Reversible Deactivation Radical Polymerization: Mechanisms and Synthetic Methodologies 9780841233171, 0841233179, 9780841233188, 9780841233232

Table of contents :
Content: Machine generated contents note: 1. Reversible Deactivation Radical Polymerization: State-of-the-Art in 2017 / Krzysztof Matyjaszewski --
2. Mechanistic Insights into Lewis Acid Mediated Sequence- and Stereo-Control in Radical Copolymerization / Michelle L. Coote --
3. Time-Resolved Electron Spin Resonance Observations of the Initial Stages of Conventional and Controlled Radical Polymerization Processes / Atsushi Kajiwara --
4. Elements of RAFT Navigation / Graeme Moad --
5. Reducing the Hydrogen Atom Abstraction Efficiencies of Benzophenone-Based Photosensitive Alkoxyamines / Didier Gigmes --
6. Catalyzed Radical Termination (CRT) in the Metal-Mediated Polymerization of Acrylates: Experimental and Computational Studies / Rinaldo Poli --
7. Electrochemical Procedures To Determine Thermodynamic and Kinetic Parameters of Atom Transfer Radical Polymerization / Armando Gennaro --
8. Insights into the Reactivity of Epoxides as Reducing Agents in Low-Catalyst-Concentration ATRP Reactions / Nicolay V. Tsarevsky --
9. Toward Butadiene-ATRP with Group 10 (Ni, Pd, Pt) Metal Complexes / Alexandra D. Asandei --
10. Reversible Deactivation Radical Polymerization of Vinyl Chloride / Jorge F. J. Coelho --
11. Photoinduced Metal Free Strategies for Atom Transfer Radical Polymerization / Y. Yagci --
12. Recent Developments in External Regulation of Reversible Addition Fragmentation Chain Transfer (RAFT) Polymerization / Krzysztof Matyjaszewski --
13. How Can Xanthates Control the RAFT Polymerization of Methacrylates? / Samir Z. Zard --
14. Hydroxyl Radical Activated RAFT Polymerization / Greg G. Qiao --
15. Vinyl Ether/Vinyl Ester Copolymerization by Cationic and Radical Interconvertible Simultaneous Polymerization / Masami Kamigaito --
16. Catalytic Chain Transfer Polymerization and Reversible Deactivation Radical Polymerization of Vinyl Acetate Mediated by Cobalt(II) Phenoxy-imine Complexes / Chi-How Peng --
17. Tailor-Made Poly(vinylamine)s via Thermal or Photochemical Organometallic Mediated Radical Polymerization / Antoine Debuigne --
18. Alkyl Bromide as Precursor of Initiating Dormant Species in Organocatalyzed Living Radical Polymerization / Atsushi Goto --
19. Biocatalytic ATRP / Nico Brans.

Citation preview

Reversible Deactivation Radical Polymerization: Mechanisms and Synthetic Methodologies

ACS SYMPOSIUM SERIES 1284

Reversible Deactivation Radical Polymerization: Mechanisms and Synthetic Methodologies Krzysztof Matyjaszewski, Editor Carnegie Mellon University, Pittsburgh, Pennsylvania

Haifeng Gao, Editor University of Notre Dame, Notre Dame, Indiana

Brent S. Sumerlin, Editor University of Florida, Gainesville, Florida

Nicolay V. Tsarevsky, Editor Southern Methodist University, Dallas, Texas

Sponsored by the ACS Division of Polymer Chemistry, Inc.

American Chemical Society, Washington, DC Distributed in print by Oxford University Press

Library of Congress Cataloging-in-Publication Data Names: Matyjaszewski, K. (Krzysztof), editor. | Gao, Haifeng, editor. | Sumerlin, Brent S., editor. | Tsarevsky, Nicolay V., editor. | American Chemical Society. Division of Polymer Chemistry. Title: Reversible deactivation radical polymerization / Krzysztof Matyjaszewski, editor (Carnegie Mellon University, Pittsburgh, Pennsylvania), Haifeng Gao, editor (University of Notre Dame, Notre Dame, Indiana), Brent S. Sumerlin, editor (University of Florida, Gainesville, Florida), Nicolay V. Tsarevsky, editor (Southern Methodist University, Dallas, Texas) ; sponsored by the ACS Division of Polymer Chemistry, Inc. Description: Washington, DC : American Chemical Society, [2018] | Series: ACS symposium series ; 1284, 1285 | Includes bibliographical references and index. Identifiers: LCCN 2018032910 (print) | LCCN 2018038460 (ebook) | ISBN 9780841233171 (ebook) | ISBN 9780841233188 (v. 1) | ISBN 9780841233232 (v. 2) Subjects: LCSH: Polymerization. | Crosslinking (Polymerization) Classification: LCC QD281.P6 (ebook) | LCC QD281.P6 R445 (print) | DDC 547/.28--dc23 LC record available at https://lccn.loc.gov/2018032910

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences—Permanence of Paper for Printed Library Materials, ANSI Z39.48n1984. Copyright © 2018 American Chemical Society Distributed in print by Oxford University Press All Rights Reserved. Reprographic copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Act is allowed for internal use only, provided that a per-chapter fee of $40.25 plus $0.75 per page is paid to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. Republication or reproduction for sale of pages in this book is permitted only under license from ACS. Direct these and other permission requests to ACS Copyright Office, Publications Division, 1155 16th Street, N.W., Washington, DC 20036. The citation of trade names and/or names of manufacturers in this publication is not to be construed as an endorsement or as approval by ACS of the commercial products or services referenced herein; nor should the mere reference herein to any drawing, specification, chemical process, or other data be regarded as a license or as a conveyance of any right or permission to the holder, reader, or any other person or corporation, to manufacture, reproduce, use, or sell any patented invention or copyrighted work that may in any way be related thereto. Registered names, trademarks, etc., used in this publication, even without specific indication thereof, are not to be considered unprotected by law. PRINTED IN THE UNITED STATES OF AMERICA

Foreword The ACS Symposium Series was first published in 1974 to provide a mechanism for publishing symposia quickly in book form. The purpose of the series is to publish timely, comprehensive books developed from the ACS sponsored symposia based on current scientific research. Occasionally, books are developed from symposia sponsored by other organizations when the topic is of keen interest to the chemistry audience. Before agreeing to publish a book, the proposed table of contents is reviewed for appropriate and comprehensive coverage and for interest to the audience. Some papers may be excluded to better focus the book; others may be added to provide comprehensiveness. When appropriate, overview or introductory chapters are added. Drafts of chapters are peer-reviewed prior to final acceptance or rejection, and manuscripts are prepared in camera-ready format. As a rule, only original research papers and original review papers are included in the volumes. Verbatim reproductions of previous published papers are not accepted.

ACS Books Department

Contents Preface .............................................................................................................................. ix 1.

Reversible Deactivation Radical Polymerization: State-of-the-Art in 2017 ....... 1 Sivaprakash Shanmugam and Krzysztof Matyjaszewski

2.

Mechanistic Insights into Lewis Acid Mediated Sequence- and Stereo-Control in Radical Copolymerization ...................................................... 41 Nicholas S. Hill, Benjamin B. Noble, and Michelle L. Coote

3.

Time-Resolved Electron Spin Resonance Observations of the Initial Stages of Conventional and Controlled Radical Polymerization Processes ................. 63 Atsushi Kajiwara

4.

Elements of RAFT Navigation .............................................................................. 77 Joris J Haven, Matthew Hendrikx, Tanja Junkers, Pieter J Leenaers, Theodora Tsompanoglou, Cyrille Boyer, Jiangtao Xu, Almar Postma, and Graeme Moad

5.

Reducing the Hydrogen Atom Abstraction Efficiencies of Benzophenone-Based Photosensitive Alkoxyamines ........................................ 105 Jason C. Morris, Jean-Louis Clément, Yohann Guillaneuf, Steven E. Bottle, Kathryn Fairfull-Smith, and Didier Gigmes

6.

Catalyzed Radical Termination (CRT) in the Metal-Mediated Polymerization of Acrylates: Experimental and Computational Studies ...... 135 Thomas G. Ribelli, S. M. Wahidur Rahaman, Krzysztof Matyjaszewski, and Rinaldo Poli

7.

Electrochemical Procedures To Determine Thermodynamic and Kinetic Parameters of Atom Transfer Radical Polymerization .................................... 161 Francesca Lorandi, Marco Fantin, Francesco De Bon, Abdirisak A. Isse, and Armando Gennaro

8.

Insights into the Reactivity of Epoxides as Reducing Agents in Low-Catalyst-Concentration ATRP Reactions ................................................. 191 David C. McLeod, Kapil Dev Sayala, and Nicolay V. Tsarevsky

9.

Toward Butadiene-ATRP with Group 10 (Ni, Pd, Pt) Metal Complexes ........ 205 Vignesh Vasu, Joon-Sung Kim, Hyun-Seok Yu, William I. Bannerman, Mark E. Johnson, and Alexandru D. Asandei

vii

10. Reversible Deactivation Radical Polymerization of Vinyl Chloride ............... 227 Carlos M. R. Abreu, Ana C. Fonseca, Nuno M. P. Rocha, James T. Guthrie, Arménio C. Serra, and Jorge F. J. Coelho 11. Photoinduced Metal Free Strategies for Atom Transfer Radical Polymerization ...................................................................................................... 263 G. Yilmaz, C. Kutahya, A. Allushi, C. Aydogan, S. Aykac, and Y. Yagci 12. Recent Developments in External Regulation of Reversible Addition Fragmentation Chain Transfer (RAFT) Polymerization ................................. 273 Sivaprakash Shanmugam, Cyrille Boyer, and Krzysztof Matyjaszewski 13. How Can Xanthates Control the RAFT Polymerization of Methacrylates? .... 291 Mathias Destarac, Dimitri Matioszek, Xavier Vila, Juliette Ruchmann-Sternchuss, and Samir Z. Zard 14. Hydroxyl Radical Activated RAFT Polymerization ......................................... 307 Thomas G. McKenzie, Amin Reyhani, Mitchell D. Nothling, and Greg G. Qiao 15. Vinyl Ether/Vinyl Ester Copolymerization by Cationic and Radical Interconvertible Simultaneous Polymerization ................................................. 323 Kotaro Satoh, Yuuma Fujiki, Mineto Uchiyama, and Masami Kamigaito 16. Catalytic Chain Transfer Polymerization and Reversible Deactivation Radical Polymerization of Vinyl Acetate Mediated by Cobalt(II) Phenoxy-imine Complexes .................................................................................. 335 Yi-Hao Chen, Hung-Hsun Lu, Jia-Qi Li, and Chi-How Peng 17. Tailor-Made Poly(vinylamine)s via Thermal or Photochemical Organometallic Mediated Radical Polymerization ........................................... 349 Pierre Stiernet, Mathilde Dréan, Christine Jérôme, Patrick Midoux, Philippe Guégan, Jutta Rieger, and Antoine Debuigne 18. Alkyl Bromide as Precursor of Initiating Dormant Species in Organocatalyzed Living Radical Polymerization ............................................. 365 Feifei Li, Longqiang Xiao, and Atsushi Goto 19. Biocatalytic ATRP ................................................................................................ 379 Jonas Pollard and Nico Bruns Editors’ Biographies .................................................................................................... 395

Indexes Author Index ................................................................................................................ 399 Subject Index ................................................................................................................ 401

viii

Preface This book and a following volume are addressed to chemists and polymer scientists interested in controlled/living radical polymerization. These volumes are intended to summarize recent mechanistic advances and innovation in the area of materials and applications. The two volumes comprise the topical reviews and specialists’ contributions presented at the American Chemical Society symposium on Controlled Radical Polymerization (the IUPAC preferred term is Reversible Deactivation Radical Polymerization and it is used in the titles of these volumes) that was held in Washington, DC, August 20-24, 2017. This most recent meeting was a sequel to several previous ACS Symposia on controlled/living radical polymerization, held in San Francisco, California (1997), New Orleans, Louisiana (1999), Boston, Massachusetts (2002), Washington, DC (2005), Philadelphia, Pennsylvania (2008), Denver, Colorado (2011), and San Francisco, California (2014). The work presented at those symposia was summarized in the ACS Symposium Series Volume 685: Controlled Radical Polymerization, Volume 768: Controlled/Living Radical Polymerization: Progress in ATRP, NMP and RAFT, Volume 854: Advances in Controlled/Living Radical Polymerization, Volume 944: Controlled/Living Radical Polymerization: From Synthesis to Materials, Volume 1023: Controlled/Living Radical Polymerization: Progress in ATRP, Volume 1024: Controlled/Living Radical Polymerization: Progress in RAFT, DT, NMP and OMRP, Volume 1100: Progress in Controlled Radical Polymerization: Mechanisms and Techniques, and Volume 1101: Progress in Controlled Radical Polymerization: Materials and Applications, Volume 1187: Controlled Radical Polymerization: Mechanisms, and Volume 1188: Controlled Radical Polymerization: Materials. The Washington, DC 2017 meeting was very successful with 88 lectures and a similar number of posters presented. It is a clear sign of the health of the field that this level of participation has continued to grow over the years and is only limited by the available time/speaking slots in the meeting. The 34 chapters submitted for publication in the ACS Symposium Series could not fit into one volume, and therefore we were asked by the ACS to divide the contents into two volumes. Similar to the volumes originating from the most recent Denver and San Francisco meetings, these two volumes are dedicated to mechanisms and techniques (the current volume consisting of 18 chapters) and materials and applications (the following volume, which consists of 16 chapters). All chapters published in these two volumes show that reversible deactivation radical polymerization has made significant progress within the last two decades. New systems have been discovered, substantial progress has been achieved in ix

understanding the mechanism and kinetics of reactions involved in all reversible deactivation radical polymerization systems. Significant progress has also been made towards developing a comprehensive relationship between molecular structure and macroscopic properties. Several commercial applications of reversible deactivation radical polymerization were announced at the Washington, DC Meeting and it is anticipated that new products made by controlled/living radical polymerization will soon be on the market. The financial support for the symposium from the following organizations is acknowledged: ACS Division of Polymer Chemistry, Inc., Army Research Office, Anton Parr, Boron Molecular, Kaneka, the National Science Foundation, PPG, Royal Society of Chemistry, Sigma Aldrich, Tosoh, and Wiley-VCH.

Krzysztof Matyjaszewski Center for Macromolecular Engineering Department of Chemistry Carnegie Mellon University 4400 Fifth Avenue Pittsburgh, Pennsylvania 15213

Haifeng Gao Department of Chemistry University of Notre Dame South Bend, Indiana 46556

Brent Sumerlin George & Josephine Butler Polymer Research Laboratory Center for Macromolecular Science & Engineering Department of Chemistry University of Florida Gainesville, Florida 32605-7200

Nicolay V. Tsarevsky Department of Chemistry Southern Methodist University 3215 Daniel Avenue Dallas, Texas 75275

x

Chapter 1

Reversible Deactivation Radical Polymerization: State-of-the-Art in 2017 Sivaprakash Shanmugam and Krzysztof Matyjaszewski* Center for Macromolecular Engineering, Department of Chemistry, Carnegie Mellon University, 4400 Fifth Avenue, Pittsburgh, Pennsylvania 15213, United States *E-mail: [email protected].

This chapter highlights the current advancements in reversible-deactivation radical polymerization (RDRP) with a specific focus on atom transfer radical polymerization (ATRP). The chapter begins with highlighting the termination pathways for acrylates radicals that were recently explored via RDRP techniques. This led to a better understanding of the catalytic radical termination (CRT) in ATRP for acrylate radicals. The designed new ligands for ATRP also enabled the suppression of CRT and increased chain end functionality. In addition, further mechanistic understandings of SARA-ATRP with Cu0 activation and comproportionation were studied using model reactions with different ligands and alkyl halide initiators. Another focus of RDRP in recent years has been on systems that are regulated by external stimuli such as light, electricity, mechanical forces and chemical redox reactions. Recent advancements made in RDRP in the field of complex polymeric architectures, organic-inorganic hybrid materials and bioconjugates have also been summarized.

Introduction The overarching goal of this chapter is to provide an overall summary of the recent achievements in reversible-deactivation radical polymerization (RDRP), primarily in atom transfer radical polymerization (ATRP), and also in reversible addition-fragmentation chain transfer (RAFT) polymerization, tellurium mediated © 2018 American Chemical Society

polymerization (TERP), iodine mediated polymerization (IMP), and nitroxide mediated polymerization (NMP). Of these techniques, progress in ATRP will be covered in more depth, since subsequent chapters in this book will provide more insights into the development of the other RDRP techniques especially RAFT polymerization. Due to the large volume of literature generated over the years, the discussion on the progress of these techniques will be focused on recent literature spanning from 2013 to 2017, since the last ACS Meeting on Controlled Radical Polymerization (1, 2). Nevertheless, readers are directed to several excellent reviews for in-depth discussions on the different areas explored (3–7). Initial sections will explore the recent discoveries in catalytic radical termination (CRT) in ATRP, design of a novel ligand that reduces CRT, and insights into ATRP in the presence of Cu0. In addition, as polymer chemists aim to design complex macromolecules similar to protein and DNA found in nature, achieving spatial, temporal, sequence and stereochemical regulations during polymer synthesis has recently become the central theme of RDRP. This book chapter will highlight the recent literature on photochemical, electrochemical, and mechanochemical means of achieving these goals. These achievements in externally regulated polymer synthesis are translated into advancing the synthesis of complex polymeric architectures, hybrid materials (polymer brushes), and bioconjugates. The advancements in these areas will also be highlighted.

Advancements in ATRP Mechanism of Catalytic Radical Termination (CRT) in ATRP In the discussion of novel mechanisms and initiation pathways, it is paramount to highlight the fundamental importance of understanding the mechanism of radical termination. Dispersity, which is a measure of polymer molecular weight distribution, relies on monomer conversion, number of monomer addition per activation/deactivation cycle, and amount of dead chains. As current literature often neglects the termination factor in calculation of dispersity for ATRP, a new dispersity expression by blending dormant and dead chain populations was recently derived (8). Bimolecular radical termination relies on two pathways – disproportionation and combination. In disproportionation, two chains, one with saturated chain end and the with unsaturated chain end, are formed. In combination, a single chain is made through C-C coupling (9). The pathway of chain termination in radical polymerization relies on the nature of the radical species involved. Styrene and acrylonitrile propagating radicals undergo chain termination primarily through combination while methacrylate radicals terminate by both disproportionation and combination (10). Different methods to study selectivity of radical termination have been proposed but these methods were unable to provide definitive conclusions. For instance, the ratios of disproportionation to combination for polymerizations of methyl methacrylate (MMA) and styrene (St) vary between reports despite identical polymerization conditions (11). A new method to study radical termination with TERP was proposed. The use of TERP allowed for activation of organotellurium dormant species by photoirradiation to generate polymer-end radical. This method was 2

used to estimate ratio of disproportionation to combination for poly(methyl methacrylate) and polystyrene radicals that approximately agreed with previous reports (11). In comparison to the well understood termination of PSt and PMMA radicals, termination of acrylate radicals was a subject of intense debate. A recent report claimed that acrylate radicals generated from organotellurium polymerizations are terminated primarily (99%) through disproportionation at room temperature (12). Ab initio molecular dynamics computations suggested that the polyacrylate radicals can be terminated through a direct disproportionation reaction as well as a new stepwise process involving initial formation of C-O coupling product followed by intramolecular rearrangement. In a subsequent study, an avenue to control the ratio of disproportionation to combination for termination of MMA and St radicals using TERP was suggested by changing reaction temperature and viscosity (13). Disproportionation was proposed to be favored over combination at lower temperature and higher viscosity. The observed viscosity effect in the selection of mode of termination was reasoned with an “advanced collision model”, a derivative of the collision model proposed by Fischer (14). The proposed model suggested that combination was more viscosity sensitive than disproportionation, and therefore, an increase in viscosity should result in a more significant retardation of the rate constant for combination than of that observed for disproportionation (13). However, a subsequent investigation revealed that alkyl radicals generated through the decomposition of diazo initiator dimethyl 2,2′-azobis(isobutyrate) (V-601), which directly generate methacrylate radicals, yielded similar Disp/Comb ratios in a large range of temperatures and viscosities (15). The discrepancies with the previously discussed results were explained by the role of alkyl tellurium radical acting as a catalyst for disproportionation. In a typical V-601 promoted termination, bimolecular chain coupling took place within the solvent cage or outside the solvent cage. In the case of TERP mediated polymerization, cage escape was needed to enable bimolecular termination as possible side reaction that involved β-H abstraction from the carbon-based radical by tellanyl radicals took place within the solvent cage. They formed alkene and Te-H species which generated saturated chain ends in a fast reaction with macroradicals. Metals such as iron have been successfully implemented in both organometallic-mediated radical polymerization (OMRP) and ATRP (16–18), but copper complexes which are widely used in ATRP have not been adapted for OMRP despite the proven existence of copper (II) organometallic complexes (19). Nevertheless, attempts have been made to characterize these organometallic intermediates. For instance, EPR measurements were carried out to characterize copper (II) species generated in CRT. By following the reaction of [CuI]+ formed in the presence of poly(butyl acrylate) radicals by EPR, a new signal distinct from [CuII(TPMA)Br]+ complex was observed (20). Organometallic complexes generated in ATRP were also investigated by electrochemical reduction and UV-Vis. These findings revealed that radicals generated from bromoacetonitrile and chloroacetonitrile (to a lesser extent ethyl α-bromoisobutyrate), react with [CuI(TPMA)]+ and [CuI(Me6tren)]+ in dimethylsulfoxide (DMSO) and acetonitrile (ACN) to form cupric complexes with alkyl moiety in the axial coordination site (19). 3

As ATRP systems contain highly active copper catalysts, additional side reactions where acrylate radicals can react with L/CuI to reversibly generate a complex (L/CuII-polyacrylate. These species can then undergo catalytic radical termination (CRT) which is proposed as the primary mode of termination of acrylate radicals in ATRP (21). The CRT process often results in termination of acrylates to form disproportionation-like products in a tris(2-pyridylmethyl)amine (TPMA)-copper catalyzed ATRP (22). Thus, a study to better understand the bimolecular termination and CRT in ATRP was designed. Poly(methyl acrylate) terminated with bromine end group was activated with copper(I) complexes with tris[2-(dimethylamino)ethyl]amine (Me6TREN), tris(2-pyridylmethyl)amine (TPMA) or tris-(3,5-dimethyl-4-methoxy-2-pyridylmethyl)amine (TPMA*3) in the absence of monomer (23). Under conditions kinetically promoting bimolecular termination, size exclusion chromatography (SEC) revealed a polymer product with double molecular weight relative to macroinitiator distribution indicating a termination pathway via radical combination. Likewise, conditions kinetically promoting CRT resulted in no shifting in macroinitiator distribution in SEC, suggesting products resembling disproportionation of polymer chains. PREDICI simulations for TPMA mediated termination reactions highlighted the importance of midchain radical generated from acrylate radical backbiting in the termination profile. Majority of the terminated chains originated from CRT and cross-termination between secondary and tertiary midchain radicals (Scheme 1). Further studies into the mechanistic pathways of acrylate termination will be discussed in another chapter.

Scheme 1. Proposed mechanistic pathways for acrylate radical termination in ATRP.

Design of New Ligands for ATRP The recent exploitation of reducing agents in ATRP significantly decreased the amount of copper complexes needed for polymerization (24). In order to 4

further reduce the amount of copper complexes in polymerizations, significant efforts were placed towards development of more active catalysts by fine tuning the ligand structure to tailor electronic properties of copper(I) center. To achieve this goal, electron-donating groups were systematically incorporated to highly active ATRP ligands to further increase the reactivity copper(I) complex. For instance, a series of novel Cu (I and II) complexes with TPMA-based ligands containing 4-methoxy-3,5-dimethylsubstituted pyridine arms were reported (25). Cyclic voltammetry measurements revealed that by increasing substitution of electron donating groups in the 4 (-OMe) and 3,5 (-Me) positions of the pyridine rings in TPMA, the reactivity of the copper(I) complexes also increases due to increased stabilization of copper(II) oxidation state (25). In addition, design of new ligands allowed for better understanding of the CRT process. Polymerization of n-butyl acrylate (BA) with azobis(isobutyronitrile) (AIBN) in the presence of copper complexes with tridentate and tetradentate ligands showed higher rates of CRT for more reducing copper complexes with higher ATRP activity. On the other hand, ligand denticity had smaller effect on polymerization kinetics but affected the rate determining step for CRT (26). Recently, tris[(4-dimethylaminiopyridyl) methyl]amine (TPMANMe2) was reported as a novel ligand for ATRP. The TPMANMe2 based copper catalyst was ~1 billion times more active than seminal bipyridine-based catalyst, ~300 000 more active than TPMA, and ~30 000 more active than Me6TREN. Polymerization of acrylates via ICAR and Ag0 ATRP were well-controlled for catalyst loadings as low as 10 ppm relative to monomer. The high values of activation rate constants of ATRP and low concentration of TPMANMe2/CuI suppressed CRT and allowed for high-end functionality (27). Progress in SARA-ATRP Traditional ATRP systems often required the use of large amounts of copper catalyst (>1000 ppm) to account for the conversion of CuI activator to CuII deactivator due to termination reactions. However, by slow and continuous reduction process, controlled polymerizations can be carried out using as low as 1:20) (56) with a low risk of multiple monomer unit insertion (i.e., oligomerization). 80

Figure 3. Synthesis of poly(3-hexylthiophene macroRAFT agent by RAFT-SUMI. Regions 2.4–3.4 and 4.9–5.9 ppm of 1H NMR spectra showing signals for vinyl-P3HT (lower trace) and P3HT macro-RAFT agent (upper trace). Spectra are of reaction mixtures for time 0 and 20 h respectively. For signal assignments and further details see ref (51). Adapted with permission from ref (51). Copyright 2011 The Royal Society of Chemistry.

Figure 4. Scope of RAFT-SUMI for insertion of a monomer into 2-cyanoprop-2-yl dithiobenzoate. 81

Prior to these studies, McLeary, Klumperman and colleagues (55, 58–63) had observed that complete conversion of an initial RAFT agent to a species incorporating only a single monomer unit (i.e., SUMI) preceded polymerization in many well-behaved RAFT polymerizations (including those of styrene (St) (58, 61), methyl acrylate (MA) (60, 63), N-vinylpyrrolidone (62) and vinyl acetate (VAc) (62)). This behavior was called selective initialization. However, similar selectivity in formation of a two unit adduct was not observed. We made similar observations for styrene polymerization and found that the phenomenon was strongly dependent on the RAFT agent and the specific polymerization conditions used (64). For example, with 4.3 M St and 0.5 M RAFT agent, selective initialization is observed with 2-cyanoprop-2-yl and cumyl dithiobenzoates, but not with benzyl dithiobenzoate (poorer homolytic leaving group) or 2-cyanoprop-2-yl dodecyl trithiocarbonate (lower transfer constant RAFT agent) (64). Selective initialization may be observed with 2-cyanoprop-2-yl dodecyl trithiocarbonate but only when higher RAFT agent to styrene ratios are used (48). Selective SUMI requires a RAFT agent with Ctr such that, on average, there is >k-add, by low relative monomer concentrations (e.g., stoichiometric with RAFT agent) and an initiator-derived radical that is identical to the RAFT agent ‘R’ group (48, 49). SUMI is an important technique for converting macromonomers into macroRAFT agents that can be later elaborated by RAFT polymerization (51). In cases where high yield SUMI is not possible, the application of separation techniques, such as preparative recycling size exclusion chromatography (65–68), or flash column chromatography (69) can enable separation of discrete oligomers from the reaction-derived oligomer mixtures. Thus, Vandenbergh et al. (67) performed four consecutive SUMI of acrylate monomers into a trithiocarbonate RAFT agent. Excess (10-fold) monomer was used in the experiments and the degree of oligomerization was controlled by limiting the monomer conversion through short (10 min) reaction times. Automated recycle size exclusion chromatography (SEC) was developed to provide a pure SUMI product after each step.

Thermally-Initiated Sequential RAFT SUMI Zard and co-workers (45) reported consecutive SUMI of NVPI followed by an allyl monomer into a xanthate. We have demonstrated high yields in consecutive SUMI for St followed by maleic anhydride (MAH) (49). Success in this case can be partly attributed to MAH being essentially inert towards the AIBN-derived 2-cyanoprop-2-yl radicals (Figure 5). A key factor contributing to success in both of these examples is the use of a non-homopolymerizable monomer (70) (i.e., kp~0) in the second SUMI step.

82

Figure 5. Scope of RAFT-SUMI for sequential insertion of styrene and a second monomer into 2-cyanoprop-2-yl dithiobenzoate. Attempted SUMI of MAH or maleimide monomers into RAFT agents with R=tertiary cyanoalkyl and AIBN initiation was unsuccessful. The RAFT agent remained largely untouched. However, SUMI of maleimide monomers into the dithiobenzoate with R=tertiary ester 1 with azobis(methyl isobutyrate) (AIBMe) initiation proceeded in very high yield to provide 2 (Figure 6, 60% isolated yield after chromatography).

Figure 6. RAFT-SUMI for insertion of NPMI into (2-methoxycarbonyl)prop-2-yl dithiobenzoate (1) to provide the insertion product 2. However, 2 was unreactive towards RAFT-SUMI of various monomers [methyl acrylate, N,N-dimethylacrylamide (DMAm), styrene (St)] and the RAFT agent being recovered unchanged. PET-RAFT-SUMI (vide infra) of monosubstituted MAMs into maleimide terminal macroRAFT agents was also unsuccessful. The explanation for low reactivity of these RAFT agents in SUMI experiments is still being investigated. We found that the RAFT agent 2 is effective in controlling, e.g., styrene polymerization (Table 1), though some irregularity for short reaction times is apparent, where the higher than expected molar masses may indicate slow utilization of RAFT agent. 83

SUMI of non-homopolymerizable monomers also provides a method of chainend functionalization for RAFT synthesized polymers. MAH (52–55), maleimides (56, 57), β-pinene (57), and ABOC (for a primary amino end-group) (71) have been used in this context.

Table 1. Molar Mass Conversion Data for Styrene Polymerizations (110 °C, Thermal Initiation, No Added Initiator) with RAFT Agent 2

a

Time (h)

Conversion

Mn

Mn(calc)a

Dispersity

1

9.0

2300

3253

1.07 (bimodal)

2

11.4

3600

4003

1.07 (trimodal)

4.5

17.2

6300

5815

1.06

17

55.0

18600

17626

1.08

Mn(calc) = [Styrene]t/[RAFT agent]0+MW(RAFT agent).

SUMI of N-isopropylacrylamide (NIPAm) into the macroRAFT agent formed by SUMI of St into a dithiobenzoate was selective but very slow. However, yields of the desired SUMI product are substantially lowered by the concurrent formation of products from initiator-derived chains (48, 49). This finding has prompted us to investigate RAFT-SUMI of acrylamides in aqueous media and PET-RAFT-SUMI.

RAFT-SUMI of Acrylamides in Aqueous Media

Trithiocarbonates 3 with R=2-carboxyeth-2-yl and Z′S=primary alkylthio (specifically, 4-7) have been shown to be very effective agents for mediating the polymerization of monosubstituted MAMs such as acrylate, acrylamides and styrene providing low dispersities and good molar mass control. The most common RAFT agents of this class are 4 (30 references in ScifinderTM in Jan 2018), 5 (190 references), 6 (24 references) and 7 (144 references). The last three RAFT agents are commercially available. 84

Tertiary cyanoalkyl trithiocarbonates 8 with Z’S=primary alkylthio (e.g., 812) are very effective agents for mediating the polymerization of both mono- and 1,1-disubstituted MAMs. They (8) have substantially higher transfer constants than 3. Amongst the most common RAFT agents of this class (8) are those with R=4-carboxy-2-cyanobut-2-yl, which include 9 (typically in aqueous media, 191 references in SciFinderTM) and 11 (typically in organic media, 340 references). These RAFT agents, along with 10 (5 references), are commercially available.

Some initial experiments on attempted RAFT SUMI in an organic solvent (CD3CN) with trithiocarbonates 5 or 12 are shown in Figure 7. Trithiocarbonate 12 had been successfully used in earlier work (48). With trithiocarbonate 5 (Figure 7a), the reaction rate is very much faster than with 12 and selective SUMI is not observed. Products from SUMI and multiple unit insertion are formed simultaneously even at very low monomer conversion. The final product is an oligomer Xn 1.4. With 5 propagation is rate determining. With trithiocarbonate 12 selective SUMI is seen and oligomeric products from multiple unit insertion are not detected. However, the reaction rate is very much slower. After 1500 min, when the conversion plateaus due to the initiator (AIBN) being largely depleted, only 60% of the initial RAFT agent has been converted to SUMI. Under these condition the rate determining step is the 2-cyanoprop-2-yl radical adding to monomer (48). Note also that relatively high initiator concentrations were used in these experiments. 85

Figure 7. Evolution of species observed by in situ NMR during attempted RAFT SUMI of NIPAm into acid functional trithiocarbonates (a) 5 or (b) 12 in CD3CN at 70 °C. [NIPAm]:[RAFT]:[AIBN]=1:1:0.2. The amount of residual RAFT agent in (b) was estimated as [RAFT est]=1.0-[SUMI]. The amounts of AIBN and TMSN are assumed to be the same as in (a).

The diacid trithiocarbonate 6 was first reported in 2005 when it was shown to be an effective RAFT agent in bulk polymerization of butyl acrylate (BA) (72). 6 was also successfully used to mediate bulk styrene polymerization.73 The RAFT agent is water soluble and suitable for controlling aqueous copolymerization of water soluble styrenic monomers (74). It has also been successfully used to mediate RAFT inverse miniemulsion of acrylamide (Am) (75), acrylic acid (AA) (76) and AA-Am copolymers at various pH in the range 3-10 (77). Partial loss of control was observed for pH >7, which was attributed to the (hydrolytic) 86

instability of the RAFT agent (76). Most recently 6 was used in the synthesis of low dispersity acrylamide multi-block copolymers using a looped flow process (78). With this background, we initially chose 6 as a candidate for performing SUMI of acrylamide monomers (DMAm and NIPAm) in aqueous media. To ensure full solubility in aqueous DMAm at the concentration desired, the RAFT agent 6 was neutralized by addition of two equivalents of Na2CO3. Surprisingly, in an attempted SUMI experiment, we found that the RAFT agent was essentially unreactive and could be observed still largely unchanged by 1H NMR at the end of the experiment, during which period a higher molar mass poly(DMAm) was formed. Under similar conditions, we found that 5- was consumed but provided an oligomeric product consistent with the transfer constant being lower than required for successful SUMI. This result is consistent with the observation made for attempted SUMI in organic media (vide infra). Out of the series of trithiocarbonates, 6=, 5- and 9-, selective SUMI was only observed for 9-, which has a tertiary R group and, consequently, a higher transfer constant in DMAm polymerization (Figure 8).

Figure 8. Scope of RAFT-SUMI for sequential insertion of NIPAm into acid functional trithiocarbonates. An explanation for this behavior is that the transfer constant of the RAFT agent increases in the series 6=kp(n) for each step. It also helps if the product macroRAFT agent formed by SUMI is less active as a photoiniferter than the initial RAFT agent. Since no thermal initiator is used, initiator-derived byproducts (initiatorderived chains, cage products) and a slow rate of reaction caused by poor initiator efficiency – vide infra) are not an issue Products from termination will be formed in amounts consistent with the radical concentrations. With an appropriately (low) rate of photo-initiation this can be controlled. For the second monomer, N-substituted maleimides (N-phenylmaleimide (PMI), N-benzylmaleimide (BMI) and N-ethylmaleimide (EMI)) were selected due to their high reactivity towards propagating radicals with a terminal styrene unit. The maleimides undergo homopolymerization very slowly (85, 86). The PET-RAFT-SUMI experiment was performed with fac-tris[2-phenylpyridinato-C2,N]iridium(III) (fac-[Ir(ppy)3]) as the photoredox catalyst under blue light irradiation. As in the thermally initiated experiments described above, this allowed a 1:1 molar ratio of monomer to RAFT agent to be utilized. Essentially complete conversion was achieved within 48 h. As the third monomer VAc was inserted into the CDTPA-St-PMI using zinc tetraphenylporphyrin (ZnTPP) as photocatalyst under red light irradiation (λmax = 630 nm, 0.4 mW/cm2) with DMSO as solvent [[VAc]:[CDTPA-St-PMI]:[ZnTPP] of 20:1:0.01]. The selection of ZnTPP over fac-[Ir(ppy)3] as photocatalyst was driven by preliminary results that indicated the C-S bond of VAc macroRAFT agents undergo (re)activation by fac-[Ir(ppy)3], which could lead to multiple monomer insertion. RAFT polymerization of VAc using trithiocarbonate RAFT agents with thermal initiation is known to be strongly inhibited, which has been attributed to slow fragmentation of the intermediate radical (87, 88). Thermally initiated RAFT and PET-RAFT of VAc has used xanthate (89, 90) or dithiocarbamate RAFT agents (91–95). In the case of SUMI, processes that disfavour propagation following SUMI are desirable. Experiments were also performed with the non-propagating monomer limonene as third monomer.. 92

The analyses that attest to the success of our sequential PET-RAFT-SUMI experiments are shown in Figure 12 (GPC traces) and Figure 13 (ESI mass spectra).

Figure 12. GPC traces for the products from sequential PET-RAFT-SUMI into CDTPA (11). Traces shown are for CDTPA-St, CDTPA-St-NPMI, CDTPA-St-NPMI-VAc and CDTPA-St-NPMI-Lim. Reproduced with permission from ref (84). Copyright 2017 Wiley-VCH.

Figure 13. ESI-MS spectra for the products from sequential PET-RAFT-SUMI into CDTPA (11). Those shown are for initial SUMI product CDTPA-St (D), dimer CDTPA-St-NPMI (C), and trimers CDTPA-St-NPMI-VAc (B) and CDTPA-St-PMI-Lim (A). Reproduced with permission from ref (84). Copyright 2017 Wiley-VCH. 93

Figure 14. Strategy for producing discrete oligomers by successive SUMI, aminolysis, thiol-Michael addition and esterification steps.

An issue in conducting sequential SUMI of monosubstituted monomers is that the macroRAFT agents formed by SUMI do not have a sufficiently high transfer constant to allow their use in a subsequent selective SUMI experiment. It is necessary to alter of the activity of the macroRAFT agent formed by SUMI. In developing an approach to iterative SUMI, we were inspired by the work of Porel et al. (96–98) who demonstrated a synthesis of discrete oligomers based on successive thiol-ene reactions. Our approach then uses successive aminolysis, thiol-ene and esterification steps to transform the secondary trithiocarbonate end produced by photo-SUMI into a more active tertiary cyanoalkyl trithiocarbonate (Figure 14). The process may be seen to lack elegance and yields are ultimately limited by the efficiency of the isolation steps. Nonetheless, we have demonstrated a new protocol for incorporating the rich functionality of available vinyl monomers into polymers whose sequence is precisely defined at the monomer level. Thus far we have taken the process through three SUMI steps. In the course of this study we demonstrated catalyst-free photo-RAFT-SUMI (green light) in to a range of monosubstituted monomers (acrylates, acrylamides and styrenes) into trithiocarbonate 11 (Table 3). Attempted photo-RAFT-SUMI MMA into 11 under similar conditions was unsuccessful and provided an oligomeric product, which is attributed to the low transfer constant of 11 in MMA polymerization. 94

Table 3. Catalyst-Free Photoinitiated (Green Light) Single Unit Monomer Insertion into Trithiocarbonate (11)a Monomer

Time (h)

[monomer]/[RAFT]

NMR Yield (%)b

Isolated Yield (%)c

MA

24

12

>94

72

HEA

24

15

>95

66

DMAm

24

15

>95

72

NIPAm

24

15

>95

68

TlaAm

22

10

>95

60

PFSt

24

15

>95

80

a Reactions were carried out in DMSO and were degassed by nitrogen sparging. MA – methyl acrylate, HEA – hydroxyethyl acrylate, DMAm – N,N-dimethylacrylamide, NIPAm – N-isopropylacrylamide, TlaAm - thiolactone acrylamide [N-(2-oxotetrahydrothiophen3-yl)acrylamide], PFSt – pentafluorostyrene. b crude yield based on RAFT agent consumed; c isolated yield after purification by silica column chromatography.

The proposed mechanism of photoRAFT-SUMI is shown in Figure 15. Photolysis of the RAFT agent causes reversible dissociation of the C-S bond (iniferter process). The radical (Rˑ) formed can add monomer (or RAFT agent, which is a degenerate process). The radical formed by addition of monomer (R-Mˑ) can combine with a thiocarbonylthio radical to form the SUMI product, It can react with RAFT agent by addition-fragmentation, also to form the SUMI product. It may also add further monomer, which is slow. The relative concentrations of the reacting species dictate that RAFT should be the dominant process.

Figure 15. Mechanism of photoRAFT-SUMI with a trithiocarbonate RAFT agent. 95

In more recent work, we have examined visible light-initiated SUMI of DMAm into trithiocarbonate 10 in aqueous solution (99). This established that selective photoSUMI could be achieved using relatively high monomer concentrations (1 M) and with stoichiometric RAFT agent and monomer. We found that that the specificity for SUMI, over formation of higher oligomers (or byproducts), was strongly dependent on the irradiation wavelength. In particular, red light provided for selective excitation of the initial RAFT agent (10) in the presence of DMAm. This was not possible with blue or green light, which when used gave a conversion plateau at ~60-70% monomer conversion. Red light provided the cleanest reaction product and a linear kinetic profile to high (>80%) conversion, albeit with a much slower rate of reaction.

Conclusions This paper has shown progress in the development of RAFT-SUMI as a pathway discrete oligomers. Thermally initiated RAFT-SUMI represents an important, potentially high yield, route to macroRAFT agents but has obvious limitations as a route to discrete oligomers. Circumstances which allow very low initiator concentrations, such as those that pertain in aqueous polymerization of acrylamides, offer some promise. A big opportunity lies with PET-RAFT-SUMI, which allows one important source of byproducts to be avoided, and we anticipate further developments will quickly follow.

Experimental The procedures and instrumentation used for thermal SUMI not detailed below are described in our previous papers (48, 49). PET-RAFT-SUMI is described elsewhere (84, 100).

Materials The RAFT agents 6 and 10 were obtained from Boron Molecular. The RAFT agents and 5 were obtained from Sigma-Aldrich. Monomers (BA, DMAm, NVP, VAc) were obtained from Sigma-Aldrich and were treated with inhibitor remover (Aldrich) and flash-distilled prior to use as appropriate. NIPAm (Sigma-Aldrich) was purified by recrystallization from hexane/Et2O 4:1. MAH (Sigma-Aldrich) was used as received. The initiator (E)-2,2′-(Diazene-1,2-diylbis(propane2,2-diyl))bis(4,5-dihydro-1H-imidazol-3-ium) chloride (Wako VA-044) was obtained from Novachem and used as received. 1,1′-Azobis(isobutyronitrile) AIBN was obtained from DuPont (VAZO64) and was recrystallized from methanol/chloroform. 4,4′’-Azobis(4-cyanopentanoic acid) was obtained from Sigma-Aldrich. Non-aqueous solvents of high purity were obtained from commercial sources (Acros or Sigma-Aldrich) and used without purification. 96

Thermally-Initiated Sequential RAFT SUMI 2-(2,5-Dioxo-1-phenyl-4-((phenylcarbonothioyl)thio)pyrrolidin-3-yl)-2methylpropanoate (2) Ethyl 2-methyl-2-(phenylthiocarbonylthio)propionate (1) (600 mg, 2.24 × 10-3 mol), N-phenylmaleimide (387 mg, 0.00224 mol), 2,2′-azobis(2methylpropionitrile) (AIBN) (73.42 mg, 0.000447 mol) and acetonitrile (11.4 mL) were combined in an ampoule. The ampoule was degassed by three freeze-pump-thaw cycles, sealed under reduced pressure (9.0 × 10-3 mbar) and subsequently heated at 75 ºC for 21 hours. After cooling the ampoule was opened and the resulting dark red solution concentrated in vacuo. The reaction mixture was further purified using column chromatography (silica gel, eluent gradient 100% pentane -> 100% methylene chloride) to give 634 mg of a crystalline red solid ethyl 2-(2,5-dioxo-1-phenyl-4-((phenylcarbonothioyl)thio)pyrrolidin-3-yl)2-methylpropanoate (2) (yield: 64%). 1H-NMR (400 MHz, CDCl3, δ): 7.92 (dd, J1 = 8.4 Hz, J2 = 1.3 Hz, 2H); 7.59 (tt, J1 = 7.6 Hz, J2 = 1.2 Hz, 1H); 7.34-7.53 (m, 7H); 5.11 (s (br), 1H); 4.21 (m, 2H), 3.46 (d, J = 6.6 Hz, 1H); 1.58 (s, 3H); 1.41 (s, 3H); 1.28 (t, J = 7.1 Hz, 3H). 13C-NMR (100 MHz, CDCl3, δ): 175.54; 174.94; 143.81; 133.50; 132.15; 129.30; 128.93; 128.73; 127.34; 126.61; 61.71; 52.43; 49.61; 45.46; 25.30; 23.83; 14.27. ESI-MS: [M-H+] calc: 442.1141, found: 442.1140.

RAFT Polymerization of Styrene Mediated by Dithiobenzoate 2 A stock solution containing of styrene (4.55 g) and dithiobenzoate 2 (64.3 mg) was prepared in a 10 mL Erlenmeyer flask. Aliquots of this solution (1 mL) were transferred to 4 ampoules and each was degassed by three freeze-pump-thaw cycles, sealed under reduced pressure (9.0 × 10-3 mbar) and heated at 110 ºC for 1, 2, 4.5 and 17 hours respectively. The resulting polymers were diluted with DCM and 3 times precipitated into methanol to obtain a pink polystyrene. The molar mass and dispersity for various monomer conversions are shown in Table 1. RAFT-SUMI of Acrylamides in Aqueous Media 4-Cyano-4-(((ethylthio)carbonothioyl)thio)Pentanoic Acid (9) A 1 L round bottom flask was charged with NaH (13.35 g, 0.334 mol; 60% in oil) and 630 mL Et2O. The suspension was cooled (ice-bath) and stirred vigorously while ethanethiol (20 g, 24 mL, 0.321 mol) was added dropwise. A colour change from grey to yellow was observed. The mixture was stirred for a further 10 min. The crude sodium S-ethyl trithiocarbonate (41.83 g) was collected by filtration, then resuspended in Et2O (500 mL). Iodine (19.95 g) was added to the suspension with vigorous stirring. A colour change from yellow to brown was observed. The solution was washed with Na2S2O3 (5% in water, 3 × 160 mL), Brine (1 × 100 mL) and dried over Na2SO4. The solvent was removed in vacuo and the crude bis-ethyl 97

trithiocarbonate (5 g) was dissolved in ethyl acetate (100 mL) and 4,4’-azobis(4cyanopentanoic acid) (7.68 g, 27.4 mmol) was added with additional ethyl acetate (30 mL). The reaction mixture was degassed by sparging with N2 for 20 min then heated under reflux for 18 h. The mixture was cooled and the solvent removed in vacuo. The product was purified by chromatography on silica gel with 70 : 30 pentane : ethyl acetate + 5% acetic acid as eluent. The solvents were evaporated to leave an orange oil. The product was further purified by recrystallization from carbon tetrachloride (4.51 g, 93.8%). 1H-NMR: δ (ppm) = 3.32 (q, S-CH2-CH3), 3.66 (t, -CH2-CH2-COOH), 2.53-2.37 (m, -CH2-CH2-COOH), 1.86 (s, -CH3), 1.34 (t, -CH2-CH3).

Attempted SUMI of NIPAm into 2-(((Butylthio)carbonothioyl)thio)Propanoic Acid (5) in CD3CN at 70 °C NIPAm (0.1132 g, 1.0004 mmol), 5 (0.2380 g, 0.9984 mmol) and CD3CN (1 mL) were combined in a 10 mL sample vial and AIBN (0.0328 mg, 0.1997 mmol) was added. 0.75 mL of this solution was transferred into a valved NMR tube [the NMR tube was a Wilmad 528-PP-7 NMR tube with a J Young valve fitted, manufactured by GPE Scientific Limited (UK)] and the solution was degassed by freeze-pump-thaw (4 cycles) with the use of dry ice. After the fourth cycle, the NMR tube remained under vacuum and was placed in the NMR probe at ambient temperature (25 °C). The lock was established at 25 °C, The NMR tube was then removed from the probe and the probe is heated to 70 °C. After achieving the desired temperature the NMR tube was loaded into the NMR spectrometer and that time was taken as time zero. After 24 hrs the reaction was quenched by removing the NMR tube from the spectrometer and exposing the mixture to air. Based on integration of the post-reaction NMR spectra the Xn of oligo(NIPAm) was estimated to be ~1.4. The kinetic data obtained in the experiment are reported in Figure 7a.

Automated Sequential Insertion of DMAm RAFT Agent 9 in D2O at 60 °C Using a Chemspeed® Robotic Platform The various stock solutions were prepared and degassed by N2 sparging for ~15 minutes. The RAFT agent stock solution (A, 10 mL) contained the RAFT agent 9 (0.35 g, 1.33 mmol, 0.133 M) and 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid sodium salt (1H NMR reference, 5 mg) in D2O/CD3CN (40%). The first monomer stock solution B (5 mL) contained DMAm (1.15 mL, 11.10 mmol, 2.22 M) in D2O. The second monomer stock solution C (5 mL) contained DMAm (0.23 mL, 2.2 mmol, 0.44 M) in D2O. The initiator stock solution (D, 5 mL) contained VA-044 (65 mg, 0.2 mmol, 0.04 M) in D2O. The stock solutions were placed in the Chemspeed® SLTII sample holder and the chamber was purged with N2 to obtain an inert atmosphere. A first reactor was charged with stock solution A (3 mL), B (0.9 mL) and C (0.1 mL), which was heated to 70 °C and vortexed for 2 h before withdrawing 0.1 mL for analysis. At this stage, stock solution C (0.9 mL) and D 98

(0.1 mL) were added. After 2h of vortexing an aliquot (0.1 mL) was withdrawn for analysis. All analysis samples were immediately dispensed into cool D2O (0.9 mL) to quench further reaction. This process was repeated for all subsequent blocks. GPC data relevant to this experiment is shown in Figure 10.

Acknowledgments Research on thermally initiated SUMI was conducted by MH, PJL and TT while they were present at CSIRO under the industrial traineeship component of their Masters degrees at Eindhoven University of Technology. Research by JJH was carried out as part of his PhD program at Hasselt University while on an exchange visit to CSIRO, which was in part funded under a CSIRO Newton-Turner Award provided to GM. We are grateful to Roger Mulder and Jo Cosgriff for assistance with NMR spectroscopy and to Ben Muir and Shaun Howard for assistance in designing and conducting the high throughput (Chemspeed®) experiments. PET-RAFT-SUMI experiments were largely conducted by JX, Changkui Fu and Sivaprakash Shanmugam (84) or Changkui Fu and Zixuan Huang (100) at CAMD UNSW under the supervision of CB as part of an informal collaboration also involving Craig Hawker (University of California Santa Barbara) and GM.

References 1. 2.

3. 4. 5. 6. 7.

8. 9. 10.

11. 12. 13.

