers (PFA, vol. 100 ml); (c, d) rack for six centrifuge tubes (PP, vol. 50 ml); and (e, f ) rack for six centrifuge tubes with cover (PP, vol. 50 ml).

a liquid handling system Biomek 2000 (Beckman Coulter, Krefeld) integrated into the automation system. The liquid handler is equipped with a single pipetting tool to enable the processing of single vials. The single vials are positioned in racks in microplate format, which allows easy handling by the robots used and positioning at the defined positions (ALPs). In order to prevent cross-contaminations, changeable tips are used in the system. According to the

4.2 Centralized Open Automation System

2

3 1 8

4 7 7 5

6 9

Figure 4.23 CAD construction of single vial liquid handler – 1 and 2: horizontal and vertical linear rails, 3: pneumatic carriage, 4: needle, 5: wash station, 6: ALPs for positioning racks in microplate format, 7: status lights (green: ready, rot: in operation), 8: ORCA laboratory robot (Beckman coulter, Krefeld), 9: linear rail for laboratory robot.

type of application, tips with a special filter are used, which prevent the raising of vapors into the pipette cannulas which could have a negative impact on the system mechanics. The Biomek 2000 can be operated in the range between 1 and 1000 μl. For handling greater volumes, a single-vial liquid handler is integrated [235]. This liquid handler has a Cartesian structure with three DOF (xyz portal) (see Figure 4.23). The vertical movement of the needle into the containers as well as the horizontal movement of the needle over the containers is realized with two linear rails, the carriage is driven by a stepper motor over drive belts. The third DOF (orthogonal to the horizontal linear rail) is realized with a pneumatic carriage. The containers positioned in racks in microplate format can be placed on two predefined positions. The dosage is realized with a dispenser ML511C (Hamilton, Bonaduz) with a syringe 1010TLL (vol. maximum 10 ml; inaccuracy (CV) of liquid handling merely ≤1%). In the described application, the single-vial liquid handler is used for the dosage of one solvent (water). More dispensers can be added to extend the system to handle different solvents. Alternatively, a multi-needle head or a multiport valve can be used. A passive wash station is integrated to avoid cross-contaminations. Using a predefined method, the needle is rinsed in the wash port from the outside with a solvent, which is collected in a waste container [235]. Sample treatment: Microwave digestion is required within the sample preparation procedure for the determination of mercury in wood. The digestion is performed automatically in a microwave device Mars5 (CEM, Kamp Lintfort). Due to safety reasons, the device has to be positioned under a hood. If it is not integrated into the automation system, a separate manual transport step

119

120

4 Automation Systems with Central System Integrator

is required. Direct integration of the microwave into the automation system requires comprehensive enhancement of the safety features of the housing as well as the integration of additional features, such as suitable exhaust units and splash back tools. Additional sample preparation steps include the homogenization of the samples or a derivatization at specific temperatures. These tasks are realized by different thermomixers (thermomixer MKR23, HLC BioTech, Bovenden; thermomixer Comfort, Eppendorf, Hamburg), which differ by the type of racks that can be used. The use of PFA-based microwave digestion vessels can lead to an electrostatic charge. Particles at the inside wall of the containers can have a negative impact on the microwave digestion due to ignition and damage to the containers. An ionization device Antistat 2000 (CEM, Kamp Lintfort) is integrated into the automation system to ensure electrostatic discharge. The device is easily accessible so that a flow of deionized air can reach the required tubes. Sample preparation processes for structural analysis often require a change of the solvent, for example, from inorganic to organic solvents. A sample concentrator with a heating block DB-3A type FD803AD (Techne, Burlington) was thus integrated into the automation system. Three blocks with 20 GC vials can be positioned in the concentrator. The sample concentration is realized with a constant nitrogen flow into the separate vials. Alternatively, the system can be modified for handling microplates by changing the cannula matrix and the heating block FSC496D (Techne, Burlington). Sample handling: This functional unit includes the opening and closing of GC vials (volume 2 ml) with screwing, snap or crimp lids. The used lids include a septum that has to be pierced by the injection needle of the automated analytical measurement system. The covers of the racks can be handled by the liquid handler or the ORCA robot, whereas screwing, capping, and crimping require a separate work station. A high-speed robot TS60 (Stäubli, Bayreuth) with four axes (SCARA robot) is integrated into the automation system for opening and closing tasks. The robot has a range of 60 cm and a stroke of 40 cm with a repeatability of ±0.01 mm. The nominal payload is 2 kg and the maximum payload is 8 kg. To decouple the main platform of the automation system from the separate workstation of the high-speed robot (which is necessary because of vibrations due to the high speed), a plate shuttle is included, which transports the labware in microplate format on a linear rail between the two platforms (see Figure 4.24). The implemented weighing station BP 211D (Sartorius, Göttingen) offers weighing in the range of 0.01 g up to 210 g with an inaccuracy of ≤0.05 mg up to 80 g and an inaccuracy of ≤0.1 mg up to 210 g. Allowing for accurate weighing processes of individual vials, the weighing station provides an internal calibration weight and is picked and placed by high precision SCARA robot motions. Moreover, allowing for data storage and sample dosage options, weights will be logged in an Excel file. An additional task of the functional unit “sample handling” is the reading of barcodes of the different samples using a webcam HQ DA-71813 (Digitus, Lüdenscheid). A large percentage of laboratory errors are especially related to errors in sample identification [236]. Hence, for certification or product registration under the ISO 9000 series of standards, specific improvements in the areas of measurement traceability and data audit trails are essential,

4.2 Centralized Open Automation System

Figure 4.24 Shuttle for the transport of microplates between two automation platforms.

which have become an important part of manufacturing quality systems documentation [237]. Thus, to ensure sample identification, the Digitus HQ Webcam USB 2.0 has been implemented into the automated system supplying 2D barcode reading. The Digitus HQ Webcam is a cost-efficient camera, which has a resolution of 1600 × 1200 megapixel with a frame rate of up to 30 frames per second. It provides driverless installation, a USB 2.0 interface, and supports Windows 8, 7, Vista, and XP. The complete system can be partially or completely surrounded by a housing. Due to the high-motion velocities and the resulting forces during the operation of the robot, the TS60 is enclosed in a full housing. Since substances are handled in the sample preparation processes that could be a health hazard, some parts of the automation system, where substances are handled in an open way, are surrounded by additional housings. One example is the liquid handling system. The housing includes an exhaust tool, which removes all vapors immediately from the system. The integration of sensors for measuring toxic solvents and gases such as nitrous gases is another safety feature in the system. Analytics: A mass spectrometric system is used for the determination of mercury in wood samples. The system consists of an ICP-MS 7700x (Agilent Technologies, Waldbronn) and an autosampler unit ASX-500 (Cetac, Neuss) for automated sample delivery. The MS has been directly integrated with the automation system. Finally, the prepared samples are transferred by the system integrator into the autosampler of the ICP-MS. Due to safety reasons, an exhaust facility is required above the system. For the determination of dental material components, an LC-TOF mass spectrometer (Agilent Technologies, Waldbronn) and a GC-MS (Agilent Technologies, Waldbronn) have been integrated into the automation system. The data is automatically evaluated after the measurements in online mode. 4.2.4 Process Description 4.2.4.1 Process Description for Determination of Mercury in Waste Wood

All necessary solutions and labware are provided manually to the automation system. In the automated processing, the samples, standard solutions, reagents, solvents, and the labware are transported by the central robot from the storage positions to the final positions on the liquid handler deck. After removal of the

121

122

4 Automation Systems with Central System Integrator

rack covers, the digestion acid (aqua regia or nitric acid) with an internal standard (for correction of variations caused by sample evaporation) is added to the milled wood samples in the microwave digestion tubes. After a pre-digestion of about 20 min in open tubes, the cover is attached. The rack is transported to the sample transfer station and is manually provided to the automated microwave digestion. (Note: This step is realized separately from the automation system due to safety reasons. A full automation including the microwave step is possible. In this case, the microwave should be integrated into the system. The samples can be transported into the microwave by the central robot.) The cooled down tubes with the digested samples are transported back to the sample transfer station and to the deck of the liquid handler. For digested material dilution, empty analysis vials are provided to the single-vial liquid handler for filling with 10 ml of ultrapure water. Subsequently, the vials are transported to the liquid handler for the addition of 1.5 ml water. Two different devices are used for water pipetting. The single-vial liquid handler can dispense higher volumes, which results in time savings compared to the Biomek 2000 liquid handler, since only one pipetting step is required. For dispensing 10 ml, the pipetting time can be reduced by 90%, since the Biomek 2000 can only pipette a maximum volume of 1 ml. The remaining 1.5 ml water and 1.0 ml of the digestion solution are added by the Biomek 2000 liquid handler. The finally prepared solutions are transported to the autosampler of the ICP-MS to be analyzed. The used labware is transported back to the hotels and home positions, from where it can be removed from the system. Figure 4.25 shows the complete process with all required process steps. 4.2.4.2 Process Description for the Determination of Methacrylates in Dental Materials

All necessary solutions and labware are provided manually to the automation system. In automated processing, the samples, standard solutions, reagents, solvents, and the labware are transported by the central robot from the storage positions to the final positions on the liquid handler deck. The samples have to be transferred into GC vials for the analytical measurements using HPLC-MS and GC-MS. The GC vials are provided on a rack in microplate format and can thus easily be handled by the robot (see Figure 4.26). The sample preparation for the analytical determination using HPLC-MS includes the preparation of stock solutions of the compounds TEGDMA, BisDMA, BisGMA, and UDMA with a concentration of 2 g/l in acetonitrile. Four calibration solutions in the concentration range of 5–100 μg/l (TEGDMA and UDMA), respectively 10–500 μg/l (BisGMA), and 10–250 μg/l (BisDMA) are prepared from the stock solutions. The sample preparation for the analytical determination using GC/MS includes the preparation of stock solutions of the compounds TEGDMA, HEMA, MMA, and MA with a concentration of 10 μg/l in dichloromethane. Four calibration solutions in the concentration range of 5–50 mg/l (TEGDMA and HEMA), 5–100 mg/l (MMA), and 1–100 mg/l (MA) are prepared. For the analytical investigation of the non-polymerized dental materials, the material is manually weighted. The automation of this step requires the

4.2 Centralized Open Automation System

Start Initial transport Supply of samples, stock solutions, reagents, solvents, and labware

Transport Storage system → Liquid handler (samples, stock solutions, reagents, solvents, and labware)

Opening Remove cover from rack (microwave digestion vessels)

Pipetting Add digestion acid to samples

Closing Put cover on rack (microwave digestion vessels)

Pre-digestion Wait 20 min with open vessels

Transport Liquid handler → Sample transfer station (microwave digestion vessels)

Transport Sample transfer station → Microwave device (microwave digestion vessels)

Microwave digestion Digestion of samples

Transport Microwave device → Sample transfer station (microwave digestion vessels)

Transport Sample transfer station → Liquid handler (microwave digestion vessels)

Transport Storage system → Single-vial liquid handler (analysis vials)

Opening Remove cover from rack (analysis vials)

Pipetting Add water

Transport Single-vial liquid handler → Liquid handler (analysis vials)

Opening Remove cover from rack (microwave digestion vessels and analysis vials)

Pipetting Add digestion solution

Transport Liquid handler → ICP-MS (analysis vials)

Analysis Measurements using ICP-MS

Transport Liquid handler → Storage system (labware used)

Transport ICP-MS → Storage system (labware used)

Closing Put cover on rack (analysis vials)

Closing Put cover on racks (microwave digestion vessels and analysis vials)

Final transport Storage/disposal/cleaning of samples, stock solutions, reagents, solvents, and labware

End

Figure 4.25 Process flow chart for the automated sample preparation for the determination of mercury in wood.

integration of a solid dispensing station. Depending on the type of the solid, ripple technologies with defined volumes or screw-conveyors can be used. To enable a maximum of flexibility for the automation system, the integration of different types of solid dispensers would be required, which involves considerable costs. After weighing, the samples are diluted with acetonitrile to a concentration of 150–200 μg/l (HPLC-MS) or 300–500 mg/l with dichloromethane (GC-MS).

123

124

4 Automation Systems with Central System Integrator

(a)

(b)

Figure 4.26 Racks in microplate format for handling of GC vials (vol. 2 ml) (a) CAD design and (b) realized rack with vials.

For the determination of rest monomers in artificial saliva, ethanol, lactic acid, or citric acid, the solvent is removed under a stream of nitrogen in the sample concentrator. The residuals are resolved in acetonitrile (HPLC-MS) or dichloromethane (GC-MS). The GC vials have to be screwed with a lid including a septum that has to be pierced by the injection needle of the HPLC autosampler or GC injector. The screwing is realized with the SCARA robot TS60, wherefore the samples are transported with the shuttle from the main automation system to the external robot work station. After screwing, the samples are homogenized for 1 min at 750 rpm on the mixer (thermomixer Comfort, Eppendorf, Hamburg). The samples have to be filtered to avoid a clogging of the measurement devices due to non-dissolved particles or salt crystals. This requires a new opening of the vials followed by filtration and re-closing using the TS60 robot. Two possibilities are available for the filter process. A filtration module is installed on the Biomek 2000, which is designed for the filtration of microplates. Thus, reformatting from the GC vials to a microplate is required. In case of HPLC-MS measurements, the resulting plates can directly be used without further reformatting. For the analytical measurement using GC-MS, the filtrated solutions have to be refilled into GC vials. Alternatively, automated filtration is possible using manual syringe filters; this requires the integration of an additional device. A suitable system will be introduced in Section 5.1. The final samples are provided to the analytical measurement systems by the central system integrator. Figure 4.27 shows the process flow chart with all required process steps. 4.2.5 Control of the Automation Process

The automation system is controlled by the scheduling software “SAMI Workstation Ex 4.0” (Beckman Coulter, Krefeld). Two SAMI methods have been developed for the addition of reagents to the wood samples including the following pre-digestion (Figure 4.28) and the final dilution of the digested samples for the following analytical measurement using ICP-MS (Figure 4.29) [230]. Figure 4.30 exemplarily illustrates the BioWorks method for the liquid handler Biomek 2000 (Beckman Coulter, Krefeld) for pipetting of dilution acid to the solid wood samples.

4.2 Centralized Open Automation System

Start Initial transport Supply of samples, reagents, solvents, and labware

Transport Storage system → Liquid handler (samples, reagents, solvents, and labware)

yes Sample extract?

Pipetting Add sample extract

Transport Liquid handler → Sample concentrator (GC vials)

Transport Sample concentrator → Liquid handler (GC vials)

Vaporizing Removal of solvent using nitrogen

no

Pipetting Add ACN

HPLC

HPLC or GC?

GC Pipetting Add CH2Cl2

Transport Liquid handler → SCARA TS60 (GC vials)

Closing Add screw caps

Transport SCARA TS60 → Shaker (GC vials)

Homogenization Mixing of sample and solvent

Transport Shaker → SCARA TS60 (GC vials)

Opening Remove screw caps

Transport SCARA TS60 → Liquid handler (GC vials)

Reformatting GC vials → 96 well MTP

Filtration

GC

Reformatting 96-well MTP → GC vials

Transport Liquid handler → SCARA TS60 (GC vials)

Closing Add screw caps

HPLC or GC?

HPLC

Transport Liquid handler → HPLC (96 well MTP)

Transport Liquid handler → Storage system (labware used)

Analysis Measurements using HPLC-MS

Transport HPLC-MS → Storage system (labware used)

Transport SCARA TS60 → GC (GC vials)

Analysis Measurements using GC-MS

Transport GC-MS → Storage system (labware used)

Final transport Storage/disposal/cleaning of samples, stock solutions, reagents, solvents, and labware

End

Figure 4.27 Process flow chart for the automated sample preparation for the investigation of dental materials.

125

126

4 Automation Systems with Central System Integrator

Figure 4.28 Automated method for sample preparation of wood samples: “SAMI Method Editor” – method for addition reagents and pre-digestion (24 vessels).

Besides the liquid handler Biomek 2000 (Beckman Coulter, Krefeld), a single vessel handler was integrated into the automation system, which is available as a device in the SAMI software (Beckman Coulter, Krefeld) and can be included in different methods. The configuration of the liquid handling processes is realized with a software callable from SAMI, which provides two dialogs [235]. Using the “Action Config” dialog, the operator can initialize all system devices (e.g., XYZ portal, pumps, valves, etc.), the approach of a park position, and the creation of the liquid handling methods. A typical transfer process includes an aspirating and a dispensing step. The used medium, the position of the needle head (fixed or dynamic, the medium following immersion depth), the transfer volume, and the flow rate can be defined. With the “Configuration Config” dialog, the parameters of the liquid handler are saved as the system configuration and calculation basis. Thus, the height of the washing blocks and the racks or the needle length are important parameters for the positioning of the XYZ portal. Other important parameters include the dimensions of the sample tubes to be handled [235]. 4.2.6 Evaluation of the Automation System

High-throughput screening methods in the field of environmental analysis require fast and economical measurement techniques. An ICP-MS in combination with fully automated sample preparation provides a very sensitive and rapid multi-element measurement technique for the determination of mercury and heavy metals in wood materials and waste wood. The measurement method was validated using microwave vessels with various volumes and standard reference material. Sample weights in the range of 0.1–0.8 g and volumes of the digestion solution in the range of 6–10 ml are used in the regulations [187, 188]. For

4.2 Centralized Open Automation System

Figure 4.29 Automated method for sample preparation of wood samples: “SAMI Method Editor” – method for dilution (1 : 12.5, v/v) of the digestion solution (24 vessels).

Figure 4.30 Automated method for sample preparation of wood samples: “BioWorks Method Editor” – method for dilution of the digestion solution (24 vessels).

127

128

4 Automation Systems with Central System Integrator

Table 4.4 Results of the repeatability and recovery rate using microwave vessels with various volumes [231]. Microwave digestion vessels Xpress 75 ml

XP1500 50 ml

Xpress 25 ml

Xpress 25 ml

Performed

Manually

Manually

Manually

Automated

Number of samples

35

33

25

25

Hg concentration (average) (mg/kg)

0.562

0.608

0.579

0.585

STDEV (mg/kg)

0.047

0.053

0.067

0.051

CV (%)

8.35

8.69

11.57

8.62

Recovery rate minimum/ maximum (%)

79–112

87–130

80–120

84–110

Recovery rate (average) (%)

94

102

96

98

efficient automation, this standard procedure was miniaturized and tested with microwave vessels with volumes between 25 ml and 75 ml. These vessels enable low sample volumes from 60 to 300 mg and low digestion volumes from 2 to 8 ml. The repeatability, recovery rate, within-laboratory precision, method stability, measurement precision, and the limit of detection and quantification were determined [64, 227–231]. Table 4.4 gives an overview about selected validation results. The limit of detection in the measurement solutions was 6.1 ng/l and the limit of quantification was 14.3 ng/l. For the solid waste wood samples, the limit of detection was 2.45 μg/kg and the limit of quantification was 5.73 μg/kg. A comparison of the measurement methods developed was finally carried out using real wood samples from environmental and waste origin. The individual techniques showed similar results [64, 227–231]. However, in the miniaturized method the sample weight and the volume of reagents were reduced by 75–88% compared to the classical method. Table 4.5 provides an overview of these significant material savings. Moreover, the results of the repeatability and the within-laboratory precision for manual and automated sample preparation were performed using the miniaturized method and compared to the calculated Horwitz criterion [238]. The coefficients of variations determined for both the manual and automated method were below the maximum value allowed. Furthermore, the value of the CV for the repeatability could be reduced with the automated method from 11.57% to 8.60% (see Figure 4.31). The processing time for sample preparation required was determined for the standard manual method, the miniaturized method, and the automated method (values rounded up to full minutes). Table 4.6 provides an overview of the results.

4.2 Centralized Open Automation System

Table 4.5 Overview of material savings using microwave vessels with different volumes and the methods presented [231]. Standard method (vessel vol. 75 ml)

Miniaturized method (vessel vol. 25 ml)

Reduction factor

Reduction (%)

Sample weight (mg)

250–350

62.5

4.0–5.6

75.0–82.1

Vol. aqua regia (ml)

8.0

2.0

4.0

75.0

Vol. digestion solution (mL)

8.0

1.0

8.0

87.5

Vol. analysis solution (mL)

100

12.5

8.0

87.5

100 Recovery rate

13.5

Repeatability

80

12

70

10.5

60

9

50

7.5

40

6

30

4.5

20

3

10

1.5

0

Manual standard (Xpress 75 ml)

Manual miniaturized Automated miniaturized (Xpress 25 ml) (Xpress 25 ml)

Repeatability (coefficient of variation) (%)

Recovery rate (%)

90

15

0

Figure 4.31 Comparison of recovery rates and repeatabilities achieved using the manual standard, the manual miniaturized, and the automated miniaturized methods [64, 229, 231, 239].

The miniaturization led to time savings due to an easier dilution procedure without volumetric flasks. The automated method sequentially prepares six racks with six samples. Compared to manual processing, the pipetting operations in automated mode are more time-consuming. The reason is the use of two liquid handler devices, which differ in their maximum volumes (maximum 1 ml and

129

130

4 Automation Systems with Central System Integrator

Table 4.6 Processing time for determination of mercury in wood with the manual standard procedure, the miniaturized, and the automated methods [228, 229, 231–233, 240]. Processing time (min) Process step

Pipetting of digestion acid

Number of samples/run

Standard manual

Miniaturized manual

Miniaturized automated

36

3

3

—

6×6

—

—

40

36

20

20

—

36

—

—

20

Close microwave vessels

36

6

6

—

36

—

—

6

Microwave digestion incl. cooling down

36

70

70

—

36

—

—

70

36

12

12

—

36

—

—

12

36

85

25

—

6×6

—

—

67

196

136

215

Pre-digestion

Open microwave vessels

Dilution

Total time (min)

maximum 10 ml) (see Section 4.2.3). In general, maximum accuracy is achieved when the pipetting volume is not less than 10% of the maximum volume of dispensing device [241]. According to the volume to be pipetted, the appropriate liquid handler is selected. However, this requires additional process steps such as placing and removing the covers on/from the racks and transport steps between the liquid handlers. A further improvement of the automated method can be reached by a simultaneous preparation of four racks with six samples (see Section 4.3). This reduces transportation tasks and tool change procedures at the liquid handler.

4.3 Decentralized Closed Automation System Centralized automation systems can be extended to decentralized systems. In this case, the different stations of the automated sample preparation and analytical measurement are locally distributed (see Figure 4.32). To ensure a completely automated procedure, the different stations can be combined with mobile robots. The determination of phosphorous and calcium in bones is used as an example for a decentralized closed automation system, including the sample preparation at two different stations as well as an ICP-MS system for the analytical measurement.

4.3 Decentralized Closed Automation System

Figure 4.32 Schematic visualization of a decentralized closed automation system with a central system integrator (SP: sample preparation, A: analytical measurement, CSI: central system integrator, CE: connection element).

Decentralized/closed SP

CSI Mobile CE A

4.3.1 Background and Applicative Scope

In general, human bones consist of 35% mineral salt. The main components are calcium and phosphorus in the form of hydroxyapatite (molecular formula: Ca10 (OH)2 (PO4 )6 ). The inorganic matrix also contains carbonate and citrate, and small amounts of magnesium, sodium, chloride, fluoride, and various trace elements. The organic matrix, representing about 20% of a bone, involves collagen, bone marrow, fat, and non-collagenous proteins. The water content of bone is approximately 45% [242, 243]. The elemental composition of human bones may provide comprehensive information on the health condition, acute or chronic diseases, poisoning, and nutritional patterns. Furthermore, the concentrations of certain elements can be used as indicators of the nutritional status and for monitoring biological functions in the field of environmental and occupational health and safety [244]. Osteoporosis is a common bone disease, which may result in decreased bone density leading to an increased risk of falls and fractures, especially in older people [245, 246]. In the United States, more than 10 million people suffer from osteoporosis, and the annual treatment costs exceed 13.5 billion dollars [246]. Studies revealed a correlation between the progressing of arteriosclerosis and the incidence of osteoporosis and bone fractures [247]. A multitude of parameters can be used to determine the properties and the state of bones: components and composition of the inorganic and organic matrix, the size of the mineral crystals and the mechanical strength [246]. The main indicator for a diagnosis of osteoporosis is the bone mineral density (BMD) [248]. The determination of the elemental composition is another way to examine the stage and the causes of this disease. Various studies showed, that the concentrations of the elements calcium and phosphorus are independent of age [245, 249]. However, a correlation between a decreased calcium concentration and the occurrence of osteoporosis disease was reported [244, 245, 249–251]. Various analytical techniques are available for the characterization of bone tissue and the determination of the bone state. Dual energy X-ray absorptiometry (DXA or DEXA) [252, 253], quantitative computed tomography (QCT), peripheral quantitative computed tomography (pQCT) [253, 254], and qualitative ultrasonic [253] are common non-invasive techniques. Gamma ray

131

132

4 Automation Systems with Central System Integrator

attenuation may be used to determine the mineral content [251, 255]. Other techniques for the study of mineralized tissues include diffraction techniques using neutrons, X-rays, or electrons as well as vibrational spectroscopy such as Fourier-transform infrared spectroscopy (FTIR) spectroscopy or Raman spectroscopy [256]. For determining the elemental composition of bones, various single-element and multi-element techniques are known, whereby the latter are often preferred [257]. Before the analytical measurement a complex sample preparation is required (pre sensoric selectivity). This includes the removal of adherent tissue and fat [249, 251, 258]. Different methods are available, whereby the use of hydrogen peroxide (H2 O2 ) is very effective [249]. After roughly crushing, drying to constant weight, and further particle size reduction by milling, the solid samples are transferred to a liquid state, wherein the organic matrix is completely decomposed. This can be done using a classical wet acid digestion or a microwave-assisted acid digestion [242, 259–261]. The most important techniques for the determination of the calcium content involve spectroscopic techniques such as F-AAS, ICP-AES, or ICP-OES [245, 249, 250, 261–266]. For the determination of the phosphorus content mostly ICP-AES and ICP-MS are used [244, 245, 249, 267]. In contrast to single-element techniques (e.g., AAS), multi-element techniques such as ICP-AES (Inductively Coupled Plasma Atomic Emission Spectroscopy), ICP-OES, and ICP-MS allow the simultaneous determination of several elements. These multi-element techniques differ in the element-dependent limits of detection. Typically, the plasma gas consumption and the required sample volume in the ICP-MS are smaller than in the ICP-AES and in the ICP-OES. For this reason, the ICP-MS is a powerful technique to determine elements both in the trace range and at higher concentrations. Ideally, a microwave digestion should prevent or eliminate the addition of compounds and elements as well as the formation of polyatomic ions, which generate spectral or non-spectral interferences [260, 268–270]. But in reality, a multitude of polyatomic interferences commonly occur in the determination of phosphorus and calcium. Table 4.7 provides an overview of the most common polyatomic interferences in ICP-MS [267, 268]. The use of a chemically inert collision gas such as helium (He) in a collision cell before the mass analyzer reduces such undesirable effects [271, 272]. To reduce the generation of nitrogen-based polyatomic interferences, the concentration of the nitric acid (HNO3 ) used for the microwave digestion should be decreased to 1% (v/v) in the final analysis solution [267]. 4.3.2 Automation Goals

The sample preparation of bone materials requires the removal of adherent tissue and fat residues. The bones are crushed to remove the bone marrow in the treatment with hydrogen peroxide. The purified bone material is dried in a drying cabinet at maximum 60 ∘ C to weight stability. Subsequently, the bone fragments are ground by an oscillating mill to fine bone meal. To transfer the solid sample into a liquid state, the bone powder is weighed into the digestion vessels. The digestion acid (concentrated nitric acid) and the internal standard are added, and this sample mixture is digested in the microwave device. To adjust the acid

4.3 Decentralized Closed Automation System

Table 4.7 Polyatomic interferences for P and Ca in the ICP-MS analysis [267, 268]. Isotope

Abundance (%)

Interferences

31

100

14

P

13

N16 O1 H+ , 15 N15 N1 H+ , 15 N16 O+ , 14 N17 O+ , C18 O+ , 12 C18 O1 H+ , 30 Si1 H+

40

Ca

96.97

40

Ar+

42

Ca

0.64

40

Ar1 H2

0.145

27

Al16 O+

2.06

12

C16 O2 , 14 N2 16 O+ , 28 Si16 O+ N16 O2 + , 32 S14 N+

43 44

Ca Ca

46

Ca

0.003

14

48

Ca

0.19

33 15

S N+ , 34 S14 N+ , 32 S16 O+

content and the analyte concentration for the subsequent analysis using ICP-MS, the samples are diluted and transferred in vessels suitable for the autosampler of the analysis instrument. After the analysis using ICP-MS, the evaluation of the measured data is performed, followed by a storage or safe disposal of the measured samples. Between the subprocesses of sample preparation, analysis, and sample storage or disposal, material transports are necessary, especially between the automation system and the external stations of the microwave device and the ICP-MS. The transports within the automation system are performed by the central system integrator. The transports between separate subsystems and stations are realized manually or by using mobile robots. In general, all subprocesses can be automated. When using classical laboratory equipment, crushing, grinding, and weighing the samples remain manual procedures. The drying step can be automated using a robot compatible drying cabinet. Due to safety issues, the microwave digestion is performed automatically at a separate station. The analysis using ICP-MS is performed on an externally located device in another laboratory. The limits of detection vary between 49 and 76 μg/l (phosphorus) and 17–64 μg/l (calcium) [273]. In addition to the determination of phosphorus and calcium, further medically relevant elements such as strontium, magnesium, or heavy metals can be integrated into a multi-element method. Figure 4.33 illustrates the general process flow with its subprocesses [64]. Bone sample

Roughly crashing, removal of tissue and fat residual, drying, fine milling, weighing Sample storage or disposal Pipetting of reagents and internal standard

Microwave digestion

Dilution, reformatting

Measurement using ICP-MS Data evaluation

Material flow Information flow

Figure 4.33 Process workflow for the determination of calcium and phosphorus in bones using ICP-MS.

