Mechanistic Investigation of the Nickel-Catalyzed Transfer Hydrocyanation of Alkynes

The implementation of HCN-free transfer hydrocyanation reactions on laboratory scales has recently been achieved by using HCN donor reagents under nickel- and Lewis acid co-catalysis. More recently, malononitrile-based HCN donor reagents were shown to undergo the C(sp3)–CN bond activation by the nickel catalyst in the absence of Lewis acids. However, there is a lack of detailed mechanistic understanding of the challenging C(sp3)–CN bond cleavage step. In this work, in-depth kinetic and computational studies using alkynes as substrates were used to elucidate the overall reaction mechanism of this transfer hydrocyanation, with a particular focus on the activation of the C(sp3)–CN bond to generate the active H–Ni–CN transfer hydrocyanation catalyst. Comparisons of experimentally and computationally derived 13C kinetic isotope effect data support a direct oxidative addition mechanism of the nickel catalyst into the C(sp3)–CN bond facilitated by the coordination of the second nitrile group to the nickel catalyst.


General information
Unless otherwise stated, reagents were used as supplied from commercial sources without any further purification. Bis (1,5-cyclooctadiene)nickel(0), Ni(cod)2, was purchased from Strem and 2,2'-bis(diphenylphosphinomethyl)-1,1'-biphenyl (BISBI) from abcr GmbH. Both reagents were used as received and stored in a glovebox at -36 °C under an argon atmosphere. Solvents were dried using an LC Technology Solutions solvent purification system under an atmosphere of N2 (H2O content < 10 ppm, as determined by Karl-Fischer titration) and stored over molecular sieves.
All glassware was dried for at least one hour in an oven set at 100 °C prior to use. All reactions using Ni(cod)2 were carried out in 4 mL screw-cap vials and were set up under an argon atmosphere glovebox (LABmaster Pro SP, MBraun).
NMR: 1 H, 13 C, 19 F, and 31 P NMR spectra were recorded on a Bruker AVIII 400 MHz, a Bruker Neo 400 MHz or a Bruker Neo 500 MHz spectrometer and are reported in parts per million (ppm). 1 H NMR spectra are calibrated with respect to the corresponding residual solvent peak (CHCl3: 7.26 ppm, toluene: 2.08 ppm, THF: 1.72 ppm). 13 C NMR spectra were recorded with broadband 1 H decoupling and are calibrated with respect to the corresponding residual solvent peak ( 13 CDCl3: 77.16 ppm, d8toluene: 20.43 ppm, d8-THF: 67.21 ppm). Multiplet signals are reported as follows: s = singlet, d = doublet, t = triplet, q = quartet, quint = quintet, sept = septet, m = multiplet, br = broad, or combinations thereof. 13 C and 19 F signals are singlets unless otherwise stated. 2 H NMR spectra were recorded on a Bruker AVIII HD 500 MHz spectrometer and are reported in parts per million (ppm). 13 C Kinetic Isotope Effects at Natural Abundance: Quantitative 13 C spectra and inversion experiments were recorded on a Bruker AVIII 600 MHz spectrometer equipped with a DCH cryoprobe optimized for 13 C detection. The interscan delay (d1 + acquisition time) was set to 90 s. For each spectrum, 67 072 complex points were acquired. The data points were extended by zero-filling to give a spectral size of 262 144 points. The spectral width was 248.5 ppm and the transmitter position was set at 65.4 ppm (o1p).
Exponential apodization was applied with a line broadening of 2 Hz for inversion experiments and 0.5 Hz for quantitative 13 C experiments.
In each reaction, the recovered crude reaction mixture was compared with the starting material of the same lot. The NMR samples of the starting and recovered material were prepared identically in d8-toluene. All spectra were manually integrated using Mestrenova. Each individual peak was independently integrated three times with slightly different phasing and integration regions to account for processing and integration errors.