Le, T. P.; Moad, G.; Rizzardo, E.; Thang, S. H. Polymerization with living characteristics. WO9801478, DuPont/CSIRO, 1998. Chiefari, J.; Chong, Y. K.; Ercole, F.; Krstina, J.; Jeffery, J.; Le, T. P. T.; Mayadunne, R. T. A.; Meijs, G. F.; Moad, C. L.; Moad, G.; Rizzardo, E.; Thang, S. H. Macromolecules 1998, 31, 5559–5562. Jenkins, A. D.; Jones, R. I.; Moad, G. Pure Appl. Chem. 2010, 82, 483–491. Szwarc, M. Nature 1956, 178, 1168–1169. Szwarc, M. J. Polym. Sci., Part A: Polym. Chem. 1998, 36, IX–XV. Penczek, S.; Moad, G. Pure Appl. Chem. 2008, 80, 2163–2193. Moad, G. RAFT Polymerization -Then and Now. In Controlled Radical Polymerization: Mechanisms; Matyjaszewski, K., Sumerlin, B. S., Tsarevsky, N. V., Chiefari, J., Eds.; American Chemical Society: Washington, DC, 2015; Vol. 1187, pp 211−246. Krstina, J.; Moad, C. L.; Moad, G.; Rizzardo, E.; Berge, C. T.; Fryd, M. Macromol. Symp. 1996, 111, 13–23. Krstina, J.; Moad, G.; Rizzardo, E.; Winzor, C. L.; Berge, C. T.; Fryd, M. Macromolecules 1995, 28, 5381–5385. Engelis, N. G.; Anastasaki, A.; Nurumbetov, G.; Truong, N. P.; Nikolaou, V.; Shegiwal, A.; Whittaker, M. R.; Davis, T. P.; Haddleton, D. M. Nat. Chem. 2016, 9, 171–178. Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2012, 65, 985–1076. Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2009, 62, 1402–1472. Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2006, 59, 669–692. 99

14. Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2005, 58, 379–410. 15. Keddie, D. J.; Moad, G.; Rizzardo, E.; Thang, S. H. Macromolecules 2012, 45, 5321–5342. 16. Moad, G.; Rizzardo, E.; Thang San, H. Fundamentals of RAFT Polymerization. In Fundamentals of Controlled/Living Radical Polymerization; Tsarevsky, N. V., Sumerlin, B. S., Eds.; Royal Society of Chemistry: Cambridge, U.K., 2013; pp 205−249. 17. Moad, G. Macromol. Chem. Phys. 2014, 215, 9–26. 18. Moad, G.; Rizzardo, E.; Thang, S. H. Polym. Int. 2011, 60, 9–25. 19. Moad, G.; Chen, M.; Häussler, M.; Postma, A.; Rizzardo, E.; Thang, S. H. Polym. Chem. 2011, 2, 492–519. 20. Moad, G. Polym. Int. 2015, 64, 15–24. 21. Moad, G. Polym. Int. 2017, 66, 26–41. 22. Moad, G. Polym. Chem. 2017, 8, 177–219. 23. Jasperse, C. P.; Curran, D. P.; Fevig, T. L. Chem. Rev. 1991, 91, 1237–1286. 24. Clark, A. J. Chem. Soc. Rev. 2002, 31, 1–11. 25. Moad, G.; Solomon, D. H. Initiation. In The Chemistry of Radical Polymerization, 2nd ed.; Elsevier: Oxford, 2006; pp 49−166. 26. Kharasch, M. S.; Jensen, E. V.; Urry, W. H. Science 1945, 102, 128. 27. Matyjaszewski, K. Curr. Org. Chem. 2002, 6, 67–82. 28. Taylor, M. J. W.; Eckenhoff, W. T.; Pintauer, T. Dalton Trans. 2010, 39, 11475–11482. 29. Cao, M.-Y.; Ren, X.; Lu, Z. Tetrahedron Lett. 2015, 56, 3732–3742. 30. Hawker, C. J. J. Am. Chem. Soc. 1994, 116, 11185–11186. 31. Moad, G.; Rizzardo, E., History of nitroxide-mediated polymerization. In Nitroxide Mediated Polymerization: From Fundamentals to Applications in Materials; Gigmes, D., Ed.; The Royal Society of Chemistry: Cambridge, U.K., 2016; pp 1−44. 32. Lutz, J.-F. Polym. Chem. 2010, 1, 55–62. 33. Lutz, J.-F.; Ouchi, M.; Liu, D. R.; Sawamoto, M. Science 2013, 341. 34. Lutz, J. F. Macromol. Rapid. Commun. 2017, 38, 1700582. 35. Soejima, T.; Satoh, K.; Kamigaito, M. ACS Macro Lett. 2015, 4, 745–749. 36. Soejima, T.; Satoh, K.; Kamigaito, M. J. Am. Chem. Soc. 2016, 138, 944–954. 37. Soejima, T.; Satoh, K.; Kamigaito, M. Polym. Chem. 2016, 7, 4833–4841. 38. Tong, X.; Guo, B.-h.; Huang, Y. Chem. Commun. 2011, 47, 1455–1457. 39. Hibi, Y.; Ouchi, M.; Sawamoto, M. Angew. Chem., Int. Ed. Engl. 2011, 50, 7434–7437. 40. Delduc, P.; Tailhan, C.; Zard, S. Z. J Chem Soc Chem Comm 1988, 308–310. 41. Corpart, P.; Charmot, D.; Biadatti, T.; Zard, S.; Michelet, D. Block polymer synthesis by controlled radical polymerization. WO9858974, Rhodia Chimie, 1998. 42. Zard, S. Z. Angew. Chem., Int. Ed. Engl. 1997, 36, 672–685. 43. Quiclet-Sire, B.; Zard, S. Z. Pure Appl. Chem. 2011, 83, 519–551. 44. Quiclet-Sire, B.; Zard, S. Z. Chimia 2012, 66, 404–412. 45. Quiclet-Sire, B.; Revol, G.; Zard, S. Z. Tetrahedron 2010, 66, 6656–6666. 100

46. Chen, M.; Ghiggino, K. P.; Mau, A. W. H.; Rizzardo, E.; Sasse, W. H. F.; Thang, S. H.; Wilson, G. J. Macromolecules 2004, 37, 5479–5481. 47. Chen, M.; Ghiggino, K. P.; Rizzardo, E.; Thang, S. H.; Wilson, G. J. Chem. Commun. 2008, 1112–1114. 48. Houshyar, S.; Keddie, D.; Moad, G.; Mulder, R.; Saubern, S.; Tsanaktsidis, J. Polym. Chem. 2012, 3, 1879–1889. 49. Moad, G.; Guerrero-Sanchez, C.; Haven, J. J.; Keddie, D. J.; Postma, A.; Rizzardo, E.; Thang, S. H. RAFT for the Control of Monomer Sequence Distribution – Single Unit Monomer Insertion (SUMI) into Dithiobenzoate RAFT Agents. In Sequence-Controlled Polymers: Synthesis, Self-assembly and Properties; Lutz, J.-F., Ouchi, M., Sawamoto, M., Meyer, T., Eds.; American Chemical Society: Washington DC, 2014; Vol. 1170, pp 133−147. 50. Adhikari, R.; Postma, A.; Li, J.; Arima, K.; Hirai, T.; Shimizu, H.; Ueno, K. J. Inst Image Inform Televis Eng 2012, 66, 370–376. 51. Chen, M.; Haeussler, M.; Moad, G.; Rizzardo, E. Org. Biomol. Chem. 2011, 9, 6111–6119. 52. Feng, X. S.; Pan, C. Y. Macromolecules 2002, 35, 4888–4893. 53. Shi, G. Y.; Tang, X. Z.; Pan, C. Y. J. Polym. Sci., Part A: Polym. Chem. 2008, 46, 2390–2401. 54. Sasso, B.; Dobinson, M.; Hodge, P.; Wear, T. Macromolecules 2010, 43, 7453–7464. 55. van den Dungen, E. T. A.; Rinquest, J.; Pretorius, N. O.; McKenzie, J. M.; McLeary, J. B.; Sanderson, R. D.; Klumperman, B. Aust. J. Chem. 2006, 59, 742–748. 56. Henry, S. M.; Convertine, A. J.; Benoit, D. S. W.; Hoffman, A. S.; Stayton, P. S. Bioconjugate Chem. 2009, 20, 1122–1128. 57. Zhou, C.; Qian, S.; Zhang, A.; Xu, L.; Zhu, J.; Cheng, Z.; Kang, E.-T.; Yao, F.; Fu, G. D. RSC Advances 2014, 4, 8144–8156. 58. McLeary, J. B.; Calitz, F. M.; McKenzie, J. M.; Tonge, M. P.; Sanderson, R. D.; Klumperman, B. Macromolecules 2004, 37, 2383–2394. 59. McLeary, J. B.; Calitz, F. M.; McKenzie, J. M.; Tonge, M. P.; Sanderson, R. D.; Klumperman, B. Macromolecules 2005, 38, 3151–3161. 60. McLeary, J. B.; McKenzie, J. M.; Tonge, M. P.; Sanderson, R. D.; Klumperman, B. Chem. Commun. 2004, 1950–1951. 61. McLeary, J. B.; Tonge, M. P.; Klumperman, B. Macromol. Rapid. Commun. 2006, 27, 1233–1240. 62. Pound, G.; McLeary, J. B.; McKenzie, J. M.; Lange, R. F. M.; Klumperman, B. Macromolecules 2006, 39, 7796–7797. 63. van den Dungen, E. T. A.; Matahwa, H.; McLeary, J. B.; Sanderson, R. D.; Klumperman, B. J. Polym. Sci., Part A: Polym. Chem. 2008, 46, 2500–2509. 64. Moad, G.; Chong, Y. K.; Mulder, R.; Rizzardo, E.; San H. Thang, New Features of the Mechanism of RAFT Polymerization. In Controlled/Living Radical Polymerization. Progress in RAFT, NMP & OMRP; Matyjaszewski, K., Ed.; American Chemical Society: Washington, DC, 2009; Vol. 1024, pp 3−18. 65. Vandenbergh, J.; Reekmans, G.; Adriaensens, P.; Junkers, T. Chem. Science 2015, 6, 5753–5761. 101

66. Haven, J. J.; Vandenbergh, J.; Kurita, R.; Gruber, J.; Junkers, T. Polym. Chem. 2015, 6, 5752–5765. 67. Vandenbergh, J.; Reekmans, G.; Adriaensens, P.; Junkers, T. Chem. Commun. 2013, 49, 10358–10360. 68. Haven, J. J.; Vandenbergh, J.; Junkers, T. Chem. Commun. 2015, 51, 4611–4614. 69. Haven, J. J.; De Neve, J. A.; Junkers, T. ACS Macro Lett. 2017, 6, 743–747. 70. Braun, D.; Hu, F. Progress Polym. Sci. 2006, 31, 239–276. 71. Isahak, N.; Gody, G.; Malins, L. R.; Mitchell, N. J.; Payne, R. J.; Perrier, S. Chem. Commun. 2016, 52, 12952–12955. 72. Wang, R.; McCormick, C. L.; Lowe, A. B. Macromolecules 2005, 38, 9518–9525. 73. Ran, R.; Chen, Z.; Wang, X. L. J. Appl. Polym. Sci. 2009, 111, 2011–2017. 74. Wang, R.; Lowe, A. B. J. Polym. Sci., Part A: Polym. Chem. 2007, 45, 2468–2483. 75. Qi, G.; Eleazer, B.; Jones, C. W.; Schork, F. J. Macromolecules 2009, 42, 3906–3916. 76. Ouyang, L.; Wang, L.; Schork, F. J. Macromol. React. Eng. 2011, 5, 163–169. 77. Ouyang, L.; Wang, L.; Schork, F. J. Macromol. Chem. Phys. 2010, 211, 1977–1983. 78. Kuroki, A.; Martinez-Botella, I.; Hornung, C. H.; Martin, L.; Williams, E. G. L.; Locock, K. E. S.; Hartlieb, M.; Perrier, S. Polym. Chem. 2017, 8, 3249–3254. 79. Gody, G.; Maschmeyer, T.; Zetterlund, P. B.; Perrier, S. Macromolecules 2014, 47, 639–649. 80. Gody, G.; Maschmeyer, T.; Zetterlund, P. B.; Perrier, S. Nat. Commun.s 2013, 4, 3505. 81. Gody, G.; Maschmeyer, T.; Zetterlund, P. B.; Perrier, S. Macromolecules 2014, 47, 3451–3460. 82. Moad, G. Prog. Polym. Sci. 2017 submitted. 83. Otsu, T. J. Polym. Sci., Part A: Polym. Chem. 2000, 38, 2121–2136. 84. Xu, J.; Fu, C.; Shanmugam, S.; Hawker, C. J.; Moad, G.; Boyer, C. Angew. Chem., Int. Ed. Engl. 2017, 56, 8376–8383. 85. Hill, D. J. T.; Shao, L. Y.; Pomery, P. J.; Whittaker, A. K. Polymer 2001, 42, 4791–4802. 86. Bezdushna, E.; Ritter, H. Macromol. Rapid Commun. 2005, 26, 1087–1092. 87. Postma, A.; Davis, T. P.; Li, G.; Moad, G.; O’Shea, M. Macromolecules 2006, 39, 5307–5318. 88. Chernikova, E.; Yulusov, V.; Mineeva, K.; Golubev, V.; Garina, E. Polym. Sci., Ser. B 2011, 53, 437–447. 89. Stenzel, M. H.; Cummins, L.; Roberts, G. E.; Davis, T. R.; Vana, P.; BarnerKowollik, C. Macromol. Chem. Phys. 2003, 204, 1160–1168. 90. Shanmugam, S.; Xu, J.; Boyer, C. Macromolecules 2014, 47, 4930–4942. 91. Benaglia, M.; Chen, M.; Chong, Y. K.; Moad, G.; Rizzardo, E.; Thang, S. H. Macromolecules 2009, 42, 9384–9386. 102

92. Benaglia, M.; Chiefari, J.; Chong, Y. K.; Moad, G.; Rizzardo, E.; Thang, S. H. J. Am. Chem. Soc. 2009, 131, 6914–6915. 93. Keddie, D. J.; Guerrero-Sanchez, C.; Moad, G.; Mulder, R.; Rizzardo, E.; Thang, S. H. Macromolecules 2012, 45, 4205–4215. 94. Gardiner, J.; Martinez-Botella, I.; Kohl, T. M.; Krstina, J.; Moad, G.; Tyrell, J. H.; Coote, M. L.; Tsanaktsidis, J. Polym. Int. 2017, 66, 1438–1447. 95. Gardiner, J.; Martinez-Botella, I.; Tsanaktsidis, J.; Moad, G. Polym. Chem. 2016, 7, 481–492. 96. Porel, M.; Alabi, C. A. J. Am. Chem. Soc. 2014, 136, 13162–13165. 97. Porel, M.; Brown, J. S.; Alabi, C. A. Synlett 2015, 26, 565–571. 98. Porel, M.; Thornlow, D. N.; Phan, N. N.; Alabi, C. A. Nat. Chem. 2016, 8, 590–596. 99. Aerts, A.; Lewis, R. W.; Zhou, Y.; Malic, N.; Moad, G.; Postma, A. Macromol. Rapid. Commun. 2018 submitted. 100. Fu, C.; Huang, Z.; Hawker, C. J.; Moad, G.; Xu, J.; Boyer, C. Polym. Chem. 2017, 8, 4637–4643.

103

Chapter 5

Reducing the Hydrogen Atom Abstraction Efficiencies of Benzophenone-Based Photosensitive Alkoxyamines Jason C. Morris,1,2 Jean-Louis Clément,1 Yohann Guillaneuf,1 Steven E. Bottle,2 Kathryn Fairfull-Smith,2 and Didier Gigmes*,1 1Aix

Marseille Univ, CNRS, ICR UMR 7273, 13397 Marseille, France of Chemistry, Physics and Mechanical Engineering, Science and Engineering Faculty, Queensland University of Technology, 2 George Street, Brisbane City, 4000, Australia *E-mail: [email protected]. 2School

The work reported herein details the synthesis, as well as the photophysical and photochemical investigation, of novel benzophenone-based photosensitive alkoxyamines for potential application in nitroxide-mediated photopolymerization. As benzophenone has found widespread application in photosensitive alkoxyamines, serving as an intramolecular photosensitizer, the deleterious effects of competitive excited state hydrogen atom abstraction processes are examined. In order to reduce the efficiency of excited state hydrogen atom abstraction processes, electron donating methoxy and pyrrolidine substituents were incorporated para to benzophenone’s carbonyl moiety to modulate the excited state character of the substituted benzophenone motifs. Whilst the incorporation of electron donating substituents reduced the efficiency of excited state hydrogen atom abstraction processes, this did not translate into increased photo-dissociation efficiencies. Greater photo-dissociation efficiencies were obtained for benzophenone-based alkoxyamines lacking the electron donating functionalities. The excited state character of the benzophenone moiety is therefore a key parameter governing both the hydrogen atom abstraction and photo-dissociation efficiencies of benzophenone-based photosensitive alkoxyamines.

© 2018 American Chemical Society

Introduction Photopolymerization has widespread impact, ranging from long-established applications such as coatings, inks and adhesives, to more contemporary applications including microelectronics, laser direct imaging technology, 3D printing as well as dental and medical applications (1–6). Whilst photopolymerization has traditionally been achieved through the use of photoinitiators (2, 5, 6), controlled photopolymerization methodologies have become the subject of increasing research interest. Controlled photopolymerization methodologies combine the spatial and temporal control of photochemical processes, with the ability to reinitiate polymer chains to furnish defined polymer architectures. Similar to well-established thermally-based controlled polymerization methodologies, the fundamental strategy of controlled photopolymerization methodologies consists of avoiding the instantaneous initiation, propagation and termination steps by reversibly deactivating the growing polymeric chains. Following this approach of reversible deactivation, an increasing array of controlled photopolymerization methodologies have been investigated. Notably, cobalt-mediated radical polymerization (CMRP) (7, 8), tellurium-mediated radical polymerization (TERP) (9) iodine-transfer polymerization (ITP) (10), atom transfer radical polymerization (ATRP) (11–13) and reversible addition-fragmentation chain-transfer (RAFT) (14) have each emerged as efficient controlled photopolymerization methodologies. Nitroxide-mediated photopolymerization (NMP2) offers a number of advantages over the preceding controlled polymerization systems. Despite the inherent difficulties associated with the synthesis of various alkoxyamines, the generally low toxicity of nitroxides as well as the simple polymerization formulation, make NMP2 desirable to a range of applications. However, while the aforementioned controlled photopolymerization methodologies benefit from the addition of exogenous chromophoric species, to facilitate intermolecular photosensitization of their respective control agents, this approach has been met with significant limitations in NMP2. The incorporation of photosensitizers with alkoxyamines derived from nitroxides such as TEMPO (2,2,6,6-tetramethylpiperidin-1-yloxyl) and SG1 (N-tert-butyl-N-(1-diethylphosphono-2,2-dimethylpropyl)-N-oxyl) has revealed poor photo-dissociation efficiencies (15, 16), restricting the initiation and propagation of polymer chains required for a controlled NMP2 process. Consequently, the focus of NMP2 research has been directed toward the synthetic fusion of chromophore and alkoxyamine components (photosensitive alkoxyamines), to enhance the photosensitization efficiency of alkoxyamine moieties. Benzophenone is the archetypical chromophore for the investigation of a range of photochemical processes and has found widespread application in photosensitive alkoxyamines. Alkoxyamines derived from some of the most prominent nitroxides used to mediate thermal NMP (TEMPO (1) (17), TIPNO (2) (18) and SG1 (3) (19)) have each been investigated with pendant benzophenone motifs (Figure 1). 106

Figure 1. TEMPO- (1), TIPNO- (2) and SG1- (3) based photosensitive alkoxyamines bearing a benzophenone chromophore.

However, whilst incorporation of the benzophenone motif presents a convenient route toward the intramolecular photosensitization of alkoxyamine moieties, the benzophenone motif also presents the opportunity for detrimental excited state hydrogen atom transfer (HAT) processes (commonly referred to as Norrish type II processes) (20–24). Within a controlled NMP2 medium, the formation of HAT derived radical species would lead to disruption of the dynamic equilibrium between reactive (radical) and dormant (alkoxyamine) states (25). Furthermore, HAT induced photochemical reduction of the benzophenone motif removes the carbonyl moiety central to the photochemistry of benzophenone (20–24). The photochemically reduced species are therefore insensitive to the applied irradiation conditions, preventing reinitiation of macroalkoxyamines under UV irradiation. Benzophenone undergoes HAT from its excited triplet state, of n-π* character, with remarkable efficiency (ΦHAT ≈ 1.0) (20–24). The efficiency of excited state HAT processes is attributable to both the character and longevity of the excited triplet state (20–24). The promotion of a non-bonding electron to an anti-bonding π orbital results in the formation of a long-lived, electron deficient diradical-like excited triplet state (20–24). HAT processes allow the electron deficient diradical to complete an unpaired orbital, centered on the oxygen atom, generating a stabilized, sterically hindered ketyl radical (20–24). As the efficiency of HAT derived radical processes are related to the electron deficiency of the ketyl oxygen in the n-π* excited triplet state, modulation of the excited state character can impart a profound impact on benzophenone HAT reactivity (26–31). Incorporation of electron donating substituents para to benzophenone’s carbonyl moiety have been shown to lead to mixing, or inversion, of excited state character between n-π*, π-π* and charge transfer (CT) excited states (26–31). Importantly, for π-π* and CT excited states, the electron deficiency of the ketyl oxygen is significantly reduced, causing a concomitant reduction in HAT reactivity (26–32). Accordingly, in the interest of examining benzophenone-based photosensitive alkoxyamines with a lower propensity for HAT processes, benzophenone-based alkoxyamines possessing electron donating methoxy and pyrrolidine substituents para to benzophenone’s carbonyl moiety were envisioned for this work. 107

Discussion Thus far, research on NMP2 has focused on the optimization of alkoxyamines derived from, or analogous to, alkoxyamines commonly employed in thermally-based NMP (17–19, 33–35). The benefit of this line of research is to utilize alkoxyamines with known abilities to afford high control over thermally-based polymerizations. However, whilst both thermally and photochemically derived NMP processes utilize a similar control process, the excited state energy transfer efficiency and photochemical stability of the generated nitroxides are of critical importance. Nitroxide degradation can proceed through a variety of photochemical processes (36). Perhaps the most common photochemical degradation process involves photochemically induced α-cleavage, leading to the formation of nitrone and carbon centered radical species (37). In the case of open-chain nitroxides, α-cleavage leads to the separation of nitrone and carbon centered radical species, reducing the probability of recombination as the probability for cross coupling of radical species and further degradation processes increase. Cyclic nitroxides such as 2,2,6,6-tetramethylpiperidin-1-yloxyl (TEMPO) (4), 2,2,5,5-tetramethylpyrrolidin-1-yloxyl (PROXYL) (5) and 1,1,3,3tetramethylisoindolin-2-yloxyl (TMIO) (6) (Figure 2) are susceptible to similar photochemically induced α-cleavage processes (38).

Figure 2. Cyclic nitroxides: TEMPO (4), PROXYL (5) and TMIO (6). However, due to the proximity of the generated nitrone and carbon centered radical species following α-cleavage, recombination to regenerate the nitroxide moiety is enhanced (38). This effect has been suggested to be of greatest effect for TMIO (6) owing to the rigidity of the fused aryl system (39). Following α-cleavage, the generated nitrone and carbon centered radical species, derived from TMIO (6), are held in a fixed proximity, greatly enhancing the probability of recombination to regenerate the nitroxide moiety (39). In addition to providing increased photochemical stability, TMIO (6) is imbued with an in-built aryl system. Whilst photosensitive alkoxyamines derived from cyclic nitroxides such as TEMPO (4) and PROXYL (5) require a linker to join nitroxide and chromophore components. The in-built aryl system of TMIO (6) allows for synthetic expansion of the isoindoline aryl system through non-cleavable carbon frameworks. Synthetic expansion of the isoindoline system provides minimal separation between nitroxide and chromophore components, which due to the distance dependence of photosensitization, would be anticipated to improve the photosensitization efficiency of corresponding alkoxyamine moieties. Accordingly, photosensitive alkoxyamines formed through synthetic expansion of the isoindoline aryl system into substituted benzophenone motifs 108

forms the subject of investigation reported herein. Moreover, styrenic (7a, 8a and 9a), methacrylic (7b, 8b and 9b) and acrylic (7c, 8c and 9c) benzophenone-based alkoxyamines were investigated (Figure 3), to determine their relevance to the photopolymerization of a range of industrially and academically relevant monomers.

Figure 3. Benzophenone-based alkoxyamines employing electron donating substituents (7a-c - 9a-c).

Alkoxyamine Synthesis The key synthetic intermediate (10), from which divergent and parallel syntheses were proposed to furnish the fused benzophenone-based alkoxyamines (7a-c - 9a-c), was prepared through known reaction conditions affording 5-bromo-2-methoxy-1,1,3,3-tetramethylisoindoline (10) with an overall yield of 21% over 5 steps (40–43). Following formation of the brominated isoindoline motif (10), lithiumhalogen exchange, followed by quenching with appropriately substituted benzaldehydes, provided the carbon framework of the fused isoindolinechromophore systems (44) (11 - 13) (78% - 95%) (Scheme 1). Subsequent oxidation of the generated diphenylmethanolic cores (11 - 13), with pyridinium chlorochromate, then furnished the desired benzophenone motifs (44) (14 - 16) in good yields (88 - 94%). Demethylation of the methoxyamine protecting groups (14 - 16) with m-CPBA, via a cope-type elimination of formaldehyde (45), was then undertaken to unveil the corresponding benzophenone-based nitroxides (17 19) (89% - 92%). Circumventing the incompatibility of the pyrrolidine substituent with m-CPBA, formation of the brominated benzophenone-based nitroxide (19) permitted incorporation of the pyrrolidine substituent under Buchwald-Hartwig reaction conditions (20) (82%) (Scheme 1). The target benzophenone-based alkoxyamines (7a-c - 9a-c) were then prepared from radicals derived from (1-bromoethyl)benzene, ethyl α-bromoisobutyrate and ethyl 2-bromopropionate using a Cu mediated ATRA system (Scheme 1). Under deoxygenated conditions, reaction of the prepared benzophenone-based nitroxides (17, 18 and 20) with the generated alkyl radicals, furnished the target styrenic (7a, 8a and 9a), methacrylic (7b, 8b and 9b) and acrylic (7c, 8c and 9c) benzophenone-based alkoxyamines in good yield (86 94%) (Scheme 1). 109

Scheme 1. Synthetic Approach Employed toward Benzophenone-Based Photosensitive Alkoxyamines (7a-c - 9a-c)

Photophysical Investigations The investigated chromophores were selected for their intrinsic photophysical and photochemical properties. To determine whether incorporation of the isoindoline motif led to a deviation of the photophysical properties, relative to their parent chromophores, from which a deviation in the photochemical properties may be inferred, the photophysical properties of the benzophenone-based nitroxides (17, 18 and 20) and corresponding alkoxyamines (7a-c - 9a-c) were investigated. The UV absorbance spectra demonstrate the lowest excited singlet state electronic transitions of each of the examined benzophenone-based nitroxides (17, 18 and 20) in polar (acetonitrile) and non-polar (tert-butylbenzene) solutions (Figure 4). The observed hypsochromic shift in the absorbance spectra of the unsubstituted benzophenone-based nitroxide (17) with increasing solvent polarity is consistent with the assignment of an n-π* transition as the lowest excited singlet state. Moreover, the absorbance maxima, vibrational structure, molar extinction coefficient and assigned excited state character of the unsubstituted 110

benzophenone-based nitroxide (17) were consistent with the parent benzophenone chromophore (46).

Figure 4. UV-visible absorbance spectra overlay of unsubstituted (17), methoxy substituted (18) and pyrrolidine substituted (20) benzophenone-based nitroxides in acetonitrile (17 = 2.7 × 10-4 mol.L-1, 18 = 8.6 × 10-5 mol.L-1, 20 = 2.1 × 10-6 mol.L-1) and tert-butylbenzene (17 = 3.7 × 10-4 mol.L-1, 18 = 6.1 × 10-5 mol.L-1, 20 = 2.8 × 10-6 mol.L-1) solvents. (see color insert)

A bathochromic shift in the second excited singlet state transition (vertical region) was observed in the absorbance spectra of the methoxy substituted benzophenone-based nitroxide (18) with increasing solvent polarity (Figure 4). The observed bathochromic shift, consistent with the expected π-π* transition, led to merging of the lowest and second excited singlet state electronic transitions. Merging of the lowest and second excited singlet state electronic transitions masked a potential solvatochromic shift of the lowest excited singlet state transition. While the absorbance maxima, vibrational structure and molar extinction coefficient (Table 1) appear consistent with the assignment of an n-π* lowest excited singlet state (Figure 4), the absence of observable solvatochromism prevents reliable determination of the lowest excited singlet state character of the methoxy substituted benzophenone-based nitroxide (18). The pyrrolidine substituted benzophenone-based nitroxide (20) displayed a bathochromic shift in the lowest excited singlet state transition with increasing solvent polarity, consistent with the assignment of a π-π* electronic transition. Furthermore, the high molar extinction coefficient obtained (Table 1) supports the π-π* transition being of charge transfer character, consistent with the incorporation of the strongly electron donating pyrrolidine substituent (47). The UV absorbance properties of the benzophenone-based alkoxyamines (7a-c - 9a-c) were then examined to determine whether incorporation of the alkoxyamine moieties led to deviation of the photophysical properties relative to their corresponding nitroxides (17, 18 and 20) (Figure 5). 111

The photophysical properties of the unsubstituted (7a-c) and methoxy substituted (8a-c) benzophenone-based alkoxyamines were consistent with their corresponding nitroxides (17 and 18), though shifts in the tail regions were observed. A similar shift in the absorbance spectra was observed for the pyrrolidine substituted benzophenone-based alkoxyamines (9a-c), relative to their corresponding nitroxide (20). Though notably, the shifts were more pronounced, leading to shifts in their absorbance maxima, relative to their corresponding nitroxide (20). However, the characteristically high molar extinction coefficient of the pyrrolidine substituted benzophenone-based nitroxide (20) was retained in the corresponding alkoxyamines (9a-c) (Table 1), consistent with a lowest singlet excited state of CT character.

Figure 5. UV-visible absorbance spectra overlay of: (a) the unsubstituted benzophenone-based nitroxide (17) and corresponding alkoxyamines (7a-c) in tert-butylbenzene (17 = 3.7 × 10-4 mol.L-1, 7a = 2.5 × 10-4 mol.L-1, 7b = 2.4 × 10-4 mol.L-1, 7c = 2.5 × 10-4 mol.L-1) (b) methoxy substituted benzophenone-based nitroxide (18) and corresponding alkoxyamines (8a-c) in tert-butylbenzene (18 = 9.8 × 10-5 mol.L-1, 8a = 1.7 × 10-4 mol.L-1, 8b = 1.6 × 10-4 mol.L-1, 8c = 2.0 × 10-4 mol.L-1) (c) pyrrolidine substituted benzophenone-based nitroxide (20) and corresponding alkoxyamines (9a-c) in tert-butylbenzene (20 = 2.0 × 10-6 mol.L-1, 9a = 1.8 × 10-6 mol.L-1, 9b = 1.7 × 10-6 mol.L-1, 9c = 2.1 × 10-6 mol.L-1). (see color insert) 112

Accordingly, despite small shifts in the absorbance spectra of the examined unsubstituted (7a-c), methoxy substituted (8a-c) and pyrrolidine substituted (9a-c) benzophenone-based alkoxyamines, the photophysical properties of the examined alkoxyamines (7a-c - 9a-c) were consistent with their corresponding nitroxides (17, 18 and 20).

Table 1. Photophysical Properties of the Examined Benzophenone-Based Systems Compound

a

tert-Butylbenzene

Acetonitrile

λmax (nm)

εa (mol.L-1.cm-1)

λmax (nm)

εa (mol.L-1.cm-1)

17

345

210

338

324

7a

345

210

-

-

7b

345

202

-

-

7c

345

191

-

-

18

338

626

338

571

8a

338

416

-

-

8b

338

450

-

-

8c

338

394

-

-

20

350

31288

355

41466

9a

344

35376

-

-

9b

344

33067

-

-

9c

344

27825

-

-

Molar extinction coefficients calculated at λmax as indicated

Photochemical Investigations The photo-dissociation efficiencies of the examined alkoxyamines were investigated by EPR spectroscopy. Initial investigations focused on determining the photo-dissociation efficiencies of the styrenic benzophenone-based alkoxyamines (7a - 9a) in order to optimize the irradiation conditions to each chromophore before assessing the methacrylic (7b - 9b) and acrylic (7c - 9c) benzophenone-based alkoxyamines. Under broad range UV irradiation (UVA, UVB and UVC) the styrenic unsubstituted (7a), methoxy substituted (8a) and pyrrolidine substituted (9a) benzophenone-based alkoxyamines reached 22%, 15% and 7% nitroxide recoveries respectively (Figure 6). In each case, after reaching maximum nitroxide recoveries, the nitroxide recoveries decreased over 113

time. Notably, the decrease in nitroxide recovery was most pronounced for the unsubstituted (7a) and methoxy substituted (8a) benzophenone-based systems.

Figure 6. Nitroxide recoveries obtained for the styrenic unsubstituted (7a), methoxy substituted (8a) and pyrrolidine substituted (9a) benzophenone-based alkoxyamines under broad range UV irradiation (UVA, UVB and UVC). (see color insert)

Figure 7. Nitroxide recoveries obtained for the styrenic unsubstituted (7a), methoxy substituted (8a) and pyrrolidine substituted (9a) benzophenone-based alkoxyamines under UVA irradiation. (see color insert) Under UVA irradiation similar nitroxide recoveries were observed for the styrenic unsubstituted (7a) (25%) and pyrrolidine substituted (9a) (7%) benzophenone-based alkoxyamines, though over increased durations (Figure 7). However, the nitroxide recovery of the styrenic methoxy substituted benzophenone-based alkoxyamine (8a) was reduced, reaching 8% nitroxide recovery (Figure 7). Similar to the trend observed under broad range UV 114

irradiation (UVA, UVB and UVC) (Figure 6), after reaching maximum nitroxide recoveries, the nitroxide recoveries decreased over time, with the decrease most pronounced for the unsubstituted (7a) and methoxy substituted (8a) benzophenone-based systems (Figure 7). Similar photo-dissociation studies were applied to the methacrylic (7b 9b) and acrylic (7c - 9c) benzophenone-based alkoxyamines. However, in each case, no discernible increase in the nitroxide recovery was observed under UVA irradiation (Figure 8).

Figure 8. Nitroxide recoveries obtained for the methacrylic and acrylic unsubstituted (7b and 7c), methoxy substituted (8b and 8c) and pyrrolidine substituted (9b and 9c) benzophenone-based alkoxyamines under UVA irradiation. (see color insert) To investigate whether the low nitroxide recoveries obtained could be attributed to photochemical decomposition of the generated nitroxide moieties, solutions of the benzophenone-based nitroxides (17, 18 and 20) in tert-butylbenzene were subjected to UVA irradiation and monitored by EPR spectroscopy. As oxygen quenching of excited states could potentially impact the photochemical stability of the nitroxide (48), both oxygenated and deoxygenated solutions of the benzophenone-based nitroxides (17, 18 and 20) in tert-butylbenzene were investigated. Despite an initial decrease in the stabilities of the unsubstituted (17) and methoxy substituted (18) benzophenone-based nitroxides under oxygenated and deoxygenated conditions, the photochemical stabilities remained constant, with a 5% reduction in the stability of both the unsubstituted (17) and methoxy substituted (18) benzophenone-based nitroxides observed after 5 hours of UVA irradiation (Figure 9). Accordingly, the photochemical stabilities of the unsubstituted (17) and methoxy substituted (18) benzophenone-based nitroxides, under oxygenated and deoxygenated conditions, indicate that photochemical decomposition of the nitroxide moieties could not account for the low nitroxide recoveries obtained. 115

Figure 9. Nitroxide stabilities obtained under oxygenated and deoxygenated conditions for the unsubstituted (17) and methoxy substituted (18) benzophenone-based nitroxides under UVA irradiation. (see color insert)

In contrast to the high photochemical stabilities of the unsubstituted (17) and methoxy substituted (18) benzophenone-based nitroxides under oxygenated and deoxygenated conditions, the photochemical stability of the pyrrolidine substituted benzophenone-based nitroxide (20) was significantly reduced, particularly in the absence of oxygen (Figure 10).

Figure 10. Nitroxide stabilities obtained under oxygenated and deoxygenated conditions for the pyrrolidine substituted (20) benzophenone-based nitroxide under UVA irradiation. (see color insert) 116

As oxygen is a known quencher of excited triplet states (48), the increased rate of photochemical decomposition of 20 in the absence of oxygen indicates that degradation may proceed through the excited triplet state. The reduction in concentration of the pyrrolidine substituted benzophenone-based nitroxide (20) (Figure 10) may be attributed to single electron transfer (SET) from the incorporated pyrrolidine substituted to the nitroxide moiety, leading to the formation of the corresponding hydroxylamine anion and pyrrolidine radical cation. Alternatively, given the known propensity for the excited triplet state benzophenone motif to undergo HAT processes (20–24), HAT could potentially account for the reduced photochemical stability of the pyrrolidine substituted benzophenone-based nitroxide (20) relative to the unsubstituted (17) and methoxy substituted (18) benzophenone-based nitroxides. The pyrrolidine substituent was incorporated to lower the quantum yield of HAT (26–32). However, a reduction in the quantum yield of HAT is potentially offset by the significantly enhanced molar absorptivity of the pyrrolidine substituted benzophenone-based nitroxide (20) with respect to the unsubstituted (17) and methoxy substituted (18) benzophenone-based nitroxides (128-fold and 73-fold respectively) (Table 1). Furthermore, as the photochemical stability of the pyrrolidine substituted benzophenone-based nitroxide (20) was assessed in tert-butylbenzene, a poor hydrogen atom donor, the hydrogen atoms of the incorporated pyrrolidine substituent present the most labile hydrogen atom donors which may be abstracted, leading to the formation of non-radical products (6). HAT from the styrenic (7a, 8a, and 9a), methacrylic (7b, 8b and 9b) and acrylic (7c, 8c and 9c) alkoxyamine moieties to the excited state carbonyl of the benzophenone motif may account for the low nitroxide recoveries obtained. Reaction of HAT derived radical species with the nitroxide may furnish photochemically insensitive alkoxyamines. Additionally, HAT induced photochemical reduction of the benzophenone motif would remove the carbonyl moiety, rendering the photochemically reduced alkoxyamines insensitive to the applied irradiation conditions (20–24). Accordingly, to assess the potential for HAT, cyclohexane, a more labile hydrogen atom donor than the previously employed tert-butylbenzene, was used as the solvent (49–52). To determine if HAT processes could account for the low nitroxide recoveries obtained for the benzophenone-based alkoxyamines (7a-c - 9a-c), deoxygenated solutions of the benzophenone-based nitroxides (17, 18 and 20) in cyclohexane were investigated under UVA irradiation to facilitate monitoring of HAT processes via EPR spectroscopy (Figure 11). In each case, a reduction in the concentration of the nitroxide was observed over time, consistent with the involvement of HAT processes. Upon cessation of UVA irradiation and introduction of air into the EPR sample, the nitroxide recovery increased. Regeneration of the nitroxide is consistent with the presence of hydroxylamine, which is known to be readily oxidized to the corresponding nitroxide by molecular oxygen (53, 54). The formation of hydroxylamine species derived from the benzophenonebased nitroxides (17, 18 and 20) is consistent with the involvement of HAT processes. HAT from cyclohexane to the excited state benzophenone moiety would lead to the formation of ketyl radicals through the well-established pinacol 117

chemistry (20–24). Addition of the nitroxide to the benzophenone-derived ketyl radical to generate the corresponding hemiketal, followed by elimination of the hydroxylamine anion, and subsequent proton transfer, could potentially account for the formation of hydroxylamine (Scheme 2, a). Moreover, this process could account for the apparent bimodal decay of the nitroxides (17, 18 and 20) (Figure 11). Quenching of the excited state benzophenone motif through electron exchange with the unpaired spin of the nitroxide, could explain the initially slow reduction of the nitroxide signal. Increasing removal of the nitroxide moiety, through formation of hydroxylamine or reaction with the generated cyclohexyl radicals, would lead to increasing benzophenone HAT reactivity, thereby accounting for the increased rate of reduction of the nitroxide signals over time.

Figure 11. Nitroxide stabilities in cyclohexane obtained under deoxygenated conditions for unsubstituted (17), methoxy substituted (18) and pyrrolidine substituted (20) benzophenone-based nitroxides under UVA irradiation followed by cessation of UVA irradiation and introduction of air. (see color insert)

An alternative explanation for the presence of hydroxylamine species derived from the benzophenone-based nitroxides (17, 18 and 20) could be the involvement of excited state energy transfer from the excited triplet state of the benzophenone moiety to the nitroxide doublet state. Energy transfer from the benzophenone motif to the nitroxide moiety could potentially produce an excited state nitroxide capable of HAT from cyclohexane (Scheme 2, b) (55). This process may also account for the apparent bimodal decay of the benzophenone-based nitroxides (17, 18 and 20) (Figure 11). Inefficient HAT from cyclohexane to an excited state nitroxide moiety could account for the initially slow reduction of the nitroxide signal. HAT to the nitroxide moiety would remove the unpaired spin of the nitroxide, resulting in the formation of hydroxylamine and cyclohexyl radicals. Upon removal of the unpaired spin of the nitroxide, through formation of a hydroxylamine or coupling with the 118

generated cyclohexyl radicals, the established benzophenone HAT reactivity would be restored. Normal benzophenone HAT processes could then account for the increased rate of reduction of the nitroxide signal over time.

Scheme 2. Proposed Formation of Hydroxylamine Involving a) Hemiketal Formation b) HAT Involving an Excited State Nitroxide Moiety

Conclusions Novel unsubstituted (7a-c), methoxy substituted (8a-c) and pyrrolidine substituted (9a-c) benzophenone-based photosensitive alkoxyamines were prepared in order to investigate their relevance to NMP2. Substitution of the benzophenone motif with electron donating substituents was undertaken in order to address the established HAT photochemistry of the benzophenone motif. UV-visible spectroscopy revealed that the unsubstituted benzophenone-based nitroxide (17) and corresponding alkoxyamines (7a-c) maintained lowest n-π* excited singlet states, characteristic of the parent benzophenone chromophore. However, incorporation of the strongly electron donating pyrrolidine substituent (20 and 9a-c) led to the formation of an excited singlet state of CT character. Consistent with previous reports of reduced HAT 119

reactivity of CT excited states of substituted benzophenones, photochemical analyses, in the presence of sufficiently labile HAT donors, revealed a reduction in the rate of HAT for the pyrrolidine substituted benzophenone-based nitroxide (20), relative to the unsubstituted (17) and methoxy substituted (18) benzophenone-based nitroxides. However, reduced involvement of HAT processes did not translate into increased photo-dissociation efficiencies. On the contrary, electron donating substituents reduced the photo-dissociation efficiency of the benzophenone-based alkoxyamines. Accordingly, the electronic distribution of the excited state benzophenone motif is a key parameter governing both the HAT reactivity and photo-dissociation efficiency of benzophenone-based photosensitive alkoxyamines. The nitroxide recovery obtained for the styrenic unsubstituted benzophenonebased alkoxyamine (7a) is an improvement over preceding photosensitive alkoxyamines employing pendant benzophenone chromophores. Combined with the high nitroxide stability, in the absence of sufficiently labile HAT donors, these results indicate that photosensitive alkoxyamines derived from the isoindoline class of nitroxide may possess a structural advantage over other classes of nitroxides. However, the involvement of HAT processes indicates that alternative chromophores should be investigated in order to avoid detrimental competitive excited state processes.

Experimental All reagents were purchased from Sigma Aldrich and used without further purification. Anhydrous solvents were dried over sodium. Reactions were followed by thin layer chromatography (Merck Silica Gel 60 F254). Reactions were purified by silica gel column chromatography (Silica gel 60 Å (230 - 400 mesh)). 1H NMR analyses were conducted at 400 MHz and 13C NMR analyses conducted at 100 MHz. Chemical shifts (δ) for 1H NMR and 13C NMR analyses, conducted in deuterated chloroform, are reported in ppm relative to their solvent residual peaks: proton (δ = 7.26 ppm) and carbon (δ = 77.16 ppm). Multiplicity is indicated as follows: s (singlet); d (doublet); t (triplet); m (multiplet); dd (doublet of doublet); br s (broad singlet). Coupling constants are reported in Hertz (Hz). Mass spectrometry analyses were conducted with electrospray as the ionization technique. Infrared spectroscopy analyses were conducted as neat samples using a Nicolet 870 Nexus Fourier Transform infrared spectrometer equipped with a DTGS TEC detector and an Attenuated Total Reflectance (ATR) accessory (Nicolet Instrument Corp., Madison, WI) using a Smart Endurance single reflection ATR accessory equipped with a composite diamond IRE with a 0.75 mm2 sampling surface and a ZnSe focusing element. Analytical HPLC were performed on a Hewlett Packard 1100 series HPLC, using an Agilent prep-C18 scalar column (10 μm, 4.6 × 150 mm) at a flow rate of 1 mL/min, purity determined by UV detection of absorbing species at 254 nm. All UV-visible spectra were recorded on a single beam Varian Cary 50 UV-visible spectrophotometer. Melting 120

points were measured on a Gallenkamp variable temperature apparatus. EPR analyses were performed with Magnettech Miniscope, Bruker EMX and Bruker ELEXSYS EPR spectrometers. Photo-dissociation analyses were performed with a Hamamatsu LC8 UV lamp as the light source.

General Procedure for EPR Photo-Dissociation Analyses Following the accurate weighing of alkoxyamine (5 decimal point balance), a 0.0001 M solution is made up in tert-butylbenzene. To confirm the concentration of alkoxyamine present, a thermal EPR experiment is performed (0.0001 M solution of alkoxyamine in tert-butylbenzene). Photo-dissociation analyses are performed under direct irradiation within the ELEXSYS EPR spectrometer with a Hamamatsu LC8 UV lamp as the light source. A baseline noise correction is performed for each experiment, reducing the EPR intensity error (< 2%). The photo-dissociation efficiencies are then calculated in relation to the expected recoveries determined through thermal EPR experiments.

5-(Hydroxy(phenyl)methyl)-2-methoxy-1,1,3,3-tetramethylisoindoline (11) A solution of 5-bromo-2-methoxy-1,1,3,3-tetramethylisoindoline (10) (1.00 g, 3.52 mmol, 1.0 eq.) in anhydrous THF (50 mL) was cooled to - 78 °C and placed under an inert atmosphere of argon. To this solution was added n-butyllithium (1.6 M in n-hexanes) (2.40 mL, 3.87 mmol, 1.1 eq.) dropwise. The reaction was allowed to stir for 15 minutes, maintaining a constant temperature of reaction (~ - 78 °C). A solution of benzaldehyde (1.12 g, 10.56 mmol, 3.0 eq.) in anhydrous THF (50 mL) was added dropwise, maintaining a constant temperature of reaction (~ - 78 °C), until the addition was complete. The reaction was allowed to return to room temperature over two hours and then quenched by the addition of water. The resulting mixture was extracted with dichloromethane and the combined organic extracts washed with a saturated solution of brine, dried over anhydrous sodium sulphate and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : dichloromethane; 1 : 19) afforded 5-(hydroxy(phenyl)methyl)-2-methoxy-1,1,3,3-tetramethylisoindoline (11) as a clear oil (1.04 g, 95% yield). Rf = 0.40, diethyl ether : dichloromethane, 1 : 19. IR (ATR) νmax = 3359 (OH), 3027 (=CH), 2972 and 2933 (–CH), 1618, 1603, 1587, 1492 and 1453 (C=C) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.40 - 7.33 (m, 4H, Harom), 7.29 - 7.26 (m, 1H, Harom), 7.19 (dd, J = 7.6, 1.6 Hz, 1H, Harom), 7.17 - 7.16 (m, 1H, Harom), 7.04 (d, J = 7.6 Hz, 1H, Harom), 5.84 - 5.83 (m, 1H, CH), 3.78 (s, 3H, OCH3), 2.23 - 2.22 (m, 1H, OH), 1.42 (br s, 12H, 4×CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 145.7, 144.8, 144.0, 143.1, 128.6, 127.7, 126.7, 125.9, 121.7, 119.8, 76.6, 67.3, 67.0, 65.6, 29.8, 25.4. HRMS: m/z calculated for C20H26NO2 [M+H]+ 312.1958; found 312.1951. Purity > 95% as determined by analytical HPLC. These data are consistent with that previously reported by Morris et al. (44) 121

5-(Hydroxy(4-methoxyphenyl)methyl)-2-methoxy-1,1,3,3tetramethylisoindoline (12) A solution of 5-bromo-2-methoxy-1,1,3,3-tetramethylisoindoline (10) (1.02 g, 3.59 mmol, 1.0 eq.) in anhydrous THF (50 mL) was cooled to - 78 °C and placed under an inert atmosphere of argon. To this solution was added n-butyllithium (1.6 M in n-hexanes) (2.47 mL, 3.95 mmol, 1.1 eq.) dropwise. The reaction was allowed to stir for 15 minutes, maintaining a constant temperature of reaction (~ - 78 °C). A solution of 4-methoxybenzaldehyde (1.47 g, 10.77 mmol, 3.0 eq.) in anhydrous THF (50 mL) was added dropwise, maintaining a constant temperature of reaction (~ - 78 °C), until the addition was complete. The reaction was allowed to return to room temperature over two hours and then quenched by the addition of water. The resulting mixture was extracted with dichloromethane and the combined organic extracts washed with a saturated solution of brine, dried over anhydrous sodium sulphate and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : dichloromethane; 1 : 19) afforded 5-(hydroxy(4-methoxyphenyl)methyl)-2-methoxy-1,1,3,3-tetramethylisoindoline (12) as a clear oil (1.12 g, 91% yield). Rf = 0.35, diethyl ether : dichloromethane, 1 : 19. IR (ATR) νmax = 3409 (OH), 3031 (=CH), 2972 and 2933 (–CH), 1611, 1585, 1491 and 1441 (C=C), 1245 and 1034 (=C–O) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.31 - 7.28 (m, 2H, Harom), 7.19 - 7.16 (m, 2H, Harom), 7.04 (d, J = 7.6 Hz, 1H, Harom), 6.89 - 6.86 (m, 2H, Harom), 5.80 - 5.79 (m, 1H, CH), 3.80 - 3.78 (m, 6H, 2×OCH3), 2.19 - 2.18 (m, 1H, OH), 1.42 (s, 12H, 4×CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 159.2, 145.6, 144.7, 143.3, 136.3, 128.0, 125.8, 121.6, 119.6, 114.0, 76.1, 67.3, 67.0, 65.6, 55.4, 29.5, 25.2. HRMS: m/z calculated for C21H28NO3 [M+H]+ 342.2064; found 342.2055. Purity > 95% as determined by analytical HPLC. 5-((4-Bromophenyl)(hydroxy)methyl)-2-methoxy-1,1,3,3tetramethylisoindoline (13) A solution of 5-bromo-2-methoxy-1,1,3,3-tetramethylisoindoline (10) (1.03 g, 3.62 mmol, 1.0 eq.) in anhydrous THF (50 mL) was cooled to - 78 °C and placed under an inert atmosphere of argon. To this solution was added n-butyllithium (1.6 M in n-hexanes) (2.49 mL, 3.99 mmol, 1.1 eq.) dropwise. The reaction was allowed to stir for 15 minutes, maintaining a constant temperature of reaction (~ - 78 °C). A solution of 4-bromobenzaldehyde (2.01 g, 10.86 mmol, 3.0 eq.) in anhydrous THF (50 mL) was added dropwise, maintaining a constant temperature of reaction (~ - 78 °C), until the addition was complete. The reaction was allowed to return to room temperature over two hours and then quenched by the addition of water. The resulting mixture was extracted with dichloromethane and the combined organic extracts washed with a saturated solution of brine, dried over anhydrous sodium sulphate and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : dichloromethane; 1 : 19) afforded 5-((4-bromophenyl)(hydroxy)methyl)-2-methoxy-1,1,3,3-tetramethylisoindoline (13) as a clear oil (1.10 g, 78% yield). Rf = 0.45, diethyl ether : dichloromethane, 1 : 19. IR (ATR) νmax = 3384 (OH), 2973 and 2932 (–CH), 1616, 1591, 1573, 122

1486 and 1460 (C=C), 1070 and 1009 (=C–Br) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.48 - 7.45 (m, 2H, Harom), 7.27 - 7.24 (m, 2H, Harom), 7.15 (dd, J = 7.6, 1.2 Hz, 1H, Harom), 7.11 (s, 1H, Harom), 7.05 (d, J = 7.6 Hz, 1H, Harom), 5.77 - 5.76 (m, 1H, CH), 3.78 (s, 3H, OCH3), 2.34 - 2.33 (m, 1H, OH), 1.42 (s, 12H, 4×CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 145.9, 145.2, 142.9, 142.7, 131.7, 128.4, 125.9, 121.9, 121.5, 119.8, 76.0, 67.3, 67.1, 65.7. HRMS: m/z calculated for C20H2579BrNO2 [M+H]+ 390.1063; found 390.1061. Purity > 95% as determined by analytical HPLC. 5-Benzoyl-2-methoxy-1,1,3,3-tetramethylisoindoline (14) To a solution of 5-(hydroxy(phenyl)methyl)-2-methoxy-1,1,3,3-tetramethylisoindoline (11) (910 mg, 2.92 mmol, 1.0 eq.) in dichloromethane (100 mL) was added pyridinium chlorochromate (756 mg, 3.51 mmol, 1.2 eq.). The reaction was stirred for 2.5 hours at room temperature. The reaction mixture was filtered through celite, eluting with dichloromethane, and concentrated in vacuo. Purification via silica gel column chromatography (dichloromethane) afforded 5-benzoyl-2-methoxy-1,1,3,3-tetramethylisoindoline (14) as a white solid after extensive concentration in vacuo (850 mg, 94% yield). Rf = 0.38, dichloromethane. M.p. 69 - 70 °C. IR (ATR) νmax = 3056 (=CH), 2977 and 2935 (–CH), 1655 (C=O), 1612, 1597, 1576, 1484 and 1445 (C=C) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.80 (d, J = 7.6 Hz, 2H, Harom), 7.66 (d, J = 8.0 Hz, 1H, Harom), 7.61 - 7.57 (m, 2H, Harom), 7.51 - 7.47 (m, 2H, Harom), 7.18 (d, J = 8.0 Hz, 1H, Harom), 3.80 (s, 3H, OCH3), 1.47 (s, 12H, 4×CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 196.7, 150.3, 145.8, 138.0, 137.0, 132.4, 130.1, 130.0, 128.4, 123.5, 121.4, 67.4, 67.2, 65.7, 29.8, 25.0. HRMS: m/z calculated for C20H24NO2 [M+H]+ 310.1802; found 310.1789. Purity > 95% as determined by analytical HPLC. These data are consistent with that previously reported by Morris et al.44 5-(4-Methoxybenzoyl)-2-methoxy-1,1,3,3-tetramethylisoindoline (15) To a solution of 5-(hydroxy(4-methoxyphenyl)methyl)-2-methoxy-1,1,3,3tetramethylisoindoline (12) (718 mg, 2.10 mmol, 1.0 eq.) in dichloromethane (100 mL) was added pyridinium chlorochromate (544 mg, 2.52 mmol, 1.2 eq.). The reaction was stirred for 2.5 hours at room temperature. The reaction mixture was filtered through celite, eluting with dichloromethane, and concentrated in vacuo. Purification via silica gel column chromatography (dichloromethane) afforded 5-(4-methoxybenzoyl)-2-methoxy-1,1,3,3-tetramethylisoindoline (15) as a white solid (656 mg, 92% yield). Rf = 0.30, dichloromethane. M.p. 80 82 °C. IR (ATR) νmax = 3050 (=CH), 2973 and 2933 (–CH), 1654 (C=O), 1612, 1599, 1574, 1508 and 1463 (C=C), 1248 and 1030 (=C–O) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.82 (d, J = 8.4 Hz, 2H, Harom), 7.62 (d, J = 8.0 Hz, 1H, Harom), 7.53 (s, 1H, Harom), 7.18 (d, J = 8.0 Hz, 1H, Harom), 6.97 (d, J = 8.0 Hz, 2H, Harom), 3.90 (s, 3H, OCH3), 3.80 (s, 3H, OCH3), 1.47 (s, 12H, 4×CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 195.5, 163.2, 149.7, 145.6, 137.7, 132.6, 130.5, 129.6, 123.3, 121.3, 113.6, 67.3, 67.2, 65.6, 55.6, 29.6, 25.0. HRMS: m/z 123

calculated for C21H25NO3Na [M+Na]+ 362.1727; found 362.1724. Purity > 95% as determined by analytical HPLC.