133

134

4 Automation Systems with Central System Integrator

The subprocess “liquid handling” involves the pipetting of standard solutions, addition of the digestion acid, and the internal standard to the solid sample. The subprocesses “dilution” and “reformatting” perform the final dilution of the digestion solution to a final acid content of 1% (v/v) and the transfer of the solutions into sample containers (vol. 8 ml). The subprocess “analysis” is characterized by short analysis times using a multi-element method. The subprocess “transport” is performed by the central system integrator and mobile robots. The automation of this process eliminates human errors (pipetting errors, contamination of the digestion solutions, errors in dilution, etc.). The amount of bone samples, the volumes of the digestion acid, internal standards, and solvent can be reduced due to the miniaturization. This is of great importance in the study of human bone samples since here usually very little sample material is available. In addition, occupational safety is increased with the automation of this process. 4.3.3 System Design

The designed automation systems consist of the functional units “sample storage,” “liquid handling,” “sample treatment,” “sample handling,” “analytics,” and “mobile robotics” besides the central system integrator. The functional units are similar as described in Section 4.2.3. Differences exist regarding the analytical measurement devices. The analytical devices are separated from the sample preparation system in different laboratories. Thus, the devices can also be used by other applications and operators; in addition, the spatial restrictions that are connected with a centralized system can be eliminated. To set up a real distributed system, the functional unit “mobile robotics” has been added to the automation system. Using mobile robots enables the integration of the external stations ICP-MS (functional unit “analytics”) and microwave (functional unit “sample treatment”) and their connection to the automated sample preparation system. The microwave is positioned in a separate laboratory under a hood. Analytics: The analytical determination of phosphorous and calcium in bone materials is realized using ICP-MS (elemental analytics). The system setup is mainly similar to that discussed in Section 4.2. The ICP-MS system is not integrated to ensure the use of the system for different analytical applications. The automation system is isolated in a housing and equipped with an exhaust system. A special sample transfer station has been designed to enable the sample transfer between the system and the mobile robots. The analytical device is located in a separate laboratory for trace analysis/elemental analytics, which provides suitable climatic and air conditions (very little air contamination). The system is installed with an extensive gas supply, whereas argon and an argon/oxygen mixture are used as plasma and cooling gases for inorganic or organic samples. Helium is used as collision cell gas. The integration into a decentralized system enables a complete automation in combination with a high amount of flexibility. The separation of the analytical devices results in more space, which can be used for the integration of new stations to increase the functionality of the system.

4.3 Decentralized Closed Automation System

Mobile robotics: A mobile robot H20 (Dr Robot Inc., Markham) connects the different distributed stations. The H20 is a humanoid dual arm robot on a platform driven with two 7′′ wheels. The arms have six DOF each; the optimized grippers have two DOF. The range of the arms is about 60 cm, the payload is 800 g. The robots can be operated with a maximum speed of 75 cm/s. The robot head, which has two cameras with a resolution of 640 × 480 pixels installed, can be moved in six DOF. The robot is equipped with 5 ultrasonic and 10 infrared sensors and uses a StarGazer module HSG-A-02 (Hagisonic, Daejeon) for the navigation of the robots using an infrared camera and passive landmarks at the ceilings (Indoor GPS) [274, 275]. The landmarks are designed as a matrix with 4 × 4 dot-shaped markers made of reflecting material; a maximum of 4095 different waypoints (IDs) can be built. This number is suitable for the generation of maps in life science laboratories, to mark the required paths for the robots [275, 276]. In addition, a Kinect sensor (Microsoft, Redmond), which can detect obstacles at a distance between 0.5 and 5 m is used for collision avoidance. The positioning for picking and placing tasks is realized with the passive landmarks and the StarGazer module. Two ultrasonic sensors in the robot platform enable fine positioning; picking and placing is realized as a blind operation without additional sensors or cameras [277]. Higher precision and flexibility can be achieved using camera-based methods for the labware identification and picking/placing tasks. Special sample transfer stations have been developed and integrated with the different external stations. Another feature is the monitoring of the battery power to avoid failures during the transportation steps [278]. 4.3.4 Process Description

All required samples, stock solutions, reagents, solvents, and labware are provided manually or with mobile robots to the dedicated storing positions in the automated sample preparation system. The two ORCA robots transport all materials to the defined positions (ALPs) on the liquid handler deck. After removing the covers, the digestion acid (nitric acid) with internal standard (correction of measured values due to evaporation) is added to the milled bone samples in the microwave tubes. After a pre-digestion of about 20 min, the tubes are closed with the cover and the rack is transported to the sample transfer station. The mobile robot picks the rack and transports it to the station for the automated microwave digestion. The final and cooled samples are transported back by a mobile robot to the sample transfer station on the sample preparation station. The ORCA robot as the central system integrator transports the samples to the liquid handler Biomek 2000. Empty sample vials (vol. 50 ml) are provided at the single vial liquid handler for filling with 49.5 ml of ultrapure water. Subsequently, the vials are transported to the liquid handler for the addition of 0.5 ml digestion solution (dilution 1 : 100). This dilution step decreases the acid concentration of the nitric acid to 1% (v/v) for the analytical measurement using ICP-MS. A second dilution step (1 : 10) is required to decrease the content of phosphorus and calcium to the working range of the ICP-MS. Therefore, empty analysis vials (vol. 8 ml) are transported from the sample storage system to the liquid handler Biomek 2000. A volume of 5.4 ml nitric acid (1%, v/v) and 0.6 ml diluted digestion solution was added. Again, both

135

136

4 Automation Systems with Central System Integrator

liquid handling systems, the single vial liquid handler and the Biomek 2000 are involved in this dilution to ensure a total dilution of 1 : 1000. After closing the samples, the final diluted solutions for the analytical measurements are transferred to the sample transfer station, where the mobile robot picks up the rack and transports it to the ICP-MS laboratory. Once the analytical measurement is completed, the samples have to be transported back to the sample preparation system from where they can be removed from the system together with the used labware. Figure 4.34 shows the resulting process flow chart including all necessary process steps. 4.3.5 Control of the Automation Process

The automation system is controlled with the scheduling software “SAMI Workstation Ex 4.0” (Beckman Coulter, Krefeld). Two SAMI methods have been developed for the addition of the reagents to the bone samples followed by pre-digestion as well as for the final three-stage dilution of the digestion solutions for the following analytical measurement using ICP-MS [230]. Figure 4.35 shows the SAMI method for the three-stage dilution; Figure 4.36 illustrates exemplarily the related BioWorks method of the liquid handler Biomek 2000 for the dilution of the digestion solutions. Three control levels have been developed for controlling the mobile robots [279, 280]. The transportation request level is the highest level, which accepts transportation requests and distributes the transportation tasks to the mobile robots. The second level enables the transport management (transportation managing level). On this level, a suitable robot is selected and provided with a route guaranteeing the shortest distance. The third level includes the execution of the transportation task (transportation execution level) by one or more selected robots along a defined path [281]. With respect to the execution of the movement of the robot in the room and the movement of the robotic arms, this level can further be divided into a motion execution level and an arm execution level [276]. The control systems were developed according to these levels. The first level is the process management system (PMS) or a hierarchical workflow management system (HWMS) (see Section 4.4). The robot remote center (RRC) realizes the second level and the robot board center (RBC) located on the local computer of each mobile robot is the third level (see Figure 4.37) [281]. Data transfer between the PMS and the RRC is realized in a local network (LAN). The robots communicate over a wireless local network (WLAN) with IEEE 802.11g standard (frequency band 2.4 GHz). The doors between the different laboratories and corridors can be opened via WLAN [280]. Figure 4.38 shows an overview of the used network architecture [276, 281]. 4.3.6 Evaluation of the Automation System

The focus was to optimize the standard method, to miniaturize this optimized method, and finally to automate it. The evaluation of the automation system was performed with a comparison of the results of validation parameters for

4.3 Decentralized Closed Automation System

Start Initial transport Supply of samples, stock solutions, reagents, solvents, and labware

Transport I Storage system → Liquid handler (samples, stock, solutions, reagents, solvents, and labware)

Opening Remove cover from rack (microwave digestion vessels)

Pipetting Add digestion acid to samples

Transport I Liquid handler → Sample transfer station (microwave digestion vessels)

Closing Put cover on rack (microwave digestion vessels)

Pre-digestion Wait 20 min with open vessels

Transport II Sample transfer station → Microwave device (microwave digestion vessels)

Microwave digestion Digestion of samples

Transport II Microwave device → Sample transfer station (microwave digestion vessels)

Transport I Sample transfer station → Liquid handler (microwave digestion vessels)

Transport I Storage system → Single-vial liquid handler (dilution vessels)

Opening Remove cover from rack (dilution vessels)

Pipetting Add water

Closing Put cover on rack (dilution vessels)

Transport I Single-vial liquid handler → Liquid handler (dilution vessels)

Opening Remove cover from rack (microwave digestion and dilution vessels)

Pipetting Add digestion solution

Closing Put cover on rack (microwave digestion vessels)

Transport I Storage system → Liquid handler (analysis vials)

Opening Remove cover from rack (analysis vials)

Pipetting Add diluted acid

Pipetting Add diluted digestion solution

Closing Put cover on rack (dilution vessels and analysis vials)

Transport I Liquid handler → Sample transfer station (analysis vials)

Transport II Sample transfer station → External ICP-MS (analysis vials)

Transport I Liquid handler → Storage system (used labware)

Analysis Measurements using ICP-MS

Transport I Sample transfer station → Storage system (analysis vials)

Transport II External ICP-MS → Sample transfer station (analysis vials)

Final transport Storage/disposal/cleaning of samples, stock solutions, solvents, and labware

End

Figure 4.34 Process flow chart for the automated sample preparation and analysis for the determination of calcium and phosphorus in bones (Transport I: transport using the central system integrator, Transport II: transport using the mobile robot).

137

138

4 Automation Systems with Central System Integrator

Figure 4.35 Automated method for sample preparation of bone samples: “SAMI Method Editor” – Method for dilution (1 : 1000, v/v) of the digestion solution (24 vessels on 4 racks).

Figure 4.36 Automated method for sample preparation of bone samples: “BioWorks Method Editor” – Method for dilution of the digestion solution (last dilution step, 24 vessels).

all methods developed. The validation parameters involve the determination of the repeatability, the within-laboratory precision, the measurement precision and stability, and the limits of detection and quantification. The results achieved with the manual standard method and the miniaturized manual and automated method using microwave vessels with a volume of 25 ml were published in [273]. Figure 4.39 shows the calibration curves for the determination of phosphorus and calcium using ICP-MS, and a typical mass spectrum of a pig bone sample. Repeatability was checked with homogenous samples of pig bones and the NIST reference material SRM 1486. In all methods, the data was distributed normally (confirmed using the David test) and fulfilled the HORWITZ criterion.

4.3 Decentralized Closed Automation System

Process management system (PMS)

Transportation tasks

Robot status Executive results

Robot remote center (RRC) ... ...

Selected path Control commands

Robot board center (RBC) ... ...

Robot status (power, localization) ... ... Recognized waypoint

Robot board center (RBC) Motion and arm commands

Mobile robot

Robot board center (RBC)

Robot board center (RBC)

... ...

... ...

Mobile robot

Mobile robot

Data acquisition of sensors, images, and motors

Mobile robot

... ...

Figure 4.37 Communication model for mobile robotics. (Redrawn from [276, 281]).

Power module WLAN module 1 Motion and head module Process management system (PMS)

Arm servo module

Robot remote center (RRC)

WLAN module 2 Indoor GPS module Wireless bridge

Cameras Inside switch board of robot Kinect sensor

Router/Access point

Wireless bridge

Robot board center (RBC)

Inside robot

Outside robot

Figure 4.38 Network architecture for communication with mobile robots. (Redrawn from [276]).

The average values obtained were in the range from 93.5 to 100.5 g/kg (phosphorus) and 191.2 to 203.2 g/kg (calcium) and corresponded to the values given in the literature [249]. Table 4.8 gives an overview about the repeatability results achieved. The average values depend on the sample (pig bones, reference material) and show only minor differences. The coefficients of variation could be decreased in the automated methods compared to the manual methods (see Figure 4.40). The recovery rate was determined using spiked pig bones [273] and alternatively NIST reference material SRM 1486. Table 4.9 gives an overview about the results achieved with the reference material. The recovery rate mean values for calcium are in the range from 92% to 103% and for phosphorus from 99% to 104%. Figure 4.41 demonstrates a good agreement of the measured values between the different methods.

139

4 Automation Systems with Central System Integrator ×10−1

31

×10−1

P

6

Ratio

4

y = 4.0259 ⋅ 10−5 + 0.0014 R = 0.9999 DL = 15.47 μg/l BEC = 34.19 μg/l

Ratio

140

44

−5

y = 4.9789 ⋅ 10 + 0.0014 R = 1.0000 DL = 18.02 μg/l BEC = 28.87 μg/l

Ca

4

2 2

0

0

5000.0 10000.0 Conc(ug/I)

(a)

5000.0 10000.0 Conc(ug/I) 185

Lu (ISTD I)

×105 4

175

Re (ISTD II)

2

44 31

(b)

Ca

P

0 20

30

40

180

190

Figure 4.39 Calibration curves for the determination of phosphorus and calcium by ICP-MS (a), mass spectrum of a sample solution from pig bones (b).

Table 4.8 Results of the repeatability rate using microwave vessels with various volumes. Microwave digestion vessels Xpress 75 ml

Xpress 25 ml

Xpress 25 ml

Xpress 10 ml

Performed

Manually

Manually

Automated

Automated

Samples

Pig bones

Pig bones

Pig bones

NIST SRM 1486

Number of samples

39

46

46

24

Ca concentration (average) (g/kg)

203.24

191.16

200.55

244.57

Ca STDEV (g/kg)

5.94

4.21

3.14

5.18

Ca CV (%)

2.92

2.20

1.56

2.12

P concentration (average) (g/kg)

100.47

93.47

98.85

121.28

P STDEV (g/kg)

3.18

2.34

2.01

2.63

P CV (%)

3.17

2.50

2.04

2.17

Within-laboratory reproducibility was determined with 10 pig bone samples on 5 consecutive days. The mean values obtained using the manual standard method were 94.1–128.22 g/kg (phosphorus) and 189.8–229.0 g/kg (calcium). The coefficients of variation CV ranged from 1.47% to 8.8%. The heterogeneity

4.3 Decentralized Closed Automation System

3.5

Repeatability (coefficient of variation) (%)

3

2.5

2

1.5

1

0.5

0

Manual standard (Xpress 75 ml)

Manual miniaturized Automated Automated (Xpress 25 ml) miniaturized (Xpress miniaturized (Xpress 25 ml) 10 ml) Repeatability Ca

Repeatability P

Figure 4.40 Comparison of the coefficients of variation of the repeatability achieved with manual and automated methods. Table 4.9 Results of recovery rate using microwave vessels with various volumes and the NIST reference material SRM 1486. Microwave digestion vessels Xpress 75 ml

Xpress 25 ml

Xpress 25 ml

Xpress 10 ml

Recovery rate Ca minimum/maximum (%)

101–106

98–100

90–101

87–95

Recovery rate Ca (average) (%)

103

99

93

92

Recovery rate P minimum/maximum (%)

99–108

101–103

99–110

94–102

Recovery rate P (average) (%)

104

102

103

99

of the natural samples used may be a cause for the fluctuating values [250]. In the manual miniaturized method, averages of 95.7–103.6 g/kg (phosphorus) and 188.08–204.8 g/kg (calcium) were determined. The CV values did not exceed 2.2%. The results obtained with the automated method showed comparable concentrations of 92.7–109.2 g/kg (phosphorus) and 193.3–224.2 g/kg (calcium); the CV values were in the range of 0.66–3.50%. The method stability was determined according to Chapter 3. The samples were not frozen between measurements, but

141

4 Automation Systems with Central System Integrator

100

80 Recovery rate (%)

142

60

40

20

0

Manual standard (Xpress 75 ml)

Manual miniaturized (Xpress 25 ml)

Automated miniaturized (Xpress 25 ml)

Recovery rate Ca

Automated miniaturized (Xpress 10 ml)

Recovery rate P

Figure 4.41 Mean values of the recovery rate for manual and automated methods for determination of Ca and P in NIST reference material SRM 1486.

stored in a dry and dark place at room temperature. The results showed stable values within the expected range on the different test days. CV values were maximum 1.68% per sample within the 5 days for the automated method. Analogous results were obtained with the standard manual method and the manual miniaturized method. The coefficients of variation of the measurement precision were in the range 1.07–2.08% for all methods. The limit of detection in the solutions for the methods presented were detected at 49–76 g/l (phosphorus) and 17–64 g/l (calcium). The limits of quantification were 140–312 g/l (phosphorus) and 141–181 g/l (calcium). These values are in good agreement with similar studies described in the literature [245, 249, 260, 273]. The processing time for sample preparation required was determined for the standard manual method, the miniaturized, and the automated method (values rounded up to full minutes). Table 4.10 gives an overview of the results. The miniaturization led to time savings and a reduction in starting materials (see Section 4.2.6). The automated method simultaneously prepares four racks with six samples each. This reduces transportation tasks and tool change procedures at the liquid handler. Compared to manual processing, the pipetting operations in automated operation are more time consuming. The reason is the use of two liquid handler devices, which differ in their maximum volumes (maximum 1 ml and maximum 10 ml) (see Section 4.2.6). The appropriate liquid handler is

4.4 Decentralized Open Automation System

Table 4.10 Processing time for determination of Ca and P in bones with the manual standard procedure, the miniaturized, and the automated methods [230, 239, 273]. Processing time (min) Process step

Number of samples/run

Vol. of microwave vessel (ml) Pipetting of digestion acid

Pre-digestion

Close microwave vessels

Standard manual

Miniaturized manual

Automated

75

25

10

36

3

3

—

24

—

—

12

36

20

20

—

24

—

—

20

36

6

6

—

24

—

—

4

Microwave digestion incl. cooling down

36

75

75

—

24

—

—

75

Open microwave vessels

36

12

12

—

24

—

—

8

Dilution (3 steps)

36

150

62

—

Dilution (2 steps)

24

Total time (min)

—

—

72

266

178

191

selected according to the volume to be pipetted. This requires additional process steps such as placing and removing the covers on/from the racks and transport steps between the liquid handlers.

4.4 Decentralized Open Automation System The centralized closed automation system introduced in Section 4.3 can be extended to a decentralized open automation system (see Figure 4.42). The different stations for the automated sample preparation and analytical measurements are distributed in different laboratories. To ensure full automated operation, they are connected by mobile robots. The principal features of an open system are the flexible combination of different system components as well as the possibility of (also simultaneous) execution of any applications. In addition, new substations can be integrated into the complex automation system. Besides the hardware-based realization of the different stations, the real implementation of a complete open system has enormous demands regarding the control components. Flexible sample preparation for the determination of mercury in wood (see Section 4.2) and calcium/phosphorous in bones (see Section 4.3) as well as the analytical investigation of dental materials including sample preparation

143

144

4 Automation Systems with Central System Integrator

Decentralized/open SP 1

SP 2

SP X

CSI

CSI

CSI

Figure 4.42 Schematic visualization of a decentralized open automation system with a central system integrator (SP 1,2,…,X: sample preparation, A 1,2,…,X: analytical measurement, CSI: central system integrator, CE: connection element).

Mobile CE A1

A2

AX

and analytical measurement (see Section 4.2) will be described as exemplarily applications for a decentralized open system. 4.4.1 System Design

The system design matches the system described in Section 4.3.3 of a decentralized closed automation system. Besides the central system integrator, it consists of the functional units “sample storage,” “liquid handling,” “sample treatment,” “sample handling,” “analytics,” and “mobile robotics.” The functional units are identical to the system in Section 4.3. To enable maximum flexibility, additional stations were integrated. Analytics: Additional measuring systems can be integrated into the automation systems to extend the analytical functionalities. This includes the integration of an HPLC-TOF mass spectrometer for high resolution mass spectrometric investigations, which are used especially for the investigation of complex samples and matrices. A GC-Triple-Quad mass spectrometer was integrated especially for structural investigations. Besides the detection of classical mass spectra, this enables the determination of daughter ions due to a dedicated fragmentation of different ions. Simple measuring problems can be realized with simple detection systems coupled to GC and HPLC devices. Thus, a GC with flame ionization detection (FID) and an HPLC with DAD have been integrated. For the determination of elemental compositions, atomic absorption, and atomic emission spectrometer are integrated as additional stations. The analytical measurement systems are automated islands with manual sample introduction. The mobile robots cannot directly provide the samples to the analytical measurement systems. Thus, specific sample transfer stations have to be designed. Sample Treatment: Besides the units for sample processing described in Sections 4.2 and 4.3, additional stations can be integrated, which can perform different tasks. This also includes the central closed automation system described in Section 4.1, which can also be connected with other stations by the mobile robots. 4.4.2 Process Description

If different applications should be performed in one complex automation system, a first decision has to be made regarding the type of sample preparation required for the specific process. Secondly, a decision is required regarding the type of analytical measurements suitable for the specific application. The samples,

4.4 Decentralized Open Automation System

Start

Type of sample preparation ?

Determination of mercury in wood Microwave digestion and dilution (1:12.5)

Determination of phosphorus and calcium in bones Microwave digestion and dilution (1:1000)

Determination of methacrylates in dental materials Extraction and/or dilution

Measurement technique ?

Transport Sample transfer station → ICP-MS

Analysis Measurements using ICP-MS

Transport Sample transfer station → GC-MS

Analysis Measurements using GC-MS

Transport Sample transfer station → HPLC-MS

Analysis Measurements using HPLC-MS

End

Figure 4.43 Process flow chart for flexible parallel execution of three applications in a decentralized open automation system.

reagents, and the required labware have to be provided in the storage system of the automated sample preparation system. Subsequently, the process follows the execution of the selected methods. Figure 4.43 shows the process flow chart for the parallel execution of three different applications. Although the methods for the determination of phosphorous and calcium in bones have already been designed for decentralized operation including transportation by mobile robots (see Chapter 4.3), the methods for the determination of mercury in wood and of methacrylates in dental materials (see Chapter 4.2) have to be adapted to the decentralized system. Especially, the sample transfer to the decentralized located analytical measurement instruments has to be changed. The ORCA as central system integrator has to provide the samples to a sample transfer station (instead of directly transferring them to the analysis devices) to enable the pick-up and transfer of the samples to the analytical stations by the mobile robots. Figures 4.44 and 4.45 show the adapted procedures for the process execution on a decentralized open system. 4.4.3 Control of the Automation System

The system control in centralized systems is realized by a process control system specially configured for the specific system. Process control systems in life science applications are usually proprietary software solutions. If different stations in a decentralized automation system should be connected, suitable solutions have to be developed to combine the different process control systems. Therefore, the use of an HWMS is a possibility. The goal of the workflow management system is

145

146

4 Automation Systems with Central System Integrator

Start Initial transport Supply of samples, stock solutions, reagents, solvents, and labware

Transport I Storage system → Liquid handler (samples, stock solutions, reagents, solvents, and labware)

Opening Remove cover from rack (microwave digestion vessels)

Pipetting Add digestion acid to samples

Closing Put cover on rack (microwave digestion vessels)

Pre-digestion Wait 20 min with open vessels

Transport I Liquid handler → Sample transfer station (microwave digestion vessels)

Transport II Sample transfer station → Microwave device (microwave digestion vessels)

Microwave digestion Digestion of samples

Transport II Microwave device → Sample transfer station (microwave digestion vessels)

Transport I Sample transfer station → Liquid handler (microwave digestion vessels)

Transport I Storage system → Single-vial liquid handler (analysis vials)

Opening Remove cover from rack (analysis vials)

Pipetting Add water

Closing Put cover on rack (analysis vials)

Transport I Single-vial liquid handler → Liquid handler (analysis vials)

Opening Remove cover from rack (microwave digestion vessels and analysis vials)

Pipetting Add digestion solution

Closing Put cover on racks (microwave digestion vessels and analysis vials)

Transport I Liquid handler → ICP-MS (analysis vials)

Analysis Measurements using ICP-MS

Transport I Liquid handler → Storage system (labware used)

Transport I ICP-MS → Storage system (labware used)

Final transport Storage/disposal/cleaning of samples, stock solutions, reagents, solvents, and labware

End

Figure 4.44 Process flow chart for the automated execution of sample preparation for the analytical determination of mercury in wood using mobile robots for transportation tasks (Transport I: central system integrator, Transport II: mobile robot).

the combination of integrated life science system islands to an internal automated network. The physical combination between the different islands is possible using mobile robots, which realize the transport of samples and labware between the systems. The principle tasks of the hierarchical organized workflow management system include data administration and management by the user up to the automated flow control with all subsystems (see Section 7.3).

