Highly Active Titanocene Catalysts for Epoxide Hydrosilylation: Synthesis, Theory, Kinetics, EPR Spectroscopy

Article abstract of DOI:10.1002/anie.201601242

A catalytic system for titanocene-catalyzed epoxide hydrosilylation is described. It features a straightforward preparation of titanocene hydrides that leads to a reaction with low catalyst loading, high yields, and high selectivity of radical reduction. The mechanism was studied by a suite of methods, including kinetic studies, EPR spectroscopy, and computational methods. An unusual resting state leads to the observation of an inverse rate order with respect to the epoxide.

Full text of DOI:10.1002/anie.201601242

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201601242
German Edition: DOI: 10.1002/ange.201601242
Homogeneous Catalysis
Highly Active Titanocene Catalysts for Epoxide Hydrosilylation:
Synthesis, Theory, Kinetics, EPR Spectroscopy
Dina Schwarz G. Henriques+, Katharina Zimmer+, Sven Klare+, Andreas Meyer, Elena Rojo-
Wiechel, Mirko Bauer, Rebecca Sure, Stefan Grimme, Olav Schiemann,* Robert A. Flowers II,*
and Andreas Gansäuer*
Abstract: A catalytic system for titanocene-catalyzed epoxide
hydrosilylation is described. It features a straightforward
preparation of titanocene hydrides that leads to a reaction
with low catalyst loading, high yields, and high selectivity of
radical reduction. The mechanism was studied by a suite of
methods, including kinetic studies, EPR spectroscopy, and
computational methods. An unusual resting state leads to the
observation of an inverse rate order with respect to the epoxide.
E
poxide hydrosilylations are virtually unexplored reactions
that offer a simple access to alcohols.[1] Epoxidation and
titanocene-catalyzed[2] epoxide hydrosilylation (Scheme 1)
provide a two-step approach to the formal anti-Markovnikov
addition of H2O to olefins that circumvents hydroboration.
However, it also highlights the difficulties associated with the
reactive and relatively unstable titanocene(III) hydrides
([Cp2TiH] hereafter). Their generation is capricious and
requires either harsh conditions[3] or the syntheses of sensitive
complexes.[4] Herein, we report a reliable and easy procedure
for catalyst generation from titanocene dichlorides, mecha-
nistic key features of the reaction, and EPR data for the
species involved.
Scheme 1. Titanocene-catalyzed epoxide hydrosilylation.[1a]
DG298 values for the s-bond metathesis indicate that
[Cp2TiH] formation is favorable for R = Me, thermoneutral
for R = Allyl, and endergonic for R = Ph with PhSiH3.
(EtO)3SiH was predicted to be inferior. In practice,
[Cp2TiAllyl] is especially attractive since it can be prepared
in situ by stirring Cp2TiCl2 (1 equiv) and AllylMgBr
(2.2 equiv) in THF (1–10 min, purple reaction mixture).[5]
After the addition of PhSiH3 (1.5 equiv with respect to the
epoxide), the generation of [Cp2TiH] (the “allyl activation”)
is complete in 1–10 min (green reaction mixture).
Our starting point was the in situ generation of [Cp2TiH].
[Cp2TiR] complexes are attractive in this context because of
the weak Ti C bond.[5] We studied s-bond metathesis of these
À
complexes with PhSiH3 and (EtO)3SiH computationally
(Figure 1).[6]
Examples of the hydrosilylation are shown in Scheme 2.