5-(4-Bromobenzoyl)-2-methoxy-1,1,3,3-tetramethylisoindoline (16) To a solution of 5-((4-bromophenyl)(hydroxy)methyl)-2-methoxy-1,1,3,3tetramethylisoindoline (13) (1.03 g, 2.64 mmol, 1.0 eq.) in dichloromethane (100 mL) was added pyridinium chlorochromate (0.68 g, 3.17 mmol, 1.2 eq.). The reaction was stirred for 2.5 hours at room temperature. The reaction mixture was filtered through celite, eluting with dichloromethane, and concentrated in vacuo. Purification via silica gel column chromatography (dichloromethane) afforded 5-(4-bromobenzoyl)-2-methoxy-1,1,3,3-tetramethylisoindoline (16) as a white solid (0.90 g, 88% yield). Rf = 0.50, dichloromethane. M.p. 76 - 78 °C. IR (ATR) νmax = 3070 (=CH), 2971 and 2931 (–CH), 1654 (C=O), 1613, 1584, 1565, 1483 and 1460 (C=C), 1068 and 1012 (=C–Br) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.68 - 7.61 (m, 5H, Harom), 7.55 (d, J = 1.2 Hz, 1H, Harom), 7.19 (dd, J = 7.6, 0.4 Hz, 1H, Harom), 3.79 (s, 3H, OCH3), 1.46 (br s, 12H, 4×CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 195.6, 150.7, 146.0, 136.8, 136.6, 131.7, 131.6, 129.9, 127.5, 123.4, 121.6, 67.4, 67.2, 65.7, 29.7, 25.6. HRMS: m/z calculated for C20H2279BrNO2Na [M+Na]+ 410.0726; found 410.0725. Purity > 95% as determined by analytical HPLC.

5-Benzoyl-1,1,3,3-tetramethylisoindolin-2-yloxyl (17) To a solution of 5-benzoyl-2-methoxy-1,1,3,3-tetramethylisoindoline (14) (600 mg, 1.94 mmol, 1.0 eq.) in dichloromethane (100 mL) was added m-CPBA (77% purity) (957 mg, 4.27 mmol, 2.2 eq.) in portions. The reaction was stirred for one hour at room temperature. Sodium hydroxide (2M, 100 mL) was added and the resulting mixture stirred vigorously for 15 minutes. The reaction mixture was extracted with dichloromethane. The combined organic extracts were washed with water followed by a saturated solution of brine, dried over anhydrous sodium sulphate and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : dichloromethane; 1 : 19) followed by recrystallization from methanol afforded 5-benzoyl-1,1,3,3-tetramethylisoindolin-2-yloxyl (17) as small orange needles (526 mg, 92% yield). Rf = 0.35, diethyl ether : dichloromethane, 1 : 19. M.p. 149 - 150 °C. IR (ATR) νmax = 3040 (=CH), 2973 and 2928 (–CH), 1652 (C=O), 1613, 1597, 1577, 1488 and 1449 (C=C), 1430 (NO•) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.86 (br s, 2H, Harom), 7.67 - 7.63 (m, 1H, Harom), 7.53 (br s, 2H, Harom). HRMS: m/z calculated for C19H20NO2Na [M+Na]+ 317.1386; found 317.1384. EPR (DCM): αN = 1.424 mT. Purity > 95% as determined by analytical HPLC. These data are consistent with that previously reported by Morris et al. (44)

124

5-(4-Methoxybenzoyl)-1,1,3,3-tetramethylisoindolin-2-yloxyl (18) To a solution of 5-(4-methoxybenzoyl)-2-methoxy-1,1,3,3-tetramethylisoindoline (15) (550 mg, 1.62 mmol, 1.0 eq.) in dichloromethane (100 mL) was added m-CPBA (77% purity) (800 mg, 3.57 mmol, 2.2 eq.) in portions. The reaction was stirred for one hour at room temperature. Sodium hydroxide (2M, 100 mL) was added and the resulting mixture stirred vigorously for 15 minutes. The reaction mixture was extracted with dichloromethane. The combined organic extracts were washed with water followed by a saturated solution of brine, dried over anhydrous sodium sulphate and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : dichloromethane, 1 : 19) followed by recrystallization from methanol afforded 5-(4-methoxybenzoyl)-1,1,3,3-tetramethylisoindolin-2-yloxyl (18) as a yellow crystalline solid (484 mg, 92% yield). Rf = 0.35, diethyl ether : dichloromethane, 1 : 19. M.p. 108 - 109 °C. IR (ATR) νmax = 3064 (=CH), 2978 and 2932 (–CH), 1643 (C=O), 1614, 1600, 1575, 1506 and 1455 (C=C), 1431 (NO•), 1254 and 1028 (=C–O) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.90 (br s, 2H, Harom), 7.02 (br s, 2H, Harom), 3.94 (s, 3H, OCH3). HRMS: m/z calculated for C20H22NO3Na [M+Na]+ 347.1492; found 347.1485. EPR (DCM): αN = 1.422 mT. Purity > 95% as determined by analytical HPLC. 5-(4-Bromobenzoyl)-1,1,3,3-tetramethylisoindolin-2-yloxyl (19) To a solution of 5-(4-bromobenzoyl)-2-methoxy-1,1,3,3-tetramethylisoindoline (16) (0.80 g, 2.06 mmol, 1.0 eq.) in dichloromethane (100 mL) was added m-CPBA (77% purity) (1.02 g, 4.53 mmol, 2.2 eq.) in portions. The reaction was stirred for one hour at room temperature. 2M sodium hydroxide (100 mL) was added and the resulting mixture stirred vigorously for 15 minutes. The reaction mixture was extracted with dichloromethane. The combined organic extracts were washed with water followed by a saturated solution of brine, dried over anhydrous sodium sulphate and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : dichloromethane, 1 : 19) afforded 5-(4-bromobenzoyl)-1,1,3,3-tetramethylisoindolin-2-yloxyl (19) as a yellow solid (0.68 g, 89% yield). Rf = 0.45, diethyl ether : dichloromethane, 1 : 19. M.p. 147 - 148 °C. IR (ATR) νmax = 3053 (=CH), 2970 and 2924 (–CH), 1659 (C=O), 1612, 1583, 1565, 1480 and 1461 (C=C), 1436 (NO•), 1068 and 1010 (=C–Br) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.73 - 7.68 (m, 4H, Harom). HRMS: m/z calculated for C19H1979BrNO2Na [M+Na]+ 395.0491; found 395.0502. EPR (DCM): αN = 1.426 mT. Purity > 95% as determined by analytical HPLC. 5-(4-(Pyrrolidin-1-yl)benzoyl)-1,1,3,3-tetramethylisoindolin-2-yloxyl (20) To a solution of 5-(4-bromobenzoyl)-1,1,3,3-tetramethylisoindolin-2-yloxyl (19) (900 mg, 2.41 mmol, 1.0 eq.) in anhydrous toluene (50 mL) was added Pd(OAc)2 (54 mg, 0.24 mmol, 0.1 eq.), BINAP (225 mg, 0.36 mmol, 0.15 eq.) and Cs2CO3 (1.33 g, 4.08 mmol, 1.7 eq.). The reaction was deoxygenated by argon bubbling for 20 minutes and stirred under an inert atmosphere of argon. 125

Pyrrolidine (293 mg, 4.12 mmol, 1.7 eq.) was added and the reaction heated to 110 °C and stirred for 16 hours. After cooling to room temperature, the reaction mixture was placed directly onto a short silica column, eluting with a gradient of n-hexanes to diethyl ether : n-hexanes (1 : 3), to remove toluene, salts and non polar components. Further purification via silica gel column chromatography (diethyl ether : dichloromethane, 1 : 19) afforded 5-(4-(pyrrolidin-1-yl)benzoyl)1,1,3,3-tetramethylisoindolin-2-yloxyl (20) as a yellow solid (702 mg, 80% yield). Rf = 0.25, diethyl ether : dichloromethane, 1 : 19. M.p. 175 - 176 °C. IR (ATR) νmax = 3047 (=CH), 2972 and 2928 (–CH), 1640 (C=O), 1596, 1574, 1483 and 1461 (C=C), 1397 (=C–N), 1436 (NO•) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.85 (br s, 2H, Harom), 6.58 (br s, 2H, Harom), 3.42 (br s, 4H, 2×CH2), 2.07 (br s, 4H, 2×CH2). HRMS: m/z calculated for C23H27N2O2Na [M+Na]+ 386.1985; found 386.1989. EPR (DCM): αN = 1.436 mT. Purity > 95% as determined by analytical HPLC. (1-((5-Benzoyl-1,1,3,3-tetramethylisoindolin-2-yl)oxy)-ethyl) benzene (7a) 5-Benzoyl-1,1,3,3-tetramethylisoindolin-2-yloxyl (17) (150 mg, 0.51 mmol, 1.0 eq.) and copper turnings (32 mg, 0.51 mmol, 1.0 eq.) were placed in a round bottom flask, sealed with a rubber septum and placed under an inert atmosphere of argon. Acetonitrile (5 mL), previously deoxygenated by argon bubbling (1 hour), was added via syringe and the reaction mixture further deoxygenated by argon bubbling (15 minutes). N,N,N′,N′,N″-pentamethyldiethylenetriamine (PMDETA) (88 mg, 0.51 mmol, 1.0 eq.) and (1-bromoethyl)benzene (118 mg, 0.64 mmol, 1.25 eq.) were added via syringe and the reaction stirred for 16 hours under an inert atmosphere of argon. The reaction mixture was filtered over celite, eluting with dichloromethane, and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : n-hexanes, 1 : 4) afforded (1-((5-benzoyl-1,1,3,3tetramethylisoindolin-2-yl)oxy)-ethyl) benzene (7a) as a clear oil (191 mg, 94% yield). Rf = 0.45, diethyl ether : n-hexanes, 1 : 4. IR (ATR) νmax = 3063 (=CH), 2973 and 2929 (–CH), 1657 (C=O), 1612, 1598, 1578, 1494 and 1447 (C=C) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.80 - 7.76 (m, 2H, Harom), 7.66 - 7.27 (m, 10H, Harom), 7.20 - 7.07 (m, 1H, Harom), 4.87 (q, J = 6.8 Hz, 1H, CH), 1.65 - 1.64 (m, 3H, CH3), 1.59 - 1.58 (m, 3H, CH3), 1.46 - 1.45 (m, 3H, CH3), 1.29 - 1.28 (m, 3H, CH3), 0.94 (br s, 3H, CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 196.5, 150.6, 150.1, 146.0, 145.6, 144.7, 144.6, 137.97, 137.95, 136.94, 136.86, 132.3, 130.0, 129.95, 129.9, 128.3, 128.2, 127.5, 127.1, 123.6, 123.4, 121.6, 121.4, 83.60, 83.56, 67.9, 67.8, 67.3, 67.2, 30.3, 30.1, 29.7, 29.5, 25.5, 25.3, 25.2, 25.1, 22.5. HRMS: m/z calculated for C27H29NO2Na [M+Na]+ 422.2091; found 422.2105. Purity > 95% as determined by analytical HPLC. Ethyl 2-methyl-2-((5-benzoyl-1,1,3,3-tetramethylisoindolin-2-yl)oxy) propanoate (7b) 5-Benzoyl-1,1,3,3-tetramethylisoindolin-2-yloxyl (17) (150 mg, 0.51 mmol, 1.0 eq.) and copper turnings (32 mg, 0.51 mmol, 1.0 eq.) were placed in a round bottom flask, sealed with a rubber septum and placed under an inert atmosphere of 126

argon. Acetonitrile (5 mL), previously deoxygenated by argon bubbling (1 hour), was added via syringe and the reaction mixture further deoxygenated by argon bubbling (15 minutes). N,N,N′,N′,N″-pentamethyldiethylenetriamine (PMDETA) (88 mg, 0.51 mmol, 1.0 eq.) and ethyl α-bromoisobutyrate (124 mg, 0.64 mmol, 1.25 eq.) were added via syringe and the reaction stirred for 16 hours under an inert atmosphere of argon. The reaction mixture was filtered over celite, eluting with dichloromethane, and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : n-hexanes, 1 : 4) afforded ethyl 2-methyl-2-((5benzoyl-1,1,3,3-tetramethylisoindolin-2-yl)oxy) propanoate (7b) as a white solid (186 mg, 89% yield). Rf = 0.30, diethyl ether : n-hexanes, 1 : 4. M.p. 44 - 46 °C. IR (ATR) νmax = 3066 (=CH), 2972 and 2929 (–CH), 1728 and 1651 (C=O), 1611, 1597, 1577, 1489 and 1447 (C=C) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.79 (dd, J = 8.4, 1.2 Hz, 2H, Harom), 7.66 (dd, J = 8.0, 1.2 Hz, 1H, Harom), 7.61 7.58 (m, 2H, Harom), 7.50 - 7.47 (m, 2H, Harom), 7.17 (d, J = 8.0 Hz, 1H, Harom), 4.23 (q, J = 7.2 Hz, 2H, CH2), 1.54 (br s, 3H, CH3), 1.46 - 1.43 (m, 12H, 4×CH3), 1.34 (t, J = 7.2 Hz, 3H, CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 196.6, 175.2, 150.2, 145.6, 138.0, 137.0, 132.3, 130.1, 130.0, 128.3, 123.6, 121.5, 81.4, 68.0, 67.8, 61.0, 29.7, 29.5, 25.3, 25.09, 25.07, 25.0, 14.3. HRMS: m/z calculated for C25H31NO4Na [M+Na]+ 432.2145; found 432.2149. Purity > 95% as determined by analytical HPLC. Ethyl 2-((5-benzoyl-1,1,3,3-tetramethylisoindolin-2-yl)oxy) propanoate (7c) 5-Benzoyl-1,1,3,3-tetramethylisoindolin-2-yloxyl (17) (150 mg, 0.51 mmol, 1.0 eq.) and copper turnings (32 mg, 0.51 mmol, 1.0 eq.) were placed in a round bottom flask, sealed with a rubber septum and placed under an inert atmosphere of argon. Acetonitrile (5 mL), previously deoxygenated by argon bubbling (1 hour), was added via syringe and the reaction mixture further deoxygenated by argon bubbling (15 minutes). N,N,N′,N′,N″-pentamethyldiethylenetriamine (PMDETA) (88 mg, 0.51 mmol, 1.0 eq.) and ethyl 2-bromopropionate (115 mg, 0.64 mmol, 1.25 eq.) were added via syringe and the reaction stirred for 16 hours under an inert atmosphere of argon. The reaction mixture was filtered over celite, eluting with dichloromethane, and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : n-hexanes, 1 : 4) afforded ethyl 2-((5-benzoyl1,1,3,3-tetramethylisoindolin-2-yl)oxy) propanoate (7c) as a clear oil (183 mg, 91% yield). Rf = 0.30, diethyl ether : n-hexanes, 1 : 4. IR (ATR) νmax = 3056 (=CH), 2973 and 2932 (–CH), 1749 and 1650 (C=O), 1612, 1598, 1578, 1489 and 1447 (C=C) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.80 - 7.78 (m, 2H, Harom), 7.66 (d, J = 7.6 Hz, 1H, Harom), 7.61 - 7.57 (m, 2H, Harom), 7.50 - 7.47 (m, 2H, Harom), 7.17 (d, J = 7.6 Hz, 1H, Harom), 4.50 (q, J = 6.8 Hz, 1H, CH), 4.28 - 4.19 (m, 2H, CH2), 1.57 - 1.56 (m, 3H, CH3), 1.49 - 1.41 (m, 12H, 4×CH3), 1.32 (t, J = 6.8 Hz, 3H, CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 196.6, 174.0, 173.9, 150.3, 149.7, 145.7, 145.2, 138.0, 137.2, 137.1, 132.4, 130.2, 130.1, 130.0, 128.4, 123.8, 123.5, 121.7, 121.5, 81.6, 81.5, 68.2, 68.1, 67.9, 67.8, 60.8, 30.5, 30.4, 29.9, 29.7, 25.5, 25.3, 25.1, 25.0, 18.1, 14.4. HRMS: m/z calculated for C24H30NO4 [M+H]+ 396.2169; found 396.2173. Purity > 95% as determined by analytical HPLC. 127

(1-((5-(4-Methoxybenzoyl)-1,1,3,3-tetramethylisoindolin-2-yl)oxy)-ethyl) benzene (8a) 5-(4-Methoxybenzoyl)-1,1,3,3-tetramethylisoindolin-2-yloxyl (18) (140 mg, 0.43 mmol, 1.0 eq.) and copper turnings (27 mg, 0.43 mmol, 1.0 eq.) were placed in a round bottom flask, sealed with a rubber septum and placed under an inert atmosphere of argon. Acetonitrile (5 mL), previously deoxygenated by argon bubbling (1 hour), was added via syringe and the reaction mixture further deoxygenated by argon bubbling (15 minutes). N,N,N′,N′,N″-pentamethyldiethylenetriamine (PMDETA) (75 mg, 0.43 mmol, 1.0 eq.) and (1-bromoethyl)benzene (100 mg, 0.54 mmol, 1.25 eq.) were added via syringe and the reaction stirred for 16 hours under an inert atmosphere of argon. The reaction mixture was filtered over celite, eluting with dichloromethane, and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : n-hexanes, 1 : 2) afforded (1-((5-(4-methoxybenzoyl)-1,1,3,3-tetramethylisoindolin-2-yl)oxy)-ethyl) benzene (8a) as a clear oil (171 mg, 92% yield). Rf = 0.38, diethyl ether : n-hexanes, 1 : 2. IR (ATR) νmax = 3060 (=CH), 2971 and 2929 (–CH), 1650 (C=O), 1614, 1602, 1576, 1493 and 1454 (C=C), 1255 and 1033 (=C–O) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.86 - 7.78 (m, 2H, Harom), 7.62 - 7.26 (m, 7H, Harom), 7.19 - 7.07 (m, 1H, Harom), 6.98 - 6.94 (m, 2H, Harom), 4.87 (q, J = 6.8 Hz, 1H, CH), 3.89 - 3.88 (m, 3H, OCH3), 1.65 - 1.64 (m, 3H, CH3), 1.59 1.58 (m, 3H, CH3), 1.46 - 1.45 (m, 3H, CH3), 1.29 - 1.28 (m, 3H, CH3), 0.94 (br s, 3H, CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 195.3, 163.1, 149.9, 149.4, 145.7, 145.4, 144.63, 144.59, 137.6, 137.5, 132.4, 130.39, 130.37, 129.4, 129.3, 128.2, 127.5, 127.0, 123.3, 123.1, 121.4, 121.2, 113.5, 83.51, 83.48, 67.8, 67.7, 67.2, 67.1, 55.4, 30.2, 30.1, 29.6, 29.5, 25.4, 25.3, 25.2, 25.0, 22.4. HRMS: m/z calculated for C28H31NO3Na [M+Na]+ 452.2196; found 452.2201. Purity > 95% as determined by analytical HPLC. Ethyl 2-methyl-2-((5-(4-methoxybenzoyl)-1,1,3,3-tetramethylisoindolin-2yl)oxy) propanoate (8b) 5-(4-Methoxybenzoyl)-1,1,3,3-tetramethylisoindolin-2-yloxyl (18) (140 mg, 0.43 mmol, 1.0 eq.) and copper turnings (27 mg, 0.43 mmol, 1.0 eq.) were placed in a round bottom flask, sealed with a rubber septum and placed under an inert atmosphere of argon. Acetonitrile (5 mL), previously deoxygenated by argon bubbling (1 hour), was added via syringe and the reaction mixture further deoxygenated by argon bubbling (15 minutes). N,N,N′,N′,N″-pentamethyldiethylenetriamine (PMDETA) (75 mg, 0.43 mmol, 1.0 eq.) and ethyl α-bromoisobutyrate (105 mg, 0.54 mmol, 1.25 eq.) were added via syringe and the reaction stirred for 16 hours under an inert atmosphere of argon. The reaction mixture was filtered over celite, eluting with dichloromethane, and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : n-hexanes, 1 : 2) afforded ethyl 2-methyl-2-((5-(4-methoxybenzoyl) -1,1,3,3-tetramethylisoindolin-2-yl)oxy) propanoate (8b) as a white solid (176 mg, 93% yield). Rf = 0.28, diethyl ether : n-hexanes, 1 : 2. M.p. 78 - 79 °C. IR 128

(ATR) νmax = 3062 (=CH), 2976 and 2934 (–CH), 1729 and 1646 (C=O), 1613, 1602, 1576, 1506 and 1455 (C=C), 1254 and 1029 (=C–O) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.82 (d, J = 8.8 Hz, 2H, Harom), 7.62 (dd, J = 8.0, 1.6 Hz, 1H, Harom), 7.51 (d, J = 1.2 Hz, 1H, Harom), 7.16 (d, J = 8.0 Hz, 1H, Harom), 6.97 (d, J = 8.8 Hz, 2H, Harom), 4.23 (q, J = 7.2 Hz, 2H, CH2), 3.89 (s, 3H, OCH3), 1.54 (br s, 6H, 2×CH3), 1.46 - 1.42 (m, 12H, 4×CH3), 1.34 (t, J = 7.2 Hz, 3H, CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 195.5, 175.3, 163.2, 149.6, 145.5, 137.7, 132.6, 130.5, 129.6, 123.4, 121.5, 113.6, 81.5, 68.0, 67.9, 61.0, 55.6, 29.7, 29.6, 25.3, 25.2, 25.12, 25.05, 14.3. HRMS: m/z calculated for C26H34NO5 [M+H]+ 440.2431; found 440.2445. Purity > 95% as determined by analytical HPLC. Ethyl 2-((5-(4-methoxybenzoyl)-1,1,3,3-tetramethyliso-indolin2-yl)oxy) propanoate (8c) 5-(4-Methoxybenzoyl)-1,1,3,3-tetramethylisoindolin-2-yloxyl (18) (140 mg, 0.43 mmol, 1.0 eq.) and copper turnings (27 mg, 0.43 mmol, 1.0 eq.) were placed in a round bottom flask, sealed with a rubber septum and placed under an inert atmosphere of argon. Acetonitrile (5 mL), previously deoxygenated by argon bubbling (1 hour), was added via syringe and the reaction mixture further deoxygenated by argon bubbling (15 minutes). N,N,N′,N′,N″-pentamethyldiethylenetriamine (PMDETA) (75 mg, 0.43 mmol, 1.0 eq.) and ethyl 2-bromopropionate (98 mg, 0.54 mmol, 1.25 eq.) were added via syringe and the reaction stirred for 16 hours under an inert atmosphere of argon. The reaction mixture was filtered over celite, eluting with dichloromethane, and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : n-hexanes, 1 : 2) afforded ethyl 2-((5-(4-methoxybenzoyl)-1,1,3,3-tetramethylisoindolin-2-yl)oxy) propanoate (8c) as a white solid (169 mg, 92% yield). Rf = 0.25, diethyl ether : n-hexanes, 1 : 2. M.p. 62 - 64 °C. IR (ATR) νmax = 3061 (=CH), 2976 and 2934 (–CH), 1728 and 1647 (C=O), 1613, 1602, 1576, 1506 and 1455 (C=C), 1253 and 1029 (=C–O) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.81 (d, J = 8.8 Hz, 2H, Harom), 7.62 (d, J = 7.6 Hz, 1H, Harom), 7.51 (d, J =4.8 Hz, 1H, Harom), 7.16 (d, J = 7.6 Hz, 1H, Harom), 6.96 (d, J = 8.8 Hz, 2H, Harom), 4.50 (q, J = 6.8 Hz, 1H, CH), 4.27 - 4.20 (m, 2H, CH2), 3.89 (s, 3H, OCH3), 1.56 - 1.55 (m, 3H, CH3), 1.49 - 1.41 (m, 12H, 4×CH3), 1.32 (t, J = 6.8 Hz, 3H, CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 195.5, 174.0, 163.3, 149.6, 149.1, 145.5, 145.1, 137.9, 137.7, 132.6, 130.5, 129.7, 129.6, 123.5, 123.3, 121.6, 121.4, 113.7, 81.57, 81.55, 68.2, 68.1, 67.9, 67.8, 60.8, 55.6, 30.5, 30.4, 29.9, 29.7, 25.5, 25.3, 25.1, 25.0, 18.1, 14.4. HRMS: m/z calculated for C25H32NO5 [M+H]+ 426.2275; found 426.2277. Purity > 95% as determined by analytical HPLC. (1-((5-(4-(Pyrrolidin-1-yl)benzoyl)-1,1,3,3-tetramethylisoindolin-2-yl)oxy)ethyl) benzene (9a) 5-(4-(Pyrrolidin-1-yl)benzoyl)-1,1,3,3-tetramethylisoindolin-2-yloxyl (20) (125 mg, 0.34 mmol, 1.0 eq.) and copper turnings (22 mg, 0.34 mmol, 1.0 eq.) were placed in a round bottom flask, sealed with a rubber septum and 129

placed under an inert atmosphere of argon. Acetonitrile (5 mL), previously deoxygenated by argon bubbling (1 hour), was added via syringe and the reaction mixture further deoxygenated by argon bubbling (15 minutes). N,N,N′,N′,N″-pentamethyldiethylenetriamine (PMDETA) (60 mg, 0.34 mmol, 1.0 eq.) and (1-bromoethyl)benzene (80 mg, 0.43 mmol, 1.25 eq.) were added via syringe and the reaction stirred for 16 hours under an inert atmosphere of argon. The reaction mixture was filtered over celite, eluting with dichloromethane, and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : n-hexanes, 1 : 1) afforded (1-((5-(4-(pyrrolidin-1-yl)benzoyl)-1,1,3,3tetramethylisoindolin-2-yl)oxy)-ethyl) benzene (9a) as a yellow solid (148 mg, 92% yield). Rf = 0.38, diethyl ether : n-hexanes, 1 : 1. M.p. 108 - 110 °C. IR (ATR) νmax = 3031 (=CH), 2973 and 2926 (–CH), 1638 (C=O), 1607, 1580, 1493 and 1451 (C=C), 1400 (=C–N) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.80 - 7.76 (m, 2H, Harom), 7.60 - 7.26 (m, 7H, Harom), 7.17 - 7.05 (m, 1H, Harom), 6.56 6.52 (m, 2H, Harom), 4.87 (q, J = 6.8 Hz, 1H, CH), 3.39 - 3.38 (m, 2H, CH2), 2.05 - 2.04 (m, 2H, CH2), 1.64 - 1.58 (m, 6H, 2×CH3), 1.451 - 1.446 (m, 3H, CH3), 1.29 - 1.28 (m, 3H, CH3), 0.93 (br s, 3H, CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 194.9, 150.8, 148.8, 148.4, 145.3, 145.0, 144.69, 144.66, 138.7, 138.6, 132.9, 129.0, 128.9, 128.1, 127.4, 127.0, 124.39, 124.37, 123.1, 122.9, 121.2, 121.0, 110.6, 83.44, 83.43, 67.8, 67.7, 67.2, 67.1, 47.5, 30.3, 30.2, 29.7, 29.5, 25.5, 25.4, 25.3, 25.2, 25.1, 22.4. HRMS: m/z calculated for C31H37N2O2 [M+H]+ 469.2850; found 469.2830. Purity > 95% as determined by analytical HPLC. Ethyl 2-methyl-2-((5-(4-(pyrrolidin-1-yl)benzoyl)-1,1,3,3tetramethylisoindolin-2-yl)oxy) propanoate (9b) 5-(4-(Pyrrolidin-1-yl)benzoyl)-1,1,3,3-tetramethylisoindolin-2-yloxyl (20) (125 mg, 0.34 mmol, 1.0 eq.) and copper turnings (22 mg, 0.34 mmol, 1.0 eq.) were placed in a round bottom flask, sealed with a rubber septum and placed under an inert atmosphere of argon. Acetonitrile (5 mL), previously deoxygenated by argon bubbling (1 hour), was added via syringe and the reaction mixture further deoxygenated by argon bubbling (15 minutes). N,N,N′,N′,N″-pentamethyldiethylenetriamine (PMDETA) (60 mg, 0.34 mmol, 1.0 eq.) and ethyl α-bromoisobutyrate (84 mg, 0.43 mmol, 1.25 eq.) were added via syringe and the reaction stirred for 16 hours under an inert atmosphere of argon. The reaction mixture was filtered over celite, eluting with dichloromethane, and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : n-hexanes, 1 : 1) afforded ethyl 2-methyl-2-((5-(4-(pyrrolidin1-yl)benzoyl)-1,1,3,3-tetramethylisoindolin-2-yl)oxy) propanoate (9b) as an off white solid (142 mg, 86% yield). Rf = 0.25, diethyl ether : n-hexanes, 1 : 1. M.p. 100 - 102 °C. IR (ATR) νmax = 3049 (=CH), 2975 and 2929 (–CH), 1729 and 1631 (C=O), 1605, 1581 and 1488 (C=C), 1399 (=C–N) cm-1. 1H NMR (400 MHz, CDCl3): (ppm) = 7.79 (d, J = 8.8 Hz, 2H, Harom), 7.60 (d, J = 8.0, 1.2 Hz, 1H, Harom), 7.47 (s, 1H, Harom), 7.14 (d, J = 7.6 Hz, 1H, Harom), 6.55 (d, J = 8.8 Hz, 2H, Harom), 4.23 (q, J = 7.2 Hz, 2H, CH2), 3.40 - 3.37 (m, 4H, 2×CH2), 2.06 - 2.03 (m, 4H, 2×CH2), 1.54 (br s, 6H, 2×CH3), 1.45 - 1.42 (m, 12H, 4×CH3), 1.34 (t, J = 7.2 Hz, 3H, CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 195.1, 130

175.3, 150.9, 148.5, 145.1, 138.8, 133.0, 129.1, 124.6, 123.2, 121.3, 110.7, 81.4, 68.0, 67.9, 61.0, 47.7, 29.8, 29.7, 25.6, 25.3, 25.2, 25.13, 25.07, 14.3. HRMS: m/z calculated for C29H39N2O4 [M+H]+ 479.2904; found 479.2907. Purity > 95% as determined by analytical HPLC. Ethyl 2-((5-(4-(pyrrolidin-1-yl)benzoyl)-1,1,3,3-tetramethylisoindolin-2yl)oxy) propanoate (9c) 5-(4-(Pyrrolidin-1-yl)benzoyl)-1,1,3,3-tetramethylisoindolin-2-yloxyl (20) (125 mg, 0.34 mmol, 1.0 eq.) and copper turnings (22 mg, 0.34 mmol, 1.0 eq.) were placed in a round bottom flask, sealed with a rubber septum and placed under an inert atmosphere of argon. Acetonitrile (5 mL), previously deoxygenated by argon bubbling (1 hour), was added via syringe and the reaction mixture further deoxygenated by argon bubbling (15 minutes). N,N,N′,N′,N″-pentamethyldiethylenetriamine (PMDETA) (60 mg, 0.34 mmol, 1.0 eq.) and ethyl 2-bromopropionate (78 mg, 0.43 mmol, 1.25 eq.) were added via syringe and the reaction stirred for 16 hours under an inert atmosphere of argon. The reaction mixture was filtered over celite, eluting with dichloromethane, and concentrated in vacuo. Purification via silica gel column chromatography (diethyl ether : n-hexanes, 1 : 1) afforded ethyl 2-((5-(4-(pyrrolidin-1-yl)benzoyl)-1,1,3,3-tetramethylisoindolin-2-yl)oxy) propanoate (9c) as an off white solid (144 mg, 90% yield). Rf = 0.25, diethyl ether : n-hexanes, 1 : 1. M.p. 88 - 89 °C. IR (ATR) νmax = 3051 (=CH), 2974 and 2931 (–CH), 1736 and 1632 (C=O), 1605, 1578 and 1488 (C=C), 1399 (=C–N) cm-1. 1H NMR (400 MHz, CDCl3): δ (ppm) = 7.79 (d, J = 8.8 Hz, 2H, Harom), 7.60 (d, J = 7.2 Hz, 1H, Harom), 7.46 (d, J = 4.4 Hz, 1H, Harom), 7.14 (d, J = 6.4 Hz, 1H, Harom), 6.55 (d, J = 8.8 Hz, 2H, Harom), 4.50 (q, J = 6.8 Hz, 1H, CH), 4.28 - 4.20 (m, 2H, CH2), 3.40 - 3.37 (m, 4H, 2×CH2), 2.06 - 2.03 (m, 4H, 2×CH2), 1.56 - 1.55 (m, 3H, CH3), 1.49 - 1.40 (m, 12H, 4×CH3), 1.32 (t, J = 6.8 Hz, 3H, CH3). 13C NMR (400 MHz, CDCl3): δ (ppm) = 195.1, 174.1, 151.0, 148.5, 148.0, 145.1, 144.7, 139.1, 138.9, 133.5, 129.3, 129.2, 124.6, 123.3, 123.0, 121.4, 121.2, 110.7, 81.58, 81.55, 68.2, 68.1, 67.9, 67.8, 60.8, 47.7, 30.5, 30.4, 29.9, 29.8, 25.6, 25.5, 25.4, 25.1, 25.0, 18.1, 14.4. HRMS: m/z calculated for C28H37N2O4 [M+H]+ 465.2748; found 465.2752. Purity > 95% as determined by analytical HPLC.

References 1. 2. 3. 4. 5. 6.

Fouassier, J. P.; Ruhlmann, D.; Graff, B.; Morlet-Savary, F. Prog. Org. Coat. 1995, 25, 235. Fouassier, J. P.; Allonas, X.; Burget, D. Prog. Org. Coat. 2003, 47, 16. Clapper, J. D.; Sievens-Figueroa, L.; Guymon, A. C. Chem. Mater. 2008, 20, 768. Crivello, J. V.; Reichmanis, E. Chem. Mater. 2014, 26, 533. Xiao, P.; Zhang, J.; Dumur, F.; Tehfe, M. A.; Morlet-Savary, F.; Graff, B.; Gigmes, D.; Fouassier, J. P.; Lalevee, J. Prog. Polym. Sci. 2015, 41, 32. Yagci, Y.; Jockusch, S.; Turro, N. J. Macromolecules 2010, 43, 6245. 131

7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.

Zhao, Y.; Yu, M.; Zhang, S.; Liu, Y.; Fu, X. Macromolecules 2014, 47, 6238. Kermagoret, A.; Wenn, B.; Debuigne, A.; Jerome, C.; Junkers, T.; Detrembleur, C. Polym. Chem. 2015, 6, 3847. Benedikt, S.; Moszner, N.; Liska, R. Macromolecules 2014, 47, 5526. Tonnar, J.; Pouget, E.; Lacroix-Desmazes, P.; Boutevin, B. Macromol. Symp. 2009, 281, 20. Dadashi-Silab, A.; Tasdelen, M. A.; Yagci, Y. J. Polym. Sci., Part A: Polym. Chem. 2014, 52, 2878. Tasdelen, M. A.; Ciftci, M.; Yagci, Y. Macromol. Chem. Phys. 2012, 213, 1391. Treat, N. J.; Sprafke, H.; Kramer, J. W.; Clark, P. G.; Barton, B. E.; Alaniz, J. R.; Fors, B. P.; Hawker, C. J. J. Am. Chem. Soc. 2014, 136, 16096. Xu, J.; Jung, K.; Atme, A.; Shanmugam, S.; Boyer, C. J. Am. Chem. Soc. 2014, 136, 5508. Sciano, J. C.; Connolly, T. J.; Mohtat, N.; Pliva, C. N. Can. J. Chem. 1997, 75, 92. Versace, D. L.; Lalevee, J.; Fouassier, J. P.; Gigmes, D.; Guillaneuf, Y.; Bertin, D. J. Polym. Sci., Part A: Polym. Chem. 2010, 48, 2910. Hu, S.; Malpert, J. H.; Neckers, D. C. Polymer 2000, 41, 445. Guillaneuf, Y.; Versace, D. L.; Bertin, D.; Lalevee, J.; Gigmes, D.; Fouassier, J. P. Macromol. Rapid Commun. 2010, 31, 1909. Versace, D. L.; Guillaneuf, Y.; Bertin, D.; Fouassier, J. P.; Lalevee, J.; Gigmes, D. Org. Biomol. Chem. 2011, 9, 2892. Demeter, A.; Horvath, K.; Boor, K.; Molnar, L.; Soos, T.; Lendvay, G. J. J. Phys. Chem. A 2013, 117, 10196. Beckett, A.; Porter, G. J. Chem. Soc. Faraday Trans. 1963, 59, 2038. Leigh, W. J.; Lathioor, E. C.; St. Pierre, M. J. J. Am. Chem. Soc. 1996, 118, 12339. Pitts, J. N.; Letsinger, R. L.; Taylor, R. P.; Patterson, J. M.; Recktenwald, G.; Martin, R. B. J. Am. Chem. Soc. 1959, 81, 1068. Scaiano, J. J. Photochem. 1973, 2, 81. Nicolas, J.; Guillaneuf, Y.; Lefay, C.; Bertin, D.; Gigmes, D.; Charleux, B. Prog. Polym. Sci. 2013, 38, 63. Demeter, A.; Horvath, K.; Boor, K.; Molnar, L.; Soos, T.; Lendvay, J. Phys. Chem. A 2013, 117, 10196. Favaro, G. Chem. Phys. Lett. 1973, 21, 401. Wagner, P. J.; Truman, R. J.; Scaiano, J. C. J. Am. Chem. Soc. 1985, 107, 7093. Ghoneim, N.; Monbelli, A.; Pilloud, D.; Suppan, P. J. Photochem. Photobiol., A 1996, 94, 145. Singh, A. K.; Bhasikuttan, A. C.; Palit, D. K.; Mittal, J. P. J. Phys. Chem. A 2000, 104, 7002. Bhasikuttan, A. C.; Singh, A. K.; Palit, D. K.; Sapre, A. V.; Mittal, J. P. J. Phys. Chem. A 1998, 102, 3470. Palit, D. K. Res. Chem. Intermed. 2005, 31, 205. Goto, A.; Scaiano, J. C.; Maretti, L. Photochem. Photobiol. Sci. 2007, 6, 833. 132

34. Guillaneuf, Y.; Bertin, D.; Gigmes, D.; Versace, D. L.; Lalevee, J.; Fouassier, J. P. Macromolecules 2010, 43, 2204. 35. Versace, D. L.; Lalevee, J.; Fouassier, J. P.; Guillaneuf, Y.; Bertin, D.; Gigmes, D. Macromol. Rapid Commun. 2010, 31, 1383. 36. Nilsen, A.; Braslau, R. J. Polym. Sci., Part A: Polym. Chem. 2005, 44, 697. 37. Anderson, D. R.; Keute, J. S.; Chapel, H. L.; Koch, T. H. J. Am. Chem. Soc. 1979, 101, 1904. 38. Kalugina, A. D.; Mel’nikov, M. Y. High Energy Chem. 2007, 41, 385. 39. Bottle, S.; Chand, U.; Micallef, A. S. Chem. Lett. 1997, 9, 857. 40. Hsieh, J. C.; Cheng, C. H. Chem. Commun. 2005, 36, 4554. 41. Griffiths, P. G.; Moad, G.; Rizzardo, E.; Solomon, D. H. Aust. J. Chem. 1983, 36, 397. 42. Micallef, A. S.; Bott, R. C.; Bottle, S. E.; Smith, G.; White, J. M.; Matsuda, K.; Iwamura, H. J. Chem. Soc., Perkin Trans. 2 1999, 1, 65. 43. Fairfull-Smith, K. E.; Bottle, S. E. Eur. J. Org. Chem. 2008, 5391. 44. Morris, J. C.; Walsh, L. A.; Gomes, B. A.; Gigmes, D.; Fairfull-Smith, K. E.; Bottle, S. E.; Blinco, J. P. RSC Adv. 2015, 5, 95598. 45. Chalmers, B. A.; Morris, J. C.; Fairfull-Smith, K. E.; Grainger, R. S.; Bottle, S. E. Chem. Commun. 2013, 49, 10382. 46. Turro, N. J.; Ramamurthy, V.; Scaiano, J. C. Principles of Molecular Photochemistry; University Science Books: Sausalito, CA, 2009. 47. Singh, A. K.; Palit, D. K.; Mittal, J. P. Res. Chem. Intermed. 2001, 27, 125. 48. Grewer, C.; Brauer, H. D. J. Phys. Chem. 1994, 98, 4230. 49. Chunsen, L.; Danovich, D.; Shaik, S. Chem. Sci. 2012, 3, 1903. 50. Droege, A. T.; Tully, F. P. J. Phys. Chem. 1987, 91, 1222. 51. Jornet, D.; Tormos, R.; Miranda, M. A. J. Phys. Chem. B 2011, 115, 10768. 52. Sakamoto, M.; Cai, X.; Hara, M.; Tojo, S.; Fujitsuka, M.; Majima, T. J. Phys. Chem. A 2004, 108, 8147. 53. Nelsen, S. F.; Bartlett, P. D. J. Am. Chem. Soc. 1966, 88, 143. 54. Hoffmann, A. K.; Henderson, A. T. J. Am. Chem. Soc. 1961, 83, 4671. 55. Johnston, L. J.; Tencer, M.; Scaiano, J. C. J. Org. Chem. 1986, 51, 2806.

133

Chapter 6

Catalyzed Radical Termination (CRT) in the Metal-Mediated Polymerization of Acrylates: Experimental and Computational Studies Thomas G. Ribelli,1 S. M. Wahidur Rahaman,2 Krzysztof Matyjaszewski,1 and Rinaldo Poli*,2,3 1Department of Chemistry, Carnegie Mellon University, 4400 Fifth Avenue, Pittsburgh, Pennsylvania 15213, United States 2CNRS, LCC (Laboratoire de Chimie de Coordination), Université de Toulouse, UPS, INPT, 205 Route de Narbonne, BP 44099, F-31077, Toulouse Cedex 4, France 3Institut Universitaire de France, 1 rue Descartes, 75231 Paris Cedex 05, France *E-mail:[email protected].

Metal complexes stabilized by appropriate ligands, particularly CuI/L systems, have proven powerful for the controlled polymerization of acrylates and other monomers by atom transfer radical polymerization (ATRP). The polymerization of acrylates by CuI/L systems, however, is haunted by interference of catalyzed radical termination (CRT), which reduces the chain-end fidelity. Other monomers do not appear to be affected by this phenomenon to any significant extent. The phenomenon appears to involve the formation of an organometallic intermediate by reversible radical trapping, as in organometallic mediated radical polymerization (OMRP). We summarize here the current knowledge and the efforts made to elucidate the CRT pathway and products.

© 2018 American Chemical Society

Transition metal complexes are quite useful for macromolecular engineering by reversible deactivation radical polymerization (RDRP), notably using atom transfer radical polymerization (ATRP) (1), but also organometallic-mediated radical polymerization (OMRP) (2). However, the reactivity of organic radicals with transition metal complexes is a complex subject, with many possible reaction pathways leading to different outcomes, as summarized in Scheme 1 (3–5). The OMRP scenario, by itself, already shows complexity, because two controlling modes are possible depending on the metal coordination environment, the radical/metal stoichiometry and the metal-carbon bond strength: reversible termination (OMRP-RT) and degenerative transfer (OMRP-DT). The OMRP-RT process requires reversible homolytic cleavage of a relatively weak metal-carbon bond in the dormant species, whereas OMRP-DT requires excess radicals and an easily available coordination site on the metal atom in order to promote the reversible, rapid and degenerative associative exchange.

Scheme 1. Possible reactivity pathways of radical chains with transition metal complexes, or relevance in RDRP. Mtx/L is a metal complex where L represents the coordination sphere and x is the formal oxidation state, Y = halogen, E = any element, and Pn• is a growing radical chain with degree of polymerization n. Dashed arrows are pathways that have been discarded by the experiments (see text). 136

If only ATRP activation/deactivation and/or OMRP activation/deactivation and/or degenerative exchange occur, a well-controlled RDRP may be obtained. Synergies between ATRP and OMRP-RT (6–8) or between the two OMRP controlling mechanisms (9–11) have been highlighted, without negative consequences (or with positive consequences) on the quality of control. However, other phenomena may interfere with controlled chain growth and it is important to understand how these occur and which factors promote them, in order to improve the polymerization control. Catalytic chain transfer (CCT) involves the same Mtx/L and Pn• partners that yield OMRP-RT deactivation, but results in a β-H atom transfer to generate a dead chain with an unsaturated chain end and the metal hydride species H-Mtx/L, which then starts a new chain by delivering the H atom to monomer, see Scheme 1. Therefore, CCT and OMRP-RT deactivation are in direct competition. This is a well-known phenomenon that has its own relevance for macromolecular engineering and for industrial production (12, 13). This chapter does not deal with this nevertheless important phenomenon. Rather, it deals with the last phenomenon shown in Scheme 1, namely catalyzed radical termination (CRT). This phenomenon entails an accelerated disappearance of the radical chains, relative to spontaneous bimolecular terminations, because of the presence of the transition metal complex. In particular, it appears to have the strongest impact on the radical polymerization of acrylates and is strongly metal-dependent. It has been described so far for CuI and FeII catalysts, but not for CoII, which preferentially lead to CCT or OMRP-DT when the [radical]:[Mtx/L] ratio is greater than one. As hinted in Scheme 1, two possible pathways for CRT are possible, via the hydride complex H-Mtx+1/L and via the organometallic complex Pn-Mtx+1/L and current evidence points to the latter option, as will be detailed below.

CRT Discovery and Initial Studies The story begins with the report by Matyjaszewski and Woodworth back in 1998 (14) of the styrene and methyl acrylate polymerization under OMRP conditions, i.e. AIBN-initiated, in the presence of CuI and CuII triflates coordinated by substituted 2,2’-bipyridine (bpy) ligands and in the absence of halides. While CuII had no noticeable effect, the presence of CuI significantly retarded the MA (but not the styrene) polymerization, see upper part of Figure 1. These results were taken as evidence that, at least for styrene, the interaction between the growing polymer chain and the metal center does not contribute to chain growth control in ATRP. Otherwise stated, there is no ATRP/OMRP-RT interplay for styrene. For the MA polymerization, on the other hand, the CuI complex is able to act as an OMRP-RT deactivator. However, the pure OMRP-RT mechanism is not sufficient to control the polymerization because the Mn does not evolve linearly with conversion and the molecular weight distributions (MWDs) are broad, see Figure 1 (lower part). Note that this study made use of excess copper relative to initiator, rather than a catalytic amount as in later studies. It was stated that “polymer molecular weights are mostly unaffected” by the presence of CuI, although close inspection of Figure 1 indicates that a slight 137

decrease may be present. On the basis of the current knowledge, an Mn decrease is expected in the presence of CRT activity (vide infra). It is also worth noting that the CuI triflate complex was produced in situ by comproportionation of CuII triflate and Cu0 and, as established in a later contribution (15), Cu0 can also promote termination. Therefore, it is not possible to conclude on the presence of CuI-catalyzed termination in this system. As we now know, CRT requires formation of the organometallic intermediate, thus the absence of CRT activity for this system (or a weak one at the most) is probably the consequence of a rather weak CuI-PMA bond.

Figure 1. Kinetics and evolution of Mn and Mw/Mn with Conversion in AIBN-Initiated Polymerization of Methyl Acrylate in Toluene (3 mL) at 60°C. Conditions: MA/AIBN = 33:0.04 mmol with either no copper, or Cu(OTf)2 (0.016 mmol) or Cu(OTf)2 and Cu0 (0.033 mmol of each). Reproduced with permission from ref. (14). Copyright 1998 American Chemical Society.

Clear evidence for the presence of CuI-CRT was obtained for the first time in a study of the butyl acrylate polymerization in the presence of [CuI(TPMA*3)]+, initiated by AIBN in anisole (once again, OMRP conditions) (16). The CuI complex was made in situ from [CuI(MeCN)4][BF4] and the TPMA*3 ligand, which is shown in Scheme 2 together with other ligands used in later studies. This ligand imparts a very negative reduction potential to the CuII/CuI redox couple, promoting OMRP-RT deactivation (as well as ATRP activation (17)). Hence, greater stability is predicted for the organometallic dormant chains [PBA-CuII(TPMA*3)]+. 138

Scheme 2. TPMA and Substituted Derivatives Used in CuI-CRT Investigations Instead of controlled polymer growth by OMRP-RT, the study revealed an unexpected decrease of the polymerization rate in the presence of substoichiometric amounts of CuI. Furthermore, this decrease was proportional to [CuI]0 (see Figure 2). Correspondingly, greater [CuI]0 led to polymers with a lower Mn. PREDICI simulations were consistent with a rate-limiting reaction between CuI and the radical chain, followed by rapid reaction of the generated intermediate with a second radical, but could not discriminate between the hydride and the organometallic pathways shown in Scheme 1.

Figure 2. Kinetic plots for BA polymerizations with various [CuI] loadings. Conditions: [BA]:[AIBN]:[TPMA*]:[CuI(MeCN)4][BF4] = 160:0.2:0.06−0.24:0.016−0.064, [BA] = 5.6 M, 20% (v/v) anisole, T = 60 °C. Adapted with permission from ref. (16). Copyright 2012 American Chemical Society. In subsequent work (18), Buback, Matyjaszewski et al. further investigated the CuI-CRT mechanism from the kinetic point of view using [CuI(TPMA)]+ as an ATRP activator. Normal ATRP of BA was conducted and the chain-end functionality was monitored via the accumulation of [X-CuII(TPMA)]+ as defined by the persistent radical effect (PRE). At 45% monomer conversion, the CEF was found to be 97% whereas 99.95% was expected if conventional bimolecular radical termination (RT) was the only source of radical termination. The difference was attributed to CRT. If the CEF resulted only from bimolecular terminations (vt = kt[Pn•]2) regulated by the PRE (19), the deactivator accumulation in solution should yield a linear growth of the F(Y) function (equation 1, where Y = [CuII/L] and I0 = [RX]0 and C0 = [CuI/L]0) (20): F(Y) = 2ktKATRP2t. On the other hand, under the assumption that CRT occurs by the rate law vt = ktCu(I)[CuI][Pn•] and that 139

CRT is the dominant termination process, the deactivator accumulation should lead to the linear dependence of another function, G(Y), defined in equation 2: G(Y) = ktCu(I)KATRPt.

Figure 3. Time dependence of F(Y) and G(Y) in termination experiments of the model acrylate radical generated from the activation of MBrP with CuI/TPMA. Reproduced with permission from ref. (18). Copyright 2013 American Chemical Society. Working with the methyl 2-bromopropionate (MBrP) initiator as a model system (in the absence of monomer) in MeCN at 25°C, a reasonably linear growth of F(Y) was observed when using a 10-fold excess of MBrP relative to the CuI complex (Figure 3a), indicating that conventional bimolecular radical termination dominates. However, when the reaction was conducted under stoichiometric conditions, a positive curvature of the F(Y) plot (Figure 3b) was observed, indicating that the accumulation of X-CuII was faster than expected for RT only. On the other hand, the G(Y) plot was linear after an initial negatively curved region (Figure 3c), which results from the initial non-negligible contribution of conventional bimolecular terminations. A linear G(Y) plot was also observed for the model study run with complex [CuI(Me6TREN)]+ as activator, as well as for the same two activators under BA polymerization conditions. The slope (TPMA) and 2.2·104 analysis gave the apparent CRT rate constants -1 -1 (Me6TREN) M s . It is to be noted that these investigations were conducted in MeCN as solvent, which can potentially interfere with the termination mechanism due to the relatively weak C-H bonds. The previous study in anisole (16) used [CuI(MeCN)4]+ as precursor, where MeCN is also present. We’ll come back on this point at the end of the chapter. This contribution (18) also highlighted for the first time that CRT gives a much larger contribution to the termination 140

of acrylates than to those of methacrylates and styrene. This contrasts with the CCT behavior, where methacrylates are more active than acrylates, hinting to the involvement of a different key intermediate in the two mechanisms. A combination of low-temperature EPR and near-infrared investigations for the TPMA system attempted to identify this intermediate, with evidence for the accumulation of an unstable species at -40°C, but the identity of this species could not be elucidated (a more recent investigation with model radicals will be detailed in a later section). CRT is not limited to copper(I) complexes. In a 2014 contribution (21), Schröder and Buback highlighted a FeII-catalyzed termination of acrylates, using the single-pulse pulsed laser polymerization (SP-PLP) technique to generate high concentration of photoinitiated radical chains, in combination with EPR monitoring. Comparison of the radical disappearance rate in the presence and absence of FeBr2/[nBu4N]Br showed a combination of conventional bimolecular and FeII-catalyzed termination. Both the secondary propagating radicals (SPR) and the midchain radicals (MCR; formed via backbiting of SPR), could be followed in time because they give two distinct EPR signals. The FeII-CRT was found to be the dominant termination pathway at high catalyst concentrations, with a rate law vt = ktFe(II)[FeII][PBA•] and a mean value of ktFe(II) = 2.3·104 M-1 s-1 at -60°C, whereas the backbiting process played a greater role at low catalyst loadings and at higher temperatures. The kinetic analysis also showed that the FeII-CRT is not an important process for MCRs and that, as also previously shown for the CuI-CRT, FeII-CRT plays no major role for the termination of methacrylates. Note that all these investigations were conducted in bulk monomer, not in MeCN or other solvents with weak C-H bonds. It is also useful to underline that the nature of the termination products (disproportionation or combination) was not investigated.