S3
Fourier Transformed Infrared spectroscopy (FT-IR) measurements were carried out using a Bruker INVENIO-R FT-IR Spectrometer equipped with a diamond ATR. Selected bands are reported. Bands arising from atmospheric CO2 are sometimes visible around 2361 cm -1 and 2335 cm -1 due to incomplete background subtraction.

2-Isopropylmalononitrile (2a)
A solution of malononitrile (3.30 g, 50.0 mmol) and acetone (7.30 mL, 100 mmol, 2.0 equiv) in i PrOH (100 mL) was cooled to 0 °C and NaBH4 (1.89 g, 50.0 mmol, 1.0 equiv) was added portionwise. The reaction mixture was stirred at 0 °C for 2 hours and was then acidified by addition of 1 M HCl solution. The aqueous phase was extracted with CH2Cl2 and the combined organic extracts were washed with brine, then dried over Na2SO4. After filtration, the solvents were removed under reduced pressure, and the residue was purified by flash column chromatography (SiO2, 0 -15% Et2O in pentanes) to afford the product 2a as a colorless oil (yield = 2.54 g, 47%).
Note: Prior to the use of the donor molecule in the catalytic transformations, the pure donor product was filtered over a plug of activated neutral alumina in the glovebox. 2-(p-Tolyl)malononitrile 2c was synthesized following a modified procedure. 7 In a flame-dried and nitrogen filled 100 mL Schlenk flask, K2CO3 (5.53 g, 40.0 mmol, 4.0 equiv), CuI (190 mg, 1.00 mmol, 10 mol%), (S)-proline (0.23 g, 2.0 mmol, 20 mol%) and malononitrile (1.98 g, 30.0 mmol, 3.0 equiv) were suspended in dry DMSO (20 mL). After the addition of 4-methyliodobenzene S6 (1.12 mL, 10.0 mmol), the flask was sealed and the reaction mixture was stirred for 18 h at 90 °C. After cooling to room temperature, the suspension was poured into aq. HCl (2 M, 100 mL). The aqueous layer was extracted with EtOAc three times.
The combined organic layers were washed with water and brine, dried over Na2SO4, filtered, and the solvent was removed under reduced pressure. The product was purified by column chromatography (SiO2, 0 -50% CH2Cl2 in hexanes) and obtained as a

Optimization of the transfer hydrocyanation reaction of alkynes
To an oven-dried 4 mL screw-cap vial, 2-isopropylmalononitrile 2a (27 mg, 0.25 mmol, 1.0 equiv), anhydrous toluene (0.05 mL), and 4-octyne 1a (37 L, 0.25 mmol) were added under an argon atmosphere in a glovebox. In another oven-dried vial, BISBI (14 mg, 0.025 mmol, 10 mol%) and Ni(cod)2 (6.9 mg, 0.025 mmol, 10 mol%) were dissolved in anhydrous toluene (0.45 mL) and the mixture was stirred until complete dissolution (solution turned dark red). The precatalyst was then added to the starting material in one portion. The vial was sealed and removed from the glovebox then heated at 100 °C for 18 hours. After cooling to room temperature, n-dodecane (20 μL) was added as an internal standard, the crude mixture was diluted with EtOAc, an aliquot (0.2 mL) was filtered through a plug of cellulose and then subjected to GC-FID analysis.

Synthesis on a 3-mmol scale
To an oven-dried 16 mL screw-cap vial, 2-isopropylmalononitrile 2a (389 mg, 3.60 mmol, 1.2 equiv), anhydrous toluene (0.48 mL), and 4-octyne 1a (0.44 mL, 3.0 mmol) were added under an argon atmosphere in a glovebox. In two separate oven-dried vials, BISBI (165 mg, 0.300 mmol, 10 mol%) was dissolved in anhydrous toluene (1.5 mL) and Ni(cod)2 (82.5 mg, 0.300 mmol, 10 mol%) was dissolved in anhydrous toluene (2.5 mL), respectively. After stirring at room temperature until complete dissolution, the nickel precatalyst and the ligand solution were combined and then added to the starting material mixture in one portion. The vial was sealed and removed from the glovebox then heated at 100 °C for 18 hours. After cooling to room temperature, the crude reaction mixture was purified by flash column chromatography (SiO2, 0-5% EtOAc in hexanes) to give the product 3a as a colorless oil (yield = 350 mg, 85%). were dissolved in anhydrous toluene (2.35 mL and 2.82 mL, respectively) and the mixtures were stirred until complete dissolution.

