Formic acid dehydrogenation catalysed by a novel amino-di(N-heterocyclic carbene) based Ru-CNC pincer complex

Abstract

A new Ru(II) complex featuring a novel amino-di(N-heterocyclic carbene) CNC pincer ligand, ^iPrCNC-RuCl₂(CO) (Ru-1), has been developed and characterised in depth. Ru-1 forms an efficient and durable catalytic formic acid dehydrogenation system in combination with the ionic liquid 1-ethyl-3-methylimidazolium diethylphosphate (EMIM PO₂(OEt)₂).

---

Formic acid (FA) is considered a promising candidate for the long-term, safe, and practical chemical storage of hydrogen, serving as a platform compound to connect renewable energy and hydrogen fuel cells.^1–3 Formic acid dehydrogenation (FADH),⁴i.e., the transformation of FA to H₂ and CO₂, has been demonstrated under mild conditions by numerous groups using a combination of transition metal (pre)catalyst, solvent, and additive.^5–11 For instance, using a Ru-PNP (PNP = phosphorous–nitrogen–phosphorous) pincer catalyst, in N,N′-dimethyl formamide (DMF) with amine bases, Pidko was able to achieve FADH with TON values of up to 706 500 at 90 °C.⁶ A related mono Ir-complex was reported by Himeda to similarly facilitate FADH of an aqueous solution of FA with TON values of up to 10 000 000.⁸ Sponholz, Junge, and Beller used the triazine-based Kempe Mn-PNP complex⁹ for FADH in the presence of stoichiometric amounts of lysine in a 1 : 1 H₂O/THF mixture, achieving total TON values of 600 000 at 90 °C.¹⁰ Milstein showed the first example of a homogeneous catalytic system capable of conducting FADH in neat FA.¹¹

Apart from the latter example, many systems utilise (volatile) organic solvents in combination with additives, such as amine bases. While achieving high TONs, the presence of any volatile components necessitates further purification of the generated gas-mixtures. In this context, ionic liquids (ILs) represent a non-volatile alternative media for catalytic FADH,^12,13 and we recently started exploring their solvent properties for Ru-PNP^13a and Ru-POP^13b (POP = phosphorous–oxygen–phosphorous) catalysed FADH. ILs are a wide family of salts characteristic of a low melting point, a negligible vapor pressure, and a high chemical and thermal stability. Thus, using the commercially available Ru-MACHO-BH as catalyst in combination with the IL 1-butyl-3-methylimidazolium acetate (BMIM OAc) for FADH reached a TON exceeding 18 million and a durability of more than 110 days. Moreover, a large volume ratio between FA and IL exceeding 3600 was achieved, a desirable feature for application. When switching to Ru-POP pincer congeners, similar catalytic activities are observed for the short-term batch FADH systems.^13b However, their durability is inferior, showing catalyst deactivation within several hours.

N-Heterocyclic carbenes (NHCs) are a group of ligands that are analogous to phosphines but exhibit certain desirable properties, such as the formation of thermodynamically stable metal–ligand bonds, which suggest that replacing phosphine groups with NHCs may lead to complexes demonstrating enhanced stability.^14–17 Especially pincer ligands bearing an amino-bridgehead are particularly interesting, as the bifunctionality of the amino-moiety renders metal–ligand cooperativity (MLC) possible. While a few other metal complexes bearing an (aliphatic) amino-based CNC (CNC = carbon (NHC)–nitrogen–carbon (NHC)) pincer ligand exist,^18–22 only two Ru-based complexes have been published by Pidko²³ and Masahiro and Yusuke,²⁴ both demonstrating high catalytic activity in carbonyl hydrogenation at low catalyst loadings.

