Photoinduced Electron Transfer Reactions of Hexarhenium(III) Cluster: Oxidative Quenching of [Re₆S₈Cl₆]⁴⁻ Emission by Electron Acceptors

Abstract

The photoredox ability of a hexarhenium(III) cluster, [Re₆(μ ₃-S)₈Cl₆]⁴⁻, in acetonitrile was examined on the basis of the emission quenching experiments of the cluster by a series of neutral organic electron acceptors (A). The photoinduced electron transfer quenching rate constants and the relevant activation parameters demonstrated that the hexarhenium(III) cluster acted as a useful photoredox sensitizer towards various A and the system gave a high charge separation yield owing to electrostatic repulsion between the product ions: [Re₆(μ ₃-S)₈Cl₆]³⁻ and A ⁻. These results were proved explicitly by nanosecond transient absorption spectroscopy.

Introduction

Photoinduced electron transfer (PET) reactions are the fundamental basis of chemical conversion of light energy as well as of synthetic organic photochemistry, and the related studies have been conducted extensively among the past decades [1, 2]. For photocatalytic reactions by redox sensitizers, transition metal complexes are widely employed. In particular, polypyridine ruthenium(II) complexes (RuL₃ ²⁺, L = polypyridine or polydiazine ligand) have been studied extensively [1–4], since the complexes in general possess absorption in the visible region and long excited state lifetime in solution at room temperature. Furthermore, it is well known that some of RuL₃ ²⁺ possesses strong reduction and oxidation abilities towards various substrates: electron acceptors (A) and donors (D) [1–4]. In practice, various RuL₃ ²⁺ complexes represented by Ru(bpy) ²⁺₃ (bpy = 2,2′-bipyridine) have been shown to be potential photoredox sensitizers [1–7]. It is worth emphasizing that, however, it is certainly true that PET reactions of RuL₃ ²⁺ with D give rise to high ion product yields, while those with A do not afford free ions efficiently (i.e., RuL₃ ³⁺ and A ⁻). This is simply because that oxidative quenching of RuL₃ ²⁺ by A produces electrostatically-attractive ion pair (RuL₃ ³⁺ and A ⁻), while reductive quenching affords an electrostatically repulsive RuL₃ ⁺–D ⁺ pair [5–7]. For constructing efficient photoredox systems, therefore, the electrostatic interaction in a product ion pair is worth considering and studying in detail.

We reported previously, on the other hand, that hexarhenium(III) clusters, [Re₆(μ ₃-S)₈ X ₆]⁴⁻ (X = Cl⁻, Br⁻, I⁻, and so forth), showed relatively intense room temperature emission in the visible ~ near infrared region (600–1,000 nm) with the emission lifetime (τ) and quantum yield (Φ) in acetonitrile being τ = 4–11 μs and Φ = 0.015–0.056, respectively [8]. Later, several research groups have reported that various [Re₆ E ₈ L ₆]^z (E = S, Se, or Te; L = X [8–11], CN [11–13], OH/H₂O [12–14],  NCS [15], N-heteroaromatics [16–21], phosphine derivatives [17, 18, 22, 23], HCOO [24], CH3COO [25], and N₃ [26],) clusters showed intense and long-lived emission in solution and solid phases at room temperature. It is worth noting that the τ values of the some of [Re₆E₈L₆]^z clusters in deaerated acetonitrile are much longer (τ = 4–11 μs) than that of Ru(bpy) ²⁺₃ (τ = 850 ns in acetonitrile at 298 K). As another unique characteristics, furthermore, [Re₆E₈L₆]^z possesses a strong reducing ability. As a typical example, the oxidation potential of [Re₆S₈Cl₆]⁴⁻ (=[Re₆]⁴⁻) in acetonitrile is +0.27 V (vs. saturated calomel electrode (SCE)) and, thus, the redox ability of [Re₆]⁴⁻ is remarkable [20, 27, 28]. Owing to the long excited-state lifetime and the strong redox ability, [Re₆]⁴⁻ is thus expected to act as an efficient photoredox sensitizer in various systems. For oxidative quenching of the excited triplet state of [Re₆]⁴⁻ by a neutral electron acceptor (A), in particular, the negative charges on the cluster suggest very strong electrostatic repulsion between the product ions, [Re₆]³⁻ and A ⁻, and, thus, the system enables a high charge separation yield. A detailed study on PET emission quenching of [Re₆]⁴⁻ by electron acceptors would thus contribute to further progresses in the photoredox chemistry of transition metal complexes. In this paper, we demonstrate that [Re₆S₈Cl₆]⁴⁻ undergoes efficient PET reactions with various electron acceptors in acetonitrile [29].

