Mechanism of the Quenching of the Tris(bipyridine)ruthenium(II

B. Limburg , E. Bouwman , and S. Bonnet ... Mechanistic Insight into Peroxydisulfate Reactivity: Oxidation of the cis,cis-[Ru(bpy)2(OH2)]2O “Blue Di...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/JPCA

Mechanism of the Quenching of the Tris(bipyridine)ruthenium(II) Emission by Persulfate: Implications for Photoinduced Oxidation Reactions A. Lewandowska-Andralojc† and D. E. Polyansky* Chemistry Department, Brookhaven National Laboratory, Upton, New York 11973-5000, United States S Supporting Information *

ABSTRACT: A revised mechanism for the oxidation of the excited state of Ru(bpy)32+ with the persulfate anion is described in this work. The formation of the precursor complex in the electron transfer reaction involves ion pairing between the metal complex in ground and excited states and S2O82‑. The equilibrium constant for the ion-pair formation (KIP = 2.7 M−1) was determined from electrochemical measurements and analysis of thermal reaction between Ru(bpy)32+ and persulfate. It was found to be consistent with the calculated value estimated from the Debye−Hückel model. The analysis of rate constants for reactions between persulfate and various metal complexes indicates that thermal and photochemical reactions most likely proceed through a common pathway. Extremely high reorganization energy (ca. 3.54 eV) for the electron transfer obtained from fitting experimental data with the Marcus equation is indicative of significant nuclear reorganization during the electron transfer step. In view of these results the electron transfer can be described as dissociative probably involving substantial elongation or complete scission of the O−O bond. The proposed model accurately describes experimental results for the quenching of *Ru(bpy)32+ over a wide range of persulfate concentrations and resolves some discrepancies between the values of KIP and ket previously reported. The implications of various factors such as the ionic strength and dielectric constant of the medium are discussed in relation to measurements of the quantum yields in photodriven oxidation reactions employing the Ru(bpy)32+/ persulfate couple.

■

INTRODUCTION The reactivity of Ru(bpy)33+ as a one-electron oxidizing agent has prompted its widespread use in studies of various oxidation processes,1 including photodriven catalytic water oxidation.2−5 The key properties making Ru(bpy)33+ an attractive sacrificial oxidant include its ability to react as a pure one-electron oxidant, the wide range of reduction potentials accessible through ring substitution (Supporting Information (SI), Table S1), pH independent potentials, high quantum yield of photogeneration, and characteristic absorption spectra of both Ru(bpy)33+ and Ru(bpy)32+ which allows the reaction progress to be followed using UV−vis spectroscopy. Production of Ru(bpy)33+ (E0 = 1.26 V, also see the SI, Table S1) in high quantum (∼60% for the photochemical step,4,6−8 Scheme 1) and chemical yields can be conveniently achieved through a photoexcitation of Ru(bpy)32+ and subsequent quenching of its excited state with a sacrificial electron acceptor such as S2O82‑ ion in aqueous solution. The ability to rapidly generate Ru(bpy)33+ by the so-called flash-quench technique7,9−13 has permitted kinetic studies of fast oxidation reactions, not always accessible by stopped-flow techniques. While the process of photoinduced production of Ru(bpy)33+ may appear to be straightforward (Scheme 1), understanding the detailed mechanism is crucial for meaningful interpretation of results of photodriven oxidation reactions. Moreover, the value of the quantum yield for photochemical production of the sacrificial © 2013 American Chemical Society

Scheme 1. Individual Steps in the Photochemical Generation of Ru(bpy)33+

oxidant under a variety conditions turns out to be essential in determining accurate quantum yields of photodriven catalytic water oxidation reactions and optimization of the catalytic conditions.5,8,14 Mechanistic studies of the quenching of the excited state of Ru(bpy)32+ by S2O82‑ date back several decades.15−19 The work by White et al.19 provided one of the most comprehensive early mechanistic models and has since been widely used for the analysis of photodriven oxidation reactions using the Ru(bpy)32+|S2O82‑ pair. In brief, the model proposed by White et Received: July 30, 2013 Revised: September 13, 2013 Published: September 16, 2013 10311

dx.doi.org/10.1021/jp407573d | J. Phys. Chem. A 2013, 117, 10311−10319

The Journal of Physical Chemistry A

Article

Scheme 2. Mechanism of the Quenching of the Excited State of Ru(bpy)32+ by S2O82‑