4.4 Decentralized Open Automation System

Start Initial transport Supply of samples, reagents, solvents, and labware

Transport I Storage system → Liquid handler (samples, reagents, solvents, and labware)

yes Sample extract?

Pipetting Add sample extract

Transport I Liquid handler → Sample concentrator (GC vials)

Transport I Sample concentrator → Liquid handler (GC vials)

Vaporizing Removal of solvent using nitrogen

no

Pipetting Add ACN

HPLC

HPLC or GC?

GC Pipetting Add CH2CI2

Transport I Liquid handler → SCARA TS60 (GC vials)

Closing Add screw caps

Transport I SCARA TS60 → Shaker (GC vials)

Homogenization Mixing of sample and solvent

Transport I Shaker → SCARA TS60 (GC vials)

Opening Remove screw caps

Transport I SCARA TS60 → Liquid handler (GC vials)

Reformatting GC vials → 96-well MTP

Filtration

GC

Reformatting 96-well MTP → GC vials

Transport I Liquid handler → SCARA TS60 (GC vials)

Closing Add screw caps

HPLC or GC?

HPLC

Transport II Liquid handler → HPLC (96-well MTP)

Transport I Liquid handler → Storage system (labware used)

Analysis Measurements using HPLC-MS

Transport II HPLC-MS → Storage system (labware used)

Transport I SCARA TS60 → GC (GC vials)

Analysis Measurements using GC-MS

Transport II GC-MS → Storage system (labware used)

Final transport Storage/disposal/cleaning of samples, stock solutions, reagents, solvents, and labware

End

Figure 4.45 Process flow chart for the automated execution of the sample preparation for the analytical investigation of dental materials using mobile robots for transportation tasks (Transport I: central system integrator, Transport II: mobile robot).

147

148

4 Automation Systems with Central System Integrator

Database

User interface Workflow control layer

Scheduler Execution controller

Process control layer

Transportation management system

Robot remote center

Process control adapter system

PCS 1

PCS 2

ICS

Instrument layer Mobile robots

Mobile device

Hierarchical workflow management system (HWMS) Process management system (PMS)

Automated workstation 1

Transportation system

Automated workstation 1

Integrated system

Laboratory integration system

Figure 4.46 Architecture of the automated laboratory system (PCS: process control system, ICS: instrument control system). (Redrawn from [282].)

Figure 4.46 illustrates the components distributed in the hierarchical system of laboratory automation. This system comprises several automated workstations and integrated systems, mobile robots, and mobile devices for human laboratory assistants. As the highest level in this architecture, HWMS is located in the workflow control layer. It manages the components of automated laboratories according to the information, which is obtained from the administrator. On the process control layer, the transportation management system with the RRC and the laboratory automation systems with several process control systems (PCS) are located to adapt and integrate subsystems on the layers below. The lowest level is the instrument layer that includes mobile robots, mobile devices, automated workstations, and integrated systems with instrument control system (ICS). The HWMS consists of four function modules: user interface (UI), database, scheduler, and workflow execution controller (WEC). As the base of HWMS, the database manages the material core data of laboratory and workflow detail data. A workflow of heterogeneous subsystems can be modeled and modified by an abstracted material flow diagram on the UI. According to the detailed data of the workflow, an optimized task list is created by the scheduler and is executed by the WEC.

References 1 Hauptmann, S. and Mann, G. (1996) Stereochemie, Spektrum Akademischer

Verlag, Heidelberg, Berlin, Oxford. 2 Smith, S.W. (2009) Chiral toxicology: it’s the same thing…only different. Tox-

icol. Sci., 110 (1), 4–30.

References

3 Wynberg, H. and Feringa, B. (1976) Enantiomeric recognition and interac-

tions. Tetrahedron, 32 (22), 2831–2834. 4 Brunner, H. (1999) Rechts oder links. In der Natur und anderswo, Weinheim,

Wiley-VCH Verlag GmbH. 5 Cintas, P. (2007) Ursprünge und Entwicklung der Begriffe Chiralität und

Händigkeit in der chemischen Sprache. Angew. Chem., 119 (22), 4090–4099. 6 Podlech, J. (2001) Origin of organic molecules and biomolecular homochiral-

ity. Cell. Mol. Life Sci., 58 (1), 44–60. 7 Hellwich, K.-H. (2002) Stereochemie – Grundbegriffe, Springer, Berlin, Heidel-

berg. 8 Roth, H.J. (1998) Stereochemie und Arzneistoffe: Grundlagen-Betrachtungen-

Auswirkungen, Wissenschaftliche Verlagsgesellschaft mbH, Stuttgart. 9 Francotte, E. and Lindner, W. (2006) Chirality in Drug Research, Wiley-VCH

Verlag GmbH, Weinheim. 10 Federsel, H.-J. (1993) Chirale Arzneimittel. Scale-up, Kontrolle und deren

11 12 13 14 15 16 17 18 19

20

21

22 23

Bedeutung für die pharmazeutische Forschung. Chemie in unserer Zeit, 27 (2), 78–87. Frost & Sullivan (2001) World Chiral Separation Technology Markets: Market Study, Frost & Sullivan, London. Stinson, S.C. (2001) Chiral chemistry. Chem. Eng. News, 79 (20), 45–56. Global Industry Analysts (2008) Chiral Technology – Executive Summary: Market Study, Global Industry Analysts, San Jose, CA. Silverman, W.A. (2002) The schizophrenic career of a ’monster drug’. Pediatrics, 110 (2 I), 404–406. Fabro, S., Smith, R.L., and Williams, R.T. (1967) Toxicity and teratogenicity of optical isomers of thalidomide. Nature, 215 (5098), 296. Winter, W. and Frankus, E. (1992) Thalidomide enantiomers. Lancet, 339 (8789), 365. Roth, K. (2005) Eine unendliche chemische Geschichte. Chemie in Unserer Zeit, 39 (3), 212–217. Robinson, T. (1976) D-Amino acids in higher plants. Life Sci., 19 (8), 1097–1102. Forsum, O., Svennerstam, H., Ganeteg, U., and Nasholm, T. (2008) Capacities and constraints of amino acid utilization in Arabidopsis. New Phytol., 179 (4), 1058–1069. Sasamura, T., Matsuda, A., and Kokuba, Y. (1998) Tumor growth inhibition and nutritional effect of D-amino acid solution in AH109A hepatoma-bearing rats. J. Nutr. Sci. Vitaminol., 44 (1), 79–87. Esteban, S., Garau, C., Aparicio, S., Moranta, D., Barcelo, P., Fiol, M.A. et al. (2010) Chronic melatonin treatment and its precursor L-tryptophan improve the monoaminergic neurotransmission and related behavior in the aged rat brain. J. Pineal Res., 48 (2), 170–177. Bhushan, R. and Kumar, R. (2010) Enantioresolution of DL-penicillamine. Biomed. Chromatogr., 24 (1), 66–82. Tsai, G.C., Yang, P.C., Chung, L.C., Lange, N., and Coyle, J.T. (1998) D-Serine added to antipsychotics for the treatment of schizophrenia. Biol. Psychiatry, 44 (11), 1081–1089.

149

150

4 Automation Systems with Central System Integrator

24 Stinson, S.C. (1998) Counting on chiral drugs. Chem. Eng. News, 76 (38),

83–104. 25 Stinson, S.C. (2000) Chiral drugs. Chem. Eng. News, 78 (43), 55–78. 26 Stinson, S.C. (2001) Chiral pharmaceuticals. Chem. Eng. News, 79 (40),

79–97. 27 Rouhi, A.M. (2002) Chiral roundup – As pharmaceutical companies face

28 29 30

31

32

33

34

35 36 37

38

39 40

41

bleak prospects, their suppliers diligently tend the fertile fields of chiral chemistry in varied ways. Chem. Eng. News, 80 (23), 43–50. Friedman, M. (1999) Chemistry, nutrition, and microbiology of D-amino acids. J. Agric. Food, 47 (9), 3457–3479. Cherkin, A., Eckardt, M.J., and Gerbrandt, L.K. (1976) Memory – proline induces retrograde-amnesia in chicks. Science, 193 (4249), 242–244. Cherkin, A. and van Harreveld, A. (1978) L-Proline and related compounds – correlation of structure, amnesic potency and anti-spreading depression potency. Brain Res., 156 (2), 265–273. Cherkin, A., Davis, J.L., and Garman, M.W. (1978) D-Proline: stereospecificity and sodium chloride dependence of lethal convulsant activity in the chick. Pharmacol. Biochem. Behav., 8 (5), 623–625. Hamasu, K., Shigemi, K., Kabuki, Y., Tomonaga, S., Denbow, D.M., and Furuse, M. (2009) Central L-proline attenuates stress-induced dopamine and serotonin metabolism in the chick forebrain. Neurosci. Lett., 460 (1), 78–81. Hamasu, K., Shigemi, K., Tsuneyoshi, Y., Yamane, H., Sato, H., Denbow, D.M. et al. (2010) Intracerebroventricular injection of L-proline and D-proline induces sedative and hypnotic effects by different mechanisms under an acute stressful condition in chicks. Amino Acids, 38 (1), 57–64. Hudson, C., Hudson, S.P., Hecht, T., and MacKenzie, J. (2005) Protein source tryptophan versus pharmaceutical grade tryptophan as an efficacious treatment for chronic insomnia. Nutr. Neurosci., 8 (2), 121–127. Boman, B. (1988) L-Tryptophan: a rational anti-depressant and a natural hypnotic? Aust. N. Z. J. Psychiatry, 22 (1), 83–97. Faber, K. (1997) The ‘enantiomeric ratio’ – beware of confusion!. Enantiomer, 2 (5), 411–414. Selke, R., Facklam, C., Foken, H., and Heller, D. (1993) Application of the term relative enantioselectivity as useful measure for comparison of chiral catalysts, demonstrated on asymmetric hydrogenation of amino-acid precursors. Tetrahedron:Asymmetry, 4 (3), 369–382. Selke, R. (1997) Importance of the term ‘enantiomeric ratio’ for comparison of chiral catalysts in prochiral substrate reactions. Enantiomer, 2 (5), 415–419. Schurig, V. (1996) Terms for the quantitation of a mixture of stereoisomers. Enantiomer, 1 (2), 139–143. Pernot, P., Mothet, J.P., Schuvailo, O., Soldatkin, A., Pollegioni, L., Pilone, M. et al. (2008) Characterization of a yeast D-amino acid oxidase microbiosensor for D-serine detection in the central nervous system. Anal. Chem., 80 (5), 1589–1597. Brückner, H., Jaek, P., Langer, M., and Godel, H. (1992) Liquid chromatographic determination of D-amino acids in cheese and cow milk. Implication

References

42

43

44

45 46

47

48

49 50

51

52

53

54 55

of starter cultures, amino acid racemases, and rumen microorganisms on formation, and nutritional considerations. Amino Acids, 2 (3), 271–284. Brückner, H., Langer, M., Lüpke, M., Westhauser, T., and Godel, H. (1995) Liquid chromatographic determination of amino acid enantiomers by derivatization with o-phthaldialdehyde and chiral thiols: Applications with reference to food science. J. Chromatogr. A, 697 (1–2), 229–245. Brückner, H., Haasmann, S., and Friedrich, A. (1994) Quantification of D-amino acids in human urine using GC-MS and HPLC. Amino Acids, 6 (2), 205–211. Brückner, H., Westhauser, T., and Godel, H. (1995) Liquid chromatographic determination of D- and L-amino acids by derivatization with o-phthaldialdehyde and N-isobutyryl-L-cysteine: Applications with reference to the analysis of peptidic antibiotics, toxins, drugs and pharmaceutically used amino acids. J. Chromatogr. A, 711 (1), 201–215. Gördes, D., Kolukisaoglu, Ü., and Thurow, K. (2011) Uptake and conversion of D-amino acids in Arabidopsis thaliana. Amino Acids, 40 (2), 553–563. Gördes, D., Koch, G., Thurow, K., and Kolukisaoglu, Ü. (2013) Analyses of Arabidopsis ecotypes reveal metabolic diversity to convert D-amino acids. SpringerPlus, 2 (1), 1–11. Pätzold, R. and Bruckner, H. (2005) Mass spectrometric detection and formation of D-amino acids in processed plant saps, syrups, and fruit juice concentrates. J. Agric. Food. Chem., 53 (25), 9722–9729. Waldhier, M.C., Gruber, M.A., Dettmer, K., and Oefner, P.J. (2009) Capillary electrophoresis and column chromatography in biomedical chiral amino acid analysis. Aanl. Bioanal. Chem., 394, 695–706. Kirschner, D.L. and Green, T.K. (2009) Separation and sensitive detection of D-amino acids in biological matrices. J. Sep. Sci., 32 (13), 2305–2318. Grigorean, G., Ramirez, J., Ahn, S.H., and Lebrilla, C.B. (2000) A mass spectrometry method for the determination of enantiomeric excess in mixtures of D,L-amino acids. Anal. Chem., 72, 4275–4281. Tao, W.A., Zhang, D.X., Nikolaev, E.N., and Cooks, R.G. (2000) Copper(II)-assisted enantiomeric analysis of D,L-amino acids using the kinetic method: Chiral recognition and quantification in the gas phase. J. Am. Chem. Soc., 122, 10598–10609. Zhang, D.X., Tao, W.A., and Cooks, R.G. (2001) Chiral resolution of D-and L-amino acids by tandem massspectrometry of Ni(II)-bound trimeric complexes. Int. J. Mass spectrom., 204, 159–169. Schug, K.A. and Lindner, W. (2005) Stereoselective discrimination and quantification of arginine and N-blocked arginine enantiomers by formation and dissociation of calcium-mediated diastereomeric trimer complexes with a chiral reference compound using electrospray ionization-ion trap tan. J. Am. Soc. Mass. Spectrom., 16, 825–834. Vedejs, E. and Chen, X.H. (1997) Parallel kinetic resolution. J. Am. Chem. Soc., 119 (10), 2584–2585. Guo, J.H., Wu, J.Y., Siuzdak, G., and Finn, M.G. (1999) Measurement of enantiomeric excess by kinetic resolution and mass spectrometry. Angew. Chem. Int. Ed., 38 (12), 1755–1758.

151

152

4 Automation Systems with Central System Integrator

56 Horeau, A. (1977) in Stereochemistry: Fundamentals and Methods, vol. 3 (ed.

H.B. Kagan), Thieme, Stuttgard, pp. 51–94. 57 Fleischer, H. and Thurow, K. 2013 Innovative Software Solution for Special

58 59

60

61 62 63

64 65

66

67

68

69 70 71 72 73

Data Evaluation in Mass Spectrometry: Rapid Determination of Enantiomeric Excess of Chiral Compounds. IEEE International Instrumentation and Measurement Technology Conference, I2MTC 2013. p. 1624–1629. Bhushan, R. and Brückner, H. (2004) Marfey’s reagent for chiral amino acid analysis: a review. Amino Acids, 27 (3–4), 231–247. Fleischer, H., Gördes, D., and Thurow, K. (2009) High-throughput screening application for enantiomeric excess determination using ESI-MS. Am. Lab., 41 (7), 21–24. Fleischer, H. and Thurow, K. (2013) Fast mass spectrometry-based enantiomeric excess determination of proteinogenic amino acids. Amino Acids, 44 (3), 1039–1051. Fleischer, H. 2011 Ein hochdurchsatzfähiges Analysenmesssystem für die Bestimmung chiraler Komponenten. PhD thesis. Universität Rostock. Fleischer, H. and Thurow, K. (2011) Rapid enantiomeric excess determination of D- and L-proline using ESI-MS. Am. Lab., 43 (9), 32–36. Hentschel, M. 2011 Hochdurchsatzbestimmung multifunktionaler chiraler Wirkstoffe mittels ESI-Massenspektrometrie. Master thesis. Universität Rostock. Fleischer H. 2016 Automatisierungsstrategien in der analytischen Messtechnik. Habilitation thesis. Universität Rostock. Kim, S., Rodgers, R.P., Blakney, G.T., Hendrickson, C.L., and Marshall, A.G. (2009) Automated electrospray ionization FT-ICR mass spectrometry for petroleum analysis. J. Am. Soc. Mass. Spectrom., 20 (2), 263–268. Allwardt, A., Stoll, N., Wendler, C., and Thurow, K. (2004) New developments in reaction engineering – The HPMR 50-96 multiparallel reactor. Chem. Ing. Tech., 76 (11), 1679–1682. Allwardt, A., Stoll, N., Wendler, C., and Thurow, K. (2005) New developments in the field of reaction technology: the multiparallel reactor HPMR 50–96. J. Autom. Methods Manage. Chem., 2005, 26–30. Allwardt, A., Holzmüller-Laue, S., Wendler, C., and Stoll, N. (2007) Multiparallel high-pressure reaction in a 384 microplate. J. Assoc. Lab. Autom., 12 (6), 368–373. HighTech Business Decisions 2002 High-throughput Screening 2002: New Strategies and Technologies – Report; Market Study. Moraga; CA. Moll, F. (1930) Große Männer der Holzimprägnierungstechnik. Angew. Chem., 43 (37), 830–834. Moll, F. (1913) Der künstliche Schutz des Holzes durch Ätzsublimat (Kyanisierung). Angew. Chem., 26 (67), 459–463. Nowak, A. (1952) Die Kyanisierung von Leitungsmasten. Eur. J. Wood Wood Prod., 10 (1), 12–15. Theden, G. (1952) Über die Tiefe der Wirksamkeit einer Holzschutzbehandlung mit Quecksilberchlorid: 1. Mitteilung aus dem Materialprüfungsamt

References

74 75 76 77 78 79

80

81

82

83

84

85

86 87

88

89

Berlin-Dahlem über Quecksilberchlorid als Holzschutzmittel. Holzforschung – International Journal of the Biology, Chemistry, Physics, and Technology of Wood, 6 (1), 1–6. Moll, F. (1914) Zur Frage der Vergiftungsgefahr durch verdunstendes Sublimat. Angew. Chem., 27 (80–81), 559–560. Moll, F. (1927) Die Bedeutung des Sublimats als Holzimprägnationsmittel. Angew. Chem., 40 (41), 1137–1140. Daunderer, M. (1990) Handbuch der Umweltgifte – Klinische Umwelttoxikologie für die Praxis, Ecomed, Landsberg/Lech. Daunderer, M. (1981) Klinische Toxikologie: Giftinformation, Giftnachweis, Vergiftungstherapie, Ecomed, Landsberg/Lech. Jitaru, P. and Adams, F. (2004) Toxicity, sources and biogeochemical cycle of mercury. J. Phys. IV , 121, 185–193. Bakir, F., Damliji, S.F., Aminzaki, L., Murtadha, M., Khalidi, A., Alrawi, N.Y. et al. (1973) Methylmercury poisoning in Iraq: an interuniversity report. Science, 181 (4096), 230–241. Tatton, J.O.G. and Wagstaffe, P.J. (1969) Identification and determination of organomercurial fungicide residues by thin-layer and gas chromatography. J. Chromatogr., 44 (2), 284–289. Chatterjee, S., Pillai, A., and Gupta, V.K. (2002) Spectrophotometric determination of mercury in environmental sample and fungicides based on its complex with o-carboxy phenyl diazoamino p-azobenzene. Talanta, 57 (3), 461–465. Sunderland, E.M. and Chmura, G.L. (2000) An inventory of historical mercury emissions in Maritime Canada: implications for present and future contamination. Sci. Total Environ., 256 (1), 39–57. Taylor, C.G. and Tickle, W. (1969) Radiometric studies of mercury loss from fungicidal paints: 1. loss of phenyl mercuric acetate. J. Appl. Chem., 19 (1), 1–7. Taylor, C.G., Tickle, W., and Dwyer, A. (1969) Radiometric studies of mercury loss from fungicidal paints: 2. Comparison of 3 phenyl mercury compounds. J. Appl. Chem., 19 (1), 8–13. Taylor, C.G. and Hunter, G.H. (1972) Radiometric studies of mercury loss from fungicidal paints: 3. loss from 3 paints with various mercury contents. J. Appl. Chem. Biotechnol., 22 (6), 711–718. Stock, A. (1938) Endlich quecksilberfreie Haarfilzhüte!. Angew. Chem., 51 (2), 33–35. Blair, A.M.J.N., Clark, B., Clarke, A.J., and Wood, P. (1975) Tissue concentrations of Mercury after chronic dosing of squirrel monkeys with thiomersal. Toxicology, 3 (2), 171–176. Mutkus, L., Aschner, J.L., Syversen, T., Shanker, G., Sonnewald, U., and Aschner, M. (2005) In vitro uptake of glutamate in GLASTand GLT-1-transfected mutant CHO-K1 cells is inhibited by the ethylmercury-containing preservative thimerosal. Biol. Trace Elem. Res., 105 (1–3), 71–86. Tornow, E. (1932) Eine Schnellmethode zur Prüfung von Salben und Chemikalien auf Quecksilber. Angew. Chem., 45 (13), 707–708.

153

154

4 Automation Systems with Central System Integrator

90 Kijewski, H. (2012) Probleme bei der Beurteilung von Quecksilber-(Hg-)

Vergiftungen und Hg-Belastungen. Toxichem Krimtech, 79 (2), 61–70. 91 Lacerda, L.D. (1997) Global mercury emissions from gold and silver mining.

Water Air Soil Pollut., 97 (3–4), 209–221. 92 Lacerda, L.D. (2003) Updating global Hg emissions from small-scale gold

93

94

95

96

97 98 99

100 101 102 103

104

105 106

107

mining and assessing its environmental impacts. Environ. Geol., 43 (3), 308–314. Marvin-DiPasquale, M.C., Agee, J.L., Bouse, R.M., and Jaffe, B.E. (2003) Microbial cycling of mercury in contaminated pelagic and wetland sediments of San Pablo Bay, California. Environ. Geol., 43 (3), 260–267. Miller, J.R. and Lechler, P.J. (2003) Importance of temporal and spatial scale in the analysis of mercury transport and fate: an example from the Carson River system, Nevada. Environ. Geol., 43 (3), 315–325. Sexauer Gustin, M., Coolbaugh, M.F., Engle, M.A., Fitzgerald, B.C., Keislar, R.E., Lindberg, S.E. et al. (2003) Atmospheric mercury emissions from mine wastes and surrounding geologically enriched terrains. Environ. Geol., 43 (3), 339–351. Hojdová, M., Navrátil, T., Rohovec, J., Zak, K., Vanek, A., Chrastný, V. et al. (2011) Changes in mercury deposition in a mining and smelting region as recorded in tree rings. Water Air Soil Pollut., 216 (1–4), 73–82. Rytuba, J.J. (2003) Mercury from mineral deposits and potential environmental impact. Environ. Geol., 43 (3), 326–338. Cheng, H. and Hu, Y. (2012) Mercury in municipal solid waste in China and its control: a review. Environ. Sci. Technol., 46 (2), 593–605. Roussat, N., Mehu, J., Abdelghafour, M., and Brula, P. (2008) Leaching behaviour of hazardous demolition waste. Waste Manage. (Oxford), 28 (11), 2032–2040. Quick, J.C., Brill, T.C., and Tabet, D.E. (2003) Mercury in US coal: observations using the COALQUAL and ICR data. Environ. Geol., 43 (3), 247–259. Leiße, B. (1992) Holzschutzmittel im Einsatz: Bestandteile, Anwendungen, Umweltbelastungen, Bauverlag, Wiesbaden. Lay, J.-P. and Marutzky, R. (2003) Kompendium der Holzschutzmittelanalytik, Springer-VDI-Verl, Düsseldorf. Huber, J. and Leopold, K. (2013) Umweltmonitoring von Quecksilber in Gewässern: Reagenzienfreie und schnelle Bestimmung von Hg im Spurenbereich. GIT Labor Fachz., 57 (5), 306–309. Bundesministerium der Justiz Deutschland 2015 Altholzverordnung: Verordnung über Anforderungen an die Verwertung und Beseitigung von Altholz, http://www.gesetze-im-internet.de/bundesrecht/altholzv/gesamt.pdf (accessed 01 December 2016). Kolker, A., Orem, W., and Lechler, P. (2003) Introduction to mercury – special issue. Environ. Geol., 43 (3), 245–246. Varekamp, J.C. and Kreulen, B. (2003) Buchholtz ten Brink M.R., Mecray E.L. Mercury contamination chronologies from Connecticut wetlands and Long Island Sound sediments. Environ. Geol., 43 (3), 268–282. Loukola-Ruskeeniemi, K., Kantola, M., Halonen, T., Seppänen, K., Henttonen, P., Kallio, E. et al. (2003) Mercury-bearing black shales and

References

108

109 110

111 112 113 114 115

116

117 118

119

120 121 122 123 124 125

human Hg intake in eastern Finland: impact and mechanisms. Environ. Geol., 43 (3), 283–297. Gray, J.E., Greaves, I.A., Bustos, D.M., and Krabbenhoft, D.P. (2003) Mercury and methylmercury contents in mine-waste calcine, water, and sediment collected from the Palawan Quicksilver Mine, Philippines. Environ. Geol.., 43 (3), 298–307. Harada, M. (1995) Minamata disease – methylmercury poisoning in Japan caused by environmental pollution. Crit. Rev. Toxicol., 25 (1), 1–24. Enkelmann, R., Eschnauer, H., May, K., and SToeppler, M. (1984) Quecksilberspuren in Most und Wein. Fresenius Zeitschrift für Analytische Chemie, 317 (3–4), 478–480. Katz, S.A. and Katz, R.B. (1992) Use of hair analysis for evaluating mercury intoxication of the human body: a review. J. Appl. Toxicol., 12 (2), 79–84. Hammond, A.L. (1971) Mercury in the environment: natural and human factors. Science, 171 (3973), 788–789. Tölg, G. and Lorenz, I. (1977) Quecksilber – Ein Problemelement für den Menschen? Chemie in unserer Zeit, 11 (5), 150–156. Stock, A. (1928) Die Gefährlichkeit des Quecksilbers und der Amalgam-Zahnfüllungen. Angew Chem., 41 (24), 663–672. Yorifuji, T., Tsuda, T., Takao, S., Suzuki, E., and Harada, M. (2009) Total mercury content in hair and neurologic signs historic data from Minamata. Epidemiology, 20 (2), 188–193. Yorifuji, T., Tsuda, T., Inoue, S., Takao, S., and Harada, M. (2011) Long-term exposure to methylmercury and psychiatric symptoms in residents of Minamata. Japan. Environ. Int., 37 (5), 907–913. Stock, A. (1939) Die chronische Quecksilber- und Amalgamvergiftung. Zahnärztliche Rundschau., 48, 371–407. Jensen, G., Ruzickova, K., and Cameron, P. (2007) Quecksilber – Globale Belastungen und Gefährdung kindlicher Fähigkeiten. Bund für Umwelt und Naturschutz e.V. (BUND), Health Care Without Harm (HCWH) Europe, Health & Environment Alliance (HEAL), Berlin, Prag. Keil, D.E., Berger-Ritchie, J., and McMillin, G.A. (2011) Testing for toxic elements: a focus on arsenic, cadmium, lead, and mercury. Labmedicine, 42 (12), 735–742. Stock, A. (1935) Die Wirkung von Quecksilberdampf auf die oberen Luftwege. Die Naturwissenschaften, 23 (26/28), 453–456. Stock, A., Cucuel, F., and Kohle, H. (1933) Die Bestimmung kleinster Quecksilbermengen in organischem Material. Angew. Chem., 46 (13), 187–191. Stock, A. (1931) Die Bestimmung kleinster Quecksilbermengen und ihre Bedeutung. Die Naturwissenschaften, 19 (23/25), 499–502. Stock, A., Lux, H., Cucuel, F., and Kohle, H. (1933) Zur mikrometrischen Bestimmung kleinster Quecksilbermengen. Angew. Chem., 46 (3), 62–65. Stock, A. and Cucuel, F. (1934) Die Bestimmung des Quecksilbegehaltes in der Luft. Ber. Dtsch. Chem. Ges., 67, 122–127. Stock, A., Gerstner, F., and Kohle, H. (1932) Über die Oxydation des Quecksilbers durch Luft. Die Naturwissenschaften, 20 (51), 954–955.

