Mechanistic investigation of the formation of H₂ from HCOOH with a dinuclear Ru model complex for formate hydrogen lyase

Abstract

We report the mechanistic investigation of catalytic H₂ evolution from formic acid in water using a formate-bridged dinuclear Ru complex as a formate hydrogen lyase model. The mechanistic study is based on isotope-labeling experiments involving hydrogen isotope exchange reaction.

Graphical Abstract

Keywords:

• Mechanistic investigation
• H₂ evolution
• formic acid
• formate hydrogen lyase model
• dinuclear Ru complex

Classification:

• 30 Bio-inspired and biomedical materials
• 106 Metallic materials

1. Introduction

Formate hydrogen lyase (FHL) is an enzyme complex that catalyzes the conversion of HCOOH to evolve CO₂ and H₂ (Eq. 1) [1–3]. This enzyme complex is composed of a formate dehydrogenase (FDH) [4,5] and a [NiFe]hydrogenase ([NiFe]H₂ase) [6–9], and catalyzes oxidation of HCOOH (Eq. 2) and evolution of H₂, respectively (Eq. 3).(1) $$\text{HCOOH}\stackrel{\text{FHL}}{\leftrightarrows }{\text{CO}}_2+{\text{H}}_2$$(1) (2) $$\text{HCOOH}\stackrel{\text{FDH}}{\leftrightarrows }{\text{CO}}_2+2{\text{H}}^{+}+2{\text{e}}^{-}$$(2) (3) $$2{\text{H}}^{+}+2{\text{e}}^{-}\stackrel{\left[\text{NiFe}\right]{\text{H}}_{\text{2}}\text{ase}}{\leftrightarrows }{\text{H}}_2$$(3)

We have previously reported a Ni^IIRu^II model complex [Ni^II(X)(HCOO)(μ-H)Ru^II(C₆Me₆)] (X = N,N′-dimethyl-3,7-diazanonane-1,9-dithiolato) that can catalyze fast conversion of HCOOH to H₂ and CO₂ [10] [turnover frequency {TOF = (mol of evolved H₂/mol of catalyst) per hour} = 857 h⁻¹]. This extremely fast reaction rate, however, prevented us from investigating the mechanism of H₂ formation.

As part of our efforts to investigate the mechanism of H₂/CO₂ formation from HCOOH, we now report a dinuclear Ru^I complex [Ru^I₂(CO)₄(μ-HCOO)₂(DMSO)₂] (1) that catalyzes the above reaction at a significantly slower reaction rate (TOF = 13.1 h⁻¹). Here we disclose the detailed mechanism of H₂ formation from HCOOH based on isotope-labeling experiments.

2. Experimental details

2.1. Materials and methods

All experiments were carried out under N₂ or Ar atmosphere by using standard Schlenk techniques and a glovebox. Tetrahydrofuran (THF) was distilled from Na/benzophenone under N₂ atmosphere prior to use. HCOOH, DCOOD, 40% NaOD/D₂O, and dimethylsulfoxide (DMSO) were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). HCOOD and DCOOH were purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). These materials were used without further purification. Proton nuclear magnetic resonance (¹H NMR) spectra were recorded on a JEOL JNM-AL300 spectrometer (JEOL, Tokyo, Japan) at 25 °C, in which the chemical shifts were referenced to tetramethylsilane (TMS) in chloroform-d₁ and DMSO in DMSO-d₆. UV–vis absorption spectra were recorded on a JASCO V–670 UV–Visible–NIR spectrophotometer (the light pass length was 1.0 cm). An IR spectrum was recorded on a Thermo Nicolet NEXUS 870 Fourier transform infrared (FTIR) instrument (Thermo Fisher Scientific, Massachusetts, USA) at 25 °C. Gas chromatographic (GC) analyses were conducted by a Shimadzu GC–8A (He carrier) (Shimadzu, Kyoto, Japan) with a MnCl₂–alumina column (model: Shinwa OGO–SP) at –196 °C (liquid N₂) for quantitative analyses of H₂, HD, and D₂ and by a Shimadzu GC-2014 (Ar carrier) (Shimadzu, Kyoto, Japan) with activated charcoal at 100 °C for quantitative analyses for H₂, HD, D₂, and CO₂ using a thermal conductivity detector. Elemental analysis data were obtained by a PerkinElmer 2400II series CHNS/O analyzer (PerkinElmer, Massachusetts, USA) using Ar as the carrier gas. Dynamic light scattering measurements were conducted with a Malvern Zetasizer Nano (Malvern, Worcestershire, UK). X-band electron spin resonance (ESR) spectra were measured by a JEOL JES-FA200 spectrometer (JEOL, Tokyo, Japan) at –150 °C.

