Estimation of Electron Distribution over Dinuclear Organometallic Molecular Wires by “IR Tag” Analysis of Ancillary Acyl-Cp Ligands

Abstract

Understanding the details of electronic properties of mixed-valence (MV) states of organometallic molecular wires is essential to gain insights into electron transfer phenomena. Although the field of MV chemistry is matured, there remain issues to be solved, which cannot be accessed by the conventional analytical methods. Here, we describe the synthesis and properties of diruthenium bridging (diethynylbenzene)diyl wires, (μ-p and m-C≡C–C₆H₄–C≡C){RuCp^R(dppe)}₂1, with the acyl-substituted cyclopentadienyl rings [Cp^R: Cp^e; η⁵-C₅H₄COOMe (a-series: ester derivatives), Cp^a; η⁵-C₅H₄COMe (b-series: acetyl derivatives)], which are installed as IR-tags to estimate electron densities at the metal centers in the MV species. The electrochemical and IR/near IR-spectroelectrochemical studies reveal that the two metal centers in the para-isomers p-1a,b⁺ interact with each other more strongly than those in the meta-isomers m-1a,b⁺. Electron-spin resonance study also supports the radicals being delocalized over the Ru-(p-C≡C–C₆H₄–C≡C)–Ru moieties in p-1a,b⁺. The spectroelectrochemical IR study shows significant higher-energy shifts of the ν(C=O) vibrations brought about upon 1e-oxidation. Spectral simulation on the basis of the Bloch equations allows us to determine the electron transfer rate constants (k_et) between the two metal centers being in the orders of 10¹² s^–1 (p-1⁺) and 10⁹ s^–1 (m-1⁺). The shifts of the ν(C=O) bands reveal that the charge densities on the para-isomers p-1a,b⁺ are widely delocalized over the Ru-(p-C≡C–C₆H₄–C≡C)–Ru linkages in contrast to the meta-isomers m-1a,b⁺, where the electron densities are mainly localized on the metal fragments, as supported by the density functional theory and time-dependent density functional theory studies as well as comparison with the reference mononuclear acetylide complexes, C₆H₅–C≡C–RuCp^R(dppe) 2. We have successfully demonstrated that the carbonyl groups (>C=O) in the ancillary Cp ligands also work as IR-tags to report detailed information on the electron densities at the metal centers and the electron distribution over MV complexes as well.

Introduction

Mixed-valence (MV) chemistry of inorganic and organometallic complexes has been continuously developed to deepen the fundamental understanding of the electron transfer processes involved since the discovery of the monumental Creutz–Taube ion, [Ru(NH₃)₅-μ-pyrazine-Ru(NH₃)₅]⁵⁺, in 1969, and the complex is regarded as the first MV coordination complex.¹ A central issue of this area is the understanding of the electronic properties of the MV state, that is, the extent of charge delocalization over the molecule and the electron transfer rate between the two redox sites. Robin and Day classified the MV complexes into three categories, that is, a fully delocalized state (class III), a fully localized state (class I), and an intermediate state between them (class II).² For the determination of the class, various physicochemical methods such as electrochemistry, electron-spin resonance (ESR), Mössbauer, X-ray photoelectron spectroscopy (XPS), and UV–vis–NIR (near IR) absorption spectroscopy have been used. However, discussion of the classification is not always straightforward. For example, the comproportionation constant (K_C), which indicates the thermodynamic stability of the MV species, can be readily obtained from the CV study, but the redox potentials are strongly influenced by the environment such as the counter anions and the solvents used for the measurements.^3,4 The intervalence charge transfer (IVCT) band in the NIR region observed for the MV species is commonly used for their classification. According to the Marcus–Hush theorem,⁵ the strength of electronic coupling (V_ab) can be determined from the IVCT band parameters. Because, however, inorganic and organometallic MV complexes frequently show multiple and overlapped NIR bands, identification of the IVCT band is frequently ambiguous.⁶

IR spectroscopy is another powerful method to evaluate the electronic structure of MV complexes because the timescale of IR spectroscopy is short enough (∼10^–12 s) to evaluate electron transfer processes of MV complexes. In the seminal work by Kubiak and Ito using the IR tag technique, they reported the electron transfer rates for a series of MV complexes composed of the two μ₃-oxotriruthenium clusters bridged by a pyrazine linker on the basis of the detailed analysis of the IR data of the terminal CO ligands.⁷ The successful IR analysis involving line shape analysis afforded the electron transfer rate constants (k_et) of the MV complexes, which were in the order of 10¹² s^–1. On the other hand, studies on organometallic MV systems by the IR-tag technique are rare. Winter reported dinuclear MV complexes with the vinylruthenium fragments having supporting CO ligands,^8,9 and on the basis of the spectroelectrochemical IR study, the localized and delocalized charge states were successfully distinguished. While the methods are relatively simple and provide a variety of information, the studies reported to date are mainly limited to the MV complexes with CO ligands showing intense IR absorptions.¹⁰⁻¹³

Multinuclear complexes bearing the group 8 metal acetylide fragments, CP(dppe)M–C≡C [M = Fe, Ru; CP = cyclopentadinyl (Cp) or 1,2,3,4,5-pentamethylcyclopentadienyl (Cp*); dppe = 1,2-bis(diphenylphosphino)ethane], have been extensively studied due to the thermodynamic stability of their 1e-oxidized monocationic MV states caused by the electron-rich metal fragments, and such a feature allows studies of a wide variety of one- and two-dimensional multinuclear species.¹⁴⁻¹⁹ Deep understanding of electronic structures of MV is important. For example, Low proposed another MV state called “bridge-localized state”. 1e-Oxidation of Cp*(dppe)Ru–C≡C–Ar–C≡C–Ru(dppe)Cp* (Ar = 1,4-benzenediyl, 1,4-naphthalenediyl, and 9,10-anthracenediyl) formed the MV species, where the positive charge is mainly localized on the organic bridging part rather than on the metal centers, and the trend became remarkable as the π-system of the bridging aromatic ring was enlarged.²⁰ Furthermore, details of the electronic structures of such ‘‘bridge-localized’’ states have been investigated on the basis of a theoretical study, which showed that the conformation of the Ru fragments is a dominant factor for the charge localization/delocalization. Detailed density functional theory (DFT) analyses by Kaupp suggested that the parallel and antiparallel conformational isomers with respect to the two Cp*Ru(dppe) fragments exhibit the bridge-localized state, while the twisted conformation causes the metal-localized state (Scheme 1).²¹

Scheme 1. Schematic Descriptions of Metal-Localized and Bridge-Localized States for CP(dppe)Ru–C≡C–Ar–C≡C–Ru(dppe)CP Complexes Associated with Their Conformational Isomers.

To experimentally clarify the charge distribution, the electron density, that is, the oxidation states, at the ruthenium centers should be determined. We envisaged that new metal fragments with IR-active pendants should provide information on the electron density at the metal centers. Herein, we report the synthesis and characterization of dinuclear ruthenium complexes bridged by the p- and m-(diethynylbenzene)diyl ligands (1) with the Cp ligands having the IR-active acyl groups, Cp^RRu(dppe) [Cp^R: Cp^e; η⁵-C₅H₄COOMe (a-series: ester derivatives), Cp^a; η⁵-C₅H₄COMe (b-series: acetyl derivatives)], together with the reference mononuclear complexes 2. The electronic structures of the MV species thereof are analyzed on the basis of the ν(C=O) vibrations (Figure 1). Dinuclear MV complexes with (diethynylbenzene)diyl bridges have been studied extensively^22,23 and thus would be suitable motifs to verify the usefulness of IR-tags.

Figure 1.

Structures of mono- and di-nuclear complexes with IR-active ruthenium fragments studied in the present work.

Results and Discussion

Synthesis and Characterization

Ruthenium chloride precursors, Cp^RRu(dppe)Cl [R = COOCH₃ (3a), COCH₃ (3b)], were synthesized by the reaction of the Cp^RNa derivatives with RuCl₂(PPh₃)₃, followed by the ligand substitution with dppe in good yields in a manner similar to the synthesis of the Cp derivative.²⁴ Sequential treatment of 3a and 3b with AgBF₄ and the appropriate terminal alkyne in the presence of DBU afforded the corresponding mono- and di-nuclear complexes 1 and 2 in 45–75% yields (Scheme 2). It is noted that the conventional synthetic method using Na⁺/K⁺ PF₆^–25 instead of the silver salt left the starting chloride precursors 3 untouched probably due to the stronger Ru–Cl bonds caused by the electron-withdrawing acyl groups.

Scheme 2. Synthetic Routes to 1 and 2.

