Iron-Catalyzed H/D Exchange of Primary Silanes, Secondary Silanes, and Tertiary Siloxanes.

A synthetic study into the catalytic hydrogen/deuterium (H/D) exchange of 1° silanes, 2° silanes, and 3° siloxanes is presented, facilitated by iron-β-diketiminato complexes (1a and 1b). Near-complete H/D exchange is observed for a variety of aryl- and alkyl-containing hydrosilanes and hydrosiloxanes. The reaction tolerates alternative hydride source pinacolborane (HBpin), with quantitative H/D exchange. A synthetic and density functional theory (DFT) investigation suggests that a monomeric iron-deuteride is responsible for the H/D exchange.

Deuterium labeling is of fundamental importance in medicinal[1−5] and synthetic chemistry.[6−10] Deuterosilanes are frequently deployed as reagents for isotopic labeling, including but not limited to, hydrosilylation reactions,[11−16]ortho-silylations,[17] reductions of carbon-halogen[18,19] and carbon–carbon multiple bonds,[20−22] and are a key mechanistic probe when investigating the deuterium kinetic isotope effect of a reaction. Deuterosilanes are traditionally generated by the reaction of hazardous NaBD4 or LiAlD4 with corresponding chlorosilane, generating stoichiometric quantities of waste metal salts.[23] With increasing demand for deuterium-labeled products, catalysis provides a promising solution because of its inherently improved selectivity, functional group tolerance, and reduced waste.[24−26] Therefore, finding catalysts that facilitate the hydrogen/deuterium (H/D) exchange of hydrosilanes is desirable.

Thus far, catalytic deuteration of silanes has been limited to precious metals[27−40] and photocatalysis.[26,41] Furthermore, most of these reports are limited to the H/D exchange of 3° silanes. Three examples are known where activity is retained for 1° and 2° silanes, and with the exception of Carmona and co-workers’ study, these are limited to aromatic hydrosilanes.[32,33,38] In 2010, Carmona and co-workers reported the H/D exchange of silanes catalyzed by a cationic rhodium complex and D2 (Scheme a, left). This study included 1° and 2° silanes PhSiD3, Ph2SiD2, and Et2SiD2, all reaching ≥99% H/D exchange. Three loadings of D2 were required to achieve full deuterium incorporation. In 2011, Nolan and co-workers described an iridium-catalyzed H/D isotopic exchange of hydrosilanes with D2 (Scheme a, middle). The reactivity of one 2° silane was reported, Ph2SiD2, achieving 95% deuterium exchange. In 2017, Apeloig and co-workers reported silane deuteration catalyzed by a simple platinum complex and D2 (Scheme a, right). 1° and 2° silanes were limited to PhSiD3 and MePhSiD2, both achieving 90% H/D exchange. Finding a consistent method for the catalytic H/D exchange of 1° and 2° silanes remains a challenge. Furthermore, to the best of our knowledge, a method utilizing earth-abundant metals has not been reported.

Scheme 1

(a) Catalysts That Facilitate H/D Exchange of 1° and 2° Silanes Are Based around Precious Metals (Rh, Ir, Pt). (b) This Work Focuses on an Fe(II) Precatalyst (1a) for the Deuteration of a Range of 1° and 2° Hydrosilanes

The vast natural abundance of iron makes it a cheaper and more sustainable alternative to precious metals.[42−44] Its reduced toxicity makes its use in the pharmaceutical industry appealing.

The potential for the reactive, three-coordinate, iron-β-diketiminato complex (dippBDKFeCH2TMS, 1a) to facilitate silane H/D exchange was evidenced during our previous study of dehydrocoupling.[45] When 1a, aniline, and methylphenylsilane were exposed to 1 atm of D2, H/D exchange was observed at silicon in the generated silazane. Herein, we report the first example of iron-catalyzed H/D exchange of silanes (Scheme b).

