Lewis acid protection turns cyanide containing [FeFe]-hydrogenase mimics into proton reduction catalysts

Sustainable sources of hydrogen are a vital component of the envisioned energy transition. Understanding and mimicking the [FeFe]-hydrogenase provides a route to achieving this goal. In this study we re-visit a molecular mimic of the hydrogenase, the propyl dithiolate bridged complex [Fe2(μ-pdt)(CO)4(CN)2]2−, in which the cyanide ligands are tuned via Lewis acid interactions. This system provides a rare example of a cyanide containing [FeFe]-hydrogenase mimic capable of catalytic proton reduction, as demonstrated by cyclic voltammetry. EPR, FTIR, UV-vis and X-ray absorption spectroscopy are employed to characterize the species produced by protonation, and reduction or oxidation of the complex. The results reveal that biologically relevant iron-oxidation states can be generated, potentially including short-lived mixed valent Fe(i)Fe(ii) species. We propose that catalysis is initiated by protonation of the diiron complex and the resulting di-ferrous bridging hydride species can subsequently follow two different pathways to promote H2 gas formation depending on the applied reduction potential.


SI_3 Treatment of complex 3 2− with DCl
To a 3 mM solution of complex 3 2− in acetonitrile, 2 equivalents of DCl were added. There was no difference in the FTIR of complex 4 − and the isotopologue (i.e., 3 2− [-D]).

SI_4 Determination of protonation rate constant by stopped flow rapid mixing FTIR
Experimental procedure Stopped-flow rapid-mixing rapid-scan FTIR was used to monitor the protonation reaction of 3 2− to form 4 − . In the glovebox, one syringe was filled with a 0.1 mM solution of 3 2− in acetonitrile, making sure to remove all gas bubbles from the syringe. The second syringe was loaded with a solution of HCl. The HCl solution had been diluted from 1 M HCl in diethyl ether to yield x mM solutions. Where x = 0.4, 0.6, 0.8 or 1.6. The FTIR cell was filled with acetonitrile. The stopped flow apparatus was removed from the glovebox and brought to the FTIR spectrometer. The FTIR cell was put in the sample chamber which was subsequently purged with nitrogen for 10 minutes before collecting a background spectrum.
Next, the syringes were manually compressed using the push-plate. The contents of both syringes travelled to the FTIR cell and were mixed. The contents of the FTIR cell from the previous measurement were transferred to the waste syringe pushing the trigger so that the control software could start the rapid scan measurement. Due to the large size of the sample syringes compared to the waste syringe, it was possible to measure up to 5 repeats from the same syringes.  (5) and this pushes the trigger plate (6) which then triggers the instrument software to measure IR spectra for monitoring of the reaction. The deadtime of the instrument is ~50 ms and the time resolution is also ~50 ms.

Baseline treatment
A baseline shift is observed during the experiment. Thus, prior to kinetic analysis we performed a baseline treatment for each individual carbonyl band. In order to carry out the baseline treatment, a carbonyl band was selected to be analysed, for example 1956 cm -1 . The two points equidistant on either side of the peak at 1956 cm -1 were also selected. A representative example using the peak at 1956 cm -1 is shown in Fig. S8, using 1975 cm -1 and 1937 cm -1 for baseline estimates.

Kinetic analysis
Attempts to fit the resulting data for a pseudo first order chemical reaction were unsuccessful. Consequently, we carried out a second order analysis to determine a second order rate constant.
Where x = 4, 6, 8 or 16 Where: This equation was inserted into the fitting function in the Origin program. SI figure 8 (right) shows that the difference between the first and second points is negligible. Therefore the first point was removed in order to improve the reliability of the fit. Only the initial 8 points of the reaction were fitted in order to plot the initial rate.
This analysis was carried out for each carbonyl band of 3 2− (i.e. 1922, 1956 and 1988 cm -1 ) and for each concentration of HCl (i.e 4mM, 6 mM, 8 mM, 16 mM). Each measurement was repeated at least twice.
Fit data Figure S9. Three representative time traces taken from the stopped flow data and used to extract second order rate constants. The rate constants for all of the data collected are given below in Table  S5.  Figure S10. Cyclic voltammogram of ferrocene, showing how the internal reference potential was determined and also demonstrating that the well-known reversible couple has a peak separation of 81 mV in acetonitrile in the employed system. The mean value of the two peak voltages corresponds to a midpoint potential of 0.29 V vs. Ag/AgCl.