measurements showed that the emission intensity decreased by less than 2% during a second run. The ion pair between Ru(bpy)32+ and persulfate (or sulfate) was prepared as a solid-state sample. About 5 mL each of 0.5 mM solutions of Ru(bpy)32+ and persulfate (or sulfate) in THF-H2O (9/1, v/v) mixtures were mixed together allowing almost immediate precipitation. The precipitate was isolated on a filter paper by vacuum filtration in the dark. Emission lifetime experiments of Ru(bpy)32+|S2O82‑ and Ru(bpy)32+|SO42‑ ion pairs in the solid state were conducted with a transient emission apparatus: excitation was provided by a YAG:Nd3+ laser (355 nm; Continuum, Powerlite 7010, 10 ns, 5 Hz, 2 mJ/pulse). A Tektronix DPO4032 digital phosphor oscilloscope (350 MHz, 2.5 GS/s) was used to digitize the transient signals from a Hamamatsu R928 PMT detector. The emission lifetimes were estimated by single- or biexponential fits to the kinetic trace at 610 nm. Electrochemical Measurements. Electrochemical measurements for Ru(bpy)32+ were conducted with a BAS 100b electrochemical analyzer from Bioanalytical Systems. Square wave voltammograms were measured using 0.1 mM solutions of Ru(bpy)32+ in pure water or mixtures: THF-H2O (23% v/v; 33% or 53% of H2O) in the presence of various concentrations of S2O82‑ (0.05 − 5 mM) at constant ionic strength using sodium perchlorate as an electrolyte. A glassy carbon disk was used as a working electrode, a platinum wire as a counter electrode, and Ag/AgCl as a reference electrode. Bulk electrolysis (applied potential 1400 mV vs NHE) of a 0.55 mM solution of Ru(bpy)32+ in water with 0.01 M TFMS (pH 2) was conducted to obtain Ru(bpy)33+. In bulk electrolysis experiments, glassy carbon was replaced with a platinum gauze electrode. DFT Calculations. The structures of persulfate anion, its corresponding one-electron reduced form as well as the sulfate radical and sulfate anion were calculated using the Gaussian 09 program package (for complete reference see the SI),22 using the B3LYP hybrid functional with a cc-pvdz basis set. The effect of the bulk solvent was included in all geometry optimization and vibrational frequency calculations of species with standard states in solution through the use of the C-PCM method using UAKS radii. One explicit water molecule hydrogen-bonded to persulfate oxygen atoms was included in the calculations.

al. is based on two parallel pathways for the production of Ru(bpy)33+: “unimolecular” and “bimolecular” (SI, Figure S1). The unimolecular mechanism involves excitation of the groundstate ion pair between Ru(bpy)32+ and S2O82‑ and proceeds via electron transfer within the excited ion pair. The bimolecular mechanism involves the reaction between S 2 O 8 2‑ and *Ru(bpy)32+ or an excited ion pair. While the proposed model can explain experimental results in a limited range of S2O82‑ concentrations, some of its main features differ significantly from other formalisms in the field.20 For example, the equilibrium constant for formation of the ground-state ion pair is several orders of magnitude higher compared to similar systems reported earlier.15,21 In addition it was necessary to invoke an unusually long lifetime of the photoexcited ion pair for it to be quenched by a second persulfate. These and other inconsistencies prompted us to reexamine the kinetics and mechanism of the oxidative quenching of *Ru(bpy)32+ by S2O82‑ using a wider range of experimental methods and conditions. The mechanism we propose is more consistent with the classical view of electron transfer and is supported by independent determinations of some of the key parameters. The present studies also provide further insight into the reductive dissociation of persulfate.

