Easy Ligand Activation in the Coordination Sphere of Ru inside the [PW₁₁O₃₉]^7– Backbone

Abstract

Irradiation of the Keggin-type [PW₁₁O₃₉{Ru(NO)}]⁴⁻ (Ru-NO) polyoxometalate in CH₃CN results in rapid NO ligand elimination with the formation of [PW₁₁O₃₉{Ru^III(CH₃CN)}]⁴⁻ (Ru-CH₃CN). This complex offers an easy entry into the Ru-based chemistry of the {PW₁₁Ru} complex. Attempts to substitute N₃⁻ for CH₃CN in the presence of an NaN₃ excess lead a variety of products: (i) [PW₁₁O₃₉{Ru^III(N₃)}]⁴⁻ (Ru-N₃); (ii) [PW₁₁O₃₉{Ru^III(N₄HC-CH₃)}]⁴⁻ (Ru-Tz) as a click-reaction product; and (iii) [PW₁₁O₃₉{Ru^II(N₂)}]⁵⁻ (Ru-N₂). UV-VIS, CV, and HR-ESI-MS techniques were used for the reaction monitoring and characterization of the products.

1. Introduction

Chemistry of noble-metal-substituted polyoxometalates, apart from general chemical interest of bringing noble metal centers in coordination to what can be regarded as noninnocent multidentate oxygen-donor ligand, is important for the search of new pathways of coordinated molecules activation and design of new catalysts, including photo- and electrocatalysts.

Of the polyoxometalates (POMs) incorporating noble metals, the most studied are Ru-substituted Keggin-type heteropolytungstates, where a {W=O}⁴⁺ moiety in the plenary Keggin archetype, [XW₁₂O₄₀]ⁿ⁻ (X = P (n = 3), Si and Ge (n = 4)), is substituted by a {RuL} group. L can be neutral or charged, such as O²⁻ or N³⁻, and Ru in POMs is known to exist in at least five oxidation states, ranging from Ru(II) to Ru(VI). Incorporation of Ru confers unique redox and catalytic properties on the resulting POMs [1,2]. Catalytic activities of [PW₁₁O₃₉Ru^III(H₂O)]⁴⁻, [PW₁₁O₃₉Ru^II(DMSO)]⁵⁻ (DMSO: dimethyl sulfoxide), [SiW₁₁O₃₉Ru^III(H₂O)]⁵⁻, [SiW₁₁O₃₉Ru^III(DMSO)]⁵⁻, and [GeW₁₁O₃₉Ru^III(H₂O)]⁵⁻ in oxidation of olefins [3,4,5], water [6,7,8], DMSO [9,10], and alcohols [11,12]; reduction of DMSO [10] and carbon dioxide [13]; and oxidative C–C bond formation [14] have been reported. The H₂O ligand attached to Ru is exchangeable with other organic and inorganic donor molecules to form Ru-pyridine [10,15,16], Ru-pyrazine [12], Ru-DMSO [10,17,18], Ru-NO [19,20,21], Ru-Cl [22], Ru-CO [23,24,25], Ru-olefin [10], and Ru-O-Ru derivatives [26,27].

In 2013 we reported incorporation of a {Ru(NO)}³⁺ group into a Keggin type POM, [PW₁₁O₃₉]⁷⁻, and versatile reactivity of the coordinate NO ligand [20,21]. Easily available [Ru(NO)Cl₅]²⁻ offers serious advantages as Ru source for preparation of Ru-substituted Keggin-type heteropolytungstates due to the redox stability of the {Ru(NO)}³⁺ unit, hydrolytic inertness, and consequent lack of uncontrollable hydrolytic oligomerization, though the inertness of coordinated Cl ligands in this low-spin d⁶-Ru(II) complex requires drastic reaction conditions. The coordinated NO can be destroyed or eliminated, leaving a vacant coordination site, which can open a way for activation of different substrates in catalytical applications [28,29,30]. Recently, we reported the reactions of [Ru(NO)Cl₅]²⁻ with pseudotrivacant B-α-[XW₉O₃₃]⁹⁻ (X = As^III, Sb^III) at 160 °C resulting in rearrangement of the polyoxometalate backbones into {XM₁₈} structures [31]. In the case of [α-B-AsW₉O₃₃]⁹⁻, oxidation of As(III) to As(V) took place, accompanied by rearrangement into Dawson type [As₂W₁₇{Ru(NO)}O₆₁]⁷⁻. In the case of [α-B-SbW₉O₃₃]⁹⁻, the product [SbW₁₇{Ru(NO)}O₅₉]¹⁰⁻ was isolated as (DMAH)₁₀[SbW₁₇{Ru(NO)}O₅₉]∙11H₂O.

