Perspective www.ijc.wiley-vch.de doi.org/10.1002/ijch.202200002 Computational Chemistry as a Conceptual Game Changer: Understanding the Role of London Dispersion in Hexaphenylethane Derivatives (Gomberg Systems) Sören Rösel[a] and Peter R. Schreiner*[a] Abstract: The present personal perspective sheds light on the bonding by mesomeric resonance, and challenged the purely checkered history of hexaphenylethane (HPE) and some of repulsive view on substituent effects. Understanding the its key derivatives, including successes and failures in origin of the instability of HPE on the one hand and the interpreting experimental as well as computational data. HPE stability of the sterically much more crowded all-meta tBu- has become a testing ground for chemical theory since the substituted HPE on the other hand may well be considered groundbreaking work of Gomberg in 1900. It sparked the an important turning point in the appreciation of London growth of theoretical carbon chemistry beyond tetravalency, dispersion (LD) and the analytical power of computational forced chemists to improve theory of color and chemical chemistry. Keywords: Bond Lengths · Hydrocarbons · Noncovalent Interactions · London Dispersion · Radicals Hexaphenylethane: What a Story! bei einer bestimmten Zersetzung gerade unangegriffen blei- benden Reste” (In our opinion, radicals are simply unchanged Gomberg arguably is the godfather of organic radical residues during decompositions).[6] Free radicals like methyl chemistry and the discoverer of the trivalent carbon.[1] As a “C2H3“ were assumed, but proven to be dimers, here, ethane very skilled and meticulous chemist,[2] he was vexed by his „C H [3]4 6“. In 1858, Kekulé recognized and postulated the initial failure to synthesize hexaphenylethane (HPE) and principal tetravalency of carbon.[7] This view hardened in the arrived at the conclusion that this is due to HPE being a very chemical community and deviation from it was often reactive unsaturated hydrocarbon. In a scientific world in dismissed per se without further consideration.[3] Even though which the concept of free radicals was thought to be a mere the radical notion as stated by Kekulé remained, it was “speculative invention” and deviation from Kekulé’s axiom of increasingly faced with doubt. the tetravalent carbon equaled heresy,[3] Gomberg trusted his During this time, Gomberg obtained his chemical educa- analyses and bravely concluded that “Die […] mitgetheilten tion. After escaping from Russia, he arrived in the USA and experimentellen Ergebnisse zwingen mich zu der Annahme, worked hard to pursue his education. He enrolled in the dass in dem ‘ungesättigten Kohlenwasserstoff’ das Radical University of Michigan in 1886, obtained his PhD in 1894, Triphenylmethyl vorliegt” (the experimental findings force me and gained a solid background in analytical and organic to assume that the “unsaturated hydrocarbon” presents the chemistry.[2] In 1896/97, Gomberg conducted research in triphenylmethyl radical).[1a] From thereon, a fervid debate Munich with Adolf von Baeyer and in Heidelberg with Victor about the structure and properties of “triphenylmethyl”, its Meyer, where he synthesized the long elusive dimers and derivatives began, which is still vital and ongoing tetraphenylmethane.[8] Lacking modern spectroscopic tools, more than 120 years later. structural elucidations were based solely on elemental analy- The Years Before 1900 [a] S. Rösel, P. R. Schreiner Institute of Organic Chemistry, Justus Liebig University, Heinrich- Lavoisier first mentioned the term “radical” in 1789 as the Buff-Ring 17, 35392 Giessen, Germany oxidized element in an oxygen acid.[4] In the following years Twitter: @prsgroupjlu until 1840, it was unpopular to call an element a radical. Still, E-mail: prs@uni-giessen.de the term persevered as a description accounting for repeating © 2022 The Authors. Israel Journal of Chemistry published by units in sum formulas, e. g., the etherin radical “C H “[5] was Wiley-VCH GmbH. This is an open access article under the terms of4 4 the Creative Commons Attribution Non-Commercial NoDerivs Li- present in ethanol “C4H4,H2O” or ethyl acetate cense, which permits use and distribution in any medium, provided “C4H4,C2H4O [3]2”. Kekulé summarized radicals appropriately the original work is properly cited, the use is non-commercial and as: “Nach unserer Ansicht sind Radicale nichts weiter als die no modifications or adaptations are made. Isr. J. Chem. 2022, 62, e202200002 (1 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective sis, changes in melting (cryoscopy) or boiling points (ebullio- rated hydrocarbon corresponds to the triphenylmethyl radical scopy) of mixtures as well as chemical behavior. One of (H-1, Figure 1). Gomberg’s main structural proofs for tetraphenylmethane was The possibility of the formation of H-12 (=HPE) was the lack of color upon treatment of the tetranitro derivative dismissed because different chemical behavior would be with aqueous KOH.[9] Back in Michigan, Gomberg envisioned expected of a perphenylated hydrocarbon. Gomberg also the synthesis of the next higher homologue in the series of vouched for a new type of trivalent carbon, unlike that found perphenylated alkanes HPE (H-12) to repeat this structural in unsaturated systems. This unsaturated trivalency might arise proof, ensure its validity, and back up the success of his former from the impossibility to add a second, “zu viel des Raumes synthesis of tetraphenylmethane. […] beanspruchenden” (bulky) triphenylmethyl group to form H-12. Note that literally the same paper was published (just translated) in the Journal of the American Chemical Society Gomberg and the First Organic Radical soon thereafter.[1b] This was not only acceptable but even customary at the time, as was the claim at the end of the paper In 1900, Gomberg published his “Triphenylmethyl, ein Fall “This work will be continued and I wish to reserve the field for von dreiwerthigem Kohlenstoff“ in the famous Berichte der myself”. Deutschen Chemischen Gesellschaft,[1a] thereby also describing his attempts to synthesize H-12. Wurtz coupling of triphenylmethyl bromide and chloride with sodium in benzene The Early “Triphenylmethyl” Period proved unsuccessful. The same reaction with silver, mercury or zinc yielded an insoluble, crystalline solid which he found These findings were soon challenged by Norris and Sanders,[10] to be the peroxide after meticulous examination via elemental by Markovnikoff,[11] and later by Ullmann and Borsum.[12] analysis and comparison to an authentic reference compound. Norris and Sanders first argued against the originality of Molecular weights were accessible neither by cryoscopy due Gomberg’s publication as they also had synthesized the same to poor solubility, nor by ebullioscopy due to the temperature hydrocarbon previously.[13] A second provocative paper named sensitivity of the compound. Exclusion of oxygen by conduct- “On The Non-Existence of Trivalent Carbon” discarded ing the reaction with freshly reduced silver (or better, zinc) in Gomberg’s idea of a trivalent carbon by postulating a quinoid a CO atmosphere yielded a solid hydrocarbon that was readily structure of a monomeric “triphenylmethyl” (TMP)[14]2 – a soluble in benzene and stable over weeks in the absence of structural element that was of high importance for the next oxygen. This hydrocarbon behaved as if highly unsaturated 15 years of vital discussions. Gomberg defended the origi- because it reacted “mit größter Begierde” (with eagerness) nality of his work[15] and answered the second Norris/Sanders with O2, Cl2, Br2 and even with diluted solutions of I2 to give publication thoroughly by disproving it point by point.[16] the corresponding peroxide, chloride, bromide, or iodide, Another disturbance was introduced as Ullmann and Borsum respectively. Gomberg therefore concluded that the unsatu- claimed to have synthesized HPE.[12] Gomberg found the same Sören Rösel currently develops lab syntheses of Peter R. Schreiner is professor of organic active pharmaceutical ingredients into scalable chemistry and Liebig-Chair at the Justus Liebig processes at the R&D department of Thermo University Giessen, Germany. He studied Fisher Scientific’s CDMO site in Linz, Austria. chemistry in his native city at the University of He earned his doctoral degree in organic Erlangen-Nürnberg, Germany, where he re- chemistry at the Justus Liebig University in ceived his Dr. rer. nat. (1994) in Organic Giessen, Germany, under the supervision of Chemistry. Simultaneously, he obtained a PhD Prof. P. R. Schreiner. Here he established (1995) in Computational Chemistry from the hexaphenylethane derivatives as models for University of Georgia, USA. He completed his large dispersion stabilized molecular structures habilitation at the University of Göttingen as demonstrated through experimental explora- (1999). P. R. Schreiner is a member of the tion of their physicochemical and structural Leopoldina – German National Academy of features. Photo credit: Rösel Sciences and received an Arthur C. Cope Scholar Award (2021), the Academy Award of the Berlin-Brandenburg Academy of Science (2020), the Adolf-von-Baeyer Memorial Award (2017), and the Dirac Medal (2003). His research interests include organic reaction dynamics and reactive intermediates, quantum mechanical tunneling as well as London dis- persion interactions as probed in the realm of nanodiamonds and organocatalysis. Photo credit: Schreiner Isr. J. Chem. 2022, 62, e202200002 (2 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective Figure 1. Gomberg’s first visualization of the triphenylmethyl radical. Reprinted with from ref.[1a] Copyright (2021) John Wiley and Sons. Figure 3. Gomberg’s representation of the zinc(II)chloride salt of trityl chloride. Reprinted from ref.[22] Copyright (2021) John Wiley and Sons. stable hydrocarbon after treatment of “triphenylmethyl” with acids and also concluded that this was HPE.[17] Not until two years later, in 1904, Tschitschibabin brominated this hydro- with ZnCl2, to which the following structure was attributed carbon and hydrolyzed it to the corresponding alcohol. This (Figure 3): Shortly thereafter, Gomberg re-advertised the TPM alcohol only could arise from para-benzhydryltetraphenyl- idea based on the isolation of trityl iodide.[23] While the ethane and therefore disproved the HPE claim.[18] reaction of iodine with “triphenylmethyl” was already found A substantial debate about the structure of “triphenylmeth- two years earlier, the isolation was impeded by its similar yl” arose and until 1905, when all structures considered reactivity in solution towards oxygen akin to “triphenylmeth- relevant from today’s viewpoint were already postulated, yl”. Gomberg keenly recognized that “in ungesättigten namely monomeric triphenylmethyl (H-1)[1] and its quinoid Verbindungen zwei Affinitäten unter Bildung einer Doppelbin- mesomer (H-1’),[10] -note that the concept of resonance had not dung verbraucht werden [und] wenn eine solche Verbindung been established at the time- as well as the dimers hexaphenyl- unter gewissen Umstanden durch Halogene gesattigt wird […] ethane (H-1 /HPE),[11] a singly (2)[19]2 and a doubly (3)[20] lagert sich […] an [beide] Kohlenstoffatome ein Halogenatom quinoid isomer (Figure 2). an. Aber in dem hier beschriebenen Kohlenwasserstoff addirt The Quinoid structures. Quinoid structures were initially sich alles Iod nur an ein einziges Kohlenstoffatom - und proposed to account for reactivity and color. The yellow color daraus muss geschlossen werden, dass der Zustand des was first mentioned by Norris and Sanders in 1901.