SI_6 Randles Sevcik analysis of oxidation of complex 4 −
For the Randles-Sevcik analysis of the oxidation of 4 − at −0.48 V vs Fc +/0 , blank CVs at various scan rates were collected as background data and CVs of the complex were also collected in a scan window of −0.2 -0.8 V vs. Fc +/0 ). The background CVs were subtracted from the CVs of complex 4 − .
For each CV, the peak and trough currents were recorded (ip,ox and ip,red) and plotted against the square root of the scan rate. This gave a plot that was linearly dependent on the square root of the scan rate. The slope of the linear fit and the Randles-Sevcik equation ( In a graph of ip vs (scan rate) ½ the slope represents the ratio of the two variables:

SI_7 Trumpet plot analysis of oxidation of complex 4 −
A Trumpet plot was used to calculate the standard rate constant for heterogeneous electron transfer (ks). The peak and trough potentials of the oxidation wave of 4 − were recorded at various scan rates (Ep,ox and Ep,red). Values for Ep,ox/red -E½ were plotted against the log of the scan rate (log(v)). This plot was overlaid with a working curve generated in the electrochemistry fitting software (Digielch) using Dsim = 1 x10 -5 cm 2 s -1 and ks,sim = 1 cm s -1 . The x-axes of the two plots were shifted until the y-axes overlap. At this point the following holds true Λs,Fe = Λs,sim, where Λs is a dimensionless parameter defined in the following equation  Figure S15. FTIR spectra demonstrate that only 1 eq of AgNO3 is required to oxidise 3 2− (5 mM, thick blue spectrum) to 5 (thick magenta spectrum). As Ag + is added to 3 2− (thick blue spectrum), the FTIR signal begins to decrease and shifts by 40 cm -1 to higher wavenumbers, approximately 46 % of the signal amplitude at 1956 cm -1 is lost. P a g e 17 | 25 SI_11 Reduction of 5 by NaBH4 or CoCp* observed by FTIR spectroscopy Figure S16. FTIR spectra in MeCN show that treatment of 5 (magenta spectrum) with NaBH4 (left; green spectrum) or CoCp* (right; purple spectrum) results in partial recovery of complex 3 2− . After treatment with NaBH4 and CoCp*, only 32 % of the original carbonyl signal of 3 2− is recovered. Figure S17. EPR spectra of 0.25 mM 3 2− (black spectrum), mixed with AgNO3 at room temperature to make 5 (red spectrum), and mixed with AgNO3 at -70 o C to yield a species with an isotropic signal with g = 2.022 (green spectrum), tentatively attributed to 3 -. Spectra were measured at 10 K and 1 mW microwave power. The signal at ~317 mT is attributed to a small amount of an unknown EPR active substance in the starting material (3 2− ).

SI 13 EPR spectra of 4 − and CoCp* mixed at -40 o C, power and temperature dependence, and simulations
Mixing of 4 − with CoCp* at -40 o C resulted in a complicated EPR spectrum. At least two signals are observed in the g≈2 region, denoted (rhombic) g1,2,3 = 2.039, 2.015 and 2.004 and (axial) g⊥= 2.033, g‖ = 2.027, as determined through simulations (Fig. S20). The amplitudes of both components show a maximum at 10 K and both diminish in a similar fashion at elevated temperatures (Fig. S18). The power dependence of the amplitudes of 1-3 (as indicated in Fig. S19, panel A) is shown in Fig. S19, panels B-D.        Figure  4 (main text). *Coordination numbers were fixed in the simulations, ms refers to a multiple-scattering shell, the given error sum, RF, was calculated for 1-3 Å of reduced distance in the Fourier-transforms *Coordination numbers were fixed in the simulations, ms = multiple scattering, the error sum, R F , was calculated for 1-3 Å of reduced distance.
.  Table S4 (the legend on the right annotates respective interatomic interactions; ms, multiple scattering).