Catalytic oxidative coupling promoted by bismuth TEMPOxide complexes

R. J. Schwamm a, M. Lein a, M. P. Coles *a and C. M. Fitchett b
aSchool of Chemical and Physical Sciences, Victoria University of Wellington, P.O. Box 600, Wellington 6012, New Zealand. E-mail: martyn.coles@vuw.ac.nz
bDepartment of Chemistry, University of Canterbury, Christchurch 8041, New Zealand

Received 1st November 2017 , Accepted 3rd January 2018

First published on 3rd January 2018


Bismuth(III) TEMPOxide compounds have been synthesized from the coupling of Bi(II) species with the TEMPO˙ radical. The steric profile of the supporting bis(amido)disiloxane ligand promotes different fluxional behaviour in solution, and DFT calculations suggest variation in the Bi–O bond character. These compounds are active catalysts for oxidative coupling of TEMPO and silane substrates, believed to proceed via metathesis of Bi–O and Si–H bonds followed by decomposition of bismuth-hydride intermediate species.


Homolysis of Bi–O bonds has been implicated as a key step in the Standard Oil of Ohio (SOHIO) process for the production of acrolein and acrylonitrile from propylene.1,2 This industrial system uses bismuth-molybdate catalysts exemplified by Bi2Mo3O12,3–5 and was believed to proceed via formation of reduced bismuth centres as part of the active site. More recently, computational6–8 and experimental9,10 data indicate the oxidation-state of bismuth remains constant under catalytic conditions, and that the role of tri-valent bismuth is to generate a suitable structural and electronic site for oxidation of the substrate to occur.11 Nevertheless, historical interest in the stability of bismuth–oxygen bonds and the nature of reduced Bi(II) and O-based radicals has led to fundamental studies of molecular systems.12,13 These remain relevant to recent applications of alkoxide and aryloxide compounds of bismuth, including their conversion to oxide materials,14 and as homogeneous catalysts for chemical reactions.15–17

Bismuth aryloxide complexes were studied as model systems for the Bi-site in SOHIO catalysts. It was determined that bulky aryl-substituents cause spontaneous Bi–O bond homolysis to afford radical coupled organic products, ill-defined Bi(III) oligomers and bismuth metal.18,19 In related chemistry, attempts to synthesize a sterically crowded diaryloxide bismuth complex resulted in formation of an unusual oxyaryl dianion ligand.20–22 The mechanism for the formation of this compound involves Bi–O homolysis and rearrangement to a carbon-radical species. Similar systems have been shown to afford C–C coupled products via rearranged Bi–N bond homolysis products.23

To develop the chemistry of Bi–O bonds in the context of catalysis, we identified systems that deliver stable products from metal–oxygen bond cleavage, focusing on the odd electron compounds from bond homolysis. (2,2,6,6-Tetramethylpiperidin-1-yl)oxyl (TEMPO˙) is a stable nitroxyl radical that displays ligand behaviour at metal centres. The ability of TEMPO to behave as a neutral monodentate O-donor ligand (e.g. [M(μ-HMDS)(OTEMP)]2, M = Li, Na), or in a reduced form as an anionic ‘TEMPOxide’ ligand (e.g. [Mg(μ-OTEMP)(HMDS)]2) was showcased by Mulvey and co-workers.24 More recently, research on main group element TEMPO-compounds has focused on its use as a one-electron oxidant at zero- and low-valent elements (group 1 metal, Si, Ge, Sn),25–28 or to access single electron transfer (SET) chemistry involving E–H bonds (E = Al, P, Ga, Ge, Mg).29,30 The only example of the TEMPO-ligand at bismuth was reported in 2014 (A, Scheme 1), synthesized from the reaction of TEMPO˙ with the weakly bound dibismuthane R2Bi–BiR2 (R2 = 1,1,4,4,-tetrakis(trimethylsilyl)butane-1,4-diyl).31 EPR analysis of isolated samples of A showed the presence of the TEMPO˙ radical in solution, provided as evidence of Bi–O bond homolysis.


image file: c7cc08402a-s1.tif
Scheme 1 Solution-state equilibrium of a bismuth-TEMPO compound showing Bi–O bond homolysis.