[10] They Ungesättigtseins sich nur auf ein Kohlenstoffatom erstreckt. attributed the reactivity and especially the color to the quinoid Von diesem Gesichtspunkte aus sollte daher dem Kohlenwas- substructure in the “tautomer” H-1’. This view was also held serstoff die Formel [H-1] zugeschrieben werden.” (two by Kehrmann and Wentzel.[21] Gomberg argued against this affinities are consumed for one double bond in unsaturated view, gave colorless examples of such quinoid structures and compounds and if such compounds are saturated with had even obtained the solid “triphenylmethyl” as colorless halogens, halogen atoms are added on both carbons. Instead, crystals.[16] Only the connection between the colored solutions only one iodine atom is added in the described hydrocarbon – and the quinoid structure remained a point of discussion for we therefore have to conclude that the state of unsaturation him. By 1902, it had become evident that “triphenylmethyl” belongs only to one carbon atom. From this point of view the had two modifications. Gomberg could now show that solid structure [H-1] shall be assumed for the investigated hydro- “triphenylmethyl” was indeed colorless, while solutions were carbon). always yellow.[22] Gomberg switched from the opposing In the meantime, the discussion about the molecular Kehrmann-Wentzel/Norris-Sanders theory to Baeyer’s recent structure was enriched by Heintschel’s double quinoid dimer theory of halochroism and attributed the color to a heterolysis structure 3,[20] another instance of the quinoid mesomer to of “triphenylmethyl” into “pseudoions”. This agreed with the account for reactivity and color. Subsequently, the dimeric appearance of the same color upon treatment of trityl chloride nature of “triphenylmethyl” received support in 1904 when Figure 2. Suggested structures in the discussion on radicals and HPE (H-12): the monomeric triphenylmethyl radical (H-1), its quinoid mesomer (H-1’), and the dimers HPE, singly (2) and a doubly (3) quinoid isomers. Isr. J. Chem. 2022, 62, e202200002 (3 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective Figure 4. Ullmann’s hydrocarbon 4 is obtained after treatment of “triphenylmethyl” with dry HCl.[12] Tschitschibabin subsequently obtained alcohol 5 via bromination and hydrolysis.[18] Jacobson took the acid catalyzed re-aromatization as evidence for 2.[19] (after initial problems with the cryogenic determination of the molecular weight of “triphenylmethyl”)[24] Gomberg published a final molecular weight of 477, which is almost double the weight of the monomer (243).[25] This clearly pointed towards a dimeric structure. Hence, Gomberg gave up a solely monomeric “triphenylmethyl” and mentioned an equilibrium Figure 5. Abstraction of the para-chloride from a benzoic structure between an associated and dissociated “triphenylmethyl”, as in 12 was thought to be difficult while the abstraction from the which could not be 3. Higher temperatures yielded a deeper quinoid position was assumed to proceed readily. Today, this would color as well as a lower molecular mass. Both pieces of be explained by the difference in radical stability between a phenyl evidence pointed towards a temperature-dependent equilibrium radical localized in an sp 2-orbital and a stabilized allylic radical in a delocalized π-orbital. of a colored and an uncolored form. Based on Tschitschibabin’s publication on the wrong “HPE”,[18] Jacobson immediately suggested an unsymmetric dimer, namely 1-diphenylmethylen-4-triphenylmethyl- The Hexaphenylethane (HPE) Structure. Another group cyclohexa(2,5)diene 2.[19] He based this structure on the acid preferred H-12 as the structure of “triphenylmethyl”. Already catalyzed rearrangement of “triphenylmethyl” to Ullmann’s in 1902, Markownikoff attributed the high reactivity of hydrocarbon (Figure 4). This was the first solid evidence that “triphenylmethyl” to a very unstable H-1 .[11]2 Later, Flürscheim could have led to the assignment of the correct structure. Yet, – an advocate for a “mesomeric resonance” – argued that this it was questioned.[26] “resonance” might lead to the abnormal formation of In 1905, Gomberg formulated his dilemma:[26d] On the one Jacobson’s hydrocarbon 2.[26b] Nonetheless, the direct forma- hand, Jacobson’s structure 2 (the Jacobson-Nauta structure, tion of a bond between the carbon atoms of the highest affinity JNS) accounts for a dimer, the rearrangement to Ullmann’s and thus the formation of HPE seemed more likely. hydrocarbon, and evades the postulation of a trivalent carbon. “Triphenylmethyl” therefore could be colorless H-12 in It does not explain the observed reactivity and color (now equilibrium with colored H-1. Tschitschibabin also preferred Gomberg invoked the lack of color of known quinoid the HPE structure of “triphenylmethyl”, but stated that one structures). On the other hand, TPM explained its chemical could not exclude Jacobson’s structure, for the same reasons behavior perfectly, but contradicted the experimental mass. A Jacobson had named.[26c] This was immediately challenged by true equilibrium was not yet recognized and Gomberg instead Gomberg in 1905, who excluded HPE because of the made use of an alternative explanation by postulating an presumably “saturated” character of this hydrocarbon, the association of H-1, similar to the dimerization of carboxylic impossibility to form similar, less phenylated ethanes at the acids to account for the odd molecular weight.[26d] time[26d] and, in 1906, due to the experimental para-halogen Even more evidence towards a quinoid dimeric structure lability.[27] After tetra- and pentaphenylethane had been came in 1906 by Gomberg himself. In an attempt to solve the synthesized, Gomberg argued the opposite (but still against H- dimeric structure of “triphenylmethyl” experimentally, he 12): “So dürfen wir doch aus der Existenz […] des treated para-halogenated triphenylmethyl chlorides with Tetraphenyl- und Pentaphenyl-Aethans, schliessen, dass auch silver.[27] The amount of extracted chloride then should provide Hexaphenyläthan existenzfahig ist, und dass es, einmal hints regarding the binding mode of the corresponding carbon gebildet, sich als beständiger und wenig reactionsfähiger (Figure 5) and therefore whether H-12, 2 or 3 was present; the Kohlenwasserstoff erweisen wird.” (We have to conclude from results excluded H-12 but did not allow for a differentiation the existence of tetra- and pentaphenylethane that HPE, once between 2 or 3. Later results also excluded an ortho-quinoid formed, will be a stable and inert hydrocarbon).[27] structure.[28] In 1907, Gomberg concluded that “triphenylmeth- The last chemical evidence (beside the acid catalyzed yl” mesomerizes in solution into 2 which is in equilibrium rearrangement and para-halogen lability) was noticed early in with H-1. “[…] die Konstitution des in Lösung befindlichen 1901, but only brought to completion 70 years later: A liquid Kohlenwasserstoffes [ist] schon jetzt als mit befriedigender autooxidation byproduct referred to as “the oil”.[30] The yield Sicherheit aufgeklärt [als] Jacobson zugeschriebene of bis(trityl)peroxide upon quenching of a “triphenylmethyl” Formel[…]” (the constitution of the dissolved hydrocarbon is solution with oxygen was only about 80%, even though a full already characterized securely as Jacobson’s suggested equivalent of oxygen was consumed. The remaining 20% formula).[29] contained mainly an oily oxidation byproduct of Isr. J. Chem. 2022, 62, e202200002 (4 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective “triphenylmethyl”. This caught the attention of Gomberg, who proved the existence of a free radical and made Gomberg’s reinvestigated “the oil”,[24] and Norris, who stated that “[the hypothesis of “triphenylmethyl” being the trityl radical fully oil] has not received in the past the attention that it plausible. Secondly, by demonstrating the possibility of deserves.”[31] Gomberg examined the oxidation of isolating a stable HPE derivative, namely 1,1’-diphenyl- “triphenylmethyl” to the peroxide in detail and found bisfluorenyl 7 (Figure 6).[39] Its synthesis was erroneously increased amounts of “the oil” by usage of absolute oxygen claimed by Gomberg in 1906,[27] but later 7 was obtained as a instead of air. However, “Wir haben diese öligen Producte colorless solution which evidently consumed oxygen.[27] This übrigens keiner eingehenden Prüfung unterzogen und uns auf back-bound HPE derivative dissociated only upon heating. die Feststellung beschränkt, dass [die Öle] etwa ebensoviel Further evidence was introduced by Piccard who recog- Sauerstoff wie das Peroxyd selbst enthalten” (We did not nized, based on deviations from Beer’s law, that the equili- analyze “the oil” further and leave it with the notion that it brium must be between a colored monomeric and a colorless contains as much oxygen as the peroxide).[25] Norris’ opinion dimeric “triphenylmethyl”.[40] By disproving Schmidlin’s “As an oil [is] at times present, an investigation of the cause “tautomerism” thesis,[35] further support for the H-12)*H-1 of the formation of [this substance] must be made before the equilibrium was thought to have been found. reaction,[32] which is evidently a very complex one, can be Gomberg, however, warned: “Es ist in den letzten Jahren exactly interpreted” would finally prove correct. Deeper üblich geworden, die dimolekulare Form des “Triphenylmeth- insights were gained in the 1930s[33] and the problem was yls” als Hexaphenyläthan zu bezeichen. So plausibel diese solved in the 1970s,[30a] as will be shown later. Ansicht auch von vornherein erscheinen mag, so dürfte es By the end of 1907, Gomberg admitted, regarding the doch angebracht sein, daran zu erinnern, daß sie des constitution of solid “triphenylmethyl”: “eine Auswahl zwi- experimentellen Beweises bis jetzt noch entbehrt” (It became schen [den Möglichkeiten[34]] zu treffen, […] mehr Sache des common to consider the dimolecular “triphenylmethyl” as persönlichen Gefühls als der logischen Deduktion” (a HPE. As plausible as this might look, it is advisable to remind distinction between the options is rather a matter of personal oneself that this view lacks proper experimental evidence).[41] flavor than logical deduction).[29] He still adhered to the quinoid structures and postulated the More light was shed on the “triphenylmethyl” puzzle, “equilibrium formula” in which all possible forms – the when Schmidlin observed a temporary decolorization of a mesomers of TPM and HPE – were taken into account. This “triphenylmethyl” solution if shortly exposed to oxygen.[35] view is explained in depth in Gomberg’s 1914 review about This was rationalized by an equilibrium between an unreactive the initial years of “triphenylmethyl”.[42] The existence of free colorless and a reactive colored form in accordance with radicals is established as “many analogues of triphenylmethyl Gomberg’s earlier hypothesis[25] of such an equilibrium. are capable to exist”. He recapitulates the most striking Subsequently, Flürscheim advertised his earlier view of an features of these radicals, namely color, autooxidation, and equilibrium between colorless H-12 and colored H-1 again and acid catalyzed re-aromatization: The color is indicative of two saw it confirmed by Schmidlin’s observation.