Article Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX
pubs.acs.org/IC
CO-Photolysis-Induced H‑Atom Transfer from MnIO−H Bonds Karthika J. Kadassery,† Komal Sethi,† Paul M. Fanara, and David C. Lacy* Department of Chemistry, University at Buffalo, State University of New York, Buffalo, New York 14260, United States
Downloaded via UNIV OF TOLEDO on March 18, 2019 at 01:15:46 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
S Supporting Information *
ABSTRACT: The formation of TEMPOH from a mixture of [Mn(CO)3(μ3-OH)]4 (1) and (2,2,6,6-tetramethylpiperidin1-yl)oxyl (TEMPO) is shown to occur through a lightinitiated CO photolysis from 1 (illumination at 300−375 nm). One hypothesis is that the loss of carbon monoxide (CO) causes significant O−H bond weakening to render proton-coupled electron transfer (PCET) to TEMPO favorable. For instance, the ground-state O−H bond dissociation free energy (BDFEO−H) of 1 (computed with density functional theory and estimated using effective BDFE reagents) is too high to transfer an H-atom to TEMPO. We also demonstrate that TEMPO and 1 interact in the dark through a hydrogen-bonded “precomplex” (1···TEMPO). We suggest that the PCET reaction that forms TEMPOH is the result of a H-atom-transfer reaction that occurs immediately after photolysis of a CO ligand(s).
■
ening.11,12 Hence, the importance of this O−H bondweakening phenomenon to water reduction, and its emerging utility in synthesis,13,14 warrants additional investigations. To this end, we previously reported a water-reduction process mediated by the tetramer [Mn(CO)3(μ3-OH)]4 (1) and provided preliminary evidence that H2 formation involved H-atom transfer (HAT) processes from water-derived ligands (Scheme 1).5,15 Specifically, we found that radical traps like
INTRODUCTION The scalable production of dihydrogen (H2) from water and sunlight is a worthy pursuit for the purpose of clean energy. An interesting strategy, if at least to gain fundamental knowledge about chemical reactivity, is one that uses single-site molecular compounds for direct solar photoconversion of water into fuel.1,2 While not yet practical, it is reported that certain molecular organometallic complexes split water (H2O) through reduction and oxidation reactions, producing H2 and hydrogen peroxide (H2O2) as the initial products.3−5 Analogous photochemical HY splitting (Y = halogen)6 and catalytic photochemical alkane dehydrogenation7 provide insight into the elementary reactions that might be involved in photochemical water splitting mediated by similar complexes. However, water has unique coordination chemistry and thus deserves special attention. For instance, the first two steps in a water-splitting reaction with CpMn(CO)3 is carbon monoxide (CO) photolysis and subsequent water coordination to CpMn(CO)2.4 Fan and coworkers demonstrated substantial O−H bond weakening in a methanol (MeOH) adduct coordinated to CpMn(CO)2.8 It follows that similar O−H bond weakening may also occur in the CpMn(CO)2(OH2) species. In a related study, Beweries and co-workers showed that an ansa-titanocene-aqua complex can facilitate cyclic “water splitting”, where the O−H bond in the aqua ligand is cleaved homolytically and transferred to (2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO).9,10 These two cases exemplify how O−H bond weakening upon coordination to a metal center is likely an important component in the water-reduction step for organometallic complexes. Chirik and co-workers unambiguously demonstrated this possibility through spontaneous H2 production resulting from water-coordination-induced O−H bond weak© XXXX American Chemical Society
Scheme 1. Summary of Photochemical H2 Formation with 1
5,5-dimethyl-1-pyrroline N-oxide (DMPO) and TEMPO inhibit H2 production from 1 under irradiation in toluene and that {DMPO-H}• was an observed product. Herein we confirm that TEMPO is a competitive inhibitor of H2 production and that TEMPOH is the product in yields that are comparable to that of H2 formation when TEMPO is absent. Furthermore, we show that 1 and TEMPO do not react in the dark, other than to form a H-bonded complex (1··· TEMPO). The lack of reactivity in the dark is consistent with the high O−H bond dissociation free energy (BDFEO−H) of 1 Received: February 2, 2019
A
DOI: 10.1021/acs.inorgchem.9b00322 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry that we estimate using density functional theory (DFT) for this report. Therefore, we posit that the OH ligands in 1 undergo O−H bond weakening as a result of photolysis of a CO ligand, similar to findings with CpMn(CO)2(MeOH) by Fan and coworkers.8 The findings with TEMPO are briefly contextualized within the framework of developing new strategies to produce H2 from water and light.