Experimental

Chemicals

[Re₆S₈Cl₆]⁴⁻ abbreviated as [Re₆]⁴⁻ used in this study was the same sample reported earlier [8, 29]. Acetonitrile (Wako Pure Chemicals Co. Ltd.) as a medium throughout the study was purified by distillation over CaH₂. Organic electron acceptors (A) were purchased from Tokyo Kasei Co. Ltd.: for the numbering of A, see Table 1. 4-Nitroanisole (1), 3-nitroanisole (5), methyl 3-nitrobenzoate (6), or 1,4-benzoquinone (12) was recrystallized from n-hexane, an aqueous methanol, methanol, or petroleum ether, respectively. 2,3,5,6-Tetramethyl-1,4-dinitrobenzene (2), 3,4-dimethylnitrobenzene (3), 3,5-dimethylnitrobenzene (4), methyl 4-nitrobenzoate (7), and 1,3-dinitrobenzene (8) were purified by recrystallization from ethanol. Tetrachlorophathalic anhydride (9) and tetrachloro-1,4-benzoquinone (13) were recrystallized from acetone. 1,4-Naphthoquinone (10) and pyromellitic dianhydride (11) were purified by recrystallization from ethanol and 1,4-doxane, respectively, followed by vacuum sublimation. Pyrene (Wako Pure Chemicals Co. Ltd.) was purified successively by column chromatography (silica gel, n-hexane:chloroform = 1:2) and vacuum sublimation.

Table 1 PET emission quenching of [Re₆S₈Cl₆]⁴⁻ by neutral electron acceptors (A) in acetonitrile

Emission Spectroscopy and Other Measurements

Steady-state emission quenching experiments of [Re₆]⁴⁻ by A in CH₃CN at a given temperature (T) were performed with a Hitachi F-4500 spectrofluorometer. Emission lifetime measurements were conducted by using a Nd³⁺:YAG laser (Continum Surelite II, 355 nm, ~6 ns pulse-width) as an excitation light source and a streak camera system (Hamamatsu Photonics, C4334) as a photodetector. In both cases, sample temperature was controlled within 1 °C (13–58 °C) by circulating thermostated water to a sample cell holder by using a thermo-electronic controller (Hitachi 250-0346 or Yamato-Komatsu COOLNICS CTE42A). Sample solutions were deaerated thoroughly by purging an Ar-gas stream over 20 min and sealed at the constriction of a cell prior to the experiments. Cyclic voltammetry of [Re₆]⁴⁻ or A in CH₃CN was carried out by using an electrochemical analyzer (ALS, model 701A) with tetra-n-butylammonium hexafluorophosphate (0.1 M) as a supporting electrolyte.

Nanosecond Time-Resolved Absorption Spectroscopy

Nanosecond 355-nm laser pulses mentioned above were used for excitation of a sample solution. Timing between the laser and a pulsed Xe lamp as a monitoring light source (150 W, Tokyo Instruments Inc.) was controlled by a digital delay generator (DG535, Stanford Research Systems Inc.) and the transient absorption spectrum was measured by a gated multichannel photodiode detector (SMA, Princeton Instruments Inc.) equipped with a polychromator (Jovin-Yvon HR-320).