Herein, we present the synthesis and characterisation of a Ru(II)-carbonyl dichlorido complex, ^iPrCNC-RuCl₂(CO) (Ru-1), bearing an amino-di(NHC) pincer ligand. We demonstrate that Ru-1 in the IL 1-ethyl-3-methylimidazolium diethylphosphate (EMIM PO₂(OEt)₂) produces a highly active (and stable) catalytic system that enables full conversion of 200 equivalents of FA to CO₂ and H₂ within an hour at 100 °C, while also being tolerant towards exposure to air and multiple FA recharge additions without losing its catalytic activity. Furthermore, we show that the IL EMIM PO₂(OEt)₂ is a superior reaction medium to the acetate-based IL 1-ethyl-3-methylimidazolium acetate (EMIM OAc) for Ru-1 catalysed FADH.

Starting from N-benzyl-bis(2-chloroethyl) amine hydrochloride, [^iPrCNC][Cl]₂ is isolable in excellent yield (90%, Scheme 1). Transmetalation of [^iPrCNC][Cl]₂via an Ag-masked carbene with RuCl₂(CO)(PPh₃)₂(dmf)²⁵ in THF affords Ru-1 as yellow crystals in an acceptable yield of up to 40%. The NHC precursor was obtained by reacting [^iPrCNC][Cl]₂ to Ag₂O in DCM, which yields a light-sensitive orange/brown material that features a ¹H nuclear magnetic resonance (NMR) and ¹³C-NMR spectra consistent with a heteroleptic NHC-Ag–Cl species, see Fig. S5–S7 (ESI†).¹⁹Ru-1 was characterised in the solid state by single-crystal X-ray diffraction (SC-XRD) and infrared-spectroscopy (IR), and in solution by high-resolution mass-spectrometry (HRMS) as well as 1D and 2D multinuclear NMR spectroscopy, including ¹H, ¹³C, ³¹P, {¹H–¹H} COSY, as well as {¹H–¹³C} HSQC and HMBC. Ru-1 crystallises in the monoclinic P2₁/c space group and exhibits a slightly distorted octahedral coordination environment of Ru, as depicted in Scheme 1. The CNC pincer ligand coordinates in a meridional fashion and features a carbonyl ligand trans to the amino-bridgehead. The Ru–C bond lengths belonging to the two NHCs are 2.096(3) and 2.105(3) Å, respectively, and the Ru–C bond length of Ru–CO is 1.817(4) Å. These bond lengths are comparable to those reported by Masahiro and Yusuke.²⁴ Solution-state ¹H and ¹³C NMR of Ru-1, Fig. S9–S14 (ESI†), are consistent with the solid-state structure. In the ¹³C-NMR spectrum, the two most downfield-shifted signals are consistent with the C atom of the carbonyl ligand (208.8 ppm) and the C₂ atom of the NHCs (184.0 ppm). {¹H–¹³C} HMBC delineates these two signals from one another, as only the latter signal demonstrates a cross-peak with the imdazole-2-ylidene C₄/C₅-protons. Importantly, no residual triphenylphosphine is observed via³¹P-NMR spectroscopy.

Scheme 1 Synthesis and solid-state of Ru-1. Thermal ellipsoid set at 50% probability level. All H atoms but the amino-bridgehead have been omitted for clarity. Colour scheme: H white, C grey, O red, N blue, Cl dark green, Ru dark blue. Selected interatomic distances (Å) and angles (°): Ru–C1 1.817(4); Ru–C2 2.105(3); Ru–N 2.247(3); Ru–C3 2.096(3); Ru–Cl1 2.4188(7); Ru–Cl2 2.4319(7); C1–O 1.149(5); C1–Ru–N 173.62(14); C2–Ru–C3 173.21(11); Cl1–Ru–Cl2: 175.35(3).

