|Home | About | Journals | Submit | Contact Us | Français|
In order to understand how the separation between the electron and proton-accepting sites affects proton-coupled electron transfer (PCET) reactivity, we have prepared ruthenium complexes with 4′-(4-carboxyphenyl)terpyridine ligands, and studied reactivity with hydrogen atom donors (H-X). RuII(pydic)(tpy-PhCOOH) (RuIIPhCOOH), was synthesized in one pot from [(p-cymene)RuCl2]2, sodium 4′-(4-carboxyphenyl)-2,2′:6′,2″-terpyridine ([Na+]tpy-PhCOO−), and disodium pyridine-2,6-dicarboxylate (Na2pydic). RuIIPhCOOH plus nBu4NOH in DMF yields the deprotonated Ru(II) complex, nBu4N[RuII(pydic)(tpy-PhCOO)] (RuIIPhCOO−). The Ru(III) complex (RuIIIPhCOO) has been isolated by one-electron oxidation of RuIIPhCOO− with triarylaminium radical cations (NAr3•+). The bond dissociation free energy (BDFE) of the O–H bond in RuIIPhCOOH is calculated from pKa and E1/2 measurements as 87 kcal mol-1, making RuIIIPhCOO a strong hydrogen atom acceptor. There are 10 bonds and ca. 11.2 Å separating the metal from the carboxylate basic site in RuIIIPhCOO. Even with this separation, RuIIIPhCOO oxidizes the hydrogen atom donor TEMPOH (BDFE = 66.5 kcal mol-1, ΔG°rxn = -21 kcal mol-1) by removal of an electron and a proton to form RuIIPhCOOH and TEMPO radical in a concerted proton-electron transfer (CPET) process. The second order rate constant for this reaction is (1.1 ± 0.1) × 105 M-1 s-1 with kH/kD = 2.1 ± 0.2, similar to the observed kinetics in an analogous system without the phenyl spacer, RuIII(pydic)(tpy-COO) (RuIIICOO−). In contrast, hydrogen transfer from 2,6-di-tert-butyl-p-methoxyphenol [tBu2(OMe)ArOH] to RuIIIPhCOO is several orders of magnitude slower than the analogous reaction with RuIIICOO.
Coupling electron transfer to proton transfer is key to a wide range of chemical and biochemical processes, such as converting solar energy to chemical fuels.1 While many of the fundamentals of electron transfer (ET) are well understood, the principles of proton-coupled electron transfer (PCET) are still being developed. The effects of increasing the distance between reaction centers has been much studied for ET,2 and we have started to explore PCET systems in which the electron- and proton-accepting sites are increasingly separated.3 PCET processes with large separations appear to be important in a number of biological systems, such as ribonucleotide reductases and photosystem II.4 They may also be involved in charge injection into oxide semiconductors from ruthenium polypyridyl-carboxylate complexes.5 In our previously reported ruthenium terpyridine-4′-carboxylate complex RuIIICOO (Scheme 1), the Ru is six bonds and 6.9 Å removed from the basic carboxylate oxygen atoms.3 Despite this separation, the reported reactions occur with H+ and e− transferring in the same kinetic step, by concerted proton-electron transfer (CPET).1,3,6-8 In this report, the distance between the metal and basic sites is extended further, by inserting a phenyl spacer between the terpyridine and the carboxylate. Reactivity is contrasted between the two systems, which are the first studies of long, well-defined separations.
The new protonated Ru(II) complex, RuII(pydic)(tpy-PhCOOH) (RuIIPhCOOH), was prepared from [(η6-cymene)RuCl(μ-Cl)]2 and the known ligands, 4′-(4-carboxyphenyl)-2,2′:6′,2″-terpyridine ([Na+]tpy-PhCOO−) and pyridine-2,6-dicarboxylate (Na2pydic).3,9 Deprotonation of this complex with nBu4NOH in DMF yields nBu4N[RuII(pydic)(tpy-PhCOO)] (RuIIPhCOO−). Both compounds have been characterized by 1H NMR and UV-Visible spectroscopies, electrochemistry, ESI-MS, and elemental analyses. Based on the structure of the related homoleptic complex RuII(tpy-PhCOO−)2, the distance between the Ru and the carboxylate oxygens is 11.2 ± 0.1 Å.10 The optical spectra of RuIIPhCOOH and RuIIPhCOO− in the visible region are very similar, as shown in Figure 1a.9 The differences in the spectra are small, but they are consistent and reversible upon additions of acid and base.9 Titration of RuIIPhCOO– with benzoic acid (pKa = 20.7 ± 0.111) in MeCN gives a pKa for RuIIPhCOOH of 20.5 ± 0.2. Cyclic voltammograms of RuIIPhCOOH and RuIIPhCOO− in DMF show almost identical chemically reversible oxidations, with E1/2 = 0.081 ± 0.006 V and 0.083 ± 0.019 V vs FeCp2+/0, respectively.