The reaction of 1 with 3 leads to an excellent result. With
1 mol% 3, 97% 5 was obtained (86% 5 with 0.5 mol%
(Cp2TiOEt)2, that is 1 mol% of Ti, and 78% 5 with 1 mol%
Cp2TiMe2).[1a]
[*] D. S. G. Henriques,[+] K. Zimmer,[+] S. Klare,[+] E. Rojo-Wiechel,
Prof. Dr. A. Gansäuer
KekulØ-Institut für Organische Chemie und Biochemie
Universität Bonn
We then turned our attention to mechanistic issues and
investigated whether EPR-active species[7] are present in the
solution after “allyl activation”. To this end, the activation
Gerhard-Domagk-Straße 1, 53121 Bonn (Deutschland)
E-mail: andreas.gansaeuer@uni-bonn.de
M. Bauer, Dr. R. Sure, Prof. Dr. S. Grimme
Mulliken Center for Theoretical Chemistry
Institut für Physikalische Chemie und Theoretische Chemie
Universität Bonn
Beringstraße 4, 53115 Bonn (Deutschland)
A. Meyer, Prof. Dr. O. Schiemann
Institut für Physikalische Chemie und Theoretische Chemie
Universität Bonn
Wegelerstraße 12, 53115 Bonn (Deutschland)
Prof. Dr. R. A. Flowers II
Department of Chemistry, Lehigh University
Bethlehem, PA 18015 (USA)
[+] These authors contributed equally.
Figure 1. Computed DG at 298.15 K for [Cp2TiH] generation by s-bond
metathesis in THF in kcalmolÀ1. Values in parentheses refer to
(EtO)3SiH. For comparison, data for two dispersion-corrected density
functionals are given.
Supporting information for this article can be found under:
http://dx.doi.org/10.1002/anie.201601242.
Angew. Chem. Int. Ed. 2016, 55, 7671 –7675
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7671

Angewandte
Communications
Chemie
porting Information for details). Calculations show that this
process is strongly exergonic. The calculated hfc constants are
in qualitative agreement with the experimental values (see
the Supporting Information). The other signal shows a triplet
hfc structure with an hfc constant of 20.50 MHz and is
centered at giso,triplet = 1.9930. The hfc structure of both signals
is absent when PhSiD3 is used. For the species resulting in the
observed triplet, we propose a complex of solvated HMgBr
with Cp2TiH or its dimer (see Figure 2 and the Supporting
Information for details). The groups of Brintzinger and Mach
have postulated, isolated, and characterized closely related
Scheme 2. Titanocene-catalyzed epoxide hydrosilylation after “allyl acti-
vation” (“allyl activation”: Titanocene dichloride (1 equiv) and
AllylMgBr (2.2 equiv with respect to [Ti]) in THF for 1–10 min (purple
reaction mixture), then PhSiH3 (30–150 equiv with respect to [Ti],
1.5 equiv with respect to the epoxide) for 1–10 min (green reaction
mixture).
complexes.[5b,8] Therefore, all species detected contain Ti H
À
bonds.
To understand the mechanism of this reaction, a series of
rate experiments were performed on the hydrosilylation of 4
through “allyl activation” of 6 with PhSiH3. The reaction was
monitored in real time through in situ IR Spectroscopy or Vis
spectroscopy. Each experiment was repeated at least once.
Initial experiments were designed to examine the stability of
the catalyst under synthetically relevant conditions.[9,10] The
conversion of 4 into 5 in the presence of 10 mol% catalyst was
monitored by following [5] and subsequently plotting [4] as
a function of time (Figure 3). A higher catalyst loading was
was conducted under the standard reaction conditions with
either PhSiH3 or PhSiD3. The resulting X-band EPR spectra
and their simulations are shown in Figure 2. Two dominating
signals were observed when using PhSiH3. The first signal is
centered at giso,quintet = 1.9970 and shows a quintet hyperfine
coupling (hfc) structure with an hfc constant of 8.75 MHz. We
suggest that the observed quintet is the result of adduct
formation between PhSiH3 and Cp2TiH (Figure 2 and Sup-
Figure 3. “Same excess” experiment. Profile of Run 1 plotted as [4] vs.
time and the time-adjusted profile of Run 2 as [4] vs. adjusted time
(see the Supporting Information for details).
employed to ensure reproducible detection of the titanocene
species and to obtain fast turnover. When the reaction reaches
the first half-life with respect to [4], an equivalent amount of
PhSiH3 should be consumed and [Ti] should remain constant
under ideal conditions. A second run was initiated under
identical conditions to the first half-life of the first reaction.
When the concentration data for the second run are plotted
versus time and time-adjusted accordingly, the overlay is
good, showing only a very slight inhibition or deactivation at
the early stages of the reaction.[11]
The rate orders with respect to the catalyst, 4, and PhSiH3
were determined by using the “initial rate” method. The
reaction was approximately first order with respect to the
catalyst, but surprisingly showed inverse order with respect
to 4.