pH Adjustment. The pH of the solution was adjusted by using HCOOH and NaOH/H₂O (1.0–7.0). The pD of the solution was adjusted by using HCOOH, DCOOH, HCOOD, DCOOD, and 40% NaOD/D₂O (1.0–7.0), in which H⁺ concentration of HCOOH and DCOOH is negligible quantity compared to D⁺ concentration of D₂O. In a pH (or pD) range of 1.0–7.0, the pH (or pD) values of the solutions were determined by a pH meter (model: TOA HM20 J; DKK-TOA, Tokyo, Japan) equipped with a pH combination electrode (model: TOA GST-5725C; DKK-TOA, Tokyo, Japan). Values of pD were corrected by adding 0.4 to the observed values (pD = pH meter reading + 0.4) [11,12].

[Ru^I₂(CO)₄(μ-HCOO)₂(DMSO)₂] (1). A THF solution of [Ru⁰₃(CO)₁₂] (1.00 g, 1.56 mmol) and HCOOH (12 mL, 0.32 mol) was refluxed for 3.5 h. The resulting solution was cooled to room temperature, to which was added DMSO (335 μL, 4.72 mmol) followed by its stirring for 30 min. The solvent was removed under reduced pressure to yield a yellow powder of 1, which was collected and dried in vacuo {yield: 85% based on the amount of [Ru⁰₃(CO)₁₂] added}. ¹H NMR (300 MHz, in chloroform-d₁, referenced to TMS): δ 8.23 (s, 2H, HCOO), 3.06 (s, 12H, (CH₃)₂SO). Anal. Calcd for 1 (C₁₀H₁₄O₁₀Ru₂S₂): C, 21.43; H, 2.52%. Found: C, 21.45; H, 2.22%.

Typical procedure of H₂ evolution from HCOOH catalyzed by 1. In a 3.0 mL vial capped with a septum, a solution of HCOOH (2.60 mmol) in H₂O (1.0 mL) was added to 1 (1.25 μmol). The pH of the resulting solution was adjusted to 1.0–7.0 and the solution was heated at 80 °C for 1 h. The gas above the solution within the vial was sampled with a gas-tight syringe (500 μL) and analyzed for H₂ and CO₂ by GC. No CO was observed. No nanoparticles were formed in the catalytic reaction, which was confirmed by dynamic light scattering measurements.

Isotope-labeling experiments for catalytic H₂, HD, and D₂ evolution. In a 3.0 mL vial capped with a septum, DCOOH (2.60 mmol) in H₂O (1.0 mL), HCOOD (2.60 mmol) in D₂O (1.0 mL), or DCOOD (2.60 mmol) in D₂O (1.0 mL) was added to 1 (1.25 μmol). The pH or pD of the resulting solution was adjusted to 3.5 by the addition of NaOH or NaOD, respectively, and it was stirred at 80 °C for 1 h. The gas above the solution within vial was sampled with a gas-tight syringe (500 μL) and analyzed for H₂, HD, D₂, and CO₂ by GC. No CO was observed. No nanoparticles were formed in the catalytic reaction, which was confirmed by a dynamic light scattering measurements.

The initial rate of the catalytic H₂ evolution against the catalyst concentration. In a 3.0 mL vial capped with a septum, a solution of HCOOH (2.60 M) in H₂O was added into 1 (0.63, 1.25, 2.5, or 5.0 mM). The pH of the resulting solution was adjusted to 3.5 and the solution was heated at 80 °C for 300 s. The gas above the solution within the vial was sampled with a gas-tight syringe (500 μL) and analyzed for H₂ by GC. The catalytic reaction is first-order against the catalyst concentration.

