INTRODUCTION

An extension of a number of drugs for chemotherapy of oncological diseases is one of the topical tasks in the synthesis of new metal complexes. As noted in [1], it is ruthenium complexes with N-heterocycles that became the objects which managed to break the existing paradigm of the irreplaceability of platinum-containing drugs [2-5]. Ruthenium nitrosyl complexes with N-heterocycles have not only cytotoxicity comparable with that of cisplatin against different cancer cell lines [6], but also can regulate cell reparation processes due to photochemical activation with breaking the Ru–NO bond in solutions [7, 8]. Another, no less interesting aspect in the chemistry of ruthenium nitrosyl complexes is their ability to bond isomerization in a solid [9-11] for designing information storage devices and materials with nonlinear optical properties. Thus, the extension of a number of nitrosyl complexes with biomimetic N-donor ligands is a relevant research field.

In this work, we report the synthesis, characterization, structural data, and photoisomerization of the mer-[RuNO(3-PyCN)2Cl3] complex. The obtained experimental data are compared with the calculations using different basis sets and functionals to get an idea which method is mostly accurate in the description of these systems.

EXPERIMENTAL

Synthesis of mer-[RuNO(3-PyCN)2Cl3]. To a solution of 3-PyCN (200 mg, 1.8 mmol) in 6 mL of DMF K2[RuNOCl5] (250 mg, 0.6 mmol) was added. The obtained suspension was boiled for 1 h until the starting ruthenium-containing compound dissolved. After the reaction completion, the suspension was cooled to room temperature and potassium chloride was filtered out. The filtrate was evaporated to dryness and the product was reprecipitated from a solution in acetonitrile via diethyl ether. The product precipitate was filtered off, washed with three 2mL portions of diethyl ether, three 2 mL portions of water, and then dried on a filter in the air flow. The product yield was 58%. Crystals suitable for the single crystal X-ray diffraction (XRD) analysis were obtained by slow diffusion of diethyl ether into an acetonitrile solution of the complex. For С12H8N5O1Cl3Ru calculated (%): C 32.2, H 1.8, N 15.7. Found (%): C 32.8, H 2.0, N 15.2. Main bands in the IR spectrum (КBr, cm–1): 3134, 3038, 2850 – ν(С–H), 2251 – ν(С–N) of the nitrile group, 1880 – ν(NO), 565 – δ(Ru–N–O).

Single crystal XRD data for the complex were measured on a Bruker APEX DUO diffractometer with a 4K СDD detector and graphite monochromatized MoKα radiation (λ = 0.71073 Å) at 150 K. Absorption correction was applied using the SADABS program [12]. The structure was solved by direct methods and refined by the full-profile least-squares technique using the SHELXTL program [12]. All non-hydrogen atoms were refined as anisotropic, hydrogen atoms were calculated in ideal positions and refined with the riding model. Crystals of the complex (С12H8N5O1Cl3Ru, M = 445.65) are red prisms of the monoclinic crystal system (space group P21/c). At 150 K the unit cell parameters are: a = 8.1487(3) Å, b = 14.8118(6) Å, c = 13.0364(5) Å, β = 101.966(1)°, V = 1539.26(10) Å3, Z = 4, μ = 1.546 mm–1, F(000) = 872. In total, 12645 reflections were measured, out of which 3826 independent (Rint = 0.0736), 199 refinement parameters: R1 = 0.0256, wR2 = 0.0669. Atomic coordinates and other structural parameters have been deposited with the Cambridge Crystallography Data Center (№ 2015158; deposit@ccdc.cam.ac.uk or http://www.ccdc.cam.ac.uk/data_request/cif).

