Emerging 2D Materials and Devices for the Internet of Things: Information, Sensing and Energy Applications [1 ed.] 0128183861, 9780128183861

Emerging 2D Materials and Devices for the Internet of Things: Information, Sensing and Energy Applications summarizes st

741 52 39MB

English Pages 348 [340] Year 2020

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Emerging 2D Materials and Devices for the Internet of Things: Information, Sensing and Energy Applications [1 ed.]
 0128183861, 9780128183861

Table of contents :
Emerging 2D Materials and Devices for the Internet of Things
Copyright
Contents
List of contributors
1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches
1.1 Introduction to two-dimensional nonvolatile resistive memory
1.2 Two-dimensional materials preparation and memory device fabrication
1.2.1 Preparation and characterization of two-dimensional monolayers
1.2.2 Fabrication of memory devices
1.2.2.1 Crossbar
1.2.2.2 Litho-free and transfer-free
1.2.2.3 Exfoliation
1.3 Two-dimensional nonvolatile resistive memory
1.3.1 Nonvolatile resistive memory based on different device conditions
1.3.1.1 Crossbar device
1.3.1.2 Litho-free and transfer-free device (no polymer contamination)
1.3.1.3 Single crystalline device (no grain boundary)
1.3.1.4 Using different metal as electrodes
1.3.2 Memory performance
1.3.2.1 Reliability
1.3.2.2 Pulse operation
1.3.2.3 Flexibility
1.4 Switching mechanics
1.4.1 Factors influencing resistive switching
1.4.1.1 Temperature dependence
1.4.1.2 Device area dependence
1.4.1.3 Compliance current dependence
1.4.1.4 Voltage sweep rate and MoS2 layer number dependence
1.4.2 Possible switching mechanics based on ab initio simulation
1.5 MoS2 radio frequency switches
1.5.1 Introduction to radio frequency switch
1.5.2 Fabrication and measurement of MoS2 radio frequency switch
1.5.3 MoS2 radio frequency switch performance
1.6 Summary
Acknowledgment
References
2 Two-dimensional materials-based radio frequency wireless communication and sensing systems for Internet-of-things applica...
2.1 Introduction
2.2 Radio frequency performance of two-dimensional transistors
2.3 Frequency mixers and signal modulators based on two-dimensional transistors
2.4 Integrated wireless Internet-of-things sensors
2.5 Radio frequency energy harvesting using two-dimensional electronic devices
2.6 Conclusion
References
3 Graphene electronic tattoo sensors for point-of-care personal health monitoring and human–machine interfaces
3.1 Introduction
3.2 Theoretical background
3.2.1 Elastic membrane-skin conformability
3.2.2 Electrical model of skin-conformal and skin-nonconformal dry sensors
3.3 Fabrication of graphene electronic tattoo sensors
3.4 Applications of graphene electronic tattoo sensors and effects of the thickness on performance
3.4.1 Skin temperature sensing
3.4.2 Skin hydration sensing
3.4.3 Electrocardiography
3.4.4 Electroencephalography
3.4.5 Electromyography
3.4.6 Electrooculography
3.4.7 Human–machine interface
3.5 Conclusion
References
4 Transition metal dichalcogenides as ultrasensitive and high-resolution biosensing nodes
4.1 New opportunities for biosensing devices
4.2 Electronic biosensors made from transition metal dichalcogenides
4.3 Biosensors based on optical and optoelectronic properties of transition metal dichalcogenides
4.4 Biosensors based on structural properties of transition metal dichalcogenides
4.5 Final remarks
References
5 Nanophotonics and optoelectronics based on two-dimensional MoS2
5.1 MoS2-based nanoplasmonics
5.1.1 Exciton–plasmon interactions in MoS2
5.1.2 Plasmonic hot-electron injection
5.1.3 Surface plasmons in highly doped MoS2
5.1.4 Nanofabrication of plasmonic-MoS2 structures
5.2 MoS2-based optoelectronics
5.2.1 MoS2-based photodetectors
5.2.2 MoS2-based solar cells
5.2.3 MoS2-based light-emitting diodes
5.2.4 MoS2-optical cavity systems for enhanced light-emitting performance
5.3 Summary
References
6 Graphene-based anode materials for lithium-ion batteries
6.1 Introduction
6.2 Lithium-ion batteries and anode materials
6.2.1 Fundamentals of lithium-ion batteries
6.2.2 Challenges on anode materials
6.3 Graphene and graphene-based composites as anode materials
6.3.1 Graphene anodes
6.3.1.1 Porous graphene anodes
6.3.1.2 Doped graphene anodes
6.3.2 Graphene-based nanocomposite anodes
6.3.2.1 Graphene/insertion-type anodes
6.3.2.2 Graphene/alloy-type anodes
6.3.2.3 Graphene/conversion-type anodes
6.4 Conclusion and outlook
References
7 Two-dimensional materials as photoelectrodes in water reduction devices for energy applications
7.1 Basic mechanism of solar water splitting
7.2 Design principles of photoelectrochemical cells for water splitting
7.3 Two-dimensional materials as conducting channels
7.4 Two-dimensional materials as charge mediator/separator
7.5 Two-dimensional materials as cocatalysts
7.6 Two-dimensional materials as other roles
7.7 Summary and perspectives
References
8 Two-dimensional Xenes and their device concepts for future micro- and nanoelectronics and energy applications
8.1 Introduction
8.2 First-generation Xenes
8.2.1 Silicene
8.2.2 Germanene
8.2.3 Stanene
8.2.4 Plumbene
8.3 Second-generation Xenes
8.3.1 Borophene
8.3.2 Gallenene
8.3.3 Phosphorene
8.3.4 Arsenene
8.3.5 Antimonene
8.3.6 Bismuthene
8.3.7 Selenene
8.3.8 Tellurene
8.4 Perspectives and conclusion
References
9 Piezoelectric one- to two-dimensional nanomaterials for vibration energy harvesting devices
9.1 Introduction
9.2 Preparation and characterization of piezoelectric 1–2D nanomaterials
9.2.1 BZT-BCT nanofilm
9.2.2 Piezoelectric nanofibers
9.3 Piezoelectric 1–2D nanomaterial for energy harvesting
9.3.1 Nanogenerator
9.3.2 Self-charging power cell
9.3.3 Strain sensor
9.3.4 Dye degradation
9.4 Conclusion
Acknowledgment
References
10 Nanocomposite materials for nano-electronic-based Internet of things sensors and energy device signaling
10.1 Introduction
10.2 Nanocomposite materials for chemical sensory devices and Internet of things
10.2.1 Composite materials based on carbon nanotube/graphene and functional building blocks
10.2.1.1 Organic polymer–functionalized carbon nanomaterials
10.2.1.2 Metal oxide–functionalized carbon nanomaterials
10.2.1.3 Metal-functionalized carbon nanomaterials
10.2.1.4 Interface between carbon nanomaterials and the functionalization layer
10.2.2 Nano-electronic-based sensory devices
10.2.2.1 Device configuration
10.2.2.2 Device performance: chemiresistor and field-effect transistor
10.2.3 Chemical sensors from single-walled carbon nanotube-based composites and their applications in breath analysis
10.2.3.1 Single-walled carbon nanotube/PAni core/shell composites for chemical sensing
10.2.3.2 Breath analysis
10.2.4 Perspectives and challenges of nano-electronic sensors in Internet of things technology
10.3 Electronic sensing and signaling for sustainable energy devices
10.3.1 Principles and methodology of nano-electronic approach for chemical signaling
10.3.2 Application of electrical transportation spectroscopy for energy device investigations
10.3.2.1 New insights on various energy conversion reactions
10.3.2.2 Monitoring of anionic chemisorption and interfacial competition with reactive intermediates in oxygen reduction re...
10.3.2.3 Application in the bioelectrochemical system
10.3.3 Benefits and challenges of energy device signaling in Internet of things
References
11 Prospects and challenges in low-dimensional materials and devices for Internet of things
11.1 Flexible and wearable devices for Internet of things
11.1.1 Substrates
11.1.2 Two-dimensional materials as a functional layer
11.1.2.1 Graphene
11.1.2.2 Transition metal dichalcogenides
11.1.3 Progress and challenges
11.2 Human–machine interface devices for Internet of things
11.2.1 Internet of things and human–machine interface
11.2.2 A wearable sensor in human–machine interface system
11.2.3 Electronic skin
11.2.3.1 Pressure sensor for electronic skin
11.2.3.2 Human–machine interface enabled by triboelectric nanogenerator
11.2.3.3 Electric signal recordings for human–machine interface
11.2.3.4 Multifunctional human–machine interface sensors
11.2.4 Summary and prospective
11.3 Two-dimensional multifunctional device node for Internet of things
11.3.1 Sensor, modulator, and memory multifunction
11.3.2 Amplitude, frequency, and phase position hybrid modulation
11.3.3 Sensing, radio frequency, and energy collection simultaneously
11.3.4 Prospective and challenges
11.4 Sustainable energy devices for Internet of things
11.4.1 Fuel cell
11.4.2 Supercapacitors
11.4.3 Energy harvesting systems
11.4.3.1 Solar cells
11.4.3.2 Thermoelectric generators
11.4.4 Summary
11.5 5G/6G technology engaging with Internet of things
11.5.1 High bandwidth
11.5.2 Low-latency, real-time data communication
11.5.3 Artificial intelligence
11.5.4 Outlook
Acknowledgments
References
Index
Пустая страница

Citation preview

EMERGING 2D M AT E R I A L S A N D DEVICES FOR THE INTERNET OF THINGS

EMERGING 2D M AT E R I A L S A N D DEVICES FOR THE INTERNET OF THINGS Information, Sensing and Energy Applications

Edited by LI TAO School of Materials Science and Engineering, Southeast University, Nanjing, P.R. China

DEJI AKINWANDE Department of Electrical and Computer Engineering, University of Texas – Austin, United States

Elsevier Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States Copyright © 2020 Elsevier Inc. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress ISBN: 978-0-12-818386-1 For Information on all Elsevier publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Matthew Deans Acquisitions Editor: Simon Holt Editorial Project Manager: Isabella C. Silva Production Project Manager: Prasanna Kalyanaraman Cover Designer: Christian J. Bilbow Typeset by MPS Limited, Chennai, India

Contents List of contributors ................................................. xi

1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches .......................................................................1 Ruijing Ge, Xiaohan Wu, Myungsoo Kim, Jack C. Lee and Deji Akinwande 1.1 Introduction to two-dimensional nonvolatile resistive memory........................ 1 1.2 Two-dimensional materials preparation and memory device fabrication .................... 3 1.3 Two-dimensional nonvolatile resistive memory........................................................... 7 1.4 Switching mechanics ................................... 12 1.5 MoS2 radio frequency switches .................. 18 1.6 Summary ...................................................... 25 Acknowledgment ................................................ 25 References ........................................................... 25

2 Two-dimensional materials-based radio frequency wireless communication and sensing systems for Internet-of-things applications ............................................................... 29 Liang Zhu, Mohamed Farhat, Khaled Nabil Salama and Pai-Yen Chen 2.1 Introduction .................................................. 29 2.2 Radio frequency performance of two-dimensional transistors........................ 32 2.3 Frequency mixers and signal modulators based on two-dimensional transistors ....... 45 2.4 Integrated wireless Internet-of-things sensors .......................................................... 48 2.5 Radio frequency energy harvesting using two-dimensional electronic devices ........... 51 v

vi

Contents

2.6 Conclusion .................................................... 52 References ........................................................... 53