Reactivity of hydride species 2 toward proton of HCOOH. In a 3.0 mL vial capped with a septum, 10 equivalents of HCOOH (9.4 μL, 250 μmol) was added into a DMSO (2.0 mL) solution of 2 (vide infra), which was prepared from the reaction of 1 (14 mg, 25 μmol) with HCOONa (1.7 mg, 25 μmol) at 80 °C for 1 h. No H₂ gas was formed, as confirmed by GC analysis. The same reaction was conducted using DMSO-d₆ (450 μL) instead of DMSO (2.0 mL), which was monitored by ¹H NMR spectroscopy. No decrease of the hydride-derived peak of 2 was observed.

H⁺/D⁺ exchange of hydride ligand of 2. In an NMR sample tube, a DMSO-d₆ solution (400 μL) of 1 (14 mg, 25 μmol) with HCOONa (1.7 mg, 25 μmol) was heated at 80 °C for 1 h to form 2, which was confirmed by ¹H NMR spectroscopy. Then, D₂O (50 μL) was added into the resulting DMSO-d₆ solution under N₂ atmosphere. The H⁺/D⁺ exchange of hydride ligand of 2 was confirmed by ¹H NMR spectroscopy with CH₂Br₂ as an internal standard to investigate the intensities of hydride- and formate-derived peaks.

X-ray crystallographic analysis of 1. A single crystal of 1 suitable for X–ray analysis was obtained from the diffusion of diethyl ether into its THF solution. Measurements were performed on a Rigaku/MSC Saturn CCD diffractometer (Rigaku, Tokyo, Japan) with confocal monochromated Mo-Kα radiation (λ = 0.7107 Å). Data were collected and processed using the CrystalClear program. All calculations were performed using the CrystalStructure crystallographic software package except for refinement, which was performed using SHELXL–97. Crystallographic data for 1 have been deposited at the Cambridge Crystallographic Data Centre as Supplementary Publication No. CCDC 1556459. Copies of the data can be obtained free of charge on application to CCDC, 12 Union Road, Cambridge CB2 1EZ, UK {fax.: [phone]; e-mail: [email protected]}.

3. Results and discussion

A bis(μ-formate) Ru^I₂ complex, [Ru^I₂(CO)₄(μ-HCOO)₂(DMSO)₂] (1) was synthesized from the reaction of [Ru⁰₃(CO)₁₂] with HCOOH and DMSO in THF under a N₂ atmosphere, which was then characterized by X-ray analysis, and ¹H NMR, UV-vis, and IR spectroscopies (Figures 1–4). Both Ru atoms adopt a distorted octahedral geometry in which the Ru ion is ligated by two C(CO), an S(DMSO), two O(μ-formate), and an adjacent Ru atom (Figure 1). Two Ru atoms are tethered by a Ru–Ru bond and two formate ligands. The bond distance of Ru1–Ru2 {2.6654(3) Å} is similar to those of other formate-bridged dinuclear Ru complexes (2.679 and 2.720 Å) [13,14]. The C–O bond distances of each HCOO^– ligand {C1–O1 = 1.258(2), C1–O2 = 1.262(2), C2–O3 = 1.263(2), and C2–O4 = 1.259(2) Å} are almost the same, which indicates that this could be explained by the delocalization of a double bond over two C–O bonds. The C–O bond distances of four CO ligands {1.142(2)–1.148(2) Å} are slightly longer than that of free CO (1.128 Å), suggestive of back donation of electron density from the t_2 g orbital of low-valent Ru^I to the π*-antibonding orbital of CO.

Figure 1. An ORTEP drawing of 1 with thermal ellipsoids at 50% probability. The hydrogen atoms of DMSO are omitted for clarity. Selected interatomic distances (l/Å) and angles (ϕ/°): Ru1–Ru2 = 2.6654(3), Ru1–O1 = 2.1152(13), Ru1–O3 = 2.1290(13), Ru1–S1 = 2.4107(5), Ru1–C3 = 1.855(2), Ru1–C5 = 1.853(2), Ru2–O2 = 2.1374(13), Ru2–O4 = 2.1221(13), Ru2–S2 = 2.4134(5), Ru2–C4 = 1.855(2), Ru2–C6 = 1.849(2), C1–O1 = 1.258(2), C1–O2 = 1.262(2), C2–O3 = 1.263(2), C2–O4 = 1.259(2), C3–O5 = 1.145(2), C4–O6 = 1.142(2), C5–O7 = 1.148(2), C6–O8 = 1.146(2), O1–C1–O2 = 126.67(18), O3–C2–O4 = 126.76(17).