Unsuccessful substrates
Then the ligand and catalyst solution were combined and mixed, which led to a dark red precatalyst solution. Then 0.22 mL of the precatalyst solution was added to each of the 22 reaction vials. The vials were sealed and removed from the glovebox then heated at 100 °C for the indicated time. After cooling to room temperature, n-dodecane (20 μL) was added as an internal standard, the crude mixture was diluted with EtOAc, filtered through a plug of silica, and then subjected to GC-FID analysis to evaluate the reaction progress by determining the yield of the hydrocyanation product 3a. Under an argon atmosphere in a glovebox in an oven-dried 4 mL screw-cap vial, a stock solution was prepared containing 2- were dissolved in anhydrous toluene (2.35 mL and 2.82 mL, respectively) and the mixtures were stirred until complete dissolution.
Then the ligand and catalyst solution were combined and mixed, which led to a dark red precatalyst solution. Then 0.22 mL of the precatalyst solution was added to each of the 22 reaction vials. The vials were sealed and removed from the glovebox then heated at 100 °C for the indicated time. After cooling to room temperature, n-dodecane (20 μL) was added as an internal standard, the crude mixture was diluted with EtOAc, filtered through a plug of silica and then subjected to GC-FID analysis to evaluate the reaction progress by determining the yield of the hydrocyanation product 3a. added to each of the 22 reaction vials. The vials were sealed and removed from the glovebox then heated at 100 °C for the indicated time. After cooling to room temperature, n-dodecane (20 μL) was added as an internal standard, the crude mixture was diluted with EtOAc, filtered through a plug of silica and then subjected to GC-FID analysis to evaluate the reaction progress by determining the yield of the hydrocyanation product 3a.

b) Order in reagents
The reaction orders in catalyst, alkyne and donor substrate were determined by using both initial rate kinetics as well as variable time normalization graphical analysis (VTNA) described by Blackmond 12,13 and Burés 14 .

General procedure:
Under an argon atmosphere in a glovebox in an oven-dried 4 mL screw-cap vial, a stock solution was prepared containing 2-   The order in precatalyst was then analyzed by using initial rate kinetics. suggesting that the reaction rate is approximately first-order in catalyst concentration (b = 1.23 ± 0.02).

S18
The same data set was also analyzed using non-linear fitting parameters.  (D) Assuming a 2 nd -order dependence of the initial reaction rate on the catalyst concentration.
Due to catalyst degradation over time and the limited amount of data points, analysis of the data by VTNA is difficult and not appropriate as the active catalyst concentration can not easily be determined experiementally. 14 ii. Order in alkyne 1a:  The initial rate dependence on the alkyne concentration was then analyzed by using initial rate kinetics. suggesting that the reaction rate is approximately inverse first-order in alkyne concentration (b = -1.04 ± 0.43).

S21
The same data set was also analyzed using non-linear fitting parameters.

S22
Analysis of the kinetic data by the VTNA method.

Figure S8
VTNA of the transfer hydrocyanation of alkynes constructed from the data provided in Table S5, S9, and S10.

(A)
Assuming an inverse 2 nd -order of the initial reaction rate on the alkyne concentration. iii. Order in donor substrate 2a:  The initial rate dependence on the donor 2a concentration was then analyzed by using initial rate kinetics.

