Stereochemistry and Global Connectivity: The Legacy of Ernest L. Eliel [2] 0841232407, 9780841232402

Citation preview

Stereochemistry and Global Connectivity: The Legacy of Ernest L. Eliel Volume 2

ACS SYMPOSIUM SERIES 1258

Stereochemistry and Global Connectivity: The Legacy of Ernest L. Eliel Volume 2 H. N. Cheng, Editor U.S. Department of Agriculture, New Orleans, Lousiana

Cynthia A. Maryanoff, Editor Baruch S. Blumberg Institute, Doylestown, Pennsylvania

Bradley D. Miller, Editor ACS International Activities, Washington, DC

Diane Grob Schmidt, Editor University of Cincinnati, Cincinnati, Ohio

Sponsored by the ACS Division of Organic Chemistry

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

Library of Congress Cataloging-in-Publication Data Names: Cheng, H. N., editor. | American Chemical Society. Division of Organic Chemistry. Title: Stereochemistry and global connectivity : the legacy of Ernest L. Eliel Volume 2 / H.N. Cheng, editor (U.S. Department of Agriculture, New Orleans, Lousiana) [and three others] ; sponsored by the ACS Division of Organic Chemistry. Description: Washington, DC : American Chemical Society, [2017]- | Series: ACS symposium series ; 1258 | Includes bibliographical references and index. Identifiers: LCCN 2017045040 (print) | LCCN 2017050615 (ebook) | ISBN 9780841232372 (ebook) | ISBN 9780841232389 (v. 1) | ISBN 9780841232402 (v. 2) Subjects: LCSH: Stereochemistry. | Communication in science. | Conformational analysis. | Eliel, Ernest L. (Ernest Ludwig), 1921-2008. Classification: LCC QD481 (ebook) | LCC QD481 .S7574 2017 (print) | DDC 547/.1223--dc23 LC record available at https://lccn.loc.gov/2017045040

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 © 2017 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 Foreword .......................................................................................................................... ix In Memory of Ernest Eliel .............................................................................................. xi Preface ............................................................................................................................ xiii

Stereochemistry of Organic Compounds 1.

Theoretical Evidence for the Relevance of n(S) → σ*(C-P), σ(C-S) → σ*(C-P), and n(F) → σ*(C-X) (X = H, C, O, S) Stereoelectronic Interactions .... 3 Eusebio Juaristi and Rafael Notario

2.

The Importance of Electrostatic Interactions on the Conformational Behavior of Substituted 1,3-Dioxanes: The Case of 5-Phenyl-1,3-dioxane ...... 19 William F. Bailey and Kyle M. Lambert

3.

Asymmetric Autocatalysis and the Origin of Homochirality ............................. 27 Kenso Soai and Arimasa Matsumoto

4.

Interplay between Organocatalysis and Multicomponent Reactions in Stereoselective Synthesis ........................................................................................ 49 Daniel G. Rivera and Márcio W. Paixão

5.

Synthetic Approaches to the Stereochemically Complex Antitumor Drug Ecteinascidin-743: A Marine Natural Product by the Name Yondelis® or Trabectidin .............................................................................................................. 61 Plato A. Magriotis

6.

Condensation Reaction between Phenanthroline-5,6-diones and Ethylenediamine and Its Optimization through Dialogue between Theory and Experiment ...................................................................................................... 79 Luis Sanhueza, Diego Cortés, Iván González, and Bárbara Loeb

7.

Solvent Effects on Electronic Circular Dichroism Spectra ................................ 91 Aguinaldo Robinson de Souza, Valdecir Farias Ximenes, and Nelson Henrique Morgon

vii

NMR Applications 8.

Ernest L. Eliel, a Physical Organic Chemist with the Right Tool for the Job: Rapid Injection Nuclear Magnetic Resonance .................................................. 105 Andy A. Thomas and Scott E. Denmark

9.

Characterization of Materials with NMR Spectroscopy .................................. 135 Cecil Dybowski

10. From NMR Spectra to Molecular Structures and Conformation ................... 161 Alan E. Tonelli 11. Solution NMR Structure and Conformation of Silk Fibroins Stored in Bombyx mori and Samia cynthia ricini Silkworms ............................................ 191 Tetsuo Asakura, Yu Suzuki, and Akio Nishimura 12. Novel Polymeric Products Derived from Biodiesel ........................................... 207 Atanu Biswas, Zengshe Liu, Roselayne Furtado, Carlucio R. Alves, and H. N. Cheng Editors’ Biographies .................................................................................................... 221

Indexes Author Index ................................................................................................................ 225 Subject Index ................................................................................................................ 227

viii

Foreword A major strength of American Chemical Society (ACS) is the large number of volunteers who help to grow and sustain the organization, from local sections to technical divisions, from regional to national meetings, from task forces to national committees, and from conducting research to writing and reviewing manuscripts for journals. Some of them spend literally thousands of hours on behalf of ACS and the global chemistry enterprise, helping students or fellow scientists, organizing meetings and symposia, reaching out to the local communities, publicizing the benefits of science, and connecting with global colleagues and partners. One of the people who excelled in these efforts was the late Prof. Ernest L. Eliel. For many years he taught at the University of Notre Dame and the University of North Carolina and was an acknowledged leader in organic stereochemistry and conformational analysis. He was also a leader at ACS, serving as ACS President in 1992 and Chair of ACS Board of Directors in 1987-89. He initiated several international activities, particularly with respect to Latin America. I had worked with him on many occasions, and he was always courteous, judicious, analytical, and thoughtful. Unfortunately Prof. Eliel left us in 2008. As ACS Immediate Past President in 2016, it was my pleasure to be associated with a symposium organized by Dr. Cynthia Maryanoff and me at the ACS National Meeting in Philadelphia in honor of Prof. Eliel. Because of the success of the symposium, Dr. Maryanoff and I decided to team up with Dr. H. N. Cheng and Dr. Bradley Miller to edit a book highlighting stereochemistry and global connectivity, which represented two of the key legacies of Prof. Eliel. We were very pleased that so many of our academic colleagues were willing to participate in this project in honor of Prof. Eliel. Thanks are due to all of the authors and my co-editors for their wonderful efforts. Because stereochemistry is a fundamental chemistry concept, ongoing research is carried out in different subfields of chemistry (such as organic, medicinal, carbohydrates, polymers), using various analytical techniques (such as NMR, X-ray crystallography, and circular dichroism). The two volumes of this book contain many research papers that represent cutting-edge research in all the above areas. Because chemistry is now a world-wide enterprise, global connectivity is important to chemistry practitioners, and the chapters on international activities should be of great interest as well. Hopefully the readers will find these two book volumes useful to them in their research or professional work, and the example of Prof. Eliel will motivate them to volunteer their time and talent to ACS and the global chemistry enterprise!

Diane Grob Schmidt 2015 ACS President ix

In Memory of Ernest Eliel

(Phtot Credit: Reproduced from Chem. & Eng. News, 2008, 86(14), 62-67. Peter Cutts Photography)

Ernest Ludwig Eliel was an eminent organic chemist known for his seminal contributions to organic stereochemistry and conformational analysis. He was born in Cologne, Germany and moved to Scotland, then Canada, then Cuba. He received his B.S. from the University of Havana in 1946. He moved to the USA in 1946 and obtained his PhD at the University of Illinois in two years with eight publications that resulted from his work. He taught at the University of Notre Dame in 1948-1972 and then moved to the University of North Carolina at Chapel Hill. He was well-known as an outstanding teacher and researcher. He died in Chapel Hill in 2008. Eliel was active in the ACS and served in numerous capacities. He was Chair of the ACS Board of Directors in 1987-89 and ACS President in 1992. His scientific contributions and professional involvements were widely recognized. Among his awards and recognition were membership of the National Academy of Sciences (NAS) (1972), the George C. Pimentel Award in Chemical Education xi

(1995), the ACS Priestley Medal (1996) and the NAS Award for Chemistry in Service to Society (1997). We dedicate this book to the memory of his contributions to chemistry and his efforts to connect scientists worldwide through its practice.

xii

Preface The two volumes of this book are partially based on a symposium held at the ACS Fall National Meeting in Philadelphia in August 2016 on Connectivity and the Global Reach of Chemistry: Honoring the Life and Scientific Contributions of Ernest L. Eliel. The symposium speakers consisted of friends and former colleagues of Prof. Eliel, and they provided reminiscences of Eliel as well as updates of their own research work in honor of Eliel. The symposium was well-received and encouraged us to edit this two-volume symposium book. In addition to the symposium speakers, we also invited several other colleagues in the global chemistry enterprise to contribute papers related to stereochemistry and global connectivity. Prof. Eliel was well-known scientifically for his work on organic stereochemistry and conformational analysis. His book, Stereochemistry of Carbon Compounds, was a classic that had a profound influence on the field. He was also active in the ACS and served as Chair of the ACS Board of Directors in 1987-89 and ACS President in 1992. As an ACS leader, he recognized the importance of global connectivity and promoted the strengthening of international activities in chemistry. He started several valuable programs that facilitated international exchange, particularly those that related to Cuba, Mexico and the rest of Latin America. In view of Prof. Eliels outstanding contributions to chemistry and its global practice, we are featuring two topics in this book that reflect his legacy: stereochemistry and global connectivity. Thus, the aim of this book is to recognize the contributions of Prof. Eliel and to highlight the latest developments in stereochemistry and global connectivity so that younger people who may not know Eliel can benefit from the information given in the chapters of this book. A total of 22 chapters are included in the two volumes of this book. Volume 1 consists of 10 chapters, including an overview (Chapter 1), two chapters on Tribute to Ernest Eliel, three chapters on Global Connectivity, and four chapters on Carbohydrates. In this volume (Volume 2), 12 chapters are gathered and grouped in two sections: 1) Stereochemistry of Organic Compounds 2) NMR Applications All 12 chapters (including those under NMR Applications) involve stereochemistry in various organic, inorganic, biochemical, and polymer systems

xiii

and are excellent examples of the importance of stereochemistry to the practice of chemistry today. This book is targeted to all chemists and chemical engineers, particularly those with an interest in stereochemistry or international relations. Since stereochemistry is an important basic concept for chemistry, biochemistry, polymer science, and other related areas, hopefully many chemistry professionals and students find the chapters useful. Moreover, the chemistry enterprise is becoming increasingly globalized. Thus, global connectivity and networking may be of interest as well. This book is also suitable as a reference work for libraries. We appreciate the efforts of the authors who took time to prepare their manuscripts and the reviewers for their cooperation during the peer review process. We also thank Arlene Furman, Tracey Glazener, Elizabeth Hernandez and their colleagues at ACS Books for their patient and effective handling of the manuscripts.

H. N. Cheng USDA − Agricultural Research Service Southern Regional Research Center New Orleans, LA 70124, USA

Cynthia A. Maryanoff Baruch S. Blumberg Institute Drug Discovery and Development Doylestown, PA 18902, USA

Bradley D. Miller ACS International Activities External Affairs & Communications Office of the Secretary and General Counsel American Chemical Society Washington, DC 20036, USA

Diane Grob Schmidt Department of Chemistry University of Cincinnati Cincinnati, OH 45221, USA

xiv

Stereochemistry of Organic Compounds

Chapter 1

Theoretical Evidence for the Relevance of n(S) → σ*(C-P), σ(C-S) → σ*(C-P), and n(F) → σ*(C-X) (X = H, C, O, S) Stereoelectronic Interactions Eusebio Juaristi*,1,2 and Rafael Notario*,3 1Departamento

de Química, Centro de Investigación y de Estudios Avanzados, Instituto Politécnico Nacional 2508, Colonia Zacatenco, 07360 Ciudad de México, México 2El Colegio Nacional, Luis González Obregón 23, Centro Histórico, 06020 Ciudad de México, México 3Instituto de Química Física “Rocasolano”, CSIC, c/ Serrano 119, 28006 Madrid, Spain *E-mail: [email protected]; [email protected]

Three decades after the discovery of a strong S-C-P anomeric effect in 2-diphenylphosphinoyl-1,3-dithiane (1), a suitable interpretation was pending; nevertheless, very recent DFT geometry optimization of 1-ax and 1-eq did reproduce the S-C-P anomeric effect in 1, worth 5.45 kcal/mol (in chloroform solvent). Furthermore, NBO computational analysis suggests the involvement of n(X) → σ*(C-P(O)Ph2) stereoelectronic interactions that stabilize the axial conformer. Along similar lines, theoretical calculations on r-1,c-3,c-5trifluorocyclohexane (2), r-2,c-4,c-6-trifluoro-1,3,5-trioxane (3) and r-2,c-4,c-6-trifluoro-1,3,5-trithiane (4) confirm the relevance of n(F) → σ*(C-X)gem hyperconjugative interactions, where X = H, C, O, S. Thus, contrary to generally accepted concepts, fluorine is a good lone pair electron donor towards geminal sigma bonds.

© 2017 American Chemical Society

Introduction Since its discovery six decades ago, the so-called ‘anomeric effect’ has turned into one of the most frequently used concepts employed to explain conformational preferences, structural properties, and even the reactivity of organic molecules (1). Nevertheless, the origin of the anomeric effect is still a matter of controversy (2), and it is thus apparent that further studies of this important effect are necessary. As demonstrated by E. L. Eliel in the 1960s and 1970s (3–5), the presence of lone electron pairs in substituted saturated heterocyclic compounds can have pronounced effects on their conformation. In this context, the interaction of electron-withdrawing anomeric substituents [electronegative groups localized at C(1)] with endocyclic lone electron pairs induces a preference by these substituents to adopt the axial instead of the equatorial orientation. This conformational effect was initially discovered by Edward (6) and Chü and Lemieux (7), and attracted much attention. This phenomenon became to be known as the anomeric effect (Scheme 1).

Scheme 1. Counter-intuitive (according to prevailing concepts in the early times of conformational analysis) preference of electronegative substituents at the anomeric position to adopt the axial orientation. Reproduced with permission from ref. (8). Copyright [2015] ACS.

In an imaginative, remarkable interpretation of this conformational effect, a stabilizing interaction between a lone electron pair on the ring heteroatom “X” and the low-energy antiperiplanar antibonding orbital of the bond connecting the axial electronegative substituent “Y” at the anomeric carbon [n(X) → σ*(C-Y)app hyperconjugation] was proposed. Because of the double bond-no bond canonical structure associated with this hyperconjugative interaction, a lengthening of the axial C-Y bond, as well as a shortening of the endocyclic C-X bond are anticipated (9, 10). Entirely by chance, while working in the very first research project undertaken in our laboratory, proton NMR spectroscopic data showed significant deshielding of the syn-axial protons at C(4,6) in 2-diphenylphosphinoyl-1,3-dithiane (1), which suggested an axial conformation of the diphenylphosphinoyl group (Scheme 2) (11). In order to quantitate this conformational effect, conformationally fixed diastereomeric models were synthesized, and their chemical equilibration under basic catalysis (ethanolic EtO−Na+) afforded ΔG° = +1.0 kcal/mol, the axial conformer being more stable than the equatorial conformer (11). 4

Scheme 2. Predominance of the axial conformation in the conformational equilibrium of 2-diphenylphosphinoyl-1,3-dithiane (1). Reproduced with permission from ref. (11). Copyright [2015] ACS.

This conclusion was very intriguing in view of the rather large size of the diphenylphosphinoyl group. Nevertheless, X-ray diffraction data from suitable crystals of 2-diphenylphosphinoyl-1,3-dithiane (1) provided the structure and conformation shown in Figure 1.

Figure 1. X-Ray diffraction structure of 1-ax, exhibiting the axial orientation of the diphenylphosphinoyl group. Reproduced with permission from ref. (11). Copyright [2015] ACS.

Surprisingly, comparison of the structural data of 1-ax (Figure 1) and its conformationally-fixed equatorial analog did not exhibit the expected (in terms of a n(S) → σ*(C-P)app hyperconjugative interaction, see above) contraction of the endocyclic S-C(2) bond and lengthening of the exocyclic C(2)-P bond in 1-ax (11). With the advent of powerful computational equipment and software in recent years, it was decided to explore whether theoretical calculations could reproduce the anomeric effect manifested experimentally in the conformational behavior of 2-diphenylphosphinoyl-1,3-dithiane (1-ax ⇌ 1-eq, Scheme 2). Furthermore, the question was posed as to whether natural bond order (NBO) calculations could provide support for stereoelectronic interactions at the origin of this conformational equilibrium. 5

Computational Methods All calculations were carried out with Gaussian 09 programs (12). The structures of interest were fully optimized at the B3LYP/6-311+G(d,p) level of theory (13). Electronic structures were examined with Natural Bond Orbital (NBO) analysis (14), and hyperconjugative interactions were evaluated by means of the NBO program (version 3.1) (15). Simulation of solvent was accomplished according to the polarizable continuum model developed by Tomasi and co-workers (16).

Results and Discussion Anomeric Effect in the S-C-P Segment The lowest energy geometries of axial and equatorial 1 at the B3LYP/6311+G(d,p) level of theory, are shown in Figure 2 and Table 1.

Figure 2. B3LYP/6-311+G(d,p)-optimized structures of 2-diphenylphosphinoyl1,3-dithiane, in the axial conformation, 1-ax, and in the equatorial conformation, 1-eq. Reproduced with permission from ref. (8). Copyright [2015] ACS.

According to the calculations summarized in Table 1 (8), the C(2)-P(O) bond in the axial isomer 1-ax (1.867 Å) is exactly similar to the C(2)-P(O) bond in equatorial 1, 1.867 Å. By the same token, the endocyclic C(2)-S bonds in 1-ax and 1-eq are essentially identical, 1.836 ± 0.002 Å in both isomers. Experimentally (11), the X-ray crystallographic data afforded 1.825 Å for axial C(2)-P and 1.840 Å for equatorial C(2)-P. On the other hand, the C(2)-S bond lengths are 1.809 Å in 1-ax and 1.809 Å in the anancomeric (conformationally fixed) equatorial model. As indicated in the Introduction, both the experimental and calculated structural data are not in line with the anticipated consequences of n(S) → σ*(C-P) stereoelectronic interaction; i. e., one would expect a substantial shortening of the S-C(2) endocyclic bond as well as a lengthening of the C(2)-P(O) bond in 1-ax relative to 1-eq.

6

Table 1. B3LYP/6-311+G(d,p)-optimized geometrical parameters of 2-diphenylphosphinoyl-1,3-dithiane, in the axial conformation, 1-ax, and in the equatorial conformation, 1-eq. Bond distances in Å, and bond angles in degrees. Bond

1-ax

1-eq

C(2)-P

1.867

1.867

C(2)-S

1.835-1.836

1.837-1.838

P=O

1.505

1.497

C(4)-S

1.841-1.842

1.837-1.838

C(4)-C(5)

1.528-1.529

1.529

S-C(2)-S

114.7

114.0

C(2)-P-O

113.0

114.8

C(2)-S-C(4)

100.8-100.9

98.0

SC(4)-C(5)

114.1-114.2

114.4-114.5

C(4)-C(5)-C(6)

113.4

113.8

In the most relevant result from the theoretical calculations (8), DFT calculations do reproduce the S-C-P anomeric effect in diphenylphosphinoyl1,3-dithiane 1; that is, the preference of the phosphorus substituent to adopt the axial orientation instead of the equatorial orientation. Indeed, as revealed by B3LYP/6-31G(d) and B3LYP/6-311+G(d,p) calculations, in ethanol solvent at 294 K the conformer with the diphenylphosphinoyl group in the axial position (1-ax in Scheme 2) is lower in energy, ΔG° +1.36 kcal/mol and +1.30 kcal/mol, respectively. These values are rather close to the experimentally observed ΔG°294 K = +0.99 kcal/mol in ethanol (11). Nevertheless, at the B3LYP/6-311+G(d,p) level of theory, the calculated ΔG°298 K = +3.80 kcal/mol in the gas phase seems too large, probably as the result of overestimated hydrogen bonding interactions between the phosphoryl oxygen and the axial hydrogens at C(4,6), which stabilize the axial isomer. As it has been shown by Alabugin (17), the NBO method developed by Weinhold (18) is a quite useful for the study of hyperconjugation. In our work, NBO analysis afforded an estimate of the magnitude of the delocalizing interaction that weakens the axial C-P bond. In particular, the energies of delocalization (Edel) were obtained by deletion of the corresponding Fock elements, followed by the recalculation of the wave function. Table 2 lists Edel for the main hyperconjugative interactions in dithianes 1-ax and 1-eq. Table 2 includes also the energy difference between the corresponding donor and acceptor orbitals, which (as anticipated) shows an inverse relationship between energy gap and the magnitude of the two-electron/two-orbital hyperconjugative interaction.

7

Table 2. Selected hyperconjugative interactions (Edel) in axial and equatorial 2-diphenylphosphinoyl-1,3-dithiane, 1-ax and 1-eq.

1-ax

1-eq

Donor orbital

Acceptor orbital

Edel (kcal/mol)

ΔEdonor/acceptor (Hartrees)

n(Sax)

σ*(C-Pax)

3.86

0.41

n(Seq)

σ*(C-Pax)

1.65

0.81

n(Seq)

σ*(C(2)-S)

5.31

0.37

n(Sax)

σ*(C(2)-S)

1.97

0.77

σ(C(4,6)-S

σ*(C-P)

---

---

n(S)

σ*(C-Peq)

---

---

n(Seq)

σ*(C(2)-S)

6.88

0.37

n(Sax)

σ*(C(2)-S)

1.29

0.77

σ(C(4,6)-S)

σ*(C-Peq)

1.87

0.76

The most salient observations are the following: 1) n(S) → σ*(C-P)app stereoelectronic interactions are observed in 1-ax but not in 1-eq. This observation is in agreement with anticipation in terms of an efficient stereoelectronic interaction in the axial conformation, where the donor and acceptor interacting orbitals are antiperiplanar to each other. Such antiperiplanar orientation of the donor/acceptor orbitals in not possible in equatorial 1. As discussed in the Introduction, this stereoelectronic interaction is responsible for the S-C-P anomeric effect, i.e., the axial predominance of the phosphorus substituents at C(2) in the 1,3-dithiane ring in 2-diphenylphosphinoyl-1,3-dithiane 1. 2) Interestingly, n(S) → σ*(C(2)-S)app stereoelectronic interactions are present both in the axial and equatorial isomers, so that this two orbitals-two electrons stabilizing interaction is equally effective in both orientations of the phosphorus group and has no consequence in the conformational free energy difference of the 2-P-substituted 1,3-dithianes 1. 3) Most interestingly, antiperiplanar σ(C(4,6)-S) → σ*(C-P)app stereoelectronic interactions are effective in equatorial 1-eq, but are not operative in 1-ax. This hyperconjugative stereoelectronic interaction 8

apparently weakens the equatorial C(2)-P bonds, which are rendered longer. This may help explain the “anomalous” structural observation discussed in the Introduction, that the C(2)-P bond distances are of the same length in the axial and equatorial isomers of 1. That is, n(S) → σ*(C-P)app stereoelectronic interactions are responsible for the longer C(2)-P axial bonds, but σ(C(4,6)-S) → σ*(C-P)app hyperconjugation gives rise to longer C(2)-P equatorial bonds. The interpretations advanced above are supported by deletion of the key hyperconjugative interactions followed by reoptimization of the geometries with those interactions switched off by means of NBODEL (15). In all cases, application of NBODEL while switching off the key hyperconjugative interactions resulted in lengthening of the C(2)-S bonds and simultaneous shortening of the antiperiplanar C(2)-P bonds, as anticipated in terms of n(S) → σ*(C-Pax)app in 1-axial. By the same token, NBODEL results were in line with σ(C(4,6)-S) → σ*(C-Peq)app and n(S) → σ*(C(2)-S)app stereoelectronic interactions operative in the equatorial analog. In summary, DFT calculations do reproduce the S-C-P anomeric effect in diphenylphosphinoyl-1,3-dithiane 1, i.e., the preference of the phosphorus substituent to adopt the axial orientation. The Natural Bond Orbitals (NBO) method developed by Weinhold and co-workers (15) turned out to be a very useful theoretical strategy for the study of the hyperconjugative interactions present in 1. In particular, NBO analysis afforded the estimated energies of the delocalizing interactions that weaken the axial C-P bonds of interest. Specifically, n(S) → σ*(C-P)app stereoelectronic interactions are observed in 1-ax but not in 1-eq, as anticipated in terms of an efficient hyperconjugative interaction in the conformation where the donor and acceptor interacting orbitals are antiperiplanar to each other (Scheme 3a). This stereoelectronic interaction gives rise to the S-C-P anomeric effect, that is manifested as the axial predominance of the diphenylphosphinoyl substituent in 1. On the other hand, antiperiplanar σ(C(4,6)-S) → σ*(C-P)app stereoelectronic interactions are only effective in equatorial 1-eq (Scheme 3b). The combination of the two stereoelectronic effects helps explain the “anomalous” structural observation that the C(2)-P bond distances are quite similar in length in the axial and equatorial isomers of 1.

n(F) → σ*(C-X) (X = H, C, O, S) Stereoelectronic Interactions In contrast to orbitals associated with lone pairs of electrons at the electronegative elements nitrogen and oxygen, which turn out to be rather good donor orbitals, the orbitals associated with lone electron pairs on fluorine usually exhibit poor electron donating power (1). Interestingly, O’Hagan and coworkers recently provided structural and theoretical data supporting the existence of significant n(F) → σ*(C-X)gem stereoelectronic interactions in the geminal F-C-H and F-C-C segments present in all-cis-1,2,3,4,5,6-hexafluorocyclohexane (19). 9

Scheme 3. a) n(S) → σ*(C-P)app stereoelectronic interaction in 1-ax. b) σ(C(4,6)-S) → σ*(C-P)app stereoelectronic interaction in 1-eq.

Intrigued by this report, we carried out theoretical calculations of the conformational behavior of r-1,c-3,c-5-trifluorocyclohexane (2-ax ⇌ 2-eq, Scheme 4), r-2,c-4,c-6-trifluoro-1,3,5-trioxane (3-ax ⇌ 3-eq, Scheme 4) and r-2,c-4,c-6-trifluoro-1,3,5-trithiane (4-ax ⇌ 4-eq, Scheme 4). We were able to confirm the importance of n(F) → σ*(C-X)gem stereoelectronic interactions in 2, 3 and 4. Furthermore, the relevance of “anomeric” type n(X) → σ*(C-F)app (X = O, S) stereoelectronic interactions in the axial conformations of heterocycles 3 and 4 was also established.

r-1,c-3,c-5-Trifluorocyclohexane, 2 The optimized geometries of axial and equatorial 2 at the MP2/6-311+G(d,p) level of theory, are presented in Figure 3 and Table 3.

10

Scheme 4. Conformational equilibria of fluorinated compounds examined in the present study: r-1,c-3,c-5-trifluorocyclohexane (2-ax ⇌ 2-eq), r-2,c-4,c-6-trifluoro-1,3,5-trioxane (3-ax ⇌ 3-eq), and r-2,c-4,c-6-trifluoro-1,3,5-trithiane (4-ax ⇌ 4-eq). Reproduced with permission from ref. (20). Copyright [2015] ACS.

Figure 3. MP2/6-311+G(d,p)-optimized structures of r-1,c-3,c-5trifluorocyclohexane, in the axial conformation, 2-ax, and the equatorial conformer, 2-eq. Reproduced with permission from ref. (20). Copyright [2015] ACS.

11

Table 3. MP2/6-311+G(d,p)-optimized geometrical parameters of r-1,c-3,c-5-trifluorocyclohexane, in the axial, 2-ax, and equatorial, 2-eq, conformations. Bond distances in Å.

a

2-axial

2-equatorial

C-Fax

1.395

---

C-Feq

---

1.397

C-C

1.522

1.521

C-Hax

1.098

1.096(1.097)a

C-Heq

1.094 (1.096)a

1.094

H geminal to F.

As it could be anticipated in terms of strong dipole-dipole repulsion originated from the 1,3-diaxial orientation of the C-F bonds, 1-axial is calculated to be 3.6 kcal/mol less stable than 1-equatorial. This result is in line with the value calculated recently by Schaefer, et al. (ΔE = 3.45 kcal/mol) (21). Table 4 summarizes the delocalization energies (E(2)) for the most relevant hyperconjugative interactions operative in trifluorocyclohexanes 2-ax and 2-eq. Table 4 presents also the calculated energy difference (energy gap, ΔE) between the donor and acceptor orbitals of interest. As anticipated, an inverse relationship between donor/acceptor energy gap and the magnitude of the two-electron/twoorbital hyperconjugative interaction is observed. Salient observations are the following: n(Fax) → σ*(C-Heq)gem with an interaction energy of 9.17 kcal/mol; n(Feq) → σ*(C-Hax)gem with an interaction energy of 8.93 kcal/mol; n(Fax) → σ*(C-C)gem with an interaction energy of 5.21 kcal/mol; and n(Feq) → σ*(C-C)gem with an interaction energy of 4.82 kcal/mol. These results seem to confirm the suggestion advanced by O’Hagan et al. (19) in the sense that the donor character of the fluorine lone pair towards the geminal sigma bonds is significant. In this regard, in our calculations the σ*(C-H) antibonding orbital is estimated to be a better acceptor orbital relative to σ*(C-C), as it could be anticipated on the basis of the accepted interpretation of the gauche effect (1). The calculations also show quite strong σ(C-Hax) → σ*(C-Fax)app hyperconjugative interactions, worth 6.12 kcal/mol (Table 4).

12

Table 4. Selected hyperconjugative interactions in r-1,c-3,c-5trifluorocyclohexane, 2-ax and 2-eq. 2-axial

2-equatorial

E(2)/kcal/mol

ΔE/Hartrees

E(2)/kcal/mol

ΔE/Hartrees

n(Fax) → σ*(C-Heq)

9.17

1.23

---

---

n(Feq) → σ*(C-Hax)

---

---

8.93

1.22

σ(C-Hax) → σ*(C-Fax)

6.12

1.19

---

---

n(Fax) → σ*(C-C)

5.21

1.24

---

---

n(Feq) → σ*(C-C)

---

---

4.82

1.25

σ(C-Heq) → σ*(C-C)

3.59a 3.38b

1.33a 1.31b

3.22

1.31

3.50c 3.29d

1.27c 1.32d

3.55

1.33

σ(C-Hax) → σ*(C-Hax) σ(C-C) → σ*(C-Feq) a Heq geminal to Fax.

--b Heq geminal to Hax.

---

c Hax geminal to Heq.

d Hax geminal to Feq.

r-2,c-4,c-6-Trifluoro-1,3,5-trioxane, 3 The lowest energy structures of axial and equatorial 3 at the MP2/6311+G(d,p) level of theory, are presented in Figure 4 and Table 5.

Figure 4. MP2/6-311+G(d,p)-optimized structures of r-2,c-4,c-6-trifluoro-1,3,5trioxane, in the axial, 3-ax, and equatorial, 3-eq, conformations. Reproduced with permission from ref. (20). Copyright [2015] ACS.

13

Table 5. MP2/6-311+G(d,p)-optimized structural parameters of r-2,c-4,c-6-trifluoro-1,3,5-trioxane, in the axial, 3-ax, and equatorial, 3-eq, conformations. Bond distances in Å. 3-axial

3-equatorial

C-Fax

1.363

---

C-Feq

---

1.334

C-O

1.390

1.395

C-Hax

---

1.099

C-Heq

1.086

---

In strong contrast with the conformational energies calculated for r-1,c-3,c5-trifluorocyclohexane (2), the all-axial conformer 3-ax was predicted to be 7.0 kcal/mol lower in energy (more stable) than the equatorial isomer 3-eq. That is, replacement of the methylene groups in 2 for oxygen atoms in 3 is accompanied by 10.6 kcal/mol stabilization of the axial conformer, in spite of the strong dipoledipole repulsive interactions between syn-diaxial C-F bonds. Very likely, this conformational is the consequence of the n(O) → σ*(C-Fax)app and σ(C-Hax → σ*(C-Fax)app stereoelectronic interactions present in 3-ax but not in 3-eq. In line with this interpretation, the structural data estimated for the optimized structures of 3-ax and 3-eq (Table 5) show that the axial C-F bonds are significantly longer (1.363 Å) than the equatorial C-F bonds (1.334 Å), as anticipated in terms of the proposed interactions that weaken the axial C-F bonds (1). Application of the NBO analysis gives evidence for the salient hyperconjugative interactions present in 3-ax and 3-eq (Table 6). Quite significant are the rather large interaction energies involving fluorine as a lone electron pair donor. As it can be appreciated in Table 6, the strongest hyperconjugative interaction results from n(O) → σ*(C-Fax)app, with 19.07 kcal/mol interaction energy. Quite strong is also the hyperconjugation involving the fluorine lone electron pairs: n(Fax) → σ*(C-O)gem (10.75 kcal/mol interaction energy) and n(Feq) → σ*(C-O)gem (11.03 kcal/mol interaction energy). Most important is also the “anomeric-type” n(O) → σ*(C-O)app hyperconjugative interaction, with 7.1 to 13.1 kcal/mol interaction energies (Table 6).

r-2,c-4,c-6-Trifluoro-1,3,5-trithiane, 4 The optimized structure of axial and equatorial 4 at the MP2/6-311+G(d,p) level of theory, are presented in Figure 5 and Table 7.

14

Table 6. Selected hyperconjugative interactions in r-2,c-4,c-6-trifluoro-1,3,5trioxane, 3-ax and 3-eq. 3-axial

3-equatorial

E(2)/kcal/mol

ΔE/Hartrees

E(2)/kcal/mol

ΔE/Hartrees

n(O) → σ*(C-Fax)

19.07

1.03

---

---

n(O) → σ*(C-Feq)

---

---

6.88

1.39

n(Fax) → σ*(C-O)

10.75 4.04

1.19 1.19

---

---

n(Feq) → σ*(C-O)

---

---

11.03 4.34

1.19 1.19

n(Fax) → σ*(C-Heq)

9.24

1.24

---

---

n(Feq) → σ*(C-Hax)

---

---

10.20

1.20

n(O) → σ*(C-O)

7.10 3.71

1.08 1.39

13.10

1.08

σ(C-Heq) → σ*(C-O)

3.92

1.31

n(O) → σ*(C-Hax)

---

---

7.58

1.08

n(O) → σ*(C-Heq)

3.46

1.43

---

---

σ(C-O)→σ*(C-Fax)gem

1.24

1.65

---

---

σ(C-O)→σ*(C-Feq)gem

---

---

2.48

1.66

Figure 5. MP2/6-311+G(d,p)-optimized structures of r-2,c-4,c-6-trifluoro-1,3,5trithiane, in the axial, 4-ax, and equatorial, 4-eq, conformations. Reproduced with permission from ref. (20). Copyright [2015] ACS.

15

Table 7. MP2/6-311+G(d,p)-optimized geometrical parameters of r-2,c-4,c-6-trifluoro-1,3,5-trithiane, in the axial, 4-ax, and equatorial, 4-eq, conformations. Bond distances in Å. 4-axial

4-equatorial

C-Fax

1.351

---

C-Feq

---

1.339

C-S

1.804

1.815

C-Hax

---

1.080

C-Heq

1.080

---

Table 8. Selected hyperconjugative interactions in r-2,c-4,c-6-trifluoro-1,3,5trithiane, 4-ax and 4-eq. 4-axial

4-equatorial

E(2)/kcal/mol

ΔE/Hartrees

E(2)/kcal/mol

ΔE/Hartrees

n(S) → σ*(C-Fax)

15.38

0.82

---

---

n(S) → σ*(C-Feq)

---

---

1.06

1.28

n(Fax) → σ*(C-S)

8.31 2.91

0.98 1.02

---

---

n(Feq) → σ*(C-S)

---

---

6.96 2.66

0.99 1.01

n(Fax) → σ*(C-Heq)

9.84

1.24

---

---

n(Feq) → σ*(C-Hax)

---

---

9.53

1.23

n(S) → σ*(C-S)

5.41 3.38

0.72 1.18

9.74

0.73

σ(C-Heq) → σ*(C-S)

1.82

1.10

n(S) → σ*(C-Hax)

---

---

4.24

0.94 ---

n(S) → σ*(C-Heq)

0.83

1.39

---

σ(C-S) →σ*(C-Fax)gem

1.50

1.29

---

---

σ(C-S) →σ*(C-Feq)gem

---

---

2.18

1.27

Most interesting is the much larger stability of the axial conformer 4-axial relative to 4-equatorial. The calculated structural data, specially the longer C-Fax (1.351 Å) by comparison with C-Feq (1.349 Å), as well as the shorter encocyclic C-S bond in 3-axial (1.804 Å) relative to the same bond in 4-equatorial (1.815 Å) are in line with n(S) → σ*(C-F)app hyperconjugation. 16

Indeed, application of the NBO analysis provides evidence for the relevant hyperconjugative interactions (Table 8), which confirm the rather strong n(S) → σ*(C-Fax)app stereoelectronic interaction, worth 15.38 kcal/mol. While the magnitude of this stereoelectronic interaction involving sulfur tends to be smaller than the corresponding one with oxygen [E(2) = 19.07 kcal/mol, Table 6] it contrasts with early theoretical studies from the Schleyer group suggesting that sulfur is an ineffective donor in S-C-X anomeric segments (22). The data collected in Table 8 also support the existence of significant interactions involving fluorine as a lone pair donor in n(Fax) → σ*(C-S)gem, n(Feq) → σ*(C-S)gem, n(Fax) → σ*(C-Heq)gem, and n(Feq) → σ*(C-Hax)gem stereoelectronic interactions.

Conclusions Theoretical calculations confirm the donor ability of fluorine lone electron pairs in hyperconjugative interactions involving geminal sigma bonds as acceptors. This is in agreement with the recent proposal by O’Hagan and coworkers from examination of the structural characteristics of all-cis-1,2,3,4,5,6-hexafluorocyclohexane. In particular, compelling evidence supporting the importance of n(F) → σ*(C-C)gem, n(F)→σ*(C-H)gem, n(F)→σ*(C-O)gem, and n(F)→ σ*(C-S)gem was gathered from the study of r-1,c-3,c-5-trifluorocyclohexane (2), r-2,c-4,c-6-trifluoro-1,3,5-trioxane (3) and r-2,c-4,c-6-trifluoro-1,3,5-trithiane (4). The Natural Bond Orbital (NBO) method developed by Weinhold and co-workers (18) was a convenient theoretical method for the study of the hyperconjugative interactions present in 2-4. MP2/6-311+G(d,p) calculations on the conformational equilibria of 2-4 indicate that 2-ax ⇌ 2-eq is shifted to the right by 3.6 kcal/mol, while 3-ax ⇌ 3-eq and 4-ax ⇌ 4-eq equilibria strongly favor the axial isomer by 7.0 and 10.1 kcal/mol, respectively. The strong axial preference for 3-ax and 4-ax originates from dominant n(X) → σ*(C-F)app hyperconjugative interactions, where X = O or S.

References 1. 2. 3. 4. 5. 6. 7.

8.

Juaristi, E.; Cuevas, G. The Anomeric Effect; CRC Press: Boca Raton, 1995. Wang, C.; Ying, F.; Wu, W.; Mo, Y. J. Am. Chem. Soc. 2011, 133, 13731–13736. Eliel, E. L. Acc. Chem. Res. 1970, 3, 1–8. Eliel, E. L. Pure Appl. Chem. 1971, 25, 509–525. Eliel, E. L. Angew. Chem., Int. Ed. 1972, 11, 739–750. Edward, J. T. Chem. Ind. (London) 1955, 1102–1104. Lemieux, R. U.; Chü, P. Abstracts of Papers; 133rd National Meeting of the American Chemical Society; American Chemical Society: Washington, DC, 1958. Juaristi, E.; Notario, R. J. Org. Chem. 2015, 80, 2879–2883. 17

9. 10. 11. 12.

13. 14. 15. 16. 17. 18.

19. 20. 21. 22.

Altona, C.; Romers, C.; Buys, H. R.; Havinga, E. Top. Stereochem. 1969, 4, 39–97. See, also: Bailey, W. F.; Eliel, E. L. J. Am. Chem. Soc. 1974, 96, 1798–1806. Juaristi, E.; Valle, L.; Valenzuela, B. A.; Aguilar, M. A. J. Am. Chem. Soc. 1986, 108, 2000–2005. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Keith, T.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision C.01; Gaussian, Inc.: Wallingford, CT, 2010. Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789. Reed, A. E.; Weinhold, F. Isr. J. Chem. 1991, 31, 277–285. Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; Weinhold, F. NBO, Version 3.1; University of Wisconsin: Madison, WI, 1988. Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105, 2999–3094. Alabugin, I. V. J. Org. Chem. 2000, 65, 3910–3919. Weinhold, F. Natural Bond Orbital Methods. In Encyclopedia of Computational Chemistry; Schleyer, P. v. R., Allinger, N. L., Clark, T., Gasteiger, J., Kollman, P. A., Schaefer, H. F., III, Schreiner, P. R., Eds.; Wiley: Chichester, U.K., 1998; Vol. III, pp 1792−1811. Keddie, N. S; Slawin, A. M. Z.; Lebl, T.; Philp, D.; O’Hagan, D. Nat. Chem. 2015, 7, 483–488. Juaristi, E.; Notario, R. J. Org. Chem. 2016, 81, 1192–1197. Luo, Q.; Randall, K. R.; Schaefer, H. F. RSC Adv. 2013, 3, 6572–6585. Schleyer, P. v. R.; Jemmis, E. D.; Spitznagel, G. W. J. Am. Chem. Soc. 1985, 107, 6393–6394.

18

Chapter 2

The Importance of Electrostatic Interactions on the Conformational Behavior of Substituted 1,3-Dioxanes: The Case of 5-Phenyl-1,3-dioxane William F. Bailey* and Kyle M. Lambert Department of Chemistry, University of Connecticut, Storrs, Connecticut 06269-3060, United States *E-mail: [email protected]

Seminal, fundamental investigations by Ernest Eliel’s group in the 1960’s and 70’s of conformational equilibria in substituted 1,3-dioxanes contributed significantly to our current understanding of the etiology of stereoelectronic effects in saturated heterocyclic systems. Following Eliel’s lead, we have explored the conformational behavior of 5-phenyl-1,3-dioxanes that bear remote substituents in an effort to probe the effect of the distant substituent on the conformational equilibria of such systems.

Introduction The conformational energy of 5-phenyl.1,3-dioxane was determined, as illustrated in Scheme 1, by Eliel and Knoeber in 1968 (1). The energy difference, 1.03 ± 0.02 kcal / mol favoring the equatorial isomer, is notably smaller than the ~2.8 kcal/mol conformational energy of phenylcyclohexane (2). The lower conformational energy of 5-phenyl-1,3-dioxane vis-à-vis phenylcyclohexane was attributed at the time to, “a diminution of the [syn-] axial interactions because, where in cyclohexane there are axial hydrogens in positions 1 and 3, in 1,3-dioxane there are, instead, electron pairs on oxygen” (1). However, given the results of recent investigations of the conformational behavior of phenylcyclohexane, this explanation must be reconsidered.

© 2017 American Chemical Society

Scheme 1. Conformational energy of 5-phenyl-1,3-dioxane. Reproduced from reference (3). Copyright 2016, American Chemical Society.

The main interactions responsible for the rather large conformational energy of phenylcyclohexane, in which the plane of the axial phenyl ring is perpendicular to the benzylic C–H bond, were shown by Allinger and Tribble (4) to be steric repulsion between the equatorial hydrogens at the C(2) and C(6) positions of the cyclohexane ring and the ortho- hydrogens of the axial phenyl group. Thus, syn-axial interactions contribute little to the conformational energy of phenylcyclohexane. It seems clear that the same repulsive steric interactions present in phenylcyclohexane should also be in 5-phenyl-1,3-dioxane were it also to adopt a rotameric conformation having the plane of the ring perpendicular to the C(5) benzylic hydrogen. Given this context, the question is: why is the conformational energy of a phenyl group in 5-phenyl-1,3-dioxane only a third as large as that of phenylcyclohexane? An unexpected result of a MP2/6-311+G*computational investigation of the rotameric conformations of phenyl rings in a series of axially and equatorially substituted 1,3-dioxanes and tetrahydropyrans (5) suggested an answer to this question. As shown in Figure 1, the computed lowest energy rotamer of axial 5-phenyl-1,3-dioxane is one in which the plane of the phenyl ring bisects the 1,3-dioxane ring and is parallel to the benzylic C(5)–H bond. This rotameric arrangement leads to an attractive CH…O Coulombic interaction between an ortho-hydrogen and the oxygens of the 1,3-dioxane ring. We reasoned that the strength of such an interaction should respond to substituents placed remotely on the phenyl ring. That is: electron-withdrawing groups should strengthen the attractive CH…O interaction and electron-donating groups should lessen the interaction. The results presented below demonstrate that this is indeed the case.

20

Figure 1. Calculated minimum energy rotamer of axial 5-phenyl-1,3-dioxane. Reproduced from reference (5). Copyright 2015, American Chemical Society.

Results and Discussion A representative series of anancomeric 2-alkyl-5-aryl-1,3-dioxanes, 1 – 14, bearing remote substituents on the phenyl ring were prepared by acid catalyzed condensation of isobutyraldehyde or pivaldehyde with 2-aryl-1,3-propane diols (3). The diastereoisomeric pairs of 5-aryl-1,3-dioxanes were then separated chromatographically and fully characterized by 1H NOESY analysis (3). An X-ray crystallographic analysis of cis-2-t-butyl-5-p-chlorophenyl-1,3dioxane (3), shown in Figure 2, validates the computational result noted above: the axial phenyl ring in adopts a rotameric arrangement in which an otho-hydrogen is in close proximity to one of the oxygen atoms of the dioxane (3). Following the protocol pioneered by Eliel and Knoeber (1), each of the 5aryl-1,3-dioxane pairs (1-14) were equilibrated at room temperature as solutions in either cyclohexane or diethyl ether over dry Amberlyst-15 resin. After the solutions were neutralized by shaking with anhydrous K2CO3, the area ratio of the isomeric mixture was determined by capillary GC analysis. It was judged that equilibrium had been reached when the same area ratios were obtained from initially pure samples of each isomer. Area ratios for each equilibration, which reflect the equilibrium constant for the process, were taken as the average of 5−14 independent determinations from each side, and the free energy difference for the equilibrium was calculated in the normal way: ΔG° = −RT ln K. The results of these studies are summarized in Table 1.

21

Figure 2. Crystal structure of cis-2-t-butyl-5-(p-chlorophenyl)-1,3-dioxane (3). The top view is from the side; the lower view, from above, shows the CH…O interaction. Reproduced from reference (3). Copyright 2016, American Chemical Society.

22

Table 1. Equilibria in 5-aryl -1,3-dioxanes. Reproduced from reference (3). Copyright 2016, American Chemical Society.

23

Inspection of the data presented in Table 1 demonstrates that substituents on the phenyl ring, as remote as the para position, affect the conformational energy of a phenyl group. Electron-withdrawing substituents (p-Cl, p-Br, p-CF3, and 3,5-bisCF3) stabilize the cis-isomer while electron-donating groups (p-OMe and p-TMS) have a destabilizing effect. Amazingly and unexpectedly, a 3,5-bis-CF3 substituted 5-phenyl group actually displays a pronounced preference for the axial orientation (Table 1, entries 9 and 10). This is, to our knowledge, an unprecedented result. It would appear that substituents affect the strength of the non-classical CH…O hydrogen bond (6) between an ortho-hydrogen and a dioxane oxygen. This, in turn, is reflected in the conformational energy of a substituted 5-phenyl-1,3-dioxane. The origin of these substituent effects is almost certainly electrostatic. A Hammett plot of the experimental ΔG° values in cyclohexane solution from Table 1 versus σm constants, derived from the pKa’s of substituted benzoic acids (7), presented in Figure 3, displays a very linear correlation (r = 0.98) having a slope (ρ) of +1.5. In this connection, it should be noted that para-substituents are meta with respect to the ortho hydrogen of the phenyl ring that interacts with a ring oxygen. This linear correlation strongly implies that the effect of substituents on the ΔG° of a 5-phenyl-1,3-dioxane has the same origin as the effect of those substituents on the acidity of benzoic acid: namely, an inductive, electrostatic phenomenon.

Figure 3. Hammett plot of experimental ΔG° values (Table 1) determined in cyclohexane solution vs. σm values. Reproduced from reference (3). Copyright 2016, American Chemical Society.

In short, electron-withdrawing substituents render the ortho-hydrogens of the axial C(5) phenyl group more positive thus increasing the attractive non-classical 24

CH…O hydrogen bond of that hydrogen with an oxygen of the 1,3-dioxane. The conclusion is rendered pictorially in Figure 4.

Figure 4. Effect of electron-withdrawing p-substituents on the conformational energy of a 5-phenyl-1,3-dioxane.

Conclusions In summary, the minimum energy rotameric conformation of an axial 5-phenyl-1,3-dioxane has been demonstrated to be one that positions the aryl ring such that an ortho-hydrogen is in close proximity to one of the dioxane ring oxygens. The results described above demonstrate that the strength of this non-classical CH…O hydrogen bond may be tuned in response to the electron-withdrawing or electron-donating ability of substituents positioned remotely on the aryl ring: electron-withdrawing substituents decrease the conformational energy of the phenyl group while electron-donating substituents increase the conformational energy of the group. The results of this investigation of the conformational behavior of 5-phenyl-1,3-dioxanes strengthens the notion that non-classical CH…X hydrogen bonds are often relevant to an understanding of broader conformational issues involving heterocyclic systems bearing aryl groups. A full account of the study summarized above is available (3).

References 1. 2. 3.

4.

5.

6.

