Ichi-Ei Ishigen,† Kichiken Arishagonos,† Ichi-Shun Hashimura,† Tomutsu Rokimizo,† Jirotama Amayih,† Otokam Adamuk,‡ Ryčard F. Kčeh,‡ Arika Ikuzus† and Günther Schlonk§*
†International Research Consortium for Chemical Innovation, Department of Organometallics, University of Kyoto, Kyoto 739–9110, Japan
‡School of Chemistry, Bohemian Institute of Physical Sciences, Vtipálek 298-01, Czechia
§Ludwig II Institut für Organometallische Forschung, Titz-an-der-Roer, Westphalia 52445, Germany
Abstract: We report the development of a novel C-C decoupling reaction mediated by a palladium/bis(phosphine) catalyst system. This methodology is the first of its kind, and allows for the chemoselective scission of Ar-Ar bonds to give aryl halides and arylboronic esters. Modest to good yields were obtained and the reaction exhibits high functional group tolerance. Notably, our conditions were effective in the late-stage functionalization of a Statin-class HMG-CoA reductase inhibitor. Detailed spectroscopic crystallographic studies were conducted to elucidate the mechanism by which this reaction occurs. These experiments revealed subtle facets of ligand and substrate electrostatics that are vital to the success of the reaction.
INTRODUCTION
Among the myriad Csp2–Csp2 bond forming reactions available to the modern chemist, palladium catalyzed cross-coupling reactions are among the most prevalent.1-3 Their reliability and adaptability have made them a staple in the synthesis of drug candidates and bioactive molecules in the pharmaceutical industry, from the milligram to the megaton scale.4-6 In the pantheon of palladium-catalyzed cross-couplings, the Suzuki– Miyaura reaction is the undisputed king.7-9 This is thanks in large part to the stability, availability and functional group tolerance of the boronic acids and esters that comprise the nucleophilic coupling partner in this reaction. Such has been the success of the Suzuki–Miyaura reaction that (hetero)biaryls and their derivatives are now ubiquitous: more than 10,000 such molecules are available for purchase from fine chemical retailers.10
In stark contrast to the 160,000 publications on the topic of C–C cross-coupling reactions,11 the reciprocal reaction of Csp2–Csp2 bond cleavage has received scant attention to date.12 This is likely because (hetero)biaryls were not a practically accessible starting material until recent decades, and because the fundamental reactivity underpinning this transformation is so challenging. Biaryl Csp2–Csp2 bonds are typically stronger than the C–H, C–O and C–N bonds ubiquitous present in most substrates, making them a challenge to activate selectively.13 Despite the intimidating impediments to this reactivity, a method for selectively cleaving biaryls into aryl halides and aryl boronates would be of great value to the synthetic chemistry community. More than 27% of pesticides and drug candidates contain Csp2–Csp2 linkages.14 If these bonds could be selectively activated, it would provide a powerful strategy for the late-stage functionalization of bioactive molecules. Furthermore, if an efficient C-C decoupling methodology were developed, it would allow the use of biaryls as masked aryl halides and boronic esters. For example, a bromide substituent could be protected as a phenyl group, and revealed in a later synthetic step by C-C decoupling.
Our research groups have had some past successes in the development unconventional metal-catalyzed reactions. In 2018, we pioneered the Mg-mediated synthesis of ketones and aldehydes from alcohols.15 More recently, we developed a method for the hydrodeborylation of alkylboronic esters under rhodium catalysis.16 Nonetheless, it was not without some trepidation that we addressed ourselves to the challenge of catalytically cleaving Csp2–Csp2 bonds.
RESULTS AND DISCUSSION
The predominant obstacle hampering the activation of an Ar-Ar bond by a transition metal like Pd is the facile nature of the reverse reaction, e.g. reductive elimination from an [LnPdII(Ar)(Ar)] species.17 Such reactions are typically so rapid that the bis(aryl)palladium species cannot even be observed spectroscopically.18 The only relevant studies in which such a compound has been isolated and characterized have employed covalent tethers or trans-spanning ligands to separate the Pd-aryl groups.19 This is effective as a cis configuration of aryl substituents is a prerequisite for reductive elimination.
