Transition metal hydride complexes serve as essential intermediates in a myriad of chemical reactions that produce fuels and commodity chemicals. In particular, transition metal hydride complexes play a key role in a variety of reactions, including hydrogenation catalysis, C-H bond activation, N 2 reduction, and CO 2 reduction reactions. [1][2][3][4] The formation of transition metal hydride complexes using visible light energy offers the opportunity to use solar energy to drive many of society's most valuable chemical transformations. Specifically, the chemical bonds formed upon reaction of a transition metal hydride complex with a substrate will encapsulate some of the energy of the photon that initiated the entire process. 5,6 Thus, gaining a better understanding of light-driven hydride formation provides a means to develop future transition metal catalysts for efficient production of solar fuels.
There are two general approaches to initiate the formation of transition metal hydride complexes with light. The first approach utilizes the excitation of a photoreductant or photoacid to drive a proton-coupled electron transfer (PCET) reaction with a ground state transition metal complex to afford the target transition metal hydride complex. 5,6 A handful of examples that make use of photoreductants 3,7 and photoacids 8 to drive the production of transition metal hydride complexes with light have been reported. The second approach employs a photoactive transition metal complex that when excited, reacts with an acid and a reductant in its ground state to generate a transition metal hydride. Examples of coordination complexes which, upon photoexcitation, are reduced and protonated to generate metal hydride complexes are much rarer. 9,10 [CpM(CO) 3 ] 2 (M = Cr, Mo, W; Cp = cyclopentadienyl) complexes are known to undergo photochemical disproportionation reactions when irradiated with visible light in the presence of Lewis basic ligands (L) to yield monomeric, 18e -disproportionation products [CpM(CO) 3 L] + and [CpM(CO) 3 ] -. [11][12][13][14][15][16] Two mechanisms for the photochemical disproportionation of [CpM(CO) 3 ] 2 complexes have previously been considered, a radical chain mechanism (Scheme 1A) and an in-cage disproportionation mechanism (Scheme 1B). Based on bulk photolysis studies of [CpM(CO) 3 ] 2 (M = Mo) varying the identity of the Lewis base and measuring quantum yields, Tyler and coworkers concluded that the disproportionation reaction occurred via a light-initiated radical chain pathway proceeding through a CpM(CO) 3 L • 19e -intermediate. [11][12][13] In this mechanism, absorption of a visible light photons results in homolysis of the metal-metal σ bond to form two 17e -radical CpM(CO) 3
• complexes (Scheme 1A). The geminate 17e -radical species can quickly recombine, 14 but upon cage escape the 17e - species can diffuse into bulk solution and coordinate a Lewis base (such as an amine, nitrile, or phosphine) to form a highly reducing 19e -radical species CpM(CO) 3 L • . . Subsequent M-M bond rupture of the anionic dimer produces the 18e -anion [CpM(CO) 3 ] -and another equivalent of the 17e -radical CpM(CO) 3
. 11 Tyler's observation of quantum yield measurements greater than unity (under some conditions) and observed correlations between quantum yield and I -1/2 (I = photon flux on the sample) are consistent with this radical chain mechanism. 11 Transient absorption experiments by Cahoon et al. on the picosecond timescale provide evidence for in-cage disproportionation of the 17e -radicals CpM(CO) 3
• (M = W) formed upon photo-induced homolysis of the [CpM(CO) 3 ] 2 dimer (Scheme 1B). 14,15,18 Coordination of a phosphine or phosphite ligand to one of the caged radicals yields the 19e -intermediate CpM(CO) 3 L • which subsequently transfers an electron to the co-caged 17e -radical CpM(CO) form in the solvent cage. 14 Intrigued by the light-driven production of [CpM(CO) 3 ] -, we hypothesized that under the appropriate conditions, this anionic species would react with a proton source to produce the corresponding CpM(CO) 3 H species (Scheme 2). Indeed, we find that photolysis of a dimeric tungsten complex [CpW (CO) 3 ] 2 in the presence of pyridinium tetrafluoroborate ([PyH][BF 4 ]), using acetonitrile (CH 3 CN) as both the Lewis base and solvent, leads to the near quantitative formation of the tungsten hydride complex CpW(CO) 3 H (with 1 equiv. CpW (CO) 3 H generated per [CpW(CO) 3 ] 2 dimer). The photolysis reaction was monitored in situ using 1 H NMR spectroscopy; analysis shows that the overall rate law is acid concentration inde-
Dalton Transactions pendent. Quantum yield measurements as a function of photon flux suggest the primary reaction pathway toward tungsten hydride formation is a radical chain mechanism likely accessed by the 19e -radical, CpW(CO) 3 L • , that escapes the solvent cage, providing new insight to the dominant mechanism of radical chain disproportionation.