Eliel, E. L.; Knoeber, M. C. Conformational Analysis. XVI. 1, 3-Dioxanes. J. Am. Chem. Soc. 1968, 90, 3444–3458. Eliel, E. L.; Wilen, S. H. Stereochemistry of Organic Compounds; Wiley: New York, 1994; pp 697−698. Bailey, W. F.; Lambert, K. M.; Stempel, Z. D.; Wiberg, K. B.; Mercado, B. Q. Controlling the Conformational Energy of a Phenyl Group by Tuning the Strength of a Nonclassical CH···OHydrogen Bond: The Case of 5-Phenyl1,3-dioxane. J. Org. Chem. 2016, 81, 12116–12127. Allinger, N. L.; Tribble, M. T. Conformational Analysis. LXXVIII. The Conformation of Phenylcyclohexane and Related Molecules. Tetrahedron Lett. 1971, 12, 3259–3262. Wiberg, K. B.; Lambert, K. L.; Bailey, W. F. The Role of CH···OCoulombic Interactions in Determining Rotameric Conformations of Phenyl Substituted 1,3-Dioxanes and Tetrahydropyrans. J. Org. Chem. 2015, 80, 7884–7889. For a review of non-classical hydrogen bonds, see: Takahashi, O.; Kohno, Y.; Nishio, M. Relevance of Weak Hydrogen Bonds in the Conformation 25

7.

of Organic Compounds and Bioconjugates: Evidence from Recent Experimental Data and High-Level ab Initio MO Calculations. Chem. Rev. 2010, 110, 6049–6076. Hansch, C.; Leo, A.; Taft, R. W. A Survey of Hammett substituent Constants and Resonance and Field Parameters. Chem. Rev. 1991, 91, 165–195.

26

Chapter 3

Asymmetric Autocatalysis and the Origin of Homochirality Kenso Soai* and Arimasa Matsumoto Department of Applied Chemistry, Tokyo University of Science, Kagurazaka, Shinjuku-ku, Tokyo, 162-8601 Japan *E-mail: [email protected]

Asymmetric autocatalysis is a reaction in which chiral product acts as chiral catalyst for its own production. Pyrimdin-5-yl-iso-propylcarbinol (pyrimidyl alkanol) was found to be a highly efficient asymmetric autocatalyst in the enantioselective addition of diisopropyl zinc to pyrimidine-5-carbaldehyde to produce more of itself (Soai reaction). The process is the automultiplicaion, i.e., catalytic self-replication, of a chiral compound. Pyrimidyl alkanol with extremely low enantiomeric excess (ee) automultiplies and is amplified during three consecutive asymmetric autocatalyses to reach >99.5% ee. By using asymmetric autocatalysis, the origin of homochirality has been examined. Circularly polarized light, inorganic chiral crystals such as quartz, enantiotopic surface of achiral inorganic chiral such as gypsum, chiral crystals of achiral organic compounds, chiral compounds of isotope (hydrogen, carbon, oxygen, nitrogen) substitution, acting as chiral initiators in asymmetric autocatalysis, can be correlated to the highly enantioenriched products. Spontaneous absolute asymmetric synthesis was achieved without the intervention of any chiral factor.

1. Introduction Living organisms on Earth are composed of highly enantioenriched compounds such as L-amino acids and D-sugars (Figure 1). Enzymes composed of random mixtures of D- and L-amino acids would not operate, and neither © 2017 American Chemical Society

would DNA and RNA composed of D- and L-(deoxy)ribose. The origin of chiral homogeneity of biomolecules, often called as biological homochirality, has been a research subject of broad interest, because the biological homochirality is considered to be essential for the origin and evolution of life. Several theories have been proposed for the origins of the chirality of organic compounds (1–7). However, the enantiomeric excesses of organic compounds induced by the proposed mechanisms have been very low (99.5% ee as an asymmetric autocatalyst, (S)-2-alkynyl-5-pyrimidyl alkanol 4 with >99.5% ee, was obtained in a yield of >99% (Scheme 3) (27). The product 4 in one round was used as an asymmetric autocatalyst for the following round. Even after the tenth round, the yield of newly formed 4 was >99% and the ee was >99.5%. Thus, 2-alkynylpyrimidyl alkanol 4 was found to be a practically perfect asymmetric autocatalyst.

Scheme 3. Practically perfect asymmetric autocatalysis.

3. Amplification of Enantiomeric Excess in Asymmetric Autocatalysis An asymmetric non-autocatalytic reaction with amplification of ee was first reported by Kagan (28). In asymmetric autocatalysis, if the amplification of ee is observed, the reaction would become a more powerful method for amplifying a very low ee to a very high ee by applying consecutive asymmetric autocatalysis. Indeed, we found that the ee of pyrimidyl alkanol increased during the asymmetric autocatalysis (26). We used the product of one round as the asymmetric autocatalyst for the following round. 30

When (S)-pyrimidyl alkanol 2 with only 2% ee was employed as the initial asymmetric autocatalyst, the reaction afforded (S)-2 with an increased ee of 10% (Scheme 4) (26). The (S)-2 with 10% ee was then used as the asymmetric autocatalyst for the following consecutive asymmetric autocatalysis. The ee was further amplified step-by-step to 57, 81, and 88% ee. The overall process is the asymmetric autocatalysis of (S)-2 starting from a low 2% ee and with significant amplification of ee to 88% ee. This is the first asymmetric autocatalysis with amplification of ee. Regarding the amplification of ee, in sharp contrast to non-autocatalytic amplification reactions, One of the striking features of the amplification of ee in the present asymmetric autocatalysis is that there is no need for any chiral auxiliary other than the asymmetric autocatalyst itself.

Scheme 4. Asymmetric autocatalysis with amplification of ee. After examining the effect of the substituent at 2 position of pyrimidyl alkanol, 2-alkynylpyrimidyl alkanol 4 was finally found to exhibit the most efficient amplification of ee. Thus, starting from alkanol 4 with ee as low as ca. 0.00005% ee, three consecutive asymmetric autocatalyses amplified ee to 57, 99, and >99.5% ee (Figure 2) (29). The initially slightly major (S)-enantiomer of 4 automultiplied by a factor of ca. 630,000 during these three consecutive asymmetric autocatalyses, whereas the initially slightly minor (R)-enantiomer of 4 automultiplied by a factor of less than 1,000. As described, we found highly efficient asymmetric autocatalysts with significant amplification of ee.

4. The Origin of Chirality Examined by Using Asymmetric Autocatalysis The origins of chirality proposed so far can induce chiral compounds with very low enantiomeric excess. A chiral compound with low ee may act as a chiral initiator for the asymmetric autocatalysis of pyrimidyl alkanol (30). The subsequent amplification of ee during the asymmetric autocatalysis affords highly enantioenriched pyrimidyl alkanol with the corresponding absolute configuration to that of the chiral initiator (31). Thus, asymmetric autocatalysis can act as a link between the low ee induced by the origin of chirality and compounds with high ee observed in nature (Scheme 5). 31

Figure 2. Automultiplication of (S)- and (R)-4 in consecutive asymmetric autocatalysis.

Scheme 5. Schematic correlation of the origin of chirality to chiral compound with high enantiomeric excess. 32

4.1. Circularly Polarized Light 4.1.1. Chiral Compounds Induced with Circularly Polarized Light Initiate Asymmetric Autocatalysis Circularly polarized light (CPL) has long been considered one of the physical chiral factors of the origin of chirality (4, 6, 32). Asymmetric photolysis of racemic leucine by right-handed CPL (r-CPL, 213 nm) induces only 2% ee in residual L-leucine (33). Hexahelicene with low ee (99.5% ee (Scheme 7) (37). On the contrary, upon the irradiation of righthanded (r)-CPL, (R)-4 with >99.5% ee was formed. The rationalization regarding the relationship between the handedness of CPL and alkanol 4 can be made as follows: (R)- and (S)-pyrimidyl alkanols 4 exhibit positive and negative circular dichroism (CD) spectra at 313 nm, respectively. This means that (R)-4 absorbs l-CPL preferentially. Then the direct irradiation of l-CPL to racemic alkanol 4 33

would induce the asymmetric photodegradation of (R)-4 preferentially and leave the (S)-4 with low ee. Thus, even if the ee of the remaining (S)-4 was extremely low after the asymmetric photodegradation by CPL, the remaining compound 4 itself would serve as an asymmetric autocatalyst in the asymmetric autocatalysis with significant amplification of ee.

Scheme 7. A route to obtain a near enantiopure compound by CPL irradiation in conjunction with asymmetric autocatalysis.

4.2. Chiral Inorganic Material 4.2.1. Chiral Inorganic Crystals as the Origins of Chirality in Conjunction with Asymmetric Autocatalysis Quartz is a chiral inorganic crystal and exhibits either a dextrorotatory (d) or a levorotatory (l) enantiomorph. The rotation of plane polarized light, i.e., optical activity, was found with quartz. Thus, quartz has been considered for many years the origin of chirality in nature; however, it is unclear whether quartz induces the apparent enantiomeric imbalance in organic compounds. Thus, it is a challenge to use quartz as a chiral initiator in the asymmetric autocatalysis of pyrimidyl alkanol. When 2-alkynylpyrimidine-5-carbaldehyde 3 was reacted with i-Pr2Zn in the presence of d-quartz powder as a chiral initiator, (S)-pyrimidyl alkanol 4 with 97% ee was formed in 95% yield (Scheme 8) (38). In sharp contrast, in the presence of l-quartz powder, (R)-4 with 97% ee was formed in 97% yield. These results clearly show that the chirality of quartz controls the absolute configurations of the obtained pyrimidyl alkanol 4. Thus, the first experimental realization was achieved that a chiral inorganic crystal as the origin of chirality is correlated to that of a chiral organic compound with high ee.

34

Scheme 8. Asymmetric autocatalysis triggered by quartz.

Cinnabar (HgS) and Retgersite (NiSO4•6H2O) are other examples of chiral inorganic crystal. These chiral inorganic crystals were also found to trigger the asymmetric autoctalysis (39, 40). In addition to the chiral inorganic crystals, chiral inorganic nano-structure also trigger the asymmetric autocatalysis, and helical silica and helical mesoporous silica can direct the enantiomeric outcome of the asymmetric autocatalysis (Scheme 9) (41, 42).

Scheme 9. Asymmetric autocatalysis triggered by helical silica gel.

35

4.2.2. Enantiotopic Face of Achiral Inorganic Crystal Surface structures also have chirality and the crystal surfaces sometimes exhibit chirality even the whole crystal structure itself is achiral. Gypsum (CaSO4•2H2O) is an achiral inorganic mineral, but is known to have enantiotopic faces. The asymmetric autocatalysis performed on the enantiotopic surface of gypsum afforded enantio-enriched chiral compounds. The addition reaction of diisopropyl zinc vapor to the adsorbed aldehyde 3 gives the enantioenriched alkanol 4 with the corresponding absolute structure to the adsorbed crystal surface of the gypsum (Scheme 10). This result is the first example of the enantioselectivity control of the reaction product by the surface chirality of achiral inorganic crystal (43).

Scheme 10. Asymmetric autocatalysis on the enantiotopic surface of achiral Gypsum.

4.3. Asymmetric Autocatalysis Triggered by Chiral Organic Crystals Composed of Achiral Organic Compounds Some of the achiral organic compounds crystallize in chiral forms. It was unclear whether chiral crystals of achiral compounds can act as a chiral inducer in enantioselective reactions. We found an enantioselective reaction using these chiral crystals as chiral inducers of asymmetric autocatalysis. Cytosine is an essentially flat achiral molecule and a nucleobase of cytidine and deoxycytidine. Cytosine is known to form a chiral crystal 5 (space group: P212121) when it is crystallized from methanol. It was found that achiral cytosine when crystallized from methanol with stirring and without adding any seed crystal affords powder-like crystals that exhibit either a plus or minus Cotton effect in solid-state CD spectra at ca. 310 nm in Nujol mulls (44). The distribution of the formation of [CD(+)310Nujol]-5 and [CD(–)310Nujol]-5 was stochastic. Next, the chiral crystals 5 were used as chiral triggers for asymmetric autocatalysis (Scheme 11) (44). When pyrimidine-5-carbaldehyde 3 and i-Pr2Zn were reacted in the presence of [CD(+)310Nujol]-5, enantioenriched (R)-pyrimidyl alkanol 4 was obtained after the subsequent asymmetric autocatalysis. On the 36

other hand, [CD(–)310Nujol]-5 afforded (S)-4. These results clearly show that the cytosine crystal acts as the origin of chirality in asymmetric autocatalysis.

Scheme 11. Asymmetric autocatalysis triggered by chiral crystal of achiral cytosine.

Cytosine crystallizes from water as an achiral monohydrate (space group: P21/ c). It was found that the enantioselective formation of a chiral crystal of cytosine takes place through the dehydration of crystallization water by heating or in vacuo from the enantiotopic faces (Figure 3) (45, 46).

Figure 3. Chirality generation by dehydration of crystallization water.

In addition to cytosine, various achiral organic compound which crystallized in chiral structure act as chiral initiator of asymmetric autocatalysis (Figure 4) (47–53). 37

Figure 4. Enantiomorphs of achiral organic compounds, which can act as chiral initiators of asymmetric autocatalysis. 4.4. Asymmetric Autocatalysis Triggered by Isotope Chirality 4.4.1. Chiral H/D Isotopomers Isotope substitution sometimes make a chirality in achiral compounds (54). Glycine and α-methylalanine are known to be achiral amino acids. However, deuterium substitution of one of the hydrogen atoms of the methylene group of glycine and one methyl group of α-methylalanine makes these compounds chiral: glycine-α-d 17 and α-methyl-d3-alanine 18 (Scheme 12).

Scheme 12. Chiral hydrogen isotopomers of glycine and α-methylalanine trigger asymmetric autocatalysis. 38

The chiral glycine-α-d 6 and α-methyl-d3-alanine 7 resulting from isotope substitution were found to act as chiral initiators in asymmetric autocatalysis to afford pyrimidyl alkanol 4 with high ee (Scheme 12) (55). In the presence of (S)-6, (S)-5-pyrimidyl alkanol 4 was formed with high ee. On the other hand, (R)-4 was formed in the presence of (R)-6 instead of (S)-6. It was also found that (R)- or (S)-α-methyl-d3-alanine 7 act as the chiral initiator of asymmetric autocatalysis. Thus, (R)-7 afforded (S)-7, while (S)-7 afforded (R)-4. These results are the first examples of a highly enantioselective reaction induced by chirally deuterated amino acids.

4.4.2. Chiral Carbon Isotope (13C/12C) Chirality Many achiral organic molecules may become chiral by carbon isotope substitution (Figure 5). However, because the chirality originates from the very small difference between the carbon isotopes (13C/12C), it has been experimentally difficult to discriminate carbon isotope chirality. It has been a question whether chiral carbon chirality can induce chirality in some reactions. We found that carbon isotope (13C/12C) chiral compounds trigger asymmetric autocatalysis (56).

Figure 5. Carbon isotope (12C/13C) substitution generates chirality.

The carbon isotopically chiral compound methyl-13C-methylphenyl methanol 8 arising from 13C substitution of the methyl group was used as a chiral trigger. The preparation of alkanol 8 is shown in Scheme 13. In the presence of (R)-alkanol 8, when i-Pr2Zn and pyrimidine-5-carbaldehyde 3 were reacted, (S)-pyrimidyl alkanol 4 was formed with high ee (Scheme 14). On the contrary, (S)-8 afforded (R)-4. Chiral alcohols 9 and 10 resulting from 13C substitution act as chiral triggers of asymmetric autocatalysis to afford pyrimidyl alkanols 4 with high ee, and which have the corresponding absolute configurations of the isotopically substituted carbon chirality of 9 and 10. 39

Scheme 13. Asymmetric synthesis of 13C-labeled dimethylphenylmethanol 8.

Scheme 14. Asymmetric autocatalysis triggered by chiral carbon isotopomer.

4.4.3. Asymmetric Autocatalysis Triggered by Oxygen (16O/18O) and Nitrogen (14N/15N) Isotope Chirality Similar to the carbon, other atoms in typical organic compounds such as oxygen and nitrogen have stable isotopes. Recognition of chirality only from isotopes becomes much more difficult when the isotope atoms become heavier due to the small relative mass difference. 40

Asymmetric autocatalysis can recognize these isotope chiralities by oxygen (18O/16O) and nitrogen (15N/14N) (Scheme 15).

Scheme 15. Asymmetric autocatalysis triggered by chiral oxygen and nitrogen isotopomers. (1R,2S)-1,2-Diphenylethandiol-1,2-diol 11 is meso diol. However, introduction of one oxygen isotope 18O results in the isotopically chiral [18O](R)-11 and [18O](S)-11. Similarly, the achiral triol glycerin 12 becomes oxygen isotopically chiral glycerin by substitution of oxygene at 1- or 3- position by oxygen isotope 18O. These oxygen isotopically chiral compounds act as chiral initiators of asymmetric autocatalysis and the [18O](R)-11 afforded (S)-pyrimidyl alkanol 4 and the [18O](S)-11 gave the opposite enantiomer (57, 58). Nitrogen isotope substitution generates an isotope chirality in achiral diamine. Tetraethyl-3,4-butandiamine is meso diamine but the 15N substitution on one nitrogen reduces its symmetry, and isotopically chiral diamine [15N](S)-13 and [15N](R)-13 were obtained (59). These nitrogen isotopically chiral compounds also act as a chiral initiator in asymmetric autocatalysis. As mentioned, these results are the first examples of asymmetric induction by carbon, oxygen and nitrogen isotopically chiral compounds.

5. Absolute Asymmetric Synthesis based on Statistical Fluctuation and Amplification Spontaneous absolute asymmetric synthesis, i.e., the formation of enantioenriched compounds from achiral reagents without the intervention of any chiral factor, has been proposed as the origin of chirality (1, 60). However, in the usual reaction, the enantiomeric excesses of the products are far below the detection level, i.e., racemate. As described, asymmetric autocatalysis can amplify enantiomeric excess from extremely low to very high. We expected that when i-Pr2Zn was treated with pyrimidine-5-carbaldehyde without adding any chiral substance, initial statistically generated slight enantiomeric excess in the formation of (the zinc 41

alkoxide of) pyrimidyl alkanol, and that the amplification of enantiomeric excess by subsequent asymmetric autocatalysis would afford the pyrimidyl alkanol with detectable enantiomeric excess (Figure 6).

Figure 6. Absolute asymmetric synthesis of pyrimidyl alkanol without the intervention of chiral factor in conjunction with asymmetric autocatalysis. Indeed, when 2-alkynylpyrimidine-5-carbaldehyde 3 was reacted with i-Pr2Zn, and the subsequent asymmetric autocatalysis with amplification of ee gave (S)- or (R)-pyrimidyl alkanol 4 with enantiomeric excess above the detection level (61, 62). The absolute configurations of the product pyrimidyl alkanol 4 exhibit an approximate stochastic distribution of S- and R-enantiomers. In the presence of achiral silica gel (63) or achiral amine (64), it was also found that (S)and (R)-4 were formed with approximate stochastic distributions in the reaction of aldehyde 3 with i-Pr2Zn in toluene (Figure 7).

Figure 7. Histograms of the absolute configuration and ee of pyrimidyl alkanols. 42

These results fulfill one of the conditions necessary for spontaneous absolute asymmetric synthesis. Thus, absolute asymmetric synthesis between pyrimidine5-carbaldehyde and diisopropylzinc was achieved in conjunction with asymmetric autocatalysis.

6. Reaction Models and Crystal Structures of Asymmetric Autocatalysis Several groups have published papers on the reaction model of the present asymmetric autocatalysis (65–84). The enantiomeric amplification in the usual asymmetric catalysis is sometimes explained by the formation of dimer. However, in the mechanism of the amplification of enantiomeric excess from very low ee, an additional mechanism such as more aggregation may be involved. Actually, kinetic analysis of pyrimidyl alkanol suggested that the reaction is mainly of the first order in the zinc alkoxide of pyrimidyl alkanol, i.e., catalyst and second order in aldehyde and show an interesting temperature dependence. NMR studies also support the existance of higher order structures. The X-ray analysis of the single crystals of asymmetric autocatalyst, i.e., alkylzinc alkoxide of pyrimidyl alkanol, gave the following information: isopropylzinc alkoxide of enantiopure pyrimidyl alkanol with excess amount of i-Pr2Zn form tetrameric structure with two 4-membered rings of Zn-O-Zn-O, one 12-membered ring, and with the coordination of 6 molecule of i-Pr2Zn per one tetramer. When the amount of i-Pr2Zn is small, oligomeric crystal is formed. The results indicate that various aggregation status should exist in the reaction system and this accessibility to the higher aggregation state may play an important role in this high enantiomeric excess amplification (85, 86).

7. Conclusions Pyrimidyl alkanol was discovered as asymmetric autocatalyst with amplification of enantiomeric excess in the reaction between pyrimidine-5carbaldehyde and i-Pr2Zn. The process is a catalytic self-replication of a chiral molecule. The reaction exhibits significant amplification of enantiomeric excess, from extremely low ee to >99.5% ee. Thus, we proved that there is a chemical reaction in which the initial low ee can become >99.5% ee by asymmetric autocatalysis with amplification of ee. We apply asymmetric autocatalysis with amplification of ee for the correlation between the origin of chirality and chiral organic compounds with high enantiomeric excesses. Various chiral compounds and chiral factors trigger asymmetric autocatalysis. Irradiation of CPL to the racemate of pyrimidyl alkanol and the subsequent asymmetric autocatalysis affords a highly enantioenriched compound. Chiral inorganic crystals such as d- and l-quartz and cinnabar act as chiral triggers to afford highly enantioenriched compounds. Spontaneous absolute asymmetric synthesis was achieved for the first time in the reaction of pyrimidine-5-carbaldehyde and i-Pr2Zn in conjunction with asymmetric autocatalysis without the intervention of chiral factor. 43

Acknowledgments The authors are grateful to the coworkers whose names appear in the papers. Financial support from Japan Society for the Promotion of Science (JSPS) is 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.

24. 25. 26.

Mislow, K. Collect. Czech. Chem. Commun. 2003, 68, 849–864. Eschenmoser, A. Science 1999, 284, 2118–2124. Pályi, G., Zucchi, C., Caglioti, L., Eds. Fundamentals of Life; Accad. Naz. Sci. Lett. Arts; Elsevier Life Science: Paris, 2002. Feringa, B. L.; van Delden, R. A. Absolute Asymmetric Synthesis: The Origin, Control, and Amplification of Chirality. Angew. Chem., Int. Ed. 1999, 38, 3418–3438. Weissbuch, I.; Lahav, M. Chem. Rev. 2011, 111, 3236–3267. Inoue, Y. Chem. Rev. 1992, 92, 741–770. Bonner, W. A. Origins Life Evol. Biospheres 1991, 21, 59–111. Frank, F. C. Biochim. Biophys. Acta 1953, 11, 459–463. Bolm, C.; Bienewald, F.; Seger, A. Angew. Chem., Int. Ed. 1996, 35, 1657–1659. Avalos, M.; Babiano, R.; Cintas, P.; Jiménez, J. L.; Palacios, J. C. Chem. Commun. 2000 (11), 887–892. Blackmond, D. G. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 5732–5736. Podlech, J.; Gehring, T. Angew. Chem., Int. Ed. 2005, 44, 5776–5777. Caglioti, L.; Zucchi, C.; Pályi, G. Chem. Today 2005, 23, 38–43. Soai, K.; Shibata, T.; Sato, I. Acc. Chem. Res. 2000, 33, 382–390. Soai, K.; Sato, I.; Shibata, T. Chem. Record 2001, 1, 321–332. Soai, K. Asymmetric autocatalysis and the origin of chiral homogeneity of biologically relevant molecules. In Fundamentals of Life; Pályi, G., Zucchi, C., Caglioti, L., Eds.; Elsevier: Paris, 2002; pp 427–435. Soai, K.; Shibata, T.; Sato, I. Bull. Chem. Soc. Jpn. 2004, 77, 1063–1073. Soai, K.; Kawasaki, T. Chirality 2006, 18, 469–478. Soai, K.; Kawasaki, T. Top. Curr. Chem. 2008, 284, 1–33. Soai, K.; Kawasaki, T. In Organometallic Chirality; Pályi, G., Zucchi, C., Caglioti, L., Eds.; Accad. Naz. Sci. Lett. Arts; Mucchi Editore: Modena, 2008; pp 107–125. Kawasaki, T.; Soai, K. J. Fluor. Chem. 2010, 131, 525–534. Soai, K.; Kawasaki, T. Chem. Today 2009, 27 (6, Suppl.), 3–7. Soai, K.; Kawasaki, T.; Shibata, T. Asymmetric Amplification and Autocatalysis in Catalytic Asymmetric Synthesis, 3rd ed.; Ojima, I., Ed.; John Wiley & Sons, Inc.: Hoboken, NJ, 2010; Chapter 12, pp 891–930. Soai, K.; Kawasaki, T.; Matsumoto, A. Acc. Chem. Res. 2014, 47, 3643–3654. Soai, K.; Kawasaki, T.; Matsumoto, A. Chem. Rec. 2014, 70–83. Soai, K.; Shibata, T.; Morioka, H.; Choji, K. Nature 1995, 378, 767–768. 44

27. Shibata, T.; Yonekubo, S.; Soai, K. Angew. Chem., Int. Ed. 1999, 38, 659–661. 28. Girard, C.; Kagan, H. B. Angew. Chem., Int. Ed. 1998, 37, 2922–2959. 29. Sato, I.; Urabe, H.; Ishiguro, S.; Shibata, T.; Soai, K. Angew. Chem., Int. Ed. 2003, 42, 315–317. 30. Maioli, M.; Micskei, K.; Caglioti, L.; Zucchi, C.; Pályi, G. J. Math. Chem. 2008, 43, 1505–1515. 31. Shibata, T.; Yamamoto, J.; Matsumoto, N.; Yonekubo, S.; Osanai, S.; Soai, K. J. Am. Chem. Soc. 1998, 120, 12157–12158. 32. Noorduin, W. L.; Bode, A. a C.; van der Meijden, M.; Meekes, H.; van Etteger, A. F.; van Enckevort, W. J. P.; Christianen, P. C. M.; Kaptein, B.; Kellogg, R. M.; Rasing, T.; Vlieg, E. Nat. Chem. 2009, 1, 729–732. 33. Nishino, H.; Kosaka, A.; Hembury, G. A.; Aoki, F.; Miyauchi, K.; Shitomi, H.; Onuki, H.; Inoue, Y. J. Am. Chem. Soc. 2002, 124, 11618–11627. 34. Kagan, H.; Moradpour, A.; Nicoud, J. F.; Balavoine, G.; Tsoucaris, G. J. Am. Chem. Soc. 1971, 93, 2353–2354. 35. Sato, I.; Ohgo, Y.; Igarashi, H.; Nishiyama, D.; Kawasaki, T.; Soai, K. J. Organomet. Chem. 2007, 692, 1783–1787. 36. Sato, I.; Yamashima, R.; Kadowaki, K.; Yamamoto, J.; Shibata, T.; Soai, K. Angew. Chem., Int. Ed. 2001, 40, 1096–1098. 37. Kawasaki, T.; Sato, M.; Ishiguro, S.; Saito, T.; Morishita, Y.; Sato, I.; Nishino, H.; Inoue, Y.; Soai, K. J. Am. Chem. Soc. 2005, 127, 3274–3275. 38. Soai, K.; Osanai, S.; Kadowaki, K.; Yonekubo, S.; Shibata, T.; Sato, I. J. Am. Chem. Soc. 1999, 121, 11235–11236. 39. Shindo, H.; Shirota, Y.; Niki, K.; Kawasaki, T.; Suzuki, K.; Araki, Y.; Matsumoto, A.; Soai, K. Angew. Chem., Int. Ed. 2013, 52, 9135–9138. 40. Matsumoto, A.; Ozawa, H.; Inumaru, A.; Soai, K. New J. Chem. 2015, 39, 6742–6745. 41. Sato, I.; Kadowaki, K.; Urabe, H.; Jung, J. H.; Ono, Y.; Shinkai, S.; Soai, K. Tetrahedron Lett. 2003, 44, 721–724. 42. Kawasaki, T.; Araki, Y.; Hatase, K.; Suzuki, K.; Matsumoto, A.; Yokoi, T.; Kubota, Y.; Tatsumi, T.; Soai, K. Chem. Commun. 2015, 51, 8742–8744. 43. Matsumoto, A.; Kaimori, Y.; Uchida, M.; Omori, H.; Kawasaki, T.; Soai, K. Angew. Chem., Int. Ed. 2017, 56, 545–548. 44. Kawasaki, T.; Suzuki, K.; Hakoda, Y.; Soai, K. Angew. Chem., Int. Ed. 2008, 47, 496–499. 45. Kawasaki, T.; Hakoda, Y.; Mineki, H.; Suzuki, K.; Soai, K. J. Am. Chem. Soc. 2010, 132, 2874–2875. 46. Mineki, H.; Kaimori, Y.; Kawasaki, T.; Matsumoto, A.; Soai, K. Tetrahedron: Asymmetry 2013, 24, 1365–1367. 47. Kawasaki, T.; Jo, K.; Igarashi, H.; Sato, I.; Nagano, M.; Koshima, H.; Soai, K. Angew. Chem., Int. Ed. 2005, 44, 2774–2777. 48. Kawasaki, T.; Suzuki, K.; Hatase, K.; Otsuka, M.; Koshima, H.; Soai, K. Chem. Commun. (Camb). 2006, 2, 1869–1871. 49. Kawasaki, T.; Harada, Y.; Suzuki, K.; Tobita, T.; Florini, N.; Pályi, G.; Soai, K. Org. Lett. 2008, 10, 4085–4088. 45

50. Kawasaki, T.; Nakaoda, M.; Kaito, N.; Sasagawa, T.; Soai, K. Origins Life Evol. Biospheres 2010, 40, 65–78. 51. Carter, D. J.; Rohl, A. L.; Shtukenberg, A.; Bian, S.; Hu, C.; Baylon, L.; Kahr, B.; Mineki, H.; Abe, K.; Kawasaki, T.; Soai, K. Cryst. Growth Des. 2012, 12, 2138–2145. 52. Kawasaki, T.; Uchida, M.; Kaimori, Y.; Sasagawa, T.; Matsumoto, A.; Soai, K. Chem. Lett. 2013, 42, 711–713. 53. Matsumoto, A.; Ide, T.; Kaimori, Y.; Fujiwara, S.; Soai, K. Chem. Lett. 2015, 44, 688–690. 54. Barabás, B.; Caglioti, L.; Micskei, K.; Zucchi, C.; Pályi, G. Origins Life Evol. Biospheres 2008, 38, 317–327. 55. Kawasaki, T.; Shimizu, M.; Nishiyama, D.; Ito, M.; Ozawa, H.; Soai, K. Chem. Commun. 2009, 4396–4398. 56. Kawasaki, T.; Matsumura, Y.; Tsutsumi, T.; Suzuki, K.; Ito, M.; Soai, K. Science 2009, 324, 492–495. 57. Kawasaki, T.; Okano, Y.; Suzuki, E.; Takano, S.; Oji, S.; Soai, K. Angew. Chem., Int. Ed. 2011, 50, 8131–8133. 58. Matsumoto, A.; Oji, S.; Takano, S.; Tada, K.; Kawasaki, T.; Soai, K. Org. Biomol. Chem. 2013, 11, 2928–2931. 59. Matsumoto, A.; Ozaki, H.; Harada, S.; Tada, K.; Ayugase, T.; Ozawa, H.; Kawasaki, T.; Soai, K. Angew. Chem., Int. Ed. 2016, 55, 15246–15249. 60. Caglioti, L.; Barabas, B.; Faglioni, F.; Florini, N.; Lazzeretti, P.; Maioli, M.; Micskei, K.; Rabai, G.; Taddei, F.; Zucchi, C.; Pályi, G. Chem. Today 2008, 26 (suppl.), 30–32. 61. Soai, K.; Shibata, T.; Kowata, Y. Jpn. Kokai Tokkyo Koho JP 19960121140 19960418, 1996. An abstract is readily available as JPH09268179 from the European Patent Office (http://worldwide.espacenet.com). 62. Soai, K.; Sato, I.; Shibata, T.; Komiya, S.; Hayashi, M.; Matsueda, Y.; Imamura, H.; Hayase, T.; Morioka, H.; Tabira, H.; Yamamoto, J.; Kowata, Y. Tetrahedron: Asymmetry 2003, 14, 185–188. 63. Kawasaki, T.; Suzuki, K.; Shimizu, M.; Ishikawa, K.; Soai, K. Chirality 2006, 18, 479–482. 64. Suzuki, K.; Hatase, K.; Nishiyama, D.; Kawasaki, T.; Soai, K. J. Syst. Chem. 2010, 1, 5. 65. Sato, I.; Omiya, D.; Igarashi, H.; Kato, K.; Ogi, Y.; Tsukiyama, K.; Soai, K. 66. Sato, I.; Omiya, D.; Tsukiyama, K.; Ogi, Y.; Soai, K. Tetrahedron: Asymmetry 2001, 12, 1965–1969. 67. Blackmond, D. G.; McMillan, C. R.; Ramdeehul, S.; Schorm, A.; Brown, J. M. J. Am. Chem. Soc. 2001, 123 (41), 10103–10104. 68. Gridnev, I. D.; Serafimov, J. M.; Brown, J. M. Angew. Chem., Int. Ed. 2004, 43, 4884–4887. 69. Islas, J. R.; Lavabre, D.; Grevy, J.-M.; Lamoneda, R. H.; Cabrera, H. R.; Micheau, J.-C.; Buhse, T. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 13743–13748. 70. Lavabre, D.; Micheau, J.-C.; Rivera Islas, J.; Buhse, T. Top. Curr. Chem. 2008, 284, 67–96. 71. Saito, Y.; Hyuga, H. Top. Curr. Chem. 2008, 284, 97–118. 46

72. Lente, G. J. Phys. Chem. A 2005, 109, 11058–11063. 73. Micskei, K.; Rábai, G.; Gál, E.; Caglioti, L.; Pályi, G. J. Phys. Chem. B 2008, 112, 9196–9200. 74. Crusats, J.; Hochberg, D.; Moyano, A.; Ribó, J. M. ChemPhysChem 2009, 10, 2123–2131. 75. Barabas, B.; Caglioti, L.; Micskei, K.; Palyi, G. Bull. Chem. Soc. Jpn. 2009, 82, 1372–1376. 76. Klankermayer, J. J.; Gridnev, I. D.; Brown, J. M. Chem. Commun. 2007, 3151–3153. 77. Schiaffino, L.; Ercolani, G. Angew. Chem., Int. Ed. 2008, 47, 6832–6835. 78. Ercolani, G.; Schiaffino, L. J. Org. Chem. 2011, 76, 2619–2626. 79. Quaranta, M.; Gehring, T.; Odell, B.; Brown, J. M.; Blackmond, D. G. J. Am. Chem. Soc. 2010, 132, 15104–15107. 80. Dóka, É.; Lente, G. J. Am. Chem. Soc. 2011, 133, 17878–17881. 81. Gridnev, I. D.; Vorobiev, A. K. ACS Catal. 2012, 2, 2137–2149. 82. Gehring, T.; Quaranta, M.; Odell, B.; Blackmond, D. G.; Brown, J. M. Angew. Chem., Int. Ed. 2012, 51, 9539–9542. 83. Gridnev, I. D.; Vorobiev, A. K. Bull. Chem. Soc. Jpn. 2015, 88, 333–340. 84. Barabás, B.; Zucchi, C.; Maioli, M.; Micskei, K.; Pályi, G. J. Mol. Model. 2015, 21, 33. 85. Matsumoto, A.; Abe, T.; Hara, A.; Tobita, T.; Sasagawa, T.; Kawasaki, T.; Soai, K. Angew. Chem., Int. Ed. 2015, 54, 15218–15221. 86. Matsumoto, A.; Fujiwara, S.; Abe, T.; Hara, A.; Tobita, T.; Sasagawa, T.; Kawasaki, T.; Soai, K. Bull. Chem. Soc. Jpn. 2016, 89, 1170–1177.

47

Chapter 4

Interplay between Organocatalysis and Multicomponent Reactions in Stereoselective Synthesis Daniel G. Rivera1,* and Márcio W. Paixão2 1Center

for Natural Products Research, Faculty of Chemistry, University of Havana, Zapata y G, 10400, La Habana, Cuba 2Departamento de Química, Universidade Federal de São Carlos, São Carlos, SP, 13565-905, Brazil *E-mail: [email protected]

Multicomponent reactions (MCRs) serve as powerful approaches for the synthesis of bioactive heterocyclic compounds and natural products. Owing to their high chemical efficiency, atom economy and diversity-generating character, these processes have been intensively exploited in drug discovery and chemical biology programs. As chiral ketones and aldehydes are key components in stereocontrolled MCRs, methods for the asymmetric functionalization of carbonyl compounds are relevant for the development of novel stereoselective multicomponent approaches. In the last decade, organocatalysis has been successful in the α-, β-, γand even ε-asymmetric functionalization of carbonyls, thus showing promise for its combination with MCRs in the pursuit of highly stereoselective cascade sequences. This chapter highlights a recent international endeavor that combines the diversity and complexity-generating character of MCRs with the high stereoselection of organocatalysis for the synthesis of enantiomerically pure compounds. The reaction sequences comprise the asymmetric aminocatalytic functionalization of α,β-unsaturated aldehydes followed by isocyanide-MCRs with such oxo-components as chiral inputs. Methods described herein provide a convergent and stereoselective

© 2017 American Chemical Society

way of producing natural product-like compounds such as hydroquinolines, chromenes, epoxy- and depsi-peptides.

Introduction Enantiomerically pure or enriched small organic molecules are key starting materials in organic synthesis. Chiral building blocks are in high demand in the total synthesis of complex natural products and in drug discovery and development, as well as in the production of pesticides, fragrances and advanced materials (1). Asymmetric organocatalysis is considered a cutting-edge tool for the synthesis of bioactive molecules and chiral building blocks, in most cases achieving similar success as metal-based asymmetric catalysis (2–6) Similarly, multicomponent reactions (MCRs) (7) serve as powerful approaches for the synthesis of bioactive heterocyclic compounds (8–10) and natural products (11). MCRs are procedures wherein more than two building blocks react in one pot to afford a structure including moieties from all reactants (7, 12). Among the MCRs, those employing isocyanides are generally considered as type II MCRs (13), as they involve a sequence of reversible mono- and bimolecular events that proceed sequentially until an irreversible step traps the final product, i.e., the exothermic oxidation of the CII of isocyanides to CIV. The divalent carbon atom of isocyanides shows a particular reactivity toward nucleophiles and electrophiles, a feature – that along with carbenes – renders isocyanides as exceptional chemical species. In general, MCRs are considered a subclass of domino reactions, as all transformations are performed in one pot under similar reaction conditions and in a time-resolved manner (14). Accordingly, several chemical bonds are formed with high chemical efficiency, thus allowing the generation of high levels of structural diversity and complexity. Such intrinsic characteristics of MCRs may be further improved by combination with pre- and post-MCR modifications, although most examples have been described for the latter approach (15, 16). In the past few years, various research groups have recognized the potential of combining the capacity of organocatalysis for generating enantiomerically enriched compounds with the synthetic power of MCRs in the diversification of such chiral pools (17–21). This strategy is conceptually different from both organocatalytic asymmetric MCRs (22–24) and organocatalytic multicomponent domino reactions (25, 26). In these latter approaches, the chiral small molecule directly catalyses the multicomponent process, while the first approache relies on the organocatalytic asymmetric functionalization of carbonyl compounds that are subsequently used in diastereoselective MCRs, either in one-pot protocols or not. As organocatalysis has proven great efficiency in the α-, β- (27–30), γ- and ε-asymmetric (31, 32) functionalization of ketones and aldehydes, its combination with the available repertoire of MCRs – incorporating such carbonyl components as chiral inputs – provides a manifold of synthetic possibilities. A recent review by Banfi et. al. (33) covered the literature reports up to 2014 on MCRs including enantioenriched components arising from both organocatalytic and biocatalytic approaches. This chapter highlights reports 50

from our own laboratories dealing with the design of approaches comprising the asymmetric aminocatalytic functionalization of α,β-unsaturated aldehydes followed by isocyanide-MCRs. We illustrate how this combined strategy enables the stereoselective synthesis of complex scaffolds, including natural product-like compounds such as epoxy-peptidomimetics, hydroquinolines, piperidines, chromenes and cyclic depsipeptides. As part of an international endeavor to honour the contribution of Prof. Ernest Eliel to Organic Chemistry, we hope not only to encourage further progress in the fields of MCRs and organocatalysis, but especially to advocate their combination as a powerful tool in the stereoselective synthesis of complex enantioenriched molecular scaffolds.

Results and Discussion Aminocatalytic Epoxidation of α,β-Unsaturated Aldehydes Followed by Passerini Three-Component Reaction Among the several organocatalytic approaches known for the asymmetric functionalization of carbonyl compounds, our initial effort focused on the development of an eco-friendly, one-pot aminocatalytic epoxidation α,β-unsaturated aldehydes followed by a Passerini three-component reaction (3CR) to furnish epoxy-peptidomimetics (34). For this, we implemented a modification of a protocol developed by Jørgensen’s group (35) for the synthesis of enantioenriched epoxy-aldehydes using a diarylprolinol silyl ether catalyst. As depicted in scheme 1, after a series of parallel experiments with varied organocatalysts, it was found that aryl-modified diarylprolinol silyl ether 2 is an effective catalyst for the epoxidation of a variety of α,β-unsaturated aldehydes in aqueous conditions. The Passerini-3CR is the condensation of an aldehyde or ketone, a carboxylic acid and an isocyanide to produce an α-acyloxy carboxamide skeleton (36). This type of scaffold is of frequent occurrence in natural products such as the depsipeptides, and in this case is combined with an epoxide functionality at the vicinal position. Importantly, the procedure could be implemented in one-pot by carrying out the initial organocatalytic step in the solvent mixture EtOH/H2O 3:1 (v/v) followed by addition of the carboxylic acid and the isocyanide. Scheme 1 illustrates just a selection of the epoxy-peptidomimetics obtained by this protocol, in which the substrate scope includes the use of aliphatic and aromatic conjugated aldehydes, as well as N-protected amino acids, aromatic and aliphatic carboxylic acids. In all cases, final compounds were obtained in good to excellent yields in a diastereoselectivity of about 3:1. However, the poor diastereoselection is due to the multicomponent step, as it was initially proven that the organocatalytic step rendered the epoxy-aldehydes in excellent enantio and diastereoselectivity. On the other hand, the Passerini-3CR turned out to be poorly diastereoselective, despite the fact that it was undertaken with a chiral aldehyde functionalized at α and β positions. Overall, this work comprised the first report of a one-pot organocatalytic multicomponent reaction sequence based on an asymmetric epoxidation reaction and a Passerini-3CR. 51

Scheme 1. One-pot synthesis of epoxy-peptidomimetics by an eco-friendly aminocatalytic epoxidation of α,β-unsaturated aldehydes followed by the Passerini-3CR.

Aminocatalytic Conjugate Addition of 1,3-Cycloalkanediones to α,β-Unsaturated Aldehydes Followed by a Novel Isocyanide-MCR This second section describes the development of a highly stereoselective one-pot approach for the synthesis of complex molecular hybrids incorporating moieties such as hydroquinolines, chromenes, piperidines, peptides, lipids and glycosides. The strategy involved the implementation of an asymmetric organocatalytic conjugate addition of dicarbonyl compounds to α,β-unsaturated aldehydes, followed by an intramolecular isocyanide-based MCR including a chiral bifunctional intermediate, an amine and an isocyanide component. Owing to low diastereoselectivity of intermolecular isocyanide-based MCRs, we aimed at utilizing an intramolecular MCR, as these typically provide better stereocontrol as compared to their intermolecular versions (24, 33). Similarly, the chosen organocatalytic process should provide an enantiomerically enriched chiral intermediate bearing a pair of reactive functionalities suitable to exert stereocontrol during the intramolecular MCR. As shown in Scheme 2, we implemented an organocatalytic procedure developed independently by the groups of Rueping (37, 38) and Jørgensen (39) followed by a novel isocyanide-MCR including a chiral bifunctional substrate (40). The initial cascade process comprises the asymmetric conjugate addition of dimedone (5) to 2-pentenal (6) catalyzed by diarylprolinol silyl ether 7, followed by acetalization to cyclic hemiacetal 8. This intermediate fulfills the requisite of being a biologically relevant chiral substrate as well as having two functionalities suitable for the MCRs, i.e. an aldehyde and a conjugated enol (41). 52

Scheme 2. One-pot organocatalytic conjugate addition/isocyanide-MCR sequence to 2-amido-hydroquinolin-6-ones. Since the multicomponent step did not work well in CH2Cl2, we sought to implement the whole sequence in one pot by addition of trifluoroethanol (TFE) after formation of organocatalytic product 8. In this way, the second step is carry out in the solvent mixture CH2Cl2/TFE (1:1, v/v) leading to the 2-amido-hydroquinolin-6-ones 10a-j. It must be noticed that the presence of 10 mol% of organocatalyst 7 (a secondary amine) did not interfere in the isocyanide-MCR, as no product including this fragment was detected. A key feature of this approach is the different stereochemical outcome derived from variation of primary amine. As shown in Scheme 2, the use of benzyl amine led to almost no diastereoselectivity in the multicomponent step, while the chiral (S)-α-methylbenzyl amine provided the product 10b in enantiopure form with an excellent diastereoselectivity (>99:1). The relative configuration of hydroquinolin-6-ones 10a,b was determined by NMR analysis, proving the cis configuration of the two substituents. After the scope of this reaction was addressed, the focus was posed on the variation of the four different components to produce structurally diverse tetrahydroquinoline scaffolds 10c-j (scheme 2). Thus, the 1,3-dicarbonyl and aldehyde was varied during the initial organocatalytic step while the amine and isocyanide stayed the same in the isocyanide-MCR. As before, the one-pot processes were performed without isolation of intermediate 8, but the subsequent MCR was simply carried out by addition of TFE, the amine and isocyanide components immediately after completion of the organocatalytic step. The stereoselectivity of the multicomponent sequence leading to hydroquinolinones 10c-j proved once more to be excellent using chiral amines (α-MeBn and amino acids), while unbranched n-butyl and benzyl amines provided poor stereocontrol. The use of either S or R-α-methylbenzyl amine as well as either D- or L-amino acid methyl esters provided the same stereo-differentiation to the cis isomers. Intriguingly, an experiment with bulky achiral amines such as cyclohexyl and t-butyl amines resulted in moderate diastereoselectivity for 10h but excellent one for 10i, proving also achiral amines with bulky substituents at 53

α-position can induce stereoselection in the MCR. To expand the scope of this diversity-oriented approach, we sought to implement the one-pot protocol for the synthesis of piperidinocoumarine hybrids substituted at positions 1, 2 and 4. Scheme 3 illustrates the one-pot approach leading to piperidinocoumarine hybrids 14 utilizing enantiomerically enriched chromenone 11 and piranocoumarine 12. As before, enantiomerically pure hybrids were produced in an excellent diastereomeric ratio, with the cis isomers as major products as proven by NMR. Finally, we were able to generate even higher structural complexity in a one-pot organocatalytic/MCR protocol with the incorporation of natural product fragments of peptidic, lipidic and saccharidic nature into piperidine-based hybrid architectures. Scheme 4 depicts the implementation of the multicomponent step with very complex substrates such as di and tripeptidic isocyanide and glucosyl amine. These substrates proved to react readily with the chiral cyclic hemiacetals arising from the organocatalytic step, showing that the combination of the two processes encompasses useful complexity-generation characteristics.

Scheme 3. One-pot organocatalytic conjugate addition/isocyanide-MCR sequence to piperidinocoumarine hybrids.