Spectroscopic data of the complexes are summarized in Table 1. ¹H NMR spectra of the mononuclear complexes 2a and 2b contain the methyl signals for the ester and acetyl groups at 3.60 and 1.95 ppm, respectively. ³¹P{¹H} NMR signals for the dppe ligands are observed at 85.0 (2a) and 83.9 (2b) ppm, which are comparable to those for the acetylides bearing the CpRu(dppe) (δ_p ∼ 87 ppm) and Cp*Ru(dppe) (δ_p ∼ 81 ppm) groups.²² Symmetric structures of the dinuclear complexes 1 in solution are confirmed by the single sets of the ¹H and ³¹P NMR signals for the Cp^R (R = COOCH₃, COCH₃) and dppe ligands. The ν(C≡C) and ν(C=O) vibrations are observed as single bands around 2080 and 1650–1705 cm^–1, respectively. In general, diethynylbenzene compounds show single ν(C≡C) vibrations due to the asymmetric coupling of the two C≡C vibrations, while those of the symmetric modes are usually silent.²⁶ The ν(C≡C) vibrational stretching frequencies for the ester and acetyl derivatives 1–2 are virtually identical, suggesting that the electron densities at the Ru centers are similar. Molecular structures of 1–2 determined by X-ray crystallography are shown in Figure 2. The Ru atoms in 1 and 2 adopt the three-legged piano stool geometries and the Ru–C≡ distances ranging between 2.003 and 2.020 Å are comparable to that for CpRu(dppe)–C≡C–C₆H₅ (4; Ru–C: 2.001 Å, Ru–P: 2.238 Å, Ru–C_cp: 2.257 Å).^27,28 However, the slightly longer Ru–P bonds (Δd = 0.009–0.019 Å) and the shorter Ru–C_cp bonds (Δd 0.005–0.018 Å) compared to those in 4 should be due to the decreased and increased back donation from the ruthenium centers, respectively. The Cp_centroid–Ru–Ru–Cp_centroid dihedral angles are 180° (p-1a/b), 117° (m-1a), and 125° (m-1a). The through-space Ru–Ru distances are determined for the para- and meta-isomers to be ∼12.1 and ∼10.7 Å, respectively.

Table 1. Selected Spectroscopic Data for 1–3.

	1H NMR/δHa			IR/cm–1e
complex	C5H4-Rb	CH3 in Rc	31P NMR/δPa,d	ν(C≡C)	ν(C=O)
p-1a (R = COOMe)	5.44, 4.57	3.43	84.9	2081	1703
p-1b (R = COMe)	5.51, 4.52	1.90	83.9	2081	1650
m-1a (R = COOMe)	5.48, 4.58	3.41	84.9	2075	1703
m-1b (R = COMe)	5.53, 4.53	1.92	84.1	2074	1650
2a (R = COOMe)	5.47, 4.60	3.47	85.0	2085	1703
2b (R = COMe)	5.54, 4.54	1.95	83.9	2085	1650
3a (R = COOMe)	5.24, 4.14	3.60	79.1		1705f
3b (R = COMe)	5.30, 4.14	1.95	78.2		1653f

^aIn CDCl₃.

^bMultiplet signals, 4H for each signal of 1 and 2H for each signal of 2 and 3

^cSinglet signals (3H).

^d1H-decoupled spectra, singlet signals.

^eIn CH₂Cl₂.

^fAs KBr pellets.

Figure 2.

ORTEP diagrams of mono- (2) and di-nuclear complexes (1) drawn with the thermal ellipsoids at 50% level. Hydrogen atoms and solvents are omitted for clarity.

Electrochemical Study

Redox properties of the Ru complexes were examined by cyclic voltammetry (Figure 3 and Table 2). CV charts of the mononuclear complexes 2 contain single partially reversible redox waves with half-wave potential of around 40 mV (vs Cp₂Fe/Cp₂Fe⁺ couple), which are anodically shifted by 120 mV compared to that of Cp(dppe)Ru–C≡C–C₆H₅4, as a result of the decreased electron densities at the metal centers caused by the electron-withdrawing acyl substituents on the Cp rings. The E_1/2 values of 2a,b are almost identical, suggesting the comparable electron-withdrawing effects of the ester and acetyl groups. For the dinuclear para-isomers p-1a,b, two fully reversible waves around −190 and 70 mV (vs FeCp₂/FeCp₂⁺ couple, 100 mV/s) are observed. The first oxidation processes are shifted cathodically compared to the mononuclear complexes 2 due to the electronic interactions between the two ruthenium centers and also inductive effects due to the attachment of the two electron-rich CpRu(dppe) donors to the bridging ligand. From the separation of the two redox waves (ΔE), the comproportionation constants (K_C) are determined to be 1.35 × 10⁴ (p-1a) and 9.00 × 10³ (p-1b). In contrast to the para-isomers p-1, the meta-isomers m-1 show two irreversible redox waves in the range of 0–400 mV when scanned at 100 mV/s.

Figure 3.

Cyclic voltammograms of 1–2 ([complex] = ∼1.0 mM in CH₂Cl₂, [NBu₄][PF₆] = 0.1 M, W.E. and C.E. Pt, R.E. Ag/Ag⁺, scanned at 100 mV/s).

Table 2. Selected Electrochemical Data for 1–2 and Related Complexes^a.

compound	E1/21/mV [ipc/ipa]	E1/22/mV [ipc/ipa]	ΔE/mV	KC
p-1a	–190 (−195) [1.0]	55 (50) [1.0]	245 (245)	1.35 × 104
p-1b	–160 (−165) [1.0]	70 (65) [1.0]	230 (230)	9.0 × 103
m-1a	21 (5) [0.4]	(170) [0.4]	(165)	6.9 × 102
m-1b	75 (55) [0.5]	238 (225) [0.3]	163 (170)	8.4 × 102
2a	42 (25) [0.6]
2b	45 (41) [0.9]
RuC≡CC6H5 (4)15	–80
(μ-C≡C–p-C6H4–C≡C) Ru2 (5)15	–320	–90	230	9.0 × 103
(μ-C≡C–m-C6H4–C≡C) Ru2* (6)23	–360	–200	160	6.7 × 102

^aData obtained by DPV measurements are shown in parentheses. Ru = Ru(dppe)Cp, Ru* = Ru(dppe)Cp*. Potentials vs Cp₂Fe/Cp₂Fe⁺ couple.

Upon increasing the scan rate to 1000 mV/s, reversibility of the waves was improved. When the range of −500 to 200 mV was swept at a rate of 100 mV/s, partially reversible 1e-oxidation processes were observed, indicating that the monocationic species of the meta-isomers m-1a,b are stable on the timescale of the electrochemical measurements. Then, differential pulse voltammetry (DPV) measurements (see the Supporting Information) was carried out to obtain the K_C values of 6.9 × 10² (m-1a) and 8.4 × 10² (m-1b), which are considerably smaller than those of the para-isomers p-1a,b due to the connection through the cross-conjugated meta-linkages. In spite of the small K_C values for the meta-isomers, the monocationic species turned out to be populated enough in solutions.

The redox potentials and K_C values of the ester and acetyl derivatives p-1 (K_C ∼ 1 × 10⁴) and m-1 (K_C ∼ 7–8 × 10²) are very similar, suggesting that the electronic interactions between the metal centers are akin to each other. The K_C values of the dinuclear para-isomers p-1 are comparable to that of the related Cp analogue [p-5, (μ-p-C≡C–C₆H₄–C≡C){RuCp^R(dppe)}₂]. Similarly, the K_C values of the meta-isomers m-1 are very close to that of the Cp* derivative 6 [μ-m-C≡C–C₆H₄–C≡C){RuCp*(dppe)}₂].^28,29 These facts suggest that the thermodynamic stability of the MV species of the dinuclear complexes is rather insensitive to the electron-donating properties of the substituted Cp rings.

Spectroelectrochemical IR Study

Spectroelectrochemical IR studies were performed for CH₂Cl₂ solutions in the presence of 0.1 M [NBu₄][PF₆] as an electrolyte by using an OTTLE cell (Figures 4 and 5 and Table 3). We first examined the mononuclear reference complexes 2a,b, which showed single C=O stretching vibrations around 1700 cm^–1 (for the ester groups) and around 1650 cm^–1 (for the acetyl groups) (Figure 4a,b). In addition, the intense bands observed around 2100 cm^–1 were unequivocally assigned to the C≡C stretching vibrations. Upon oxidation of 2, the C≡C and C=O bands were replaced by the bands at around 1930 cm^–1 [ν(C≡C)] and 1728 (for 2a) and 1686 cm^–1 (for 2b) [ν(C=O)] with the decreased intensities. It is noted that isosbestic points were observed, indicating the occurrence of a two-component system involving direct conversions from 2 to 2⁺. In addition to the significant shifts of the C≡C stretching bands to lower energies (Δν ∼ 150 cm^–1), the C=O bands also shifted to higher energies by 25 cm^–1 for 2a (1728 cm^–1) and 36 cm^–1 for 2b (1686 cm^–1) (Figure 5a,b), suggesting that the carbonyl substituents are sensitive to the electron densities at the ruthenium centers.