■

EXPERIMENTAL SECTION Materials. All chemicals were of reagent grade quality and used without further purification, except for the following. Trifluoromethanesulfonic acid (TFMS) was ReagentPlus grade (Aldrich), distilled under vacuum and immediately used to prepare a 1.0 M aqueous stock solution. The solution was stored under refrigeration (ca. 10 °C). Tris(2,2′−bipyridine)ruthenium(II)chloride hexahydrate ([Ru(bpy)3]Cl2·6H2O) was obtained from Strem Chemicals Inc. 2,2,2-Trifluoroethanol (TFE), tetrahydrofuran (THF), potassium persulfate, and sodium perchlorate were supplied by Sigma-Aldrich. Aqueous solutions were prepared with distilled water that had been passed through a Millipore ultrapurification system. Spectroscopic Measurements. UV−vis spectra were measured on a Varian Cary 500 dual-beam spectrophotometer or an Agilent Technologies 8453 diode-array spectrophotometer. The samples were held in a thermostatted cell holder, which was maintained at 25.0 ± 0.1 °C. Single wavelength kinetic traces at 670 nm were recorded for determination of the Ru(bpy)33+ formation rate in the thermal reaction and for monitoring of Ru(bpy)33+ decay at pH 2. Steady-state luminescence quenching was studied using a PTI spectrofluorimeter. Aqueous solutions of Ru(bpy)32+ (50 μM) and K2S2O8 (1−40 mM) with NaClO4 or NaNO3 as an additional electrolyte were prepared in the dark to avoid photoreactions. Similar solutions were prepared in a mixture of H2O and TFE (4/1, v/v) with phosphate buffer (20 mM, pH 7.0). All solutions were degassed with argon before measurements. For steady-state fluorescence measurements, samples were excited at 470 nm, and emission intensity data were collected at 480−800 nm at room temperature. Spectral

■

RESULTS AND DISCUSSION The proposed model for the quenching of the excited state of Ru(bpy)32+ by S2O82‑ is based on the classical description of electron transfer between a donor and acceptor capable of ion pairing.20,23 The mechanism (Scheme 2) assumes that electron transfer occurs within the excited or ground state ion pair (precursor complex). The excited-state ion pair can be formed via excitation of the ground-state ion pair or through the diffusion-controlled association of *Ru(bpy)32+ and S2O82‑. In Scheme 2, τ0 and τIP 0 are the lifetimes due to radiative and nonradiative (except for the electron transfer) decay of the free 10312

dx.doi.org/10.1021/jp407573d | J. Phys. Chem. A 2013, 117, 10311−10319

The Journal of Physical Chemistry A

Article

and paired excited state, respectively, ket is the rate constant for the photoinduced electron transfer within the excited-state ion pair (precursor complex), ktet is the rate constant for electron transfer within the ground-state ion pair, kd and k−d are the diffusion-controlled rate constants for formation and dissociation of the ion pair, and KIP is the formation constant of the ground-state ion pair. In this and subsequent discussions we will assume that KIP is equal to kd/k−d (see the SI for details). Based on this model the relationship between (I0/I) (where I0 and I are the emission intensity values in the absence and presence of persulfate, respectively) and S2O82‑ = Q is given by (1 + τ0kq[Q])(1 + KIP[Q]) I0 = I 1 + C[Q]

and r2, estimated from the radius of a sphere of equivalent volume, was ∼3.4 Å.21 Based on the above estimates the value of r0 was assumed ∼9.9 Å and, in common with general practice, δr ≈ 1 Å. Based on the above analysis it is evident that KIP is significantly lower than the values reported by White et al.19 and is also very dependent on the ionic strength (Figure 1).

(1)

where C=

(τ0−1 + k −d)KIP + kd(1 + KIP[Q ]) (τ0−1 + ket + k −d)

(2)

and kq is given by kq =

kd(τ0−1 + ket) τ0−1 + ket + k −d

(3) Figure 1. Calculated KIP for Ru(bpy)32+ and S2O82‑ in water as a function of S2O82‑ concentration in the absence of added electrolyte; [Ru(bpy)32+] = 50 μM; ionic strength changes from 0.003 to 0.3 M.

The above equations were derived under the assumptions that τIP 0 = τ0 and that the emission rate constants are the same for the excited ion pair and free ion (see details in SI). Below we discuss the individual steps of the above model and how they compare with the scheme considered by White et al.19 Measurements of KIP. The value of KIP obtained by White et al.19 through the fitting of their quenching data (see the SI, Figure S1 and eqs S1a and S1b) was ∼1800 M−1 in water. As noted in the Introduction, this value seems unusually high compared to similar systems reported in earlier studies. In addition, the model proposed by White et al.19 does not take into account the ionic strength dependence of KIP. In view of the relatively large number of unknown parameters we searched for an independent and more direct means of determining KIP, other than interpretation of the excited state quenching data, since it contains other unknown variables. First, KIP can be easily estimated on the basis of the Debye−Hückel theory. Equations 4-624,25 were used to estimate KIP defined as the equilibrium constant for the formation of close-contact reactant pairs in which the separation of the centers of the reactants (assumed spherical) is between r0 and (r0 + δr) where r0 = (r1 + r2), the sum of the radii of the two reactants.