[36] This interpre- forms of “triphenylmethyl”, a colored and an uncolored one in tation was backed by Wieland who recognized the similarity equilibrium with each other. “Triphenylmethyl” takes up between the temperature dependency of the N2O4)*NO2 oxygen with the eager of yellow phosphorus (the impure white equilibrium and the temperature dependency of the molecular allotrope). The acid catalyzed rearrangement of weight of “triphenylmethyl”. He postulated that only the “triphenylmethyl” to Ullmann’s hydrocarbon is depicted as a radical character of H-1 is responsible for the color and single step from H-1 to 4 (Figure 7) without discussion of the reactivity and makes the idea of mesomerism to a quinoid mechanism or noticing its importance. form unnecessary.[37] Gomberg reminds us of his dilemma: While the trivalent The structural debate was settled in 1910, when Schlenk carbon was hesitantly accepted and the free TPM radical entered the field. Firstly, by the synthesis of the tri(4,4’- explains all the reactivity, the weight of “triphenylmethyl” biphenyl)methyl radical 6 (Figure 6), which was the first fully corresponds to a dimer and gets lower with raising temper- dissociated, persistent, free organic radical;[38] this finally ature. He concluded: “Hexaphenylethane still remains a figment of the imagination. And so the idea began to take root more and more generally that, after all, there is no difference between the so-called “triphenylmethyl” and the hexaphenyl- ethane, that the former is in reality the latter. […] Henceforth Figure 6. The first fully dissociated free radical tri(4,4’-biphenyl)meth- Figure 7. Gomberg’s representation of the acid catalyzed re-aromati- yl radical 6 and the first stable HPE analog 1,1’-diphenyl-bisfluorenyl zation. Reprinted with permission from ref.[42] Copyright (2021) 7. American Chemical Society. Isr. J. Chem. 2022, 62, e202200002 (5 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective the hexaphenylethane constitution for triphenylmethyl began pathway, which leads to “the oil”: The scavenging of oxygen to appear in the literature more frequently[…]” But “It is self- by the radical and, and this was new, the direct oxidation of evident that such a constitution could at best account for the triphenylmethyl.[33b] Three years later, the concentration colorless triphenylmethyl alone.” He summarized dependence of the direct oxidation of the dimer was “triphenylmethyl” therefore as (Figure 8): While the equili- discovered.[33a] Ziegler showed an inhibition of product brium formula “is in full harmony with the facts known up to formation by high concentrations of pyrogallol in the oxidation the present, […] the explanation hexaphenylethane *) to the bis(trityl)peroxide. Only tritylhydroperoxide formed triphenylmethyl [radical] is finding a wider acceptance largely and, accordingly, one equivalent of O2 had been consumed per because of its simplicity.” This point of view ultimately would trityl chloride. This revealed a stepwise oxidation. Unfortu- be accepted for the next five decades: Schmidlin’s book about nately, Ziegler made no attempt to characterize the product of “triphenylmethyl” broadly reviews every aspect and treats direct oxidation and the interpretation of “triphenylmethyl” as solid “triphenylmethyl” as H-12 which upon dissolution “most H-12 remained. probably” partly dissociates to the H-1 radical.[43] Jacobson’s The ability to measure the magnetic susceptibility sparked structure 2 seemed out of question, even though he mentioned attempts to gain new insights into the behavior of it as one of the other possibilities. Schmidlin saw nothing “triphenylmethyl” and its derivatives. Roy and Marvel found arguing against H-12. In 1916, Lewis explained organic that with increasing “weight” of alkyl substituents in “all-para radicals based on the electronic structure as molecules with an alkyl HPE” the degree of dissociation increased.[49] These odd number of electrons.[44] He correlated the color of the finding are hardly explicable if one assumed a hexaphenyl- radicals with the last, odd electron and rendered the quinoid ethane structure as steric and resonance effects are negligible mesomer of the “triphenylmethyl” monomer unnecessary. He in the para position. Indeed, the change in dissociation was agreed with an equilibrium between H-12 and H-1 in benzene. later attributed to substituent effects on the molar magnetic In Gomberg’s 1924 review, a quinoid mesomer of the susceptibility.[50] When Ziegler wrote a “Festschrift” about his monomeric species was still assumed to account for the 25th anniversary in radical chemistry in 1948, the H-12)*H-1 color.[45] In the whole review, 2 was not mentioned, rather, an equilibrium and the successful synthesis of HPE had been equilibrium between H-12, H-1, and quinoid 1’. The structure fully accepted. In the review, the question was raised why the of solid “triphenylmethyl” seemed accepted as HPE. The acid central carbon-carbon bond in HPE is so weak and resonance catalyzed rearrangement now occurred as a two-step process vs. steric effects were discussed. Both effects were said to as postulated by Tschitschibabin and therefore the strongest work together: Resonance stabilizes the radicals while steric chemical hint for the true structure of “triphenylmethyl” was clashes lead to an elongated and therefore weak(er) “ethane” disregarded. Soon after, even Gomberg accepted the structure bond.[51] Around the time when the whole academic knowledge of H-12 for solid “triphenylmethyl” because of its lack of of the synthesis of HPE imploded, Schwartz synthesized a color.[30a] further back-bound, stable HPE derivative, bitriptycyl (8),[52] and Wittig and Schoch synthesized hexabenzo- [b.d.g.i.Z.n][4.4.4]propellamancene (9), an all-ortho cross Physical Properties of “Triphenylmethyl” bridged HPE (Figure 9).[53] No further questions on “triphenylmethyl” were asked. In the following years, the question shifted from structure to properties. In 1929, Ziegler found the dissociation of “triphenylmethyl” to be in agreement with Ostwald‘s law of A New Race Towards Hexaphenylethane dilution.[46] The determined degree of dissociation of 3.6% corresponds to a heat of dissociation (ΔH 293d ) of A landmark in the “triphenylmethyl” history was the 1968 11.5 kcalmol 1. For the first time, a quantification of the publication of Lankamp, Nauta and MacLean, which once and instability of dimeric “triphenylmethyl” was given. This value for all proved the structure of dimeric “triphenylmethyl” as 1- was later confirmed by Müller[47] and much later with the diphenylmethylen-4-triphenylmethyl-cyclohexa(2,5)diene 2,[54] correct structural assignment by Neumann.[48] the structure suggested by Jacobson in 1905.[19] They In 1930, experimental rates of the oxidation of reinvestigated “triphenylmethyl” in terms of UV and NMR “triphenylmethyl” revealed the existence of a second reaction spectroscopy. Hence, “triphenylmethyl” showed peaks at 5.8 Figure 8. The “equilibrium formula” from Gomberg’s 1914 review. Reprinted with permission from ref.[42] Copyright (2021) American Chemical Society. Isr. J. Chem. 2022, 62, e202200002 (6 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective Figure 9. Perspective skeletal formulas of the back bonded or clamped HPE derivatives: 1,1’-diphenyl-bisfluorenyl 7, bitriptycyl 8 and hexabenzo-[b.d.g.i.Z.n][4.4.4]propellamancene 9. to 6.4 ppm, assigned to the olefinic hydrogens, and a signal at biphenylyl)methyl radical (10) and also found “no significant 5 ppm, assigned to the para-methine hydrogen, in contrast to additional stabilization due to π-conjugation relative to [the the expectation of a signal-free olefinic region in the 1H NMR all-meta tBu TPM radical]”.[61] for true HPE.[54] The interpretation was confirmed one year In 1978, Rieker et al. crystallized the first unbridged HPE later when Brunner synthesized an α-13C derivative of derivative entitled “Hexakis(2,6-di-tert-butyl-4- triphenylmethyl chloride and investigated the corresponding biphenylyl)ethane [102] – The First Unbridged dimer via 13C NMR. Two resonances [62.6 (sp3-C) and 137.3 Hexaarylethane”.[62] Baer et al. found monomeric crystals (sp2-C) ppm] were found indicative for the unsymmetric grown from benzene;[59] Rieker chose cyclohexane and structure 2. Further, 1H NMR studies gave integrals of 4 :1 for obtained colorless crystals of 102. While the determined SC- the signals at 6.4–5.8 and 5.05 ppm, respectively. All-para XRD proved the HPE structure (Figure 10), an oddly short deuteration eliminated the resonance at 5.05 ppm and gave a ethane bond of rCC=1.47 Å was determined. This contradicted clear AB system at 6.3 and 5.9 ppm with J=10.5 Hz, in line previous views of Ziegler[51] and empirical force field (EFF) with structure 2.[55] Subsequently, the structure of “the oil” (the computations of Mislow,[63] who predicted an elongated bond. byproduct of the autooxidation of “triphenylmethyl”) was Yet, Rieker explained the existence of 102 purely based on determined as para-(hydroperoxidobenzhydryl)-tetra-phenyl- steric reasons: “Steric hindering of formation of the Jacobson- methane, which must originate from 2.[30b] Ultimately, single Nauta structure by incorporation of bulky groups [in meta and crystal X-ray diffraction (SC-XRD) structures of 2 were para positions lead to the following picture: the TPM radical] determined by Allemand[56] and later by Blom.[57] While evidence from spectroscopic tools like NMR might have been available only since the 1950s, pure chemical evidence – acid catalyzed re-aromatization, para-halogen lability and, if it had been analyzed, the structure and origin of “the oil” – pointed directly towards structure 2.[30a] Hitherto, hexaphenylethane had never been synthesized and remained elusive. Therefore, the race towards the synthesis of an unbridged HPE had been reset by Nauta in 1968. The First Unbridged Hexaphenylethane In the 1970s, the structural properties of radicals were predominately determined via electron paramagnetic reso- nance (EPR) spectroscopy. The EPR spectrum of TPM was problematic to solve because of odd contributions at the meta positions. Based on previous findings that the introduction of tBu groups drastically simplified EPR spectra,[58] Baer et al. synthesized and investigated all-meta tBu TPM.[59] They found a negligible electronic influence by meta tBu substitution and a Figure 10.Molecular structure of hexa(2,6-di-tert-butyl-4- twisting of 24 to 26° of the phenyl rings. These findings were biphenylyl)ethane (102) determined via SC-XRD and as depicted by later confirmed by Sakurai and coworkers.[60] Following the Rieker. Hydrogens were not shown. The central rCC was determined EPR resolution improving ability of all-meta tBu substitution, as 1.47(2) Å. Reprinted with permission from ref.[62] Copyright (2021) Stein and Rieker synthesized the tri(2,6-di-tert-butyl-4- John Wiley and Sons. Isr. J. Chem. 2022, 62, e202200002 (7 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective must then either be persistent as the monomer – or dimerize […] to [HPE].” Moreover, they found a decay of the HPE signals over six days with a final degree of dissociation of 63% for 102 and 55% for all-meta tBu HPE (tBu-12). The authors took this decay as a sign of a slow equilibration and the assumed high activation barrier as a hint toward the existence of an HPE derivative. Few spectroscopic data were given; NMR spectroscopy gave no definitive conclusion because of “the high degree of dissociation and the sensitivity of the radical towards traces of oxygen”. Still, a chemical shift of 70.7 ppm was specified for the ethane carbons. In the end it was claimed that “The high melting point and slow dissocia- tion of [102] would suggest that the unsubstituted hexaphenyl- ethane should also be synthesizable.” In contrast to an unusual short central C C-bond, Mislow predicted a bond elongation in HPE on the basis of extended force-field (EFF) computations.