Scheme 4. One-pot synthesis of hydroquinolin-6-one and piperidinocoumarine hybrids by an organocatalytic conjugate addition/isocyanide-MCR sequence. 54

Aminocatalytic Conjugate Addition of Nitroethanol to α,β-Unsaturated Aldehydes Followed by the Ugi Multicomponent Reaction The Ugi four-component reaction (42) – i.e., the condensation of a primary amine, a carboxylic acid, an aldehyde/ketone and an isocyanide – is a powerful synthetic tool to produce N-substituted peptides and peptidomimetics (13, 43–45). An important variation of the Ugi reaction is the so-called Ugi five-center four-component reaction (Ugi-5C-4CR) developed in 1996 (46). The Ugi’s concept behind this remarkable process was the utilization of α-amino acids as bifunctional scaffolds, which lead to a six-membered ring α-adducts that evade the classic Mumm rearrangement enabling the attack of the solvent methanol. A diastereoselective version of this reaction employing Lewis acid catalysts has been also reported (47). A modification of this reaction was developed by Ugi himself using trifunctional scaffolds like the amino acid lysine, thus leading to cyclic scaffold (48). In this new variant, named Ugi five-center three-component reaction (Ugi-5C-3CR), the α-adduct evolves through an intramolecular acylation of the side chain amino group leading to an α-amino-ε-lactam derivative. α-Homoserine has also been used as trifunctional building block for this reaction by Kim and coworkers during the synthesis of α-aminobutyrolactones (49). In this report, the intramolecular acylation of the α-homoserine primary alcohol is the key step in this new variant of the Ugi-5C-3CR. Interestingly, an external amine can be used as external nucleophile instead of a nuclephilic alcohol (50). Recently, unprotected carbohydrates and α-amino acids were employed as chiral bifunctional substrates of this type of isocyanide-MCR to enable the diastereoselective formation of novel cyclic glycopeptidomimetics (51). This third section describes a novel approach consisting of the preparation of chiral 4,5-disubstituted 2-hydroxytetrahydropyrans and their utilization as inputs for the Ugi-multicomponent synthesis of cyclic depsipeptide mimics. The overall strategy involves an initial asymmetric organocatalytic conjugate addition of nitroethanol to α,β-unsaturated aldehydes, followed by an Ugi-5C-3CR including the chiral cyclic hemiacetal and a variety of amino acids and isocyanides. The rationale of using the 2-hydroxytetrahydropyran scaffold lies at its bifunctional character, as the aldehyde group may react with the other Ugi components to form the α-adduct, while the appendage primary hydroxyl group undertakes the intramolecular acylation leading to a new type of depsipeptide mimic (Scheme 5). An efficient approach previously developed by Hayashi and co-workers (52) was chosen as initial organocatalytic step towards cyclic hemiacetals. The process comprised the asymmetric conjugate addition of nitroethanol to α,β-unsaturated aldehydes catalyzed by a diphenylprolinol silyl ether 7, which after cyclization rendered the enantioenriched 4,5-disubstituted 2-hydroxy-tetrahydropyrans 20. After having access to a pool of enantioenriched 4,5-disubstituted 2-hydroxy-tetrahydropyrans (20), we turned to the synthesis of the cyclic depsipeptide mimics 21 by the Ugi-5C-3CR with such chiral hemiacetals. Isocyanide-MCRs have been used to produce macrocyclic lactams resembling naturally occurring depsipeptide (53), but in most cases the MCR is not responsible for the ring closure step and it has never been used to produce 55

medium-size lactone rings. As depicted in scheme 5, hemiacetal 20 effectively reacted with a variety of amino acids (Val, Leu, Phe, PhGly, Met, His and Trp) and isocyanides (including peptidic, lipidic and glycosidic ones) at room temperature to afford nine-membered ring depsipeptides 21a-q. These compounds were produced with low diastereoselectivity, meaning that the stereogenic centers at the hemiacetal exert no stereocontrol over the multicomponent reaction. As we aimed at improving the diastereoselectivity of this reaction, parallel experiments using various Lewis acid catalysts were carried out, as a previous report had shown that such catalysts enhance the diastereoselection of the classic Ugi-5C-3CR between amino acids, isocyanides and aromatic aldehydes (47). However, after several attempts and screening of conditions, there was no improvement in the diastereoselectivity of this system.

Scheme 5. Multicomponent synthesis of cyclic depsipeptides by an organocatalytic conjugate addition/Ugi-5C-3CR sequence.

Conclusions Herein we have covered examples from our laboratories showing the development of stereoselective – eventually one-pot – sequences leading to complex hybrid molecules. These approaches enable the incorporation of different molecular fragments into a single skeleton, and with very low synthetic cost. Overall, the reports confirm that the asymmetric aminocatalytic functionalization of carbonyl compounds is an effective pre-MCR process capable of providing enantiomerically enriched building blocks for subsequent multicomponent diversification. Three isocyanides-MCRs were successfully employed after organocatalytic funcionalization of aldehydes: a) the classic Passerini-3CR, b) 56

a new three-component reaction of cyclic hemiacetals (derived from conjugated enols) with amines and isocyanides and c) a variant of the Ugi-5C-3CR of cyclic hemiacetals with unprotected amino acids and isocyanides. The versatility of this diversity-oriented strategy relies on the vast number of organocatalytic steps capable of producing chiral carbonyl compounds to be next reacted with other components for generating high levels of molecular complexity. We envisage that other iminium, enamine and related organocatalytic processes may be combined with varied MCRs, hence expanding the repertoire of stereoselective synthesis.

Acknowledgments We are grateful to CNPq, FAPESP (14/50249-8 and 15/17141-1) and CAPES (CAPES-MES/Cuba Program) for financial support to the projects herein reviewed.

References Carreira, E. M.; Kvaerno. L. Classics in Stereoselective Synthesis; WileyVCH: Weinheim, 2009. 2. List, B. The ying and yang of asymmetric aminocatalysis. Chem. Commun. 2006, 819–824. 3. Beeson, T. D.; Mastracchio, A.; Hong, J.-B.; Ashton, K.; MacMillan, D. W. C. Enantioselective Organocatalysis Using SOMO Activation. Science 2007, 316, 582–585. 4. Dalko, P. I.; Moisan, L. In the Golden Age of Organocatalysis. Angew. Chem., Int. Ed. 2004, 43, 5138–5175. 5. Jensen, K. L.; Dickmeiss, G.; Jiang, H.; Albrecht, L.; Jørgensen, K. A. The Diarylprolinol Silyl Ether System: A General Organocatalyst. Acc. Chem. Res. 2012, 45, 248–264. 6. Narayanaperumal, S.; Rivera, D. G.; Silva, R. C.; Paixão, M. W. Terpene-Derived Bifunctional Thioureas in Asymmetric Organocatalysis. ChemCatChem 2013, 5, 2756–2773. 7. Multicomponent Reactions; Zhu, J., Bienyamé, H., Eds.; Wiley-VCH: Weinheim, 2005. 8. Rotstein, B. H.; Zaretsky, S.; Rai, V.; Yudin, A. K. Small Heterocycles in Multicomponent Reactions. Chem. Rev. 2014, 114, 8323–8359. 9. Ruijter, E.; Scheffelaar, R.; Orru, R. V. A. Multicomponent Reaction Design in the Quest for Molecular Complexity and Diversity. Angew. Chem., Int. Ed. 2011, 50, 6234–6246. 10. Wessjohann, L. A.; Rhoden, C. R. B.; Rivera, D. G.; Vercillo, O. E. Cyclic Peptidomimetics and Pseudopeptides from Multicomponent Reactions. Top. Heterocycl. Chem. 2010, 23, 199–226. 11. Touré, B. B.; Hall, D. G. Natural Product Synthesis Using Multicomponent Reaction Strategies. Chem. Rev. 2009, 109, 4439–4486. 12. Brauch, S.; van Berkel, S. S.; Westermann, B. Higher-order multicomponent reactions: beyond four reactants. Chem. Soc. Rev. 2013, 42, 4948–4963. 1.

57

13. Dömling, A.; Ugi, I. Multicomponent Reactions with Isocyanides. Angew. Chem., Int. Ed. 2000, 39, 3168–3210. 14. Tietze, L. F.; Brasche, G.; Gericke, K. M. Domino Reactions in Organic Synthesis; Wiley-VCH: Weinheim, 2006. 15. Kumar, A.; Vachhani, D. D.; Modha, S. G.; Sharma, S. K.; Parmar, V. S.; Van der Eycken, E. V. Post-Ugi gold-catalyzed diastereoselective domino cyclization for the synthesis of diversely substituted spiroindolines. Beilstein J. Org. Chem. 2013, 9, 2097–2102. 16. Marcaccini, S.; Torroba, T. In Multicomponent Reactions; Zhu, J., Bienyamé, H., Eds.; Wiley-VCH: Weinheim, 2005; pp 33−75. 17. Moni, L.; Banfi, L.; Basso, A.; Galatini, A.; Spallarossa, M.; Riva, R. Enantio- and diastereoselective synthesis of highly substituted benzazepines by a multicomponent strategy coupled with organocatalytic and enzymatic procedures. J. Org. Chem. 2014, 79, 339–351. 18. Morana, F.; Basso, A.; Riva, R.; Rocca, V.; Banfi, L. The homo-PADAM Protocol: Stereoselective and Operationally Simple Synthesis of α-Oxo- or α-Hydroxy-γ-acylaminoamides and Chromanes. Chem. Eur. J. 2013, 19, 4563–4569. 19. Albrecht, L.; Jiang, H.; Jørgensen, K. A. A Simple Recipe for Sophisticated Cocktails: Organocatalytic One-Pot Reactions—Concept, Nomenclature, and Future Perspectives. Angew. Chem., Int. Ed. 2011, 50, 8492–8509. 20. Riguet, E. Enantioselective Organocatalytic Friedel-Crafts Alkylation Reaction of Indoles with 5-Hydroxyfuran-2(5H)-one: Access to Chiral γ-Lactones and γ-Lactams via a Ugi 4-Center 3-Component Reaction. J. Org. Chem. 2011, 76, 8143–8150. 21. Umbreen, S.; Brockhaus, M.; Ehrenberg, H.; Schmidt, B. Norstatines from Aldehydes by Sequential Organocatalytic α-Amination and Passerini Reaction. Eur. J. Org. Chem. 2006, 4585–4595. 22. van Berkel, S. S.; Bögels, B. G. M.; Wijdeven, M. A.; Westermann, B.; Rutjes, F. P. J. T. Recent Advances in Asymmetric Isocyanide-Based Multicomponent Reactions. Eur. J. Org. Chem. 2012, 3543–3559. 23. de Graaff, C.; Ruijter, E.; Orru, R. V. A. Recent developments in asymmetric multicomponent reactions. Chem. Soc. Rev. 2012, 41, 3969–4010. 24. Ramón, D. J.; Yus, M. Asymmetric Multicomponent Reactions (AMCRs): The New Frontier. Angew. Chem., Int. Ed. 2005, 44, 1602. 25. Enders, D.; Grondal, C.; Hüttl, M. R. M. Asymmetric Organocatalytic Domino Reactions. Angew. Chem., Int. Ed. 2007, 46, 1570–1581. 26. Pellissier, H. Stereocontrolled Domino Reactions. Chem. Rev. 2013, 113, 442–524. 27. Melchiore, P.; Marigo, M.; Carlone, A.; Bartole, G. Asymmetric Aminocatalysis – Gold Rush in Organic Chemistry. Angew. Chem., Int. Ed. 2008, 47, 6138–6171. 28. Mukherjee, S.; Yang, J. W.; Hoffmann, S.; List, B. Asymmetric Enamine Catalysis. Chem. Rev. 2007, 107, 5471–5569. 29. Erkkilä, A.; Majander, I.; Pihko, P. M. Iminium Catalysis. Chem. Rev. 2007, 107, 5416–5470. 58

30. Trost, B. M.; Brindle, C. S. The direct catalytic asymmetric aldol reaction. Chem. Soc. Rev. 2010, 39, 1600–1632. 31. Nielsen, M.; Worgull, D.; Zweifel, T.; Gschwend, B.; Bertelsen, S.; Jørgensen, K. A. Mechanisms in aminocatalysis. Chem. Commun. 2011, 47, 632–649. 32. Melchiorre, P. Cinchona-based Primary Amine Catalysis in the Asymmetric Functionalization of Carbonyl Compounds. Angew. Chem., Int. Ed. 2012, 51, 9748–9770. 33. Banfi, L.; Basso, A.; Moni, L.; Riva, R. The Alternative Route to Enantiopure Multicomponent Reaction Products: Biocatalytic or Organocatalytic Enantioselective Production of Inputs for Multicomponent Reactions. Eur. J. Org. Chem. 2014, 2005–2015. 34. Deobald, A. M.; Correa, A. G.; Rivera, D. G.; Paixão, M. W. Organocatalytic asymmetric epoxidation and tandem epoxidation/Passerini reaction under eco-friendly reaction conditions. Org. Biomol. Chem. 2012, 10, 7681–7684. 35. Marigo, M.; Franzén, J.; Poulsen, T. B.; Zhuang, W.; Jørgensen, K. A. Asymmetric Organocatalytic Epoxidation of α,β-Unsaturated Aldehydes with Hydrogen Peroxide. J. Am. Chem. Soc. 2005, 127, 6964–6289. 36. Banfi, L.; Riva, R. The Passerini Reaction. Org. React. 2005, 65, 1–140. 37. Rueping, M.; Sugiono, E.; Merino, E. Asymmetric Iminium Ion Catalysis: An Efficient Enantioselective Synthesis of Pyranonaphthoquinones and βLapachones. Angew. Chem., Int. Ed. 2008, 47, 3046–3049. 38. Rueping, M.; Sugiono, E.; Merino, E. Asymmetric Organocatalysis: An Efficient Enantioselective Access to Benzopyranes and Chromenes. Chem. Eur. J. 2008, 14, 6329–6332. 39. Franke, P. T.; Richter, B.; Jørgensen, K. A. Organocatalytic Asymmetric Synthesis of Functionalized 3,4-Dihydropyran Derivatives. Chem. Eur. J. 2008, 14, 6317–6321. 40. Echemendía, R.; de la Torre, A. F.; Monteiro, J. L.; Pila, M.; Corrêa, A. G.; Westermann, B.; Rivera, D. G.; Paixão, M. W. Highly Stereoselective Synthesis of Natural Product-like Hybrids by an Organocatalytic/ Multicomponent Reaction Sequence. Angew. Chem., Int. Ed. 2015, 54, 7621–7625. 41. This type of isocyanide-MCR has been previously reported in an intermolecular manner, see: Castellano, T. G.; Neo, A. G.; Marcaccini, S.; Marcos, C. F. Enols as Feasible Acid Components in the Ugi Condensation. Org. Lett. 2012, 14, 6218–6221. 42. Marcaccini, S.; Torroba, T. The use of the Ugi four-component condensation. Nat. Protocols 2007, 2, 632–639. 43. Dömling, A. Recent Developments in Isocyanide Based Multicomponent Reactions in Applied Chemistry. Chem. Rev. 2006, 106, 17–89. 44. Wessjohann, L. A.; Rivera, D. G.; Vercillo, O. E. Multiple Multicomponent Macrocyclizations (MiBs): A Strategic Development Towards Macrocycle Diversity. Chem. Rev. 2009, 109, 796–814. 45. Dömling, A.; Wang, W.; Wang, K. Chemistry and Biology Of Multicomponent Reactions. Chem. Rev. 2012, 112, 3083–3136. 59

46. Demharter, A.; Hörl, W.; Herdtweck, E.; Ugi, I. Synthesis of Chiral 1,l′-Iminodicarboxylic Acid Derivatives from a-Amino Acids, Aldehydes, Isocyanides, and Alcohols by the Diastereoselective Five-CenterFour-Component Reaction. Angew. Chem., Int. Ed. Engl. 1996, 35, 173–175. 47. Godet, T.; Bonvin, Y.; Vincent, G.; Merle, D.; Thozet, A.; Ciufolini, M. A. Titanium catalysis in the Ugi reaction of α-amino acids with aromatic aldehydes. Org. Lett. 2004, 6, 3281–3284. 48. Ugi, I.; Demharter, A.; Hörl, W.; Schmidt, T. Ugi Reactions with Trifunctional α-Amino Acids, Aldehydes, Isocyanides and Alcohols. Tetrahedron 1996, 52, 11657–11664. 49. Park, S. J.; Keum, G.; Kang, S. B.; Koh, H. Y.; Kim, Y.; Lee, D. H. A Facile Synthesis N-Carbamoylmethyl-α-aminobutyrolactones by the Ugi Multicomponent Condensation Reaction. Tetrahedron Lett. 1998, 39, 7109–7110. 50. Khoury, K.; Sinha, M. K.; Nagashima, T.; Herdtweck, E.; Dömling, A. Efficient Assembly of Iminodicarboxamides by a “Truly” Four-Component Reaction. Angew. Chem., Int. Ed. 2012, 51, 10280–10283. 51. Voigt, V.; Mahrwald, R. Multicomponent Cascade Reactions of Unprotected Carbohydrates and Amino Acids. Org. Lett. 2015, 17, 2606–2609. 52. Gotoh, H.; Okamura, D.; Ishikawa, H.; Hayashi, Y. Diphenylprolinol Silyl Ether as a Catalyst in an Asymmetric, Catalytic, and Direct Michael Reaction of Nitroethanol with α,β-Unsaturated Aldehydes. Org. Lett. 2009, 11, 4056–4059. 53. León, F.; Rivera, D. G.; Wessjohann, L. A. Multiple Multicomponent Macrocyclizations Including Bifunctional Building Blocks (MiBs) Based on Staudinger and Passerini Three-Component Reactions. J. Org. Chem. 2008, 73, 1762–1767.

60

Chapter 5

Synthetic Approaches to the Stereochemically Complex Antitumor Drug Ecteinascidin-743: A Marine Natural Product by the Name Yondelis® or Trabectidin Plato A. Magriotis* Laboratory of Pharmaceutical Chemistry, Department of Pharmacy, University of Patras, Rio, 26504, Greece *E-mail: [email protected]

Ecteinascidin-743 (Et-743) isolated from the Caribbean tunicate Ecteinascidia turbinate, is arguably the most potent cytotoxin known as indicated by its evaluation against the National Cancer Institute’s human in vitro cell line panel including melanoma, non-small-cell lung, ovarian, renal, prostate, and breast cancers, demonstrating potencies ranging from 1 pM to 10 nM. In fact, the antiproliferative activity of Et-743 is greater than that of Taxol, camptothecin, adriamycin, mitomycin C, cisplatin, bleomycin, and etoposide by 1-3 orders of magnitude, propelling Yondelis or Trabectidin to become the first marine anticancer drug to be approved, intact, in the European Union (EU, October 2007) and more recently in the US as a first-line treatment for soft tissue sarcomas. It is also undergoing clinical trials for breast, prostate, ovarian, and pediatric sarcomas in other countries. The complexity of molecular architecture, the remarkable biological activities, and the restricted natural availability (1.0 g from about 1.0 ton of tunicate) made Et-743 an exceedingly attractive synthetic target for total synthesis. The successful approaches toward the total synthesis of Et-743 are reviewed in this article..

© 2017 American Chemical Society

Introduction Ecteinascidin-743 (1), isolated from the Caribbean tunicate Ecteinascidia turbinate (1), is probably the most potent cytotoxin known as indicated by its evaluation against the National Cancer Institute’s human in vitro cell line panel including melanoma, non-small-cell lung, ovarian, renal, prostate, and breast cancers, demonstrating potencies ranging from 1 pM to 10 nM (2). In fact, the antiproliferative activity of Et-743 is greater than that of Taxol, Camptothecin, Adriamycin, Mitomycin C, Cisplatin, Bleomycin, and Etoposide by 1-3 orders of magnitude (3, 4). Its proposed, unique mechanism of action involves binding to the N2 position of guanine in the minor groove demonstrating a preference for sequences containing 5′-PuGC and 5′PyGC motifs. Subsequent alkylation of DNA, via an iminium ion intermediate generated from an intramolecular acid-catalyzed activation and dehydration of the carbinolamine functionality present in the all important piperazine ring (box a, 1), induces a curvature of the DNA toward the major groove that ultimately disrupts the binding of transcription factors involved in cell proliferation (1, 5–7).

Another way of looking at Et-743 structurally is that it consists of three fused tetrahydroisoquinoline rings, two of which (subunits A and B) provide the framework for covalent interaction within the minor groove of the DNA double helix (8). The third ring (subunit C) protrudes from the DNA duplex and interacts with adjacent nuclear proteins, an interaction that is thought to account for the cytotoxicity of Et-743. Specifically, the cytotoxicity of Et-743 appears to be associated with the DNA repair mechanism. In vitro studies have demonstrated inhibition of transcription-dependent nucleotide excision repair pathways and therefore inhibition of cell cycle progression leading to p53-independent apoptosis. The transcription-coupled nucleotide excision repair process (TC-NER) involves recognition of DNA damage and recruitment of various nucleases at the site of DNA damage. At micromolar concentrations, Et-743 has been shown tο trap these nucleases in a malfunctioning nuclease-(Et-743)-DNA adduct complex and thereby inducing irreparable single-strand breaks in the DNA. This is supported by the fact that mammalian cell lines deficient in TC-NER show resistance to Et-743 (9). The complexity of molecular architecture, the remarkable biological activities, and the restricted natural availability (1.0 g from about 1.0 ton of 62

tunicate) made 1 an exceedingly attractive target for total synthesis. Three total syntheses and two formal total syntheses have been reported by Corey (10), Fukuyama (11), Zhu (4), Danishefsky (12), and Williams (13), respectively. More specifically, Corey and co-workers reported the first total synthesis of 1 in 36 steps and an overall yield of 0.72% (10). A second-generation synthesis improved the overall yield to 2.04% but still required 36 steps (14). Subsequently, Fukuyama and co-workers achieved a total synthesis of 1 in 50 steps and 0.56% overall yield (11). Later, Zhu and co-workers reported a 31 step synthesis in 1.7% overall yield (4), while Danishefsky and co-workers reported a formal total synthesis via a pentacyclic compound that intercepted a late-stage intermediate of Fukuyama’s route (vide infra (12)). Finally, Williams and Fishlock reported another formal total synthesis in still more than 30 steps (13). A second generation, more efficient, and robust synthesis of the natural product was reported four years ago by Fukuyama (15). Despite these advancements in total synthetic approaches to 1, the clinical supply of this complex drug is semisynthetically derived from natural cyanosafracin B obtained by fermentation as described by PharmaMar researchers (16), and shown in eq 1.

Notably, the original Corey and co-workers total synthesis has been applied to the described above semisynthesis of 1 from cyanosafracin B. The analysis and discussion of the successful approaches toward the total synthesis of 1 will be based on the construction of its central piperazine ring ((17), structure 1, box a; see also Scheme 2 in reference (17)) which not only comprises both tetrahydroisoquinoline subunits A and B but also is considered a privileged pharmacophore in medicinal chemistry since it is more than frequently found in biologically active compounds across a number of therapeutic areas (18). Actually, a recent survey of more than a thousand orally administered drugs showed that about 6% of these contained a piperazine fragment (19).

Total Syntheses of 1: Construction of the Piperazine Subunit and Its Surrogates The first enantioselective total synthesis of 1, achieved by Corey and co-workers, featured an internal Mannich bisannulation for the construction of the central piperazine ring as shown in the conversion of 4 to 5 (Scheme 1) which was effected by treatment of 4 with DIBAL, KF, and 20 equiv of trifluoromethanesulfonic acid in that order. Selective trifluoromethanesulfonation of the least hindered phenolic hydroxyl of 5 was followed by: (1) selective silylation of the primary hydroxyl, (2) protection of the remaining phenolic 63

group as the MOM ether, (3) double deallylative deprotection, (4) reductive N-methylation, and (5) Stille methylation to provide 6 in 83% yield and set up the stage for the installation of the ten-membered lactone ring and the spiro tetrahydroisoquinoline subunit C that will be described in the next section dealing with the completion of the total syntheses of 1.

Scheme 1. Piperazine construction in Corey’s total synthesis. Interestingly, 4 was prepared by the union of hοmochiral tetrahydroisoquinoline lactone 2 and α-amino aldehyde 3 both of which the chirality was set by a catalytic enantioselective hydrogenation, for example the reduction of 7 to furnish 8 in 97% yield and 96% ee (eq 2). The conversion of 8 to 2 demonstrates a method for control of stereochemistry in the tetrahydroisoquinoline series (10).

The second enantioselective total synthesis of 1, accomplished by Fukuyama and co-workers in Japan, proceeded through the piperazine surrogate 12 made by an intramolecular Heck reaction of the cyclic enamide 11 which was in turn obtained from diketopiperazine 10 that was assembled in short order by the powerful Ugi four-component condensation reaction of homochiral fragments 7 and 8 as well as isocyanide 9 and acetaldehyde as shown in Scheme 2 (11). The chirality of aryl glycinol derivative 7 was set by the development of novel methodology (20), whereas that of 8 was installed by a similar catalytic 64

enantioselective hydrogenation as shown in eq 2. Finally, conversion of 12 to the pentacyclic intermediate 13 was realized in sixteen conventional steps (11).

Scheme 2. Piperazine construction in Fukuyama’s total synthesis. The third total synthesis of 1, achieved by Zhu and co-workers in France (4), was centred on piperazine 21 (Scheme 3) that was diastereoselectively prepared from homochiral Garner’s aldehyde 14 and L-3-hydroxy-4-methoxy-5-methyl phenylalanol (15) which in turn was efficiently made in eight steps based on Corey’s enantioselective alkylation of N-(diphenylmethylene) glycine tert-butyl ester with a 2-methyl catechol derivative (21). Specifically, condensation of 14 and 15 provided, under optimized conditions shown in Scheme 3, the desired tetrahydroisoquinoline 16 as the only isolable product at the expense of other regio- (C-19 vs C-15, Et-743 numbering) and diastereoisomers. Interestingly, the stereochemistry at C-11 was controlled solely by the absolute configuration of amino alcohol 15 since condensation of 15 and (R)-14 gave also the C-11-C-13 cis diastereomer in excellent yield. It seems reasonable to assume, therefore, that both C-11 and C-13 substituents adopt pseudoequatorial positions leading to the observed cis selectivity after ring closure. The coupling of 16 and 17 (1:1 ratio) furnished products 18 and 19 in good yield and 1:3 ratio respectively. The observed 50% de can be explained by a SN1 mechanism via an ortho-quinone methide intermediate (Scheme 3). Accordingly, acetate 19 was protected, deprotected, and advanced to tricyclic piperazine 21 via an intramolecular Strecker reaction of oxidized primary alcohol 20, a transformation that resembles the intermolecular conversion of 2+3 to 4 in Corey’s synthesis (Scheme 1). 65

Scheme 3. Piperazine construction in Zhu’s total synthesis.

As mentioned above, the fourth reported total synthesis of 1 by Danishefsky and co-workers was typically a formal one since it defined the alcohol corresponding to acetate 13, an advanced intermediate in Fukuyama’s total synthesis of 1 (Scheme 2), as a milestone which was achieved by a novel vinylogous Pictet-Spengler cyclization employing an unusual o-hydroxystyrene moiety (vide infra (12)). Thus, coupling of tetrahydroisoquinoline 22, derived from Borchardt’s alcohol (12), and homochiral L-tyrosine derivative 23, the chirality of which was set by a catalytic enantioselective hydrogenation of the corresponding dehydro methyl ester (cf eq 2), under the agency of BOP-Cl afforded amide 24 in excellent yield (Scheme 4). After oxidative cleavage of the PMB group and dehydration of the benzylic alcohol, the primary alcohol was oxidized to the aldehyde 25 suitable for the crucial intramolecular/vinylogous Pictet-Spengler/Mannich cyclization that was effected by the action of difluoroacetic acid in benzene to deliver pentacyclic compound 26 (Scheme 4) including the piperazine surrogate pharmacophore. The fifth reported total synthesis of 1 by Williams and co-workers was a formal one as well (13), by virtue of the conversion of a pentacyclic compound 31 (Scheme 5) into a related intermediate which constitutes a formal total synthesis by relay through the Danishefsky (12) and then Fukuyama (11) syntheses, respectively (vide infra (13)).

66

Scheme 4. Piperazine surrogate construction in Danishefsky’s total synthesis.

Scheme 5. Piperazine surrogate construction in Williams’ formal total synthesis.

67

Specifically, acylation of tetrahydroisoquinoline 27, constructed employing a novel radical cyclization of a glyoxalimine derived from Borchardt’s catachol and Garner’s aldehyde, with the N-Fmoc protected amino acid chloride 28 prepared utilizing the oxazinone template technology developed in the William’s laboratory (22), furnished amide 29 without epimerization (Scheme 5). Interestingly, the use of the N-Boc free acid with a variety of coupling agents (DCC, HOBt, HATU) all resulted in very sluggish reactions with poor isolated yields, as did the attempted use of the N-Boc acid fluoride. Removal of the acetonide and the TBS group from 29 was followed by Swern oxidation to yield the piperazinone derivative 30 which upon treatment with trifluoroacetic acid gave rise to 31 and 32 as an approximately 0.72:1 ortho: para mixture of regioisomers in 72% combined yield (Scheme 5). This synthetic approach has been recently applied to the total synthesis of the related marine natural product Renieramycin T (23). Fukuyama and co-workers described their second generation total synthesis of 1 starting from N,N′-diacetylated diketopiperazine 33 prepared from inexpensive L-glutamic acid as the source of chirality (15). Accordingly, Perkin condensation of 33 with aldehyde 34 proceeded regioselectively to give 35. Convrersion to 36 was highly diastereo-selective consisting of protection, hydrogenolysis, hydrazinolysis, and selective reduction (Scheme 6). Cyclizaton of the latter proceeded smoothly via the corresponding N-acyliminium ion to furnish 37 after bistriflation under basic conditions in good yield. Finally the key intermediate 38 including the piperazine surrogate was accessed uneventfully in excellent overall yield as shown in Scheme 6.

Scheme 6. Piperazine surrogate construction in second generation Fukuyama’s total synthesis.

68

Earlier, Kubo and co-workers had reported (24) the synthesis of tetrahydroisoquinoline A (1) including piperazine surrogate 45 starting from 1,4-diacetyl-2,5-piperazinedione (39) and aryl alehyde 40 (25). Specifically, condensation of 39 and 40, according to the procedure of Gallina and Liberatori (26), gave 3-arylidene-2,5-piperazinedione 41 in 74% yield (Scheme 7).

Scheme 7. Piperazine surrogate construction in Kubo’s synthetic approach.

Alkylation of 41 with 4-methoxybenzyl chloride followed by treatment with hydrazine hydrate, methylation, deprotection with TFA and concentrated H2SO4, as well as hydrogenation over 20% palladium hydroxide on carbon in ethanol, provided 42 in very good overall yield (Scheme 7). Protection of the phenol with benzyl bromide in DMF afforded the O-benzylated compound, which was then converted to the corresponding isopropyl carbamate. Deprotection of the latter by hydrogenolysis followed by bromination and chemoselective reduction with an excess of lithium-tert-butoxyaluminum hydride in THF, gave a diastereomeric mixture of alcohol 43 in excellent overall yield. Treatment of 43 with TFA delivered tricyclic compound 44 (96% yield). Carbamate hydrolysis of 44 followed by N-methylation and debromination achieved their final goal including the piperazine surrogate structure 45 (Scheme 7). The conversion of 45 to 1 has not been reported yet.

69

Completion of the Total Syntheses of 1 Completion of the first total synthesis of 1 was accomplished by Corey and co-workers in the following manner. Oxidation of phenol 6 (Scheme 1) with (PhSeO)2O effected position-selective angular hydroxylation to yield, after desilylation and esterification of the resulting primary hydroxyl function with (S)-N-((allyloxy)carbonyl)-S-(9-fluorenyl-methyl)cysteine, ester 46 in very good yield (Scheme 8). Compound 46 was then transformed in one flask to the bridged lactone 47 in 79% overall yield by the following remarkable operations: (1) reaction of 46 with the in situ-generated Swen reagent from excess triflic anhydride and DMSO at -40°C for 30 min to convert the tertiary hydroxyl group of 46 to the O-dimethylsulfonium derivative, (2) addition of i-Pr2NEt and warming to 0°C for 30 min to form the quinone methide by cycloelimination of the Swern type oxosulfonium ylide intermediate, (3) quenching with tert-butyl alcohol to destroy excess Swern reagent, (4) addition of excess N-tert-butyl-N′,N′,N′′,N′′-tetramethylguanidine to convert the 9-fluorenylmethyl thiolether to the thiolate ion and to promote nucleophilic addition of sulfur to the quinone methide and therefore generate the 10-membered lactone bridge, and (5) addition of excess Ac2O to acylate the resulting pnenoxide ion and generate 47 (Scheme 8).

Scheme 8. Completion of Corey’s total synthesis.

Finally, the N-((allyloxy)carbonyl) group of 47 was cleaved and the resulting α-amino lactone was oxidized to the corresponding α-keto lactone 48 with the methiodide of pyridine-4-carboxaldehyde. Reaction of 48 with 2-[3-hydroxy-4-methoxyphenyl]ethylamine in the presence of silica gel generated the spiro tetrahydroisoquinoline C (Structure 1) stereospecifically which was 70

then subjected to MOM cleavage and replacement of CN by OH to yield ecteinascidin-743 (1, Scheme 8). The last steps of the second total synthesis of 1, reported by Fukuyama (10) are described in Scheme 9. Thus, hydrolysis of acetate 13 (Scheme 2) and condensation of the resultant alcohol with L-cysteine derivative 49 furnished ester 50. Chemoselective hydrazinolysis of the thioacetate gave the thiol, which, upon exposure to TFA in 2,2,2-trifluoroethanol under high dilution conditions (0.009 M), smoothly underwent cyclization to the ten-membered sulfide 51 after acetylation of the phenol group. Cleavage of the Troc group and reductive alkylation afforded the corresponding N-methyl amine of which the alloc carbamate and allyl ether were simultaneously cleaved with palladium catalyst to give rise to aminophenol 52 that was biomimetically oxidized according to Corey and co-workers to the respective α-ketolactone. Pictet-Spengler reaction of the latter with amine hydrochloride salt 53 furnished ecteinascidin-770 (54, Structure 1). Finally, generation of the labile hemiaminal from the aminonitrile was effected with silver nitrate to provide ecteinascidin -743 (1, Scheme 9).

Scheme 9. Completion of Fukuyama’s total synthesis. The Zhu and co-workers total synthesis of 1 was accomplished from intermediate 21 (Scheme 3) by reduction of the ester function and subsequent acetylation of the resulting primary alcohol affording the respective acetate. O-Desilylation of the latter followed by Dess-Martin oxidation gave rise to the corresponding aldehyde the Pomerantz-Fritsch-type cyclization of which took place smoothly to provide polyheterocyclic compound 55 (Scheme 10) with concomitant removal of the phenolic MOM-protecting group. Although of no consequence, the cyclization is highly stereoselective (dr > 20:1) and the 71

configuration at C4 of the major isomer was tentatively assigned as S in analogy to the work of Fukuyama and co-workers (11, 15). Saponification of 55 followed by coupling of the resulting alcohol with R-Troc-(S-4,4′,4′′-trimethoxytrityl)cysteine (56) under standard conditions yielded 57 in excellent yield. Gratifyingly, by simply dissolving 57 in TFE containing 1% of TFA, the bridged macrocycle 58 was produced in good isolated yield after masking the phenol as the corresponding acetate. In this operationally simple experiment a complex reaction sequence involving S-trityl deprotection, 1,4-β elimination leading to ortho-quinone methide and macrocyclization via an intramolecular Michael addition occurred in a highly ordered manner to accomplish the key C-S bond-forming process. Simultaneous removal of N-Alloc and O-allyl functions under Guibé’s conditions (27) followed reductive N-methylation to produce 59 after removal ot the N-Troc protective group and biomimetic oxidation according to Corey’s protocol (Scheme 10).

Scheme 10. Completion of Zhu’s total synthesis. 72

Pictet-Spengler reaction of 59 with 3-hydroxy-4-methoxy-phenethylamine (60) was followed by treatment with AgNO3 in MeCN/H2O to provide Ecteinascidin-743 (1) in excellent overall yield for these last two steps. The conversion of 26 (Scheme 4) in Danishefsky’s formal total synthesis of 1 to primary alcohol 67, corresponding to acetate 13 (Scheme 2) in Fukuyama’s total synthesis, was accomplished as described in Scheme 11 below. Specifically, protection of phenol 26 was followed by a McCluskey reaction (28) of the N-Me amine, thereby providing 61 after deprotection ot the TBS ether and reprotection with MOM-Cl. Treatment of the latter with dimethyldioxirane (DMDO) led to epoxidation of the C3-C4 double bond affording the presumed epoxide 62 which upon treatment with 5 equivalents of sodium cyanoborohydride, provided ketone 64 along with a alcohol 63 in a 6:1 ratio.

Scheme 11. Completion of Danishefsky’s formal total synthesis. 73

Two mechanistic hypotheses were put forth to account for the formation of the ketone 64. One could envision a concerted rearrangement with hydride migration, from C4 to C3, to afford ketone 64. Alternatively, the nitrogen atom of the lactam could open the epoxide to produce amidonium alkoxide 65 (Scheme 12). This intermediate could then undergo 1,2-hydride migration to give 64 or competitive reduction by external hydride to provide 63. Apparently, this duality is in fact operating since recourse to a large excess of sodium cyanoborohydride led to alcohol 63 as the predominant product (Scheme 11). Finally, the two benzyl groups were removed giving rise to a triol which upon treatment with a 1:1 ate complex of n-BuLi and DIBAL-H underwent partial reduction of the lactam to provide oxazolidine 66, the phenol group of which was selectively protected as its allyl ether to furnish milestone aminonitrile 67 after exposure to KCN in acetic acid and deprotection of the MOM ether with TFA (12).

Scheme 12. Mechanistic hypotheses accounting for the formation of ketone 64.

74

As mentioned, the total synthesis of 1 by Williams and co-workers is also a formal one by virtue of the conversion of intermediate 31 (Scheme 5) to 61 (Scheme 11), an advanced intermediate in Danishefsky and co-workers total synthesis, which in turn was advanced to 67 (Scheme 11) corresponding to acetate 13 (Scheme 2) described in the Fukuyama and co-workers total synthesis (eq 3).

Finally, The second generation synthesis of Fukuyama and co-workers was completed as follows. After treatment of aromatic amine 68 with tert-butyl nitrite and BF3 etherate, the resulting diazonium salt was reacted with enamide 38 (Scheme 6) in the presence of a palladium catalyst to perform the crucial, intermolecular in this case, Heck reaction. As expected the reaction occurred exclusively from the less hindered face of the enamide to produce coupling product 69 (Scheme 13) with the desired stereo- and regiochemistry. It is noted that this intermolecular Heck reaction was carried out on a multigram scale in excellent yield. An osmium-mediated dihydroxylation of the resulting highly hindered double bond in 69 was accomplished by using K3[Fe(CN)6] as a co-oxidant in the presence of quinuclidine and methanesulfonamide (15). Oxidative cleavage of the resulting 1,2-diol with H5IO6 formed a dialdehyde which underwent facile hydration to afford 70 as a single isomer the stereochemistry of which could not be determined. Hydrogenolysis of the benzyl group in 70 gave the corrersponding phenol heating of which in m-xylene promoted liberation of the dialdehyde that was trapped intramolecularly by the electron-rich B-ring (1) to furnish aldehyde 71 as a 5:1 mixture of diastereomers. Subsequent reduction of 71 with Red-Al afforded an oxazolidine similar to 66 (Scheme 11) which was treated with KCN in acetic acid to produce aminonitrile 72 (Scheme 13). Condensation of the primary hydroxyl group in 72 with cysteine derivative 49 (Scheme 9) followed by selective cleavage of the S-acetyl group with hydrazine produced the respective thiol 73 which upon treatment with TFA led to the cyclic sulfide 74 after acetylation of the phenolic hydroxyl group. Eventually, 74 was transformed to 52 (Scheme 9) the conversion of which to 1 was described in the first synthesis of Fukuyama and co-workers (11).

75

Scheme 13. Completion of Fukuyama’s second generation total synthesis.

Conclusion and Outlook As discussed above, most of the synthetic approaches, that culminated in the total syntheses of 1 outlined, relied on the Mannich reaction for the construction of the critical for antitumor activity piperazine ring. The latter is actually the common section of the tetrahydroisoquinoline subunits A and B (Structure 1). A more direct synthetic approach to this central piperazine ring could be provided by the unprecedented DiAza Diels-Alder Reaction (DADAR) between a 4-trialkysilyloxy-2-azadiene and an appropriately substituted imine as an integral 76

part of our retrosynthetic analysis shown in Scheme 14. Research efforts toward the development of the first DADAR as well as the implementation of the latter analysis are ongoing in the author’s Pharmaceutical Chemistry Laboratory at the University of Patras.

Scheme 14. Magriotis’ retrosynthetic analysis of Ecteinascidin-743. It remains to be seen whether one or a combination of the synthetic approaches discussed in this chapter will in fact find any use, besides the Corey synthesis, in the commercial production of 1.

References 1.

Rinehart, K. L.; Holt, T. G.; Fregeou, N. L.; Stroh, J. G.; Keifer, P. A.; Sun, F.; Li, L. H.; Martin, D. G. J. Org. Chem. 1990, 55, 4512–4515. 2. Hedge, S.; Schmidt, M. Ann. Rep. Med. Chem. 2008, 43, 492–493. 3. Molinski, T. F.; Dalisay, D. S.; Lievens, S. L.; Saludes, J. P. Nat. Rev. Drug Discovery 2009, 8, 69–85. 4. Chen, J.; Chen, X.; Bois-Choussy, MN.; Zhu, J. J. Am. Chem. Soc. 2006, 128, 87–89. 5. Pommier, Y.; Kohlhagen, G.; Bailly, C.; Waring, M.; Mazumder, A.; Kohn, K. W. Biochemistry 1996, 35, 13303–13309. 6. Moore, B. M., II; Seaman, F. C.; Wheelhouse, R. T.; Hurley, L. H. J. Am. Chem. Soc. 1998, 126, 2490–2491. 7. Rinehart, K. L. Med. Drug Rev. 2000, 1–27. 8. Zewail-Foote, M.; Hurley, L. H. J. Med. Chem. 1999, 42, 2493–2497. 9. Zewail-Foote, M.; Li, V.-S.; Kohn, H.; Bears, D.; Guzman, M.; Hurley, L. H. Chem. Biol. 2001, 8, 1033–1049. 10. Corey, E. J.; Gin, D. Y.; Kania, R. S. J. Am. Chem. Soc. 1996, 118, 9202–9203. 11. Endo, A.; Yanagisawa, A.; Abe, M.; Tohma, S.; Kan, T.; Fukuyama, T. J. Am. Chem. Soc. 2002, 124, 6552–6554. 77

12. Zheng, S.; Chan, C.; Furuuchi, T.; Wright, B. J. D.; Zhou, B.; Guo, J.; Danishefsky, S. J. Angew. Chem., Int. Ed. 2006, 45, 1754–1759. 13. Fishlock, D.; Williams, R. M. J. Org. Chem. 2008, 73, 9594–9600. 14. Martinez, E. J.; Corey, E. J. Org. Lett. 2000, 2, 993–996. 15. Kawagishi, F.; Toma, T.; Inui, T.; Yokoshima, S.; Fukuyama, T. J. Am. Chem. Soc. 2013, 135, 13684–13687. 16. Cuevas, C.; Pérez, M.; Martin, M. J.; Chicharro, J. L.; Fernández-Rivas, C.; Flores, M.; Francesch, A.; Gallego, P.; Zarzuelo, M.; delaCalle, F.; Garcia, J.; Polanco, C.; Rodriguez, I.; Manzanares, I. Org. Lett. 2000, 2, 2545–2548. 17. Sakai, R.; Jares-Erijman, E. A.; Manzanares, I.; Silva Elipe, M. V.; Rinehart, K. L. J. Am. Chem. Soc. 1996, 118, 9017–9023. 18. Shaquiquzzaman, M.; Verma, G.; Marella, A.; Akhter, M.; Akhtar, W.; Khan, M. F.; Tasneem, S.; Alam, M. M. Eur. J. Med. Chem. 2015, 102, 487–529. 19. Vieth, M.; Siegel, M. G.; Higgs, R. E.; Watson, I. A.; Robertson, D. H.; Savin, K. A.; Durst, G. L.; Hipskind, P. A. J. Med. Chem. 2004, 47, 224–232. 20. Thoma, S.; Endo, A.; Kan, T.; Fukuyama, T. Synlett 2001, 1179–1181. 21. De Paolis, M.; Chen, X.; Zhu, J. Synlett 2004, 729–731. 22. Jin, W.; Williams, R. M. Tetrahedron Lett. 2003, 44, 4635–4639. 23. Yokoya, M.; Toyoshima, R.; Suzuki, T.; Le, V. H.; Williams, R. M.; Saito, N. J. Org. Chem. 2016, 81, 4039–4047. 24. Saito, N.; Tachi, M.; Seki, R.-i.; KaMayachi, H.; Kubo, A. Chem. Pharm. Bull. 2000, 48, 1549–1557. 25. Saito, N.; Tachi, M.; Seki, R.-i.; Sugawara, Y.; Takeuchi, E.; Kubo, A. Synth. Commun. 2000, 30, 2407–2421. 26. Gallina, C.; Liberatori, A. Tetrahedron 1974, 30, 667–673. 27. Guibé, F. Tetrahedron 1998, 54, 2967–3042. 28. Hobson, J. D.; McCluskey, J. G. J. Chem. Soc. (C) 1967, 2015–2017.

78

Chapter 6

Condensation Reaction between Phenanthroline-5,6-diones and Ethylenediamine and Its Optimization through Dialogue between Theory and Experiment Luis Sanhueza,1 Diego Cortés,1 Iván González,1,2 and Bárbara Loeb1,* 1Facultad

de Química, Pontificia Universidad Católica de Chile, Vicuña Mackenna 4860, Santiago, Chile 2Facultad de Ciencias de la Salud, Universidad Central de Chile, Lord Cochrane 417, Santiago, Chile *E-mail: [email protected]

The synthetic route to obtain the pyrazino polypyridinic type of ligands generally involves a condensation reaction between a diaminne and a dione. In the case of pyrazino[2,3-f][1,10]phenanthroline, ppl, and pyrazino[2,3-f][4,7]phenanthroline, ppz, a notoriously lower yield for the condensation reaction of the later has been observed. In this work, experimental results along with DFT methods allowed us to elucidate and improve the synthetic pathways involved in these ligands. Intermediary molecules for the corresponding condensation of dione and ethylenediamine were detected. By a continuous dialogue between theory and experiments the limiting reaction step was established as the formation of a “non-aromatic” intermediate, which was shown to be the cause for the lower yield observed for ppz. This intermediate was theoretically and experimentally characterized, thereby permitting us to facilitate its conversion to the desired product and obtain close to quantitative yield for the reaction.

© 2017 American Chemical Society

Introduction The use of polypyridines as organic reactants and/or coordinating ligands in transition metal complexes has been widely investigated. These versatile aromatic compounds permit us to modify the structural and electronic properties of coordination complexes and use them in applications such as solar cells (1, 2), OLED (3, 4) and nonlinear optical devices (5). Among polypyridinic ligands, dipyrido[3,2-a:2′,3′-c]phenazine (dppz), has been widely studied since the early discovery of its behavior as a spectroscopic probe of [Ru(bpy)2dppz]2+ in DNA intercalation (6). The analogous ligand, pyrazino[2,3-f][1,10]- phenanthroline (ppl), has been less studied (7, 8). The synthesis of both of these ligands by condensation of the corresponding dione with ethylene diamine is straightforward (9).

It has been found that the use of the structural “isomers” of the aforementioned ligands, dipyrido[2,3-a:3′,2′-c]phenazine (dbq′), and pyrazino[2,3-f][4,7]phenanthroline, (ppz or dpp′), affect considerably the spectroscopic properties and the electronic behavior of coordination complexes of the type [Ru(R2-bpy)2(L)]2+, where L is dbq´ and ppz. For example, the absorption of [Ru(bpy)2(ppz)]2+ is noticeably red-shifted when compared to [Ru(bpy)2(ppl)]2+, while its lifetime is markedly enhanced (10, 11), making the compound promising for device applications. However, the main limitation for their use, particularly for ppz, involves overcoming the synthetic challenge. As mentioned above, ppl can be obtained by simple condensation between ethylenediamine and 1,10-phendione with more than 90% yield (12), but the synthesis of ppz under the same reaction conditions was reported to give significantly lower yields (13) (Scheme 1). It has been suggested that the low yield of this synthesis is due to a disproportionation reaction, obtaining an aromatic species (the desired product) and also a "non-aromatic" derivative, H2ppz (14). It should be mentioned that an alternative synthetic method gave ppz with good yield, but with a longer and more complex route involving potentially unstable or air-sensitive intermediates (14). In order to overcome the difference in behavior for the synthetic process, a continuous dialogue between theory and experiments was undertaken to evaluate the stages involved in the condensation reaction and the variables so as to control and improve this reaction. A separate analysis of the starting materials, the intermediate products, and the final products was undertaken, that as a whole permitted us to understand and control the reaction.

80

Scheme 1. Condensation reactions for ppl and ppz compounds under the same experimental conditions.

Analysis of Starting Materials As can be seen in Scheme 1, under similar experimental conditions, after 13 h stirring, ppl precipitates as a yellow pale solid and is obtained in 85 % yield, while in similar conditions ppz is obtained in 42 % yield. The yields are in agreement with previously reported results (12, 13). As a first stage looking for variables that could explain the difference in yield, an analysis in regard to the starting materials was considered. Specifically, the idea was to understand if the presence of the N groups in 4,7 positions affect the reactivity of the phendione when compared with 1,10 phendione. Conceptual density functional theory (DFT) was applied to evaluate the global and local reactivity properties. The global reactivity indexes, molecular hardness η (15) and electrophilicity ω (16), given in Table 1, show that 1,10-phendione and 4,7-phendione have almost the same electronic reactivity and electrophilic character. The differences of η and ω between 1-10-phendione and 4,7-phendione are negligible in this context (1.78 and 2.34 kcal/mol, espectively). Moreover, taking into account that condensation is a frontier molecular orbital-controlled step at local level, the Fukui Functions (FF) (17, 18) for susceptible sites to electrophilic (f+) and nucleophilic (f-) attacks were obtained by frozen core approximation. Results (Figure 1) show that no significant difference can be found; in both molecules the larger values of f+ and f are located on the C and O atoms that take part in the condensation step. FF by finite difference approximation using the Yang-Mortier scheme was also calculated, with analogous results. Therefore, according to the global and local indexes, a difference in reactivity of the phendione precursors seems not likely to be the reason for the difference in yields observed.

81

Table 1. Molecular hardness (η) and electrophilicity (ω) for 1,10 and 4,7 phendione precursors. Molecule

η (kcal/mol)

ω (kcal/mol)

1-10 phendione

41.49

193.84

4-7 phendione

39.71

191.50

Figure 1. Condensed Fukui Functions for electrophilic (f+) and nucleophilic (f-) attack calculated for 4,7-phendione and 1,10-phendione.