Oxidation of RuIIPhCOO− by [(p-tol)3N•+]PF6− in MeCN gives the neutral, deprotonated Ru(III) carboxylate complex, RuIII(pydic)(tpy-PhCOO) (RuIIIPhCOO, Scheme 1). This zwitterionic complex can be precipitated as a brown solid with CH2Cl2 but it is difficult to handle without some decomposition in solution,9 so it is more conveniently generated in situ from RuIIPhCOO− plus [(p-BrC6H4)3N•+][B(C6F5)4−] in MeCN. The 1H NMR spectrum of RuIIIPhCOO shows all nine paramagnetic peaks, and the optical spectrum has a shoulder at 435 nm (ε ~ 9,000 M-1 cm-1). Reduction of in situ-generated RuIIIPhCOO with (C5Me5)2Fe rapidly regenerates Ru(II), with a yield of ~95% based on the absorption at 531 nm, along with other product(s). Addition of base to this solution causes a shifting of the 400 nm peak, suggesting that the product mixture is ca. 70/30 RuIIPhCOO− / RuIIPhCOOH.9 The data indicate that RuIIIPhCOO is predominantly deprotonated in solution but may contain some RuIIIPhCOOH+.12
A formal O–H bond dissociation free energy (BDFE) can be defined for RuIIPhCOOH, despite the 11.2 Å distance between Ru and O. ΔG° for RuIIPhCOOH → RuIIIPhCOO + H• in MeCN = 23.1E1/2 + 1.37pKa + CG13 = 87 ± 1 kcal mol-1, using the pKa given above and E1/2 = 0.17 ± 0.03 V vs FeCp2+/0 for RuIIPhCOO− in 90/10 MeCN/DMF.9 The BDFE is 6 kcal mol-1 higher than that found for the complex without the phenyl spacer, RuIICOOH (81 ± 1 kcal mol-1). Thus, RuIIIPhCOO is a strong hydrogen atom acceptor,3 which may partly explain its instability.
RuIIIPhCOO reacts with the hydroxylamine TEMPOH within seconds to form the nitroxyl radical TEMPO and predominantly RuIIPhCOOH, (eq 1), as indicated by 1H NMR and UV-visible
spectra.9 The protonated RuIIPhCOOH product is implicated by the peak at 400 nm in the optical spectrum, which shifts to 394 nm upon addition of base (Figure 1a).9,14 The reaction is quite downhill, ΔG°1 = -21 kcal mol-1, based on the BDFE(TEMPOH) = 66.5 kcal mol-1.15 Stopped flow rapid-scanning spectrophotometry under pseudo-first order conditions of excess TEMPOH shows a ca. 80% yield of RuIIPhCOOH (Figure 1b). It is possible that the low yield is due to the presence of protonated Ru(III) in solution, which reacts more slowly with TEMPOH.14 The pseudo-first order rate constants vary linearly with [TEMPOH] (Figure S109) indicating a simple second-order rate law, with k1H = (1.1 ± 0.1) × 105 M-1 s-1 and ΔG1‡ = 10.6 ± 0.1 kcal mol-1.9 Using rate data from 17 – 52 °C, the activation parameters are ΔH1‡ = 6.8 ± 1.1 kcal mol-1 and ΔS1‡ = -13 ± 4 cal mol-1 K-1. The reaction with TEMPOD gives k1D = (5.6 ± 0.3) × 104 M-1 s-1. The small primary isotope effect, k1H/k1D = 2.1 ± 0.2, indicates that the proton is transferred in the rate limiting step.16
Reaction 1 could occur by (i) initial proton transfer (PT) to form RuIIIPhCOOH+ and TEMPO− followed by electron transfer (ET), (ii) ET followed by PT, or (iii) CPET with no intermediates. The pKas of RuIIIPhCOOH+ and TEMPOH are 20.5 and 39, respectively, so the initial step in path (i) has ΔG°PT = 25.3 kcal mol-1.9 Similarly, using the reduction potentials of the reactants, ΔG°ET = 12.5 kcal mol-1. Since these are the minimum barriers for initial PT and ET (ΔG‡ > ΔG°), and they are larger than the observed ΔG1‡ (10.6 ± 0.1 kcal mol-1), neither of the stepwise pathways can be occurring. Even with the large separation between Ru and COO−, the reaction still occurs by CPET. This is also indicated by the primary kinetic isotope effect (KIE) of 2.1.