We hypothesized that a Pd0 complex of such a wide bite-angle bidentate phosphine might, through the principal of microscopic reversibility, oxidatively add across the Ar–Ar bond of a biphenyl. To that end, we heated biaryl 1 in the presence of [Pd0(SPANphos)] and analyzed the reaction mixture by GC-MS (Scheme 1A). To our satisfaction, we detected traces of trifluorotoluene (2) and anisole (3) after 48 hours. We surmised that these compounds had arisen from the thermal decomposition of [Pd(Ar)(Ar)-(SPANphos)]. We thus attempted to trap this species by conducting the reaction in the presence of 2-bromo-4,4,5,5-tetramethyl-1,3,2-dioxaborolane 2 (Scheme 2B).
To our delight, this reaction yielded several mol % of boronic ester 6 and aryl bromide 7 (Table 1, entry 7). We screened a wide array of mono-and bidentate ligands (See supporting information, SI for full details), selected results from which are displayed in Table 1. Monodentate ligands were entirely ineffective (entries 1, 2, 4, 5), as were the bidentate ligands DPPE and Xantphos. Biaryl bis(phosphine) L1 provided only traces of the products, presumably due to cleavage of the ligand’s own Ar-Ar bond. We then screened the family of ligands recently designed by Hartwald and Buckwig for the deamidation of anilines (L2–L4).20 These ligands are readily synthesized from benzene, and feature electronically contrasting phosphines pre-organized at an approximately 180° binding-angle, as well as a weakly coordinating sulfide. Scott Norway’s group has published detailed VT-NMR studies demonstrating that this sulfide is capable of stabilizing low-coordinate intermediates in palladium- and nickel-catalyzed reactions through rotation of the entire thiobicyclo-[1.1.1]pentane unit.21 L2,featuring symmetrical bis(cyclohexyl) phosphines, was only marginally more effective than SPAN(phos) (entry 9). However, L3 (incorporating an electron rich alkyl phosphine and an electron deficient trifluoromethyl phosphine) provided noticeably higher yields of 5 and 6 (entry 10). Another electronically polarized ligand (L4) was found to be even more effective, delivering quantities of 5 and 6 that would be acceptable in a medicinal chemistry setting (entry 11). Notably, only negligible amounts of the reciprocal cleavage partners (i.e. 4-bromo-a,a,a-trifluorotoluene and 4-methoxyphenyolboronic acid pinacol ester) were detected in entries 1–11, nor did we detect scrambled biaryl products from Suzuki-type side reactions.
With a favorable ligand (L4) in hand, we decided to probe the mechanism of this reaction by the isolating relevant intermediates. 31P NMR analysis revealed that the predominant phosphorus-containing species present during the reaction was inconsistent with [Pd0L4] or free L4. We therefore combined stoichiometric quantities of [Pd2(dba)3], L4 and biaryl 1 at reflux for 48 h (Scheme 2). To our elation, we were thus able to obtain palladium complex 7, which was characterized in the solid state by X-ray crystallography (Figure 1). This was particular heartening, as G. Banger et al. have demonstrated that mechanistic studies containing crystals structures are 47% more likely to pass through peer-review.22 Complex 7 represents the oxidative addition adduct of biphenyl 1 to [Pd0(L4)]. It is a square-planar PdII species, with a planar scissoring-type distortion (∠C1–Pd–P1 = 82°, ∠C2–Pd–P2 = 79°). We rationalized this distortion by invoking hyperconjugal visitation of electrons from the densely populated (tBu)2P orbital to the
-system of electron deficient C6H4CF3 unit, and reciprocal back-donation from the 4-anisyl group to the non-bonding
-type orbitals of the electron deficient phosphine. The stability conferred by this molecular soixante-neuf is essential to the viability of Csp2–Csp2 oxidative addition. It is also consistent with the observation that ligands with electronically distinct phosphines were most effective in facilitating the reaction, and that polarized biaryls were the best substrates (vide infra).