All syntheses were performed under N 2 , either on a Schlenk line or in an inert-atmosphere glovebox. Synthesis of CpW(CO) 3 H, 19 triphenylmethyl radical 20 (Moses Gomberg's dimer), and pyridinium tetrafluoroborate 21 were performed according to literature methods. Acetonitrile (Fisher Scientific, HPLC grade, >99.9%), diethyl ether (Fisher Scientific, >99%), and dichloromethane (Fischer Scientific, >99%) were degassed with argon and dried using a Pure Process Technology solvent system. Deuterated acetonitrile (Cambridge Isotope Laboratories, >99.8%) was stirred over CaH 2 for 24 hours, degassed using a freeze-pump-thaw technique, vacuum transfered to a stauss flask, and stored under the inert-atmosphere glovebox. All NMR spectra were recorded on either a VIII 500 MHz spectrometer with a standard 5 mm broad band probe (characterization) or a Bruker NEO 600 MHz spectrometer with a QNP cryoprobe or with a BBFO probe (reaction monitoring). 1 H NMR spectra were acquired using an externally calibrated 30 degree pulse width and referenced to residual proteo solvent chemical shifts. UV-vis absorption measurements were taken on an Agilent Cary 60 UV-vis spectrophotometer using 1 cm path length quartz cuvettes.
Preparation of [CpW(CO) 3 ] 2 was adapted from literature procedures. [22][23][24] Within an inert atmosphere glovebox, in separate scintillation vials, recrystallized CpW(CO) 3 H (0.140 g, 0.421 mmol, 2.05 equiv.) and trityl bromide (0.100 g, 0.205 mmol, 1 equiv.) were dissolved in neat diethyl ether and then combined to yield a cloudy, red solution. After stirring for 30 min, the diethyl ether solvent was removed by vacuum and the resulting maroon powder was rinsed several times with diethyl ether for purification (0.079 g, 0.119 mmol, 61% yield). 1 H NMR (500 MHz, CD 3 CN) δ 5.56 (s, 10H, gauche), 5.46 (s, 10H, anti). The anti and gauche isomers are present in a 5 : 3 ratio.
Preparation of [CpW(CO) 3 (NCCH 3 )][PF 6 ] was adapted from literature procedures. 23,25,26 Within the inert atmosphere glovebox, recrystallized CpW(CO) 3 H (0.080 g, 0.24 mmol, 1.1 equiv.) and acetonitrile (5 equiv.) were combined in dichloromethane. A solution of tritylium hexafluorophosphate (0.085 g, 0.22 mmol, 1 equiv.) in dichloromethane, was added to the CpW(CO) 3 H solution to yield a bright orange solution instantaneously. After stirring for 20 min, solvent was removed by vacuum. The resulting orange powder was rinsed several times and subsequently recrystallized in dichloromethane and diethyl ether. Bright orange crystals suitable for X-ray diffraction were obtained (0.093 g, 0.18 mmol, 80% yield). 1 H NMR (500 MHz, CD 3 CN) δ 6.05 (s, 5H), 2.57 (s, 3H).