Based on our investigation into dehydrocoupling of silanes with amines, we began by reacting MePhSiH2 (2a-H) with D2 in the presence of 5 mol % 1a under 1 atm of D2 (Table , Entry 1).[45] After 16 h, almost complete consumption of the Si–H peak is observed in the 1H NMR spectrum. Deuteration was confirmed by the introduction of the corresponding peak in the 2H NMR spectrum and the appearance of a quintet in the 29Si NMR spectrum, arising from 1JSi-D coupling to two quadrupolar deuterium atoms. Nuclear magnetic resonance (NMR) analysis reveals a deuterium incorporation of 92%. 1a is required to catalyze the H/D exchange (Entry 2). Lewis-acidic Fe(acac)2 is unable to affect deuteration (Entry 3) while FeCl2·(THF)1.5, dippBDKFe(μ-Cl)2Li(THF)2, and dippBDK (precursors to 1a) are unable to impart any H/D exchange (Entries 4–6). Replenishing deuterium in the vessel after 16 h gives no improvement in deuterium incorporation (Entry 7). Finally, optimal deuterium incorporation is achieved when filling the ampoule over liquid N2, achieving 97% after 16 h (Entry 8). We estimate that this is equivalent to 4 atm. D2 in the vessel.

Table 1

Optimization of Silane Deuterationa

entry	catalyst	D incorporation (%)b	spec. yield (%)c
1	1a	92	93
2	none	0	0
3	Fe(acac)2	0	0
4	FeCl2·THF1.5	0	0
5	dippBDKFe(μ-Cl)2 Li(THF)2	0	0
6	dippBDK	0	0
7d	1a	87	96
8e	1a	97	95

Conditions: 60 mL ampoule containing methylphenylsilane (0.25 mmol), catalyst (5 mol %), D2 (after free-pump-thaw cycle), C6H6 or C6D6 (0.5 mL), RT, 16 h.

Determined by 1H NMR spectra comparing residual Si–H to Si–CH3 or mAr–H after vacuum distillation and 2H NMR spectroscopy.

Spectroscopic yield determined by 1H NMR spectroscopy by comparing Si–CH3 or mAr–H to 1,3,5-trimethoxybenzene (TMB, 0.25 mmol) as the internal standard after vacuum distillation.

Freeze-pump-thawed after 16 h, refilled with D2 (1 atm), and stirred for a further 4 h.

D2 filled over liquid nitrogen (4 atm).

With the optimal conditions in hand (Table , Entry 8), we began exploring the substrate scope (Table ). 1° and 2° aromatic silanes react well under the optimized conditions, with deuterium incorporation of 95 and 97%, respectively (Entries 2 and 3a). After a 10-fold scale-up, 2c is isolated in 92% yield with corresponding deuterium incorporation of 96% (Entry 3b). The reaction also tolerates the aliphatic hydrocarbon solvent (pentane) with deuterium incorporation of 97% (Entry 3c). 1° and 2° aliphatic deuterosilanes 2d, 2e, and 2f are generated with 98, 96, and 96% deuterium incorporation, respectively (Entries 4 to 7). 3° siloxane undergoes H/D exchange with 98% deuterium incorporation, while poly(methylhydrosiloxane), PMHS, undergoes quantitative deuteration to generate 2g and 2h, respectively. Notably, the selectivity toward 1° and 2° hydrosilanes compliments previous investigations into metal-catalyzed H/D exchange; sterically demanding di-tert-butylsilane and 3° silanes are not tolerated under our optimized conditions (see the Supporting Information).

Table 2

Substrate Scopea

Conditions: 60 mL ampoule containing silane (0.25 mmol), catalyst (5 mol %), D2 (filled over liquid N2, 4 atm), C6H6 or C6D6 (0.5 mL), RT, 16 h.

Determined by 1H NMR spectroscopy comparing residual Si–H to C–H.

Determined by 1H NMR spectroscopy using 1,3,5-trimethoxybenzene (0.25 mmol) as the internal standard.

2.5 mmol scale, 300 mL ampoule.

Pentane (0.5 mL) instead of C6H6.

Complete loss of the Si–H signal observed by 1H NMR spectroscopy, spectroscopic yield determined by 2H NMR spectroscopy relative to toluene-d8.

Determined by 1H, 11B, and11B{1H} NMR spectroscopy.

In an attempt to expand the scope to include bulky 2° and 3° silanes in our substrate scope, sterically less encumbered 1b was synthesized (dmpBDKFeCH2TMS, Figure ). Screening of multiple crystals reveals a mixture of mono- and dinuclear 1b. Previously, 1b had been prepared and isolated as the four-coordinate tetrahedral THF adduct, which had showed decreased reactivity compared to 1a.[46]

Figure 1

Single-crystal X-ray structure of the dimeric form of 1b. Hydrogen atoms have been omitted and dmp groups represented in wireframe mode, for clarity. Ellipsoids are represented at 30% probability. Symmetry operations: 1 1 – x, 1 – y, 1 – z; 2 1 – x, y, z; 3x, 1 – y, 1 – z.