IR spectra were measured on a sample in a KBr pellet on a VERTEX 80v spectrometer. Photochemical isomerization was performed by means of a blue diode laser (445 nm, 100 mV) at 80 K. The pellet with the sample under study was cooled to 80 K in a cryostat and laser radiated for 20 min, and the spectra were measured with an interval of 5 s to determine the achieved equilibrium occupancy of the metastable state. Then the pellet with the sample was heated with a constant rate and IR spectra were measured. To determine the activation parameters for the decomposition of the metastable isomer the observed drop in the intensity of the respective absorption band of the nitrosyl group was modeled. The temperature sensor was calibrated by the known reference nitrosyl complex in the previous work [13].

Quantum chemical calculations by DFT of stable and metastable states of the trans-[RuNO(3-PyCN)2Cl3] complex were carried out using the ADF2017 program package [14]. In the calculations the BP and PBE functionals were used [15, 16], each of which was combined DZP, TZP, and TZ2P basis sets [17]. All calculations were performed in the gas phase for the uncharged closed-shell complex and zero total spin. After the primary optimization (PBE-DZP) the structural parameters were refined in the other calculations presented. From the geometrical parameters obtained bond lengths and the Ru–N–O angle in the nitrosyl group were calculated. To gain the idea about the accuracy of the calculation techniques used we calculated a sum of square differences: experimental and calculated bond lengths. The vibrational spectra of the compound under study were calculated for the optimized structures with the respective functional and basis set.

RESULTS AND DISCUSSION

Synthesis and structure of the [RuNO(3-CNPy)2Cl3] complex. A similar mer-[Ru(NO)(4-CNPy)2Cl3] complex was previously synthesized from Ru(NO)Cl3·2H2O by the reaction with 4-cyanopyridine in an ethanol solution with an yield of 60% [18]. The drawback of this procedure is the use of trichloronitrosyl ruthenium as the starting compound because it is difficult to obtain it in the individual form [19]. A more preferred starting nitrosyl complex is K2RuNOCl5 prepared from RuCl3 with a practically quantitative yield because it is convenient to store it since it is non-hygroscopic, stable and easily synthesized. Just this complex was taken to synthesize mer- and fac-[Ru(NO)Py2Cl3]: they were prepared from an aqueous solution by the reaction of pentachloronitrosyl ruthenate with an excess of pyridine with yields of 90% and 70% respectively [20]. The fac-isomer is a kinetic substitution product while for the synthesis of the mer-isomer a tetrapyridine complex was first obtained with subsequent reverse substitution of two pyridine ligands in a hydrochloric solution. Yet another way to obtain the mer-isomer is thermal isomerization of the fac-isomer. The replacement of the solvent (water by DMF) allows an increase in the reaction temperature, which results in a one-pot synthesis of the more stable mer-isomer at a thermodynamic control of the reaction [21].

According to the single crystal XRD data, obtained mer-[Ru(NO)(3-CNPy)2Cl3] crystallizes in the triclinic space group P21/c; the coordination polyhedron of the ruthenium atoma [RuN3Cl3] is a distorted octahedron. In the equatorial plane of the complex there are two 3-cyanopyridine molecules at trans-position to each other, and the other two positions are occupied by chlorine atoms, the nitrosyl group and the third chlorine atom are at axial positions (Fig. 1).

Cyanopyridine molecules are coordinated to ruthenium via the nitrogen atom of the aromatic ring with a Ru–N distance of 2.109 Å, which is slightly shorter than that in the similar mer-[Ru(NO)Py2Cl3] (2.121 Å) complex [20]. In the complex compound there are two types of chlorine atoms, and correspondingly, two Ru–Cl bond lengths: the chlorine atom coordinated at trans-position to the nitrosyl group is at a distance of 2.339 Å from ruthenium, and the other two atoms in the plane with cyanopyridine ligands are removed from ruthenium at a distance of 2.375 Å. A difference in ruthenium–chlorine bond lengths is characteristic of cis- and trans-coordinates in various nitrosyl complexes, both uncharged mer-[Ru(NO)Py2Cl3] (2.377 Å and 2.342 Å) and mer-[Ru(NO)(NH3)2Cl3] (2.372 Å and 2.336 Å) [20, 22] and charged mer-[Ru(NO)(NH3)3Cl2]Cl (2.398 Å and 2.394 Å) [23].

