Download PDF
Synlett 2018; 29(16): 2122-2125
DOI: 10.1055/s-0037-1610908
cluster
© Georg Thieme Verlag Stuttgart · New York

Prologue: Atropisomerism

Jay S. Siegel*
Further Information

Publication History

Received: 26 August 2018

Accepted: 28 August 2018

Publication Date:
21 September 2018 (online)

 
 PDF Download Permissions and Reprints All articles of this category


Jay S. Siegel received his Ph.D. from Princeton (1985), was a Swiss Universities Fellow at ETH Zurich (1983-4), and NSF–CNRS postdoctoral fellow at the University of Louis Pasteur in Strasbourg (1985-6). He began as Assistant Professor of Chemistry (1986) at UCSD, was promoted to Associate Professor (1992) and Full Professor (1996). In 2003, he was appointed as Professor and co-director of the Organic chemistry institute of the University of Zurich (UZH) and Director of its laboratory for process chemistry research (LPF). He served as Dean of Studies and Head of the Research Council for the Faculty of Sciences at UZH. He moved to Tianjin University in 2013 as dean, and joined the Schools of Pharmaceutical and Life Sciences into a new Health Science Platform. His research is in the area of Stereochemistry and Physical Organic Chemistry.


#

This special issue of Synlett is focused on the theme of “Atropisomerism” in a context relevant to modern molecular design and synthesis. For additional details see the preface to this issue. The historical roots of atropiso­merism reach back to beginnings of structural organic chemistry. Its use over the years embodies many of the issues that plagued the development of modern stereochemistry and vestiges of these issues appear even today.

Given that “atropos-” is a Greek combining form, its Greek roots should provide some reference point for its appropriate use. Atropos was one of the oldest of the three fates, who was generally characterized by her inflexibility and resistance to change.[1] She is also the fate who cuts the thread of life and therefore stops the thread from being spun, rendering the dynamic inanimate. Accordingly, atropisomerism is well suited to a concept comprising isomers that arise generally from restricted motion, not only rotations, but also inversions and other dynamic processes as well. Thus, “not turning into” or “not changing” would be a better use of atropos than simply “not turning”. Freeing atropisomerism from the classical isomerism due to restricted rotation about a covalent single bond broadens its applicability and removes the constraints of a valence bond description of the molecule. It is with this modern perspective that the reader should critique for her/himself the contributions to this issue, many of which retain a classical, if not anachronistic, use of atropisomerism.[2]

The general reader may ponder why organic chemists prefer to talk of the nonequivalence of structures derived from a lack of motion rather than starting from static isomeric components and discussing the rendering of them equivalent due to the onset of specific motions. The historical role of dynamic behavior in the development of organic structural theory may provide some insight. In the early days of structural organic chemistry two pioneers, Kekulé and van’t Hoff proffered structural theories for benzene,[3] and tetrahedral carbon,[4] respectively. Both structural ­hypotheses met with difficulties due to their inability to ­account for isomers in static substituted derivatives.[5] The static Kekulé benzene had issues because it predicted two ortho-dichlorobenzenes, when only one was found.[6] Even before van’t Hoff’s fundamental paper, Paterno observed that there was only one vic-dibromoethane, when gauche and anti isomers would be predicted from a static connected tetrahedral model (Figure [1]).[7]

Zoom Image
Figure 1 Paterno's Dilemma: Models for 1,2-dibromoethane predict static isomers (proto-atropisomers)