[63] A chiral D3 eclipsed conformation was found to be the ground state, being 2.55 kcalmol 1 lower in energy than an achiral S staggered Figure 11.Molecular structure of hexa(3,5-di-tert-butylphenyl)ethane6 t [66] conformation that had been found experimentally. Severe Bu-12 determined via SC-XRD by Mislow taken from the CSD (No. 1153536).[68] The rCC was determined as 1.67(3) Å. Hydrogens andstrain between the trityl moieties was made responsible for a disorders were omitted for clarity. The odd angles around the ipso- drastically elongated rCC of 1.64 Å in both conformers. Steric carbons arise from the disorder. arguments were also used by Rieker to explain the formation of the ethane bond. As there were several bridged HPE structures available,[39,52–53] Mislow commented: “Neverthe- unambiguously determined to be 1.67(3) Å in line with less,[…] all compounds […] with bond lengths greater than previous computations. In 1988, Mislow and Kahr again 1.6 Å are bridged to avoid […] the molecule to fall apart. published new evidence to account for the lack of biphenyl Thus, the molecules survive despite the presence of what is groups in earlier studies. Because they were unable to obtain most likely a severely weakened bond. HPE does not enjoy suitable crystals for SC-XRD, nutation NMR spectroscopy such bridging.” Clearly a correlation between bond length and was performed on powders. α-13C Derivatives of 10 t2 and Bu- strength is made: “The stability of polyarylethanes correlates 12 were synthesized and 13C 13C distances of 1.64–1.65 Å well with rCC[…]”. Rieker argued that the tBu substitution were determined in both cases. With this, Mislow summarized might lead to the bond shrinkage instead elongation.[62] As it that “[…] two independent experimental and one computa- turned out later, vide infra, there is some truth in this statement tional determination of [rCC] comprise a set of values with but with an entirely different causality. surprisingly little variance. […] These results remove any last As a result to Rieker’s argument, Mislow revised his doubt that the previously reported central bond distance is former computations on HPE because of “the dismaying grossly in error.”[67] Hence, theoretically well-founded compu- prospect of a major failure of the EFF method in the prediction tations provided the impetus to redo an experiment and of a structure of central importance in chemistry.”[64] At the eventually correct a very significant error. One key question time the trust in theory was in many instances not high enough was not asked by Mislow and Kahr: why is tBu-12 isolable to question an experiment. The efforts were extended to the while much less crowded H-12 is not? What is the origin of the MMI, MMPI, and MM2 empirical force-fields as well as stabilization through the tBu groups? semiempirical MNDO quantum mechanical computations on It was not until 14 years later, when HPE found its way HPE. Even EFF computations of the very large 102 were back into the literature by a computational analysis of Vreven performed; “the largest molecule ever calculated by a full- and Morokuma, who re-computed the structure of H-12 with a relaxation EFF method”[64] – and so any computational method three-level hybrid quantum mechanical ONIOM method.[69] at that time.[65] The computational results all agree on an With a computed BDE of 16.6 kcalmol 1, the synthesis of elongated central C C bond of more than 1.6 Å. Mislow HPE seemed feasible. But, as Lewars later put it: “The thoroughly confirmed his former results and declared a problem with making hexaphenylethane likely lies not with the reinvestigation of the SC-XRD to eliminate any doubts. instability of this molecule, but rather with the preference of Five years later, Mislow and Kahr crystallized a truncated the triphenylmethyl radical to dimerize faster to the meth- 10 , namely all-meta t2 Bu HPE tBu-12 (Figure 11).[66] The ylenecyclohexadiene than to the hexaarylethane. This is a crystallization took over two months. Still, the crystals showed kinetic effect probably arising from less steric hindrance in the severe disorder similar to other hexa-homosubstituted ethanes. transition state for formation of the methylenecyclohexadiene: The outer phenyl substituents were heavily influenced by this this dimerization mode is less disfavored by nonbonded (steric) disorder (Figure 11). Nonetheless, the central bond length was Isr. J. Chem. 2022, 62, e202200002 (8 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective interactions than is the ‘end-on’ dimerization of Ph3C+ opinion of the present authors that it is much more satisfying CPh3.”[65] to identify stabilizing interactions than destabilizing factors; it Vreven and Morokuma found a rCC of 1.72 Å which is is also didactically easier to convey. longer than the reported experimental values for the more A step in the right direction was taken by Dames et al. in crowded HPEs.[66–67] Referring to strain as the major reason for 2010.[72] For the first time, parent H-12 was examined with a the long rCC and therefore the instability of HPE, they were computational method that included considerable mid-range unable to rationalize this deviation in rCC and attribute it to the electron-correlation effects at the M06-2X/6-31+G(d,p) level “soft central bond” being compressed by crystal packing of theory, thereby also capturing an appreciable amount of effects. London Dispersion (LD). A rCC of 1.70 Å and a BDE of 11.6 kcalmol 1 was determined (11.3 kcalmol 1 via an iso- desmic equation) in agreement with Vreven and Morokuma.[69] London Dispersion Interactions in Hexaphenylethanes In contrast, B3LYP (which is not geared toward inclusion of medium to long-range electron dispersion effects, including Mislow et al. as well as Vreven and Morokuma attributed a LD) failed to describe the BDE ( 23.8 kcalmol 1).[73] In light “great internal strain and long (therefore weak) central C C of this, “The central C C bond strength [is] a result of bond” to HPE; seemingly steric clashes are responsible for the competition between steric repulsion and dispersive attrac- failure of all synthetic attemps.[63,65,69] According to Rieker, tion.” steric clashes are also responsible for the ethane bond formation as ”Steric hindering of formation of the Jacobson- Nauta structure [10] must then either be persistent […] or Insertion: Long Known but Long Neglected London dimerize to 10 .“[62]2 Another instance of steric clash was found Dispersion Interactions when Neumann et al. disproved earlier findings[49] that increased bulkiness of all-para alkyl substituents at TPM The importance of attractive intermolecular forces was already radicals leads to more facile dissociation of the “HPEs”: “Eine noticed by Johannes Diderik van der Waals (vdW), a Dutch Dimerisierung der Radikale findet also nicht statt. Denkbar physicist, in 1873. Severe deviations from the ideal gas law war zumindest in einem Teil der Beispiele eine Chinoidbildung were apparent at that time. Therefore, vdW developed a simple […]. Das ist nicht der Fall. Selbst bei kleinen Resten wie F but intelligible correction through the incorporation of micro- oder CN wird dem ein unüberwindbarer Widerstand entgegen- scopic properties to the macroscopic ideal gas law to account gesetzt. […] Weiterhin hätte, wie im Falle des Tris(3,5-di-tert- for the deviations to model real gases. The vdW equation butyl-4-phenylphenyl)methyls 102 bei einfacher para-Substitu- allows, in contrast to, e. g., the virial expansion, for a direct tion nun doch vielleicht das ethanartige α,α-Dimer auftreten comprehension of molecular expansions and interactions. können. Auch das ist nicht der Fall” (The radicals do not These interactions and the behavior of compounds in all states dimerize. At least in part of the examples a quinoid can be understood on the basis of electrostatic interactions dimerization can be envisioned. This is not the case. Even between (in shape and charge distribution anisotropic) mole- small substituents such as F or CN introduce an insurmount- cules. Hence, Coulomb, Keesom, and Debye interactions are able resistance against dimerization. Further, in analogy to 102, readily explained by static charges, dipoles or higher multi- simple para substitution could lead to an ethane-like α,α- poles. The existence of condensed matter for most molecules dimer. Again, this is not the case).[50] is rationalized on the basis of these classic electrostatic Steric clashes refer to Pauli (exchange) repulsion of interactions. Still, none of these interactions explain the molecules or moieties in finite proximity. This is a destabiliz- appearance of a condensed phase for, e.g., noble gases. ing effect. In this sense, steric strain destabilizes HPE and an A novel, non-classic, quantum mechanic, attractive, long equilibrium between the less strained JNS and free radicals range interaction (~ r 6ab ) was found in a perturbative was observed. Steric shielding of the para-positions by para description of two hydrogen atoms at “large” distances by substitution also destabilizes the JNS through increased strain London and Eisenschitz in 1930.[74] “Man wird diese […] and persistent TPM radicals are observed. Steric shielding of Anziehungskräfte in Zusammenhang mit den van der Waal- the para-positions by all-meta substitution leads to an isolable schen a-Kräfte bringen“ (One will connect these attractive HPE. This observation could arise from a destabilization of forces with the parameter a from the van der Waals forces). In not only the JNS but also of the radicals. Destabilization of the their rigorously derived perturbation matrix the second off- TPM radicals could occur by a decreased delocalization of the diagonal elements were attributed to a resonance effect. Here, unpaired electron via increased twisting angles of the phenyl a virtual double excitation takes place: Transitions have dipole rings or electronic substituent effects. Both are experimentally moments which mutually attract each other.[75] These excita- not observed as the twist angles in TPM (26.5°)[70] or all-meta tions arise from “virtuelle periodische Bewegungen” (virtual tBu TPM (24–26°)[59] are similar and meta alkyl substitution periodic movements), later referred to as zero-point motion.[76] has a minor effect on the radical stability.[71] Therefore, the In the subsequent publications London developed an question arises whether there is a stabilizing effect that could extensive systematization of the vdW interactions – Keesom, explain the increased stability of tBu-12. Note that it is the Debye, and dispersion – and introduced further approxima- Isr. J. Chem. 2022, 62, e202200002 (9 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective tions for the latter.[77] The magnitude of the transition dipole large attention recently.