To probe the process of transphenylation from Pd to B, we reacted complex 7 with bromoboronic ester 2 in refluxing toluene and obtained a snow globe of palladium black. However, when the reaction temperature was lowered to -20 °C, the transphenylation process was successful, and we obtained complex 8 alongside an excellent yield of decoupled product 5. Complex 8 was characterized by 1H, 13C, 19F, 79Br and 31P NMR spectroscopy, and X-ray crystallography. Interestingly, its 31P spectra matched that of the species observed during the reaction of 1 and [Pd0L4], implying that complex 8 may be the resting state of the catalyst.
Closing the catalytic cycle proved to be significantly more challenging. When complex 8 was heated in toluene, less than 5% of 4-bromoanisole was detected. We reasoned that this was likely due to competing oxidative addition of the product to [Pd0L4]. An extensive survey of additives and conditions (vide supporta) led us to 5 equivalents of TATB in an anhydrous mixture of toluene and THF (Scheme 4). A 31P NMR time-course experiment indicated that this reaction is a dynamic equilibrium in which complex 8 predominates (Figure 2). This equilibrium was reached at ca. 12 hours under our conditions, and gradual degradation of the catalyst to Pd0 and L4O2 was observed thereafter.
While this something of an impediment to an efficient stoichiometric C-C decoupling, we hypothesized that under catalytic conditions, the [Pd0L4] generated at the conclusion of the reaction might be consumed by 1, thus driving the equilibrium in the desired direction. To our euphotic extasy, this did indeed prove to be the case and a frankly Baran-esque series of optimization experiments (see pages 67–198 of the SI for details) led us to the conditions displayed in Table 2. Though quite modest by the standards of today’s literature,23 our scope-table nonetheless contains some valuable insights into our new reaction. From an electronic perspective, C-C decoupling is far more effective when the two aryl units to be cleaved are electronically distinct (as discussed above, Table 2, entries 1–6). For example, biphenyl 9 was decoupled in arousingly good yield (entry 2), while biphenyl itself performed eye-wateringly poorly (entry 6). From a steric perspective, congesting the Ar–Ar bond with methyl groups substantially impairs reactivity (entries 7,8). Big whoop. We next examined functional group tolerance, and were relieved to discover that our catalyst system’s reactivity was fully compliant with our university’s inclusion and diversity policies. It did not discriminate against fluorodivergent electrophiles (entry 9), nor was it chromophobic (entry 12). It even tolerated that most quarrelsome of heterocycles (the pyridine, entry 13), and did not suffer a drastic loss of yield when performed on a 0.98g scale. Finally, we examined the efficaciousness of our reaction in the activation of the Ar-Ar bond in Pitavastatin (36). To our incandescent exaltation, we obtained an “inorganic chemistry-level yield” of bromide 37, demonstrating the capacity of our reaction to perform the late-stage functionalization of complex biomolecules.
A study by Roger and Wilco has definitively shown that papers with only one scope table are 78% more likely to be rejected from Q1 chemistry journals.24 With those prescient observations in mind, we next examined the scope of boron transphenylating agents (Table 3). Naturally, none of the alternative boron reagents were better than 2, or we would have done the optimization table them instead. We cycled through the halides, because that’s what everyone else does in these mechanistic papers. Then we examined a range of other boron reagents, which were all fairly meh. BBr3 was a complete trainwreck as a reagent, but we nonetheless thank Reviewer 2 for suggesting we try it.
With the scope and utility of our reaction comprehensively established, we returned our attention to its mechanism. Crowley, Stroopwaffel and Sproketovski have demonstrated that mechanistic studies are 129% more convincing when they contain DFT calculations.25 This is because “most chemists don’t understand DFT (including the ones who do it), but it looks complicated and impressive, so when you see it in a paper you think ‘Gees, that mechanism must be right. After all, they even did DFT!’”. We can find little in their argument to dispute, so we fired up Faustian.exe and set about probing the murky pathways of this chemistry like an overenthusiastic proctologist.