Figure 4.

IR spectral changes of 1 and 2 observed in an OTTLE cell upon application of bias voltages ([complex] ∼ 10 mM, 0.1 M [NBu₄][PF₆] in CH₂Cl₂, W.E. and C.E. Pt, and R.E. Ag/Ag⁺). The spectra for the three oxidation states are drawn in red (neutral), blue (monocation), and green (dication). The isosbestic points are indicated with the arrows in the insets.

Figure 5.

Parts of IR spectra [ν(C=O) region] for 1ⁿ⁺ and 2ⁿ⁺ (n = 0–2) ([complex] = ∼10 mM, [NBu₄][PF₆] = 0.1 M in CH₂Cl₂, W.E. and C.E. Pt, and R.E. Ag/Ag⁺). The spectra for the three oxidation states are drawn in red (neutral), blue (monocation), and green (dication).

Table 3. IR Data for 1ⁿ⁺–2ⁿ⁺ (n = 0–2) Obtained with an OTTLE Cell^a.

compound	n	ν(C≡C)/cm–1	ν(C=O)/cm–1
2an+	0	2085 (s)	1703 (s)
	1	2081 (w), 1934 (w)	1728 (s)
2bn+	0	2085 (s)	1650 (s)
	1	2083 (w), 1932 (w)	1686 (s)
p-1an+	0	2081 (s)	1703 (s)
	1	2073 (w), 2009 (s), 1972 (s), 1920 (w)	1712 (s)
	2	2001 (w), 1938 (w)	1727 (m)
p-1bn+	0	2081 (s)	1651 (s)
	1	2072 (w), 2009 (s), 1969 (s), 1919 (w)	1667 (s)
	2	1995 (w), 1938 (w)	1685 (m)
m-1an+	0	2075 (s)	1703 (s)
	1	2073 (s), 2001 (w)	1704 (m), 1726 (m)
	2	2072 (w), 1934 (w)	1728 (m)
m-1bn+	0	2074 (s)	1650 (s)
	1	2071 (w), 2001 (w)	1651 (m), 1686 (m)
	2	1931 (w)	1686 (m)

^a([Complex] = ∼10 mM, [NBu₄][PF₆] = 0.1 M in CH₂Cl₂, W.E. and C.E. Pt, and R.E. Ag/Ag⁺).

Encouraged by these results, the spectroelectrochemical IR measurements were extended to the dinuclear complexes (1). For the para-isomers p-1a,b (Figures 4c and 5c,d), two-step oxidation processes were observed. Upon application of bias voltages, the weak ν(C≡C) vibrations of the neutral species p-1a,b were replaced by the intense bands around 1970–2020 cm^–1 and the weak bands around 2070 and 1920 cm^–1, which were assignable to p-1a,b⁺. The four ν(C≡C) vibrations observed for p-1a,b⁺ are similar to those reported for the Cp* analogues and are derived from the occurrence of rotamers (Figure 1).³⁰ Further oxidation caused the second process resulting in the replacement of the bands around 2000 cm^–1 by the new weak bands around 1940 cm^–1, which were ascribed to the dioxidized species p-1a,b²⁺. For the first step, similar observations were reported for the monocationic dinuclear ruthenium complexes with the Cp* and Cp rings.^20,28 The large changes of the intensities may be caused by the coupling of two C≡C bonds. Along with the changes of the ν(C≡C) bands, the C=O vibrations were also shifted in a stepwise manner (Figure 5c). For p-1a, the ν(C=O) bands appeared at 1703 (p-1a), 1712 (p-1a⁺), and 1727 cm^–1 (p-1²⁺). Similarly, for p-1b, the sequential and stepwise shifts of the ν(C=O) band of the acetyl groups was observed: 1651 cm^–1 (p-1b) → 1667 cm^–1 (p-1b⁺) → 1685 cm^–1 (p-1b²⁺) (Figure 5d). It is noted that the ν(C=O) vibration energies of the neutral and dicationic species for p-1a,b are very similar to those of the neutral and monocationic species of the mononuclear complexes 2a,b, suggesting that the electron densities at the ruthenium centers in the neutral and cationic species are similar with each other. Interestingly, during the oxidation processes, the intensities of the ν(C≡C) bands drastically changed, while those of the ν(C=O) bands were not affected significantly. This is because the ν(C≡C) vibration is symmetry forbidden, and the degree of charge localization/delocalization strongly affects the symmetry of the molecules, leading to the drastic changes of the intensity. In addition, vibrations of the C≡C bonds in a molecule (such as 2) may couple to show, for example, symmetrically and antisymmetrically coupled vibrations and make analysis difficult. Thus, analysis based on the ν(C=O) bands provides another merit.

Spectroelectrochemical IR measurements of the meta-isomers m-1 were also performed (Figures 4e,f and 5e,f), and the following changes were noted for ν(C≡C) [∼2075 cm^–1 (m-1) → ∼2000 cm^–1 (very weak: m-1⁺) → ∼1931, 2000, 2070 cm^–1 (m-1²⁺)] and for ν(C=O) [1703 cm^–1 (m-1a) → 1726, 1704 cm^–1 (m-1a⁺) → 1728 cm^–1 (m-1a²⁺); 1650 cm^–1 (m-1b) → 1651, 1686 cm^–1 (m-1b⁺) → 1686 cm^–1 (m-1b²⁺)]. The ν(C=O) bands of m-1a⁺ were overlapped with those of m-1a and m-1a²⁺.

To obtain electron transfer rate constants (k_et) of the MV species, simulation with the VIBEX GL program, which is based on the Bloch equation, was performed (Figures 6 and 7).^7,31 As shown in Figure 7, upon changing the k_et values in the range between 10¹¹ and 10¹³ s^–1, the shapes of the simulated spectra apparently changed, and the k_et values were obtained by the fitting analysis of the experimental and calculated spectra. The band shapes are successfully reproduced for p-1a⁺, p-1b⁺, and m-1a⁺, although the shoulder band at 1728 cm^–1 should be taken into account for p-1a⁺.³² Overlapping with the stretching bands of the Ph groups hindered the line fitting analysis of m-1b⁺.

Figure 6.

Experimental and simulated IR spectra [ν(C=O) region] for 1⁺ (red dashed lines: the sum of the simulated curves, green dashed lines: individual simulated curves, and blue lines; experimentally obtained curves).

Figure 7.

Band shape analysis of the ν(C=O) stretching vibrations for 1⁺.

The k_et values of p-1a⁺ and p-1b⁺ are determined to be 6 × 10¹² and 6 × 10¹² s^–1, respectively. The k_et value for m-1a⁺ was estimated to be at most 1 × 10¹⁰ s^–1 because in the k_et region below 1 × 10¹⁰ s^–1, the spectral change is rather insensitive to k_et, and therefore, k_et cannot be determined accurately on the basis of the line shape analysis.

Spectroelectrochemical UV–Vis–NIR Study

To determine electronic coupling V_ab on the basis of the NIR spectra of 1⁺, spectroelectrochemical measurements in an OTTLE cell were performed for CH₂Cl₂ solutions of 1 containing 0.1 M [NBu₄][PF₆] as an electrolyte (Figure 8 and Table 4). For the neutral species 1, UV absorption bands were observed around 290–360 nm, which were characteristic π–π* bands admixed with ligand-to-metal charge transfer bands.²⁰ Upon application of the bias voltage to the para-isomers p-1a,b, the UV bands disappeared and new bands assignable to the monocationic species p-1a,b⁺ appeared in the visible (500 nm, π–π*) and NIR regions (1000–2000 nm) (for the assignments, see below). Upon increment of the bias voltage, the bands due to p-1a,b⁺ were replaced by the new bands around 750 nm, which were ascribed to the dicationic species p-1a,b²⁺. On the other hand, the absorption band for the neutral meta-species m-1a at 298 nm was weakened upon 1e-oxidation and the new bands for m-1a⁺ appeared at 796 and 1326 nm. Further oxidation led to the replacement of the NIR band by that at 1460 nm (for m-1a²⁺) together with the increase of the intensity of the band at 796 nm, suggesting sequential generation of the mono- and di-cationic species in accord with the IR measurements discussed above. For m-1b, we could not obtain reproducible results probably due to the instability of its cationic species, and thus, in the following section, discussion for the meta-isomers will be focused on m-1a.³³

Figure 8.