Figure 2. Proposed structures present in solution and corresponding
experimental (solid lines) and simulated (dashed lines) EPR spectra
after “allyl activation” at X-band (a) and Q-band (b). At X-band,
PhSiH3 (A) and PhSiD3 (B) were used during the activation. Free
energies are given in kcalmolÀ1
.
7672
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 7671 –7675

Angewandte
Communications
Chemie
The rate order with respect to the silane was established
Our calculations (Figure 4) show that [Cp2Ti-OEt] forms
by varying [PhSiH3] and monitoring [Cp2TiH] through in situ
Vis spectroscopy at 601 nm because of interfering silane
bands in the IR spectrum. The overlap of the graphs was
consistent with a rate order of zero with respect to the silane.
To confirm these findings, two further sets of experiments
were carried out. First, the rate order with respect to silane
was additionally determined using “different excess” experi-
ments (Supporting Information). The experiment examining
the impact of [PhSiH3] confirmed that the rate order with
respect to the silane is indeed zero. Second, an experiment on
the hydrosilylation of 7 (Scheme 3) also established an inverse
rate order with respect to the epoxide (Table 1).
a more stable epoxide complex than [Cp2Ti-H]. The opening
of the [Cp2Ti-OEt] epoxide complex to the b-metaloxy radical
is distinctly slower. In agreement with this prediction, we
found that [Cp2TiOEt]2 does not open 4 (see the Supporting
Information). Therefore, it seems that an epoxide complex of
titanocene(III) alkoxides is indeed the resting state of this
catalyst.
Clearly, the mechanism of epoxide hydrosilylation is
rather intricate. Neither epoxide opening, nor the intra-
molecular hydrogen atom transfer (HAT), nor s-bond meta-
thesis can be the rate-determining step. The formation of
a resting state of the catalyst that reversibly binds the epoxide
substrate without inducing ring opening must thus be
responsible for the observed reaction orders. A plausible
scenario is epoxide binding by Cp2TiOR (OR is formed
through epoxide opening followed by HAT, Scheme 1).
Figure 4. Computational analysis of epoxide complexation and open-
ing by [Cp2Ti-H] and [Cp2Ti-OEt]. Free energies are given in kcalmolÀ1
.
To validate this conclusion, a solution of Cp2TiAllyl was
mixed with PhSiH3 and 4 and immediately shock-frozen with
liquid N2 (Figure 5, black lines). Sample preparation was
repeated using PhSiH3 and 2,2-D2-4 (red lines), PhSiD3 and 4
(green lines), and PhSiD3 and 2,2-D2-4 (blue lines). The
frozen solutions were subjected to pulsed Q-band EPR and
electron nuclear double resonance (ENDOR) spectroscopy
(Figure 5a).[12] The EPR spectra of the different samples are
all similar and cover a field range of approximately 30 mT,
with intense maxima at g-values of 1.9857 and 1.9651 (see the
Supporting Information). The values are in the typical range
for titanocenes.[4,13] The spectra are the result of the presence
of several paramagnetic species.
Scheme 3. Titanocene-catalyzed epoxide hydrosilylation with 3 after
“allyl activation” in the presence of 1.5 equiv (second example
2.3 equiv) PhSiH3.
1
ENDOR spectroscopy of these samples yielded the H
hyperfine coupling patterns shown in Figure 5a. When 2,2-
D2-4 was used instead of 4, the proton resonance at an
approximate hfc constant of 2.5 MHz was no longer observed.
Similarly, resonances at an approximate hfc constant of
1.0 MHz were not detected with PhSiD3, although the
magnitude of the observed effect appeared to be weaker.
The smaller hfc constant indicates a larger electron–nucleus
separation. Finally, both resonances are suppressed when
both reagents are deuterated. Additionally, deuteron reso-
nances at corresponding hfc constants were observed when
deuterated reagents were used (Figure 5b). The observed
effects in the ENDOR spectra indicate that one of the ligands
at the TiIII center is the alkoxide formed after epoxide opening
and HAT. Whether or not an additional epoxide is also bound
to the TiIII center cannot conclusively be answered by using
ENDOR spectroscopy. However, the obtained spectra do not
Table 1: Rate orders with respect to the different reactants for the
titanocene-catalyzed hydrosilylation of 4 after “allyl activation”. Values in
parentheses refer to 7.