The Ru–NO distance is 1.744 Å; the N–O distance is 1.138 Å; the Ru–N–O angle is 174.4°. Table 1 compares these three parameters for complexes with 4-CNPy, pyridine, and ammonia. The ruthenium–nitrosyl group bond length insignificantly changes when 4-CNPy substitutes for 3-CNPy and significantly differs from the distance in the complex with ammonia. The N–O bond length is in the typical range 1-1.2 Å. The geometry of the Ru–N–O moiety is close to linear, however, with a significant difference as compared to the [Ru(NO)(NH3)2Cl3] complex: 174.4° for the complex under study against 180° for the complex with ammonia.

Fig. 1
figure 1

Structure of the trans-[Ru(NO)(3-CNPy)2Cl3] complex. Thermal vibration ellipsoids are drawn at a 50% probability.

Full size image
Table 1 Comparison of Parameters of the Nitrosyl Group of trans-[Ru(NO)(3-CNPy)2Cl3] with the Other Nitrosyl Complexes with N-Donor Ligands at the trans-Position
Full size table

The N(NO)–Ru–L angles between the nitrosyl group and ligands in the equatorial plane of [Ru(NO)(3-CNPy)2Cl3] are close to 90°: for cyanopyridine ligands the N(NO)–Ru–N(heterocycle) angles are 94.1° and 89.9° and the N(NO)–Ru–Cl(equat. pl) angles are 92.9° and и 90.9°. The angle between the nitrosyl nitrogen atom, the ruthenium group, and the chlorine atom at trans-position to NO is 178.4°. The torsion angle between the aromatic ring planes is 3.2°; the ring with the N11 atom is turned at 32.6° relative to the axial plane, and the ring with the N21 atom is turned at 35.9°.

The ruthenium–ruthenium distances between the neighboring molecules in the crystal lattice are in the range 7.03-9.69 Å. Cyanopyridine rings of the neighboring molecules are located above each other at a distance (between the ring centroids) of 3.631 Å (С24 and С14 carbon atoms are above N21 and N11 nitrogen atoms of the neighboring molecules) and make pairs. The planes of these rings are at an angle of 5.7°, which may indicate the occurrence of weak stacking interactions (Fig. 2). The volume of the reaction cavity around the nitrosyl group, which was calculated in Voronoi–Dirichlet polyhedra, is 28.82 Å3 and this is a quite typical value of nitrosyl complexes [24]. The shortest distances from nitrosyl group atoms to atoms of the neighboring molecules are 3.003 Å, 3.004 Å (nitrogen atoms of cyano groups).

Metal–ligand bond lengths and angle values are convenient parameters for the comparison of calculations and experimental data. The most indicative values of nitrosyl complexes are bond lengths between the ruthenium atom and the nitrosyl group and ligands as well as the Ru–N–O angle. For the ground state of the complex (GS, nitrosyl group is coordinated via the nitrogen atom) Table 2 lists ruthenium–ligand bond lengths, bonds in the nitrosyl group, and the Ru–N–O angle in calculations performed with the use of PBE or BP functionals and basis sets (DZP, TZP, TZ2P). All six combinations make it possible to obtain comparable data on ruthenium–ligand and nitrogen–oxygen bond lengths in the nitrosyl group. The Ru–NO bond length of 1.752 Å is closest to the experimental one in the calculations with the TZ2P basis set and the PBE functional and 1.755 Å with the BP functional, at the experimental value being 1.744 Å and average 1.768 Å for all the other calculations. All calculations of the N–O bond of the nitrosyl group result in values in the range 1.163-1.169 Å, whereas the experimental bond length is noticeably shorter – 1.138 Å. Data on bond between the ruthenium and chlorine atoms at the cis-position to the nitrosyl group are practically the same; the average calculated length of this bond is 2.418 Å, with the experimental one being 2.375 Å, and the most accurate results are again observed for the TZ2P basis set. Moreover, in the calculations the Ru–Nheterocycle bond length is also overestimated; in the most accurate PBE-TZ2P calculation the bond lengths are 2.120 Å against experimental of 2.109 Å. Thus, the best agreement between the calculated and experimental geometric parameters for the studied complex is observed for the PBE functional and the TZ2P basis set. Calculated wavenumbers of characteristic vibrations vary in the range 1822-1846 cm–1 for ν(NO) of the nitrosyl group and 2239-2249 cm–1 for ν(CN) of the cyano group at experimental values of 1880 cm–1 and 2251 cm–1.