Hydride or Organometallic Intermediate? As already mentioned in the previous section, the greater CRT activity for acrylates than for methacrylates, when compared to the opposite relative susceptibility of these two radicals for CCT, hints to a different key intermediate in the two processes, hence to the probable involvement of the organometallic dormant species Pn-Mtx+1/L in CRT. As can be noted in Scheme 1, generation of a hydride intermediate would necessarily lead to termination by disproportionation (Disp). The pathway via the organometallic intermediate, on the other hand, could result in either Disp, combination (Comb) or chain-end saturation depending on how the organometallic species catalyzes termination (more on this later). It is therefore crucial to establish the nature of the termination products generated from CRT. We must point out here that the nature of the termination products for acrylates has been controversial, even for conventional bimolecular termination. Earlier work has pointed to the dominance of either Comb (22, 23) or Disp (24), although the former seems to attract greater favors within the community. Recent work by Asua et al. (25) concludes that the SPRs terminate preferentially by Comb, but Disp competes to a greater extent at higher temperatures following 141

the production of MCRs by backbiting. Additional contributions on this issue, related to the CRT mechanism, will be highlighted in the next section. A first study of the products obtained from the CRT of polyacrylate radicals is in the above-mentioned [CuI(TPMA)]+-CRT study (18). Polystyrene (PSt-Br) and poly(methyl acrylate) (PMA-Br) ATRP macroinitiators were activated by CuIBr/TPMA in the absence of monomer, monitoring the Br-CuII/L deactivator accumulation by UV-visible spectroscopy. The resulting polymers were analyzed by size exclusion chromatography (SEC) and compared to the corresponding macroinitiator (see reaction sequence in Scheme 3, where the organometallic species is assumed as the key intermediate).

Scheme 3. Activation and Termination of ATRP Macroinitiators For the PSt-Br macroinitiator, there was 50% termination after 18 h at 40°C in toluene/DMSO and the SEC analysis of the recovered polymer showed a prominent distribution at twice the molecular weight (MW) of the macroinitiator (Figure 4b), consistent with the dominance of Comb. The low Mn distribution mostly corresponds to the unreacted macroinitiator. As stated above, the CuI complex has negligible CRT activity for PSt radicals, which are known to preferentially terminate bimolecularly by Comb (24). For the PMA-Br macroinitiator, there was 83% termination after only 2.5 h at 25°C in MeCN and the SEC of the recovered polymer shows the absence of Comb products (Figure 4a). Note that these investigations were carried out in MeCN solution. On the basis of the widespread perception that the polyacrylates bimolecular termination leads to Comb, it was then suggested that the CuI-CRT of polyacrylates leads to Disp (18). Although the experimental evidence that CRT is more efficient for acrylates than for methacrylates militates against involvement of the hydride intermediate, the result of the above study is insufficient to discard the hydride pathway, because transit via this intermediate would also lead to Disp (Scheme 1). Additional evidence in favor of the organometallic pathway was gathered via DFT calculations. A first contribution (26) evaluated the aptitude of the [CuI(TPMA)]+ complex to abstract a β-H atom from an acrylate radical, using the small model CH3CH•(COOCH3), to generate the hydride complex [H-CuII(TPMA)]+ and methyl acrylate. The reaction profile was compared with that of [CoII(porphyrin)], a model of the bulkier tetramesitylporphyrin complex that efficiently catalyzes chain transfer to monomer. As shown in Figure 5, both complexes are predicted to favorably trap the acrylate radical to yield the OMRP-RT dormant species, (CH3)(COOCH3)CH-Mtx+1/L. Conversely, the β-H atom transfer process, which occurs via the van der Waals adduct •CH(COOCH3)(CH2-H···Mtx/L), is less energetically favorable and has a much higher activation barrier for the CuI catalyst than for the CoII system. This result agrees with the known CCT activity of the CoII system. Incidentally, for both 142

systems, the calculations also suggest that further quenching of the hydride intermediate by a radical to complete the CRT process has essentially no energetic barrier. The concentration difference ([M] >> [R•]) rationalizes the preference of [H-CoIII/L] to deliver the H atom to a monomer molecule to complete the CCT cycle, rather than to a radical to complete the CRT cycle.

Figure 4. GPC curves of PMA-Br (a, Mn = 4200, Đ = 1.20) and PSt-Br (b, Mn = 12000, Đ = 1.10) macroinitiators and of terminated polymers after activation by [CuI(TPMA)]+ in MeCN at 25°C for the indicated time. Reproduced with permission from the Supporting Information of ref. (18). Copyright 2013 American Chemical Society.

Figure 5. Gibbs energy profile for the competing OMRP-RT trapping and β-H atom transfer involving the model acrylate radical R = CH3CH•(COOCH3) and complex [CuI(TPMA)]+ or [CoII(porphyrin)]. 143

The hydride intermediate hypothesis was further probed experimentally with new systems where the tetradentate TMPA ligand was replaced with a tridentate one (either BPMAMe or BPMA*Pr, see Scheme 2) (27). This was motivated by the idea that β-H elimination could only occur in the presence of a vacant coordination site cis to the CuII-C bond, which would require dissociation of one of the TPMA pyridine arms. In the same contribution, a more detailed investigation was carried out for the full [CuI(TMPA*n)]+ series of complexes (n = 0, 1, 2, 3, Scheme 2) in order to relate redox potential (E1/2) to CRT activity. Each substitution of the pyridine ring resulted in an approximately -60 mV shift in E1/2. By comparing the polymerization rates and polymer MWs, it was established that the CRT activity increases with more negative E1/2, whereas the tridentate ligands yield less active catalysts: (BPMAMe), 20 (BPMA*Pr), 29 (TPMA), 45 (TPMA*1), 55 (TPMA*2), 96 (TPMA*3). These results further contribute to discard the hydride pathway. In parallel, DFT calculations of the CuII-CH(CH3)(COOCH3) homolytic bond dissociation free energy (BDFE) revealed a linear correlation between this parameter and KATRP and the standard reduction potential of the CuI/CuII redox couple, see Figure 6. Interestingly, the ATRP activation equilibrium is most sensitive, the OMRP-RT activation equilibrium has intermediate sensitivity, and the CRT activity is least sensitive to the CuII/L reduction potential. A further point of interest of that study was to reveal two limiting scenarios. For the BPMA*Pr system, the polymer Mn scaled inversely with the catalyst loading, whereas for the TPMA-based systems the Mn decreased only marginally. Subjecting the data to kinetic analysis, the first scenario is consistent with a rate-determining formation of the organometallic intermediate, whereas the second one suggests a rate-determining reaction between the organometallic intermediate and the second radical.

Figure 6. Relationship between ln KATRP (red squares) for MBrP in MeCN at 25 °C, ln KOMRP for acrylate radical (blue circles), and ln of apparent rate coefficient of Cu-mediated termination, ln (green triangles), with the CuII/L reduction potential E1/2. Reproduced with permission from ref. (27). Copyright 2016 American Chemical Society. 144

A more recent contribution by Zerk and Bernhardt has provided evidence for the generation of an organometallic intermediate, under conditions related to Cu-CRT, although not with acrylate radical models. It reports the cyclic voltammetric investigations of [CuII(TPMA)Br]Br and [CuII(Me6TREN)X]X (X = Cl or Br) in the presence or variable amounts of ethyl 2-bromoisobutyrate (EBiB) or bromoacetonitrile (28). The [CuII(TPMA)Br]+ reduction in DMSO or MeCN yields an electrochemically reversible wave at E°X in the absence of RBr, but addition of the latter induces the appearance of a second irreversible reduction process at a more negative potential E°R. This is more clearly visible as an irreversible wave when using BrCH2CN. The same behavior is also exhibited by the voltammogram of [CuII(Me6TREN)X]X in the presence of the corresponding XCH2CN. The second wave is attributed to the reduction of the organometallic derivative [R-CuII/L]+, which is generated by the sequence of processes summarized in Scheme 4: (i) reduction of [X-CuII/L]+ to [X-CuI/L] at E°X; (ii) halide dissociation to produce [CuI/L]+, (iii) ATRP activation to regenerate [X-CuII/L]+ and produce R•; (iv) R• trapping by the electrogenerated [CuI/L]+ complex in the diffusion layer to yield [R-CuII/L]+; (v) reduction of [R-CuII/L]+ at E°R. The implication of radicals is consistent with the elimination of the reduction at E°R upon addition of TEMPO to the systems. The reason for the absence of a distinct wave at E°R for the EBiB substrate is consistent with the expected weaker Cu-R bond produced in that case. In other words, 1/KOMRP = kda/ka in step (iv) of Scheme 4 is too small for the EBiB system to produce a sufficiently stable organocopper(II) intermediate. The authors have attributed the irreversibility of the process at E°R to spontaneous dissociation of the carbanion R- from the CuI center followed by probable protonation by traces of water. It is important to note that the addition of 17 eq. of water relative to [RX]0 had little to no effect on the stability of the R-CuII/L intermediate, indicating stability towards hydrolysis for this intermediate on the time-scale of the electrochemical sweep. A spectroelectrochemical experiment with UV-visible detection, while maintaining a constant potential E such that E°R < E < E°X, was carried out on the [CuII(TPMA)Br]Br/BrCH2CN system in MeCN. At this potential, the organocopper(II) species is not reduced and accumulates in the medium. The UV-visible monitoring showed a hypochromic shift of the electronic transition and an isosbestic behavior, suggesting the relative stability of the organometallic product. The possible formation of a hydride complex, [H-CuII/L]+, is excluded because the •CH2CN radical cannot furnish a β-H atom. The proposed [CuII(TPMA)(CH2CN)]+ product was also characterized by EPR spectroscopy.

145

Scheme 4. Sequence of Reactions Occurring in the Cyclic Voltammetry of [X-CuII/L]+ + RX

Combination or Disproportionation? A series of thought-provoking contributions were published by Yamago et al. on the acrylate radical termination, both with and without a copper complex as a catalyst. It is more appropriate to first highlight the studies carried out in the absence of copper complexes because they are relevant for the subsequent work on CRT. The strategy used for these studies is similar the Buback/Matyjaszewski one (Scheme 3) (18) with radicals generated from well-defined macroinitiators. In this case, the radicals were generated from R0-Mn-TePh (M = methyl or n-butyl or t-butyl acrylate), produced by tellurium-mediated radical polymerization (TERP). Small model radicals (n = 0, 1) were also used, see Scheme 5 (29).

Scheme 5. Activation and Termination of TERP Macroinitiators and Small Models The Disp/Comb ratio for the termination products was determined by SEC for the polymers or by NMR for the small model radicals. Photolysis in benzene at 25°C yielded a Disp/Comb ratio > 98/2, although a greater impact of Comb was recorder at higher temperatures (e.g. 52/48 for a PMA-TePh with Mn = 3200 g/mol at 120°C). In a previous publication, the same method applied to the study of polymethacrylate and polystyrene radicals (30) gave results (i.e. Disp/Comb = 67/37 and 13/87 at 100 °C for PMMA and PSt, respectively) in agreement with previous reports. The results for polyacrylates, on the other hand, are in stark contrast with some of the older and the more recent reports, as mentioned above. Even more surprisingly, in a follow-up study (31) the Disp/Comb ratio for the termination of PMMA and PSt radicals at various temperatures and in different solvents was found to steadily increase as the medium viscosity increased. For instance, a 97/3 Disp/Comb ratio was recorded for a PMMA-TePh macroinitiator in PEG 400 at 25°C and the same ratio was also recorded for 146

a PSt-TePh in polystyrene (Mn = 96000 g/mol) at 60°C. Although a dominant Disp for PMMA radical is not surprising, PSt radicals were known to undergo preferred termination by Comb, under any conditions, prior to this contribution. As will be detailed below, we have found a more logical interpretation of these results, not requiring to question the established dogma on the PSt, PMMA and PMA bimolecular termination mechanisms, but before detailing this alternative interpretation, we need to come back to CRT. In a subsequent contribution (32), Yamago et al. reported radical termination results using PMA-Br, PMMA-Br and PS-Br macroinitiators activated by CuBr/ Cu0/Me6TREN, i.e. the same strategy of Scheme 3. It was shown that, while the PMMA and PSt radical terminations were as expected on the basis of the known bimolecular radical pathways (i.e. CRT played no role), the outcome of the polyacrylate radical termination was altered by the copper system. Notably, using a PMA-Br with Mn = 2900 and Đ = 1.06 in the presence of CuIBr (1 equiv), Cu0 (4 equiv), and Me6TREN (2 equiv) in toluene at 70 °C, the termination was complete in 1 h and the SEC analysis of the isolated polymer indicated a bimodal peak with a 73/27 ratio of identical and double MWDs, relative to that of the macroinitiator (Figure 7a). This was considered surprising by the authors because, on the basis of their above-mentioned study of the termination from PMA-TePh (29), the PMA radicals should predominantly give the Disp products under these conditions. Therefore, these authors concluded that the higher MW product must result from CRT. Namely, according to these authors, Cu-CRT of polyacrylates leads to Comb, which is opposite to the conclusion of the previous study by Buback, Matyjaszewski et al. (18) Furthermore, the polymer analysis by NMR and MALDI-TOF-MS revealed that the polymer contains only saturated chain ends (PMA-H), whereas Disp would also yield an equimolar amount of macromolecules with an unsaturated chain end. The offered explanation was that, following activation, the PMA radicals are reduced to anions, most likely as an organocopper species, and then quenched to PMA-H by the moisture present in the solvent. This conclusion was further supported by an additional termination study, carried out in the presence of CH3OD (20 equiv) under otherwise identical conditions. This experiment led to the almost exclusive formation of PMA-D (Figure 7b), as confirmed by SEC and MALDI-TOF-MS. This result also implies that the putative organocopper species is quenched faster by CH3OD than terminated according to Cu-CRT. Interestingly, these opposite conclusions concerning the nature of the polymer produced by Cu-CRT by Buback/Matyjaszewski (18) and by Yamago (32) were reached on the basis of the same strategy for the experiments (ATRP activation of a PMA-Br macroinitiator). However, they were based on different assumptions on the nature of the conventional bimolecular termination products: Buback, Matyjaszewski et al. considered that conventional bimolecular termination leads mostly to Comb and concluded that Cu-CRT gives Disp because they only observed Disp-like products, whereas Yamago et al. concluded that Cu-CRT leads to Comb because they also observed a significant extent of Comb, while they had found that the same radicals generated from PMA-TePh in the absence of copper yields essentially only Disp. The two studies, while following the 147

same strategy, are however characterized by certain differences: the Yamago study used Me6TREN as ligand, Cu0 and (wet) toluene as solvent, whereas the Buback/Matyjaszewski study used TPMA as ligand, no Cu0 and MeCN as solvent. As already mentioned above, Cu0 has also been shown to promote, by itself, the termination of polyacrylate radicals (15), whereas the presence of protic species obviously played a role in the termination study reported by Yamago et al. As noted above, water had no effect on the stability of the R-CuII/L organometallic intermediate for primary and tertiary radicals on the time-scale of CV. However, the observed termination with protic species would be indistinguishable from a Cu0 catalyzed termination pathway which has been shown to kinetically dominate both CuI-catalyzed and conventional radical termination in SARA ATRP (15). Further investigations are needed to assess the contribution of protic species on the termination pathways.

Figure 7. SEC profiles of the CuI/Cu0-mediated reaction of PMA-Br without additive (a) and with CH3OD (b). Reproduced with permission from ref. (32). Copyright 2016 American Chemical Society. In order to elucidate the reasons for these observed differences and shed more light onto polyacrylate radical termination, with or without copper catalysts, we have run additional investigations using TPMA, TPMA*3 and Me6TREN in carefully dried MeCN and in the absence of Cu0 (33). The use of Cu0 was avoided, not only to eliminate any contribution of Cu0-CRT, but also to accurately determine the extent of termination from the UV-visible monitoring of the CuII signal. In the presence of Cu0, the deactivator would regenerate CuI by comproportionation. Furthermore, in order to obtain a well-defined [CuI/L]+ activator, this was generated from [CuI(MeCN)4][BF4] rather than from CuIBr, because the CuIBr/L system is known to lead to a complex mixture of different species with different activity in ATRP (34). Finally, MeCN was preferred as solvent because it is known to disfavor [CuI/L]+ disproportionation. An additional key feature of this investigation was the use of variable amounts of added CuII deactivator, which plays no role in CRT but moderates the ATRP activation pseudo-equilibrium and thus alters the radical concentration. The ratio between 148

the conventional (RT) and catalyzed (CRT) termination rates is given by equation 3, showing that an increase of CuII deactivator affects the termination process by favoring CRT. The rate ratio given by equation 3 decreases with time because of the [Pn-X] decrease and the [X-CuII/L] increase. Thus, the relative contribution of CRT to termination keeps increasing as the process goes on.

Using a PMA-Br macroinitiator with Mn = 3300 and Đ = 1.09, Me6TREN as ligand, and the same [PMA−Br]0/[Me6TREN/CuI]0 ratio (1:5) used in the previous study (32), the CuII-free termination led to quantitative (>99%) termination after 30 min at room temperature and the polymer had the SEC bimodal red trace shown in Figure 8 (left). The deconvolution of this trace, shown in Figure 8 (right), yields a 73/27 ratio of low and high MWs in very good agreement with the Yamago contribution (32). However, the polymers recovered from termination experiments run in the presence of increasing amounts of CuII deactivator, i.e. with an increased CRT contribution (Equation 3), led to a decreased proportion of the Comb product. This clearly indicates that the high molecular peak attributed by Yamago et al. to CRT coupling is actually the result of conventional termination (RT) due to the initially high radical concentration, which is caused by the high ATRP activity (high KATRP) of the CuI/Me6TREN activator. The same trend (lower Comb/Disp for lower RT/CRT) was obtained when the macroinitiator was activated by CuI/ TPMA and CuI/TPMA*3 (33). The relative proportion of Comb product under the same concentration conditions was greater in the order TPMA < TPMA*3 < Me6TREN. PREDICI simulations, which included consideration of backbiting, could fit the deactivator concentration and Disp fraction evolution only under the assumption that RT and CRT yield respectively Comb and Disp, whereas no fit was possible when using the opposite assumption. A minor contribution of Disp only occurs via the MCR resulting from backbiting. It thus seems confirmed that the conventional termination of polyacrylate secondary (chain-end) radicals occurs overwhelmingly by combination. At this point, it becomes necessary to clarify the origin of the preferential disproportionation observed from the termination of polyacrylate radicals photogenerated from PMA-TePh at low temperature (29). A good hint is the observed high Disp/Comb for PSt radicals in high viscosity media (31). For a given solvent, the viscosity decreases upon raising the temperature, thus the results of the polyacrylate termination (higher Disp/Comb at lower temperature) may also be determined by the medium viscosity. The Pn-TePh photolysis generates the Pn•/PhTe• radical pair, but in order to achieve radical termination of Pn• (and dimerization of PhTe• to Ph2Te2, e.g. as shown for PMA-TePh in Scheme 5), the radicals must first diffuse away from the solvent cage. Radical escape from solvent cages is known to be highly viscosity dependent (35). It is thus conceivable that the Disp products arise from a side reaction between Pn• and PhTe• within the radical pair cage. 149

Figure 8. (A) SEC profiles of the PMA termination products obtained by the activation of PMA-Br with [CuI(Me6TREN)]+ and variable amounts of [CuIIBr(Me6TREN)]+; (B) deconvolution of the SEC trace for the 1:5:0 experiment. Reproduced with permission from ref. (33). Copyright 2017 American Chemical Society.

In order to probe this hypothesis, we have investigated the Disp/Comb ratio by generating radicals in an alternative way (36). One of the model radicals used in the previously study (31), Me2C•(COOMe), can also be generated thermally or photochemically from the commercially available diazo initiator V-601 as shown in Scheme 6. Contrary to Me2C(COOMe)(TeMe), which forms a Me2C•(COOMe)/PhMe• radical pair, decomposition of V-601 produces two identical radicals within the same radical pair cage, thus the termination product distribution should not greatly depend on the rate of radical escape from the cage. The results of the experiments are shown in Figure 9, in comparison with those of the previous Me2C(COOMe)-TeMe study (31).

Scheme 6. Two Independent Ways to Generate the Same Me2C•(COOMe) Radical 150

Figure 9. Fraction of disproportionation vs. (A) temperature and (B) viscosity for the isobutyryl radical in benzene. Reproduced with permission from ref. (36). Copyright 2017 Wiley-VCH.

Quite clearly, the strong solvent, temperature, and viscosity dependence of Disp/Comb when the radical is generated from the organotellurium precursor is not present when it is generated from V-601. Notably, high viscosity media yield a lower Disp/Comb ratio for the radicals generated from V-601. The greater Disp fraction obtained in the presence of the tellanyl radical is proposed to result from a tellanyl-catalyzed process, see Scheme 7. DFT calculations on model systems, carried out with both a methacrylate and an acrylate radical, show that the two key steps, the β-H atom transfer to yield the H-TeR intermediate (a) and the H atom transfer from H-TeR to a second organic radical (b), are energetically downhill processes with very small energy barriers (36).

Scheme 7. Proposed Mechanism for C-Based Radical Disproportionation Catalyzed by RTe•. (i) Radical pair generation; (ii) solvent cage escape; (iii) bimolecular radical termination. 151

Effect of the Metal Nature We have wondered about the fundamental reason making certain metal centers (i.e. CuI, FeII) catalytically active in CRT, whereas CoII leads to CCT or OMRP-DT when exposed to an overstoichiometric amount of radicals (cf. OMRP-DT and CRT pathways in Scheme 1). One fundamental difference between these two classes of metal centers is that those with CRT activity lead to a paramagnetic organometallic intermediate (L/CuII-Pn: S = 1/2; L/FeIII-Pn: S = 5/2), whereas the cobalt system leads to a diamagnetic (S = 0) L/CoIII-Pn dormant species. A DFT calculation of models for these dormant species for the methyl acrylate polymerization (see Figure 10, upper part) shows that the spin density is not completely localized on the metal center. The weakness of the bond leaves a substantial amount of spin density on the alkyl chain, partially delocalized on the ester carbonyl O atom (by resonance) and on the β-H atom located anti relative to the metal atom (by hyperconjugation). The other two β-H atoms do not bear any spin density. Of course, in the real dormant species the polymer chain replaces one of the three β-H atoms, preferably the anti one for steric reasons, but there may be a certain probability to have the chain in a gauche position and an anti H atom. Thus, although the radical chain trapping (OMRP-RT equilibrium) disfavors RT by lowering the radical concentration, the chains radical reactivity is maintained, opening access to new pathways for termination. The dormant species can be drawn as a resonance of four limiting structures, with the unpaired electron localized on the metal, carbon, carbonyl oxygen and β-hydrogen atom, respectively (lower part of Figure 10). This leads to four possible positions of attack by the second radical chain, hence four pathways for CRT. For the diamagnetic [L/CoIII-Pn] system, on the other hand, the second radical may only attack a metal orbital and since the two chains are located in a relative trans position, they can only lead to degenerative exchange and not to coupling by reductive elimination. Of the four possible pathways, the first two lead to coupling. The first one, because the new bond formed between L/Mtx+1(Pn) and Pm•, yielding a [L/Mtx+2(Pn)(Pm)] intermediate, opens access to reductive elimination of Pn-Pm if the two chains occupy cis relative position in the metal coordination sphere. We can therefore imagine that control over CRT vs. OMRP-DT, even for a diamagnetic system, depends on the geometrical details of the metal coordination sphere. The second pathway involves direct bond formation between Pn• and Pm•. It is to be underlined that these two reactivity pathways have documented precedents, or have been proposed, in organometallic radical reactivity (37).

152

Figure 10. Upper part: Mulliken spin densities on selected atoms in the DFT-optimized geometries of [L/Mtx+1]-CH(CH3)COOCH3 ([L/Mtx+1] = [CuII(TPMA)] +, left; [FeIIIBr3]-, center; [CoIII(porphyrin)], right. Lower part: possible pathways leading to CRT.

The third pathway seems unprecedented in organometallic reactivity. However, a previous molecular dynamics computational study has shown this as a possibility for the polyacrylate termination in free-radical polymerization (29). For PMA, the C-O bonded intermediate corresponds to Pn-1CH2CH=C(OMe)O-Pm, from which a facile intramolecular rearrangement, leading to the Disp products Pn-H and Pm-1CH=CHCOOMe, can be envisaged. Finally, the fourth pathway in Figure 10 would lead directly to the Disp products Pn-1CH=CHCOOMe and Pm-H. Apparently, this reaction type is also unprecedented in organometallic chemistry. Since the studies detailed in the previous sections seem to suggest that CRT leads to Disp products, the first two pathways appear excluded and the choice is restricted to the 3rd and 4th pathway in Figure 10 (however, see next section). In order to distinguish between these two possibilities, we have conceived an isotope labelling experiment. Use of a deuterated macroinitiator obtained from CD2=CDCOOCH3 (d3-MA), namely p(d3-MA)-Br, would lead to the same products (PMA-D and PMA-CD=CDCOOCH3), independently of the atom being attacked (O or β-D). Therefore, the two pathways cannot be distinguished from the product analysis. However, the β-D attack is expected to be accompanied by a significant kinetic isotope effect (KIE), whereas the O-attack should not, under the assumption that the rate-determining step of CRT is the reaction between the organometallic intermediate and a second radical. 153

Involvement of Solvent Moisture Before undertaking an accurate kinetic study of the isotope effect, the nature of the terminated polymers was analyzed by SEC, 1H NMR and MALDI-TOF-MS, giving unexpected results using [CuI(TPMA)][Br] as catalyst (similar conditions used in the previous studies (18, 27, 33)). At complete termination of pMA-Br in MeCN, SEC revealed that 98% of chains had the same molecular weight as the pMA-Br macroinitiator with only 2% of chains showing coupling (Figure 11A), whereas the MALDI-TOF MS analysis of this polymer showed that it contains essentially only saturated chain-ends, without unsaturations (Figure 11B). This result is similar to that reported by Yamago et al. for their study in wet toluene, thus we reasonably wondered about the presence of water in the MeCN solvent, although we also considered the possible involvement of the weak C-H bond of MeCN in a hydrogen atom transfer (HAT) mechanism (Scheme 1). A termination of pMA-Br in d3-MeCN and DMF under the same conditions also showed only pMA-H saturated chains with only a small presence of coupled chains and the termination of p(d3-MA)-Br gave p(d3-MA)-H with no unsaturated species or deuterium-capped chains, as indicated by MALDI-TOF-MS. This excludes the HAT mechanism and confirms that residual solvent moisture may play an important role under these diluted conditions (needed to accurately measure the CuII concentration by UV-visible spectroscopy).

Figure 11. Termination of pMA-Br in A) MeCN or B) d3-MeCN under the initial conditions [pMA-Br]:[CuBr]:[TPMA] = 1:2:2.1 at 25°C; [pMA-Br]0 = 10 mM.

Additional termination experiments were also conducted with the ATRP initiator methyl 2-bromopropionate (MBrP) as an acrylate model under the same conditions as the polymeric analog in MeCN. This species does not undergo backbiting, allowing to remove one possible complication. The NMR analysis of the products in both d3-MeCN and d7-DMF showed that 80% of the terminated radicals were coupled, as a result of bimolecular RT, while the rest were saturated, CH3CH2COOMe. There was no evidence of disproportionation either from vinyl peaks or the formation of oligomers. In this case, possibly because of a different amount of residual moisture in the solvent, the impact of hydrolytic decomposition 154

was lower. Also of relevance is the absence of succinonitrile, NCCH2CH2CN, and of the coupling product between the cyanomethyl and acrylate radicals, CH3CN(CH2CN)COOMe, once again discarding any Cu-catalyzed HAT from the solvent (Scheme 1). These results may suggest that CRT is either less effective for the unimer radical generated from MBrP, or also leads preferentially to combination. Note also that the bimolecular termination rate constant, kt, is 10 times slower for polymeric radicals than for small molecules (38). Incidentally, this result further supports the notion that the RT of acrylates proceeds via combination (33). These results clearly show that there are multiple pathways for the termination of acrylate radicals in ATRP: 1) bimolecular termination of chain-end radicals, predominantly by combination; 2) bimolecular termination of mid-chain radicals (from backbiting), mostly by disproportionation; 3) CRT, proceeding via the L/ CuII-Pn organometallic species by a yet unknown mechanism and 4) hydrolysis of the L/CuII-Pn species by adventitious moisture. These pathways are summarized in Scheme 8. The intimate details of the Cu-CRT are currently being investigated.

Scheme 8. Pathways of Acrylate Radical Termination in Copper-Catalyzed ATRP

Critical Evaluation of CRT The recent discovery of the moisture involvement in the CuI-CRT studies, detailed in the previous section, leads us to reconsider all previous conclusions reached by ourselves and others on the CRT mechanism and on the nature of the dead chains (Comb vs. Disp). Most previous studies of CRT carried out under 155

both catalytic and stoichiometric conditions in our own laboratories were done in the presence of MeCN, either as solvent or from the CuI salt (16, 18, 27, 33). The MWD of the terminated polymers obtained from the activation of Pn-Br under stoichiometric conditions shows that the major component has a MWD identical to that of the macroinitiator, initially interpreted as resulting from Disp (18). In light of the more recent findings by Yamago et al. (32) and ourselves (previous section), all those MWDs were probably resulting from a stoichiometric hydrolytic decomposition of the organocopper(II) intermediate with the sole formation of PnH. The existence of CRT, however, is clearly proven by the polymerization studies carried out under OMRP conditions with catalytic quantities of copper complex and must lead to bimolecular radical products (Disp or Comb). The presence of adventitious water has no effect on Cu-catalyzed ATRP, it only consumes the copper catalyst by hydrolyzing the organocopper(II) intermediate to generate a putative CuII(OH) species, which may however be reinjected into the ATRP system in the presence of suitable reducing agents. The FeII-CRT of polyacrylate radicals reported by Buback et al., investigated in bulk monomer, also proves the presence of CRT. Unfortunately, that study gave no information on the product MWD (21). Clearly, more investigations are necessary on the nature of the polymer (MWD) obtained in the presence of stoichiometric amounts of activator, under conditions in which moisture is completely avoided or at least drastically reduced relative to copper, in order to fully elucidate the CRT mechanism (these experiments are ongoing in our laboratories).

Conclusion The present account outlines the complicated scenario related to the important CEF problem in the ATRP of acrylate monomers, resulting from radical terminations. A contribution of Cu-CRT has clearly been illustrated by recent reports. Progress in this area has been somewhat polluted by contrasting evidences and subsequent debates on the nature of the termination products from both the spontaneous bimolecular RT and CRT (combination or disproportionation), as well as by water contamination of the solvents, leading to hydrolytic decomposition of the organometallic intermediate. While the spontaneous bimolecular RT was shown to consist of combination for secondary chain-end radicals and disproportionation for tertiary mid-chain radicals that result from back-biting processes, the contribution of solvent moisture in the generation of saturated chains during stoichiometric experiments has haunted so far the determination of the Cu-CRT mechanism. An initially suspected HAT from the MeCN solvent has not been substantiated and does not seem to occur to a significant extent. We hope that the present account, as well as further investigations currently ongoing in our laboratories, will contribute to elucidate all intimate details of the radical termination in the acrylate ATRP, allowing optimization of the catalyst and of the operating conditions in order to obtain polymers with the highest possible CEF. 156

Abbreviations AIBN ATRP BA BDFE BPMA CCT CEF Comb CRT DFT Disp EBiB HAT KIE MA MALDI-TOFMS MBrP MCR Me6TREN MW MWD MMA NMR OMRP-RT OMRP-DT PEG RDRP-DT SEC SPR St TPMA

Azobis(isobutyronitrile) Atom transfer radical polymerization n-Butyl acrylate Bond dissociation free energy Bis(2-pyridylmethyl)amine Catalytic chain transfer Chain-end functionality Combination Catalyzed radical termination Density functional theory Disproportionation Ethyl 2-bromoisobutyrrate Hydrogen atom transfer Kinetic isotope effect methyl acrylate Matrix-assisted laser desorption ionization – time-of-flight – mass spectrometry Methyl 2-bromopropionate Midchain radical tris(2-dimethylaminoethyl)amine Molecular weight Molecular weight distribution methyl methacrylate Nuclear magnetic resonance Organometallic-mediated radical polymerization by reversible termination Organometallic-mediated radical polymerization by degenerative transfer Poly(ethylene glycol) Reversible deactivation radical polymerization Size exclusion chromatography Secondary propagating radical Styrene tris(2-pyridylmethyl)amine

Acknowledgments We thank the Centre National de la Recherche Scientifique (CNRS) for funding our collaborative project in the form of a PICS grant (No. 6782, 2015-17) and then for the establishment of a “Laboratoire Internation Associé” (LIA) called “Laboraty of Coordination Chemistry for Controlled Radical Polymerization” (2018-21). The computational work was granted access to the HPC resources of IDRIS under the allocation 086343 made by GENCI (Grand Equipement National de Calcul Intensif) and to the resources of the CICT (Centre Interuniversitaire de Calcul de Toulouse, project CALMIP). Additional support from the French 157

Embassy in Washington D.C. (Chateaubriand fellowship to TGR), from the ANR grant FLUPOL (ANR-14-CE07-0012, 2015-18), from the Mational Science Foundation (CHE 1707490), and from the Institut Universitaitre de France (IUF) is also gratefully acknowledged.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

Matyjaszewski, K. Macromolecules 2012, 45, 4015–4039. Poli, R. Organometallic Mediated Radical Polymerization. In Reference Module in Materials Science and Materials Engineering; Elsevier, 2016. Poli, R. Angew. Chem., Int. Ed. 2006, 45, 5058–5070. Poli, R. Eur. J. Inorg. Chem. 2011, 1513–1530. Poli, R. Chem. Eur. J. 2015, 21, 6988–7001. Le Grognec, E.; Claverie, J.; Poli, R. J. Am. Chem. Soc. 2001, 123, 9513–9524. Braunecker, W. A.; Brown, W. C.; Morelli, B.; Tang, W.; Poli, R.; Matyjaszewski, K. Macromolecules 2007, 40, 8576–8585. Poli, R.; Shaver, M. P. Inorg. Chem. 2014, 53, 7580–7590. Wayland, B. B.; Peng, C.-H.; Fu, X.; Lu, Z.; Fryd, M. Macromolecules 2006, 39, 8219–8222. Maria, S.; Kaneyoshi, H.; Matyjaszewski, K.; Poli, R. Chem. Eur. J. 2007, 13, 2480–2492. Santhosh Kumar, K. S.; Gnanou, Y.; Champouret, Y.; Daran, J.-C.; Poli, R. Chem. Eur. J. 2009, 15, 4874–4885. Gridnev, A. A.; Ittel, S. D. Chem. Rev. 2001, 101, 3611–3659. Slavin, S.; McEwan, K.; Haddleton, D. M. Cobalt-Catalyzed Chain Transfer Polymerization: A Review. In Polymer Science: A Comprehensive Reference; Matyjaszewski, K., Möller, M., Eds.; Elsevier BV: Amsterdam, 2012; Vol. 3, pp 249−274. Matyjaszewski, K.; Woodworth, B. E. Macromolecules 1998, 31, 4718–4723. Krys, P.; Wang, Y.; Matyjaszewski, K.; Harrisson, S. Macromolecules 2016, 49, 2977–2984. Schröder, K.; Konkolewicz, D.; Poli, R.; Matyjaszewski, K. Organometallics 2012, 31, 7994–7999. Kaur, A.; Ribelli, T. G.; Schroeder, K.; Matyjaszewski, K.; Pintauer, T. Inorg. Chem. 2015, 54, 1474–1486. Wang, Y.; Soerensen, N.; Zhong, M.; Schroeder, H.; Buback, M.; Matyjaszewski, K. Macromolecules 2013, 46, 683–691. Fischer, H. Chem. Rev. 2001, 101, 3581–3610. Tang, W.; Tsarevsky, N. V.; Matyjaszewski, K. J. Am. Chem. Soc. 2006, 128, 1598–1604. Schroeder, H.; Buback, M. Macromolecules 2014, 47, 6645–6651. Ayrey, G.; Humphrey, M. J.; Poller, R. C. Polymer 1977, 18, 840–844. Szablan, Z.; Junkers, T.; Koo, S. P. S.; Lovestead, T. M.; Davis, T. P.; Stenzel, M. H.; Barner-Kowollik, C. Macromolecules 2007, 40, 6820–6833. 158

24. Bamford, C. H.; Jenkins, A. D. Nature 1955, 176, 78–78. 25. Ballard, N.; Hamzehlou, S.; Ruiperez, F.; Asua, J. M. Macromol. Rapid Commun. 2016, 37, 1364–1368. 26. Rahaman, S. M. W.; Matyjaszewski, K.; Poli, R. Polym. Chem. 2016, 7, 1079–1087. 27. Ribelli, T. G.; Rahaman, S. M. W.; Daran, J.-C.; Krys, P.; Matyjaszewski, K.; Poli, R. Macomolecules 2016, 49, 7749–7757. 28. Zerk, T. J.; Bernhardt, P. V. Inorg. Chem. 2017, 56, 5784–5792. 29. Nakamura, Y.; Lee, R.; Coote, M. L.; Yamago, S. Macromol. Rapid Commun. 2016, 37, 506–513. 30. Nakamura, Y.; Yamago, S. Macromolecules 2015, 48, 6450–6456. 31. Nakamura, Y.; Ogihara, T.; Hatano, S.; Abe, M.; Yamago, S. Chem. Eur. J. 2017, 23, 1299–1305. 32. Nakamura, Y.; Ogihara, T.; Yamago, S. ACS Macro Lett. 2016, 5, 248–252. 33. Ribelli, T. G.; Augustine, K. F.; Fantin, M.; Krys, P.; Poli, R.; Matyjaszewski, K. Macromolecules 2017, 50, 7920–7929. 34. De Paoli, P.; Isse, A. A.; Bortolamei, N.; Gennaro, A. Chem. Commun. 2011, 47, 3580–3582. 35. Khudyakov, I. V. Res. Chem. Intermed. 2013, 39, 781–804. 36. Ribelli, T. G.; Rahaman, S. M. W.; Matyjaszewski, K.; Poli, R. Chem. Eur. J. 2017, 23, 13879–13882. 37. Crossley, S. W. M.; Obradors, C.; Martinez, R. M.; Shenvi, R. A. Chem. Rev. 2016, 116, 8912–9000. 38. Fischer, H.; Radom, L. Angew. Chem., Int. Ed. 2001, 40, 1340–1371.

159

Chapter 7

Electrochemical Procedures To Determine Thermodynamic and Kinetic Parameters of Atom Transfer Radical Polymerization Francesca Lorandi,1 Marco Fantin,2 Francesco De Bon,1 Abdirisak A. Isse,1 and Armando Gennaro*,1 1Department

of Chemical Sciences, University of Padova, Via Marzolo 1, 35131 Padova, Italy 2Department of Chemistry, Carnegie Mellon University, 4400 Fifth Avenue, Pittsburgh, Pennsylvania 15213, United States *E-mail: [email protected].

Electrochemical investigation provides information about the stability, activity and halidophilicity of catalysts for Atom Transfer Radical Polymerization (ATRP). Moreover, several electrochemical tools were developed to measure thermodynamic and kinetic parameters of ATRP. These techniques enabled to determine activation rate constants spanning over 12 orders of magnitude. ATRP equilibrium constant and relevant side reactions concerning the catalyst were also examined by electrochemical methods. As such, electrochemistry enables to build a database of kinetic and thermodynamic constants of ATRP and related reactions of copper complexes.

Introduction In atom transfer radical polymerization (ATRP), the equilibrium between propagating radicals and dormant species is governed by the redox couple XCuIIL+/CuIL+ (L = polydentate amine, X = halogen). CuIL+ generates radicals by reductive cleavage of the C−X bond in the initiator (RX) or dormant chain (PnX). The C−X cleavage is coupled with atom transfer to the Cu complex (Scheme 1). The ensuing oxidized form of the catalyst XCuIIL+ deactivates radicals after they add to few monomer (M) molecules (1). © 2018 American Chemical Society

Scheme 1. General ATRP Mechanism

The equilibrium constant of ATRP is defined by the ratio between the activation and deactivation rate constants, i.e. KATRP = kact/kdeact. These quantities are related to the rate of polymerization (Rp) and to the dispersity of the obtained polymer (Eq. 1 and Eq. 2, respectively).

where kp is the propagation rate constant, DP is the degree of polymerization and p is the monomer conversion. kp depends on temperature and on the nature of monomer and solvent. kp is typically measured by well-established techniques such as pulsed laser initiated polymerization coupled with size exclusion chromatography (PLP-SEC), whereas PLP combined with time-resolved near-infrared spectroscopy (2), in the single pulse (SP)-PLP method, provides accurate values of the termination rate constant (kt) of polymer chains (3). KATRP, kact, and kdeact values are affected by temperature and by nature of catalyst, initiator, solvent, monomer, and the monomer/solvent ratio. ATRP equilibrium constants were determined by GC or UV-Vis spectroscopy (4, 5). Activation rate constants were measured by gas-chromatography, HPLC, NMR, or UV-Vis spectroscopy (6–8). These techniques are compatible only with moderately slow reactions, thus stopped-flow methods were adopted for faster systems (9, 10). kdeact values are generally very high, approaching the diffusion-controlled limit, therefore only few experimental methods were developed for their direct determination. Some kdeact values were obtained from EPR measurements (11) or from radical clock reactions with radical trapping agents (7). More commonly, kdeact was determined by independently measuring kact and KATRP. In the last decade, several electrochemical tools were developed to measure kact, kdeact, and KATRP over a large range of values spanning several orders of magnitude. Minimal amount of reactants are required, and highly accurate values are generally determined within a short measurement time. The most modern and active complexes can be easily investigated by voltammetric techniques. These methods are presented in this chapter. The contribution of electrochemistry to ATRP has created a large database of kinetic and thermodynamic data that has enhanced the comprehension of the polymerization process, providing tools to 162

select the appropriate experimental conditions for the desired polymerization result.

Mechanism of ATRP Activation In ATRP, three possible pathways can describe the formation of the propagating radical: i) outer-sphere electron transfer generating intermediate RX•− (OSET-SW), ii) OSET concerted with C−X bond rupture (OSET-C), and iii) inner-sphere electron transfer or atom transfer (ISET-AT) (inset in Figure 1).

Figure 1. Comparison of free energies during ISET-AT and OSET-C processes for the reaction of bromoacetonitrile with CuITPMA+ in acetonitrile at 25 °C Inset: possible mechanisms of electron transfer in ATRP. Reprinted with permission from ref. (12). Copyright 2008 American Chemical Society.

Theoretical and electrochemical analysis of the activation process of a wide series of alkyl halides, mimicking the chain end of the macromolecular dormant species, enabled to define the real mechanism of the activation reaction (12). The OSET-SW pathway was discarded because it was experimentally proved that reductive cleavage of alkyl halides used as initiators in ATRP follows a concerted mechanism. ISET-AT and OSET-C have similar thermodynamic requirements; therefore, they were discriminated by considering their relative transition state energies. The rate constant of the hypothetical OSET-C reaction between bromoacetonitrile and CuITPMA+ complex was estimated using Marcus theory for ET with correction for in-cage interactions of ion-dipole fragments (the sticky model) (13, 14). The calculated kOSET-C ≈ 10-11 M-1 s-1 was more than 12 orders of magnitude smaller than the experimental value (kact ≈ 82 M-1 s-1 at 25 °C in CH3CN), pointing out that the correct activation mechanism is via ISET-AT. 163

This conclusion was confirmed by measuring the activation rate constants of reduction of some alkyl halide initiators by outer sphere electron donors (i.e. electrogenerated organic radical anions, A•−), and by comparing them with kact by some CuIL+ complexes (15).

Figure 2. A comparison between rate constants, k, of RX activation by aromatic radical ions (dots) and CuIL+ complexes (squares), measured in CH3CN + 0.1 M Et4NBF4 at 25 °C. The lines are the best-fitting curves for dissociative electron transfer by A•−. Reprinted with permission from ref. (15). Copyright 2013 Elsevier.

Figure 2 shows a comparison of kact values measured obtained for three CuIL+ complexes. In all cases, activation orders of magnitude faster than activation by outer-sphere the same standard potential of the copper catalysts, which mechanism for Cu complexes.

for A•− with those by CuIL+ was 7-10 electron donors of confirms the ISET

Cyclic Voltammetry of ATRP Catalysts Cyclic voltammetry (CV) is a versatile screening tool for ATRP catalysts and initiators. Binary Cu/L and ternary X/Cu/L complexes exhibit a quasi-reversible voltammetric peak couple from which their standard reduction potential is measured as the semi-sum of cathodic (Epc) and anodic (Epa) peak potentials: E ≈ E1/2 = (Epc + Epa)/2. E of X/Cu/L is generally more negative than E of the corresponding Cu/L (Figure 3a), and both are considerably more negative than E of the solvated Cu(II) salt:

. 164

As a rule, the more negative E, the higher the catalytic activity of the complex in ATRP (5). E generally shifts to more positive values if the temperature is raised, resulting in a decrease of the reducing power of the catalyst (16). However, monomer propagation is favored at high T, thus ATRP rate increases with T. CV conducted in water at different pHs showed that common amine ligands may be protonated by decreasing the pH, thus losing their ability to stabilize the copper center. Instead, the CuIIL(OH)+ complex forms at basic pH (Figure 3b) (17) because OH– binds copper more strongly than X−.

Figure 3. (a) Cyclic voltammetry (CV) of 1 mM CuIIMe6TREN2+ in the absence and presence of 2 mM Et4NBr or Et4NCl, in DMSO + 0.1 M Et4NBF4, v = 0.2 V s-1, T = 25 °C. (b) CV of 1 mM CuIITPMA2+ at different pH, in water + 0.1 M Et4NBF4, v = 0.2 V s-1, T = 25 °C. Adapted with permission from ref. (17). Copyright 2015 American Chemical Society. (c) CV of 1 mM CuIIMe6TREN2+, CuIITPMA2+, CuIIPMDETA2+ in [BMIm][OTf], v = 0.1 V s-1, T = 50 °C. Adapted with permission from ref. (18). Copyright 2017 Elsevier. (d) CV of 1 mM BrCuIITPMA2+ with increasing amount of sodium dodecyl sulfate (SDS) in water + 0.1 M NaBr. v = 0.1 V s-1, T = 65 °C. Adapted with permission from ref. (19). Copyright 2017 American Chemical Society. 165

Electrochemical analysis in Ionic Liquids (ILs) can be performed without adding a supporting electrolyte, because of the relatively high ionic conductivity of these media. CV of common ATRP catalysts and initiators in 1-butyl-3-methylimidazolium trifluoromethanesulfonate showed similar features to the ones observed in traditional organic solvents (Figure 3c) (18). Voltammetric studies were carried out also in an oil-in-water miniemulsion, formed with sodium dodecyl sulfate (SDS) surfactant. E of BrCuIITPMA+ in the miniemulsion shifted to more negative values, compared to pure water, while the peak currents decreased (19). A similar effect was found in water, when adding increasing amounts of SDS (Figure 3d), suggesting that this surfactant interacted with the catalyst, enhancing the stabilization of the CuII species. CVs of alkyl halides used as ATRP initiators show an irreversible bi-electronic reduction peak, with a peak potential Ep (Figure 6), followed by a reversible peak couple due to the reduction of the remaining CuIIL2+ molecules not involved in the catalytic process. The position of the prepeak is directly correlated to kact (Eq. 14).

By applying Eq. 14, kact was calculated directly from CVs recorded at various CRX and/or scan rates. The effect of these two parameters on the voltammetric response is depicted in Figure 6. The intensity of the prepeak increased and Ep shifted toward more negative values by either increasing CRX or v. Under the total catalysis regime, a slope of about -30 mV was obtained by plotting Ep vs. logCRX or vs. logv, in agreement with Eq. 14.

Figure 6. Cyclic voltammetry of 10−3 M CuIITPMA2+ recorded in H2O + 0.1 M Et4NBF4 at: a) 0.05 V s−1 in the absence and presence of 2-hydroxyethyl-2-bromoisobutyrate (HEBiB) with CHEBiB = 2 × 10−4, 3 × 10−4, 5 × 10−4 and 7 × 10−4 M; b) different scan rates 0.01, 0.02, 0.03, 0.05, 0.07, and 0.1 V s−1 in the presence of 5 × 10−4 M HEBiB. Insets: variation of prepeak potential, Ep, with log CHEBiB or log v. Adapted with permission from ref. (20). Copyright 2017 American Chemical Society.

The total catalysis technique does not require a radical scavenger, because all RX molecules are rapidly converted to radicals in a thin layer close to the electrode surface, thus the termination rate is enhanced and the activation process becomes irreversible. However, this simple method requires clear and well-reproducible voltammetric responses. 177

CV under total catalysis conditions was used in water (Table 4) to measure kact values as high as 2.6 × 107 M-1 s-1, and very recently to study the activation of ethyl α-bromoisobutyrate in acetonitrile by the extremely active, new ATRP catalyst CuII(TPMANMe2) (30).

Table 4. Activation Rate Constants of Various Catalyst/Initiator Systems Measured by Cyclic Voltammetry under Total Catalysis Conditions, T = 25 °C RX

Ligand

a

Solvent

kact (M-1 s-1)

Ref.

Me6TREN

HEBiBa

H2O

2.6 × 107

(20)

TPMA

HEBiBa

H2O

5.4 × 106

(20)

TPMANMe2

ethyl α-bromoisobutyrate

CH3CN

7.2 × 106

(30)

2-hydroxyethyl-2-bromoisobutyrate.

Figure 7. Background-subtracted experimental and simulated voltammograms of CuIIMe6TREN2+ in the presence of bromoacetonitrile recorded at 0.5 and 1, in CH3CN + 0.1 M Et4NBF4 at 25 °C. Adapted with permission from ref. (24). Copyright 2017 Elsevier. 178

d. Cyclic Voltammetry with Digital Simulation kact of both slow and fast ATRP systems was measured by digital simulations of CV (20, 24, 31). Experimental and digital CVs have been overlapped, having previously subtracted the background in the experimental signals (20, 24). The set of reactions required to simulate the electrochemical response is identical to the one used for the HRC method (Scheme 2), thus all parameters, except kact, should be measured independently, or be available in the literature. If some parameters are not easily accessible, they can be left as adjustable variables in the simulation together with kact, although at the expense of a greater uncertainty over the obtained kact values. This technique requires the presence of TEMPO, at least for slow activation processes. An example of a good match obtained by overlapping experimental and digital CVs is reported in Figure 7.

Table 5. Activation Rate Constants of Various Catalyst/Initiator Systems Measured by Cyclic Voltammetry with Digital Simulation, T = 25 °C Ligand

TPMA

Me6TREN

RX

Solvent

H2O

oligo(ethylene oxide) 2bromopropionate ethyl α-bromoisobutyrate

1.3 × 105

CH3CN

3.7 × 104

DMSO

8.7 × 104

(20)

(31)

2.4 × 103

methyl 2-bromopropionate

DMSO

2.5 × 102

(32)

4.5 × 101

benzyl bromide

Me6TREN

Ref.

5.4 × 106

2-hydroxyethyl-2bromoisobutyrate

ethyl α-bromoisobutyrate PMDETA

kact (M-1 s-1)

bromoacetonitrile

CH3CN

1.0 × 105

bromoacetonitrile

DMSO

3.6 × 105

chloroacetonitrile

CH3CN

5.9 × 102

CH3CN

4.0 × 104

DMSO

1.8 × 104

CH3CN

3.6 × 104

DMSO

2.4 × 104

bromoacetonitrile TPMA ethyl α-bromoisobutyrate

(33)

(33)

Digital simulation of CV has been applied to both slow and fast activation reactions, measuring kact in the range 10-108 M-1s-1, in water, acetonitrile, and DMSO (Table 5). In contrast, the total catalysis method was valid in a limited range of kact, CRX, and v values. 179

Electrochemical Determination of kdeact The accurate measurement of the ATRP deactivation rate constant is difficult because of its high value, generally close to the diffusion-controlled limit. An electrochemical method based on comparing experimental catalytic CVs with simulated voltammograms was used to measure few kdeact values in DMSO (32). First, cyclic voltammetry in the presence of TEMPO, coupled with digital simulation enabled to determine kact, as described in the previous section. The set of reactions reported in Scheme 2 was used in the simulations, with kact as the only unknown parameter. The value of kdeact had no effect on this experiment, because TEMPO quenched all available radicals. The same experiment was then repeated in the absence of TEMPO and simulated by using the just calculated kact, while refining the kdeact value. Indeed, when no radical scavenger was added, the deactivation reaction occurred and the recorded catalytic current Ip decreased, due to the backward reaction between XCuIIL2+ and radicals. The same set of scan rates and ratios was used in both experiments, thus the only difference was the absence or presence of the deactivation process. The procedure was used on 3 different initiators, with BrCuIIPMDETA2+ as catalyst, in DMSO (Table 6) (32).