Reaction component
Rate order
[Ti]
PhSiH3
Epoxide 4 (and 7)
1.2Æ0.3[a]
0.1Æ0.4[b]
À1.4Æ0.3[c] (À1.3Æ0.2)[d]
[a] Conditions: Initial rate method. [Ti] 3.3–13 mm, 67 mm epoxide 4,
103 mm PhSiH3. [b] Owing to interfering silane bands in the IR spectrum,
this rate order was determined by in situ monitoring of the [Ti] species
through Vis spectroscopy. Conditions: Initial rate method. [Ti] 2 mm,
37.5 mm epoxide 4, 38.0–47.6 mm PhSiH3. [c] Conditions: Initial rate
method. [Ti] 6.7 mm, 67–267 mm epoxide 4, 103 mm PhSiH3. [d] Con-
ditions: Initial rate method. [Ti] 6.7 mm, 67–183 mm epoxide 7, 103 mm
PhSiH3.
Angew. Chem. Int. Ed. 2016, 55, 7671 –7675
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7673

Angewandte
Communications
Chemie
ments with theoretical methods, we have shown that the
catalytic cycle is rather unusual in having a resting state that
leads to an inverse reaction order with respect to the epoxide
(Figure 6).
Acknowledgements
We thank the DFG (SFB 813 “Chemistry at Spin Centers”
and Ga 619/11-1), the Jürgen Manchot Stiftung (fellowships
to S.K. and D.S.G.H.), and the National Science Foundation
(CHE-1123815, R.A.F.) for financial support.
Keywords: homogeneous catalysis · hydrogen-atom transfer ·
hydrosilylation · radicals · titanium
How to cite: Angew. Chem. Int. Ed. 2016, 55, 7671–7675
Angew. Chem. 2016, 128, 7801–7805
[1] a) A. Gansäuer, M. Klatte, G. M. Brändle, J. Friedrich, Angew.
Chem. Int. Ed. 2012, 51, 8891 – 8894; Angew. Chem. 2012, 124,
9021 – 9024; b) Y.-Q. Zhang, N. Funken, P. Winterscheid, A.
Gansäuer, Angew. Chem. Int. Ed. 2015, 54, 6931 – 6934; Angew.
Chem. 2015, 127, 7035 – 7038.
Figure 5. Q-band Davies ENDOR spectra of the protons[12b] (a) and
Mims ENDOR spectra of the deuterons[12c] (b) of the reaction mixture
using different degrees of deuteration. Black: no deuteration, red: 2,2-
D2-4, green: PhSiD3, blue: 2,2-D2-4, PhSiD3. The asterisks in (b) indi-
cate higher harmonics of the proton resonances.
[2] Recent examples of titanocene(III)-catalysis: a) X. Zheng, J. He,
H. H. Li, A. Wang, X. J. Dai, A. E. Wang, P. Q. Huang, Angew.
Chem. Int. Ed. 2015, 54, 13739 – 13742; Angew. Chem. 2015, 127,
13943 – 13946; b) J. Streuff, M. Feurer, G. Frey, A. Steffani, S.
Kacprzak, J. Weweler, L. H. Leijendekker, D. Krakert, D. A.
Plattner, H. Dietmar, J. Am. Chem. Soc. 2015, 137, 14396 –
14405; c) J. Streuff, A. Gansäuer, Angew. Chem. Int. Ed. 2015,
54, 14232 – 14242; Angew. Chem. 2015, 127, 14438 – 14448.
[3] a) C. Aitken, J. F. Harrod, J. Organomet. Chem. 1985, 279, C11 –
C13; b) C. A. Willoughby, S. L. Buchwald, J. Am. Chem. Soc.
1994, 116, 11703 – 11714; c) C. A. Willoughby, S. L. Buchwald, J.