UV–vis–NIR spectra of (a) p-1aⁿ⁺, (b) p-1bⁿ⁺, and (c) m-1aⁿ⁺ (n = 0–2) obtained with an OTTLE cell ([p-1a/b] = ∼0.1 mM, [m-1a] = ∼1 mM, [NBu₄][PF₆] = 0.1 M in CH₂Cl₂, W.E. and C.E. Pt, and R.E. Ag/Ag⁺).

Table 4. UV–Vis–NIR Data for p-1ⁿ⁺ and m-1aⁿ⁺ (n = 0–2) Obtained with an OTTLE Cell^a.

compound	absorption maxima/nm (ε × 10–4/cm–1 M–1)
p-1a	356 (4.14)
p-1a+	530 (3.00), 1242 (2.12), 1542 (2.37),
p-1a2+	742 (4.33)
p-1b	348 (4.28)
p-1b+	532 (3.01), 1266 (2.23), 1564 (2.58)
p-1b2+	750 (4.83)
m-1a	298 (3.11)
m-1a+	796 (0.197), 1326 (0.127)
m-1a2+	756 (0.305), 1460 (0.208)

^a[p-1a/b] ∼ 1 mM, [m-1a] = ∼10 mM, [NBu₄][PF₆] = 0.1 M in CH₂Cl₂, W.E. and C.E. Pt, and R.E. Ag/Ag⁺.

IVCT bands of organometallic MV complexes in the NIR region are frequently observed as overlapped multiple bands.³⁰ Therefore, the deconvolution analysis of the bands is essential to obtain V_ab values (Figure 9). For p-1a⁺, the NIR bands were deconvoluted into the three Gaussian curves (bands A–C) as summarized in Table 5. For p-1b⁺, essentially the same result as p-1a⁺ was obtained. Similarly, the IVCT bands of m-1a⁺ were deconvoluted into the two bands (bands A and B) in a manner similar to the Cp* analogue 6⁺.^28,29 The bands for the para-isomers p-1⁺ (ε ∼ 17,000–20,000 M^–1 cm^–1) were more intense than those for the meta-isomer m-1a⁺ (ε ∼ 1000–5000 M^–1 cm^–1). For para-isomers p-1⁺, the band widths calculated according to the Hush’s theory [ν_calc = (2310 × ν_max)^1/2] were much larger than those of the bands A–C (Table 5), indicating stronger metal–metal interaction. The insensitivity of the NIR bands for para-isomers p-1a,b⁺ to the solvent polarity (acetone/CH₂Cl₂ = 98:2 vs CH₂Cl₂) (Figure 10) suggests the assignment of p-1⁺ as a class II/III border or class III species.³⁴ On the other hand, the value for the meta-isomer m-1a⁺ was comparable to the theoretical value, suggesting that m-1a⁺ is a class II species.

Figure 9.

NIR spectra of (a) p-1aⁿ⁺, (b) p-1bⁿ⁺, and (c) m-1aⁿ⁺ obtained with an OTTLE cell and their deconvoluted spectra ([p-1a/b] = ∼1 mM, [m-1a] = ∼10 mM, [NBu₄][PF₆] = 0.1 M in CH₂Cl₂, W.E. and C.E. Pt, and R.E. Ag/Ag⁺).

Table 5. NIR Spectral Data for p-1⁺ and m-1a⁺ and Related Data^a.

	deconvoluted IVCT bands
compound	band	νmax/cm–1	εmax/M–1 cm–1	ν1/2calcb/cm–1	ε/M–1 cm–1	Vab (classIII)c/Vab (classII)d/cm–1 (dRu–Ru)
p-1a+	A	6290	17050	1100	3812	3145/585
	B	7857	19521	1956	4260	3929/932
	C	9851	4878	2400	2355	4926/573
p-1b+	A	6200	21138	1100	3784	3100/646
	B	7850	21156	1871	4258	3925/949
	C	9851	5184	2400	2355	4926/591
m-1a+	A	5500	1457	3560	3564	288 (12.1 Å)e, 323 (10.8 Å)f
	B	7760	5015	3500	4234	623 (12.1 Å)e, 698 (10.8 Å)f

^a[p-1a/b] = ∼1 mM, [m-1a] = ∼10 mM, [NBu₄][PF₆] = 0.1 M in CH₂Cl₂, W.E. and C.E. Pt, and R.E. Ag/Ag⁺.

^bν_1/2^calc = (2310 × ν_max)^1/2.

^cV_ab(class III) = ν_max/2.

^dV_ab(class II) = 2.06 × 10^–2 × (ν_max ε_max ν_1/2)/d_Ru–Ru.

^eThrough bond distances.

^fThrough space distances.

Figure 10.

Normalized NIR spectra of (a) p-1a⁺ and (a) p-1b⁺ observed in CH₂Cl₂ (solid blue lines) and a mixed solvent of acetone/CH₂Cl₂ (98:2) (dashed red lines).

DFT and TD–DFT Calculations

In order to clarify the electronic structures of the MV species and to assign the NIR bands (Figures 11–13), DFT and TD-DFT calculations on 1ⁿ⁺ and their Cp derivatives, p-5ⁿ⁺ and m-5ⁿ⁺ ([(μ-p/m-C≡C–C₆H₄–C≡C){RuCp(dppe)}₂]ⁿ⁺) (n = 0, 1), were performed. The structural optimization of 1 was started with the antiparallel conformations observed by the X-ray structural analysis. The Kohn–Sham orbitals of the neutral species are summarized in Figure 11. Bond lengths and angles well coincide with those obtained by the X-ray analysis. The highest occupied molecular orbitals (HOMOs) of the para-isomers p-1a,b are delocalized over the Ru–C≡C–C₆H₄–C≡C–Ru frameworks, whereas the lowest unoccupied molecular orbitals (LUMOs) are localized on the dppe ligands. These results are in accord with the results for the related dinuclear bridging acetylide complexes reported previously.²⁰ Similar features are noted for the meta-isomers m-1a,b. The HOMOs for the para-isomers p-1 are higher in energy by about 0.3 eV compared to those of the meta-isomers m-1, and the result is consistent with the trend of their redox potentials. On the other hand, the energy levels of the LUMOs are less sensitive to the connectivity of the central ring because the dppe ligands mainly constitute them. The HOMO and LUMO features for p/m-1a are very similar to those for p/m-5. The HOMO and LUMO levels of p/m-5 are higher by about 0.2 and 0.4 eV, respectively, than those of the acyl derivatives 1, and the differences should derive from the electron-withdrawing acyl substituents.

Figure 11.

Frontier orbitals of p-1a/b, m-1a,b, and p/m-5 and their energies.

Figure 13.

β-HOSO and β-LUSO orbitals for 1⁺.

The spin density plots for the para-isomers of the monocationic species p-1⁺ reveal the full and symmetrical charge-delocalization over the Ru–C≡C–C₆H₄–C≡C–Ru linkages (Figure 12), suggesting the class III nature. On the other hand, the spin density of the meta-isomers m-1⁺ is substantially localized on one of the two ruthenium fragments, as is consistent with the charge-localized class II behavior. The spin density distribution is similar to those of p/m-5⁺, indicating again that the acyl substituents do not strongly influence the charge delocalization.

Figure 12.

Spin density plots of p-1⁺, m-1⁺, and p/m-5⁺. The α and β spin densities are represented in blue and red, respectively.

TD-DFT analysis of 1⁺ shows that the lowest energy bands arise from the transitions from β-HOSOs (the highest occupied spin orbitals) to β-LUSOs (the lowest unoccupied spin orbitals) (Figure 13 and Table 6). The computed energies well coincide with the experimental data, although the sub-bands (bands B and C) of para-isomers 1⁺ cannot be reproduced because different rotational isomers may be present.²⁰ For the para-isomers p-1a,b⁺, both of the β-HOSO and β-LUSO show the fully delocalized Ru–C≡C–C₆H₄–C≡C–Ru character, which is in accord with the assignment as a class III IVCT transition. On the other hand, the β-HOSOs and β-LUSOs of m-1a,b⁺ show clear charge-transfer characters from the dπ and pπ orbitals located on one of the Ru–C≡C–C₆H₄ fragments to those on the other metal fragment. These transition characters can be assignable to typical class II type IVCT bands. On the basis of the TD-DFT results and the assignment of the lowest energy band to the IVCT bands, the V_ab values of p-1a⁺, p-1b⁺, and m-1a⁺ estimated from the Marcus–Hush two state model are 3145 (p-1a⁺), 3100 (p-1b⁺) and 288 cm^–1 (m-1a⁺).³⁵⁻³⁷

Table 6. TD–DFT Data of NIR Bands for 1⁺.