S24
The same data set was also analyzed using non-linear fitting parameters.

S25
Analysis of the kinetic data by the VTNA method.

Figure S11
VTNA of the transfer hydrocyanation constructed from the data given in Table S5, S11, S12.

Reaction #1
Under an argon atmosphere in a glovebox in an oven-dried 4 mL screw-cap vial, a stock solution was prepared containing 1Hcyclopentylmalononitrile 2b (220 mg, 0.150 mmol per reaction, 1.0 equiv), anhydrous toluene (0.55 mL), and 4-octyne 1a (0.36 mL, 0.23 mmol per reaction, 1.5 equiv). The mixture was then divided by taking each time 0.08 mL of the stock solution and adding it to oven-dried 4 mL screw-cap vials containing a magnetic stirring bar (set-up for ten reactions). In two separate vials equipped with stirring bars, BISBI (174 mg, 0.0150 mmol per reaction, 10 mol%) and Ni(cod)2 (86 mg, 0.015 mmol per reaction, 10 mol%) were dissolved in anhydrous toluene (2.1 mL and 2.52 mL, respectively) and the mixtures were stirred until complete dissolution. The ligand and catalyst solution were combined and mixed, which lead to a dark red precatalyst solution.
Then 0.22 mL of the precatalyst solution was added to each of the ten reaction vials. The vials were sealed and removed from the glovebox then heated at 80 °C for the indicated time. After cooling to room temperature, tetradecane (20 μL) was added as an internal standard, the crude mixture was diluted with EtOAc, filtered through a plug of cellulose and then subjected to GC-FID analysis to evaluate the reaction progress by determining the yield of the hydrocyanation product 3a.

Reaction #2
Under an argon atmosphere in a glovebox in an oven-dried 4 mL screw-cap vial, a stock solution was prepared containing 1D- Then 0.22 mL of the precatalyst solution was added to each of the ten reaction vials. The vials were sealed and removed from the glovebox then heated at 80 °C for the indicated time. After cooling to room temperature, tetradecane (20 μL) was added as an internal standard, the crude mixture was diluted with EtOAc, filtered through a plug of cellulose and then subjected to GC-FID analysis to evaluate the reaction progress by determining the yield of the hydrocyanation product 3a.  Figure S12 The KIE of the -H atom in the donor substrate was determined to be kH/kD = 1.20538  0.00007 considering 98% (determined by quantitative 1 H NMR analysis) of D incorporation in the donor substrate.

Possible explanation for observed secondary KIE:
As a deviation from unity for the KIE of the -H atom is observed, an effect of the -H atom on the nitrile activation step is proposed. Similar effects were previously reported in literature and referred to as hyperconjugation effects, 15  were dissolved in anhydrous toluene (2.1 mL and 2.52 mL, respectively) and the mixtures were stirred until complete dissolution.
The ligand and catalyst solution were combined and mixed, which led to a dark red precatalyst solution. Then 0.22 mL of the precatalyst solution was added to each of the ten reaction vials. The vials were sealed and removed from the glovebox then heated at 80 °C for the indicated time. After cooling to room temperature, tetradecane (20 μL) was added as an internal standard, the crude mixture was diluted with EtOAc, filtered through a plug of cellulose and then subjected to GC-FID analysis to evaluate the reaction progress by determining the yield of the hydrocyanation product 3a.