[80,84] This renewed interest in LD is moments should be derived from the oscillator strength f supported by computational developments which enable us to within the framework of the dispersion of light because dissect noncovalent interactions and reveal the significance of “[…]diese Wirkungen, die man selbst für die einfachsten LD. Among these are symmetry adapted perturbation theory Moleküle kaum je wird direct berechenen können[… ]” (these (SAPT),[85] switchable DFT dispersion corrections (DFT-D),[86] dipole moments are barely directly calculable, even for the and the local energy decomposition (LED),[87] to only name smallest molecules). Hence, London called these interactions but a few. dispersion forces. Since even values for f had not been While it is generally accepted that LD governs the determined then, the isotropic polarizability α was used interaction of uncharged molecules in the gas phase and solid instead. Further, as a workaround for the unknown “virtual” state, there are questions about the meaning and importance of excitations energies for the noble gases, London assumed an LD in solution. While Hunter questioned stabilizing LD interval between the first excitation energy and the ionization solute-solute interactions in general,[88] Cockroft et al. state energy IE. It appeared that the calculated a-values correspond- regarding the difference between measured and theoretical ing to the IE agreed well with the experimental a-values and unfolding energies of a double alkyl substituted Wilcox the dispersion energy EDisp was reasonably described for noble balance[89] that: “The most likely explanation for the order of gases by: magnitude difference between the very small alkyl-alkyl interaction energies measured in this study and the large 3 IEa 2 energies derived from enthalpies of vaporization and computa- EDisp ¼ (1)4 r6 tional methods is that dispersion forces are effectivelyab cancelled by competitive dispersion interactions with the Already at this early stage, London noticed that for many solvent,[…].”[90] This difference is most likely caused by the molecules, EDisp dominates intermolecular interactions. London multiple conformers of the long alkyl chains which are in states: “It is seen that the induction effect is in all cases equilibrium with each other at r.t. that are not taken into practically negligible, and that even in such a strong dipole account of in Cockroft’s computational analysis. This cancel- molecule as HCl the permanent dipole moments give no ation also contradicts Wilcox’s earlier notion of the importance noticeable contribution to the van der Waals attraction. Not of LD in his balance.[91] Shimizu found not a cancelation but earlier than with NH3, does the orientation effect become just an attenuation and “Aromatic surfaces still form attractive comparable with the dispersion effect, which […] seems in no dispersion interactions in solution just as they do in vacuo.”[92] case to be negligible.”[76] A recent experimental comparison of the very same species in Unfortunately, these insights and rationalizations were by gas phase and solution by Chen and co-workers did not show and large not accepted or even acknowledged in the chemistry a cancelation but also an attenuation. The study of solvent community. With the appearance of SC-XRD, CPK models, effects on intramolecular dispersion were also pursued in the force fields and highly approximate quantum mechanical Schreiner group by using the valence bond isomer equilibra- methods the focus continually shifted to the “directly” tion of 1,4- and 1,6-di-tBu-cyclooctatetraene in 16 different accessible interactions like steric effects via bond length and solvents.[93] In the folded 1,6-isomer, the two tBu-groups are at angle deformations as well as electrostatic interactions like an H···H distance of about 2.5 Å but they are well separated in hydrogen bonding or “π-π stacking” attributed to certain the unfolded 1,4-isomer (H···H distance �7 Å). Temperature- functional groups. Hunter cunningly points this out when dependent nuclear magnetic resonance measurements on the writing: “When molecular scientists obtain an unexpected equilibrium positions in these solvents revelead that the folded result in a system, they tend to invoke the mythical powers of isomer always is preferred. Energy decomposition analyses at the ‘π-π interaction’, ‘π-stacking’, ‘charge transfer’ (CT), ‘π- the density functional and ab initio levels emphasize the acid/π-base’ or ‘electron donor acceptor (EDA) predominance of LD interactions enthalpically, to arrive at the interaction’.”[78] Note that the interaction of benzenes is conclusion that intramolecular LD interactions not cancel in governed predominantly by LD.[79] The notion of LD as a solution. small force between small particles supported its massive Hence, a significant amount of LD transfers from the gas underestimation.[80] Long accepted and often used density phase into solution;[94] this is a general observation.[95] functional theory (DFT) implementations (e. g., B3LYP), Astonishingly, LD even significantly contributes in highly frequently employed because of their moderate scaling of polar media[96] or even in dimers of equally charged species.[97] computational demand whilst providing direct insights into In light of the importance of LD for molecular structures it is chemical properties,[81] inherently truncate electron correlation striking how a purely steric view on HPE could remain for effects and therefore lack the inclusion of LD.[82] It was over 100 years. noticed early that certain popular combinations might cause favorable error compensations and give “the right answer for the wrong reasons”.[83] The failure of such widely applied now considered early DFT computations surfaced and the importance of LD gained Isr. J. Chem. 2022, 62, e202200002 (10 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective Solving the Hexaphenylethane Riddle and Schreiner concluded that “the overall repulsive phenyl- phenyl interactions in HPE are overcompensated in all-meta The question about the singularity of isolable all-meta tBu tBu HPE by addition [of] tert-butyl groups that serve as substituted HPE derivatives 10 t2 and Bu-12 or stable, unbridged ‘dispersion energy donors’ […].” These are now commonly and yet very elongated sp3-sp3 C C bonds in general was referred to as “DEDs” (Figure 12). approached by Schreiner and Fokin in 2011, when they discussed the bond length-bond strength correlation:[98] “[…]shorter bonds are considered stronger, and vice versa. Dissociation of Hexaphenylethanes However, there are many exceptions […]” and “The general recipe for elongating chemical bonds involves steric crowding. We published our first experimental results about tBu-12 at the [It] reaches its limit of applicability with the highly crowded end of 2016.[102] The goal was to gain deeper experimental ‘classic riddle’ hexaphenylethane, which has not yet been insights into the bonding situation between the two TPM realized […],” and, finally, “Such compounds can be realized halves in tBu-12, especially considering the influence of by shifting the energy balance in favor of attractive dispersion intramolecular LD interactions. As Vreven and Morokuma put interactions that outweigh to a large degree the repulsive it 14 years earlier: “A key question in HPE chemistry is the dispersion contributions leading to C C bond elongation.” binding energy of the symmetric dimer”.[69] Remarkably, the This was very true for the isolable all-meta tBu HPE. This was hindered rotation of the phenyl rings in tBu-12 leads to a the starting point for Grimme and Schreiner reconsidering the distinctly different chemical environment for the “off-ipso- work of Dames et al. to answer the ultimate question why para-axis” moieties. In the NMR spectrum, the ortho-hydro- especially all-meta tBu TPM, but not all-para tBu TPM or gens as well as the meta-tBu groups split into doublets, TPM itself α,α-dimerizes: “How can the derivative of a whereby the former experience a large 1038 Hz and the latter molecule that dissociates owing to steric hindrance become a smaller 18 Hz split. We observed a chemical shift of δ= stable by increasing steric bulk?”[99] At TPSS-D3/TZV(2d,2p) 71.5 ppm (C6D12) for the central α-carbons, in line with (including LD “D3”-corrections) HPE is computed to sponta- Rieker’s 70.7 ppm[62] (C6D6) and the extrapolation of the α-13C neously dissociate (ΔG 298d = 9 kcalmol 1) and recombine as shifts of 1,1,1-triphenylethane, 1,1,1,2-tetraphenylethane and JNS. A stability gain via the tBu in all-para tBu HPE (112) is 1,1,1,2,2-pentaphenylethane (69(3) ppm[103]).[104] “practically absent”, the molecule also spontaneously disso- To obtain the first experimental value for the free ciates (ΔG 298d = 7 kcalmol 1) but para substitution prevents dissociation energy of tBu-12 and probe the earlier computa- JNS formation and therefore all-para tBu TPM is fully tional predictions, we performed variable temperature NMR dissociated and persistent as a radical. experiments. The corresponding van ‘t Hoff plot showed a Contradictory to Mislow’s statement that “the tert-butyl strong temperature dependence on the tBu-12 and tBu-1 groups have no special effect on the bonding parameters of concentrations and gave ΔG 298d = 1.60(6) kcalmol 1, in good hexa(2,6-di-tert-butyl-4-biphenylyl)ethane”,[64] the computa- agreement with true computational predictions of Grimme and tional analysis revealed that the major reason for the stability Schreiner of 2011.[99] Theory had reached a stage where it of tBu-1 (ΔG 2982 d = +13.7 kcalmol 1) is LD between the tBu could make a bold prediction that could be tested experimen- groups (Edisp=40 kcalmol 1).[100] More elaborate PWPB95- tally and confirmed to be correct. D3/QZVP(g,f) computations with solvent corrections predicted A computational study (B3LYP-D3(BJ)/cc-pVDZ) on all- (in the true sense of the word, i. e., foresaw) a ΔG 298d in the meta alkyl substituted HPE derivatives (R-12) – not corrected range of 3 to +1 kcalmol 1. Further, the dissociation curve for the possible dynamic transitions of close lying low energy of tBu-12 revealed a second minimum, revealing a vdW structures due to rotation of lower symmetry alkyl moieties – complex (tBu-1)2 similar to the Schlenk radicals.[101] Grimme revealed a good correlation between the polarizability α, hence Figure 12.Molecular structures of HPE H-12, all-para tBu HPE 11 , all-meta t2 Bu HPE tBu-12 and the vdW triplet complex of two all-meta tBu TPM radicals (tBu-1)2. Isr. J. Chem. 2022, 62, e202200002 (11 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective Figure 13. Correlation between the increasing free dissociation energies (ΔG 298d ) of R-12 to 2 R-1 determined by B3LYP-D3(BJ)/cc-pVDZ and the increasing computed polarizabilities (α) with substituent size is found (R2=0.98). The opposite, but slightly worse correlation is found for dispersion-uncorrected B3LYP/cc-pVDZ computations. The grey area indicates the Edisp acting between the molecular moieties. correlating LD interactions as well, and ΔG 298d (Figure 13). The increase in LD interactions was also visually confirmed by NCI[105] plots. These derivatives therefore might serve as excellent molecular balances[106] to experimentally probe LD between the meta alkyl moieties presumably acting as DEDs.[99] In the following year, we found all-meta tBu triphenyl- methane (12), formally the hydrogenation product of tBu-12, to crystallize in a head-to-head dimeric arrangement 122 (Fig- ure 14).