complex	νmaxexp/cm–1	νmaxDFT/cm–1 (nm)	oscillator strength	main transition	assignment
p-1a+	6290	5435 (1840)	1.0929	β-HOSO → β-LUSO (∼100%)	IVCT
p-1b+	6200	5714 (1750)	1.035	β-HOSO → β-LUSO (∼100%)	IVCT
m-1a+	5500	4675 (2135)	0.0523	β-HOSO → β-LUSO (74%)	IVCT
	7760	5167 (1913)	0.0031	β-HOSO – 3 → β-LUSO (40%)	dπ–dπ transition (ligand field)
				β-HOSO – 4 → β-LUSO (25%)
m-1b+		5000 (2000)	0.0118	β-HOSO – 3 → β-LUSO (70%)	dπ–dπ transition (ligand field)
		6897 (1450)	0.0579	β-HOSO → β-LUSO (89%)	IVCT

ESR Study

ESR spectroscopy is another powerful tool to obtain environmental information of an unpaired electron on radical species such as the monocationic species, which was generated by in situ oxidation of the neutral species by the addition of [FeCp₂](PF₆) in CH₂Cl₂ and was subjected to the measurements as glasses after immediate freezing at 77 K (Figure 14). While intense and broadened isotropic or partially anisotropic signals with no hyperfine splitting were observed for the para-isomers p-1⁺,³⁸ very weak signals were found for m-1⁺. This suggests for p-1⁺ that the unpaired electrons are delocalized over the bridging ligands on the timescale of ESR spectroscopy (∼10^–8 s). The spectra are in contrast to that observed for [Cp*(dppe)Ru–C≡C–C₆H₅]⁺7⁺ with a rhombic g-tensor and an isotropic g value (g_iso = g₁ + g₂ + g₃) of 2.09.³⁹ The g values of p-1⁺ (∼2.06) are smaller than that of 7⁺ but much larger than those of organic free radicals (g = 2.0023). These factors further support the charge-delocalization over the Ru–C≡C–C₆H₄–C≡C–Ru moieties in p-1⁺.

Figure 14.

ESR spectra of p-1a⁺ and p-1b⁺ recorded as CH₂Cl₂ glasses at 77 K.

Effects of Substituents on the Cp Ligands

As expected, strong metal–metal interactions are observed for the para-isomers p-1ⁿ⁺ as judged by their K_C, V_ab, and k_et values, which are significantly larger than those of the meta-isomers m-1ⁿ⁺ with the different connectivities at the central benzene ring. Furthermore, the electrochemical and spectroscopic features of the core parts of the Ru complexes with/without the acyl functional groups are virtually the same. In fact, the theoretical DFT and TD-DFT studies support that the charge-delocalization properties are not perturbed by the Cp^R ligands so much when compared to the nonsubstituted Cp derivatives, although redox potential is sensitive to the electron-withdrawing substituents. Thus, these results support that Cp^R ligands (R = COOMe, COMe) can be used to evaluate the electronic structure of MV complexes without drastic electronic perturbation.

Charge Distribution Parameter Δρ

The charge distribution parameter, Δρ, which was first introduced by Geiger and Gleiter, provides a degree of charge delocalization between two IR-active centers in mixed valence species (MV) of a compound × [2·X(MV) ⇄ X⁺(ox) + X^– (red)].¹² Δρ is defined by eq 1

1

where Δν_ox = ν_ox – ν_ox^MV, Δν_red = ν_red – ν_red. In the present IR system, ν_ox and ν_red represent vibrational frequencies of the di-oxidized and neutral species, respectively, and ν_ox^MV and ν_red represent those of the higher-energy and lower-energy bands of the mono-oxidized species, respectively (Figure 15). When Δρ = 0 or 0.50, the complex is a localized class I species or a delocalized class III species, respectively. A compound with an intermediate Δρ value (0 < Δρ < 0.50) is categorized into class II. According to the equation, the p-isomers p-1a,b⁺ with Δρ = 0.50 turned out to be class III species, whereas the m-isomer m-1a⁺ with Δρ = 0.06 is characterized as a class II species and the charge is localized on one of the two metal fragments (Table 7). Due to the overlapped bands for m-1b⁺, its Δρ value cannot be determined accurately, but the value appears to be comparable to that of m-1a⁺. When the Δρ values are compared to the spin densities at the Ru atoms obtained by the DFT calculations (see above), they are well consistent with the charge distribution estimated by the Δρ values. Thus, the IR-tag serves as a practical indicator of charge delocalization over MV species.

Figure 15.

Schematic representation of IR stretching vibrations of a mixed valence system [2·X(MV) ⇄ X⁺(ox) + X^– (red)].

Table 7. Parameters Derived from IR, UV–Vis–NIR, and DFT Studies.

compound	Δνox/cm–1	Δνred/cm–1	νox – νred	Δρa	charge distribution on the two metal centers (%)	spin densities at the two metal centersb (%)	ΔG*/cm–1	ket/s–1
p-1a+	9	15	24	0.50	50:50	50:50		6 × 1012
p-1b+	16	18	34	0.50	50:50	50:50		6 × 1012
m-1a+	1	2	25	0.06	94:6	100:0	1100	<1.0 × 1010
m-1b+	1	0	36	c	c	100:0		<1.0 × 1010

^aρ = (Δν_ox + Δν_ox)/0.5(ν_ox – ν_red).

^bEstimated from the DFT calculations.

^cNot determined due to overlapped bands.

Comparison of k_et and V_ab Values

Electron transfer rates of MV complexes are sensitive to the environments. As the rate of solvent reorientation timescale is considered to be in the range from 10^–11 to 10^–12 s, IVCT bands of class II species should be affected by the polarity of the solvent. Therefore, k_et of class II and class III species falls below (<10^–11 s) and above (>10^–12 s) this range, respectively. As discussed above, the NIR bands of the para-isomers p-1⁺ showed little dependence on the solvents, and their k_et values turned out to be ∼6 × 10¹² s^–1, supporting their class III nature.

On the other hand, the k_et values of the meta-isomers m-1⁺ are less than 10¹⁰ s^–1 in accord with the assignment as class II classification. For a weakly coupled class II system, the k_et value can also be evaluated by using the following equations in case the electronic coupling V_ab is small and the thermal ET process passes through the diabatic regime.⁴⁰

2

3

where ΔG* is the free energy of activation needed for the thermally induced ET transfer, ν_max is the energy of the IVCT transition band at the maximum extinction coefficient, h is Planck’s constant, k_B is Boltzmann’s constant, and T is the temperature. On the basis of eqs 2 and 3, ΔG* and k_et for m-1a⁺ are determined to be 1100 cm^–1 and 1.26 × 10¹¹ s^–1, respectively. The k_et value is overestimated by an order of magnitude compared to that obtained by the IR methods (<1.0 × 10¹⁰ s^–1). The deviation may derive from (i) the erroneous estimation of V_ab values, which are determined by the deconvolution analysis of overlapped bands, (ii) the ground-state energy curves deviating from the ideal diabatic states, or (iii) the presence of rotamers as discussed above.

Scope and Limitation of the IR-Tag Strategy on the Cp Ligand

As we have discussed above, the application of the IR-tags strategy toward organometallic MV complexes with Cp ligands is useful, especially, for the estimation of the local charges on the metal fragments and of the k_et values obtained by the digital simulations. On the other hand, the following limitation was noted in this study. The k_et values can be determined only when the values are within the range that the simulation allows (10^–10 to 10^–12 s). Furthermore, in some case, the C=O vibrational bands can be overlapped with the other bands to hamper the analysis. Nevertheless, the present IR-tag method is promising because it can be facilely applied to MV complexes with various types of metal species, Cp ligands, and supporting ligands.

Conclusions

We have synthesized mono- and di-ruthenium alkynyl complexes with cyclopentadienyl ligands bearing the IR-active acyl groups and examined their electronic structures by electrochemical, spectroelectrochemical UV–vis–NIR and IR, ESR, and theoretical methods. It is clear that the para-isomers p-1 show metal–metal interactions stronger than the meta-isomers m-1. In contrast to the determination process of the V_ab value involving ambiguous factors, k_et can be unambiguously determined by the line shape analysis of IR spectra of the acylated Cp ligands attached to the metal centers and thus helps us fully understand the whole pictures of the electronic structures of the metal parts in MV states. The ν(C=O) frequencies of the acyl groups are sensitive to the electron densities at the metal centers and, thus, afford information on the extent of charge-delocalization. As described herein, we have successfully extended the IR-tag technique toward metal alkynyl MV complexes with half-sandwich cyclopentadienyl structural motifs, which have been extensively studied for these 3 decades. This study should provide new insights into the electronic structures of metal alkynyl MV complexes.