Reaction #2
Under an argon atmosphere in a glovebox in an oven-dried 4 mL screw-cap vial, a stock solution was prepared containing 6Dcyclopentylmalononitrile da-2b (220 mg, 0.150 mmol per reaction, 1.0 equiv), anhydrous toluene (0.55 mL), and 4-octyne 1a (0.36 mL, 0.23 mmol per reaction, 1.5 equiv). The mixture was then divided by taking each time 0.08 mL of the stock solution and adding it to oven-dried 4 mL screw-cap vials containing a magnetic stirring bar (set-up for ten reactions). In two separate vials equipped with stirring bars, BISBI (174 mg, 0.015 mmol per reaction, 10 mol%) and Ni(cod)2 (86 mg, 0.015 mmol per reaction, 10 mol%) were dissolved in anhydrous toluene (2.1 mL and 2.52 mL, respectively) and the mixtures were stirred until complete dissolution. The ligand and catalyst solution were combined and mixed, which led to a dark red precatalyst solution.
Then 0.22 mL of the precatalyst solution was added to each of the ten reaction vials. The vials were sealed and removed from the glovebox then heated at 80 °C for the indicated time. After cooling to room temperature, tetradecane (20 μL) was added as an internal standard, the crude mixture was diluted with EtOAc, filtered through a plug of cellulose and then subjected to GC-FID analysis to evaluate the reaction progress by determining the yield of the hydrocyanation product 3a.

Figure S14
The KIE of the -H atom in the donor substrate was determined to be kH/kD = 0.9167  0.0001 considering 65% (determined by quantitative 1 H NMR analysis) of D incorporation in the donor substrate.

d) 13 C kinetic isotope effect
Under an argon atmosphere in a glovebox in an oven-dried 4 mL screw-cap vial, a stock solution was prepared containing 2- complete dissolution. The ligand and catalyst solution were combined and mixed, which led to a dark red precatalyst solution.
Then 0.22 mL of the precatalyst solution was added to each of the two reaction vials. The vials were sealed and removed from the glovebox then heated at 100 °C for 7 hours. After cooling to room temperature, the crude mixture was filtered four times over a plug of cellulose using d8-toluene as an eluent. The filtered crude mixture was subjected to quantitative 13 C NMR measurements. 13 C measurements were carried out for a total of 4 reactions. The integral of the two methyl groups (Cint. std.) in each spectrum was set to 200. All spectra were integrated manually. For each run, the same batch of donor substrate was used and the standard results corresponding to this batch are shown in the Table S15. The raw data of each run was processed independently and integrated three times with slightly different phasing and integration regions to account for processing and integration errors.
Since all integrals are referenced internally, experimental errors introduced by instrumental instabilities of the NMR spectrometer can be neglected.
To evaluate the 13 C KIE of the transfer hydrocyanation of alkynes, the following considerations have been made according to Singleton and co-workers. 17 During the course of the reaction, the more slowly reacting isotopologue of the starting material accumulates if it is involved in the rate-determing step of the reaction. The following relation between the ratio of the abundance of this isotope in the recovered starting material (R) compared to its initial abundance (R0), the conversion of the starting material (F) and the relative rates for the two isotopologues (kinetic isotope effect: KIE) can be used. Based on this, the KIE can be calculated as follows.
The fractional conversion of the starting material (F) is defined by the amount of recovered starting material (Sr) and product (Pr).
The integral of the isopropyl methyl groups in the starting material 2a was used as internal reference and set to 200 (see above).
Thus the following expression could be derived, where IP,Me1 and IP,Me2 refer to the integrals of the two methyl groups of the alkenyl nitrile byproduct.
All integrations were performed manually. Every sample (run) was measured once. To monitor differences in relaxation behavior between samples, potentially leading to systematic changes in signal intensities (e.g. due to the presence of paramagnetic impurities in a subset of the samples), a single delay adiabatic 13 C inversion recovery experiment was carried out for each sample (recovery delay: 12 s). According to an initial determination of 13 C T1 values in starting material and byproduct (see Figure S16

d8-toluene)
The following peaks were considered in the evaluation of the recovered starting material as well as the alkenyl nitrile byproduct. Based on these assignments, R/R0 and F were calculated relative to the methyl peaks of 2a (Cint. stand.) after the reaction and in the initial donor batch, as it was assumed that these were not involved in the rate-determining step of the reaction.
Four independent runs were performed and the filtered samples were submitted to quantitative NMR analysis. First, the 13 C content of the initial donor batch was evaluated, then the 13 C content after the reaction was analyzed. For each sample, the raw data were independently processed and integrated three times to account for processing and integration errors. Based on these initial integrations, we also assumed that the error for the initial integration of the starting material can be neglected, and the mean initial content was used as a reference (R0), thus the following R/R0 values were calculated. Based on these results, the KIEcalc for the 13 C at natural abundance were calculated based on the equation shown above. These results were then used to determine the mean KIE and the standard deviation of the mean KIE. Below the integrated raw data are presented for the initial donor batch as well as all the four independent runs.