Another aspect highlighted here is reactions of coordinated ligands [32] or ligand activation in the coordination sphere of noble metals. Numerous research papers have been published in this field. Specifically we would like to highlight the reactivity of coordinated nitriles in the Pt coordination sphere [33,34,35,36,37] or reactivity of triosmium clusters toward a wide range of substrates [38,39,40,41]. UV-irradiation (or photolysis) is a well-known technique for a lot of chemical processes, including (i) ligand exchange reactions (e.g., low-valence iron, osmium clusters) [42,43], (ii) ligand transformation [44,45,46], and (iii) CO₂ capture [47].

Here, we report an easy way to remove the NO ligand from [PW₁₁O₃₉(Ru(NO)}⁴⁻, quantitative generation of [PW₁₁O₃₉Ru^III(CH₃CN)]⁴⁻, and its reactions with N₃^– which follow three different pathways.

2. Results

Coordinated NO (formally, NO⁺) ligand in the {Ru(NO)}³⁺ complexes typically can play three roles: (i) stabilizing low-spin d⁶-Ru(II) and imparting stability and inertness on the coordination sphere; (ii) acting as spectator ligand, capable of taking up or releasing extra electron density on Ru due to π-backbonding; (iii) easy transformations or elimination (releasing a coordination site) upon irradiation or in the presence of reducing agents or nucleophiles. In this work we used mercury lamp irradiation to eliminate NO ligand from [PW₁₁O₃₉{Ru(NO)}]⁴⁻ (Ru-NO) (Figures S1–S3, Table S1) in CH₃CN solution producing corresponding [PW₁₁O₃₉{Ru^III(CH₃CN)}]⁴⁻ anion. The reaction can be viewed as dissociation of neutral NO molecule with the formation of coordinatively unsaturated species [PW₁₁O₃₉Ru^III]⁴⁻, followed by rapid solvation with CH₃CN, quantitatively yielding [PW₁₁O₃₉{Ru^III(CH₃CN)}]⁴⁻ (Ru-CH₃CN). This process was followed by IR, UV-VIS, and CV techniques.

After 90 min irradiation of a solution of Ru-NO in CH₃CN, the UV-VIS spectrum changed (Figure 1) by disappearance of the band at 412 nm (ε = 1059 M^–1cm^–1), and appearance of two new bands at 356 nm (ε = 2843 M⁻¹cm⁻¹) and 410 nm (ε = 1720 M⁻¹cm⁻¹). Cyclic voltammetry (CV) can be used for the monitoring of the substitution of CH₃CN for NO by following the change in the corresponding redox processes. The CV plot of Ru-CH₃CN in acetonitrile is shown in Figure 2. Three redox processes at E_1/2 = 0.92, −0.33 and −1.64 V (vs. Ag/AgCl) were detected. The difference between the anodic and cathodic peak potentials (ΔE) does not exceed 80 mV, and the ratio between the anodic and cathodic peak currents (I_a/I_c) is close to 1 for each of these processes. All this indicates the reversibility of the processes. The oxidation process at 0.92 V corresponds to Ru(II)/Ru(III) couple. It was previously found that a similar Ru-NO complex reversibly oxidizes at 1.29 V (vs. Ag/AgCl) [20]. Two redox processes at −0.33 and −1.64 V should correspond to the reduction of W(VI) in the {PW₁₁} framework, as is well documented.

The Ru-CH₃CN complex can be isolated as TBA-salt by diethyl ether diffusion for crystalline sample or bulk ether precipitation of the crude product. IR-spectroscopy confirms absence of NO ligand by disappearance of the tell-tale NO band at 1846 cm⁻¹ (Figures S7 and S8).