3 Graphene electronic tattoo sensors for point-of-care personal health monitoring and human machine interfaces ..................................59 Shideh Kabiri Ameri and Liu Wang 3.1 Introduction .................................................. 59 3.2 Theoretical background ............................... 61 3.3 Fabrication of graphene electronic tattoo sensors .......................................................... 65 3.4 Applications of graphene electronic tattoo sensors and effects of the thickness on performance ................................................. 71 3.5 Conclusion .................................................... 85 References ........................................................... 85

4 Transition metal dichalcogenides as ultrasensitive and high-resolution biosensing nodes......................................................87 Xiaogan Liang 4.1 New opportunities for biosensing devices .......................................................... 87 4.2 Electronic biosensors made from transition metal dichalcogenides ................ 94 4.3 Biosensors based on optical and optoelectronic properties of transition metal dichalcogenides...........................................103 4.4 Biosensors based on structural properties of transition metal dichalcogenides...........107 4.5 Final remarks ...............................................112 References ..........................................................113

5 Nanophotonics and optoelectronics based on two-dimensional MoS2 .........................................121 Zilong Wu, Linhan Lin and Yuebing Zheng 5.1 MoS2-based nanoplasmonics ....................121

Contents

5.2 MoS2-based optoelectronics ......................127 5.3 Summary .....................................................133 References ..........................................................134

6 Graphene-based anode materials for lithium-ion batteries ..............................................139 Hui Xu, Zhengming Sun and Jian Chen 6.1 Introduction .................................................139 6.2 Lithium-ion batteries and anode materials ......................................................140 6.3 Graphene and graphene-based composites as anode materials..................143 6.4 Conclusion and outlook ..............................158 References ..........................................................159

7 Two-dimensional materials as photoelectrodes in water reduction devices for energy applications .............................................................165 Li Ji, Xingli Zou, Hsien-Yi Hsu, Kai Huang, Na Gao, Hao Zhu, Lin Chen, Qingqing Sun, Peng Zhou and David Wei Zhang 7.1 Basic mechanism of solar water splitting ........................................................167 7.2 Design principles of photoelectrochemical cells for water splitting ...............................168 7.3 Two-dimensional materials as conducting channels ...................................170 7.4 Two-dimensional materials as charge mediator/separator......................................172 7.5 Two-dimensional materials as cocatalysts ...................................................174 7.6 Two-dimensional materials as other roles..............................................................176 7.7 Summary and perspectives........................177 References ..........................................................177

vii

viii

Contents

8 Two-dimensional Xenes and their device concepts for future micro- and nanoelectronics and energy applications .......................................181 Carlo Grazianetti, Christian Martella and Alessandro Molle 8.1 Introduction .................................................181 8.2 First-generation Xenes................................183 8.3 Second-generation Xenes ..........................195 8.4 Perspectives and conclusion ......................207 References ..........................................................208

9 Piezoelectric one- to two-dimensional nanomaterials for vibration energy harvesting devices .................................................221 Ruijian Zhu and Zengmei Wang 9.1 Introduction .................................................221 9.2 Preparation and characterization of piezoelectric 1 2D nanomaterials .............223 9.3 Piezoelectric 1 2D nanomaterial for energy harvesting .......................................229 9.4 Conclusion ...................................................240 Acknowledgment ...............................................240 References ..........................................................240

10 Nanocomposite materials for nano-electronic-based Internet of things sensors and energy device signaling .............243 Congyue Liu, Bailin Tian and Mengning Ding 10.1 Introduction .............................................243 10.2 Nanocomposite materials for chemical sensory devices and Internet of things .....................................245 10.3 Electronic sensing and signaling for sustainable energy devices ....................263 References ..........................................................285

Contents

11 Prospects and challenges in low-dimensional materials and devices for Internet of things .......................................................................291 Anhan Liu, Siyao Jiang, Zhengrui Zhu, Sixin Zhang, Dingxuan Kang and Li Tao 11.1 Flexible and wearable devices for Internet of things .....................................292 11.2 Human machine interface devices for Internet of things ...............................297 11.3 Two-dimensional multifunctional device node for Internet of things..........305 11.4 Sustainable energy devices for Internet of things .....................................313 11.5 5G/6G technology engaging with Internet of things .....................................319 Acknowledgments .............................................323 References ..........................................................324 Index..................................................................... 329

ix

List of contributors Deji Akinwande Microelectronics Research Center, Department of Electrical and Computer Engineering, University of Texas at Austin, Austin, TX, United States Jian Chen Jiangsu Key Laboratory of Advanced Metallic Materials, School of Materials Science and Engineering, Southeast University, Nanjing, P.R. China Lin Chen State Key Laboratory of ASIC and System, School of Microelectronics, Fudan University, Shanghai, P.R. China Pai-Yen Chen Department of Electrical and Computer Engineering, University of Illinois at Chicago, Chicago, IL, United States Mengning Ding Key Laboratory of Mesoscopic Chemistry, School of Chemistry and Chemical Engineering, Nanjing University, Nanjing, P.R. China Mohamed Farhat Computer, Electrical, and Mathematical Science and Engineering (CEMSE) Division, King Abdullah University of Science and Technology (KAUST), Thuwal, Saudi Arabia Na Gao Collaborative Innovation Center for Optoelectronic Semiconductors and Efficient Devices, Department of Physics, Xiamen University, Xiamen, P.R. China Ruijing Ge Microelectronics Research Center, Department of Electrical and Computer Engineering, University of Texas at Austin, Austin, TX, United States Carlo Grazianetti CNR-IMM, Unit of Agrate Brianza, Agrate Brianza, Italy Hsien-Yi Hsu Department of Materials Science and Engineering, School of Energy and Environment, City University of Hong Kong, Kowloon Tong, P.R. China Kai Huang Collaborative Innovation Center for Optoelectronic Semiconductors and Efficient Devices, Department of Physics, Xiamen University, Xiamen, P.R. China

xi

xii

List of contributors

Li Ji State Key Laboratory of ASIC and System, School of Microelectronics, Fudan University, Shanghai, P.R. China Siyao Jiang Jiangsu Key Laboratory of Advanced Metallic Materials, School of Materials Science and Engineering, Southeast University, Nanjing, P.R. China Shideh Kabiri Ameri Department of Electrical and Computer Engineering, Queen’s University, Kingston, ON, Canada Dingxuan Kang Jiangsu Key Laboratory of Advanced Metallic Materials, School of Materials Science and Engineering, Southeast University, Nanjing, P.R. China Myungsoo Kim Microelectronics Research Center, Department of Electrical and Computer Engineering, University of Texas at Austin, Austin, TX, United States Jack C. Lee Microelectronics Research Center, Department of Electrical and Computer Engineering, University of Texas at Austin, Austin, TX, United States Xiaogan Liang Mechanical Engineering Department, University of Michigan, Ann Arbor, MI, United States Linhan Lin State Key Laboratory of Precision Measurement Technology and Instruments, Department of Precision Instrument, Tsinghua University, Beijing, P.R. China Anhan Liu Jiangsu Key Laboratory of Advanced Metallic Materials, School of Materials Science and Engineering, Southeast University, Nanjing, P.R. China Congyue Liu Key Laboratory of Mesoscopic Chemistry, School of Chemistry and Chemical Engineering, Nanjing University, Nanjing, P.R. China Christian Martella CNR-IMM, Unit of Agrate Brianza, Agrate Brianza, Italy Alessandro Molle CNR-IMM, Unit of Agrate Brianza, Agrate Brianza, Italy Khaled Nabil Salama Computer, Electrical, and Mathematical Science and Engineering (CEMSE) Division, King Abdullah University of Science and Technology (KAUST), Thuwal, Saudi Arabia Qingqing Sun State Key Laboratory of ASIC and System, School of Microelectronics, Fudan University, Shanghai, P.R. China

List of contributors

Zhengming Sun Jiangsu Key Laboratory of Advanced Metallic Materials, School of Materials Science and Engineering, Southeast University, Nanjing, P.R. China Li Tao Jiangsu Key Laboratory of Advanced Metallic Materials, School of Materials Science and Engineering, Southeast University, Nanjing, P.R. China Bailin Tian Key Laboratory of Mesoscopic Chemistry, School of Chemistry and Chemical Engineering, Nanjing University, Nanjing, P.R. China Liu Wang Department of Mechanical Engineering, Massachusetts Institute of Technology, Boston, MA, United States Zengmei Wang Jiangsu Key Laboratory of Construction Materials, School of Materials Science and Engineering, Southeast University, P.R. China Xiaohan Wu Microelectronics Research Center, Department of Electrical and Computer Engineering, University of Texas at Austin, Austin, TX, United States Zilong Wu Walker Department of Mechanical Engineering, Materials Science and Engineering Program, Texas Materials Institute, The University of Texas at Austin, Austin, TX, United States Hui Xu Jiangsu Key Laboratory of Advanced Metallic Materials, School of Materials Science and Engineering, Southeast University, Nanjing, P.R. China David Wei Zhang State Key Laboratory of ASIC and System, School of Microelectronics, Fudan University, Shanghai, P.R. China Sixin Zhang Jiangsu Key Laboratory of Advanced Metallic Materials, School of Materials Science and Engineering, Southeast University, Nanjing, P.R. China Yuebing Zheng Walker Department of Mechanical Engineering, Materials Science and Engineering Program, Texas Materials Institute, The University of Texas at Austin, Austin, TX, United States Peng Zhou State Key Laboratory of ASIC and System, School of Microelectronics, Fudan University, Shanghai, P.R. China Hao Zhu State Key Laboratory of ASIC and System, School of Microelectronics, Fudan University, Shanghai, P.R. China

xiii

xiv

List of contributors

Liang Zhu Department of Electrical and Computer Engineering, University of Illinois at Chicago, Chicago, IL, United States Ruijian Zhu Jiangsu Key Laboratory of Construction Materials, School of Materials Science and Engineering, Southeast University, Nanjing, P.R. China; School of Materials Engineering, Nanjing Institute of Technology, Nanjing, P.R. China; Jiangsu Key Laboratory of Advanced Structural Materials and Application Technology, Nanjing, P.R. China Zhengrui Zhu Jiangsu Key Laboratory of Advanced Metallic Materials, School of Materials Science and Engineering, Southeast University, Nanjing, P.R. China Xingli Zou State Key Laboratory of Advanced Special Steel, Shanghai Key Laboratory of Advanced Ferrometallurgy, School of Materials Science and Engineering, Shanghai University, Shanghai, P.R. China

1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches Ruijing Ge, Xiaohan Wu, Myungsoo Kim, Jack C. Lee and Deji Akinwande Microelectronics Research Center, Department of Electrical and Computer Engineering, University of Texas at Austin, Austin, TX, United States

Chapter Outline 1.1 Introduction to two-dimensional nonvolatile resistive memory 1 1.2 Two-dimensional materials preparation and memory device fabrication 3 1.2.1 Preparation and characterization of two-dimensional monolayers 3 1.2.2 Fabrication of memory devices 3 1.3 Two-dimensional nonvolatile resistive memory 7 1.3.1 Nonvolatile resistive memory based on different device conditions 7 1.3.2 Memory performance 11 1.4 Switching mechanics 12 1.4.1 Factors influencing resistive switching 12 1.4.2 Possible switching mechanics based on ab initio simulation 16 1.5 MoS2 radio frequency switches 18 1.5.1 Introduction to radio frequency switch 18 1.5.2 Fabrication and measurement of MoS2 radio frequency switch 20 1.5.3 MoS2 radio frequency switch performance 21 1.6 Summary 25 Acknowledgment 25 References 25

1.1

Introduction to two-dimensional nonvolatile resistive memory

Worldwide advancement in wireless communication and connectivity systems for Internet-of-things (IoT) has resulted in an ever-increasing demand for memories [1]. Over the past few Emerging 2D Materials and Devices for the Internet of Things. DOI: https://doi.org/10.1016/B978-0-12-818386-1.00001-1 © 2020 Elsevier Inc. All rights reserved.