[107] The SC-XRD featured a short C H···H C contact of the central C H moiety. Using neutron diffraction (NRD) measurement of a large crystal (3.6×4.4×4.6 mm) of 12 at temperatures as low as 20 K confirmed the hydrogen atom positions in 122 accurately. The NRD structure revealed an RH···H of 1.566(5) Å – the shortest C H···H C contact reported to date. This head-to-head arrangement was recently also confirmed via ionization loss stimulated Raman spectroscopy in molecular beam experiments to resolve structure sensitive vibrations. Hence, the arrangement is not result of crystal packing and as it prevails in the gas phase,. The head-to-head arrangement is maintained even under isolated molecular beam conditions in the absence of packing effects of the solid state. The peculiar head-to-head arrangement therefore must result from extraordinarily strong LD interactions. The central Figure 14. Top: Skeletal structures of dimeric 122 (left) with its short Raman-active aliphatic C D vibration of methine-deuterated H···H contact and former record holder “Half-Cage” 13 (right, R1: tBu-1 is associated with an unusually short C D···D C OBn).[110]2 Bottom: Hits for H···H contacts determined by NRD in the distance as revealed by the strong blue-shift compared to the Cambridge Crystal Structure Database (V5.38+2 Updates) increase unperturbed case. As a counterexample, the dimer of unsub- quickly starting from distances around the sum of the van-der-Waals stituted triphenylmethane displays an approximately S -sym- (vdW) radii of hydrogen. The high energetic range below R H···H R6 metric tail-to-tail arrangement. <1.9 Å contains only seven non-bonded H···H contacts in sixmolecules revealing the difficulties to obtain such short contacts. The massive mutual penetration of the electron density – Reprinted from the Annual Report of ILL 2017, Institut Max von Laue the contact is 35% shorter than the sum of their vdW radii - Paul Langevin (ILL). (2.4 Å[108]) and more than half way to covalent H H – causes large Pauli repulsion, which must predominantly be counter- balanced by the only available attractive interaction, namely Recently, we expanded our efforts and synthesized a series LD. It was shown that the tBu groups again act as DEDs and of all-meta hydrocarbyl substituted TPM radicals (R-1; R= contribute decisively to the compression of the vdW complex. Me, iPr, tBu, Cy, Ph, Adamantyl=Ad) to experimentally This peculiar dimer gained widespread attention.[109] corroborate our previous computations on the stability of the Isr. J. Chem. 2022, 62, e202200002 (12 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective corresponding HPE derivatives (R-12).[111] Over the course of derivatives can be generated, which also gain stability by these studies we also synthesized and characterized the dispersion forces.”[114] alcohols [TPM-OH (R-14)], hydrocarbons [TPM-H (R-15)], Yet, there is plenty of evidence of the importance of LD halides [TPM-X (R-16)] and peroxides [(TPM-O)2 (R-17)]. regarding the stability of 102 and tBu-12. Corminbœuf and Several head-to-head dimers akin to 122 were found in SC- Sherrill tested their intramolecular symmetry adapted perturba- XRD for alcohols R-14 and hydrocarbons R-15. The hydro- tion theory (ISAPT) on H-12 and tBu-1 .[115]2 They performed carbon dimers Me-152, 12 (= t2 Bu-152), and Cy-152 feature an energy decomposition and used it for comparison of these short C H···H C contacts. In Me-152 the contact is only 13% two molecules at fixed C C bond distances of rCC(H-12)= shorter than the sum of the vdW radii (2.4 Å[108]), in Cy-152, 1.713 Å or r tCC(Bu-12)=1.661 Å and found that in both cases the contact is even 24% shorter.[112] The record holder 122 (short and long rCC) the gain in dispersion overcomes the displays a distance which is decreased by 35%. This supports repulsion and therefore tBu-12 is more stable. Meitei and the proficiency of the tBu moiety as a DED. Size, shape, and Heßelmann used H-12 and tBu-12 as test systems for their fitting effects where indicated by NCI plots of the molecular incremental molecular fragmentation (IMF) method and structures of R-17. obtained results very similar to those of Grimme and A large energetic deviation from the initial qualitative Schreiner, and Dames et al.,[116] namely, that the gain in LD computational estimations to experiment compelled us to from parent to substituted HPE is larger than the increase in expand and improve the computations. The application of a repulsion. Ackermann, Breugst, and coworkers point out: “For triple-ζ basis set reduced superposition errors and a possible a long time, large and bulky substituents have intuitively been vdW radical complex (R-1)2 as well as the JNS (R-2) were considered to act through unfavorable steric interactions, taken into account. A ΔG 298d of 1.8 kcalmol 1 for tBu-12 was although London dispersion – the attractive part of the van- determined, in excellent agreement with experiment. der-Waals interaction – is known for more than 100 years. The Besides the known tBu-12, we were able to identify Ad-12 stabilizing nature of C H···H C interactions and their via NMR by induction. Because the radical peak of Ad-1 was importance for organic transformations has only been fully obscured, a mathematical solution was established to directly realized within the last decades. Among others, these inter- determine ΔH 298d via the temperature dependence of solely the actions explain the hexaarylethane riddle and [the] very short Ad-12 peaks; ΔG 298d was obtained a posteriori. The method H···H contacts in tris(3,5-di-tert-butylphenyl)methane.”[117] was validated with known tBu-12 and showed reasonable Ultimately, Boeré et al. state very appropriately: agreement. Ad-12 is more stable (ΔG 298d =2.1(6) kcalmol 1), “[…]Dispersion interactions […] between 3,5-di-tert-butyl- qualitatively in line with our computational predictions and on phenyl groups […] supplying more energy towards holding the the basis that Ad is larger and more polarizable than tBu. The sterically challenged all-meta tert-butyl hexaphenylethane deviations between experiment and computations were attrib- together than the C C covalent bond.”[118] uted mostly to having to approximate the solution phase with a continuum solvent model. As no other R-12 was observed, only tBu-12 and Ad-12 are stabilized enough by their Proclaimed Syntheses of Hexaphenylethane substituents. Because tBu and, even more so, Ad are expected to act as excellent DEDs and computations show a decisive It is clear that none of the scientist who claimed to have role of LD in these R-1 , we must conclude that LD indeed synthesized HPE before Nauta’s publication[54]2 in 1968 could governs the stability in all-meta hydrocarbyl hexaphenyl- have known to have synthesized in fact another structure. ethanes. This realization should have consequences in the way However, there are several proclaimed syntheses of parent we look at many other crowed structures and it forces us to HPE after 1968 without convincing or even fabricated reconsider many of the common “steric strain” interpretations spectroscopic evidence. Common to these publications is that to interpret structures and mechanisms. they synthesized HPE en passant without giving any This view had not been generally accepted before our indication of the importance to the synthesis of such a work appeared. Allinger tested his MM4(2015) molecular historically important molecule. mechanics force field with overcrowded molecules including In this regard, Alper and Prince mentioned HPE in 1980 in 11 .[113]2 While it is recognized “[…]that the forces that a very short communication about the mild desulfurization of determine these bond lengths [are] i. e., stretching, bending, trityl thiol.[119] As a byproduct a dimer written as “RR” is torsional, van der Waals, and electrostatic […]” it follows that obtained in 8% yield of isolated product! This could also “Long C C bonds are […] mainly the result of steric e ects correspond to the JNS (which still would be very sensitive to […].” It is surprising that the question was not asked why handle). But the CAS No.1117854-07-8 named “RR” in the these molecules exist after all, not to speak of an explanation follow-up publication by Alper three years later[120] corre- what stabilized them. Suzuki et al. retained the notion that sponds to HPE that was produced in 38% yield. Neither steric shielding of the para-positions in HPE is most important publication comments on the “HPE” synthesis nor is a for the synthesis of all-meta tBu HPE, with LD is only an add- spectroscopic analysis provided. In another mild desulfuriza- on: “When the formation of the JNS is prevented by the tion in 1998 by Yu and Verkade no room for doubt was left as attachment of bulky substituents on the aryl moieties, HPE “Substrate 16 underwent a similar reaction to generate Isr. J. Chem. 2022, 62, e202200002 (13 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective triphenylmethane and hexaphenylethane”, the latter in 27% manipulations as some apparently unwanted peaks were yield.[121] Again, no analysis of the product was given. The simply cut. This was discovered and communicated to us by latter procedure was used to synthesize this “HPE” and found Robert Mayer from the Ludwig-Maximilians-Universität Mün- its way into the literature again in “Mesolysis of Radical chen who submitted a comment of Liu’s paper to the same Anions of Tetra-, Penta-, and Hexaphenylethanes” by Majima journal. The handling editor did not allow this comment to be in 2013.[122] The bond dissociation process of this “HPE” is published and instead urged us to contact the authors directly. broadly discussed and in the experimental section is stated that Ultimately a corrigendum was published which corrects the “1,1,1,2,2-pentaphenylethane (Ph5E) and 1,1,1,2,2,2-hexaphe- HPE structure (H-12) to Ullmann’s hydrocarbon 4.[127] The nylethane (Ph6E) are known compounds ”. Incidentally, the corrigendum finishes with a remarkable comment: “The quoted reference for the synthesis of these two does not spectra editing does not affect the integrity of the research and mention HPE.[123] conclusions of the published paper.” We learn that for some Rheingold and Trogler published the synthesis of azidobis- authors and journal editors it now is apparently acceptable to (pentamethylcyclopentadienyl)-vanadium(III) [η5- edit NMR spectra to remove unwanted peaks. C5Me ) 55 2VN3] by treatment of (η -C5Me5)2V with sterically While the synthesis was unproven proclaimed several hindered azides.[124] “For the case of N3CPh3, Pasteur times after 1968, to the best of our knowledge, parent H-12 separation and IR analysis (comparison with authentic remains elusive. samples) proved the blue crystals to be [η5-C5Me5)2VN3] and the pale-yellow crystals were Ph3CCPh3.” No spectral data were provided. Similarly, “HPE” was obtained in a three-step Outlook on Making Hexaphenylethane procedure in 9% yield as a side product in the construction of tetrasubstituted carbons from carbonyl compounds with low- The synthesis of HPE is still an open challenge. Its story is valent vanadium complexes.[125] Again, no spectral proofs one of serendipity, wishful thinking, and human failure. Some were provided. synthetic attempts beyond a Wurtz type coupling of trityl The most recent claim of having made HPE was published halides were performed by Anschütz[128] and Kahr.[129] An- in 2017 by Liu et al..[126] In this publication Table 2, entry 16, schütz thermally extruded CO2 from the ester 18 (Figure 15a), depicts the synthesis of HPE by a nickel catalyzed reductive but most likely observed Ullmann’s hydrocarbon 4.