Ru-CH₃CN was prepared to study reactivity at the Ru site in various reactions. We tried to exchange the coordinated acetonitrile molecule with azide ligand in order to make the [PW₁₁O₃₉Ru^III(N₃)]⁵⁻ (Ru-N₃) complex and to study the reactions of the coordinated azide ligand. The reaction was monitored by CV. Heating of Ru-CH₃CN with a large excess of NaN₃ results in the changing of the UV-VIS spectrum (Figure 1). The spectrum differs from that of Ru-CH₃CN or Ru-NO and looks essentially featureless. The implicit absorption bands could be identified by calculating wavelength derivatives [48]. Two bands at 366 nm (ε = 2888 M⁻¹cm⁻¹) and 415 nm (ε = 1633 M⁻¹cm⁻¹) were detected (Figure S9). Substitution of acetonitrile leads to a noticeable cathodic shift in the Ru(II)/Ru(III) potential from 0.92 V (red line in Figure 3) to 0.57 V (black line in Figure 3). Additionally two small irreversible shoulders were detected. However, within the concentrations of NaN₃ employed in this work, we were unable to achieve complete substitution of CH₃CN by azide by working in CH₃CN. However, with large excess in NaN₃, the cyclic voltammogram plot shows the presence of three reversible Ru(II)/Ru(III) waves in the positive region (Figure 4). The first one corresponds to the formation of the azide complex, and the second one corresponds to the unreacted acetonitrile complex.

The composition of the complex was confirmed with elemental analysis and HR-ESI-MS (Figure 5, Figures S4–S6, Table S2). We also used HR-ESI-MS techniques to study these reactions. The products were isolated by diethyl ether precipitation as mixtures of TBA salts and redissolved in CH₃CN for the measurements. The spectra (Figure 5, Figures S10–S15, Table S3) indicate the presence of both Ru-N₃ and Ru-CH₃CN complexes. Moreover, there are other species, such as {2H⁺ + Ru-N₃ + CH₃CN}³⁻ (m/z 954.451), {Na⁺ + H⁺ + Ru-N₃ + CH₃CN}³⁻ (m/z 961.779), or {2Bu₄N⁺ + Ru-N₃ + CH₃CN}³⁻ (m/z 1115.424) (Table S3). Since an acetonitrile molecule is routinely required for matching calculated and experimental m/z values, this can be attributed to the formation of N₄HC-CH₃ (Tz) tetrazole or tetrazolate (Tz⁻) in the coordination sphere of ruthenium as a click reaction result ([2 + 3] dipolar addition of azide to the nitrile bond). We have to conclude that the formation of Ru-N₃ and [PW₁₁O₃₉{Ru^III(N₄HC-CH₃)}]⁴⁻ (Ru-Tz) anion occurred during the reaction.

When the reaction is carried out in a weakly coordinating CH₃NO₂, the HR-ESI-MS data (Figure 5, Figures S16–S21, Table S4) reflect different pathways of ligand transformation. There are two products, with one of them being the above-mentioned Ru-Tz complex. This [PW₁₁O₃₉{Ru^III(N₄HC-CH₃)}]⁴⁻ (Ru-Tz) was detected in the composite peaks {2Bu₄N⁺ + Ru-Tz}²⁻ (m/z 1673.633) and {2Bu₄N⁺ + Ru-Tz + H₂O}²⁻ (m/z 1682.642).

We can conclude that activation of azide ligand in the coordination sphere of Ru leads to a fast reaction with CH₃CN into the tetrazole molecule. We proposed the following mechanism of this reaction (Figure 6).

In the CH₃CN solution, and in the presence of excess azide, there are competing reactions leading to Ru-N₃ and Ru-CH₃CN complexes. These are not possible in CH₃NO₂ when stoichiometric amounts of azide are employed.

The azide–alkyne Huisgen cycloaddition is a 1,3-dipolar cycloaddition between an azide and a terminal or internal alkyne to give a 1,2,3-triazole [49]. Typically, such reactions can be catalyzed by Cu(I) [50], Ru [51,52], and Ag(I) [53,54]. The Ru-based catalysts are typically low-valent organometallic complexes like [Cp*RuCl(PPh₃)₂], [Cp*RuCl(COD)], and [Cp*RuCl(NBD)] [52]. The ruthenium-catalyzed azide−alkyne cycloaddition proceeds by an oxidative coupling of the azide and alkyne to give a six-membered ruthenacycle intermediate, in which the first new carbon−nitrogen bond is formed between the more electronegative carbon of the alkyne and the terminal, electrophilic nitrogen of the azide. In our case POM plays the role of a bulky ligand with “hard” oxo-environment, and ruthenium has 3+ oxidation state. This nonfavorable arrangement can induce partial leaving of a noble metal atom from the POM lacuna to achieve the suitable geometry for oxidative addition. The above-mentioned mechanism does not include changing of the Ru coordination environment and can be more favorable.