155

156

4 Automation Systems with Central System Integrator

126 Stock, A. (1934) Über Verdampfung, Löslichkeit und Oxydation des metallis-

chen Quecksilbers. Z. Anorg. Allg. Chem., 217 (3), 241–253. 127 Stock, A. and Cucuel, F. (1934) Aufnahme und Verteilung des Quecksilbers

im Organismus. Angew. Chem., 47 (49), 801–806. 128 Stock, A. and Cucuel, F. (1934) Der Quecksilbergehalt der menschlichen

129

130 131

132 133

134

135

136

137

138

139

140

Ausscheidungen und des menschlichen Blutes. Angew. Chem., 47 (37), 641–647. Stock, A. and Neuenschwander-Lemmer, N. (1938) Die mikroanalytische Quecksilberbestimmung (XXVII. Mitteil. über Wirkung und Verbreitung des Quecksilbers). Ber. Dtsch. Chem. Ges., 71 (3), 550–562. Chilov, S. (1975) Determination of small amounts of Mercury. Talanta, 22 (3), 205–232. Theden, G. and Engelbrecht, G. (1952) Chemische Verfahren zum Sichtbarmachen von Quecksilberchlorid im Holz: 2. Mitteilung aus dem Materialprüfungsamt Berlin-Dahlem über Quecksilberchlorid als Holzschutzmittel. Holzforschung – Int. J. Biol., Chem., Phys., Technol. Wood, 6 (1), 1–6. Reimers, R.S., Burrows, W.D., and Krenkel, P.A. (1973) Total mercury analysis: review and critique. J. Water Pollut. Control Fed., 45 (5), 814–828. Segade, S.R. and Tyson, J.F. (2003) Determination of inorganic mercury and total mercury in biological and environmental samples by flow injection-cold vapor-atomic absorption spectrometry using sodium borohydride as the sole reducing agent. Spectrochim. Acta, Part B, 58 (5), 797–807. Serafimovski, I., Karadjova, I., Stafilov, T., and Cvetkovic, J. (2008) Determination of inorganic and methylmercury in fish by cold vapor atomic absorption spectrometry and inductively coupled plasma atomic emission spectrometry. Microchem. J., 89 (1), 42–47. Depoi, F.S., Bentlin, F.R.S., and Pozebon, D. (2010) Methodology for Hg determination in honey using cloud point extraction and cold vapour-inductively coupled plasma optical emission spectrometry. Anal. Methods, 2 (2), 180–185. Millan, R., Schmid, T., Sierra, M.J., Carrasco-Gil, S., Villadóniga, M., Rico, C. et al. (2011) Spatial variation of biological and pedological properties in an area affected by a metallurgical mercury plant: Almadenejos (Spain). Appl. Geochem., 26 (2), 174–181. Antes, F.G., Duarte, F.A., Mesko, M.F., Nunes, M.A.G., Pereira, V.A., Mueller, E.I. et al. (2010) Determination of toxic elements in coal by ICP-MS after digestion using microwave-induced combustion. Talanta, 83 (2), 364–369. Kaercher, L.E., Goldschmidt, F., Paniz, J.N.G., Flores, E.M.M., and Dressler, V.L. (2005) Determination of inorganic and total mercury by vapor generation atomic absorption spectrometry using different temperatures of the measurement cell. Spectrochim. Acta, Part B, 60 (5), 705–710. Wuilloud, J.C.A., Wuilloud, R.G., Silva, M.F., Olsina, R.A., and Martinez, L.D. (2002) Sensitive determination of mercury in tap water by cloud point extraction pre-concentration and flow injection-cold vapor-inductively coupled plasma optical emission spectrometry. Spectrochim. Acta, Part B, 57 (2), 365–374. Wuilloud, J.C., Wuilloud, R.G., Salonia, J.A., Olsina, R.A., and Martinez, L.D. (2002) Modified gas-liquid separator for the determination of mercury

References

141

142

143 144

145

146

147

148

149

150

151

152

153

154

with cold vapor flow injection-inductively coupled plasma optical emission spectrometry. J. Anal. Chem., 57 (9), 799–801. Lin, C., He, M., Zhou, Y., Hu, L., Guo, W., Quan, X. et al. (2007) Mercury contamination and dynamics in the sediment of the second Songhua river, China. Soil Sediment Contam., 16 (4), 397–411. Hamilton, M.A., Rode, P.W., Merchant, M.E., and Sneddon, J. (2008) Determination and comparison of heavy metals in selected seafood, water, vegetation and sediments by inductively coupled plasma-optical emission spectrometry from an industrialized and pristine waterway in Southwest Louisiana. Microchem. J., 88 (1), 52–55. Ure, A.M. (1975) Determination of Mercury by non-flame atomic-absorption and fluorescence spectrometry – review. Anal. Chim. Acta, 76 (1), 1–26. United States Environmental Protecion Agency (EPA) 2002 Mercury in Water by Oxydation, Purge and Trap, and Cold Vapor Atomic Fluorescence Spectrometry. EPA-821-R-02-019, Method 1631, Revision E. Deutsches Institut für Normung (2007) e.V. Wasserbeschaffenheit – Bestimmung von Quecksilber – Verfahren mittels Atomabsorptionsspektrometrie. DIN EN 1483, Beuth Verlag GmbH, Berlin. Silva, S.G., Oliveira, P.V., and Rocha, F.R.P. (2010) A green analytical procedure for determination of copper and iron in plant materials after cloud point extraction. J. Braz. Chem. Soc., 21 (2), 234–239. Bezerra, M.A., Arruda, M.A.Z., and Ferreira, S.L.C. (2005) Cloud point extraction as a procedure of separation and pre-concentration for metal determination using spectroanalytical techniques: A review. Appl. Spectrosc. Rev., 40 (4), 269–299. Deutsches Institut für Normung (1998) e.V. Bestimmung von Quecksilber – Verfahren nach Anreicherung durch Amalgamierung. DIN EN 12338, Beuth Verlag GmbH, Berlin. United States Environmental Protecion Agency (EPA) 2007 Mercury in Solids and Solutions by Thermal Decomposition, Amalgamation, and Atomic Absorption Spectrophotometry. Method 7473, Revision 0. Costa-Fernandez, J.M., Lunzer, F., Pereiro-Garcia, R., and Sanz-Mendel, A. (1995) Direct coupling of high-performance liquid-chromatography to microwave-induced plasma atomic emission spectrometry via volatile-species generation and its application to Mercury and Arsenic speciation. J. Anal. At. Spectrom., 10 (11), 1019–1025. Janzen, R., Sperling, M., Buscher, W., Bösel, A., Ehlbeck, J., Lischka, S. et al. (2011) Spezies-Analyse von Quecksilber-, Zinn- und Bleiverbindungen: Neues GC-MIP-OES-System. GIT Labor Fachz., 2011 (01), 22–27. Deutsches Institut für Normung (2004) e.V. Massenspektrometrie mit induktiv gekoppletem Plasma (ICP-MS) – Teil 1: Allgemeine Grundlagen und Begriffe. DIN 51002-1, Berlin, Beuth Verlag GmbH. Riss, P. and Lumet, L. (2013) Schleichender Tod durch flüssiges Silber?: Quecksilber in natürlichen Gewässern per ICP-MS bestimmen. Laborpraxis, 37 (7/8), 56–58. Millour, S., Noël, L., Kadar, A., Chekri, R., Vastel, C., and Guérin, T. (2011) Simultaneous analysis of 21 elements in foodstuffs by ICP-MS after

157

158

4 Automation Systems with Central System Integrator

155

156

157

158

159 160

161

162

163

164

165

166

closed-vessel microwave digestion: Method validation. J. Food Compos. Anal., 24 (1), 111–120. Nardi, E.P., Evangelista, F.S., Tormen, L., Sain´tPierre, T.D., Curtius, A.J., Souza, S.S. et al (2009) The use of inductively coupled plasma mass spectrometry (ICP-MS) for the determination of toxic and essential elements in different types of food samples. Food Chem., 112 (3), 727–732. Deutsches Institut für Normung (2007) e.V. Wasserbeschaffenheit – Anwendung der induktiv gekoppelten Plasma-Massenspektrometrie (ICP-MS) – Teil 1: Allgemeine Anleitung. DIN EN ISO 17294-1, Beuth Verlag GmbH, Berlin. Deutsches Institut für Normung (2005) e.V. Wasserbeschaffenheit – Anwendung der induktiv gekoppelten Plasma-Massenspektrsometrie (ICP-MS) – Teil 2: Bestimmung von 62 Elementen. DIN EN ISO 17294-2, Beuth Verlag GmbH, Berlin. Djedjibegovic, J., Larssen, T., Skrbo, A., Marjanovic, A., and Sober, M. (2012) Contents of cadmium, copper, mercury and lead in fish from the Neretva river (Bosnia and Herzegovina) determined by inductively coupled plasma mass spectrometry (ICP-MS). Food Chem., 131 (2), 469–476. Broekaert, J.A.C. (1992) Einsatz der ICP-Atomspektrometrie in der Wasseranalytik. Tech. Mess., 59 (4), 147–153. Khouzam, R.B., Lobinski, R., and Pohl, P. (2011) Multi-element analysis of bread, cheese, fruit and vegetables by double-focusing sector-field inductively coupled plasma mass spectrometry. Anal. Methods, 3 (9), 2115–2120. Zhao, Y., Zheng, J., Fang, L., Lin, Q., Wu, Y., Xue, Z. et al. (2012) Speciation analysis of mercury in natural water and fish samples by using capillary electrophoresis-inductively coupled plasma mass spectrometry. Talanta, 89, 280–285. Deutsches Institut für Normung (2006) e.V. Prüfung von Holzschutzmitteln – Nachweis von Holzschutzmitteln im Holz – Probenahme aus verbautem Holz. DIN 52161-1, Beuth Verlag GmbH, Berlin. Deutsches Institut für Normung (2007) e.V. Dauerhaftigkeit von Holz- und Holzprodukten – Mit Holzschutzmitteln behandeltes Vollholz – Teil 1: Klassifizierung der Schutzmitteleindringung und – aufnahme. DIN EN 351-1, Beuth Verlag GmbH, Berlin. Deutsches Institut für Normung (2007) e.V. Dauerhaftigkeit von Holz- und Holzprodukten – Mit Holzschutzmitteln behandeltes Vollholz – Teil 2: Leitfaden zur Probenentnahme für die Untersuchung des mit Holzschutzmitteln behandelten Holzes. DIN EN 351-2, Beuth Verlag GmbH, Berlin. United States Environmental Protecion Agency (EPA) (2009) Identification and Listing of Hazardous Waste 40 CFR §261.4(b): Exclusions: Solid Wastes which are Not Hazardous Wastes, United States Environmental Protecion Agency. United States Environmental Protecion Agency (EPA) 2001 EPA 260-B-01-004 – Emergency Planning and Community Righ-to-Know Act – Section 313: Guidance for Reporting Toxic Chemicals: Mercury and Mercury Compounds Category. EPA 260-B-01-004. Washington, DC: Office of Enviranmental Information.

References

167 Thy, P. and Jenkins, B.M. (2010) Mercury in biomass feedstock and combus-

tion residuals. Water Air Soil Pollut., 209 (1), 429–437. 168 United States Environmental Protecion Agency (EPA) 1997 Mercury Study

169

170

171 172

173

174

175

176 177

178 179 180 181

Report to Congress – Volume I: Executive Summary. EPA-452/R-97-003: Office of Air Quality Planning & Standards and Office of Research and Development. Kotz, L., Kaiser, G., Tschöpel, P., and Tölg, G. (1972) Aufschluss biologischer Matrices für die Bestimmung sehr niedriger Spurenelementgehalte bei begrenzter Einwaage mit Salpetersäure unter Druck in einem Teflongefäß. Fresenius Zeitschrift für Analytische Chemie, 260 (3), 207–209. Kotz, L., Henze, G., Kaiser, G., Pahlke, S., Veber, M., and Tölg, G. (1979) Wet mineralization of organic matrices in glassy carbon vessels in a pressure-bomb system for trace element analysis. Talanta, 26 (8), 681–691. Feng, Y. and Barrat, R.S. (1994) Digestion of dust samples in a microwave oven. Sci. Total Environ., 143 (2–3), 157–161. Fordham, P.J., Gramshaw, J.W., Castle, L., Crews, H.M., Thompson, D., Parry, S.J. et al. (1995) Determination of trace elements in food contact polymers by semi-quantitative inductively coupled plasma-mass spectrometry. Performance evaluation using alternative multi-element techniques and in-house polymer reference materials. J. Anal. At. Spectrom., 10 (4), 303–309. Pereira, J.S.F., Moraes, D.P., Antes, F.G., Diehl, L.O., Santos, M.F.P., Guimaraes, R.C.L. et al. (2010) Determination of metals and metalloids in light and heavy crude oil by ICP-MS after digestion by microwave-induced combustion. Microchem. J.., 96 (1, Special Issue SI), 4–11. Sucharova, J. and Suchara, I. (2006) Determination of 36 elements in plant reference materials with different Si contents by inductively coupled plasma mass spectrometry: Comparison of microwave digestions assisted by three types of digestion mixtures. Anal. Chim. Acta, 576 (2), 163–176. Zhou, C.Y., Wong, M.K., Koh, L.L., and Wee, Y.C. (1996) Comparison of acid mixtures in high-pressure microwave digestion methods for the determination of the total mercury in sediments by cold-vapor atomic absorption spectrometry. Anal. Sci., 12 (3), 471–476. Tölg, G. (1977) Zur Analytik von Spurenenlementen in biologischem Material. Fresenius Zeitschrift für Analytische Chemie, 283 (4), 257–267. Hassan, N.M., Rasmussen, P.E., Dabek-Zlotorzynska, E., Celo, V., and Chen, H. (2007) Analysis of environmental samples using microwave-assisted acid digestion and inductively coupled plasma mass spectrometry: maximizing total element recoveries. Water Air Soil Pollut., 178 (1–4), 323–334. Zunk, B. (1990) Mikrowellenaufschluss zur Bestimung von Spurenelementen in Pflanzenmaterial. Anal. Chim. Acta, 236 (2), 337–343. Gerdes, M. and Sengutta, U. (1998) Bestimmung von Quecksilber in Blut. Urin und Abwasser. Laborpraxis, 1998 (07/08), 32–36. Sengutta, U. (2003) Probenvorbereitung mit Mikrowellentechnologie. Laborpraxis, 2003 (06), x2–x3. Klinger, R. (2005) Der Mikrowellenaufschluss als richtungsweisendes Totalaufschlussverfahren in der geochemischen Analytik. Umweltwissenschaften und Schadstoff Forschung, 17 (3), 151–158.

159

160

4 Automation Systems with Central System Integrator

182 Krachler, M., Radner, H., and Irgolic, K.J. (1996) Microwave digestion meth-

183

184

185

186

187

188

189

190 191

192 193 194 195 196

197

ods for the determination of trace elements in brain and liver samples by inductively coupled plasma mass spectrometry. Fresenius J. Anal. Chem., 355 (2), 120–128. ISO 16772. 2004 Soil quality – Determination of mercury in aqua regia soil extracts with cold-vapour atomic spectrometry or cold-vapour atomic fluorescence spectrometry. ISO International Organization for Standardization, Vernier, Geneva, Switzerland. ISO 11466. 1995 Soil quality – Extraction of trace elements soluble in aqua regia. ISO International Organization for Standardization, Vernier, Geneva, Switzerland. ISO 12846. 2012 Water quality – Determination of mercury – Method using atomic absorption spectrometry (AAS) with and without enrichment. ISO International Organization for Standardization, Vernier, Geneva, Switzerland. United States Environmental Protecion Agency (EPA) 2015 EPA Method 7471B: Mercury in Solid or Semisolid Waste (Manual Cold-Vapor Technique). Deutsches Institut für Normung (2003) e.V. Charakterisierung von Abfällen – Aufschluss zur anschließenden Bestimmung des in Königswasser löslichen Anteils an Elementen in Abfällen. DIN EN 13657, Beuth Verlag GmbH, Berlin. Deutsches Institut für Normung (2014) e.V. Prüfung von Holzschutzmitteln – Nachweis von Holzschutzmitteln im Holz – Teil 7: Bestimmung des Gehaltes von kupferhaltigen Holzschutzmitteln. DIN 52161-7, Beuth Verlag GmbH, Berlin. Ahl, H., Fromm, J., and Melcher, E. (2013) Microwave assisted digestion as a preliminary method for the chemical analysis of inorganic wood preservative components. Eur. J. Wood Wood Prod., 71 (4), 533–535. Bowen, R.L. (1956) Use of epoxy resins in restorative materials. J. Dent. Res., 35 (3), 361–369. Paffenbarger, G.C., Nelsen, R.J., and Sweeney, W.T. (1953) Direct and indirect filling resins: a review of some physical and chemical properties. J.Am. Dent. Assoc., 47 (5), 516–524. Buonocore, M.G. (1955) A simple method of increasing the adhesion of acrylic filling materials to enamel surfaces. J. Dent. Res., 34 (6), 849–853. Bowen, R.L. (1963) Properties of a silica-reinforced polymer for dental restorations. J. Am. Dent. Assoc., 66 (1), 57–64. Bowen, R.L. (1979) Compatibility of various materials with oral tissues. I: the components in composite restorations. J. Dent. Res., 58 (5), 1493–1503. Tyas, M.J. and Burrow, M.F. (2004) Adhesive restorative materials: a review. Austr. Dent. J., 49 (3), 112–121. Buruiana, T., Melinte, V., Costin, G., and Buruiana, E.C. (2011) Synthesis and properties of liquid crystalline urethane methacrylates for dental composite applications. J. Polym. Sci., Part A: Polym. Chem., 49 (12), 2615–2626. Mjör, I.A., Moorhead, J.E., and Dahl, J.E. (1999) Selection of restorative materials in permanent teeth in general dental practice. Acta Odontol. Scand., 57 (5), 257–262.

References

198 Rogalewicz, R., Voelkel, A., and Kownacki, I. (2006) Application of HS-SPME

199 200

201

202

203

204

205

206

207 208

209 210

211

212

in the determination of potentially toxic organic compounds emitted from resin-based dental materials. J. Environ. Monit., 8 (3), 377. Roulet, J.-F. (1997) Benefits and disadvantages of tooth-coloured alternatives to amalgam. J. Dent., 25 (6), 459–473. Mjör, I.A. and Jokstad, A. (1993) Five-year study of Class II restorations in permanent teeth using amalgam, glass polyalkenoate (ionomer) cermet and resin-based composite materials. J. Dent., 21 (6), 338–343. Opdam, N.J.M., Bronkhorst, E.M., and Loomans, B.A.C. (2010) Huysmans M-CDNJM. 12-year survival of composite vs. amalgam restorations. J. Dent. Res., 89 (10), 1063–1067. Pallesen, U., van Dijken, J.W.V., Halken, J., Hallonsten, A.-L., and Höigaard, R. (2013) Longevity of posterior resin composite restorations in permanent teeth in Public Dental Health Service: a prospective 8 years follow up. J. Dent., 41 (4), 297–306. Tanaka, K., Taira, M., Shintani, H., Wakasa, K., and Yamaki, M. (1991) Residual monomers (TEGDMA and Bis-GMA) of a set visible-light-cured dental composite resin when immersed in water. J. Oral Rehabil., 18 (4), 353–362. Polydorou, O., Huberty, C., Wolkewitz, M., Bolek, R., Hellwig, E., and Kuemmerer, K. (2012) The effect of storage medium on the elution of monomers from composite materials. J. Biomed. Mater. Res. B: Appl. Biomater., 100B (1), 68–74. Finer, Y. and Santerre, J.P. (2004) The influence of resin chemistry on a dental composite’s biodegradation. J. Biomed. Mater. Res. Part A, 69A (2), 233–246. Geurtsen, W., Spahl, W., Muller, K., and Leyhausen, G. (1999) Aqueous extracts from dentin adhesives contain cytotoxic chemicals. J. Biomed. Mater. Res., 48 (6), 772–777. Geurtsen, W. (1998) Substances released from dental resin composites and glass ionomer cements. Eur. J. Oral Sci., 106 (2p2), 687–695. Söderholm, K.-J., Zigan, M., Ragan, M., Fischlschweiger, W., and Bergman, M. (1984) Hydrolytic degradation of dental composites. J. Dent. Res., 63 (10), 1248–1254. Geurtsen, W. (2000) Biocompatibility of resin-modified filling materials. Crit. Rev. Oral Biol. Med., 11 (3), 333–355. Polydorou, O., Koenig, A., Hellwig, E., and Kümmerer, K. (2009) Long-term release of monomers from modern dental-composite materials. Eur. J. Oral Sci., 117 (1), 68–75. Borchers, L., Eichner, K., and Kappert, H.F. (2005) Zahnärztliche Werkstoffe und ihre Verarbeitung, Band 1: Grundlagen und Verarbeitung, 8th.edn edn, Georg Thieme Verlag, Stuttgart. Fleischer, H., Vorberg, E., Warkentin, M., and Thurow, K. (2014) Qualitative and quantitative determination of methacrylates in dental filling materials. Am. Lab., 46 (2), 10–15.

161

162

4 Automation Systems with Central System Integrator

213 Song, J., van Velde, J.W., Vertommen, L.L.T., Smith, D.F., Heeren, R.M.A.,

214 215

216

217

218

219

220 221

222

223

224

225

226

227

and van den Brink, O.F. (2011) End-group analysis of methacrylic (co)polymers by LC-ESI-MS2. Macromolecules, 44 (6), 1319–1326. Stanley, H.R., Swerdlow, H., and Buonocore, M.G. (1967) Pulp reactions to anterior restorative materials. J. Am. Dent. Assoc., 75 (1), 132–141. Suarez, C.L., Stanley, H.R., and Gilmore, H.W. (1970) Histopathologic response of the human dental pulp to restorative resins. J. Am. Dent. Assoc., 80 (4), 793–800. Stanley, H.R., Going, R.E., and Chauncey, H.H. (1975) Human pulp response to acid pretreatment of dentin and to composite restauration. J. Am. Dent. Assoc., 91 (4), 817–825. Stanley, H.R., Bowen, R.L., and Folio, J. (1979) Compatibility of various materials with oral tissues. II: pulp responses to composite ingredients. J. Dent. Res., 58 (5), 1507–1517. Geurtsen, W., Spahl, W., and Leyhausen, G. (1998) Residual monomer/additive release and variability in cytotoxicity of light-curing glass-ionomer cements and compomers. J. Dent. Res., 77 (12), 2012–2019. Geurtsen, W., Lehmann, F., Spahl, W., and Leyhausen, G. (1998) Cytotoxicity of 35 dental resin composite monomers/additives in permanent 3T3 and three human primary fibroblast cultures. J. Biomed. Mater. Res., 41 (3), 474–480. Andrews, L.S. and Clary, J.J. (1986) Review of the toxicity of multifunctional acrylates. J. Toxicol. Environ. Health, 19 (2), 149–164. Hanks, C.T., Craig, R.G., Diehl, M.L., and Pashley, D.H. (1988) Cyto-toxicity of dental composites and other materials in a new in vitro device. J. Oral Pathol. Med., 17 (8), 396–403. Wataha, J.C., Hanks, C.T., Strawn, S.E., and Fat, J.C. (1994) Cytotoxicity of components of resins and other dental restorative materials. J. Oral Rehabil., 21 (4), 453–462. Spahl, W., Budzikiewicz, H., and Geurtsen, W. (1998) Determination of leachable components from four commercial dental composites by gas and liquid chromatography mass spectrometry. J. Dent., 26 (2), 137–145. Thurow, K., Koch, A., Maletz, R., and Behrend, D. (2004) Determination of monomer residues in dental materials using thermal desorption – mass spectrometry. Biomaterialien, 5 (S1), 48–49. Rogalewicz, R., Batko, K., and Voelkel, A. (2006) Identification of organic extractables from commercial resin-modified glass-ionomers using HPLC-MS. J. Environ. Monit., 8 (7), 750–758. Vorberg E. 2012 Entwicklung und Evaluation von analysenmesstechnischen Verfahren zur Bestimmung von Monomerrückständen in Dentalmaterialien. Master thesis. Universität Rostock. Fleischer, H., Marquardt, D., and Thurow, K. (2013) Quecksilberrückstände in Holzproben – Gehaltsbestimmung mittels ICP-OES und Kaltdampftechnik. GIT Labor Fachz., 57 (3), 168–170.

References

228 Fleischer, H. and Thurow, K. (2013) Determination of total mercury content

229

230

231

232

233

234

235

236 237

238 239

240

241 242

and trace metal analysis in wood materials Part 1: ICP-OES using mercury cold vapor. Am. Lab., 45 (8), 6–9. Fleischer, H. and Thurow, K. (2013) Determination of total mercury content in wood materials – Part 2: ICP-MS – a multi-element method. Am. Lab., 45 (9), 6–11. Vorberg, E 2015 Process automation for analytical measurements providing high precise sample preparation in life science applications. PhD thesis. Universität Rostock. Fleischer, H., Vorberg, E., and Thurow, K. (2014) Determination of total mercury content in wood materials – Part 3: miniaturization using ICP-MS. Am. Lab., 46 (6), 16–20. Vorberg, E, Thurow, K, Fleischer, H, Junginger, S and Stoll, N 2014 Automated Sample Preparation for Mercury Analysis in Wood Materials. 5th IMEKO TC19 Symposium on Environmental Instrumentation and Measurement:66–70. Vorberg, E., Fleischer, H., Junginger, S., Stoll, N., and Thurow, K. (2016) Automated sample preparation for mercury analysis in wood materials. IET Sci. Meas. Technol., 10 (5), 398–404. Vorberg, E., Fleischer, H., Junginger, S., Liu, H., Stoll, N., and Thurow, K. (2016) A highly flexible, automated system providing reliable sample preparation in element- and structure-specific measurements. J. Lab. Autom., 21 (5), 682–692. Kleinwächter, A 2012 Entwicklung einer flexiblen Roboterplattform für die Laborautomatisierung im Einzelgefäß-Handling. PhD thesis. Universität Rostock. Plebani, M. (2007) Laboratory medicine: Value for patients is the goal. Clin. Chem., 53 (10), 1873–1874. Jansen, K.L., Higgins, D.B., Wolf, D.P., and Bates, W.A. (1992) Laboratory automation. A strategic step in laboratory integration. Chemom. Intell. Lab. Syst., 17 (2), 171–180. Horwitz, W. (1982) Evaluation of analytical methods used for regulation of foods and drugs. Anal. Chem., 54 (1), 67A–76A. Blitti, K. 2016 Fully automated robotic sample preparation for elementand structure-analytic measurements. Master thesis. Rostock: University of Rostock. Blitti, K. 2016 Fully Automated Sample Preparation for the Determination of Calcium and Trace Metals in Biliary Endoprosthesis. Project thesis. Rostock: University of Rostock. Syringe Selection Guide 2012 Company Brochure Hamilton, Bonaduz (CH), http://www.hamiltoncompany.com/ (accessed 01 December 2016). Galiová, M.V., Fišáková, M.N., Kynický, J., Prokeš, L., Neff, H., Mason, A.Z. et al. (2013) Elemental mapping in fossil tooth root section of Ursus arctos by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). Talanta, 105, 235–243.