1

2

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

decades, tremendous efforts have been made to develop highdensity, low-cost, and nonvolatile memory devices [2]. Compared with volatile memory such as dynamic random access memory and static random access memory dissipating both dynamic and static energy, nonvolatile memory with zero-static power is attractive considering energy efficiency. One representative nonvolatile memory, Flash, has the largest solid-state nonvolatile memory market [3]. However, since it faces its scalability limit due to short channel effect and channel boosting leakage, researchers have been working on some emerging memories to meet the new requirements [4]. One emerging memory alternative is phasechange memory (PCM) based on phase-change materials, which can be reversibly switched between crystalline and amorphous phases through thermal processes, resulting in a change in resistivity. However, PCM is primarily limited due to thermal proximity effects. Another major emerging nonvolatile memory candidate, resistive random-access memory (RRAM), does not depend on thermal processes and has lower power consumption compared with PCM [5]. Conventional RRAM structure is simply an oxide material sandwiched between metal electrodes, called metal-insulatormetal (MIM) structure. It is worthwhile to notice that RRAM has not only simple structure and materials, but also low operating voltage, low energy consumption, high operating speed, high density, long retention, and wide-range state modulation. RRAM works based on nonvolatile resistive switching (NVRS) between a high-resistance state (HRS) and a low-resistance state (LRS). The switching event from HRS to LRS is called the “SET” process, while the transition from LRS to HRS is called the “RESET” process. Usually the fresh samples require a higher voltage than SET to trigger on the resistive switching behavior and this activation step is called “forming” process. The resistive switching can be generally classified into two modes: unipolar and bipolar. As for unipolar, SET/RESET can occur at the same bias polarity, while bipolar means that the SET/RESET requires opposite bias polarity. To continue scaling and reducing cell size, our future toward high-density memory would be made up of cells with several atoms or a cluster of molecules. Thus two-dimensional (2D) materials are promising candidates to overcome vertical scaling obstacle in NVRS [6,7]. Recently, NVRS has been observed in various multilayer 2D materials including graphene oxide, partially degraded black phosphorus, functionalized MoS2 and composites, transition metal dichalcogenide (TMD)-based hybrids and multilayer hexagonal boron nitride (h-BN) [712]. However, it was believed that NVRS phenomenon was not accessible in

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

single-layer atomic sheets [7,13] due to excessive leakage current that is known to prevent nanometer scaling in conventional oxide-based vertical MIM configuration. While Sangwan et al. discovered that grain boundaries in single-layer MoS2 can produce resistive switching based on planar structures and it has been attributed to the defect migration at certain grain boundaries [14]. However, the planar structure without three-dimensional (3D) stacking ability has limitation of low integration density. In this chapter, we demonstrated the application of 2D monolayer atomic sheets (TMDs and h-BN) in nonvolatile resistive memory using MIM vertical structure. These devices can be labeled as “atomristor,” which means the memristor effect in atomically thin nanomaterials. Among 2D memory devices, atomristor stands out due to the atomic thinness of the active layer, low switching voltage, forming-free characteristic, large ON/OFF current ratio, and fast switching speed. In the last section, another major application based on atomristor, monolayer MoS2-based radio frequency (RF) switch will be presented. The results discussed in this chapter have been organized and reproduced with permissions based on several representative publications in this field [1518].

1.2 1.2.1

Two-dimensional materials preparation and memory device fabrication Preparation and characterization of two-dimensional monolayers

Monolayer MoS2 was prepared via various methods, namely chemical vapor deposition (CVD) [19], metal-organic chemical vapor deposition (MOCVD) [20], and exfoliation on SiO2/Si substrates or Au foil [21]. CVD-grown monolayer h-BN was synthesized on Ni foil [22,23] and Au foil [24] substrates. Several materials characterizations for TMDs and h-BN have been conducted to verify the quality, thickness, and uniformity as shown in Figs. 1.1 and 1.2.

1.2.2

Fabrication of memory devices

The process flows and optical images of (1) crossbar, (2) lithofree and transfer-free, and (3) exfoliated memristors based on monolayer MoS2 are shown in Fig. 1.3. Other materials-based memristors can be constructed using similar processes and stackings. To avoid effects due to metal oxides, inert metal, gold, is

3

4

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

Figure 1.1 (A) SEM image of as-grown MOCVD monolayer MoS2 indicating uniformity. (B) Raman and (C) PL spectrum of MOCVD-grown monolayer MoS2, MoSe2, WS2, and WSe2. (D) Atomically resolved scanning tunneling microscope image of monolayer MoS2 grown on Au (100). The S vacancy defects (B1012/cm2) are indicated by the dashed circle. (E) High-resolution TEM image of MoS2 grown on Au foil on a folded edge showing the monolayer feature of the MoS2 film. (F) PL intensity mapping of exfoliated monolayer MoS2 flake. Inset: Optical image of the same flake. Source: (AE) Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society. (F) Reprinted with permission from R. Ge, X. Wu, M. Kim, P.-A. Chen, J. Shi, J. Choi, et al., Atomristors: memory effect in atomically-thin sheets and record RF switches, in: 2018 IEEE International Electron Devices Meeting (IEDM), 15 December 2018, IEEE, 2018, pp. 22.6.122.6.4. doi:10.1109/IEDM.2018.8614602. Copyright (2018) IEEE.

Figure 1.2 (A) AFM images of transferred samples. Inset: the height profile analysis yields an average height of B0.49 nm relative to the substrate, suggesting monolayer h-BN. (B) Atomic-resolution scanning tunneling microscope image of as-grown samples, which shows the representative honeycomb structure of h-BN with a lattice constant of B0.25 nm, again suggesting the formation of a high-quality h-BN layer on Au foil. (C) SEM images of as-grown samples on Au foil indicating fully coverage with monolayer CVD-grown h-BN films. The wrinkles are derived from the thermal expansion coefficient difference between Au foils and h-BN films. Source: Reprinted with permission from X. Wu, R. Ge, P.-A. Chen, H. Chou, Z. Zhang, Y. Zhang, et al., Thinnest nonvolatile memory based on monolayer h-BN, Adv. Mater. 31 (2019). doi:10.1002/adma.201806790. Copyright (2019) John Wiley and Sons.

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

5

Figure 1.3 Process flows and optical images of crossbar, litho-free, and transfer-free, and exfoliation samples. Source: (AD and F) Reprinted with permission from R. Ge, X. Wu, M. Kim, P.-A. Chen, J. Shi, J. Choi, et al., Atomristors: memory effect in atomically-thin sheets and record RF switches, in: 2018 IEEE International Electron Devices Meeting (IEDM), 15 December 2018, IEEE, 2018, pp. 22.6.122.6.4. doi:10.1109/IEDM.2018.8614602. Copyright (2018) IEEE. (E) Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society.

used as electrodes (if not specified) and thus ensures that 2D materials play the active role in the resistive switching behavior.

1.2.2.1

Crossbar

Fabrication started with bottom electrode (BE) patterning by electron beam lithography (EBL) and 2 nm Cr (as adhesion

6

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

Figure 1.4 Process flow of PDMS transfer. Source: Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society.

layer)/60 nm Au metal stack deposition on an SiO2/Si (285 nm) substrate. Single-layer TMD was then transferred onto the fabricated substrate using polydimethylsiloxane (PDMS) stamp transfer method (Fig. 1.4). During PDMS transfer process, monolayer TMD was brought into conformal contact with PDMS. Then substrate-TMD-PDMS system was soaked into deionized water. Since the original substrate (SiO2) is hydrophilic, it is easy for water to diffuse into the TMDsubstrate interface, which helps separate the two layers. The PDMS-TMD film was then brought into contact with the target-fabricated substrate. The PDMS stamp was subsequently peeled off to leave monolayer TMD films on the target substrate. While CVD h-BN was transferred onto BE from the Ni foil substrate using a poly(methyl methacrylate) (PMMA)-assisted wet transfer method. A thin layer of PMMA was spin-coated onto the h-BN/Ni and then the Ni was etched away in 0.5 M ammonia persulfate solution. The PMMA/h-BN was rinsed in deionized water to remove any etchant by-product prior to lifting by the target substrate with BE. The PMMA was then removed by immersing in acetone. Top electrode (TE), using the same fabrication process as BE, was patterned and deposited.

1.2.2.2 Litho-free and transfer-free Monolayer MoS2 and h-BN was directly grown on Au foils [22,24]. TE was then deposited through shadow mask without any lithography. Importantly, the litho- and transfer-free device represents a near-ideal clean device and serves to confirm the memory effect is an intrinsic property.

1.2.2.3 Exfoliation Monolayer MoS2 was mechanically exfoliated from bulk crystal onto the deposited Au film on SiO2/Si substrate. TE was patterned using EBL.

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

1.3

Two-dimensional nonvolatile resistive memory

1.3.1

Nonvolatile resistive memory based on different device conditions

1.3.1.1

Crossbar device

DC electrical measurements were performed on as-fabricated crossbar devices consisting of atomic sheets with Au BE and TE and revealed nonvolatile resistance switching in the monolayer MX2 and h-BN active layers (Fig. 1.5). For instance, MoS2, the prototypical TMD, featured low currents corresponding to an HRS until the application of B1 V, which SET the atomic-layer switch to an LRS that persists until a negative voltage is applied to RESET it (Fig. 1.5A). A compliance current is applied at SET process to prevent irreversible breakdown, while no compliance current is applied at RESET process. Interestingly, the single-layer nonvolatile switch required no electroforming step, a prerequisite in transition metal oxides (TMOs) that initialize a soft dielectric breakdown to form a conductive filament for subsequent NVRS operation [5,25]. Although it has been shown that electroforming can be avoided with thickness scaling into the nm-regime, excessive leakage current from trap-assisted tunneling is a limiting consequence [5,26]. Here, an ON/OFF ratio above 104 can be achieved, highlighting a defining advantage of crystalline monolayers over ultrathin amorphous oxides. Certain single-layer MoS2 devices of the same MIM construction feature unipolar switching where voltage of the same polarity is used for both SET and RESET programming (Fig. 1.5B). Motivated by the observation of NVRS in MoS2, the quartet of single-layer MX2 (MoSe2, WS2, and WSe2) was investigated and all showed similar intriguing characteristics of bipolar switching and unipolar switching as demonstrated in Fig. 1.5CH. Similar qualitative results were achieved with a typical 2D insulator material h-BN, in lieu of TMDs, in the MIM configuration (Fig. 1.5I and J). These collective results of NVRS in representative atomic sheets allude to a universal effect in nonmetallic monolayers, which opens a new avenue of scientific research on defects, charge, and interfacial phenomena at the atomic scale and the associated materials design for diverse applications. It should be noted that the unipolar devices can also work at opposite bias, which in some cases is referred to the nonpolar switching mode. We will mainly use MoS2 as the active layer in the following experiments and discussions owing to its greater material maturity.