Experimental Section

Instruments

NMR spectra were recorded on Bruker biospin AVANCE III 400 MHz (¹H 400 MHz for 2D NMR measurements) and Bruker biospin ASCEND-500 spectrometers (¹H 500 MHz, ³¹P 202 MHz ¹³C NMR 126 MHz) in CDCl₃. NMR chemical shifts were referenced to the residual protio impurities in the deuterated solvents. ESI-TOF-MS analysis was performed on a Bruker micrOTOF II spectrometer. UV–vis and IR spectra (KBr pellets and CH₂Cl₂ solution) were recorded on JASCO V670DS and FTIR 4200 spectrometers, respectively. Electrochemical measurements (CV and DPV) were made with a Hokuto Denko HZ-5000 [observed in CH₂Cl₂; [complex] = ca. 1 × 10^–3 M; [NBu₄PF₆] = 0.1 M; working electrode: Pt, counter electrode: Pt, reference electrode: Ag/AgNO₃; and scan rate was 100 mV/s (CV)]. After the measurement, ferrocene (Fc) was added to the mixture and the potentials were calibrated with respect to the Fc/Fc⁺ redox couple. ESR spectra for the monocationic species were recorded on a JEOL JES-FA100 spectrometer, and samples were prepared by the addition of small amounts of oxidant [tris(4-bromophenyl)aminium hexachloroantimonate] to CH₂Cl₂ solutions of the complexes in ESR tubes under a N₂ atmosphere just before cooling in a liquid nitrogen bath. The spectra were recorded at the liquid nitrogen temperature. X-ray crystal analysis was performed on a Bruker Smart Apex II Ultra diffractometer. CCDC numbers 1861281–1861286 contain the supplementary crystallographic data for p-1a (1861281), p-1b (1861282), m-1a (1861283), m-1b (1861284), 2a (1861285), and 2b (1861286).

Spectroelectrochemistry

All measurements were conducted under a N₂ atmosphere. Spectroelectrochemical UV–vis–NIR measurements were performed in a home-built 1 mm quartz cell equipped with working (Pt mesh), counter (Pt wire), and reference electrodes (Ag/AgNO₃). A CH₂Cl₂ solution of complex (0.1 mM for p-1a,b, 1 mM for m-1a,b) was subjected to the measurements in the presence of NBu₄PF₆ (0.1 M), a supporting electrolyte. The spectroelectrochemical IR measurements were performed with an OTTLE cell, purchased from the University of Reading, which comprised a Pt-mesh working and counter electrodes and a thin silver wire as a pseudo-reference electrode. The optical beam was passed through CaF₂ windows of a conventional liquid IR cell. The conditions were the same as those for the spectroelectrochemical UV–vis–NIR measurements except the concentration of the complexes (10 mM for 1).

Materials

Reactions were performed under a N₂ atmosphere using the standard Schlenk tube technique unless stated otherwise. Purchased dry THF, CH₂Cl₂, and hexane were further purified by the Grubbs solvent purification system.⁴¹ Dry acetone purchased from Kanto Chemical Co., Inc. was used without further purification. Triethylamine was predried over KOH pellets and distilled from CaH₂. DBU was distilled from CaH₂. Other regents, silica gel (Kanto chemical Co Inc. Silica Gel 60N), and alumina (Merck Aluminum oxide 90 standardized) were used as received. RuCl₂(PPh₃)₃,⁴² sodium methoxycarbonylcyclopentadienide,²⁴ and sodium acetylcyclopentadienide²⁴ were synthesized according to the literature procedures.

Synthesis of 3a

A THF (50 mL) solution of RuCl₂(PPh₃)₃ (3.00 g, 3.13 mmol) and sodium methoxycarbonylcyclopentadienide (487 mg, 3.33 mmol) was stirred for 1 h at room temperature. The mixture was concentrated and hexane (80 mL) was added to form red precipitates, which were collected by filtration, washed with hexane, and dried under reduced pressure to afford crude products of 3a. To the obtained solid was added dppe (1.247 g, 3.13 mmol) and toluene (50 mL), and the mixture was refluxed for 4 h. After the mixture was cooled to room temperature, the volatiles were evaporated, and the resulting solid was subjected to silica gel column chromatography (CH₂Cl₂/Et₂O = 9:1) to yield 3a (824 mg, 1.25 mmol, 40%) as a yellow powder. ¹H NMR: δ 7.83–7.76 (m, 4H, Ph in dppe), 7.40–7.36 (m, 6H, Ph), 7.33–7.26 (m, 6H, Ph), 7.12 (virtual t, J = 8.7 Hz, 4H, Ph), 5.24 (t, J = 2.0 Hz, 2H, η⁵-C₅H₄-R), 4,14 (t, J = 2.0 Hz, 2H, η⁵-C₅H₄-R), 3.60 (s, 3H, COOCH₃), 2.78–2.61 (m, 2H, CH₂), 252–2.35. (m, 2H, CH₂). ³¹P NMR: δ 79.1 (s). ¹³C{¹H} NMR: δ 167.8 (s, COOMe) 140.5 (virtual dt, J_C–P = 43.6, 12.9 Hz, ipso-Ph in dppe), 134.2 (t, J_C–P = 4.9 Hz, o-Ph in dppe), 134.5–133.8 (m, ipso-Ph in dppe), 131.7 (t, J_C–P = 4.8 Hz, o-Ph in dppe), 130.0 (s, p-Ph in dppe), 129.5 (s, p-Ph in dppe-Ph), 128.3 (t, J_C–P = 4.4 Hz, m-Ph in dppe), 128.1 (t, J_C–P = 4.7 Hz, m-Ph in dppe), 89.1, 79.7, 76.7 (s × 3, η⁵-C₅H₄-R), 51.5 (s, COOMe), 27.5 (m, CH₂ in dppe). HRESI-TOF MS (m/z): calcd for C₃₃H₃₁P₂O₂Ru, 623.0846. Found: 623.0845 [M–Cl]⁺. IR(KBr/cm^–1): 1705 ν(C=O). Anal. Calcd for C_32.1H_31.2Cl_1.2O₂P₂Ru (3a· (CH₂Cl₂)_0.1): C, 58.90; H, 4.80. Found: C, 59.02; H, 4.28.

Synthesis of 3b

Synthesis of 3b was previously reported,²⁴ but we prepared 3b in a manner analogous to the synthesis of 3a using RuCl₂(PPh₃)₃ (3.00 g, 3.13 mmol), sodium acetylcyclopentadienide (429 mg, 3.30 mmol), and dppe(1.247 g, 3.13 mmol). 3b was obtained as a yellow powder (987 mg, 1.54 mmol, 49%). The spectral data were consistent with the value reported previously. ¹H NMR: δ 7.74–7.68 (m, 4H, o-Ph in dppe), 7.42–7.38 (m, 6H, Ph), 7.34 (t, J = 7.3 Hz, 2H, Ph), 7.30 (t, J = 7.3 Hz, 4H, Ph), 7.12 (virtual t, J = 8.0 Hz, 4H, o-Ph in dppe), 5.30 (m, 2H, η⁵-C₅H₄-R), 4.18–4.10 (m, 2H, η⁵-C₅H₄-R), 2.78–2.60 (m, 2H, CH₂ in dppe), 2.42–2.23 (m, 2H, CH₂ in dppe), 1.95 (s, 3H, COCH₃). ³¹P NMR: δ 78.2 (s).¹³C{¹H} NMR: δ 196.5 (s, COMe), 139.5 (virtual dt, J_C–P = 43.8, 12.6 Hz, ipso-Ph in dppe), 133.6 (t, J_C–P = 5.0 Hz, o-Ph in dppe), 131.8 (t, J_C–P = 5.0 Hz, o-Ph in dppe), 129.9 (s, p-Ph in dppe), 129.7 (s, p-Ph in dppe), 128.2 (t, J_C–P = 4.8 Hz, m-Ph in dppe), 127.9 (t, J_C–P = 5.0 Hz, m-Ph in dppe), 88.4, 88.0, 76.7 (s × 3, η⁵-C₅H₄-R), 28.2 (s, COMe), 26.7 (m, CH₂ in dppe). IR(KBr/cm^–1): 1653 ν(C=O).

General Procedure for the Synthesis of Alkynylruthenium Complexes

To a mixture of Cp^RRuCl(dppe) (1 equiv, 0.15–0.5 mmol scale) and AgBF₄ (1 equiv) was added CH₂Cl₂ (5–10 mL), and the resultant mixture was stirred at room temperature for 30 min. The suspension was filtered through a Celite pad to remove precipitated AgCl and the filtrate was added to a solution of the appropriate ethynylbenzene (1 and 0.5 equiv for mono- and di-nuclear complexes, respectively) in CH₂Cl₂ (5–10 mL). After the mixture was stirred for 1 h, several drops of DBU (<∼50 μL) was added to the reaction mixture, which was further stirred for 1 h. The solution was evaporated and the obtained residue was subjected to silica gel column chromatography (CH₂Cl₂) to afford the product.