Figure 2. An IR spectrum of 1 in a KBr disk.

Figure 3. A ¹H NMR spectrum of 1 in chloroform-d₁.

Figure 4. A UV-vis absorption spectrum of 1 (0.060 mM) in H₂O.

The IR spectrum of 1 shows the stretching frequencies of CO coordinated to Ru^I centers at 1950, 1994, and 2041 cm⁻¹ (Figure 2), which are lower than the free CO stretching frequency (2143 cm⁻¹). This weakened CO bond is also caused by the back donation from the low-valent Ru^I to the CO ligand. The ¹H NMR signals observed in the diamagnetic region revealed that 1 is diamagnetic, which originates from an antiferromagnetic exchange interaction between two Ru^I centers through metal–metal bonding (Figure 3). The UV-vis absorption spectrum of 1 in H₂O shows a sharp band at 280 nm (8480 M⁻¹ cm⁻¹) and a broad band at 420 nm (550 M⁻¹ cm⁻¹) (Figure 4).

The bis(μ-formate) Ru^I₂ complex 1 can convert to a (hydride)(formate) species 2 with evolution of CO₂ in DMSO at 80 °C in the presence of 1 equivalent of HCOONa. β-Hydrogen elimination is expected to be assisted by the metal centers to release CO₂ and form the hydride ligand [15]. The ¹H NMR spectrum of 2 in DMSO-d₆ shows hydride- and formate-derived signals at –12.4 and 7.76 ppm, respectively, with the same intensities (Figure 5(a)). The hydride-derived peak is typical of those found in hydride-bridged Ru complexes [16–19]. The dimer structure of 2 with an antiferromagnetic exchange interaction between two Ru^I centers, was suggested by the signals observed in the diamagnetic region of the ¹H NMR spectrum and its ESR silent character.

Figure 5. ¹H NMR spectra of 2 (a) before and (b) after addition of D₂O, in which 2 was obtained from heating of 1 at 80 °C in DMSO-d₆ for 1 h in the presence of 1 equivalent of HCOONa. †: the peak of CH₂Br₂ as an internal standard.

We have confirmed that the hydride ligand of 2 has a protic character rather than hydridic character based on the following investigations. We observed 2 being unreactive toward proton of HCOOH, i.e. dihydrogen gas was not formed via protonation of the hydride ligand, which was confirmed by ¹H NMR spectroscopy and GC analysis. Then, we observed an H⁺/D⁺ exchange of the hydride ligand of 2, as confirmed by ¹H NMR spectroscopy (Figure 5). The intensity of hydride-derived peak of 2 only decreased by the addition of D₂O into DMSO-d₆ solution of 2 (Figure 5), meaning that the hydride ligand underwent the H⁺/D⁺ exchange with D⁺.

Complex 1 is a precursor to catalyze the pH-dependent conversion of HCOOH to H₂ and CO₂ in water at pH 1.0–7.0 (Figures 6–10 and Table 1). H₂ and CO₂ gases were detected by GC. No nanoparticles were formed in the catalytic reaction, which was confirmed by dynamic light scattering measurements. Figure 6 shows the time-dependent profile of turnover numbers (TONs, mol of H₂ evolved/mol of 1) of H₂ evolution in the reaction of 1 with an excess amount of HCOOH in water at 80 °C. The pH-dependent TON shows a maximum around pH 3.5 (Figure 7). We investigated the dependence of initial rate for H₂ production against the concentration of 1 (0.63–5.0 mM). This linear correlation clearly indicates that the catalytic reaction is first-order against the catalyst concentration (Figure 8).

Figure 6. A plot of turnover numbers (TONs) vs. time for the H₂ evolution from HCOOH catalyzed by 1 (1.25 mM) with HCOOH (2.60 M) in H₂O at 80 °C and pH 3.5.

Figure 7. A pH-dependent H₂ evolution from HCOOH catalyzed by 1 (1.25 mM) with HCOOH (2.60 M) in H₂O for 1 h at 80 °C. The maximum turnover number (TON) is 13.1 at pH 3.5.