With Ru-1 in hand, we investigated its potential use as a (pre)catalyst for FADH. While Ru-1 does not catalyse FADH in neat FA (Table 1, entries 1 and 2), using EMIM OAc as solvent afforded 83% FA conversion (TOF_max = 670 h⁻¹) at 80 °C after 3 hours with a low loading of Ru-1 (0.1 mol%, entry 3). Increasing the temperature to 90 °C and to 100 °C, afforded conversions of 91% and 94%, respectively, and significantly enhanced TOF_max values of 1480 h⁻¹ and 2650 h⁻¹, respectively (entries 4 and 5). As depicted in Fig. S15 (ESI†), FADH conversion with Ru-1 in EMIM OAc operating at temperatures below 100 °C clearly has an induction phase. Despite this induction period, Ru-1 continues to effectively catalyse FADH even when the loading is reduced from 0.1 mol% to 0.05 mol% and further to 0.025 mol%, resulting in 86% and 73% conversion after 3 hours, respectively, at 90 °C (entries 6 and 7). Despite the incomplete FA consumption within the set time frame, high TON (2940) and TOF_max (2010 h⁻¹) values were recorded. The same behaviour was observed at 100 °C when decreasing the loading to 0.5 mol%, with TON = 1800 and TOF_max = 3500 h⁻¹ (entry 8).

Table 1 Optimisation of reaction parameters for Ru-1 catalysed FADH in EMIM OAc^a

Entry	IL	Ru-1 [mol%]	Temp. [°C]	Conv.^b [%]	TON	TOFmax [h^−1]
a Standard reaction conditions: Ru-1 (0.025–0.1 mol%), IL (1 mL), FA (0.5 mL, 13.25 mmol), 3 h under gentle flow of Ar. Gas composition is analysed by GC-TCD. b Determined by ^1H-NMR.
1	None	0.1	80	<5	—	—
2	None	0.1	90	<5	—	—
3	EMIM OAc	0.1	80	83	830	670
4	EMIM OAc	0.1	90	91	910	1480
5	EMIM OAc	0.1	100	94	930	2650
6	EMIM OAc	0.05	90	86	1730	1630
7	EMIM OAc	0.025	90	73	2940	2010
8	EMIM OAc	0.05	100	90	1800	3500

---

Contrasting our previous work on FADH where acetate-based ILs were employed with various catalysts,¹³ using EMIM OAc in combination with Ru-1 does not afford complete conversion within a reasonable timeframe such as 3 hours. We speculate that this is due to Ru-1 forming off-cyclic acetato-adducts in acetate-based ILs when FA concentration is sufficiently low (vide infra). Hence, we turned our attention to EMIM PO₂(OEt)₂, and pleasingly observed full FA conversion at 90 °C, see Table 2, entry 3. However, as evident in the TOF_max values, there is a marked difference in the initial rate between the two ILs, see Fig. S16 (ESI†).

Table 2 Ionic liquid screening for FADH optimisation with Ru-1^a

Entry	IL	Ru-1 [mol%]	Temp. [°C]	Conv.^b [%]	TON	TOFmax [h^−1]
a Standard reaction conditions: Ru-1 (0.05–0.1 mol%), IL (1 mL), FA (0.5 mL, 13.25 mmol), 3 h under gentle flow of Ar. Gas composition is analysed by GC-TCD. b Determined by ^1H-NMR.
1	EMIM OAc	0.1	90	91	910	1480
2	EMIM OAc	0.05	90	86	1730	1630
3	EMIM PO2(OEt)2	0.1	90	>99	1000	400
4	EMIM PO2(OEt)2	0.1	100	>99	1000	1100
5	EMIM PO2(OEt)2	0.05	100	>99	2000	820

---

Subsequent screening of reaction conditions of the phosphate-based IL system revealed that at 100 °C, full conversion was achieved within one hour resulting in a TOF_max of 1100 h⁻¹ (entry 4). Lowering the loading of Ru-1 to 0.05 mol% still achieved full conversion albeit only after 3 hours, providing a TON of 2000 and TOF_max of 820 h⁻¹ (entry 5).