Table 6. Deactivation Rate Constants Measured by Cyclic Voltammetry with Digital Simulation in DMSO, T = 25 °C RX

Ligand

PMDETA

kdeact (M-1 s-1)

Ethyl α-bromoisobutyrate

1.8 × 106

Methyl 2-bromopropionate

7.6 × 106

Benzyl bromide

8.6 × 105

It must be considered that side reactions may hamper these kinetic analysis. In particular, the CuI catalyzed radical termination (CRT, Scheme 3) should be considered (34).

Scheme 3. Mechanism of CuI Catalyzed Radical Termination (CRT) The CRT mechanism involves the reversible formation of an organometallic adduct, R-CuII, followed by termination with a second radical species. If TEMPO is present, CRT has no effect, because radicals are trapped before forming the organometallic intermediate. However, in the absence of any radical scavenger, CRT can significantly alter the shape and peak current of the voltammetric 180

response. Therefore, reported kdeact values should be refined considering the influence of CRT on primary and secondary alkyl radicals, whereas tertiary radicals are substantially unaffected by this reaction (33, 35).

Electrochemical Determination of KATRP Chronoamperometry at a rotating disk electrode was used to determine KATRP, by monitoring the disappearance of CuIL+ during the reaction between CuIL+ and RX, and by applying a modified Fischer’s equation (23). This equation relates the concentration of XCuIIL+ to the ATRP equilibrium constant:

where Y is the concentration of the “persistent radical” XCuIIL+, C0 and I0 are the initial concentrations of CuIL+ and RX, and c′ and c″ are constants. Eq. 15 is valid for C0 ≠ I0, whereas it reduces to Eq. 16 if C0 = I0 (36). The experimental procedure was identical to the one reported for kact determination via RDE, except that TEMPO was not added to the solution. Therefore, activation, deactivation, and radical termination occur during the measurement of KATRP. Side reactions, and particularly CRT, must be avoided. The ratio between the rates of radical-radical termination (RT) and CRT is given by Eq. 17.

Since kt and kCRT are generally of the order of 109 M-1s-1 and 104 M-1s-1 (34), respectively, whereas KATRP for active catalysts in polar solvents is in the range ratio is required to suppress CRT. Therefore, a large 10-4-10-6, a high CRX excess of RX with respect to the initial amount of CuI was used to minimize the = 50, contribution of CRT. KATRP values in Table 7 were obtained using which ensured CRX > 50, especially during the initial stages of the reaction (36). KATRP values of CuIMe6TREN+ and CuITPMA+ with methyl 2-bromopropionate (acrylate mimic) were measured in CH3CN, DMF, and mixtures with butyl acrylate (Table 7). The presence of the monomer did not alter the reaction scheme because with such active catalysts and no deactivators most radicals terminated immediately and monomer conversion was negligible. Interestingly, for relatively slow systems it was possible to determine KATRP and kact in one experiment, by monitoring first the reaction between CuIL+ and RX, and then adding a concentrated solution of TEMPO to trap all radicals and isolate the activation step by suppressing deactivation. 181

It should be noticed that literature values for KATRP (36) were generally smaller than the ones determined by RDE (23). This is likely due to the presence of Br− or Cl− ions in previous measurements. These halide ions bind to CuI by formation of species, thus decreasing the amount of the active catalyst, inactive XCuIL or CuIX2−. The presence of these species slows down the process. Therefore, absolute values of KATRP and kact should be measured in the absence of halide ions. The simplicity and high reproducibility (5-10 % error) of this technique offer a valuable tool to investigate various catalyst/initiator systems, thus increasing the understanding of the ATRP mechanism.

Table 7. ATRP Equilibrium Constants of Various Systems Measured by Rotating Disk Electrode, T = 25 °C Ligand

Me6TREN

RX

methyl 2-bromopropionate

ethyl 2-chloropropionate

TPMA

methyl 2-bromopropionate

Solvent

KATRP

CH3CN

1.1 × 10-5

CHCH3CN /BA 1/1 v/v

7.2 × 10-6

DMF

1.3 × 10-4

DMF/BA 1/1 v/v

1.9 × 10-5

CH3CN

2.3 × 10-6

CH3CN

3.5 × 10-6

CHCH3CN/BA 1/1 v/v

1.8 × 10-6

DMF

1.6 × 10-5

DMF/BA 1/1 v/v

1.7 × 10-6

Estimation of Halidophilicity Constants from ATRP Measurements If halogen ions are present during kinetic analysis of the ATRP process, the measured parameter is an apparent constant, whose value depends on . were measured with increasing under Indeed, decreasing values of otherwise identical conditions (Figure 8a). The drop in the overall rate of current decay is due to the reduced availability of CuIL+ owing to the speciation of ratios, CuI involving both active CuIL+ and inactive XCuIL (at higher additional inactive halogenated CuI species are formed). Therefore, the decrease values can be used to measure the halidophilicity constant for CuIL+, of by correlating KATRP to This procedure was validated by = 186 M-1, previously measured by potentiometry for using the value of CuIMe6TREN+ in CH3CN (22). Then it was applied to CuITPMA+ in CH3CN, obtaining = 389 M-1 (Figure 8b). 182

Figure 8. a) vs. time for the reaction of CuIL+ with 2-methyl bromopropionate in the presence of different amounts of Br−, measured at an KATRP RDE in CH3CN + 0.1 M Et4NBF4, T = 25 °C. b) Experimental values, in the presence of increasing and fitting of data to . Adapted with permission from ref. (23). Copyright 2018 Elsevier. obtain

Electrochemical Determination of kdisp The rotating disk electrode was also used to measure the rate constant of CuI disproportionation, kdisp (17, 27):

The experimental procedure is the same as previously described for kact determination, except that neither RX nor radical scavenger is added. The which is easily calculated by Eq. recorded current, IL, is proportional to 10. The extent of disproportionation mainly depends on the nature of ligand and solvent, and halide ions if present. Indeed, in the presence of X-, additional equilibria for CuI speciation and disproportionation of XCuIL, CuIX2- and CuIX should be considered. The equilibrium constant of CuIL+ disproportionation in Eq. 18 can be expressed by the following relation (21):

or alternatively

where

is the disproportionation constant of solvated Cu+ ions:

Thus, the disproportionation process strongly depends on the extent of stabilization provided by the ligand to the metallic center. 183

values between 10-4 and 102 were calculated from Eq. 20 in water (17), with TPMA providing the greatest stability for CuI, whereas smaller values were reported in organic solvents (10-2-10-7) (29). In acetonitrile, the disproportionation of CuI is negligible. Regarding disproportionation kinetics, relatively high kdisp values, ranging from 1 to 102 M-1s-1, were measured in water, generally increasing with pH and exhibiting a good correlation with the corresponding Kdisp (17). kdisp values between 10-2 and 1 M-1s-1 were measured in organic solvents and solvent/monomer combinations (27).

The Interplay between Activation and Disproportionation: SARA-ATRP or SET-LRP? The measurement of kdisp gave an important contribution to the mechanistic analysis of Reversible-Deactivation Radical Polymerizations (RDRP) in the presence of Cu0 (Scheme 4). Percec et al. proposed a mechanism, called Single Electron Transfer – Living Radical Polymerization (SET-LRP), which i) considered Cu0 as the only activator of RX (or PnX), through an outer sphere electron transfer process, and ii) assumed fast and complete disproportionation of CuI species (37). In contrast, Matyjaszewski et al. proposed the Supplemental Activator Reducing Agent (SARA)-ATRP mechanism, based on CuI/CuII activation/deactivation, with Cu0 acting as a supplemental activator of RX, and reducing agent by comproportionation with CuII complex, thus re-generating CuI (38).

Scheme 4. SET-LRP versus SARA-ATRP Mechanism. Adapted with permission from ref. (27). Copyright 2015 Elsevier. The RDE technique provided low kdisp values in DMSO and DMSO/methyl acrylate mixtures, for both CuI/TPMA and CuI/Me6TREN catalysts (10-2-100 M1s-1) (27). The presence of the monomer and/or halide ions decreased the value of kdisp for a fixed complex. In the presence of a Cu0 wire the disproportionation was only slightly faster than in the absence of Cu0, with kdisp increasing with the length of the wire. kact of RX with different reactivity was measured via RDE for the aforementioned systems. The obtained values (1 < kact < 103 M-1s-1) show that generally kact > kdisp. In particular, under typical conditions of RDRP in the 184

presence of Cu0 (i.e. Cu/Me6TREN in DMSO/methyl acrylate, with very reactive RX), kact was more than 3 orders of magnitude higher than kdisp. It follows that CuI is mainly consumed by the activation reaction, thus the extent of its disproportionation is negligible. The ratio between the rates of activation and disproportionation can be evaluated (Eq. 22). Considering that generally CRX , it results that ract >> rdisp. >>

Hence, SARA-ATRP mechanism was confirmed because i) CuI disproportionation was slow, even negligible, ii) CuI rapidly activated RX, thus Cu0 cannot be the sole activator. Similarly, electrochemical determination of high kact values in aqueous media, by using CV and digital simulations, provided further confirmation of the SARA mechanism (29). Indeed, in pure water CuIMe6TREN+ activates OEOBP and HEBiB with rate constants kact = 6.6 × 105 M-1s-1 and 2.6 × 107 M-1s-1, respectively, whereas kdisp = 1.25 × 102 M-1s-1. In water/monomer mixtures (e.g. 18 vol% OEOA), kact decreased by about one order of magnitude, but the reactivity of CuI toward RX activation was still much higher than that of Cu0. It was estimated that at least 400 m of Cu0 wire were required to match the activity of 10-6 M CuIMe6TREN+, in the same system.

Electrochemical Investigation of Organo-Copper Complexes in ATRP A complete mechanistic understanding of ATRP requires to identify side-reactions and measure their kinetic and thermodynamic parameters. CuI catalyzed radical termination (CRT) plays an important role in ATRP, particularly for acrylate monomers. However, the mechanism of CRT is not fully defined. As previously reported in Scheme 3, CuI species reacts with radicals by forming CuII-R species, with an equilibrium constant KOMRP (where OMRP = organometallic mediated radical polymerization) (35). Then, free radicals react with the organometallic complex, with a rate constant kCRT, leading to radical termination and re-generation of CuI. It is likely that termination proceeds by a disproportionation-like pathway, giving products with the same molecular weight of the original polymeric radicals (39). Zerk and Bernhardt exploited the use of cyclic voltammetry combined to digital simulations to analyze the interplay between ATRP and OMRP (33). TPMA and Me6TREN were used as ligands, CH3CN and DMSO as solvents, and bromoacetonitrile (BrAN) and ethyl α-bromoisobutyrate (EBiB) as initiators. The voltammetric analysis was conducted in the absence of radical traps such as TEMPO. The two initiators gave distinct responses. When RX was EBiB, a single reduction peak was observed in most cases (Figure 9a,c). In contrast, in the presence of BrAN a second reduction peak appeared at more negative potentials (Figure 9b,d). 185

Figure 9. Cyclic voltammetry of 1.0 mM BrCuIITPMA+ + RX in DMSO (A and B) or CH3CN (C and D) + 0.1 M Et4NClO4, at v = 200 mV s−1. Experimental data in solid curves, simulated data in dashed curves. Adapted with permission from ref. (33). Copyright 2017 American Chemical Society.

The second peak was attributed to the reduction of the organometallic complex R−CuIITPMA+ that formed after the generation of radicals by the activation process. When TEMPO was present, no second peak was observed; the scavenger immediately trapped the radicals, thus suppressing their reaction with the catalyst. The steric hindrance of tertiary radicals, such as that generated from EBiB, strongly disfavors the formation of the organometallic species. The formation of the organometallic species was supported by spectroelectrochemistry of BrCuIITPMA+ catalyst, in the presence of an 8-fold excess of BrAN initiator. A constant potential of about -900 mV vs Fc+/Fc was applied to generate R−CuIITPMA+, during which Vis-NIR spectra were recorded. The intensity of the peak corresponding to BrCuIITPMA+ progressively decreased, while a new peak appeared at a shorter wavelength. This new peak was assigned to CuII(TPMA)(CH2CN)+. EPR spectroscopy was used to prove the presence of two distinct copper complexes with trigonal-bipyramidal geometry, CuII(TPMA)(CH2CN)+ and BrCuIITPMA+. Spectroelectrochemistry was also applied to BrCuIIMe6TREN+ with BrAN in CH3CN and DMSO, and with ClAN in CH3CN. A similar hypsochromic shift was observed in each case. 186

CVs in Figure 9 were then simulated by using the set of reactions in Scheme 2 in addition to the reactions in Scheme 5, with the corresponding kinetic and thermodynamic parameters.

Scheme 5. Reactions and parameters required to simulate the OMRP equilibrium involving the ATRP catalyst: formation of the organometallic species (a), reduction of R-CuIIL+ (b), and dissociation of R-CuIIL+ (c). All kinetic and thermodynamic parameters in Scheme 5 as well as kdeact were set as adjustable during the CV simulation. Interestingly, the values obtained from the simulation indicated that the formation of the organometallic species was quite fast (1.6 × 107 < kd,OMRP < 3.6 × 107 M-1s-1), whereas its dissociation was much slower (ka,OMRP ≈ 10-1 s-1). The only exception was BrCuIITPMA+/EBiB in DMSO, where ka,OMRP ≈ 103 s-1 was found, which is in agreement with the larger steric hindrance of the ensuing tertiary radical. Therefore, the thermodynamic constant of the OMRP process was in the range 107-108 M-1 for the primary radical, whereas KOMRP = 104 M-1 for the tertiary one. It follows that a combination of high kact (i.e. fast radical generation) and high KOMRP (i.e. fast radical trapping) was required to observe an effect on the voltammetric response. Summarizing, CVs combined to digital simulations provided kinetic and thermodynamic information about the influence of the OMRP equilibrium on ATRP. Spectroelectrochemistry allowed a direct observation of R−CuIIL+ generation. However, it should be noticed that the series of reactions used in the simulation did not take into account the “true” CRT process, i.e. the termination reaction between R−CuIIL+ and a free radical (Eq. 23). Thus, more work is needed to obtain information about CRT in the overall mechanism of ATRP/OMRP.

Conclusions Information about the activity and stability of ATRP catalysts are readily accessible by performing cyclic voltammetry in the polymerization environment. The affinity of Cu species for halide ions can be also investigated. Other electrochemical tools, including chronoamperometry at a rotating disk electrode and voltammetric analysis combined with digital simulation, enabled the measurement of kact, kdeact, and KATRP. Slow catalytic systems as well as super active catalysts were successfully studied with these techniques. Finally, the disproportionation of CuI complexes and the formation of organocopper intermediates can be observed and quantified by electrochemical methods. In 187

brief, electrochemical analysis emerges as an essential tool to better understand ATRP, and predict the polymerization outcomes. Several efforts are being devoted to the synthesis of new ligands that convey specific properties to copper catalysts. Therefore, a database of redox properties of Cu/L complexes can guide the selection of appropriate catalyst to target specific macromolecular properties and structures. At present, electrochemical methods appear as the most versatile, allowing to analyze both low and extremely active ATRP catalysts.

Acknowledgments Financial support from the University of Padova (Grant CPDA150001) is gratefully acknowledged.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

Matyjaszewski, K. Macromolecules 2012, 45, 4015–4039. Drawe, P.; Buback, M. Macromol. Theory Simul. 2016, 25, 74–84. Barth, J.; Buback, M. Macromol. React. Eng. 2010, 4, 288–301. Wang, Y.; Kwak, Y.; Buback, J.; Buback, M.; Matyjaszewski, K. ACS Macro Lett. 2012, 1, 1367–1370. Tang, W.; Kwak, Y.; Braunecker, W.; Tsarevsky, N. V.; Coote, M. L.; Matyjaszewski, K. J. Am. Chem. Soc. 2008, 130, 10702–10713. Goto, A.; Fukuda, T. Macromol. Rapid Commun. 1999, 20, 633–636. Matyjaszewski, K.; Paik, H.-J.; Zhou, P.; Diamanti, S. J. Macromolecules 2001, 34, 5125–5131. Tang, W.; Matyjaszewski, K. Macromolecules 2007, 40, 1858–1863. Pintauer, T.; Braunecker, W.; Collange, E.; Poli, R.; Matyjaszewski, K. Macromolecules 2004, 37, 2679–2682. Schröder, K.; Mathers, R. T.; Buback, J.; Konkolewicz, D.; Magenau, A. J.; Matyjaszewski, K. ACS Macro Lett. 2012, 1, 1037–1040. Soerensen, N.; Barth, J.; Buback, M.; Morick, J.; Schroeder, H.; Matyjaszewski, K. Macromolecules 2012, 45, 3797–3801. Lin, C. Y.; Coote, M. L.; Gennaro, A.; Matyjaszewski, K. J. Am. Chem. Soc. 2008, 130, 12762–12774. Pause, L.; Robert, M.; Savéant, J.-M. J. Am. Chem. Soc. 2000, 122, 9829–9835. Cardinale, A.; Isse, A. A.; Gennaro, A.; Robert, M.; Savéant, J.-M. J. Am. Chem. Soc. 2002, 124, 13533–13539. Isse, A. A.; Bortolamei, N.; De Paoli, P.; Gennaro, A. Electrochim. Acta 2013, 110, 655–662. Lanzalaco, S.; Fantin, M.; Scialdone, O.; Galia, A.; Isse, A. A.; Gennaro, A.; Matyjaszewski, K. Macromolecules 2017, 50, 192–202. Fantin, M.; Isse, A. A.; Gennaro, A.; Matyjaszewski, K. Macromolecules 2015, 48, 6862–6875. 188

18. Lorandi, F.; De Bon, F.; Fantin, M.; Isse, A. A.; Gennaro, A. Electrochem. Commun. 2017, 77, 116–119. 19. Fantin, M.; Chmielarz, P.; Wang, Y.; Lorandi, F.; Isse, A. A.; Gennaro, A.; Matyjaszewski, K. Macromolecules 2017, 50, 3726–3732. 20. Fantin, M.; Isse, A. A.; Matyjaszewski, K.; Gennaro, A. Macromolecules 2017, 50, 2696–2705. 21. Fantin, M.; Lorandi, F.; Gennaro, A.; Isse, A. A.; Matyjaszewski, K. Synthesis 2017, 49, 3311–3322. 22. Bortolamei, N.; Isse, A. A.; Di Marco, V. B.; Gennaro, A.; Matyjaszewski, K. Macromolecules 2010, 43, 9257–9267. 23. Lorandi, F.; Fantin, M.; Isse, A. A.; Gennaro, A.; Matyjaszewski, K. Electrochim. Acta 2018, 260, 648–655. 24. Fantin, M.; Isse, A. A.; Bortolamei, N.; Matyjaszewski, K.; Gennaro, A. Electrochim. Acta 2016, 222, 393–401. 25. Zerk, T. J.; Bernhardt, P. V. Dalton Trans. 2013, 42, 11683–11694. 26. De Paoli, P.; Isse, A. A.; Bortolamei, N.; Gennaro, A. Chem.Commun. 2011, 47, 3580–3582. 27. Lorandi, F.; Fantin, M.; Isse, A. A.; Gennaro, A. Polymer 2015, 72, 238–245. 28. Dos Santos, N. A. C.; Lorandi, F.; Badetti, E.; Wurst, K.; Isse, A. A.; Gennaro, A.; Licini, G.; Zonta, C. Polymer 2017, 128, 169–176. 29. Konkolewicz, D.; Krys, P.; Góis, J. R.; Mendonça, P. V.; Zhong, M.; Wang, Y.; Gennaro, A.; Isse, A. A.; Fantin, M.; Matyjaszewski, K. Macromolecules 2014, 47, 560–570. 30. Ribelli, T. G.; Fantin, M.; Daran, J.-C.; Augustine, K. F.; Poli, R.; Matyjaszewski, K. J. Am. Chem. Soc. 2018, 140, 1525–1534. 31. Bell, C. A.; Bernhardt, P. V.; Monteiro, M. J. J. Am. Chem. Soc. 2011, 133, 11944–11947. 32. Zerk, T. J.; Bernhardt, P. V. Inorg. Chem. 2014, 53, 11351–11353. 33. Zerk, T. J.; Bernhardt, P. V. Inorg. Chem. 2017, 56, 5784–5792. 34. Wang, Y.; Soerensen, N.; Zhong, M.; Schroeder, H.; Buback, M.; Matyjaszewski, K. Macromolecules 2013, 46, 683–691. 35. Ribelli, T. G.; Wahidur Rahaman, S.; Daran, J.-C.; Krys, P.; Matyjaszewski, K.; Poli, R. Macromolecules 2016, 49, 7749–7757. 36. Tang, W.; Tsarevsky, N. V.; Matyjaszewski, K. J. Am. Chem. Soc. 2006, 128, 1598–1604. 37. Percec, V.; Guliashvili, T.; Ladislaw, J. S.; Wistrand, A.; Stjerndahl, A.; Sienkowska, M. J.; Monteiro, M. J.; Sahoo, S. J. Am. Chem. Soc. 2006, 128, 14156–14165. 38. Konkolewicz, D.; Wang, Y.; Krys, P.; Zhong, M.; Isse, A. A.; Gennaro, A.; Matyjaszewski, K. Polym. Chem. 2014, 5, 4396–4417. 39. Ribelli, T. G.; Augustine, K. F.; Fantin, M.; Krys, P.; Poli, R.; Matyjaszewski, K. Macromolecules 2017, 50, 7920–7929.

189

Chapter 8

Insights into the Reactivity of Epoxides as Reducing Agents in Low-Catalyst-Concentration ATRP Reactions David C. McLeod,*,1,2 Kapil Dev Sayala,1 and Nicolay V. Tsarevsky1 1Department

of Chemistry and Center for Drug Discovery, Design, and Delivery, Southern Methodist University, 3215 Daniel Avenue, Dallas, Texas 75275, United States 2Department of Chemistry and Biochemistry, Heidelberg University, 310 East Market Street, Tiffin, Ohio 44883, United States *E-mail: [email protected]. E-mail: [email protected].

Modern atom transfer radical polymerization (ATRP) techniques employ relatively-low concentrations of copper complexes as the polymerization mediators. Typically, the catalyst is initially added in the higher oxidation (deactivating) state, so these systems require a reducing agent to generate in situ the lower oxidation state (activating) complex, able to react with an alkyl halide initiator, thereby initializing the polymerization. Epoxides can serve in this function and in this work, ethyl acrylate, methyl methacrylate, and styrene were homopolymerized in a well-controlled manner from an alkyl bromide initiator under low-catalyst-concentration ATRP conditions in the presence of an equimolar amount (vs monomer) of epoxides such as styrene oxide or phenyl glycidyl ether. A study on the free radical polymerization of ethyl acrylate, which occurred in the presence of styrene oxide and CuBr2 but only in the absence of radical traps/inhibitors, further demonstrated that the reduction of the deactivator by epoxides proceeds via the formation of a radical derived from an alkoxide anion originating from the epoxide.

© 2018 American Chemical Society

Introduction Epoxide-containing polymers are very useful in the synthesis of functional materials because the epoxide groups can be modified in a facile manner via nucleophilic substitution reactions in order to attach other functional groups onto the polymers (1). The ability to produce well-defined, epoxide-containing polymers directly from various epoxide-containing vinyl monomers, such as 4-vinylphenyloxirane (4VPO) and glycidyl methacrylate (GMA), using reversible-deactivation radical polymerization techniques is of special interest. Atom transfer radical polymerization (ATRP) is a robust and widely utilized form of reversible-deactivation radical polymerization that enables the production of macromolecules with pre-determined molecular weights, diverse well-defined architectures, and high degrees of chain-end functionality from a wide variety of vinyl monomers (2–6). In the classical ATRP process, the halogen atom of an alkyl halide initiator is homolytically cleaved by a redox-active, metal complex in the lower oxidation (activating) state. The resulting alkyl radicals initiate the polymerization of vinyl monomers, and the growing chains briefly propagate until they are reversibly capped by a halogen atom, donated by the higher oxidation state halide complex, dubbed deactivator. This dynamic equilibrium favors the production of the dormant, non-propagating (halogen-capped) chains, thereby limiting the amount of irreversibly terminated “dead” chains. A significant drawback of traditional ATRP is that the catalyst is added to the reaction in the lower-oxidation state, which requires special handling due to the sensitivity of the catalyst towards oxygen. The catalyst is also often present in an amount stoichiometrically equivalent to the initiator. Given that the metal component of the catalyst, which is most often copper, is relatively expensive, highly colored, and potentially toxic, using high concentrations of catalyst is also undesirable, and necessitates time-consuming post-polymerization purification procedures. In addition, redox reactions between the propagating radicals and the catalyst (especially pronounced when the latter is present at high concentrations) limit the molecular weights that can be attained (7, 8). However, the amount of catalyst in traditional ATRP reactions cannot be significantly decreased because the deactivating complex accumulates over time, due to the inevitable occurrence of irreversible radical termination events (a kinetic feature known as the persistent radical effect) (9, 10). An overabundance of deactivating complex can slow or even stop ATRP reactions at low monomer conversion when relatively small amount of catalyst is initially present. To remedy these deficiencies, low-catalyst-concentration (LCC) ATRP techniques, such as “Activators ReGenerated by Electron Transfer” (ARGET) (11) and “Initiators for Continuous Activator Regeneration” (ICAR) ATRP (12), which are more facile, cost-effective, and environmentally friendly than traditional ATRP, have been developed (5). Well-controlled polymerizations of vinyl monomers are often possible using these LCC ATRP techniques, even when the amount of catalyst is 98%, ATRP Solutions), CuBr2 (99%, Aldrich), diethyl 2-bromo-2-methylmalonate (DEBrMM; 98%, Aldrich), phenyl glycidyl ether (PGE, 99%, Aldrich), styrene oxide (StyOx, 97+%, Acros), 2,2,6,6-Tetramethyl-1-piperidinyloxy (TEMPO, 98%, Aldrich) tetramethylsilane (TMS; >99.9%, Aldrich), CDCl3 (99.8% D, Cambridge Isotope Laboratories), N,N-dimethylformamide (DMF; >99.8%, EMD). Analyses Monomer conversions were determined by periodically withdrawing samples from the reaction mixtures using a nitrogen-purged syringe, diluting a portion of the samples with CDCl3 (with TMS as a chemical shift reference), and then analyzing the samples by 1H NMR spectroscopy on a JEOL ECA-500 spectrometer operating at 500 MHz. The remaining portion of the withdrawn samples was diluted with THF, filtered through Acrodisc 0.2 µm PTFE filters, and injected into a size exclusion chromatography (SEC) system to determine apparent number-average molecular weights (Mn) and molecular weight distribution (MWD) dispersities (Đ = Mw/Mn). SEC data was collected on a Tosoh EcoSEC 195

HLC-8320 system equipped with a series of four columns (TSK gel guard Super HZ-L, Super HZM-M, Super HZM-N and Super HZ2000) with THF as the eluent at a flow rate of 0.35 mL min–1 and at 40 °C. The SEC calibration curve was based on linear polystyrene standards. UV/vis spectra were recorded on a Shimadzu UV-3600 spectrophotometer equipped with a Peltier heating device (accurate to 0.01 °C) using 1-cm quartz cuvettes. Synthetic Procedures LCC ATRP of EA, MMA, or Sty in the Presence of Epoxides In the following procedure, the value of DPn, targ (i.e., [monomer]0 / [RBr]0) was kept at 100, with MMA acting as the monomer, and PGE acting as the epoxidebased reducing agent. MMA (1.000 g, 9.988 mmol) was weighed in a 10 mL vial equipped with a stir bar, to which PGE (1.500 g, 9.988 mmol) was added. A stock solution of CuBr2 (0.0089 g, 4.0 × 10–2 mmol) and TPMA (0.0116 g, 4.0 × 10–2 mmol) in DMF (1 mL) was prepared. Then, 0.1 mL of the catalyst stock solution (containing 4.0 × 10–3 mmol of CuBr2 and TPMA) was injected into the reaction vial, followed by the initiator DEBrMM (0.019 mL, 0.010 mmol). The vial was sealed with a rubber septum, and a steady flow of nitrogen was used to sparge the solution for 30 min, after which the vial was placed in an oil bath heated to 70 °C. Similar procedures were conducted using StyOx (1.200 g, 9.988 mmol) as the reducing agent. Procedures using EA (1.000 g, 9.988 mmol) or styrene (1.040 g, 9.988 mmol) as the monomer, and PGE or StyOx (in the same relative molar amounts as listed above) as the reducing agent were also carried out.

Polymerization of EA in the Presence of Epoxides and CuBr2 To a vial equipped with a magnetic stir bar were added EA (500 eq., 2 mL, 18.7mmol), StyOx (20 eq., 84 μL, 0.72 mmol), and 200 µL of CuBr2 (1 eq., 0.0084 g, 0.0375 mmol) in DMF (taken from a stock solution containing 0.2090 g CuBr2 in 5 mL of DMF). The tube was capped with a rubber septum, placed in an ice bath, and sparged with nitrogen for 15 minutes. It was then transferred in an oil bath preheated to 60 °C. Samples were withdrawn throughout the reaction for analysis by NMR. Identical reactions were performed in the presence of air (no nitrogen sparging was conducted) or with TEMPO (5 eq., 0.029 g, 0.187mmol) added to the reaction mixture. Control reactions were also performed in which either StyOx or CuBr2 were omitted.

Reduction of CuBr2/TPMA by PGE or StyOx CuBr2 (0.0223 g, 0.1 mmol) and TPMA (0.0290 g, 0.1 mmol) were added to a Schlenk flask equipped with a stir bar. The flask was sealed with a ground glass joint attached to a quartz cuvette and was then evacuated and back-filled with nitrogen three times. Through the side arm of the flask, 10 mL of deoxygenated 196

DMF was added via a nitrogen-purged syringe. The reactants were stirred until a solution (light green in color) was formed. The epoxide (4 mmol, 40 eq. vs CuBr2/TPMA) – either PGE (0.54 mL) or StyOx (0.45 mL), was deoxygenated by sparging with nitrogen and injected into the Schlenk flask via a nitrogen-purged syringe. The cuvette attached to the flask was inserted in the UV/vis spectrometer thermostatted at 70 °C. The decrease in the absorbance at the wavelength corresponding to the λmax of the CuBr2/TPMA complex in DMF (978 nm; ε978 = 229.7 M−1cm −1) was monitored as a function of time.

Results and Discussion Efficiency of Reduction of CuBr2/TPMA by Epoxides Given that various epoxides (or, more specifically, the alkoxide products of epoxide ring-opening with bromide anion) can reduce copper catalysts, and given that the well-controlled LCC ATRP of epoxide-containing methacrylic and styrenic monomers has already been successfully demonstrated (15, 17, 19), it should be possible to polymerize monomers that do not have attached epoxide groups, so long as epoxides are added to the reaction mixture. Glycidyl butyrate and GMA, two low-molecular-weight, alkyl-type epoxides, were previously studied (17) for the ability to reduce the complex of CuBr2/TPMA in acetonitrile; however, the latter epoxide was unsuitable for the current work (i.e., as additive to polymerization reaction mixtures) because it is also radically polymerizable, while the former compound is relatively expensive. PGE and StyOx also contain epoxide groups and are much more affordable than glycidyl butyrate. It was found by UV/vis spectroscopy that both PGE and StyOx reduced the complex of CuBr2/TPMA in DMF at 70 °C (i.e., conditions close to those used in the polymerizations described below), as shown in Figure 1. The reduction by StyOx was relatively slow, in agreement with the previous study, conducted in acetonitrile (17).

Figure 1. Reduction of the catalyst complex (0.01 M in DMF) at 70 °C by PGE or StyOx (40 eq. vs CuBr2/TPMA). 197

LCC ATRP of EA, MMA, and Sty Using Epoxides as Reducing Agents Homopolymerizations of MMA and Sty were therefore conducted under LCC ATRP conditions using either an alkyl- or aryl-substituted oxirane (PGE or StyOx, respectively) as the reducing agent, the complex of CuBr2/TPMA as the catalyst, and DEBrMM as the alkyl bromide initiator (DPn, targ = 100). Previously reported experiments utilized epoxide-containing monomers, so the epoxide groups were initially in equal concentration to the vinyl groups. Similarly, all polymerizations in this work were also conducted with the epoxides present in equal molar concentration to the monomer. Other than a relatively small amount of DMF used to solubilize and transfer the catalyst complex to the reaction mixture, no other solvents were added in order to avoid diluting the reaction mixtures, which would slow down the polymerizations. MMA polymerized at a higher rate than Sty and, in both cases, the reactions showed first-order kinetics (Figures 2a and 3a for the two respective monomers) with no significant difference in polymerization rate between the reactions conducted in the presence of PGE or StyOx.

Figure 2. LCC ATRP of MMA (90% w/v in DMF) at 70 °C initiated by DEBrMM (DPn, targ = 100) with an equimolar amount of PGE or StyOx as the reducing agent and 4 mol % CuBr2/TPMA vs initiator as the catalyst: (a) kinetics; (b) evolution of molecular weights and Đ with monomer conversion; (c, d) SEC traces of the polymers with monomer conversion shown at each curve. 198

Polymerizations of MMA and Sty reached high monomer conversions of ca. 90%, with similar control over the molecular weights and comparatively narrow and symmetrical MWDs (Figures 2b and 3b). The molecular weights of both polyMMA and polySty increased linearly with monomer conversion, and the experimentally determined molecular weights closely aligned with the theoretically expected values. The Đ values of these polymers were in the range of 1.2−1.3 and increased slightly with conversion, but more noticeably so at higher conversions. SEC traces of the polymers produced by these reactions (Figures 2c-d and 3c-d) showed a small amount of tailing in the low-molecular-weight region.

Figure 3. LCC ATRP of styrene (90% w/v in DMF) at 70 °C initiated by DEBrMM (DPn, targ = 100) with an equimolar amount of PGE or StyOx as the reducing agent and 4 mol % CuBr2/TPMA vs initiator as the catalyst: (a) kinetics; (b) evolution of molecular weights and Đ with monomer conversion; (c, d) SEC traces of the polymers with monomer conversion shown at each curve. In addition to MMA and Sty, LCC ATRP of an acrylate monomer, EA, was explored under identical conditions using PGE as the reducing agent. Interestingly, the polymerization control (in terms of values of Đ) was better than when MMA or Sty were polymerized. The first-order kinetics were not perfectly linear (Figure 4a), for reasons that are unclear but the polymerization proceeded to reach 86% conversion in 7 h. The reaction with PGE also showed all of the typical hallmarks of controlled/”living” radical polymerization, with a linear increase in the molecular weight of the polymers with monomer conversion (Figure 4b). Unlike the reactions with MMA or Sty, the values of Đ of these 199

polymers steadily decreased throughout the polymerization, reaching a value of 1.12 at 86% monomer conversion. SEC traces of the polymers (Figure 4c) showed no noticeable amount of tailing in the low-molecular-weight region.

Figure 4. LCC ATRP of EA (90% w/v in DMF) at 70 °C initiated by DEBrMM (DPn, targ = 100) with an equimolar amount of PGE as the reducing agent and 4 mol % CuBr2/TPMA vs initiator as the catalyst: (a) kinetics; (b) evolution of molecular weights and Đ with monomer conversion; (c) SEC traces of the polymers with monomer conversion shown at each curve.

No further effort was made to optimize these initial reaction conditions, since these experiments were only intended to serve as a proof-of-concept to show that LCC ATRP can also be successfully conducted for non-epoxide-containing monomers using extraneous epoxide-type reducing agents. In future research it would be interesting to decrease the amount of epoxide and ascertain what is the minimum amount that can be used while still maintaining satisfactory control over the reactions and an acceptable reaction rates. Studies to determine how the structure of the epoxide affects the reactivity of the resulting alkoxide anion would be of interest as well. 200

Free Radical Polymerization of EA in the Presence of CuBr2 and Epoxides In the original study (17) on the use of epoxides as reducing agents in LCC ATRP, it was proposed that bromide anions open the epoxide rings, producing alkoxide anions, which can then reduce the copper catalyst complex by electron transfer (Scheme 2). This would produce an alkoxy radical, which could abstract a hydrogen atom from the solvent or some reactant present in the reaction mixture. The epoxide is thus converted into β-bromoalcohol (identified by NMR spectroscopy), i.e., the same product that would be formed by the ionic ring-opening of epoxides by “HBr”. This study aimed in part to clarify the exact mechanism of formation of β-bromoalcohol and therefore of the reduction of Cu(II) halides by epoxides. The production of alkoxy radicals was confirmed by the ability to carry out (uncontrolled) free radical polymerization of EA in the presence of StyOx and CuBr2 (Figure 5). Ligands such as TPMA were not added to the reaction mixtures because CuBr2 is a markedly stronger oxidant than CuBr2/TPMA and its use would guarantee faster polymerizations. If either the epoxide or the Cu(II) halide were removed from the reaction, no polymerization occurred. Likewise, if oxygen was not removed from the reaction mixture, or if a radical inhibitor such as TEMPO was added to it, no polymerization took place. Both the produced alkoxy radicals and the radicals formed by hydrogen transfer from the solvent or other additives to the alkoxy radicals, are likely to react with the vinyl monomers, initiating the growth of new polymer chains.

Figure 5. Kinetics of the free radical polymerization of EA (60 °C) in the presence of various additives. The polymerization kinetics of EA in the presence of CuBr2 and epoxides are rather complex because the former compound can play two roles, both having a marked impact on the polymerization rate. On the one-hand, CuBr2 is responsible for the generation of alkoxide anions and for their consecutive oxidation, which yields radicals. In other words, sufficiently high concentration of the Cu(II) halide is needed to ensure the continuous formation of radicals at sufficient concentration. Thus, when the Cu(II) halide is depleted, the reaction could be expected to stop. On 201

the other hand, CuBr2 can serve as very efficient chain transfer agent, transferring irreversibly (in contrast to ATRP reactions conducted in the presence of suitable ligands) a bromine atom to the propagating chains and essentially “killing” them. A notable drawback of ICAR ATRP is that that the radical initiators will generate new chains throughout the reaction, contaminating the product with a small amount of low-molecular-weight polymers, which lack a functionality at the α-chain end derived from the alkyl halide initiator. It was noted in the previous section that some low-molecular-weight polymer was seen in the SEC traces of polyMMA and polySty, which broadened the MWD distributions of these polymers. Given that alkoxy radicals are generated in this electron transfer process, and given that these alkoxy radicals can initiate polymerization of vinyl monomers, it is reasonable to conclude that these low-molecular-weight polymers are initiated by alkoxide radicals. Therefore, when epoxides are used as reducing agents for LCC ATRP, it is likely that a combination of ICAR and ARGET ATRP is actually occurring throughout the reaction to generate and then continuously regenerate the activator.

Conclusions This work demonstrated that well-controlled LCC ATRP of acrylic, methacrylic, and styrenic-type monomers can be achieved using epoxides as reducing agents. It was further demonstrated by the uncontrolled free radical polymerization of an acrylic monomer in the presence of CuBr2 and epoxides that the reduction of Cu(II) halide complexes proceeds via the formation of radical derived from the epoxide (or more precisely, the product of its ring-opening by a nucleophile).

References Benaglia, M.; Alberti, A.; Giorgini, L.; Magnoni, F.; Tozzi, S. Polym. Chem. 2013, 4, 124–132. 2. Matyjaszewski, K.; Xia, J. Chem. Rev. 2001, 101, 2921–2990. 3. Kamigaito, M.; Ando, T.; Sawamoto, M. Chem. Rev. 2001, 101, 3689–3746. 4. Ouchi, M.; Terashima, T.; Sawamoto, M. Chem. Rev. 2009, 109, 4963–5050. 5. Tsarevsky, N. V.; Matyjaszewski, K. In Fundamentals of Controlled/Living Radical Polymerization; Tsarevsky, N. V., Sumerlin, B. S., Eds.; The Royal Society of Chemistry: Cambridge, UK, 2013; pp 287−357. 6. Matyjaszewski, K.; Tsarevsky, N. V. J. Am. Chem. Soc. 2014, 136, 6513–6533. 7. Tsarevsky, N. V.; Braunecker, W. A.; Matyaszewski, K. J. Organomet. Chem. 2007, 692, 3212–3222. 8. Tsarevsky, N. V. Isr. J. Chem. 2012, 52, 276–287. 9. Fischer, H. J. Polym. Sci., Part A: Polym. Chem. 1999, 37, 1885–1901. 10. Fischer, H. Chem. Rev. 2001, 101, 3581–3610. 11. Jakubowski, W.; Min, K.; Matyjaszewski, K. Macromolecules 2006, 39, 39–45. 1.

202

12. Matyjaszewski, K.; Jakubowski, W.; Min, K.; Tang, W.; Huang, J.; Braunecker, W. A.; Tsarevsky, N. V. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 15309–15314. 13. Jonsson, M.; Nyström, D.; Nordin, O.; Malmström, E. Eur. Polym. J. 2009, 45, 2374–2382. 14. Hansson, S.; Östmark, E.; Carlmark, A.; Malmström, E. ACS Appl. Mater. Interfaces 2009, 1, 2651–2659. 15. Tsarevsky, N. V.; Jakubowski, W. J. Polym. Sci., Part A: Polym. Chem. 2011, 49, 918–925. 16. Woodruff, S. R.; Davis, B. J.; Tsarevsky, N. V. ACS Symp. Ser. 2012, 1100, 99–113. 17. Woodruff, S. R.; Davis, B. J.; Tsarevsky, N. V. Macromol. Rapid Commun. 2014, 35, 186–192. 18. Tsarevsky, N. V.; Woodruff, S. R.; Wisian-Neilson, P. J. J. Chem. Educ. 2016, 1452–1459. 19. McLeod, D. C.; Tsarevsky, N. V. Polym. Int. 2014, 63, 868–875. 20. Woodruff, S., R.; Tsarevsky, N. V. ACS Symp. Ser. 2015, 1188, 149–167. 21. Yang, Q.; Balverde, S.; Dumur, F.; Lalevee, J.; Poly, J. Polym. Chem. 2016, 7, 6084–6093.

203

Chapter 9

Toward Butadiene-ATRP with Group 10 (Ni, Pd, Pt) Metal Complexes Vignesh Vasu, Joon-Sung Kim, Hyun-Seok Yu, William I. Bannerman, Mark E. Johnson, and Alexandru D. Asandei* Institute of Materials Science and Department of Chemistry, University of Connecticut, 97 North Eagleville Road, Storrs, Connecticut 06269-3139, United States *E-mail: [email protected].

The ligand effect in a series of Group 10 (Ni, Pd, Pt) complexes was investigated in the ICAR ATRP of butadiene initiated from bromoesters in toluene at 110 °C. While a few trends such as (CO)2Ni(PPh3)2 >> Ni(COD)2, NiBr2/L: L = bpy ≥ PMDETA > MeO-bpy > TPMA, Cp2Ni ≥ NiTPP, (PPh3)2NiCl2 >> (PBu3)2NiCl2 > (PCy3)2Ni(Napht)Cl >~ (PCy3)2NiCl2 , Pd(PPh3)4 > (PPh3)2PdCl2 and (PPh3)2NiCl2 >> (PPh3)2PdCl2 ~ (PPh3)2PtCl2 as are apparent in each group, only (PPh3)2NiCl2 > (PPh3)2Ni(CO)2 stand out in terms of polymerization control and in their ability to activate the initiator and provide polybutadiene with high (> 65%) Br chain end functionality.

Introduction Polymers based on conjugated 1,3-dienes (butadiene, (BD), isoprene (ISO), dimethylbutadiene, (DMBD), chloroprene (ClP)) are produced industrially to the tune of billions of pounds/year, which testifies to the importance of their synthesis (1, 2). However, while emulsion radical polymerization can be employed in the production of adhesives, coatings, rubbers, and high impact materials based on random diene copolymers with styrene (St) (1) or acrylonitrile (AN), the corresponding thermoplastic elastomer block copolymers are prepared by © 2018 American Chemical Society

expensive, water and air sensitive coordination (3) or anionic polymerizations which demand strict reaction conditions, but only offer a limited selection of chain end and initiator functionalities (1). Therefore, low-cost, water amenable reversible deactivation/controlled radical polymerizations (CRPs, which proceed with a linear dependence of Mn on conversion, narrow polydispersity (PDI) and high chain end functionality (CEF)) (4–11), are desirable. However, the radical polymerization of dienes suffers from a series of disadvantages including low monomer boiling , Diels-Alder (DA) thermal dimerization points (12) (e.g. BD to 4-vinyl cyclohexene (12–14) and ISO to limonene) and chain transfer to the weak allylic Hs (branching/crosslinking at high conversion). In addition, the distinctive diene radical propagation, based on the equilibrium of the dominant primary 1,4-radical with other allylic delocalized resonance contributors (15), generates mixtures of constitutionally isomeric 1,2-, 3,4- or 1,4-cis/trans linkages (1) leading to the lowest radical propagation rate constants (kp) of all typical radical monomers (1, 4, 15, 16) , thus demanding the use of high pressure and high temperature (13) metal reactors. This is challenging experimentally, since by contrast with St or St CRPs, which can be conveniently sampled from Schlenk reaction tubes on a g scale, diene kinetics comprise multiple one data point experiments. As a consequence, contrasting with the massive literature on the CRP for non-gaseous monomers (4–11), there is very little work on dienes, and mostly on the more convenient, higher boiling ISO. Examples include CRPs enabled by nitroxides (17), RAFT agents (18), Te (19), poor iodine degenerative transfer (IDT) telomerizations (14), and Co selective dimerizations (20). Here, our reports on the Cp2TiCl (21–39) mediated CRPs of BD (21–25), ISO (26–34) and DMBD (35–39) initiated by the radical ring opening of epoxides (40, 41), SET reduction of aldehydes (42–45) and halides (46–49) remain the only examples of transition metal mediated (5) diene-CRPs and of subsequent diene block copolymer (22–39) synthesis. While successful, the Ti-mediated CRP remains a water-sensitive, organometallic protocol, likely unsuitable for emulsion polymerization. The known advantages of ATRP (4–11) (inexpensive and commercially available reagents, simplicity, catalytic nature, rational fine-tuning, water tolerance) (4–11), vs. all other typical CRP protocols, and those of emulsion (no solvents, lower cost, high rate increase) (1) vs. anionic/coordination polymerizations, render emulsion ATRP the most appropriate for industrial scale-up. However, over two decades and > 10,000 (11) articles since the original reports (50–52), while the mechanistic understanding (4–11, 53) of ATRP has considerably advanced for (meth)acrylates and St (4–11), its extension to VAc (54), VDF (54–64), ethylene and dienes has proven problematic. Indeed, previous diene-ATRP attempts (65–68), are at best qualitative, without evidence of control, and lacking detail on specific features (kinetics, mechanism, effect of reaction parameters etc.) and of the complex relationship between the reaction variables and the side reactions.

206

To address this problem, we set up a research program aimed at providing the first in-depth, quantitative study of the scope and limitations of diene-ATRP and of its applications in the synthesis of complex dienes architectures (69). Following our preliminary investigations on CuX initiated diene radical polymerizations (70–76) we subsequently demonstrated (69, 77) that while diene-ATRP failure was previously ascribed to BD/CuX catalyst coordination (11, 78), this not the real reason for poor polymerization results. Indeed, unlike the case of polar AN (79), or even that of the weak coordination (80) of CuX by octene, St, MA or MMA (81) which does not affect the respective ATRPs, CuX-μ-(1,3-Diene)-CuX (82) complexes are only stable at T < -78 °C, and the poor Lewis base dienes are no match for typical Cu N-ligands at the high (T > 100 °C) BD-ATRP temperatures. Most importantly, the essential pitfall of butadiene ATRP is rather the low stability of the weak allylic, 1,2 Pn-CH2-CH(CH=CH2)-X and 1,4 Pn-CH2-CH=CH-CH2-X halide chain ends, and BD-ATRP fails primarily by loss of its halide chain end functionality (CEF). Indeed, the primary (1,4-PBD) or secondary (1,2-PBD) allyl-X are the weakest of all polymer chain ends, irrespective of the CRP capping agent (83) (the Pn-X bond dissociation energy (BDE) order is Pn = allyl < AN < MMA < St < MA < VAc < VDF < Et) (78). As such, they implicitly also are the weakest of all ATRP-derived halide chain ends (15, (78). Thus, 16) considering known allyl-X ATRP data (84),

(85), (86),

(9,

86),

(86)), we suggested that polydiene halide chain ends (PBD-X) and the corresponding propagating allyl radicals (PBD•) have among the highest kact and lowest kdeact of all (4–11) monomers. Consequently, they display the largest reversible dissociation equilibrium constants in CRPs mediated by the persistent radical effect

(78), as well as the fastest exchange rates in degenerative transfer (DT) CRPs. Unfortunately, the easy activation of PBD-X and the slow deactivation of PBD• also enables competing processes which decrease halide CEF and broaden PDI (4–11). Indeed, our investigation of the effect of the nature of the halide, initiator, solvent, catalyst, ligand and ATRP procedure has revealed (69) that in addition to low bp, DA cycloaddition (12), the lowest kp (15, 16), and typical termination by recombination (87), the allyl PBD-X/PBD• are the most prone of all polymer chain ends to side reactions including CuX/CuX2/L oxidations/reductions (5, 88), beta-H eliminations (89) or catalytic termination (90) of propagating radicals and respectively, to thermal and base assisted PBD-X dehydrohalogenation via quaternization (91) with basic/nucleophilic N- or P-ligands in polar/basic solvents, followed by onium elimination/fragmentation (92) driven by formation of conjugated double bonds or allenes. Thus, due to the use of much less of a nucleophilic ligand, catalytic ATRP processes, especially 207

ICAR, are preferred and work well with tertiary Br vs. Cl ester initiators, and with non-nucleophilic bpy and MeO-bpy vs. highly activating but very nucleophilic ligands (TREN, PMDETA, etc.) affording poly(butadiene) (PBD) with high chain end functionality suitable for block copolymer synthesis. However, besides the well-developed Cu catalysts, many other transition metal complexes with variety of ligand frameworks mediate ATRP, and some of the other more successful systems for (meth)acrylates and styrene include group 8 and group 10 derivatives (5). We decided to thereby to extend our study and we are presenting below our current results with group 10 Ni, Pd and Pt complexes as BD-ATRP catalysts.

Experimental Materials 2,2′-Bipyridyl (bpy, 99 %), 4-4′-dimethoxy-2-2′-bipyridine (MeO-bpy, 97%), N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine (TPMA, 98 %), N,N,N′,N′,N″-pentamethyldiethylenetriamine (PMDETA, 98 %), nickel(II) bromide (NiBr2, 98 %), toluene anhydrous (99.9 %), bis(1,5-cyclopentadienyl)nickel (Cp2Ni, nickelocene, 99 %), 5,10,15,20-tetraphenyl-21H,23H-porphinenickel(II) (NiTPP, 95 %), bis(triphenylphosphine)dicarbonylnickel ((PPh3)2NiCO2, 98%), bis(triphenylphosphine)nickel(II)chloride ((PPh3)2NiCl2, 98%), 1,3-butadiene (> 99 %), dichlorobis(triphenyl-phosphine)palladium(II) ((PPh3)2PdCl2, 98 %), dichlorobis(triphenyl-phosphine)platinum(II) ((PPh3)2PdCl2), 98 %), ethyl 2-bromoisobutyrate (EBiB) all from Sigma Aldrich, bis(1,5-cyclooctadienyl)nickel(0) (Ni(COD)2, 98 %), bis (tricyclohexylphosphine)nickel(II)chloride ((PCy3)2NiCl2, 98%), bis (tributylphosphine)nickel(II)chloride ((PBu3)2NiCl2), 98%), and tetrakis triphenylphosphine palladium(0) ((PPh3)4Pd, 98 %) all from Strem Chemicals, were used as received. Bis(tricyclohexylphosphine)nickel(II)(napthyl)chloride (Cy3P)2NiCl(Napthyl) (93) and 1,1′-biphenylyl-1,4-bis(2-bromopropanoate) (Br-C(CH3)2-CO-O-C6H4-C6H4-O-CO-C(CH3)2-Br, DB3) (69) were synthesized as described in the literature. Techniques 1H

NMR (500 MHz) spectra were obtained on a Bruker DRX-500 at 24ºC in chloroform-d. Gel Permeation Chromatography (GPC) was performed on a Waters GPC equipped with a PL-ELS1000 evaporative light scattering (ELS) and a Jordi 2 mixed bed columns setup at 40°C. THF (Fisher, 99.9% HPLC grade) was used as eluent at a flow rate of 1 mL/min. Number-average (Mn) molecular weights were determined from calibration plots vs. PS standards. Since low Mn polydienes are quite soluble, precipitation in MeOH leads to artificially lower PDI, and we are reporting herein the values of raw samples. As in the Te-CRP (19), of isoprene, PSt calibrated GPC overestimates Mn. Thus, the initiator efficiency (IE = Mntheor/MnGPC) values are underestimated. 208

Butadiene Polymerization As described for gaseous VDF (55) and diene Cp2TiCl2-CRPs (21–35), BDATRP was performed in low pressure glass tubes which enable faster optimization and reproducible sampling under Ar after cooling the tube with dry ice/acetone. In a typical ICAR-ATRP procedure, the catalyst ((PCy3)NiCl2, 0.0123 g, 0.018 mmol) and toluene (3mL) were added to a 35 mL Ace Glass pressure tube, which was purged with Ar, cooled to ~-80 ºC using a dry ice/acetone bath and the R-BRr initiator (DB3, 0.1719 g, 0.36 mmol) and the free radical coinitiator (TBPO, 0.0104 g, 0.071 mmol) were added. BD (1.92 g, 36 mmol) was condensed on top of the frozen reaction mixture which was then degassed by a series of vacuum/Ar refill cycles, and the tube was places in an oil bath at 110 ° C. Sampling was performed under Ar, after cooling in a dry ice/acetone bath, and the system was degassed after each sampling. In most cases, visual inspection indicated that these polymerizations were homogeneous. The BD conversion was determined by integrating the polymer alkene resonances vs. the methyl resonance of toluene, which was used both as solvent as well as an internal standard. The Br chain end functionality was determined by integrating the allyl Br resonances at δ ~ 4 ppm vs. either the alkoxy resonance of EBIB at δ ~ 4.2 ppm vs. the aromatic resonances of DB3 at δ ~ 7.1 or 7.6 ppm.