7

8

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

Figure 1.5 Typical IV curves of monolayer TMD atomristors. (A) Representative IV curve of bipolar resistive switching behavior in monolayer MoS2 crossbar device with lateral area of 2 3 2 μm2. Step 1: voltage increases from 0 to 1.2 V. At B1 V, the current abruptly increases to compliance current, indicating a transition (SET) from high-resistance state (HRS) to low-resistance state (LRS). Step 2: voltage decreases from 1.2 to 0 V. The device remains in LRS. Step 3: voltage increases from 0 to 21.5 V. At 21.25 V, the current abruptly decreases, indicating a transition (RESET) from LRS to HRS. Step 4: voltage decreases from 21.5 to 0 V. The device remains in HRS until next cycle. (B) Representative IV curve of unipolar resistive switching behavior in monolayer MoS2 crossbar device with lateral area of 2 3 2 μm2. For unipolar operation, both SET and RESET transitions are achieved under positive bias. Representative IV curves of bipolar and unipolar resistive switching behavior in monolayer (C and D) MoSe2, (E and F) WS2, (G and H) WSe2, and (I and J) h-BN crossbar MIM devices alluding to a universal nonvolatile phenomenon in nonmetallic atomic sheets. The areas of the crossbar devices are 0.4 3 0.4 μm2 for MoSe2, 2 3 μm2 for WS2, 2 3 2 μm2 for WSe2, and 1 3 1 μm2 for h-BN. Source: (AH) Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society. (I and J) Reprinted with permission from X. Wu, R. Ge, P.-A. Chen, H. Chou, Z. Zhang, Y. Zhang, et al., Thinnest nonvolatile memory based on monolayer h-BN, Adv. Mater. 31 (2019). doi:10.1002/adma.201806790. Copyright (2019) John Wiley and Sons.

1.3.1.2 Litho-free and transfer-free device (no polymer contamination) To rule out the undesirable effect of polymer contamination from microfabrication, very clean devices, lithography-free and transfer-free devices, were made, which also produced the NVRS

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

9

Figure 1.6 Representative IV curves of resistive switching behavior in litho-free and transfer-free MIM devices based on (A) monolayer MoS2 grown on Au foil and monolayer h-BN grown on (B) Au foil and (C) Ni foil. Source: (A) Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society. (B and C) Reprinted with permission from X. Wu, R. Ge, P.-A. Chen, H. Chou, Z. Zhang, Y. Zhang, et al., Thinnest nonvolatile memory based on monolayer h-BN, Adv. Mater. 31 (2019). doi:10.1002/adma.201806790. Copyright (2019) John Wiley and Sons.

effect (Fig. 1.6), alluding to an intrinsic origin. The lithographyfree and transfer-free devices are based on monolayer MoS2 grown directly on gold foil [21] and monolayer h-BN on Au foil and Ni foil [22,24]. Subsequently, gold TE is deposited using ebeam evaporation via laser shadow mask. Thus no transfer process or lithography process is used, excluding possible residues from transfer and lithography.

1.3.1.3

Single crystalline device (no grain boundary)

It has been previously reported that line or grain boundary defects in polycrystalline monolayer MoS2 or multilayer h-BN play an intrinsic role in switching [14,27]. However, it is not an exclusive factor from a MIM device realized on a single-crystal (boundary-free) CVD-grown MoS2 flake (Fig. 1.7A) and exfoliated MoS2 flake (Fig. 1.7B), highlighting the potential role of localized effects.

1.3.1.4

Using different metal as electrodes

The switching behavior is not only dependent on the active layer but also related to the metal electrodes and their interfacial properties. In the majority of experiments, Au was selected as an inert electrode to rule out any switching effect that might arise from interfacial metal oxide formation. However, the phenomenon is not restricted to inert electrodes since monolayer MoS2 with electrochemically active (Ag) electrodes (Fig. 1.8A) [8,25] and monolayer h-BN using Ni foil as a global BE (Fig. 1.6C) also produce NVRS. Moreover, monolayer graphene is also a

10

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

Figure 1.7 Typical IV curves of resistive switching behavior of single-crystal MoS2 flakes from (A) CVD growth and (B) exfoliation. The optical image of the CVD-grown boundary-free flake is shown in the inset, where the dashed line is an outline of the single-crystal MoS2 triangle. Source: (A) Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society. (B) Reprinted with permission from R. Ge, X. Wu, M. Kim, P.-A. Chen, J. Shi, J. Choi, et al., Atomristors: memory effect in atomically-thin sheets and record RF switches, in: 2018 IEEE International Electron Devices Meeting (IEDM), 15 December 2018, IEEE, 2018, pp. 22.6.122.6.4. doi:10.1109/IEDM.2018.8614602. Copyright (2018) IEEE.

Figure 1.8 Representative IV curve of MoS2 (A) litho-free device using silver as top and bottom electrodes and (B) crossbar device using graphene as the top electrode and gold as bottom electrode with an area of 1 3 1 μm2. Source: Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society.

suitable electrode option (Fig. 1.8B). Interestingly, these results open the design space for electrode engineering (work-function and interface redox) from inert to active metals to 2D semimetals, the latter offering the potential of atomically thin subnanometer switches for ultraflexible and dense nonvolatile computing fabrics.

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

1.3.2

Memory performance

1.3.2.1

Reliability

11

Reliability properties, such as cycling and retention, are critical when it comes to the applications, like resistive memories and RF switches. Presently, B120 manual DC cycling (Fig. 1.9A) is not yet sufficient to meet the stringent requirements for solid-state memory, a reflection of the nascent state of atomristors compared to TMO memristors, which had similar endurance (,103 cycles) in early research but has now advanced above 106 cycles [5]. Oxygenation by interface engineering or doping may improve endurance, similar to what was observed for amorphous carbon memory devices [28]. Retention of nonvolatile states tested up to a week (Fig. 1.9B) is already sufficient for certain neuromorphic applications involving short- and medium-term plasticity [29].

1.3.2.2

Pulse operation

Compared to DC cycling, pulse operation is used more commonly in industry. Pulse test can help voltage apply in a timecontrolled manner (often on the order of microseconds and nanoseconds) to prevent the device under test from overheating or overstress. Fast switching speed (,15 ns) is demonstrated via pulse SET operation of monolayer MoS2 as shown in Fig. 1.10. The read IV curves before and after applying 15 ns pulses clearly show the switching from OFF state to ON state.

Figure 1.9 (A) DC cycling of MoS2 crossbar MIM device with 120 manual DC switching cycles. (B) Time-dependent measurements of MoS2 crossbar switch featuring stable retention over a week at room temperature. The resistance of the HRS and LRS is determined by measuring the current at a small bias of 0.1 V, which cannot trigger switching. Source: Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society.

12

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

Figure 1.10 Fifteen-nanosecond pulse SET demonstration. The IV characteristics before and after pulse driver clearly show the switching from HRS to LRS. Source: Reprinted with permission from R. Ge, X. Wu, M. Kim, P.-A. Chen, J. Shi, J. Choi, et al., Atomristors: memory effect in atomically-thin sheets and record RF switches, in: 2018 IEEE International Electron Devices Meeting (IEDM), 15 December 2018, IEEE, 2018, pp. 22.6.122.6.4. doi:10.1109/IEDM.2018.8614602. Copyright (2018) IEEE.

1.3.2.3 Flexibility A major contemporary challenge concerns the choice of a switching layer material suitable for high-performance flexible memories, providing a flexible platform for the IoT. The MoS2 memristor on the polyimide (PI) flexible substrate reveals that both HRS and LRS afford retention after 1000 bending cycles with reproducible switching curves (Fig. 1.11). The high breaking strain and ease of integration of 2D materials on soft substrates can afford flexible nonvolatile digital that can endure mechanical cycling, beneficial to the applications in flexible IoT system.

1.4 1.4.1

Switching mechanics Factors influencing resistive switching

In order to gain insight into the underlying mechanism(s), electrical measurements over five degrees of freedom, namely, temperature, area scaling, compliance current, voltage sweep rate, and layer thickness, were investigated.

1.4.1.1 Temperature dependence The IV characteristics at different temperatures are analyzed to explain the electron transport mechanisms at LRS and

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

13

Figure 1.11 (A) Stable resistance of both HRS and LRS and (B) typical switching IV curve before and after 1000 cycles at 1% strain of MoS2 crossbar device. Inset: schematic of flexible device. These results show a promising application for flexible memory devices based on MoS2 atomristors. Source: Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society.

HRS. At LRS (Fig. 1.12A and D), metallic Ohmic conduction can be deduced for both MoS2 and h-BN since the current decreases as the temperature increases, and the normalized conductance Gn 5 ðdI=dV Þ=ðI=V Þ is pffiffiffiffiffiffiffiffiffiffiffiffi unity, a signature of linear  approximately transport, J~KV exp 2 4πd 2m ϕ=h , where J is the current density, m* is the effective mass, ϕ is the tunnel barrier height, h is Planck’s constant, and K is proportional to the lateral area (A) and dependent on the barrier parameters (m, ϕ, d) [30]. d is the 2D barrier thickness. At HRS (Fig. 1.12B), nonlinear IV characteristics are observed, with the current increasing as the temperature increases. Considering different transport models, the MoS2 HRS data were best-fitted by the Schottky emission model with good agreement (Fig. 1.12C) [30]; pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi     J~A T 2 exp 2 qðφB 2 qE=4πεr ε0 Þ=kT , A 5 120m =m0 , where A* is the effective Richardson constant, m0 is the free electron mass, T is the absolute temperature, q is the electronic charge, φB is the Schottky barrier height, E is the electric field across the dielectric, k is Boltzmann’s constant, ε0 is the permittivity in vacuum, and εr is the optical dielectric constant. The effective thickness of B1 nm is used and m*/m0 is B1. The extracted barrier height is B0.47 eV at 300K, and the refractive index pffiffiffiffi n 5 εr is 6.84. While the HRS data based on h-BN devices are best-fitted pby the ffi PooleFrenkel emission model ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi J~Eexp½2 qðφT 2 qE=πεr ε0 Þ=kT  [31,32], where J is the current density, T is the absolute temperature, q is the electronic charge, qφT is the trap energy level, E is the electric field across the

14

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

Figure 1.12 Temperature dependence of (AC) MoS2 and (DF) h-BN nonvolatile switching devices. (A and D) The “READ” IV characteristics at LRS based on MoS2 and h-BN crossbar devices at different temperatures indicating a metallic character. The inset shows the normalized conductance Gn. (B and E) The “READ” IV characteristics at HRS based on MoS2 and h-BN crossbar devices at different temperatures. The current increases as the temperature increases. The inset shows the normalized conductance Gn. (C) Fitted data using Schottky emission model for MoS2 HRS. (F) Fitted data using PooleFrenkel model for h-BN HRS. Source: (AC) Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society. (DF) Reprinted with permission from X. Wu, R. Ge, P.-A. Chen, H. Chou, Z. Zhang, Y. Zhang, et al., Thinnest nonvolatile memory based on monolayer h-BN, Adv. Mater. 31 (2019). doi:10.1002/adma.201806790. Copyright (2019) John Wiley and Sons.

h-BN layer, k is Boltzmann’s constant, ε0 is the permittivity in vacuum, and εr is the optical dielectric constant.

1.4.1.2 Device area dependence Area-scaling studies clearly show distinct profiles with the LRS relatively flat, while the HRS features a more complex relation (Fig. 1.13). The LRS profile is consistent with the theory of a single (or few) localized conductive link(s) [5,25,33]. Below 100 μm2, the HRS resistance scales inversely with area due to uniform conduction. For larger sizes, the resistance becomes area-invariant and is attributed to the presence of localized grain boundaries. We note that the average domain size of typical CVD MoS2 monolayer is B102103 μm2.

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

15

Figure 1.13 Area dependence of low- and high-resistance states with Au/1L-MoS2/Au structure. The resistances of each state are determined at a low voltage of 0.1 V. The line curves are visual guides. Source: Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society.