We then studied the durability of Ru-1 over multiple FADH cycles with various loadings of FA. Successively supplying the setup with FA still led to full conversion after six cycles, after which no more cycles were attempted (see Fig. 1 and Table S21, ESI†). Furthermore, leaving the reaction medium standing for three days between cycles 2 and 3 has practically no influence on its catalytic performance. In addition, in cycle 4 the amount of supplied FA was doubled which considerably slowed down the FADH activity in the first hour (approximately 20% conversion) but nevertheless led to 85% conversion after 3 hours and full conversion after 4 hours. Hence, there seems to be a saturation point of FA in EMIM PO₂(OEt)₂ after which the beneficial effect of the IL on the catalytic activity of activated Ru-1 diminishes significantly but can be regained once sufficient FA has been converted (at a slower rate). Notably, full FA conversion is repeatedly reached in the following two cycles as well (cycle 5 and 6), and all six cycles proceed with similar TOF_max values (490–650 h⁻¹), indicating that the combination of Ru-1 and EMIM PO₂(OEt)₂ comprises a stable and durable catalytic system for FADH.

Fig. 1 FA conversion over time in batch reactions conducted in EMIM PO₂(OEt)₂ using Ru-1 as (pre)catalysts. In cycle 4 an excess of FA is purposefully added to the system. It shows a similar conversion to the first cycle.

In situ HR-MS and ¹H-NMR spectroscopy (Fig. S17–S19, ESI†) show early formation of a formato species in cycle 1 during the slow onset of catalysis and the appearance of a hydrido species after two cycles. This is in line with the expectation that Ru-1 is a precatalyst, the formato complex an off-cycle species, and the hydrido complex the resting state. Ru-1 is proposed to turn into the amido congener by HCl elimination, likely facilitated by the IL anion. In the same manner, the formato complex turns into the amido congener by FA elimination. Furthermore, FA dehydrogenation is proposed to occur on the amido complex via an outer sphere mechanism. Moreover, EMIM OAc seems more efficient than EMIM PO₂(OEt)₂ at generating the amido species, but also coordinates to the catalyst, a particular obstacle at low FA concentration. On the other hand, the use of phosphate-based ILs might be superior to acetate-based ones due to a lower propensity to compete with formato complex formation. This also explains the slow FADH with doubled FA addition in EMIM PO₂(OEt)₂.

Having demonstrated the resilience of Ru-1 and EMIM PO₂(OEt)₂, we sought to assess the influence of the CNC pincer ligand. To this end, we carried out the successive cycles using two related Ru(II) carbonyl complexes typically considered as simple metal precursors, i.e., RuCl₂(CO)(PPh₃)₂(dmf) and RuHCl(CO)(PPh₃)₃.

Fig. 2 shows the FA conversion for four of six successive cycles, comparing Ru-1 to the two precursors where the inserts A, B, C, and D, show the first, second, fifth, and sixth batch cycle, respectively. See Fig. S25–S29 and Tables S15–S34 (ESI†) for relevant data on gas composition, comparison of FADH efficacy of the individual precursor over six batches as well as compared with Ru-1, and listings of TON/TOF values. Surprisingly, in the first cycle (insert A), both precursors, but particularly RuCl₂(CO)(PPh₃)₂(dmf), demonstrated markedly better catalytic activity than Ru-1. Interestingly, in situ¹H NMR analysis before complete FA consumption (45 min reaction time) revealed the formation of several hydrido species in the two precursor systems, with one species appearing to be dominant in both systems, see Fig. S18a (ESI†). In addition, in situ HR-MS analysis suggests the incorporation of an EMIM-based NHC-ligand into the ligand system of both precursors, see Fig. S19a–d (ESI†). However, these catalytically active species seem to degrade over time, possibly due to an instability at elevated temperature in the absence of FA. This is also evident from the second FADH cycle (insert B), where RuHCl(CO)(PPh₃)₃ exhibited poor catalytic conversion and RuCl₂(CO)(PPh₃)₂(dmf) has become similar to Ru-1 in activity. Furthermore, oversaturating the systems with FA in cycle 4 leads to much lower conversions in cycles 5 and 6 (inserts C and D, respectively), demonstrating that the precursors are significantly inferior to Ru-1. These observations support that the use of a CNC pincer ligand imposes stability for FADH catalysis in EMIM PO₂(OEt)₂.