Employing 1b as a precatalyst in the H/D exchange of 2b shows that the level of deuteration is maintained (96% D, 78% spectroscopic yield, compared to 95% D, 97% spectroscopic yield using 1a). Despite this, 1b is unable to affect the deuteration of di-tert-butylsilane and 3° silanes. Finally, the protocol can be extended to HBpin, with 2i deuterated quantitively as determined by 1H and 11B NMR spectroscopy (Table , Entry 9). Synthesis of 2i is notoriously challenging, requiring complex reaction setup[47] or precious metal catalysts.[40,48−50] Only two reports of catalytic deuteration of HBpin are known utilizing earth-abundant metal complexes.[51,52] The H/D exchange of alternative commercially available boranes, catecholborane (HBcat), and 9-borabicyclo(3.3.1)nonane (9-BBN) was also investigated. The reaction of 1a and 1b with HBcat yields only partial H/D exchange. However, the reaction of 1a and 1b with 9-BBN leads to the formation of a pink solution and no H/D exchange. Performing this reaction stoichiometrically with two equivalents of 9-BBN yields complexes 3a and 3b, respectively. Single crystal X-ray diffraction reveals the formation of bridging hydride complexes in the solid state (Scheme ). It is plausible that complexes 3a and 3b arise from the sequestration of hydridic iron species.

Scheme 2

Reaction of Complexes 1a and 1b with 9-BBN Result in the Formation of Bridged Hydride Complexes 3a and 3b, Respectively

Hydrogen atoms (with the exception of those bonded to boron centers) are omitted while dmp and dipp groups are shown in wireframe mode, for clarity. Ellipsoids are represented at 30% probability. Symmetry operation (3a): 1 x, 1/2 – y, z

Our efforts then turned to elucidating the mechanism for the iron-catalyzed silane H/D exchange. First, the reversibility of the reaction was examined (Scheme a). Under 4 atm of H2, 2a is converted to 2a-H, indicating that the reaction is highly reversible and dependent on excess D2 in the system. The stoichiometric reaction between phenylsilane and 1a or 1b reveals no change in the 1H NMR spectrum. However, under catalytic conditions with a large excess of silane, reactivity is observed. In situ 1H NMR reaction monitoring experiments using 1a or 1b both show rapid consumption of PhSiH3 between 0.0 and 0.5 h (P1). The rate of PhSiH3 consumption then reduces (P2), before accelerating again until the reaction reaches completion (P3). These rate trends are reproducible, and the process is accompanied by the introduction of a peak at 0.04 ppm, assigned to PhSiH2(CH2TMS), A.[53,54] This suggests a reaction between 1a or 1b, and PhSiH3 occurs during P1 to generate an iron-hydride intermediate. Previous work has shown that 1a does not react with H2 at 20 atm, indicating that D2 is not involved in the initial catalyst activation and further supporting the assertion that precatalyst activation involves reaction with silane.[55] Furthermore, isolable complexes 3a and 3b provide evidence for the cleavage of the precatalyst Fe–C bond with a hydride source and the generation of an iron-hydride intermediate. The production of PhSiH2(CH2TMS) gives an approximation of the percentage catalyst activation (Scheme b, bottom). Interestingly, only partial activation of complex 1a occurs, reaching 20% after 16 h. Conversely, complex 1b activates to a greater extent, reaching 80% after 16 h. We postulate that the reduced steric bulk of the 2,6-dimethylphenyl flanking groups within complex 1b facilitates the reaction with PhSiH3. This improved activation is mirrored by an increase in the rate of H/D exchange for 1b over 1a, reaching 25 and 9% H/D exchange after 14 h, respectively. A lag phase follows (P2). Holland and co-workers have shown previously that [(BDK)FeH]2 complexes readily exchange with D2 to form [(BDK)Fe(D/H)]2.[56] We suggest that P2 arises from dimerization and H/D exchange at the iron center. The synthesis of complex 4a and use in a standard deuteration reaction (Scheme c)[45,56] show 97% deuterium incorporation. This indicates that complex 4a is an active species in catalysis. Reaction monitoring experiments were undertaken using 4a as the catalyst. Importantly, no peak at 0.04 ppm is observed, and the reaction proceeds rapidly with no activation or lag phase (P1 and P2 are not observed). Furthermore, the rate of reaction is significantly improved when catalyzed by 4a instead of 1a, reaching 20 and 10% H/D exchange after 16 h, respectively [Scheme b, top (for full reaction profile, see the Supporting Information, Figure S1)]. The concentration of 4a was varied to determine the order in catalyst (see the Supporting Information, Figures S10 and S12). The reaction is predicted to be 0.5th order with respect to 4a.[57] This indicates that the monomeric form of 4a participates in the rate-determining step, requiring the dissociation of the [(BDK)Fe(D/H)]2 dimer. Monitoring the reaction between 4a and PhSiH3 at various temperatures generates an Eyring plot with ΔS = −164 ± 28.2 J mol–1 K–1 (see the Supporting Information, Figure S7). We therefore hypothesize that the reaction proceeds via a highly ordered four-membered transition state, in a σ-bond metathesis-type process.