Preparation of Na[CpW(CO) 3 ] was modified from literature procedures. 27,28 CpW(CO) 3 H (0.050 g, 0.15 mmol, 1.2 equiv.) and sodium hydride (0.003 g, 0.013 mmol, 1 equiv.) were dissolved in neat acetonitrile to yield a transparent and colorless solution. After stirring for 25 min, solvent was removed by vacuum. The resulting white powder was washed 3 times with diethyl ether (0.038 g, 0.11 mmol, 76% yield) 1 H NMR (500 MHz, CD 3 CN) δ 5.11 (s, 5H).
All DFT and TD-DFT calculations were carried out using the program package Gaussian 29,30 and the B3LYP density functional. 30 The basis set used for all other atoms besides tungsten atoms was 6-31G**. 31 For the two tungsten atoms, the LANL2DZ basis set, which includes a relativistic effective core potential, was used for all calculations. 32 Hessian matrices were calculated at each stationary point in order to ensure that true minima along the potential energy hypersurfaces had been found.
To monitor photolysis in situ, we utilized a New Era photoNMR sampling device equipped with a concentric tube whose interior held a heat resistant, 0.6 mm polymer optical fiber connected to a Prizmatrix ultra high brightness LED lamp centered at a wavelength of 450 nm (FWHM = 25 nm). The fiber coating was removed for the length for the NMR tube and the fiber itself scuffed with sandpaper only in the region of the detection coil in the probe; this allows the entire length of the sampling device to be irradiated. The light intensity of the LED light is modulated using an analog inputa ninety-nine turn dialfor precise power control with LED power stabilizer. The dial was set to a power setting of 3 for all in situ photoreactions. All samples were prepared in a nitrogen filled glovebox as 1 mL solutions with biphenyl as an internal standard (0.130 mmol, 20 mg) and the reagents specific to that reaction. Samples containing [CpW(CO) 3 ] 2 were filtered to remove any undissolved dimer. A New Era photoNMR sampling device was loaded with 600 μL of solution. One spectrum was obtained prior to turning the light on, followed by 59 single scan spectra with the light on, unless specified otherwise. The D 1 time for all experiments was set to 60 s with a fixed delay of 0 s between scans.
All quantum yield measurements were completed with a ThorLabs multi-wavelength LED lamp (455 nm, FWHM = 18 nm) set-up on the lab bench. The lamp was turned on an hour before irradiation to allow the power to stabilize. All samples were prepared in a nitrogen filled glovebox.
Where I λ is the photon flux of the 455 nm ThorLabs LED lamp source on our sample tube (mol photon per s) at a given current output determined by actinometry with potassium ferrioxalate (see ESI † for additional details), t is time of irradiation, and A λ is the absorbance of the sample at the irradiation wavelength (ε 455 = 1240 M -1 cm -1 ). [CpW(CO) 3 ] 2 (Cp = cyclopentadienyl) was synthesized as previously reported. [22][23][24] Based on previous work investigating the photochemistry of [CpM(CO) 3 ] 2 (M = Cr, Mo, W) and the work reported herein, the anticipated products of the photolysis of [CpW(CO) 3 ] 2 in the presence of a proton source are CpW(CO) 3 H and [CpW(CO) 3 (NCCH 3 )] + . Thus for comparison purposes, authentic samples of CpW(CO) 3 H and [CpW(CO) 3 (NCCH 3 )][PF 6 ] were prepared. 19,25,26 [CpW(CO) 3 ] 2 , CpW(CO) 3 H and [CpW(CO) 3 (NCCH 3 )][PF 6 ] were all characterized by 1 H NMR spectroscopy to confirm product identity. [CpW(CO) 3 ] 2 has a diagnostic feature for the cyclopentadienyl ring, with anti (5.46 ppm) and gauche (5.56 ppm) isomers observed in a 5 : 3 ratio. The Cp resonance for CpW(CO) 3 H is observed downfield of the dimer (5.63 ppm) and the species also has a diagnostic resonance for the hydride (-7.44 ppm). For comparison, the Cp peak of [CpW(CO) 3 (NCCH 3 )][PF 6 ] is downfield of the dimer and the hydride (6.05 ppm). X-ray quality crystals of [CpW (CO) 3 (NCCH 3 )][PF 6 ] (Table S1 †) were obtained by recrystallization in dichloromethane and diethyl ether. Structural characterization confirms identity of this cationic product (Fig. 1).