Fig. 2
figure 2

Crystal packing of trans-[Ru(NO)(3-CNPy)2Cl3].

Full size image
Table 2 Data for GS of the mer-[RuNO(3-PyCN)2Cl3] Complex
Full size table

Metastable states of coordination nitrosyl complexes have been known in which the nitrosyl group is coordinated via the oxygen atom. For the complex with 3-cyanopyridine it is found that the isonitrile complex with Ru–ON coordination is indeed a local minimum in this system. The global minimum corresponding to GS of the complex lies lower than the minimum of the metastable state by 1.62 eV. After the geometric optimization we calculated IR spectra for the optimized structure. The calculated and experimental data for the metastable state (Table 3) were compared by juxtaposing ν(N–O) and ν(C–N) because the structural data for the metastable state of the complex are absent. For the metastable isomer the calculated wavenumber for the stretching vibration of the isonitrosyl group varies from 1744 cm–1 to 1767 cm–1, with the experimental value being 1749 cm–1, and that for the cyano group remains in the range 2239-2248 cm–1, with the experimental value being 2254 cm–1. Thus, the vibrational wavenumbers of the nitrosyl and cyano groups, which are closest to the experimental values, in both cases, were obtained in the calculation with the TZ2P basis set and the PBE functional.

Table 3 Data for the Metastable State of the mer-[RuNO(3-PyCN)2Cl3] Complex
Full size table

Photoisomerization of [RuNO(3-CN-Py)2Cl3] in the crystalline state. The formation of bond isomers during photo activation of solid nitrosyl complexes is accompanied by a change in the absorption band corresponding to the characteristic stretching vibration of the nitrosyl group. Therefore, IR spectroscopy is one of the most convenient techniques to determine parameters of the formation and inverse transformation of metastable bond isomers. For the MS1 isomer in which the nitrosyl group is coordinated through the oxygen atom (Ru–ON) a shift of the ν(NO) absorption band to lower energies is usually 130-150 cm–1 [25]. In the complex studied, a new vibrational band of the nitrosyl group arises at 1749 cm–1 with a shift of GS by 131 cm–1. The dependence of its intensity on the irradiation time (Fig. 3) shows that photochemical equilibrium is attained in 400-600 s. The reached maximum occupancy of the metastable state, which was determined from a decrease in the integrated intensity of the ν(NO) band of GS in the experiment with a decreased amount of the compound in the pellet, is 17%.

An elevation of the sample temperature causes an inverse transformation of the bond isomer to GS with coordination of the nitrosyl group by the nitrogen atom (Fig. 4). Curves of changes in the band intensity regularly shift to higher temperatures with increasing heating rate. Combined processing of all three curves in a model of the first order reaction enables the determination of the activation energy and pre-exponent for the MS1–GS transition: Ea = 46.5(1.7) kJ/mol and lgk0 = 10.8(1.0). A complex parameter governing the thermal stability of metastable bond isomers in nitrosyl complexes is considered to be the decomposition temperature Td (at that the reaction rate constant is 10–3 s–1 [26]), which is 176.3(7) K for the complex under study. It was shown earlier that distinctions in the spatial environment of the nitrosyl group did not affect the stability of metastable states [27]. Thus, the comparison with similar data obtained for the related mer-[RuNO(Py)2Cl3] complex [20] allows a hypothesis that the introduction of an electron-acceptor substituent into the pyridine ring causes a decrease in the thermal stability of the metastable isomer.

