Muonium Chemistry at Diiron Subsite Analogues of [FeFe]‐Hydrogenase

Abstract The chemistry of metal hydrides is implicated in a range of catalytic processes at metal centers. Gaining insight into the formation of such sites by protonation and/or electronation is therefore of significant value in fully exploiting the potential of such systems. Here, we show that the muonium radical (Mu.), used as a low isotopic mass analogue of hydrogen, can be exploited to probe the early stages of hydride formation at metal centers. Mu. undergoes the same chemical reactions as H. and can be directly observed due to its short lifetime (in the microseconds) and unique breakdown signature. By implanting Mu. into three models of the [FeFe]‐hydrogenase active site we have been able to detect key muoniated intermediates of direct relevance to the hydride chemistry of these systems.

Developing new approaches to gaining insight into catalytic systems is of central importance.W ea re now exploring the exciting possibility of using muonium radicals as surrogates for HC in the study of catalytic and electrocatalytic reactivity at metal centers.The system we have chosen is the active site of the [FeFe]-hydrogenase.
The[ FeFe]-and [NiFe]-hydrogenases catalyze the reversible reduction of protons to dihydrogen at high turnover frequencies and at low overpotentials.Onaper catalytic site basis,t he turnover frequencies of certain of these enzymes immobilized on electrodes can rival the best conventional electrocatalyst, carbon-supported platinum. [1][2][3] However,the high molecular mass and large geometric footprint of the native enzymes result in rather low current densities,a tb est ca. 3mAcm À2 at 20 8 8C. Given that hydrogen fuel or producer cells typically operate at current densities some two or three orders of magnitude greater,i ti sa rguable whether hydrogenase-based electrodes are likely to be useful materials in practical devices.N evertheless,i ti sw ell established that the relatively small metallosulfur centers within these enzymes are responsible for their high catalytic activity and this has prompted extensive research on synthetic analogues of these active sites. [1,3,4] This research is driven both by the need to provide mechanistic insights into the biological chemistry and the possibility of creating new materials for electrocatalysis based on abundant 3d metals.
Mechanistically,t he formation of more or less transient hydride intermediates is central to both hydrogen evolution or uptake by the [FeFe]-and [NiFe]-hydrogenases and electrocatalysis by synthetic analogues of their active sites. [5] We now describe the first application of muon spin spectroscopy to probe hydride chemistry at metallosulfur sites related to that within ahydrogenase,specifically the diiron subsite of [FeFe]-hydrogenase.
When an energetic muon (m + )beam passes through asolid sample some of the sub-atomic m + ions capture an electron to form muonium radicals (MuC); [6] these are sufficiently slowed to react with the bulk target species,i no ur case the diiron subsite analogue.M uonium can be considered chemically as acting as alight isotope of hydrogen atom, having about 1/9 of its mass;MuC thus provides asurrogate for hydrogen radicals. Thei mplantation of am uonium radical on ad iiron-subsite analogue can be viewed as the equivalent of concerted addition of ap roton and an electron;i tt akes place on the nano-to microsecond timescale giving rise to paramagnetic species. [6] Analogous to the proton, the muon has an uclear spin of one-half;a tt he end of its life (lifetime 2.197 ms) it decays to give apositron which is emitted preferentially along the spin direction at the moment of decay.B ecause it is possible to produce almost 100 %spin-polarized muon beams, detection of the direction of emission of the decay positrons allows the study of the evolution of the muon(ium) spin within the implanted sample.T his gives information on the system under scrutiny via the magnetic fields local to the muon site, mSR (muon spin rotation/relaxation/resonance) spectroscopy. [7] To date,t here have been only av ery small number of studies on application of mSR to organometallic systems. [8] These have been focused on silylenes, [9] or cyclopentadienyl or arene (half)sandwich systems; [10][11][12][13] in these compounds protonation and proton-coupled electron transfer are of limited relevance to the key chemical reactions of the metal complexes.I nc ontrast, the spectroscopic data we have so obtained can be modelled by ab initio DFT calculations and we show that this provides compelling evidence for the