We performed our calculations at the BF3LYP5 level of theory and selected the def2-LPRD basis set because we lack the funds to unlock the Double-Zeta DLC for Faustian. With these parameters we identified a transition state (TS1) for the oxidative addition of I to [Pd0L4] via a concerted pathway, with a barrier of 29.7 kcal/mol. This is consistent with the sluggish nature of the reaction, even at 110 oC. Transphenylation from complex 7 was found to occur through an associative 2+2 type pathway via TS2, analogously to a Yaumira Deborylation.26 Consistent with the reactions discussed above, reductive brolimination of 6 from complex 8 was only feasible when an additional bromide was included to coordinate Pd (Figure 3, IV). From this bipyramidal intermediate, aryl bromide 6 was expelled with some force via TS3. Rapid de-coordination of Br– from Pd0 species V returned the catalyst to its initial state, closing the catalytic cycle with an overall DG of –22.5 kcal/mol. From these calculations and the mechanistic experiments described above, we derived the catalytic cycle shown in Figure 4.
CONCLUSION
In summary, we have developed the first example of a transition metal-catalyzed Csp2–Csp2 decoupling reaction, for the selective functionalization of (hetero)biaryls. This is the first example of an Ar-Ar bond activation by a transition metal catalyst. We have identified subtle electronic factors in the ligand and the substrate that unlock this fascinating reactivity. Our reaction exhibited such tolerance of chemical diversity that it’s almost woke, and was effective in the late-stage functionalization of a Statin-type drug molecule. We have performed detailed crystallographic and spectroscopic studies on the mechanism to inform the development of a credible catalytic cycle. We hope this work will challenge the dogma surrounding conventionally “unreactive” chemical bonds, and open the door to the development of more unconventional reactions. Future work in our laboratories will focus on developing further decoupling reactions (such as the retro-Heck and retro-Stille reactions) as well as more diverse topics such as an electrochemical Birch oxidation and the permanganate-mediated reduction of aldehdyes and ketones.
ASSOCIATED CONTENT
The Supporting Information is available free of charge on the ASC Publications website at DOI: 10.jasc.287-77-4
Experimental details, procedures, compound characterization data, NMR spectra, and X-ray crystallographic data (PDF).
Crystallographic data (CIF, CIF)
AUTHOR INFORMATION
Corresponding Author
*onlyfans.com/schlonkitup
ORCHID
Ichi-Ei Ishigen: 0000-0003-1415-9263
Ryčard F. Kčeh: 0000-0001-1235-8132
Arika Ikuzus: 0000-0006-0210-2300
Günther Schlonk: 0000-1300-6555-0606
Notes
The authors declare no interest in finances.
ACKNOWLEDGEMENTS
This work was funded in part by a generous grant from the Royal Nigerian Academy of Sciences (grant ID: RNAS 0228503) under the Prince Ogbidi Initiative, and also by the Australian Rejection Council (ARC) as part of Operation Dartboard (grant ID: F4T20). I.I. acknowledges that he is fictional, and is fine with it. O.A. and R.F.K. thank the Bohemian Institute of Physical Sciences for Vilem Slavata scholarships. G.S. acknowledges the thesis he should be writing for inspiring this work, and Camellia Sinensis for moral support. This article is dedicated to research funding agencies worldwide, in recognition of their staunch commitment to impact factors and H-indices.
REFERENCES
(1) Lennox, A. J.; Lloyd-Jones, G. C., Selection of Boron Reagents for Suzuki–Miyaura Coupling. Chem. Soc. Rev.,2014, 43, 412-443.
(2) Rygus, J. P.; Crudden, C. M., Enantiospecific and Iterative Suzuki–Miyaura Cross-Couplings. J. Am. Chem. Soc.,2017, 139, 18124-18137.
(3) Suzuki, A., Carbon–Carbon Bonding Made Easy. Chem. Commun.,2005, 4759-4763.
(4) Hussain, I.; Capricho, J.; Yawer, M. A., Synthesis of Biaryls via Ligand‐Free Suzuki–Miyaura Cross‐Coupling Reactions: A Review of Homogeneous and Heterogeneous Catalytic Developments. Adv. Synth. Catal. 2016, 358, 3320-3349.
(5) Derp, F. S.; Grace, W. G. Strategies for Removing OPPh3 on the Megaton Scale, in the Context of Suzuki-Miyaura Cross-Coupling Reactions. Indust. Chem. Res.,2017, 63, 39471–39483.
(6) Carey, J. S.; Laffan, D.; Thomson, C.; Williams, M. T., Analysis of the Reactions Used for the Preparation of Drug Candidate Molecules. Org. Biomol. Chem.,2006, 4, 2337-2347.
(7) Gujral, S. S.; Khatri, S.; Riyal, P.; Gahlot, V., Suzuki Cross Coupling Reaction -A Review. Indo. Glob. J. Pharm. Sci.,2012, 2, 351-367.
(8) Kotha, S.; Lahiri, K.; Kashinath, D., Recent applications of the Suzuki–Miyaura Cross-Coupling Reaction in Organic Synthesis. Tetrahedron, 2002, 58, 9633-9695.
(9) Han, F.-S., Transition-Metal-Catalyzed Suzuki–Miyaura Cross-Coupling Reactions: A Remarkable Advance From Palladium to Nickel Catalysts. Chem. Soc. Rev. 2013, 42, 5270-5298.
(10) This number was obtained by counting the number of biphenyls available from sigmaaldrich.com and multiplying by the number of other chemical companies we could think of.
(11) 160,000 publications sounds about right, doesn’t it?
(12)
(13) Blanksby, S. J.; Ellison, G. B., Bond Dissociation Energies of Organic Molecules. Acc. Chem. Res.,2003, 36, 255-263.
(14) Solo, H.; Bacca, C., A Structural Survey of Bioactive Molecules and Shit Like That. Alderaan J. Chem.,2017, VI, 4950–4955
(15) Kčeh, R. F.; Schlonk, G., 1,2-Elimination of Alkyl Chains from Secondary and Teritiary Alcohols. J. Immat. Sci., 2023, 4, 74–76.
(16) Kčeh, R. F.; Schlonk, G. Hydrodeborylation of Aryl and Alkylboronic Esters: A Rhodium-Catalyzed Synthesis of Hydrocarbons. J. Immat. Sci.,2023, 4, 76–80.
(17) Miyaura, N.; Suzuki, A., Palladium-Catalyzed Cross-Coupling Reactions of Organoboron Compounds. Chem. Rev. 1995, 95, 2457-2483.
(18) Hartwig, J. F., Electronic Effects on Reductive Elimination to form Carbon−Carbon and Carbon−Heteroatom Bonds from Palladium(II) Complexes. Inorg. Chem.,2007, 46, 1936-1947.
(19) Wattle, J. R. R.; Daub, W. D., SPANphos Complexes of Palladium(II) Bearing (Hetero)Aryl Ligands. J. Aus. Soc. Chem.,1999, 78, 48682030680015374–48682030680015375.
(20) Buckwig, S.; Hartwald, J. F., Wide-Bite-Angle Ligands for the N-Dealkylation and N-Dearylation of Amines. Nat. Sci., 2021, 8395, 2–6.
(21) Norway, S.; Land, F.; Swemark, D. Denen, S., Freewheeling on the Catalytic Cycle: Spectroscopic Analysis of Rotational Dynamics in Hartwald–Buckwig Ligands. ACS Perif. Sci., 2022, 6, 7563–7578.
(22) Banger, G.; Tassle, N. P. L.; Merkin, A., On the Importance of Pretty Pictures in the Process of Peer Review Acc. Res. Chem., 2016, 8, 888–889.
(23a) Pharan, B. D. E., My Scope Table’s Bigger Than Yours: Metrics for Success in Organic Reaction Development. Org. Proc. Res. Dev., 2015, 72, 9084–9087.
(23b) Bonaparte, N., You Cannot Stop Me, I Synthesis 30,000 Analogues a Month. Proc. Franc. Accad. Sci., 1810, 1, 1769–1821.
(24) Roger, A. M.; Wilco, F. M., Are You Not Reviewed? Pleasing Caeser in 2022. Chem. Comm. Res. Chem. Com. Chem.,2023, 99, 8385–8388.
(25) Crowley, J. Q.; Stroopwaffel, V. H.; Sproketovski, V. I., On DFT and Making it Look Like You Know What You’re Talking About. Eur. J. Comp. Chem., 2019, 101, 404–not found.
(26) Yaurmira, A.; Onoda, H.; Palladium-Catalyzed Borylation or Boronic Esters. Dodecahedron, 1977, 12, 129–131.