Magnetic and Electrochemical Properties of Lantern-Type Dinuclear Ru(II,III) Complexes with Axial Chloride Ions or Water Molecules

: By using [Ru 2 (O 2 CC 3 H 7 ) 4 Cl] n ( 1 ) as a starting material, n Bu 4 N[Ru 2 (O 2 CC 3 H 7 ) 4 Cl 2 ] ( n Bu 4 N + = tetra( n -butyl)ammonium cation) ( 2 ) and [Ru 2 (O 2 CC 3 H 7 ) 4 (H 2 O) 2 ]BF 4 ( 3 ) were prepared. The lantern-type dinuclear structures with axial chloride ions or water molecules were conﬁrmed for 2 and 3 by X-ray crystal structure analyses. The crystal structures of 2 and 3 were compared with that of 1 . In the crystal of 2 , there were three crystallographically different dinuclear units; the Ru–Ru distances of each unit were 2.3094(3), 2.3046(4), and 2.3034(4) Å, respectively, which were longer than those of 1 (2.281(4) Å) and 3 (2.2584 (7) Å). Temperature dependent magnetic susceptibility measurements were performed for 1 and 2 as well as 3 . The effective magnetic moments ( µ eff ) at 300 K were 3.97 (for 1 ), 4.00 (for 2 ), and 3.97 µ B (for 3 ), respectively. The decreases in the µ eff value were conﬁrmed for all of the complexes due to the large zero-ﬁeld splitting ( D ): D = 68 cm − 1 for 1 , 78 cm − 1 for 2 , and 60 cm − 1 for 3 . Cyclic voltammograms measured in CH 2 Cl 2 with a electrolyte of n Bu 4 N(BF 4 ) showed the Ru 25+ /Ru 24+ process at − 0.2– − 0.4 V (vs. SCE) and the Ru 26+ /Ru 25+ one at 1.3–1.4 V (vs. SCE), of which potentials were conﬁrmed by the DFT calculation for n Bu 4 N[Ru 2 (O 2 CC 3 H 7 ) 4 Cl 2 ]. 2 O were conﬁrmed by X-ray crystal structure analyses. Temperature dependent magnetic susceptibility measurements were performed to show that all of the complexes ([Ru 2 (O 2 CC 3 H 7 ) 4 Cl] n , n Bu 4 N[Ru 2 (O 2 CC 3 H 7 ) 4 Cl 2 ], and [Ru 2 (O 2 CC 3 H 7 ) 4 (H 2 O) 2 ]BF 4 ) had an S = 3/2 ground state, with a large zero-ﬁeld splitting ( D = 60–80 cm − 1 ). No important magnetic interaction was observed between the dinuclear units for the complexes. Cyclic voltammograms (measured in CH 2 Cl 2 with an electrolyte of n Bu 4 N(BF 4 )) showed the Ru 25+ /Ru 24+ process at − 0.2– − 0.4 V (vs. SCE) and the Ru 26+ /Ru 25+ one at 1.3–1.4 V (vs. SCE), where the potentials were conﬁrmed by the DFT calculation for [Ru 2 (O 2 CC 3 H 7 ) 4 Cl 2 ] − .


Synthesis and Characterizations
The addition of an excess amount of n Bu4NCl to the dinuclear units of [Ru2(O2CC3H7)4Cl]n (1) with stirring at room temperature for 24 h in dichloromethane solution gave a dichloridodiruthenium(II,III) complex n Bu4N[Ru2(O2CC3H7)4Cl2] (2) with a yield of 85% (based on Ru2 5+ unit of 1). On the other hand, the axial chloride ligand of 1 could be removed by the reaction with AgBF4 in THF at room temperature for 24 h with stirring to give a diaquadiruthenium(II,III) complex [Ru2(O2CC3H7)4(H2O)2]BF4 (3) with a yield of 69%. Their chemical formula were confirmed by elemental analyses. The IR spectra of 2 and 3 showed the COO vibrations as a set of two distinctive bands (νasym (COO) 1465 cm −1 and νsym (COO) 1426 cm −1 for 2; νasym (COO) 1455 cm −1 and νsym (COO) 1428 cm −1 for 3) in a similar energy region to those of 1 (νasym (COO) 1462 cm −1 and νsym (COO) 1425 cm −1 ). These facts suggest that the dinuclear skeleton is preserved in the above-mentioned reactions to 2 and 3. The stretching vibrations of BF4 − appear as a broad band around 1090 cm −1 in 3, which indicate no coordination of the ion to the Ru2 5+ core [9]. 10 with the support of EPR (electron paramagnetic resonance) spectral data [7]. In the study, they did not mention the interaction through the axial Cl − ions between the paramagnetic dinuclear units (S = 3/2), although we have lately described the importance of the axial bond through interaction to understand the magnetic behaviors of the polymer complexes of Ru 2 5+ units linked by axial ligands [5]. An

Synthesis and Characterizations
The addition of an excess amount of n which indicate no coordination of the ion to the Ru 2 5+ core [9].