Next, the stability of [CpW(CO) 3 (NCCH
The UV-Vis absorption spectrum of [CpW(CO) 3 ] 2 exhibits two absorption bands, one at 488 nm (ε = 2017 M -1 cm -1 ) and one at 356 nm (ε = 15 791 M -1 cm -1 ) (Fig. 2). Wrighton and Ginley 33 assigned these transitions as dπ → σ* and σ → σ* transitions, respectively, similar to assignments of [M(CO) 5 ] 2 (M = Mn, Re) by Levenson, Gray, and Ceasar. 34 Time-dependent density functional theory (TD-DFT) calculations of [CpW (CO) 3 ] 2 were used to calculate excitations for [CpW(CO) 3 ] 2 in acetonitrile using implicit solvation (Fig. 2). The dominant low energy feature at 488 nm is attributed to a dπ → σ* transition, and the intense absorption feature at 356 nm is best characterized by a σ → σ* transition, in agreement with Wrighton and Ginley's original assignments. Both transitions populate the ] were obtained using a New Era photoNMR tube with an LED light source introduced to the tube with a fiber optic cable that enables in situ NMR monitoring of photochemical reactions (λ ex = 450 nm). The concentrations of [CpW(CO) 3 ] 2 , CpW(CO) 3 H, [CpW(CO) 3 (NCCH 3 )] + , and [CpW(CO) 2 (NCCH 3 ) 2 ] + were monitored over the photolysis period with reference to the internal standard, biphenyl. The decay of the reactant and formation of the product are observed to occur with first-order kinetics (Fig. 4).
There are no reaction intermediates detected via 1 H NMR upon photolysis of [CpW(CO) 3 ] 2 in the presence of [PyH][BF 4 ]. This observation suggests that on the NMR timescale, [CpW (CO) 3 ] -reacts rapidly with acid to form CpW(CO) 3 H, as anticipated. These data, along with the first-order kinetics observed for both [CpW(CO) 3 ] 2 consumption and CpW(CO) 3 H formation, suggest that the rate-determining step is not [CpW (CO) 3 ] -protonation, but a step involved in the disproportionation reaction. To examine this hypothesis further, the photochemical reaction kinetics of [CpW(CO) 3 ] 2 were measured at 62.4, 67.9, 131, and 264 mM [PyH][BF 4 ] at a fixed photon intensity (Fig. S9 †). Initial rates of CpW(CO) 3 H formation were determined to be independent of acid concentration, further supporting the conclusion that the rate limiting step for CpW (CO) 3 H formation is not [CpW(CO) 3 ] -protonation (Fig. S9 †).
A 30 min bulk photolysis (λ = 455 nm) of [CpW(CO) 3 ] 2 with [PyH][BF 4 ] generates CpW(CO) 3 H in 110% yield (Fig. 3), determined via reference of the 1 H NMR signals to an internal standard. The observed yield of 110% is greater than anticipated for disproportionation of the radical species followed by protonation of the resulting [CpW(CO) 3 ] -, indicating a distinct pathway to form CpW(CO) 3 H exists outside those shown in Schemes 1 and 2. Further, the disproportionation mechanism predicts the combined yield of the cationic products [CpW (CO) 3 (NCCH 3 )] + and [CpW(CO) 2 (NCCH 3 ) 2 ] + should also yield 100% assuming no competing pathways are present. However, their combined yield is only 32.9%.
The observation of a CpW(CO) 3 H yield greater than unity suggests that additional [CpW(CO) 3 ] -is formed beyond the 1 equiv. produced from disproportionation. Further, our finding that the cationic species [CpW(CO) 3 (NCCH 3 )] + and [CpW (CO) 2 (NCCH 3 ) 2 ] + are formed in sub stoichiometric yield suggests these products react further under photolysis with (2.2 × 10 -3 mmol), to 0.040 × 10 -3 mmol of CpW(CO) 3 H is formed and 0.55 × 10 -3 mmol of the starting [CpW (CO) 3 (NCCH 3 )][PF 6 ] is lost, presumably to form unidentified decomposition products (Fig. S7 †). The mass loss and observation of decomposition products by 1 H NMR spectroscopy indicates that these products have limited photostability.
To ensure thermal decomposition was not contributing to the observed reactivity, [CpW(CO) 3 ] 2 was simultaneously heated in the presence of [PyH][BF 4 ] and the absence of light ([CpW (CO) 3 ] 2 left in ambient light reacts slowly, Fig. S3 †). The sample was heated at 80 °C for 12 h, 1 H NMR spectrum indicated no reaction and retention of [CpW(CO) 3 ] 2 . This data confirms that photons, not incidental heating, are responsible for dimer disproportionation and subsequent hydride formation (Fig. S1 †). Notably, no carbonyl ligands are lost under these conditions. The photolysis of [CpW(CO) 3 ] 2 (1.4 × 10 -3 mmol, 2.8 × 10 -3 mmol W atoms) was also conducted in the absence of acid to better understand the reaction before protonation and to ensure that CpW(CO) 3 H is protonated by [PyH][BF 4 ] and not an adventitious proton source. When the sample was irradiated for 30 min with 455 nm light, the anticipated products [CpW(CO) 3 (NCCH 3 )] + , [CpW(CO) 2 (NCCH 3 ) 2 ] + and [CpW(CO) 3 ] - were detected via 1 H NMR spectroscopy (Fig. S4 †). Surprisingly, CpW(CO) 3 H was also observed. We determined that [CpW(CO) 3 ] -, [CpW(CO) 3 (NCCH 3 )] + , [CpW (CO) 2 (NCCH 3 ) 2 ] + , and CpW(CO) 3 H are present in the photolysis reaction mixture in 0.65 × 10 -3 , 0.16 × 10 -3 , 0.21 × 10 -3 , and 0.50 × 10 -3 mmol yields respectively, 0.26 × 10 -3 mmol of [CpW(CO) 3 ] 2 are left unreacted, and 1.0 × 10 -3 mmol W atoms are unaccounted for. This yields a net ratio of the tungsten anion plus hydride to cationic species of 3.1 : 1. The sub stoichiometric yield of cationic products is qualitatively consistent with what is observed in the presence of [PyH][BF 4 ]. The appearance of [CpW(CO) 3 ] -in the 1 H NMR spectrum in the absence of acid after irradiation confirms that [CpW(CO) 3 ] -is a reaction intermediate and further supports that conclusion that protonation is fast and quantitative when [PyH][BF 4 ] is present. The surprising observation of CpW(CO) 3 H as a photolysis product in the absence of acid suggests an adventitious proton source is present that protonates the anion. While all solvents were dried over calcium hydride and stored over sieves, adventitious water is likely still present at low concentrations and was hypothesized to be the proton source for experiments conducted in the absence of [PyH][BF 4 ]. To confirm our hypothesis, a control experiment was conducted in which [CpW(CO) 3 ] 2 was photolyzed with intentionally added water (1.1 mM). Compared to the same experiment in the absence of water or acid, an increase in conversion of [CpW(CO) 3 ] 2 to CpW (CO) 3 H is observed, from 5.4% with no proton source added to 13% with water added (Fig. S5 †). Additionally, the amount of [CpW(CO) 3 ] -present after irradiation decreases from 4.9% to 2.8%, supporting the conclusion that water can act as a proton source in this reaction. While the conversion of CpW(CO) • radicals disproportionate within the solvent cage depends on both the Lewis base identity and its concentration (Scheme 1B), acknowledging these conclusions are restricted to the processes that occur within 150 picoseconds after excitation. 14,15,18 This work also established that for very strong Lewis bases, the equilibrium between CpW(CO) 3