Moreover, analysis of the HR-ESI-MS data reveals another product in the reaction mixture in CH₃NO₂. The m/z calculations suggest coordination of the {PW₁₁O₃₉Ru} moiety with a ligand of molecular weight between 28 and 30, which can be addressed to CO, NO, or N₂. However, CH₃NO₂ cannot be a source of NO or CO under reaction conditions and most likely is not involved in the reaction. We are to assume, therefore, that this ligand comes from azide, and can be nothing except N₂, attached to Ru^II (Figure 5). To confirm this, we studied the bulk solid mixture with Raman spectroscopy and found a band at 2135 cm⁻¹ (Figure S22). Stable Ru^II complexes with coordinated N₂ molecule have been reported since the 1960s, including the first examples of dinitrogen complexes, isolated as [Ru(NH₃)₅N₂]X₂ (X = Br⁻, I⁻, BF₄⁻, PF₆⁻) [55]. According to the Raman data, the N₂ stretches can lie between 2103 cm⁻¹ (for trans-[RuN₃(N₂)(en)₂]PF₆ [56]) and 2140 cm⁻¹ (for [Ru₂H₆N₂(PPh₃)₄] [57]). Therefore, based on the HR-ESI-MS and Raman data, we can suggest formation of [PW₁₁O₃₉{Ru^II(N₂)}]⁵⁻ complex following N₃⁻ activation upon coordination to the {PW₁₁O₃₉Ru}. The formation of the [PW₁₁O₃₉{Ru^II(N₂)}]⁵⁻ can be rationalized by the following simplified scheme:

{Ru-CH₃CN} + N₃^– = {Ru-N₃} + CH₃CN

(1)

{Ru-N₃} = {Ru-N} + N₂

(2)

{Ru-N} + {Ru-N₃} = 2{Ru-N₂}

(3)

After detection of Ru-N₂ complex in the mixture of products isolated after the reaction in CH₃NO₂, we checked a mixture isolated from acetonitrile and found peaks from the same complex with nitrogen. This means that heating of Ru-CH₃CN with azide induces two parallel reactions of activated azide, resulting in tetrazole formation or nitrogen stabilization. The [PW₁₁O₃₉Ru^III(N₃)]⁵⁻ complex must be considered as a highly reactive intermediate.

3. Materials and Methods

K₂[RuNOCl₅] was prepared according to the literature [58]. IR spectra were recorded on a FT-801 FT-IR spectrometer (Simex, Russia). Elemental analysis was carried out on a Eurovector EA 3000 CHN analyzer. The TGA measurements were performed on a NETZSCH TG 209 F3 thermobalance in aluminum crucibles while heating the samples from 30 to 300 °C at a step of 10 °C. A 1.5 kW full-spectrum Hg lamp (GO Pnik, Iskitim, Russia) was used for the bulk solution photolysis.

The cyclic voltammograms (CV) were recorded with a 797 VA Computrace system (Metrohm, Zurich, Switzerland). All measurements were performed with a conventional three-electrode configuration consisting of glassy carbon working and platinum auxiliary electrodes and an Ag/AgCl/KCl reference electrode. The solvent used in all experiments was CH₃CN, which was deoxygenated before use. Tetra-n-butylammonium hexafluorophosphate (0.1 M solution) was used as a supporting electrolyte. The concentration of the complexes was approximately 10⁻³ M. The potential scan rate was 100 mV/s. The half-wave potential (E_1/2) values were determined as (E_a + E_c)/2, where E_a and E_c refer to anodic and cathodic peak potentials, respectively. Ferrocene was used as an internal standard, and the Fc/Fc⁺ potential was 0.43 V.

The high-resolution electrospray ionization mass spectrometric (HR-ESI-MS) measurements were performed at the Center of Collective Use “Mass Spectrometric Investigations” SB RAS (Novosibirsk, Russia). Spectra were obtained with a direct injection of liquid samples on an ESI quadrupole time-of-flight (ESI-q-TOF) high-resolution mass spectrometer Maxis 4G (Bruker Daltonics, Bremen, Germany). The spectra were recorded in the 300–3000 m/z range in negative mode.