With these experimental and computational data, the complete reaction of phendione and ethylenediamine was divided in two steps: condensation between the oxygens of phendione and the NH2 groups of ethylenediamine, and aromatization reaction to generate the respective ppl and ppz aromatic-type compounds.

82

If the second step involving an aromatization (oxidation) reaction is the ratelimiting step, the presence of O2 should have a role to obtain the desired aromatic product. Therefore, the same previous reactions were repeated under identical conditions as in Scheme 1, but under N2 atmosphere. After 17 hours stirring, no precipitation was observed, and the solutions for both reactions kept their deep red color. The solvent of these reddish solutions was eliminated by evaporation, and the corresponding red-to-orange solids isolated. It was assumed that the solids correspond to the “non-aromatic” intermediates for ppl and ppz respectively.

Analysis of Intermediates In order to prove this assumption, a complete analysis of these products was done. The influence of these intermediates on the difference in reaction yield was additionally analyzed. UV-Vis Spectra A comparison of the absorption spectra between the “non-aromatic” intermediates (isolated by the reaction in N2 atmosphere) and the corresponding final products was conducted. The first spectra show mainly a red-shifted transition at 284 nm and 257 nm, displaced by 32 nm and 5 nm respectively for “non-aromatic” ppz and ppl from the related aromatic compound, where all transitions are the π → π * type. Moreover, new bands appearing at 376 and 398 nm can be observed for ppl and ppz intermediates respectively. In contrast, the final products are completely transparent in this region. Two different structures for each intermediate are feasible, differing in the protonation of the pyrazinic N, as shown in Scheme 2.

Scheme 2. Possible structures for the non-aromatic intermediates. In order to determine the most probable intermediate, experimental and theoretical absorption spectra for both intermediates were compared, and are shown in Figure 2. For this analysis, the goal is to determine if the pyrazinic 83

nitrogens of the intermediates are protonated, i.e., to understand if the N is of sp3(A) or sp2(B) type.

Figure 2. Experimental and calculated absorption spectra in MeCN for the “A proposed intermediates” (with C and N sp3 configuration). i) ppl and ii) ppz compounds. When comparing theoretical and experimental UV-Vis spectra for the “A” type intermediates (scheme 2), i.e., with N protonated, the same tendency is observed for the ppl and the ppz derivatives. For ppl intermediates, good correlation can be observed from 270 to 400 nm, where intensities and absorption positions are well simulated. This correlation is quite clear for the 276 nm and 298 nm bands. In addition, the calculation for this “A” type structure predicts a band at 363 nm, with a difference of 0.12 eV compared to the experimental band, at 376 84

nm (Figure 2-A). In the case of the “A” type intermediate for ppz, the same good correlation between experimental and theoretical information is observed (Figure 2-B). Specifically, the transitions at 285 nm and 233 nm correlate perfectly well. Regarding the lower energy band for this intermediate, a difference of 0.19 eV is observed between the experimental (398 nm) and theoretical (425 nm) bands. Although greater than that of the ppl derivative, it is still in an acceptable range. Regarding the “B” type structure for the ppz intermediate, the theoretical calculations determine a relatively intense band at 262 nm that is absent in the experimental spectra. Moreover, the calculation for the lower energy band, although closer to the experimental value, appears with a negligible oscillator strength (4,81x10-3) compared to 0.081 for A-ppz. According to the previous analysis, the intermediate structure type “A” seems to be the more probable. NMR characterization sheds more light about this assumption. NMR Characterization The four compounds: ppl, ppz, and the corresponding experimental intermediates were fully characterized by 1H-NMR spectroscopy. Twodimensional measurements were carried out when necessary. For ppl and ppz products the positions and couplings were in agreement with previously reported characterizations (7–10) (Scheme 3 and Table 2). In order to corroborate the nature (A or B) of the intermediates, the experimental products were also fully characterized (Table 2).

Table 2. 1H-NMR characterization for ppl and ppz compounds and their corresponding intermediates

a

Moleculea

Ha

Hb

Hc

Hd

NH

ppla

9.22

7.92

9.43

9.13

--

ppza

8.94

7.83

9.27

9.25

--

ppl-intermediateb

8.15

7.48

8.91

3.43

--

ppz-intermediateb

8.30

7.07

8.29

3.48

4.89

In CDCl3.

b

In CD3OD.

13C

DEPT spectra in deuterated methanol for the ppl reaction intermediate shows a strong signal at 45.02 ppm that is unequivocally assigned to CH2 (Figure 3). On the contrary, for the ppz intermediates, only a trace of a possible aromatic impurity can be observed in this region. Additionally, for the ppz reaction intermediates, HSQC and HMBC NMR spectra in MeOD were measured. They showed the coupling between H-d(3,48 ppm) and C-7 at 41,71 ppm to be direct, and the same proton with C-6 126,79 at ppm, confirming that proton signal at 3.48 ppm corresponds effectively to CH2 from the saturated fragment in the molecule, where carbons are of the sp3 type. 85

Scheme 3. NMR characterization for ppl and ppz compounds, same labels can be used for intermediates.

Figure 3. NMR DEPT spectra in deuterated methanol for the intermediates for ppl (upper plot) and ppz (lower plot) 1H-NMR

spectrum in DMSO-d6 of the ppz-intermediate shows a signal at 3.54 ppm, attributable to the aliphatic “d” protons. Additionally, a signal at 6.14 ppm appears. According to HSQC spectrum, this signal is not related to a C atom, indicating indirectly that this H should therefore be coordinated to a N atom. This conclusion was supported by HMBC spectroscopy, where a coupling between H and C4 (quaternary) at 139.2 ppm was observed, confirming the presence of a direct NH bond after condensation and therefore an “A” type structure. 86

In order to check the solvent effect in the condensation reaction, the synthesis for ppl and ppz was repeated in the same conditions as in Scheme 1, but in dry CDCl3. Effectively, for both ppl and ppz, after 4 h no product and/or red solution was observed. The same was observed when CHCl3 and THF were used. These results reflect the nature of the solvent used, which is quite important for the condensation reaction, with a protic solvent needed for the reaction to occur. Moreover, in aprotic solvents, after ending the reaction, the starting materials can be observed by NMR analysis, without any change. The keto-enol tautomerism is well known for this type of systems (19), and the protic solvent could be playing an important role in it. This tautomerism would occur before the nucleophilic attack involved in the condensation reaction.

Theoretical Reactivity Indexes To understand the preference of an “A” type structure, and specially to answer the original question about the reasons of a low yield in the ppz synthesis compared to ppl, we returned to theoretical calculations, checking local reactivity properties for each “non-aromatic” compound. The same global reactivity indices were determined for the four intermediaries (Table 3).

Table 3. Molecular hardness (η) and electrophilicity (ω) for non-aromatic intermediates and products Molecule

η (kcal/mol)

ω (kcal/mol)

A-ppl-intermediate

45.57

70.07

A-ppz-intermediate

36.53

71.35

B-ppl-intermediate

43.98

134.58

B-ppz-intermediate

43.20

128.57

The η index of electronic reactivity is similar between non-aromatic intermediaries, with a difference of ~7 kcal/mol. However, ω for “A-type intermediates” is less than for “B-type” ones by ~60 kcal/mol; this difference would make them preferable candidates for the aromatization to ppl and ppz by oxidation because A-ppl-intermediate and A-ppz-intermediate are better nucleophiles than B-ppl-intermediate and B-ppz-intermediate. Regarding the reason for the difference in yields of both reactions, the second step of the synthesis, i.e., the oxidation of the intermediate to generate the final product, was analyzed. The total energy for the intermediate compounds was determined; A-ppz-intermediate is ~30 kcal/mol more stable than the A-ppl-intermediate. This can be caused by the interaction between the hydrogen atoms of the NH groups with the nitrogen atoms in “Y” position of A-ppz-intermediate, establishing a higher energetic barrier to form the final product with respect to A-ppl-intermediate. In addition, the susceptible oxidation sites through an 87

electrophilic attack of the intermediaries were analyzed. This is a process controlled by the frontier molecular orbitals. However, in terms of the local reactivity, it is not possible to find significant differences because in aromatic and non-aromatic compounds the sites with larger contribution to f- are on the nitrogen atoms adjacent to the CH2 groups.

Complete Conversion of Products The analysis described so far shows that the lower yield in the condensation reaction to obtain ppz is not attributable to the reactivity of the starting material. The reason seems to be related to the properties of the intermediate product, identified as an “A” type ppz-structure. Theoretical calculations show that the A-ppz-intermediate has a lower tendency to donate electrons than the A-ppl-intermediate. As mentioned above, the electronic pairs from the N atoms in 4 and 7 positions interact with NH from the piracinic group, hindering the aromatization into the ppz compound. Therefore, it can be thought that the strength of the oxidant is important, and that it should have the capacity to aromatizate the corresponding intermediate. In the case of the A-ppz-intermediate, a stronger oxidant than oxygen from air would be needed, avoiding at the same time decomposition. To test this assumption, the same reaction to obtain ppz shown in Scheme 1 was carried out, but adding MnO2. Specifically, 8 eq. of this oxidant were added to the mixture and the solution stirred for additional 2 hours. The reaction was followed by UV-Vis spectroscopy until the disappearance of the 398 nm band, characteristic of the intermediate. The product was precipitated by addition of an excess of diethyl ether. The 1H-NMR spectrum of this product corresponded to ppz, with no evidence of any by-product. The reaction yield was close to be quantitative. Elemental analysis gave also satisfactory results.

Conclusion The analysis of all previous experimental and theoretical data showed that the low yield for the condensation reaction to obtain ppz cannot be attributed to the corresponding starting material, but rather to the nature of the non-aromatic intermediate. Through identification of its structure and local reactivity analysis, the need of a different oxidant was revealed and applied. To our knowledge, this is the first time this simple condensation reaction could be conducted for ppz in close to quantitative yield. This was possible by a dynamic dialogue methodology between theory and experiments that allowed us to identify, characterize and isolate the relevant intermediaries, and to optimize the reaction to achieve the product in almost quantitative yield.

Acknowledgments This work was financially supported by FONDECYT Chile by Project Nº 1110991. The authors gratefully acknowledge Dr. Enrique Castro and Dr. David Moreno for helpful discussions. 88

To the memory of Dr. Ernest Eliel, for the long hours of interesting discussions during my visits to Chapel Hill, regarding scientific policies and the way to increase the development of chemistry in latin american countries. Definitely, they changed my way to conduct my academic career.

References 1. 2. 3. 4.

5. 6. 7. 8. 9.

10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

Mitsopoulou, C. A.; Veroni, I.; Philippopoulos, A. I.; Falaras, P. J. Photochem. Photobiol., A 2007, 191, 6–12. Bomben, P. G.; Robson, K. C.; Sedach, P. A.; Berlinguette, C. P Inorg. Chem. 2009, 48, 9631–9643. Ulbricht, C.; Beyer, B.; Friebe, C.; Winter, A.; Schubert, U. S. Adv. Mater. 2009, 21, 4418–4441. Dragonetti, C.; Falciola, L.; Mussini, P.; Righetto, S.; Roberto, D.; Ugo, R.; Valore, A.; De Angelis, F.; Fantacci, S.; Sgamellotti, A.; Ramon, M.; Muccini, M. Inorg. Chem. 2007, 46, 8533–8547. Le Bozec, H.; Renouard, T. Eur. J. Inorg. Chem. 2000, 229–239. Friedman, A. E.; Chambron, J. C.; Sauvage, J. P.; Turro, N. J.; Barton, J. K. J. Am. Chem. Soc. 1990, 112, 4960–4962. Delgadillo, A.; Romo, P.; Leiva, A. M.; Loeb, B. Helv. Chim. Acta 2003, 86, 2110. Aldrich-Wright, J. R.; Greguric, I.; Vagg, R. S.; Vickery, K.; Williams, P. A. J. Chromatogr A. 1995, 718, 436–443. Arias, M.; Concepción, J.; Crivelli, I.; Delgadillo, A.; Díaz, R.; Francois, A.; Gajardo, F.; López, R.; Leiva, A. M.; Loeb, B. Chem. Phys. 1996, 326, 54–70. Treadway, J. A.; Loeb, B.; Lopez, R.; Anderson, P. A.; Keene, F. R.; Meyer, T. J. Inorg. Chem. 1996, 35, 2242–2246. Fuchs, Y.; Lofters, S.; Dieter, T.; Shi, W.; Morgan, R.; Strekas, T. C.; Gafney, H. D.; Baker, A. D. J. Am. Chem. Soc. 1987, 109, 2691–2697. Collins, J. G.; Sleeman, A. D.; Aldrich-Wright, J. R.; Greguric, I.; Hambley, T. W. Inorg. Chem. 1998, 37, 3133–3141. Schmidt, P.; Druey, J. Helv. Chim. Acta 1957, 42, 350–355. Abeywickrama, C.; Baker, A. D. J. Org. Chem. 2004, 69, 7741–7744. Parr, R. G.; Pearson, R. G. J. Am. Chem. Soc. 1983, 105, 7512–7516. Parr, R. G.; Szentpaly, L. V.; Liu, S. J. Am. Chem. Soc. 1999, 121, 1922–1924. Fukui, K. Science 1982, 218, 747–754. Fukui, K.; Yonezawa, T.; Shingu, H. J. Chem. Phys. 1952, 20, 722–725. Yang, W.; Mortier, W. J. J. Am. Chem. Soc. 1986, 108, 5708–5711.

89

Chapter 7

Solvent Effects on Electronic Circular Dichroism Spectra Aguinaldo Robinson de Souza,*,1 Valdecir Farias Ximenes,*,1 and Nelson Henrique Morgon*,2 1Department

of Chemistry, São Paulo State University, São Paulo, 17033-360, Brazil 2Institute of Chemistry, Campinas State University, Campinas, 13083-970, Brazil *E-mail: [email protected], [email protected], [email protected]

The Electronic Circular Dichroism Spectrum (ECD) is a valuable tool to study the unknown absolute configuration of an optically active molecule. And the comparison between experimental data and theoretical computational calculations has been a successful strategy for this study. However, the ECD spectrum is very sensitive to solvent effects that significantly change the character of the results obtained. This chapter is focused on the study of the solvent effects and their application in both experimental and computational chemistry of ECD of the compound 3,3′-dibromo-1,1′-bi-2-naphthol.

Introduction The study of the chiroptical properties has as one of the challenges the elucidation of the absolute conformation of a molecule. And the comparison of the results obtained experimentally with those derived from computational calculations has been a successful strategy employed. Another motivation in this area is the determination of the conformation of flexible molecules that can adopt specific conformations when interacting with chiral environments. The chiroptical properties of a molecule are very sensitive to the medium polarity such as, for example, the variation of the dielectric constant of the medium in the presence © 2017 American Chemical Society

of a solvent or local polarity and conformational changes at the active site of proteins. This phenomenon occurs in the experiments of optical rotation (OR), electronic or magnetic circular dichroism (ECD or MCD), vibrational CD (VCR), and Raman optical activity (ROA). From this point of view, understanding the effects of the medium where the molecule is located is of fundamental importance for the calculation and experimental determination of chiroptical properties. In the present chapter, we present some experimental and computational simulation results in order to understand the influence of the solvent on the chiroptic properties of the enantiomeric pair (R)-(+)-3,3′-dibromo-1,1′-bi-2naphthol and (S)-(−)-3,3′-dibromo-1,1′-bi-2-naphthol. Next, we will present some methodologies used in the elaboration of models for the study of the effect of the solvent on chiroptic properties. A detailed discussion of the effects on optical rotation and electronic circular dichroism can be found in the works of Pecul et al. (1, 2), and Pescitelli et al. (3). These studies present a more elaborate discussion on the subject and should be consulted for a greater understanding on the part of the readers. Another interesting study by Polavarapu et al. (4) on the conformational sensitivity of 6,6′-dibromo-1,1′-bi-2-naphtol predicted that the electronic absorption and ECD spectra do not show significant variations for different conformations of the hydroxyl group present in the molecule. Although the extensive study on this topic is found in the literature, we have not found a study, both experimental and theoretical, on the influence of the solvent on the chiroptic properties of the enantiomeric pair of 3,3′-dibromo-1,1′-bi-2-naphthol. We also present some considerations about the phenomenon of chirality and atropisomerism in the context of its importance in obtaining the absolute conformation of molecules of interest.

Chirality and Atropisomerism The omnipresence of molecules that constitute the living matter and which have as the major characteristics to differ from each other only by the three-dimensional arrangements of their constituent elements can be thought of as a medium that evolution “chose” to increase the levels of differentiation of matter. In others words, the formation of different substances from the few elements that account for the most of the mass of organic matter, i.e., carbon, hydrogen, oxygen, and nitrogen atoms. It would therefore be reasonable to dispose of the levels of differentiation of matter, starting from the following possible levels: First: different combinations of protons, neutrons, electrons and other fundamental particles forming different chemical elements; Second: different combinations in number and types of chemical elements forming different molecules; Third: same chemical elements in number and types, but different connectivity forming constitutional isomers, and; Fourth: same elements, same connectivity, but different spatial arrangements forming stereoisomers. 92

If we consider that particles, except the electron, are formed by the combination of other subatomic particles, like quarks, and these in turn, formed by different combinations of other elements present in the string theory, we can the raise the hypothesis that all the known matter with their different physical and chemical properties are nothing but expressions of one or a few forming units of the whole and also of quantum vacuum fluctuations. It is therefore no surprise that molecules differing only by the spatial arrangement of the chemical elements may have different chemical properties. Among the molecules that keep the stereoisomeric relationship between them, we highlight those that are non-overlapping mirror images. These are the enantiomers, molecules that exist in pairs in nature, that is, molecules that can be thought as a “right” and “left”. As well-established, the condition for a molecule to exist in pairs is the total absence of symmetry in its three-dimensional structure. For organic compounds, the simplest condition for achieving this asymmetry is the existence of a single carbon atom bearing four different substituents, known as chiral or asymmetric carbon. However, the presence of a chiral carbon is not a sufficient or necessary condition, because even in its absence, a molecule may exist in pairs. As an example, we have some biphenyl systems, which will the object of the present chapter. This phenomenon is known as atropisomerism, and takes place when the energetic barrier for rotation around the single bond that connects the aryl groups is sufficiently high, leading to enantiomeric conformations sufficiently stable to be isolated. The word atropos comes from the Greek and means “a” (not) and “tropos” (rotation), that is, impeded rotation in the axial axis that joins the aryl groups. The first molecule synthesized and exhibiting this characteristic was 6,6′-dinitro-2-2′-diphenyl acid obtained by Christie and Kenner in 1922 (5). These are biaryl systems where the presence of substituents in the ortho position prevents the free rotation around the single bond and, consequently, the non-interconversion of one enantiomer into the other.

Computational Models for Solvent The solvent can be considered, from the standpoint of computational modeling, in distinct ways: 1) the solvent is considered as a source of potential around the molecule, and in this case the molecule is treated from a quantum mechanical point of view and the solvent can be modeled as a homogeneous and isotropic continuous dielectric model, the polarizable continuum model (PCM) is commonly used method in computational chemistry to model these effects, or as a collection of discrete charges, and 2) the solvent and the molecule are treated at the same level of calculation. In the PCM model the molecule of interest is placed in a cavity in the continuous medium and the interactions are described by the inclusion of a term describing these interactions in the wave equation. The shape of the cavity may vary depending on the model chosen, and may be a sphere or have a more complex shape. The cavity model developed in the research group of Professor Tomasi is one of the most used and the cavity is formed by spheres 93

centered in the solute´s nucleus and the surface of the same is divided into charge finite elements (6, 7). Another approach used is 3) to consider the solvent in an explicit manner, together with the molecule of interest forming a super molecule, where the solute molecule and the solvent molecules are treated at the same level of theory. The effect of the solvent molecules is obtained from the difference between the results obtained with the molecular complex and those obtained with the isolated molecule. This approach uses a combination of quantum-mechanical and molecular mechanics methods (QM/MM) where the most important part of the system is treated by ab initio methods and the effect of the environment is treated using MM (8).

Basic Theory for Solvated Systems The theory involved in the Polarizable Continuum Model in the study of solvent effect in the spectrum of electronic circular dichroism was developed by Tomasi et al. The effects of the solvent are incorporated into a version of the integral equation formalism (IEF). In the formulation of a theory for the effect of the solvent we must take into account the quantum mechanical methodology used and the model chosen for the solvent. One of the most used methodologies that lead to results with more precision is the Density Functional Theory (DFT) and in relation to the solvent model one of the choices is to use a continuous atomistic model for the solvent. The complete Hamiltonian for the solute molecule can be written as (9):

where H0 is the Hamiltonian in vacuum and V′(t) is the time-dependent perturbation. In this formulation a surface charge density was introduced which represents the response of the solvent to the external field after the creation of the solute cavity in the solvent. The surface charge density was partitioned into small portions called "tesserae" of area equal to k. The equation for V′(t) describes the time-dependent perturbation in the solute molecule in terms of the external electric field and from this equation we can calculate the effective polarizability of the molecule. Approximate solutions from the effective Hamiltonian can be obtained from a procedure analogous to that obtained for isolated molecules. The modification that must be made is in the Fock operator which must now include the effects of the solvent in the presence of an oscillating field. It has been shown (9) that the orbitals and their energies are modified in comparison to the isolated molecule. This modification was represented as:

where Q is the dielectric matrix that defines the apparent bulk of the solvent and depends on the geometry and the dielectric constant. The PCM model solute94

solvent interactions are of electrostatic nature and due to this fact the correlation between the results obtained in computer simulations are in better agreement with polar solvent than in relation to non-polar solvent. The method derived from the DFT / PCM formalism provides reliable model in relation to electrostatic effects in the solvent chiroptical properties of molecules and also has a dependence on the nature of the solvent (9).

Solvent Effects on ECD Spectra (R)-(+)-3,3′-dibromo-1,1′-bi-2-naphthol (1) and (S)-(−)-3,3′-dibromo-1,1′-bi2-naphthol (2) (Figure 1) were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO, USA). Stock solutions of 1 and 2 (10 mM) were prepared in ethyl alcohol. The solvents were of analytical grade. Circular Dichroism (CD) spectra were recorded with a Jasco J-815 spectropolarimeter (Jasco, Japan) equipped with a thermostatically controlled cell holder. The spectra were obtained with 1 nm step resolution, response time of 1 s and scanning speed of 50 nm/min. The spectra were recorded at a bi-naphthol concentration of 15 mM over a wavelength range of 200–350 nm at 20°C. A 3 mL quartz cuvette with a 10 mm path length and a magnetic stirrer were used for the measurements. The baseline (solvents) was subtracted from all measurements.

Figure 1. Chemical Structures of (R)-(+)-3,3′-Dibromo-1,1′-bi-2-naphthol (1) and (S)-(−)-3,3′-dibromo-1,1′-bi-2-naphthol (2). In Figures 2a and 2b we present the ECD spectra of the binaphthols 1 and 2 in different solvents. Although some alteration can be seen depending on the solvent, the (R) enantiomer always showed a negative band centered around 240 nm and a positive one around 226 nm. How could be expected, the (S) enantiomer presented opposite ECD bands. Interestingly, a correlation was obtained between the dielectric constants of the solvent and the ECD intensity. The results depicted in Figure 3 show that the increase in the polarity of the solvent provoked a decrease in ECD intensity. 95

Figure 2. ECD spectra of binaphthols 1 (a) and 2 (b) in different solvents.

Figure 3. Dielectric constant effects on ECD intensity of binaphthols 1 and 2. The solvents used were: tetrahydrofuran (T), ethanol (E), methanol (M), acetonitrile (A) and water (W).

96

Computational Methodology The torsional potential energy curves of the stereoisomers (R,S) -3,3′-dibromo-1,1′-bi-2-naphthol (BiNaphthol) were obtained at the ab initio level in a time dependent Density Functional Theory approach in search to obtain reliable results as compared with the experimental ones. The search for the most stable conformation adopted by BiNaphthol, in the solvent phase was performed in a relaxed scan of δ1 dihedral angle, 0 – 360(30°), at B3LYP/6-31G (d,p) level of theory (Figure 4). In Figure 4 the Carbon, Hydrogen, Bromide and Oxygen atoms are represented by colors brown, white, purple and red, respectively.

Figure 4. Definition of the torsional angle δ1.

Figure 5 shows two minimum energy conformations of BiNaphthol using water and ethanol described by PCM solvation model. This procedure was applied to the others solvents: methanol, acetonitrile and tetrahydrofuran. On the basis of the low energy conformers, the stationary points for each curve were confirmed by the frequency analysis minima with all real frequencies. As we can observe in the Figure 5 the energetic barrier at 180 degrees for the R-S interconversion of BiNaphthol is much higher than the thermal energy, kT, at 25°C (0.59 kcal/mol). This barrier is more pronounced in the presence of water as solvent than with ethanol; this behavior can be explained due to the higher value of the dielectric constant of water as compared with ethanol. Another interesting result is the same value for the minimum of the torsion angle of BiNaphthol in both solvents. According to the Figure 5 we can infer that the molecule can adopt two stable conformations at the torsion angle of approximately 90 and 270 degrees. The excited-state behavior in the Franck–Condon (FC) region of the BiNaphthol was investigated in different solvents by means of time-dependent Density Functional Calculations, including solvent effects (Water, Methanol, Ethanol, Acetonitrile, and Tetrahydrofuran) by the Polarizable Continuum Model and using the CAM-B3LYP functional and 6-311++G(2df,p) basis sets at molecular geometry obtained employing B3LYP/6-31G(d,p). The ECD spectra obtained are shown at Figures 6 and 7. All calculations were performed using the Gaussian09 program (10).

97

Figure 5. Potential Energy Surface for BiNaphthol at B3LYP/6-31G(d,p) + PCM.

Figure 6. Calculated ECD spectra of (R,S) BiNaphthol in Water and Ethanol at the CAM-B3LYP/6-311+G(2df,p)//B3LYP/6–31G(d,p) level (a), and its experimental ECD results (b) Ethanol and (c) Water. 98

Figure 7. Simulated ECD spectra of BiNaphthol in Water, Methanol, Ethanol, Acetonitrile, and Tetrahydrofuran at the CAM-B3LYP/6-311+G(2df,p)//B3LYP/6–31G(d,p) level. In Figure 6 (a) we present the results for BiNaphthol two solvents: water and ethanol. The red line symbolizes the R-isomer and the black line the S-isomer. We can observe that there is an inversion of the ECD signal to the antipodes R-S as observed experimentally. We also observed that the ECD spectrum consists of three bands and in the case of the R isomer we have two positive bands around 212 and 300 nm and a negative one around 240 nm. In Figure 6 (b) and 6 (c) we can observe a good agreement for both computed spectra. We can observe that the three bands in the indicated regions and also the shift to the lower energy region for the ethanol solvent when compared to the solvent water. The second positive band, around 300 nm, does not have its position modified when comparing the two solvents used. In Figure 7 we present the spectrum of ECD for the BiNaphthol molecule in the studied solvents: water, methanol, ethanol, acetonitrile and tetrahydrofuran. We can observe the presence of three bands: two positive around 212 and 300 nm and one negative around 240 nm. The highest intensity is verified for the band at 212 nm followed by the band at 240, and the lowest intensity located at around 300 nm. The effect of the solvent is more pronounced for the band located around 240 nm (the negative band), and the other two bands do not show a significant variation with the solvent change. This result can also be verified experimentally when compared with the experimental results presented in Figure 3 unless for the water solvent which shows a significant variation also for the band located around 212 nm; this effect was not verified in the calculations. The band located around 240 nm shows the following order of decreasing value for the intensities: IAcetonitrile < IEthanol < Iwater < Imethanol < ITHF. The intensities, 99

relative to the solvents acetonitrile, ethanol, and THF, of the calculated negative band are in good agreement with the values obtained experimentally and presented in Figure 3. The results for the solvents water, and methanol, however, are in not agreement with the experimental results. Further calculations are been carried out to overcome these findings as, for example, the elaboration of a model with explicit solvent molecules around the solute.

Concluding Remarks The effects of the solvent on the ECD spectrum are an important research area due to the possibility of obtaining information on the absolute conformation of a molecule of interest. The PCM model is applicable in the study of solvent effect for those with low and high polarity, and for the case of solvents were hydrogen bonds play a role this methodology is not yet extensively applicable, however, some results are quite close to those obtained experimentally. The molecule under study has a solvent-dependent ECD spectrum with a pronounced negative band around 240 nm and has the following decreasing order for the intensity value: IAcetonitrile t-BuLi > n-BuLi) reported by Hsieh (Figure 12). The initiation of polymerization of 3-methylstyrene with n-BuLi display a brief inhibition period that is not observed with either s-BuLi or t-BuLi. This behavior results in a higher molecular weight distribution for the polymerization with n-BuLi because chain propagation is faster than initiation. The computationally derived order of polymerization is calculated to be (n-BuLi)1 > (s-BuLi)1 > (t-BuLi)1 > (n-BuLi)2 in which the subscripts refer to the aggregation state. Both the computational and the experimental results suggest that the bulk of the organic group is the governing factor in the polymerization because the monomeric form of n-BuLi is predicted to be the fastest computationally. This also suggests that the dimeric form of n-BuLi reacts directly with styrene and does not dissociate to a monomer in solution because it was observed to be the slowest.

114

Figure 9. 1H-RI-NMR spectra following the injections of cyclopentadiene (A) and benzaldehyde (B) into BuLi in THF at –86 °C. R = reference, S = solvent, T = BuLi tetramer, D = BuLi dimer, C = cyclopentadiene, C– = lithium cyclopentadiene, B = benzaldehyde, A = alcoholate. Reproduced from (31). Copyright 1985, American Chemical Society.

115

Figure 10. 7Li RI-NMR spectra following the injection of n-BuLi into a THF/Me2O (1:3) solution of 28 at –135 °C. The spheres are lithium atoms. Reproduced from (23). Copyright 2007, American Chemical Society.

Figure 11. Mechanism for styrene polymerization initiated by butyl lithium.

116

Figure 12. Consumption of 3-methylstyrene as initiated by isomeric butyllithium reagents. Reproduced from (42). Copyright 1995, American Chemical Society.

Mechanism of Organocuprate Reactions In the 1950’s Gilman described the generation of organocuprates “Gilman reagents” by combining, e.g., two equivalents of methyllithium with copper(I) iodide (44) (Figure 13). However, organocuprates did not develop into synthetically useful reagents until the pioneering work by House and Whitesides (conjugate addition) and Corey and Posner (alkylation) a decade later (45, 46). These reagents have since been widely employed into three classes of reactions including: (1) conjugate addition, (2) alkylation and (3) allylation. Copper(III) complexes had long been proposed as intermediates in all of these reactions and their existence had eluded chemists until, Bertz and Ogle initiated an RI-NMR study. It should be noted that this manuscript focuses on the RI-NMR investigations by Bertz and Ogle with the conjugate addition (47–54) reaction which shares intermediates observed during RI-NMR investigations in both the alkylation (55, 56) and allylation reactions (57–59).

117

Figure 13. Organocuprate preparation and reactions.

Mechanism of Conjugate Addition Reactions with Organocuprates In 2002, Bertz and Ogle investigated the reaction of 2-cyclohexanone (39) with Me2CuLi•LiI at –100 °C in THF by RI-NMR (47). Remarkably, two π-complexes 45 and 45•LiI are observed and are stable for hours at –100 °C, and only upon warming to –80 °C is 1,4-addition product 46 observed (Figure 14). The identity of complex 45•LiI was confirmed by adding 1.0 equiv of LiI to Me2CuLi•LiI prior to the injection of 39 which results in a increased ratio of 45•LiI relative to 45 (0.78 to 1.42). Furthermore, complex 45•LiI is the initial product at –100 °C (300 ms) (from 39 and Me2CuLi•LiI); only after the system has reached equilibrium (~20 s) is 45 observed. Moreover, these two π-complexes are be in equilibrium with one another by EXchange SpectroscopY (EXSY). Upon measuring the rates of formation of 45 and 45•LiI and using the obtained equilibrium data, the structure of Me2CuLi•LiI was suggested to be 47 and not a solvent separated ion pair 48 (The solution structures of Gilman reagents continue to be debated (see review (60) by Gschwind). Additionally, at higher ratios of LiI to Me2CuLi•LiI, the formation of addition product 46 is seen even at –100 °C suggesting that the LiI is incorporated in the complex that undergoes the carbon-carbon bond forming event. The π-complexes described above are the first step in the mechanism proposed for the conjugate addition reaction. The following step invokes an elusive copper(III) complex (Figure 14) (48). With the help of RI-NMR Bertz and Ogle found that upon the injection of (CH3)3SiCN into a freshly generated solution of π-complexes 45 and 45•LiI an η1-Cu(III) complex 49 is observed. Complex 49 is also stable (i.e. no product formation) at –100 °C allowing for it to be fully characterized by 1D and 2D-NMR spectroscopy (48). To assure that the CN group is indeed bound to the copper atom in complex 49 the authors prepared 13(CH3)2Cu•Li13CN (from Cu13CN and 2.0 equiv of 13CH3Li) and reacted it with 118

39. The observation of 13C coupling across the square planar copper atom reveals that the CN is bound (Figure 15). Warming the solution of 49 to –80 °C results in the direct formation of reductive elimination product 46.

Figure 14. Mechanism of conjugate addition reaction

Figure 15. Selected 13C NMR traces for labeled 49 (+ solvent peaks). The bottom spectra are for RCu(13CH3)213CNLi. The upper spectrum is for RCu(CH3)213CNLi. Reproduced from (48). Copyright 2007, American Chemical Society.

Interestingly, Bertz and Ogle found that the addition of (CH3)3SiCl into a solution of a chalcone π-complex 51 (from 50 and Me2CuLi•LiI) does not result in the observation of a η1-Cu(III) complex but rather an η3-Cu(III) complex 52 (Figure 16) (50). However the η3-Cu(III) complex 52 does convert to exclusively one η1-Cu(III) complex 53 after the addition of methyllithium. This type of intermediate had also been long proposed but never detected (61). 119

Figure 16. Observation of η2-Cu(III) complexes, η3-Cu(III) complexes η1-Cu(III) complexes in the conjugate addition reaction (50)

Mechanism of Conjugate Addition Reactions with Mixed Organocuprates Although organocuprates are widely used in organic synthesis these reagents are used in stoichiometric amounts and only one of the alkyl groups can be transferred from the organocuprate. Atom economy becomes a serious issue if the alkyl group is precious. In the early 1970’s Corey (62) and House (63) demonstrated that a 1-pentynyl group served as a non-transferable ligand in the conjugate addition and alkylation reactions. Non-transferable ligands “dummy ligands” (RNT) such as phenylthio (64), thienyl (65), dialkylamido (66), diarylphosphido (67), cyano (68), and trimethylsilylmethyl groups (69) have since been demonstrated to be effective. In 2012, Bertz and Ogle found that mixed organocuprates (MeCuRNT) formed π-complexes with various CC and CS double bonded groups in which a remarkable orientation effect is observed in which the transferable group is cis to the β-position on the enone (53). Specifically, the combination of chalcone 50 with a solution of MeCuSPhLi 54 “Posner’s cuprate” at –100 °C, π-complex 55, was observed in equilibrium with the starting materials (Figure 17). The authors report that upon warming the solution of π-complex 55 to –80 °C the corresponding homoorganocuprate π-complex 56 is observed. Through kinetic analysis it was found that the exchange involved a second-order reaction of the mixed cuprate in respect to the mixed-olefin complex (k1 = 0.016 min−1 mM−2, k−1 = 3 min−1 mM−2, k2 = 0.0035 min−1 mM−2) (Figure 17) (51).

120

Figure 17. Conversion of mixed π-complex 55 (□) to homo π-complex 56 (○). Chalcone is represented by (▵). Posner’s cuprate is represented by (◊). Reproduced from (51). Copyright 2012, American Chemical Society.

Cram Chelates; Are They Real? In 1959, Cram and Kopecky rationalized that the high diastereoselectivity observed during reactions between organometallic reagents and chiral, α-alkoxy ketones was the result of a chelation event “Cram-Chelate model” between both the carbonyl and alkoxy substituents with the organometallic reagent (Figure 18) (10).

Figure 18. Cram’s chelate rule.

121

During this time several investigations demonstrated that chelates are indeed formed between Lewis acids and α-alkoxy carbonyl compounds (70). However, these experiments were performed under equilibrium conditions which did not reveal if chelates are present in the transition state (see ref. (4) for a review on the Curtin-Hammett Principle). To unambiguously determine if chelates or if a nonchelated Lewis acid adduct is present during the carbon-carbon forming event a detailed kinetic study was needed (Figure 19).

Figure 19. Reaction coordinate for nonchelated vs. chelated pathways (16).

In 1987 Eliel and co-workers initiated an RI-NMR kinetic investigation between Me2Mg and α-alkoxy ketones which was reported in full in 1992 (13–16). Because Lewis acids are known to activate carbonyl groups to accept nucleophiles a simple Lewis acid adduct that does not form a chelate could be responsible for the carbon-carbon forming event. To determine which pathway is operative the rates for both pathways needed to be determined. The rates of product formation from the reaction of 2-hexanone 64 and α-silyloxy ketone 62 with Me2Mg in THF at –70 °C are (0.47 ± 0.01) x 102 M–1 s–1 and (0.45 ± 0.04) x 102 M–1 s–1, respectively (Table 1). The similarity of the rate constants for reactions of 62 and 64 suggests that silyloxy ketone 62 is not capable of forming a chelate. Moreover, the diastereomeric composition of the product from ketone 62 was 42:58 (chelated/nonchelated) signifying that the faces of the ketone are not effectively biased by the α-silyloxy group. Next, the rates of reactions of α-alkoxy ketones 57-60 along with the diastereomeric ratio of the products were measured (Table 1). The authors were able to derive a quantitative relationship between the reactivity and stereoselectivity based on the following assumptions. Assuming that the nonchelated rate of knonchelated = (0.45 ± 0.04) x 102 M–1 s–1 is the same for all of the ketones used in study and that kchelated =kobs – knonchelated. Further assuming that the nonchelated pathway gives rise to 42/58 chelate/ nonchelated product from ketone 62 and the perfect chelate pathway always gives a 100% (i.e. ketone 57). On the basis of the measured rate data and the fact that no appreciable pre-equilibrium exists between the chelated and nonchelated pathways then the relative amounts of the products should only depend on the activation energy of each competing pathway. Thus, the diastereoselectivity of the reactions could be predicted based on the obtained rate data using eq 1 (16). 122

Remarkably, the measured product distributions match the derived values from the kinetic analysis. This analysis was crucial to unambiguously determine that chelates are involved in the transition state. Although Eliel and co-workers could not detect an actual chelate intermediate during the RI-NMR experiments, the kinetic investigation provided substantial evidence. However, Reetz and co-workers subsequently performed a similar RI-NMR investigation in which chelates could be observed and this study is discussed below (24, 71–74). Reetz and co-workers demonstrated the first direct observation of a “Cram chelate” by combining ethyl ketone 64 with MeTiCl3 in dichloromethane at –45 °C where two diastereomeric complexes 65 and 66 are observed by 13C RI-NMR spectroscopy (Figure 20, 1) (72). Similarly, injection of aldehyde 68 into a solution of MeTiCl3 also results in the observation of two diastereomeric complexes 69 and 70 (Figure 20, 2) (74). However the ability to observe an intramolecular methyl transfer from the titanium to the carbonyl group could not be unambiguously proven. The RI-NMR investigations by Eliel and Reetz unambiguously demonstrate that chelates are intermediates during the addition reactions described, and are responsible for the diastereoselectivities observed.

Table 1. Analysis for Cram chelates. Reproduced from (16). Copyright 1992, American Chemical Society.

123

Figure 20. Reaction coordinate for nonchelated vs. chelated pathways (71, 74). Lewis Base Catalyzed Aldol Reactions The “traditional” aldol reaction involves the union of two carbonyl compounds that have the ability to form two new adjacent stereogenic centers (75). Historically, the aldol reaction was executed by combining two different carbonyl compounds (aldehydes or ketones) along with a catalytic amount of an acid or base that frequently resulted in mixtures of homo- and cross-aldol products. An elegant solution to this problem of chemoselectivity was reported in 1973 by Mukaiyama and co-workers. These authors demonstrated that silyl enol ethers could serve as enolate surrogates in reactions with aldehydes in the presence of a Lewis acid (76, 77). This process enabled the development of catalytic enantioselective variants with chiral Lewis acids (78). However, the Mukaiyama aldol reaction is stereoconvergent and the control of relative diastereoselectivity was variable. In the late 1990’s Denmark and co-workers introduced the concept of Lewis base activation of Lewis acids to address this shortcoming (26). One of the first manifestations of the concept was the introduction of trichlorosilyl enol ethers as aldol nucleophiles in combination with chiral Lewis bases (79–82). In this variant, the silicon serves as an organizational center to enable closed transition states that could control both diastereo- and enantioselectivity. For example, phosphoramide catalysts 72 and 73 proved to be excellent promoters for the enantioselective aldol addition reaction between cyclohexanone trichlorosilyl enol ether 74 and benzaldehyde 26 (Figure 21) (83). The N-methyl phosphoramide catalyst 72 affords high enantioselectivities (96:4) and anti-diastereoselectivity (99:1) whereas the N-phenyl phosphoramide catalyst 73 provides modest enantioselectivity (76.5:23.5) but with opposite syn-diastereoselectivity (99:1). The origins for the observed change in diastereoselectivity rests on the hypothesis that two of the smaller N-methyl phosphoramide catalysts 72 are present during the carbon-carbon bond forming event (i.e. chair-like transition state 75) while the later, boat-like transition state 76 incorporates only one of the larger N-phenyl phosphoramide catalysts 73. To substantiate this hypothesis, the Denmark 124

RI-NMR apparatus was constructed and deployed. The outlined proposal was elegantly shown by demonstrating that the order in catalysts (Lewis base) 72 and 73 to be 2nd and 1st order respectively confirming the proposal (84, 85).

Figure 21. (Top) Reaction and proposed transition structures for the reaction of 74 with 26. (Bottom) Various chiral Lewis bases used in the aldol reactions.

In these reactions the Lewis base serves two roles, to activate the nucleophile and the electrophile in an organized, closed transition state. Another variation of the Mukiayama aldol addition was developed to employ the more common and easily handled trialkylsilyl enol ethers in which the weak Lewis acid silicon tetrachloride is activated by chiral Lewis bases. Preliminary kinetic analysis of the reaction reveal second-order dependence on concentration of catalyst 79 which led to the development of chiral bis-phosphoramide catalysts such as 80 (86, 87). These catalysts provide high enantioselectivity and unprecedented, stereoconvergent anti-diastereoselectivity. The development and optimization of the silicon tetrachloride/silyl ketene acetal aldol reaction proved to be more challenging which led to a substantial amount of mechanistic work in which rapid injection NMR played a crucial role during catalyst development (88). 125

A survey of various phosphoramide Lewis bases (See Figure 21 above) produced a modest range of er values in the reaction between 1-napthaldehyde 82 and 83 (Figure 22a). To gain mechanistic insight into this transformation the rate equation was required. Therefore, partial orders in all of the components in the reaction were determined by 1H RI-NMR. Not surprisingly, the reaction is first order in both substrates (82 and 83) and zero order with respect to silicon tetrachloride. The lack of a rate dependence on silicon tetrachloride is expected because the active silicon-Lewis-base catalyst is saturated. However, the order in Lewis base hexamethylphosphoramide 81 resulted in a partial order of 2/3. This result is surprising and led to further order determinations with other Lewis bases such as 79, 80a, and 80c where 2nd, 1st and ½ orders are observed respectively (Figure 22a). Because these results are ambiguous without the knowledge of the catalysts resting states they were explored next with 29Si NMR (Figure 22b). Interestingly, the combination of silicon tetrachloride with HMPA affords a silicon adduct 85 that incorporates three molecules of HMPA. Moreover, silicon tetrachloride is found to incorporate two molecules of phosphoramide 79 creating a proposed resting state 86. Furthermore, bis-phosphoramide 80a serves as a chelate while catalyst 80c forms a dimeric species (i.e. 280c/2SiCl4). The analysis of the combined data allowed a generalized mechanism to be proposed where the active silicon Lewis base catalyst incorporates two Lewis base units “P=O” (Figure 22c). For example, the 2/3 partial order observed with HMPA can be explained because the resting state must lose a molecule of HMPA to form the active catalyst species, such as 89. This generalized mechanism is also found to be amenable to Lewis base catalyzed meso-epoxide openings. Is should also be noted that various RI-NMR investigations with lithium enolate aldol additions have been performed by Reich (89). Transmetalation from Tin to Lithium Organolithium compounds have served as work horse reagents in organic chemistry for many years (90). While these reagents are effective they are also highly reactive making them difficult to isolate in pure form. Therefore, methods have been developed to generate highly functionalized organolithium reagents in situ, which include “tin-lithium” exchange reactions (91). Unlike organolithium reagents, organotin compounds are often air and water stable and in certain cases can even be distilled or subjected to column chromatography (92, 93). Typically the reagent of choice to perform the tin-lithium transmetalation reaction is n-BuLi. This method is attractive because the only by-product is a tetraalkyltin compound, which is inert and non-polar and thus easily removed from reaction mixtures. In 2007 Klein and Gawley became interested in why stannylpiperidine 94 under goes transmetalation whereas 95 is virtually unreactive (Figure 23) (94). As transmetalation reactions are fast even at low temperatures (i.e. −78 °C), they utilized 119Sn RI-NMR to perform a detailed kinetic study to determine which conformation results in favorable transmetalation from various stannylpiperidines. Following the 119Sn RI-NMR injection of n-BuLi into a THF/Et2O solution of 96 at −82.15 °C provides a transmetalation rate of (5.4 ± 0.2 x 10–4) s–1 (Figure 23). Interestingly, the rate for product formation from the injection of n-BuLi into a 126

THF/Et2O solution compound 94 (equatorial tin) is marginally increased (14.8 ± 0.1 x 10–4) s–1 compared to 96.

Figure 22. (A) Reaction of 82 with 83 as catalyzed by SiCl4 with various Lewis bases. (B) Proposed resting state structures of various Lewis bases adducts with SiCl4. (C) Generalized mechanism for the Lewis base catalyzed aldol reaction.

Figure 23. Transmetalation from organotin to organolithium (94). 127

Finally, complex 95 (axial tin) affords a rate of (0.3 ± 0.09 x 10–4) s–1 indicating that the axial tin substituent undergoes slower transmetalation. The observed reactivity trend of 94 > 96 > 95 suggests that conformationally labile system 96 (equal distribution of 96a and 96b) undergoes transmetalation through conformer 96a, as understood in terms of Winstein-Holness kinetic behavior (4). Transmetalation from Boron to Palladium The formation of sp2-sp2 carbon-carbon bonds is of profound importance in view of the ubiquity of this connection in both synthetic and naturally occurring substances. This fundamental process in organic synthesis is frequently accomplished by the combination of organic nucleophiles and electrophiles with a metal catalyst or organometallic reagent. Specifically, palladium has emerged as an excellent metal catalyst for an entire class of reactions known as cross-coupling reactions (95). These reactions have fundamentally changed the practice of organic synthesis in both the academic and industrial settings alike. Among these, the Nobel prize sharing Suzuki-Miyaura cross-coupling reaction is the premier process (96). Interestingly the mechanisms for all of palladium catalyzed cross-coupling reactions share three fundamental steps (Figure 24). While the oxidative addition and reductive elimination steps are common to all cross-coupling variants only the transmetalation step preparatively and mechanistically differentiates these processes. Interestingly, the identity of the often-invoked pre-transmetalation species containing a Pd-O-B linkage had remained elusive until an RI-NMR investigation was initiated in these laboratories (19, 20).

Figure 24. Proposed Suzuki-Miyaura reaction mechanism (19). Last year we demonstrated for the first time that pre-transmetalation intermediates containing a Pd-O-B linkage are active in the Suzuki-Miyaura reaction. By combining trans-(i-Pr3P)2(4-FC6H4)Pd(OH) (106) with 4fluorophenylboronic acid (107) in THF at 30 °C with an additional 2.0 equiv 128

of i-Pr3P we observed complex 108 which could be fully characterized by 1D and 2D NMR spectroscopy with a NOESY experiment establishing the bonding connectivity (Figure 25). The sterically congested environment about palladium (two i-Pr3P ligands) results in the elimination of a molecule of water creating the observed tri-coordinate boron center. Thus, to enable the saturation of the boron valences the removal of a i-Pr3P ligand was required.

Figure 25. Formation of Pd-O-B linkage 108 from 106 and 107 (19). This hypothesis led to the investigation of [(i-Pr3P)(4-FC6H4)Pd(OH)]2 (109) with 4-fluorophenylboronic acid (107) (1.0 equiv/Pd) at –100 °C which allowed a bis-arylpalladium boronate complex 110 to be detected (Figure 26, 1). Interestingly, the formation of a 1:1 Pd/B complex, such as 111, was not successful upon the addition (2.0 equiv/Pd) of boronic acid to 107. The inability to form a 1:1 Pd/B stoichiometry is attributed to the thermochemical preference for the Pd-(µOH)-Pd moieties, which is observed in other mixed-hydroxide complexes (97). However, the RI-NMR injection of methanol into a solution of 110 with 1.0 equiv of 107 results in the formation of a 1:1 (Pd/B) complex 111 (Figure 26, 2).