[65] Kahr coupling. Most remarkably, the melting point as well as NMR attempted to make use of the crystal structure of trityl iodide, data were provided in the supporting information: “white which crystallizes in a shifted head-to-head fashion. Solid state solid; m.p. 230–233 °C (lit. 2227[recte 222.7]-230 °C);10 1H decomposition was expected to lead to HPE via a simple NMR (500 MHz, CDCl3) δ 6.97–7.30 (m, 30H); 13C NMR Walden inversion (Figure 15c). Unfortunately, thermal decom- (125 MHz, CDCl3) δ 131.2, 129.4, 128.4, 128.2, 127.3, 126.2, position started only beyond 100 °C and HPE could not be 125.8, 56.4.” While the 1H NMR chemical shifts appear identified. Lewars also named further extrusion precursors,[65] plausible, the number of 13C signals is dubious. Named are but the problem of a rearrangement of labile H-12 to the JNS 2 eight, while five or seven are expected, depending on whether under common reaction conditions remains. there is free rotation of the phenyl groups or not, respectively. A way to exclude rearrangements and facilitate the Further, the chemical shift of 56.4 ppm appears to be shifted isolation of H-12 would perhaps be encapsulation. At the upfield relative to tetra- and penta-phenylated ethanes.[104] conditions of thermal extrusion of CO2 in solution, more than Also, the reference for the melting point is given as “[10] R. enough energy is present in the system to split, dissociate and binaghi[sic!], Gazzetta Chimica Italiana 1923, 53, 879–887” – rearrange H-12 to 2. If 18 is volatile enough to be evaporated which clearly does not describe HPE. and isolated in, e.g., a noble gas matrix under cryogenic But the story continues. The corresponding HPE NMR conditions, extrusion could be induced by irradiation. Full spectra in the Supplementary Information revealed crude dissociation of the reaction products is excluded by the solid Figure 15. Synthetic attempts at hexaphenylethane by a) thermal extrusion, b) reductive coupling by metals or c) Walden inversion. Isr. J. Chem. 2022, 62, e202200002 (14 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective matrix. The change in shape and size during the extrusion is positions, however, can be spectroscopically identified even in small it should proceed readily. Nonetheless, the large shape solution, parent HPE is, in a cheeky way, “not sterically change from H-12 to 2 might impede the rearrangement crowded enough”, thereby lacking the positive aspects of large (Figure 16). The low temperatures as well as the fast relaxation bulky groups: London dispersion. This key aspect was of hot states should promote the bound HPE state. Extrusion proposed by Grimme and Schreiner on the basis of DFT of N2, CO, or SO2 from bis(trityl)diazene 19, hexaphenylace- computations with and without dispersion corrections, ulti- tone 20, or bis(trityl)sulfone 21, respectively, would proceed mately leading to the concept of dispersion energy donors in the same fashion, but these are themselves elusive.[130] (DEDs). The unknown 19 might be stable in an LD-shell akin the We were able to measure experimentally the dissociation stable all-meta alkyl substituted bis(trityl)peroxide derivatives. energy of tBu-12. This was the necessary measure to validate The weak O O bond is stabilized by the surrounding DEDs computational methods which reveal the divided contributions against Wieland[131] rearrangement. Similarly, the labile of inter- or intramolecular interactions. In remarkable con- N=N fragment would experience stabilization. fluence of experiment and theory, we demonstrated that LD This stabilizing effect of the LD-shell could also lead to a interactions between the tBu groups are the decisive contribu- quantification of LD in all-meta alkyl substituted tion counterbalancing the repulsive forces responsible for the bis(trityl)peroxides or bis(trityl)diazenes. The onset of the spontaneous dissociation of HPE. Most importantly, we Wieland rearrangement or N2 extrusion should depend on the demonstrated the strength of these LD interactions with the strength of the LD interactions between the substituents all-meta tert-butyl triphenylmethane dimer, in which HPE’s similar to R-12. Lastly, EPR investigations of the radical central C C bond is formally substituted by a linear and solutions of R-1 in subcooled toluene similar to experiments extremely short C H···H C arrangement. Not only does the by Broser[101] might reveal the dimeric vdW complexes (R-1)2 overall structure of tBu-1 survive, but the central C H···H C and provide evidence for the predicted “bond length group is compressed to the shortest H···H contact reported to isomerism”.[99] date (1.567 Å) – without any external help! Our spectroscopic investigations of a series of alkyl substituted triphenylmethyl derivatives excluded stereoelectronic substituent effects as a Conclusions cause for the increased stability of tBu-12. tBu-12 and Ad-12 are in equilibrium with their radical Hexaphenylethane has a multifarious history, but has remained monomers in solution. Therefore, tBu-12 and Ad-12 are elusive to date primarily because of its instability due to the thermodynamically more stable against dissociation than other steric clashes between the trityl moieties and the resulting R-12 species and it becomes evident that large, spherical and weak central C C bond. As very crowded HPE derivatives rigid substituents act as excellent DEDs,[99] even in solution. such as those with tBu and adamantyl groups in the all-meta These hexaphenylethanes can be used as models for thermody- Figure 16. Comparison of the shape change during the extrusion of a) N2 from 19 (Ph3C N=N CPh3) or b) CO2 from 18 (Ph3CC(O)OCPh3) to yield H-12 and the rearrangement of the latter to 2. Isr. J. Chem. 2022, 62, e202200002 (15 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective namic LD stabilization and, in this regard, the obtained [12] F. Ullmann, W. Borsum, Ber. Dtsch. Chem. Ges. 1902, 35, insights – the importance of size, fit, and rigidity to maximize 2877–2881. the DED efficiency to stabilize peculiar structures – might be [13] J. F. Norris, W. W. Sanders, Am. Chem. J. 1901, 25, 54–62. transferred to other fields like inorganic coordination [14] As there was a lot of confusion about the structure of this chemistry,[132] supermolecular chemistry,[133] and synthetic compound in the years before 1968, we will refer to the [134] compound as “triphenylmethyl” in quotation marks. Theorganic chemistry, in particular, catalysis. molecular structures (after 1968 also the compounds) will The HPE story reveals the importance of synergy between nonetheless be named without quotation marks as experiment and theory: While Gomberg trusted his experi- triphenylmethyl radical (TPM; H-1), hexaphenylethane (HPE; ments more than the best available theory at the time, the H-12) and 1-diphenylmethylen-4-triphenylmethyl- reverse is true for Mislow who eventually trusted multiple cyclohexa(2,5)diene (JNS; 2). computations more than the experimental results, thereby [15] M. Gomberg, J. Am. Chem. Soc. 1901, 23, 109–110. prompting a re-determination of a critical X-ray structure. [16] M. Gomberg, J. Am. Chem. Soc. 1901, 23, 496–502. Finally, experiment and theory agreed. [17] M. Gomberg, Ber. Dtsch. Chem. Ges. 1902, 35, 3914–3920.[18] A. E. Tschitschibabin, Ber. Dtsch. Chem. Ges. 1904, 37, 4709– Subsequently, Grimme and Schreiner proposed on the basis 4715. of dispersion-corrected DFT computations that the bulky [19] P. Jacobson, Ber. Dtsch. Chem. Ges. 1905, 38, 196–199. substituents in tBu-12 are responsible for its stability through [20] a) E. Heintschel, Ber. Dtsch. Chem. Ges. 1903, 36, 320–322; mutual LD attraction. This gave birth to the concept of b) E. Heintschel, Ber. Dtsch. Chem. Ges. 1903, 36, 579–579. dispersion energy donors (DEDs) that are now increasingly [21] F. Kehrmann, F. Wentzel, Ber. Dtsch. Chem. Ges. 1901, 34, used as strategic design elements in catalytic processes (vide 3815–3819. supra). It is only a question of (long?) time until textbooks [22] M. Gomberg, Ber. Dtsch. Chem. Ges. 1902, 35, 2397–2408. will also begin to appreciate a quantum-chemical based [23] M. Gomberg, Ber. Dtsch. Chem. Ges. 1902, 35, 1822–1840.[24] M. Gomberg, Ber. Dtsch. Chem. Ges. 1901, 34, 2726–2733. viewpoint on effects that are absolutely essential for under- [25] M. Gomberg, L. H. Cone, Ber. Dtsch. Chem. Ges. 1904, 37, standing structures and chemical reactivities. 2033–2051. [26] a) K. Auwers, Ber. Dtsch. Chem. Ges. 1907, 40, 2154–2159; b) B. Flürscheim, J. Prakt. Chem. 1904, 71, 497–539; c) A. E. Acknowledgments Tschitschibabin, Ber. Dtsch. Chem. Ges. 1905, 38, 771–773; d) M. Gomberg, L. H. Cone, Ber. Dtsch. Chem. Ges. 1905, 38, This work was generously supported by the Deutsche 2447–2458. Forschungsgemeinschaft, Priority Program “Dispersion” (SPP [27] M. Gomberg, L. H. Cone, Ber. Dtsch. Chem. Ges. 1906, 39,1461–1470. 1807). Open Access funding enabled and organized by Projekt [28] M. Gomberg, D. D. V. Slyke, J. Am. Chem. Soc. 1911, 33, 531– DEAL. 549. [29] M. Gomberg, Ber. Dtsch. Chem. Ges. 1907, 40, 1847–1888. [30] a) J. M. McBride, Tetrahedron 1974, 30, 2009–2022; b) K. J. Skinner, H. S. Hochster, J. M. McBride, J. Am. Chem. Soc. 1974, 96, 4301–4306. References [31] J. F. Norris, L. R. Culver, Am. Chem. J. 1903, 29, 129–140. [32] “The reaction“ refers to the autooxidation of “triphenylmethyl“. [1] a) M. Gomberg, Ber. Dtsch. Chem. Ges. 1900, 33, 3150–3163; [33] a) K. Ziegler, L. Ewald, Liebigs Ann. Chem. 1933, 504, 162– b) M. Gomberg, J. Am. Chem. Soc. 1900, 22, 757–771. 181; b) R. C. Mithoff, G. E. K. Branch, J. Am. Chem. Soc. 1930, [2] a) J. J. C. Bailar, Biogr. Mem. 1970, 141–172; b) C. Rüchardt, 52, 255–268. Nachr. Chem. 2000, 48, 904–910. [34] H-12 or 2. [3] A. J. Ihde, Pure Appl. Chem. 1967, 15, 1–14. [35] J. Schmidlin, Ber. Dtsch. Chem. Ges. 1908, 41, 2471–2479. [4] A.-L. Lavoisier, Elements of Chemistry, Dover Press, New [36] B. Flürscheim, Ber. Dtsch. Chem. Ges. 1908, 41, 2746–2747. York, 1965. [37] H. Wieland, Ber. Dtsch. Chem. Ges. 1909, 42, 3020–3030. [5] This was before Gerhardt and finally Cannizzaro (1860) [38] a) W. Schlenk, T. Weickel, Liebigs Ann. Chem. 1909, 368, 295– corrected the wrong mass determined by Berzelius between 304; b) W. Schlenk, T. Weickel, A. Herzenstein, Liebigs Ann. 1808 and 1818. Here, H2 was counted by mass 1 so that C had Chem. 1910, 372, 1–20. an atomic mass of 6, sulphur of 8, etc. The given chemical [39] W. Schlenk, A. Herzenstein, T. Weickel, Ber. Dtsch. Chem. Ges. formula would then correspond to C2H4. All chemical formulas 1910, 43, 1753–1758. in quotation marks correspond to wrong pre-Cannizzaro [40] J. Piccard, Liebigs Ann. Chem. 1911, 381, 347–351. formulas. [41] M. Gomberg, Ber. Dtsch. Chem. Ges. 1913, 46, 225–228. [6] A. Kekulé, Liebigs Ann. Chem. 1857, 104, 129–150. [42] M. Gomberg, J. Am. Chem. Soc. 1914, 36, 1144–1170. [7] A. Kekulé, Liebigs Ann. Chem. 1858, 106, 129–159. [43] J. Schmidlin, Chemie in Einzeldarstellungen, VI. Bd.: Das [8] M. Gomberg, Ber. Dtsch. Chem. Ges. 1897, 30, 2043–2047. Triphenylmethyl., F. Enke, Stuttgart, 1914. [9] Less phenylated methanes yield upon nitration acidic benzylic [44] G. N. Lewis, Proc. Nat. Acad. Sci. 1916, 2, 586–592. hydrogens, which can be deprotonated by aqueous KOH. The [45] M. Gomberg, Chem. Rev. 1924, 1, 91–141. resulting carbanions are colored. [46] K. Ziegler, L. Ewald, Liebigs Ann. Chem. 1929, 473, 163–193. [10] J. F. Norris, W. W. Sanders, Am. Chem. J. 1901, 25, 117–123. [47] E. Müller, I. Müller-Rodloff, Liebigs Ann. Chem. 1936, 521, [11] V. W. Markownikoff, J. Russ. Phys. Chem. Soc. 1902, 34, 140. 89–94. Isr. J. Chem. 2022, 62, e202200002 (16 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective [48] W. P. Neumann, W. Uzick, A. K. Zarkadis, J. Am. Chem. Soc. [76] F. London, Trans. Faraday Soc. 1937, 33, 8b–26. 1986, 108, 3762–3770. [77] a) F. London, Z. Phys. B 1930, B11, 222–251; b) F. London, Z. [49] M. F. Roy, C. S. Marvel, J. Am. Chem. Soc. 1937, 59, 2622– Phys. 1930, 63, 245–279. 