Figure S17
Quantitative 13 C NMR of the initial donor batch.

S36
Figure S18 13 C NMR spectrum of run 1. For peak assignments see above.
Figure S19 13 C NMR spectrum of run 2. For peak assignments see above.
S37 Figure S20 13 C NMR spectrum of run 3. For peak assignments see above.
Figure S21 13 C NMR spectrum of run 4. For peak assignments see above.

Theoretical derivation of the rate law for the transfer hydrocyanation of alkynes
The theoretical rate law for the transfer hydrocyanation of alkynes in the presence of nickel catalyst was derived by considering the following pathway (see Scheme d-1).

Scheme d-1
Overview of the proposed intermediates in the transfer hydrocyanation of alkynes.
Initially, the rate of the production formation can be described as follows, [D] and [A] refers to donor and alkyne concentration respectively.
As the rate-determining step of the reaction is the C-CN bond activation (k2), the concentration of intermediate 10 and 11 can be assumed to be neglectable. Therefore, this assumption can be made: The rate of the reaction can then be expressed as follows: Assuming 5 is the resting state of the reaction, the following steady-state approximation can be applied (k1 < k-1).

Synthesis of [(BISBI)Ni(cod)]
In a glovebox under an argon atmosphere, Ni(cod)2 (6.9 mg, 0.025 mmol, 1.0 equiv) and BISBI (14 mg, 0.025 mmol, 1.0 equiv) were dissolved in d8-toluene (0.5 mL) in an NMR tube. Then the tube was sealed with a rubber septum. The NMR tube was removed from the glovebox and sonicated for 2 minutes to ensure complete dissolution. Then the NMR spectra were recorded. X-ray quality crystals were obtained from this saturated toluene solution kept at room temperature overnight.
Then the NMR spectra were recorded. X-ray quality crystals were obtained by layering the saturated toluene solution with hexane and slow diffusion and evaporation at room temperature over several days in the glovebox under an argon atmosphere.

Large-scale reaction:
In an oven-dried 50 mL Schlenk tube equipped with a magnetic stirring bar, Ni(cod)2 (100 mg, 0.364 mmol, 1.00 equiv) and BISBI (200 mg, 0.364 mmol, 1.00 equiv) were dissolved in toluene (10 mL). Then 4-octyne (1.10 mL, 0.750 mmol, 2.06 equiv) were added. The flask was sealed and removed from the glovebox. The mixture was stirred at room temperature for 30 minutes. The reaction mixture was filtered into another 50 mL Schlenk tube using a filter cannula.
The toluene was evaporated until approx. 1 mL of toluene was left. Then dried and deoxygenated hexane (10 mL) was added, resulting in the formation of a precipitate. The crude mixture was vigorously stirred for 10 minutes. Stirring was stopped and the residual solvent was filtered into another flask using a filter cannula. Hexane was added again to the filtrate (2 x 10 mL) and was again filtered into the separate flask. All hexane washes were combined and concentrated under reduced pressure using Schlenk technique. The residual beige solid was dried under high vacuum and transferred into the glovebox to yield the product as a light-yellow solid (134 mg, 51%).