■
RESULTS AND DISCUSSION Estimating the BDFEO−H in 1. This investigation began with an attempt to experimentally measure the ground-state BDFEO−H in 1. Such information is important for an assertion that proton-coupled electron transfer (PCET) processes with 1, such as HAT, are the result of CO photolysis and not a reaction involving 1. In an ideal case, one can measure the reduction potential of XH ([XH]0/+) and the pKa value of the resulting cation (B + [XH]+ = X + [BH+]) and use Hess’s law to obtain BDFE X−H with the equation in Scheme 2 Scheme 2. Thermochemistry of 1 (Units in kcal/mol) Figure 1. Cyclic voltammogram of 0.001 M 1 in MeCN with and without TEMPO (0.1 M [Bu4N][PF6], glassy carbon working electrode, Ag/AgNO3 reference electrode with a CoralPor separator, and platinum auxiliary electrode). For TEMPO, E1/2 = 0.22 V; for 1, Epa = 0.76 V and Epa/2 = 0.66 V (200 mV/s).
Supporting Information (SI) for a detailed description of the method].22 This high value is consistent with the O−H bond strength of water and hydroxide; for example, the BDFEO−H values of H2O and OH− in liquid water are 122.7 and 117.5 kcal/mol, respectively.16 Because there are no reported single-component PCET reagents that are capable of abstracting an H atom with high BDFE ∼ 102 kcal/mol, we used base/oxidant reactant pairs with defined effective BDFE values to further test the computed BDFEO−H.16,21 Namely, the pKa value of a base (conjugate acid) and the E1/2 value of an oxidant have a defined effective BDFE(MeCN) determined with the equation in Scheme 2. Mayer and co-workers have recently shown that effective BDFE base/oxidant pairs can mediate PCET reactions via a multisite concerted proton−electron-transfer (MS CPET) mechanism if the base forms a hydrogen-bond adduct with a “H-atom” donor molecule (e.g., TEMPOH··· pyridine).21 Without this adduct, a termolecular reaction between the H-atom donor, base, and oxidant is required for MS CPET to occur.23−25 Conveniently, 1 is known to form various hydrogen-bonding adducts with aromatic solvent molecules, OPPh3, and water and thus is amenable to this MS CPET approach.19 We confirmed that 1 forms a hydrogenbonded adduct with DBU, which is apparent from a shift of δO−H in the 1H NMR spectrum of 1 + DBU in MeCN-d3 or toluene-d8 (Figure S2).26 The neutral oxidant Mn(hfacac)3 paired with DBU has an effective BDFE(MeCN) = 109.0 kcal/mol (Table 1), which is above the DFT-computed BDFEO−H value of 105 kcal/mol. Unfortunately, the background reaction between DBU and Mn(hfacac)3 is fast, and Mn(hfacac)3 reacts with MeCN too rapidly to be of use in this scenario.27 Mn(hfacac)3 is stable in toluene for several days,27 but it reacts rapidly with DBU. This side reaction appears too fast to allow for productive PCET chemistry with 1. For example, the 1H NMR spectrum of a mixture of 1 and DBU in toluene-d8 treated with Mn(hfacac)3 confirms that [{DBU}H]+ does not form and essentially
(alternatively one can determine the pKa value of XH and the redox potential of the resulting anion). CG accounts for the free energy of formation and solvation of the H atom; the value CG = 54.9 kcal/mol for MeCN is exclusively used for this report.16 Epa and Epa/2 for irreversible 1/[1]+ are 0.76 and 0.66 V versus [FeCp2]+/0 in acetonitrile (MeCN) at 200 mV/s, respectively (Figure 1).17 Unfortunately, isolation of the [1]+ species has been unsuccessful.18 Attempts to establish equilibrium with a base and 1 to evaluate the pKa value were also unsuccessful, consistent with Zaworotko and co-workers’ findings.19 Besides hydrogen bonding between 1 and 1,8diazabicyclo[5.4.0]undec-7-ene (DBU; confirmed by 1H NMR and Fourier transform infrared (FTIR) spectroscopy, vide infra), no reaction was observed between 1 and excess DBU (0.001 M 1; up to 2000 equiv of DBU; pKa = 24.3).20 The use of stronger bases like KH, NaH, or K[N(SiMe3)2] (pKa = 34)21 furnished brown solutions that rapidly decomposed into 1 and other unidentified insoluble brown materials. In summary, the reactions between 1 and a base or oxidant generate unstable species that disproportionate into intractable solids and 1, as evidenced from recovery of 1 from these reactions. The current lack of success to access either 1+ or [1H]− prevents our ability to determine a BDFEO−H value of 1 using Hess’s law with traditional square scheme methods. Hence, DFT was used to compute BDFEO−H(MeCN) = 105 kcal/mol and BDFEO−H(toluene) = 102 kcal/mol [see the B
DOI: 10.1021/acs.inorgchem.9b00322 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry Table 1. Base/Oxidant Pairs with Their Effective BDFEsa 1 2 3 4 5 6 7
base (pKa)b
oxidant (E1/2)c
DBU (24.3) DBU DMAP (18.0) DMAP DBU
Mn(hFacac)3 (0.9) Mn(acac)3(0.7) Mn(hFacac)3 Mn(acac)3 [FeCp2]+ (0)
radical
BDFE
t Bu3ArO• TEMPO•
109.0 104.4 100.2 95.6 88.2 77.1d 66.5d
a
BDFE values are in kcal/mol in MeCN [BDFE(MeCN)]. bpKa values from ref 20. cV vs FeCp20/1+ (see ref 27 for entries 1−4). d BDFE obtained from ref 16. Note that entries 1−4 are sufficiently strong to cleave the C−H bonds in MeCN (ref 28).
matches the control reaction between Mn(hfacac)3 and DBU (Figure S2). Moreover, by using the Epa/2 and DFT-computed BDFE(MeCN) values of 1, the expected pKa of [1]+ is ≥1 pKa unit higher than that of DBU,22 so even if PCET does occur, the equilibrium will not favor [{DBU}H]+; it is worth noting here that the estimated high pKa of [1]+ is one possible cause of the failure to isolate it.29 Various combinations of base/oxidant pairs with weaker effective BDFE(MeCN) were also used (Table 1). For instance, when the weaker oxidant Mn(acac)3 was added to a solution of 1 and DBU, there was no formation of [{DBU}H]+ [effective BDFE(MeCN) = 104.4 kcal/mol; Figure S3]. Unfortunately, in the absence of a positive reaction, it is impossible to confirm the limits of BDFEO−H in 1. Nevertheless, these experiments, in addition to the DFTcomputed value, support a high BDFEO−H value of 1 in MeCN that is ≥104.4 kcal/mol [104.4 = effective BDFE of the Mn(acac)3/DBU pair]. Formation of the 1···TEMPO Precomplex. The high BDFE of 1 indicates that 1 should not be able to transfer H atoms to substrates like TEMPO [BDFE(MeCN) = 66.5 kcal/ mol] or 2,4,6-tritert-butylphenoxyl radical [tBu3ArO•; BDFE(MeCN) = 77.1 kcal/mol].16 Indeed, there is no reaction between 1 and either of these radicals. Despite the lack of HAT in the dark, TEMPO interacts with 1 to form a hydrogenbonded adduct, as indicated by solution [dichloromemthane (DCM)] transmission FTIR spectroscopy (Figure 2). The FTIR spectrum of a solution of 1 and TEMPO in DCM exhibits a new feature at 3390 cm−1 that is attributed to a 1:1 hydrogen-bonded adduct of TEMPO and 1 (1···TEMPO; Scheme 3). Plotting the ratio of the concentration of 1··· TEMPO to 1 against the concentration of TEMPO furnishes a straight line with a slope equal to the equilibrium constant (Keq = 5.7 ± 1.1 M−1).30 CO-Photolysis-Induced Formation of TEMPOH. As discussed briefly in the Introduction, DMPO and TEMPO inhibit H2 production when 1 is irradiated in toluene (300− 375 nm).5 For this report, we confirm that stoichiometric TEMPO completely inhibits H 2 production and that TEMPOH is the product in yields as high as 65% for 10 equiv of TEMPO and 35% for 1 equiv of TEMPO (Figure S5). Because TEMPO and 1 do not engage in PCET chemistry in the dark and the BDFEO−H value of 1 is high, we posit that TEMPOH forms via O−H bond weakening induced by CO photolysis of 1. An important empirical element that supports this hypothesis is an apparent competitive formation of TEMPOH and H2 in the irradiation of toluene solutions containing 1 and
Figure 2. (top) FTIR spectra of 0.78 mM 1 (black) in DCM with increasing concentration of TEMPO. The decrease in the transmittance at 3611 cm−1 is due to the fall in the concentration of adduct-free 1 and the increase in the transmittance at 3390 cm−1 assigned as the 1···TEMPO hydrogen-bonded adduct. (bottom) Representative plot of [1···TEMPO]/[1] versus [TEMPO] (triplicate Keq = 5.7 ± 1.1 M−1).
Scheme 3. Proposed Structure of 1···TEMPO (Keq = 5.7 ± 1.1 M−1)
TEMPO. For example, when 1 equiv or more of TEMPO is included in the irradiation of 1, H2 formation is completely suppressed (Table 2). Alternatively, when