1.4.1.3

Compliance current dependence

The current and resistance dependence on compliance current (Fig. 1.14) reveals a linear scaling that can be credited to an increase in the cross-sectional area of a single filament or to the formation of multiple filaments [33]. From an application perspective, the programmable resistance states are suitable for multilevel nonvolatile memory that can store more than one bit per cell. In addition, the intrinsic low-resistance values, approaching B5 Ω (Fig. 1.14B), open a new application for low-power nonvolatile electronic RF switches, and we will discuss it in Section 1.5.

1.4.1.4

Voltage sweep rate and MoS2 layer number dependence

The dependence of the SET/RESET voltages on sweep rate (Fig. 1.15A) suggests that slower rates afford more time for ionic diffusion resulting in reduced voltages, an important consideration for low-voltage operation. Layer-dependent studies up to four layers demonstrate that the switching phenomena persists (Fig. 1.15A), with a distinction that the LRS resistance increases with layer number (Fig. 1.15B).

16

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

Figure 1.14 (A) Dependence of the READ current on the compliance current after SET process in MoS2 litho-free atomristor. Four separate resistance states (three ON states and one OFF state) are obtained in a single device by varying the compliance current at 20 mA (ON state I), 40 mA (ON state II), and 80 mA (ON state III). (B) Relationship between LRS resistance and compliance current indicating a sub-10 Ω resistance is achievable for RF switch applications. The fitting curve is obtained with an inverse quadratic model, y ~ x22. Source: Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society.

Figure 1.15 (A) Dependence of the SET and RESET voltages on the sweep rate. The area of this litho-free device is 15 3 15 μm2. (B) Layer-dependent IV characteristics of MoS2 litho-free MIM switches, each with an area of 15 3 15 μm2. (C) Relationship between LRS resistance and layer number of few-layer MoS2 litho-free devices. The straight line is a visual guide. For layer-dependent studies, the preparation method for monolayer is MOCVD and PDMS transfer, while few-layer devices are CVD-grown and wet transfer. Source: Reprinted with permission from R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. Copyright (2018) American Chemical Society.

1.4.2

Possible switching mechanics based on ab initio simulation

Based on the temperature-dependent conduction experiments, the fitted results indicate that trapping states and Schottky barrier are involved in the electron transportation at HRS [34,35]. While

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

area-dependent studies, the resistance switching of MoS2 devices can be explained by the proposed model that, in the SET process, the electrons are transported through a filamentary-like onedimensional (1D) conductive link, and in the RESET process, the conductive path is broken, resulting in a Schottky barrier at the device interfaces. Considering the structure of monolayer MoS2 and h-BN, intrinsic sulfur vacancies in MoS2 and boron vacancies in h-BN are energetically favorable and may serve as the localized trapping centers for electrons. While at LRS, the vacancies can be substituted by metal ions, leading to a more conductive Ohmic transport through the links formed by the metal ions. This proposal is further supported by ab initio simulation results. The simulation package, Atomistix ToolKit (ATK) from QuantumWise, is applied for ab initio simulation, including building optimization system and electronic analysis. The electronic properties are evaluated with density functional theory. From defect studies, intrinsic vacancies commonly exist in 2D materials, such as MoS2 and h-BN sheets, and have significant influence on the electron conduction [12,36,37]. The simulated configuration consists of a single-layer MoS2 sheet with a sulfur vacancy and a positive gold ion is placed over the sheet in Fig. 1.16A. After the optimization process, the positively charged gold ion is reduced at the sulfur vacancy and then chemisorbs on the MoS2 layer, which is shown as the final state in Fig. 1.16B. The simulated result for h-BN (Fig. 1.16C and D) shows the similar trend that positively charged Au ion can occupy boron vacancy. Moreover, negatively charged gold ion with sulfur vacancy (or boron vacancy) is investigated. However, no movement of the ion is observed during the optimization process, which indicates that positive ions are energetically preferable for substitution. In order to further verify the existence of localized conductive path, a small MoS2 and h-BN systems are studied with different ratios of gold ions replacing the sulfur and boron vacancies (Fig. 1.17A and C as an example). Based on the density of states calculation (Fig. 1.17B and D), the bandgap of the monolayer MoS2 and h-BN shrinks as the replacement percentage increases. At high percentage, the Fermi level is embedded within the induced states, indicating a switching to LRS with metallic character. The occupation of the metal ions into vacancies suggests a local conductivebridge-like mechanism in the MIM sandwich. It can be inferred that at HRS, electrons transport through the 2D atomic sheets with intrinsic vacancies on it, where the vacancies serve as trapping centers. During the SET process, the gold atoms located at the positive biased electrodes lose their electrons and become positively charged gold ions (Au-Au1 1 e2). Then, the ions are attracted by

17

18

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

Figure 1.16 The ab initio simulation results of (A) the initial state and (B) the final state of optimization process for Au1 ion and sulfur vacancies (Vs) in monolayer MoS2. Similarly, (C) the initial and (D) the final state of optimization process for Au1 ion and boron vacancies in monolayer h-BN. The result shows that the Au1 tends to move to the sulfur or boron vacancies, possibly resulting in the conductive bridge formation and SET process, while same simulations based on neutral Au atom and Au12 ions are unfavorable for occupation of vacancies. Source: (A and B) Reprinted with permission from R. Ge, X. Wu, M. Kim, P.-A. Chen, J. Shi, J. Choi, et al., Atomristors: memory effect in atomicallythin sheets and record RF switches, in: 2018 IEEE International Electron Devices Meeting (IEDM), 15 December 2018, IEEE, 2018, pp. 22.6.122.6.4. doi:10.1109/IEDM.2018.8614602. Copyright (2018) IEEE. (C and D) Reprinted with permission from X. Wu, R. Ge, P.-A. Chen, H. Chou, Z. Zhang, Y. Zhang, et al., Thinnest nonvolatile memory based on monolayer h-BN, Adv. Mater. 31 (2019). doi:10.1002/adma.201806790. Copyright (2019) John Wiley and Sons.

the vacancies and subsequently reduced (Au1 1 e2-Au), forming a conductive path in the vertical direction to establish the LRS. Temperature-dependent and area-dependence characteristics further corroborate the conductive-bridge-like behavior for resistance switching in monolayer atomic sheets.

1.5 1.5.1

MoS2 radio frequency switches Introduction to radio frequency switch

The rapid development of the IoT has introduced a consideration that cannot be overlooked: the need for the device to have RF switches, which can route signals from one band to another

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

19

Figure 1.17 An illustration for density of states calculation, where 25% (A) sulfur atoms or (C) boron atoms are replaced by gold atoms. Density of states for different Au replacement percentage of (B) sulfur and (D) boron (dashed line: Fermi level). The results indicate Au replacement can lead to switching to LRS. Source: (A and B) Reprinted with permission from R. Ge, X. Wu, M. Kim, P.-A. Chen, J. Shi, J. Choi, et al., Atomristors: memory effect in atomicallythin sheets and record RF switches, in: 2018 IEEE International Electron Devices Meeting (IEDM), 15 December 2018, IEEE, 2018, pp. 22.6.122.6.4. doi:10.1109/IEDM.2018.8614602. Copyright (2018) IEEE. (C and D) Reprinted with permission from X. Wu, R. Ge, P.-A. Chen, H. Chou, Z. Zhang, Y. Zhang, et al., Thinnest nonvolatile memory based on monolayer h-BN, Adv. Mater. 31 (2019). doi:10.1002/adma.201806790. Copyright (2019) John Wiley and Sons.

[3847]. Conventional switches are realized with solid-state diode or transistor devices [3840], which are volatile and consequently dissipate both dynamic and static energy. The former is due to a switching event and the latter, a consequence of the required bias voltage. To reduce leakage current, switches based on microelectromechanical systems (MEMS) have been investigated [4143], however, MEMS devices require rather large switching voltages (B10100 V) and are difficult to integrate onto arbitrary platforms owing to complex fabrication and packaging. For the purpose of improving energy efficiency, nonvolatile switches are attractive because they require no DC voltage for

20

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

operation and, as a benefit, consume zero-static power or energy. Toward this end, nonvolatile memory devices such as RRAM and PCM have been recently considered for RF switch applications. As discussed in Section 1.1, NVRS afford resistance modulation between HRS (ROFF) and LRS (RON) and subsequently retain the previous state without power consumption [5,48]. RRAM devices are typically realized with amorphous transitional metal oxides that have LRS values above 1 kΩ, making them unsuitable for RF switching owing to system requirements for RON to be much less than 50 Ω in order to avoid excessive insertion losses. Moreover, a large forming voltage is typically required to initiate the RRAM device [5]. On the other hand, RF switches based on PCM have shown promising results with low RON, high endurance, and decent fc figure of merit (FOM) [4446]. However, the high-temperature melting requirements and slow switching times have limited their utility. The RF switch based on 2D monolayer NVRS can overcome aforementioned limitations of RRAM and PCM devices with characteristics including forming-free, low ON-state resistance values (,10 Ω), fast switching (B15 ns), and room-temperature operation capability. This section focuses on RF switches based on MoS2 NVRS.

1.5.2

Fabrication and measurement of MoS2 radio frequency switch

Fig. 1.18. shows the MIM device structure of a monolayer MoS2 RF switch. Its fabrication starts with patterning and deposition of ground-signal-ground (GSG) pad structure using EBL and e-beam evaporation consisting of Cr (2 nm, as adhesion layer)/Au (6070 nm). All metal layers are formed by lift-off after patterning and the CVD-grown monolayer MoS2 is then transferred to the bottom signal electrode using dry PDMS transfer or wet PMMA etching transfer as carefully presented in Section 1.2.1. The top signal electrode was prepared using the same fabrication process as BE. The overlap between the TE and BE (as shown in Fig. 1.18B) defines the switch area ranged from 0.15 μm 3 0.2 μm to 1 μm 3 1 μm. Fig. 1.18A depicts the devices structure allowing (ON) or rejecting (OFF) the transmission of RF signal depending on the switch DC status. Small-signal performance of the RF switch was characterized by calibrated network analyzer measurements aided by an equivalent circuit model. The switch FOM cutoff frequency (fc 5 1/2πRONCOFF) was determined using the

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

21

Figure 1.18 (A) Simplified illustration of the signal transmission and device structure of the RF switches based on monolayer MoS2. (B) Zoomed-in top-view SEM image of a MoS2 RF switch with lateral area of 1 3 1 μm2. The dashed box in (B) marks the area covered with MoS2. Source: Reprinted with permission from M. Kim, R. Ge, X. Wu, X. Lan, J. Tice, J.C. Lee, et al., Zero-static power radio-frequency switches based on MoS2 atomristors, Nat. Commun. 9 (2018), 2524. Copyright (2018) Springer Nature.

de-embedded ON-state resistance (RON) and OFF-state capacitance (COFF) from the circuit model. For RF measurements, GSG device configuration was employed to facilitate S-parameter characterization that is essential for analyzing the insertion loss and isolation at GHz frequencies.