Magnetic Properties
The temperature dependences of magnetic susceptibilities and effective magnetic moments of 1, 2, and 3 are shown in Figures 5-7, respectively. The magnetic moments at 300 K were 3.97 (for 1), 4.00 (for 2), and 3.97 µB (for 3), respectively, indicating the existence of three unpaired electrons within the

Magnetic Properties
The temperature dependences of magnetic susceptibilities and effective magnetic moments of 1, 2, and 3 are shown in Figures 5-7, respectively. The magnetic moments at 300 K were 3.97 (for 1), 4.00 (for 2), and 3.97 µ B (for 3), respectively, indicating the existence of three unpaired electrons within the dinuclear Ru 2 5+ cores like the other lantern-type tetrakis(carboxylato)diruthenium(II,III) with a formal electron configuration of σ 2 π 4 δ 2 (δ*π*) 3 (i.e., S = 3/2 ground state) [1]. All of the complexes showed decreases in the moments by lowering the temperature due to the strong zero-field splitting (D). The magnetic behaviors were simulated using the Equations (1)-(3) [3,5,7,14]: where χ is the magnetic susceptibility and χ // and χ ⊥ are magnetic susceptibility terms defined as follows: Figure 5. Temperature dependences of magnetic susceptibility χ M (circles) and moment µ eff (triangles) for [Ru 2 (O 2 CC 3 C 7 ) 4 Cl] n (1). The red and blue solid lines were calculated and drawn with the parameter values described in the text. (2). The red and blue solid lines were calculated and drawn with the parameter values described in the text. The simulation results provided the following parameter values: g = 2.06, D = 68 cm −1 for 1, g = 2.08, D = 78 cm −1 for 2, and g = 2.05, D = 60 cm −1 for 3. Large D values are common for the lantern-type Ru 2 5+ complexes [1][2][3][4][5] and are in the range of D = 50-100 cm −1 . Although the magnetic interactions between dinuclear units can be estimated using zJ, which means the exchange energy multiplied by the number of interacting neighboring units and is defined by χ' = χ/{1 − (2zJ/Ng 2 µ B 2 )χ}, when a molecular field approximation is applied [3,5,7,14], the temperature-dependent profiles of magnetic susceptibilities and moments of 1-3 could all be reproduced well without the zJ term. That is, the magnetic interactions were negligible for the complexes (zJ = 0 cm −1 ). This is reasonable for 2 and 3 because the X-ray crystal structural data showed that the Ru 2 5+ units were distant from each other without any axial linkers mediating the interaction. As for 1, the antiferromagnetic interaction could be possible through an axial chloride linker ligand. The negligible interaction may be due to the zig-zag chain structure with a smaller Ru-Cl ax -Ru bond angle (125.4 • ). According to an empirical linear relationship between zJ and the structural parameter Ru-X/Ru-X-Ru, which was proposed for lantern-type Ru 2 5+ complexes with axial halide (X − ) linkers by Delgado-Martinez et al., smaller Ru-X-Ru bond angles decrease the antiferromagnetic interaction [15]. The EPR spectra measured at 5 K in solid for 1-3 are given in Figure 8, Figures S3 and S4. The signal intensities were strong enough for 2 and 3 to analyze the spectra. Despite the weak signal intensities for 1, the g values were barely estimated. The estimated g values were g // = 2.040 and g ⊥ = 4.390 for 1; g // = 1.980 and g ⊥ = 4.385 for 2; and g // = 1.975 and g ⊥ = 4.335 for 3. For the S =3/2 system with D >> gβH, the estimated effective g values (g e = hν/βH) are g // e ≈ g // and g ⊥ e ≈ 2g ⊥ [7,17]. Thus obtained g values (g // = 2.040 and g ⊥ = 2.195 for 1; g // = 1.980 and g ⊥ = 2.1925 for 2; g // = 1.975 and g ⊥ = 2.168 for 3) are typical of the lantern-type Ru 2 5+ complexes [3,7,8,10,17]. The axial signal pattern was observed in 1:1 toluene/CH 2 Cl 2 at 3.4 K for 1 (g // = 1.9465 and g ⊥ = 4.400) [7].