163

164

4 Automation Systems with Central System Integrator

243 Tandon, L., Iyengar, G.V., and Parr, R.M. (1998) A review of radiologi-

244

245

246 247

248

249

250

251

252 253 254

255

256

257 258

cally important trace elements in human bones. Appl. Radiat. Isot., 49 (8), 903–910. Yoshinaga, J., Suzuki, T., and Morita, M. (1989) Sex- and age-related variation in elemental concentrations of contemporary Japanese ribs. Sci. Total Environ., 79 (3), 209–221. Helliwell, T.R., Kelly, S.A., Walsh, H.P.J., Klenerman, L., Haines, J., Clark, R. et al. (1996) Elemental analysis of femoral bone from patients with fractured neck of femur or osteoarthrosis. Bone, 18 (2), 151–157. Faibish, D., Ott, S.M., and Boskey, A.L. (2006) Mineral changes in osteoporosis: a review. Clin. Orthop. Relat. Res., 443, 28–38. Gonnelli, S., Caffarelli, C., Tanzilli, L., Pondrelli, C., Lucani, B., Franci, B.M. et al. (2014) Effects of intravenous zoledronate and ibandronate on carotid intima-media thickness, lipids and FGF-23 in postmenopausal osteoporotic women. Bone, 61, 27–32. Srivastava, M., Mandal, S.K., Sengupta, S., Arshad, M., and Singh, M.M. (2008) Quantification of bone mineral density to define osteoporosis in rat. J. Endocrinol. Invest., 31 (5), 393–399. Roberts, N.B., Walsh, H.P.J., Klenerman, L., Kelly, S.A., and Helliwell, T.R. (1996) Determination of elements in human femoral bone using inductively coupled plasma atomic emission spectrometry and inductively coupled plasma mass spectrometry. J. Anal. At. Spectrom., 11 (2), 133–138. Burnell, J.M., Baylink, D.J., Chestnut, C.H. III, Mathews, M.W., and Teubner, E.J. (1982) Bone matrix and mineral abnormalities in postmenopausal osteoporosis. Metabolism, 31 (11), 1113–1120. Lappalainen, R., Knuuttila, M., Lammi, S., and Alhava, E.M. (1983) Fluoride content related to the elemental composition, mineral density and strength of bone in healthy and chronically diseased persons. J. Chronic Dis., 36 (10), 707–713. Boeing, H. and Müller, M.J. (2007) Ernährungsmedizinische Praxis: Methoden – Prävention – Behandlung, 2. edn, Springer, Heidelberg. Burckhardt, P. (2004) Screening für Osteoporose. Journal für Menopause, 11 (4), 5–8. Sievänen, H., Koskue, V., Rauhio, A., Kannus, P., Heinonen, A., and Vuori, I. (1998) Peripheral quantitative computed tomography in human long bones: Evaluation of in vitro and in vivo precision. J. Bone Miner. Res., 13 (5), 871–882. Karjalainen, P. (1973) A method for determination of the mineral content and mineral density of the distal radius using gamma ray attenuation. Ann. Clin. Res., 5 (4), 231–237. Bazin, D., Chappard, C., Combes, C., Carpentier, X., Rouzière, S., André, G. et al. (2009) Diffraction techniques and vibrational spectroscopy opportunities to characterise bones. Osteoporosis Int., 20 (6), 1065–1075. Zwanziger, H. (1989) The multielemental analysis of bone. A review. Biol. Trace Elem. Res., 19 (3), 195–232. Alhava, E.M., Olkkonen, H., Puittinen, J., and Nokso-Koivisto, V.M. (1977) Zinc content of human cancellous bone. Acta Orthop. Scand., 48 (1), 1–4.

References

259 Hisanaga, A., Hirata, M., Tanaka, A., Ishinishi, N., and Eguchi, Y. (1989)

260

261

262

263

264

265 266

267

268 269

270 271

272

273

Variation of trace metals in ancient and contemporary Japanese bones. Biol. Trace Elem. Res., 22 (3), 221–231. Ashoka, S., Peake, B.M., Bremner, G., Hageman, K.J., and Reid, M.R. (2009) Comparison of digestion methods for ICP-MS determination of trace elements in fish tissues. Anal. Chim. Acta, 653 (2), 191–199. Šˇcanˇcar, J., Milaˇciˇc, R., Benedik, M., and Bukovec, P. (2000) Determination of trace elements and calcium in bone of the human iliac crest by atomic absorption spectrometry. Clin. Chim. Acta, 293 (1–2), 187–197. Lambert, J.B., Vlasak, S.M., Thometz, A.C., and Buikstra, J.E. (1982) A comparative study of the chemical analysis of ribs and femurs in woodland populations. Am. J. Phys. Anthropol., 59 (3), 289–294. Yoshinaga, J., Suzuki, T., Morita, M., and Hayakawa, M. (1995) Trace elements in ribs of elderly people and elemental variation in the presence of chronic diseases. Sci. Total Environ., 162 (2–3), 239–252. Riond, J.-L., Goliat-von Fischer, I., Küffer, B., Toromanoff, A., and Forrer, R. (1998) Influence of the dosing frequency of parathyroid hormone-(1-38) on its anabolic effect in bone and on the balance of calcium, phosphorus and magnesium. Eur. J. Nutr., 37 (2), 183–189. Kuo, H.-W., Kuo, S.-M., Chou, C.-H., and Lee, T.-C. (2000) Determination of 14 elements in Taiwanese bones. Sci. Total Environ., 255 (1–3), 45–54. Kosugi, H., Hanihara, K., Suzuki, T., Himeno, S.-I., Kawabe, T., Hongo, T. et al. (1986) Elemental composition of ancient Japanese bones. Sci. Total Environ., 52 (1–2), 93–107. Becker, J.S., Boulyga, S.F., Pickhardt, C., Becker, J., Buddrus, S., and Przybylski, M. (2003) Determination of phosphorus in small amounts of protein samples by ICP-MS. Anal. Bioanal.Chem., 375 (4), 561–566. May, T.W. and Wiedmeyer, R.H. (1998) A table of polyatomic interferences in ICP-MS. At. Spectrosc., 19 (5), 150–155. Vanhoe, H., Goossens, J., Moens, L., and Dams, R. (1994) Spectral interferences encountered in the analysis of biological materials by inductively coupled plasma mass spectrometry. J. Anal. At. Spectrom., 9 (3), 177–185. Tan, S.H. and Horlick, G. (1987) Matrix-effect observations in inductively coupled plasma mass spectrometry. J. Anal. At. Spectrom., 2 (8), 745–763. McCurdy, E. and Woods, G. (2004) The application of collision/reaction cell inductively coupled plasma mass spectrometry to multi-element analysis in variable sample matrices, using He as a non-reactive cell gas. J. Anal. At. Spectrom., 19 (5), 607–615. Vanhaecke, F., Vandecasteele, C., and Dams, R. (1992) Inductively coupled plasma mas-spectrometryfor the determination of aluminium, calcium, chlorine, iron, magnesium; manganese, and sodium in fresh water. Anal. Lett., 25 (5), 919–936. Fleischer, H., Vorberg, E., Warkentin, M., Behrend, D., and Thurow, K 2014 Determination of Calcium and Phosphor in Bones Using Microwave Digestion and ICP-MS: Comparison of Manual and Automated Methods using ICP-MS. 5th IMEKO TC19 Symposium on Environmental Instrumentation and Measurement: 94–99.

165

166

4 Automation Systems with Central System Integrator

274 Ghandour, M., Liu, H., Stoll, N., Thurow and K. 2015 Improving the Nav-

275 276 277

278

279

280 281

282

igation of Indoor Mobile Robots Using Kalman Filter. IEEE International Instrumentation and Measurement Technology Conference, I2MTC 2015. p. 1434–1439. Liu, H. 2016 Intelligent strategies for mobile robotics in laboratory automation. Habilitation thesis. Universität Rostock. Liu, H., Stoll, N., Junginger, S., and Thurow, K. (2013) Mobile robot for life science automation. Int. J. Adv. Rob. Syst., 10, 1–14. Liu, H., Stoll, N., Junginger, S., and Thurow, K. (2014) A fast approach to arm blind grasping and placing for mobile robot transportation in laboratories. Int. J. Adv. Rob. Syst., 11 (1), 1–12: Article number 43. Liu, H., Stoll, N., Junginger, S. and Thurow, K. 2014 A new approach to battery power tracking and predicting for mobile robot transportation using wavelet decomposition and ANFIS networks. IEEE International Conference on Robotics and Biomimetics, ROBIO 2014. p. 253–258. Liu, H., Stoll, N., Junginger, S. and Thurow, K. 2012 A common wireless remote control system for mobile robots in laboratory. IEEE International Instrumentation and Measurement Technology Conference, I2MTC 2012. p. 688–693. Liu, H. 2013 Mobile robot transportation in laboratory automation. PhD thesis. Universität Rostock. Liu, H., Stoll, N., Junginger, S. and Thurow, K. 2013 A fast method for mobile robot transportation in life science automation. IEEE International Instrumentation and Measurement Technology Conference, I2MTC 2013. p. 238–242. Gu, X., Neubert, S., Stoll, N. and Thurow, K. 2016 Intelligent Scheduling Method for Life Science Automation Systems. IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems, MFI 2016. p. 156–161.

167

5 Automation Systems with Flexible Robots The use of classical laboratory devices within an automation system can be realized with the help of one or more flexible robots. In contrast to a robot, which acts only as a system integrator, the flexible robot performs transportation as well as manipulation tasks. This chapter introduces four different automation concepts based on flexible robots, including realized applications. These concepts include centralized and decentralized systems, which are customized for a specific application (closed) or which can be used for different applications due to additional system adaptations (open). For each of the four automation concepts, a detailed description of the system design is given including the used robot types, functional units, and peripheral devices. In addition, the different processes are introduced with the single process steps. A very brief outline is given regarding data evaluation. Detailed information regarding the concepts and realized software for a flexible, automated evaluation of measurement data is included in Chapter 6.

5.1 Centralized Closed Automation System In this chapter, a centralized closed automation system is described, which uses a fix positioned dual arm robot as a working robot (see Figure 5.1). The robot is surrounded by a workbench containing all classical laboratory equipment required for the process. Besides the transportation tasks, the robot also performs the direct manipulation of the samples as well as the sample transfer to the integrated analytical measurement systems. The system is proprietarily configured and programmed for the application. The automated determination of chiral compounds (see Sections 4.1.1 and 4.1.2) will be described as an example of the application. 5.1.1 System Design

Besides the flexible robot, the automation system contains the functional units “sample storage,” “analytics,” and “safety.” The functional units “liquid handling” and “sample treatment,” which are included in automation systems with central system integrators, are realized by the flexible robot, which does not only integrate the different devices, but also manipulates the samples. Figure 5.2 shows the automation system from different views. Automation Solutions for Analytical Measurements: Concepts and Applications, First Edition. Heidi Fleischer and Kerstin Thurow. © 2018 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2018 by Wiley-VCH Verlag GmbH & Co. KGaA.

168

5 Automation Systems with Flexible Robots

Centralized/closed

SP

Figure 5.1 Schematic visualization of a centralized closed automation system with a flexible robot (SP: sample preparation, A: analytical measurement, FR: flexible robot).

FR A

Flexible robot: A dual-arm robot SDA10F with a control unit FS100 (Yaskawa, Kitaky¯ush¯u) performs the manipulation and transport tasks. The humanoid design enables a great spectrum of application possibilities for this industrial robot, including assembly and transfer, feeding of machines, packing, and other handling tasks. Due to the possibility of human-like motion sequences, this robot is also interesting for applications in laboratory automation. Each arm has 7 degrees of freedom (DOF) resulting in a total of 15 degrees of freedom including the basic rotation axis (see Table 5.1). The maximum payload for each arm is 10 kg. Each arm can be used for different tasks; in addition, synchronized motions of both arms are possible (synchronized move). The repeatability is ±0.1 mm. Each arm is equipped with a gripper of type LEHF 20K2-48-R86P5 (SMC Pneumatik, Egelsbach). Different motion modes are available for robot programming. The joint motion is a synchronous point-to-point motion (PTP), where the movement times of the axes are adjusted to the slowest axis. This ensures that all axes stop their motion simultaneously. Velocities of 0.01–100% (joint velocity) can be used. This mode is not suitable in a laboratory environment with numerous devices and tools. The tool center point (TCP) reaches exactly the destination point, the pathway is uncontrolled. The pathway depends on the selected joint velocity and is – similar to axis velocities and acceleration – not predictable. This can cause collisions with the laboratory equipment. Cartesian pathway control (continuous path, CP) in the form of linear or circular motions is suitable for work in narrow environments, in workpiece machining, or for the manipulation and handling of samples in laboratory automation. The TCP is moving on a predefined line or circular orbit, whereas path way velocities between 0.1 and 1500 mm/s (or 0–9000 cm/min) are possible. In contrast to PTP motions, the CP motions require more joint movements. The dual arm robot SDA10 can perform transportation as well as manipulation tasks. For the use of classical laboratory equipment, which is also used in manual processes, special adapters might be necessary. An example is pipetting with manual pipettes. To enable secure handling of the pipettes by the gripper, a special adapter has been designed for each pipette, which is used for pipette gripping by the robot and as a holder of the pipette on the automation system. Figure 5.3 shows the adapter for the manual pipettes, the pipette holder on the automation system, and the pipetting process using both robot arms.

5.1 Centralized Closed Automation System

3 2 5a

1

6 5c 4

5b

(a)

8

1

9 6

13 11 7 10 12 (b)

Figure 5.2 Centralized closed automation system with flexible robot – (a) overall view, (b) top ¯ u), ¯ 2: LC-MS system (Agilent Technologies, view, 1: dual-arm robot SDA10F (Yaskawa, Kitakyush Waldbronn), 3: GC-MS-MS system (Agilent Technologies, Waldbronn), 4: workbench with chemical resistant coating, 5: light curtain (a: transmitter, b: deflection mirror, c: receiver), 6: storage system with two levels, 7: ultrasonic bath RK31 (Bandelin Electronic, Berlin), 8: thermomixer Comfort (Eppendorf, Hamburg), 9: ALPs for samples, reagents, solvents, and labware, 10: shaker Teleshake (Variomag, Daytona Beach), 11: holder for HPLC auto sampler tray, 12: waste container, 13: holder for pipettes [1]. (Reproduced with permission of Sage Publications.)

169

170

5 Automation Systems with Flexible Robots

Table 5.1 Work range and velocities of the main and hand axes of the SDA10 [2, 3]. Axis

Maximum work range (∘ )

Maximum velocity (∘ /s)

Main axis Rotation axis (basis axis)

±170

130

S axis (lifting)

±180

170

L axis (lower arm)a)

±110

170

E axis (lower arm, twist)

±170

170

U axis (upper arm)a)

±135

170

±180

200

Hand axis R axis (upper arm, twist)a) B axis (wrist pitch/yaw)

±110

200

T axis (wrist twist)

±180

400

a) According to the position the work range of the L, U, and R axis can be limited.

(a)

(b)

(c)

Figure 5.3 Pipetting with the dual-arm robot SDA10 – (a) manual pipette with adapter, (b) pipettes on the holder of the automation system, (c) pipetting process using both robot arms [1]. (Reproduced with permission of Sage Publications.)

5.1 Centralized Closed Automation System

(a)

(b)

Figure 5.4 Gripper with different fingers – (a) finger for grabbing microplates and single vials at left robot arm, (b) finger for grabbing single vials, lids, manual pipettes, and glass pipettes at right robot arm [1]. (Reproduced with permission of Sage Publications.)

The grippers of the robot have to be adapted according to the tasks. The two grippers have been equipped with different fingers to ensure handling of standardized plates and racks in microplate format as well as non-standardized laboratory equipment such as syringes, filters, and many more (see Figure 5.4). The left hand is composed of two metal fingers on a plastic socket and is used for handling of microplates and single vials. The right hand with two plastic fingers can grab single vials and their lids, variable pipettes, and glass pipettes, and can operate the laboratory equipment. This includes removing and applying the thermomixer lid, the opening and closing of the HPLC auto sampler door, and the transport of the HPLC auto sampler tray. The tray of the HPLC auto sampler (removable drawer for two microplates) is equipped with an adapter to enable the robot-based transfer of the samples in the 96 well microplate format into the HPLC system. Figure 5.5 shows a CAD layout of the realized tray and the dual arm robot during tray insertion into the HPLC system. Sample storage: All required standard and sample solutions, reagents, and solvents for the determination of the chiral compounds are provided to the system in glass containers or microplates. The stock solutions and reagent cannot be provided in microplates due to the quantity required during sample processing. They have to be stored in single vials with volumes of 4–22 ml. Special racks have been developed in the microplate format, which hold the different vessels.

171

172

5 Automation Systems with Flexible Robots

(a)

(b)

(c)

Figure 5.5 Insertion of samples into the HPLC auto sampler with the dual-arm robot SDA10 – (a) CAD layout of microplate tray with adapter and gripper, (b) realized tray with adapter, microplate, and cover, (c) insertion process [1]. (Reproduced with permission of Sage Publications.)

In addition, racks for glass pipettes have been designed to provide new pipettes for the processes and as waste positions for used glass pipettes. Figure 5.6 shows a variety of different racks for the analytical determination of chiral compounds. All labware, standard and sample solutions, reagents and solvents are positioned on predefined positions in the storage system. This includes two levels. The lower level on the workbench has different shelves, whereas on the second level, positions are available for providing pipette tips on automated laboratory positioners (ALPs) (see Figure 5.2). Analytics: The analytical measurement of the enantiomeric excess was realized with an LC-TOF mass spectrometer (Agilent Technologies, Waldbronn). The samples to be analyzed are injected with an HPLC auto sampler as discussed in Section 4.1. The sample transfer to the HPLC system is performed by the robot. The data is automatically evaluated in online mode (see Chapter 6). Figure 5.7 shows an overview of the setup of the analytical measurement system. Safety: Since the movement of the robot during automation can cause damage to humans and subject matters, the whole system is covered with a light

5.1 Centralized Closed Automation System

(a)

(c)

(e)

(b)

(d)

(f)

Figure 5.6 Racks in microplate format for the handling of sample vessels with different volumes and glass pipettes – above: CAD design, below: realized racks with containers; (a,b) rack for 12 glass vials with screw-on lid (vol. 4 ml); (c,d) rack for 12 glass vials with screw-on lid (vol. 22 ml); (e,f ) rack for 12 glass pipettes with dispensing bulb [1]. (Reproduced with permission of Sage Publications.)

1 2 3 8

4 5

7

6

Figure 5.7 Measuring system HPLC-TOF-MS (1: solvent storage system, 2: vacuum degasser, 3: binary pump, 4: high-performance auto sampler, 5: column oven, 6: optical diode array detector, 7: time-of-flight mass spectrometer, 8: computer with control software) [1]. (Reproduced with permission of Sage Publications.)

curtain as a safety feature (see Figure 5.2). The optical sensors used in the light curtain enable a touchless function (contactless active protective device). The used light curtain OY245S (ifm electronic, Essen) includes an optical transmitter and receiver unit, whereas there are several transmitting elements in the transmitter unit and a corresponding number of receiver elements in the receiver

173

174

5 Automation Systems with Flexible Robots

unit. Infrared light with 950 nm wavelength is used. The protection height is 760 mm, the resolution reaches 30 mm [4]. If one or more light beams are interrupted, safe outputs are shut off and the robot is stopped (output signal switching device, OSSD). Transmitter and receiver unit are combined over a deflection mirror, which results in two orthogonal radiation fields. 5.1.2 Process Description

All required samples, standard solutions, reagents, solvents, and labware are provided manually to the storing positions of the system. The flexible robot transports them to predefined ALPs on the workbench (see Figure 5.2). The dual arm robot executes the preparation of five calibration solutions of pure amino acid enantiomers. Therefore, the standard solutions are diluted (1 : 50, v/v) and five solutions with −100 ee%, −50 ee%, 0 ee%, +50 ee%, and +100 ee% are pipetted. From the weighted solid compounds (L-FDVA and D-FDLA), an equimolar auxiliary solution is prepared by fourfold addition of solvents and a transfer into a vessel. Similarly to the manual process, glass pipettes are used. Subsequently, the preparation of the reaction solutions on a 96 well microplate is executed by sequential addition of the chiral compounds (amino acids), pseudo enantiomeric mass tagged auxiliaries (equimolar mixture of L-FDVA and D-FDLA), and sodium hydrogen carbonate (NaHCO3 ). After dosing is completed, the derivatization is performed on a thermomixer Comfort (Eppendorf, Hamburg). After 1 h of reaction time, quenching is realized with concentrated hydrochloric acid (HCl). The solutions are homogenized in the microplate with the thermomixer or alternatively on a non-temperated shaker Teleshake (Variomag, Daytona Beach). The sample plates are transferred by the flexible dual arm robot to the HPLC auto sampler, whereby the door of the auto sampler is opened and closed by the robot (Figure 5.5) [5]. Data acquisition and data evaluation are executed as discussed in Section 4.1; a detailed description can be found in Chapter 6. Figure 5.8 shows the process description chart with all necessary process steps. 5.1.3 Control of the Automation System

In general, robots can be programmed online or offline. In online programming mode, the desired positions (coordinates) can be directly or with the help of a separate teach arm approached and saved (teaching). The created programs are quite inflexible, since the different motion steps are predefined with the specific positions. In case of changes in the process description or the position of different samples, this can lead to a high effort for program adaptation. In contrast, offline programming using a programming environment and a 3D model of the robot enables higher flexibility, if the programs are created in a modular way [6–8]. A modular structure means the encapsulation of different repeating motion sequences using position variables and relative position relations. Fixed positioning of the different samples, labware, chemicals, and laboratory devices in the robot periphery on the workbench is the requirement for using this type of programming. The goal is flexible programming, which enables the repeated use

5.1 Centralized Closed Automation System

Start Initial transport Supply of samples, stock solutions, reagents, solvents, and labware

Transport Storage system → ALPs on workbench (samples, stock solutions, reagents, solvents, and labware)

Pipetting to single vials Dilution of stock solutions

Pipetting to single vials Calibration solutions

Pipetting to microplate Calibration solutions, sample solutions, auxiliary solutions, sodium bicarbonate

Closing Put cover on microplate

Transport ALP → Thermoshaker (microplate)

Derivatization 1 h, 20 °C/68 °F, 750 rpm

Transport Thermoshaker → ALP (microplate)

Opening Remove cover from microplate

Transport ALP → Shaker (microplate)

Homogenization 1 min, 750 rpm

Transport Shaker → ALP (microplate)

Opening Remove cover from microplate

Transport ALP → ALP with tray (microplate)

Closing Put cover on microplate

Transport ALP with tray → Autosampler (microplate on tray)

Analysis Measurements using ESI-TOF-MS

Pipetting to microplate Reaction quenching using hydrochloric acid

Pipetting to single vials Auxiliary solutions

Closing Put cover on microplate

Pipetting to microplate Sample dilution

Transport Autosampler → ALP with tray (microplate on tray)

Transport ALP with tray → Storage system (used microplate) ALPs → Storage system (used labware)

Final transport Storage/disposal/cleaning of samples, stock solutions, reagents, solvents, and labware

End

Figure 5.8 Process flow chart for the automated analytical determination of chiral amino acids.

175

176

5 Automation Systems with Flexible Robots

of often required tasks at different positions. This includes, for example, pipetting with manual flexible pipettes and glass pipettes, the transport of microplates, the removing and applying of lids and caps, the dropping of the used tips and the transport of labware from and to the ALPs on the workbench. The described automation system is controlled with the scheduling software “SAMI Workstation Ex Version 4.1” (Beckman Coulter, Krefeld). Therefore, the programs (jobs) for the robot control, which have to be flexible, must be integrated. This is explained with the example of a simple pipetting process. The hardcoded pipetting is only possible for fixed positions. A change in the source or destination labware, the pipette type, the positions of the microplates and glass vials or the number of pipetting steps is not possible. Thus, new jobs have to be created for different parameters, leading to an enormous number of robot jobs. If these parameters are transferred as variables in flexible, modular programming, the highest possible flexibility and reusability of the job is guaranteed. Thus, one module only is necessary for the pipetting process, in which the requested parameters (source and destination labware, pipette type, position of microplate wells or glass vials, number of pipetting steps) have to be defined. To enable this flexibility, the dual arm robot is programmed on the controller FS100 (Yaskawa, Kitaky¯ush¯u) in the menu-driven programming language Inform III and with the “Motoman SDK” in the programming language C. Task-specific robot motion modules have been defined and implemented on the basis of motion frames [8]. A motion frame is a sequence of motion steps; whose parameters can be defined with position variables. Using relative variables enables the execution at any position (e.g., on different ALPs). This requires a hardware-based reference point (e.g., on an ALP) to initialize the motion frame and calculate the absolute positions from the relative positions. Figure 5.9 shows the structure of a motion frame. With the help of motion frames and modular programming, transport and manipulation tasks can be flexibly realized. For automated processing, these program modules are integrated into the scheduling software “SAMI Workstation EX Version 4.1” (Beckman Coulter, Krefeld). Figure 5.10 shows a SAMI method for the sample preparation of chiral amino acids. The automated control of the measuring instrument requires the integration of the HPLC-MS system into the modular software “SAMI Workstation Ex.” Thus,

Motion frame

Step 1

Motion steps Step 2 Variable interface

Reference point

Step x

Figure 5.9 General structure of a motion frame.

5.1 Centralized Closed Automation System

Figure 5.10 Example for an automated method in “SAMI Method Editor” (sample preparation of chiral amino acids).

a software interface for the communication between the process control software and the measuring instrument was developed. The communication between the “SAMI Workstation Ex” and the interface is realized over TCP/IP. Since the interface is directly installed on the control PC of the measuring instrument, the communication between the interface und the instrument software “MassHunter Acquisition” (Agilent Technologies, Waldbronn) can be realized over a Windows API. The communication between “MassHunter” and the measuring instrument is also realized over TCP/IP. Figure 5.11 visualizes this principle. The software interface consists of three modules (see Figure 5.12) [5]. The XML module enables the establishment of a socket connection between SAMI and

177

178

5 Automation Systems with Flexible Robots

SAMI Workstation Ex

Figure 5.11 Communication between the process control software and the measuring instrument LC-MS.