Scheme 3

(a) Deuterium Incorporation Is Reversible with a Hydrosilane Formed from a Deuterosilane. (b) Reaction Monitoring Shows That There Are Three Distinct Steps When 1a (triangle) and 1b (square) Are Employed As Precatalysts, Whereas the Use of Iron Hydride Dimer 4a (diamond) Shows a Steady Increase in the Product (Top Chart, Measured As a Fraction of 0.25 mmol), 1b Undergoes Higher Levels of Activation To Form the Corresponding Dimer (4b, See the Supporting Information) Compared to 1a (Bottom Chart). (c) 4a Gives Near-Quantitative H/D Exchange

Next, we studied the ease of catalyst activation and H/D exchange using density functional theory (DFT) (Figure , see the Supporting Information for full computational details). 1b was chosen as the model precatalyst. The transformations were modeled on the quintet and triplet energy surface to investigate any possible spin-crossover pathway. The triplet surface was calculated to be much higher in energy and is therefore not discussed further here (see the Supporting Information for details).[58] The activation of with PhSiH3 proceeds with an appreciable barrier of +18.5 kcal mol–1 () via an associative σ-bond metathesis reaction. is generated with a small thermodynamic gain of −0.2 kcal mol–1. The large barrier to catalyst activation, accompanied by a small thermodynamic driving force, agrees with the slow activation and lack of stoichiometric reactivity observed experimentally for 1a and 1b. Therefore, it is unsurprising that excess PhSiH3 is required for reactivity to be observed. Next, using the experimentally obtained Eyring data as a foundation, the H/D exchange of PhSiH3 was investigated based on the highly ordered transition state, . H/D exchange occurs with a barrier of +10.3 kcal mol–1. The reduced barrier of compared to agrees with the experimentally observed increased rate of reaction for 4a, compared to 1a (Scheme b). Sequential deuteration steps to access phenyl(silane-d2) and phenyl(silane-d3) proceed with negligible thermodynamic gain. This indicates that any secondary isotope effect is insignificant (for the full energy-level diagram, see the Supporting Information, Figure S74). These findings suggest that under catalytic conditions, reversible precatalyst activation competes with a fast H/D exchange process.

Figure 2

DFT-derived free energy profile for catalyst activation and H/D exchange. Energies are calculated at the B3PW91-D3BJ/Def2-TZVP/IEF-PCM(C6H6)//BP86/BS1-level of theory. All energies are reported in kcal mol–1 and referenced to and reactants.

Based on our mechanistic findings, we propose the following mechanism (Scheme ). Slow reaction of C1 with silane generates C2 via sequential σ-bond metathesis and dimerization. C2 can activate D2, generating C3.[56] Dissociation of C3 to monomeric C4 facilitates the silane H/D exchange, regenerating C2 after dimerization. The partially deuterated silane may re-enter the catalytic cycle until complete deuteration is achieved.

Scheme 4

Postulated Catalytic Cycle Based on Experimental and Computational Results

We have reported the first iron-catalyzed H/D exchange of hydrosilanes and hydrosiloxanes. Excellent deuterium incorporation was observed for a variety of 1° and 2° silanes and 3° siloxanes. The conditions also tolerated HBpin, reaching quantitative deuterium incorporation. Mechanistic studies revealed that activity could be enhanced with the replacement of 2,6-diisopropyl (1a) flanking groups with 2,6-dimethyl (1b). Iron-hydride complex 4a is suggested to be the active catalyst in the system. The order in the catalyst and Eyring analysis suggest that the dissociation of hydride dimer 4a is required to facilitate silane H/D exchange. DFT studies reveal a slow and reversible catalyst activation competes with silane deuteration.