Formation of [(BISBI)Ni(donor)] complex 6:
In a glovebox under an argon atmosphere, Ni(cod)2 (6.9 mg, 0.025 mmol, 1.0 equiv) and BISBI (14 mg, 0.025 mmol, 1.0 equiv) were dissolved in d8-toluene (0.5 mL) in an NMR tube. Then 2-isopropylmalononitrile 2a (5.4 mg, 0.050 mmol, 2.0 equiv) and 4octyne 1a (3.75 L, 0.025 mmol, 1.0 equiv) were added. The tube was sealed with a rubber septum. The NMR tube was removed from the glovebox and sonicated for 2 minutes to ensure complete solvation. Then the NMR spectra were recorded. When mixing the nickel pre-catalyst with the diphosphine as well as the HCN donor, the formation of four distinct peaks in the 31 P{ 1 H} NMR can be detected. The sharp peak refers to the [(BISBI)Ni(COD)] complex while the three broader peaks were tentatively assigned to a side-on (symmetric) and a potential end-on donor coordination complex (unsymmetric, see above ).

b) VT-NMR analysis of standard reaction conditions
In a glovebox under an argon atmosphere, Ni(cod)2 (15 mg, 0.055 mmol, 20 mol%) and BISBI (30 mg, 0.055 mmol, 20 mol%) were dissolved in d8-toluene (0.50 mL) in an NMR tube. Then 2-isopropylmalononitrile 2a (30 mg, 0.28 mmol, 1.0 equiv) and 4octyne 1a (61 L, 0.42 mmol, 1.5 equiv) were added. The tube was sealed with a rubber septum and parafilm. The NMR tube was removed from the glovebox and sonicated for 5 minutes to ensure complete dissolution. Then the NMR spectra were recorded at the given temperatures.

c) Attempts to identify oxidative addition complexes
In initial experimental attempts, the identity of the proposed oxidative addition complexes [(BISBI)Ni(CN)(alkyl)] was examined.
Stoichiometric experiments were run to either detect the formation of these complexes in-situ by 1 H or 31 P{ 1 H} NMR analysis or by crystallization to identify any of these complexes via single-crystal X-ray analysis.
The identity of any additional signals hinting towards the formation of oxidative addition complexes could not be detected in the 31 P{ 1 H} spectra, while the formation of byproduct 4a was confirmed by 1 H NMR analysis.

Isolation of oligomeric metal cluster
In a glovebox under an argon atmosphere, Ni(cod)2 (6.9 mg, 0.025 mmol, 1.0 equiv), BISBI (14 mg, 0.025 mmol, 1.0 equiv), 2isopropylmalononitrile 2a (2.7 mg, 0.025 mmol, 1.0 equiv) were dissolved in d8-toluene (0.50 mL) in an NMR tube. The NMR tube was sealed, removed from the glovebox, and sonicated for 2 minutes. Then the NMR tube was placed in an oil bath and heated at 80 °C for several hours. After cooling down to room temperature, the NMR tube was kept at room temperature under an inert atmosphere for several days until the formation of crystals was observed, which were subsequently submitted to X-ray analysis.
After cooling down to room temperature, the NMR tube was kept at room temperature for several days under an inert atmosphere until the formation of crystals was observed, which were subsequently submitted to X-ray analysis.
In another oven-dried vial, BISBI (28 mg, 0.050 mmol, 10 mol%), and Ni(cod)2 (14 mg, 0.050 mmol, 10 mol%) were dissolved in d8-toluene (1.30 mL) and the mixture was stirred until complete dissolution (solution turned dark red). The precatalyst solution was then added to the starting material in one portion. The vial was sealed and removed from the glovebox then heated at 100 °C for 18 hours. After cooling to room temperature, 1,1,2,2,-tetrachloroethane (53 μL) was added as an internal standard, and the mixture was filtered over a plug of silica into an NMR tube. The tube was sealed and subjected to NMR analysis of the crude reaction mixture.