Figure 8. A plot of initial rate of H₂ production against the concentration of 1 (0.63–5.0 mM) in the reaction of 1 with HCOOH (2.60 M) at 80 °C and pH 3.5. The initial rate of H₂ production was determined for 300 s.

Figure 9. A proposed reaction mechanism for the conversion of HCOOH to H₂ and CO₂ catalyzed by dinuclear Ru complexes in H₂O.

Figure 10. A proposed reaction mechanism for the hydrogen isotope exchange reaction in the conversion of HCOOD to H₂, HD, D₂, and CO₂ in D₂O catalyzed by dinuclear Ru complexes (entry 3 of Table 1).

Table 1. Dehydrogenation of HCOOH, DCOOH, HCOOD, or DCOOD in H₂O or D₂O by using 1 as a precursor^a

Entry	Substrate	Solvent	TON^b	Ratio of H2: HD: D2 (%)
				H2	HD	D2
1	HCOOH	H2O	13.1	100	0	0
2	DCOOH	H2O	5.2	7	93	0
3	HCOOD^c	D2O	6.7	1	91	8
4	DCOOD^c	D2O	2.4	0	2	98

^a Reaction conditions: 1 (1.25 mM) and HCOOH (2.60 M), DCOOH (2.60 M), or DCOOD (2.60 M) in H₂O or D₂O at 80 °C at pH or pD 3.5.

^b Turnover number: mol of evolved dihydrogen gases (H₂, HD, and D₂)/mol of 1.

^c HCOOD and DCOOD contain 2% H atom.

In order to investigate the reaction mechanism of the dehydrogenation of HCOOH, deuterium isotope-labeling experiments were conducted as shown in Table 1. Entries 2 (DCOOH in H₂O at pH 3.5) and 3 (HCOOD in D₂O at pD 3.5) show the TONs of 5.2 and 6.7, respectively, which are lower than that of 13.1 in the entry 1 (HCOOH in H₂O at pH 3.5). These results suggest that the cleavage of C–H bond in HCOOH and O–H bond in H₂O should be involved in the rate-determining step. The lowest TON value of 2.4 in entry 4 (DCOOD in D₂O at pD 3.5) can be expected by the results of entries 2 and 3. According to the results of entries 1–4, we propose the reaction mechanism shown in Figure 9. The bis(μ-formate) complex 1 is converted to the (hydride)(formate) species 2 with the release of CO₂. The subsequent decarboxylation of 2 is expected to result in formation of a dihydride species A, which can be proposed owing to the previously reported analogous structures [20–24]. Then reductive elimination of H₂ can yield a low-valent species B. Protonation of B would provide a monohydride species C, which can bind HCOO^– to return to 2. This proposed mechanism involves protonation step of B and binding step of HCOO^–, with a pK_a of 3.75, to C, which causes a maximum TON around pH 3.5 because acidic conditions facilitate the protonation step and basic conditions deprotonating HCOOH facilitate the binding step of HCOO^– to C.

We also investigated the hydrogen isotope exchange reaction by observation of evolved dihydrogen gases in Table 1. The entry 2 shows H₂ (7%) and HD (93%) are evolved in the reaction of 1 with DCOOH in H₂O. This result clearly indicates that dihydrogen gases (H₂ and HD) originate from H of H₂O and D of DCOOH. The evolved H₂ is a key product to elucidate the reaction mechanism since H₂ comes from only H⁺ in H₂O, which enables us to exclude the possibility of protonation of the hydride ligand to evolve dihydrogen gas. The same is true of entry 3. In this context, we can propose the reaction mechanism of the hydrogen isotope exchange reaction, involving the H⁺/D⁺ exchange of the protic hydride ligand, causing reductive elimination of dihydrogen gas, as shown in Figure 10 [25].

4. Conclusions

Isotope-labeling experiments suggest that the formation of dihydrogen gas from formic acid takes place via H⁺/D⁺ exchange of the protic hydride ligands and reductive elimination from the protic hydride ligands of the dinuclear Ru complex in H₂O/D₂O.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This work was supported by Japan Society for the Promotion of Science (JSPS), [grant numbers JP26000008 (Specially Promoted Research); JP16K05727; JP15K05566]; World Premier International Research Centre Initiative (WPI), Japan.

Supplemental data

The supplemental material for this paper is available online at https://doi.org/10.1080/14686996.2017.1379857