Results and Discussion Initial Considerations While widely used in organometallic chemistry (94), group 10 catalysts are dwarfed by Cu in ATRP literature. Several P- and N- ligated Ni and Pd complexes were previously used in ATRP of MMA and St (5). These include phosphine systems such (PPh3)2NiX2/Al(iOPr)3 (95) X = Cl, Br, Ni(PPh3)4/Al(iOPr)3 (96) or (PPh3)2NiBr2 (97) for acrylates at 60-85 °C, the more thermally stable and soluble (Bu3P)2NiBr2 (98) at up to 120 °C, and (PPh3)2Ni(naphthyl)Cl (99) for MMA, St and p-ClSty at 90 °C. N- ligands such as Ni{C6H3(CH2NMe2)2-2,6}Br (100) were unsuccessful for St at 130 °C (101), but worked for MMA (102) at 80 °C, whereas NiX2 diimines (103) and Ni(II) tetraphenyl porphyrin (NiTPP) (104, 105) mediated the ATRP of St at 110-120 °C. Similarly, Pd(PPh3)4 (106) enabled the ATRP of MMA at 70 °C. However, while ligand and solvent are very important parameters with a combined influence of > 6 orders of magnitude for the KATRP of Cu-mediated ATRP of St and (meth)acrylates (4–11), to the best of our knowledge, there is no Ni, Pd or Pt data for kact, kdeact or KATRP, and no group 10 complexes were ever used in ICAR-ATRP of any monomer or diene-ATRP. Dienes should parallel St trends, but they are masked by high propensity of weak allyl Br chain ends to dehydrohalogenations quaternizations side reactions with nucleophilic/basic and by catalyst mediated oxidations/reductions and terminations of the propagating radical (88–90). The results are summarized in Table 1, Figures 1-7 and the mechanism is illustrated in Scheme 1. While Ni(0) coordinates and oligomerizes 1,3-dienes at low temperature and high pressure (107), and activated Ni(0)/RX initiates a free 209

radical polymerization (108), such coordination is likely irrelevant at the high T > 100 °C ATRP temperatures (109) as dienes are poorer ligands than phosphines or amines. Therefore, ligands are needed, and the structures of the Group 10 complexes tested in this work include (PPh3)2MtCl2 (Mt = Ni, Pd, Pt), Ni(COD)2, (PPh3)2Ni(CO)2, (PR3)2NiCl2 (R = nBu, Cy, Ph), (PCy3)2Ni(naphthyl)Cl, NiBr2/L (L = bpy, MeO-bpy, PMDETA, TPMA), NiCp2, NiTPP and Pd(PPh3)4, Scheme 2. Here, due to the TBPO reduction or interconversion (con/disproportionation) of various oxidation states, the starting valence of the metal may not be essential, and they are presented based on initial oxidation state. Butadiene initiation proceeds with the generation of a relatively less reactive, allyl delocalized PBD• radical and thus occurs easily from any typical ATRP initiators (69), as they all provide more reactive radicals (78). Since our previous Cu investigations indicated a clear Br >> Cl preference (69), the R-Br initiators used herein were Ebib and a difunctional tertiary bromide analog, DB3. In addition, as the halide chain in ICAR is derived primarily (> 90 %) (5) from the initiator, the nature of the initial halide in the metal complex is less important. A representative 1H NMR of polybutadiene (PBD) from Ni-mediated ICAR ATRP is shown in Figure 1 and demonstrates the DB3 initiation (aromatic resonances of at δ ~7-7.5 ppm), the polymerization by a halide atom transfer (allyl Br chain ends, δ ~ 4-4.5 ppm, which also allow the calculation of the Br chain end functionality, Br-CEF) and the free radical nature of the polymerization (typical (1) ~80/20 free radical distribution of the isomeric main chain 1,4 and 1,2-BD units). While the values of kact for these metal complexes and R-Br are not known, for good control, it is desirable that the rate of initiation is faster than that of propagation. This is also relevant for the Br-CEF calculation by integration of the ally Br chain ends vs. the initiator resonances. For Ebib, the initiator activation kinetics can be determined due to the different position of the CH3-CH2-O- in the pure initiator (δ ~ 4.2 ppm) vs. in the chain end (δ ~ 4.1 ppm) and are additionally illustrated in some cases. However, the aromatic DB3 resonances are the same in both starting and polymer-bound initiator. While the R-CH2-CH=CH-CH2connectivity at δ ~2.5 ppm could be used, this resonance overlaps with the toluene CH3 signal, which serves as an in-situ reference for BD conversion determination. Thus, the Br-CEF values form DB3 initiated, unprecipitated samples may be underestimated if the initiation is slow. Since excess free ligand may alkylate the weak PBD-Br chain ends (69), typical stoichiometric R-X/cat = 1/1 N-ATRP ratios were not used, but a comparison of N-ATRP with substoichiometric ratios and ICAR (e.g. [BD]/[RBr]/[TBPO]/[Mt] = 100/1/0/0.2 vs. 100/1/0.2/0.05 or 100/1/0.2/0.2) is provided in several cases. While DB3 was used in most ICAR polymerizations, the relative comparisons are meaningful, and so are those where only Ebib was employed. However, when both initiators were tested at the same Mt/initiator ratios, the Mt/Br ratio is twice as large for Ebib than for DB3, but remains catalytic, and we do not expect this to significantly alter the ligand trends. To minimize potential halide chain end alkylations/quaternizations and other side reactions, the apolar toluene was used as solvent. Polymerizations were performed at 110 °C where < 10 % of BD undergoes Diels-Alder dimerization 210

(69), and where the TBPO half lifetime is long enough to continuously supply radicals over a period of at least a week (1, 110).

Scheme 1. Mechanism of BD-ATRP Catalyzed by Group 10 Metal Complexes

Scheme 2. Structure of Ligands and Group 10 Complexes Tested in BD-ATRP 211

Table 1. Group 10 Mediated BD-ATRP Ex.

Catalyst

212

TBPO]/[Cat]/ [L]*

Mn x10-3

PDI

I.E. /Mn0x10-3

1

Ni(COD)2a

0.2/0.05/0.00

45.1

1.5

0.03/43

213

19

2

(PPh3)2Ni(CO)2a

0.2/0.05/0.00

10.5

1.8

0.20/5.3

257

30

3

(PPh3)2Ni(CO)2a

0.2/0.20/0.00

8.1

1.4

0.29/3.9

260

34

4

(PPh3)2Ni(CO)2a

0.0/0.20/0.00

2.7

1.3

0.40/0.8

144

11

5

NiBr2/bpya

0.2/0.05/0.15

20

1.6

0.13/1.2

144

34

6

NiBr2/bpyb

0.2/0.05/0.15

11.3

1.4

0.09/1.8

160

24

7

NiBr2/MeO-bpya

0.2/0.05/0.15

37.8

1.4

0.02/26

144

39

8

NiBr2/MeO-bpyb

0.2/0.05/0.15

8.0

2.2

0.20/2.1

144

26

9

NiBr2/TPMAa

0.2/0.05/0.15

55.5

1.4

0.04/22

144

29

10

NiBr2/PMDETAa

0.2/0.05/0.15

28.4

2.0

0.07/3.5

196

37

11

(PPh3)2NiCl2a

0.2/0.05/0.00

41.5

1.6

0.07/6.0

120

46

12

(PPh3)2NiCl2b

0.2/0.05/0.00

67.7

1.5

0.04/15

160

48

13

(PBu3)2NiCl2a

0.0/0.20/0.00

2

1.6

0.40/0.7

144

6

14

(PBu3)2NiCl2a

0.2/0.05/0.00

11.6

1.6

0.17/1.1

144

28

15

(PCy3)2NiCl2a

0.0/0.20/0.00

9.4

1.4

0.13/7.7

144

13

16

(PCy3)2NiCl2a

0.2/0.05/0.00

13.2

1.5

0.14/8.9

144

24

17

(Cy3P)2NiCl (Naph.)a

0.0/0.20/0.00

7.1

1.5

0.24/2.8

72

22

Time (h)

Conv. (%)

213

*

Ex.

Catalyst

18

TBPO]/[Cat]/ [L]*

Mn x10-3

PDI

I.E. /Mn0x10-3

(Cy3P)2NiCl (Naph.)a

0.2/0.05/0.00

11.2

1.6

0.18/8.6

144

28

19

Cp2Nia

0.0/0.20/0.00

150

1.2

0.01/66

144

10

20

Cp2Nia

0.2/0.05/0.00

79.9

1.4

0.03/78

72

14

21

NiTPPa

0.2/0.05/0.00

83.1

2.3

0.03/84

169

34

22

Pd(PPh3)4b

0.2/0.05/0.00

11.8

1.8

0.18/6.6

96

37

23

Pd(PPh3)4b

0.2/0.20/0.00

61.2

1.8

0.03/12

144

33

0.2/0.05/0.00

109

1.9

0.01/85

120

25

41

1.6

0.05/25

144

30

Cl2b

24

(PPh3)2Pd

25

(PPh3)2PtCl2b

0.2/0.20/0.00

All reactions are [BD]/[R-Br] = 100/1, R-Br = DB3

a,

EBIB

b.

T = 110 °C, in toluene.

Mn0

Time (h)

= Mn intercept at zero conversion.

Conv. (%)

Figure 1. 1H-NMR of polybutadiene from a Ni mediated ICAR-ATRP reaction: [BD]/[DB3]/[TBPO]/[(PCy3)2NiCl2] = 100/1/0.2/0.05, toluene, 110 ºC.

Ni(0) Complexes: Ni(COD)2 and (PPh3)2Ni(CO)2. Most likely due to its poor thermal stability at 110 °C (94), Ni(COD)2 (Figure 2) presents a free radical polymerization (FRP) with a low IE, and a rapidly vanishing Br-CEF. By contrast, the more stable (PPh3)2Ni(CO)2, especially at higher levels (0.2 vs. 0.05) promotes a pseudo-CRPs with a Mn0 intercept of ~4,500, but with an increase of Mn with conversion as well as with lower PDI, and higher Br-CEF (20 to 70 %) upon increasing [Ni]. As expected for ICAR, the kinetics are controlled only by the TBPO decomposition rate and are thus similar. Conversely, for the same catalyst level (0.2), N-ATRP in the absence of TBPO proceeds to much lower conversion, but with a low Mn intercept and relatively high Br-CEF of > 65 %. 214

Figure 2. Ni(0) complexes in BD-ATRP. (a) Dependence of Mn, PDI and Br-CEF on conversion, (b) kinetics. [BD]/[DB3]/[TBPO]/[Ni(COD)2] = 100/1/0.2/0.05 ( ), [BD]/[DB3]/[TBPO]/[(PPh3)2Ni(CO)2] =100/1/0.0/0.2 ( ) 100/1/0.2/0.05 ( ), 100/1/0.2/0.2 ( ), T = 110 °C, toluene.

Figure 3. NiBr2-Ligand effect in BD-ICAR ATRP. (a) Dependence of Mn, PDI and CEF on conversion, (b) kinetics. [BD]/[R-Br]/[TBPO]/[NiBr2]/[L] = [100/1/0.2/0.05/0.15]. R-Br = DB3 or EBiB: L = MeO-bpy ( , ), bpy ( , ), TPMA ( , none), PMDETA ( , none). Ebib activation kinetics: MeO-bpy, bpy ( , ). T = 110 °C, toluene. 215

Ni(II) Complexes NiBr2 with N-Ligands: L = bpy, MeO-bpy, PMDETA, and TPMA Here, some features of control are observed (Figure 3) in all ICAR cases. Here, for DB3, the more reactive MeO-bpy > TPMA provide a much higher (Mn0 ~ 25,000) intercept than bpy > PMDETA, and all have similar PDI ~1.5 except higher for PMDETA. Conversely, TPMA affords ~30 % Br-CEF, while all others are ~ 10 %. Doubling the amount of Ni/Br for EBiB, leads as expected to a faster polymerization with broader PDI for MeO-bpy > bpy. The higher PDI values of PMDETA and MeO-bpy are consistent with the allyl halide alkylation trends in observed in BD Cu-ATRP (5, 69). Moreover, the Ebib activation kinetics confirm the much higher activity of MeO-bpy (111) vs.

Cp2Ni and NiTPP Both ICARs (Figure 4) proceed with similar kinetics and FRP Mn profiles, no Br-CEF, but with lower PDI for Cp2Ni. Interestingly, the Cp2Ni mediated N-ATRP shows slower kinetics, but PDI as low as 1.25 in conjunction with an increase in Mn with conversion. However, both ICAR and N-ATRP are devoid of Br-CEF, indicating that Cp2Ni is a very poor halide transfer deactivator. Nonetheless, the weak control in conjunction with a lack of halide chain end may indicate a potential contribution from an organometallic mediated reversible deactivation mechanism (OMRP) (112–114), albeit with a substoichiometric use of Ni.

(PR3)2NiCl2 (R = Bu, Cy, Ph) and (PCy3)2Ni(Naphthyl)Cl A comparison of the phosphine effect in ICAR using the DB3 initiator indicates a clear Ph >> Bu > Cy trend, with better Mn vs. conversion profiles with an origin intercept for both Ph and Bu, but a higher (Mn0 ~ 8,000) intercept for Cy and with similar Br-CEF (< 0.15) and PDI values (~1.4 - 1.6) for R = Bu > Cy in both ICAR and in the slower N-ATRP. Replacing one of the Cls of (PCy3)2NiCl2 with a naphthyl ligand leads to only small comparative improvements in ICAR, but a better Mn profile, higher Br-CEF (~0.25), higher conversion, and slightly faster kinetics in N-ATRP.

216

Figure 4. Ni(II) complexes in BD-ATRP. (a) Dependence of Mn, PDI and CEF on conversion; (b) kinetics. [BD]/[DB3]/[TBPO]/[Ni] = [100/1/X/Y]. X/Y = 0.0/0.2 or 0.2/0.05; Cp2Ni: ( , ); NiTPP: (none, ).

Figure 5. Ni(II)phosphine complexes in BD-ATRP. (a) Dependence of Mn, PDI and CEF on conversion, (b) kinetics. [BD]/[R-Br]/[TBPO]/[Ni] = [100/1/X/Y]. R-Br = DB3, X/Y = 0.0/0.2 or 0.2/0.05; (PBu3)2NiCl2: ( , ), (PCy3)2NiCl2: ( , ), (PCy3)2Ni(Napht)Cl ( , ). (PPh3)2NiCl2: X/Y = 0.2/0.05, R-Br = EBiB, DB3 ( , ) and Ebib activation kinetics ( ). 217

Figure 6. Pd complexes in ICAR-ATRP. (a) Dependence of Mn, PDI and Br-CEF on conversion, (b) kinetics. [BD]/[EBiB]/[TBPO]/[Pd] = 100/1/0.2/(0.05 and 0.2) (PPh3)4Pd ( , ), (PPh3)2PdCl2 ( , none). Ebib activation kinetics ( , , ).

Figure 7. Group 10 Mt(II) complexes in BD-ICAR-ATRP. (a) Dependence of Mn, PDI and Br-CEF on conversion, (b) kinetics. [BD]/[EBiB]/[TBPO]/ [(PPh3)2>MtCl2] = 100/1/0.2/X; X = 0.05 Ni: ( ), Pd: ( ); X = 0.2, Pt ( ). Ebib activation kinetics ( , , ) 218

Although initially higher for Ph, all PDI values converge to ~1.6. Most importantly, Ph provides by far the fastest polymerization kinetics to the highest conversion (~50 % vs. ~25 %), the largest Br-CEF (~0.7 vs. ~ 0.15) and the fastest Ebib activation rate in the whole series of group 10 catalysts. While these trends should parallel the ligand effect in previous N-ATRP studies with Ni phosphines (95–99) in terms of catalyst stability and reactivity, a quantitative comparison is not possible due to the low ICAR ratios, the different temperatures, monomer and lack of Al(iOPr)3 in this study.

Pd(0) and Pd(II) Complexes The (PPh3)2PdCl2 mediated ICAR (Figure 6) leads to a poor FRP/CRP with high Mn0 ~90,000 intercept, PDI ~ 2, and only trace Ebib activation. By contrast, the less stable and more reactive Pd(PPh3)4 and clearly activates Ebib (Figure 6b), and decreasing the Pd from 0.2 to 0.05 affords better Mn control with a low intercept, higher IE, faster kinetics and similar PDIs. However, in all cases due to poor deactivation and possible dissociation of free PPh3 which quaternizes with the allyl chain end, no Br-CEF can be seen.

Ni versus Pd versus Pt A comparison of (PPh3)2MtCl2 in (Figure 7) confirms the superiority of Ni >> Pd and Pt for the PPh3 ligands in terms of Mn control, Br-CEF, as well as polymerization and Ebib activation rates.

Br-CEF Trends Finally, a comparison of the dependence of the Br-CEF on conversion for the higher CEF systems (Figure 8) reveals a clear 1,4-trans >> 1,4-cis >> 1,2 trend not only in the PBD microstructure, but also in the initiator addition to BD, and in the Br chain ends, where consistent with their respective stabilities, and similar to PVAc-I and PVDF-I chain ends (55), while the total Br-CEF remains constant, the mole fraction the 1,4-trans Br increases and that of the 1,4-cis Br decreases with conversion, and where, consistent with known superiority of divs. monofunctional halide initiators (55), a slightly better halide CEF is seen for the DB3 vs. Ebib.

219

Figure 8. Dependence of total, 1,4-trans, 1,4-cis and 1,2 Br-CEF on conversion. [BD]/[DB3]/[TBPO]/[(PPh3)2NiCl2] = 100/1/0.2/0.05 ( , , , ); [BD]/[Ebib]/[TBPO]/[(PPh3)2NiCl2] = 100/1/0.2/0.05 ( , , , ); [BD]/[DB3]/[TBPO]/[(PPh3)2Ni(CO)2] = 100/1/0.2/0.20 ( , , , ).

Conclusions Due to the low bp, very low kp, Diels-Alder dimerization, and the poor stability of the labile allyl halide chain ends, the ATRP of dienes remains a challenging topic for controlled radical polymerizations. The current kinetic investigations of the ligand effect in the ICAR and pseudo N-ATRP of butadiene initiated from bromoesters with group 10 complexes in toluene at 110 °C indicates that under similar reaction conditions, the effects seen here embed the ability of these catalysts to activate the R-Br initiator and subsequently the allyl-Br PBD chain end, and most importantly, to reversibly deactivate the PBD radical. While these observations should parallel the ATRP of styrene, they are also masked by the competition of specific diene side reactions such as PBD-Br alkylation with excess nucleophilic ligands, as well as by the thermal stability of the catalyst in some cases. Nonetheless, a few trends such as (CO)2Ni(PPh3)2 >> Ni(COD)2, NiBr2/L: L = bpy ≥ PMDETA > MeO-bpy > TPMA, Cp2Ni ≥ NiTPP, (PPh3)2NiCl2 >> (PBu3)2NiCl2 > (PCy3)2Ni(Napht)Cl >~ (PCy3)2NiCl2, Pd(PPh3)4 > (PPh3)2PdCl2 and for the (PPh3)2NiCl2 >> (PPh3)2PdCl2 ~ (PPh3)2PtCl2 as are apparent in each class. 220

However, even in unoptimized experiments, only two catalysts, i.e. (PPh3)2NiCl2 > (PPh3)2Ni(CO)2 stand out in terms of polymerization control and in their ability to activate the initiator and provide a quite high (> 65%) Br-CEF. Thus, while not on par with Cu systems, BD-Ni-ATRP still appears feasible and the rational selection of the reaction parameters (R-Br initiators, T = 110 °C, apolar solvents (toluene), low nucleophilicity/basicity ligands (phosphines > amines) and catalytic ATRP procedures ICAR > N-ATRP) enables minimization of side reactions and the successful synthesis of well-defined PBD with acceptable Br-CEF, a wide range of molecular weights and reasonably narrow PDI, eventually suitable for the preparation of complex macromolecular architectures. As criteria for successful BD-ATRP are emerging, they can guide optimization for other dienes and the elaboration of their industrially important emulsion polymerizations. In additions, while these solution polymerizations were performed at pressures of < few atm. which barely affect rate constants (115), better quality ATRPs should result (115) from high pressure emulsion polymerizations due to the large kp/kt (116), increase and the faster emulsion kinetics. Research along these lines is in progress and will be reported soon.

Acknowledgments Financial support from grant NSF CHE-1508419 and the University of Connecticut is gratefully acknowledged.

References Odian, G. Principles of Polymerization, 4th ed.; John Wiley & Sons, Hoboken, NJ, 2004; pp 311. 2. Schöps, M.; Leist, H.; DuChesne, A.; Wiesner, U. Macromolecules 1999, 32, 2806–2809. 3. Evans, W. J.; Giarikos, D. G.; Allen, N. T. Macromolecules 2003, 36, 4256–4257. 4. Braunecker, W. A.; Matyjaszewski, K. J. Mol. Catal. A: Chem. 2006, 254, 155–164. 5. Braunecker, W. A.; Matyjaszewski, K. Progr. Polym. Sci. 2007, 32, 93–146. 6. Pintauer, T.; Matyjaszewski, K.; Guan, Z. Top Organomet. Chem. 2009, 26, 221–251. 7. Tsarevsky, N. V.; Matyjaszewski. K. Atom Transfer Radical Polymerization (ATRP) In Fundamentals of Controlled/Living Radical Polymerization; Tsarevsky, N. V., Sumerlin. B. S., Eds.; RSC: Cambridge, 2013; pp 287−357. 8. di Lena, F.; Matyjaszewski, K. Progr. Polym. Sci. 2010, 35, 959–1021. 9. Ouchi, M.; Terashima, T.; Sawamoto, M. Chem. Rev. 2009, 109, 4963–5050. 10. Matyjaszewski, K.; Spanswick, J. In Polymer Science: A Comprehensive Reference; Matyjaszewski, K., Möller, M., Eds.; Elsevier BV: Amsterdam, 2012; Vol. 3, pp 377−428. 1.

221

11. Matyjaszewski, K. Macromolecules 2012, 45, 4015–4039. 12. Huybrechts, G.; Luyckx, L.; Vandenboom, T.; Van Mele, B. Int. J. Chem. Kinet. 1977, 9, 283–293. 13. Weerts, P. A.; German, A. L.; Gilbert, R. G. Macromolecules 1991, 24, 1622–1628. 14. Lebreton, P.; Ameduri, B.; Boutevin, B.; Corpart, J.; Juhue, D. Macromol. Chem. Phys. 2000, 201, 1016–1024. 15. Kamachi, M.; Kajiwara, A. Macromolecules 1996, 29, 2378–2382. 16. Beuermann, S.; Buback, M. Prog. Polym. Sci. 2002, 27, 191–254. 17. Benoit, D.; Harth, E.; Fox, P.; Waymouth, R. M.; Hawker, C. J. Macromolecules 2000, 33, 363–370. 18. Moad, G. Polym. Int. 2017, 66, 26–41. 19. Nakamura, Y.; Arima, T.; Tomita, S.; Yamago, S. J. Am. Chem. Soc. 2012, 134, 5536–5539. 20. Debuigne, A.; Jerome, C.; Detrembleur, C. Angew. Chem., Int. Ed. 2009, 48, 1422–1424. 21. Asandei, A. D.; Yu, H. S.; Simpson, C. P. Polym. Mater.: Sci. Eng. 2010, 103, 511–512. 22. Asandei, A. D.; Yu, H. S.; Simpson, C. P. Polym. Mater.: Sci. Eng. 2010, 102, 68–69. 23. Asandei, A. D.; Yu, H. S. Polym. Prepr. 2009, 50 (2), 601–602. 24. Asandei, A. D.; Yu, H. S; Adebolu, A. Polym. Mater.: Sci. Eng. 2009, 101, 1377–1378. 25. Asandei, A. D.; Yu, H. S; Simpson, C. P. Polym. Mater.: Sci. Eng. 2009, 101, 1379–1380. 26. Asandei, A. D.; Simpson, C. P.; Yu, H. S.; Adebolu, O. I.; Saha, G.; Chen, Y. ACS Symp. Ser. 2009, 1024, 149–166. 27. Asandei, A. D.; Simpson, C. P. Polym. Prepr. 2008, 49 (2), 75–76. 28. Asandei, A. D.; Yu, H. S.; Simpson, C. P. Polym. Prepr. 2010, 51 (2), 584–585. 29. Asandei, A. D.; Yu, H. S.; Simpson, C. P. Polym. Prepr. 2010, 51 (1), 545–546. 30. Asandei, A. D.; Adebolu, O.; Yu, H. S; Simpson, C. P.; Gilbert, M. Polym. Prepr. 2009, 50 (1), 177–178. 31. Asandei, A. D.; Yu, H. S; Adebolu, O.; Simpson, C. P.; Duong, O. Polym. Mater.: Sci. Eng. 2009, 100, 366–367. 32. Asandei, A. D.; Simpson, C. P. Polym. Prepr. 2008, 49 (1), 452–453. 33. Asandei, A. D.; Simpson, C. P.; Yu, H. S. Polym. Prepr. 2008, 49 (2), 73–74. 34. Asandei, A. D.; Saha, G. Polym. Prepr. 2005, 46 (2), 474–475. 35. Asandei, A. D.; Simpson, C. P.; Olumide, A.; Yu, H. S. Polym. Prepr. 2010, 51 (1), 553–554. 36. Asandei, A. D.; Yu, H. S.; Simpson, C. P. Polym. Prepr. 2010, 51 (2), 586–587. 37. Asandei, A. D.; Simpson, C. P.; Olumide, A.; Yu, H. S. Polym. Mater.: Sci. Eng. 2010, 102, 425–426. 38. Asandei, A. D.; Simpson, C. P.; Olumide, A.; Yu, H. S. Polym. Prepr. 2010, 51 (1), 498–499. 222

39. Asandei, A. D.; Yu, S. H.; Adebolu, O.; Simpson, C. P. Polym. Prepr. 2011, 52 (2), 470–471. 40. Asandei, A. D.; Moran, I. W. J. Am. Chem. Soc. 2004, 126, 15932–15933. 41. Asandei, A. D.; G Saha, G. Macromol. Rapid Commun. 2005, 26, 626–631. 42. Asandei, A. D.; Chen, Y.; Saha, G.; Moran, I. W. Tetrahedron 2008, 64, 11831–11838. 43. Asandei, A. D.; Chen, Y.; Adebolu, O. I.; Simpson, C. P. J. Polym. Sci., Part A: Polym. Chem. 2008, 46, 2869–2877. 44. Asandei, A. D.; Chen, Y.; Moran, I. W.; Saha, G. J. Organomet. Chem. 2007, 692, 3174–3182. 45. Asandei, A. D.; Chen, Y. Macromolecules 2006, 39, 7459–7554. 46. Asandei, A. D.; Chen, Y.; Simpson, C.; Gilbert, M.; Moran, I. W. Polym. Prepr. 2008, 49 (1), 489–490. 47. Asandei, A. D.; Chen, Y.; Adebolu, O. Polym. Mater.: Sci. Eng. 2008, 98, 370–371. 48. Asandei, A. D.; Saha, G. Polym. Prepr. 2007, 48 (2), 272–273. 49. Asandei, A. D.; Chen, Y. Polym. Mater.: Sci. Eng. 2007, 97, 450–451. 50. Wang, J.; Matyjaszewski, K. Macromolecules 1995, 28, 7901–7910. 51. Percec, V.; Barboiu, B. Macromolecules 1995, 28, 7970–7972. 52. Kato, M.; Kamigaito, M.; Sawamoto, M.; Higashimura, T. Macromolecules 1995, 28, 1721–1723. 53. Percec, V.; Guliashvili, T.; Ladislaw, J. S.; Wistrand, A.; Stjerndahl, A.; Sienkowska, M. J.; Monteiro, M. J.; Sahoo, S. J. Am. Chem. Soc. 2006, 128, 14156–14165. 54. Xia, J.; Paik, H.; Matyjaszewski, K. Macromolecules 1999, 32, 8310–8314. 55. Asandei, A. D.; Adebolu, O. I.; Simpson, C. P. J. Am. Chem. Soc. 2012, 134, 6080–6083. 56. Asandei, A. D.; Adebolu, O. I.; Simpson, C. P.; Kim, J. Angew. Chem., Int. Ed. 2013, 52, 10027–10030. 57. Asandei, A. D.; Adebolu, O. I.; Simpson, C. P. ACS Symp. Ser. 2012, 1106, 47–63. 58. Asandei, A. D.; Adebolu, O. I.; Simpson, C. P. In Handbook of Fluoropolymer Science and Technology; Smith. D. W., Iacono, S. T., Iyer, S. S., Eds.; John Wiley & Sons, Inc., 2014; pp 21−42. 59. Simpson, C. P.; Adebolu, O. I.; Kim, J. S.; Vasu, V.; Asandei, A. D. ACS Symp. Ser. 2015, 1187, 183–209. 60. Cernoch, P.; Petrova, S.; Cernochova, Z.; Kim, J. S.; Simpson, C. P.; Asandei, A. D. Eur. Polym. J. 2015, 68, 460–470. 61. Sun, F. C.; Dongare, A. M.; Asandei, A. D.; Alpay, S. P.; Nakhmanson, S. J. Mater. Chem. C 2015, 3, 8389–8396. 62. Simpson, C. P.; Adebolu, O. I.; Kim, J. S.; Vasu, V.; Asandei, A. D. Macromolecules 2015, 48, 6404–6420. 63. Asandei, A. D. Chem. Rev., 116, 2244–2274. 64. Ghosh, A.; Louis, L.; Asandei, A. D.; Nakhmanson, S. Soft Mater. 2018, 14, 2484–2491. 65. Wootthikanokkhan, J.; Peesan, M.; Phinyocheep, P. Eur. Polym. J. 2001, 37, 2063–2071. 223

66. Heo, S. U.; Rhee, G. H.; Lee, D. H.; Kim, J. H.; Choi, D. S.; Lee, D. W. J. Ind. Eng. Chem. 2006, 12, 241–247. 67. Hua, J.; Xu, H.; Geng, J.; Deng, Z.; Xu, L.; Yu, Y. J. Pol. Res. 2011, 18, 41–48. 68. Li, J.; El-harfi, J.; Howdle, S. M.; Carmichael, K.; Irvine, D. J. Polym. Chem. 2012, 3, 1495–1501. 69. Asandei, A. D. Polymerization of Diene Monomers. US Patent 9,862,789 B2, 2018. 70. Asandei, A. D.; Yu, S. H. Polym. Prepr. 2011, 52 (2), 584–585. 71. Asandei, A. D.; Yu, S. H. Polym. Prepr. 2011, 52 (2), 564–565. 72. Asandei, A. D.; Simpson, C. P.; Adebolu, O.; Chen, Y. Polym. Prepr. 2011, 52 (2), 759–560. 73. Asandei, A. D.; Yu, H. S. Polym. Prepr. 2011, 52 (1), 415–416. 74. Asandei, A. D.; Yu, H. S. Polym. Prepr. 2011, 52 (1), 413–414. 75. Asandei, A. D.; Yu, H. S. Polym. Mater.: Sci. Eng. 2011, 104, 619–620. 76. Asandei, A. D.; Yu, H. S. Polym. Mater.: Sci. Eng. 2011, 104, 627–628. 77. Vasu, V.; Kim, J. S.; Yu, H. S.; Bannerman, W. I.; Johnson, M. E.; Asandei, A. D. Polym. Chem. 2018 In Press, DOI:10.1039/C8PY00463C. 78. Gillies, M. B.; Matyjaszewski, K.; Norrby, P.; Pintauer, T.; Poli, R.; Richard, P. Macromolecules 2003, 36, 8551–8559. 79. Matyjaszewski, K.; Jo, S. M.; Paik, H.; Shipp, D. A. Macromolecules 1999, 32, 6431–6438. 80. Wang, X.; Zhao, H.; Li, Y.; Xiong, R.; You, X. Top. Catal. 2005, 35, 43–61. 81. Braunecker, W. A.; Tsarevsky, N. V.; Pintauer, T.; Gil, R. R.; Matyjaszewski, K. Macromolecules 2005, 38, 4081–4088. 82. Hakansson, M.; Brantin, K.; Jagner, S. J. Organomet. Chem. 2000, 602, 5–14. 83. Lin, C. Y.; Marque, S. R. A.; Matyjaszewski, K.; Coote, M. L. Macromolecules 2011, 44, 7568–7583. 84. Seeliger, F.; Matyjaszewski, K. Macromolecules 2009, 42, 6050–6055. 85. De Paoli, P.; Isse, A. A.; Bortolamei, N.; Gennaro, A. Chem. Commun. 2011, 47, 3580–3582. 86. Jakubowski, W.; Tsarevsky, N. V.; Higashihara, T.; Faust, R.; Matyjaszewski, K. Macromolecules 2008, 41, 2318–2323. 87. Deibert, S.; Bandermann, F. Makromol. Chem. 1993, 194, 3287–3299. 88. Tsarevsky, N. V.; Braunecker, W. A.; Matyjaszewski, K. J. Organomet. Chem. 2007, 692, 3212–3222. 89. Lutz, J.; Matyjaszewski, K. J. Polym. Sci., Part A: Polym. Chem. 2005, 43, 897–910. 90. Wang, Y.; Soerensen, N.; Zhong, M.; Schroeder, H.; Buback, M.; Matyjaszewski, K. Macromolecules 2013, 46, 683–691. 91. Anastasaki, A.; Waldron, C.; Wilson, P.; McHale, R.; Haddleton, D. M. Polym. Chem. 2013, 4, 2672–2675. 92. Ko, E. C. F.; Leffek, K. T. Can. J. Chem. 1972, 50, 1297–1302. 93. Jezorek, R. L.; Zhang, N.; Leowanawat, P.; Bunner, M. H.; Gutsche, N.; Pesti, A. K. R.; Olsen, J. T.; Percec, V. Org. Lett. 2014, 16, 6326–6329. 224

94. Rosen, B. M.; Quasdorf, K. W.; Wilson, D. A.; Zhang, N.; Resmerita, A. M.; Garg, N. K.; Percec, V. Chem. Rev. 2011, 111, 1346–1416. 95. Uegaki, H.; Kotani, Y.; Kamigaito, M.; Sawamoto, M. Macromolecules 1998, 31, 6756–6761. 96. Uegaki, H.; Kamigaito, M.; Sawamoto, M. J. Polym. Sci., Part A: Polym. Chem. 1999, 37, 3003–3009. 97. Moineau, G.; Minet, M.; Dubois, Ph.; Teyssié, P.; Senninger, T.; Jérôme, R. Macromolecules 1999, 32, 27–35. 98. Uegaki, H.; Kotani, Y.; Kamigaito, M.; Sawamoto, M. Macromolecules 1998, 31, 6756–61. 99. De Roma, A.; Yang, H. J.; Milione, S.; Capacchione, C.; Roviello, G.; Grassi, A. Inorg. Chem. Commun. 2011, 14, 542–544. 100. van de Kuil, L. A.; Grove, D. M.; Gossage, R. A.; Zwikker, J. W.; Jenneskens, L. W.; Drenth, W.; van Koten, G. Organometallics 1997, 16, 4985–4994. 101. Percec, V.; Barboiu, B.; Neumann, A.; Ronda, J. C.; Zhao, M. Macromolecules 1996, 29, 3665–3668. 102. Granel, C.; Dubois, Ph.; Jerome, T.; Teyssie, Ph. Macromolecules 1996, 29, 8576–8582. 103. O’Reilly, R. K.; Shaver, M. P.; Gibson, V. C. Inorg. Chim. Acta 2006, 359, 4417–4420. 104. Asandei, A. D.; Moran, I. W.; Brückner, C. Polym. Prepr. 2003, 44 (2), 827–828. 105. Asandei, A. D.; Moran, I. W.; Brückner, C. Polym. Prepr. 2003, 44 (1), 833–834. 106. Lecomte, Ph.; Drapier, I.; Dubois, Ph.; Teyssié, Ph.; Jérôme, R. Macromolecules 1997, 30, 7631–7633. 107. Reed, H. W. B. J. Chem. Soc. 1954, 1931–1941. 108. Otsu, T.; Yamaguchi, M. J. Polym. Sci., Part A.: Polym. Chem. 1969, 7, 387–394. 109. Buch, H. M.; Binger, P.; Goddard, R.; Kruger, C. J. Chem. Soc., Chem. Commun. 1983, 648–650. 110. Duh, Y. S.; Kao, C. S.; Lee, W. L. J. Therm. Anal. Calorim. 2017, 127, 1071–1087. 111. Magenau, A. J. D.; Kwak, Y.; Schröder, K.; Matyjaszewski, K. ACS Macro Lett. 2012, 1, 508–512. 112. Poli, R. Chem. Eur. J. 2015, 21, 6988–7001. 113. Poli, R. Eur. J. Inorg. Chem. 2011, 1513–1530. 114. Poli, R. Angew. Chem. Int. Ed. 2006, 45, 5058–5070. 115. Buback, M.; Morick, J. Macromol. Chem. Phys. 2010, 211, 2154–2161. 116. Wang, Y.; Schroeder, H.; Morick, J.; Buback, M.; Matyjaszewski, K. Macromol. Rapid. Comm. 2013, 34, 604–609.

225

Chapter 10

Reversible Deactivation Radical Polymerization of Vinyl Chloride Carlos M. R. Abreu,1 Ana C. Fonseca,1 Nuno M. P. Rocha,2 James T. Guthrie,3 Arménio C. Serra,1 and Jorge F. J. Coelho*,1 1CEMMPRE,

Department of Chemical Engineering, University of Coimbra, Rua Sílvio Lima-Pólo II, 3030-790 Coimbra, Portugal 2INEGI – Institute of Mechanical Engineering and Industrial Management, University of Porto, Rua Dr. Roberto Frias 400, 4200-465 Porto, Portugal 3Department of Colour Science, School of Chemistry, University of Leeds, LS2 9JT, United Kingdom *E-mail: [email protected].

Poly(vinyl chloride) (PVC) is one of the higher consumed polymers (more than 40 million tons per year) and can only be prepared on an industrial scale by free-radical polymerization (FRP). Several intrinsic limitations of FRP have triggered interest in synthesizing this polymer by reversible deactivation radical polymerization (RDRP) methods. Despite the many achievements that have been made, the RDRP of nonactivated monomers, such as vinyl chloride (VC), presents several challenges to the scientific community. Several features of VC make its control by RDRP techniques particularly difficult. The most recent developments on RDRP of VC are critically discussed.

Introduction Presently, poly(vinyl chloride), also known as PVC is the second most used polymer. PVC use has gone through steady and continuous growth, despite the various public debates that have been held because of the need to find halogen free polymers, for several health and environmental reasons. PVC has many important characteristics amongst which are low cost, general versatility and good flame retardancy. These have caused it to be widely used in applications © 2018 American Chemical Society

from packaging to construction. Also, as a result of its industrial production, PVC plays an important role because of the incorporation of chlorine (approximately 57% of the VC (vinyl chloride) mass), making use of the chlorine that is obtained as a secondary product of chlor-alkali industry. This factor means that VC supply has been less affected by fluctuation in oil prices compared with other fossil-based monomers. The presence of polar chloro groups improve the options provided by PVC in its blend with other compounds (plasticizers) or polymers. Currently, the only available way to synthesize PVC on a large scale is by free radical polymerization (FRP). This process has some side reactions that inevitably lead to structural defects. These unwanted structural defects are the cause of the low thermal stability of the polymer from which some major technological limitations arise (1, 2). In the PVC using industries, other issues faced include the need to make use of stabilizers and the restricted temperature range in which PVC can be processed, even when there are stabilizing compounds in the formulation. The impossibility of processing PVC at higher temperatures entails some problems. These include products’ incompatibility, incomplete fusion of PVC crystallites and the high melting viscosities that result in the need for specialized equipment. In addition, FRP makes it impossible to adopt the use of macromolecular engineering in creating PVC-based controlled structures. With the newly developed reversible deactivation radical polymerization (RDRP) methods, there are now some options when considering the synthesis of polymers that possess a controlled molecular weight, functional chain ends, complex architectures and designed topologies (3). The possibility of using a radical-based method that is capable suppressing most of the side reactions is very important when trying to prevent the occurrence of structural defects, especially if the PVC is to be used in subsequent macromolecular engineering (4). Before RDRP methods were adopted, the majority of the innovations that arose in the PVC industry were largely devoted to polymerization process optimizations, which aimed to improve the production efficiency at more reduced costs. Little has been achieved as far as understanding the polymerization mechanism is concerned. Previous efforts to polymerize VC with the use of ionic initiators, such as alkyl lithium, or metal atom-containing systems (e.g. metallocenes) have either not produced a PVC-based material that has the desired structures, or has made use of technologies that were impossible to replicate on an industrial scale.

Polymerization of Vinyl Chloride VC (CAS number 75-01-4) is an organo-chloro compound whose formula is H2C=CHCl. It has Q- and e- values of 0.44 and 0.2 respectively which indicates the low reactivity of the VC monomer and the very high reactivity of the corresponding radical. VC is categorized as a non-conjugated, weak-electron-withdrawing vinyl monomer (5). The free radical polymerization of VC gives one of the higher values of chain transfer (C) to monomer amongst the more commonly used monomers. For instance at 60 ºC, the values are 1.0 x 10-397%), vinyl acetate (TCI, >99%), and vinyl pivalate (TCI, >99%) were distilled on calcium hydride under reduced pressure before use. ZnCl2 (Aldrich, 1.0 M in Et2O) were used as received. 2,2-Azobis(4-methoxy-2,4-dimethylvaleronitrile) (V-70) (Wako, 95%) was purified by washing with acetone at –15 °C and evaporated until dry under reduced pressure. S-1-isobutoxyethyl O-ethyl xanthate (BEEX) and S-1-isobutoxyethyl N,N-diphenyl dithiocarbamate (BEDPDC) were synthesized according to the literatures (26). Ethyl acetate (KANTO; >99%) was distilled from calcium hydride before use. Cationic and Radical Interconvertible Simultaneous Polymerization A typical procedure for the cationic and radical interconvertible copolymerization is given below for CEVE and VPv. The reaction was initiated by addition of the prechilled ethyl acetate solution (0.5 mL) of V-70 (0.18 mmol) and ZnCl2 (0.0075 mmol) via dry syringes into the monomer solution (2.5 mL) containing VPv (1.33 mL, 9.0 mmol), CEVE (0.91 mL, 9.0 mmol), and BEDPDC (0.18 mmol) in ethyl acetate at 20 °C. The total volume of the reaction mixture was 3.0 mL. In predetermined intervals, the polymerization was terminated with methanol (1.0 mL) and then by cooling the reaction mixtures to –78 ˚C. The monomer conversions were determined from the concentration of the residual monomer measured by 1H NMR using ethyl acetate as an internal standard (CEVE; 55%, VPv; 21% for 18 h, and CEVE; 74%, VPv; 68% for 94 h, respectively). The quenched reaction mixture was washed with diluted hydrochloric acid, and distilled water to remove the catalyst residues, evaporated until dry under reduced pressure to obtain the product polymers (Mn = 4200, Mw/Mn = 1.25 for 18 h and Mn = 6400, Mw/Mn = 1.46 for 94 h). The incorporated ratio of CEVE/VPv in the copolymer was determined by 1H NMR (CEVE/VPv = 69/21 for 18 h and 52/48 for 94 h, respectively). Measurements 1H

and 13C NMR spectra were recorded on a JEOL ECS-400 spectrometer, operated at 400 and 100 MHz. The number-average molecular weight (Mn) and dispersity (Mw/Mn) of the copolymers were analyzed by size-exclusion chromatography (SEC) in THF at 40 °C through two polystyrene gel columns [Shodex KF-805 L (pore size: 20–1000 Å; 8.0 mm i.d. × 30 cm) × 2; flow rate 1.0 332

mL/min], which were calibrated against 10 standard polystyrene samples [Varian; peak-top molecular weight (Mp) = 575–2783000, Mw/Mn = 1.02–1.23], connected to a JASCO PU-2080 precision pump and a JASCO RI-2031 detector.

Acknowledgments This work was partially supported by Precursory Research for Embryonic Science and Technology (PRESTO) from the Japan Science and Technology Agency (JST) (No. JPMJPR14K8) for K.S., a Grant-in-Aid for Scientific Research (A) (No. 26248032) by the Japan Society for the Promotion of Science for K.S., and Program for Leading Graduate Schools “Integrative Graduate Education and Research Program in Green Natural Sciences”.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

Jenkins, A. D.; Jones, R. G.; Moad, G. Pure Appl. Chem. 2010, 82, 483–491. Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2005, 58, 379–410. Moad, G.; Rizzardo, E.; Thang, S. H. Acc. Chem. Res. 2008, 41, 1133–1142. Destarac, M. Polym. Rev. 2011, 51, 163–187. Handbook of RAFT Polymerization; Barner-Kowollik, C., Ed.; Wiley-VCH: Weinheim, Germany, 2008. Boyer, C.; Stenzel, M. H.; Davis, T. P. J. Polym. Sci., Part A: Polym. Chem. 2011, 49, 551–595. McKenzie, T. G.; Fu, Q.; Uchiyama, M.; Satoh, K.; Xu, J.; Boyer, C.; Kamigaito, M.; Qiao, G. G. Adv. Sci. 2016, 3, 1500394. Aoshima, H.; Uchiyama, M.; Satoh, K.; Kamigaito, M. Angew. Chem., Int. Ed. 2014, 53, 10932–10936. Satoh, K.; Hashimoto, H.; Kumagai, S.; Aoshima, H.; Uchiyama, M.; Ishibashi, R.; Fujiki, Y.; Kamigaito, M. Polym. Chem. 2017, 8, 5002–5011. Yagci, Y.; Tasdelen, M. A. Prog. Polym. Sci. 2006, 31, 1133–1170. Kumagai, S.; Nagai, K.; Satoh, K.; Kamigaito, M. Macromolecules 2010, 43, 7523–7531. Charmot, D.; Corpart, P.; Adam, H.; Zard, S. Z.; Biadatti, T.; Bouhadir, G. Macromol. Symp. 2000, 150, 23–32. Destarac, M.; Charmot, D.; Franck, X.; Zard, S. Z. Macromol. Rapid Commun. 2000, 21, 1035–1039. Stenzel, M. H.; Davis, T. P.; Barner-Kowollik, C. Chem. Commun. 2004, 1546–1547. Malepu, V.; Petruczok, C. D.; Tran, T.; Zhang, T.; Thopasridharan, M.; Shipp, D. A. ACS Symp. Ser. 2009, 1024, 37–47. Harrisson, S.; Liu, X.; Ollagnier, J.-N.; Coutelier, O.; Marty, J.-D.; Destarac, M. Polymers 2014, 6, 1437–1488. Hwang, J.; Lee, H.-C.; Antonietti, M.; Schmidt, B. V. K. J. Polym. Chem. 2017, 8, 6204–6208. Mishra, A. K.; Choi, C.; Maiti, S.; Seo, Y.; Lee, K. S.; Kim, E.; Kim, J. K. Polymer 2018, 139, 68–75. 333

19. Hedir, G. G.; Arno, M. C.; Langlais, M.; Husband, J. T.; O’Reilly, R. K.; Dove, A. P. Angew. Chem., Int. Ed. 2017, 56, 9178–9182. 20. Greenley, R. Z. In Polymer Handbook, 4th ed.; Brandrup, J., Immergut, E. H., Grulke, E. A., Eds.; Wiley-Interscience: New York, 1999; pp II-181–II-308. 21. Koumura, K.; Satoh, K.; Kamigaito, M.; Okamoto, Y. Macromolecules 2006, 39, 4054–4061. 22. Higashimura, T.; Law, Y. M.; Sawamoto, M. Polym. J. 1984, 16, 401–406. 23. Lipscomb, C. E.; Mahanthappa, M. K. Macromolecules 2009, 42, 4571–4579. 24. Roy, D.; Sumerlin, B. S. Polymer 2011, 52, 3038–3045. 25. Polyvinyl Alcohol–DeVelopments, 2nd ed.; Finch, C. A., Ed.; Wiley: Chichester, 1992. 26. Uchiyama, M.; Satoh, K.; Kamigaito, M. Angew. Chem., Int. Ed. 2015, 54, 1924–1928.

334

Chapter 16

Catalytic Chain Transfer Polymerization and Reversible Deactivation Radical Polymerization of Vinyl Acetate Mediated by Cobalt(II) Phenoxy-imine Complexes Yi-Hao Chen, Hung-Hsun Lu, Jia-Qi Li, and Chi-How Peng* Department of Chemistry and Frontier Research Center on Fundamental and Applied Sciences of Matters, National Tsing Hua University, Hsinchu, Taiwan 30013 *E-mail: [email protected].

Organo-cobalt complexes have demonstrated unique properties in the mediation of radical polymerizations such as catalytic chain transfer polymerization (CCTP) of methacrylates and reversible deactivation radical polymerization (RDRP) of unconjugated monomers particularly vinyl acetate (VAc). Two pathways of reversible termination (RT) and degenerative transfer (DT) have been rationalized as the major control mechanisms in cobalt mediated RDRP. In this chapter, cobalt(II) phenoxy-imine complexes mediated radical polymerization of vinyl acetate is reported. The increasing of electron donating property of phenoxy-imine ligand could switch the polymerization mechanism from RDRP to CCTP, demonstrating the potential of this system to combine CCTP and RDRP techniques.

© 2018 American Chemical Society

Introduction Chain transfer reaction was described as a side reaction in the chain growth polymerization process (1–6). The active species generated by initiators (usually carbon radicals, anions, or cations) not only can propagate with monomers but also react with other components in the polymerization such as solvent, monomer, or impurity via hydrogen abstraction to transfer the active centers from propagating chains to small molecules and thus causes the decreasing of average molecular weight of polymeric products. The schematic comparison of free radical polymerization and chain transfer polymerization was shown in Figure 1a. In 1975, Smironov and Marchenko et al. reported the application of cobalt porphyrin complex (Figure 2a) for the acceleration of chain transfer reaction in methyl methacrylate polymerization. The average molecular weight of poly(methyl methacrylate) decreased linearly with the equivalent of cobalt complex. The correlation of cobalt concentration and degree of polymerization was schematically shown in Figure 1b (7). The cobaloxime (Figure 2b) was then found as a highly efficient chain transfer agent for methyl methacrylate radical polymerization (8–11) and was commercialized as an industrial technique to produce polymeric products with low molecular weight for the convenience of following processing (1). Recently Haddleton et al. expanded the application of cobaloxime to the synthesis of sequential-controlled multiblock copolymers (12). Poli et al. also contributed to this field by developing new cobalt complexes for catalytic chain transfer polymerization (CCTP) of vinyl acetate (VAc) (13). More details of CCTP could be found in the review article published by Heuts et al. (1)

Figure 1. (a) Illustration of the difference between free radical polymerization and polymerization with chain transfer reaction; (b) Correlation of cobalt concentration and degree of polymerization in catalytic chain transfer polymerization. 336

Living polymerization was first reported by Szwarc (14) and defined as a polymerization process with no chain breaking reactions (chain transfer or chain termination), in which the molecular weight of polymeric products increases linearly with conversion, the molecular weight distribution is narrow, and the block copolymers can be obtained by sequential addition of second monomers. However, in radical polymerization, the radicals termination is inevitable so that the term of “living radical polymerization” has been revised to “controlled radical polymerization” or “controlled/living radical polymerization”, and then “reversible deactivation radical polymerization (15, 16)”. Therefore, in this chapter, “reversible deactivation radical polymerization (RDRP)” will be used to refer this technique. Cobalt complexes mediated RDRP was first reported by Wayland et al. in 1994 using cobalt(II) tetramesitylporphyrin (CoII(TMP), Figure 3a) to control the radical polymerization of methyl acrylate (17). At the same time, Harwood et al. used alkylcobaloximes (CoII(dmgH)2, Figure 3b) to mediate the light initiated radical polymerization of methyl acrylate (18). Another milestone of cobalt complexes mediated C/LRP was achieved by Jérôme et al. using cobalt(II) bis-acetylacetonate (CoII(acac)2, Figure 3c) to efficiently control the vinyl acetate radical polymerization (6). Afterward, The β-diketonato and β-ketiminato analogues of CoII(acac)2 (Figure 3d,e) (19), 1,3-Bis(2-pyridylimino) isoindolatocobalt(II) complexes (CoII(acac)(bpi), Figure 3f) (20), cobalt(II) [N,N’-bis(3,5-di-tert-butylsalicylidene)-1,2-cyclohexanediamine] (CoII(Salen*), Figure 3g) (21), and cobalt(II) bipyridine bisphenolate (CoII(bpybph), Figure 3h) (22) have been used to control the radical polymerization of acrylates, vinyl acetate, or other vinyl monomers.

Figure 2. Catalytic transfer agent: (a) cobalt tetramethoxy hematoporphyrin-IX and (b) cobaloxime. 337

Figure 3. Cobalt complexes used to mediate the reversible deactivation radical polymerization.