An important issue in a PCET reagent is the amount of interaction or communication between the redox and basic sites.3 One measure of this is the thermodynamic coupling, in this system how much the pKa shifts depending on the Ru oxidation state (ΔpKa), and equivalently1d how much the E° shifts with the protonation state (ΔE1/2). Cyclic voltammograms of RuIIPhCOO− and RuIIPhCOOH in DMF are the same within error (ΔE1/2 = 2 ± 20 mV), indicating that there is essentially no communication between the redox and basic sites. This is also indicated by the close similarity of pKas of RuIIPhCOOH and benzoic acid. For comparison, ΔE1/2 for RuIICOO(H) (with no Ph spacer) is 0.13 V (other systems have larger values3). CPET may be the favored mechanism for reaction 1 because of the large ΔE1/2 for TEMPO(H), ca. 2.6 V.9 In a related intermetal PCET system without this large ΔE1/2, concerted transfer was not observed.17
RuIIIPhCOO also rapidly oxidizes 2,6-di-t-butyl-4-methoxy-phenol, (ArOH), to give the aryloxyl radical ArO• 18 and RuIIPh-COOH.9 Pseudo-first order kinetic studies give kArOH = (1.0 ± 0.1) × 103 M-1 s-1 and kH/kD = 2.6 ± 0.4.9,16 In this case, thermochemical analyses do not rule out initial PT or ET mechanisms since ΔG°PT and ΔG°ET (10.8 ± 0.5 and 11.8 ± 0.8 kcal mol-1)9 are both lower than the observed ΔG‡ = 13.4 ± 0.1 kcal mol-1. Therefore kArOH is an upper limit for the CPET rate constant. The mechanism is still likely to be CPET because ΔG°PT and ΔG°ET are much larger than ΔG°CPET and because the KIE of 2.6 is larger than would be expected for ET or for PT between oxygen atoms.
Table 1 compares the rate constants and ΔG° values for the reactions of RuIIIPhCOO vs. RuIIICOO. For the reactions with TEMPOH, the rate constants are within a factor of 2, even though CPET to RuIIIPhCOO is 6 kcal mol-1 more exoergic.3 For oxidation of ArOH, the RuIIIPhCOO rate constant is a thousand times slower despite the 6 kcal mol-1 larger driving force. Thus the larger driving force for the reactions of RuIIIPhCOO is offset by the decreased communication and longer distance.
In conclusion, we have designed and prepared a system with a well-defined separation of ten bonds and 11.2 Å between the metal (Ru) and basic (carboxylate) sites. At this distance, there is almost no interaction between the redox and basic sites, as indicated by thermodynamic and spectroscopic measurements. Despite this lack of communication, the reaction of RuIIIPhCOO with TEMPOH (eq 1) occurs by concerted transfer of H+ and e− (CPET). However, the more exoergic reactions of RuIIIPhCOO proceed more slowly than those of RuIIICOO. Thus the increased distance and decreased communication do appear to affect the reaction rates, as will be discussed in a future report focused on the dependence of CPET rate constants on driving force20 and on the position and interaction of redox and acid/base sites.
We gratefully acknowledge support from the U.S. National Institutes of Health (GM050422) and the University of Washington.