Figure 26. Formation of Pd-O-B linkages 110 and 111 (19) 129

The ability to form intermediates containing Pd-O-B linkages provided the unique opportunity to probe the requirements for transmetalation. Generation of complex 111 at −30 °C in THF/CH3OH leads to the formation of cross-coupling product 113 (Figure 27). The similarity of rate constants suggests that the ratedetermining step is the transfer of the organic fragment from boron to palladium. Next the transfer of the bis-arylpalladium boronate complex 110 was explored where a first order-decay and formation of cross-coupling product 113 is observed. The similarity of the rate constants for both the THF and THF/CH3OH reactions suggest that complex 110 forms a 1:1 Pd/B adduct 112 prior to transmetalation.

Figure 27. Kinetic analysis for product formation from intermediates 107 and 112 containing a single phosphine ligand (19). Finally complex 108 was found to be thermally stable at −30 °C; however, warming to 20 °C results in the formation of cross-coupling product 113 (Figure 28). The decay of complex 108 displayed S-shaped kinetics which were indicative of an auto-catalytic process. Additionally, an inverse order in i-Pr3P was determined which indicates that a prior dissociation event is required before the transmetalation event can take place. These combined kinetic results indicate that a requirement for transmetalation is an empty coordination site on the palladium atom.

Figure 28. Kinetic analysis of tricoordinate boron intermediate 108. 130

Final Remarks Ernest L. Eliel is widely considered to be the godfather of organic stereochemistry in the United States. His 1962 publication of “The Stereochemistry of Organic Compounds” was the first comprehensive treatise dedicated to the advanced topic and it served as the “Bible” of stereochemistry for generations of organic chemists. Ernest was passionate about stereochemistry and was one of the founding editors of the landmark series “Topics in Stereochemistry” which was dedicated to chronicling the most significant advances in the field. He was also one of the driving forces behind the creation of the Gordon Research Conference on Stereochemistry which started in 1975 (coincidentally the same year that the Nobel Prize was awarded to Vladimir Prelog and John W. Cornforth for studies in stereochemistry). Ernest’s love of stereochemistry, his commitment to scholarship and service to the chemical community and his enduring connection to the University of Illinois served as a powerful example to one of us (S.E.D.) of what it meant to be a leader in academic science. We would do well to have more “Eliels” around today.

References 1.

2. 3. 4. 5. 6. 7. 8. 9.

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

Eliel, E. L. From Cologne to Chapel Hill. In Profiles, Pathways, and Dreams. Autobiographies of Eminent Chemists; Seeman, J. I., Ed.; American Chemical Society: Washington, DC, 1990. Eliel, E. L. Experientia 1953, 9, 91–93. Read, J.; Grubb, W. J. J. Chem. Soc. 1934, 1779–1783. Seeman, J. I. Chem. Rev. 1983, 83, 83–134. Becker, E. D. Anal. Chem. 1993, 65, 295A–302A. Corey, E. J.; Seebach, D. Angew. Chem., Int. Ed. Engl. 1965, 4, 1075–1077. Eliel, E. L.; Hartmann, A. A. J. Am. Chem. Soc. 1971, 93, 2572–2573. Eliel, E. L. Phosphorus Sulfur 1985, 24, 73–95. Eliel, E. L. From Cologne to Chapel Hill. In Profiles, Pathways, and Dreams. Autobiographies of Eminent Chemists; Seeman, J. I., Ed.; American Chemical Society: Washington, DC, 1990; p 86. Cram, D. J.; Kopecky, K. R. J. Am. Chem. Soc. 1959, 81, 2748–2755. Eliel, E. L. In Asymmetric Synthesis; Morrison, J. D., Ed.; Academic Press: New York, 1983; Vol. 2, Chapter 5. Cherest, M.; Felkin, H.; Prudent, N. Tetrahedron Lett. 1968, 2199–2204. Frye, S. V.; Eliel, E. L.; Cloux, R. J. Am. Chem. Soc. 1987, 109, 1862–1863. Chen, X.; Hortelano, E. R.; Eliel, E. L.; Frye, S. V. J. Am. Chem. Soc. 1990, 112, 6130–6131. Eliel, E. L.; Frye, S. V.; Hortelano, E. R.; Chen, X.; Bai, X. Pure Appl. Chem. 1991, 63, 1591–1598. Chen, X.; Hortelano, E. R.; Eliel, E. L.; Frye, S. V. J. Am. Chem. Soc. 1992, 114, 1778–1784. McGarrity, J. F.; Prodolliet, J.; Smyth, T. Org. Magn. Reson. 1981, 17, 59–65. 131

18. Denmark, S. E.; Williams, B. J.; Eklov, B. M.; Pham, S. M.; Beutner, G. L. J. Org. Chem. 2010, 75, 5558–5572. 19. Thomas, A. A.; Denmark, S. E. Science 2016, 352, 329–332. 20. Thomas, A. A.; Wang, H.; Zahrt, A. F.; Denmark, S. E. J. Am. Chem. Soc. 2017, 139, 3805–3821. 21. McGarrity, J. F.; Prodolliet, J. W. Tetrahedron Lett. 1982, 23, 417–420. 22. McGarrity, J. F.; Prodolliet, J. J. Org. Chem. 1984, 49, 4465–4470. 23. Jones, A. C.; Sanders, A. W.; Bevan, M. J.; Reich, H. J. J. Am. Chem. Soc. 2007, 129, 3492–3493. 24. Reetz, M. T.; Raguse, B.; Marth, C. F.; Huegel, H. M.; Bach, T.; Fox, D. N. A. Tetrahedron 1992, 48, 5731–5742. 25. Mok, K. N.; Nagashima, T.; Day, I. J.; Jones, J. A.; Jones, C. J. V.; Dobson, C. M.; Hore, P. J. J. Am. Chem. Soc. 2003, 125, 12484–12492. 26. Denmark, S. E.; Beutner, G. L. Angew. Chem., Int. Ed. 2008, 47, 1560–1638. 27. Reich, H. J. Chem. Rev. 2013, 113, 7130–7178. 28. Degennaro, L.; Giovine, A.; Carroccia, L.; Luisi, R. In Lithium Compounds in Organic Synthesis; Luisi, R., Capriati, V., Eds.; Wiley-VCH: Weinheim, 2014; Chapter 18. 29. Margerison, D.; Newport, J. P. Trans. Faraday Soc. 1963, 59 (Pt. 9), 2058–2063. 30. McGarrity, J. F.; Ogle, C. A. J. Am. Chem. Soc. 1985, 107, 1805–1810. 31. McGarrity, J. F.; Ogle, C. A.; Brich, Z.; Loosli, H. R. J. Am. Chem. Soc. 1985, 107, 1810–1815. 32. Seebach, D.; Hassig, R.; Gabriel J. Helv. Chim. Acta 1983, 66, 308–337. 33. Jones, A. C.; Sanders, A. W.; Sikorski, W. H.; Jansen, K. L.; Reich, H. J. J. Am. Chem. Soc. 2008, 130, 6060–6061. 34. Plessel, K. N.; Jones, A. C.; Wherritt, D. J.; Maksymowicz, R. M.; Poweleit, E.; Reich, H. J. Org. Lett. 2015, 17, 2310–2313. 35. Reich, H. J. J. Org. Chem. 2012, 77, 5471–5491. 36. Jones, A. C. In Spectroscopic Advances in Organolithium Reactivity: the Contribution of Rapid-Injection NMR (RINMR); Wiley-VCH: Weinheim, 2014; Chapter 3. 37. Scott, N. D. U.S. Patent, 2181,777, 1939. 38. Stavely, F. W. Ind. Eng. Chem. 1956, 48, 778–783. 39. Maul, J., Frushour, B. G., Kontoff, J. R., Eichenauer, H., Ott, K.-H.; Schade, C. Polystyrene and Styrene Copolymers. Ullmann’s Encyclopedia of Industrial Chemistry; Wiley: 2007; pp 475−522. 40. Hsieh, H. L. J. Polym. Sci., Part A 1965, 3, 163–172. 41. Bauer, W.; Winchester, W. R.; Schleyer, P. v. R. Organometallics 1987, 6, 2371–2379. 42. Ogle, C. A.; Johnson, H. C. I. V.; Wang, X. L.; Strickler, F. H.; Bucca, D.; Gordon, B. , III Macromolecules 1995, 28, 5184–5191. 43. Ogle, C. A.; Wang, X. L.; Carlin, C. M.; Strickler, F. H.; Gordon, B. , III J. Polym. Sci., Part A: Polym. Chem. 1999, 37, 1157–1168. 44. Gilman, H.; Jones, R. G.; Woods, L. A. J. Org. Chem. 1952, 17, 1630–1634. 45. House, H. O.; Respess, W. L.; Whitesides, G. M. J. Org. Chem. 1966, 31, 3128–3141. 132

46. Corey, E. J.; Posner, G. H. J. Am. Chem. Soc. 1968, 90, 5615–1516. 47. Bertz, S. H.; Carlin, C. M.; Deadwyler, D. A.; Murphy, M. D.; Ogle, C. A.; Seagle, P. H. J. Am. Chem. Soc. 2002, 124, 13650–13651. 48. Bertz, S. H.; Cope, S.; Murphy, M.; Ogle, C. A.; Taylor, B. J. J. Am. Chem. Soc. 2007, 129, 7208–7209. 49. Bertz, S. H.; Moazami, Y.; Murphy, M. D.; Ogle, C. A.; Richter, J. D.; Thomas, A. A. J. Am. Chem. Soc. 2010, 132, 9549–9551. 50. Bertz, S. H.; Hardin, R. A.; Murphy, M. D.; Ogle, C. A.; Richter, J. D.; Thomas, A. A. J. Am. Chem. Soc. 2012, 134, 9557–9560. 51. Bertz, S. H.; Browder, K. L.; Hardin, R. A.; Murphy, M. D.; Ogle, C. A.; Thomas, A. A. Organometallics 2012, 31, 7809–7811. 52. Bertz, S. H.; Cope, S. K.; Hardin, R. A.; Murphy, M. D.; Ogle, C. A.; Smith, D. T.; Thomas, A. A.; Whaley, T. N. Organometallics 2012, 31, 7827–7838. 53. Bertz, S. H.; Hardin, R. A.; Murphy, M. D.; Ogle, C. A.; Richter, J. D.; Thomas, A. A. Angew. Chem., Int. Ed. 2012, 51, 2681–2685. 54. Bertz, S. H.; Hardin, R. A.; Murphy, M. D.; Ogle, C. A. Chem. Commun. 2013, 49, 3010–3012. 55. Bertz, S. H.; Cope, S.; Dorton, D.; Murphy, M.; Ogle, C. A. Angew. Chem., Int. Ed. 2007, 46, 7082–7085. 56. Bartholomew, E. R.; Bertz, S. H.; Cope, S.; Dorton, D. C.; Murphy, M.; Ogle, C. A. Chem. Commun. 2008, 10, 1176–1177. 57. Bartholomew, E. R.; Bertz, S. H.; Cope, S.; Murphy, M.; Ogle, C. A. J. Am. Chem. Soc. 2008, 130, 11244–11245. 58. Bartholomew, E. R.; Bertz, S. H.; Cope, S. K.; Murphy, M. D.; Ogle, C. A.; Thomas, A. A. Chem Commun. 2010, 46, 1253–1254. 59. Bertz, S. H.; Murphy, M. D.; Ogle, C. A.; Thomas, A. A. Chem. Commun. 2010, 46, 1255–1256. 60. Gschwind, R. M. Chem. Rev. 2008, 108, 3029–3053. 61. Corey, E. J.; Boaz, N. W. Tetrahedron Lett. 1985, 26, 6015–6018. 62. Corey, E. J.; Beames, D. J. J. Am. Chem. Soc. 1972, 94, 7210–7211. 63. House, H. O.; Umen, M. J. J. Org. Chem. 1973, 38, 3893–3901. 64. Posner, G. H.; Ting, J.-S. Synth. Commun. 1974, 4, 355–366. 65. Lindstedt, E. L.; Nilsson, M. Acta Chem. Scand., Ser. B 1986, B40, 466–469. 66. Ghozland, F.; Luche, J. L.; Crabbe, P. Bull. Soc. Chim. Belg. 1978, 87, 369–374. 67. Bertz, S. H.; Dabbagh, G. J. Chem. Soc., Chem. Commun. 1982, 18, 1030–1032. 68. Gorlier, J. P.; Hamon, L.; Levisalles, J.; Wagnon, J. J. Chem. Soc., Chem. Commun. 1973, 3, 88. 69. Bertz, S. H.; Eriksson, M.; Miao, G.; Snyder, J. P. J. Am. Chem. Soc. 1996, 118, 10906–10907. 70. Keck, G. E.; Castellino, S. Tetrahedron Lett. 1987, 28, 281–284. 71. Reetz, M. T.; Huellmann, M.; Seitz, T. Angew. Chem. 1987, 99, 478–480. 72. Reetz, M. T. Angew. Chem. 1984, 96, 542–555. 73. Reetz, M. T. Acc. Chem. Res. 1993, 26, 462–468. 74. Reetz, M. T.; Raguse, B.; Seitz, T. Tetrahedron 1993, 49, 8561–8568. 133

75. Nielsen, A. T.; Houlihan, W. J. Org. React. 2011, 16, 1–438. 76. Mukaiyama, T.; Inomata, K.; Muraki, M. J. Am. Chem. Soc. 1973, 95, 967–968. 77. Mukaiyama, T.; Narasaka, K.; Banno, K. Chem. Lett. 1973, 9, 1011–1014. 78. Carreira, E. M.; Fettes, A.; Marti, C. Org. React. 2006, 67, 1–216. 79. Denmark, S. E.; Fujimori, S. In Modern Aldol Reactions; Mahrwald, R., Ed.; Wiley-VCH: Weinheim, 2004; Vol. 2, Chapter 7. 80. Fu, J.; Fujimori, S.; Denmark, S. E. In Lewis Base Catalysis in Organic Synthesis; Vedejs, E., Denmark S. E., Eds.; Wiley-VCH: Weinheim, 2016; Vol. 1, Chapter 9. 81. Denmark, S. E.; Stavenger, R. A. Acc. Chem. Res. 2000, 33, 432–440. 82. Denmark, S. E.; Fujimori, S.; Pham, S. M. J. Org. Chem. 2005, 70, 10823–10840. 83. Denmark, S. E.; Winter, S. B. D.; Su, X.; Wong, K.-T. J. Am. Chem. Soc. 1996, 118, 7404–7405. 84. Denmark, S. E.; Pham, S. M. Helv. Chim. Acta 2000, 83, 1846–1853. 85. Denmark, S. E.; Pham, S. M.; Stavenger, R. A.; Su, X.; Wong, K.-T.; Nishigaichi, Y. J. Org. Chem. 2006, 71, 3904–3922. 86. Beutner, G. L.; Denmark, S. E. Angew. Chem., Int. Ed. 2013, 52, 9086–9096. 87. Rossi, S.; Denmark, S. E. In Lewis Base Catalysis in Organic Synthesis; Vedejs, E., Denmark S. E., Eds.; Wiley-VCH: Weinheim, 2016; Vol. 3, Chapter 21. 88. Denmark, S. E.; Eklov, B. M.; Yao, P. J.; Eastgate, M. D. J. Am. Chem. Soc. 2009, 131, 11770–11787. 89. Kolonko, K. J.; Wherritt, D. J.; Reich, H. J. J. Am. Chem. Soc. 2011, 133, 16774–16777. 90. Degennaro, L.; Giovine, A.; Carroccia, L.; Luisi, R. Practical Aspects of Organolithium Chemistry. In Lithium Compounds in Organic Synthesis; Luisi, R., Capriati, V., Eds.; Wiley-VCH: Weinheim, 2014; Chapter 18. 91. Pereyre, M.; Quintard, J.-M.; Rahm, A. Tin in Organic Synthesis; Butterworths: London, 1987; pp 149−184 92. Gielen, M. Tin chemistry: fundamentals, frontiers, and applications; Wiley: Hoboken, 2008. 93. Farina, V.; Krishnamurthy, V.; Scott, W. J. Org. React. 1997, 50, 1–652. 94. Klein, R.; Gawley, R. E. J. Am. Chem. Soc. 2007, 129, 4126–4127. 95. Metal-Catalyzed Cross-Coupling Reactions and More, 3rd ed.; de Meijere, A., Bräse, S., Oestreich, M., Eds.; Wiley-VCH: Weinheim, Germany, 2014; Vols. 1−3. 96. Suzuki, A. Angew. Chem., Int. Ed. 2011, 50, 6722–6737. 97. Driver, M. S.; Hartwig, J. F. Organometallics 1997, 16, 5706–5715.

134

Chapter 9

Characterization of Materials with NMR Spectroscopy Cecil Dybowski* Department of Chemistry and Biochemistry, University of Delaware, Newark, Delaware 19716, United States *E-mail: [email protected]

NMR spectroscopy is a major characterization technique. With modern technology addressing the state of solids, it provides a tool for understanding the character and function of materials. Examples are given from work at the University of Delaware.

Introduction Beginning in the 1940s and continuing after World War II, an era of application of new technology to characterize substances brought a panoply of new ways to address the identities of molecules. Among the spectroscopic techniques to which chemists had access was the newly discovered technique of nuclear magnetic resonance (NMR) spectroscopy (1, 2). The dependence of NMR parameters on chemical state pointed the way to the ultimate importance of the technique for chemical analysis (3, 4). The use of NMR spectroscopy of protons revolutionized chemical analysis, particularly of organic materials (5). Later, the application of 13C NMR spectroscopy added a new dimension to organic analytical capabilities in the late 1960s and early 1970s. Over the decades, Professor Eliel contributed materially to the use of NMR spectroscopy, particularly 13C NMR spectroscopy, to study conformation and dynamics of organic molecules in solutions (6). In the 1960s and thereafter, chemists began to study other nuclei than 1H and 13C, which added to the versatility of analysis with NMR spectroscopy. Today the analytical capabilities of NMR spectroscopy are applied to problems involving a wide variety of chemical systems because of the ease of studying many different nuclei.

© 2017 American Chemical Society

In the early period, most applications of NMR spectroscopy by chemists involved either neat liquids or dilute solutions of solids. Although physicists had continued to study the NMR spectroscopy of solids since its discovery, chemists tended to focus on the NMR spectroscopy of liquids and solutions during that era. By the early 1970s, however, chemists had begun to study the chemical properties of solids with NMR methods (7–10). Examples began to be seen in the chemical literature of applications of NMR spectroscopy to technologically relevant solid-like materials, such as polymers, zeolites, and catalysts (11–18). The distinction, from the viewpoint of NMR analysis, between a solid and a solution is the time scale on which the major dynamics of constituent molecules/atoms occur. In a solution at room temperature, fast isotropic random motions on the sub-nanosecond time scale produce substantial dynamical averaging of magnetic interactions, with the result that spectra contain only dynamically averaged information. The relatively fixed positions of the confined molecules/atoms of a solid (or the anisotropic rapid motions) limit that averaging. Thus, NMR spectra of soluton and solid samples often appear noticeably different, even though they arise from the same interactions. Starting in the 1960s and continuing to the present, the use of computational techniques in chemistry also has had a considerable impact on NMR spectroscopic analyses. The prediction of NMR parameters such as magnetic shielding and spin-spin coupling from known or expected structures was at first limited by speed and memory of computers, but as the computational techniques became more efficient and reliable, the ability to characterize materials by comparing experimental results to theoretical predictions has added to the arsenal of techniques for understanding chemical identity with NMR spectroscopy. Very often, calculations of NMR properties have been performed on isolated-molecule models, with the extra-molecular local environment being treated, if at all, in some mean-field approximation. In this article, I describe the application of NMR spectroscopy to materials, which are naturally mostly solid or solid-like. Many materials are more readily characterized by NMR spectroscopy of nuclei other than protons and carbons. As a consequence, the spectroscopy appropriate to “unusual” nuclei and to solid-state structures is important in these experiments. These problems further demonstrate how, to make a computational prediction based on known structure, one must treat nuclei embedded in a solid, rather than in an isotropic fluid or an isolated molecule.

Solid-State NMR Spectroscopy With NMR spectroscopy, one determines the magnitudes of magnetic interactions of nuclei with their environments (19). The various interactions are specified by an expansion of the nuclear magnetic Hamiltonian, with terms appropriate to the interactions, as indicated in equation 1.

136

where the various terms represent typical interactions of the nuclear spins with the environment. The first is the Zeeeman coupling to the applied magnetic field, followed by the coupling to the radio-frequency magnetic field, the magnetic shielding of nuclei by coupling to electronic motion, the indirect spin-spin coupling, the coupling of the quadrupole moment of the nucleus (if it is non-zero) to the electric-field gradient, and the direct through-space dipole-dipole coupling of nuclear spins. In some cases, terms may be added to represent other interactions that may affect the NMR spectrum, but these in Eq. 1 are the common interactions typically affecting NMR spectroscopy. Each is characterized by a parameter or parameters that define the engagement of the nucleus with its environment, and it is these characteristic parameters for a particular material that give information on the state of the material. Not all terms in the Hamiltonian in Eq. 1 are accessible in all NMR experiments. In particular, the fast, isotropic random motion of liquids often results in averaging of effects for common experiments on solutions. The effects of the dipolar and quadrupolar terms average (in first order) to zero, and their effects are not apparent in first-order solution-state NMR spectra. Similarly, under fast, isotropic, random motion, the effects of other terms are limited only to those parts that are non-zero. The magnetic shielding and the spin-spin coupling only display isotropic average effects. (In experiments at sufficiently high magnetic fields, the influence of the field can be strong enough to encourage the rapid motion to be anisotropic, the result of which is reintroduction of the effects of averaged tensor properties such as magnetic shielding and dipolar coupling, but we shall not discuss these in this chapter.) To detect the effects of these anisotropic interactions, one must carry out the spectroscopy on nuclei whose nuclear environment is fixed during the time of the experiment, and that usually means examination of solids. There are two basic means of examining solids: (1) with a sample that is a single crystal; and (2) with a powder of microcrysallites. In the first case, one detects only signals of molecules at one or a few specific orientations of the molecular framework relative to the magnetic field; in the second case, one simultaneously detects the signals from molecules having many different orientations of the molecular axes relative to the magnetic field. In either case, one is able to extract information on the parameters describing the interactions, including how they depend on the structural relation between the molecular framework and the magnetic field. An example of the spectroscopy of a solid subject only to the full magnetic shielding Hamiltonian is given in Figure 1. The 13C NMR spectrum of a solid consisting of microcrystals of CO2 is a band, the shape of which is characteristic of materials having the nuclear center sitting on an axis of cylindrical symmetry. Each point on the spectrum corresponds, in this case, to a specific orientation of the cylinder axis relative to the magnetic field. The intensity at each point is determined by the number of ways in a random distribution of orientations that particular chemical shift is found. From this band, one determines two characteristic parameters of the interaction, the two unique principal components of the 13C chemical-shift tensor. (The magnetic-shielding tensor and the chemical-shift tensor are related, and differ by an additive constant, determined by the reference point for defining the chemical shift.) 137

Figure 1. Schematic depiction of the 13C NMR spectrum of a random powder of microcryalline CO2. At the top is shown a depiction of how molecules in the sample are oriented relative to the applied magnetic field, B0, for the two extreme chemical shifts.

For a general solid, all of the sub-Hamiltonians of Eq. 1 may be active. The spectrum may be much more complex than that of CO2, in which the magnetic shielding is, by far, the dominant effect. Such situations often call for more complex experiments to extract information. For example, a common organic material is likely to contain many more than a single carbon type. As a consequence, the spectrum of a static powdered sample consists of the overlap of the signals from the various sites, which may mean loss of information. Performing the spectroscopy with magic-angle spinning (MAS) causes the dispersion of the spectral line shape to be suppressed, resulting in a spectrum that is similar to that of a solution for identification of materials, as seen in Figure 2 for the 13C in α-glycine (11). One technical issue is that the 13C nuclei are usually strongly affected by the direct dipole-dipole interaction with adjacent protons in many organic materials. The spectrum in Figure 2 was not only taken while spinning the sample, but also with concurrent irradiation of the protons to suppress the effects of that direct dipole-dipole interaction, allowing observation of the average magnetic shielding that distinguishes each resonance from the other (dipolar decoupling, DD). 138

Figure 2. The 13C CP-MAS-DD NMR spectrum of α-glycine. There are two carbon resonances. The small resonances are the sidebands of the carboxyl resonance, a phenomenon that occurs in the MAS experiment. A second technical issue is the fact that the natural abundance of 13C is low. To enhance that 13C signal, one uses a technique in which the magnetic moment of the carbon is made larger by transfer of magnetization from the protons in a process called cross polarization (CP) (20). The spectrum in Figure 2 has been obtained by enhancement of the 13C magnetization. This process is made possible by the non-zero direct dipole-dipole interaction with the protons. Such experimental techniques are commonly used in NMR spectroscopy. There is another technical issue that must be resolved when obtaining NMR spectra of other nuclei such as 207Pb. For a nucleus like 13C the whole range of resonance positions spans only about 300 ppm. With current techniques, it is possible to excite all of the various 13C resonances with commonly used techniques using reasonably short pulses, but for nuclei like 207Pb that have ranges of several thousand ppm, that may not be possible (21). In these cases, it is necessary to design excitation protocols to take account of that facet of the spectroscopy. There are several ways to carry out excitation. For example, one can obtain the overall spectrum as a series of subspectra taken to ensure coverage of part of the NMR line shape. This technique is sometimes automated by summing responses with the so-called variable-offset-chemical-shift technique. One may attempt to excite over a much wider range than usual by incorporating elements into the excitation scheme that spread the energy. The success of the application of a technique largely depends on the specific situation. In Figure 3 are shown three 207Pb NMR spectra of a random powder of lead acetate. In Figure 3a, the spectrum is determined while spinning the sample about the axis canted at the magic angle to the magnetic field. The sideband pattern arises from the fact that the spinning rate is low compared to the width of the spectrum in hertz. From fitting of the intensity profile, one obtains the three principal components of the 207Pb chemical-shift tensor. In Figure 3b, the 139

spectrum is obtained by use of the wideband-uniform-rate-smooth-truncationwith-Carr-Purcell-Meiboom-Gill-detection (WURST-CPMG) experiment, which spreads the excitation to excite across a range of several thousand ppm uniformly. The envelope of the spikelets clearly indicates the three principal components of the chemical-shift tensor. In Figure 3c is shown the result of a similar experiment, called the Carr-Purcell-Meiboom-Gill (CPMG) experiment. In this spikelet spectrum, one sees excitation across a reasonably wide range, but the shape shows that the excitation is not as uniform as it is in the WURST-CPMG experiment of Figure 3b. Of note is the fact that the excitation trails off quickly in the region between -2800 and -3000 ppm, with the result that one might misinterpret the position of the sharp edge and, therefore misstate the value of the δ33 component. Additionally, the line shape in the CPMG experiment depends on where the central carrier frequency of the NMR experiment is set.

Figure 3. 207Pb NMR spectra of a powdered sample of lead acetate. (a) Taken with MAS spectrometry; (b) taken with the WURST-CPMG experiment; (c) taken with the CPMG experiment. For mild distortion of the excitation, one may characterize the excitation profile independently. Knowing the empirical distortion function, one may correct for the non-uniform excitation with the “transfer function” (22, 23). Although 140

the procedure has been shown to work well when the excitation profile is only slightly non-uniform, the use of wideband excitation experiments, such as the WURST-CPMG experiment, is preferred (24). Even with WURST-type excitation, if the NMR spectrum is sufficiently wide, the WURST excitation can be non-uniform over the spectral window. Under such conditions, one may take recourse to the idea of variable-offset-cumulativespectroscopy (25). The idea is straightforward. For very wide bands, one uses an experiment in which a portion of the data is taken with one setting of the carrier frequency. A second portion is taken with a different setting of the carrier frequency to excite another portion of the spectrum. That procedure is repeated until all areas of the NMR line shape are excited. Combining these subspectra into a single spectrum produces the full spectrum. The accumulation may be done before or after Fourier transformation. In our laboratory, we proceed after Fourier analysis of the individual subspectra, requiring matching of intensities at the overlap of two regions. There are many “tricks” used to obtain information in NMR spectroscopic experiments. The use of specific excitation sequences to suppress the action of certain parts of the Hamiltonian is typical, yielding a spectrum that emphasizes only part of the information available. The development and use of multidimensional NMR techniques allows one to correlate separate pieces of information about a system, such as how close two nuclei are with what sites they are. A discussion of the various ways in which experiments are designed to emphasize particular information or deal with the effects of unusual interactions could easily fill a book. The important point is that the parameters extracted from any spectrum give a picture of the state of the system. In the end, one must rationalize these bits of information with other ways of describing the material to provide as complete a description of the system under study as possible.

Computational Chemistry and NMR Spectroscopy One facet of NMR spectroscopy (as with other forms of analysis) is the assignment problem. One may detect NMR features, such as the two peaks in Figure 2 arising from the two different carbon sites of α-glycine. But, how does one a priori link a specific parameter (in this case, the chemical shift) to a specific site? In the early days (and to some extent, still today), assignment was carried out by analogy. The spectra of many different carboxylic acids with 13C NMR spectroscopy always showed an absorption peak in the region between 170 and 180 ppm relative to tetramethylsilane, which was assigned to a transition involving carbon nuclei in that environment. Over the years, many assignments have been made so that there are now libraries that show consistent trends with variation of environment. Such assignments are aided, in some cases, by observation of the spin-spin coupling to adjacent nuclei. Almost immediately after their development, two-dimensional NMR techniques for solutions had shown the benefits of correlation of various resonances through the spin-spin coupling. Whatever the NMR technology, the 141

assignment of resonance positions in unknown materials has been empirical, relying on analogies to NMR absorptions in molecules of similar known local structure, i.e. the “group effect”. NMR parameters are descriptors of the quantum mechanical state of a system consisting of nuclei and electrons. A knowledge of the quantum mechanical wave function of such a system allows one to relate those NMR descriptors to other parameters such as positions of atoms and strengths of bonds. The theory of the effects of electrons on NMR magnetic shielding was, for example, given only a few years after the discovery of NMR. In principle, calculation of magnetic shielding for known structures should be a means to define systems. Although there had been efforts to provide the link between the measured quantities and computationally derived quantities, the full capability of calculation had to await the development of fast, multiprocessor computers to begin to provide a meaningful implementation of calculation. For molecular systems, one may model the molecule as an isolated entity. In principle, this model of a gas-phase molecule works rather well for molecules in the gas phase or in dilute solution. If the environment outside of the molecule in solution is included, it is often treated in some mean-field approximation such as the conductor-like screening model (COSMO) or the polarizable continuum model (PCM), particularly when there is evidence of interaction with the solvent. For molecular solids, the nucleus-bearing molecule is embedded in a fixed structure of surrounding molecules, and the structure persists for a significant time, rather than being dynamically averaged to some sort of continuous distribution throughout the surrounding space. In this case, a proper model must include those intermolecular interactions when making predictions of the electronic state (from which one predicts the NMR parameters such as magnetic shielding). There are at least two ways to model such systems. The first type of model is based on the known periodicity of the solid lattice. A widely used method is the gauge-including projector augmented wave (GIPAW) approach of Mauri and Pickard in the density functional theory formalism (26). In this approach, the orbitals are expanded in a plane-wave basis that describes the lattice structure. Calculations of NMR parameters follow directly from the knowledge of the electron density as a function of the position. The second type is the cluster model (27–29). In this formalism, the lattice is described relative to the nucleus of interest. A cluster of atoms describing the environment is determined relative to the nucleus of interest. Ultimately, the cluster ends at some terminal atoms (usually forming a molecule for molecular solids). The important point is that NMR effects such as the magnetic shielding and the quadrupolar coupling are relatively short-range, and as one increases the size of the cluster modeling the environment, the predicted NMR parameters converge to a limit. Knowing the local symmetry of the nuclear site in a solid, one may choose as models clusters that have symmetry elements in common with the site in an infinite solid. For example, in Figure 4 is a cluster of 15 β-D-fructopyranose molecules which has the point symmetry elements at the central site in common with the space group P212121. Such symmetry-adapted structures are used to ensure that calculations reflect the known structural features of a solid object. 142

Figure 4. Cluster of 15 molecules of β-D-fructopyranose, such that the central point has the point symmetry of a similar point in the space group P212121. (Reproduced, with permission, from reference (29). Copyright 2017 by Sean T. Holmes.) In certain cases, the material is a network solid. In such materials, although the stoichiometry may be simple, e.g. PbO, the local structure of the material may have a general Pb site with strong interactions with more O sites than implied by the stoichiometry. Often, under these conditions, it is not possible to choose a truncation that creates an uncharged cluster and also maintains the point symmetry at the nuclear site. Approximations that allow one to carry out calculations must be made for such materials. By a judicious choice of computational method, including in some cases treatment of some or all electrons as relativistic objects (e.g. those associated with heavy nuclei like Pb or Hg), one may reliably predict the NMR properties of solid materials, essentially in agreement with measured NMR parameters. The combination of computational chemistry with NMR spectroscopy of solids allows one to specify the material and its structure to the point that some have suggested that NMR analysis with structure incorporation is a means of specifying structure, in a technique complementary to traditional diffraction-based techniques.

Examples of NMR Characterization of Materials Orientation in Polymeric Materials In processes like extrusion, extension, or compression, a material that originally had no macroscopic orientation of molecular or crystallographic axes 143

can commonly be induced to form an oriented material. Whether the orientation is transitory or permanent or what the extent of the orientation turns out to be depends on many variables. Thus, measures of the orientation of molecular structures in such a macroscopic sample are aids in determining the effects of these processes. One may address the question of the orientation with a variety of structure-sensitive techniques, such as X-ray analysis or electron microscopy or optical birefringence. The strong dependence of the NMR response on the orientation of a molecule in the magnetic field provides a means to use NMR to infer the distribution of oirentations in an oriented material (30). The band shape of a nucleus in a polycrystalline material, such as that of CO2 in Figure 1, depends on the manner in which the principal axes of the chemical-shift tensor (which specify how the position of resonance depends on the orientation of molecular axes relative to the direction of the magnetic field) are laid out in space. The schematic band shape in Figure 1 results when there is a random distribution of these molecular axes for CO2 relative to the magnetic field direction. Should the sample be macroscopically oriented, the shape of the band is distorted from that ideal shape in a manner characteristic of the orientation process. By careful analysis of the band shape and how it depends on the orientation of macroscopic sample axes to the magnetic field, one may specify the orientation distribution of the chemical-shift axes relative to the macroscopic axes (which are generally determined by how forces are applied in the process that produces orientation). As an example, Figure 5 shows two theoretical spectra. The chemical-shift tensor has only two unique principal components in this case. In (a), the characteristic spectrum of a random polycrystalline sample, i.e. the usual case for a system not subject to orienting forces, is shown. In (b) is shown the spectrum of a sample that has been subject to some uniaxial orienting force. Generally, for a sample subject to a uniaxial stress, the distribution can be described relative to the direction of that uniaxial stress. Thus, the spectral shape depends on the orientation of the axis of stress relative to the magnetic field. Figure 5b shows the theoretical spectrum of a uniaxially deformed sample for one particular orientation of the stress axis relative to the magnetic field direction. Placing the sample in the magnetic field at some other orientation of the stress axis results in a different, but related, spectrum (31, 32). An analysis of the band shape for a known orientation of the two axes such as Figure 5b can, in principle, allow a specification of how the stress produces orientation of the chemical-shift principal axes. However, there are other effects that may also cause dispersion of the resonance that are folded into this spectrum. To compensate for other non-orientation-dependent effects, a study of the variation of the band shape with the orientation of a macroscopic axis of the sample (often the direction of stress in a uniaxially deformed material) relative to the magnetic field direction is often performed, as was done in a study of the uniaxial orientation of poly(tetrafluoroethylene) (30). In that case, the chemical shifts were known for the situations in which the microcrystallite axes were oriented parallel to and perpendicular to the magnetic field. The analysis produced a specification of the orientation distribution function (relative to the stress direction) in several uniaxially deformed materials (Figure 6). 144

Figure 5. Simulated spectra of an NMR band for an axially symmetric chemical-shift tensor. (a) The band shape for a random distribution of microcrystallite axes relative to the magnetic field direction. (b) The band shape for a nonrandom distribution of microcrystallites in a uniaxially oriented object. Note that the non-random distribution tends to overemphasize certain resonance positions relative to the random distribution, and that others are de-emphasized relative to the random distribution. The NMR-based orientation analysis can also be applied to cases in which the material is subject to a more complex force field. For example, one may interrogate the distribution due to biaxial orientation of a film. The major requirement is that one must link the orientation of the chemical-shift principal axes to axes of interest that define the molecular frame. It is often the case that the chemical shift symmetry reasonably follows the symmetry of the electronic structure, thereby allowing one to specify the orientational effects on these electronic co-ordinates. It is also possible to address orientational effects at various points in a molecule by use of specific NMR probes at those points of interest in the molecule. One may study other processes such as disorientation, for example by subjecting the sample to heating with free ends (32). Identification of Species at Catalytic Surfaces The principal use of NMR spectroscopy has been the identification of molecular species. It has been used repeatedly over the last seven decades as a means to specify the outcome of organic reactions because of the straightforward relationship between NMR parameters and structure. In the last 40 years or so, the development of multidimensional NMR techiques that correlate parameters has added tools to allow the study of ever-more-complicated materials (33). 145

Initially, and perhaps still usually, the material of interest is dissolved in some solvent and dynamical averaging limits one to the study of dynamically averaged NMR parameters. However, the application of NMR to other forms of matter and other situations also provides information on structure and electronic states, such as molecules in crystalline solids, molecules dissolved in ordered liquids (liquid crystals), and species in surface phases (34–36). Since the early days of proton NMR spectroscopy, it has been applied to solid phases, as well. The line shape is often dominated by dipolar interactions, and this may be used to give information on internuclear distances, such as the measurement of the proton-proton distance in gypsum (37). The presence of dynamics in solids also affects NMR line shapes, because the averaging of the dipolar effects depends on the geometry and time scale of such motions (38, 39). The application of NMR spectroscopy to surface phases, such as one might find in studies of catalytic processes, has a number of disadvantages. For one, the species of interest are often dilute because they comprise only a small volume of the sample. For another, the other material may have interfering resonances that appear in the spectrum. For yet another, the electronic structure at the surface of particle (usually where the phases of interest reside) may give rise to a large variation of the magnetic environment, including inhomogeneties that wipe out the distinctions normally expected among different functional group that are the key to distinguishing various phases with NMR spectroscopy.

Figure 6. Simulations of NMR-derived orientation distribution functions for poly(tetrafluoroethylene). The three colors represent samples stretched to three different extension ratios. The distribution function can be divided into two parts, one that aligns with a specific distribution width and the other being a disordered random phase. Upon longer extension, the amount of ordered phase increases and the amount of disordered phase becomes smaller. [After reference (31).] 146

These disadvantages notwithstanding, it has been possible to analyze surface phases with NMR spectroscopy. For example, early on, hydrogen species on the surface of supported metal catalysts were extensively studied with NMR spectroscopy (14, 15, 40, 41). The use of carbon NMR spectroscopy, usually with 13C-enriched materials extended the usefulness of the technique to address questions about the involvement of organic materials in catalytic processes (42–51). These studies often involved the observation of the conversion of one organic species into a second species, as well as the observation of species unique to the surface environment. Because of the binding at surfaces that may affect orientational information about the species at the surface, NMR properties such as dipolar couplings (48, 52) and quadrupolar couplings (53) have been used to determine geometric factors. For example, in chemisorbed benzene at room temperature, the resonance line shape is consistent with bonding onto small platinum clusters in which the centroid of the benzene ring lies over a platinum atoms, as opposed to lying over the bridge site or the threefold-hollow site, and the magnitude of the coupling suggested that, at room temperature, the benzene ring was rapidly spinning about its six-fold axis (52). Adsorption of phenanthrene into a porous zeolite at low concentrations, showed the presence of a two-phase system in which one phase was bound and the other was highly mobile, as if in a gas phase. The temperature dependences of the relative amounts of the two phases suggested that an equilibrium was established been the two phases, from which one could estimate the enthalpy of transition (53). Comparison to gas-phase binding to ions suggested that the process was the binding of phenanthrene molecules to counterions in the zeolite. Reactions can be probed because NMR spectroscopy is sensitive to the local environment (54–56). The power of obtaining structural information from dipole-dipole interaction strengths of nuclei at surfaces provides imilar information to that for a solid or anisotropic liquid (57, 58). For example, specification of distances determined with NMR spectroscopy in adsorbed benzene relied heavily on the interaction between the protons of benzene and 195Pt nuclei in supported platinum catalysts (52). The effect of motion on the dipolar coupling affecting 13C in enriched benzene, indicated that benzene was sorbed on a supported platinum catalyst in a variety of energetically distinct sites (48). The use of the SEDOR technique to determine proximity of sites that contain NMR-active nuclei, as was the case in measurement of Pb-Al distances in Pb-exchanged zeolites, can allow the determination of structure or placement of ions on catalytic structures (59).

Use of NMR as a Probe To Study Porous Materials A commonly used technique for investigation of materials is through the use of a probe that senses the local environment, but does not participate too strongly to perturb the system. One probe that gained popularity for studying porous materials such as zeolites or clathrates is sorbed xenon gas (60–64). The weak interaction of xenon atoms with the local environment is reflected by changes in the NMR parameters of 129Xe. For example, the chemical shift is sensitive 147

to the collision rate and the type of partner with which the xenon atom collides. For example, xenon sorbed in a faujasite at low concentrations has a resonance frequency that is about 60 ppm from the low-pressure resonance position of xenon in a macroscopic space (64). The study of the influence of the environment on the xenon NMR parameters has led to its use in characterizing porous materials (62). The resonance position relative to the resonance position of the gas in a macroscopic space depends on the size of the space in which the xenon is trapped, first shown for zeolite adsorption of xenon. Similarly, xenon trapped in small cages in clathrates show a dramatic shift relative to low-pressure xenon in a macroscopic space (65). Of course, other effects may be present. In certain cases, paramagnetic counterions in the zeolite also affect the NMR parameters such as resonance position (66). The temperature at which spectroscopy is carried out may affect the spectroscopic parameters, giving information on the dynamical state (67). In cases where dynamics of the porous material are important, one may investigate the change in NMR parameters of sorbed xenon to determine elements of the dynamics of the porous material (68). Xenon can probe the void volumes of other materials, such as polymers to yield information on the space (69). In fact, one may use the NMR of noble gases like 129Xe and 3He in porous spaces such as lungs to investigate the pore spaces in such materials by MRI (70).

Study of Chemical Changes in Art Masterworks The study of solids with NMR spectroscopy is particularly well exemplified by the investigation of a problem in art conservation (71–73). It has been noticed that, in many masterworks, slow reactions between free fatty acids and pigmentderived ions produce soaps (often lead soaps, when the pigments are lead-based) that cause the appearance of protrusions (soap aggregates), clear spots in paintings, and crazes. A combination of 13C, 207Pb and 119Sn ssNMR has given information on the reactivity of lead-tin yellow type I with palmitic acid (74, 75). Comparison of 207Pb spectra show that the lead coordination in lead palmitate is similar to that of lead stearate, but the structures of both are different from that of lead azelate, and that carboxylates with chain lengths from C6 to C8 are in one structural class, whereas those of C9 or higher fall into a separate class (76). X-ray measurements of lead nonanoate indicate that the local structure around the lead site is quite different from that around lead in lead azelate and lead heptanoate (77). Figure 7 shows a comparison of the 207Pb NMR spectra of three of the pure lead soaps implicated in the deterioration of masterworks. The spectra had to be obtained with a NMR technique specially designed to spread the excitation as uniformly as possible across the region where resonance absorption occurs. Hence, the spikelet appearance. 13C NMR spectra of the lead soaps (Figure 8) also indicate a difference between the short-chain and the long-chain lead carboxylates, in the magnitude of the separation of the two carboxylate resonances. For the long-chain carboxylates (C9, C10, C11, C16, and C18), the separation of the two carboxylate peaks is in 148

the range of 1.13-1.25 ppm. For the short-chain carboxylates (C6, C7, and C8) and lead azelate, the separation of the two carboxylate peaks is in the range of 0.50 – 0.69 ppm. For the α-carbon, the resonance is doubled, with chemical-shift separations in the range of 2.0-2.1 ppm for C9, C10, C11, C16 and C18, and in the range of 0.50 – 0.69 ppm for C6, C7, and C8. The separation of the peaks for lead azelate is similar to that of the short-chain lead carboxylates. Another spectroscopic distinction between the two groups of lead monocarboxylates is that the 13C methyl resonance is doubled for the short-chain materials, whereas for the long-chain group, the resonance is a singlet.

Figure 7. The local lead coordination environment and the 207Pb WURST-CPMG spectra for (a) lead heptanoate, (b) lead azelate and (c) lead palmitate. Shown at the left of each is a representation of the local Pb co-ordination in each. 149

Figure 8.

13C

ssNMR spectra of (a) lead heptanoate, (b) lead azelate, (c) lead palmitate.

These observations imply the existence of two conformations of the chain end for the short-chain soaps, whereas for the long-chain soaps there is only a single conformation. The structure of a lead soap depends on the length and the saturation of the fatty acid chain, as well as other factors that affect the local environment (76). In particular, the presence of other materials such as the binding medium may have a measurable effect on the dynamic state. Some studies of soap formation suggest that water can increase reactivity (78). The reaction can readily be monitored with ssNMR 13C spectroscopy because the carboxyl resonances of free palmitic acid and lead palmitate can be distinguished easily (Figure 8). One may observe the slow transformation from the free acid to the soap, as shown in Figure 9. The appearance of lead palmitate in this reaction is a result of diffusion of the reactants, plus reaction. In Figure 10 is shown the half-time, T50, for samples conditioned by contact with different relative humidities before contact with the free palmitic acid. Although there is quite a bit of scatter, the figure shows that the reaction depends on the exposure to humidity. 150

Figure 9. Product build-up by reaction of free palmitic acid with a lead white paint layer as a function of time for a sample with ~10% (w/w) of the acid. The sample was equilibrated with water at 60% relative humidity before contact with the free acid.

Figure 10. T50 versus relative humidity for the formation of lead palmitate from palmitic acid in a lead white paint film. The trend line is presented to aid the eye in following the data, and does not represent a prediction of the dependence on humidity. Calculation of NMR Chemical Shifts of Solid Materials The exact resonance frequency of a nucleus (whether 1H, 13C, 31P, 207Pb or some other nucleus) in a magnetic field is determined by, among other things, the local electronic structure in the vicinity of the nucleus (79). This facet of the 151

technique is what gives NMR spectroscopy such a strong analytical utility. In principle, precise knowledge of the electronic structure in the region of a nucleus can be used to predict the resonance frequency (or what is often described as the “local field”) through relationships described in the early years of NMR study (5, 80). However, the resolution of experimental chemical shifts is so great that it only becomes possible to predict the resonance frequencies sufficiently accurately with the extremely precise computational protocols that rival the resolution of the experiment. Differences in experimental frequencies of a part per million require extremely accurate depictions of the electronic structure to allow similarmagnitude predictions of magnetic shieldings. The interpretation of NMR spectra necessarily requires this level of calculational accuracy to allow the association of the NMR parameters with other descriptors of the molecular state. In the early days of NMR spectroscopy, the association was often done by analogy with known spectroscopic characteristics of similar materials. With the development of computational chemistry, it has become ever more possible to link the experimental results to model structures by combining experimental NMR spectroscopy with calculations of NMR parameeters based on the structures. The agreement, for example, between experimental chemical-shift tensor components and calculations of them give confidence that the model structure represents the real structure, to the point that some practitioners refer to the combination of NMR experiment with calculated results as NMR crystallography. There are a number of problems that must be overcome to ensure that the the calculation is sufficiently accurate to allow reasonable comparison with experimental data. Prediction of NMR magnetic shielding requires a precise knowledge of the electronic state, at least in the near vicinity of the nuclear species. That requirement involves several issues, including the requirements of gauge invariance as well as properly including the relativistic nature of electrons, particularly in systems containing heavy nuclei such as 207Pb. The reference problem (making a connection between the theoretical magnetic shielding and the practical chemical-shift scale) also must be addressed. Chemical-shift tensors represent not only the local environment, but also the extended structure of a material, because the extended structure may influence the local electronic structure, as well as have direct effects through susceptibility effects. Inclusion of solid-state structure in calculations of NMR properties is important to model the solid state. For example, it has been shown that NMR chemical shifts are sufficiently sensitive to local environment that agreement of calculated parameters with experimental NMR data on materials like naphthalene allow one to specify a more precise set of structural parameters (81). A major development that opened the possibility of accurate predictions of magnetic-shielding tensors was the implementation of gauge-including atomic orbitals with density functional theory (DFT) (82). The same group would also later include relativistic effects through the inclusion of the zero-order regular approximation (ZORA) in DFT, which opened the possibility of calculation of magnetic shielding of heavy NMR-active nuclei (83). The intermolecular contributions to NMR magnetic shielding have long been appreciated, but it has been difficult to include them because of the limitations 152

of handling models of the electronic system that contain many electrons (84). For modelling a solid, there are two principal means of including the effects of structure. The first is to use periodic boundary conditions to specify the arrangement of atoms in the unit cell of a solid (26). The second involves the construction of a cluster that represents the region around a nucleus of interest (84, 85). The cluster may be chosen to emphasize a particular effect, such as hydrogen bonding, which may result in a relatively small cluster. To include more intermolecular effects systematically, one creates ever-larger clusters, but there is a practical limit to increasing the size of the cluster for model calculations (27). A careful choice of cluster properties should include such factors as the local rotational symmetry of the nuclear site (29) and the compensation of charge on the cluster, if any. The primary effects come from the electron distribution near the site of the nucleus, so adding or modifying elements further from the nuclear site may not affect the results of magnetic-shielding calculations as much as neglecting the near atoms in a cluster (28). For network solids such as PbO, the bonding is essentially an infinite network. Truncation of the network at any point, for example to create a cluster having the rotational symmetry elements, may leave charge on the cluster that gives a systematic error in calculations of the magnetic shielding. In such cases, it is necessary to compensate for the excess charge on the cluster to create a model that more closely represents the electronic state. This can be accomplished systematically through a bond-valence method developed by Brown (86), which can be applied to reduce the charge on a cluster by modification of the charge on the terminal atoms of the cluster (28, 87). In DFT calculations that provide NMR properties, the inclusion of exchange and relativistic effects have a great effect on the predicted values. For heavy nuclei or lighter nuclei bonded to heavy nuclei, one must take account of the fact that the electrons must be treated as relativistic particles. The inclusion of these relativistic effects by the zero order regular approximation, ZORA, has provided a means to estimate resonance frequencies of nuclei like 43Ca, 207Pb, and 199Hg with a relative accuracy sufficient to distinguish various sites in a solid (27, 28, 88–90). Of course, the amount of exchange incorporated into a calculation has noticeable.effect on the values of parameters like chemical shifts. Holmes has pointed out this effect is particularly strong for fluorine (91). The developments in calculational methods in recent years, both in computational power and in improvements in algorithms, has led to a situation in which comparison of predicted values of NMR parameters with experiment, when carefully done, can form the basis for specifying structure, or to refinement of structure determined by other means.