• and CpW(CO) 3
• L favors the 19e -species so that any 17e -radicals not consumed by disproportionation are converted to CpW(CO) 3 • L. While Cahoon's study focused on phosphine and phosphite ligands, acetonitrile is reported to have even stronger donating ability for this family of complexes. 35 Separately, Tyler and coworkers utilized the observation of quantum yields of greater than unity (observed under some conditions) and observed correlations between quantum yield and I -1/2 (I = photon flux on the sample) to support their assignment of a radical chain mechanism for the homolysis of [CpM(CO) 3 ] 2 complexes. 11,16,35,36 In this limiting mechanism, radicals that escape the solvent cage before disproportionation drive a radical chain reaction (Scheme 1A). While Tyler's reports suggest it is the 17e -radicals CpM(CO) 3
• that escape the solvent cage, and then bind a Lewis base in solution to form the highly reducing 19e -radical CpM(CO) 3 L • (that then reduces unreacted [CpM(CO) 3 ] 2 ), we suggest here that L binds to the CpW(CO) 3
• radicals in the solvent cage forming the 19e - radicals, CpW(CO) 3 L • , invoked by Cahoon's findings (that strong donors in high concentration can bind to both 17e - radicals in the solvent cage) and that these 19e -radicals can also provide entry to this cycle (Scheme 3). This revises the radical chain pathway initially proposed by Tyler, as entry to
the radical chain cycle via a 19e -radical has not previously been postulated.
In our work, the disproportionation product [CpW(CO) 3 ] - is the critical precursor to CpW(CO) 3 H. Eager to interrogate the underlying mechanism of this photon-driven hydride formation reaction, we determined quantum yields for CpW(CO) 3 H formation upon photolysis of [CpW(CO) 3 ] 2 with 455 nm light. Quantum yields were measured at 9 different light intensities ranging from 6.18 × 10 -9 -3.45 × 10 -8 mol photons per s per cm 2 , with light intensities determined for each setting of the LED power supply using actinometry (see ESI † for details). These data (Table 1) show that quantum yields for CpW(CO) 3 H formation range from 2.6-7.0%. While Tyler has reported quantum yields of greater than unity for the disproportionation of [(MeCp)Mo(CO) 3 ] 2 (MeCp = η 5 -CH 3 C 5 H 4 ) in acetone, it's notable his reported quantum yields for the same species in acetonitrile are more consistent with what we observe for the CpW(CO) 3 H formation at similar photon fluxes (Φ = 0.02 at 1.77 × 10 -8 mol photon per s). 13 A more telling test to assess the mechanism is to plot the quantum yield as a function of photon flux (Fig. S13 †). We observed a correlation between quantum yield and the inverse square root of photon flux on our sample tube (I -1/2 ), consistent with what is predicted by the steady state approximation for the radical chain mechanism (see ESI †).
The correlation observed between quantum yield and I -1/2 , together with Cahoon's observation that strong donor ligands
drive the formation of the highly reducing 19e -species CpW (CO)
Published on 17 February 2023. Downloaded by University of North Carolina at Chapel Hill on 3/6/2023 4:50:30 AM.View Article Online
This journal is © The Royal Society of Chemistry 2023 Dalton Trans. Published on 17 February 2023. Downloaded by University of North Carolina at Chapel Hill on 3/6/2023 4:50:30 AM.View Article Online