Determination of the Quenching Rate Constant and Other Parameters

The PET reaction of [Re₆]⁴⁻ with A in acetonitrile is assumed to proceed by the several steps as shown in Scheme 1 [6, 29, 30]. Upon photoexcitation of [Re₆]⁴⁻, the excited triplet state of [Re₆]⁴⁻ (*[Re₆]⁴⁻) undergoes diffusional encounter with A (the rate constant of k ₁₂) and this produces the encounter complex, [*[Re₆]⁴⁻···A]. k ₂₁ is the dissociation rate constant of the encounter complex and, k ₂₃ and k ₃₂ are the forward and backward electron transfer rate constants within the contact ion pair, [[Re₆]³⁻···A⁻], respectively. [[Re₆]³⁻···A⁻] as the PET product undergoes either dissociation to free ions (k ₃₄, [Re₆]³⁻ and A⁻) or back electron transfer to the ground state (k ₃₀).

Scheme 1

PET reaction of [Re₆S₈Cl₆]⁴⁻ (=[Re₆]⁴⁻) with A in a polar solvent

A bimolecular quenching rate constant (k _o) in CH₃CN at a given temperature (T) was determined by the slope value of the Stern–Volmer plot on the basis of emission intensity or lifetime measurements. For the emission intensity quenching experiments, the emission lifetime of [Re₆]⁴⁻ in the absence of A (τ) at a given T was determined by separate experiments. Every emission decay determined at around 700 nm obeyed a single exponential function in the T range of 13–58 °C, which was used to calculate k _o at a given T. Since k _o determined by the emission intensity measurements agreed very well with that by the emission lifetime measurements irrespective of A, the quenching reaction of [Re₆]⁴⁻ by A was concluded to proceed in a dynamic pathway.

Since k _o includes the contributions from both the diffusional and electron transfer steps, a correction for the diffusional effect on k _o was made by Eq. 1,

$$k_{q}^{ - 1} = k_{\text{o}}^{ - 1} - k_{12}^{ - 1}$$

(1)

where k _q represent the corrected activation-controlled quenching rate constant. The k ₁₂ value was calculated by the Smoluchowski equation, Eq. 2,

$$k_{12} = \frac{2RT}{3000\eta }\left[ {2 + \frac{{r_{\text{Re}} }}{{r_{A} }} + \frac{{r_{A} }}{{r_{\text{Re}} }}} \right]$$

(2)

where r _A and r _Re are the radii of A (3.8 Å) and [Re₆]⁴⁻ (5.5 Å), respectively. R and η represent the gas constant and the viscosity of CH₃CN, respectively.

Results and Discussion

Kinetic Analysis of PET Emission Quenching of [Re₆S₈Cl₆]⁴⁻ by Neutral Organic Electron Acceptors

The emission from [Re₆]⁴⁻ in acetonitrile was quenched efficiently in the presence of A. As the k _q data at 298 K were summarized in Table 1, the value increased from 10⁷ to 10¹⁰ M⁻¹s⁻¹ in the decreasing order of the reduction potential of A. To discuss further the point, we evaluated the Gibbs free energy change for the forward electron transfer step, ∆G ₂₃, on the basis of the oxidation potential of [Re₆]⁴⁻ (E([Re₆]³⁻/[Re₆]⁴⁻) = + 0.27 V vs. SCE), the reduction potential of A (E(A/A ⁻) vs. SCE, see Table 1), and the excited triplet state energy of [Re₆]⁴⁻ (E ^em = 1.61 eV, the emission energy at 298 K) as in Eq. 3,

$$\Delta G_{23} = E([\text{Re}_{6} ]^{3 - } /[\text{Re}_{6} ]^{4 - } ) - E(A/A^{ - } ) - E^{em} + w_{p} - w_{r}$$

(3)

where w _p or w _r is the electrostatic work necessary to bring two product ions or two initial reactants together to the close-contact distance (d = r _A + r _Re), respectively, and is given by Eq. 4.