1.5.3

MoS2 radio frequency switch performance

Current reports focused on DC electrical studies, which were conducted on the nonvolatile memory devices consisting of single-layer MoS2 sandwiched between Au BE and TE. Fig. 1.19 shows the intrinsic experimental RF characteristics of monolayer MoS2 switch with promising results of B0.25 dB insertion loss in the ON state and isolation ,29 dB in the OFF state up to 67 GHz. The extracted cutoff frequencies for monolayer RF switches is 70 THz for 0.03-μm2 device. The FOM is used to evaluating RF switches [46,49]. Notably, the monolayer MoS2 RF switch achieved a record cutoff frequency value compared to emerging solid-state, MEMS, and phase-change (PC) material-based switches, with the added benefit of smaller feature size and frequency scalability without compromising insertion loss. Besides record cutoff frequency, the unique combination of (approximately) area-independent LRS resistance and areadependent HRS capacitance yields higher FOM by reduction of

22

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

Figure 1.19 Measured and extracted S-parameter data in both the ON state (insertion loss) and OFF state (isolation) of an RF switch based on 0.15 3 0.2-μm2 monolayer MoS2 atomristor. The extracted RON, COFF, and fc values are 2.69 Ω, 0.84 fF, and 70 THz respectively. Source: Reprinted with permission from R. Ge, X. Wu, M. Kim, P.-A. Chen, J. Shi, J. Choi, et al., Atomristors: memory effect in atomically-thin sheets and record RF switches, in: 2018 IEEE International Electron Devices Meeting (IEDM), 15 December 2018, IEEE, 2018, pp. 22.6.122.6.4. doi:10.1109/IEDM.2018.8614602. Copyright (2018) IEEE.

Figure 1.20 Device performance of the RF switches based on MoS2 atomristor. The equivalent lumped element circuit model extracts (A) ON-state resistance, (B) OFF-state capacitance, and (C) cutoff frequency dependence on area featuring a normalized figure of merit, fc  AB1 THz μm2. While ON-state resistance is area-independent, OFFstate capacitance is dependent on the lateral area of the device and has a normalized capacitance of B23.5 fF/ μm2. The lines in the figures are area-scaling guides. Source: Reprinted with permission from M. Kim, R. Ge, X. Wu, X. Lan, J. Tice, J.C. Lee, et al., Zero-static power radio-frequency switches based on MoS2 atomristors, Nat. Commun. 9 (2018) 2524. Copyright (2018) Springer Nature.

device area, a defining advantage over PC switches [46,49]. Almost two dozen RF switches based on monolayer MoS2 were realized and their area dependence studied for nonvolatile RF switches with the results summarized in Fig. 1.20. As expected, the ON-state resistance shows negligible area dependence affirming a 1D filamentary (or bridge) conduction mechanism (Fig. 1.20A). The lowest achieved RON is about 4 Ω, a favorable

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

23

value for reduced insertion loss switches. In contrast, COFF is area-dependent (Fig. 1.20B) owing to the parallel-plate capacitance proportionality relation (CBA). A linear fit to the capacitance-area data yields COFFB23.5 fF/μm2. Considering the effective thickness between the electrodes is about a nanometer based on the monolayer thickness and van der Waals gap at the interfaces, the extracted capacitance corresponds to an effective dielectric constant of B2.6, which is consistent with the significantly reduced value for a monolayer compared to the bulk MoS2 [50]. The resulting area-dependent statistics of the cutoff frequency FOM is displayed in Fig. 1.20C. The area-frequency expression can be derived from the FOM formula with only COFF showing area dependence; therefore fc~1/ COFF~1/A. As such, fc  A is a constant, a unique and beneficial property for atomristor MIM switches. From the inverse linear fit to the experimental fc data (Fig. 1.20C), fc  AB1 THz μm2. The extracted cutoff frequency is over 70 THz for 0.03 μm2 device as we discussed, which is in consistent with the trend. If we continue to decrease the device area, for example, 0.01 or 0.001 μm2 switches will afford a cutoff frequency of 100 THz or even 1000 THz. Other monolayer semiconductor or insulating TMDs provide an additional degree of design, particularly if they can offer lower MIM capacitance that can enable even higher cutoff frequencies.

Table 1.1 Comparison of representative 2D NVRS publications. Reference

Active layer materials

Active layer thickness (nm)

Device structure

Switching voltage (V)

Switching time

Results presented in this chapter Sangwan et al. (2018) [14] Wang et al. (2018) [51] Pan et al. (2017) [12] Zhao et al. (2017) [26] Cheng et al. (2016) [52] Hao et al. (2016) [10]

MoS2

B0.65

Vertical

Down to 0.6

15 ns

MoS2

B0.65

Planar

.20

1 ms

MoS22xOx h-BN BNOx 1T MoS2 Degraded BP

B40 B1.82.5 0.92.3 550 B10

Vertical Vertical Vertical Vertical Vertical

B1 Down to 0.4 0.61.7 B0.1 12

100 ns Not reported 1 ms Not reported Not reported

BNOx, Boron nitride oxide; BP, bottom electrode. Source: Reprinted with permission from R. Ge, X. Wu, M. Kim, P.-A. Chen, J. Shi, J. Choi, et al., Atomristors: memory effect in atomically-thin sheets and record RF switches, in: 2018 IEEE International Electron Devices Meeting (IEDM), 15 December 2018, IEEE, 2018, pp. 22.6.122.6.4. doi:10.1109/IEDM.2018.8614602. Copyright (2018) IEEE.

Table 1.2 Comparison of this work with other representative RF switch papers. Reference

Device technology

Nonvolatility Control voltage (V)

Cutoff frequency (THz)

Operating Switching Dimension (single device environment time W 3 L) (µm2)

This work

MoS2 switch

Yes

B0.51.5

70

15 ns

0.15 3 0.2

Yes

3.5

17

Ambient condition Heater needed

Leon et al. (2017) [53] Madan et al. (2015) [45] Pi et al. (2015) [48] Stefanini et al. (2011) [42]

GeTe phasechange switch VO2 phasechange switch Memristive switch MEMS switch

,0.5 μs

0.8 3 0.5

No

B1.5

26.5

25 ns

1 3 0.1

Yes

3

35.2

No

B65

3.8

Ambient condition Ambient condition Hermetic packaging

Not reported 0.11 3 0.035 2.2 μs

300 3 24

Source: Reprinted with permission from R. Ge, X. Wu, M. Kim, P.-A. Chen, J. Shi, J. Choi, et al., Atomristors: memory effect in atomically-thin sheets and record RF switches, in: 2018 IEEE International Electron Devices Meeting (IEDM), 15 December 2018, IEEE, 15 December 2018, IEEE, 2018, pp. 22.6.122.6.4. doi:10.1109/IEDM.2018.8614602. Copyright (2018) IEEE.

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

1.6

Summary

In summary, NVRS has been demonstrated based on monolayer TMDs (MoS2, MoSe2, WS2, and WSe2) and h-BN in vertical MIM device structure, with low switching voltage, high ON/OFF ratio and forming-free characteristic. This phenomenon alludes to a universal NVRS effect in nonmetallic 2D atomic sheets. The reliability, flexibility, and ultrafast pulse switching have been demonstrated and further support future applications in flexible, ultra-scaled, high-performance and energy-saving memory fabrics and RF switches for IoT systems. The following tables show representative works in 2D nonvolatile memory (Table 1.1) and RF switches (Table 1.2), exploring larger design space for the choice of active layer material, thickness, device structure, operating principles, etc.

Acknowledgment The authors appreciate Po-An Chen and Meng-Hsueh Chiang of National Cheng Kung University for ab initio simulation. We are grateful to Jesse Tice and Xing Lan of NGAS for collaborative discussion regarding RF switch and design. We acknowledge the supply of CVD samples from Jianping Shi, Zhepeng Zhang, and Yanfeng Zhang of Peking University. We thank the group of Nanshu Lu of the University of Texas at Austin for providing mechanical bending apparatus and Jo Wozniak of Texas Advanced Computing Centre (TACC) for 3D renderings. D.A. acknowledges the Presidential Early Career Award for Scientists and Engineers (PECASE) through Army Research Office (W911NF-16-1-0277) and National Science Foundation (NSF) grant #1809017. The authors acknowledge use of Texas Nanofabrication Facilities supported by the NSF NNCI award #1542159.

References [1] M.T. Ghoneim, M.M. Hussain, Review on physically flexible nonvolatile memory for Internet of everything electronics, Electronics 4 (2015) 424479. [2] C.S. Hwang, Prospective of semiconductor memory devices: from memory system to materials, Adv. Electron. Mater. 6 (1) (2015) 140056. [3] G.W. Burr, B.N. Kurdi, J.C. Scott, C.H. Lam, K. Gopalakrishnan, R.S. Shenoy, Overview of candidate device technologies for storage-class memory, IBM J. Res. Dev. 52 (4.5) (2008) 449464. [4] A. Chen, A review of emerging non-volatile memory (NVM) technologies and applications, Solid-State Electron. 125 (2018) 2538. [5] H.-S.P. Wong, H.-Y. Lee, S. Yu, Y.-S. Chen, Y. Wu, P.-S. Chen, et al., Metaloxide RRAM, Proc. IEEE 100 (6) (2012) 19511970. [6] G.R. Bhimanapati, et al., Recent advances in two-dimensional materials beyond graphene, ACS Nano 9 (12) (2015) 1150911539. [7] C. Tan, et al., Non-volatile resistive memory devices based on solutionprocessed ultrathin two-dimensional nanomaterials, Chem. Soc. Rev. 44 (2015) 26152628.

25

26

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

[8] A.A. Bessonov, M.N. Kirikova, D.I. Petukhov, M. Allen, T. Ryha¨nen, M.J.A. Bailey, Layered memristive and memcapacitive switches for printable electronics, Nat. Mater. 14 (2) (2015) 199204. [9] D. Son, S.I. Chae, M. Kim, M.K. Choi, J. Yang, K. Park, et al., Colloidal synthesis of uniform-sized molybdenum disulfide nanosheets for Waferscale flexible nonvolatile memory, Adv. Mater. 28 (42) (2016) 93269332. [10] C. Hao, F. Wen, J. Xiang, S. Yuan, B. Yang, L. Li, et al., Liquid-exfoliated black phosphorous nanosheet thin films for flexible resistive random access memory applications, Adv. Funct. Mater. 26 (12) (2016) 20162024. [11] K. Qian, R.Y. Tay, V.C. Nguyen, J. Wang, G. Cai, T. Chen, et al., Hexagonal boron nitride thin film for flexible resistive memory applications, Adv. Funct. Mater. 26 (13) (2016) 21762184. [12] C. Pan, Y. Ji, N. Xiao, F. Hui, K. Tang, Y. Guo, et al., Coexistence of grainboundaries-assisted bipolar and threshold resistive switching in multilayer hexagonal boron nitride, Adv. Funct. Mater 27 (10) (2017) 1604811. [13] C. Tan, H. Zhang, Two-dimensional transition metal dichalcogenide nanosheet-based composites, Chem. Soc. Rev. 44 (9) (2015) 27132731. [14] V.K. Sangwan, D. Jariwala, I.S. Kim, K.-S. Chen, T.J. Marks, L.J. Lauhon, et al., Gate-tunable memristive phenomena mediated by grain boundaries in single-layer MoS2, Nat. Nanotechnol. 10 (5) (2015) 403406. [15] R. Ge, X. Wu, M. Kim, J. Shi, S. Sonde, L. Tao, et al., Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides, Nano Lett. 18 (1) (2018) 434441. [16] X. Wu, R. Ge, P.-A. Chen, H. Chou, Z. Zhang, Y. Zhang, et al., Thinnest nonvolatile memory based on monolayer h-BN, Adv. Mater. 31 (2019). Available from: https://doi.org/10.1002/adma.201806790. [17] R. Ge, X. Wu, M. Kim, P.-A. Chen, J. Shi, J. Choi, et al., Atomristors: memory effect in atomically-thin sheets and record RF switches, in: 2018 IEEE International Electron Devices Meeting (IEDM), 15 December 2018, IEEE, 2018, pp. 22.6.122.6.4. doi:10.1109/IEDM.2018.8614602. [18] M. Kim, R. Ge, X. Wu, X. Lan, J. Tice, J.C. Lee, et al., Zero-static power radio-frequency switches based on MoS2 atomristors, Nat. Commun. 9 (2018) 2524. [19] H.-Y. Chang, M.N. Yogeesh, R. Ghosh, A. Rai, A. Sanne, S. Yang, et al., Large-area monolayer MoS2 for flexible low-power RF nanoelectronics in the GHz regime, Adv. Mater. 28 (9) (2016) 18181823. [20] K. Kang, S. Xie, L. Huang, Y. Han, P.Y. Huang, K.F. Mak, et al., High-mobility three-atom-thick semiconducting films with wafer-scale homogeneity, Nature 520 (7549) (2015) 656660. [21] J. Shi, D. Ma, G.-F. Han, Y. Zhang, Q. Ji, T. Gao, et al., Controllable growth and transfer of monolayer MoS2 on Au foils and its potential application in hydrogen evolution reaction, ACS Nano 8 (10) (2014) 1019610204. [22] A. Ismach, H. Chou, D.A. Ferrer, Y.P. Wu, S. McDonnell, H.C. Floresca, et al., Toward the controlled synthesis of hexagonal boron nitride films, ACS Nano 6 (2012) 6378. [23] A. Ismach, H. Chou, P. Mende, A. Dolocan, R. Addou, S. Aloni, et al., Carbon-assisted chemical vapor deposition of hexagonal boron nitride, 2D Mater. 4 (2017) 025117. [24] Z.P. Zhang, X.J. Ji, J.P. Shi, X.B. Zhou, S. Zhang, Y. Hou, et al., Direct chemical vapor deposition growth and band-gap characterization of MoS2/ h-BN van der Waals heterostructures on Au foils, ACS Nano 11 (2017) 4328. [25] D.J. Wouters, R. Waser, M. Wuttig, Phase-change and redox-based resistive switching memories, Proc. IEEE 103 (8) (2015) 12741288.