Kekulé and van’t Hoff rescued their respective proposals by adding dynamic equilibria to their theories. Kekulé ­introduced “time averaging” forms, which in one instant (Zeiteinheit) had single and double bonds in one orientation around the ring and in a second instant the compliment.[3b] Van’t Hoff assumed “free rotation” around single bonds, which effectively averaged the isomers of 1,2-dibromoethane but still predicted the residual stereoisomers of 1,2-dibromo-1,2-dichloroethane, for example. Given the state of structural theory at the time, adding a dynamic process to make the theory fit was an act of fancy even more imaginative than the specific structures. Indeed, ­Findig parodied Kekulé’s benzene by treating the atoms in benzene like a ring of monkeys swinging to and fro (Figure [2]).[8] Van’t Hoff’s initial models have substituents at the faces and in Die Lagerung der Atome in Raume he provides templates to make models with substituents at either the faces or the vertices (Figure [3]).[9] Among factors that favored the vertex model may have been that a free rotation concept is more easily conceptualized if the carbons are linked by a point connection rather than a facial fusion.

Zoom Image
Figure 2 a) Historical model of Kekulé’s benzene; b) Findig’s monkey model satire of Kekulé’s benzene.
Zoom Image
Figure 3 a) van’t Hoff’s original model with substituents at the faces and face-to-face single bonds. b) van’t Hoff’s tetrahedral templates with substituents at faces or vertices.

These critiques notwithstanding, the greater chemical community adopted the idea that a dynamic averaging of molecular structures could give rise to a singular chemical compound at a time when static structural valence theory was not yet on a firm physical footing.[10] It took years before the natural contrapositive to these specific dynamic averaging theories (i.e. that quelling those dynamics should lead to isolable isomeric compounds) would be properly tested.

Mills and Nixon believed to have found a proof for a tautomerim theory of benzene through the reaction products of annulated benzenes.[11] However, Pauling’s 1930 paper on resonance theory, which proposes the D6h structure of benzene, debunked the idea of Kekulé’s tautomers altogether.[12] The feature of equilibrating vs instantaneous contributors to a single chemical entity is the basis for distinguishing tautomerism from mesomerism (resonance theory), of which only tautomers can be isolated.[13] Thus, the realization that benzene possessed a single high symmetry structure obviated looking for any static tautomers, although Pauling’s explanation was apparently too late to avoid the confusion created by the misinterpretations of Mills’ and Nixon’s experiments.[14] No matter the details, the hypothesis of dynamically averaged benzene tautomers was simply obviated because the assumption of two elementary ortho-substituted isomers was incorrect.[15]

Zoom Image
Figure 4Depiction of Kekulé's tautomers by van 't Hoff's tetrahedra

In contrast, the assumption of “free rotation” about single bonds became effectively axiomatic in organic chemistry. When Christie and Kenner began their study of 6,6'-dinitro-diphenic acid in 1922, they were motivated to answer the question of a planar vs twisted biphenyl structure rather than to establish a counter example to “free rotation”; however, their observation of stable, resolvable enantiomers at room temperature addressed both questions.[16] A decade later in 1933, Kuhn coined the term “atropisomerism” to account for isomers resulting from violation of the “free rotation” assumption.[2b]

Kemp and Pitzer recognized in 1936 that even in ethane the archetypal free-rotor there was a varied torsional potential energy profile; and, at low temperature or at very short observation times independent isomers of anti and gauche butane could be observed (cf Paterno's dilemma).[17] [18] This finding extended the discussion of atropisomers to conformational isomerism in general, and focused many physical stereochemists to define conformation in terms of features on a potential energy surface. Important is the relationship between energy, barrier, and time-scale when discussing the observability of "static" isomers vs "dynamic" equilibria.

These “new” isomers placed emphasis on the single bond in the stereochemistry of carbon compounds. Indeed, this focus on the single valence bond sparked many a debate about stereoisomerism. When initially working out the definition of configuration vs conformation, a pivotal issue was whether the definition should be based on the concept of a potential energy surface (PES), irrespective of bonding model or tied to a valence bond model with two-electron/single-line bond connectivity.