Synthesis of 2a

The title complex was obtained from 3a (100 mg, 0.152 mmol) and 4-ethynyltoluene (17 mg, 0.15 mmol) as a yellow powder (83 mg, 0.11 mmol, 74%). ¹H NMR: δ 7.89–7.83 (m, 4H, o-Ph in dppe), 7.42–7.37 (m, 6H, Ph), 7.34–7.21 (m, 10H, Ph), 6.74 (d, 2H, J = 7.9 Hz, C₆H₅CH₃), 6.39 (d, 2H, J = 7.9 Hz, C₆H₅CH₃), 5.49 (t, J = 2.2 Hz, 2H, η⁵-C₅H₄-R), 4.60 (t, J = 2.2 Hz, 2H, η⁵-C₅H₄-R), 3.48 (s, 3H, C(O)OCH₃), 2.80–2.63 (m, 2H, CH₂ in dppe), 2.40–2.23 (virtual octet, J = 6.8 Hz, 2H, CH₂ in dppe), 2.18 (s, 3H, C₆H₅CH₃). ³¹P NMR: δ 85.0 (s). ¹³C{¹H} NMR: δ 167.5 (s, COMe), 141.6 (virtual dt, J_C–P = 39.2, 9.2 Hz, ipso-Ph in dppe), 136.2 (virtual td, J_C–P = 27.4, 4.6 Hz, ipso-Ph in dppe), 134.2 (t, J = 5.1 Hz, o-Ph in dppe), 132.9 (s, C₆H₅CH₃), 131.8 (t, J_C–P = 5.1 Hz, o-Ph in dppe), 130.4 (s, C₆H₅CH₃), 129.6 (s, p-Ph in dppe), 129.2 (s, p-Ph in dppe), 128.2 (s, C₆H₅CH₃), 128.1 (t, J = 4.7 Hz, m-Ph in dppe), 127.7 (t, J = 5.0 Hz, m-Ph in dppe), 126.9 (s, C₆H₅CH₃), 111.7 (s, Ru–C≡C), 111.0 (t, J = 25.7 Hz, Ru–C≡C), 89.3, 83.0, 82.4 (s × 3, η⁵-C₅H₄-R), 51.2 (s, COOMe), 28.4 (m, CH₂ in dppe), 21.2 (s, C₆H₅CH₃). HRESI-TOF MS (m/z): calcd for C₄₂H₃₈P₂O₂Ru, 738.1397. Found: 738.1397 [M]⁺. IR(CH₂Cl₂/cm^–1): 2085 ν(C≡C), 1703 ν(C=O). Anal. Calcd for C₄₂H₃₈O₂P₂Ru: C, 68.38; H, 5.19. Found: C, 68.38; H, 5.24.

Synthesis of 2b

The title complex was obtained from 3b (100 mg, 0.156 mmol) and 4-ethynyltoluene (17 mg, 0.15 mmol) as a yellow powder (63 mg, 0.087 mmol, 58%). ¹H NMR: δ 7.77–7.71 (m, 4H, o-Ph in dppe), 7.42–7.38 (m, 6H, Ph), 7.33 (t, J = 7.7 Hz, 2H, p-Ph in dppe), 7.28 (t, J = 7.0 Hz, 4H, Ph), 7.23 (m, 4H, Ph), 6.80 (d, J = 7.8 Hz, 2H, C₆H₅CH₃), 6.47 (d, J = 7.8 Hz, 2H, C₆H₅CH₃), 5.54 (t, J = 2.1 Hz, 2H, η⁵-C₅H₄-R), 4.56–4.52 (m, 2H, η⁵-C₅H₄-R), 2.80–2.66 (m, 2H, CH₂ in dppe), 2.30–2.13 (m, 2H, CH₂ in dppe), 2.21 (s, 3H, C₆H₅CH₃), 1.95 (s, 3H, COCH₃). ³¹P NMR: δ 83.9 (s).¹³C{¹H} NMR: δ 196.1 (s, COMe), 140.6 (virtual dt, J_C–P = 34.4, 11.1 Hz, ipso-Ph in dppe), 135.7 (virtual td, J_C–P = 26.2, 5.4 Hz, ipso-Ph in dppe), 133.5 (t, J_C–P = 4.9 Hz, o-Ph in dppe), 133.3 (s, C₆H₅CH₃), 132.0 (t, J_C–P = 5.3 Hz, o-Ph in dppe), 130.3 (s, C₆H₅CH₃), 129.4 (s, p-Ph in dppe), 129.3 (s, p-Ph in dppe), 128.2 (s, C₆H₅CH₃), 128.0 (t, J_C–P = 4.7 Hz, m-Ph in dppe), 127.5 (t, J_C–P = 5.0 Hz, m-Ph in dppe), 126.7 (s, C₆H₅CH₃), 112.7 (t, J = 24.9 Hz, Ru–C≡C), 111.5 (s, Ru–C≡C), 90.3, 89.2, 82.6 (s × 3, η⁵-C₅H₄-R), 27.8 (m, CH₂ in dppe), 27.4 (s, COMe), 21.1 (s, C₆H₅CH₃). HRESI-TOF MS (m/z): calcd for C₄₂H₃₈P₂ORu, 722.1447 [M]⁺. Found: 722.1447 [M]⁺. IR(CH₂Cl₂/cm^–1): 2085 ν(C≡C), 1650 ν(C=O). Anal. Calcd for C_42.3H_38.6Cl_0.6OP₂Ru (2b· (CH₂Cl₂)_0.3): C, 67.99; H, 5.21. Found: C, 67.77; H, 5.23.

Synthesis of p-1a

The title complex was obtained from 3a (320 mg, 0.486 mmol) and 1,4-diethynylbenzene (30 mg, 0.24 mmol) as a yellow powder (245 mg, 0.179 mmol, 75%). ¹H NMR: δ 7.79 (virtual t, J = 8.5 Hz, m, 8H, o-Ph in dppe), 7.37–7.20 (m, 32H, Ph), 6.11 (s, 4H, C₆H₄), 5.44 (t, J = 2.1 Hz, 4H, η⁵-C₅H₄-R), 4.57 (t, J = 2.1 Hz, 4H, η⁵-C₅H₄-R), 3.43 (s, 3H, C(O)OCH₃), 2.75–2.60 (m, 4H, CH₂ in dppe), 2.35–2.18 (m, 4H, CH₂ in dppe), ³¹P NMR: δ 84.9 (s). ¹³C{¹H} NMR: δ 167.4 (s, C(O)OMe), 141.1 (virtual dt, J_C–P = 39.0, 9.6 Hz, ipso-Ph in dppe), 136.1 (virtual td, J_C–P = 28.0, 4.5 Hz, ipso-Ph in dppe), 134.0 (m, o-Ph in dppe), 131.6 (m, o-Ph in dppe), 129.3 (s, p-Ph in dppe), 129.0 (s, p-Ph in dppe), 127.9 (m, m-Ph in dppe), 127.5 (m, m-Ph in dppe), 112.7 (s, Ru–C≡C), 110.7 (t, J = 26.5 Hz, Ru–C≡C), 89.2 (s, η⁵-C₅H₄-R), 83.0 (s, η⁵-C₅H₄-R), 51.2 (s, C(O)OMe) 28.4 (m, CH₂ in dppe). HRESI-TOF MS (m/z): calcd for C₇₆H₆₆P₄O₄Ru₂, 1370.1998 [M]⁺. Found 1370.1995 [M]⁺. IR(CH₂Cl₂/cm^–1): 2085 ν(C≡C), 1703 ν(C=O). Anal. Calcd for C_76.3H_66.6Cl_0.6O₄P₄Ru₂ (p-1a· (CH₂Cl₂)_0.3): C, 65.70; H, 4.81. Found: C, 65.54; H, 4.70.