Raman spectra were recorded on a LabRAM Horiba spectrometer (Horiba, Kyoto, Japan). An ion He-Ne laser (Simex, Moscow, Russia) with a wavelength of exciting light of 633 nm was used. The spectra were obtained in the backscattering geometry using a Raman microscope.

Synthesis of (Bu₄N)₄[PW₁₁O₃₉{Ru(NO)}] (Ru-NO): 1 g (0.31 mmol) K₇[PW₁₁O₃₉]·14H₂O was dissolved in 15 mL H₂O. Then, the solution of 0.132 g (0.31 mmol) K₂[RuNOCl₅] in 1 mL H₂O was added. The resulting mixture was transferred into a Teflon-lined Parr autoclave and kept at 150 °C for 18 h. After cooling to the room temperature, TBABr was added (1 g, 0.95 mmol). The precipitation was filtered on a glass filter and rinsed with 200 mL of a distilled water. Yield: 1.070 g (95%). ³¹P NMR (CH₃CN + CD₃CN): −13.88 ppm. (Figure S23). IR (ATR, cm⁻¹): 1846(s), 1482 (m), 1462 (m), 1380 (w), 1151 (w), 1089 (m), 1036 (m), 956 (s), 884 (m), 793 (vs), 658 (m), 613 (w), 606 (w), 590 (w), 578 (w), 569 (w), 554 (w). EA, found C,H,N (%): 16.4, 3.1, 1.5; calc C,H,N (%): 16.3, 3.1, 1.6.

Synthesis of (Bu₄N)₄[PW₁₁O₃₉{Ru(CH₃CN)}] (Ru-CH₃CN): 0.1 g (0.028 mmol) of (Bu₄N)₄[PW₁₁O₃₉{Ru(NO)}] was dissolved in 10 mL of CH₃CN. (Bu₄N)₄[PW₁₁O₃₉{Ru(CH₃CN)}] was generated by photolysis for 90 min. The solid product was obtained by precipitation with diethyl ether. Yield: 0.07 g (64%). ³¹P NMR (CH₃CN + CD₃CN): −13.57 ppm. (Figure S24). IR (ATR, cm⁻¹): 1482 (m), 1465 (m), 1378 (w), 1152 (w), 1079 (m), 1045 (m), 954 (s), 879 (s), 784 (vs), 626 (m), 613 (m), 598 (w), 587 (m), 578 (w), 563 (w), 556 (w). EA, found C,H,N (%): 20.8, 3.6, 2.0; calc C,H,N (%): 20.9, 3.9, 1.8.

Reaction of Ru-CH₃CN with NaN₃ in acetonitrile: Crude sodium azide (0.017 g, 0.26 mmol) was added to the solution of freshly prepared (Bu₄N)₄[PW₁₁O₃₉{Ru(CH₃CN)}] (0.100 g, 0.026 mmol) in 10 mL of CH₃CN. The reaction mixture was kept at 70 °C upon stirring for 18 h. After cooling to room temperature, white precipitate was filtered off. The resulting product was isolated by vapor diffusion of Et₂O. Yield: 0.70 g.

Reaction of Ru-CH₃CN with NaN₃ in nitromethane: 0.120 g (0.032 mmol) of (Bu₄N)₄[PW₁₁O₃₉{Ru(CH₃CN)}] was dissolved in 7 mL of CH₃NO₂. Then, 0.044 g (0.64 mmol) of solid NaN₃ was added to the solution. The resulting mixture was transferred into a Teflon-lined Parr autoclave and kept at 100 °C for 18 h. The formed precipitate was filtered off, and the reaction product was isolated by the addition of diethyl ether. Yield: 0.092 g. IR (ATR, cm⁻¹): 1477 (m), 1469 (m), 1376 (w), 1150 (w), 1077 (s), 1047 (m), 970 (s), 955 (s), 888 (sh), 880 (s), 798 (sh), 784 (vs), 773 (vs), 657 (s), 627 (s), 594 (m), 580 (m), 559 (m).

4. Conclusions

This manuscript summarized our studies of Ru-atom reactivity inside the POM backbone toward azide anion. We detected two reaction pathways resulting in (i) azide–acetonitrile click reaction and (ii) azide decomposition. The first one produces coordinated tetrazole and the second generates a complex with coordinated N₂. Such reactivity is important for organic substrate transformation and N₂ activation. Further studies in these directions are in progress.