Am. Chem. Soc. 1994, 116, 8952 – 8965; d) S. Xin, J. F. Harrod,
Can. J. Chem. 1995, 73, 999 – 1002; e) K. Rahimian, J. F. Harrod,
Inorg. Chim. Acta 1998, 270, 330 – 336; f) J. Yun, S. L. Buchwald,
J. Am. Chem. Soc. 1999, 121, 5640 – 5644.
contradict this interpretation. All of the computational and
experimental data thus hint at the epoxide complex of the
corresponding [Cp2TiOR] as the resting state of this catalyst.
Finally, we investigated the efficiency of our reaction for
the substrates shown in Scheme 3. In all cases, we obtained
high yields and diastereoselectivity (when applicable).
In summary, we have designed and developed an in situ
method for the generation of [Cp2Ti-H] from Cp2TiCl2 or
substituted titanocenes, AllylMgBr, and PhSiH3. The system
provides highly active and diastereoselective epoxide hydro-
silylation and constitutes an atom-economical radical reac-
tion.[14] By combining kinetic, EPR, and synthetic experi-
[4] E. Samuel, J. F. Harrod, D. Gourier, Y. Dromzee, F. Robert, Y.
Jeannin, Inorg. Chem. 1992, 31, 3252 – 3259.
[5] a) H. A. Martin, F. Jellinek, J. Organomet. Chem. 1967, 8, 115 –
128; b) H. H. Brintzinger, J. Am. Chem. Soc. 1967, 89, 6871 –
6877; c) E. Klei, J. H. Teuben, J. Organomet. Chem. 1980, 188,
97 – 107; d) E. Klei, J. H. Teuben, H. J. de Liefde Meijer, E. J.
Kwak, J. Organomet. Chem. 1982, 224, 327 – 339.
[6] The TURBOMOLE 7.0 program package[a] was employed for
gas-phase geometry optimizations and calculations of thermo-
statistical corrections at the TPSS[b]-D3[c,d]/def2-TZVP[e] level, as
well as single-point energies at the PW6B95[f]-D3 or B3LYP[g,h]
-
D3 level with the def2-QZVP[e] basis set. Solvation contributions
to DG were included through the COSMO-RS[i,j] continuum
solvation model. a) F. Furche, R. Ahlrichs, C. Hättig, W.
Klopper, M. Sierka, F. Weigend, WIREs Comput. Mol. Sci.
2014, 4, 91 – 100; b) J. Tao, J. Perdew, V. Staroverov, G. Scuseria,
Phys. Rev. Lett. 2003, 91, 146401; c) S. Grimme, J. Antony, S.
Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132, 154104; d) S.
Grimme, S. Ehrlich, L. Goerigk, J. Comput. Chem. 2011, 32,
1456 – 1465; e) F. Weigend, R. Ahlrichs, Phys. Chem. Chem.
Phys. 2005, 7, 3297 – 3305; f) Y. Zhao, D. G. Truhlar, J. Phys.
Chem. A 2005, 109, 5656 – 5667; g) A. D. Becke, J. Chem. Phys.
1993, 98, 5648 – 5652; h) P. J. Stephens, J. Devlin, C. F. Chaba-
lowski, M. J. Frisch, J. Phys. Chem. 1994, 98, 11623 – 11627; i) A.
Klamt, J. Chem. Phys. 1995, 99, 2224 – 2235; j) F. Eckert, A.
Klamt, AIChE J. 2002, 48, 369 – 385.
Figure 6. The catalytic cycle for epoxide hydrosilylation after “allyl
activation” of 6.
7674 www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 7671 –7675

Angewandte
Communications
Chemie
[7] Multifrequency Electron Paramagnetic Resonance, Theory and
Davies, Phys. Lett. A 1974, 47, 1 – 2; c) W. B. Mims, Proc. R.
Applications (Ed.: S. K. Misra), Wiley-VCH, Weinheim, 2011.
[8] a) H. Brintzinger, J. Am. Chem. Soc. 1966, 88, 4306 – 4307; b) V.
Kupfer, U. Thewalt, M. Horµcek, L. Petrusovµ, K. Mach, Inorg.