2625. [78] C. A. Hunter, Angew. Chem. Int. Ed. Engl. 1993, 32, 1584– [50] D. Dünnebacke, W. P. Neumann, A. Penenory, U. Stewen, 1586. Chem. Ber. 1989, 122, 533–535. [79] a) P. Hobza, H. L. Selzle, E. W. Schlag, J. Phys. Chem. 1996, [51] K. Ziegler, Angew. Chem. 1949, 61, 168–179. 100, 18790–18794; b) C. F. R. A. C. Lima, M. A. A. Rocha, [52] C. Koukotas, S. P. Mehlman, L. H. Schwartz, J. Org. Chem. L. R. Gomes, J. N. Low, A. M. S. Silva, L. M. N. B. F. Santos, 1966, 31, 1970–1971. Chem. Eur. J. 2012, 18, 8934–8943; c) W. L. Jorgensen, D. L. [53] G. Wittig, W. Schoch, Liebigs Ann. Chem. 1971, 749, 38–48. Severance, J. Am. Chem. Soc. 1990, 112, 4768–4774. [54] H. Lankamp, W. T. Nauta, C. MacLean, Tetrahedron Lett. 1968, [80] J. P. Wagner, P. R. Schreiner, Angew. Chem. Int. Ed. 2015, 54, 9, 249–254. 12274–12296; Angew. Chem. 2015, 127, 12446–12471. [55] H. A. Staab, H. Brettschneider, H. Brunner, Chem. Ber. 1970, [81] W. Kohn, Rev. Mod. Phys. 1999, 71, 1253–1266. 103, 1101–1106. [82] a) K. Müller-Dethlefs, P. Hobza, Chem. Rev. 2000, 100, 143– [56] J. Allemand, R. Gerdil, Acta Crystallogr. Sect. B 1978, 34, 168; b) S. Grimme, J. Antony, T. Schwabe, C. Mück-Lichten- 2214–2220. feld, Org. Biomol. Chem. 2007, 5, 741–758; c) Y. Zhao, D. G. [57] N. S. Blom, G. Roelofsen, J. A. Kanters, Cryst. Struct. Commun. Truhlar, Theor. Chem. Acc. 2008, 120, 215–241; d) P. Hobza, K. 1982, 11, 297–304. Müller-Dethlefs, in Non-Covalent Interactions: Theory and [58] H. G. Aurich, F. Baer, Chem. Ber. 1968, 101, 1770–1782. Experiment, The Royal Society of Chemistry, 2009, p. 47pp. [59] K. Schreiner, A. Berndt, F. Baer, Mol. Phys. 1973, 26, 929–939. [83] S. Grimme, A. Hansen, J. G. Brandenburg, C. Bannwarth, [60] H. Sakurai, H. Umino, H. Sugiyama, J. Am. Chem. Soc. 1980, Chem. Rev. 2016, 116, 5105–5154. 102, 6837–6840. [84] a) O. Takahashi, Y. Kohno, M. Nishio, Chem. Rev. 2010, 110, [61] M. Stein, A. Rieker, Tetrahedron Lett. 1975, 16, 2123–2126. 6049–6076; b) M. S. Ahlquist, P. O. Norrby, Angew. Chem. Int. [62] a) M. Stein, W. Winter, A. Rieker, Angew. Chem. Int. Ed. 1978, Ed. 2011, 50, 11794–11797; Angew. Chem. 2011, 123, 11998– 17, 692–694; b) M. Stein, W. Winter, A. Rieker, Angew. Chem. 12001; c) D. J. Liptrot, P. P. Power, Nat. Chem. Rev. 2017, 1, 1978, 90, 737–738. 0004; d) D. J. Liptrot, J. D. Guo, S. Nagase, P. P. Power, Angew. [63] W. D. Hounshell, D. A. Dougherty, J. P. Hummel, K. Mislow, J. Chem. Int. Ed. 2016, 55, 14766–14769; Angew. Chem. 2016, Am. Chem. Soc. 1977, 99, 1916–1924. 128, 14986–14989. [64] E. Osawa, Y. Onuki, K. Mislow, J. Am. Chem. Soc. 1981, 103, [85] a) K. Szalewicz, WIREs Comput. Mol. Sci. 2011, 2, 254–272; 7475–7479. [65] E. G. Lewars, in Modeling Marvels: Computational Anticipation b) G. Jansen, WIREs Comput. Mol. Sci. 2013, 4, 127–144. of Novel Molecules (Ed.: E. G. Lewars), Springer Netherlands, [86] a) S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. Dordrecht, 2008, pp. 115–129. 2010, 132, 154104–154119; b) S. Grimme, WIREs Comput. [66] B. Kahr, D. v. Engen, K. Mislow, J. Am. Chem. Soc. 1986, 108, Mol. Sci. 2011, 1, 211–228. 8305–8307. [87] W. B. Schneider, G. Bistoni, M. Sparta, M. Saitow, C. Riplinger, [67] N. Yannoni, B. Kahr, K. Mislow, J. Am. Chem. Soc. 1988, 110, A. A. Auer, F. Neese, J. Chem. Theory Comput. 2016, 12, 6670–6672. 4778–4792. [68] https://www.ccdc.cam.ac.uk/structures/. [88] C. A. Hunter, Angew. Chem. Int. Ed. 2004, 43, 5310–5324; [69] T. Vreven, K. Morokuma, J. Phys. Chem. A 2002, 106, 6167– Angew. Chem. 2004, 116, 5424–5439. 6170. [89] S. Paliwal, S. Geib, C. S. Wilcox, J. Am. Chem. Soc. 1994, 116, [70] M. Karplus, G. K. Fraenkel, J. Chem. Phys. 1961, 35, 1312– 4497–4498. 1323. [90] L. X. Yang, C. Adam, G. S. Nichol, S. L. Cockroft, Nat. Chem. [71] a) X. Creary, Acc. Chem. Res. 2006, 39, 761–771; b) T. H. 2013, 5, 1006–1010. Fisher, S. M. Dershem, M. L. Prewitt, J. Org. Chem. 1990, 55, [91] E.-i. Kim, S. Paliwal, C. S. Wilcox, J. Am. Chem. Soc. 1998, 1040–1043; c) R. A. Jackson, R. Moosavi, J. Chem. Soc. Perkin 120, 11192–11193. Trans. 2 1992, 885–888; d) W. Adam, H. M. Harrer, F. Kita, H.- [92] J. Hwang, B. E. Dial, P. Li, M. E. Kozik, M. D. Smith, K. D. G. Korth, W. M. Nau, J. Org. Chem. 1997, 62, 1419–1426; Shimizu, Chem. Sci. 2015, 6, 4358–4364. e) A. M. d. P. Nicholas, D. R. Arnold, Can. J. Chem. 1986, 64, [93] J. M. Schümann, J. P. Wagner, A. K. Eckhardt, H. Quanz, P. R. 270–276. Schreiner, J. Am. Chem. Soc. 2020, 143, 41–45. [72] E. Dames, B. Sirjean, H. Wang, J. Phys. Chem. A 2010, 114, [94] R. Pollice, M. Bot, I. J. Kobylianskii, I. Shenderovich, P. Chen, 1161–1168. J. Am. Chem. Soc. 2017, 139, 13126–13140. [73] Unfortunately, Dames was unable to disentangle the early [95] a) A. D. McLachlan, Discuss. Faraday Soc. 1965, 40, 239–245; confusions in the HPE riddle and discussed a 1951 “HPE” b) K. D. Shimizu, Nat. Chem. 2013, 5, 989–990; c) C. Zhao, P. BDE[73] in comparison to a 1986 JNS BDE[48] and found: Li, M. D. Smith, P. J. Pellechia, K. D. Shimizu, Org. Lett. 2014, “The only available gaseous phase experimental value for the 16, 3520–3523; d) L. Schweighauser, M. A. Strauss, S. Bellotto, BDE of HPE comes from Szwarc. They reported a BDE value H. A. Wegner, Angew. Chem. Int. Ed. 2015, 54, 13436–13439; of 11�2 kcalmol 1 […]. Admittedly, [this] value is close to Angew. Chem. 2015, 127, 13636–13639. that of the intradimer C C BDE in [JNS]. That C C BDE is [96] K. Fumino, V. Fossog, P. Stange, D. Paschek, R. Hempelmann, incidentally also around 11 kcal/mol.”. R. Ludwig, Angew. Chem. Int. Ed. 2015, 54, 2792–2795; Angew. [74] R. Eisenschitz, F. London, Z. Phys. 1930, 60, 491–527. Chem. 2015, 127, 2834–2837. [75] The direct diagonal elements are attributed to a short range (~ [97] S. Grimme, J.-P. Djukic, Inorg. Chem. 2011, 50, 2619–2628. e cr(ab)) interaction with an attractive and a repulsive solution. [98] P. R. Schreiner, L. V. Chernish, P. A. Gunchenko, E. Y. Tikhon- This first description of a bonding and antibonding covalent chuk, H. Hausmann, M. Serafin, S. Schlecht, J. E. P. Dahl, interaction was described earlier by London.[75]. R. M. K. Carlson, A. A. Fokin, Nature 2011, 477, 308–311. Isr. J. Chem. 2022, 62, e202200002 (17 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH. Perspective [99] S. Grimme, P. R. Schreiner, Angew. Chem. Int. Ed. 2011, 50, [118] I. D. L. Guimarães, J. R. Garcia, K. Wohnrath, R. T. Boeré, Eur. 12639–12642; Angew. Chem. 2011, 123, 12849–12853. J. Inorg. Chem. 2018, 0, 3606–3614. [100] As B3LYP lacks LD, the dispersion correction (in this case [119] H. Alper, T. L. Prince, Angew. Chem. 1980, 92, 321–321; Grimme’s dispersion correction version 3 [D3]) corresponds Angew. Chem. Int. Ed. 1980, 19, 315–316. well to the dispersion energy (Edisp). [120] H. Alper, S. Ripley, T. L. Prince, J. Org. Chem. 1983, 48, 250– [101] W. Broser, H. Kurreck, W. Niemeier, Tetrahedron 1976, 32, 252. 1183–1187. [121] Z. Yu, J. G. Verkade, Tetrahedron Lett. 1998, 39, 2671–2674. [102] S. Rösel, C. Balestrieri, P. R. Schreiner, Chem. Sci. 2017, 8, [122] S. Tojo, M. Fujitsuka, T. Majima, J. Org. Chem. 2013, 78, 405–410. 1887–1893. [103] A linear regression of (xn: 3, 4, 5 phenyl rings; ym: 53.6, 59.6, [123] W. B. Smith, M. C. Harris, J. Org. Chem. 1983, 48, 4957–4962. 63.7 ppm) yields: δ(nrings)=5.1(5) ·nrings+38(2). This linear [124] J. H. Osborne, A. L. Rheingold, W. C. Trogler, J. Am. Chem. equation gives a chemical shift of 69(3) ppm. This is not Soc. 1985, 107, 7945–7952. reported in the literature. [125] Y. Kataoka, I. Makihira, H. Akiyama, K. Tani, Tetrahedron [104] M. I. Watkins, G. A. Olah, J. Am. Chem. Soc. 1981, 103, 6566– 1997, 53, 9525–9540. 6574. [126] Y. Liu, S. Xiao, Y. Qi, F. Du, Chem. Asian J. 2017, 12, 673– [105] J. Contreras-García, E. R. Johnson, S. Keinan, R. Chaudret, J.-P. 678. Piquemal, D. N. Beratan, W. Yang, J. Chem. Theory Comput. [127] Y. Liu, S. Xiao, Y. Qi, F. Du, Chem. Asian J. 2017, 12, 2873– 2011, 7, 625–632. 2873. [106] I. K. Mati, S. L. Cockroft, Chem. Soc. Rev. 2010, 39, 4195– [128] R. Anschütz, Liebigs Ann. Chem. 1908, 359, 196–201. 4205. [129] R. Arkin, B. Cowans, B. Kahr, Chem. Mater. 1996, 8, 1500– [107] S. Rösel, H. Quanz, C. Logemann, J. Becker, E. Mossou, L. 1503. Canadillas-Delgado, E. Caldeweyher, S. Grimme, P. R. [130] a) S. Sugimori, Nippon Kagaku Zasshi 1971, 92, 426–428; Schreiner, J. Am. Chem. Soc. 2017, 139, 7428–7431. b) W. Schlenk, E. Bergmann, Liebigs Ann. Chem. 1928, 464, 1– [108] R. S. Rowland, R. Taylor, J. Phys. Chem. 1996, 100, 7384– 21. 7391. [131] H. Wieland, Ber. Dtsch. Chem. Ges. 1911, 44, 2550–2556. [109] a) J. Kemsley, Chem. Eng. News 2017, 95, 8; b) D. Yablon, J. [132] a) S. Singha, M. Buchsteiner, G. Bistoni, R. Goddard, A. Am. Chem. Soc. 2017, 139, 7665–7665; c) S. M. Bachrach, Furstner, J. Am. Chem. Soc. 2021, 143, 5666–5673; b) A. http://comporgchem.com, 2017; d) H. Rzepa, http://www.ch. Casitas, J. A. Rees, R. Goddard, E. Bill, S. DeBeer, A. Fürstner, imperial.ac.uk/rzepa/blog, 2017; e) K. Krämer, https://www. Angew. Chem. Int. Ed. 2017, 56, 10108–10113; Angew. Chem. chemistryworld.com, 2017; f) R. van der Lingen, http://chem. 2017, 129, 10242–10247; c) R. J. Witzke, D. Hait, M. Head- vander-lingen.nl, 2017; g) https://www.chemistryviews.org, Gordon, T. D. Tilley, Organometallics 2021, 40, 1758–1764; 2017. d) C. Ganesamoorthy, J. Schoening, C. Wölper, L. Song, P. R. [110] O. Ermer, S. A. Mason, F. A. L. Anet, S. S. Miura, J. Am. Schreiner, S. Schulz, Nat. Chem. 2020, 12, 608–614; e) L. Song, Chem. Soc. 1985, 107, 2330–2334. J. Schoening, C. Wölper, S. Schulz, P. R. Schreiner, Organo- [111] S. Rösel, J. Becker, W. D. Allen, P. R. Schreiner, J. Am. Chem. metallics 2019, 38, 1640–1647; f) J. Bresien, C. Hering- Soc. 2018, 140, 14421–14432. Junghans, P. Kumm, A. Schulz, M. Thomas, A. Villinger, Eur. [112] Determined by SC-XRD. The C H bond length determined by J. Inorg. Chem. 2018, 2018, 647–651. X-ray diffraction generally appears too short because the low [133] a) A. Garcia-Marquez, A. Frontera, T. Roisnel, R. Gramage- refraction of hydrogen atoms. The rCH is therefore usually set Doria, Chem. Commun. 2021, 57, 7112–7115; b) A. Elmi, S. L. to 1.00 Å for aliphatic and to 0.95 Å for aromatic C H bonds to Cockroft, Acc. Chem. Res. 2021, 54, 92–103; c) S. Spicher, S. best fit the electron density map. Instead, an rCH around 1.2 Å Grimme, J. Phys. Chem. Lett. 2020, 11, 6606–6611; d) M. is determined by neutron diffraction, which does not suffer from Giese, M. Albrecht, ChemPlusChem 2020, 85, 715–724; e) A. this drawback. Poblotzki, H. C. Gottschalk, M. A. Suhm, J. Phys. Chem. Lett. [113] N. L. Allinger, J.-H. Lii, H. F. Schaefer, J. Chem. Theory 2017, 5656–5665. Comput. 2016, 12, 2774–2778. [134] a) G. Lu, R. Y. Liu, Y. Yang, C. Fang, D. S. Lambrecht, S. L. [114] Y. Uchimura, T. Takeda, R. Katoono, K. Fujiwara, T. Suzuki, Buchwald, P. Liu, J. Am. Chem. Soc. 2017, 139, 16548–16555; Angew. Chem. Int. Ed. 2015, 54, 4010–4013; Angew. Chem. b) C. Eschmann, L. Song, P. R. Schreiner, Angew. Chem. Int. 2015, 127, 4082–4085. Ed. 2021, 60, 4823–4832; Angew. Chem. 2021, 133, 4873– [115] R. M. Parrish, J. F. Gonthier, C. Corminboeuf, C. D. Sherrill, J. 4882. Chem. Phys. 2015, 143, 0511031–0051135. [116] O. R. Meitei, A. Heßelmann, J. Comput. Chem. 2017, 38, 2500– 2508. Manuscript received: January 12, 2022 [117] E. Detmar, V. Müller, D. Zell, L. Ackermann, M. Breugst, Revised manuscript received: February 11, 2022 Beilstein J. Org. Chem. 2018, 14, 1537–1545. Version of record online: February 28, 2022 Isr. J. Chem. 2022, 62, e202200002 (18 of 18) © 2022 The Authors. Israel Journal of Chemistry published by Wiley-VCH GmbH.