$$w_{r} = w_{p} = \frac{{Z_{a} Z_{b} e^{2} }}{\varepsilon d}$$

(4)

In Eq. 4, Z _a and Z _b are the charge numbers of the two reactants for w _r and those of the two product ions for w _p and, e and ε are the elemental electron charge and the static dielectric constant of CH₃CN (37.5), respectively. In the present experiments, since we used neutral quenchers throughout the study, w _r was zero and w _p = –0.12 V. The ∆G ₂₃ value calculated for each A is included in Table 1. It is clear that the ∆G ₂₃ values are sufficiently negative (i.e., exoergonic) for all quenchers and the k _q value increases with a negative increase in ∆G ₂₃. The results are readily understood by PET emission quenching of [Re₆]⁴⁻.

According to the Scheme 1 and Bock et al. [30], k _q can be expressed as in Eq. 5,

$$k{}_{q} = K_{12} k_{23} \frac{{k_{30} }}{{k_{30} + k_{32} }}$$

(5)

where K ₁₂ is the formation constant of the encounter complex and can be calculated by the Fuoss–Eigen equation, Eq. 6,

$$K_{12} = \frac{{4\pi Nd^{3} }}{3000}\exp ( - w_{r} /RT)$$

(6)

where N is the Avogadro number and K ₁₂ is calculated to be 2.03 M⁻¹ for the Re⁴⁻—neutral quencher systems (w _r = 0 and d = 9.3 Å) at 298 K. When the transition-state theory is applied to the forward electron transfer step (k ₂₃), Eq. 5 can be rewritten as in Eq. 7,

$$k_{q} = K_{12} \nu_{23} F\exp ( - \Delta G_{23}^{*} /RT)$$

(7)

where ν ₂₃ is the frequency factor of the k ₂₃ step (assumed to be 10¹² s⁻¹) and F = k ₃₀/(k ₃₀ + k ₃₂). ∆G ₂₃* is the Gibbs activation free energy for the electron transfer step and is given by the Marcus theory [31], Eq. 8,

$$\Delta G_{23}^{*} = \frac{\lambda }{4}\left[ {1 + \frac{{\Delta G_{23} }}{\lambda }} \right]^{2}$$

(8)

where λ represents the sum of the inner—(λ _i) and outer-sphere reorganization energies (λ _o). Combining Eqs. 7 and 8, we obtain Eq. 9.

$$k_{q} = K_{12} \nu_{23} F\exp \left[ { - \frac{\lambda }{4}\left( {1 + \frac{{\Delta G_{23} }}{\lambda }} \right)^{2} /RT} \right]$$

(9)

When k ₃₀ ≫ k ₃₂, Eq. 5 can be simplified to Eq. 10.

$$k_{q} = K_{12} k_{23}$$

(10)

By combining Eqs. 9 and 10, furthermore, k _q in a natural logarithmic form is given by Eq. 11 at |∆G ₂₃| ≪ 2λ.line feed, similar to “Case II” Case I

$$RT\ln k_{q} = RT\ln (K_{12} \nu_{23} ) - \frac{\lambda }{4} - \frac{{\Delta G_{23} }}{2}$$

(11)

When k ₃₀ ≪ k ₃₂ and F = k ₃₀/k ₃₂ in Eq. 5, on the other hand, we obtain Eq. 12,

$$k_{q} = K_{12} K_{23} k_{30}$$

(12)

where K ₂₃ = k ₂₃/k ₃₂. Similar treatment of Eq. 12 with that obtaining Eq. 11 gives Eq. 13.

Case II

$$RT\ln k_{q} = RT\ln k_{q} K_{12} - \Delta G_{23}$$

(13)

Eqs. 11 and 13 indicate that the slope value of the RTlnk _q vs. ∆G ₂₃ plot should be −0.5 when the product ions, [[Re₆]³⁻···A⁻], undergoes efficient charge separation (k ₃₄) or back electron transfer to the ground state (k ₃₀) (Case I, k ₃₀ ≫ k ₃₂), while that is −1.0 if the k ₃₂ process dominates to k ₃₀ (Case II, k ₃₀ ≪ k ₃₂).