Ethernet TCP/IP

Interface Windows API MassHunter workstation

Ethernet TCP/IP Mass spectrometer

Program interface XML module • Socket connection • Parsing XML files • Generating XML files

Ethernet TCP/IP

SAMI Workstation Ex

XML files

LC-MS module • Generating worklist • Load worklist • Load analysis method • Start measurement process • Stop measurement process • Read device status

LC-MS user interface Mouse Keyboard Enter sample information using • Buttons • Menus • Text boxes

Figure 5.12 General structure of the software interface for the integration of the LC-MS system. (Redrawn from Ref. [5].)

5.1 Centralized Closed Automation System

the interface. XML files are used for communication, which include orders from SAMI to the measuring instrument as well as status messages from the analytical device to SAMI. The LC module creates the worklist (table with sample parameters for the measurement). This worklist can be transferred to the measuring instrument and can be loaded there. In addition, measuring methods (instrument parameters for the measurement) can be loaded. The main tasks of this module include the starting and stopping of a method and the readout and return of the actual device status (standby, ready, running, error). The third module is the liquid chromatography–mass spectrometry (LC-MS) user interface. This module enables the generation of a worklist using Excel (Microsoft, Redmond). This increases the flexibility, since the user is not bound to the device computer. The interface translates the Excel table into an XML file, which is compatible to the device software “MassHunter Acquisition” [5]. 5.1.4 Results

The automation system processes the sample fully automatically from the sample preparation to the analytical measurement and data evaluation. Besides sample preparation for the determination of chiral amino acids (described in Section 4.1), the system with the flexible robot is also used for further method development. The precision achieved using the automation system is comparable to the manual results [8–10]. The coefficients of variation in the repeatability for the intensity ratios M1 /M2 were determined between 0.27% and 1.45%; the respective values for the mass intensities M1 and M2 varied between 0.82% and 6.47%. The time requirement for the manual method as well as for the automated methods with a system integrator and the flexible robot were determined (rounded up to full minutes). For a better comparison, the required times for sequential processing of the samples are summarized in Table 5.2 [8–10]. In the manual method as well as in the automated method with a flexible robot all solutions are newly prepared. The sample preparation on the automation system with a system integrator uses prepared calibration and auxiliary solutions Table 5.2 Time requirement for preparation and analysis of 96 samples using manual method, automated method with system integrator, and automated method with flexible robot (no overlapping sample preparation, *samples and reagents are prepared). Process step

Time (min) Automated with flexible robot

Manual

Automated with central system integrator

Sample preparation

17

4*

58

Derivatization

60

60

60

Sample post processing

14

4*

23

Analytical measurement using TOF-MS

132

132

132

Total time (min)

223

200

273

179

180

5 Automation Systems with Flexible Robots

(see Section 4.1). In the manual method only two microplates with 96 wells can be processed in one working day (8 h) whereas the use of the automation systems enables the processing of 4–7 microplates per day (2.688–4.704 samples per week). Higher throughputs can be achieved using an intelligent scheduler by interleaving the sample preparation and analytical measurement process steps. This scheduling is realized using the “SAMI Workstation Ex.”

5.2 Centralized Open Automation System To extend the system described above for the processing of additional applications, the required laboratory devices and measuring instruments have to be flexibly integrated. This leads to a centralized open automation system (see Figure 5.13). Besides the introduced application of the determination of chiral compounds (see Section 5.1), the determination of cholesterol in biliary stents using gas chromatography–mass spectrometry (GC-MS) measurement should be performed. This requires the integration of another measuring system and different laboratory devices (e.g., an ultrasonic bath and a shaker). In addition, adaptation of the used labware is required to enable automated handling using the flexible dual arm robot SDA10. 5.2.1 Background and Applicative Scope

The treatment of diseases of the bile and the biliary tract takes up a major portion in the field of gastroenterology. These diseases are characterized by high complexity as well as high morbidity and mortality [11]. The most common biliary disorders include gallstones (cholelithiasis) and the inflammation of gallbladder (cholecystitis), biliary tract (cholangitis), and pancreas (pancreatitis). Gallstones are crystalline concretions, which are formed of bile components [12]. A distinction is made between different types of gallstones: cholesterol stones, calcium bilirubinate gallstones (also called pigment stones due to their dark color), and mixed stone types. The cholesterol content depends on the type of the stone and the values range from 10% to 80%. Gallstones can be found in approximately 10–15% of the adult population, whereby women are affected twice as frequently as men. This disease is Figure 5.13 Schematic visualization of a centralized open automation system with a flexible robot (SP 1,2,…,X: sample preparation, A 1,2,…,X: analytical measurement, FR: flexible robot).

Centralized/open SP 2 SP X

SP 1 FR

AX

A1 A2

5.2 Centralized Open Automation System

particularly common in Western industrialized countries. A higher incidence rate of up to 70% (e.g., Pima Indians) was observed in the indigenous population of America. A significantly lower incidence rate (0–10%) of this disease was found in East Asia, Sub-Saharan Africa, and in African-Americans, which reflects a different diet, and environmental and genetic factors as possible causes [12, 13]. In addition to medical treatment, the complete removal of the gall bladder (cholecystectomy) is the most common form of therapy [12]. Gallstones [14] or chronic lesions in consequence of gallstones, acute or chronic inflammation of the pancreas or biliary tract, the removal of the gall bladder, (benign) tumors or radiotherapy may result in a narrowing (stenosis) or the closure (obstruction) of the biliary tract. A study showed that bile duct carcinomas (adenocarcinomas) or carcinoma of the pancreas (pancreatic cancer) can cause malignant strictures, which occur more frequently (87% of 236 patients) than the benign stenoses (13% of 236 patients) [15]. Such stenoses or obstructions in the biliary system require the use of biliary endoprostheses or biliary stents, to ensure the drainage of the bile [16]. These have been used since the late 1970s and have become an established method in the following decade [17–21]. The endoscopic implantation of biliary stent rarely leads to complications and the mortality rate is very low [22, 23]. However, this form of therapy leads to a major problem: inside the stents, deposits may accumulate to a complete closure, which requires a replacement of the stent and thus a new medical procedure for the patient [16, 24–26]. Dowidar et al. reported the closure of stents at 65 of 196 patients, which is about 33% and the most common complication [24]. The parameters that support the formation and accumulation of incrustations inside the stent are still unclear [27]. An analysis of the stent incrustations showed that they mainly consisted of bacteria, yeast cells, and plant fibers (from food residues). For this analysis, light microscopy, scanning electron microscopy, and transmission electron microscopy were used and bacterial cultures were prepared from the stent incrustations [15]. Furthermore, certain bacteria such as Escherichia coli speed up the closure process due to the formation of the enzyme β-glucuronidase. This enzyme causes the elimination of bilirubin and hence the precipitation of calcium bilirubinate [16]. However, in contrast to the influence of bacteria, the influence of additional factors such as the design and material of the stents is not fully understood [16]. Groen et al. determined the composition of 21 occluded biliary endoprostheses and the deposits contained 25% protein and 20% non-soluble components (mainly plant fibers). The material was enriched with unconjugated bilirubin (non-water-soluble form of bilirubin). The main components of the bile (bile acids and lecithin) showed only 15%. In addition, cholesterol crystals were found, but none of the three types of gallstones. Gel electrophoresis, biological enzyme-based assays, and scanning electron microscopy were used for the determination of the composition of the stent incrustations [28]. Moesch et al. determined the composition of stent incrustations using microbiological and physico-chemical methods. The stents were endoscopically removed between 21 and 673 days after implantation. Stereomicroscopy, scanning electron microscopy, and Fourier transform infrared spectroscopy (FTIR) were used.

181

182

5 Automation Systems with Flexible Robots

The last technique was used for the determination of calcium palmitate. The results showed contents of 0–75% of the dry matter. Further components were: calcium bilirubinate with 0–75%, unconjugated calcium bilirubinate with 0–15%, calcium phosphate with 0–10%, cholesterol with 5% (in only one sample), and proteins with 10–70% [29]. Cetta et al. performed similar studies using the same measurement procedures. Comparable results were obtained. Cholesterol contents up to 10% of dry matter were determined in all samples [30]. A study from 2006 involves the examination of 31 biliary stents made of ethylene-vinyl acetate copolymer (EVA), two stents made of polyethylene (PE) and two of polytetrafluoroethylene (PTFE) [25]. Hydrophilic polymer-coated polyurethane (HCPC) is another material used for biliary endoprostheses [31]. The stents were divided into three parts: the upper segment closest to the liver, the medium segment, and the lower segment closest to the duodenum. The stent incrustations were measured as separate samples. Due to the high number of stent samples made of EVA, it was assumed that these materials support the generation of deposits more than stents made of PE or PTFE [25]. Some authors suggest that the stent closure is initiated by adsorption of one component at the stent inner wall and that a stepwise accumulation takes place up to complete closure [25, 28, 30]. Stents made of PE and EVA had a hardly detachable black layer directly on the stent inner wall, while the stents made of PTFE did not show this phenomenon. Due to the small number of investigated stents, no generalization was made by the authors [25]. The components of the closure material were determined using FTIR spectroscopy and pyrolysis derivatization gas chromatography mass spectrometry (Py-GC-MS). Similar results to those in the studies mentioned above were obtained. Additionally, the content of cholesterol of the stent material was determined as well. The results show that cholesterol not only anneals on the surface, but also diffuses into the material [25]. Since the role of cholesterol in closure of the biliary stent is not fully understood, a specific qualitative and quantitative analysis of this compound is required. Gas chromatography (GC) is the most common analytical technique used for the determination of cholesterol. A derivatization (e.g., silylation) of the samples is often required using a mass selective detector. Depending on the derivatization reagent tert-butyldimethylsilyl (TBDMS) or trimethylsilyl (TMS) derivatives are generated and analyzed. The determination of cholesterol and its precursors in plasma of rats [32] as well as of cholesterol and its oxidation products recovered from bleaching earth from oil refining operations [33] are application examples using sample derivatization. A high-temperature method for the GC-MS with a maximum temperature of 380 ∘ C using TMS derivatives is described by Son et al. [34]. Another mass spectrometry based method is the isotope dilution (ID), which is commonly used for the determination of cholesterol in human serum [35–37]. Langlois and Kuipers describe a rapid determination of cholesterol in natural waxes using silylation of the samples and gas chromatography-flame ionization detection (GC-FID) [38]. In contrast, Dinh et al. described a simple method for quantifying cholesterol in meat and meat products using GC-FID without evaporation or derivatization [39]. Such simplified and fast analytical methods are suitable for their use in fully automated systems.

5.2 Centralized Open Automation System

Cholesterol and related compounds can be also determined using liquid chromatography (LC). The LC-based methods show advantages over some GC-based methods. No sample derivatization is required before analysis. In addition, LC separation columns have a higher tolerance for critical matrices such as blood plasma, in which cholesterol is often to be determined [40]. Different detector types such as MS, UV, or evaporative light scattering detector (ELSD) are used. Fu and Joseph describe various LC-based methods for the determination of cholesterol in human plasma by LC-ELSD and LC-MS-MS using chemical ionization at atmospheric pressure chemical ionization (APCI). Cholesterol and lathosterol have identical molecular weights. To distinguish between both compounds, tandem mass spectrometry is used (LC-MS-MS) [40]. 5.2.2 Automation Goals

The biliary endoprosthesis is surgically removed from the patient and stored in formaldehyde. The incrustations to be investigated are located inside of the stent. Therefore, the stents are first cut in pieces and halved for easy removal of the content. The incrustations are dried at room temperature to constant weight, finely ground with a mortar, and stored in plastic containers. The sample preparation includes the addition of solvent, followed by the extraction of the components of the incrustations using an ultrasonic bath. Subsequently, the samples are filtered, an internal standard is added and the solutions are finally diluted for the subsequent analysis using GC-MS. Optionally, a derivatization of the sample can be carried out. Alternatively, the quantification can be performed using GC-FID for undiluted samples [41]. In this application, almost all subprocesses can be highly automated. The removal of incrustations from stents, drying, and weighing require special equipment or can be alternatively carried out manually. All other process steps can be automated. The subprocess “pipetting” includes the addition of the solvent and the internal standard. The subprocess “extraction” requires the integration of an ultrasonic bath so that the sample vials (vol. 2 ml) can be placed and removed automatically using a robot. The subprocess “filtration” can also be automated. Depending on the robot type, filtration units or common manual syringe filters can be used. The final subprocess in the sample preparation is the “dilution” of the samples prior to analysis. A reformatting is not necessary here, since one sample vial type is used in the entire sample preparation and analysis. If the analysis is performed using GC-FID, the dilution step can be omitted. Figure 5.14 illustrates the overall process. The transport tasks include the initial transport to provide ground and weighed incrustation samples, standard solutions, solvent, and labware. Further transport tasks are the transfer of the sample solutions to/from the ultrasonic bath and of the final sample solutions into the autosampler of the measurement instrument (GC-MS or GC-FID). If a derivatization is carried out, a transport to/from the thermomixer is required. The final transport includes the storage and disposal of measured sample solutions, residual solutions, residual solvents, and used labware.

183

184

5 Automation Systems with Flexible Robots

Stent sample

Cut in pieces, removal of incrustations, drying, fine milling, weighing

Pipetting solvent

Extraction in ultrasonic bath

Filtration

Data evaluation Dilution Pipetting internal standard

Pipetting reagents

Measurement using GC-MS Sample storage or disposal

Derivatization Measurement using GC-FID

Material flow Information flow

Data evaluation

Figure 5.14 Process workflow for the determination of cholesterol in biliary endoprostheses using GC-MS and GC-FID.

The automation of this process eliminates human errors such as pipetting errors or errors by inadvertent interchanging of sample vessels. Due to the integration of measurement systems in the automation system, 24/7 operation can be realized and the sample throughput per day increased. 5.2.3 System Design

Besides the flexible robot, the automation system contains the functional units “sample storage,” “analytics,” and “safety.” For the realization of an open structure, the task fields of the flexible robot and the functional units “sample storage” and “analytics” have been extended. Flexible robot and sample storage: These two functional units are described together, since the adaptation of the labware and the corresponding racks are closely related to the subprocesses, which have to be performed by the dual arm robot SDA10. The general system design is similar to the setup in Section 5.1, but was extended [1, 7]. In addition to the glass vials (4–22 ml volume), the robot also handles the much smaller GC vials (1.5 ml volume) including the screwing of the required lids. To ensure a completely automated process and efficient handling with the dual arm robot, adaptation of the corresponding racks in microplate format is required (see Figure 5.15). This includes a rack for the transportation of 12 GC vials and a rack for transportation and placing down of the lids. In addition, the solutions in the GC vials are treated in an ultrasonic bath, which requires a fixed positioning in the water-filled metal tub. Therefore, a frame was constructed, which can be fixed to the metal tub and can hold up to 12 GC vials (see Figure 5.15a,b). An additional process is the filtration using manual disposable syringes and filters. Special racks for holding and providing the required material have been realized in microplate format. The rack for manual disposable syringes allows exact alignment of the syringes using an integrated notch. This enables easy handling of the syringes by the robot using the sideways flaps for additional stabilization. In addition, racks for the cannulas and the filters have been designed. Figure 5.16 shows the CAD layout and the realized racks with the specific labware. The racks

5.2 Centralized Open Automation System

(a)

(c)

(e)

(b)

(d)

(f)

(A)

(B)

Figure 5.15 Labware holder and racks in microplate format for the handling of sample vials and lids – (A) CAD layout, (B) racks with vials and lids; (a) holder for GC vials in the ultrasonic bath, (b) ultrasonic bath with installed holder and GC vials; (c,d) rack for GC vials; (e,f ) rack for lids with several diameters. (Redrawn from Ref. [1].)

(a)

(c)

(e)

(b)

(d)

(f)

(A)

(B)

Figure 5.16 Automated filtration – racks in microplate format for the handling of disposable syringes, cannulas, and filters ((A) CAD layout; (B) realized racks with labware); (a,b) rack for syringes; (c,d) rack for cannulas; (e,f ) rack for filters. (Redrawn from Ref. [1].)

are manually equipped with the labware and stored in the required amount in the storage system of the automation system. Analytics: The analytical measurement of the prepared stent samples uses a GC-QQQ mass spectrometer (Agilent Technologies, Waldbronn). A GC auto sampler and GC injector execute the transfer of the sample material. The samples are placed into the measurement system by the flexible dual arm robot. The data

185

186

5 Automation Systems with Flexible Robots

is automatically evaluated after the completion of the measurement series (see Chapter 6). 5.2.4 Process Description

As an initial transport step, all weighted solid samples, stock solutions, reagents, solvents, and the required labware are provided on both levels of the storage system. The dual arm robot transfers them to the predefined positions (ALPs) on the workbench. Dichloromethane (CH2 Cl2 ) is added to stent incrustations and the vials are closed with appropriate lids. If a rack is available on the shaker, the single samples are transported to this rack; otherwise the whole rack with the GC vials is transferred to the shaker by the dual arm robot (see Figure 5.17a). After completion of the homogenization step, the GC vials are transferred into the ultrasonic bath. The sonication enables an additional surface enlargement of the milled solid samples in preparation for the following extraction. The ultrasonic bath is automatically operated by the right hand of the robot, which is moving the turning knob into the required direction (see Figure 5.17b). Once the sonication procedure is completed, the vials are transported to the shaker and agitated for 15 min to realize an extraction of the soluble material. Subsequently, the lids are removed from the vials and the open vials are transported to the predefined ALP. The next subprocess includes the filtration of the samples. Therefore, the same laboratory equipment which is used in manual processing (disposable syringes and syringe filters) can be used. The left hand of the robot picks the disposable syringes. The sample is aspirated with the help of the right hand, which operates the syringe piston. In a next step, the cannula is removed from the syringe at a special waste position and a filter is taken. Both robot arms are used for the filtration process followed by disposal of the syringe with the filter in the waste container. Figure 5.18 visualizes the robot based filtration. In the next step, the samples are diluted with dichloromethane (CH2 Cl2 ) for the following analytical measurement, the vials are closed with screwing caps and transported individually to the predefined positions in the GC auto sampler. From here, the samples are transported automatically to the GC injector

(a)

(b)

Figure 5.17 Operation of laboratory devices by the dual arm robot ((a) turn on/turn off of the shaker using a push button, (b) adjusting the sonication time using a turning knob).

5.2 Centralized Open Automation System

(a)

(b)

(c)

(d)

Figure 5.18 Automated filtration process (a: picking of the syringe cannulas, b: aspiration of the samples to be filtered, c: removal of the used cannula, d: picking of the filter).

and are injected into the measuring device. Figure 5.19 shows the related process flow chart. 5.2.5 Control of the Automation System

The system is controlled according to Section 5.1. A separate method was created in “SAMI Workstation Ex.” Thus, optional processing of the determination of chiral compounds using TOF-MS or the determination of the cholesterol content in incrustations of biliary stents is possible. The automation system has been extended to a central open system. Since a GC-MS system performs the determination of cholesterol, it must be integrated into the control software to enable fully automated processing. The software interface introduced in Section 5.1 was adapted for the communication between the process control software and the scheduling software “SAMI Workstation Ex.” Due to the different versions of the instrument software “MassHunter Workstation” for LC-MS and GC-MS, a new specialized software interface had to be implemented [42]. In the case of the LC-MS, the binary pump, a high-performance auto sampler, a column oven, and the mass spectrometer have to be controlled. In the case of the GC-MS, the gas chromatograph and the mass spectrometer must be controlled. The differences include not only the XML-based commands but also different notations such as worklist (LC-MS) and sequence (GC-MS) for the measurement table with sample information as well as status definitions (idle, busy, initializing, prerun, run, postrun, error).

187

188

5 Automation Systems with Flexible Robots

Start Initial transport Supply of samples, stock solutions, reagents, solvents and labware

Transport Storage system → ALPs on workbench (samples, stock solutions, reagents, solvents, and labware)

no

Pipetting Add solvent to samples

yes

Closing Screw caps on GC vials

no

yes Empty rack on ALP ?

Empty rack in shaker ?

Transport ALP → Shaker (rack with GC vials)

Transport ALP → Shaker (GC vials)

Homogenization Shake sample solutions

Transport Shaker → Ultrasonic bath (GC vials)

Transport Shaker → ALP (rack with GC-Vials)

Set variable vial_number = 1

Set variable vial_number = 1

Do while

Do while

vial_number ≤ 12 ?

no

vial_number ≤ 12 ?

yes

no

yes

Sonication

Grasping Grasp vial from rack

Grasping Grasp vial from shaker

Transport Ultrasonic bath → Shaker (GC vials)

Opening Remove caps from GC vials (GC-Vials)

Opening Remove caps from GC vials (GC-Vials)

Extraction Shake sample solutions

Drop vial Put vial in rack on ALP

Drop vial Put vial in rack on ALP

Set variable vial_number = vial_number + 1

Set variable vial_number = vial_number + 1

Filtration Using manual syringe filters

Pipetting Add internal standard, sample dilution

Transport ALP → GC autosampler (GC vials)

Closing Screw caps on GC vials

Analysis Measurements using GC-MS

Transport GC autosampler → Storage system (GC vials) Labware on ALPs → Storage system (used labware)

Final transport Storage/disposal/cleaning of samples, stock solutions, reagents, solvents, and labware

End

Figure 5.19 Process flow chart for the automated determination of cholesterol in incrustations of biliary endoprosthesis.

5.2 Centralized Open Automation System

Figure 5.20 Process flow chart for the communication between the interface and the software “SAMI Workstation Ex.”

Start

Run SAMI Workstation EX

Enter IP address and port number

no

Connected SAMI - interface ? yes Receive commands Send status messages

Disconnected SAMI - Interface ?

no

yes End

In addition, the layouts of the sample tables are different. This means that the table columns are differently arranged compared to the LC-MS software version. Figure 5.20 shows the communication process between the “SAMI Workstation Ex” and the measuring instrument GC-MS. 5.2.6 Results

The calibration for the determination of cholesterol is realized with six data points between 0.5 and 3 mg/l with three replicates each (see Figure 5.21). The calibration solutions can be prepared by the dual arm robot. In general, a calibration is only required for new methods, after changes in the system setup (e.g., changing chromatographic columns), or after long periods of using the devices for other applications. A calibration is thus not considered a daily routine procedure. To correct possible concentration variations that occur due to sample preparation and sample delivery to the analysis system, an internal standard is added to the calibration solutions and the samples (α-cholestan). Figure 5.22 shows the EI mass spectra of the standard solution with cholesterol and α-cholestan.

189

5 Automation Systems with Flexible Robots

y = 1.260363E-004 * x −0.029947 R2 = 0.99865590

× 10−1 3.5 3 Relative responses

190

2.5 2 1.5 1 0.5 500

750

1000 1250 1500 1750 2000 2250 2500 2750 3000 Concentration (ng/ml)

Figure 5.21 GC-MS calibration for cholesterol between 0.5 and 3 mg/l.

× 106

+EI scan (rt: 15.561 min)

4.5 77.9 4 3.5 3 2.5

90.9

2

144.9 118.9

1.5

206.8

158.9

66.9

275.0

1

301.0

386.0

230.9

353.0

0.5 0 60

80

100

120

140

160

180

(a)

200

220

240

260

280

300

320

340

360

380

400

420

400

420

Counts versus mass-to-charge (m/z)

× 10

6

+EI scan (rt: 12.561 min)

3.2 3 2.8

216.9

2.6 2.4 2.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2

148.9 94.9 66.9 357.0

120.9

372.1 174.9

232.0

262.0 314.8

0 60

(b)

80

100

120

140

160

180

200

220

240

260

280

300

320

340

360

380

Counts versus mass-to-charge (m/z)

Figure 5.22 EI mass spectra: (a) cholesterol and (b) α-cholestan (internal standard).

5.3 Decentralized Automation System

Table 5.3 Time requirement for the manual and automated sample preparation with flexible robot (no overlapping sample preparation, *no rinsing steps). Process step

Number of samples

Time (min) Manual

Automated with flexible robot

12

24

80

Extraction

15

15

Analytical measurement using GC-MS* [41]

12 × 26

12 × 26

351

407

Sample pre- and post-processing [7, 8]

Total time (min)

Due to the characteristic fragmentation pattern obtained with EI ionization, a database can be used for the qualitative determination of the compounds (e.g., NIST14 with NIST Mass Spectral Search, Agilent Technologies, Waldbronn). The signals of a quantifier ion and three qualifier ions are used for the quantitative determination. Using an HP5MS column (Agilent Technologies, Waldbronn, Germany) a CV value of 6.72% for repeatability was achieved. The recovery rate varied between 90% and 115%. Alternatively, a derivatization of the samples using a separation column DB17 (Agilent Technologies, Waldbronn, Germany) can also be applied. The time requirement for the manual method and the automated method with a flexible robot was determined (rounded up to full minutes). For a better comparison, the required times for sequential processing are summarized in Table 5.3 [8–10]. Higher throughputs are possible with intelligent scheduling. Scheduling is realized using the “SAMI Workstation Ex.” Reasons for the different processing times include the actual pipetting scheme of the flexible robot using manual pipettes with fixed predefined volumes. Different steps are required to pipette the requested volume. This requires the use of different pipettes including the grasping of the pipette and the tips, the pipetting process, the dropping of the tips, and the placing of the pipette. In contrast, the manual operator can adjust the required volume at one or two pipettes. Thus, the number of required pipetting steps is significantly higher for the automated processing. In order to reduce the processing time, electronically adjustable pipettes can be used [7]. In general, it should be demonstrated that full automation using classical manual laboratory equipment is possible.

5.3 Decentralized Automation System If sample transport is additionally realized to and from the automation platform of the flexible dual arm robot, a decentralized automation system can be created (see Figure 5.23). Sample preparation and analytical measurement of biliary stents will be used as an example for a centralized closed automation system. In contrast to the concepts introduced in Sections 5.1 and 5.2, the analytical measurements are extended to elemental analysis. Sample preparation and

191

192

5 Automation Systems with Flexible Robots

Decentralized/closed SP

FR

Figure 5.23 Schematic visualization of a decentralized closed automation system with a flexible robot (SP: sample preparation, A: analytical measurement, FR: flexible robot, CE: connection element).

Mobile CE A

Decentralized/open SP 1

SP 2

SP X

FR

FR

FR

Figure 5.24 Schematic visualization of a decentralized open automation system with a flexible robot (SP 1,2,…,X: sample preparation, A 1,2,…,X: analytical measurement, FR: flexible robot, CE: connection element).