Reflectance and Absorption Spectra
The diffuse reflectance spectra for the powder samples of 1-3 are given in Figure 9. All of the complexes showed a distinctive band at 430-490 nm with a discernible shoulder band at 550-690 nm and a broad band at 1030-1150 nm. These spectral features seem to be typical of lantern-type Ru2 5+ dinuclear complexes [1]. In fact, [Ru2{O2CC(CH3)3}4]BF4 has been reported as having corresponding bands; a band at 427 nm with a shoulder band at 545 nm and a band at 990 nm in the diffuse reflectance spectrum [18], which were assigned as π (Ru-O, Ru2) → π* (Ru2), δ*/π* (Ru2) → δ* (Ru-O), and δ(Ru2) → δ* (Ru2), respectively, according to their assignment in the literature [19]. Absorption spectra (measured in CH2Cl2) are shown in Figure 10. Absorption peaks are found in the near-ultraviolet (450-470 nm) and near-infrared region (1000-1150 nm) for all complexes. The similarity in the spectral features between the reflectance and absorption spectra indicates that the Ru2 5+ dinuclear skeletons were maintained in the solution.

Reflectance and Absorption Spectra
The diffuse reflectance spectra for the powder samples of 1-3 are given in Figure 9. All of the complexes showed a distinctive band at 430-490 nm with a discernible shoulder band at 550-690 nm and a broad band at 1030-1150 nm. These spectral features seem to be typical of lantern-type Ru 2 5+ dinuclear complexes [1]. In fact, [Ru 2 {O 2 CC(CH 3 ) 3 } 4 ]BF 4 has been reported as having corresponding bands; a band at 427 nm with a shoulder band at 545 nm and a band at 990 nm in the diffuse reflectance spectrum [18], which were assigned as π (Ru-O, Ru 2 ) → π* (Ru 2 ), δ*/π* (Ru 2 ) → δ* (Ru-O), and δ(Ru 2 ) → δ* (Ru 2 ), respectively, according to their assignment in the literature [19]. Absorption spectra (measured in CH 2 Cl 2 ) are shown in Figure 10. Absorption peaks are found in the near-ultraviolet (450-470 nm) and near-infrared region (1000-1150 nm) for all complexes. The similarity in the spectral features between the reflectance and absorption spectra indicates that the Ru 2 5+ dinuclear skeletons were maintained in the solution. O), and δ(Ru2) → δ* (Ru2), respectively, according to their assignment in the literature [19]. Absorption spectra (measured in CH2Cl2) are shown in Figure 10. Absorption peaks are found in the near-ultraviolet (450-470 nm) and near-infrared region (1000-1150 nm) for all complexes. The similarity in the spectral features between the reflectance and absorption spectra indicates that the Ru2 5+ dinuclear skeletons were maintained in the solution.
Hence, the axial chloride ligations of 2 could also be considered as kept in the measured CH 2 Cl 2 solution, although the reversibility of the redox couple (E pc = −0.46 V and E pa = −0.14 V) was not good when compared with that of 4 (E pc = −0.40 V and E pa = −0.28 V) [10]. We performed DFT calculations to estimate the redox potentials (E calc 1/2 ) for Ru 2 5+ → Ru 2 4+ as well as the Ru 2 6+ → Ru 2 5+

General Aspects
All reagents and solvents were used as received. The complex [Ru2(O2CC3H7)4Cl]n (1) was prepared according to a published procedure [8].
Elemental analyses for carbon, hydrogen, and nitrogen were performed using a Yanako CHN Corder MT-6. Infrared spectra (KBr pellets) were measured with a JASCO FT/IR-4600. Absorption spectra and diffuse reflectance spectra were obtained using JASCO V-670 and Shimadzu UV-3100 spectrometers, respectively. The temperature dependent magnetic susceptibilities were measured over the temperature range of 2-300 K at the constant field of 0.5 T with a Quantum Design MPMS XL-5. The measured data were corrected for diamagnetic contributions [20]. EPR spectra were measured at 5 K in solid by a BRUKER ELEXSYS E500 equipped with OXFORD ESR900 and OXFORD ITC503 attachments. The EPR simulation was conducted using the "Hyperfine Spectrum" program with spin Hamiltonian, Hs = βB•gS [21]. Cyclic voltammograms (CVs) were measured in dichloromethane containing n Bu4N(BF4) on a BAS ALS-DY2325 electrochemical analyzer. A glassy carbon disk (1.5 mm radius), platinum wire, and saturated calomel electrodes were used as the working, counter, and reference electrodes, respectively. All of the potential values are described versus SCE.