Fig. 2 FA conversion over time in batch reactions conducted in EMIM PO₂(OEt)₂ comparing Ru-1, RuCl₂(CO)(PPh₃)₂(dmf), and RuHCl(CO)(PPh₃)₃ as (pre)catalysts. Insert A and B show the first and second batch cycles, respectively. Insert C and D show the fifth and sixth cycles, which correspond to the first and second cycle after the introduction of an excess of FA.

In conclusion, we have developed a novel amino-based di(NHC) Ru pincer complex, Ru-1, and fully characterised it by a range of analytical techniques, including SC-XRD, IR, NMR, and MS. Ru-1 in combination with the IL EMIM PO₂(OEt)₂ results in a catalytic system that enables efficient and durable FADH under mild conditions, producing only H₂ and CO₂ gas. Additionally, the system can withstand exposure to an excess of FA without incurring significant loss in catalytic capabilities over several cycles, different from two Ru(II) precursors. In situ analysis of the catalytic reactions reveals the appearance of formato and hydrido species, indicating that Ru-1 undergoes initial activation, suggesting that phosphate-based ILs might be better than acetate-based ones due to a reduced tendency to generating off-cycle complexation with the catalyst.

The authors are grateful to the Independent Research Fund Denmark (8102-00330B and 1113-00027B), VILLUM FONDEN (19049 and 53069), and Carlsberg Foundation (CF20-0365) for generous funding. P. A. Thanks the Erasmus+ programme of the European Union (2023-1-DE01-KA131-HED-000120711). L. P. Thanks Otto Mønsted (22-55-0557), Brødrene Hartmanns Foundation (A37.412), and DTU Discovery Grant.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