Mobile CE

A1

A2

AX

the following analytical determination of calcium using inductively coupled plasma–mass spectrometry (ICP-MS) (see Section 4.3) and the determination of cholesterol are performed (see Section 5.2). The required microwave digestion for the calcium measurement was not integrated on the SDA10 platform due to safety aspects. The ICP-MS system is also not integrated with the flexible dual arm robot. Thus, transport of the samples from the SDA10 platform to these substations is required. The execution of more applications results in the formation of a decentralized open automation system (see Figure 5.24). The application area can be extended with the application of the analytical investigation of bones. In this case, sample preparation for the determination of calcium and phosphorus in bones using ICP-MS (elemental analysis) as well as sample preparation for the structural determination of steroid hormones using GC-MS is required [43]. A further extension is possible, if automation systems with a central system integrator are included as well. Due to the combination of all stations by a mobile robot, full automation of all samples and measuring techniques is possible. 5.3.1 System Design

Besides the flexible robot, the decentral automation system contains the functional units “sample storage,” “mobile robotics,” “analytics,” and “safety.” Some functional units exist in different subsystems due to the decentral system concept. Sample storage: The storage systems of the automation systems described in Sections 4.3 and 5.2 are used. An additional central sample transfer station

5.3 Decentralized Automation System

is necessary, from which the initial transport of the samples to the different automation systems (elemental or structural analytical measuring system) can be performed by mobile robots. In addition, the previously discussed manual initial transport for chemicals and labware provision can be realized by mobile robots. Mobile robotics: The transport using mobile robots follows Section 4.3. The integrated subsystems are distributed to different floors of the life science laboratory. While sample preparation and structural analysis are integrated on one floor, microwave and ICP-MS for elemental analysis are located on a different floor. This requires an extension of the maps for mobile robot navigation and the integration of the elevator into the mobile robot control center [44, 45]. Analytics: The described decentralized automation concept includes structural analytical measurements (GC-MS, see Section 5.2) and elemental analytical measurements (ICP-MS, see Section 4.3). If both measurement processes are required, intelligent and efficient data evaluation and data management are of special importance (see Chapter 6). 5.3.2 Process Description

In a first step, a decision has to be made regarding the type of analytical measurements required for the investigation of incrustations of biliary stents. Here, elemental analytical measurements are available (determination of the calcium content using ICP-MS) as well as structural analytical measurements (determination of cholesterol using GC-MS) (see Figure 5.25). If both measurements are required, parallel processing is possible due to the decentralized structure of the automation system. The samples, reagents, and the required labware are provided in the storage system of the automated sample preparation system. Once the preparation is completed, the selected method(s) will be executed. If additional applications are integrated into the automation, the result is an open system structure. The following process description chart shows the flexible execution of five applications, whose subsystems are distributed to different laboratories (see Figure 5.26). 5.3.3 Control of the Automation System

The different automated substations are controlled by the SAMI software (Beckman Coulter, Krefeld). Process management is realized using a hierarchical workflow management system (HWMS) analogous to Section 4.4. Due to the open system structure, in particular, a suitable sample management system is necessary. This includes the sample reception station, from where the samples can be transported to the dedicated subsystems by the mobile robots. On the other hand, logistical sample management using a database is essential. The sample data can be managed and the type of analytical measurement can be defined. This can include the analytical investigation using one or more analytical measurement systems (elemental analysis, structural analysis, high and low volatile analytes, etc.). Each analytical investigation starts usually with a weighing step that can be performed manually or automatically. This data must also be provided to the system and assigned to specific investigations, since these values are used for data

193

194

5 Automation Systems with Flexible Robots

Start

Elemental analysis

Kind of sample preparation ?

Structural analysis

Determination of calcium (elemental analysis) Microwave digestion, dilution

Determination of cholesterol (structural analysis) Extraction, dilution

Transport Sample preparation → ICP-MS

Transport Sample preparation → GC-MS

Analysis Measurement using ICP-MS

Analysis Measurement using GC-MS

Final transport Storage/disposal/cleaning of samples, stock solutions, reagents, solvents, and labware

End

Figure 5.25 Process flow chart for the flexible processing of an application (investigation of incrustations of biliary stents) with different measurement methods in a decentralized closed automation system.

evaluation at the end of the processes. Mobile devices can be used for these tasks (see Chapters 6 and 7).

5.4 Automation Systems with Integrated Robotics The next level for increasing the flexibility of automated life science processes is based on the concept of integrated robotics using a completely decentralized system structure. A mobile and flexible robot (=integrated robot) is used, which

5.4 Automation Systems with Integrated Robotics

Start Dental materials

Wood materials

Kind of sample ?

Incrustations of biliary endoprosthesis

Bone samples Chiral compounds

Elemental analysis

Kind of sample preparation ?

Elemental analysis

Structural analysis

Kind of sample preparation ?

Structural analysis

Determination of calcium (elemental analysis) Microwave digestion, dilution

Determination of cholesterol (structural analysis) Extraction, derivatization, dilution

Determination of phosphorous/ calcium (elemental analysis) Microwave digestion, dilution

Determination of steroid hormones (structural analysis) Extraction, derivatization, dilution

Transport Sample preparation → ICP-MS

Transport Sample preparation → GC-MS

Transport Sample preparation → ICP-MS

Transport Sample preparation → GC-MS

Analysis Measurement using ICP-MS

Analysis Measurement using GC-MS

Analysis Measurement usisng ICP-MS

Analysis Measurement using GC-MS

Elemental analysis

Kind of sample preparation ?

Structural analysis

Elemental analysis

Kind of sample preparation ?

Structural analysis

Determination of heavy metals (elemental analysis) Microwave digestion, dilution

Determination of methacrylates (structural analysis) Extraction, dilution

Determination of mercury (elemental analysis) Microwave digestion, dilution

Determination of organic wood protection agents (structural analysis) Extraction, dilution

Transport Sample preparation → ICP-MS

Transport Sample preparation → GC-MS Sample preparation → LC-MS

Transport Sample preparation → ICP-MS

Transport Sample preparation → GC-MS

Analysis Measurement using ICP-MS

Analysis Measurement using GC-MS, LC-MS

Analysis Measurement using ICP-MS

Analysis Measurement using GC-MS

Determination of chiral compounds (structural analysis) Derivatization, dilution

Transport Sample preparation → TOF-MS

Analysis Measurement using TOF-MS

Final transport Storage/disposal/cleaning of samples, stock solutions, reagents, solvents, and labware

End

Figure 5.26 Process flow chart for the flexible execution of different applications in a decentralized open automation system.

due to a mobile platform can approach different stations and can perform the sample manipulation as well as additional transportation tasks with great distances. Depending on the system design for a defined process or for the flexible execution of different processes, closed and open systems can be structured (see Figure 5.27). An example for an integrated closed system is the sample preparation of cell cultures using ultrasonication and solid phase

195

196

5 Automation Systems with Flexible Robots

Decentralized/closed

Decentralized/open

SP 1

SP 2

SP X

SP 1

SP 2

SP X

CSI/FR

CSI/FR

CSI/FR

CSI/FR

CSI/FR

CSI/FR

IR

A1

A2

IR

AX

A1

A2

AX

Figure 5.27 Schematic visualization of a decentralized/closed (a) and a decentralized/open automation system (b) using an integrated robot (SP 1,2,…,X: sample preparation, A 1,2,…,X: analytical measurement, CSI: central system integrator, FR: flexible robot, IR: integrated robot).

extraction (SPE) for the determination of cyclophosphamide using HPLC-UV and HPLC-MS. With the extension to any number of applications, the closed system can be changed into an integrated open automation system. 5.4.1 System Design

The automation system is based on the use of an integrated robot, which can realize the functionalities of the functional structures “liquid handling,” “sample treatment,” and “mobile robotics.” Due to the decentralized structure, these functional units can exist as standalone systems and can be used by other applications and users. Additional functional units of this automation concept are “sample storage” und “analytics.” Integrated robot: The integrated robot combines the functionalities of a central system integrator, a flexible robot, and a mobile robot. A desired dual arm robot can be used, which is installed on a mobile platform. This mobile platform has the highest possible mobility due to the use of omnidirectional drives, since the platform can move in any direction. The humanoid design enables a broad range of applications for this robot. To ensure maximum flexibility, each arm should have 7 degrees of freedom similar to the SDA10. Thus, the robot has 15 degrees of freedom, including the basic rotation axes (see Table 5.4). The maximum payload for each arm is considered to be 5 kg. Each arm can be used for different tasks. In addition, synchronized movements of both arms are possible as well (synchronized move). The required repeatability is ±0.1 mm. Each arm can be equipped with a gripper of type LEHF 20K2-48-R86P5 (SMC Pneumatik, Egelsbach). Sample storage: This functional unit is part of each substation of the decentralized system. Substations with shelf spaces in microplate format, which can

5.4 Automation Systems with Integrated Robotics

Table 5.4 Work range and velocities of the main and hand axis of the SDA5 [46]. Axis

Maximum work range (∘ )

Maximum velocity (∘ /s)

Main axis Rotation axis (basic axis)

±170

180

S axis (lifting)

−90/+270

200

L axis (lower arm)a)

±110

200

E axis (lower arm, twist)

±170

200

U axis (upper arm)a)

−90/+115

200

±180

200

Hand axis R axis (upper arm, twist)a) B axis (wrist pitch/yaw)

±110

230

T axis (wrist twist)

±180

350

a) according to the position the work range of the L-, U-, and R-axis can be limited.

be easily accessed by the mobile and integrated robots, are integrated at each substation for sample handover. In addition, racks for labware, chemicals, and samples are available at each substation. Sample treatment, liquid handling, and mobile robotics: These functional units can be separate substations or can be realized by the integrated robot. This flexibility is of great importance for the extension of the closed automation system to an open automation system. In this application, this includes the cell cultivation on a substation with a cell culturing system (Biomek Cell Workstation, celisca, Rostock), which is also equipped with a liquid handler Biomek NX (Beckman Coulter, Krefeld) and an incubator Cytomat 6001(K) (Thermo Electron Corporation, Langenselbold) positioned on top of each other and connected with a lift (vertical transport rail) for the microplates [47]. Additional subsystems are used for cell disruption, including an ultrasonic bath, which can be operated by the integrated robot analog to Section 5.2. An automated solid phase station is installed on a liquid handler Biomek NX (Beckman Coulter, Krefeld) for further sample manipulation [48]. Analytics: Two measurement systems are integrated in the introduced closed automation system, which allow the measurement of samples in different sample concentration ranges. For very low compound concentrations in the ppb region (μg/l), the determination of the cyclophosphamide is realized using an HPLC-MS method. Due to the high selectivity of the mass spectrometer, this also enables the determination of additional compounds, such as decomposition products. Higher concentrations in the ppm region (mg/l to g/l) are measured using a liquid chromatography with UV detection.

197

198

5 Automation Systems with Flexible Robots

5.4.2 Process Description

The first step in the process flow is the cultivation of the HeLa cells used for the further investigations [49]. The cells are defrosted in a water bath, transferred into cell culturing flasks followed by the addition of Dulbecco’s Modified Eagle Medium (DMEM) and fetal bovine serum. Subsequently, the cells are incubated at 37 ∘ C in a 5% CO2 atmosphere. The cell medium must be changed every 24 h. At a confluence rate of 90% (covering of the surface of cell culture flasks with adherent cells), cell splitting is realized. After an additional incubation, the adherent cells are detached with trypsin and centrifuged. The supernatant of the medium is removed and the cell pellet is dissolved in alginate solution. The alginate cell suspension is transferred to 96 well microplates and cultivated in DMEM. The entire cell cultivation is realized on a cell culturing station [49]. In a next process step, cyclophosphamide is added to the 3D cell cultures using a liquid handler Biomek NX (Beckman Coulter, Krefeld) installed on the cell culturing system. The solutions are incubated at 37 ∘ C in a 5% CO2 atmosphere with varying times according to the experimental plan. After completion of the incubation, the cell medium is removed and purified using SPE. For the determination of the compound concentration in the cells, four alginate beads are combined in an Eppendorf vessel (volume 1.5 ml). The pellets are washed and centrifuged. The supernatant is removed and the cell pellets are dissolved in fresh medium. The digestion of the cells is realized with a 10-min treatment in an ultrasonic bath. After digestion, the samples are centrifuged and the supernatant is transferred to a microplate. The cell media samples as well as the samples from the cell disruption are purified by an SPE method for the following analytical determination. Commercial SPE columns (C18) are used, which can be fixed to 24 well adapters. Alternatively, also the use of SPE columns in microplate format is possible. Analytical determination of cyclophosphamide is realized depending on the expected concentration range using HPLC-UV and HPLC-MS. Thus, sample transport to the measurement system and handover of the samples to the auto sampler are required. Depending on the type of auto sampler, the samples can either be directly measured in a microplate or reformatting into glass vials may be necessary. Figure 5.28 shows the process flow chart. 5.4.3 Process Control

The automated process is controlled using an HWMS as discussed in Section 4.4. The type of the mobile robot is of minor influence. The methods for localization and navigation used for the H20 robots can also be used for the integrated robot. Thus, the mobile platform of the integrated robot is equipped with a StarGazer module HSG-A-02 (Hagisonic, Daejeon). Navigation is realized with landmarks installed at the ceilings [50]. In contrast to the systems described so far, the difficulty level increases, since the integrated robot not only performs transportation tasks, but also executes sample manipulation tasks. Besides the transportation tasks, manipulation tasks have to be managed as well in the HWMS. Thus, the automation concept of integrated robotics is a combination of the concepts with central system integrators (see Chapter 4) and with flexible robots.

5.4 Automation Systems with Integrated Robotics

Figure 5.28 Process flow chart for the determination of cyclophosphamide in cells and cell culture medium.

Start

Cell cultivation HeLa cells in 2D or 3D cultures

Pipetting active agent Cyclophosphamide

Cell medium

Cell

Kind of sample ?

Sampling Cell medium

Cell disruption Ultrasonic treatment

Sample clean-up Solid phase extraction

Reformatting 24 well microplates → GC vials

HPLC-MS

Measurement device ?

Dilution Pipetting of methanol

HPLC-UV

Vaporizing Solvent (methanol)

Pipetting Add solvent (acetonitrile)

Analysis Measurement using HPLC-MS

Analysis Measurement using HPLC-UV

End

199

200

5 Automation Systems with Flexible Robots

References 1 Fleischer, H., Drews, R.R., Janson, J., Chinna Patlolla, B.R., Chu, X., Klos, M.

2 3 4

5

6

7

8 9 10 11 12 13 14

15

16

17

et al. (2016) Application of a dual-arm robot in complex sample preparation and measurement processes. J. Lab. Autom., 21 (5), 671–681. Yaskawa (2013) SDA10F Betriebs- und Wartungsanleitung (E1101000164DE01): Revision 1. Yaskawa (2015) MOTOMAN – SDA10D/SDA10F Instructions: Revision 3. IFM Electronic (2010) Originalbetriebsanleitung optoelektronische Sicherheitssensoren (Sicherheits-Lichtvorhang/Sicherheits-Lichtgitter: 704698/0104/2010). Chu X, Fleischer H, Roddelkopf T, Stoll N, Klos M, Thurow K (2015) A LC-MS integration approach in life science automation: Hardware integration and software integration. IEEE International Conference on Automation Science and Engineering, CASE 2015, pp. 979–984. Pan, Z., Polden, J., Larkin, N., van Duin, S., and Norrish, J. (2012) Recent progress on programming methods for industrial robots. Rob. Comput. Integr. Manuf., 28 (2), 87–94. Drews R, Ralf (2015) Einsatz eines Zweiarmroboters in komplexen Prozessen der Probenvorbereitung und chemischen Analytik. Master thesis. Universität Rostock. Chu, X. (2017) Automation strategies of sample preparation in life science automation. PhD thesis. Universität Rostock. Fleischer H (2011) Ein hochdurchsatzfähiges Analysenmesssystem für die Bestimmung chiraler Komponenten. PhD thesis. Universität Rostock. Fleischer H. (2016) Automatisierungsstrategien in der analytischen Messtechnik. Habilitation thesis. Universität Rostock. Gores, G.J. (2008) A strategic vision to promote advances in biliary tract disorders. Curr. Opin. Gastroenterol., 24 (3), 346–348. Gallagher, T.K. and Parks, R.W. (2014) Gallstones. Surgery (United Kingdom), 32 (12), 635–642. Shaffer, E.A. (2006) Epidemiology of gallbladder stone disease. Best Pract. Res. Clin. Gastroenterol., 20 (6), 981–996. Caddy, G.R. and Tham, T.C.K. (2006) Symptoms, diagnosis and endoscopic management of common bile duct stones. Best Pract. Res. Clin. Gastroenterol., 20 (6), 1085–1101. Dowidar, N., Kolmos, H.J., Lyon, H., and Matzen, P. (1991) Clogging of biliary endoprostheses: a morphologic and bacteriologic study. Scand. J. Gastroenterol., 26 (11), 1137–1144. Dowidar, N., Kolmos, H.J., and Matzen, P. (1992) Experimental clogging of biliary endoprostheses. Role of bacteria, endoprosthesis material, and design. Scand. J. Gastroenterol., 27 (1), 77–80. Soehendra, N. and Reynders-Frederix, V. (1980) Palliative bile duct drainage. A new endoscopic method of introducing a transpapillary drain. Endoscopy, 12 (1), 8–11.

References

18 Cotton, P.B. (1990) Management of malignant bile duct obstruction. J. Gas-

troenterol. Hepatol., 5 (Suppl. 1), 63–77. 19 Huibregtse, K., Katon, R.M., and Tytgat, G.N.J. (1986) Endoscopic treatment

of postoperative biliary strictures. Endoscopy, 18 (4), 133–137. 20 Ponchon, T., Gallez, J.-F., Valette, P.-J., Chavaillon, A., and Bory, R. (1989)

21

22 23

24

25

26

27

28

29

30

31

32

33

Endoscopic treatment of biliary tract fistulas. Gastrointest. Endosc., 35 (6), 490–498. Cotton, P.B., Forbes, A., Leung, J.W.C., and Dineen, L. (1987) Endoscopic stenting for long-term treatment of large bile duct stones: 2- to 5-year follow-up. Gastrointest. Endosc., 33 (6), 411–412. Cowling, M.G. and Adam, A.N. (2001) Internal stenting in malignant biliary obstruction. World J. Surg., 25 (3), 355–359. Smith, A.C., Dowsett, J.F., Russell, R.C.G., Hatfield, A.R.W., and Cotton, P.B. (1994) Randomised trial of endoscopic steriting versus surgical bypass in malignant low bileduct obstruction. Lancet, 344 (8938), 1655–1660. Dowidar, N., Moesgaard, F., and Matzen, P. (1991) Clogging and other complications of endoscopic biliary endoprostheses. Scand. J. Gastroenterol., 26 (11), 1132–1136. Costa, L., Bracco, P., Vada, S., Trossarelli, L., and Jacobson, K. (2001) A chemical analysis of the clogging process of polymeric biliary endoprostheses. Biomaterials, 22 (23), 3113–3119. Guaglianone, E., Cardines, R., Mastrantonio, P., Rosa, R.D., Penni, A., Puggioni, G. et al (2008) Role of multispecies microbial biofilms in the occlusion of biliary stents. Microb. Ecol. Health Dis., 20 (4), 207–209. Farnbacher, M.J., Kraupa, W., and Schneider, H.T. (2013) Cleaning of occluded biliary endoprostheses: is shockwave application an alternative to regular stent exchange? J. Med. Eng. Technol., 37 (1), 10–16. Groen, A.K., Out, T., Huibregtse, K., Delzenne, B., Hoek, F.J., and Tytgat, G.N. (1987) Characterization of the content of occluded biliary endoprostheses. Endoscopy, 19 (2), 57–59. Moesch, C., Sautereau, D., Cessot, F., Berry, P., Mounier, M., Gainant, A. et al. (1991) Physicochemical and bacteriological analysis of the contents of occluded biliary endoprostheses. Hepatology, 14 (6), 1142–1146. Cetta, F., Rappuoli, R., Montalto, G., Baldi, C., Gori, M., Cetta, D. et al. (1999) New biliary endoprosthesis less liable to block in biliary infections: description and in vitro studies. Eur. J. Surg., 165 (8), 782–785. van Berkel, A.M., van Marle, J., Groen, A.K., and Bruno, M.J. (2005) Mechanisms of biliary stent clogging: confocal laser scanning and scanning electron microscopy. Endoscopy, 37 (8), 729–734. Paik, M.-J., Yu, J., Hu, M.-B., Kim, S.-J., Kim, K.-R., Ahn, Y.-H. et al. (2008) Gas chromatographic-mass spectrometric analyses of cholesterol and its precursors in rat plasma as tert-butyldimethylsilyl derivatives. Clin. Chim. Acta, 396 (1–2), 62–65. Ubhayasekera, S.J.K.A. and Dutta, P.C. (2012) Assessment of sterol oxidation in oils recovered from exhausted bleaching earth by coupled capillary column GC and GC-MS methods. J. Am. Oil Chem. Soc., 89 (8), 1427–1433.

201

202

5 Automation Systems with Flexible Robots

34 Son, H.-H., Moon, J.-Y., Seo, H.S., Kim, H.H., Chung, B.C., and Choi, M.H.

35

36

37

38

39

40

41

42 43

44

45

46 47

(2014) High-temperature GC-MS-based serum cholesterol signatures may reveal sex differences in vasospastic angina. J. Lipid Res., 55 (1), 155–162. Eckfeldt, J.H., Lewis, L.A., Belcher, J.D., Singh, J., and Frantz, I.D. Jr. (1991) Determination of serum cholesterol by isotope dilution mass spectrometry with a benchtop capillary gas chromatograph/mass spectrometer: comparison with the national reference system’s definitive and reference methods. Clin. Chem., 37 (7), 1161–1165. Edwards, S.H., Kimberly, M.M., Pyatt, S.D., Stribling, S.L., Dobbin, K.D., and Myers, G.L. (2011) Proposed serum cholesterol reference measurement procedure by gas chromatography-isotope dilution mass spectrometry. Clin. Chem., 57 (4), 614–622. Thienpont, L.M., van Landuyt, K.G., Stöckl, D., and Leenheer, A.P. (1996) Four frequently used test systems for serum cholesterol evaluated by isotope dilution gas chromatography-mass spectrometry candidate reference method. Clin. Chem., 42 (4), 531–535. Langlois J, Kuipers J. (2010) Fast GC Analysis of Natural Waxes from Art and Museum Objects: Application Note SI-02454. Agilent Technologies (former Varian) Dinh, T.T.N., Blanton, J. JR., Brooks, J.C., Miller, M.F., and Thompson, L.D. (2008) A simplified method for cholesterol determination in meat and meat products. J. Food Comp. Anal., 21 (4), 306–314. Fu R, Joseph M (2012) LC/ELSD and LC/MS/MS of Cholesterol and Related Sterols on a Poroshell 120 Column: Application Note 5991-0452EN. Agilent Technologies Janson J. (2015) Messtechnische Untersuchung von Stent-Inkrustationen mittels struktur- und elementanalytischer Verfahren. Master thesis. Universität Rostock. Chinna Patlolla BR 2015 Software communication interface for complex chemical devices in automated systems. Master thesis. Universität Rostock. Zierdt H. (2005) Steroidhormone in bodengelagertem Skelettmaterial – Ein Ansatz zur Abschätzung von Fertilitätsparametern in historischen Bevölkerungen. PhD thesis. Georg-August-Universität zu Göttingen. Abdulla, A.A., Liu, H., Stoll, N., and Thurow, K. (2016) A new robust method for mobile robot multifloor navigation in distributed life science laboratories. J. Control Sci. Eng., 2016, 1–7. Article number 3589395. Abdulla AA, Liu H, Stoll N, Thurow K (2016) A robust method for elevator operation in semi-outdoor environment for mobile robot transportation system in life science laboratories. IEEE International Conference on Intelligent Engineering Systems, INES 2016, pp. 45–50. Yaskawa Europe Product Information Robotics, http://www.motoman.com/ (accessed 01 December 2016). Lehmann, R., Gallert, C., Roddelkopf, T., Junginger, S., Wree, A., and Thurow, K. (2016) 3 dimensional cell cultures: a comparison between manually and automatically produced alginate beads. Cytotechnology, 68 (4), 1049–1062.

References

48 Gallert, C., Vorberg, E., Roddelkopf, T., Junginger, S., Fleischer, H., and

Thurow, K. (2015) Evaluation of an automated solid-phase extraction method using positive pressure. Am. Pharm. Rev., 18 (3). 49 Sterna P. (2015) Untersuchungen zu Konzentrationsprofilen von Wirkstoffkomponenten in 3D Zellkulturen. Master thesis. Universität Rostock. 50 Ghandour M, Liu H, Stoll N, Thurow K (2015) Improving the navigation of indoor mobile robots using Kalman Filter. IEEE International Instrumentation and Measurement Technology Conference, I2MTC, 2015, pp. 1434–1439.

203

205

6 Automated Data Evaluation in Life Sciences High-throughput applications in compound oriented measurement techniques generate a high amount of data points in a short timeframe [1–8]. In modern laboratories, automated sample preparation systems are used and an increasing number of samples can be analyzed [9, 10]. In large modern laboratories, more than 1000 samples can be analyzed per day [11]. High-throughput screening procedures are used in clinical laboratories or in the field of toxicology [1, 2, 12], whereby the handling of more than 8000 samples per day is possible [13, 14].

6.1 Specific Tasks in Data Evaluation in Analytical Measurements In the evaluation of measurement data in the field of chemical analysis, typical tasks include routine measurements, the development of new analytical methods, and method validation. With increasing automation levels, an efficient automated data evaluation will become more and more important in daily laboratory work. Typically, the instruments software of the measurement devices provides modules for instrument control and for qualitative and quantitative data evaluation. Table 6.1 lists common examples of software modules currently used in analytical laboratories. In contrast to classical physical measurement values, compound oriented measurement data requires the conversion of primary measurement data into application-specific secondary data. There are significant differences in the software functionalities and data types due to the different manufacturers, instrument types, and style of the measurement data. A further task is the “tiding-up” of the acquired data. The desired measurements must be separated from non-relevant data. In addition to the samples, a typical measurement series includes measurements of solutions for flushing the measurement system to prevent cross-contamination and carry-over. Furthermore, measurements of blank samples to ensure the use of contamination-free solvents and measurements of quality-check (QC) samples to ensure a constant recovery rate are included in a measurement series. The calibration data is often hidden or removed, since it is not required in end customer reports. The individual definition of relevant and non-relevant data depends on the application and the measurement task, for example, customer sample determination or Automation Solutions for Analytical Measurements: Concepts and Applications, First Edition. Heidi Fleischer and Kerstin Thurow. © 2018 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2018 by Wiley-VCH Verlag GmbH & Co. KGaA.