In the former, a conformation is an individual structure representing a point anywhere on a PES and a configuration is a structure representative of a well on the PES. Even transition structures can be discussed as conformations. A definition based on PES representation highlights the importance of the time scale and conditions of observation, because different features in the surface can be overcome energetically or averaged out temporally with changes in conditions; thus, an IR experiment at 100 K resolves a different isomeric composition compared to a chromatographic procedure at ambient temperature (300 K) or an NMR spectrum at 450 K. This treatment is extremely general and powerful for explaining experimental results, and also obviates many polemical arguments by placing the discussion in the context of what is experimentally observable as a function of conditions.

PES-based definitions also highlight configurational lability and stability, which in turn underscore the importance of dynamic symmetry in dealing with the spectroscopic features observed for a chemical entity. This, more physical chemical treatment of isomerism, enables a broader and comprehensive set of principles. By freeing the molecule from a specific model for bonding, one enables a stereochemistry beyond the concept of classical stereoisomerism, something crucial for understanding modern supramolecular and materials chemistry phenomena. This definition also fits with the broader concept of atropiso­merism suggested above.

In contrast, a definition of conformation based on states of a model attainable by rotation about single bonds is limited by that bonding model and becomes irrelevant for complex bonding scenarios. The ball and stick or mechanical metaphor of valence bond theory offers a simplistic appeal, but it runs into problems when dealing with the physical and energetic characteristics of complex chemical entities. The classical discussion of atropisomerism based on cessation of rotation about a single bond, ties it precariously to a valence bond model of conformation and seriously limits the scope and utility of the concept.

The relationship between the tetrahedron described by orthogonally held biaryls and that of a tetrahedral carbon is similar to the relationship between ethane and an octahedral coordination center. Indeed, viewed along a specific axis, trans and cis isomers in the octahedron are the cognates of anti and gauche ethane. The Bailar twist mechanism about that axis renders them equivalent. Chromium arene complexes are often described either as dynamic octahedral complexes or as complexes with metal-arene bonds depicted as a single connection emanating from the “center” of the arene and connecting to the chromium atom. Is restricted rotation about this “bond” valid for a discussion of atropisomerism?[19] By abandoning the specific bonding model and focusing on the dynamics, the discussion becomes cleaner and the concepts become more easily grasped, even those where issues of bonding may be questionable.

Zoom Image
Figure 5 Atropisomerism about metal arene bonds

Atropisomers also motivate the question of configurational descriptors for their distinguishability and taxonomy. Configurations are implicitly related to stereogenic elements such as those outlined in the Cahn–Ingold–Prelog (CIP) model for the specification of molecular configuration.[20] Initially a modest nomenclature proposal,[21] the CIP model soon became wrongly associated with chirality and valence bond-dependent stereogenic elements indecorously became elements of chirality.[20] These centers, axes and planes vacillated in number and type — located on atoms or not, axes & planes or just axes, geometric or permutational — but never did the connection to valence bond scaffolds waiver. The valence bond model of tetracoordinate carbon atoms remained sacrosanct; however, the coincidence of permutation and symmetry groups for the regular tetra­hedron is not generalizable.[22] By removing the shackles of the specific bonding model, one can see the issue more clearly, and focus on that which is important regarding structure; energy and properties.

As regards atropisomerism, the problem comes from a belief by some chemists that chirality actually can be deconstructed into the CIP central, axial, and planar elements, a belief that is categorically false. These chemists believe that the ad hoc axial stereogenic elements, based on valence bond-dependent permutation scaffolds, are representative of differing underlying causes of chirality. One sees this expressed in the description of molecules (catalysts) as “axial-” or “planar-chiral”, a distinction that is patently baseless. Equally nonsensical is the description of processes as transforming “central to axial-” or “central to planar chirality” with the intent to express a transfer of chirality. Nonetheless, like many other misconceptions, maintaining the belief in the “Emperor’s new clothes” provides blissful ignorance whereby an admission of intellectual nakedness would be embarrassingly painful. As long as chemists use the language of chirality as mere jargon to push up impact factors, cult-of-personality will dominate over correctness of use.