References

1. C. Fellay, P. J. Dyson and G. Laurenczy, Angew. Chem., Int. Ed., 2008, 47, 3966–3968.
2. B. Loges, A. Boddien, H. Junge and M. Beller, Angew. Chem., Int. Ed., 2008, 47, 3962–3965.
3. C. Guan, Y. Pan, T. Zhang, M. J. Ajitha and K.-W. Huang, Chem. – Asian J., 2020, 15, 937–946.
4. (a) K. Müller, K. Brooks and T. Autrey, Energy Fuels, 2017, 31, 12603–12611; (b) J. Guo, C. K. Yin, D. L. Zhong, Y. L. Wang, T. Qi, G. H. Liu, L. T. Shen, Q. S. Zhou, Z. H. Peng, H. Yao and X. B. Li, ChemSusChem, 2021, 14, 2655–2681; (c) D. Mellmann, P. Sponholz, H. Junge and M. Beller, Chem. Soc. Rev., 2016, 45, 3954–3988; (d) A. Wang, P. He, J. Wu, N. Chen, C. Pan, E. Shi, H. Jia, T. Hu, K. He, Q. Cai and R. Shen, Energy Fuels, 2023, 37, 17075–17093; (e) L. Piccirilli, D. L. J. Pinheiro and M. Nielsen, Catalysts, 2020, 10, 773; (f) S. Patra, B. Maji, H. Kawanami and Y. Himeda, RSC Sustainability, 2023, 1, 1655–1671; (g) M. H. G. Prechtl and S. Sahler, Curr. Org. Chem., 2013, 17, 220–228; (h) K. Sordakis, C. Tang, L. K. Vogt, H. Junge, P. J. Dyson, M. Beller and G. Laurenczy, Chem. Rev., 2018, 118, 372–433.
5. E. A. Bielinski, P. O. Lagaditis, Y. Zhang, B. Q. Mercado, C. Würtele, W. H. Bernskoetter, N. Hazari and S. Schneider, J. Am. Chem. Soc., 2014, 136, 10234–10237.
6. G. A. Filonenko, R. van Putten, E. N. Schulpen, E. J. M. Hensen and E. A. Pidko, ChemCatChem, 2014, 6, 1526–1530.
7. J. F. Hull, Y. Himeda, W.-H. Wang, B. Hashiguchi, R. Periana, D. J. Szalda, J. T. Muckerman and E. Fujita, Nat. Chem., 2012, 4, 383–388.
8. N. Onishi, R. Kanega, E. Fujita and Y. Himeda, Adv. Synth. Catal., 2019, 361, 289–296.
9. S. Rösler, M. Ertl, T. Irrgang and R. Kempe, Angew. Chem., Int. Ed., 2015, 54, 15046–15050.
10. D. Wei, R. Sang, P. Sponholz, H. Junge and M. Beller, Nat. Energy, 2022, 7, 438–447.
11. S. Kar, M. Rauch, G. Leitus, Y. Ben-David and D. Milstein, Nat. Catal., 2021, 4, 193–201.
12. (a) A. Moazezbarabadi, D. Wei, H. Junge and M. Beller, ChemSusChem, 2022, 15, e202201502; (b) J. D. Scholten, M. H. G. Prechtl and J. Dupont, ChemCatChem, 2010, 2, 1177–1327; (c) X. Li, X. Ma, F. Shi and Y. Deng, ChemSusChem, 2010, 3, 71–74.
13. (a) L. Piccirilli, B. Rabell, R. Padilla, A. Riisager, S. Das and M. Nielsen, J. Am. Chem. Soc., 2023, 145, 5655–5663; (b) A. T. Nikol, B. Rabell, M. S. B. Jørgensen, R. W. Larsen and M. Nielsen, Sci. Rep., 2024, 14, 26209.
14. W. A. Herrmann, Angew. Chem., Int. Ed., 2002, 41, 1290–1309.
15. F. E. Hahn and M. C. Jahnke, Angew. Chem., Int. Ed., 2008, 47, 3122–3172.
16. H. Jacobsen, A. Correa, A. Poater, C. Costabile and L. Cavallo, Coord. Chem. Rev., 2009, 253, 687–703.
17. S. Díez-González, N. Marion and S. P. Nolan, Chem. Rev., 2009, 109, 3612–3676.
18. R. Zhong, Z. Wei, W. Zhang, S. Liu and Q. Liu, Chem, 2019, 5, 1552–1566.
19. J. Houghton, G. Dyson, R. E. Douthwaite, A. C. Whitwood and B. M. Kariuki, Dalton Trans., 2007, 3065–3073.
20. C.-F. Yang, T. Lu, X.-T. Chen and Z.-L. Xue, Chem. Commun., 2018, 54, 7830–7833.
21. W. B. Cross, C. G. Daly, R. L. Ackerman, I. R. George and K. Singh, Dalton Trans., 2011, 40, 495–505.
22. W. Wei, Y. Qin, M. Luo, P. Xia and M. S. Wong, Organometallics, 2008, 27, 2268–2272.
23. G. A. Filonenko, M. J. B. Aguila, E. N. Schulpen, R. van Putten, J. Wiecko, C. Müller, L. Lefort, E. J. M. Hensen and E. A. Pidko, J. Am. Chem. Soc., 2015, 137, 7620–7623.
24. M. Yusuke and T. Masahiro, US Pat., US2016145282A1, 2016.
25. V. Gómez-Benitez, J. Olvera-Mancilla, S. Hernández-Ortega and D. Morales-Morales, J. Mol. Struct., 2004, 689, 137–141.

---

Footnote

† Electronic supplementary information (ESI) available. CCDC 2387750. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4cc05164e

---