(

4πNAvr0 2δr exp KIP =

1000 z1z 2 w(r0) = Dsr0(1 + κr0) κ=

8πNAvμ 1000DskBT

−w(r0) kBT

In order to obtain KIP experimentally the electrochemical oxidation of Ru(bpy)32+ as a function of [S2O82−] (0.05 − 5 mM) was studied in the presence of sodium perchlorate to maintain constant ionic strength. The dependence of the redox potential on persulfate concentration can be described by eq 7,27 where KIP and KIP# are the equilibrium constants for formation of RuII(bpy)32+|S2O82‑ and RuIII(bpy)33+|S2O82‑, respectively. A detailed explanation of the electrochemical method and the analysis of the experimental results are presented in the SI. E = E 0′ + 0.059 log

cox (1 + KIP# [S2 O82 −]) − 0.059 log cred (1 + KIP[S2 O82 −]) (7)

K#IP

Only values of and their dependence on the dielectric constant of the medium could be determined at the relatively low ionic strengths of interest by the above method (Figure 2). Determination of KIP would require higher concentrations of persulfate; however, increasing the persulfate concentration necessarily increases the ionic strength thereby changing the value of KIP. As a result, only the upper limit of KIP could be estimated for solutions of low dielectric constant (Figure S3) and from those experiments K#IP/KIP was estimated to be >30. Based on the value of K#IP in pure water (μ = 0.015 M) being ca. 300 M−1 (Figure 2), the upper limit for KIP is estimated to be less than 10 M−1. The trends in the experimental data are similar to those predicted by the relatively simple Debye− Hückel model (eqs 4−6). The deviations are not surprising since the model does not take into account the changes in the solvation of ions in solvent mixtures of varying composition. Finally, the value of KIP can be estimated from a kinetic analysis of the reaction between Ru(bpy)32+ and S2O82‑ since the thermal reaction should also proceed via the same ion pair (Scheme 2).21 Kinetic measurements of the thermal reaction were performed at an ionic strength of 0.3 M under pseudo-

) (4)

(5)

(6)

w(r0) is the electrostatic work required to bring reactants to the separation distance r, NAv is Avogadro’s number, kB is Boltzmann’s constant, z1 and z2 are the charges on the reactants, Ds is the bulk dielectric constant of the medium and μ the ionic strength. In the present context r1, z1 and r2, z2 refer to Ru(bpy)32+ and S2O82‑, respectively; r1 was taken as ∼6.5 Å26 10313

dx.doi.org/10.1021/jp407573d | J. Phys. Chem. A 2013, 117, 10311−10319

The Journal of Physical Chemistry A

Article

rapidly oxidizes an additional equivalent of Ru(bpy)32+ to give Ru(bpy)33+. On the basis of the above mechanism the rate of formation of Ru(bpy)33+ is given by eq 8 and kobs by eq 9, where ktet is the rate constant for electron transfer within the precursor complex in the thermal reaction. 2k K [S O 2 −][Ru II] d[Ru III] = tet IP 2 8 2 − dt 1 + KIP[S2 O8 ] kobs =

(8)