Fig. 3
figure 3

Dependence of the intensity of the ν(NO) vibrational band corresponding to MS1 on the irradiation time (T = 80 K, λ = 445 nm, optical power is 100 mW).

Full size image
Fig. 4
figure 4

Kinetics of changes in the intensity of the ν(NO) band corresponding to the MS1 isomer at different heating rates of the sample.

Full size image
Fig. 5
figure 5

View of frontier orbitals of the [RuNO(3-CNPy)2Cl3] complex and their energies (eV, PBE-TZ2P).

Full size image

According to the models proposed previously, photochemical excitation in nitrosyl complexes must lead to an electron transfer to the antibonding orbital of the Ru–NO bond with retaining the singlet state; then there is relaxation to the triplet state [28] and a subsequent transition to the energy minimum corresponding to local (metastable isomer) or global (GS) [29, 30]. Data of quantum chemical calculations (Fig. 5) indicate that the main contribution to two nearest LUMOs (LUMO and LUMO+1) is made by antibonding π orbitals corresponding to the Ru–NO bond (80-84%), with a small contribution of p orbitals of chlorine atoms or π orbitals of benzene rings. The HOMO with the highest energy is a non-bonding orbital consisting of 3р orbitals of chlorine atoms at the cis-position to the nitrosyl group (49%) and dxz orbitals of the ruthenium atom (42%). The calculated LUMO–HOMO difference is 1.8 eV, which is substantially lower than the radiation energy used for excitation (2.75 eV). Lower lying MOs (HOMO–1–HOMO–3) are composed of 3р orbitals of chlorine atoms at the cis-position to the nitrosyl group with a contribution of π bonding orbitals corresponding to the Ru–NO bond, from 15% to 25%. HOMO-4 is composed of 3р orbitals of chlorine atoms (95%). The next in energy group of occupied MOs (see, e.g., HOMO–6) corresponds to linear combinations of π bonding orbitals of benzene rings and 3p orbitals of chlorine atoms. For HOMO–3–HOMO–2 orbitals the energy of a transition to the LUMO (LUMO+1) is in the range 2.6-3.0 eV, which is close to the energy of radiation applied. Correspondingly, it is possible to assume that for the studied complex photoexcitation must cause an electron transfer from the bonding π(Ru–NO) orbital to the anti-bonding π*(RuNO) one, thus leading to the weakening of the Ru–NO bond.

CONCLUSIONS

A novel mer-[RuNO(3-CNPy)2Cl3] complex was obtained as a result of the reaction between K2[RuNOCl5] and 3-cyanopyridine in DMF. The complex is characterized by the elemental analysis, IR spectroscopy, and single crystal XRD. It is shown that during photochemical activation (445 nm, 100 mW) a metastable bond isomer (MS1) forms with coordination of the nitrosyl group through the oxygen atom. Data of quantum chemical calculations reveal that the structural and spectral features of the studied complex are best reproduced when the PBE functional and the TZ2P basis set are used, with a deviation of main geometric parameters not exceeding 0.03 Å. The analysis of MOs of the complex with regard to the radiation wavelength applied for photoisomerization shows that primary excitation occurs from bonding orbitals of the Ru–NO coordinate to the anti-bonding orbital of the nitrosyl group, which causes the weakening of Ru–N and N–O bonds in the nitrosyl moiety. Experimental data confirm this fact; the ν(NO) absorption band corresponding to MS1 shifts relative to the main ν(NO) band (GS) to lower energies. Activation parameters of the thermally induced MS1–GS transformation are found, which allow the assumption that the introduction of an electron-acceptor substituent into the N-donor ligand leads to a small decrease in the thermal stability of MS1.