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

[26] L. Zhao, J. Zizhen, H.Y. Chen, J. Sohn, K. Okabe, B. Magyari-Ko¨pe, et al., Ultrathin (2nm) HfOx as the fundamental resistive switching element: thickness scaling limit, stack engineering and 3D integration, in: 2014 IEEE International Electron Devices Meeting, 1517 December, 2014, IEEE, 2014, pp. 6.6.16.6.4. [27] F.M. Puglisi, L. Larcher, C. Pan, N. Xiao, Y. Shi, F. Hui, et al., 2D h-BN based RRAM devices, in: 2016 IEEE International Electron Devices Meeting (IEDM), 37 December 2016, IEEE, 2016, pp. 34.8.134.8.4. [28] C.A. Santini, A. Sebastian, C. Marchiori, V.P. Jonnalagadda, L. Dellmann, W. W. Koelmans, et al., Oxygenated amorphous carbon for resistive memory applications, Nat. Commun. 6 (2015) 8600. [29] G. Indiveri, S.C. Liu, Memory and information processing in neuromorphic systems, Proc. IEEE 103 (8) (2015) 13791397. [30] F.-C. Chiu, A review on conduction mechanisms in dielectric films, Adv. Mater. Sci. Eng. 2014 (2014) 18. [31] J. Frenkel, On pre-breakdown phenomena in insulators and electronic semi-conductors, Phys. Rev. 54 (1938) 647. [32] P. Fiorenza, G. Greco, F. Giannazzo, R. Lo Nigro, F. Roccaforte, Poole-Frenkel emission in epitaxial nickel oxide on AlGaN/GaN heterostructures, Appl. Phys. Lett. 101 (2012) 172901. [33] N. Onofrio, D. Guzman, A. Strachan, Atomic origin of ultrafast resistance switching in nanoscale electrometallization cells, Nat. Mater. 14 (4) (2015) 440446. [34] J. Kwon, J.-Y. Lee, Y.-J. Yu, C.-H. Lee, X. Cui, J. Hone, et al., Thicknessdependent Schottky barrier height of MoS2 field-effect transistors, Nanoscale 9 (2017) 61516157. [35] H. Zhong, R. Quhe, Y. Wang, Z. Ni, M. Ye, Z. Song, et al., Interfacial properties of monolayer and bilayer MoS2 contacts with metals: beyond the energy band calculations, Sci. Rep. 6 (2016) 21786. [36] A. Zobelli, C.P. Ewels, A. Gloter, G. Seifert, Vacancy migration in hexagonal boron nitride, Phys. Rev. B 75 (2007) 094104. [37] B. Huang, H. Lee, Defect and impurity properties of hexagonal boron nitride: a first-principles calculation, Phys. Rev. B 86 (2012) 245406. [38] L.E. Larson, Integrated circuit technology options for RFICs-present status and future directions, IEEE J. Solid-State Circuits 33 (1998) 387399. [39] J.-L. Lee, D. Zych, E. Reese, D.M. Drury, Monolithic 2-18 GHz low loss, on-chip biased PIN diode switches, IEEE Trans. Microw. Theory Tech. 43 (1995) 250256. [40] Q. Li, Y.P. Zhang, K.S. Yeo, W.M. Lim, 16.6- and 28-GHz fully integrated CMOS RF switches with improved body floating, IEEE Trans. Microw. Theory Tech. 56 (2008) 339345. [41] R. Stefanini, M. Chatras, P. Blondy, G.M. Rebeiz, Miniature MEMS switches for RF applications, J. Microelectromech. Syst. 20 (2011) 13241335. [42] E.R. Brown, RF-MEMS switches for reconfigurable integrated circuits, IEEE Trans. Microw. Theory Tech. 46 (1998) 18681880. [43] J.J. Yao, RF MEMS from a device perspective, J. Micromech. Microeng. 10 (R9) (2000). [44] H. Madan, et al., 26.5 terahertz electrically triggered RF switch on epitaxial VO2-on-Sapphire (VOS) wafer, in: 2015 IEEE International Electron Devices Meeting (IEDM), 2015, pp. 9.3.19.3.4. [45] N. El-Hinnawy, et al., Low-loss latching microwave switch using thermally pulsed non-volatile chalcogenide phase change materials, Appl. Phys. Lett. 105 (2014) 013501.

27

28

Chapter 1 Two-dimensional materials-based nonvolatile resistive memories and radio frequency switches

[46] M. Wang, M. Rais-Zadeh, Development and evaluation of germanium telluride phase change material based ohmic switches for RF applications, J. Micromech. Microeng. 27 (2017) 013001. [47] S. Pi, M. Ghadiri-Sadrabadi, J.C. Bardin, Q. Xia, Nanoscale memristive radiofrequency switches, Nat. Commun. 6 (2015) 7519. [48] D.J. Wouters, R. Waser, M. Wuttig, Phase-change and redox-based resistive switching memories, Proc. IEEE 103 (2015) 12741288. [49] J.S. Moon, S. Hwa-Chang, D. Le, F. Helen, A. Schmitz, T. Oh, et al., 11 THz figure-of-merit phase-change RF switches for reconfigurable wireless frontends, in: 2015 IEEE MTT-S International Microwave Symposium, 1722 May 2015, IEEE, 2015, pp. 14. [50] S.L. Li, K. Tsukagoshi, E. Orgiu, P. Samori, Charge transport and mobility engineering in two-dimensional transition metal chalcogenide semiconductors, Chem. Soc. Rev. 45 (2016) 118151. [51] Miao Wang, et al., Robust Memristors based on layered two-dimensional materials, Nat. Electron. 1 (2) (2018) 130136. [52] P. Cheng, K. Sun, Y.H. Hu, Memristive behavior and ideal memristor of 1T Phase MoS2 nanosheets, Nano Letter. 16 (1) (2016) 572576. [53] A. Le´on et al., In-depth characterisation of the structural phase change of Germanium Telluride for RF switches, 2017 IEEE MTT-S International Microwave Workshop Series on Advanced Materials and Processes for RF and THz Applications (IMWS-AMP), Pavia, 2017, pp. 13.

2 Two-dimensional materialsbased radio frequency wireless communication and sensing systems for Internet-of-things applications Liang Zhu1, Mohamed Farhat2, Khaled Nabil Salama2 and Pai-Yen Chen1 1

Department of Electrical and Computer Engineering, University of Illinois at Chicago, Chicago, IL, United States 2Computer, Electrical, and Mathematical Science and Engineering (CEMSE) Division, King Abdullah University of Science and Technology (KAUST), Thuwal, Saudi Arabia

Chapter Outline 2.1 Introduction 29 2.2 Radio frequency performance of two-dimensional transistors 32 2.3 Frequency mixers and signal modulators based on two-dimensional transistors 45 2.4 Integrated wireless Internet-of-things sensors 48 2.5 Radio frequency energy harvesting using two-dimensional electronic devices 51 2.6 Conclusion 52 References 53

2.1

Introduction

Nowadays, the Internet of things (IoT) has had an increasing impact on health care, automobiles, agriculture, retail, security, specific environmental controls, and so forth [1,2]. Thanks to rapid advances in nanotechnology and microsystem engineering, as well as ever-expanding wireless technology, nanomaterial-based sensors are envisioned to drive the market of wireless IoT sensors that may be compact, low-cost, wearable, textile-based, and lightweight. When compared to bulky Emerging 2D Materials and Devices for the Internet of Things. DOI: https://doi.org/10.1016/B978-0-12-818386-1.00002-3 © 2020 Elsevier Inc. All rights reserved.

29

30

Chapter 2 Two-dimensional materials-based radio frequency wireless communication

semiconductors, low-dimensional nanomaterials may be significantly more sensitive to the exposed environments and chemical substances. Among various types of nanomaterials, twodimensional (2D) materials including graphene and beyond were the most studied ones in the past decade [35]. In general, they exist in a bulk form as stacks of strongly bonded monolayers with weak interlayer attraction and can be successfully isolated as single atomic layers by means of exfoliation or chemical vapor deposition (CVD) methods. Graphene formed by a single layer of graphite was the first 2D material ever discovered. In 2004, Geim’s group, which has been widely credited for the explosive growth of interest in 2D materials, reported the first graphene field-effect transistor (GFET), opening up a new field of 2D electronics [6]. GFETs with ultrahigh carrier mobilities and cutoff frequency (up to 840 GHz [7]) are particularly attractive for radio frequency (RF) applications, such as high-frequency power amplifiers and oscillators in modern communication systems. It was even more exciting that graphene transistors or electrodes can be flexible, highly durable, and optically transparent. Such properties make graphene become the preferred candidate material for wearable and textile electronics. Moreover, it has been demonstrated that the electrical properties of GFETs can be sensitively tuned by chemical gating effects, which enables selective detection of gas, chemical, or biomolecular agents, with sensitivity at the singlemolecule level [8,9]. Noticeably, pristine graphene does not have a bandgap—a property that is essential to maintain a good on/off current ratio for logic circuits. Due to the ambipolar transport property of the gapless graphene, GFETs typically exhibit a “V-shape” drain current-gate voltage characteristic that enables effective conversion of an applied RF signal into its second harmonic [10,11]. This exotic characteristic was not found in traditional silicon or IIIV devices and can be engineered to make the simplest possible frequency modulator. At terahertz (THz) and infrared frequencies, low-loss Dirac plasmons excited in graphene, together with tunability of Fermi energy through gating effects, make graphene plasmonics an excellent platform for building reconfigurable infrared optoelectronic components. In this regard, several new applications have been devised, such as reconfigurable optoelectronic devices [12], metasurfaces and metamaterials [13,14], phase shifters [15], and oscillators (with proper optical pumping) [16]. A patterned graphene, for example, graphene nanoribbon (GNR), can have a bandgap that depends on the width of the