2k tetKIP[S2 O82 −] 1 + KIP[S2 O82 −] −1

(9) 2‑

If KIP is high (≫1 M ) then the plot of kobs vs [S2O8 ] should be linear at low S2O82‑ concentration and deviates from linearity at higher concentrations. However, under the experimental conditions used in this work the plot was found to be linear in the entire range of persulfate concentration. Based on eq 9 it is concluded that KIP[S2O82‑] ≪ 1 and with [S2O82‑] = 0.1 M KIP is ≪10 M−1. Under these conditions only the product KIPktet = 5.4 × 10−3 M−1 s−1 can be obtained from the slope of the plot (Table 1). These results are consistent with earlier studies of the reaction between Ru(NH3)4(bpy)2+ and persulfate by Fuerholz and Heim21 (Table 1). They found departures from linearity increased with increasing charge on the complexes e.g for the Ru(NH3)5(pzMe)3+ (KIP = 50 M−1, pzMe+ = N-methylpyrazinium) and Ru2(NH3)10(pz)5+ (KIP = 270 M−1, pz = pyrazine). Based on the two independent experimental methods and supported by the Debye−Hückel model, it is evident that the equilibrium constant for the ion pair formation between RuII(bpy)32+ and S2O82‑ in water is ≪10 M−1. Our results are consistent with previously reported equilibrium constants for the formation of ion pairs between similar metal complexes and persulfate as mentioned above21 and do not support the high KIP values reported in ref 19. Lifetime of the *Ru(bpy)32+|S2O82‑ Excited Ion Pair. Due to the low KIP value, the equilibrium concentration of the ion pair is likely to remain relatively low in solvents with high dielectric constants. In an attempt to obtain solutions with higher ion-pair concentrations, [Ru(bpy)3]Cl2 and Na2S2O8 were mixed in THF-H2O (9/1, v/v) mixtures. However, the solubility of the ion pair was found to be low in this solvent mixture and precipitation was observed immediately after mixing. The precipitate was isolated on filter paper by vacuum filtration and the emission lifetime of the precipitate was measured using transient emission spectroscopy. For comparison, the ion pair of Ru(bpy)32+ and sulfate was prepared under the same conditions. The emission lifetime of the ion pair with sulfate was found to be ca. 570 ns under aerobic conditions, similar to the emission lifetime of Ru(bpy)32+ in solution (Figure 4).5,19 However, the emission lifetime of the ion pair with persulfate was too short to be measured with our transient emission apparatus (5 × 107 s−1 (see the SI for details). Emission Quenching. As shown above, the ion-pairing between ground-state Ru(bpy)32+ and S2O82‑ in water is weak

Figure 2. Dependence of KIP on the dielectric constant of the medium; μ = 0.015 M. Dielectric constants for H2O/THF mixtures was obtained from ref 28.

first-order conditions with S2O82‑ in at least 80-fold excess. The reaction between Ru(bpy)32+ and S2O82‑ was monitored at 670 nm, the absorbance maximum of Ru(bpy)33+ (ε670 = 420 M−1 cm−1).29 It is known that Ru(bpy)33+ is not stable in neutral solutions and is reduced back to Ru(bpy)32+ with a rate constant of 3 × 10−3 s−1 at pH 7.0.30 At pH 2 conversion of Ru(bpy)33+ to Ru(bpy)32+ occurs much slower with a rate constant of 5.0 × 10−5 s−1 (SI, Figure S7). In order to limit decomposition of the Ru(bpy)33+ formed in the thermal reaction, measurements were carried out in 0.01 M TFMS. The rate constants were calculated by a nonlinear least-squares fit according to A = A∞(1 − exp(−kobst)). The observed firstorder rate constants at 25 °C are plotted vs persulfate concentration in Figure 3.

Figure 3. First-order rate constant vs persulfate concentration for the thermal oxidation reaction. Inset: absorbance changes at 670 nm vs time for different initial concentration of S2O82‑ and 0.5 mM Ru(bpy)32+.

The thermal reaction between Ru(bpy)32+ and S2O82‑ can be interpreted in terms of the customary three steps: formation of a reactant ion pair, an electron transfer and the dissociation of the product ion pair. The first and third steps are assumed to be diffusion controlled with the electron transfer being rate determining. In addition, the electron transfer produces Ru(bpy)33+ and sulfate radical anion, SO4•− with the latter being a strong oxidant (E0(SO4•−/SO42−) ≈ 2.4 V)31 which 10314

dx.doi.org/10.1021/jp407573d | J. Phys. Chem. A 2013, 117, 10311−10319

The Journal of Physical Chemistry A

Article

Table 1. Rate Constants and Ion-Pair Formation Constants for the Oxidation of Ground-State Ruthenium(II) Complexes with Persulfate at 25 °C in Aqueous Solutions complex Ru(bpy)32+ Ru(NH3)5(pzMe)3+d Ru(NH3)5(pz)2+g Ru(NH3)4(bpy)2+

KIP (M−1)

μ (M) a

ktet (s−1)

0.3 0.1e 0.1h 0.1i

KIPktet (M−1 s−1)

−4c

>5.3 × 10 1.9 ± 0.3f (2.6 ± 0.8) × 102f 3.0 × 102c

b