The proponents of the link between chirality and the artifice of chirality elements are too addicted to their model to follow the path they know to be correct; take for example this passage out of Prelog and Helmchen’s 1982 ACIE treatise:[23]

Although in principle. static stereochemistry could dispense with the use of the rather ill-defined term chemical bond, we would not consider such a move as appropriate and agree with the pointed remarks made by a mathematician, H. Weyl (121 who writes (translated from original German): “one should not take arbitrary relational representations such as valence diagrams too seriously even though they have their use in serving as a rough guide through a seemingly chaotic accumulation of facts. One cannot expect a rough sketch purporting to represent reality to contain all possible shades of that reality. Nonetheless the sketcher should have the courage of his convictions and draw the lines firmly.”

More important than any perceived conviction to draw the lines firmly, is the scientific honesty to admit the limitations of these lines, no matter how boldly drawn. Specific to this discussion is the fact that no matter, how boldly one draws, the lines “chirality elements” (e.g. chiral axes) and “elements of chirality” (e.g. axial chirality) are just jargon with no bearing on geometric chirality; therefore, any connection between atropisomerism and the jargon is equally irrelevant to chirality.

A humorous example of the need for better teaching in the fundamentals of stereochemistry comes from a recent quiz that appeared in Chemical and Engineering News to test readers’ knowledge of atropisomerism.[24] The quiz setters manage to get the definition of atropisomerism wrong in question 1 and in a later true/false question they omit the stereochemical configurational information necessary for you to answer unambiguously whether Iomeprol should manifest atropisomers (Figure [6]). The latter example turns out to be illustrative of exactly the problem of imposing chirality jargon to atropisomers. In the compound in question, there are two flanking stereocenters about the stereoaxis. When the flanking centers are of the same configuration the compound is chiral and yet no atropisomerism is manifest. In contrast when the flanking configurations are opposite then two ACHIRAL atropisomers are found, obviating any possibility of attributing chirality to the stereoaxis. Following the indecorous precedent of calling such stereogenic elements “pseudo” asymmetric is as painful as jamming a left foot into a right shoe.[25] It is clearly time to leave aside the terminology of CIP and simply call things as they are.

Zoom Image
Figure 6 Iomeprol as depicted in Chem. Eng. News. [24]

Recently a high-profile case appeared that highlighted the unnecessary complications added by an overly restrictive definition of atropisomerism. A paper in Nature Chemistry purports to have discovered “akamptisomerism” and a new and final form of isomerism with the implications of atropisomerism.[26] This proposal provides a pseudo-taxonomic distinction completely subsumed under the concepts of atropisomerism and restricted inversion at a divalent atom; both concepts well discussed previously.

As a final word to the reader who has not already been converted to a concept of stereochemistry beyond the valence bond, this issue of Synlett offers a chance for you to try out a new pair goggles. When perusing these articles, ask yourself if a less conventional perspective might offer fewer paradoxes and contradictions and whether it might free your mind to think of creative new interpretations that encompass more phenomena with less conceptual baggage. As Kurt Mislow often admonished, follow the science about which you are passionate, strive for exactness in your communication with others about what you have discovered, be true to yourself, and let the rest sort out itself.


#

   Permissions and Reprints All articles of this category
Zoom Image
Figure 1 Paterno's Dilemma: Models for 1,2-dibromoethane predict static isomers (proto-atropisomers)
Zoom Image
Figure 2 a) Historical model of Kekulé’s benzene; b) Findig’s monkey model satire of Kekulé’s benzene.
Zoom Image
Figure 3 a) van’t Hoff’s original model with substituents at the faces and face-to-face single bonds. b) van’t Hoff’s tetrahedral templates with substituents at faces or vertices.
Zoom Image
Figure 4Depiction of Kekulé's tautomers by van 't Hoff's tetrahedra
Zoom Image
Figure 5 Atropisomerism about metal arene bonds
Zoom Image
Figure 6 Iomeprol as depicted in Chem. Eng. News. [24]