Chapter 2 Two-dimensional materials-based radio frequency wireless communication

GNR (w) as: Eg ðeVÞ 5 0:8=wðnmÞ [17]. Although a GNR can be used to improve on/off ratios of GFETs, it also causes drops in the intrinsic and extrinsic mobilities of graphene. In this context, many new types of 2D materials with sizable intrinsic bandgap (at the level of 12 eV) have been investigated [18,19], aiming to fabricate high-performance switching devices that meet requirements of chip-scale logic circuits. Transition metal dichalcogenides (TMDs) are a class of 2D materials with the chemical formula MX2, where M is a transition metal element from group IV (e.g., Ti, Zr, or Hf), group V (e.g., Nb or Ta), or group VI (e.g., Mo, W, and so on), and X is a chalcogen (e.g., S, Se, or Te). These materials are made up of layered composite of the form XMX, with the chalcogen atoms in two hexagonal planes separated by a plane of metal atoms. To date, 2D transistors with MoS2 and WeS2 channels have been demonstrated to have a greater than 108 on/off ratio [5]. Such a value makes them suitable for the development of low-power digital circuits for IoT applications. So far, many groups have successfully fabricated wafer-scale TMD-based transistors for analog and digital integrated circuits [20,21]. Wireless communication and sensing systems present a continuing evolution in terms of the number of users and the volume of exchange information, as well as the basis for advancing the IoT systems. However, it is always met with price constraints, an incessant miniaturization, and new specificities that make the design conditions increasingly difficult. For instance, textile and flexible technologies are not possible with traditional solid-state devices that are bulky and rigid, but may be realized with 2D electronics. This chapter will briefly review RF and microwave transistors based on different 2D materials, and their practical uses in RF circuit building blocks illustrated in Fig. 2.1. Fig. 2.1 shows, as an example, the RF architectures of wireless IoT sensors and integrated system, of which 2D electronics can be applied to several highlighted key building blocks. These basic blocks include passive RF components (e.g., filter and miniature antenna) and active RF components (e.g., RF mixer with a local oscillator (LO), signal modulator/demodulator, variable-gain amplifier, low-noise amplifier), memories, digital chip, and analog-to-digital converter. Replacing current RF front-end components with 2D-based devices and circuits offers several advantages in terms of cost, performance, power consumption, device size, and flexibility. Researchers have envisioned that 2D electronics will create a new paradigm in the design of next-generation wearable electronics, bioimplants, soft robots, and ubiquitous sensors for IoT applications. In the

31

32

Chapter 2 Two-dimensional materials-based radio frequency wireless communication

Figure 2.1 Schematic circuit diagram for a wireless sensor, of which high-performance 2D material-based devices can be exploited in several active components (highlighted with gray shadows).

following, the state-of-the-art GFETs and different 2D transistors working in the RF and microwave frequency ranges will be reviewed, followed by introduction of modeling and parameter extraction methods for these 2D transistors. Afterward, an overview of active RF components based on 2D transistors will be presented, and their benefits and trade-offs will be discussed. Finally, potential applications of 2D electronics in ubiquitous wireless sensors and IoT systems will be mentioned.

2.2

Radio frequency performance of two-dimensional transistors

Low-dimensional nanostructures and nanomaterials, such as semiconductor nanowires and carbon nanotubes (CNTs) have been extensively studied for developing field-effect transistors (FETs) with a ballistic channel, which potentially pushes the cutoff frequency to the terahertz region [22,23]. In practice, however, their operation frequencies were limited to tens of GHz due to the high contact resistance of nanotubes/nanowires with an infinitesimal diameter [24,25]. To address the

Chapter 2 Two-dimensional materials-based radio frequency wireless communication

fabrication challenge and drastically reduce the contact resistance, 2D materials have been exploited to scale up the drain current of nanotransistors by significantly increasing the device’s channel width. In 2004, the first 2D material, namely graphene, was isolated from graphite and transferred into FETs [6]. Since graphene has a linear energy-momentum (E-k) dispersion in the low-energy region, electrons and holes in graphene are considered as massless relativistic particles with an extremely high mobility (Dirac fermions) and an energyindependent velocity that arise in ballistic transports (if scattering with phonons or impurities is excluded). Further research has demonstrated that GFETs can possess an ultrahigh electronic mobility that exceeds 200,000 cm2/Vs [26], a high saturation velocity (5.5 3 107 cm/s [27]), and a great carrier density modulation by electric field (2 3 108 A/cm [28]), which make graphene appealing for high-speed electronic circuits on flexible substrates. In this section, we will review essential features of highfrequency transistors based on graphene and types of 2D materials, as well as figures of merits (FOMs) used for accessing their performance. Fig. 2.2A and B, respectively, show the schematics and the small-signal equivalent circuit model for a 2D materialbased FET, of which important elements include the gatesource capacitance Cgs, the gate-drain capacitance Cgd, the transconductance gm, the differential drain resistance rds (the inverse of the drain conductance gds), and the contact resistances and parasitics due to packages. The transport and smallsignal behavior of a transistor are generally described by the drift-diffusion model. As a representative example, we first discuss the modeling of transport properties of GFETs in Fig. 2.2A. Considering the symmetric low-energy electronic structure in graphene, electron and hole sheet densities (n and p) are derived from the integration of density of states DðEÞ weighted by the FermiDirac distribution f ðEÞ, and the total mobile charge density Qs in graphene is obtained as [30]: Ð  ÐN 0 Qs 5 q 3 ðp 2 nÞ 5 q 3 2N DðEÞ½1 2 f ðEÞdE 2 0 DðEÞf ðEÞdE 2 3 ðN 2 1 1 5dE; 5q E4 ðE1E Þ=K T 2 ðE2E Þ=K T F B 11 F B 11 e e πð¯h v F Þ2 0 ð2:1Þ where ¯h is the reduced Planck constant, EF is the Fermi level, v F is the Fermi velocity (vF 5 108 cm=s), and KB is the Boltzmann constant. Based on the quantum capacitance

33

34

Chapter 2 Two-dimensional materials-based radio frequency wireless communication

Figure 2.2 (A) Schematic of a 2D material transistor. (B) Small-signal equivalent circuit of a FET. (C) RF model of conventional FETs under test, where Yp1, Yp2, and Yp3 are parallel parasitics and ZL1, ZL2, and ZL3 are series parasitics. (They represent FETs contact pads and interconnects.) (D) Small-signal current gain h21 and unilateral power gain U of an RF FETs versus frequency. (E) Measured scattering parameters, S11, S12, S21, and S22, of a graphene transistor for deembedding. (F) The current gain h21 calculated from the measured S parameters as a function of frequency. The solid line corresponds to the ideal 1/f dependence, or equivalently, 220 dB/decade slope, of the current gain. The cutoff frequency is determined to be 4 GHz. The drain and gate voltages are 1.6 and 0.5 V, respectively. Source: (B and D) Reprinted with permission from F. Schwierz, Graphene transistors: status, prospects, and problems, Proc. IEEE 101 (7) (2013) 15671584. Copyright (2013) IEEE. (E and F) Reprinted with permission from Y.M. Lin, K.A. Jenkins, A.V. Garcia, J.P. Small, D.B. Farmer, P. Avouris, Operation of graphene transistors at gigahertz frequencies, Nano Lett. 9 (2009) 422426. Copyright (2009) American Chemical Society.

definition: Cq 5 2 @ðQs Þ=@V ch ; we may obtain an explicit expression for quantum capacitance as: [31]    2q2 KB T qV ch Cq 5 ln 2 1 1 cosh ; ð2:2Þ KB T πðh ¯ υF Þ2 where V ch 5 EF =q is the surface potential. Under the condition qV ch {KB T ; which is generally valid, the expression for quantum capacitance  can be simplified to Cq 5 KjV ch j, where K 5 2q3 = π ðh ¯ υF Þ2 : As a Ðresult, the net mobile charge density can be derived as Qs 5 Cq dV ch 5 2 1=2Cq V ch . In a back-gate GFET, since any variation in the Fermi level by an applied gate voltage is equivalent to the voltage drop across the quantum capacitance Cq , by applying Kirchhoff’s law, the surface potential is obtained as [3234] V ch ðxÞ 5

EF ðxÞ Cox 5 V GS 2 ΔV cnp 2 VðxÞ ; q Cox 1 Cq =2

ð2:3Þ

Chapter 2 Two-dimensional materials-based radio frequency wireless communication

where Cox 5 εox =tox is the oxide capacitance, and ΔV cnp is the back-gate voltage at the Dirac point where the carrier sheet density becomes minimal for zero drain and source voltages. ΔV cnp comprises work-function differences between the gates of the graphene channel, interfacial charges at the grapheneoxide interface, and the charge neutrality point shift ΔV cnp due to intentional or unintentional doping of graphene (e.g., chemical doping). The channel voltage VðxÞ, as a function of position in the channel, can be using the gradual channel

modeled approximation: VðxÞ 5 x=L V DS ; which is zero at the source end (x 5 0) and equal to the drain-to-source voltage V DS at the drain end (x 5 L). Since Cq is also a function of V ch ; Eqs. (2.2) and (2.3) must be solved self-consistently to evaluate the actual V ch ; which may have an explicit expression as: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  2 Cox 1 C2ox 6 2K V GS 2 ΔV cnp 2 V Cox : ð2:4Þ V ch 5 6K For a relatively thick gate oxide, that is, Cox {Cq , typically valid for back-gate transistors, V ch ðxÞ can be simply expressed qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  ffi

as V ch ðxÞ 5 EF ðxÞ=q 5 ¯h v F πCox V GS 2 x=L V DS =q3 : On the other hand, for an ultrathin high-K oxide,

V ch ðxÞ reduces to a simple form V ch ðxÞ 5 EF ðxÞ=q  V GS 2 x=L V DS : The driftdiffusion transport model suggests the drain-to-source current

IDS 5 qW Qs ðxÞ v drift ðxÞ; where W is the gate width and ν drift is the drift velocity. Using a soft saturation model, consistent with Monte Carlo simulations, ν drift ðxÞ that depends on the channel electric field F can be expressed as: [3234] vdrift 5 



μF

11 μF=vsat

γ 1=γ ;

ð2:5Þ

where μ is the carrier low-field mobility and v sat is the saturation velocity. Here, we assume that vsat is a material constant that is equal to vF =2 (which is close to GFET’s maximum predicted values), and γ 5 1 for GFETs. Applying F 5 2 dV=dx and integrating the resulting equation over the device length (L), the drain current becomes: [3234] ð W V DS



IDS 5 μ Qs dV; ð2:6Þ Leff 0 where the effective channel length accounting for the effect of saturation velocity. In order to have an explicit expression for the drain current, it is more convenient to solve the integral in

35

36

Chapter 2 Two-dimensional materials-based radio frequency wireless communication

(2.6) using Vch as the integration variable and consistently express Qs as a function of Vch ð

dV W V ch;S

IDS 5 μ Qs ðV ch Þ dV ch ; ð2:7Þ Leff V ch;D dV ch where the channel potential at the drain and source V ch;D and V ch;S are determined as Vch ðV 5VDS Þ and V ch ðV 5 0Þ, respectively. Finally, the explicit expression for drain current IDS ðV DS ; V GS Þ, as a function of bias voltages V DS and V GS , can be obtained as 2 3 V ch 5V ch;D μK W 4 V 3ch KV 4ch 5

2sgnðV ch Þ IDS 5 2

2 Leff 3 4Cox V ch 5V ch;S 82 3 3

KV 4ch;D μK W