General Aspects
All reagents and solvents were used as received. The complex [Ru 2 (O 2 CC 3 H 7 ) 4 Cl] n (1) was prepared according to a published procedure [8].
Elemental analyses for carbon, hydrogen, and nitrogen were performed using a Yanako CHN Corder MT-6. Infrared spectra (KBr pellets) were measured with a JASCO FT/IR-4600. Absorption spectra and diffuse reflectance spectra were obtained using JASCO V-670 and Shimadzu UV-3100 spectrometers, respectively. The temperature dependent magnetic susceptibilities were measured over the temperature range of 2-300 K at the constant field of 0.5 T with a Quantum Design MPMS XL-5. The measured data were corrected for diamagnetic contributions [20]. EPR spectra were measured at 5 K in solid by a BRUKER ELEXSYS E500 equipped with OXFORD ESR900 and OXFORD ITC503 attachments. The EPR simulation was conducted using the "Hyperfine Spectrum" program with spin Hamiltonian, H s = βB•gS [21]. Cyclic voltammograms (CVs) were measured in dichloromethane containing n Bu 4 N(BF 4 ) on a BAS ALS-DY2325 electrochemical analyzer. A glassy carbon disk (1.5 mm radius), platinum wire, and saturated calomel electrodes were used as the working, counter, and reference electrodes, respectively. All of the potential values are described versus SCE.

Crystal Structure Determination
Single crystals of 2 and 3 suitable for X-ray crystal structure analysis were obtained by the recrystallization from dichloromethane-diethyl ether and dichloromethane-benzene mixed solvents, respectively. X-ray crystallographic data (Table 1) were collected for a single crystal at 90 K on a Bruker CCD X-ray diffractometer (SMART APEX) using graphite-monochromated Mo Kα radiation (λ = 0.71073 Å) for 2 and a RIGAKU Saturn 724 CCD system equipped with a Mo rotating-anode X-ray generator with monochromated Mo Kα radiation (λ = 0.71075 Å) for 3. Diffraction data of 2 and 3 were processed using APEX2 (Bruker) and CrystalClear-SM (RIGAKU), respectively. The structures of 2 and 3 were solved by intrinsic phasing methods (SHELEX) and direct methods (SIR-2011), respectively and refined using the full-matrix least-squares technique (F 2 ) with SHELXL-2014 as part of the SAINT (Bruker) (Billerica, MS, USA) and CrystalStructure 4.2.5 (RIGAKU) (Tokyo, Japan) software, respectively. Non-hydrogen atoms were refined with anisotropic displacement parameters, and all hydrogen atoms were refined with a riding model. Selected bond distances and angles for 2 and 3 are given in Tables S1 and S2,

Computational Details
The unrestricted density functional theory (uDFT) calculations applied in this study were performed with the long-range and dispersion correlated hybrid DFT functional method, ωB97XD, on the Gaussian 09 program [23]. The Los Alamos effective core potential LANL08(f) and Pople's 6-311 + G* basis sets were applied for the Ru and other atoms, respectively. All molecular geometries were fully optimized and checked by the vibrational frequency analyses. The solvent effect of CH 2 Cl 2 was considered by the polarizable continuum model (PCM). The redox potentials were estimated by using the standard method with the Born-Harbor cycle and Gibbs free energy changes, which was defined by Noodleman [24]. In order to estimate the redox potentials (E calc 1/2 ) for the Ru 2 5+ → Ru 2 4+ and Ru 2 6+ → Ru 2 5+ processes, the atomic coordinates of optimized geometries for  of 3), respectively. All of the coordinates used for the estimations are given in Tables S3-S11. We subtracted 4.68 V (IUPAC value) [25] from the calculated absolute potentials of the Ru 2 complexes to make a direct comparison to the experimental CV data referenced to the SCE.

Conclusions
By