Figure 1 shows the relationship between the lnk _q and ∆G ₂₃ values. At ∆G ₂₃ < −0.8 eV, the RTlnk _q value levels off at the diffusion-limited rate constant value, while that increases with a negative increase in ∆G ₂₃ in −0.2 > ∆G ₂₃ > −0.7 eV. It is worth emphasizing that the slope value of the plot in −0.2 > ∆G ₂₃ > −0.7 eV is −0.48, demonstrating the PET emission quenching of [Re₆]⁴⁻ by A proceeds by the Case I mechanism (i.e., k ₃₀ ≫ k ₃₂). Since k ₃₀ includes the rate constant of charge separation to [Re₆S₈Cl₆]³⁻ and the anion radical of A (A ⁻, k ₃₄) as well as the back electron transfer rate constant to the ground state reactants (k _b, see Scheme 1), the present kinetic analysis does not necessarily provide direct evidence for the charge separation to [Re₆]³⁻ and A ⁻. To discuss further the point, therefore, we conducted the T dependence of the PET emission quenching and transient absorption spectroscopy of the [Re₆]⁴⁻ − A system.

Fig. 1

A relationship between RTlnk _q and ∆G ₂₃ for PET emission quenching of [Re₆S₈Cl₆]⁴⁻ by A in CH₃CN at 298 K

Temperature Dependence of PET Emission Quenching of [Re₆S₈Cl₆]⁴⁻ by A

We studied the T (285–331 K) dependence of PET emission quenching of [Re₆]⁴⁻ by 13 electron acceptors in CH₃CN. The activation free energy (∆G ₂₃*), activation enthalpy (∆H ₂₃*), and activation entropy (∆S ₂₃*) thus determined by the Eyring plots (ln(k _q/T) vs. T ⁻¹) are included in Table 1.

As seen in Table 1, a negative increase in ∆G ₂₃ from −4.4 to −34.1 kcal/mol rendered the decreases in both ∆G ₂₃* (from 6–7 to ~3 kcal/mol) and ∆H ₂₃* (from 3–4 to 0.5–1.0 kcal/mol), while ∆S ₂₃* was remained almost constant at −8 ~ −11 eu (=cal/mol deg) irrespective of ∆G ₂₃: T∆S ₂₃* = −2.4 ~ −3.3 kcal/mol at 298 K. Such results remind us the T dependence of reductive quenching of Ru(bpy) ²⁺₃ (= [Ru]²⁺) by aromatic amines (D) [5], in which the analogous ∆G ₂₃ dependences of the activation parameters with those determined in the present study have been observed. In the case of oxidative quenching of [Ru]²⁺ by A [6], on the other hand, a negative increase in ∆G ₂₃ resulted in apparent negative ∆H ₂₃* values and a large decrease in ∆S ₂₃*. The results for PET quenching of [Ru]²⁺ have been explained in terms of the electrostatic interaction in the product ion pair, [[R₆u]³⁺···A⁻]. For a [Ru]²⁺ − D system, namely, the electrostatic interaction between the product ions is repulsive ([[Ru]⁺···D⁺]) and, this usually leads to the ∆G ₂₃ dependences of the activation parameters as described above. On the other hand, since the PET reaction of a [Ru]²⁺ − A system produces an electrostatically–attractive ion pair ([[Ru]³⁺···A⁻]), back electron transfer from the ion pair to the excited state encounter complex proceeds very efficiently, rendering an apparent negative ∆H ₂₃* value and a large decrease in ∆S ₂₃* of the quenching, as reported in detail previously [5–7]. Since oxidative quenching of the excited triplet state of [Re₆]⁴⁻ by A produces an electrostatically–repulsive product ion pair ([[Re₆]³⁻···A⁻]), the analogous T dependences of the activation parameters with those for reductive quenching of [Ru]²⁺ is quite reasonable. Thus, it can be concluded that the PET reaction of a [Re₆]⁴⁻ − A system undergoes by the Case I mechanism (i.e., k ₃₀ ≫ k ₃₂), which agrees very well with the conclusion of the kinetic analysis of the data in Fig. 1.