Both RDRP and CCTP mediated by cobalt complexes start from the reaction of cobalt(II) complexes and radicals to form the organo-cobalt(III) species (Scheme 1) (23–25). Therefore, an induction period is usually observed in cobalt complexes mediated RDRP and is rationalized as the time required to transform cobalt(II) and radicals to organo-cobalt(III) (6). Although the RDRP directly initiated by organo-cobalt(III) was reported, it needs the synthesis of corresponding cobalt complex in advance (25). When the Co-C bond in organo-cobalt(III) complexes can dissociate via bond homolysis to reversely generate cobalt(II) complexes and radicals that initiate the polymerization without external radical source, the polymerization could be controlled by the reversible termination mechanism (Scheme 1a, RT). If the Co-C bond is too strong for self-dissociation, an external radical source such as AIBN or V-70 is required to initiate the polymerization. The propagating radicals could rapidly exchange with the organic radicals on organo-cobalt(III) so that each radical 338

has similar chance to grow. This mechanism controlling radical polymerization by fast interchange of propagating radicals and dormant radicals was known as degenerative transfer (Scheme 1b, DT). These two pathways, RT and DT, can both achieve the reversible deactivation radical polymerization but are distinguished by the source and concentration of radicals. When the radicals are mainly from the bond homolysis of Co(III)-P and the radical concentration is dominated by the equilibrium of Co(II) and Co(III)-P ([P•] = [Co(III)-P]/([Co(II)] × Keq)), the polymerization is controlled mainly by RT mechanism. If the radicals are solely generated from initiator and the radical concentration is determined by the initiator concentration ([P•] = (ki[initiator]/2kt)1/2), the control of polymerization is approached via the DT pathway (19). There is another pathway in which the metal center of organo-cobalt(III) abstracts the hydrogen from the organic group to form the Co(III)-H and corresponding vinyl species (3). It should be noticed that the possibility of coexistence of β-H transfer via intermolecular reaction between cobalt(II) and radicals is not excluded (1, 2, 5, 13, 26). The cobalt hydride is highly active and can react with monomer or other vinyl compounds to generate organo-cobalt(III) species that release radicals. Since the replacement of long propagating radicals (P•) by small radicals (m•) matches the chain transfer reaction in the textbook (1–3, 5, 26) and is catalyzed by cobalt(II) complexes, this pathway was known as catalytic chain transfer (Scheme 1c, CCT).

Scheme 1. Correlation of Reversible Termination (RT), Degenerative Transfer (DT), and Catalytic Chain Transfer (CCT) Mechanisms in Cobalt Mediated Radical Polymerization

One important perspective for polymer synthesis is to combine different polymerization methods such as CCTP, RDRP, or ring opening polymerization to generate new polymeric products (27). The cobalt(II) porphyrin complexes have been reported to mediate CCTP or RDRP depending on the steric effect of the ligand. The cobalt(II) tetra(p-methoxyphenyl)porphyrin (CoII(TAP)) can mediate the CCTP of methyl acrylate with low equivalence of initiator but cobalt(II) 339

tetramesitylporphyrin (CoII(TMP)) can control the polymerization of methyl acrylate (5, 23). Herein, we are reporting the cobalt(II) phenoxy-imine complexes (Figure 4) that mediate RDRP or CCTP when the ligand has varied electron donating properties, which could contribute to the hybridization of CCTP and RDRP.

Figure 4. Cobalt(II) phenoxy-imine complexes used to mediate the radical polymerization of vinyl acetate in this chapter.

Experimental Materials 3-tert-Butyl-4-hydroxyanisole (Acros), tributylamine (J.T. Baker), tin(IV) chloride (Aldrich, 1.0 M in heptane), paraformaldehyde (Alfa Aesar, 97%), cyclohexylamine (Alfa Aesar, 98+%), cobalt(II) acetate tetrahydrate (Acros, 97%), sodium hydroxide (SHIMAKYU’S), thiophenol (Alfa Aesar, 99+%), tetrabutylammonium perchlorate (TBAP, TCI, 98%), AgNO3 (Acros), and 2,2’-azo-bisisobutyronitrile (AIBN, Showa) were used without any further purification. Toluene were dried by CaH2 before used. Deuterated solvents (Aldrich) were dried over molecular sieves. Vinyl acetate (Merck, 99%) was distilled under reduced pressure and degassed by three freeze pump thaw cycles before use. Measurement The NMR spectroscopy was used to characterize the structures of chemicals and monomer conversion. The spectrum was recorded by a Mercury-400 and Varian-500 spectrometer at 298 K. The chemical shifts in 1H NMR were shown in ppm refer to residual protons in CDCl3 as δ 7.24 ppm. 13C NMR chemical shifts were given in ppm refer to residual solvent in CDCl3 δ 77 ppm. Gel permeation chromatography (GPC) equipped with Ultimate 3000 liquid chromatograph associated with a 101 refractive index detector and Shodex columns (Shodex KF-802, Shodex KF-803, and Shodex KF-805) was used to analyze the polymeric products using THF as the eluent at 30 °C with 1.0 mL min-1 flow rate. The 340

calibration was based on narrow linear poly(styrene) Shodex standard (SM-105) ranging in molecular weight form 1.20 × 102 to 2.61 × 106 g mol-1. The Mn, Mw and polydispersity (PDI) of the polymeric products were calculated by DIONEX chromeleon software.

Scheme 2. Synthesis of Aldehyde Precursor Phenol (9.013g, 50.0 mmol) and tributylamine (5.0 mL, 20 mmol) were dissolved in toluene (50.0 mL) and stirred for 10 minutes. The SnCl4 was added dropwise to solution at room temperature under inert atmosphere. After 15 minutes stir, paraformaldehyde (2.078 g, 66.5 mmol) was added in solution. The solution was heated to 100 °C for another 12 hours (Scheme 2). The mixture was then extracted with ether and purified by column chromatography packed with fresh silica gel (hexane/EtOAc = 9/1) to give the product as yellow liquid (73.7%). 1H NMR (400 MHz, CDCl3) : δ (ppm) 11.51 (s, 1H, -OH), 9.79 (s, 1H, CHO), 7.15 (d, J = 1.5 Hz, 1H, Ar-H), 6.78 (d, J= 1.5 Hz, 1H, Ar-H), 3.78 (s, 3H, -OCH3), 1.39 (s, 9H, -tBu). 13C NMR (100 MHz, CDCl3) : δ (ppm) 196.20 (-CH=O), 155.80 (Ar), 151.72 (Ar), 139.76 (Ar), 123.58 (Ar), 119.58 (Ar), 111.48 (Ar), 55.56 (-OCH3), 34.93 (-C(CH3)3), 29.09 (-C(CH3)3).

Scheme 3. Synthesis of Phenoxy-imine (Ligand) Methanol (30 mL) solution of aldehyde precursor (2.941g, 14.0 mmol) and amine (1.611ml for ligand a, 1.737g for ligand b, 14.0 mmol) was reflux under inert atmosphere for 6 hours (Scheme 3). The precipitate was collected by vacuum filtration and washed by iced methanol to give the yellow solid (89.5%) of phenoxy-imine ligand a (28, 29) and the orange solid (63.5%) of phenoxy-imine ligand b (30). 1H NMR of phenoxy-imine ligand a (R = -cyclohexyl)(CDCl3, 400 MHz) : δ (ppm) 8.31 (s, 1H, -HC=N), 6.94 (d, J= 1.5 Hz, 1H, Ar-H), 6.59 (d, J= 1.4 Hz, 1H, Ar-H), 3.75 (s, 3H, -OCH3), 3.15-3.24 (m, 1H, -C=NCH), 1.50-1.90 (m, 10H, -Cy-H), 1.41 (s, 9H, -CCH3). 13C NMR of phenoxy-imine 341

ligand a (R = -cyclohexyl)(CDCl3, 100 MHz) : δ (ppm) 162.46 (-CH=N), 154.80 (Ar), 150.82 (Ar), 138.70 (Ar), 117.95 (Ar), 117.61 (Ar), 111.24 (Ar), 67.74 (=NH-C), 55.82 (-OCH3), 35.03 (-C(CH3)3), 34.47 (-C(CH3)3), 29.36 (-Cy), 25.65 (-Cy), 24.57 (-Cy). 1H NMR of phenoxy-imine ligand b (R=-p-methoxybenzyl) (CDCl3, 400 MHz) : δ (ppm) 8.57 (s, 1H, -HC=N), 7.27 (d, J= 4.4 Hz, 2H, Ar-H), 7.00 (d, J= 1.6 Hz, 1H, Ar-H), 6.92 (d, J= 4.4 Hz, 2H, Ar-H), 6.72 (s, 1H, Ar-H), 3.82 (s, 3H, -OCH3), 3.79 (s, 3H, -OCH3), 1.44 (s, 9H, -CCH3). 13C NMR of phenoxy-imine ligand b (R=-p-methoxybenzyl)(CDCl3, 100 MHz) : δ (ppm) 160.65 (-CH=N), 158.41 (Ar), 154.68 (Ar), 151.17 (Ar), 141.13 (Ar), 138.95 (Ar), 122.07 (Ar), 118.68 (Ar), 118.36 (Ar), 114.40 (Ar), 111.61 (Ar), 55.75 (-OCH3), 55.47 (-OCH3), 35.08 (-C(CH3)3), 29.33 (-C(CH3)3).

Scheme 4. Synthesis of CoII(Phenoxy-imine)2 (Mediator) The methanol solution (20 mL) of cobalt acetate tetrahydrate (0.126g, 0.49 mmol) and NaOH (0.043g, 1.07 mmol) was added dropwise to a methanol solution (10 mL) of phenoxy-imine (0.310g for a, 0.335g for b, 1.07 mmol). The mixture was refluxed for 6 hours (Scheme 4) then the precipitate was collected by vacuum filtration and washed by ice methanol to give the orange powder (89.5%) complex a (R = -cyclohexyl) and the orange-red powder (73.6%) complex b (R = -p-methoxybenzyl). Mass spectrum (FAB, m/z) for complex a was 635, calc. exact mass C36H52CoN2O4 635 and for complex b was 683, calc. exact mass C38H44CoN2O6 683. Polymerization The desired amount of cobalt(II) complexes and AIBN were mixed in a 50 mL Schlenk flask with three vacuum/nitrogen cycles to remove oxygen. Dry, degassed vinyl acetate was subsequently injected into the Schlenk flask by syringe under inert atmosphere. The mixture was stirred and heated to 60 °C and the monomer conversion was followed by 1H NMR.

Results and Discussion The polymerization of vinyl acetate was mediated by cobalt(II) phenoxyimine of complex a or b in bulk at 60 °C with the condition of [Co(II)]0/ [AIBN]0/[VAc]0=1/10/700 (Table 1). An induction period, which was rationalized as the time required to transform the cobalt(II) complexes and radicals to 342

organo-cobalt(III) species (6), was observed in the polymerization mediated by complex a, followed by the linear increased conversion with time to reach 34% monomer conversion within 110 minutes (Figure 5 diamond). Increasing the equivalent of AIBN from 10 to 20 (Table 1, entry 2) shortened the induction period and elevated the polymerization rate to approach 39% conversion within 80 minutes (Figure 5 circle). These observations match the typical phenomena of cobalt complexes mediated RDRP (31).

Table 1. Polymerization of VAc Mediated by CoII(Phenoxy-imine)2 in Bulk at 60 °C Entry

Complex

Condition

Time (mins)

Conv. (%)a

Mnb

Mn,thc

PDIb

1

a

1/10/700

110

34

59000

20500

1.80

2

a

1/20/700

80

39

46700

23500

1.94

3

b

1/10/700

30

33400

18100

1.70

480 1H

Conversion was measured by NMR. Mn was determined by gel permeation chromatography (GPC) with polystyrene as standard. c Mn,th=([VAc]0/[CoII]0)×(M.W. of VAc)×Conv. a

b

Figure 5. The plots of conversion versus time for VAc polymerization mediated by complex a at 60 °C in bulk under the condition of [complex a]0/[AIBN]0/[VAc]0=1/10/700 (diamond) and [complex a]0/[AIBN]0/[VAc]0=1/20/700 (circle).

The smoothly shifted GPC traces demonstrated another feature of RDRP (Figure 6). The molecular weight of poly(vinyl acetate) showed a moderate linear relationship to monomer conversion (Figure 7). The gap between the measured molecular weight and theoretical one implied that the polymerization may be 343

mainly controlled by reversible termination mechanism, which was supported by the observation of decreased molecular weight with increased equivalent of AIBN because the RT mechanism is based on the equilibrium of Co(II), radicals, and organo-Co(III) so that the theoretical molecular weight calculated by the assumption of one chain per cobalt would be lower than real molecular weight and the increasing of radicals raises the concentration of organo-cobalt(III) and thus decreases the molecular weight (23). However, the relatively high PDI values of 1.80 and 1.94 demonstrated the experimental condition could be further improved for a better control process.

Figure 6. The GPC traces of VAc polymerization mediated by complex a at 60 °C in bulk under condition of [complex a]0/[AIBN]0/[VAc]0=1/20/700.

Figure 7. The plots of Mn and PDI versus conversion for VAc polymerization mediated by complex a at 60 °C in bulk under the condition of [complex a]0/[AIBN]0/[VAc]0=1/10/700 (diamond) and [complex a]0/[AIBN]0/[VAc]0=1/20/700 (circle). 344

The VAc polymerization mediated by complex b also displayed an induction period followed by a linearly increased monomer conversion with time but the polymerization was slow and took 480 minutes to reach 30% conversion (Figure 8). However, given by the GPC results, the polymerization illustrated the features of catalytic chain transfer polymerization rather than reversible deactivation radical polymerization. The molecular weight approached to 37000 at the early stage of polymerization and stayed constant (Figure 9). The PDI values increased gradually with conversion from 1.49 to 1.70. To clarify the mechanism, the concentration of complex b was varied in the VAc polymerization and the molecular weight at similar monomer conversion around 8% was recorded (Figure 10). The Mayo plots showed that the degree of polymerization declined properly with the increased concentration of complex b and provided the solid evidence for the catalytic chain transfer mechanism. The observation of induction period indicated the formation of organo-cobalt(III) species before the CCTP and thus suggested that complex b should mediate the polymerization via an intramolecular β-H transfer reaction.

Figure 8. The plots of conversion versus time for VAc polymerization mediated by complex b at 60 °C in bulk under the condition of [complex b]0/[AIBN]0/[VAc]0=1/10/700.

The methyl acrylate polymerization mediated by CoII(TAP) and CoII(TMP) demonstrated the steric effect of the ligand to the polymerization mechanism (5, 23). The more bulky ligand has better chance to block the β-H transfer reaction and leads to the reversible deactivation radical polymerization. In our study, the cobalt(II) phenoxy-imine complexes of a and b altered the VAc polymerization mechanism by electronic effect of the ligand. The cobalt complex with more electron donating ligand prefers to mediate the polymerization via the catalytic chain transfer pathway. The polymerization results of complex a and b could contribute to the development of mediators for the combination or the switch of different polymerization mechanisms and thus novel synthetic methods for advanced materials. Further study of how other factors such as temperature, and radical concentration affect the polymerization mechanism and theoretical calculation for quantitative understanding of these results are being processed. 345

Figure 9. The plots of Mn and PDI versus conversion for VAc polymerization mediated by complex b at 60 °C in bulk under the condition of [complex b]0/[AIBN]0/[VAc]0=1/10/700.

Figure 10. The plots of DP versus [Co]/[VAc] (×10-3) for VAc polymerization mediated by complex b at 60 °C in bulk under different quantity of complex b. The molecular weight was recorded at the monomer conversion at 8%. 346

Conclusion The cobalt(II) phenoxy-imine complexes of a and b have been developed to mediate the radical polymerization of vinyl acetate. As expected, complex a showed a moderate control efficiency, demonstrated by the linearly increased molecular weight with conversion and smoothly shifted GPC traces, to the VAc polymerization. According to the molecular weight deviation and the correlation between concentration of AIBN and molecular weight of PVAc, the control mechanism was proposed to be reversible termination. However, complex b, a cobalt(II) phenoxy-imine complex similar to a with the more electron donating ligands, mediated the VAc polymerization by catalytic chain transfer mechanism. Associated with the results of methyl acrylate polymerization mediated by cobalt(II) porphyrins (5, 23), the more bulky, less electron donating ligands should be able to suppress the catalytic chain transfer pathway and promote the control efficiency of cobalt complexes to the radical polymerization.

Acknowledgments We thank the research funding supported by Ministry of Science and Technology, Taiwan, (MOST 104-2113-M-007-012-MY3).

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

13. 14.

Heuts, J. P. A.; Smeets, N. M. B. Polym. Chem. 2011, 2, 2407–2423. Gridnev, A. J. Polym. Sci. Polm. Chem. 2000, 38, 1753–1766. Gridnev, A. A.; Ittel, S. D. Chem. Rev. 2001, 101, 3611–3660. Haddleton, D. M.; Crossman, M. C.; Hunt, K. H.; Topping, C.; Waterson, C.; Suddaby, K. G. Macromolecules 1997, 30, 3992–3998. Debuigne, A.; Poli, R.; Jérôme, C.; Jérôme, R.; Detrembleur, C. Prog. Polym. Sci. 2009, 34, 211–239. Debuigne, A.; Caille, J.-R.; Jérôme, R. Angew. Chem., Int. Ed. 2005, 44, 1101–1104. Enikolopyan, N. S.; Smirnov, B. R.; Ponomarev, G. V.; Belgovskii, I. M. J. Polym. Sci., Polym. Chem. Ed. 1981, 19, 879–889. Kukulj, D.; Davis, T. P. Macromol. Chem. Phys. 1998, 199, 1697–1708. Heuts, J. P. A.; Forster, D. J.; Davis, T. P. Macromolecules 1999, 32, 3907–3912. Heuts, J. P. A.; Forster, D. J.; Davis, T. P.; Yamada, B.; Yamazoe, H.; Azukizawa, M. Macromolecules 1999, 32, 2511–2519. Burczyk, A. F.; O’Driscoll, K. F.; Rempel, G. L. J. Polym. Sci., Polym. Chem. Ed. 1984, 22, 3255–3262. Engelis, N. G.; Anastasaki, A.; Nurumbetov, G.; Truong, N. P.; Nikolaou, V.; Shegiwal, A.; Whittaker, M. R.; Davis, T. P.; Haddleton, D. M. Nat. Chem. 2017, 9, 171–178. Bellan, E. V.; Thevenin, L.; Gayet, F.; Fliedel, C.; Poli, R. ACS Macro Lett. 2017, 6, 959–962. Szwarc, M. Nature 1956, 178, 1168–1169. 347

15. Wang, J.-S.; Matyjaszewski, K. J. Am. Chem. Soc. 1995, 117, 5614–5615. 16. Jenkins, A. D.; Jones, R. G.; Moad, G. Pure Appl. Chem. 2009, 82, 483–491. 17. Wayland, B. B.; Poszmik, G.; Mukerjee, S. L.; Fryd, M. J. Am. Chem. Soc. 1994, 116, 7943–7944. 18. Yang, K.; He, T.; Chen, X.; Cheng, S. Z. D.; Zhu, Y. Macromolecules 2014, 47, 8479–8486. 19. Peng, C. H.; Yang, T. Y.; Zhao, Y.; Fu, X. Org. Biomol. Chem. 2014, 12, 8580–8587. 20. Langlotz, B. K.; Lloret Fillol, J.; Gross, J. H.; Wadepohl, H.; Gade, L. H. Chem.−Eur. J. 2008, 14, 10267–10279. 21. Liao, C.-M.; Hsu, C.-C.; Wang, F.-S.; Wayland, B. B.; Peng, C.-H. Polym. Chem. 2013, 4, 3098–3104. 22. Lin, Y.-C.; Hsieh, Y.-L.; Lin, Y.-D.; Peng, C.-H. Macromolecules 2014, 47, 7362–7369. 23. Wayland, B. B.; Peng, C.-H.; Fu, X.; Lu, Z.; Fryd, M. Macromolecules 2006, 39, 8219–8222. 24. Poli, R. Angew. Chem., Int. Ed. 2006, 45, 5058–5070. 25. Debuigne, A.; Champouret, Y.; Jérôme, R.; Poli, R.; Detrembleur, C. Chem.−Eur. J. 2008, 14, 4046–4059. 26. Roberts, G. E.; Barner-Kowollik, C.; Davis, T. P.; Heuts, J. P. A. Macromolecules 2003, 36, 1054–1062. 27. Zhang, J.; Wang, B.; Wang, L.; Sun, J.; Zhang, Y.; Cao, Z.; Wu, Z. Appl. Organomet. Chem. 2018, 32, 12703–12705. 28. Jamjah, R.; Nekoomanesh, M.; Pourjafar, T.; Zohuri, G. H.; Afshartaromi, F.; Notash, B. Acta Cryst. E 2011, 67, o1775–o1776. 29. Matsukawa, N. Olefin polymerization catalyst and process for olefin polymerization using this catalyst. Eur. Pat. Appl. EP 1013674 A1 20000628, 2000. 30. Fujita, T.; Matsui, N.; Matsukawa, N.; Mitani, M.; Nitahara, M.; Saito, J.; Tsuru, K. Olefin polymerization catalyst and olefin polymerization method using same. Jpn. Kokai Tokkyo Koho JP 2000230010 A 20000822, 2000. 31. Braunecker, W. A.; Matyjaszewski, K. Prog. Polym. Sci. 2007, 32, 93–146.

348

Chapter 17

Tailor-Made Poly(vinylamine)s via Thermal or Photochemical Organometallic Mediated Radical Polymerization Pierre Stiernet,1 Mathilde Dréan,1 Christine Jérôme,1 Patrick Midoux,2 Philippe Guégan,3 Jutta Rieger,3 and Antoine Debuigne1,* 1Center for Education and Research on Macromolecules (CERM), CESAM Research Unit, University of Liege (ULiege), Sart-Tilman, Allée de la Chimie 3, Bat. B6a, B-4000 Liège, Belgium 2Centre Biophysique Moléculaire, UPR4301 CNRS, Rue Charles Sadron, 45071 Orléans Cedex 2, France 3Sorbonne Universités, UPMC Univ Paris 06, CNRS, Institut Parisien de Chimie Moléculaire, Equipe Chimie des Polymères, 4 Place Jussieu, F-75005 Paris, France *E-mail: [email protected].

Poly(vinylamine) is a highly valuable class of polymer used in several applications. Although free radical polymerization has been extensively exploited for its synthesis, the preparation of poly(vinylamine) with low dispersity and controlled molar mass is barely developed. Recently, a great step was made in this direction via organometallic-mediated radical polymerization (OMRP) of N-vinylacetamides followed by hydrolysis of the pendent amide groups. This chapter summarizes, completes and put in perspective the main accomplishments in the OMRP of acyclic N-vinylamides for the controlled synthesis of both primary and secondary poly(vinylamine)s. Thermal and photochemical initiating systems are compared and the controlled thermally initiated radical polymerization of N-vinylacetamide is reported for the first time. The optimal hydrolysis conditions for producing the poly(vinylamine) derivatives as well as their potential as vectors for gene transfection are also presented.

© 2018 American Chemical Society

Introduction Polymers containing amino groups in their backbone or as side chains constitute a very important class of materials which exhibit remarkable properties such as protonability, ability to complex metals or polyelectrolytes, H-bonding capacity and affinity for many supports. Among them, poly(vinylamine) (PVAm) is of particular interest due to the high density of amino groups and has been used in many applications notably as flocculating agent (1), support for carbon dioxide capture (2), superabsorbent material (3) but also for gas separation (4, 5), paper reinforcement (6–8), water purification (9, 10) or surface modification (11). It cannot be prepared by simple radical polymerization of N-vinylamine because such enamine is unstable and takes part in a tautomeric equilibrium with the corresponding imine (12). Alternative syntheses have been developed and mostly consist in two steps processes, i.e. polymerization of a vinyl monomer bearing a masked amino group followed by deprotection of the amine. For example, the sequential radical polymerization of N-vinylphthalimide (NVPI) and hydrazinolysis leads to PVAm (Scheme 1, upper part) (13). Nevertheless, this strategy is not used industrially due to low atom economy and the elevated cost of this method. In this respect, polymerizing N-vinylamides, like N-vinylformamide (NVF) or N-vinylacetamide (NVA), via a conventional radical pathway followed by hydrolysis of the amides is a more interesting method that has been explored in both academia and industry (Scheme 1, lower part) (14–19). In contrast to the NVPI route, the second strategy also gives access to secondary poly(N-alkyl vinylamine)s via polymerization of N-alkyl vinylamides, like N-methyl vinylacetamide (NMVA), and subsequent hydrolysis (19, 20). Although the vast majority of the precursors of poly(vinylamine)s are prepared by free radical polymerization, the development of controlled synthesis of PVAm with precise molar mass (Mn) and low dispersity (Đ) is attractive in order to tune their properties. Nevertheless, N-vinylphthalimide and N-vinylamides are non-conjugated monomers deprived of good stabilizing group for radicals, which renders the reactivation of the dormant species and so the control of the polymerization challenging. Despite this, well-defined poly(N-vinylphthalimide)s (PNVPIs) were prepared by reversible addition fragmentation chain transfer (RAFT) using xanthates or dithiocarbamates as chain transfer agents (CTAs) before being converted into PVAm with controlled molecular parameters (13, 21, 22). On the other hand, the controlled radical polymerization of various N-vinylamides has been achieved by organometallic-mediated radical polymerization (OMRP) (23–27) which is based on the reversible deactivation of the growing radical chains by a transition metal complex. As reported elsewhere, the bis(acetylacetonato)cobalt complex (Co(acac)2) was found particularly efficient for controlling the radical polymerization of the so-called ‘less activated monomers’ (LAMs) (23), including N-vinylamides (28, 29). The present chapter aims to complete, summarize and put in perspective the main achievements in the field of the OMRP of acyclic N-vinylamides for the preparation of tailor-made poly(vinylamine) derivatives. Both thermally and photochemically (28, 29) initiated OMRP of different N-vinylacetamides are discussed and compared from the standpoint of efficiency. The successful 350

controlled radical homopolymerization of NVA via a thermal unimolecular initiating system is notably presented for the first time. Finally, the optimized hydrolysis conditions are presented as well as the performance of well-defined poly(vinylamine) derivatives as gene transfection carriers (30).

Scheme 1. General Strategy for the Preparation of Poly(vinylamine)s

Experimental Section Materials N-vinylacetamide (NVA) (> 98%, TCI), cobalt (II) acetylacetonate (Co(acac)2) (97%, Aldrich), 2,2′-azobis(4-methoxy-2,4-dimethyl valeronitrile) (V-70, t½ = 10 h at 30 °C) (> 98%, Wako) and 2,2’-azobis[2-methyl-N-(2hydroxyethyl)propioamide] (VA-086, t½ = 10 h at 86 °C) (> 98%, Wako) were used as received. N-methylvinylacetamide (NMVA) (> 98%, Aldrich) and vinyl acetate (> 99%, Aldrich) were purified by distillation under reduced pressure and degassed by freeze-drying cycle under vacuum. Propanethiol (PrSH) (99%, Aldrich), dichloromethane (CH2Cl2) (p.a.) and methanol (MeOH) (p.a.) were degassed by bubbling argon for 30 min. Dimethylformamide (DMF) was dried on molecular thieves and degassed by bubbling argon for 30 minutes. The alkyl-cobalt (III) adduct initiator (R-Co(acac)2, [Co(acac)2-(CH(OAc)-CH2)99%, TCI), benzyl methacrylate (BzMA) (>98%, TCI), N,N-dimethylaminoethyl methacrylate (DMAEMA) (>98.5%, TCI), poly(ethylene glycol) methacrylate (PEGMA) (average molecular weight = 300) (98%, Aldrich, USA), 2hydroxyethyl methacrylate (HEMA) (>95%, TCI), and glycidyl methacrylate (GMA) (>95%, TCI) were purified through an alumina column. Ethyl 2-bromophenylacetate (EPh-Br) (97%, Alfa Aesar), ethyl 2-bromoisobutyrate (EMA-Br) (98%, TCI), ethyl 2-iodo-2-phenylacetate (EPh-I) (>97%, TCI), ethyl 2-iodoisobutyrate (EMA-I) (>97%, TCI), I2 (>98%, TCI), TBA (98%, TCI), 367

BNI (>98%, TCI), 2,2′-azobis(2-methylpropionitrile) (AIBN) (98%, Wako Pure Chemical, Japan), and 2,2’-azobis(2,4-dimethylvaleronitrile) (V65) (95%, Wako) were used as received. The structures of EPh-Br, EMA-Br, and the monomers are shown in Figure 1.

Measurement The GPC analysis was performed on a Shodex GPC-101 liquid chromatograph (Tokyo, Japan) equipped with two Shodex KF-804L mixed gel columns (300 × 8.0 mm; bead size = 7 μm; pore size = 20–200 Å). The eluent was tetrahydrofuran (THF) or dimethyl formamide (DMF) at a flow rate of 1.0 mL/min (THF) or 0.8 mL/min (DMF) (40 °C). The DMF eluent included LiBr (10 mM). Sample detection and quantification were conducted using a Shodex differential refractometer RI-101 calibrated with known concentrations of polymer in solvent. The monomer conversion was determined from the peak area of GPC. The column system was calibrated with standard poly(methyl methacrylate)s (PMMAs) or polystyrenes. The NMR spectra were recorded on an AV 300 (300 MHz) (Bruker, Germany) at ambient temperature. A mixed solvent, i.e., 15% acetonitrile-d3 and 85% toluene-d8 (Cambridge Isotope Laboratories, USA), was used as the solvent.

Reaction of R-Br, TBA, and I2 in Low-Mass System A mixture of acetone-d6 (0.3 g), toluene-d8 (1.7 g), EPh-Br (80 mM), TBA (80 mM), and I2 (40 mM) was heated in a Schlenk flask at 70 °C under an argon atmosphere with magnetic stirring. The reaction mixture was analyzed with NMR.

Polymerization In a typical run, a mixture of monomer (3 mL), R-Br, TBA, and I2 was heated in a Schlenk flask at 70-80 °C under an argon atmosphere with magnetic stirring. After a prescribed time t, an aliquot (0.1 mL) of the solution was taken out by a syringe, quenched to room temperature, diluted by THF or DMF to a known concentration, and analyzed by GPC.

Preparation of Poly(methyl methacrylate)-Iodide (PMMA-I) A mixture of MMA (3 mL, 100 eq), EPh-Br (1 eq), TBA (1 eq), and I2 (0.5 eq) (in a condition given in Table 1 (entry 1)) was heated in a Schlenk flask at 70 °C under argon atmosphere with magnetic stirring. After 4 h, the mixture was quenched to room temperature and diluted with THF. The polymer was reprecipitated in hexane, collected by filtration, and dried in vacuo to give a poly(methyl methacrylate)-iodide (PMMA-I). Monomer conversion = 24%; Mn = 2800 and PDI = 1.26 before purification; Mn = 2900; PDI = 1.16 after purification. The purified PMMA-I was used as a macroinitiator in block polymerizations. 368

Results and Discussion Study of Transformation of R-Br to R-I We studied the transformation of R-Br using TBA and I2 to generate R-I in a low-mass system. We heated a mixture of EPh-Br (Figure 1) (80 mM (1 eq)), TBA (80 mM (1 eq)), and I2 (40 mM (0.5 eq)) at 70 °C. As a model of MMA medium (dielectric constant ε = 7.9), we used a mixed solvent of 15% acetonitrile-d3 (ε = 37.5) and 85% toluene-d8 (ε = 2.4). Figure 2a shows the 1H NMR spectrum after heating for 30 min. EPh-Br was transformed to EPh-I with a 68% conversion in this short period of time. This result clearly suggests that the transformation of EPh-Br to EPh-I is fast in the MMA polymerizations studied below (MMA medium at 70-80 °C). Unlike EPh-Br, the transformation was slow for EMA-Br (Figure 1) that is a unimer model of PMMA-bromide (PMMA-Br). Virtually no reaction occurred at 70 °C even for a long time 4 h. EPh-Br bears two stabilizing groups (an ester and a phenyl group), while EMA-Br bears only one stabilizing group (an ester group), resulting in a slower radical generation from EMA-Br than EPh-Br (and hence slower transformation). Thus, we used EPh-Br in the following polymerizations.

Figure 2. 1H NMR spectra of (a) EPh-Br (80 mM), TBA (80 mM), and I2 (40 mM) at 70 °C for 30 min, (b) pure EPh-I, and (c) pure EPh-Br in the solution of acetonitrile-d3/toluene-d8 (15/85). 369

Table 1. Bulk Polymerizations of MMA (8 M) with EPh-Br, TBA, and I2

370

entry

Target DPa

[EPh-Br]0/[TBA]0/[I2]0/ [AIBN]0 (mM)b

T (°C)

t (h)

conv (%)

Mn (Mn,theo)c

PDI

1

100

80/80/40/0

70

12

65

6800 (6500)

1.32

2

100

80/100/40/0

70

12

71

7200 (7100)

1.30

3

100

80/80/40/0

80

8

76

7100 (7600)

1.33

4

100

80/80/40/10

70

5

81

7200 (8100)

1.37

5

200

40/45/20/0

70

21

76

15000 (15000)

1.38

6

200

40/45/20/4

70

6

80

14000 (16000)

1.42

7

400

20/25/10/0

70

20

75

29000 (30000)

1.40

8

400

20/25/10/4

70

5.5

83

31000 (33000)

1.43

C1

–

80/80/0/0

70

3

47

340000 (–)

2.61

Target degree of polymerization at 100% monomer conversion = [MMA]0/[EPh-Br]0. [MMA]0/[EPh-Br]0/[TBA]0/[I2]0 = 100/1/1/0.5 for entry 1, for example. c Theoretical Mn calculated with [MMA]0, [EPh-Br]0, and conversion, assuming that all of EPh-Br is converted to EPh-I and all of the generated EPh-I initiates the polymerization.

a

b

Polymerizations of MMA We studied the bulk polymerizations of MMA (8 M, 100 eq) containing EPhBr (80 mM, 1 eq) as a precursor of the dormant species, TBA (80 mM, 1 eq) as a catalyst, and I2 (40 mM, 0.5 eq) as a capping agent at 70 °C. The molar ratio of MMA to EPh-Br was 100. Assuming that all of the R-Br is transformed to the R-I initiating dormant species, the degree of polymerization (DP) expected at a 100% monomer conversion is 100. Figure 3 (filled circles) and Table 1 (entry 1) show the result. In the first 1 h, the transformation of EPh-Br to EPh-I occurred; the polymerization did not take place until all of I2 was consumed (all of the EPh-Br was transformed to EPh-I) (Figure 3a (filled circles)). In this transformation, a Br•/TBA complex is generated (Scheme 2a). Because the Br•/TBA radical is not a stable radical, it recombines with another Br•/TBA radical to form a Br2/(TBA)2 complex (Scheme 2c). A part of TBA is subsequently released from the Br2/(TBA)2 complex in the equilibrium (Scheme 2d). The regenerated TBA worked as a catalyst in the subsequent polymerization (Scheme 1).

Figure 3. Plots of (a) ln([M]0/[M]) vs t and (b) Mn and Mw/Mn vs conversion for the MMA/EPh-Br/TBA/I2/(AIBN) systems: [MMA]0 = 8 M; [EPh-Br]0 = 80 mM; [TBA]0 = 80 or 100 mM, [I2]0 = 40 mM; [AIBN]0 = 0 or 10 mM (Table 1 (entries 1-4)). The symbols and temperatures are indicated in the figure. 371

The monomer conversion reached 65% for 12 h. The number-average molecular weight (Mn) agreed with the theoretical value (Mn,theo), and the polydispersity index (PDI) (= Mw/Mn) (or dispersity) was approximately 1.2-1.3 from an early stage of polymerization, where Mw is the weight-average molecular weight. The good agreement of Mn with Mn,theo from an early stage of polymerization suggests the quantitative transformation of EPh-Br to EPh-I and the uniform initiation from the generated EPh-I. These results demonstrate the effective use of EPh-Br as a precursor of the dormant species. For comparison, Table 1 (entry C1) shows the polymerization in the same condition but without I2. In the absence of I2, the radical generated from EPhBr is not capped with iodide but just undergoes propagation (reacts with MMA), resulting in conventional radical polymerization (38, 39). As shown in Table 1 (entry C1), the polymerization proceeded, and the monomer conversion reached 47% for 3 h. This result confirms the radical generation from EPh-Br with TBA as supposed in Scheme 2a. The Mn was 340,000 and PDI was 2.61. The observed high molecular weight and large PDI value show that the radical generation from EPh-Br with TBA is virtually irreversible (Scheme 2a).

Increase in Polymerization Rate In the mentioned system (Figure 3 (filled circles) and Table 1 (entry 1)), a relatively long time was required to achieve a high monomer conversion (i.e., 65% for 12 h). To increase the polymerization rate (Rp), we increased the amount of the TBA catalyst from 80 mM to 100 mM (Figure 3 (squares) and Table 1 (entry 2)). The Rp became approximately 1.2 times larger; the monomer conversion reached 71% for 12 h. Instead of increasing [TBA], we elevated the temperature from 70 °C to 80 °C (Figure 3 (triangles) and Table 1 (entry 3)). The Rp became approximately twice larger; the monomer conversion reached 76% for 8 h. Another effective method to increase Rp was the addition of a small amount of an azo initiator, azobis(isobutyronitrile) (AIBN), at 70 °C (Figure 3 (open circles) and Table 1 (entry 4)). Azo initiators are often used to decrease the deactivator concentration and hence effectively increase Rp in other LRP systems such as ATRP and nitroxide-mediated polymerizations (40). Compared with the system without AIBN, the addition of AIBN (10 mM) increased Rp by a factor of approximately 4, while it kept good control in Mn and small PDI values. The relatively high Rp (81% monomer conversion for 5 h) as well as the small PDI (1.23-1.37) values is attractive for practical use.

Higher Molecular Weight We targeted higher DPs of 200 and 400 in the polymerizations of MMA (Figure 4 and Table 1 (entries 5-8)). The addition of a small amount of AIBN was also effective for fast polymerization (Figure 4 (open symbols) and Table 1 (entries 6 and 8)). We obtained relatively low-polydispersity (PDI = 1.42-1.43) polymers with high monomer conversions (80-83%) for 5.5-6 h. 372

Table 2. Bulk Polymerizations of St and Functional Methacrylates (Target DP = 100)a entry

Monomer

[EPh-Br]0/[TBA]0/[I2]0/ [AIBN]0 (mM)b

T (°C)

t (h)

conv (%)

Mn c (Mn,theo)d

1

St

80/90/50/30

80

12

83

6700 (8600)

1.35

2

BzMA

80/80/40/5

70

12

73

12000 (13000)

1.36

3

PEGMA

80/100/40/10

80

40

75

15000 (23000)

1.30

4

DMAEMA

80/100/55/30

70

4

87

15000 (14000)

1.45

5

HEMA

80/100/45/15

70

12

77

11000 (10000)

1.40

PDI

c

Target degree of polymerization at 100% monomer conversion = [monomer]0/[EPh-Br]0. b [monomer]0 = 8 M. c GPC values calibrated with polystyrene (PSt) (entry 1) and PMMA (entries 2-5) standards. d Theoretical Mn calculated with [monomer]0, [EPh-Br]0, and conversion, assuming that all of EPh-Br is converted to EPh-I and all of the generated EPh-I initiates the polymerization.

a

373

Figure 4. Plots of (a) ln([M]0/[M]) vs t and (b) Mn and Mw/Mn vs conversion for the MMA/EPh-Br/TBA/I2/(AIBN) systems (70 °C): [MMA]0 = 8 M; [EPh-Br]0 = 20 or 40 mM; [TBA]0 = 25 or 45 mM, [I2]0 = 10 or 20 mM; [AIBN]0 = 0 or 4 mM (Table 1 (entries 5-8))

Styrene and Functional Methacrylates Table 2 shows the polymerizations of St (entry 1) and functional methacrylates with benzyl (BzMA), polyethylene glycol (PEGMA), dimethylamino (DMAEMA), and hydroxyl (HEMA) groups (entries 2-5). In all studied cases, the polymerization was controlled with high monomer conversions (≥73%), suggesting a large monomer scope of this method.

Block Polymerizations Exploiting the living character, we performed block polymerizations. Starting from the purified PMMA-I macroinitiator (Mn = 2900 and PDI = 1.16 after purification) synthesized for 4 h in Figure 3 (filled circles) and Table 1 (entry 1), the polymerizations of BzMA and GMA using BNI (24) as a catalyst (alternative to TBA) yielded block copolymers with relatively low polydispersities (PDI = 1.33-1.44) (Table 3). Figure 5 shows the full molecular weight distributions (GPC chromatograms). A large fraction of the macroinitiator chains were extended to block copolymers, confirming high block-efficiency. 374

Table 3. Block Polymerizations from PMMA-I

a

entry

Monomer

[PMMA-I]0/[BNI]0/ [V65]0 (mM)a

T (°C)

t (h)

conv (%)

1

BzMA

40/40/10

60

4

64

40000 (26000)

1.33

2

GMA

80/80/0

70

1

63

15000 (12000)

1.44

[monomer]0 = 8 M.

b

GPC values calibrated with PMMA standards.

c

Mn

b

(Mn,theo)c

PDI

Theoretical Mn calculated with [monomer]0, [PMMA-I]0, and conversion.

b

375

Figure 5. GPC chromatograms before (dashed lines) and after (solid lines) the block polymerizations in Table 3.

Conclusions Efficient transformation of EPh-Br to EPh-I in the presence of TBA and I2 was experimentally demonstrated. Based on this proof, R-Br was employed as a starting precursor, and the R-I formed in situ was employed as an initiating dormant species in the organocatalyzed LRP. The molecular weight and its distribution were well controlled in the polymerizations of MMA, St and functional methacrylates and their block polymerizations. This result suggests the accessibility of this system to a range of polymer designs. The use of simple, stable, and inexpensive R-Br as precursors of the dormant species is an attractive feature of this system.

Acknowledgments This work was partly supported by a Academic Research Fund (AcRF) Tier 2 funding from Ministry of Education in Singapore (MOE2017-T2-1-018) and a program of Accelerated Innovation Research Initiative Turning Top Science and Ideas into High-Impact Values (ACCEL) of the Japan Science and Technology Agency (JST).

References 1. 2.

Polymer Science: A Comprehensive Reference; Matyjaszewski, K., Möller, K., Eds.; Elsevier: Amsterdam, 2012. Fundamentals of Controlled/Living Radical Polymerization; Tsarevsky, N. V., Sumerlin, B. S., Eds.; Royal Society of Chemistry: Cambridge, 2013. 376

3. 4. 5. 6.

7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

The Chemistry of Radical Polymerization; Moad, G., Solomon, D. H., Eds.; Elsevier: Amsterdam, 2006. Matyjaszewski, K.; Tsarevsky, N. V. J. Am. Chem. Soc. 2014, 136, 6513–6533. Ouchi, M.; Sawamoto, M. Macromolecules 2017, 50, 2603–2614. Boyer, C.; Corrigan, N. A.; Jung, K.; Nguyen, D.; Nguyen, T.-K.; Adnan, N. N. M.; Oliver, S.; Shanmugam, S.; Yeow, J. Chem. Rev. 2016, 116, 1803–1949. Zhang, N.; Samanta, S. R.; Rosen, B. M.; Percec, V. Chem. Rev. 2014, 114, 5848–5958. Nicolas, J.; Guillaneuf, Y.; Lefay, C.; Bertin, D.; Gigmes, D.; Charleux, B. Prog. Polym. Sci. 2013, 38, 63–235. David, G.; Boyer, C.; Tonnar, J.; Ameduri, B.; Lacroix-Desmazes, P.; Boutevin, B. Chem. Rev. 2006, 106, 3936–3962. Keddie, D. J.; Moad, G.; Rizzardo, E.; Thang, S. H. Macromolecules 2012, 45, 5321–5342. Hill, M. R.; Carmean, R. N.; Sumerlin, B. S. Macromolecules 2015, 48, 5459–5469. Yamago, S. Chem. Rev. 2009, 109, 5051–5068. Satoh, K.; Kamigaito, M. Chem. Rev. 2009, 109, 5120–5156. Monteiro, M. J.; Cunningham, M. F. Macromolecules 2012, 45, 4939–4957. Debuigne, A.; Hurtgen, M.; Detrembleur, C.; Jérôme, C.; BarnerKowollik, C.; Junkers, T. Prog. Polym. Sci. 2012, 37, 1004–1030. Yamago, S.; Nakamura, Y. Polymer 2013, 54, 981–994. Mastan, E.; Li, X.; Zhu, S. Prog. Polym. Sci. 2015, 45, 71–101. Zetterlund, P. B.; Thickett, S. C.; Perrier, S.; Bourgeat-Lami, E.; Lansalot, M. Chem. Rev. 2015, 115, 9745–9800. Pan, X.; Tasdelen, M. A.; Laun, J.; Junkers, T.; Yagci, Y.; Matyjaszewski, K. Prog. Polym. Sci. 2016, 62, 73–125. Chen, M.; Zhong, M.; Johnson, J. A. Chem. Rev. 2016, 116, 10167–10211. Pearson, S.; St Thomas, C.; Guerrero-Santos, R.; D’Agosto, F. Polym. Chem. 2017, 8, 4916–6946. Goto, A.; Zushi, H.; Hirai, N.; Wakada, T.; Tsujii, Y.; Fukuda, T. J. Am. Chem. Soc. 2007, 129, 13347–13354. Goto, A.; Suzuki, T.; Ohfuji, H.; Tanishima, M.; Fukuda, T.; Tsujii, Y.; Kaji, H. Macromolecules 2011, 44, 8709–8715. Goto, A.; Ohtsuki, A.; Ohfuji, H.; Tanishima, M.; Kaji, H. J. Am. Chem. Soc. 2013, 135, 11131–11139. Ohtsuki, A.; Lei, L.; Tanishima, M.; Goto, A.; Kaji, H. J. Am. Chem. Soc. 2015, 137, 5610–5617. Sarkar, J.; Xiao, L.; Goto, A. Macromolecules 2016, 49, 5033–5042. Wang, C. G.; Goto, A. J. Am. Chem. Soc. 2017, 139, 10551–10560. Zheng, J.; Wang, C. G.; Goto, A. Angew. Chem. Int. Ed. 2018, 57, 1552–1556. Xiao, L.; Sakakibara, K.; Tsujii, Y.; Goto, A. Macromolecules 2017, 50, 1882–1891. Finkelstein, H. Ber. Dtsch. Chem. Ges. 1910, 43, 1528–1532. 377

31. Pace, R. D.; Regmi, Y. J. Chem. Edu. 2006, 9, 1344–1348. 32. Stevenson, D. P.; Coppinger, G. M. J. Am. Chem. Soc. 1962, 84, 149–152. 33. Lautenberger, W. J.; Jones, E. N.; Miller, J. G. J. Am. Chem. Soc. 1968, 90, 1110–1115. 34. Ishibashi, H.; Haruki, S.; Uchiyama, M.; Tamura, O.; Matsuo, J. Tetrahedron Lett. 2006, 47, 6263–6266. 35. Lacroix-Desmazes, P.; Severac, R.; Boutevin, B. Macromolecules 2005, 38, 6299–6309. 36. Tonnar, J.; Lacroix-Desmazes, P. Soft Matter 2008, 4, 1225–1260. 37. Balczewski, P.; Mikolajczyk, M. New J. Chem. 2001, 25, 659–663. 38. Sen, S. R.; Dass, N. N. Eur. Polym. J. 1982, 18, 477–480. 39. Dwivedi, P. C.; Pandey, S. D.; Tandon, A.; Singh, S. Colloid. Polym. Sci. 1995, 273, 822–827. 40. For a review on the kinetic of LRP Goto, A.; Fukuda, T. Prog. Polym. Sci. 2004, 29, 329–385.

378

Chapter 19

Biocatalytic ATRP Jonas Pollard and Nico Bruns* Adolphe Merkle Institute, University of Fribourg, Chemin des Verdiers 4, 1700 Fribourg, Switzerland *E-mail: [email protected].

Enzymes, their cofactors as well as enzyme mimetics can catalyze atom transfer radical polymerization (ATRP). Catalysts for this bioATRP are heme enzymes such as horseradish peroxidase, hemoglobin, and catalase, as well as the copper-containing laccases. Moreover, hemin, PEGylated hemin, heme-peptide conjugates, and single-chain polymer nanoparticles have been explored as ATRP catalysts. This chapter reviews biocatalytic ATRP as well as biocatalytically initiated RAFT polymerizations, and shows how “ATRPases” and other biologically derived catalysts are used, e.g., to tune well-controlled surface-initiated polymerizations, to confine ATRP into nanoreactors, and to prepare non-viral gene vectors or biosensors.

Introduction Free radical polymerization is the method of choice to synthesize many of the most produced polymers worldwide due to its simplicity and robustness. The chain growth proceeds through the subsequent addition of vinylic monomer to the radical at the chain end. The chain growth ends, for instance, when two radicals react together. The formation of each polymer chain is completed within approximately one second (1). Due to the speed of this process and the unavoidable termination reactions, chains exhibit a broad dispersity of molecular weight and lack of chain end functionality. The latter is a major setback for applications which require polymers with tailor-made properties. Two decades ago, different techniques such as nitroxide mediated polymerization (2), atom transfer radical polymerization (ATRP) (3, 4) and reversible addition-fragmentation chain transfer polymerization (RAFT) (5) were developed to circumvent the inherent problems of free radical © 2018 American Chemical Society

polymerizations and were named controlled radical polymerizations, or according to IUPAC reversible-deactivation radical polymerizations (RDRP) (6). It was now possible to synthesize narrowly dispersed polymers with predictable molecular weights. Moreover, these methods allowed for macromolecular engineering which did reinvigorate the field of polymer synthesis. ATRP has been one of the most investigated methods for controlled radical polymerizations due to the broad range of monomers it is able to polymerize but also due to its robustness (7). ATRP is based on the use of a catalytic system composed of a metal center which is complexed by a ligand. This catalytic system is able to abstract a halogen from an alkyl halide initiator, creating a radical which initiates the polymerization. Polymer chains grow from the initiator species. The growing chains are then subjected to a reversible deactivation by the catalyst which creates an equilibrium between the so-called dormant state and the active state. The equilibrium is pushed toward the formation of the dormant species, lowering the radical concentration throughout the reaction. This limits the occurrence of termination reactions by radical recombination or disproportionation. Thus, the molecular weight can be predefined by the ratio of monomer to initiator, and the chains ends are functionalized. The latter feature widened the applications of radical polymerizations. For example, high quality sealant resins with methoxysilyl chain ends are made on the industrial scale by ATRP (7). Moreover, ATRP allows the synthesis of polymers with complex architectures such as block copolymers, star polymers and polymer brushes (1). A major setback of conventional ATRP is the difficulty to remove the (mildly) toxic catalysts from the produced polymers, as it was used in large concentrations. The reduction of the amount of catalyst needed for ATRP has been achieved by “initiator for continuous activator regeneration” (ICAR) and “activators regenerated by electron transfer” (ARGET) ATRP (8). The first method relies on organic molecules that are able to produce radicals to regenerate the activator, e.g. CuI, while the latter makes use of a reducing agent for the same purpose. Both approaches counteract the progressive deactivation of the catalyst due to the accumulation of oxidized catalyst species, e.g. CuII, caused by termination reactions. Thus, the metal complex concentration in polymerization mixtures could be greatly reduced to as low as few ppm. Nevertheless, purification steps may still be needed and traces of toxic metals may still be found in the polymers (9). The presence of residual metal ions in the polymer is a barrier to applications of materials that are made for human contact such as food packaging and biomedical applications (7, 10). Moreover, they make use of ATRP-derived polymers in electronic applications difficult (7). This highlights the need for novel catalysts, one possibility being catalysts derived from nature, such as enzymes and their cofactors. Enzymes catalyze biochemical reactions but have also found widespread application in fine chemical synthesis because of their stereo-, regio-, and substrate selectivity, because they are generally non-toxic and because they work under mild conditions (11–13). Since the 1990s, they have been used for the in vitro synthesis of polymers such as polysaccharides and polyesters (14, 15). Enzymes catalyze several types of polymerizations, including ring opening polymerizations (15, 16) or polycondensations (15, 17). Moreover, enzymes can efficiently and catalytically initiated free radical polymerizations of vinyl 380

monomers (15, 18, 19). Peroxidases use peroxides for this purpose, while laccases use oxygen (18). The prosthetic group of peroxidases is heme, i.e. an iron ion complexed by a protoporpyhrine. Laccases, on the other hand, have four copper atoms embedded in their structure. Their catalytic activity arises from a complex interplay of this metal network. As these biocatalysts are naturally sourced and biodegradable, they are inherently environmentally friendly. (Nevertheless, a detailed assessment of the whole life cycle of such catalysts should be carried out to evaluate that they are indeed more environmentally friendly than other catalysts (11).) Moreover, enzymes display a high catalytic turnover and can be easily separated from the product (19). Another advantage is that they perform optimally under mild conditions i.e. room temperature, ambient pressure and in aqueous solutions. However, they were, until recently, not known to have the ability to control radical polymerizations.

Enzymatic ATRP In 2011, it was shown that certain metalloenzymes can mediate controlled polymerizations via an ATRP-like mechanism, so called biocatalytic ATRP (bioATRP). Our group discovered this novel enzymatic activity by using horseradish peroxidase (HRP) (20, 21) (and hemoglobin (Hb) (22)) in the presence of ARGET ATRP reagents for the polymerization of N-isopropylacrylamide (NIPAAm). The synthesis of the polymer could be proven by the precipitation of the polymer above its lower critical solution temperature (LCST) at 40° C. Control reactions without one of the reagents, i.e. reducing agent, catalyst or initiator yielded no polymer, thus confirming that all components of ARGET ATRP were needed. The analysis of the HRP-derived polymers via COSY 1H NMR confirmed the presence of the ATRP initiator in the polymer chains. Neutron activation analysis experiments revealed that 67% of chain ends were bromine-terminated, a finding which also pointed toward an ATRP mechanism. The reaction was highly dependent on pH and the optimal conditions at pH 6 yielded polymeric chains with a dispersity as low as 1.44. As none of the other reagents were affected by pH, this constituted the first evidence of the influence of this parameter on enzyme activity to catalyze bioATRP. Stability studies of the enzyme by various spectrometric methods showed no denaturation or change in conformation of the enzyme throughout the polymerization. Despite first order kinetics of the polymerization, the number average molecular weight decreased with time while the dispersity increased. This means that small chains were formed with time and that the enzyme was not able to efficiently control the reaction under the investigated conditions. Nevertheless, radical formation from an alkyl halide initiator, relative low dispersities and bromine-terminated polymer chains proved that the enzyme could catalyze the initiation as well as the bromination of growing chains in a reversible bromine transfer reaction. Thus, the findings represent a novel enzyme activity, which we termed ATRPase activity. Figure 1 shows enzymes and proteins with ATRPase activity. 381

Figure 1. Enzymes that have been used to catalyze bioATRP. HRP, Hb and CBL are heme proteins that contain iron protoporpyhrine IX as cofactor, while LTV is an enzyme that contains four copper ions in a defined arrangement.

Hemoglobin (Hb) can also catalyze ARGET ATRP reactions of vinyl monomers (22). Hb is a heme protein that is known to have peroxidase activity (23). It is slightly more complex than HRP in radical reactions, as it features cysteines on its surface. The thiol group of cysteine acts as a chain transfer agent, which leads to a sulfur-centered radical on the protein that initiates the growth of a new chain, thus leading to protein-polymer conjugates. To avoid this side reaction, the cysteines of hemoglobin were blocked using a maleimide. Hb was used to polymerize NIPAAm under ARGET ATRP conditions. The pH profile of the polymerization demonstrated that the hemoglobin activity was optimal at pH 4 as maximum conversion was obtained. The bromine termination of the chains was confirmed by chain extension polymerization using a classical ATRP catalyst. However kinetic measurements revealed that the molecular weight did not increase with conversion and the dispersity drastically increased throughout the reaction. In contrast, the polymerization of other monomers, poly(ethylene glycol) methyl ether acrylate (PEGA) and poly(ethylene glycol) methyl ether methacrylate (PEGMA) could be controlled more successfully, displaying a first order kinetic, linear increase in molecular weight with conversion and relatively low dispersity. In the case of PEGMA, the dispersity remained below 1.17. 382

Finally, hemoglobin contained in red blood cells could successfully polymerize NIPAAm, PEGA and PEGMA albeit with a lower degree of control. Concurrently to our work on ATRPases, di Lena and coworkers found laccase from Trametes versicolor (LTV) to initiate free radical polymerizations of PEGMA from the ATRP initiators ethyl α-bromoisobutyrate and 2-bromopropionitrile. Control could be achieved by adding a RAFT agent, making this the first enzyme-initiated RAFT polymerization (24). Moreover, emulsion copolymerization of styrene, 2-hydroxyethyl methacrylate (HEMA) and divinylbenzene (DVB), as well as surface-initiated polymerizations of PEGMA were demonstrated. In a further study, the team investigated polymerizations catalyzed by catalase from bovine liver (CBL) (25). Under ARGET ATRP conditions, PEGA could be polymerized in a controlled way. The reaction followed first order kinetics, and the molecular weight increased linearly with conversion. The chain length distribution was shown to be below 1.7 throughout the polymerization. The bromination of end-chain was proven by successful chain extension experiments. While optimizing the polymerization, they observed that reducing the amount of reducing agent accelerated the polymerization while the control was retained. Moreover, by premixing the reducing agent and the catalyst for one hour prior to introduction of monomer and initiator, the reaction was slowed down but the molecular weight became closer to the expected theoretical value. This indicated that the catalyst could be more efficiently reduced and the initiation was thus accelerated. The reactions were carried out at 60 °C which could have led to some denaturation of the enzyme. However, conducting similar polymerizations at 40 °C diminished considerably the conversion. Hence, the authors surmised that the enzyme did not denature at high temperature. Finally, it was shown that laccase and HRP could also catalyze PEGA polymerization with similar degree of control. The first reports on bioATRP by our group and by di Lena triggered the interest of many other groups to investigate controlled radical polymerizations catalyzed or mediated by biological catalysts. Matyjaszewski confirmed that catalase produces narrowly dispersed polymers under ARGET ATRP conditions (26). Chanana and coworkers immobilized HRP on gold nanoparticles and used the colloidal catalyst to polymerize NIPAAm by bioATRP. The gold nanoparticles allowed to recover the biocatalyst by centrifugation and therefore to reuse it several times (27). Sun et al. used hemoglobin to grow a molecularly imprinted polymer network, consisting of acrylamide crosslinked by N,N′-methylenebis(acrylamide), on the surface of an gold electrode, applying electrochemically mediated bioATRP (28). The protein acted as catalyst and as template molecule. The resulting electrode could be used to sense the presence of Hb by differential pulse voltammetry measurements with a detection limit of 7.8 10-11 mg L-1. The sensitivity of such protein-imprinted polymer biosensors could be further decreased to 3.2 10-14 mg mL-1 by conducting the electrochemically mediated, surface-initiated bioATRP on three-dimensional gold nanodenrites (29). Ko et al. (30) investigated the catalytic efficiency of a variety of enzymes via surface initiated ARGET ATRP of NIPAAm on lignin nanofibers. They demonstrated that CBL, LTV and HRP exhibited different activities depending on the concentration of ascorbic acid, as evaluated by the resulting brush thicknesses. 383

However, the effect of the pH and the concentration of the reducing agent was not investigated. Polymerizations catalyzed by HRP at pH 6 exhibited better control in comparison to the reactions catalyzed by LTV which gave broader dispersities. Nevertheless, they managed to achieve a high degree of control as degrafted chains of NIPAAm catalyzed by LTV reached Mn of more than 102 kg mol-1 with a dispersity of 1.29.

Figure 2. Tuning of hemoglobin-catalyzed surface-initiated bioATRP by the affinity of the surface towards proteins. A) Polymerization of NIPAAm at 37°C, when the polymer brushes are collapsed and protein adsorbing, and at 25 °C, when the polymer bruches are hydrated and protein repellent. B) Polymerization of PEGA at 25 °C and at 37°C. The polymer brushes are protein repellent at both temperatures. Adapted with permission from Ref. (31). Copyright 2017 American Chemical Society.

While these reports show that surface-initiated bioATRP (SI-bioATRP) is feasible, the interaction of the enzyme with growing polymer brushes is obviously a complex and dynamic process. Of particular interest is the question, how the protein affinity of the polymer layer influences bioATRP on surfaces. Therefore, we teamed up with the Benetti group to investigate surface-initiated bioATRP in depth (31). Comparison of the Hb-catalyzed polymerization of NIPAAm and PEGA at 25 °C and at 37 °C showed that protein brushes grow better if the 384

polymer layer is protein adsorbent than if it is protein repellent (Figure 2). The polymerization of NIPAAm could be switched from an effective SI-bioATRP at 37 °C, when the polymer brushes are in a collapsed and protein adherent state, to a less effective one at 25 °C, when the polymer is hydrated and therefore protein repellent. Strikingly, the growth of the polymer brushes did not show any signs of irreversible chain termination, so that the brushes could be chain extended several times, also allowing the synthesis of block copolymers of polyPEGA and PNIPAAm. Furthermore, bioATRP offers a route to NIPAAm films of precise thickness, as each reaction cycle yielded an increase in thickness of 4.3 ± 0.2 nm. To study the influence of nanoscale confinement on biocatalytic ATRP, our group encapsulated HRP within the protein cage thermosome (32), which is an archaeal chaperonin of approx. 16 nm diameter. It features large gated pores that allow macromolecules to enter and leave the internal folding chambers (33–36). BioATRP of PEGA was conducted in the thermosome and compared to HRPcatalyzed reactions in the absence of the protein cage. The reaction yielded smaller and more narrowly dispersed polymers when conducted in the protein nanoreactor. Another type of nanoreactors are block copolymer vesicles (polymersomes) (37–39). In order to fill polymersomes with a hydrophilic polymer (e.g. to create cell-mimetic systems that allow to study enzymatic reactions in crowded microenvironments), we encapsulated HRP in polymersomes, made the block copolymer membrane permeable for monomers and initiators by photoreaction with a radical forming photoreagent, and carried out bioATRP of PEGA (40). This system took advantage of the bulkiness of HRP, which allowed it to stay within the permeable nanoreactor whereas classical ATRP catalysts might diffuse out. NMR spectroscopy, GPC and light scattering showed that poly(PEGA) was formed in the polymersomes and that it filled the lumen of the vesicles. Moreover, cryoTEM revealed that the polymersomes remained stable during the polymerization. Recently, our group overcame a challenging polymerization of classic ATRP by carrying out bioATRP of N-vinylimidazole (NVIm) (41). NVIm is an industrial monomer, and its polymers find application in washing formulations, in cosmetics and in filtration membranes that remove copper ions from beer and wine. It would be an ideal candidate as a delivery agent for RNA or DNA, but unfortunately advanced polymer architectures comprising NVIm are difficult to achieve. It is very challenging to polymerize the monomer in a controlled fashion due to its high reactivity which makes it prone to chain termination and/or transfer (42, 43). Even worse, ATRP of this monomer is hampered by its strong tendency to complex metal ions and therefore to destroy conventional ATRP catalysts. In contrast, LTV catalyzed the controlled radical polymerization of NVIm. The advantage of LTV over conventional ATRP catalysts is the fact that its copper network is buried within the protein, thus preventing complexation with the monomer. Kinetic experiments showed that molecular weights increased linearly with conversion and that the reaction followed first order kinetics. Matrix assisted laser desorption ionization – time of flight (MALDI-ToF) revealed a dispersity of poly(NVIm) of 1.07 (Figure 3). This was the first example where enzymatic ATRP surpassed classical ATRP. Moreover, the bulky biocatalyst could be easily removed from the polymer, so that metal-free polymers were obtained. 385

Figure 3. MALDI-ToF mass spectrum of PNVIm synthesized by LTV-catalyzed bioATRP (Mn = 4300 g mol−1, Đ = 1.07). Reproduced with permission from Ref. (41). Copyright 2016 Royal Society of Chemistry.

Enzyme Mimetics and Cofactors as Catalysts for ATRP Enzyme-catalyzed ATRP inspired several groups to use prosthetic groups of enzymes as catalysts for ATRP. The reasoning is that the active center on its own might be less sensitive to its environment and reaction conditions as the full enzyme. Therefore, using solely the active center would make it a more robust catalyst. The first study on this type of catalysts was carried out by Kadokawa and coworkes (44). They reported the polymerization of NIPAAm catalyzed by hemin, the cofactor of HRP, hemoglobin and catalase. They could show that the polymer was formed through an ATRP-like mechanism. However, they obtained quite broad dispersities demonstrating that the cofactor itself was not sufficient to gain control over the reaction. Matyjaszewski’s team also catalyzed ATRP with hemin (26). They focused on the monomer PEGMA. Once again, due to the low halidophilicity of the catalyst, its low solubility in water and its possible incorporation in polymer chains via its double bonds, controlled polymerization could not be achieved with the native hemin. The addition of KBr resulted in polymers with narrower dispersities. Moreover, the conversion and degree of control was considerably enhanced by hydrogenating the vinyl double bonds of the heme and by PEGylating the catalyst (Figure 4). The former modification avoided that the heme copolymerized with the polymer, while the latter modification increased the water-solubility of the catalyst. As a result, well-controlled polymerizations could be achieved. The same group could also show that PEGylated hemin is a good catalyst to directly polymerize methacrylic acid with good control (45). This is remarkable, as methacrylic acid is notoriously known to be a very difficult monomer for ATRP. In contrast to conventional Cu-based catalysts, the robust enzyme-mimetic catalyst withstands acidic conditions. Moreover, as heme is an iron complex, it circumvents the problem of copper that deactivates bromine-terminated chains by chain-end cyclization. 386

Figure 4. Enzyme mimetic catalysts of bioATRP. Single-chain polymer nanoparticle reproduced with permission from Ref. (46). Copyright 2015 Wiley-VCH Verlag GmbH & Co. KGaA.

Buback, Matyjaszewski and coworkers followed up on the use of PEGylated hemin as catalyst (47). Their study aimed to understand the mechanism and kinetic of the previously observed polymerization, specifically the contribution of ATRP 387

and organometallic radical polymerization (OMRP). Looking at the Fe species via UV-Vis, they could observe that sodium ascorbate only reduces the active center slowly and partially. Thus, they used sodium dithionite (Na2S2O4) to reduce the metal center more efficiently and tuned the conditions so that Na2S2O4 completely disappeared upon complete catalyst reduction. Bromine abstraction upon initiator addition was immediate as the spectra of the catalyst returned to its native oxidized state, suggesting that FeII was highly active. Electron paramagnetic resonance (EPR) spectroscopic measurements of the deactivation reaction revealed that this catalyst had the highest deactivation constant reported for an iron catalyst at room temperature. Moreover, they showed that the deactivation constant increased with water concentration. As the reaction constant for ATRP is higher than for OMRP, they could conclude that the polymerization was predominantly governed by an ATRP mechanism. Tang et al. decorated deuterohemin with a small sequence of 6 amino acids to make it water-soluble (Figure 4) (48). This pseudo-enzyme was used to catalyzed the controlled polymerization of PEGMA. Using a DMF-H2O mixture and the addition of KBr, they observed linear increase of molecular weight with conversion as well as very low dispersity. Using a macroinitiator, they confirmed by 1H NMR that ATRP initiation took place. They also demonstrated that their catalyst was not sensitive to changes in pH between 3 and 11, which is a benefit in comparison to full enzymes. Interestingly, they showed that at pH 2 no polymer was formed but that the polymerization was triggered by the addition of NaOH to reach pH 3. This indicates that extreme pH can inhibit the catalytic activity of heme groups, but that this inhibition is reversible. In further work they demonstrated that the same catalyst could polymerize the grafting of N,N-(dimethylamino)ethyl methacrylate (DMAEMA) from mesoporous silica nanoparticles (49). The grafting of the polymer chains on the surface of the nanoparticles was attested by IR, elemental analysis, thermogravimetric analysis (TGA) and X-ray photoelectron spectroscopy (XPS). However, no analysis was performed to characterize the livingness of the polymerization. Using the pH-responsiveness of PDMAEMA chains, the authors could tune the release of a model cargo by lowering the pH. Deuterohemin-peptide enzyme mimics were also investigated as ATRP catalyst in metal-organic frameworks (50). Finally, they recently reported the synthesis of amphiphilic block copolymers poly(ε-caprolactone)-block-poly(glycidyl methacrylate) by ring opening polymerization mediated by Novozym 435 followed by bioATRP using deuterohemin (51). The resulting polymer was explored as amphiphilic non-viral gene vector. Single-chain polymer nanoparticles can be considered as enzyme mimetics as they possess metal ions embedded within a soft macromolecular structure. Pomposo et al. synthesized single polymer chain globular core-shell nanostructures containing copper ion complexes, thus resembling to some extend the active center of LTV (Figure 4) (46). Polymerization of PEGMA under ARGET ATRP conditions yielded polymers, but with broad molecular weight distributions. In combination with a RAFT agent, very well controlled radical polymerizations resulted. 388

Enzymatic Initiation of RAFT Polymerization Even though biocatalytic RAFT polymerizations are not the prime focus of this book chapter, they will be reviewed here in brief form, since they also developed out of bioATRP. For bioRAFT, the radicals of a polymerization are formed by enzymatic catalysis, but the radical polymerization is controlled through the addition of conventional RAFT agents. Thus, the biocatalysts replace the radical initiators in RAFT polymerizations. One way to do that is to rely on the ability of metalloenzymes to create radicals from ATRP initiators. Indeed, the fist publication by di Lena, mentioned above, did exactly that (24). Polymerizations of PEGMA were carried out using LTV as a catalyst in the presence of ascorbic acid and ethyl α-bromoisobutyrate (EBIB) in water. These polymerizations yielded very high molecular weight polymers with broad molecular weight distributions. The addition of the RAFT agent 2-cyano-2-propyl dithiobenzoate to the reagents yielded polymers of much lower molecular weights and dispersities as low as 1.35. First order kinetics and linear increase of molecular weight with conversion proved the controlled character of the polymerization. Another example of ATRP-initiated bioRAFT is the work of Pomposo on single chain globules discussed above (46). As peroxidases are well known to create radicals from peroxides, e.g. to initiate free radical polymerizations (18, 19), they can also be used to initiate RAFT polymerizations. An and coworkers used HRP in combination with hydrogen peroxide and the mediator acetylacetone for this purpose and obtained narrowly dispersed polymers for a wide range of monomers including N,N-dimethylacrylamide (DMA), N-vinylpyrrolidone, 2-hydroxyethyl acrylate, and PEGA (52). Moreover, a successful aerobic controlled polymerization was carried out via the glucose oxidase/HRP reaction cascade using hydrogen peroxide and acetyl acetonate (52) as well as hydrogen peroxide and ascorbic acid (53) as the substrate combination for HRP. Furthermore, glucose oxidase was replaced with pyranose oxidase in such cascade reactions to deoxygenated the reaction mixtures while creating hydrogen peroxide that allowed the peroxidase to initiate RAFT polymerizations (54). Ultrahigh molecular weight polymers, such as poly(DMA) with molecular weights of more than 2.6 106 g mol-1 and dispersity of 1.35, as well as multiblock copolymers could be synthesized. Taking bioRAFT polymerizations a step further, the same group showed that polymers synthesized by HRP-initiated RAFT polymerization could be modified via thiol-ene and Diels-Alder reactions with the same biocatalyst (55). Thus, the work beautifully exploits the promiscuity of the peroxidase. In parallel to An´s work, Konkolewicz and coworkers also initiated RAFT polymerization by HRP using H2O2 and acetylacetonate (56). Well-defined macromolecules, including some with more complex architectures such as block copolymers and protein-polymer conjugates were obtained. Boyer´s group used the photoredox properties of chlorophyll to initiate RAFT polymerizations by light. A wide range of monomers were polymerized with excellent control (57).

389

Conclusions BioATRP has considerably evolved since its discovery 6 years ago. A variety of enzymes have been identified to be “ATRPases”, i.e. to catalyze radical polymerizations under typical ATRP conditions. Examples are HRP, hemoglobin or laccase; enzymes that are well known as promiscuous biocatalysts. Many factors such as pH, influence of reducing agents, addition of salts, etc. could be better understood. In principal, enzymes are a “green”, i.e. environmentally friendly and non-toxic alternative to conventional ATRP catalysts. It has to be stated that the degree of control is often lower than with well-established conventional ATRP catalysts, also because the enzymes have not yet been specifically engineered for polymerizations. Nevertheless, recent reports show that ATRPases can outperform conventional catalysts, e.g. in polymerizing challenging monomers such as N-vinylimidazole, or in surface-initiated polymerizations. Moreover, enzymes can be easily removed completely from the resulting polymers, opening routes to polymers that are free of metal ion residues. Enzyme mimetic catalysts such as PEGylated hemin or deuterohemin-peptide conjugates have turned out to be excellent catalysts for ATRP. Moreover, the addition of RAFT agents to enzymatically initiated radical polymerizations could broaden the range of monomers and increase control over the polymerizations. The first applications of polymers synthesized by bioATRP are emerging, which include molecularly imprinted polymers on gold electrodes for biosensing, amphiphilic block copolymers that can act as DNA delivery system, or nanosystems such as polymer-filled polymersomes. We hope that this book chapter inspires to further explore biocatalytic controlled radical polymerizations, and we are convinced that biocatalysts offer plenty of unexplored opportunities in polymer synthesis, in the creation of nanosystems, or for applications where conventional catalysts cause problems.

Acknowledgments This work was supported by the Swiss National Science Foundation through the projects PP00P2_144697 and PP00P2_172927.

References 1. 2. 3. 4. 5.

Matyjaszewski, K.; Tsarevsky, N. V. J. Am. Chem. Soc. 2014, 136, 6513–6533. Georges, M. K.; Veregin, R. P. N.; Kazmaier, P. M.; Hamer, G. K. Macromolecules 1993, 26, 2987–2988. Wang, J.-S.; Matyjaszewski, K. J. Am. Chem. Soc. 1995, 117, 5614–5615. Kato, M.; Kamigaito, M.; Sawamoto, M.; Higashimura, T. Macromolecules 1995, 28, 1721–1723. Chiefari, J.; Chong, Y. K.; Ercole, F.; Krstina, J.; Jeffery, J.; Le, T. P. T.; Mayadunne, R. T. A.; Meijs, G. F.; Moad, C. L.; Moad, G.; Rizzardo, E.; Thang, S. H. Macromolecules 1998, 31, 5559–5562. 390

6.

7. 8.

9. 10. 11. 12. 13. 14. 15. 16. 17.

18. 19. 20. 21. 22.

23. 24. 25. 26. 27. 28. 29. 30. 31.

Klumperman, B. In Encyclopedia of Polymer Science and Technology; John Wiley & Sons, Inc.: Hoboken, NJ, 2015; doi: 10.1002/0471440264.pst453.pub2. Matyjaszewski, K. Macromolecules 2012, 45, 4015–4039. Matyjaszewski, K.; Jakubowski, W.; Min, K.; Tang, W.; Huang, J.; Braunecker, W. A.; Tsarevsky, N. V. Proc. Natl. Acad. Sci. U S.A. 2006, 103, 15309–15314. Mueller, L.; Matyjaszewski, K. Macromol. React. Eng. 2010, 4, 180–185. Tsarevsky, N. V.; Matyjaszewski, K. Chem. Rev. (Washington, DC, U.S.) 2007, 107, 2270–2299. Ni, Y.; Holtmann, D.; Hollmann, F. ChemCatChem 2014, 6, 930–943. Reetz, M. T. J. Am. Chem. Soc. 2013, 135, 12480–12496. Bornscheuer, U. T.; Huisman, G. W.; Kazlauskas, R. J.; Lutz, S.; Moore, J. C.; Robins, K. Nature 2012, 485, 185–194. Kadokawa, J.-i.; Kobayashi, S. Curr. Opin. Chem. Biol. 2010, 14, 145–153. Shoda, S.-i.; Uyama, H.; Kadokawa, J.-i.; Kimura, S.; Kobayashi, S. Chem. Rev. (Washington, DC, U.S.) 2016, 116, 2307–2413. Albertsson, A. C.; Srivastava, R. K. Adv. Drug Delivery Rev. 2008, 60, 1077–1093. Uyama, H.; Kobayashi, S. In Enzyme-Catalyzed Synthesis of Polymers; Kobayashi, S., Ritter, H., Kaplan, D., Eds.; Springer-Verlag Berlin: Berlin, 2006; Vol. 194, pp 133−158. Hollmann, F.; Arends, I. W. C. E. Polymers 2012, 4, 759–793. Zavada, S.; Battsengel, T.; Scott, T. Int. J. Mol. Sci. 2016, 17, 195. Sigg, S. J.; Seidi, F.; Renggli, K.; Silva, T. B.; Kali, G.; Bruns, N. Macromol. Rapid Commun. 2011, 32, 1710–1715. Kali, G.; Silva, T. B.; Sigg, S. J.; Seidi, F.; Renggli, K.; Bruns, N. ACS Symp. Ser. 2012, 1100, 171–181. Silva, T. B.; Spulber, M.; Kocik, M. K.; Seidi, F.; Charan, H.; Rother, M.; Sigg, S. J.; Renggli, K.; Kali, G.; Bruns, N. Biomacromolecules 2013, 14, 2703–2712. Brunori, M. Trends Biochem. Sci. 1999, 24, 158–161. Ng, Y.-H.; di Lena, F.; Chai, C. L. L. Polym. Chem. 2011, 2, 589–594. Ng, Y.-H.; di Lena, F.; Chai, C. L. L. Chem. Commun. (Cambridge, U K.) 2011, 47, 6464–6466. Simakova, A.; Mackenzie, M.; Averick, S. E.; Park, S.; Matyjaszewski, K. Angew. Chem., Int. Ed. 2013, 52, 12148–12151. Kreuzer, L. P.; Männel, M. J.; Schubert, J.; Höller, R. P. M.; Chanana, M. ACS Omega 2017, 2, 7305–7312. Sun, Y.; Du, H.; Lan, Y.; Wang, W.; Liang, Y.; Feng, C.; Yang, M. Biosens. Bioelectron. 2016, 77, 894–900. Sun, Y.; Zhang, J.; Li, J.; Zhao, M.; Liu, Y. RSC Adv. 2017, 7, 28461–28468. Gao, G.; Karaaslan, M. A.; Kadla, J. F.; Ko, F. Green Chem. 2014, 16, 3890–3898. Divandari, M.; Pollard, J.; Dehghani, E.; Bruns, N.; Benetti, E. M. Biomacromolecules 2017, 18, 4261–4270. 391

32. Renggli, K.; Sauter, N.; Rother, M.; Nussbaumer, M. G.; Urbani, R.; Pfohl, T.; Bruns, N. Polym. Chem. 2017, 8, 2133–2136. 33. Ditzel, L.; Löwe, J.; Stock, D.; Stetter, K.-O.; Huber, H.; Huber, R.; Steinbacher, S. Cell 1998, 93, 125–138. 34. Bruns, N.; Pustelny, K.; Bergeron, L. M.; Whitehead, T. A.; Clark, D. S. Angew. Chem., Int. Ed. 2009, 48, 5666–5669. 35. Renggli, K.; Nussbaumer, M. G.; Urbani, R.; Pfohl, T.; Bruns, N. Angew. Chem., Int. Ed. 2014, 53, 1443–1447. 36. Rother, M.; Nussbaumer, M. G.; Renggli, K.; Bruns, N. Chem. Soc. Rev. 2016, 45, 6213–6249. 37. Renggli, K.; Baumann, P.; Langowska, K.; Onaca, O.; Bruns, N.; Meier, W. Adv. Funct. Mater. 2011, 21, 1241–1259. 38. Palivan, C. G.; Goers, R.; Najer, A.; Zhang, X.; Car, A.; Meier, W. Chem. Soc. Rev. 2016, 45, 377–411. 39. Gaitzsch, J.; Huang, X.; Voit, B. Chem. Rev. (Washington, DC, U.S.) 2016, 116, 1053–1093. 40. Dinu, M. V.; Spulber, M.; Renggli, K.; Wu, D.; Monnier, C. A.; Petri‐Fink, A.; Bruns, N. Macromol. Rapid Commun. 2015, 36, 507–514. 41. Fodor, C.; Gajewska, B.; Rifaie-Graham, O.; Apebende, E. A.; Pollard, J.; Bruns, N. Polym. Chem. 2016, 7, 6617–6625. 42. Green, M. D.; Allen, M. H., Jr; Dennis, J. M.; Cruz, D. S.-d. l.; Gao, R.; Winey, K. I.; Long, T. E. Eur. Polym. J. 2011, 47, 486–496. 43. Allen, M. H.; Hemp, S. T.; Smith, A. E.; Long, T. E. Macromolecules 2012, 45, 3669–3676. 44. Yamashita, K.; Yamamoto, K.; Kadokawa, J.-i. Polymer 2013, 54, 1775–1778. 45. Fu, L.; Simakova, A.; Fantin, M.; Wang, Y.; Matyjaszewski, K. ACS Macro Lett. 2018, 7, 26–30. 46. Sanchez-Sanchez, A.; Arbe, A.; Kohlbrecher, J.; Colmenero, J.; Pomposo, J. A. Macromol. Rapid Commun. 2015, 36, 1592–1597. 47. Smolne, S.; Buback, M.; Demeshko, S.; Matyjaszewski, K.; Meyer, F.; Schroeder, H.; Simakova, A. Macromolecules 2016, 49, 8088–8097. 48. Zhou, H.; Jiang, W.; An, N.; Zhang, Q.; Xiang, S.; Wang, L.; Tang, J. RSC Adv. 2015, 5, 42728–42735. 49. Zhou, H.; Wang, X.; Tang, J.; Yang, Y.-W. Polymers 2016, 8, 277. 50. Jiang, W.; Wang, X.; Chen, J.; Liu, Y.; Han, H.; Ding, Y.; Li, Q.; Tang, J. ACS Appl. Mater. Interfaces 2017, 9, 26948–26957. 51. Wang, X.; Yun, W.; Jiang, W.; Wang, D.; Zhang, L.; Tang, J. RSC Adv. 2017, 7, 9926–9932. 52. Zhang, B.; Wang, X.; Zhu, A.; Ma, K.; Lv, Y.; Wang, X.; An, Z. Macromolecules 2015, 48, 7792–7802. 53. Lv, Y.; Liu, Z.; Zhu, A.; An, Z. J. Polym. Sci., Part A: Polym. Chem. 2017, 55, 164–174. 54. Liu, Z.; Lv, Y.; An, Z. Angew. Chem., Int. Ed. 2017, 56, 13852–13856. 55. Liu, Z.; Lv, Y.; Zhu, A.; An, Z. ACS Macro Lett. 2018, 7, 1–6.

392

56. Danielson, A. P.; Van Kuren, D. B.; Lucius, M. E.; Makaroff, K.; Williams, C.; Page, R. C.; Berberich, J. A.; Konkolewicz, D. Macromol. Rapid Commun. 2016, 37, 362–367. 57. Shanmugam, S.; Xu, J.; Boyer, C. Chem. Sci. 2015, 6, 1341–1349.

393

Editors’ Biographies Kris Matyjaszewski Kris Matyjaszewski is the J.C. Warner University Professor of Natural Sciences at Carnegie Mellon University. He discovered Cu-mediated atom transfer radical polymerization in 1995. He has co-authored >1000 publications (>90,000 citations) and holds 56 US patents. Matyjaszewski has received the 2017 Franklin Medal in chemistry (with Prof. Sawamoto), 2015 Dreyfus Prize in Chemical Sciences, 2011 Wolf Prize in Chemistry, 2009 Presidential Green Chemistry Challenge Award, 2015 Overberger Prize, 2013 AkzoNobel North America Science Award, 2011 Applied Polymer Science Award, and the 2002 Polymer Chemistry Award. He is a member of the National Academy of Engineering, National Academy of Inventors, and the Polish and Russian Academies of Sciences.

Haifeng Gao Haifeng Gao is an Associate Professor at the Department of Chemistry and Biochemistry, University of Notre Dame. He earned his Ph.D. in 2008 in Chemistry under Prof. Matyjaszewski at Carnegie Mellon University. After two-year postdoc training at University of California Berkeley, he joined the University of Notre Dame in 2011 as an assistant professor and was promoted to associate professor with tenure in 2017. His research focuses on the design and synthesis of functional polymers with controlled nanostructures by determining their fundamental structure-property relationships. He has co-authored over 80 peer-reviewed papers and 6 book chapters cited more than 4000 times.

Brent Sumerlin Brent Sumerlin is the George Bergen Butler Professor of Polymer Chemistry at the University of Florida (UF). He obtained his Ph.D. in polymer science and engineering from the University of Southern Mississippi under Prof. McCormick. After postdoctoral training with Prof. Matyjaszewski at Carnegie Mellon University, he joined the faculty of Southern Methodist University in 2005 before moving to UF in 2012. He is a Fellow of the Royal Society of Chemistry and has won a number of awards, including the Alfred P. Sloan Research Fellowship, NSF CAREER Award, the Journal of Polymer Science Innovation Award, Biomacromolecules/Macromolecules Young Investigator Award, and the Hanwha-Total IUPAC Award. © 2018 American Chemical Society

Nick Tsarevsky Nick Tsarevsky obtained a Ph.D. in chemistry in 2005 from Carnegie Mellon University under Prof. Matyjaszewski. He was a visiting assistant professor at the Department of Chemistry, Carnegie Mellon University (2005-2006) and CSO of ATRP Solutions, Inc. (2007-2010). He joined the Department of Chemistry at Southern Methodist University in 2010 as an assistant professor and was promoted in 2016 to tenured associate professor. He has authored and coauthored over 90 peer-reviewed journal articles and book chapters, 1 textbook, and has co-edited 5 books. Research interests include polymerization techniques, functional materials, coordination chemistry, catalysis, hypervalent iodine compounds, and history of chemistry and chemical education.

396

Indexes

Author Index Abreu, C., 227 Allushi, A., 263 Asandei, A., 205 Aydogan, C., 263 Aykac, S., 263 Bannerman, W., 205 Bottle, S., 105 Boyer, C., 77, 273 Bruns, N., 379 Chen, Y., 335 Clément, J., 105 Coelho, J., 227 Coote, M., 41 De Bon, F., 161 Debuigne, A., 349 Destarac, M., 291 Dréan, M., 349 Fairfull-Smith, K., 105 Fantin, M., 161 Fonseca, A., 227 Fujiki, Y., 323 Gao, H., ix Gennaro, A., 161 Gigmes, D., 105 Goto, A., 365 Guégan, P., 349 Guillaneuf, Y., 105 Guthrie, J., 227 Haven, J., 77 Hendrikx, M., 77 Hill, N., 41 Isse, A., 161 Jérôme, C., 349 Johnson, M., 205 Junkers, T., 77 Kajiwara, A., 63 Kamigaito, M., 323 Kim, J., 205 Kutahya, C., 263 Leenaers, P., 77 Li, F., 365

Li, J., 335 Lorandi, F., 161 Lu, H., 335 Matioszek, D., 291 Matyjaszewski, K., ix, 1, 135, 273 McKenzie, T., 307 McLeod, D., 191 Midoux, P., 349 Moad, G., 77 Morris, J., 105 Noble, B., 41 Nothling, M., 307 Peng, C., 335 Poli, R., 135 Pollard, J., 379 Postma, A., 77 Qiao, G., 307 Rahaman, S., 135 Reyhani, A., 307 Ribelli, T., 135 Rieger, J., 349 Rocha, N., 227 Ruchmann-Sternchuss, J., 291 Satoh, K., 323 Sayala, K., 191 Serra, A., 227 Shanmugam, S., 1, 273 Stiernet, P., 349 Sumerlin, B., ix Tsarevsky, N., ix, 191 Tsompanoglou, T., 77 Uchiyama, M., 323 Vasu, V., 205 Vila, X., 291 Xiao, L., 365 Xu, J., 77 Yagci, Y., 263 Yilmaz, G., 263 Yu, H., 205 Zard, S., 291

399

Subject Index A Alkyl bromide as precursor experimental, 367 poly(methyl methacrylate)-Iodide (PMMA-I), preparation, 368 alkyl bromides and monomers used in this work, structures, 367f organocatalyzed LRP, reversible activation, 366s R-Br to R-I and reaction of .Br/TBA, transformation, 367s results and discussion, 369 EPh-Br (80 mM), TBA (80 mM), and I2 (40 mM), 1H NMR spectra, 369f GPC chromatograms, 376f higher molecular weight, 372 ln([M]0/[M]) vs t, plots, 371f MMA/EPh-Br/TBA/I2/(AIBN) systems, plots, 374f MMA (8 M) with EPh-Br, TBA, and I2, bulk polymerizations, 370t PMMA-I, block polymerizations, 375t St and functional methacrylates, bulk polymerizations, 373t Atom transfer radical polymerization ATRP activation, mechanism, 163 free energies, comparison, 163f rate constants, k, of RX activation, comparison between, 164f ATRP catalysts, cyclic voltammetry, 164 activation rate constants, various catalyst/initiator systems, 170t Br– association, kinetic and equilibrium constants, 167t Cu complexes, determination of stability and halidophilicity constants, 166 CuIIMe6TREN2+, backgroundsubtracted experimental and simulated voltammograms, 178f cyclic voltammetry under total catalysis conditions, activation rate constants, 178t cyclic voltammetry with digital simulation, activation rate constants, 179t homogeneous redox catalysis, reaction sequence, 174s kact, electrochemical determination, 168

1 mM CuIIMe6TREN2+, cyclic voltammetry, 165f 10–3 M [CuIITPMA]2+, background-subtracted voltammograms, 175f 10–3 M CuIITPMA2+, cyclic voltammetry, 177f oxidation, voltammograms, 169f various catalyst/initiator systems, activation rate constants, 176t general ATRP mechanism, 162s KATRP, electrochemical determination, 181 ATRP equilibrium constants, 182t CuIL+ with 2-methyl bromopropionate, reaction, 183f kdeact, electrochemical determination, 180 CuI catalyzed radical termination (CRT), mechanism, 180s cyclic voltammetry with digital simulation in DMSO, deactivation rate constants measured, 180t kdisp, electrochemical determination, 183 organo-copper complexes, electrochemical investigation, 185 1.0 mM BrCuIITPMA+ + RX in DMSO, cyclic voltammetry, 186f OMRP equilibrium, reactions and parameters required, 187s SARA-ATRP or SET-LRP?, interplay between activation and disproportionation, 184 SET-LRP versus SARA-ATRP mechanism, 184s Atom transfer radical polymerization, photoinduced metal free strategies, 263 chain extension and block copolymerization, photoinduced metal free ATRP, 266s DMSO-d6 solutions of anthracene/ ethyl-2- bromopropionate (EBP), 1H NMR spectra, 265f GPC traces of precursor PMMA, comparison, 267f metal free photoATRP, synthesis of hyperbranched polymers, 270s monomer conversion vs. time using ITX, 269f

401

photoactive compounds, general representation of photoinduced ATRP mediated, 264s photoinduced metal free ATRP, proposed mechanism, 268s photoinitiated living radical polymerization, mechanism, 267s PMMA-b-PCL through concurrent metal free controlled/living polymerizations, synthesis, 269s

B Benzophenone-based photosensitive alkoxyamines, 105 alkoxyamine synthesis, 109 benzophenone-based photosensitive alkoxyamines, synthetic approach employed, 110s discussion, 108 benzophenone-based alkoxyamines employing electron donating substituents, 109f cyclic nitroxides, 108f EPR photo-dissociation analyses, general procedure, 121 5-benzoyl-2-methoxy-1,1,3,3tetramethylisoindoline (14), 123 (1-((5-benzoyl-1,1,3,3tetramethylisoindolin-2-yl)oxy)ethyl) benzene (7a), 126 5-benzoyl-1,1,3,3-tetramethylisoindolin-2-yloxyl (17), 124 5-(4-bromobenzoyl)-1,1,3,3tetramethylisoindolin-2-yloxyl (19), 125 5-((4-bromophenyl)(hydroxy)methyl)-2-methoxy-1,1,3,3tetramethylisoindoline, 122 ethyl 2-((5-benzoyl-1,1,3,3tetramethylisoindolin-2-yl)oxy) propanoate (7c), 127 ethyl 2-((5-(4-methoxybenzoyl)1,1,3,3-tetramethyliso-indolin2yl)oxy) propanoate (8c), 129 ethyl 2-methyl-2-((5-(4methoxybenzoyl)-1,1,3,3tetramethylisoindolin-2-yl)oxy) propanoate (8b), 128 ethyl 2-methyl-2-((5-(4(pyrrolidin-1-yl)benzoyl)-1,1,3,3tetramethylisoindolin-2-yl)oxy) propanoate (9b), 130

ethyl 2-((5-(4-(pyrrolidin1-yl)benzoyl)-1,1,3,3tetramethylisoindolin-2-yl)oxy) propanoate (9c), 131 experimental, 120 benzophenone chromophore, TEMPO-, 107f photochemical investigations, 113 cyclohexane, nitroxide stabilities, 118f hydroxylamine, proposed formation, 119s hydroxylamine species, formation, 117 methacrylic and acrylic unsubstituted, nitroxide recoveries obtained, 115f oxygenated and deoxygenated conditions, nitroxide stabilities obtained, 116f oxygenated and deoxygenated conditions, pyrrolidine substituted, 116f pyrrolidine substituted, nitroxide recoveries obtained, 114f styrenic unsubstituted methoxy, nitroxide recoveries obtained, 114f photophysical investigations, 110 examined benzophenone-based systems, photophysical properties, 113t unsubstituted, methoxy substituted and pyrrolidine substituted benzophenone-based nitroxides, UV-visible absorbance spectra overlay, 111f unsubstituted benzophenone-based nitroxide, UV-visible absorbance spectra overlay, 112f Biocatalytic ATRP enzymatic ATRP, 381 ATRP, enzyme mimetics and cofactors as catalysts, 386 bioATRP, enzyme mimetic catalysts, 387f bioATRP, first reports, 383 catalyze bioATRP, enzymes, 382f hemoglobin-catalyzed surface-initiated bioATRP, tuning, 384f N-vinylimidazole (NVIm), bioATRP, 385 PNVIm, MALDI-ToF mass spectrum, 386f single-chain polymer nanoparticles, 388 RAFT polymerization, enzymatic initiation, 389

402

C Catalyzed radical termination (CRT), 135 combination or disproportionation?, 146 C-based radical disproportionation, proposed mechanism, 151s CuI/Cu0-mediated reaction, SEC profiles, 148f disproportionation, fraction, 151f generate the same Me2C°(COOMe) radical, two independent ways, 150s PMA-Br macroinitiator, 149 PMA termination products, SEC profiles, 150f polymer produced by Cu-CRT, opposite conclusions, 147 TERP macroinitiators and small models, activation and termination, 146s CRT, critical evaluation, 155 CRT discovery and initial studies, 137 BA polymerizations, kinetic plots, 139f CuI-CRT investigations, TPMA and substituted derivatives used, 139s F(Y) and G(Y), time dependence, 140f hydride or organometallic intermediate?, 141 ATRP macroinitiators, activation and termination, 142s competing OMRP-RT trapping, Gibbs energy profile, 143f cyclic voltammetry of &lsqb:X-CuII/L]+ + RX, sequence of reactions occurring, 146s ln KATRP (red squares) for MBrP in MeCN, relationship, 144f PMA-Br, GPC curves, 143f radicals, implication, 145 metal nature, effect, 152 selected atoms, Mulliken spin densities, 153f radical chains, possible reactivity pathways, 136s solvent moisture, involvement, 154 acrylate radical termination in copper-catalyzed ATRP, pathways, 155s pMA-Br, termination, 154f Cobalt(II) phenoxy-imine complexes, 335 experimental, 340 aldehyde precursor, synthesis, 360t CoII(phenoxy-imine)2 (mediator), synthesis, 283f

phenoxy-imine (ligand), synthesis, 341s catalytic transfer agent, 337f cobalt(II) phenoxy-imine complexes, 340f free radical polymerization and polymerization with chain transfer reaction, difference, 336f reversible deactivation radical polymerization, 338f reversible termination (RT), degenerative transfer (DT), and catalytic chain transfer (CCT) mechanisms, correlation, 339s results and discussion, 342 conversion versus time, plots, 343f conversion versus time for VAc polymerization, plots, 345f DP versus [Co]/[VAc], plots, 346f Mn and PDI, plots, 344f Mn and PDI versus conversion for VAc polymerization, plots, 346f VAc Mediated by CoII(Phenoxyimine)2, polymerization, 343t VAc polymerization, GPC traces, 344f

E Epoxides as reducing agents, 191 experimental procedures, 195 epoxide by bromide, ring-opening, 194s GMA using epoxides, LCC ATRP, 194s results and discussion, 197 catalyst complex, reduction, 197f EA, LCC ATRP, 200f free radical polymerization, kinetics, 201f MMA, LCC ATRP, 198f styrene, LCC ATRP, 199f

G Group 10 (Ni, Pd, Pt) metal complexes experimental, 208 results and discussion, 209 BD-ATRP, Ni(0) complexes, 215f BD-ATRP, Ni(II) complexes, 217f BD-ATRP, Ni(II)phosphine complexes, 217f BD-ATRP catalyzed by group 10 metal complexes, mechanism, 211s

403

mechanism of coheterotactic selectivity, 57s calculated kinetic, thermodynamic and polar parameters, 51t coheterotactic specificity, mechanistic origins, 52 MMA(BCl3)-Sty and Sty-MMA(BCl3) radicals, newman projections, 55f MMA(BCl3)-Sty-MMA(BCl3) and Sty-MMA(BCl3)-Sty radicals, lowest energy accessible conformations, 56f optimised geometries and binding free energies, 50f optimised geometries and relative free energies, 48f Sty and MMA(BCl3) monomer, electrostatic potential surfaces (ESPs), 55f Sty-based radicals, strength, 53f terminal-penultimate and terminal-monomer interactions for MMA(BCl3) radicals, strength, 54f three previously proposed mechanisms, 49s various unimeric MMA models, chemical structures, 48f tacticity determination, mechanism, 44s

BD-ICAR-ATRP, group 10 Mt(II) complexes, 218f BD-ICAR ATRP, NiBr2-ligand effect, 215f group 10 mediated BD-ATRP, 212t ICAR-ATRP, Pd complexes, 218f ligands and group 10 complexes tested in BD-ATRP, structure, 211s Ni(II) complexes, 216 Pd(0) and Pd(II) complexes, 219 polybutadiene, 1H-NMR, 214f total, 1,4-trans, 1,4-cis and 1,2 Br-CEF, dependence, 220f

H Hydroxyl radical activated RAFT polymerization enzyme-assisted initiation systems, 314 generate reactive hydroxyl radicals, 315f Fenton-RAFT polymerization, 309 common reactions involving hydroxyl radicals, 310f Fenton-RAFT polymerization, 312f RAFT polymerization, fenton initiation system, 311f introduction, 307 RAFT polymerization, some examples of commonly used initiation systems, 309f photocatalyst-based polymerization, 313 PET-RAFT, proposed mechanism, 314f sonochemical initiation systems, 316 polymer growth, GPC chromatograms, 318f water results, sonolysis, 317f

L Lewis acid mediated sequence- and stereo-control, 41 computational methodology, 46 deuterated analogue Sty-MMA copolymerization, 45s idealized modifications, 42s radical polymerization, stereospecificity, 43 results and discussion, 369 BCl3-mediated MMA/Sty copolymerization, simplified

M Methacrylates, RAFT polymerization experimental, 293 results and discussion, 296 emulsion RAFT block copolymerization, reaction conditions, 303t emulsion RAFT polymerization of MMA, reaction conditions, 301t MMA, chain transfer constants to xanthates 1-7, 297t MMA conversion during introduction, evolution, 300f Mn MALS and dispersity, evolution, 302f PMMA latexes synthesized, SEC/RI distributions, 300f PMMA latex synthesis, reaction conditions, 298 PMMA-7 seed and PMMA-PVAc block copolymer, SEC/RI chromatograms, 302f

404

study, xanthates, 299s

R RAFT navigation, elements, 77 acrylamides in aqueous media, RAFT-SUMI, 84 acrylamides, polymerization, 88 kinetic parameters used, 90t overlaid GPC traces, 91f RAFT-SUMI for sequential insertion, scope, 87f in situ NMR, evolution of species observed, 86f species observed by in situ NMR, evolution, 89f tertiary cyanoalkyl trithiocarbonates, 85 experimental, 96 DMAm RAFT, automated sequential insertion, 98 thermally-initiated sequential RAFT SUMI, 97 RAFT equilibria, schematic depiction, 78f PET-RAFT SUMI, 91 catalyst-free photoinitiated single unit monomer, 95t photoRAFT-SUMI with a trithiocarbonate RAFT agent, mechanism, 95f producing discrete oligomers, strategy, 94f producing discrete trimers, strategy, 92f products from sequential PET-RAFT-SUMI into CDTPA, ESI-MS spectra, 93f sequential PET-RAFT-SUMI, GPC traces, 93f RAFT single unit monomer insertion, 79 poly(3-hexylthiophene macroRAFT agent, synthesis, 81f RAFT-SUMI for insertion of a monomer, scope, 81f RAFT-SUMI indicating the terminology, schematic depiction, 80f thermally-initiated sequential RAFT SUMI, 82 NPMI, RAFT-SUMI for insertion, 83f RAFT-SUMI for sequential insertion, scope, 83f

styrene polymerizations, molar mass conversion data, 84t Reversible addition fragmentation chain transfer (RAFT) polymerization electrochemical means, external regulation of RAFT polymerization, 284 electrochemical RAFT (eRAFT), 285f RAFT polymerization via photochemical means, 274 discrete pentamer via SUMI, synthesis, 283f higher wavelength polymerizations, photocatalysts, 276 manipulation of wavelength of lights, switching polymerization mechanism and monomer selectivity, 279f photochemical means, chemoselective RAFT activation and sequence control, 281 photoinduced electron/energy transfer-reversible addition-fragmentation chain transfer (PET-RAFT), 275s polymer gels, living additive manufacturing achieved via photo-redox growth, 279f polymerizations, catalyst-free light mediated systems, 280 polymerizations, metal-free photocatalyst systems, 277 ZnTPP and dithiobenzoate by PheoA, specific activation of trithiocarbonates, 282f sonochemistry, external regulation of RAFT polymerization, 285 Reversible deactivation radical polymerization ATRP, advancements, 2 acrylate radical termination, proposed mechanistic pathways, 4s SARA-ATRP, progress, 5 bioconjugation, advances, 20 bottlebrush polymers, generation, 22s chemical permeation enhancer 1-phenylpiperazine, reducing toxicity, 20s glycopolymer conjugates, 23 grafting polymer from living cells, 24 grafting polymer from protein surfaces, novel strategy, 21s living cell surface grafting, 25s complex architecture, advancements in polymers, 12 sequence defined oligomers, 14

405

solvent selective iodine mediated polymerization, 15s organic-inorganic hybrid materials, advances, 16 nitrogen-doped porous carbons, scalable synthesis, 18 novel nitrogen-doped porous carbons, synthesis, 19s polymers through macromolecular metamorphosis, altering the topology, 17s universal tetherable initiator structurally analogous, 17s other advancements, 244 RDRP, photochemical, electrochemical, and mechanochemical regulation, 6 copper mediated polymerization, proposed mechanism, 7s electrochemical and mechanochemical ATRP, 9 MAA in water with chlorine initiator, promoting fast polymerization, 10s metal-free ATRP (o-ATRP), 8 piezoelectric barium titanate, sonochemical ATRP mediated, 11s

size exclusion chromatography (SEC), 74 TR ESR spectroscopy, 73 chain initiating radicals, TR ESR spectra, 65f compounds used, sturctures, 66s ESR techniques with different time resolution, schematic diagram, 64f RAFT polymerization systems, TR ESR, 69 adduct radicals, structures, 70s solution of BTBA with TMDPO, TR ESR spectra observed, 71f TR ESR spectra observed under different conditions, 72f results and discussion aqueous phase, TR ESR, 67 chain initiating radical of sodium methacrylate initiated with LiTMPO, TR ESR spectrum, 69f sodium acrylate initiated, TR ESR spectrum of chain initiating radical, 68f

V T Tailor-made poly(vinylamine)s, 349 experimental section, 351 hydrolysis procedure, 353 NMVA, thermal OMRP, 352 preparation of poly(vinylamine)s, general strategy, 351s results and discussion, 354 Mn (full symbols), evolution, 355f Mn MALLS (full symbols), dependence, 359f NVA, photochemically initiated organometallic mediated polymerization, 358s NVA and NMVA, thermally initiated organometallic-mediated polymerization, 355s OMRP NVA in DMF, thermal initiation, 357t poly(N-vinylamine)s, synthesis, 360t PVAm (A and B) and PMVAm (C and D) as pDNA carriers, use, 361f SEC curves in DMF, overlay, 356f Time-resolved electron spin resonance observations experimental

Vinyl chloride, reversible deactivation radical polymerization VC polymerization, RDRP methods, 232 CMPCD RAFT agent, 1H NMR spectra, 247f first block polymers, PVC-b-P2EHA-b-PVC and PVC-b-PBA-b-PVC, 237 NMP of VC initiated, general scheme and conditions, 249f PVAc-CTA macroCTA, GPC traces, 247f PVC-based block copolymer structures, diversity, 238f PVC-b-PBA-b-PVC block copolymers, stress-strain curves, 238f PVC-b-PVAc block copolymers, synthesis, 248f PVC-Br macroinitiator, GPC traces, 245f PVC resin prepared by SET-DTLRP, SEM micrograph, 243f PVC resins, SEM micrograph, 242f PVC-SG1 macroinitiator, SEC chromatograms, 250f

406

PVCs prepared by FRP and SET-DTLRP, thermal stabilities obtained, 240t RDRP techniques, structures of PVC-based copolymers prepared, 253t SET-DTLRP, diversity of PVC structures obtained, 236f SET-DTLRP, particle morphology of PVC prepared, 241 SET-DTLRP of VC, industrial implementation, 239 SET-DTLRP of VC in water, general mechanism proposed, 235f SET-DTLRP prepared in different reaction, PVC obtained, 240t SET-LRP initiated with CHBr3, PVC structure obtained, 244f VC, SET-DTLRP, 234 VC polymerization, milestones in RDRP methods, 233f VC polymerization, summary of the RDRP methods proposed, 251t VC using CMPCD as RAFT agent, RAFT polymerization, 246f vinyl chloride, polymerization, 228 PVC, allylic chloride and tertiary chloride structural defects, 231f

1835 to 2000, milestones in PVC history, 229f uncontrolled molecular weight, 230 zipper dehydrochlorination reaction, 231f Vinyl ether/vinyl ester copolymerization results and discussion, 325 CEVE and VPv, 1H NMR spectra, 329f CEVE and VPv, interconvertible simultaneous copolymerization, 330f CEVE/VPv copolymer obtained by interconvertible copolymerization, 331f IBVE/VAc copolymers, 1H NMR spectra, 327f interconvertible simultaneous copolymerization, CEVE and VPv, 329f interconvertible simultaneous copolymerization, IBVE and VAc, 326f other vinyl ether and vinyl ester, copolymerization, 328 RAFT radial copolymerization, 326f

407