206

6 Automated Data Evaluation in Life Sciences

Table 6.1 Example of a heterogeneous software environment (measurement instruments and related software) in the analytical laboratory of CELISCA (Center for Life Science Automation, Rostock, Germany) [7]. Measurement instrument

Manufacturer

Software for data acquisition

Software for qualitative analysis

Software for quantitative analysis

MassHunter Acquisition for TOF/QTOF

MassHunter Quantification for TOF

LC-QTOF

MassHunter Acquisition for TOF/QTOF

MassHunter Quantification for QTOF

LC-QQQ

MassHunter Acquisition for QQQ

GC-QQQ

GC-MS

Agilent Technologies

LC-TOF

MassHunter Qualification

MassHunter Quantification for QQQ

MassHunter Acquisition for GC-QQQ

MassHunter Quantification for GC-QQQ

MassHunter Acquisition for GC-MS

MassHunter Quantification for GC-MS

Data format

Excel pdf xml csv

GC-MS

GC-MSD ChemStation

csv

LC-MS

LC-MSD ChemStation online

LC-MSD ChemStation offline

aia csv dif

ICP-MS

MassHunter for ICP-MS

MassHunter Data Analysis for ICP-MS (on/offline)

Excel csv

ICP-OES F-AAS GF-AAS

Thermoscientific

Qtegra for iCAP

Excel csv

Agilent Technologies (former Varian)

SpectrAA Worksheet Oriented AA Software

prn csv

validation in method development. Usually, this is carried out manually using spreadsheet software such as “Microsoft Excel” (Microsoft, Redmond). Further tasks involve a mathematical measurement data correction and the calculation of validation parameters. Statistical calculations in measurement development, such as the determination of precision [15, 16] and the limits of detection and quantification [17, 18], are required. This is often manually determined by a spreadsheet software such as “Microsoft Excel” (Microsoft, Redmond) or special

6.2 Automation Goals

software solutions such as “Quality Assurance using Statistical Methods” (QSM, Hoffmann, Gießen) are used [19]. Additionally, general software (modular environment) for data analytics such as the “Konstanz Information Miner” (KNIME, University of Konstanz, Konstanz) is used in modern laboratories to mine the large amount of generated data [20–22]. A multitude of analytical instruments and software solutions from various manufacturers in an analytical laboratory result in a heterogeneity of the software used. Furthermore, the data format (see Table 6.1) and the data style may significantly vary between different software products. Another type of software used in laboratories is the electronic laboratory notebooks (ELNs) [23]. An ELN collects measurement data and documents analytical experiments to enable the rapid and effective search of laboratory and measurement data. Software like “Pipeline Pilot” or “Isentris” (Biovia/Accelrys, San Diego (CA), USA) analyze molecules to find similarities and substructures or filter them by features [24, 25]. Modern laboratories increasingly work electronically and paperless. This enables and requires automated data evaluation [23, 26–29]. Long familiarization time, high manual effort, and working time to perform consistent data validation and evaluation can be avoided. This applies to software updates or to the renewal or extension of existing systems. An efficient and uniform data evaluation in analytical laboratories is a challenging target, especially in highly automated systems with a decentralized open structure. This change in digitalization of data in analytical measurement, the way to handle the large amount of data, and the use of computational intelligence will change the work in the laboratory in the future [26, 29–31].

6.2 Automation Goals One of the main tasks in laboratory automation is the integration of multiple devices into one system to increase the flexibility and the sample throughput. High-throughput screening processes have been established not only in the field of biotechnology and bioscreening, but also in the field of analytical chemistry such as catalyst research [32], in clinical laboratories [33], or in toxicology [1, 2]. In data evaluation for special tasks, a multitude of individual software solutions have been developed. Special data analysis and evaluation programs are reported, for example, in toxicological measurements [1, 2] or material sciences [34]. In high-throughput screening applications, mass spectrometry is becoming increasingly important [4], since it enables rapid analysis times [5]. A software solution was developed for a homogenous data evaluation using various mass spectrometric or spectroscopic measurement systems [3, 5–8, 35–37]. Another important goal of automated data evaluation is the definition and usage of uniform and standardized data formats. In industries, especially in plant design, the fact that every software vendor uses a proprietary data format is one of the key challenges. Data generated by one engineering suite usually cannot be exchanged with another one [38]. The same problem can be found in the field of research and development in a multitude of sciences. In chemical

207

208

6 Automated Data Evaluation in Life Sciences

analysis, different manufacturers of analytical systems provide different software products and software solutions differ with the type of the analytical device. Otherwise, the software functionalities can also differ for one type of analytical device depending on the manufacturer. In general, a chemical analysis includes the control of the measurement devices and the measurements, followed by the qualitative and quantitative analysis of the results using the related instruments software. In many cases, the processed data can be exported in well-known file formats such as Excel (xls, xlsx), comma-separated values (csv), or portable document format (pdf). This enables further manual processing including statistical analysis. The exported data is formatted in a manufacturer-dependent style (e.g., the table layout). Thus, different software solutions deliver inhomogenous data and table styles. The principle of modular software development shows some advantages, especially in complex automation systems [39]. Each software component or module integrates data and functions to fulfill a self-contained task. The communication between the components is performed by defined interfaces, which facilitates the exchange of components [40]. Updates, extensions, maintenance, and tests can be carried out more easily and faster in modular software systems.

6.3 System Design The system design is based on the procedural methods in industrial automation. Especially in system engineering, the challenge exists to integrate subsystems into automation systems. These subsystems are controlled by device software with manufacturer-dependent, proprietary data formats. Modularly structured software solutions are suitable to overcome this problem [39]. Each software component and module involves data and functionalities, respectively, to fulfill a self-contained task. The communication between the modules is performed using interfaces, which enables an easy exchange of modules [40]. The design of automation systems in laboratory environments involves the integration of various devices such as laboratory equipment (e.g., thermomixer, centrifuge, ultrasonic bath), laboratory robots (transport, sample manipulation), and analytical instruments. Usually, the measurement data in analytical laboratories is acquired with analytical instruments from various vendors, which provide different software products depending on the type of the measurement device [7]. The format and the style of the measurement data are manufacturer-dependent (e.g., table layout, style of data records), which results in inhomogeneous data and table styles. Furthermore, the instruments software is installed on various versions of the operating systems depending on the measurement device. The data evaluation is carried out either on the laboratory workstation or at a separate work place. Mobile devices such as laptops, netbooks, tablet PCs, cell, and smartphones are increasingly used. A study found that 70% of US high school students have a laptop or netbook and about 30% use smartphones [41]. It can be observed that these numbers continue to grow in all fields of daily life. Employees in chemical laboratories also use mobile devices. In the field of automation engineering [42] and preventive medicine in automated laboratories [43], mobile devices will

6.3 System Design

also gain importance increasingly. In automated data evaluation for chemical measurements, the use of mobile devices is a promising approach. To provide a reliable software independently from the operating system, the type of computer and the location of the user web applications running on a webserver (client/server principle) are becoming more and more important [44]. The “Analytical Data Evaluation” software (ADE, University of Rostock, Rostock, Germany) was implemented with a modular structure. This web application is based on “ASP.NET 4.5” and runs on a webserver based on “Microsoft Windows Server 2008” (Microsoft, Redmond). The ADE software runs independently from the platform and operating system using standard web browsers [8]. Consequently, the operator can handle the measurement data on each workstation in the laboratory and additionally with mobile devices. The ADE software is connected to the “PubChem” database (NCBI – National Center for Biotechnology Information, NLM – U.S. National Library of Medicine, NIH – National Institutes of Health, HHS – U.S. Department of Health & Human Services) using the provided power user gateway (PUG) [45, 46]. This provides detailed information about desired analytes including molecular formula, molecular weight, and exact mass, simplified molecular input line entry specification (SMILES) [47], international chemical identifier (InChi) [48, 49], synonyms, and chemical structure. The ADE software was implemented for the evaluation of measurement data of various analytical devices. These techniques enable the elemental and structural analysis of various sample and matrix types and involve inductively coupled plasma mass spectrometry (ICP-MS), gas chromatography mass spectrometry (GC-MS), and liquid chromatography mass spectrometry (LC-MS). Additionally, the software is used in a special analytical technique for rapid enantiomeric excess determination with a different data evaluation workflow compared to classical analytical processes (Chiral MS) [50]. Due to the modular software design, an easy extension to further analytical devices, functionalities, and modules is possible. Figure 6.1 visualizes the system concept with several communication paths [37]. In general, the measurement data is acquired in the instruments software and stored as raw data often with a proprietary data format. This raw data (detector signals) can be opened and read out with the instruments software. From the raw data, calibration curves are created and the concentrations of the desired analytes in the samples solutions are calculated. In the case of solid samples, a conversion of the concentration value related to the solid must be performed. Some instrument software products already provide this functionality such as “SpectrAA” for AAS or “MassHunter” for ICP-MS (both from Agilent Technologies, Waldbronn, Germany) or “Qtegra” for ICP-OES (Thermo Fisher Scientific, Dreieich, Germany). However, there are manufacturer-dependent differences when entering the sample parameters (e.g., sample weight, sample volumes, dilution factor, units). Due to the inhomogeneity in this procedure, human errors often cannot be avoided. The automation of the sample table generation for a subsequent analytical measurement requires a separate software (middleware) for each device. The situation is similar for the measurement results to be exported. For current measurement systems, the data export is performed as Excel, xml, csv, or pdf files (see Table 6.1). However, the structures of the data tables are different.

209

210

6 Automated Data Evaluation in Life Sciences

Measurement instrument 1 (ICP-MS)

Measurement instrument 2 (LC-MS)

Web application Analytical data evaluation (ADE)

Raw data

Upload processed data

Raw data

Data evaluation

Upload processed data

Download results

Web server

Measurement instrument 3 (GC-MS) Measurement instruments

Raw data

Upload processed data

Laboratory workstations

Download results

Download results

Download results

Stationary workstations and mobile devices

Figure 6.1 System concept of automated data evaluation in analytical chemistry (ICP-MS: inductively coupled plasma mass spectrometry, LC-MS: liquid chromatography mass spectrometry, GC-MS: gas chromatography mass spectrometry).

The workflow of the data evaluation depends on the analytical task. In the elemental analysis using ICP-MS, the measurement results (element concentrations) are exported by the “MassHunter” software using an export file (xml) after finishing the measurement series (batch). In ICP-OES analysis, this is done using the csv format. One file contains the results from all samples of the batch. In the structure analysis using LC-MS and GC-MS, the raw data can be quantitatively (compound concentrations) evaluated using the instruments software “MassHunter Quantification.” The results can be exported as an xml file for a measurement series (worklist, sequence, or batch). In the special case of chromatography free chiral mass spectrometry (Chiral MS), the required masses are extracted from the total ion current and the peak areas are integrated using the instruments software “MassHunter Qualification.” The results are saved in an Excel file for each sample. The export files (xml, csv, or Excel) with the processed data are transferred by an additional software module “Data Upload” on the measurement instruments workstation to a web server. The data in these files differs depending on the measurement system. Using the software ADE, these files are opened and the data imported into a standard table structure (see Figure 6.2). The sample data (e.g., sample weight, dilution factor, etc.) is now entered in the related instrument software to unify the data analysis for all instruments. This is done if the measurement (measurement series or single measurement) is completed using the ADE software. The processed data is copied into a standardized data table, cleaned-up, analyzed, and visualized. These steps can be performed manually or automatically. The data analysis of measurement data acquired with

6.4 System Realization

ICP-MS

ICP-OES

GC-MS

LC-MS

Chiral MS

Data files with measurement values

Data import Generation of data table in standard format with measurement values Data table

Figure 6.2 Principle of data import from various measurement instruments (ICP-MS: inductively coupled plasma mass spectrometry, ICP-OES: inductively coupled plasma optical emission spectroscopy, GC-MS: gas chromatography mass spectrometry, LC-MS: liquid chromatography mass spectrometry, chiral MS: chiral mass spectrometry). (Redrawn from [8].)

Data acquisition Acquire raw data

Data extraktion Measurement values from raw data

Export Save values in transfer file (XML, CSV, Excel)

Upload Load transfer files to web server

Import Copy measurement values into data table

Filtering Data clean-up

Calculation Measurement results

Export Save results (Excel, PDF)

Figure 6.3 General process workflow of data evaluation using the “Analytical Data Evaluation” software (ADE).

different measurement instruments is performed with one software solution (ADE), which simplifies the workflow. Additionally, data evaluation is possible location-independently using end-user devices with different operating systems. Figure 6.3 summarizes the general process workflow in data evaluation [37].

6.4 System Realization 6.4.1 Software Structure

The automated procedure of analytical measurements coupled to subsequent data evaluation requires an intelligent combination of software products and software modules. The system consists of the instruments control software for data acquisition and data pre-processing, the “Data Upload” software on the workstation (client), and the web application ADE running on the webserver. ADE consists of several modules with self-contained tasks: “web application,” “web service,” “entities database,” “PubChem translator,” and “functionality.”

211

212

6 Automated Data Evaluation in Life Sciences

This modular software approach offers advantages compared to classical software solutions. The ADE software can be easily updated or expanded with additional functions. This, for example, enables the integration of new measurement instruments with new table structures in the export data files. The implemented graphical user interface (GUI) realizes the user-to-machine interaction and allows working in a multi-user environment. A further advantage is the separation of the GUI from the data evaluation. This allows to adapt and replace the GUI for software use with classic PCs and additionally on mobile devices with small-sized displays (e.g., smartphones, tablet PCs). The ADE software runs on a web server based on the “Microsoft Windows Server 2008” and ASP.NET (Active Server Pages.NET) as a web page and can be used with standard web browsers. Figure 6.4 gives an overview about the system structure and the software modules. The module “web application” involves user and role management and provides the user access to the data saved in the “entities database.” This was realized using the ASP.NET model view controller (MVC) framework. The application was divided into three independent modules for better control of the environment: “data” (model), “web page” (view), and “functionalities” (control). Optimum design and performance on mobile devices such as smartphones and tablet

Server Web application Client

Entities

Functionalities

PubChem Translator

Web service

Data upload

PMS

LES

PCS

BPMS

Measurement control

IMS

LIMS

ELN

SWMS

Figure 6.4 Modular structure of the automated data analysis system in analytical chemistry (PMS: process management system, LES: laboratory execution system, PCS: process control system, BPMS: business process management system, IMS: information management system, LIMS: laboratory information management system, ELN: electronic laboratory notebook, SWMS: scientific workflow management system). (Redrawn from [35].)

6.4 System Realization

PCs were realized using the “bootstrap framework” (twitter, San Francisco (CA), USA). Different user rights are defined: normal user and manager user. A normal user has full control about self-created projects. In contrast, the manager has full control of all processes. There is one exception: projects indicated as confidential can only be read and controlled by authorized users. The machine-to-machine interaction without the GUI is realized by the module “web service.” This module is used by the external software module “Data Upload” on the client and allows the access to the “entities database” and the “functionalities” module. The service-oriented architecture (SOA) using the PUG for a PubChem database access also supports the integration of the web application into existing process and information management systems (PSM, IMS) (see Figure 6.4). All data used and manipulated by the ADE software is managed using the “entities database.” This data involves user accounts, project information, measurement parameters and values, measurement instruments, and related control software. The information of one sample in a measurement series is saved in an xml string within the entity “sample.” Corresponding parsers are included in the “functionality” module to extract the desired data related to samples and analytes from these strings. The algorithms performing these tasks need information about the location/position of the required data. This is realized by manual integration of a specific tag inside the xml string. A representative export file (xml, csv) from the device software is connected to the corresponding entity “software” inside the “entities database” to minimize the manual workload. This requires only one initial adjustment for each measurement instruments software. This concept provides the advantage, that integration of a new analytical device, the related control software and the corresponding export files (xml, csv) is possible without programming knowledge. The “functionality” module contains algorithms and functionalities related to data evaluation such as the measurement value correction using internal standards and the calculation of final concentrations and statistical values. For correct data evaluation, the measurement values must be labeled with information not provided by the instruments software. Therefore, the sample name was extended with an information tag. This tag contains the type of calculation task, a link identifier, and the sample. If the sample is used as a calibration standard, the concentration and the unit are also included in this tag. This should be explained with an example. A sample (S) in routine measurement (R) of a measurement series (link identifier 1) is labeled in the sample name with the tag . A calibration standard (C) of the same measurement series (also link identifier 1) and the compound/element concentration of 50 ppm are labeled with the tag . If the calibration samples should be not included in the final results, the samples can be indexed as NONE. Several abbreviations were defined to keep the tag length as small as possible (see Table 6.2). One or more tags can be included in one sample name. The link identifier indicates samples that are grouped together. This naming convention facilitates an identification of measurement values for the desired data clean-up and calculation tasks by the calculation algorithms.

213

214

6 Automated Data Evaluation in Life Sciences

Table 6.2 Types of calculation tasks and the related abbreviations in the sample name tag [35]. Abbreviation

Calculation task

NONE

Not defined

R

Routine measurement

REPEAT

Repeatability

RR

Recovery rate

REPROD

Within-laboratory reproducibility

MSTAB

Method stability

MPREC

Measurement precision

LOD

Limit of detection and limit of quantification

MDEV

Method development

The module “Data Upload” runs on the workstation of the measurement instrument (client) and must be adjusted only on the first start on the client. Subsequently, the software module runs as a background process and monitors a data path with its subfolders given by the user. The user must also select the measurement instrument and the software product, which generates the export files containing the measurement values to be exported. A list of integrated devices is shown in a drop-down menu, which is realized by accessing the module “Data Upload” to the entity database using the web service interface. Additionally, this access is required to create a new measurement series or to update an existing series. 6.4.2 Software Operation

The human–machine interaction in the ADE software is realized using a multi-user web application. User registration is required for software operation. Email verification and account confirmation offer additional security. With the account created, the user logs in to start a new session (Figure 6.5). The instruments software of the measurement device, which acquires the measurement data to be evaluated, must be integrated into the ADE software. The parameters of the software can be individually configured. The output files of each measurement software have a special layout. This requires a configuration of the ADE software to extract the desired data for subsequent data evaluation. This includes the assignment of the table headers in case of csv-files or the tags in case of xml files to obtain the desired information. Therefore, an algorithm analyzes and provides the file structure (csv, xml) of the output file in a two-dimensional list format and the user can select the desired information. A distinction is made between elemental and structural analysis due to different types of analytes and parameters. The required information regarding samples, analytes, and results can be selected for data evaluation. Figure 6.6 shows the user interface for the integration of the measurement instrument software and the software details.

6.4 System Realization

(a)

(b)

Figure 6.5 Web interface of the ADE software for (a) user registration and (b) user login.

(a)

(b)

Figure 6.6 User interface for integration and adjustment of measurement instrument software: (a) overview of integrated software versions and (b) software details

215

216

6 Automated Data Evaluation in Life Sciences

The integrated software must be allocated to a measurement device, since several instruments may run with the same software version. This provides easy data management and serves as an overview for the user. The device must be created in the ADE software, which manages all integrated devices and visualizes them in a list. A detailed view shows the software and, if available, the measurements performed on this device. Figure 6.7 shows the user interface for the integration of a measurement instrument (device) and the device details. Measurements are separately presented in the GUI in a list, which shows the relationship to the related measurement device and the instruments software. If a new measurement is created, the instruments software must be selected. This enables the correct extraction of the required data from the export file, which was generated by this instruments software. The data extracted is shown in a detailed table on the GUI. Figure 6.8 gives an overview of the user interface for measurement management and the measurement series details. The data extracted from an export file is divided into several groups: “samples,” “sample preparation,” “sample naming convention,” and “analytes.” In the group “samples,” general information related to the sample is listed (see Figure 6.9). This involves the sample name, sample type, operator, acquisition date and time, data path, and the file name. The group “sample preparation” allows to enter the sample weight, the sample volume, and the dilution factor if necessary (see Figure 6.10). Some instruments

(a)

(b)

Figure 6.7 User interface for integration and allocation of measurement devices to software versions and measurements: (a) overview of integrated devices and related software versions and (b) device details.

6.4 System Realization

(a)

(b)

Figure 6.8 User interface for measurement management: (a) overview about measurements, measurement devices, and related software versions and (b) measurement details.

Figure 6.9 User interface with measurement details: information related to samples.

217

218

6 Automated Data Evaluation in Life Sciences

Figure 6.10 User interface with measurement details: information related to the sample preparation.

software allows to enter these values. In the scope of uniform data evaluation, it is more advantageous to enter these values using only one software. The group “naming convention” was introduced for easy sample identification and execution of the desired calculation steps. As explained in detail in Section 6.4.1, the user includes in the sample name an information tag that contains the type of calculation task, a link identifier, and the sample. These tags may be defined during integration of the instruments software or can be edited in the group “naming convention” in the measurement details (see Figure 6.11). The group “analytes” contains the extracted information related to the analytes (see Figure 6.12). This data can be extended or edited by the user. Elements/compounds can be indicated as internal standard if necessary. For the determination of the recovery rate, the expected values can be entered. Furthermore, the user may select the unit(s) and the analytes to be visualized

Figure 6.11 User interface with measurement details: information related to the naming convention in the sample name.

6.4 System Realization

Figure 6.12 User interface with measurement details: information related to the analytes.

in the final results. The software automatically converts between various units. Thus, the user can freely select the required unit. Often, a multitude of measurements is related to a study or an experiment series. Therefore, interrelated measurement series can be combined and managed in projects. Similar to software, devices and measurements projects are shown in a list. Figure 6.13 gives an overview of the user interface for project management and project details. On the results page, an additional navigation field is available to allow the user to switch between several projects and measurements. The user can select here the results per calculation type such as repeatability, measurement precision, recovery rate, and so on. Figure 6.14 shows the results page with measurement results of method development for the determination of the content of mercury

(a)

(b)

Figure 6.13 User interface for project management: (a) overview of projects and related users and (b) project details.

219

6 Automated Data Evaluation in Life Sciences

Figure 6.14 Results page with projects navigation field (left) and the measurement results with the concentration of selected elements. Measurement precision MERCURY 202 (Hg-He) 0.35

068SMPL.d 0.29

0.3 0.25 0.2 0.15 0.1 0.05

.d

.d 07

5S

M

M

PL

PL

.d 07 4S

M

PL

.d 3S 07

M

PL

.d 2S 07

1S

M

PL

.d 07

0S

M

PL

.d PL 07

06

9S

M

PL M 8S 06

06

7S

M

PL

.d

0 .d

Final concentration (ppb)

220

Figure 6.15 Chart with results of the determination of the measurement precision in the determination of mercury in wood materials.

and heavy metals in wood materials – in detail – the results of experiments for determination the repeatability. The samples are listed in rows and the analytes (here: elements) in columns. The results are visualized in charts. There are various displays possible: compound/element concentrations for each sample as well as averages and standard deviations for each analyte (compound/element). For a more detailed display, the results can be drawn for each analyte in a table and in a chart. Figure 6.15 shows an exemplarily chart containing results of the determination of the measurement precision in the determination of mercury in wood materials.

6.5 Process Description The process of automated data evaluation is divided into three subprocesses performed with different software solutions and modules. The first step requires

6.5 Process Description

the desired modification of the data acquisition method and the data analysis method using the instruments software. This also requires the selection of a suitable export data format (e.g., xml template). The export file (xml) can be manually uploaded to the server using the ADE web application. The file can also be automatically uploaded to the web server using the “Data Upload” software module installed on the measurement instruments workstation. This module must be previously configured by the user. The export folder, where the export file is saved by the instruments software, must be defined. Furthermore, the user must select the measurement instrument and the instruments software. The “Data Upload” module searches for a new export file by running a background process. If such a file is found, a copy is uploaded to the web server. Using the ADE web application, the user creates a new project and includes the desired measurement series. In the following steps, the analytes are selected and the sample weight, sample volume, and dilution factor are entered. It is also possible to include an internal standard for measurement value correction. In the last step, the results are calculated according to the type of measurement found in the tag of the sample name. The results are written in a data table and visualized in charts. Figure 6.16 gives an overview of the process workflow for the three subprocesses. (a)

(b)

(c)

Start

Start

Start

Prepare measurement (modify data acquisition method and data analysis method)

Prepare data upload modify export folder, device, and software

Add measurement series (to evaluation project)

Run measurement Export file detected?

no

Add sample data (sample weight, volume, dilution factor)

yes Generate export file (XML, CSV, Excel) New measurement series?

End

no

Calculation and Visualization (measurement results)

yes Create (new measurement series in data base)

Generate report file (Excel, PDF)

End Upload data (to measurement series in data base)

Extract information (samples, analytes, values)

End

Figure 6.16 Process workflow for the automated data evaluation using (a) the measurement instruments software, (b) the software module “Data Upload” on the instruments workstation, and (c) the web application “Analytical Data Evaluation” (ADE), (gray fields: executed by software, white fields: manual user interaction). (Redrawn from [35].)

221

222

6 Automated Data Evaluation in Life Sciences

In fully automated data evaluation, the sample data can be imported from a high-level workflow management system. This eliminates the manual user interaction. The calculation formulas for the correction of the measured values with an internal standard, to determine the statistical parameters (average, standard deviation, coefficient of variation) and the validation parameters (recovery rate, detection, and quantitation limit) can be hardcoded. To increase the flexibility, an editor can be integrated, which allows the operator to define his own formulas.

6.6 Application Examples 6.6.1 Automated Data Analysis in the Elemental Analysis

For a rapid and sensitive qualitative and quantitative analysis of chemical elements, an ICP-MS 7700x (Agilent Technologies, Waldbronn) was selected. The instruments software “MassHunter for ICP-MS” (Agilent Technologies, Waldbronn) is used both for data acquisition and for qualitative and quantitative data analysis. The measurement data of all samples of a measurement series (a so-called batch) can be exported into an xml format and includes the file name, the date and time of data acquisition, the sample name, the elements to be determined and the concentrations measured in liquid or gaseous samples. Using the software module “Data Upload,” the export file is uploaded to the web server. Subsequently, the measurement data is extracted and copied into a data table with a standardized layout. In addition to the data clean-up, the values may be corrected with an internal standard if required. In case of solid samples, the concentrations are determined after entering the sample weight, sample volume, and dilution factor. Finally, the results are saved in an Excel file and the graphics for data visualization can be found in the related project folder. The software system and its performance were determined with the data evaluation of the elemental analysis of wood materials, especially the determination of the mercury content. The method validation includes the determination of repeatability, recovery rate, within-laboratory precision, method stability, measurement precision, and the limits of detection and quantification (see Section 3.5). The determination of repeatability was selected for an exemplary comparison of the manual data evaluation using the spreadsheet software “Microsoft Excel” (Microsoft, Redmond) and the automatic software-based procedure using ADE. A typical measurement series for the determination of repeatability includes calibration samples, blank samples to prevent cross-contamination and samples containing standard reference material. The concentration of mercury and several heavy metals, which were determined in various instrument modes (no-gas mode without and He mode with activated collision cell), should be determined. In summary, the measurement series involves a high amount of data points. For the determination of repeatability, only the measurement values of the samples containing standard reference material are necessary. The table with measurement values created by the instruments software MassHunter for ICP-MS must

6.6 Application Examples

be filtered to reduce the data points. In the following step, the data is corrected using the internal standard. In general, the instruments software provides the use of an internal standard and the measurement value correction. In the case of this analytical task, a second internal standard is used to correct variations in the sample preparation using microwave-assisted digestion (see Section 4.2), which is not supported by the instruments software. The following step requires the manual entry of additional sample information (sample weight, sample volume, and dilution factor). The software converts the concentration values of the liquid measurement solutions (e.g., μg/l) into concentration values related to the solid sample (e.g., mg/kg). Finally, the statistical values (average, standard deviation, and coefficient of variation) were calculated and the charts for data visualization generated. Figure 6.17 shows the results of a repeatability experiment in the determination of mercury and heavy metals in wood material. In manual data evaluation, the most time-consuming steps are data clean-up, measurement value correction using the second internal standard, concentration calculation for solid samples, and even more, drawing the charts. In automated data evaluation using ADE, the data clean-up process is carried out using an algorithm, which interprets the tags included the sample name. The selection of the desired elements printed in the final table and entering the sample values (e.g., sample weight, volume, dilution factor) are a required interaction of the user. The results calculation and the charts generation are automatically performed without manual assistance. A comparison of the time required in the manual and automatic data evaluation in the determination of repeatability shows a reduction in processing time by more than 92% (see Table 6.3).

Figure 6.17 User interface of the ADE software with the results of the determination of the repeatability in elemental analysis of wood material (determination of mercury and heavy metals).

223

224

6 Automated Data Evaluation in Life Sciences

Table 6.3 Comparison of processing time required for manual and automated data evaluation [35]. Data evaluation task

Time required for processing (min) Manual

Automatic

Filter data

7

—

Select analytes for result pages