Transient Absorption Spectrocopy of [Re₆S₈Cl₆]⁴⁻: Pyromellitic Dianhydride (PDA) System

To obtain the direct evidence for the charge separation process of a [Re₆]⁴⁻ − A system, we conducted nanosecond transient absorption spectroscopy. As a typical example, the transient absorption spectrum of a [Re₆]⁴⁻—pyromellitic dianhydride (PDA) system in CH₃CN at 298 K observed during the first 500 ns after laser excitation is shown in Fig. 2. In the [Re₆]⁴⁻ − PDA system, we observed transient absorption in the wavelength (λ) region of 500–700 nm, which was assigned safely to the anion radical of PDA as reported by Hino et al. [32]. Unfortunately, however, we could not detect the absorption band of [Re₆]³⁻ in the λ region studied. The time profile of the absorption band at λ = 660 nm is shown in Fig. 3. Figure 3 demonstrates clearly that the anion radical of PDA decays very slowly with the second-order kinetics and survives even at around 70 μs after laser excitation. Knowing the molar absorption coefficient (ε) of the absorption band at λ = 660 nm to be ε = 3.8 × 10⁴ /M/cm [32], we evaluated the decay time constant of the PDA anion radical to be 2.7 × 10⁸ /M/s.

Fig. 2

Transient absorption spectrum (gate time = 500 ns) of the [Re₆S₈Cl₆]⁴⁻ (5.3 × 10⁻⁴ M) − PDA (1.9 × 10⁻³ M) system in CH₃CN at 298 K

Fig. 3

Time profile of the absorbance of the PDA anion radical at 660 nm

As a reference system producing an electrostatically-attractive product ion pair by PET, we also studied transient absorption of a pyrene (Py)—PDA system as shown in Fig. S1 in Supporting Information. As seen in Fig. S1, the absorption band of the Py cation radical was observed at around 445 nm in addition to the absorption bands of the anion radical of PDA. Both the anion radical of PDA and the cation radical of Py decayed by the second-order kinetics and the decay rate constants of the Py cation radical (ε = 4.5 × 10⁴ /M/cm at λ = 445 nm) and the PDA anion radical were 1.8 × 10¹⁰ and 1.4 × 10¹⁰ /M/s, respectively, whose values were 50 ~ 60 times faster than that of the PDA anion radical produced in the [Re₆]⁴⁻ − PDA system. These results demonstrate explicitly that the electrostatic interaction between the product ions governs the charge separation yield and the lifetimes of the free ions produced by a PET reaction. In the present [Re₆]⁴⁻ − A system, thus, the negative charges on [Re₆]⁴⁻ plays decisive role in determining the reaction mechanism (Case I) and the charge separation yield of the PET reaction.

Conclusions

The present study demonstrated that hexarhenium(III) complex, [Re₆S₈Cl₆]⁴⁻ = [Re₆]⁴⁻, possesses a high photoredox ability toward a neutral electron acceptor, as proven by the kinetic analysis of the PET emission quenching rate constant, the relevant activation parameters, and nanosecond transient absorption spectroscopy. It has been also demonstrated that the electrostatic interaction between the product ion pair produced by PET is the one of the important factors determining the charge separation yield of the product ion pair. Besides [Re₆S₈Cl₆]⁴⁻, a number of negatively-charged hexarhenium(III) clusters with different oxidation potentials have been hitherto reported and such clusters will be also acted as a strong photoredox sensitizer or catalyst.