Synthesis of p-1b

The title complex was obtained from 3b (130 mg, 0.202 mmol) and 1,4-diethynylbenzene (13 mg, 0.10 mmol) as a yellow powder (61 mg, 0.045 mmol, 45%). ¹H NMR: δ 7.71 (virtual t, J = 7.8 Hz, 8H, o-Ph in dppe), 7.38–7.25 (m, 24H, Ph), 7.20 (virtual t, J = 8.9 Hz, 8H, Ph), 6.25 (s, 4H, C₆H₄), 5.51 (t, J = 2.0 Hz, 4H, η⁵-C₅H₄-R), 4.52 (s, 4H, η⁵-C₅H₄-R), 2.79–2.60 (m, 4H, CH₂ in dppe), 2.28–2.11 (m, 4H, CH₂ in dppe), 1.91 (s, 6H, COCH₃) ³¹P NMR: δ 83.9 (s). ¹³C{¹H} NMR: δ 196.0 (s, COMe), 140.6 (virtual dt, J_C–P = 39.6, 18.6 Hz, ipso-Ph in dppe), 135.7 (virtual dt, J_C–P = 52.2, 25.6 Hz, ipso-Ph in dppe), 133.5 (m, o-Ph in dppe), 132.0 (m, o-Ph in dppe), 129.6–129.3 (m, p-Ph in dppe and C₆H₄), 128.0 (m, p-Ph in dppe), 127.5 (m, m-Ph in dppe), 124.9 (s, C₆H₄), 113.2 (t, J = 24.3 Hz, Ru–C≡C), 112.4 (s, C≡C), 90.3, 89.2, 82.6 (s × 3, η⁵-C₅H₄-R), 27.8 (m, CH₂ in dppe), 27.4 (s, COMe) HRESI-TOF MS (m/z): calcd for C₇₆H₆₆P₄O₂Ru₂, 1338.2099 [M]⁺. Found: 1338.2093 [M]⁺. IR(CH₂Cl₂/cm^–1): 2081 ν(C≡C), 1650 ν(C=O). Anal. Calcd for C₇₇H₆₈Cl₂O₂P₄Ru₂ (p-1b· CH₂Cl₂): C, 65.02; H, 4.82. Found: C, 65.67; H, 4.70.

Synthesis of m-1a

The title complex was obtained from 3a (160 mg, 0.243 mmol) and 1,3-diethynylbenzene (14 mg, 0.11 mmol) as a yellow powder (121 mg, 0.088 mmol, 80%). ¹H NMR: δ 7.80 (virtual t, J = 8.0 Hz, 8H, o-Ph in dppe), 7.32–7.17 (m, 32H, Ph), 6.54 (t, J = 7.7 Hz, 1H, C₆H₄), 6.20 (t, J = 1.6 Hz, 1H, C₆H₄), 6.10 (dd, J₁ = 7.7 Hz, J₂ = 1.6 Hz, 2H, C₆H₄), 5.48 (t, J = 2.2 Hz, 4H, η⁵-C₅H₄-R), 4.58 (t, J = 2.2 Hz, 4H, η⁵-C₅H₄-R), 3.41 (s, 6H, COOMe), 2.80–2.66 (virtual octet, J = 5.9 Hz, 4H, CH₂ in dppe), 2.39–2.24 (virtual octet, J = 5.3 Hz, 4H, CH₂ in dppe). ³¹P NMR: δ 84.9 (s). ¹³C{¹H} NMR: δ 167.4 (s, COOMe), 141.5 (virtual td, J_C–P = 26.9, 8.6 Hz, ipso-Ph in dppe), 136.4 (virtual td, J_C–P = 28.7, 4.3 Hz,, ipso-Ph in dppe), 134.0 (t, J_C–P = 4.9 Hz, o-Ph in dppe), 132.0 (s, C₆H₄), 131.9 (t, J_C–P = 5.2 Hz, o-C₆H₅ in dppe), 129.5 (s, p-C₆H₅ in dppe), 129.1 (s, p-Ph in dppe), 128.9 (s, C₆H₄), 128.1 (t, J_C–P = 4.5 Hz, m-Ph in dppe), 127.6 (t, J_C–P = 5.0 Hz, m-Ph in dppe), 126.5 (s, C₆H₄), 126.4 (s, C₆H₄), 112.4 (s, C≡C), 109.7 (t, J_C–P = 25.6 Hz, Ru–C≡C), 89.3, 83.1, 82.6 (s × 3, η⁵-C₅H₄-R), 51.2 (s, COOMe), 28.5 (m, CH₂-dppe). HRESI-TOF MS (m/z): calcd for C₇₆H₆₆P₄O₄Ru₂, 1370.1998 [M]⁺. Found: 1338.2003 [M]⁺. IR(CH₂Cl₂/cm^–1): 2075 ν(C≡C), 1703 ν(C=O). Anal. Calcd for C₇₆H₆₆O₄P₄Ru₂·0.5(CH₂Cl₂): C, 65.08; H, 4.78. Found: C, 64.72; H, 4.54.

Synthesis of m-1b

The title complex was obtained from 3b (130 mg, 0.202 mmol) and 1,3-diethynylbenzene (13 mg, 0.10 mmol) as a yellow powder (66 mg, 0.048 mmol, 48%). ¹H NMR: δ 7.73–7.67 (m, 8H, o-Ph in dppe), 7.35–7.24 (m, 24H, Ph), 7.19 (virtual t, J = 8.1 Hz, 8H, o-Ph in dppe), 6.62 (t, J = 7.7 Hz, 1H, C₆H₄), 6.32 (t, J = 1.6 Hz, 1H, C₆H₄), 6.16 (dd, J = 7.7 Hz, J₂ = 1.6 Hz, 2H, C₆H₄), 5.53 (t, J = 2.1 Hz, 4H, η⁵-C₅H₄-R), 4.54–4.52 (m, 4H, η⁵-C₅H₄-R), 2.82–2.65 (virtual octet, J = 6.3 Hz, 4H, CH₂ in dppe), 2.32–2.15 (virtual octet, J = 6.9 Hz, 4H, CH₂ in dppe), 1.91 (s, 6H, COCH₃). ³¹P NMR: δ 84.1(s). ¹³C{¹H} NMR: δ 195.9 (s, −COMe), 140.7 (virtual dt, J_C–P = 39.7, 18.1 Hz, ipso-Ph in dppe), 135.8 (virtual dt, J_C–P = 27.9, 5.9 Hz, ipso-Ph in dppe), 133.6 (t, J_C–P = 4.8 Hz, o-Ph in dppe), 132.0 (t, J_C–P = 5.2 Hz, o-Ph in dppe), 131.7 (s, C₆H₄), 129.5 (s, p-Ph in dppe), 129.4 (s, p-Ph in dppe), 128.9 (s, C₆H₄), 128.1 (t, J_C–P = 4.5 Hz, m-Ph in dppe), 127.6 (t, J_C–P = 5.0 Hz, m-Ph in dppe), 126.8 (s, C₆H₄), 126.7 (s, C₆H₄), 112.5 (t, J = 24.6 Hz, Ru–C≡C), 112.3 (s, C≡C), 90.4, 89.4, 82.9 (s × 3, η⁵-C₅H₄-R), 28.0 (m, CH₂-dppe), 27.6 (s, COMe). HRESI-TOF MS (m/z): calcd for. C₇₆H₆₆P₄O₂Ru₂, 1338.2099 [M]⁺. Found: 1338.2092 [M]⁺. IR(CH₂Cl₂/cm^–1): 2074 ν(C≡C), 1650 ν(C=O). Anal. Calcd for C_76.1H_66.2Cl_0.2O₂P₄Ru₂ (m-1b· (CH₂Cl₂)_0.1): C, 67.91; H, 4.96. Found: C, 67.56; H, 4.75.

DFT Calculations

DFT calculations were performed by using the Gaussian 09 (D01) and 16 (A01) program packages.⁴³ Optimizations of geometry for the neutral and monocationic complexes were carried out with the (U)B3LYP/LanL2DZ levels of theory with CPCM (CH₂Cl₂) solvent continuum. Single-point calculations of neutral complexes were performed at the same level of theory. TD-DFT calculations and spin density plots were conducted with UBLYP35/Def2SVP with the CPCM solvent continuum of CH₂Cl₂ levels of theory.⁴⁴ We also tried geometry optimization using the UBLYP35/Def2SVP level of theory, but resulted in the charge-localized geometry for the para-isomers, which were not consistent with our experimental results.

Glossary

Abbreviations

dppe

1,2-bis(diphenylphosphino)ethane

Cp

cyclopentadinyl

Cp*

1,2,3,4,5-pentamethylcyclopentadienyl

W.E

working electrode,

R.E.

reference electrode,

C.E.

counter electrode

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsorginorgau.2c00005.

• Spectral, electrochemical, and X-ray diffraction data and Cartesian coordinates of optimized geometries (PDF)

Author Contributions

Y.T. and H.T. contributed equally to this work. All the experiments were conducted by H.T. Y.T. contributed to the theoretical study. The manuscript was drafted by Y.T. and edited by M.A. All authors have given approval to the final version of the manuscript.

JSPS KAKENHI grant number 21K05211 ENEOS Tonengeneral Research/Development Encouragement & Scholarship Foundation. The Asahi Glass Foundation. Tokuyama Science Foundation.

The authors declare no competing financial interest.