Angewandte Chemie
Zuschriften muono-formyl Fe-C(Mu)=Oa nd bridging Fe-Mu-Feo r terminal Fe-Mu muonide transients,l ight isotopes of paramagnetic species that are implicit in hydrogen evolution and uptake. [1][2][3]14] As far as we are aware,t here are no studies using mSR spectroscopy to probe chemical processes relevant to metal hydrides in catalysis or electrocatalysis.Inthis work, we have examined three [FeFe]-hydrogenase subsite models,e ach of which illustrate ad ifferent aspect of hydride chemistry ( Figure 1). Thus,c omplex 1 is known to engage in electrocatalysis,i nw hich an electron and ap roton are added successively; [15] complex 2 protonates at the metal-metal bond [16] enabling subsequent electronation to yield am ixedvalence Fe(1.5)Fe(1.5) hydride (cf.H -atom addition); [17] whilst complex 3 possesses the bis-cyanide coordination found in the enzyme and has been shown to reconstitute an apoenzyme. [18][19][20] In the solid state (an anisotropic environment), coupling between the radical (electron) and muon spins gives rise to the D 1 transition. [6] This may be probed using "avoided level crossing muon spin resonance" (ALC-mSR). [6] ALC is al ongitudinal field mSR technique which detects the reduction in polarization at "level crossings" in the Breit-Rabi diagram.
For D 1 resonances,where only the muon spin changes sign, the resonance field is related to the hyperfine interaction by equation (1).
ALC-mSR spectra of 1-3 at 300 Ko ver the field range 2kGto18kGare shown in Figure 2. [21] Thegeneral features of all three spectra are remarkably similar:e ach room temperature spectrum exhibits abroad signal with amaximum in the range 8kGt o1 0kG. Thes ignals all show significant temperature dependence,with essentially complete loss of the signal at the lowest temperature (10 K) for 1 and 3,w hilst 2 exhibits only asmall residual signal.
Given the very broad nature of the ALC-mSR signals, accurate modelling of the background is essential. Scanning using acell containing amass of aluminum foil equivalent to the sample showed asmooth curve which could be fitted using af ourth-order polynomial ( Figure S1). Whilst the line shape for the solid-state anisotropic signals here is expected to be complex, it is possible to approximate the spectra using Gaussian curves following subtraction of the cell background ( Figure 3, Figures S5, S6). This process confirms the presence of ac ommon major signal at around 8.5 kG in all three samples.Asecond maximum at ca. 4.5 kG can clearly be seen at 300 Kf or 2 and at lower temperatures for 1 (Figure 2,  Figures S5, S6). Complex 3 does not show this signal.
Thes pecific nature of the muonium species which give rise to the ca. 8.5 kG and ca. 4.5 kG signals has been probed by computational simulation of plausible structures,f rom which D 0 values have been calculated. [22] Forthe known solidstate structure of 2 which has ab asal-basal deployment of phosphine ligands, [16] the high-field signal at 8.5 kG is consistent with the Mu being bound to iron either in abridging or terminal position (Figure 4). For 3,o nly the bridging muonide fits with the experimental value.The low-field signal at 4.5 kG can be accounted for by formation of aformyl-like radical ( Figure 5, left). We note that structures in which the rotation of at ripodal ligand group has occurred can also accommodate the observed signals ( Figure 5, right), but such

Angewandte Chemie
Zuschriften ar otation is perhaps unlikely in the solid state given the timeframe of the ALC-mSR experiment.
It is important to note that 2-Mu bridge is structurally analogous to the mixed-valence bridging hydride detected by electron paramagnetic resonance on one-electron reduction of the closed-shell cationic hydride,[ HFe 2 (pdt)(CO) 4 -(PMe 3 ) 2 ] + , [17] and that 1-Mu formyl can be regarded as an isotopomer of the formyl species observed upon reduction of 1 in presence of acid. [6,15] Figure 2s hows that the intensities of the signals are temperature-dependent. At low temperature (10 K) the response approaches that of the cell background. Figure 6 shows the Arrhenius plot for the major resonance in each of the three samples where the 10 Kdata is subtracted with the spectra pinned to zero at 14 kG (1)/18 kG (2 and 3). Fort he full range of data 2 and 3 there is an excellent linear correlation of ln I with 1/T. [23] Them ore limited data for 1 shows as imilar trend. Thee stimated activation energies obtained for the addition at the bridging position are 1 3.40-(2) kJ mol À1 , 2 2.45(2) kJ mol À1 and 3 1.58(3) kJ mol À1 ,c onsistent with the high reactivity of muonium. Theo rder obtained is notable,w ith activation energies falling as the systems become more electron rich: 1 is the least electron rich, 3 the most.
Thesecond resonance which we have assigned to Mu formyl in the data for the PMe 3 complex (2)remains resolved across the temperature range 100 Kto300 K. From the raw ln I and 1/T data we estimate an activation energy of ca. 9kJmol À1 from the Arrhenius relationship (correlation coefficient 0.965). In the lower temperature scans,t he intensity of the resonance is lost in the background.
As af urther verification that radical states are formed, repolarization experiments were also carried out (Figure 7, Figure S14). We clearly observed the recovery of polarization across the temperature range.W hilst such experiments confirm radical generation, extraction of hyperfine coupling constants is known to be difficult. [6] It is notable however that there is recovery at the lowest temperature (5 K), where the ALC-mSR spectra are essentially featureless.T hus,r adical states are still being formed at low temperature but likely broadened beyond detection in the ALC-mSR. This may be due to the existence of multiple sites of addition, consistent with the DFT calculations,and/or may be due to electron-spin relaxation.

Zuschriften
In summary,w eh ave provided evidence that muonium interacts in the solid state with the diiron subsite analogues 1-3 to give 35-electron bridging muonides (m-Mu)Fe 2 (pdt)-(CO) 4 (L) 2 .T he formation of these species is temperaturedependent with activation energies less than 4kJmol À1 .Inthe case L = PMe 3 (2)weobserved asecond resonance,assigned to the formation of am uono-formyl species,-COMu. The generation of these muoniated species have direct parallels in the protonation/electronation of substrates [17] and represent the first example of the application of mSR to electrocatalytic systems.
Thew ork described here is likely to presage aw ider application of muonium chemistry.The role of metal-hydride interactions in diverse inorganic,organometallic,and biological chemistry is extensive,r anging from b-elimination through water-gas shift chemistry,t he chemistry of the hydrogenases and nitrogenases,t oe lectrocatalytic systems for hydrogen fuel/producer cells.Muonium chemistry coupled with the fast timescale of mSR spectroscopy offers the prospect of unravelling mechanistic detail of such systems, for example,the possibility of detecting transient dihydrogen/ dihydride analogues.