Conclusions Many materials, particularly technologically important materials, are complex solids. Characterizing these materials requires analytical techniques that are sensitive to the local environment. Solid-state NMR spectroscopy, with its multiple experimental techniques, is particularly well-suited to probe chemical 153

structure, geometric structure, and dynamics of solid materials. As with liquids or solutions, the identity of a solid may be inferred from the NMR parameters and their similarity to parameters of similar materials. With the advent of multinuclear NMR, the technique can address questions about complex situations. More importantly, the combination of measurements with the predictive ability of calculational chemistry allows on to provide detailed information about structure, for example of active pharmaceutical ingredients. In this article, I have concentrated on examples from work at the Universtiy of Delaware, often in collaboration with others. Whether it is the state of a catalytic material, the effects of macroscopic deformation on polymers, the probing of porous solids, or reaction in a film such as one finds in masterworks of art, NMR spectroscopy provides a means to answer questions of identity and activity. One may find many other examples of the use of these techniques in the literature. The application of solid-state NMR techniques to the study of complex biological systems has provided many examples (92). Applications to such technologically important materials as metal-organic-framework systems has led to a more complete understanding of the nature of these materials, as well (93). Applications to electronic materials have increased understanding of the electronic state of these materials (94). The list continues to expand with time. It appears that what started as an NMR novelty 40 years ago now provides a significant amount of information on solids.

Acknowledgments I am indebted to many students, postdoctoral associates, and collaborators who have provided me with ongoing enjoyment as we have strived to understand nature. I am particularly indebted to funding agencies like the Research Corporation, the Petroleum Research Fund of the American Chemical Society, and the National Science Foundation for support over my career.

References 1. 2. 3. 4. 5. 6. 7.

Purcell, E. M.; Torrey, H. C.; Pound, R. V. Resonance Absorption by Nuclear Magnetic Moments in a Solid. Phys. Rev. 1946, 69, 37. Bloch, F.; Hansen, W. W.; Packard, M. Nuclear Induction. Phys. Rev. 1946, 69, 127. Proctor, W. G.; Yu, F. C. The Dependence of a Nuclear Magnetic Resonance Frequency Upon Chemical Compound. Phys. Rev. 1950, 77, 717. Dickinson, W. C. Dependence of the F-19 Nuclear Resonance Position of Chemical Compound. Phys. Rev. 1950, 77, 736. Pople, J. A.; Schneider, W. G.; Bernstein, H. J. High-Resolution Nuclear Magnetic Resonance; McGraw-Hill: New York, 1959. Seeman, J. I. Ernest L. Eliel: A Life of Purpose, Deermination, and Integrity. Chirality 2002, 14, 98–109. Waugh, J. S.; Huber, L. M.; Haeberlen, U. Approach to High-Resolution Nmr in Solids. Phys. Rev. Lett. 1968, 20, 180–182. 154

8.

9. 10. 11. 12.

13.

14. 15.

16. 17. 18.

19. 20.

21. 22.

23.

24.

25.

Mehring, M.; Pines, A.; Rhim, W.-K.; Waugh, J. S. Spin Decopuling in the Resolution of Chemical Shifts in Solids by Pulsed NMR. J. Chem. Phys. 1971, 54, 3239–3240. Pines, A.; Rhim, W.-K.; Waugh, J. S. C-13 Chemical Shielding Anisotropy in Solids. CS2 and CaCO3. J. Chem. Phys. 1971, 54, 5438–5440. Vaughan, R. W.; Elleman, D. D.; Rhim, W.-K.; Stacey, L. M. F-19 Chemical Tensor in Group Ii Difluorides. J. Chem. Phys. 1872, 57, 5383–5390. Schaefer, J.; Stejskal, E. O.; Buchdahl, R. Magic-Angle C-13 NMR Analysis of Motion in Solid Glassy Polymers. Macromolecules 1977, 10, 384–404. Lippmaa, E.; Magi, M.; Samoson, A.; Engelhardt, G.; Grimmer, A.-R. Structural Studies of Silicates by Solid-State High-Resolution Si-29 NMR. J. Am. Chem. Soc. 1980, 102, 4889–4893. Melchior, M. T.; Vaughan, D. E. W.; Jacobson, A. J. Characterization of the Silicon-Aluminum Distribtuion in Synthetic Faujasites by High-Resolution Solid-State Si-29 NMR. J. Am. Chem. Soc. 1982, 104, 4859–4864. Apple, T. M.; Dybowski, C. Effect of Coadsorption of CO on the Adsorption of H2 on Rh/TiO2. J. Catal. 1981, 71, 316–319. Apple, T. M.; Gajardo, P.; Dybowski, C. NMR, Electron-Spin-Resonance, and TPD Study of H2 Adsorption on Rh/TiO2 Catalyst. J. Catal. 1981, 68, 103–108. Apple, T. M.; Dybowski, C. TPD/NMR Study of Hydrogen Irreversibly Adsorbed on Rh/TiO2. Surf. Sci. 1982, 121, 243–248. Schlup, J. R.; Vaughan, R. W. A Nuclear Magnetic Resonance Investigation of Fluorinated Oxide Catalysts. J. Catal. 1984, 85, 311–323. Duncan, T. M.; Reimer, J. A.; Winslow, P.; Bell, A. T. The Characterization of Alkyl Intermediates on Silica-Supported Ruthenium with C-13 Nuclear Magnetic Resonance Spectroscopy. J. Catal. 1985, 95, 305–308. Gerstein, B. C.; Dybowski, C. Transient Techniques in the Nmr of Solids; Academic Press: Orlando, FL, 1985; p 295. Pines, A.; Gibby, M. G.; Waugh, J. S. Proton-Enhanced Nuclear Induction Spectroscopy. A Method for High Resolution NMR of Dilute Spins in Solids. J. Chem. Phys. 1972, 56, 1776–1777. Dybowski, C.; Neue, G. Solid State Pb-207 NMR Spectroscopy. Prog. Nucl. Magn. Reson. Spectrosc. 2002, 41, 153–170. Neue, G.; Dybowski, C.; Smith, M. L.; Barich, D. H. Fitting of Low-Intensity Wide-Line Spectra Dominated by Chemical-Shift Anisotropy. Solid State Nucl. Magn. Reson. 1994, 3, 115–119. Neue, G.; Dybowski, C.; Smith, M. L.; Hepp, M. A.; Perry, D. L. Determination of Pb-207(2+) Chemical Shift Tensors from Precise Powder Lineshape Analysis. Solid State Nucl. Magn. Reson. 1996, 6, 241–250. MacGregor, A. W.; O’Dell, L. A.; Schurko, R. W. New Method for the Acquisition of Ultra-Wideline Solid-State Nmr Spectra of Spin-1/2 Nuclides. J. Magn. Reson. 2011, 208, 103–113. Catalano, J.; Yao, Y.; Murphy, A.; Zumbulyadis, N.; Centeno, S. A.; Dybowski, C. Nuclear Magnetic Resonance Spectra and Pb-207 Chemical-Shift Tensors of Lead Carboxylates Relevant to Soap Formation in Oil Paintings. Appl. Spectrosc. 2014, 68, 280–286. 155

26. Pickard, C. J.; Mauri, F. All-Electron Magnetic Response with Pseudopotentials: NMR Chemical Shifts. Phys. Rev. B 2001, 63, 245101. 27. Alkan, F.; Dybowski, C. Calculation of Chemical-Shift Tensors of Heavy Nuclei: A Dft/Zora Investigation of Hg-199 Chemical-Shift Tensors in Solids, and the Effects of Cluster Size and Electronic-State Approximations. Phys. Chem. Chem. Phys. 2014, 16, 14298–14308. 28. Alkan, F.; Dybowski, C. Chemical-Shift Tensors of Heavy Nuclei in Network Solids: A DFT/ZORA Investigation of Pb-207 Chemical-Shift Tensors Using the Bond-Valence Method. Phys. Chem. Chem. Phys. 2015, 17, 25014–25026. 29. Holmes, S. T.; Iuliucci, R. J.; Mueller, K. T.; Dybowski, C. Density Functional Investigation of Intermolecular Effects on C-13 NMR Chemical-Shielding Tensors Modeled with Molecular Clusters. J. Chem. Phys. 2014, 141. 30. Brandolini, A. J.; Apple, T. M.; Dybowski, C.; Pembleton, R. G. Solid-State F-19 Nmr of Deformed PTFE. Polymer 1982, 23, 39–42. 31. Brandolini, A. J.; Alvey, M. D.; Dybowski, C. The Crystallite Orientation Distribution-Functions of Deformed Polytetrafluoroethylene - MultiplePulse F-19-Nmr Spectroscopy. J. Polym. Sci., Part B: Polym. Phys. 1983, 21, 2511–2524. 32. Brandolini, A. J.; Rocco, K. J.; Dybowski, C. Solid-State F-19 Nmr Investigations of Annealed Poly(Tetrafluoroethylene). Macromolecules 1984, 17, 1455–1458. 33. Bax, A. Two-Dimensional Nuclear Magnetic Resonance in Liquids; D. Riedel: Dordrecht, 1982. 34. Dybowski, C. R.; Gerstein, B. C.; Vaughan, R. W. Proton Chemical-Shift Tensor in Trichloroacetic-Acid. J. Chem. Phys. 1977, 67, 3412–3415. 35. Duncan, T. M.; Dybowski, C. Chemisorption and Surfaces Studied by Nuclear Magnetic Resonance Spectroscopy. Surf. Sci. Rep. 1981, 1, 157–250. 36. Emsley, J. W.; Lindon, J. C. NMR Spectroscopy in Liquid Crystal Solvents; Pergramon: Oxford, 1975. 37. Pake, G. E. Nuclear Resonance Absorption in Hydrated Crystals. Fine Structure of the Proton Line. J. Chem. Phys. 1948, 16, 327–336. 38. Gutowsky, H. S.; Pake, G. E. Structural Investigations by Means of Nuclear Magnetism II. Hindered Rotation in Solids. J. Chem. Phys. 1950, 18, 162–170. 39. Andrew, E. R.; Eades, R. G. A Nuclear Magnetic Resonance Investigation of Three Solid Benzenes. Proc. R. Soc. (London) 1953, A218, 537–552. 40. Gajardo, P.; Apple, T. M.; Dybowski, C. Observation of Surface-States of Hydrogen on a Rh/TiO2 Catalyst by NMR Spectroscopy. Chem. Phys. Lett. 1980, 74, 306–308. 41. Sheng, T. C.; Gay, I. D. Proton-Resonance of Hydrogen Adsorbed on Supported Platinum Metals. J. Catal. 1982, 77, 53–56. 42. Chang, J. J.; Pines, A.; Fripiat, J. J.; Resing, H. A. Qualitative Analysis of Chemisorbed Molecular Species Via C-13 NMR. Surf. Sci. 1975, 47, 661–665. 156

43. Kriz, J. F.; Gay, I. D. C-13 Nuclear Magnetic-Resonance Observations of Butenes Adsorbed on Alumina. J. Phys. Chem. 1976, 80, 2951–2953. 44. Liang, S. H. C.; Gay, I. D. Measurement of Surface Acid Site Concentration by C-13 NMR. J. Catal. 1980, 66, 294–300. 45. Bell, V. A.; Carver, R. F.; Dybowski, C.; Gold, H. S. Aldol Condensation of Butan-2-one and Pentan-3-one on an Activated Alumina as Monitored Via In situ C-13 Nuclear Magnetic-Resonance Spectroscopy. J. Chem. Soc., Faraday Trans. I 1984, 80, 831–839. 46. Tsiao, C.; Corbin, D. R.; Dybowski, C. Investigation of Methanol Adsorbed on Zeolite H-ZSM-5 by C-13-NMR Spectrometry. J. Am. Chem. Soc. 1990, 112, 7140–7144. 47. Engelsberg, M.; Yannoni, C. S.; Jacintha, M. A.; Dybowski, C. Geometry and Dynamics of Benzene Chemisorbed on a Pt/Eta-Al2O3 Catalyst - a C-13 Dipolar NMR Study. J. Am. Chem. Soc. 1992, 114, 8319–8320. 48. Engelsberg, M.; Yannoni, C. S.; Jacintha, M. A.; Dybowski, C.; Desouza, R. E. C-13 Dipolar NMR Study of Benzene Chemisorbed on High-Surface Alumina-Supported Platinum. J. Phys. Chem. 1994, 98, 2397–2402. 49. Gay, I. D. C-13 Chemical Shifts of Pyridinium Ion in Adsorbed State. J. Catal. 1977, 48, 430–432. 50. Kaplan, S.; Resing, H. A.; Waugh, J. S. C-13 NMR Chemical-Shift Anisotropy for Benzene Adsorbed on Charcoal and Silica-Gel. J. Chem. Phys. 1973, 59, 5681–5687. 51. Lamb, H. H.; Hasselbring, L. C.; Dybowski, C.; Gates, B. C. Synthesis and Reactivity of Anionic Tetraosmium Clusters on Highly Dehydroxylated MgO. J. Mol. Catal. 1989, 56, 36–49. 52. Tirendi, C. F.; Mills, G. A.; Dybowski, C.; Neue, G. Platinum Proton Coupling in the NMR Spectrum of Benzene on an Alumina-Supported Platinum Catalyst. J. Phys. Chem. 1992, 96, 5045–5048. 53. Hepp, M. A.; Ramamurthy, V.; Corbin, D. R.; Dybowski, C. H-2 NMR Investigations of Ion Molecule Interactions of Aromatics Included in Zeolites. J. Phys. Chem. 1992, 96, 2629–2632. 54. Kye, Y. S.; Wu, S. X.; Apple, T. M. C-13 NMR Studies of Ethylene Adsorbed on Ru-Y Zeolite. J. Phys. Chem. 1992, 96, 2632–2636. 55. Decanio, S. J.; Kirlin, P. S.; Foley, H. C.; Dybowski, C.; Gates, B. C. H1 NMR Investigation of a Working Catalyst. Benzene Hydrogenation on Alumina-Supported Rhodium. Langmuir 1985, 1, 243–245. 56. Decanio, S. J.; Foley, H. C.; Dybowski, C.; Gates, B. C. Reduction of Silica-Supported Bis-Allyl Rhodium by H-2 - Characterization by H-1 NMR Spectroscopy. J. Chem. Soc., Chem. Commun. 1982, 1372–1373. 57. Tirendi, C. F.; Mills, G. A.; Dybowski, C. R. Solid-State NMR Spectroscopy of Benzene Adsorbed on Eta-Al2O3. J. Phys. Chem. 1989, 93, 3282–3286. 58. Das, N.; Dai, J.; Hung, I.; Rajagopalan, M.; Zhou, H.-X.; Cross, T. A. Structure of CRGA, a Cell Division Structure; and Regulatory Protein From Mycobacterium Tuberculosis, in Lipid Bilayers. Proc. Natl. Acad. Sci. U. S. A. 2015, 112, E119–E126.

157

59. Niessen, H. G.; Van Buskirk, M.; Dybowski, C.; Corbin, D. R.; Reimer, J. A.; Bell, A. T. Solid-State NMR Studies of Lead-Containing Zeolites. J. Phys. Chem. B 2001, 105, 2945–3950. 60. Dybowski, C.; Bansal, N.; Duncan, T. M. NMR Spectroscopy of Xenon in Confined Spaces - Clathrates, Intercalates, and Zeolites. Annu. Rev. Phys. Chem. 1991, 42, 433–464. 61. Davidson, D. W.; Handa, Y. P.; Ripmeester, J. A. Xe-129 NMR and the Thermodynamic Parameters of Xenon Hydrate. J. Phys. Chem. 1986, 90, 6549–6552. 62. Demarquay, J.; Fraissard, J. Xe-129 NMR of Xenon Adsorbed on Zeolites Relationship between the Chemical-Shift and the Void Space. Chem. Phys. Lett. 1987, 136, 314–318. 63. Ito, T.; Fraissard, J. Characterization of Void Spaces and Location of Cations in Zeolite Rho by Means of Xe-129-NMR. Zeolites 1987, 7, 554–558. 64. Fraissard, J. NMR Study of Xe-129 Adsorbed on Zeolites. Z. Phys. Chem. (Leipzig) 1988, 269, 657–670. 65. Ripmeester, J. A.; Davidson, D. W. Xe-129 Nuclear Magnetic-Resonance in the Clathrate Hydrate of Xenon. J. Mol. Struct. 1981, 75, 67–72. 66. Scharpf, E. W.; Crecely, R. W.; Gates, B. C.; Dybowski, C. Characterization of NiNa-Y Zeolite by Xe-129 NMR Spectroscopy. J. Phys. Chem. 1986, 90, 9–11. 67. Smith, M. L.; Corbin, D. R.; Dybowski, C. Effects of Temperature on Xe-129 NMR-Spectroscopy of Xenon in Zeolite Rho. J. Phys. Chem. 1993, 97, 9045–9047. 68. Smith, M. L.; Corbin, D. R.; Abrams, L.; Dybowski, C. Flexibility of ZeoliteRho - Xe-129 NMR Studies of Entrapped Xenon in Cd-Rho. J. Phys. Chem. 1993, 97, 7793–7795. 69. Walton, J. H.; Miller, J. B.; Roland, C. M. Xe-129 NMR as a Probe of Polymer Blends. J. Polym. Sci., Part B: Polym. Phys. 1992, 30, 527–532. 70. Yablonsky, D. A.; Sukstanskii, A. L.; Quirk, J. D.; Woods, J. C.; Conradi, M. S. Probing Lung Microstructure with Hyperpolarized Noble Gas Diffusion Mri: Theoretical Models and Experimental Results. Magn. Reson. Med. 2014, 71, 486–505. 71. Capitani, D.; Proietti, N. NMR in Cultural Heritage. Magn. Reson. Chem. 2015, 53, 1. 72. Capitani, D.; Di Tullio, V.; Proietti, N. Nuclear Magnetic Resonance to Characterize and Monitor Cultural Heritage. Prog. Nucl. Magn. Reson. Spectrosc. 2012, 64, 29–69. 73. Lambert, J. B.; Shawl, C. E.; Stearns, J. A. Nuclear Magnetic Resonance in Archaeology. Chem. Soc. Rev. 2000, 29, 175–182. 74. Catalano, J.; Murphy, A.; Yao, Y.; Alkan, F.; Zumbulyadis, N.; Centeno, S. A.; Dybowski, C. Pb-207 and Sn-119 Solid-State NMR and Relativistic DFT Studies of the Historic Pigment Lead-Tin Yellow Type I and Its Reactivity in Oil Paintings. J. Phys. Chem. A 2014, 18, 7952–7958. 75. Catalano, J.; Yao, Y.; Murphy, A.; Zumbulyadis, N.; Centeno, S. A.; Dybowski, C. Analysis of Lead Carboxylates and Lead-Containing Pigments 158

76.

77.

78.

79. 80. 81.

82.

83.

84. 85.

86. 87.

88.

89.

90.

in Oil Paintings by Solid-State Nuclear Magnetic Resonance; Materials Research Society: Warrendale, PA, 2014. Catalano, J.; Yao, Y.; Murphy, A.; Zumbulyadis, N.; Centeno, S. A.; Dybowski, C. Nuclear Magnetic Resonance Spectra and Pb-207 Chemical Shift Tensors of Lead Carboxylates Relevant to Soap Formation in Oil Paintings. Appl. Spectrosc. 2014, 68, 280–286. Catalano, J.; Murphy, A.; Yao, Y.; Yap, G. P.; Zumbulyadis, N.; Centeno, S. A.; Dybowski, C. Coordination Geometry of Lead Carboxylates Spectroscopic and Crystallographic Evidence. Dalton Trans. 2015, 44, 2340–7. Cotte, M.; Checroun, E.; Susini, J.; Dumas, P.; Tchoreloff, P.; Besnard, M.; Walter, P. Kinetics of Oil Saponification by Lead Salts in Ancient Preparations of Pharmaceutical Lead Plasters and Painting Lead Mediums. Talanta 2006, 70, 1136–42. Ramsey, N. F. Magnetic Shielding of Nuclei in Molecules. Phys. Rev. 1950, 78, 699–703. Pople, J. A. The Theory of Chemical Shifts in Nuclear Magnetic Resonance I. Induced Current Densities. Proc. R. Soc. (London) 1957, A239, 541–549. Harper, J. K.; Iuliucci, R.; Gruber, M.; Kalakewich, K. Refining Crystal Structures with Experimental C-13 NMR Shift Tensors and Lattice-Including Electronic Structure Methods. Cryst. Eng. Commun. 2013, 15, 8693–8704. Schreckenbach, G.; Ziegler, T. Calculation of NMR Shielding Tensors Using Gauge-Including Atomic Orbitals and Modern Density Functional They. J. Chem. Phys. 1995, 99, 606–611. Ziegler, T.; Wolff, S. K.; van Lenthe, E.; Baerends, E. J. Density Functional Calculations of Nuclear Magnetic Shieldings Using ZORA for Relativistic Effects. J. Chem. Phys. 1999, 110, 7689–7699. Lau, K. F.; Vaughan, R. W. Pressure Dependence of Fuorine Chemical Shift in CaF2. J. Chem. Phys. 1976, 65, 4825–4829. Hu, J. Z.; Facelli, J. C.; Alderman, D. W.; Pugmire, R. J.; Grant, D. M. N-15 Chemical Shift Tensors in Nucleic Acid Bases. J. Am. Chem. Soc. 1998, 120, 9863–9869. Brown, I. D. Bond Valences - Simple Structural Model for Inorganic Chemistry. Chem. Soc. Rev. 1978, 7, 359–376. Holmes, S. T.; Alkan, F.; Iuliucci, R. J.; Mueller, K. T.; Dybowski, C. Analysis of the Bond-Valence Method for Calculating Si-29 and P-31 Magnetic Shielding in Covalent Network Solids. J. Comput. Chem. 2016, 37, 1704–1710. Holmes, S. T.; Bai, S.; Iuliucci, R. J.; Mueller, K. T.; Dybowski, C. Calculations of Solid-State Ca-43 NMR Parameters: A Comparison of Periodic and Cluster Approaches and an Evaluation of Dft Functionals. J. Comput. Chem. 2017, 38, 949–956. Taylor, R. E.; Carver, C. T.; Larsen, R. E.; Dmitrenko, O.; Bai, S.; Dybowski, C. Revisiting Hgcl2: A Solution and Solid-State Hg-199 NMR and ZORA-DFT Computational Study. J. Mol. Struct. 2009, 930, 99–109. Dybowski, C.; Gabuda, S. P.; Kozlova, S. G.; Neue, G.; Perry, D. L.; Terskikh, V. V. Correlation and Relativistic Effects in Beta-PbO and Other 159

91.

92.

93.

94.

Lead (II) Oxides: A Quantum Ab Initio Explanation of Pb-207 NMR and XANES Spectra. J. Solid State Chem. 2001, 157, 220–224. Holmes, S. T. DFT Calculations on Magnetic Shielding and Quadrupolar Coupling Ordered Systems: Methods and Applications to NMR Crystallography; University of Delaware: Newark, DE, 2017. Quinn, C. M.; Polenova, T. Structural Biology of Supramolecular Assemblies by Magic-Angle Spinning NMR Spectroscopy. Q. Rev. Biophys. 2017, 50, 1–44. Morris, W.; Stevens, C. J.; Taylor, R. E.; Dybowski, C.; Yaghi, O. M.; GarciaGaribay, M. A. NMR and X-ray Study Revealing the Rigidity of Zeolitic Imidazolate Frameworks. J. Phys. Chem. C 2012, 116, 13307–13312. Koumoulis, D.; Chasapis, T. C.; Leung, B.; Taylor, R. E.; Stoumpos, C. C.; Calta, N. P.; Kanatzidis, M. G.; Bouchard, L.-S. Site-Specific Contribution to the Band Inversion in a Topological Crystalline Insulator. Adv. Electron. Mater. 2015, 1, 1500117.

160

Chapter 10

From NMR Spectra to Molecular Structures and Conformation Alan E. Tonelli* Fiber & Polymer Science Program College of Textiles, North Carolina State University, Campus Box 8301, 1020 Main Campus Drive, Raleigh, North Carolina 27606-8301, United States *E-mail: [email protected]

Chemists are interested in establishing the connections between the structures of molecules and the properties of materials made from them. Nearly 30 years ago I wrote a book, “NMR Spectroscopy and Polymer Microstructure, which was subtitled “The Conformational Connection”. Its purpose was to describe how 13C-NMR spectra of polymers, currently the most sensitive NMR probe, can be assigned to their polymer microstructures. At that time and now, even the most advanced quantum mechanical methods cannot estimate 13C resonance frequencies accurately enough to delineate the detailed molecular structures that produced them. Instead empirical nuclear shielding effects produced by substituents α, β, and γ to a carbon atom were successfully used to make the connections between polymer 13C-NMR spectra and their microstructures. Principal among these are the nuclear shieldings produced by γ substituents, which were demonstrated to have a conformational origin, i.e., a γ substituent could only shield a 13C nucleus if the central bond between them produced a proximal arrangement by adopting a gauche or cis conformation. In this review, honoring the memory of the late Prof. Ernest Eliel, we update this approach by presenting a few examples of more recent attempts to utilize the conformational characteristics of flexible polymers to characterize their microstructures, as well as the conformations of rigid solid polymers, using 13C-NMR.

© 2017 American Chemical Society

Introduction The purpose of this selective review is to assess whether or not significant advances have been made, since the publication of ref. (1), in the use of 13C-NMR spectroscopy, currently the method of choice, to characterize the microstructures and conformations of polymers and other flexible molecules (1). The short answer is NO! On the theoretical side of this query, even though Quantum Mechanical theory and calculation methods have advanced over this period, it is still not possible to predict NMR resonance frequencies accurately enough to completely characterize their microstructures (See refs. (2–9)). This is particularly true for flexible molecules like polymers, because the magnetic shielding of nuclei must not only be predicted for particular microstructures, but also must be averaged over all the conformations appropriate to each microstructure. Instead the nuclear shielding effects produced by substituents α, β, and γ to and separated by one, two, and three bonds from, respectively, a carbon atom were successfully used to make the connections between 13C-NMR spectra and the contributing microstructures (1). Principal among these substituent effects were the nuclear shieldings produced by γ substituents, which were demonstrated to have a conformational origin, i.e., a γ substituent could only shield a 13C nucleus if the central bond between them produced their proximal arrangement by adopting a gauche or cis conformation. α-, β-, and especially the conformationally sensitive γ-effects were used to assign the NMR spectra and determine the microstructures of polymers in solutions and melts, where they are conformationally flexible, and to characterize their rigid conformations in solid samples. For example, in the exceptional case of polypropylene (PP), its resonances observed in solution by high resolution 13C-NMR exhibit a sensitivity to stereosequences at the undecad level. This means that eleven repeat unit fragments of PP different only in whether their terminal diads are meso (m) or racemic (r) evidence distinct resonance frequencies for the methyl carbons in their central (6th) repeat unit. [Busico and Cipullo (10), using high-field 13C-NMR (150 MHz for 13C), have in fact detected a 0.03 ppm difference in the resonance frequencies of the methyl carbons in the central repeat units of mmmmmmmrmr and mmmmmmmrmm PP undecads.] Though the sensitivity of 13C-NMR to the long-range stereochemistry of PPs is exceptional, more typically 13C-NMR is usually only sensitive to tetrad and pentad stereosequences in homopolymers and triad comonomer sequences in copolymers. The structural sensitivity of 13C-NMR presents a challenge to the chemist, i.e., how can the multitude of observed 13C resonances be assigned to the specific microstructures that generated them? As an example, the 0.03 ppm difference in resonance frequencies observed (10) between the methyl carbons in the central repeat units of mmmmmmmrmr and mmmmmmmrmm PP undecads mentioned above is reproduced by the small differences in the local populations of the central repeat unit conformations in these two PP undecads and their resultant γ-gauche shieldings (11). Here we update this approach by presenting a few examples of more recent attempts to utilize the conformational characteristics of flexible polymers to assign their observed 13C-NMR resonance frequencies and determine their microstructures and to characterize their rigid solid conformations, as well. 162

NMR Resonance Frequencies (Chemical Shifts) The numbers and types of atoms and groups of atoms attached to or near a magnetically active nucleus determines the degree of shielding it experiences from the static applied magnetic field Bo. This is because they affect the electron cloud surrounding each nucleus, which produce orbital currents accompanied by small local magnetic fields proportional to but in the opposite direction (diamagnetic) of Bo. Thus the local magnetic field actually experienced by the nucleus is Bloc = Bo(1-σ), where σ is the electronic screening constant, which is normally a tensor due to the typically anisotropic distribution of electrons that shield nuclei. It is this dependence of σ upon molecular structure that lies at the heart of NMR’s utility as a probe of molecular structure. Of the common NMR active spin ½ nuclei most abundantly present in organic molecules, 1H and 13C, the latter has an ~ order of magnitude greater sensitivity to molecular structure (σ), compared with 1H, and is unaffected by direct through bond scalar homonuclear coupling, which for 1H nuclei produce multiple resonance for identical microstructures. For these two reasons 13C-NMR remains the most effective experimental probe of molecular structure despite the greater natural abundance of 1H nuciei. However, as mentioned above, sufficiently accurate calculations of the dependence of σ upon the molecular structure surrounding magnetically active nuclei are currently not feasible (2–9). Instead we rely on the empirical effects of directly bonded nearest neighbor (α) and next nearest neighbor (β) substituent effects, and the conformationally dependent effects of γ-substituents separated by three intervening bonds from 13C nuclei (γ-gauche shielding). 13C NMR studies of paraffinic hydrocarbons (12–17) have led to the following substituent effect rules. Carbon substituents attached at α, β, and γ positions to an observed carbon produce approximate deshielding of +9 ppm downfield , deshielding of +9 ppm downfield, and a shielding of – (2 to 3) ppm upfield, respectively, compared to an unsubstituted carbon. For example, when these substituent effects are applied to the CH3, CH2, and CH carbons in each polypropylene (PP) repeat unit their calculated resonance frequencies closely follow the order observed in their 13C-NMR spectra and shown in Figure 1. In atactic PP the extensive splitting of resonances belonging to the same carbon type must be produced by the presence of different stereosequences, because the numbers of α, β, and γ substituents possessed by each carbon type are independent of stereosequence. However, we know the local conformations in vinyl polymers, such as atactic PP, are sensitive to stereosequence (18). The local magnetic field Bloc(i) experienced by a carbon nucleus i must be dependent upon the local conformation in its vicinity, Thus,

We need to know the dependence of the local magnetic field Bloc(i) on the local conformation before the connection between polymer microstructures and resonance frequencies (δ13Cis) can be made. The source of the dependence of the local magnetic field Bloc(i) on the local conformation turns out to be the effect of γ-substituents. An observed carbon Co and its γ-substituent Cγ are separated by 163

three intervening bonds (-Co–C-↻φ-C-Cγ-), and, depending on the conformation (φ) of the central bond, their mutual distance and orientation maybe varied. For example, changing the conformation from trans (φ = 0º) to gauche± (φ = ±120º) reduces their separation from 4 to 3Å. We have suggested that for a γ-substituent to shield a carbon nucleus they must be in a gauche arrangement. The methyl carbons of butane and the higher n-alkanes have a single γ-substituent, unlike the methyl carbons in propane, but they have the same number and kinds of α- and β-substituents as propane. The methyl carbons in liquid butane and higher n-alkanes resonate at ~13 ppm, while in liquid propane the methyls resonate at ~15 ppm (17). Because n-alkanes crystallize in their lowest energy, fully extended, all trans conformation, the methyl carbons of solid butane and higher n–alkanes are not gauche to their γ-methyl or methylene carbon substituents. We therefore expect that δCH3(solid CnH2n+2, n ≥ 4) = δCH3(liquid propane), which has in fact been observed (19).

Figure 1. 25 MHz 13C NMR spectra of (a) isotactic, (b) atactic, and (c) syndiotactic PPs.

164

We know how much gauche character, Pg, the –C-C- bonds in n-alkanes have (Pg = fractional population of Φ = ±120º conformations) (1, 18), so we can estimate the γ-gauche shielding (γC-C) produced at the methyl carbons in butane, for example. When the observed shielding ΔδCH3 = δCH3(butane) δCH3(propane) = 13.2 – 15.6 = -2.4 ppm is divided by the gauche character of the intervening bond (Pg = 0.46), γC-C = ΔδCH3/Pg = -2.4/0.46 = -5.2 ppm, as shown in Figure 2. When this procedure is similarly applied to n-butane,1-propanol, and 1-chloropropane, the following γ-gauche shielding effects are derived: γC-C = -5.2 ppm, γC-O = -7.2 ppm, and γC-Cl = -6.8 ppm (1). Along with the Rotational Isomeric States (RIS) conformational descriptions (18) determined for polymers containing all carbon backbones and side-chains containing C, O, and/or Cl atoms, these γ-gauche shielding effects can be used to determine their microstructure, as we will now demonstrate.

Figure 2. Derivation of the γ-gauche shielding on 13C nuclei produced by the γ substituents, C, OH, and Cl (See text).

165

Polymer Microstructures E-VAc Copolymers Both poly(viny1acetate) (PVAc) and ethylene-vinyl acetate (E-VAc) copolymers have commercial significance. Combination of solution 13C-NMR observations made at high magnetic fields with comparisons to the spectra recorded for an extensive series of E-VAc model compounds (mono- and diacetates of aliphatic alcohols and diols, etc.), have served to establish their microstructures i.e., their comonomer and stereosequence distributions. (See refs. (18, 20–24) and refs. therein). In addition, a conformational description (RIS model) of E-VAc copolymers was developed (22) by merging the RIS models of the constituent homopolymers PE (23) and PVAc (24). This permitted evaluation of the conformationally sensitive γ-gauche shielding effects for E-VAc comonomer and stereosequences. In addition, comparison of the observed spectra to the 13C chemical shifts calculated via the γ-gauche effect method using the RIS model developed for E-VAc copolymers provided an assessment of its the validity. This comparison is shown below in Figures 3 and 4, where the generally close agreement between observed and calculated 13C chemical shifts of the backbone methylene and methine carbons validates the RIS conformational model developed for E-VAc copolymers (22).

Figure 3. Comparison of observed and calculated 13C chemical shifts for the methylene carbons in E-VAc copolymers.

166

Figure 4. Comparison of observed and calculated 13C chemical shifts for the methylene carbons in E-VAc copolymers.

A similar comparison of observed and calculated 13C chemical shifts of the backbone methylene and methine carbons for atactic poly(vinyl acetate) (PVAc) are presented in Figures 5 and 6. The RIS conformational model of PVAc developed by Sundararajan (24) was used to estimate the γ-gauche contributions to the calculated 13C chemical shifts. Note in Figure 5 that both the observed and calculated order of resonances is rxr, mxr (rxm), mxm from low to high field in the methylene carbon regions. The m-centered tetrads in the spectrum are, however, observed upfield from the corresponding r-centered tetrads, while the calculated methylene carbon chemical shifts evidence the opposite behavior. The calculated tetrad methylene carbon chemical shifts are averaged over all hexads containing common tetrad stereosequences. So it may not be too surprising that the tetrad methylene carbon chemical shifts are more sensitive to the terminal diad stereochemistry than the m or r character of the central diad. This is because it is the terminal diads that contain the backbone bonds which govern their gauche:trans conformational ratios and therefore should be highly sensitive to the terminal diad stereochemistry.

167

Figure 5. Comparison of observed and calculated 13C chemical shifts for the methylene carbons in atactic PVAc.

Figure 6 compares the observed and calculated backbone methine carbon chemical shifts, which appear to generally agree. The mrmr, rrrr, and mrrm stereosequence pentads were unable to be unambiguously assigned by Sung and Noggle (21), so we took the liberty to assign them in this order with increasing field strength to achieve greater consistency with our calculated methine carbon chemical shifts. The assignment of vinyl polymer spectra, particularly when they are stereochemically random like the atactic PVAc sample observed by Sung and Nogglel (21) (Pm = 0.481), and the resonances of all stereosequences appear with nearly equal intensities in the spectrum, are substantially aided by the ability to calculate the stereosequence dependent13C chemical shifts in vinyl polymers.

168

Figure 6. Comparison of observed and calculated 13C chemical shifts for the methine carbons in atactic PVAc. The generally good overall agreement presented here between the 13C-NMR spectra observed for E-VAc copolymers and their calculated 13C chemical shifts lends strong support to the conformational description/RIS model employed for the E-VAc copolymers (22). Our ability to calculate the microstructurally sensitive 13C- NMR chemical shifts for E-VAc copolymers not only aided the assignment of their spectra, but illustrated the potential for testing and/or deriving conformational RIS models for vinyl homo- and copolymers by comparing their observed 13C-NMR spectra with 13C chemical shifts calculated via the conformationally sensitive γ-gauche effect method (Also see ref. (25)). Similar success using substituent effects to assign 13C-NMR spectra of polymers in solution utilizing, in particular the conformationally averaged γ-gauche shielding effects, were achieved for additional polymers (26–28). However, in the case of poly (methyl methacrylate)s (PMMAs) we observed the “Failure of the γ-Gauche Effect Method To Predict the Stereosequence-Dependent 13C NMR Spectrum of the Disubstituted Vinyl Polymer Atactic Poly(methy1 methacrylate)” (29), even though PMMA is similar in structure to PVAc. Since 169

this study a revised RIS conformational model has been developed for PMMA (30). For the first time, the conformations (χ) of the side chains and their effect on backbone conformations were considered. This revised conformational model may eventually lead to improvement between the observed 13C-NMR spectra of PMMAs and the 13C chemical shifts estimated with γ-gauche shielding effects conformationally averaged using it.

Polydiacetylenes Because their solid-state polymerization from crystallized monomers can produce macroscopic single crystals of polymer, polydiacetylenes (PDAs) are an unusual class of polymers. 1,6-addition polymerization results in the conjugated backbones shown in Figure 7a. Along the backbone, extensive delocalization of π-electrons reduces absorption of light in the visible range and also gives PDAs interesting optical properties (31–34), which can be influenced by mechanical or thermal stresses through alteration of the amount of π-electron delocalization. As a result, PDAs are well known for their characteristic chromism.

Figure 7. (a) Schematic representation of the solid-state synthesis of polydiacetylenes. (b) Chemical structure of poly(4BCMU).

170

Most PDAs are insoluble, but several form solutions in common organic solvents. One such soluble PDA was first synthesized in 1978 (35, 36) from 5,7-dodecadiyne-1,12-diol bis[ (( butoxy-carbony1)methyl) urethane] and is commonly called poly(4BCMU) (See Figure 7b). Poly(4BCMU)’s long side chains apparently contribute enough conformational entropy to make it soluble in solvents like chloroform, nitromethane, and toluene (31). Poly(4BCMU) crystals and solution-cast films undergo chromic transitions, as well its solutions in organic solvents. The absorption spectrum and color of the as-polymerized polymer change from blue to red at ca. 110° C, and beyond 120° C its red crystals melt to a yellow isotropic liquid. When poly(4BCMU) is dissolved in a thermodynamically good solvent, like chloroform, it forms a yellow solution, but its solution becomes red upon addition of a nonsolvent like hexane (35). However, molecular mechanisms responsible for the chromism observed in poly(4BCMU) are not well understood (37–41). In an attempt to determine the mechanism(s) for its thermo- and solvato-chromism, we undertook a 13C-NMR examination of poly(4BCMU) (42). Tables 1 and 2 present the 13C-NMR resonance frequecies observed for poly(4BCMU) in its yellow chloroform solution and red toluene gel. Based on their γ-gauche shielding analyses we were able to conclude that similar mechan-isms are operating for the solid-state and solution chromism observed for poly(4BCMU). Its backbone is transformed to an increasingly nonplanar conformation, with a greater degree of π-electron localization, leading to a color change from blue to red in the solid as the temperature is raised and from red to yellow in solution when a good solvent is changed to a poor one.

Table 1. Chemical Shifts of Poly(4BCMU) in CDCl3a

171

Table 2. Chemical Shifts of Poly(4BCMU) in Toluene-d8a

In addition to changes in the backbone conformation with temperature and/or solvent, it was concluded that the first 3 CH2–CH2 side chain bonds separating the α, β, γ, and δ methylene groups (See Figure 7b) are g+, t, and g- in the blue and red low temperature solid and poor solution states, respectively, and transform to the extended ttt conformation at high temperature in the solid, or in solution with a poor solvent. Thermal energy drives the transition in the solid, while in solution the solvent quality produces the transition.

Solid State Polymer Conformations Here we will present some examples of analyzing the conformations of solid polymers via comparison of their high resolution solid-state 13C-NMR spectra with the 13C resonance frequencies obtained from γ-gauche shieldings expected from their rigid conformations. It has been observed many times in the high resolution solid-state 13C-CPMAS-DD-NMR spectra (43) of polymer samples that their resonance frequencies are generally a consequence of and dominated by their rigid chain conformations and are not very sensitive to their inter-chain packing. Poly(phenylene sulphide) and Its Solid Model Compounds Poly(p-phenylene sulphide) (PPS) is a high-performance polymer often used as a matrix for fiber-filled composites. Based on X-ray diffraction observed from stretched fibers, Tabor et al. (44) reported a crystal structure for PPS, which was more recently confirmed by electron diffraction from single PPS crystals obtained from solution and thin molten films (45). The PPS crystalline conform-ation is illustrated in Figure 8, where sulphur atoms are in the all-trans arrangement and located in the same plane, while successive phenyl rings are inclined at +45° with respect to the backbone plane. 172

Figure 8. Schematic drawing of the crystalline conformation of PPS.

The structurally related polymers, poly(p-phenylene oxide) (PPO) and poly(2,6-dimethyl-1,4-phenylene oxide) (PDMPO), adopt the same crystalline conformation (46, 47). Schaefer and Stejskal (43) noted that two resonances were observed for the protonated carbons (P) in the the high resolution 13C-CPMAS-DD-NMR spectrum of solid PDMPO. They suggested that their nonlinear C-O-C bonds produced non-equivalent magnetic environments for the protonated carbons, which become equivalent only when the phenyl rings rotate rapidly or are fixed at 90° out of the plane of the oxygen atoms. Doubling of P resonances in PDMPO crystals results because neither occurs. Because PPS adopts the same crystalline conformation as PDMPO, we expected, but in fact did not observe (48, 49) a doubling of P resonances in PPS (See Figure 9). The CP resonances corresponding to θ1 = 45° and θ2 = 135 ° in crystalline PPS likely remain unresolved due to the ~2 ppm line width observed in the 13C-CPMAS /DP-NMR spectrum in Figure 9, because Shaefer and Stejskal (43) were able to resolve the splitting of CP resonances in solid PDMPO, which assumes the same crystalline conformation as PPS. This implies that the short C-O bonds (~1.4Å) in PDMPO lead to a greater conformational sensitivity of 13C chemical shifts than do the longer C-S bonds (~1.8 Å) in PPS. 173

Figure 9. (a) CPMAS/DD 13C n.m.r, spectrum of PPS recorded at room temperature, with CQ downfield and CP upfield (b) Same as (a) except with a short 100 μs delay without spin-locking in the 1H channel after the Hartmann-Hahn match (49), causing the absence of the CP resonance.

The cyclic pentameter of PPS [c-(PS)5] crystallizes in a conformation (See Figure 10b) with the relative orientations of all five phenyl ring pairs bonded to common sulphur atoms significantly different from each other and also different from the orientation of phenyl rings in crystalline PPS (50–52) . Table 3 presents the dihedral angles θPQ between P and Q phenyl ring carbons connected on either side of the same S. Figure 11 presents The 13C-CPMAS/DD-NMR spectra recorded at room temperature for c-(PS)5 without and with dipolar dephasing. Both carbon types show at least six distinct resonances with overall spreads of 8 and 18 ppm for the Q and P resonances. The dihedral angles between CP and CQ carbons observed in the crystal structure of c(PS)5 (See Figure 10b) span the full range θPQ = 0 to 180 °, from the cis (0°) to trans(180°) arrangements of CP and CQ across the S atom. If different arrangements of phenyl rings about the C-S bonds was the principal source of the multiple resonances observed for the Cp and CQ carbons in crystalline c-(PS)5, then the different shielding between a cis (θPQ = 0 °) and a trans (θPQ = 180 °) arrangement of Cp and CQ carbons must be 8 and 18 ppm, respectively. Relative to their trans arrangement, a cis arrangement of the methyl and protonated ring carbons in solid di- and tri-methoxy benzenes produces (53–56) a 6 ppm 174

shielding, strongly suggesting that the 18 ppm spread in observed CP chemical shifts observed for c-(PS)5 cannot be solely a consequence of its crystalline conformation. It was hoped that observation of the13C-CPMAS/DD-NMR spectra of diphenyl sulphide (DPS) shown in Figure 10c and their comparison to the spectra of PPS and c-(PS)5 would suggest a possible crystalline conformation for DPS, which is presently unknown. In Figure 12 the liquid state 13C-NMR spectrum of DPS recorded at room temperature is displayed, while in Figure 13 the solid-state 13C-NMR spectra observed for DPS at -60°C are presented. To produce a quantitative spectrum, the spectrum in Figure 13a was recorded without CP and with a 420 s delay between decoupling pulses. The CPMAS/DD spectrum of solid DPS in Figure 13b was recorded with dipolar dephasing, and only shows CQ resonances.

Figure 10. a. Projection along the CQ–S bond in the crystallineconformation of PPS. b. Crystalline conformation of c-(PS)5. c. Crystalline conformation of DPS suggested by solid-state 13C-NMR. 175

Table 3. Dihedral angles (θPQ) between the P and Q ring carbons in crystalline c-(PS)5 (52)

Figure 11. (a) 13C-CPMAS/DD-NMR spectrum recorded for c-(PS)5 at room temperature. (b) CPMASS/DD ~3C n.m.r, spectrum recorded at room temperature with a 100 μs delay (without spinlocking) in the 1H channel after the Hartmann-Hahn match (49).

176

Figure 12.

13C-MAS-NMRspectrum

of liquid DPS recorded at room temperature.

Because the crystalline conformation of DPS is not expected to significant-ly influence their chemical shifts the m and p resonances are singlets. The o-CP carbon, on the other hand, shows a resonance doublet centred at 135.5 ppm with a 2:1 ratio of intensities, and a singlet resonance ~6ppm upfield at 129.9 ppm corresponding in intensity to a single o-CP carbon. The crystalline conformation of DPS, which has yet to be determined by X-ray diffraction, may be estimated from the pattern of o-CP resonances. A DPS conformation consistent with the observed pattern of CQ and o-CP resonances is drawn in Figure 8c, where one phenyl ring is nearly coplanar with CQ-S-CQ (~φ1 =0°), while the other phenyl ring is rotated (φ2 = 30 - 40°) out of this plane. The dihedral angles between o-CP and CQ are θ1 = 0°, θ2 = 180° and θ1 = 30-40°, θ2 = 140-150°, respectively, for these two distinct phenyl ring orientations. The least shielded resonances at 136.1 ppm would correspond to the 2 o-CP carbons with θ2 = 140-150 and 180°. The resonance at 135.1 ppm would correspond to the single o-CP carbon with θ1 = 30-40°, and the most shielded resonance at 129.9 ppm could be assigned to the o-CP carbon in the cis arrangement (θ1 =0 °) with CQ. Also consistent with θ1 = 30-40 and 0°, as proposed here for crystalline DPS, is the resonance doublet for the CQ carbons and their ~4 ppm separation.

177

Figure 13. (a) 13C-MAS-NMR spectrum of DPS recorded at - 60°C with a 420 s delay between decoupling pulses. (b) 13C-CPMAS-NMR spectrum recorded at -60° C with a100 μs delay (without spin-locking) in the 1H channel after the Hartmann-Hahn match (49). (c) Their difference spectrum showing only CP carbon resonances.

X-ray diffraction studies have been reported for bis(4-mercaptophenyl) sulphide (57) and 1,4-bis(phenylthio) benzene (58), which assume asymmetric conformations in their crystals. The asymmetric crystalline conformations of these mono- and disulphides do not differ appreciably from the asymmetric conformer proposed for crystalline DPS (See Figure 8c) based on solid-state 13C-NMR results. “State-of-the-art” quantum mechanical methods of that time were applied to estimate the 13C-NMR chemical shifts for crystalline c-(PS)5 (59). The calculated spectra were obtained by assuming all resonance peaks obtained from the calculated ab initio nuclear shieldings had Lorentzian line shapes with widths at half-height of 1 ppm. Note in Figure 14 that both the total spread and distribution of resonance peaks in the observed and calculated spectra of c-(PS)5 are reasonably to some-what similar. However, to achieve even the limited favorable comparison shown in Figure 14, the observed spectrum had to be uniformly shifted 10 ppm upfield.This example once again illustrates the present limitations of theoretically estimating the NMR resonance frequencies of molecules as a function of their molecular structures and conformations. 178

Figure 14. Comparison of observed (Exp) and calculated 13C NMR spectra of crystalline c-(PS)5 (59). Syndiotacic-Polystyrene and Its Solid Model Compounds Ishihara et al. (60) reported the synthesis of highly stereoregular, crystalline, syndiotactic polystyrene (s-PS), which Zambelli et a1. (61) confirmed. X-ray (60, 62) and electron diffraction (63) performed on oriented fibers and films indicated a chain axis repeat distance of 5.1Å consistent with s-PS chains adopting an alltrans,planar zigzag conformation (form I crystals). When cast from dilute chloroform or 1,2-dichlorobenzene solutions (64, 65) or when melt-quenched films or fibers were swollen in chloroform, dichloromethane, 1,2-dibromo- or dichloroethane, or cyclohexane, a different crystalline polymorph (form II) was found (62, 66). A fiber repeat of 7.5Å was 179

obtained by X-ray diffraction patterns observed (62) for oriented, swollen form II fibers, and is consistent with a ... ttggttgg ... chain conformation observed previously for syndiotactic polypropylene (s-PP) (67). Figure 15 presents the high resolution solid-state CPMAS/DD 13C NMR spectra of forms I and I1 s-PS. A single CH2 carbon resonance at 48.4 ppm is observed in the form I spectrum, while two CH2 resonances at 49.1 and 38.1 ppm are seen in the form I1 spectrum. These 13C chemical shifts are consistent with those expected (1) from γ-gauche shielding effects if the chains adopt the planar zigzag, ... tttt.. . and 21 helical, ... ttggttgg ... conformations, respectively. In the ... ttggttgg ... conformation half of the CH2 carbons are gauche to two γ substituents (CHs),while the remaining half are trans to both γ-CH substituents.

Figure 15. CPMAS/DD 13C NMR spectra of form I (a) and form I1 (b) s-PS. The form I1 sample of (b) was obtained by absorption of dichloro-methane into an amorphous, melt-quenched film of s-PS.

180

As was observed (68) in the CPMAS/DD 13C NMR spectrum of s-PP, we expected and see (69–72) two CH2 resonances separated by two γ-effects, or ca. (2)(-5 ppm) = -10 ppm. The separation observed in Figure 15 is 11 ppm. In form I crystals, the s-PS chains assume the all trans .... tttt ... conformation, so all methylene carbons are in the trans arrangement with their γ substituents (CH’s). As expected we do see a single CH2 at nearly the same chemical shift as the most downfield of the two CH2 resonances observed in the spectrum of the form II polymorph. The application of high resolution 13C-NMR (CPMASS/DD) to crystalline s-PS has clearly resulted in confirming the X-ray derived chain conformations adopted in the form 1 and 11 crystalline polymorphs.

Poly(tri-methylene terephthalate) and Solid Model Compounds

As we have attempted to indicate here, it has been repeatedly demonstrated (1) that the resonance frequencies observed for carbon nuclei in the high resolution solid-state 13C NMR spectra of organic polymers depend principally upon and can be analyzed by means of the conformationally sensitive γ-gauche effect (1). This whether polymers are constrained in their crystals to adopt single rigid conformations or are molten, mobile, and free to interconvert rapidly on the megahertz time scale between conformations (73). Though I hesitate to end this review chapter on a negative note, the failure to confirm the crystalline chain conformation of poly (trimethylene terephthalate) (PTT) by the γ-gauche analysis of its high resolution13C-NMR spectra should not be glossed over (74). In Figure 16 the crystalline conformations and resonance frequencies (in ppm vs TMS) of the central methylene carbons in the butylene glycol fragments of poly(butylene terephthalate) (PBT), along with those of several of its model compounds, are presented (75, 76). Because the model compounds were single crystal samples, their crystalline conformations are firmly established (76). Consistent with the γ-gauche effect, their observed 13C resonance frequencies have shown that the central methylene carbons in the butylene glycol fragment that are gauche to their γ-substituent ester oxygens are shielded (Figure 16) and resonate 3-4 ppm upfield from those that are in a trans arrangement (75). Since here want to compare the high-resolution solid-state13C NMR spectra and conformations of the closely related aromatic polyester PTT, whose structure is shown in Figure 17, the comparison to crystalline PBT is very relevant.

181

Figure 16. Conformations and 13C NMR resonances observed for the central methylene carbons in crystals of PBT and its model compounds (75, 76).

Figure 17. PTT repeat unit. The propylene glycol fragments in crystalline PTT are believed to adopt alternating tg±g±t (... tg+g+ttg-g-t ...) conformations, where all –O—CH2–and –CH2—O– bonds are trans (t) and –CH2—CH2– bonds are gauche (g±) (77). In the proposed crystalline conformation of PTT, the central methylene carbons in the propylene glycol fragments would be trans (φ1,4 = t) to both of their γ-substituent carbonyl carbons. In the absence of the requisite gauche shielding arrangement, the central methylene carbons would not be expected to be shielded by their neighboring carbonyl carbons, as seen below.

In solution, in the melt, and in the amorphous solid regions of PTT, on the other hand, the –O—CH2– and –CH2—O– bonds in PTT would adopt both trans and gauche conformations (φ1,4 = t and g±). There the central methylene 182

carbons (CH2)o would be expected to experience significant shielding from their immediately adjacent C=O’s [~5 ppm for each γ-(C=O) in a gauche arrangement] causing them to resonate upfield from their crystalline counterparts. In the high resolution solid-state 13C-NMR spectra of amorphous and semicrystalline PTTs (See Figure 18) the central CH2 carbons (CH2)o, to the contrary, show resonances at δ(13C) = 28.8(amorphous) and 26.1 ppm (crystalline). In other words, the (CH2)o methylene carbons in the crystalline regions of PTT are shielded in comparison to those in the amorphous regions and so resonate upfield. This in direct opposition to expectation based on the crystalline conformation proposed for PTT, with φ1,4 = t –O—CH2– and –CH2—O–bonds and therefore no γ-(C=O)s would be in a gauche shielding arrangement.

Figure 18. 13C-CP/MAS-NMRspectra of PTT films: (A) amorphous, (B) film (A) annealed at 160°C for 30 min, and (C) film (B) drawn to a DR= 3. 183

Single crystal X-ray diffraction of the PTT model compound trimethylene glycol dibenzoate (TMGDB) has revealed a crystalline conformation (78) essentially identical to that proposed for crystalline PTT. Though the spectra are not presented here, we have observed (74) the following resonance frequencies for the central methylene carbons (CH2)o inTMGDB: 30.0 ppm (melt, 60° C), 28.7 ppm (solution), 27.4 ppm (crystal).

In agreement with PTT, but again unexpectedly, the central methylene carbons in crystalline TMGDB resonate substantially (3 ppm) upfield from those in its melt or solution, as do the chemical shifts observed for the central methylene carbons in crystalline and amorphous regions of PTT. These solid-state 13C-NMR observations lead to the following conclusions: 1. 2.

As previously concluded by X-ray (77, 78), PTT and TMGDB have closely similar crystalline conformations. Both of their central methylene (CH2)o δ(13C) resonances seem not to be influenced by the nuclear shielding usually produced by gauche conformational arrangements between carbon nuclei and their γ-carbon substituents.

Conclusion 2. appears valid despite the fact that γ-gauche shielding successfully explains the 13C resonances observed in the crystals of the closely related aromatic polyester PBT and its model compounds (75, 76), in addition to those of many other crystalline polymers (1). Figure 19 illustrates the crystalline conformation of PTT. Unlike the related terephthalate polyesters PBT and poly (ethylene terephthalate), the PTT chain is not fully extended. Instead the PTT trimethylene glycol segments gently serpentine back and forth about the line connecting the centers of their phenyl rings. This prompted an investigation (74) of how ring currents generated by the π-electrons of the phenyl rings in PTT might possibly affect the resonance frequencies of their “central” (CH2)o carbons (79–81). Molecular modeling was used to evaluate the distances between the central methylene carbons and the centers of the phenyl rings in plane (ρ) and perpen-dicular to the plane (z) of the phenyl rings in their crystalline unit cells. These distances were obtained for all phenyl rings from the x, y, z coordinates obtained from the crystalline unit cells of PTT and TMGDB (77, 78). We repeated the distance calculations for a dynamic TMGDB molecule in vacuum (74) to account for its conformational flexibility in the melt or solution by obtaining conformationally weighted average distances. Because shielding from ring-currents vanish for ρ > 4Å or z > 3Å, Table 4 makes clear that all of the central methylenes (CH2)o in both PTT and TMGDB are too far removed from their nearest phenyl rings to be significantly shielded by their ring-currents. 184

Figure 19. PTT crystalline conformation (77).

The potential effects of the magnetic susceptibility anisotropy of neighboring carbonyl groups might have on the nuclear shielding of the central methylene carbons were also investigated using a similar calculation. In this case, (CH2)o in-plane and perpendicular to the plane distances of the C=O group were similarly determined for both rigid crystalline and flexible isolated TMGDB. 185

Table 4. Distances (Å) of “Central” CH2s to the Center of (ρ) and above or below (z) the Phenyl Rings in PTT and TMGDBa

The affects of neighboring carbonyl group anisotropies were estimated for TMGDB according to a previously implemented procedure (82), that adopts a point-dipole approximation (83, 84) for magnetically anisotropic groups, such as the C=O group. The known susceptibility value of the formaldehyde carbonyl group (85) was adopted. These calculations performed on both the rigid crystalline and conformationally averaged of the TMGDB model compound resulted in a maximum possible neighboring C=O group shielding contribution of 2.2 ppm. Remember that the central methylene carbons in TMGDB resonate at 30.0 ppm in the melt, at 28.7 ppm in solution), and at 27.4 ppm in the crystal. In other words, the flexible TMGDB resonances come 1.3 - 2.4 ppm downfield from rigid crystall-ine TMGDB, while analyses of γ-gauche conformational shielding suggests they be more shielded than crystalline TMGDB by 3 - 4 ppm. Our estimate of a maximum ~ 2.2 ppm δ(rigid crystal) - δ(flexible liquid) shielding of 186

“central” methylene carbons produced by the anisotropic magnetic susceptibility of neighboring C=O groups in TMGDB is far short of that necessary to overcome the larger 5 - 7 ppm downfield shift expected from the apparent lack of γ-gauche conformational shielding in the TMGDB crystal. Since PTT and TMGDB have closely similar crystalline conformations (77, 78), with all of their (CH2)o methylene carbons and their γ-substituent carbonyl carbons in a trans conformational arrangement, the resonances observed in their melts and solutions were expected to come several ppm upfield from the their crystalline resonances, and not several ppm downfield. This means that the discrepancies between observed resonance frequencies and those anticipated from consideration of conformationally sensitive γ-gauche effects are very large, in the range 5 - 7 ppm. Accounting for the potential effects of phenyl ring currents and anisotropic carbonyl group magnetic susceptibilities did not significantly reduce this large discrepancy, and so its cause currently remains perplexing.

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

16.

17. 18.

Tonelli, A. E. NMR Spectroscopy and Polymer Microstructure:The Conformational Connection; VCH: New York, 1989. Beran, G. J. O.; Hartman, J. D.; Heit, Y. N. Acc. Chem. Res. 2016, 49, 2501–2508. Pradeepa, S. J.; Tamilvendan, D.; Boobalan, M. S.; Sundaraganesan, N. J. Mol. Struct. 2016, 1112, 33–44. Sahin, Z. S.; Kantar, G. K.; Sasmaz, S. J. Mol. Struct. 2015, 1103, 156–165. Maerker, K.; Pingret, M.; Mouesca, J.-M.; Gasparutto, D.; Hedigerand, S.; De Paëpe, G. J. Am. Chem. Soc. 2015, 137, 13796–13799. Li, L.; Cai, T.; Wang, Z. Spectrochim. Acta, Part A 2014, 120, 106–118. Moore, K. W.; Tibbetts, R. L.; Pelczer, I.; Herschel Rabitz, H. Chem. Phys. Lett. 2013, 572, 1–12. Muthu, S.; Ramachandran, G.; Maheswari, J. U. Spectrochim. Acta, Part A 2012, 93, 214–222. Muthu, S.; Maheswari, J. U. Spectrochim. Acta, Part A 2012, 92, 154–163. Busico, V.; Cipullo, R. Prog. Polym. Sci. 2001, 26, 443–533. Wei, M.; Ivey, D. T.; Tonelli, A. E. Macromolecules 2002, 35, 1976–1979. Spiesecke, H.; Schneider, W. G. J. Chem. Phys. 1961, 35, 722–730. Grant, D. M.; Paul, E. G. J. Am. Chem. Soc. 1964, 86, 2984–2990. Lindeman, L. P.; Adams, J. Q. Anal. Chem. 1971, 43, 1245–1252. Dorman, D. E.; Carhart, R. E.; Roberts, J. D. In Proceedings of the International Symposium on Macromolecules, Rio de Janerio, July 26-31, 1974; Mano, E. B., Ed.; Elsevier: New York, 1974. Bovey, F. A. In Proceedings of the International Symposium on Macromolecules, Rio de Janerio, July 26-31, 1974; Mano, E. B., Ed.; Elsevier: New York, 1974; p 169. Stothers, J. B. C-13 NMR Spectroscopy; Academic Press: New York, 1972. Statistical Mechanics of Chain Molecules; Flory, P. J. Wiley Interscience: New York, 1969 187

19. VanderHart, D. L. J. Magn. Reson. 1981, 44, 117–125. 20. Tart, E.; Wood, G.; Wernsman, D.; Sangwatanaroj, U.; Howe, C.; Zhou, Q.; Zhang, S.; Tonelli, A. E. Macromolecules 1993, 26, 4283–4286. 21. Sung, H. N.; Noggle, J. H. J. Polym. Sci., Polym. Phys. Ed. 1981, 19, 1593–1602. 22. Rathke, T. D.; Frey, M. W.; Guthrie, D.; Graham, R.; Simendinger, W.; Wang, B.-C.; Shepard, T.; Jones, R.; Tonelli, A. E. Computational Polym. Sci. 1993, 61, 3–6. 23. Abe, A.; Jernigan, R. L.; Flory, P. J. J. Am. Chem. Soc. 1966, 88, 631–639. 24. Sundararajan, P. R. Macromolecules 1978, 11, 256–263. 25. Polymer Spectroscopy; Fawcett, A. H., Ed.; Wiley: New York, 1996; Chapters 2 and 3. 26. Rusa, C. C.; Bridges, C.; Ha, S.-W.; Tonelli, A. E. Macromolecules 2005, 38, 5640–5646. 27. Uyar, T.; Rusa, M.; Tonelli, A. E. Macromol. Rapid Commun. 2004, 25, 1382–1386. 28. Wei, M.; Ivey, D. T.; Tonelli, A. E. Macromolecules 2002, 35, 1976–1979. 29. Tonelli, A. E. Macromolecules 1991, 24, 3065–3068. 30. Zhou, Z.; Abe, A. Polymer 2004, 45, 1313–1320. 31. Chance, R. R. Diacetylene Polymers. Encyclopedia of Polymer Science and Engineering; Wiley: New York, 1986; Vol. V. 32. Rao, D. N.; Chopra, P.; Ghoshal, S. K.; Swiatkiewicz, J.; Prasad, P. N. J. Chem. Phys. 1986, 84, 7049–7050. 33. Bloor, D.; Ando, D. J.; Normal, P. A.; Obhi, J. S.; Kolinowskv, P. V.; Movaghar, B. Phys. Scr. 1987, T19, 226–231. 34. Greene, B. I.; Orenstein, J.; Millard, R. R.; Williams, L. R. Chem. Phys. Lett. 1987, 139, 381–385. 35. Patel, G. N.; Chance, R. R.; Witt, J. D. J . Polym. Sci., Polym. Lett. Ed. 1978, 16, 607–614. 36. Chance, R. R.; Patel, G. N.; Witt, J. D. J. Chem. Phys. 1979, 71, 206–211. 37. Patel, G. N.; Miller, G. G. J. Macromol. Sci., Part B 1981, B20, 111–131. 38. Xu, R.; Chu, B. Macromolecules 1989, 22, 3153–3160, 4523–4528. 39. Lim, K. C.; Fincher, C. R., Jr.; Heeger, A. J. Phys. Rev. Lett. 1983, 50, 1934–1937. 40. Wenz, G.; Muller, M. A.; Schmidt, M.; Wegner, G. Macromolecules 1985, 17, 837–850. 41. Lim, K. C.; Heeger, A. J. J. Chem. Phys. 1985, 82, 522. 42. Nava, A. D.; Thakur, M.; Tonelli, A. E. Macromolecules 1990, 23, 3055–3063. 43. Schaefer, J.; Stejskal, E. O. In Topics in Carbon-13 NMR Spectroscopy; Levy, G. C., Ed.; Wiley-Interscience: New York, 1979; Vol. 4, p 283. 44. Tabor, B. J.; Magre, E. P.; Boon, J. Eur. Polym. J. 1971, 7, 1123–1133. 45. Lovinger, A. J.; Padden, F. J., Jr.; Davis, D. D. Polymer 1988, 29, 229–232. 46. Boon, J.; Magre, E. P. Makromol. Chem. 1969, 126, 130. 47. Boon, J.; Magre, E. P. Makromol. Chem. 1970, 136, 267. 48. Gomez, M. A.; Tonelli, A. E. Polymer 1991, 32, 796–801. 49. Opella, S. J.; Frey, M. H. J. Am. Chem. Soc. 1979, 101, 5854–5856. 188

50. Kaplan, M. L.; Reents, W. D., Jr. Tetrahedron Lett. 1982, 23, 373–374. 51. Reents, W. D., Jr.; Kaplan, M. L. Polymer 1982, 23, 310–313. 52. Kaplan, M. L.; Reentz, W. D., Jr.; Day, C. S. Cryst. Struct. Commun. 1982, 11−18, 1751. 53. Schaefer, J.; Stejskal, E. O.; Buchdall, R. Macromolecules 1977, 10, 384–405. 54. Marciq, M.; Waugh, J. S. Chem. Phys. Lett. 1977, 47, 327–329. 55. Lipmaa, E.; Alia, M. A.; Pehk, T. J.; Engelhardt, G. J. Am. Chem. Soc. 1978, 100, 1929–1931. 56. Steger, T. R.; Stejskal, E. O.; McKay, R. A.; Stults, B. R.; Schaefer, J. Tetrahedron Lett. 1979, 20, 295–296. 57. Garbarcyzk, J. Makromol. Chem. 1986, 187, 2489–2495. 58. Andreetti, G. D.; Garbarcyzk, J.; Krolikowska, M. Cryst. Struct. Commun. 1981, 10, 57–62. 59. Tonelli, A. E.; Chesnut, D. B. Macromolecules 1996, 29, 2537–2542. 60. Ishihara, N.; Seimiya, T.; Kuramoto, N.; Uoi, M. Macromolecules 1986, 19, 2464–2465. 61. Pellecchia, C.; Longo, P.; Grassi, A.; Ammendola, P.; Zambelli, A. Makromol. Chem., Rapid Commun. 1987, 8, 277–279. 62. Immirzi, A.; de Candia, F.; Ianelli, P.; Zambelli, A.; Vittoria, V. Makromol. Chem., Rapid Commun. 1988, 9, 761–764. 63. Greis, X. Y.; Arsano, T.; Petermann, J. Polymer (Br.) 1989, 30, 590–594. 64. Nyquist, R. A. Appl. Spectrosc. 1989, 43, 440–442. 65. Reynolds, N. M.; Savage, J. D.; Hsu, S. L. Macromol. Chem., Rapid Commun. 1989, 9, 765–769. 66. Vittoria, V.; de Candia, F.; Ianelli, P.; Immirzi, A. Macromol. Rapid Commun. 1989, 26, 1590–1593. 67. Natta, G.; -Pasquon, I.; Corradini, P.; Peraldo, M.; Pegoraro, M.; Zambelli, A. Atti Accad. Naz. Lincei, C1. Sci. Fis., Mat. Nat., Rend. 1960, 28, 539–541. 68. Bunn, A.; Cudby, E. A.; Harris, R. K.; Packer, K. J.; Say, B. J. Chem. Soc., Chem. Commun. 1981, 15–16. 69. Gomez, M. A.; Tonelli, A. E. Macromolecules 1990, 23, 3385–3386. 70. Gomez, M. A.; Jasse, B.; Cozine, M. H.; Tonelli, A. E. J. Am. Chem. Soc. 1990, 112, 5881–5882. 71. Gomez, M. A.; Tonelli, A. E. Macromolecules 1991, 24, 3533–3536. 72. Grassi, A.; Longo, P.; Guerra, G. Makromol. Chem., Rapid Commun. 1989, 10, 687–690. 73. Spiess, , H. W. Flory Prize Lecture: The Role of Conformations in the Interplay of Structure and Dynamics in Macromolecular and Supramolecular Systems. Macromol. Symp. 2010, 298, 10–16. 74. Vasanthan, N.; White, J. L.; Gyanwali, G.; Shin, I. D.; Majikes, J.; Pasquinelli, M. A.; Tonelli, A. E. Macromolecules 2011, 44, 7050–7055. 75. Gomez, M. A.; Cozine, M. H.; Tonelli, A. E. Macromolecules 1988, 21, 388–392. 76. Grenier-Loustalot, M.-F.; Bocelli, G. Eur. Polym. J. 1984, 20, 957. 77. Poulin-Dandurand, S.; Perez, S.; Revol, J.-F.; Brisse, F. Polymer 1976, 20, 419–426. 189

78. 79. 80. 81. 82. 83. 84. 85.

Perez, S.; Brisse, F. Acta Crystallogr. 1977, B33, 3259–3262. Johnson, C. E., Jr.; Bovey, F. A. J. Chem. Phys. 1958, 29, 1012–1014. Waugh, J. S.; Fessenden, R. W. J. Am. Chem. Soc. 1958, 80, 6697–6699. Bovey, F. A. Nuclear Magnetic Resonance Spectroscopy, 2nd ed.; Academic Press: New York, 1988; Chapter 3. White, J. L.; Beck, L. W.; Haw, J. F. J. Am. Chem. Soc. 1992, 114, 6182–6189. McConnell, H. M. J. Chem. Phys. 1957, 27, 226–229. Pople, J. A.; Schneider, W. G.; Bernstein, H. J. High Resolution Nuclear Magnetic Resonance. McGraw-Hill: New York, 1959. Flygare, W. H. J. Chem. Phys. 1965, 42, 1563–1568.

190

Chapter 11

Solution NMR Structure and Conformation of Silk Fibroins Stored in Bombyx mori and Samia cynthia ricini Silkworms Tetsuo Asakura,1,* Yu Suzuki,2 and Akio Nishimura1 1Department

of Biotechnology, Tokyo University of Agriculture and Technology, 2-24-16, Nakacho, Koganei, Tokyo 184-8588, Japan 2Tenure-Track Program for Innovative Research, University of Fukui, 3-9-1 Bunkyo, Fukui, Fukui 910-8507, Japan *E-mail: [email protected]

The structure of the typical tandem repeated sequence, GSGSGA, of silk fibroin stored in the middle silk glands of Bombyx mori was determined with solution NMR at the atomic level. Type II β-turn structure was found, similar to that determined for the same sequence in Silk I* in the solid state as defined in the text. The structure assumes an approximately random coil conformation but exists in aggregated states. Likewise, the structure of another typical tandem repeated sequence, YGGDGG(A)12GGAG, of silk fibroin stored in the middle silk glands of a wild silkworm, Samia cynthia ricini, was also determined with solution NMR. The polyalanine region assumes the typical α-helix conformation, but at the N-terminal and C-terminal sites of the sequence there is additional intramolecular hydrogen bonding which helps to stabilize the helical structure of the polyalanine region.

Introduction Silks are produced mainly by moth larvae and spiders and are very attractive materials because of their excellent mechanical properties, especially their high strength and toughness (1). It is useful to develop a better understanding of the silk fiber formation mechanism as a means of improving the properties of man-made natural fiber (2). The silk fibroin from domesticated silkworm, Bombyx mori (B. © 2017 American Chemical Society

mori), is the most well-studied silk in the world, and much information on its structure before and after spinning (including fiber formation mechanism) has been acquired. However, recent developments in spectroscopic methods have supplied more detailed information on silk structure and sometimes even prompted revisions of some aspects of the structure which have been widely accepted for a long time. Nuclear magnetic resonance, NMR, is one of these spectroscopic methods that can be used to study polymer structure at the atomic level. For example, in our previous paper (3) and review (4), we have pointed out that the Marsh-Pauling model (5), which was proposed (and accepted) for the structure of B. mori silk fibroin fiber, should be revised. In particular, the inter-molecular chain arrangement in the silk structure should be changed from polar to anti-polar β-sheet structure. Here, the methyl groups of Ala residues are only on one side of the sheet in the polar structure (Marsh-Pauling model), but the methyl groups alternately point to both sides of the sheet along the hydrogen bonding direction in the antipolar structure (our model). Thus, whereas in the past the inter-molecular hydrogen bonding pair between two silk fibroin chains was regarded as having anti-parallel β-sheet structure (3, 6), we proposed a heterogeneous β-sheet structure instead (3, 4). As for the B. mori silk fibroin structure before spinning, we recently proposed new type II β-turn coordinates for the main repeated sequence GSGAGA in the polypeptide chain (7). A number of factors are known to affect the silk fiber formation in B. mori silkworm (8). These include changes in pH, ion concentrations, silk concentration, water content, shear stress, and stretching force. We mainly studied stretching force and shear stress because we believe these are the primary external forces for fiber formation. We based our assumptions on the known cocooning behavior of the silkworm (i.e., the larva spinning the silk fiber via head movement to depict the number eight (9)) and the significant change in the cross-sectional area and shapes of the press part in the anterior silk gland just before spinning (10). For example, previously the pH of the B. mori silk gland lumen had been reported to change from approximately 6.9 in the posterior silk gland to 4.9 in the anterior silk gland (11). This significant pH change from neutral to acidic was considered important factors in the fiber formation. However, recent pH determination in silk fibroin in silk gland using concentric ion selective microelectrodes by Domigan et al (12) reported different pH results. They observed a pH gradient from pH 8.2 to 7.2 in the posterior silk gland, a constant pH 7 throughout the middle silk gland, and a pH gradient from 6.8 to 6.2 in the beginning of the anterior silk gland where silk-to-fiber processing occurs. In this paper the concentrated aqueous solution of silk fibroin stored in the middle silk gland is called “liquid silk.” It is important therefore to have a better understanding of the structure of liquid silk, which is a good starting point to figure out the fiber formation mechanism. In vivo 13C solution NMR spectra of B. mori and a wild silkworm, Samia cynthia ricini (S.c. ricini) silk fibroins stored in living silkworms were first reported by us in 1983 together with the spectrum of pupa (Figure 1) (13, 14). The shape of these fifth-larval-stage silkworms and pupa is suitable for direct observation in a 13C NMR sample tube with a 10-mm outer diameter. High-resolution NMR spectra from the liquid silks (Figure 1a,c) could be 192

obtained and assigned easily by comparison with the pupa spectrum (Figure 1b). Because the structure of the liquid silk changes easily due to small external forces during the process of preparation from silk glands (as described later), structural determination of liquid silk without any external forces is essential.

Figure 1. 13C solution NMR spectra (expanded spectrafrom 0 ppm to 100 ppm) of the middle silk gland portions of both living S. c. ricini (a) and B. mori (c) mature larvae and of the abdomen of S. c. ricini pupa (b) together with the assignments. This figure was adapted from ref. (13), but with a minor change. Copyright 1983 American Chemical Society.

The conformation and dynamics of the regenerated silk fibroin together with liquid silk prepared from silk gland directly were studied as a function of their concentration in aqueous solution using 13C conformation-dependent chemical shifts (15). The 13C chemical shifts of all carbons of the Gly, Ala, Ser, Tyr, and Val residues (comprising 92.5% of all amino acids) did not change from dilute aqueous solution (at 2.1% concentration) to liquid silk (at a concentration of ca. 30%) and are almost the same as the 13C chemical shifts of random coil Ac-X-NHMe (X = 193

G, A, S, Y, V) in the aqueous solution. The 13C T1 spin-lattice relaxation times and NOEs were also observed for these aqueous solutions from 2.1% to 14.5% for regenerated silk fibroin and liquid silk. The mean correlation time of the backbone of the silk fibroin calculated from both 13C T1 and NOE values increased gradually from 0.10 ns (2.1%) to 0.22 ns (liquid silk) with increasing concentration. This suggests that some aggregation of the polypeptide chains occurs, causing the chain motion to decrease as the concentration increases. In addition, the plots of ln(M∞ − Mτ) vs τ for the liquid silk stored in living silkworms were essentially single exponentials for all carbons of the Ala, Gly, Ser, and Tyr residues. This observation suggests the presence of only one dynamic component in the liquid silk (16). (M∞ is the equilibrium amplitude of the fully relaxed spectrum and Mτ is the amplitude of a partially relaxed spectrum at delay time τ.) Thus, these data indicate that the local structure of liquid silk is close to that of a random coil, which undergoes fast segmental motion in aqueous solution even in the silk gland in spite of chain aggregation (16). In this paper, we provide a review of our recent research relating to the solution structures and conformations of the silk fibroins stored in the silk glands of B. mori and S.c. ricini. The solution structure of the S.c. ricini silk fibroin is particularly interesting because its primary structure is very different from that of B. mori. Whereas B. mori silk fibroin is an approximately alternating copolymer of Ala and Gly residues (17), S.c. ricini has both a tandem poly-Ala repeat motif and a Gly-rich motif (18). As a result, S.c. ricini takes on an α-helical structure in the aqueous solution (19–22). The tandem poly-Ala repeat motif and the Gly-rich motif are similar to the primary structure of spider drag-line silk (1, 8, 23, 24) although the length of poly-Ala here is longer than that of the spider silk.

Experimental The silk glands containing liquid silk were prepared from the fifth instar larvae of B. mori and S.c. ricini silkworms. The glands were soaked in a dish filled with distilled H2O so that the liquid silk could be gently removed from the silk glands. The liquid silk was then soaked in fresh distilled H2O to remove another silk protein, silk sericin, from the liquid silk surface in a mild manner. The liquid silk was carefully placed in a 5-mm NMR tube together with a sealed capillary containing D2O for NMR 2H lock while avoiding the application of any external forces to the liquid silk. There is no significant difference between two silk fibroins in these sample preparations for the solution NMR although the shapes of the silk gland are considerably different. A uniformly 13C-labeled liquid silk was also prepared by feeding U−13C D-glucose (99% enrichment, CIL, USA) in addition to an artificial diet fed to silkworm larvae from the fourth to sixth days of the fifth instar (4, 23). The liquid silks prepared for NMR measurements were stored at 4 °C prior to their use for the NMR experiments. All NMR experiments were carried out at 10 °C on a Bruker Avance 750 spectrometer equipped with a 5-mm inverse triple-resonance probe head with three-axis gradient coils, operating at a 1H frequency of 750 MHz. The chemical 194

shifts were indirectly referenced to TSP used as an external reference for 1H and 13C and to liquid NH3 for 15N. Resonance assignments for the 1H signals were initially accomplished by 2D TOCSY and 2D NOESY experiments and confirmed by hetero-nuclear NMR experiments. 1H, 13C, and 15N sequential resonance assignments were obtained by 2D double resonance and 3D double and triple resonance through bond-correlation experiments. Torsion angle constraints for the main chain were derived from database analysis of the chemical shifts (13Cα, 13Cβ, 13CO, 1Hα, 1HN, and 15N) of the backbone atoms using TALOS-N program (25, 26). Energy minimization using MOPAC was performed in order to obtain a final structural model.

Results Bombyx mori For the determination of the solution structure of silk fibroin at atomic level, we determined the torsion angles of the liquid silk using solution NMR (7). Torsion angle constraints for the main chain were derived from the backbone chemical shifts with the program TALOS-N. Moreover, inter-residue NOE cross-peaks, whose intensities are inversely proportional to the sixth power of inter-proton distances, were analyzed to obtain spatial inter-proton distance information.

Spectral Assignments for the Repeated Sequence Motifs Figure 2a shows the TOCSY spectrum used to assign the peaks of natural abundant B. mori liquid silk to the Gly, Ala, Ser, Tyr and Val residues. Figure 2b provides the sequential assignments in the HN/Hα region of the NOESY spectrum for the tandem partial sequence, GAGSGAG, with selected portions of the 1H15N HSQC and 1H-13C HSQC spectra of the natural-abundance liquid silk. The spectra show relatively sharp and fairly well-separated cross-peaks despite the very high molecular weight of the silk fibroin. This feature and the repetitive sequence made assignments of the 1H, 15N, and 13C peaks possible for the residues in the repetitive motifs by multidimensional NMR. Sequential assignments for each repetitive motif were also accomplished by analyzing the Hα/HN region in the NOESY spectrum. The 15N and 13C chemical shifts were assigned by 1H-15N and 1H-13C HSQC, and the consistency of the 1H chemical shift assignments was confirmed. The H-chain of the B. mori silk fibroin with 5263 residues consists of 12 tandem repeats in the primary structure (17). There are several repeated sequence motifs including 428, 144, and 31 copies of the hexapeptides GAGSGA, GAGYGA, and GAGVGA, respectively, and 41 copies of tetrapeptide GAAS. For convenience, a more detailed structural determination has been performed for the most abundant sequence, GAGSGA. 195

Figure 2. (a) Assignments for the Gly, Ala, Ser, Tyr, and Val peaks in the TOCSY spectrum and (b) sequential assignments in the HN/Hα region of the NOESY spectrum for GAGSGAG with portions of the 1H−15N HSQC and 1H−13C HSQC spectra of B. mori liquid silk. This figure is reproduced from ref. (7). Copyright 2014 American Chemical Society.

Torsion Angle (Φ, φ) Map Obtained Using TALOS-N The chemical shift values of Ala 13Cα and 13Cβ carbons indicated that these motifs have neither the typical α-helix nor the β-sheet structures (27). To evaluate the unique conformation of the repetitive motifs, we employed the backbone torsion angles prediction program, TALOS-N from NIH (25, 26). TALOS-N is a database system for empirical prediction of backbone torsion angles (Φ, φ) using a combination of six types of backbone chemical shifts and sequence information. Figure 3a shows the (Φ, φ) plots of the 25 closest database matches predicted for the GAGSGAG motif obtained using TALOS-N program. This result shows that the torsion angles (Φ, φ) of the residues (A2-G7) fall into the average torsion angles of the two center residues (residues i+1 and i+2) of the type II β-turn structure observed in the structure database, (Φi+1, φi+1) = (-61 ±13˚, 136 ± 11˚) and (Φi+2, φi+2) = (80 ±16˚, 5 ± 20˚) (25). Figure 3b shows the structural models constructed using the average torsion angles for the best matches (Φ, φ) 196

for the repeated GAGSGAG motif. The models for GAGSGAG show hydrogen bond formation between the HN of i-th and the CO of (i+3)th residues, which characterize the β-turn structure.

Figure 3. (a) the 25 best matches for torsion angles (Φ, ψ) for the GAGSGAG motifs obtained using the TALOS-N program; and (b) a structural model constructed using the averaged (Φ, ψ) in the circle for each motif. Hydrogen bonds are assumed to exist between the HN of the i-th and the CO of the (i+3)-th residues for the GAGSGAG motifs. This figure is reproduced from ref. (7). Copyright 2014 American Chemical Society. Thus, all the data are consistent with a strong preference for repeated type II β-turn structures for the typical tandem repeated sequences (GAGSGA)n of B. mori silk fibroin. This type II β-turn structure is essentially the same as Silk I* structure in the solid state (Figure 4) (4, 27–29), which is locally similar to the conformations preferred by random coil peptides in the aqueous solution (30) but is more ordered; in particular, it contains many hydrogen bonds that serve to hold it together (Figure 4) (27–29). Interestingly, molecular dynamics (MD) simulations of model peptides in explicit water indicate that the torsion angles of the Gly, Ala, and Ser residues in a repeated AGSGAG sequence with type II β-turn structure in water are close to the values expected at energy minima (30). Therefore, the amplitude and width of the structural fluctuations of Gly, Ala, and Ser residues may be considered relatively large for random coil conformation in the dilute aqueous solution, but the fluctuation decreases due to aggregation of the silk fibroin molecules with increasing concentration (16), consistent with an increased population of Silk I*. In other words, the highly concentrated silk 197

solution contained in the middle silk gland has residues in energetically favored conformations close to average random coil values but forms a hydrogen-bonded network that keeps it in a repeated type II β-turn structure.

Figure 4. Packing structure of poly(AG) chains with type II β-turn as a model for Silk I*form. Dotted lines denote intra- and intermolecular hydrogen bonds. The unit lattice values, a, b, and c, were obtained from X-ray diffraction data. This figure is reproduced from ref. (28). Copyright 2005 American Chemical Society.

Samia cynthia ricini The S.c. ricini silk fibroin consists of 93 tandem repeats of a polyalanine (poly-Ala) region flanked by Gly-rich regions. The length of the poly-Ala is distributed from 10 to 14, indicating a narrow distribution of length (18). The conformation of the liquid silk has been studied by 13C in vivo and solution NMR (13, 14, 19–22), and the poly-Ala region gave a single peak due to rapid conformational exchange between the α-helix and random coil. As the temperature was increased, the main Ala CO and Ala Cα peaks shifted to higher field; meanwhile, the main Ala Cβ peak shifted to lower field due to an increase in the random coil fraction (19–21). These were the same shift tendency observed for urea-induced helix-coil transition of S.c. ricini liquid silk (13). The proportion of 198

helix component obtained from the NMR spectrum was quantitatively consistent with that obtained from circular dichroism (CD) (19). A typical amino acid sequence from the Gly-rich region is YGGDGG, which appears 36 times at the N-terminal side, and GGAG appears 65 times at the C-terminal side of the poly-Ala region (17). Thus, the sequence Y1GGDGG6(A)12G19GAG22 can be used as a representative repeat sequence for S.c. ricini silk fibroin. Assignments of 1H, 15N, and 13C resonances for the residues in the repetitive motif were obtained by the same strategy as that applied to B. mori. The sequential assignment in the HN/Hα and HN/HN regions of the NOESY spectrum for the N terminal part, YGGDGGA, was shown in Figure 5.

Figure 5. Sequential assignments in the HN/Ha and HN/HN regions of NOESY spectrum for the N-terminal part, Y1GGDGGA7 of S. c. ricini liquid silk. This figure is reproduced from ref. (31). Copyright 2015 American Chemical Society.

Torsion Angles Derived from the Chemical Shifts The torsion angles for each residue were predicted by TALOS-N as shown in Figure 6. The Φ and φ angles of the poly-Ala region were − 62 ± 2° and −30 ± 11°, respectively (31). These values were similar to those of a typical α-helix, 199

where (Φ, φ) = (−60°, −45°). This result confirms that the poly-Ala region of S.c. ricini silk fibroin formed an α-helix in liquid silk. The torsion angles of N and C-terminal residues did not indicate typical secondary structure formation. A model structure was built and energy minimized based on the torsion angles obtained from MOPAC, as shown in Figure 7a (31). The poly-Ala region formed an α-helical structure. In an α-helix in general, the first four NH groups and last four CO groups necessarily lack intra-helical hydrogen bonds and instead are often capped by alternative hydrogen bond partners. At the N-terminal region of the model structure, hydrogen bonds were formed between CO of Gly6 and NH of Ala9 and between the CO of Gly6 and NH of Ala10. These findings demonstrate that Gly6 serves as the N-cap-like motif for the poly-Ala α-helix. At the C-terminal region of the model structure, hydrogen bonds are formed between the CO of Ala16 and NH of Ala21 and between the CO of Ala17 and NH of Gly20, as shown in Figure 7b. These hydrogen bond combinations are characteristic of the Schellman C-cap motif (32). Specifically, the Schellman C-cap motif is defined as a six-residue fragment that is located at the end of an α-helix and exhibits a double hydrogen bond pattern, namely NH of (i + 5) and CO of (i), NH of (i + 4) and CO of (i + 1). The six-residue fragment from Ala16 to Ala21 in the S.c. ricini model structure contains the same combination of hydrogen bonds as the Schellman C-cap motif. The N- and C-cap motifs stabilize the poly-Ala α-helix. Indeed, the capping motifs may play a role in preventing the structural transition from α-helix to β-sheet and fibril formation inside the silkworm body, which would be fatal to the silkworm.

Figure 6. A plot of torsion angles for each residue of the S.c. ricini repeated motif in liquid silk derived from TALOS-N. ♦ and □ denote f and y, respectively. Error bars show the estimated standard deviation provided by TALOS-N. This figure is reproduced from ref. (31). Copyright 2015 American Chemical Society.

200

Figure 7. (a) An energy minimized model structure of YGGDGG(A)12GGAG built from estimates of the backbone torsion angles using the program TALOS-N. (b) The C-terminal part of the structure forms intra-molecular hydrogen bonds between Ala16 CO and Ala21 HN, Ala17 CO and Gly20 HN, which are characteristic of the Schellman C-cap motif. Details are described in ref. (31). Copyright 2015 American Chemical Society.

Discussion As shown in Figure 1c, the chemical shifts of the well-resolved sharp peaks from Ala Cα, Cβ, and CO carbons of B. mori liquid silk are quite different from the corresponding main Ala peaks of liquid silk stored in the silk gland of S.c. ricini wild silkworm. The latter peaks clearly come from its poly-Ala region; in addition, there is also a small Ala peak coming from isolated Ala residues which is evident in the Ala CO region. The poly-Ala stretches in S.c. ricini are α-helical, whereas the isolated Ala residues are random coil. The chemical shift of the main Ala peak of this latter liquid silk is a good chemical shift reference for α-helix, and the small peak a good chemical shift reference for random coil conformation (15, 19, 20). Thus, the spectral comparison provides clear evidence against the presence of α-helix in the liquid silk from B. mori silkworm. Actually, the X-ray diffraction data of liquid silk dried without using any external force are also inconsistent with the presence of α-helix (33, 34). However, infrared (IR) spectroscopy appeared to demonstrate the presence of α-helix in B. mori liquid silk using automatic analysis carried out by commercial software such as Opus 6.5 software (Bruker Optics Corp., Billerica, MA) (35). Raman spectral analysis of liquid silk reached a similar conclusion (36). The CD spectrum of 201

liquid silk from the middle silk gland supported the random coil conformation, but a pattern similar to α-helix was observed at more than about 10% level (37, 38). From the concentration dependence of the Moffitt-Yang parameter b0 in the ORD measurement, Kobayashi et al. (39) reported that the helical content in diluted liquid silk is negligible up to a concentration of 0.1%, but at higher concentration, the helical content increases rapidly. This means that with increasing concentration, the generation of helical structure appears at about 10% concentration, and the percentage increases with increasing silk fibroin concentration. Thus, the assignments of the IR and Raman spectra of the silk fibroin should be done very carefully. Theoretical approach using the co-ordinates of (AGSGAG)n with Silk I* form reported in this work should help provide more exact assignments in the spectroscopic analyses of B. mori silk fibroin stored in the silk gland. As mentioned above, it is difficult to determine the content of Silk I* from the solution NMR. Therefore, this information was determined by solid state NMR for Ser and Ala residues of B. mori liquid silk in the solid state after drying the sample (Figure 8) (40). The determination of the fraction of Ser residues contributing to Silk I* conformation is especially important with respect to the discussion on fiber formation because 68% of all Ser residues are present as the sequence (AGSGAG)n in whole silk fibroin (17). In addition, in the Silk I* model (Figures 3 and 4), the Ser OH group contributes to the stabilization of the Silk I* conformation through intra-molecular hydrogen bond formation (9, 30). The selective 13C labeling of Ser Cβ carbons of B. mori silk fibroin makes it possible to analyze the local conformation of Ser residues in detail as shown in Figure 8a (40, 41). The sharp peak at 60.2 ppm is assigned to Silk I* (broken line) and its content can be determined to be 27%. The percentage of other conformation was 59% for distorted β-turn/random coil and 14% for β-sheet. Because there is no β-sheet in B. mori silk fibroin stored in the silk gland as mentioned above (13, 14), we propose that the observed β-sheet comes from structural transition of Silk I* due to small external forces in the preparation stage to take out the silk fibroin gel from the middle silk gland or at the drying stage to prepare the solid sample. Conversion of Silk I* to β-sheet requires very little external force, implying that the 14% β-sheet observed here is very likely to be Silk I* in the silk gland. Thus, the fraction of Silk I* is 41% (27 + 14) in the absence of external forces. Even this larger estimate is considerably lower than the fraction (68%) of Ser residues in the sequence (AGSGAG)n calculated from the primary structure (17). Thus, many Ser residues in the sequence (AGSGAG)n cannot contribute to Silk I*, implying that most probably only longer (AGSGAG)n sequences contribute to Silk I*. The peak deconvolution of solid state NMR spectrum of Ala residue in B. mori liquid silk after sample drying is also shown in Figure 8b (40). In our previous papers (3, 4, 40, 42, 43), the Ala Cβ peak was deconvoluted by assuming the presence of five peaks. Details of the assignment were described in the published papers. The black solid line in Figure 8b is random coil peak from the non-crystalline domain (45% of whole silk fibroin). For the crystalline domain (55% of whole silk fibroin) deconvoluted as three peaks, the broken line is Silk I* peak, and other two peaks (the grey solid lines) are r: random coil and β: β-sheet structure, respectively. The appearance of Silk I* was also discussed (40). 202

Figure 8. The deconvolutions of Ser Cβ (a) and Ala Cβ (b) peaks in the 13C CP/MAS NMR spectra of B. mori silk fibroin before spinning in the solid state. The solid lines are the observed and deconvoluted peaks. The Ser Cb peak can be deconvoluted to three peaks (the broken line, Silk I*; the grey solid lines, r: random coil and b: b-sheet structure, respectively). For the Ala Cb peak, the black solid line is the random coil peak from the non-crystalline domain (45% of whole silk fibroin). For the crystalline domain (55% of whole silk fibroin), the broken line is Silk I* peak, and the other two peaks (the grey solid lines) are r: random coil and b: b-sheet structure, respectively. Details of the assignments and deconvolution are described in ref. (40). Copyright 2015 American Chemical Society. Furthermore, we determined the solution structure of S.c. ricini silk fibroin before spinning in native liquid silk (31). The assignments of the 13C, 15N, and 1H solution NMR spectra for the repetitive sequence motif, YGGDGG(A)12GGAG, were achieved, and the corresponding chemical shifts obtained (31). We used the program TALOS-N to predict the backbone torsion angles from the chemical shifts for this motif. The torsion angles obtained indicate: (i) an α-helical structure for the poly Ala region, (ii) an N-cap residue for Gly6 associated by a type I β-turn 203

from Gly3 to Gly6 in the N-terminal region, and (iii) a Schellman C-cap motif for the C-terminal region. Amide proton temperature coefficients also confirmed the proposed structure. The amide proton coefficients are more positive than that of a random coil for poly-Ala residues, Gly6, Gly19, and Gly20. NHs from these Gly residues form hydrogen bonds characterized N- and C-cap motifs in the proposed structure. In addition, we studied the fiber formation mechanism of S.c. ricini silk fibroin (44). The three-dimensional architecture of the spinneret of S.c. ricini was reconstructed, and the structural change in the silk fibroin that occurs exclusively at the silk press part was elucidated by the molecular dynamics simulation (44, 45). Readers interested in the experimental details may consult the appropriate references.

Conclusion In this work the structures and the conformations of silk fibroins stored in two silkworms were determined by solution NMR. The torsion angles in the polypeptides were estimated from NMR data (7). Torsion angle constraints for the main chain were derived from the backbone chemical shifts (13Cα, 13Cβ, 13CO, 1Hα, 1HN, and 15N) using the program TALOS-N. Moreover, inter-residue NOE cross-peaks, whose intensities are inversely proportional to the sixth power of inter proton distances, were examined to obtain spatial inter-proton distance information. The sequence Y1GGDGG6(A)12G19GAG22 can be used as a representative repeat sequence of S.c. ricini silk fibroin. Resonance assignments of 1H, 15N, and 13C resonances for the residues in the repetitive motif were obtained using the same strategy as that applied to B. mori. The precise S.c. ricini silk fibroin structure of the repeat motif in liquid silk reported in the present work will hopefully contribute to a better understanding of the mechanism of fibroin processing.

Acknowledgments T.A. acknowledges support by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Culture and Supports of Japan (25620169, 26248050) and Impulsing Paradigm Change through Disruptive Technologies Program (ImPACT). We also thank Dr. H. N. Cheng (Southern Regional Research Center, USDA Agricultural Research Service, New Orleans, LA 70124, USA) for discussions.

References 1. 2.

Asakura, T.; Miller, T., Eds. Biotechnology of Silk; Springer: 2014; pp 123−268. O’Brien, J. P.; Fahnestock, S. R.; Termonia, Y.; Gardner, K. C. H. Adv. Mater. 1998, 10, 1185–1195. 204

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.

Asakura, T.; Ohata, T.; Kametani, S.; Okushita, K.; Yazawa, K.; Nishiyama, Y.; Nishimura, K.; Aoki, A.; Suzuki, F.; Kaji, H.; Ulrich, A. S.; Williamson, M. P. Macromolecules 2015, 48, 28–36. Asakura, T.; Okushita, K.; Williamson, M. P. Macromolecules 2015, 48, 2345–2357. Marsh, R; Corey, R. B.; Pauling, L. Biochim. Biophys. Acta 1955, 16, 1–34. Takahashi, Y.; Gehoh, M.; Yuzuriha, K. Int. J. Biol. Macromol. 1999, 24, 127–138. Suzuki, Y.; Yamazaki, T.; Aoki, A.; Shindo, H.; Asakura, T. Biomacromolecules 2014, 15, 104–112. Fu, C.; Shao, Z.; Vollrath, F. Chem. Commun. 2009, 6515–6529. Yamane, T.; Umemura, K.; Asakura, T. Macromolecules 2002, 35, 8831–8838. Asakura, T.; Umemura, K.; Nakazawa, Y.; Hirose, H.; Higham, J.; Knight, D. P. Biomacromolecules 2007, 8, 175–181. Magoshi, J.; Magoshi, Y.; Nakamura, S. Mechanism of fiber formation of silkworm. In Silk Polymers; ACS Symposium Series 544; American Chemical Society: 1994; pp 292−310. Domigan, L. J.; Anderson, M.; Alberti, K. A.; Chesler, M.; Xu, Q.; Johansson, J.; Rising, A; Kaplan, D. L. Insect Biochem. Mol. Biol. 2015, 65, 100–106. Asakura, T.; Suzuki, H.; Watanabe, Y. Macromolecules 1983, 16, 1024–1026. Asakura, T. JEOL News 1987, 23A, 2–6. Asakura, T.; Watanabe, Y.; Uchida, A.; Minagawa, H. Macromolecules 1984, 17, 1075–1081. Asakura, T. Makromol. Chem., Rapid Commun. 1986, 12, 755–759. Zhou, C. Z.; Confalonieri, F.; Jacquet, M.; Perasso, R.; Li, Z. G.; Janin, J. Proteins Struct., Funct. 2001, 44, 119–122. Sezutsu, H.; Yukuhiro, K. J. Insect Biotechnol. Sericol. 2014, 83, 59–70. Asakura, T.; Murakami, T. Macromolecules 1985, 18, 2614–2619. Asakura, T.; Kashiba, H.; Yoshimizu, H. Macromolecules 1988, 21, 644–648. Asakura, T.; Yoshimizu, H.; Yoshizawa, F. Macromolecules 1988, 21, 2038–2041. Nakazawa, Y.; Asakura, T. FEBS Lett. 2002, 529, 188–192. Asakura, T.; Suzuki, Y.; Nakazawa, Y.; Yazawa, K.; Holland, G. P.; Yarger, J. L. Prog. Nucl. Magn. Reson. Spectrosc. 2013, 69, 23–68. Vollrath, F.; Knight, D. P. Nature 2001, 410, 541–548. Shen, Y.; Bax, A. J. Biomol. NMR 2012, 52, 211–232. Shen, Y.; Bax, A. J. Biomol. NMR 2013, 56, 227–241. Asakura, T.; Ashida, J.; Yamane, T.; Kameda, T.; Nakazawa, Y.; Ohgo, K.; Komatsu, K. J. Mol. Biol. 2001, 306, 291–305. Asakura, T.; Ohgo, K.; Komatsu, K.; Kanenari, M.; Okuyama, K. Macromolecules 2005, 38, 7397–7403. Asakura, T.; Suzuki, Y.; Yazawa, K.; Aoki, A.; Nishiyama, Y.; Nishimura, K.; Suzuki, F.; Kaji, H. Macromolecules 2013, 46, 8046–8050. Yamane, T.; Umemura, K.; Nakazawa, Y.; Asakura, T. Macromolecules 2003, 36, 6766–6772. 205

31. Suzuki, Y.; Kawanishi, S.; Yamazaki, T.; Aoki, A.; Saito, H.; Asakura, T. Macromolecules 2015, 48, 6574–6579. 32. Aurora, R.; Rose, G. D. Protein Sci. 1998, 7, 21–38. 33. Lotz, B.; Cesari, F. C. Biochimie 1979, 61, 205–214. 34. Konishi, T.; Kurokawa, M. Sen’i Gakkaishi 1968, 24, 550–554. 35. Hu, X.; Shmelev, K.; Sun, L.; Gil, E. S.; Park, S. H.; Cebe, P.; Kaplan, D. L. Biomacromolecules 2011, 12, 1686–1696. 36. Edwards, H. G. M.; Farwell, D. W. J. Raman Spectrosc. 1995, 26, 901–909. 37. Kataoka, T.; Kobayashi, Y.; Fujiwara, T.; Kyogoku, Y. Abstract in Symposium of study and utilization of non-mulberry silkworms; 1981; p 71. 38. Tsukada, M.; Hirabayashi, H. Sen’i Gakkaishi 1983, 39, 67–70. 39. Kobayashi, Y.; Fujiwara, T.; Kyogoku, Y.; Kataoka, T. Abstract in the 19th NMR meeting in Sapporo; 1980; p 149. 40. Asakura, T.; Sato, Y.; Aoki, A. Macromolecules 2015, 48, 5761–5769. 41. Okushita, K.; Asano, A.; Williamson, M. P.; Asakura, T. Macromolecules 2014, 47, 4308–4316. 42. Asakura, T.; Yao, J.; Yamane, T.; Umemura, K.; Ulrich, A. S. J. Am. Chem. Soc. 2002, 124, 8794–8795. 43. Asakura, T.; Yao, J. Prot. Sci. 2002, 11, 2706–2713. 44. Asakura, T.; Yao, J. M.; Yang, M. Y.; Zhu, Z. H.; Hirose, H. Polymer 2007, 48, 2064–2070. 45. Moriya, M.; Roschzttardtz, F.; Nakahara, Y.; Saito, H.; Masubuchi, Y.; Asakura, T. Biomacromolecules 2009, 10, 929–935.

206

Chapter 12

Novel Polymeric Products Derived from Biodiesel Atanu Biswas,1,* Zengshe Liu,1 Roselayne Furtado,2 Carlucio R. Alves,3 and H. N. Cheng4,* 1National

Center for Agricultural Utilization Research, USDA Agricultural Research Service, 1815 N. University St., Peoria, Illinois 61604, United States 2Embrapa Tropical Agroindustry, Fortaleza-CE 60511-110, Brazil 3Departamento de Química, Universidade Estadual do Ceará, Fortaleza-CE 60740-903, Brazil 4Southern Regional Research Center, USDA Agricultural Research Service, 1100 Robert E. Lee Blvd., New Orleans, Louisiana 70124, United States *E-mail: [email protected]; [email protected]

Biodiesel (produced by reacting a triglyceride with an alcohol) is increasingly being used as diesel fuel and heating oil, especially in Europe. Because of its availability and favorable environmental profile, it may be useful as a renewable feedstock for new polymers. In this work we introduced the epoxide functionality into biodiesel and converted it into a polymer through cationic polymerization with fluorosulfonic acid. Because of the stereochemistry involved, both linear and cyclic products were found. Copolymerization with epoxidized soybean oil produced polymers that ranged from liquids to solids. The use of a comonomer (e.g., diepoxides and propylene oxide) produced more diverse polymer structures. The NMR spectra of these materials provided helpful information on the reaction mechanism. These products may perhaps find applications as additives in lubricants, specialty elastomers, thickeners in coatings, and ingredients in oil-based commercial formulations.

© 2017 American Chemical Society

Introduction Products from agro-based raw materials are attractive because potentially they have the benefits of sustainability, environmental friendliness, and price stability relative to petroleum-based products (1–4). One of the more successful biobased products in recent years is biodiesel, which is typically produced through transesterification of triglyceride oil with methanol in the presence of a catalyst (5, 6). Common triglycerides being used include edible oils from plant and animal sources and waste cooking oils. Because it is renewable and has a favorable environmental profile, it can be used a “green” alternative in pure form or blended with petroleum diesel for diesel engines or in heating oil. One of the biodiesels derived from soybean oil is commonly known as “methyl soyate” (7). Because it is available, relatively low-priced, renewable, and biodegradable, we are examining its applications as a renewable feedstock for specialty polymers. Thus far, there seem to be very few papers on the use of methyl soyate or epoxidized methyl soyate as a renewable monomer for polymerization. Previously a direct polymerization methodology has been developed to convert the olefins to epoxides in plant oils and to polymerize the epoxides via cationic means; a range of polymeric materials have been obtained (8–13). The catalyst for most of the work done so far has been BF3 etherate; however, the polymers derived from such polymerization are crosslinked and insoluble (9, 10). In an earlier work (14, 15), we have shown that fluorosulfonic acid as a catalyst can polymerize epoxidized soybean oil and produce products that are soluble in organic solvents. In this article, we have carried out preliminary work to use methyl soyate as the starting material for the synthesis of polymeric products in order to increase the range of utility of this biobased raw material.

Experimental Section Materials Methyl soyate was obtained from Cooperative Producers, Inc. (Hastings, Nebraska) as SoyGold® 1000. It was made via alkaline methanolysis of soybean oil. Soybean oil came from a local grocery store. The following reagents were acquired from Sigma Aldrich (Milwaukee, WI): ethyl acetate, fluorosulfonic acid, 1,2,7,8-diepoxyoctane, bisphenol A diglycidyl ether, poly(ethylene glycol) diglycidyl ether (average Mn 500), and propylene oxide. Deuterochloroform came from Cambridge Isotope Laboratories, Inc., Andover, MA. Reactions The epoxidation reaction was carried out using formic acid and hydrogen peroxide. The reaction was followed by NMR to ensure complete conversion of the olefins. More details on this reaction are available elsewhere (16, 17). A typical polymerization procedure involved the addition of 1 g epoxidized methyl soyate and 2 mL ethyl acetate in a glass vial with stir bar, screw cap and 208

septum. About 5-20 mg of fluorosulfonic acid was added to the vial. The vial was placed in a React-ThermTM reactor set at 25-35ºC for 24 hours with the reaction under nitrogen and stirring. As a precaution, an extra needle in the cap was inserted to prevent pressure buildup. After 24 hours, water was added to the vial, mixed, and decanted off. Sodium bicarbonate solution (5%) was then added, mixed, and decanted off. The product was washed twice with deionized water and then dried in vacuo at 60ºC to remove ethyl acetate. In this procedure product recovery was almost quantitative except for transfer loss.

NMR Analysis NMR spectra were acquired on a Bruker DRX 400 spectrometer (Karlsruhe, Germany). The NMR solvent used was d-chloroform; tetramethyl-silane served as the 13C chemical shift reference at 0 ppm. Standard operating conditions were used with 30° pulse angle and 3 s between pulses. Spectral assignments were made using empirical chemical shifts rules and automated shift prediction software (18, 19).

High Frequency Reciprocating Rig (HFRR) Test The HFRR test is often used to evaluate the lubricity of diesel fuel. We followed the procedure as given in ASTM D6079-99 (20). Basically a 2-mL test specimen was placed in the test reservoir of an HFRR unit and heated to 60°C. A test steel ball was immersed in the specimen and then vibrated against a test disk at a frequency of 50 Hz for 75 min. The wear scar on the ball was then measured under a microscope. A smaller number is considered more desirable.

Results and Discussion Epoxidized soybean oil is available commercially, but epoxidized methyl soyate is apparently not. In this work we chose to epoxidize soybean oil and methyl soyate ourselves. They were then used for polymerization.

Polymerization of Epoxidized Methyl Soyate We first carried out homopolymerization of epoxidized methyl soyate (EMS) with a cationic initiator. Whereas several methods could be used (9–13), we chose, in light of our previous studies (14, 15), to use fluorosulfonic acid (FSA) in nitrogen with ethyl acetate as solvent. The FSA level varied from 5 mg to 20 mg. The reaction temperature varied from 25° to 35°C. In all cases, liquid products were obtained, indicating the formation of oligomers. A list of selected runs is given in Table 1. 209

Table 1. Polymerizationa of epoxidized methyl soyate and amountsb of minor reaction products sample

Wt of FSA (mg)

product form

ketone %

furan %

dioxolane %

dioxane %

A1

5

liquid

5

3

0

0

A2

10

liquid

4

5

0

0

A3

15

liquid

3

4

1

1

A4

20

liquid

3

5

2

1

A5

5

liquid

4

3

0

0

A6

10

liquid

2

4

2

1

A7

20

liquid

2

4

2

1

a All reactions were conducted with 1 g of EMS in 2 g ethyl acetate solvent at 35°C for 24 hours under nitrogen. b Determined via NMR.

The 13C NMR spectrum of sampleA4 is shown in Figure 1. Since methyl soyate consists primarily of methyl oleate and methyl linoleate, the spectrum is similar to what we found with the cationic polymerization of epoxidized methyl oleate and epoxidized methyl linoleate (14, 15). The large number of peaks at 14-40 ppm correspond to the fatty acid moiety in methyl soyate. The peaks at 51 ppm and 174 ppm come from methoxy and ester carbons, respectively. The cluster of peaks at 70-85 ppm (partly overlapping the CDCl3 lines) belong to the oligomers. In addition, there are many smaller peaks in the spectrum due to several minor reactions. These reactions are summarized in Scheme 1. For monoepoxides (like epoxidized oleate), the reactions with FSA lead to propagation to form a linear polymer and also rearrangements to form ketone, dioxolane, and dioxane (top part of Scheme 1). In Figure 1, the peaks at 81 ppm (ring CH) and 111 ppm (quaternary carbon on the ring) correspond to the dioxolane structure, and the peaks at 212 ppm (C=O) and 42.7 ppm (carbon α to ketone) are characteristic of the ketone. Dioxane is not observed in Figure 1 above the signal-to-noise of the spectrum. Because of the different stereochemistry involved, the reactions of a diepoxide (like epoxidized linoleate) with FSA give different products (lower part of Scheme 1). As the epoxide functionality is protonated by FSA, one reaction pathway polymerizes diepoxides to form a cyclic polymer containing a tetrahydrofuran ring. Another pathway involving elimination and isomerization gives the polymeric furan structure. In Figure 1, the furan ring structure can be clearly seen at 105 ppm (C3 and C4) and 154 ppm (C1 and C4). Note that the 70-85 ppm region gives many peaks, showing a mixture of oligomeric structures (both linear and tetrahydrofuran-type), with different tacticity and chain ends. Because of structural complexity, it is not possible at present to assign the peaks in the 70-85 ppm region. 210

Figure 1. 13C NMR spectrum of FSO4-initiated oligomer from methyl soyate 19560-49-3; f = furan, l = dioxolane, k = ketone, s = CDCl3.

Quantitative data on the minor reaction products are given in columns 4-7 of Table 1. These have been calculated from the integrated areas of appropriate 13C NMR peaks by setting ω1 carbon (methyl, at 14 ppm) or ω2 carbon (methylene, at 22 ppm) on the fatty acid chain as 100%. In the NMR spectra, ketone and furan peaks are more easily noticeable, whereas the concentration levels for dioxolane and dioxane tend to be low and their peaks are sometimes lost in the noise. Note that the total amount of all minor reaction products is less than 10% in all the samples studied.

Copolymers of EMS and Epoxidized Soybean Oil We next made copolymers of EMS with epoxidized soybean oil (ESO) using FSA as the catalyst. Different ratios of EMS and ESO were attempted. Polymerization was achieved in all cases, with the products that ranged from solid to liquid (Table 2). The reason for the solid formation at higher ESO levels is due to the multiple epoxy groups present in ESO, which tend to crosslink the resulting polymer. The 13C NMR spectrum of sample B5 is given in Figure 2. Since EMS and ESO are structurally similar, the copolymerization likely proceeded with the same reaction mechanism as Scheme 1. Indeed both furan and ketone can be found in Figure 2. In addition, methoxy and glycerol moiety are observed because of the presence of methyl ester and glycerate esters in EMS and ESO, respectively. As before, the peaks that correspond to the polymer can be found in the 70-85 ppm region. 211

Table 2. Copolymerization of EMS and ESO in the presence of fluorosulfonic acid in nitrogena sample

feed ratio EMS:ESO

obsd EMSb %

obsd EMSc %

product form

ketone (%)

furan (%)

B1

0:100

0

0

solid

12

8

B2

10:90

9

13

solid

15

7

B3

25:75

22

22

soft solid

13

8

B4

40:60

40

35

liquid

9

7

B5

50:50

50

49

liquid

11

6

B6

75:25

77

72

liquid

3

6

B7

100:0

100

100

liquid

3

3

a

All reactions were conducted with 1 g of total starting material(s), 20 mg FSA catalyst and 2 g ethyl acetate solvent at 35°C for 24 hours. b Calculated from the areas of methoxy and glycerate peaks in the NMR spectra. c Calculated from the ester peaks in the NMR spectra.

It is of interest that three distinct ester peaks are observed (Figure 2, inset). The peak at 174.2 ppm corresponds to the ester carbon of EMS, whereas the two peaks at 173.2 and 172.8 ppm come from ester carbon of ESO. These two ESO peaks are due to the two types of esters formed from glycerol: the ester for the 1,3 positions at 173.2 ppm, and the ester for the 2 position at 172.8 ppm. From the integrated intensities of these three ester peaks we can estimate the % EMS in the copolymer. Another measure of the copolymer composition is to take the ratio of the methoxy carbon peak (from EMS at 51 ppm) and glycerol peaks (from ESO at 68 and 63 ppm). The copolymer compositions derived from both calculations are shown in Table 2. These agree well with each other and with the feed ratios of EMS and ESO. The quantitative NMR data for ketone and furan are also given in Table 2. As the amount of ESO in the copolymerization increases, there seems to be a slight increase in the levels of ketone and furan. This may be due to the increase in the viscosity of the reaction medium as the proportion of ESO increases, thereby slightly favoring the minor reactions. Note that the total amount of ketone and furan is higher ( >10%) for the EMS/ESO samples relative to the EMS homopolymer.

212

Scheme 1. Reactions between epoxides and fluorosulfonic acid; x,y = 1 or 2.

213

Figure 2. 13C NMR spectrum of the copolymer of EMS and ESO-19560-37-2; f = furan, k = ketone, g = glycerol moiety, s = CDCl3. Inset: Expanded 172 -195 ppm region, showing the ester peaks.

Other Copolymers of EMS We also sought to make other copolymers of EMS. The first group of comonomers consisted of diolefins that could serve as crosslinkers, such as 1,2,7,8-diepoxyoctane (DEO, Mn 142), bisphenol A diglycidyl ether (BADGE, Mn 340), and poly(ethylene glycol) diglycidyl ether (PEGDE, Mn 500). The results for six copolymers are shown in Table 3. At 10 wt % diolefin level, all three EMS copolymers were liquids. At 30 wt % diolefin level, the copolymers involving DEO and BADGE were solids at room temperature, but the copolymer involving PEGDE was still a liquid. The 13C NMR spectra for the 70:30 EMS : diolefin copolymers are given in Figure 3. The NMR signals for the copolymer appear as a broad complex feature just above the baseline at 70-85 ppm. The presence of EMS can be clearly seen from the ester peak (174 ppm), methoxy peak (51 ppm), and the aliphatic peaks (14-40 ppm). For the diolefins, the epoxy functionality has almost fully reacted and the epoxy peaks (ca. 46 and 51 ppm) are not observed. The peaks for the rest of the carbons in the diolefins have been labelled in Figure 3. Thus, the 13C NMR spectra confirm the successful synthesis of these three copolymers.

214

Table 3. Copolymerization of EMS and diolefin in the presence of fluorosulfonic acid in nitrogena sample

diolefin

wt ratio EMS:diolefin

mole ratio EMS:diolefin

product form

C1

DEO

90 : 10

81:19

liquid

C2

DEO

70 : 30

53:47

solid

C3

BADGE

90 : 10

91:9

liquid

C4

BADGE

70 : 30

73:27

solid

C5

PEGDE

90 : 10

94:6

liquid

C6

PEGDE

70 : 30

80:20

liquid

a All reactions were conducted with 1 g of total starting material(s),

20 mg FSA catalyst and

2 g ethyl acetate solvent at 35°C for 24 hours.

Figure 3. 13C NMR spectrum of the (70:30) copolymers of EMS and three diepoxides: A) diepoxyethane, B) bisphenol A diglycidyl ether, C) poly(ethylene glycol) diglycidyl ether. f = furan, k = ketone, s = CDCl3, E = ethanol. The epoxy moiety has been fully reacted because no epoxy peaks (ca. 46 and 51 ppm) can be found.

215

Next, we attempted the copolymerization of EMS with propylene oxide (PO). Under our reaction conditions with FSA, PO by itself gave a liquid product. However, when PO and EMS were copolymerized with diepoxyethane, the reaction products were solids at room temperature (Table 4). The 13C NMR spectrum of sample D4 is given in Figure 4.

Figure 4. 13C NMR spectrum of the copolymers of EMS, propylene oxide, and diepoxyoctane (20:50:30) (sample 19640-9-2); p = PO peaks, d = diepoxyoctane peaks, s = CDCl3. The epoxy peaks (at 47 and 51 ppm) are absent, indicating complete reaction.

Table 4. Copolymerizationa of EMS, propylene oxide, and diepoxyoctane sample

wt ratio EMS:PO:DEO

mole ratio EMS:PO:DEO

product form

D1

70:0:30

53:0:47

solid

D2

50:20:30

24:47:29

solid

D3

43:30:30

16:60:24

solid

D4

20:50:30

6:76:18

gel

D5

0:100:0

0:100:0

liquid

All reactions were conducted with 1 g of total starting material(s), 20 mg FSA catalyst and 2 g ethyl acetate solvent at 35°C for 24 hours under nitrogen. a

In Figure 4, the presence of EMS is evidenced by the ester peak at 174 ppm, methoxy peak at 51 ppm, and the aliphatic peaks at 14-40 ppm. The unreacted epoxy peaks for diepoxyoctane (at 47 and 52 ppm) are not observed, indicating their complete conversion to polymers. The peaks for the other two carbons in reacted diepoxyoctane are found as broad lines at 25.6 and 32.4 ppm. In agreement with the literature (21), the poly(propylene oxide) (PPO) peaks occur at about 75.5 ppm (CH), 73.5 ppm (CH2), and 17.8 ppm (CH3). The PPO terminal methylene 216

peaks occur at 72.0 and 75.6 ppm, methine at 64-68 ppm, and methyl at 17.3 and 19.3 ppm . Thus, the 13C NMR data confirm the incorporation of EMS, PO, and DEO in the polymers made. Possible Applications In recent years, there has been a lot of interest in converting plant oil to polymers, and many different approaches have been adopted (22, 23). Examples of direct polymerization of soybean oil include thermal (24–26) and air blown polymerization (27, 28), cationic polymerization (29–31), pericyclic reactions (32, 33), and others (34). Furthermore, the plant oil can be modified in order to add more functional groups and increase the range of possible polymerizations. A number of review articles have appeared on this topic (35–37). Possible applications of these triglyceride-based polymers include lubricants, paints, coatings, adhesives, plastics, composites, and biomedical materials. In this work, epoxidized methyl soyate has been made into homopolymers and copolymers using cationic polymerization. This family of new polymers can probably be used for many of the same aforementioned applications. For example, one obvious application is to use the new materials in lubricant formulations. A preliminary study of epoxidized methyl oleate, epoxidized methyl linoleate, and EMS was carried out using the high frequency reciprocating rig (HFRR) test (Table 5). According to the results, all six samples gave low values of wear scars. In comparison, kerosene gives a HFRR value of 675 μm, and commercial D1 and D2 fuels 578 and 376 μm, respectively (38). Thus, the polymers made in this work may perhaps be regarded as possible additives for lubricants.

Table 5. High frequency reciprocating rig (HFRR) test for polymers made from several fatty acid methyl estersa sample

starting material

wt. FSA (mg/g)

wear scars (μm)

E1

epoxidized methyl oleate

7.5

224

E2

epoxidized methyl oleate

15

190

E3

epoxidized methyl linoleate

10

127

E4

epoxidized methyl linoleate

15

95

E5

EMS

5

194

E6

EMS

15

129

a Polymerization was achieved with 1 g of starting material and 2 g ethyl acetate solvent at 35°C for 24 hours under nitrogen.

Other possible applications of these polymers may include their use as bioplasticizers, specialty elastomers, and thickeners in oil-based commercial formulations. Further work is needed in order to develop the full product development potential of these materials. 217

Conclusion In this work we have used biodiesel as a natural renewable raw material for the synthesis of polymers. In particular, a cationic catalyst, fluorosulfonic acid, has been shown to catalyze the polymerization of epoxidized methyl soyate. Furthermore, copolymers can be made with epoxidized methyl soyate, thereby generating a new family of polymers. The polymers may perhaps find application in a number of areas, such as lubricants, coatings, elastomers, and other oil-based commercial formulations. It may be noted that 13C NMR was found to be useful in elucidating the reaction mechanism. Because of the stereochemistry involved, monoepoxide and diepoxide engendered different reactions and produced different products when polymerized. Thus, if biodiesels from different sources (e.g., soybean oil, palm oil, cottonseed oil, animal fat, and waste cooking oil) are used for this reaction, somewhat different products may be obtained. This product diversity may be an opportunity that permits a range of different polymers to be made with different end-use performance.

Acknowledgments Thanks are due to Janet Berfield and Gary Kuziar for expert technical assistance, Dr. Karl Vermillion for the NMR spectra, and Daniel Knetzer for helpful support. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer.

References 1.

2. 3.

4.

5.

6.

Green Polymer Chemistry: Biobased Materials and Biocatalysis; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1192; American Chemical Society: Washington, DC, 2015. Mulhaupt, R. Green polymer chemistry and bio-based plastics: Dreams and reality. Macromol. Chem. Phys. 2013, 214, 159–174. Green Polymerization Methods: Renewable Starting Materials, Catalysis and Waste Reduction; Mathers, R. T., Meier, M. A. R., Eds.; Wiley-VCH: Weinheim, Germany, 2011. Gandini, A. The irruption of polymers from renewable resources on the scene of macromolecular science and technology. Green Chem. 2011, 13, 1061–1083. Atabani, A. E.; Silitonga, A. S.; Badruddin, I. A.; Mahlia, T. M. I.; Masjuki, H. H.; Mekhilef, S. A comprehensive review on biodiesel as an alternative energy resource and its characteristics. Renewable Sustainable Energy Rev. 2012, 16, 2070–2093. Leung, D. Y. C.; Wu, X.; Lueng, M. K. H. A review on biodiesel production using catalyzed transesterification. Appl. Energy 2010, 87, 1083–1095. 218

7. 8.

9. 10.

11. 12.

13.

14. 15.

16.

17.

18. 19. 20. 21.

22. 23. 24. 25.

United Soybean Board. http://soynewuses.org/solvents/ (accessed May 14, 2017). Raghavachar, R.; Sarnecki, G.; Baghdachi, J.; Massingill, J. Cationic, thermally cured coatings using epoxidized soybean oil. J. Coat. Technol. 2000, 72, 126–133. Liu, Z. S.; Erhan, S. Z. Ring-opening polymerization of epoxidized soybean oil. J. Am. Oil Chem. Soc. 2010, 87, 437–444. Liu, Z. S.; Doll, K. M.; Holser, R. A. Boron trifluoride catalyzed ring-opening polymerization of epoxidized soybean oil in liquid carbon dioxide. Green Chem. 2009, 11, 1774–1780. Liu, Z. S. Preparation of biopolymers from plant oils in green media. Bioenergy Res. 2013, 6, 1230–1236. Liu, Z.; Knetzer, D. A. Catalyzed ring-opening polymerization of epoxidized soybean oil by hydrated and anhydrous fluoroantimonic acids. Green Mater. 2013, 1, 87–95. Liu, Z. S.; Biswas, A. Fluoroantimonic acid hexahydrate (HSbF6 · 6H2O) catalysis: The ring-opening polymerization of epoxidized soybean oil. Appl. Catal., A 2013, 453, 370–375. Biswas, A.; Liu, Z.; Cheng, H. N. Polymerization of epoxidized triglyceride with fluorosulfonic acid. Int. J. Polym. Anal. Charact. 2016, 21, 85–93. Biswas, A.; Liu, Z.; Berfield, J. L.; Cheng, H. N. Synthesis of novel plant oil derivatives: Furan and Diels-Alder reaction products. Int. J. Agric. Sci. Technol. 2015, 3, 28–35. Sharma, B. K.; Doll, K. M.; Erhan, S. Z. Oxidation, friction reducing, and low temperature properties of epoxy fatty acid methyl esters. Green Chem. 2007, 9, 469–474. Liu, Z.; Chen, J.; Knothe, G.; Nie, X.; Jiang, J. Synthesis of epoxidized cardanol and its antioxidative properties for vegetable oils and biodiesel. ACS Sustainable Chem. Eng. 2016, 4, 901–906. Cheng, H. N. 13C NMR spectral simulation and shift prediction. TrAC Trends Anal. Chem. 1994, 13, 95–104; DOI: 10.1016/0165-9936(94)87073-X. NMRShiftDB. http://nmrshiftdb.nmr.uni-koeln.de/ (accessed May 14, 2017). ASTM D 6079-99. Standard test method for evaluating lubricity of diesel fuels by the High Frequency Reciprocating Rig (HFRR), December 1999. Schilling, F. C.; Tonelli, A. E. Carbon-13 NMR Determination of Poly(propy1ene oxide) Microstructure. Macromolecules 1986, 19, 1337–1343. Xia, Y.; Larock, R. C. Vegetable oil-based polymeric materials: synthesis, properties and applications. Green Chem. 2010, 12, 1893–1909. Espinosa, L. M.; Meier, M. A. R. Plant oils: The perfect renewable resource for polymer science. Eur. Polym. J. 2011, 47, 837–852. Bernstein, I. M. Heat polymerization of nonconjugated vegetable oils. J. Polym. Sci. 1946, 1, 495–528. Wang, C.; Erhan, S. Z. Studies of thermal polymerization of vegetable oils with a differential scanning calorimeter. J. Am. Oil Chem. Soc. 1999, 76, 1211–1216. 219

26. Arca, M.; Sharma, B. K.; Price, N. P. J.; Perez, J. M.; Doll, K. M. Evidence contrary to the accepted Diels-Alder mechanism in the thermal modification of vegetable oil. J. Am. Oil Chem. Soc. 2012, 89, 987–994. 27. Christianson, R.; Noble, M.; Rourke, A.; Geiger, E.; Mahlum, L. U.S. Patent 7262311, August 28, 2007. 28. Geiger, E. J.; Becker, N. M.; Armbruster, L. A. U.S. Patent 7989647, August 2, 2011. 29. Liu, Z.; Sharma, B. K.; Erhan, S. Z. From oligomers to molecular giants of soybean oil in supercritical carbon dioxide medium. Biomacromolecules 2007, 8, 233–239. 30. Li, F.; Hanson, M. V.; Larock, R. C. Soybean oil-divinylbenzene thermosetting polymers: Synthesis, structure, properties and their relationships. Polymer 2001, 42, 1567–1579. 31. Ionescu, M..; Petrovic, Z. S. U.S. Patent 7,501,479, March 10, 2009. 32. Biswas, A.; Sharma, B. K.; Willett, J. L.; Erhan, S. Z.; Cheng, H. N. Room-Temperature Self-Curing Ene Reactions Involving Soybean Oil. Green Chem. 2008, 10, 290–295. 33. Biswas, A.; Cheng, H. N.; Kim, S.; Liu, Z. Modified Triglyceride Oil through Reactions with Phenyltriazolinedione. J. Am. Oil Chem. Soc. 2014, 91, 125–131. 34. Ronda, J. C.; Lligadas, G.; Galia, M.; Cadiz, V. Vegetable oils and platform chemicals for polymer synthesis. Eur. J. Lipid Technol. 2011, 113, 46–58. 35. Biswas, A.; Sharma, B. K.; Willett, J. L.; Erhan, S. Z.; Cheng, H. N. Energy Environ. Sci. 2008, 1, 639–644. 36. Sharma, V.; Kundu, P. P. Addition polymers from natural oils. Prog. Polym. Sci. 2006, 31, 983–1008. 37. Guner, F. S.; Yagci, Y.; Erciyes, A. T. Polymers from triglyceride oils. Prog. Polym. Sci. 2006, 31, 633–670. 38. Van Gerpen, J. H.; Soylu, S.; Chang, D. Y. Z. Evaluation of the lubricity of soybean oil-based additives in diesel fuel. http://biodiesel.org/reports/ 19980201_gen-070.pdf (accessed May 14, 2017).

220

Editors’ Biographies H. N. Cheng H. N. Cheng (Ph.D., University of Illinois) is currently a research chemist at Southern Regional Research Center of the U.S. Department of Agriculture in New Orleans, where he works on projects involving improved utilization of commodity agricultural materials, green chemistry, and polymer reactions. Prior to 2009 he worked for Hercules Incorporated where he was involved at various times with new product development, team and project leadership, new business evaluation, pioneering research, and supervision of analytical research. Over the years, his research interests have included green polymer chemistry, biocatalysis and enzymatic reactions, pulp and paper chemistry, functional foods, polymer characterization, and NMR spectroscopy. He is an ACS Fellow and a POLY Fellow and has authored or co-authored 230 papers, 25 patent publications, co-edited 16 books, and organized or co-organized 30 symposia at national ACS meetings since 2000. He is active in the ACS and serves in various capacities at national, division, and local levels.

Cynthia A. Maryanoff Cynthia A. Maryanoff (Ph.D., Princeton) is Foundation Distinguished Professor at the Baruch S. Blumberg Institute, in Doylestown, PA. She began her career in 1977 at Smith Kline & French Laboratories and joined Johnson & Johnson in 1981. She advanced through various Johnson & Johnson pharmaceutical units to the highest scientific position in the company. She retired from J&J in 2013. Her publication record includes more than 100 scientific papers, several books, and 67 U.S. or European patents. The more than 1,000 drug candidates she has been involved in developing include the anti-epileptic Topamax, and an atypical analgesic, Ultram/Tramadol, for treating pain, and the drug-eluting stent CYPHER. She has been active in ACS and served in many positions including national committees and division leadership. She organized and chaired 30 award symposia and organized or co-chaired 15 other symposia at national meetings. She is the recipient of numerous awards, including ACS Fellow, AAAS Fellow, the ACS Garvan–Olin Medal, the Earle B. Barnes Award for Leadership in Chemical Research Management, the Henry F. Whalen, Jr. Award for Business Development, and the Perkin Medal for outstanding work in applied chemistry from SCI.

© 2017 American Chemical Society

Bradley D. Miller Bradley D. Miller (Ph.D., University of Arizona) is the Director of ACS International Activities. He has worked for ACS since 1999, developing programs, products, and services to advance chemical sciences through collaborations in Africa, Asia, Europe, Latin America and the Middle East. He works with ACS staff and different governance units to create opportunities for chemistry to address global challenges through in-person and web-based scientific network development, research collaborations, and educational exchange. Miller serves on the U.S. National Commission for UNESCO and in 2009 was appointed to co-chair the ACS 2011 International Year of Chemistry Staff Working Group. He is also the long-time ACS staff liaison to the ACS International Activities Committee. A world traveler and an internationalist, he speaks English, French, Spanish and Portuguese.

Diane Grob Schmidt Diane Grob Schmidt (Ph.D., University of Cincinnati), the 2015 ACS President, was an R&D Executive at Procter & Gamble, where she served as section head for 17 years. Her P&G career covered 1981-2014. She is currently an Adjunct Professor in the Department of Chemistry at the University of Cincinnati. She holds a number of patents and played key roles in such brands as Tide®, Head & Shoulders®, Pert Plus® and Safeguard®. She has received many awards, including ACS Fellow, AAAS Fellow, National Academy of Inventors Fellow, ACS national Henry Hill Award, and Distinguished Scientist of Cincinnati from the Engineers and Scientists of Cincinnati (first woman so honored). She has served on the editorial boards of Chemical & Engineering News, the Journal of the Society of Cosmetic Chemists and the Journal of Chemical Health & Safety. She has been an ACS member for many years and held a wide variety of ACS positions, including three consecutive terms on the Board of Directors. As 2015 ACS President, her presidential theme was “Inspiring and Innovating for Tomorrow.” Her legacy as ACS President includes: championing U.S. and Global Grand Challenges [Nanotechnology, Energy, Neuroscience/BRAIN Initiative] via impactful programming, establishment of the American Association of Chemistry Teachers, and a focus on industry and ACS members.

222

Indexes

Author Index Alves, C., 207 Asakura, T., 191 Bailey, W., 19 Biswas, A., 207 Cheng, H., xiii, 207 Cortés, D., 79 de Souza, A., 91 Denmark, S., 105 Dybowski, C., 135 Furtado, R., 207 González, I., 79 Juaristi, E., 3 Lambert, K., 19 Liu, Z., 207 Loeb, B., 79 Magriotis, P., 61

Maryanoff, C., xiii Matsumoto, A., 27 Miller, B., xiii Morgon, N., 91 Nishimura, A., 191 Notario, R., 3 Paixão, M., 49 Rivera, D., 49 Sanhueza, L., 79 Schmidt, D., xiii Soai, K., 27 Suzuki, Y., 191 Thomas, A., 105 Tonelli, A., 161 Ximenes, V., 91

225

Subject Index B

intermediates, observation and kinetic measurements, 110f Kel-F sleeve, 111f original McGarrity RI-NMR apparatus, 109f rapid injection NMR, invention, 108 Reich RI-NMR apparatus, 112f final remarks, 131 introduction, 105 Cram’s rules, 107f neomenthol and neoisomenthol, conformational analysis, 106f total synthesis, development of 1,3-oxathianes, 107f organic chemistry, applications of RI-NMR spectroscopy carbon acids, deprotonation, 113 Cram’s chelate rule, 121f Cram chelates, analysis, 123t cyclopentadiene, 1H-RI-NMR spectra, 115f labeled 49, selected 13C NMR traces, 119f Lewis base catalyzed aldol reactions, 124 7Li RI-NMR spectra, 116f 3-methylstyrene, consumption, 117f mixed organocuprates, mechanism of conjugate addition reactions, 120 mixed π-complex, conversion, 121f n-butyllithium, aggregates, 113f n-butyllithium aggregates, reactivity, 112 nonchelated vs. chelated pathways, reaction coordinate, 124f organocuprate preparation and reactions, 118f organocuprate reactions, mechanism, 117 organocuprates, mechanism of conjugate addition reactions, 118 organolithium reagents, polymerization of styrene, 114 organotin to organolithium, transmetalation, 127f Pd-O-B linkage, formation, 129f Pd-O-B linkages, 129f product formation, kinetic analysis, 130f proposed Suzuki-Miyaura reaction mechanism, 128f

Bombyx mori and Samia cynthia ricini silkworms conclusion, 204 discussion, 201 Ser Cβ and Ala Cβ, deconvolutions, 203f solution NMR, 202 experimental, 194 introduction, 191 middle silk gland portions, 13C solution NMR spectra, 193f results, 195 HN/Ha and HN/HN regions, sequential assignments, 199f poly(AG) chains, packing structure, 198f TOCSY spectrum, assignments, 196f torsion angles, 25 best matches, 197f torsion angles for each residue, plot, 200f YGGDGG(A)12GGAG, energy minimized model structure, 201f

E Electronic circular dichroism spectra, solvent effects chirality and atropisomerism, 92 computational methodology, 97 BiNaphthol, simulated ECD spectra, 99f BiNaphthol at B3LYP/6-31G(d,p) + PCM, potential energy surface, 98f calculated ECD spectra, 98f torsional angle σ1, definition, 97f concluding remarks, 100 ECD spectra, solvent effects, 95 binaphthols and in different solvents, ECD spectra, 96f chemical structures, 95f ECD intensity of binaphthols, dielectric constant effects, 96f introduction, 91 solvated systems, basic theory, 94 solvent, computational models, 93 Eliel, Ernest L. background Denmark RI-NMR apparatus, 110

227

(S)- and (R)-4 in consecutive asymmetric autocatalysis, automultiplication, 32f conclusions, 43 enantiomeric excess, amplification, 30 amplification of ee, asymmetric autocatalysis, 31s introduction, 27 asymmetric autocatalysis, general scheme, 29s enantiomers of α-amino acids and sugars, structures, 28f statistical fluctuation and amplification, 41 absolute configuration and ee of pyrimidyl a, histograms, 42f pyrimidyl alkanol, absolute asymmetric synthesis, 42f

reaction and proposed transition structures, 125f reaction coordinate, nonchelated vs. chelated pathways, 122f SiCl4, reaction, 127f styrene polymerization initiated by butyl lithium, mechanism, 116f tin to lithium, transmetalation, 126 tricoordinate boron intermediate, kinetic analysis, 130f

H Homochirality, asymmetric autocatalysis and the origin asymmetric autocatalysis, 29 practically perfect asymmetric autocatalysis, 30s pyrimidyl alkanol, asymmetric autocatalysis, 30s asymmetric autocatalysis, reaction models and crystal structures, 43 chirality examined, origin, 31 achiral cytosine, chiral crystal, 37s achiral organic compounds, asymmetric autocatalysis, 36 achiral organic compounds, enantiomorphs, 38f carbon isotope (12C/13C) substitution generates chirality, 39f chiral carbon isotopomer, asymmetric autocatalysis, 40s chiral hydrogen isotopomers, 38s chiral inorganic material, 34 chiral oxygen and nitrogen isotopomers, asymmetric autocatalysis, 41s 13C-labeled dimethylphenylmethanol, asymmetric synthesis, 40s CPL irradiation, near enantiopure compound, 34s dehydration of crystallization water, chirality generation, 37f enantiotopic surface of achiral Gypsum, asymmetric autocatalysis, 36s helical silica gel, asymmetric autocatalysis, 35s L-leucine with low ee, asymmetric autocatalysis, 33s origin of chirality, schematic correlation, 32s quartz, asymmetric autocatalysis, 35s

M Molecular structures and conformation, from NMR spectra, 161 introduction, 162 NMR resonance frequencies, 163 γ-gauche shielding, derivation, 165f 25 MHz 13C NMR spectra, 164f polymer microstructures E-VAc copolymers, 166 E-VAc copolymers, comparison, 166f methylene carbons in atactic PVAc, comparison, 168f methylene carbons in E-VAc copolymers, comparison, 167f observed and calculated 13C chemical shifts, comparison, 169f poly(4BCMU) in CDCl3, chemical shifts, 171t poly(4BCMU) in toluene-d3, chemical shifts, 172t solid-state synthesis of polydiacetylenes, schematic representation, 170f solid state polymer conformations CPMAS/DD 13C n.m.r, spectrum, 174f crystallineconformation, projection, 175f crystalline c-(PS)5, 13C-MAS-NMR spectrum, 179f dichloromethane, absorption, 180f DPS recorded, 13C-MAS-NMR spectrum, 178f

228

liquid DPS recorded at room temperature, 13C-MAS-NMR spectrum, 177f molecular modeling, 184 P and Q ring carbons, dihedral angles, 176t PBT and its model compounds, methylene carbons in crystals, 182f poly(phenylene sulphide), 172 poly(tri-methylene terephthalate), 181 PPS, schematic drawing of the crystalline conformation, 173f PTT and TMGDB, phenyl rings, 186t PTT crystalline conformation, 185f PTT films, 13C-CP/MAS-NMR spectra, 183f PTT repeat unit, 182f room temperature, 13C-CPMAS/DDNMR spectrum, 176f syndiotacic-polystyrene and its solid model compounds, 179

N NMR spectroscopy, characterization of materials computational chemistry and NMR spectroscopy, 141 15 molecules, cluster, 143f conclusions, 153 introduction, 135 materials, examples of NMR characterization, 143 accurate predictions, possibility, 152 art masterworks, study of chemical changes, 148 catalytic surfaces, identification of species, 145 chemical-shift tensor, 144 13C ssNMR spectra, 150f free palmitic acid, product build-up by reaction, 151f local lead coordination environment, 149f NMR band, simulated spectra, 145f NMR-derived orientation distribution, simulations, 146f polymeric materials, orientation, 143 study porous materials, use of NMR as a probe, 147 T50 versus relative humidity, 151f solid-state NMR spectroscopy, 136 α-glycine, 13C CP-MAS-DD NMR spectrum, 139f

13C

NMR spectrum, schematic depiction, 138f powdered sample of lead acetate, 207Pb NMR spectra, 140f Novel polymeric products, 207 conclusion, 218 experimental section materials, 208 reactions, 208 introduction, 208 results and discussion, 209 EMS, 13C NMR spectrum of the copolymers, 216f EMS, (70:30) copolymers, 215f EMS, propylene oxide, and diepoxyoctane, copolymerization, 216t EMS and diolefin, copolymerization, 215t EMS and ESO, copolymerization, 212t EMS and ESO-19560-37-2, 13C NMR spectrum of the copolymer, 214f epoxides and fluorosulfonic acid, reactions, 213s epoxidized methyl soyate, polymerization, 210t FSO4-initiated oligomer, 13C NMR spectrum, 211f high frequency reciprocating rig (HFRR) test, 217t N(S) → σ*(C-P), σ(C-S) → σ*(C-P), and n(F) → σ*(C-X) (X = H, C, O, S) stereoelectronic interactions, 3 computational methods, 6 conclusions, 17 introduction, 4 1-ax, x-ray diffraction structure, 5f conformational equilibrium of 2-diphenylphosphinoyl-1,3dithiane, predominance of the axial conformation, 5s electronegative substituents, counter-intuitive preference, 4s results and discussion axial, 4-ax, and equatorial, 4-eq, conformations, 16t axial and equatorial 2-diphenylphosphinoyl-1,3dithiane, selected hyperconjugative interactions (Edel), 8t B3LYP/6-311+G(d,p)-optimized geometrical parameters, 7t B3LYP/6-311+G(d,p)-optimized structures, 6f

229

phendione precursors, molecular hardness and electrophilicity, 82t 5-Phenyl-1,3-dioxane, case conclusions, 25 introduction, 19 axial 5-phenyl-1,3-dioxane, calculated minimum energy rotamer, 21f 5-phenyl-1,3-dioxane, conformational energy, 20s results and discussion, 21 5-aryl-1,3-dioxanes, equilibria, 23t cis-2-t-butyl-5-(p-chlorophenyl)-1,3dioxane, crystal structure, 22f electron-withdrawing p-substituents, effect, 25f Hammett plot, 24f

fluorinated compounds, conformational equilibria, 11s hyperconjugative interactions, 13t MP2/6-311+G(d,p)-optimized structures, 11f r-1,c-3,c-5-trifluorocyclohexane, MP2/6-311+G(d,p)-optimized geometrical parameters, 12t r-2,c-4,c-6-trifluoro-1,3,5-trioxane, MP2/6-311+G(d,p)-optimized structural parameters, 14t r-2,c-4,c-6-trifluoro-1,3,5-trioxane, MP2/6-311+G(d,p)-optimized structures, 13f r-2,c-4,c-6-trifluoro-1,3,5-trithiane, 15f S-C-P segment, anomeric effect, 6 selected hyperconjugative interactions, 15t selected hyperconjugative interactions, r-2,c-4,c-6-trifluoro1,3,5-trithiane, 16t stereoelectronic interaction, n(S) → σ*(C-P)app, 10s stereoelectronic interactions, 9

S

P Phenanthroline-5,6-diones and ethylenediamine, condensation reaction, 79 conclusion, 88 intermediates, analysis corresponding intermediates, 1H-NMR characterization, 85t MeCN, experimental and calculated absorption spectra, 84f molecular hardness and electrophilicity, 87t NMR characterization, 85 NMR DEPT spectra, 86f non-aromatic intermediates, possible structures, 83s theoretical reactivity indexes, 87 used for intermediates, NMR characterization for ppl and ppz compounds, 86s UV-Vis spectra, 83 introduction, 80 ppl and ppz compounds, condensation reactions, 81s products, complete conversion, 88 starting materials, analysis, 81 condensed Fukui functions, 82f

Stereochemically complex antitumor drug ecteinascidin-743, 61 1, total syntheses, 63 Corey’s total synthesis, piperazine construction, 64s Danishefsky’s total synthesis, piperazine surrogate construction, 67s Fukuyama’s total synthesis, piperazine construction, 65s Kubo’s synthetic approach, 69s second generation Fukuyama’s total synthesis, 68s Williams’ formal total synthesis, piperazine surrogate construction, 67s Zhu’s total synthesis, piperazine construction, 66s conclusion and outlook, 76 Ecteinascidin-743, Magriotis’ retrosynthetic analysis, 77s introduction, 62 total syntheses of 1, completion, 70 Corey’s total synthesis, completion, 70s Danishefsky’s formal total synthesis, completion, 73s formation of ketone 64, mechanistic hypotheses, 74s Fukuyama’s second generation total synthesis, completion, 76s Fukuyama’s total synthesis, completion, 71s second generation synthesis, 75 Zhu’s total synthesis, completion, 72s

230

Stereoselective synthesis, interplay between organocatalysis and multicomponent reactions, 49 conclusions, 56 introduction, 50 results and discussion aminocatalytic epoxidation, 51 cyclic depsipeptides, multicomponent synthesis, 56s 1,3-cycloalkanediones, aminocatalytic conjugate addition, 52

231

epoxy-peptidomimetics, one-pot synthesis, 52s hydroquinolin-6-one, one-pot synthesis, 54s one-pot organocatalytic conjugate addition, 53s piperidinocoumarine hybrids, 54s Ugi multicomponent reaction, 55