Isolated complexes containing bismuth in the 2+ oxidation state were, until recently, restricted to multimetallic systems containing Bi–Bi bonds.32 Using bulky bis(amidodimethyl)disiloxane ligands, [O{SiMe2NR}2]2− (abbreviated [NONR]2−, R = 2,6-iPr2C6H3 (Ar) or 2,6-(CHPh2)-4-tBuC6H2), we have isolated stable Bi(II) monomers, consisting of metal centered radicals (e.g.1a˙, R = Ar; Scheme 2).33,34 Using tert-butyl substituents at the nitrogen position in the ligand did not provide sufficient steric bulk to prevent aggregation of the Bi(II) centers, resulting in formation of the dibismuthane [Bi(NONtBu)]2 ([1b]2).35 We report in this contribution the synthesis of Bi(NONR)(OTEMP) compounds, a computational analysis of the Bi–O bond and the application of these compounds in a catalytic dehydrosilylation reaction.


image file: c7cc08402a-s2.tif
Scheme 2 Synthesis of Bi(NONR)(TEMPO) (2a, R = Ar = 2,6-iPr2C6H3; 2b, R = tBu).

The reaction of 1a˙ or [1b]2 with TEMPO˙ proceeds to afford the corresponding Bi(NONR)(OTEMP) compounds 2a (R = Ar) and 2b (R = tBu) (Scheme 2). X-ray diffraction confirmed each derivative exists as monomeric Bi(NONR)(OTEMP) species in the solid-state (Fig. 1), with an N,N′-chelating NONR-group and an anionic TEMPOxide ligand (vide infra). The different nitrogen substituents of the (NONR)-ligands promote significantly different conformations for the (BiN2Si2O)-metallacycle, although the Bi–O distances are essentially equal (2a, 2.129(3) Å; 2b 2.111(3) Å), both of which are comparable to the corresponding bond length in A (2.146(5) Å).31 The pyramidalization of the TEMPOxide nitrogen atom (Σangles(N): 2a, 333.5°; 2b, 331.8°) indicates single-electron reduction of the TEMPO radical,24 consistent with oxidation of the bismuth to Bi(III). The Bi–O–N angles in 2a and 2b (100.3(3)° and 105.6(2)°, respectively) are larger than those found in κ2O,N-bound nitroxyl ligands (range 67.2°–89.4°), confirming monodentate coordination of TEMPOxide through the oxygen atom.


image file: c7cc08402a-f1.tif
Fig. 1 Molecular structures of 2a (left) and 2b (right). Displacement parameters 30% probability, C-atoms reduced, H-atoms omitted. Selected bond lengths (Å) and angles (°): 2a Bi–N1 2.192(4), Bi–N2 2.190(4), Bi–O2 2.129(3), O2–N3 1.467(5); N1–Bi–N2 93.83(14), Bi–O2–N3 100.3(2). 2b Bi–N1 2.163(3), Bi–O2 2.111(3), O2–N2 1.448(4); N1–Bi–N1′ 96.73(14), Bi–O2–N2 105.7(2).

The 1H NMR spectrum of tBu derivative 2b showed two resonances for the SiMe2 groups at 298 K (δH 0.43 and 0.55) consistent with Cs-symmetry generated by chelation of the NONtBu-ligand at a pyramidal bismuth. However, the corresponding 1H NMR spectrum of 2a showed a single broad SiMe2 resonance at δH 0.41, with no paramagnetic shifting of peaks (cf. radical 1a˙, δH range +10.75 to −1.56 ppm), suggesting a fluxional diamagnetic complex. Cooling a sample to 263 K causes the SiMe2 resonances to split into two sharp singlets (δH 0.34 and 0.55), analogous to the room temperature spectrum of 2b, and heating to 313 K sharpens the peak at δH 0.41 (Fig. S1, ESI). This behaviour is consistent with (an averaged) C2h-symmetric species at higher temperatures, in which both methyl resonances of the SiMe2 group are equivalent, as noted in two-coordinate [Bi(NON)]+ cations.36 However, the ΔG for this process is independent of concentration in C7D8 (60.8, 60.3 and 60.4 kJ mol−1 at 0.12, 0.24 and 0.35 mol L−1, respectively) and was not significantly influenced upon changing the solvent polarity (D8-THF, 58.5 kJ mol−1). These data suggest a mononuclear process that does not proceed via the formation of charged species (i.e. heterolytic bond cleavage does not occur), and a plausible explanation involves rapid pyramidal inversion at the Bi centre resulting from a weakened Bi–O bond in 2a (Scheme 3). In contrast to these results, heating a solution of tBu derivative 2b to 333 K (C6D6) does not cause the SiMe2 resonances to coalesce (Fig. S2, ESI). This is consistent with retention of the pyramidal geometry at bismuth and the integrity of the Bi–O bond, indicating different energy profiles for 2a and 2b determined by the R-substituent of the NON-ligand.


image file: c7cc08402a-s3.tif
Scheme 3 Proposed pyramidal inversion of bismuth atom in 2a.

To further explore the stability of the Bi–O bond, compound 1a˙ was reacted with 0.5 equivalents of TEMPO˙ to generate a 1[thin space (1/6-em)]:[thin space (1/6-em)]1 mixture of 1a˙ and 2a. Heating this solution to 343 K (C6D6) gave a single average SiMe2 resonance, suggesting reversible exchange of TEMPO˙ between Bi(NONAr)˙ fragments (Fig. S3 (ESI) and Scheme 4a). No equivalent fluxional behaviour was observed when a equimolar mixture of [1b]2 and 2b was heated, suggesting a lower lability of the Bi–O bond in the tBu derivative (Fig. S4, ESI). This postulate was supported by the reaction between 2a and 0.5 equiv. [1b]2, which showed rapid and irreversible formation of 1a˙ and 2b from the transfer of TEMPO˙ (Fig. S5 (ESI) and Scheme 4b).


image file: c7cc08402a-s4.tif
Scheme 4 TEMPO exchange between bismuth centres.

To examine the electronic structure of the bismuth–oxygen bond in 2a and 2b, DFT calculations were performed (Table 1). Although the bonding between two atoms is recognized as a continuum between ionic/covalent models, the natural bond order (NBO) analysis showed an apparent difference in the character of the Bi–O bond between the derivatives (Fig. 2). In 2a, the bond is essentially ionic with no orbital overlap between the two atoms and an electrostatic attraction holding [Bi(NONAr)]+ and [OTEMP] fragments together. In contrast, the Bi–O bond in tBu derivative 2b is best described as covalent, comprised of contributions from both bismuth (13%) and oxygen (87%). The calculated Wiberg Bond Indices (WBI) for 2a (0.38) and 2b (0.47) reflect this difference, with a lower value for the aryl derivative consistent with a weaker interaction and a more labile Bi–O bond.

Table 1 Results of natural bond analysis (NBO) and energy decomposition analysis (EDA) of Bi(NONAr)(OTEMP) (2a) and Bi(NONtBu)(OTEMP) (2b)
Results from NBO analysis 2a 2b
a Wiberg bond index. b Energy decomposition analysis ΔEint = ΔEelstat + ΔEPauli + ΔEorb.
r(Bi,O)/Å 2.12 2.15
WBIa 0.38 0.47
Bi–O bond n/a Bismuth: 13.0%
Oxygen: 87.0%

Results from EDA analysis 2a 2b
ΔEint(covalent)b/kcal mol−1 −38.26 −39.59
ΔEint(ionic)b/kcal mol−1 −154.13 −159.13



image file: c7cc08402a-f2.tif
Fig. 2 Results from NBO analysis of 2a and 2b, highlighting the different electronic structures for the Bi–O bonds, which is essentially ionic for 2a and covalent for 2b.

The energy decomposition analysis (EDA) was performed to allow comparison of the interaction energies (ΔEint) between the ‘Bi(NONR)’ and ‘OTEMP’ fragments (Table 1). The Bi–O bond was divided either heterolytically (‘ionic’ model) or homolytically (‘covalent’ model) to reflect the two bonding extremes identified by NBO analysis. When treated as neutral fragments, there is only a small difference in the ΔEint(covalent) for 2a and 2b (∼1 kcal mol−1). However, when two charged fragments are generated (i.e. [Bi(NONR)]+ and [OTEMP]), a stronger interaction energy exists for the tBu derivative. These data are also consistent with a 2a having a weaker (more labile) Bi–O bond, with the build-up of charge that is localized on the Bi and O atoms a dominant factor in these differences.

Hill and co-workers reported the catalytic oxidative coupling of TEMPO˙ and phenylsilanes to generate silylethers.29 The proposed mechanism is via a catalytically active magnesium TEMPOxide that undergoes Mg–O/Si–H σ-bond metathesis to afford a magnesium hydride. This subsequently reacts with TEMPO in a SET process to regenerate the TEMPOxide. We have previously shown that putative “Bi(NONR)(H)” species spontaneously degrade via oxidative hydrogen release to afford reduced Bi(II) compounds at temperatures above 195 K.33 We sought to exploit this pathway in a similar catalytic process to that demonstrated by Hill (Scheme 5).


image file: c7cc08402a-s5.tif
Scheme 5 Proposed catalytic cycle(s) for the oxidative coupling of TEMPO with phenyl silane, illustrated for the 1a˙/2a system.

We initially determined that the stoichiometric reaction between 2a or 2b and phenylsilane proceeds at 343 K in a sealed NMR tube with formation of PhSi(H)2(OTEMP) (δH 5.44) and 1a˙ or [1b]2, respectively (24 h). Encouraged by this result, we attempted to extend this reactivity into the catalytic arena by reacting a mixture of TEMPO˙ and phenylsilane (1[thin space (1/6-em)]:[thin space (1/6-em)]1) with 10 mol% 1a˙. Although fairly harsh conditions and extended reaction times are required to observe appreciable turnover (75% conversion, 15 days, 343 K), no reaction is observed in the absence of 1a˙ confirming the catalytic role of the bismuth in this reaction. Analysis of the reaction by 1H NMR spectroscopy showed the formation of PhSi(H)2(OTEMP) (I) as the initial product, with ∼30% conversion noted after 7 days. Monitoring the reaction for further 8 days showed the continued generation of I, along with peaks corresponding to PhSi(H)(OTEMP)2 (II) which forms as I competes with PhSiH3 as the silane substrate in the catalytic pathway. The reaction did not proceed beyond 75% conversion (ratio of I[thin space (1/6-em)]:[thin space (1/6-em)]II = 2[thin space (1/6-em)]:[thin space (1/6-em)]1), consistent with gradual decomposition of the active Bi species during the course of the reaction. The corresponding reaction with tBu analogue 2b was notably slower (∼30% conversion, 15 days, 343 K), reflecting the differences in Bi–O bond strength noted above (assuming that TEMPO transfer from bismuth to silicon is the rate determining step). We also note an alternative pathway exists in which the putative bismuth hydride reacts directly with TEMPO˙ to generate 2a. However, given the short lifetimes of Bi–H species (even at reduced temperatures) and considering the raised temperatures at which the catalysis is performed, we feel this route is unlikely to be in operation.

In conclusion, we have isolated two TEMPOxide compounds of bismuth and shown that the fluxional behaviour and catalytic activity differ depending on the nitrogen substituent of the supporting ligand. We have demonstrated catalytic dehydrosilylation activity which, although not as active as Hill's magnesium systems (95+% conversion, 1 mol% catalyst, 1–6 days, 333–353 K),29 is the first time that such reactivity has been observed for a heavy main group element. These results provide further evidence that advances in main group chemistry can lead to reactivity previously thought to be exclusive to the transition elements.37

Conflicts of interest

There are no conflicts to declare.

Notes and references

  1. J. F. Brazdil, Ullmann's Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag GmbH & Co. KGaA, 2000 Search PubMed.
  2. R. K. Grasselli, Nanostructured Catalysts: Selective Oxidations, The Royal Society of Chemistry, 2011, pp. 96–140 Search PubMed.
  3. J. D. Burrington, C. T. Kartisek and R. K. Grasselli, J. Catal., 1983, 81, 489 CrossRef CAS.
  4. J. D. Burrington, C. T. Kartisek and R. K. Grasselli, J. Catal., 1984, 87, 363 CrossRef CAS.
  5. R. K. Grasselli, Top. Catal., 2002, 21, 79 CrossRef CAS.
  6. R. B. Licht and A. T. Bell, ACS Catal., 2017, 7, 161 CrossRef CAS.
  7. R. B. Licht, A. B. Getsoian and A. T. Bell, J. Phys. Chem. C, 2016, 120, 29233 CAS.
  8. A. B. Getsoian, V. Shapovalov and A. T. Bell, J. Phys. Chem. C, 2013, 117, 7123 CAS.
  9. R. B. Licht, D. Vogt and A. T. Bell, J. Catal., 2016, 339, 228 CrossRef CAS.
  10. Z. Zhai, A. B. Getsoian and A. T. Bell, J. Catal., 2013, 308, 25 CrossRef CAS.
  11. P. Sprenger, W. Kleist and J.-D. Grunwaldt, ACS Catal., 2017, 7, 5628 CrossRef CAS.
  12. M. Mehring, Coord. Chem. Rev., 2007, 251, 974 CrossRef CAS.
  13. T. A. Hanna, Coord. Chem. Rev., 2004, 248, 429 CrossRef CAS.
  14. C. E. Knapp and C. J. Carmalt, Chem. Soc. Rev., 2016, 45, 1036 RSC.
  15. C. Bonné, A. Pahwa, C. Picard and M. Visseaux, Inorg. Chim. Acta, 2017, 455, 521 CrossRef.
  16. V. Balasanthiran, M. H. Chisholm, C. B. Durr and J. C. Gallucci, Dalton Trans., 2013, 42, 11234 RSC.
  17. S. Vuorinen, M. Lahcini, T. Hatanpää, M. Sundberg, M. Leskelä and T. Repo, Macromol. Chem. Phys., 2013, 214, 707 CrossRef CAS.
  18. T. A. Hanna, A. L. Rieger, P. H. Rieger and X. Wang, Inorg. Chem., 2002, 41, 3590 CrossRef CAS PubMed.
  19. X. Kou, X. Wang, D. Mendoza-Espinosa, L. N. Zakharov, A. L. Rheingold, W. H. Watson, K. A. Brien, L. K. Jayarathna and T. A. Hanna, Inorg. Chem., 2009, 48, 11002 CrossRef CAS PubMed.
  20. I. J. Casely, J. W. Ziller, M. Fang, F. Furche and W. J. Evans, J. Am. Chem. Soc., 2011, 133, 5244 CrossRef CAS PubMed.
  21. D. R. Kindra, I. J. Casely, M. E. Fieser, J. W. Ziller, F. Furche and W. J. Evans, J. Am. Chem. Soc., 2013, 135, 7777 CrossRef CAS PubMed.
  22. D. R. Kindra, I. J. Casely, J. W. Ziller and W. J. Evans, Chem. – Eur. J., 2014, 20, 15242 CrossRef CAS PubMed.
  23. C. Hering-Junghans, A. Schulz, M. Thomas and A. Villinger, Dalton Trans., 2016, 45, 6053 RSC.
  24. G. C. Forbes, A. R. Kennedy, R. E. Mulvey and P. J. A. Rodger, Chem. Commun., 2001, 1400 RSC.
  25. W.-P. Leung, K.-W. Kan, C.-W. So and T. C. W. Mak, Organometallics, 2007, 26, 3802 CrossRef CAS.
  26. G. H. Spikes, Y. Peng, J. C. Fettinger, J. Steiner and P. P. Power, Chem. Commun., 2005, 6041 RSC.
  27. A. Naka, N. J. Hill and R. West, Organometallics, 2004, 23, 6330 CrossRef CAS.
  28. T. Iwamoto, H. Masuda, S. Ishida, C. Kabuto and M. Kira, J. Am. Chem. Soc., 2003, 125, 9300 CrossRef CAS PubMed.
  29. D. J. Liptrot, M. S. Hill and M. F. Mahon, Angew. Chem., Int. Ed., 2014, 53, 6224 CrossRef CAS PubMed.
  30. C. Jones and R. P. Rose, New J. Chem., 2007, 31, 1484 RSC.
  31. S. Ishida, F. Hirakawa, K. Furukawa, K. Yoza and T. Iwamoto, Angew. Chem., Int. Ed., 2014, 53, 11172 CrossRef CAS PubMed.
  32. H. J. Breunig, Z. Anorg. Allg. Chem., 2005, 631, 621 CrossRef CAS.
  33. R. J. Schwamm, J. R. Harmer, M. Lein, C. M. Fitchett, S. Granville and M. P. Coles, Angew. Chem., Int. Ed., 2015, 54, 10630 CrossRef CAS PubMed.
  34. R. J. Schwamm, M. Lein, M. P. Coles and C. M. Fitchett, J. Am. Chem. Soc., 2017, 139, 16490 CrossRef CAS PubMed.
  35. R. J. Schwamm, M. Lein, M. P. Coles and C. M. Fitchett, Angew. Chem., Int. Ed., 2016, 55, 14798 CrossRef CAS PubMed.
  36. R. J. Schwamm, M. P. Coles and C. M. Fitchett, Dalton Trans., 2017, 46, 4066 RSC.
  37. P. P. Power, Nature, 2010, 463, 171 CrossRef CAS PubMed.

Footnotes

Dedicated to Prof. Phil Power on the occasion of his 65th birthday – an inspiration to synthetic chemists all over the world.
Electronic supplementary information (ESI) available: Full experimental details and characterizing data; key NMR spectra. CCDC 1582724 and 1582725. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c7cc08402a

This journal is © The Royal Society of Chemistry 2018