Chem. Commun. 1999, 2, 540 – 544; c) S. I. Troyanov, V. Varga,
K. Mach, J. Organomet. Chem. 1993, 461, 85 – 90; d) R. Gyepes,
K. Mach, I. Císarovµ, J. Loub, J. Hiller, P. Sindelµr, J. Organomet.
Chem. 1995, 497, 33 – 41.
[9] Selected reviews on RPKA: a) D. G. Blackmond, Angew. Chem.
Int. Ed. 2005, 44, 4302 – 4320; Angew. Chem. 2005, 117, 4374 –
4393; b) D. G. Blackmond, J. Am. Chem. Soc. 2015, 137, 10852 –
10866.
[10] Recent examples of RPKA in practice: a) A. J. Wagner, S. D.
Rychnovsky, Org. Lett. 2013, 15, 5504 – 5507; b) M. Scott, A. Sud,
E. Boess, M. Klussmann, J. Org. Chem. 2014, 79, 12033 – 12040;
c) N. R. Patel, R. A. Flowers II, J. Org. Chem. 2015, 80, 5834 –
5841; d) Y. N. Ji, R. E. Plata, C. S. Regens, M. Hay, M. Schmidt,
T. Razler, Y. P. Qiu, P. Geng, Y. Hsiao, T. Rosner, M. D.
Eastgate, D. G. Blackmond, J. Am. Chem. Soc. 2015, 137, 13272 –
13281; e) P. Ruiz-Castillo, D. G. Blackmond, S. L. Buchwald, J.
Am. Chem. Soc. 2015, 137, 3085 – 3092.
Soc. London Ser. A 1965, 283, 452 – 457.
[13] a) W. W. Lukens, Jr., M. R. Smith III, R. A. Andersen, J. Am.
Chem. Soc. 1996, 118, 1719 – 1728; b) A. Cangçnül, M. Behlen-
dorf, A. Gansäuer, M. van Gastel, Inorg. Chem. 2013, 52, 11859 –
11866.
[14] a) A. Gansäuer, A. Fleckhaus, M. A. Lafont, A. Okkel, K.
Kotsis, A. Anoop, F. Neese, J. Am. Chem. Soc. 2009, 131, 16989 –
16999; b) A. Gansäuer, M. Behlendorf, D. von Laufenberg, A.
Fleckhaus, C. Kube, D. V. Sadasivam, R. A. Flowers II, Angew.
Chem. Int. Ed. 2012, 51, 4739 – 4742; Angew. Chem. 2012, 124,
4819 – 4823; c) A. Gansäuer, C. Kube, K. Daasbjerg, R. Sure, S.
Grimme, G. D. Fianu, D. V. Sadasivam, R. A. Flowers II, J. Am.
Chem. Soc. 2014, 136, 1663 – 1671; d) A. Gansäuer, D. von Lau-
fenberg, C. Kube, T. Dahmen, A. Michelmann, M. Behlendorf,
R. Sure, M. Seddiqzai, S. Grimme, D. V. Sadasivam, G. D. Fianu,
R. A. Flowers II, Chem. Eur. J. 2015, 21, 280 – 289; e) A.
Gansäuer, S. Hildebrandt, A. Michelmann, T. Dahmen, D.
von Laufenberg, C. Kube, G. D. Fianu, R. A. Flowers II, Angew.
Chem. Int. Ed. 2015, 54, 7003 – 7006; Angew. Chem. 2015, 127,
7109 – 7112.
[11] a) R. D. Baxter, D. Sale, K. M. Engle, J.-Q. Yu, D. G. Black-
mond, J. Am. Chem. Soc. 2012, 134, 4600 – 4606; b) J. J.
Devery III, J. J. Douglas, J. D. Nguyen, K. P. Cole, R. A. Flower-
s II, C. R. J. Stephenson, Chem. Sci. 2015, 6, 537 – 541.
[12] a) A. Schweiger, G. Jeschke, Principles of electron paramagnetic
resonance, Oxford University Press, Oxford, 2005; b) E. R.
Received: February 3, 2016
Published online: April 29, 2016
Angew. Chem. Int. Ed. 2016, 55, 7671 –7675
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7675