Figure S29
1 H NMR of the crude reaction mixture after adding 1 equiv of TEMPO to the reaction (400 MHz, CDCl3). The peak at 5.95 ppm corresponds to the internal standard.
When TEMPO was used as a radical scavenger in the transfer hydrocyanation of alkynes, the formation of the desired transfer hydrocyanation product could not be observed.
S47 Figure S30 1 H-NMR of the crude reaction mixture after using 1 equiv of BHT in the reaction (400 MHz, CDCl3). The peak at 5.96 ppm corresponds to the internal standard.
When BHT was used as a radical scavenger in the transfer hydrocyanation of alkynes, the formation of the desired transfer hydrocyanation product could be observed. Based on the internal standard, 42% of hydrocyanation product 3a along with 1 equiv of the dehydrocyanation byproduct 4a were obtained.
Based on these experiments, the formation of any radical intermediates during the transfer hydrocyanation reaction could not be identified.

Computational details a) General information
Unless stated otherwise, all proposed intermediates were based on X-ray structures or modelled in ChemCraft 18 . For the structures in the main text conformer searches were run with CREST version 2.11 19,20 and xTB version 6.4.0. 21,22 All subsequent calculations were run with ORCA 5.0.3. 23 Geometry optimizations and frequency calculations were run with the PBE0 functional, 24 the def2-TZVP 25 basis set for Ni, and the def2-SVP 25 basis set for all other atoms. Grimme's atom-pairwise dispersion correction with the Becke-Johnson damping scheme was used, 26,27 as was the RIJCOSX algorithm, 28 and the def2/J auxiliary basis set 29 was chosen. The cpcm solvent model was used for toluene. The temperature for the frequency calculations was set to 373.15 K.
Transition states were located from relaxed potential energy surface scans and were confirmed by the number of imaginary frequencies (Nif=1) and by either an IRC calculation or by manual displacement of the imaginary frequency. Structures 8 and TS6 have one additional small imaginary frequency, but due to the magnitude, ~3i in both cases, this should not greatly influence the energy of the system. The depicted structures in the main text and the SI were visualized with CYLView20. 30 The cartesian coordinates of all optimized structures are given in a separate text file.

b) Activation of 3-methylbutanenitrile donor
To assess the stabilizing effect of the second nitrile group, we computed the reaction energy profile for the activation of the donor with only one nitrile. The proposed intermediates were based on X-ray structures or modelled in ChemCraft, starting from the structures obtained in the main computational study. Geometry optimizations, frequency calculations, and single point calculations were run with the same level of theory as the calculations in the main computational study.
S49 Figure S31 Calculated reaction pathway for the activation of the 3-methylbutanenitrile donor.

c) Kinetic isotope effect
The kinetic isotope effects were calculated using PyQuiver, 31  We used the Wigner KIE for our study (for clarity, the non-corrected and the inverted parabola KIEs that are typically also printed are omitted from the table above), and all the listed KIEs are referenced to "isotopologue C_gamma1", whose absolute KIE is 1.0011.
As there are two nitriles in the donor which cannot be distinguished spectroscopically, the apparent KIE that is derived experimentally cannot be directly compared to the calculated KIE. To enable such a comparison between the apparent (experimental) KIE and the individual calculated KIEs for the two nitriles, a number of assumptions were made: • We do not account for doubly labeled 13 C-donor as the natural abundance for this case is very low, the results would not differ significantly, and we thus assume that the error introduced by this is negligible.
• As the barrier for the subsequent β-H elimination is significantly lower (by more than 5 kcal/mol) than the reverse reaction (reductive elimination to reform the malononitrile) and the overall transformation is strongly exothermic, we assume that there is a negligible reversible character of C-CN bond activation step.
The rates for the different isotopologues of the donor are defined below.
The rate of the light isotopologue is simply k12.
The rate of the heavy isotopologue is the sum of the rate when the 13 CN is the reacting nitrile (CCN2) and the rate when the 13 Inserting the values obtained from PyQuiver gives the apparent KIE for the nitriles: