3-Methylindole-Based Tripodal Tetraphosphine Ruthenium Complexes in N₂ Coordination and Reduction and Formic Acid Dehydrogenation

Abstract

The ruthenium(II) complexes RuCl₂L1^H, RuCl₂L1^CF3, RuCl₂L1^OMe and RuCl₂L2^H were synthesized from [Ru(η⁶-benzene)Cl(μ-Cl)]₂ and the corresponding tripodal tris-3-methylindolephosphine-based ligands L1^H, L1^CF3, L1^OMe, and L2^H. Stoichiometric reduction of these complexes with KC₈ yielded the corresponding ruthenium(0) dinitrogen complexes. The latter complexes were studied in the N₂ reduction with chlorosilanes and KC₈, yielding stoichiometric amounts of the silylamines. The synthesized ruthenium(II) complexes are also active catalysts for the formic acid dehydrogenation reaction.

1. Introduction

Electrocatalytic reduction reactions that store energy by the conversion of H⁺, CO₂, or N₂, are important in the context of the transformation to a society based on sustainable energy. Therefore, the development of efficient catalysts that enable these reactions is needed. Homogeneous tripodal ligand transition metal complexes have shown beneficial reactivity over mono- or bidentate ligands in many organic transformations [1,2,3]. In recent years, tripodal ligands coordinated to molybdenum, iron, and cobalt were shown to lead to complexes that are capable of the reduction of N₂ to ammonia or silylamines [4,5,6,7,8,9,10]. Also, reversible hydrogen storage in CO₂ was shown to be possible with iron and cobalt based on tripodal ligands [11,12,13]. Previously, we reported the coordination of ligand L1^H (Figure 1) to rhodium [14,15] and ruthenium [16]. This ligand, which features three 1-(3-methylindolyl)diphenylphosphine groups tethered to a central phosphorus atom, stabilizes ruthenium in the oxidation states +II, +I, and 0. The isolated low oxidation state Ru^IClL1^H and Ru⁰(N₂)L1^H complexes revealed interesting one-electron reactivity, as they were able to abstract chlorine atoms from either chloroform or dichloromethane, leading to formation of the corresponding Ru^II complex. In this paper, we expand the coordination chemistry of ruthenium and 3-methylindole-based ligands [17,18,19,20,21,22,23] and investigate how modification of the ligand scaffold and its electronic properties influence the stability and activity of these complexes in dinitrogen coordination and subsequent reduction as well as in formic acid dehydrogenation. Ligand L1^H was modified using functionalized P(Ar)₂ units, wherein the arene carries either electron-withdrawing –CF₃ groups (L1^CF3) or electron-donating –OMe groups (L1^OMe) in the para-position (Figure 1). Additionally, the known structural isomer, L2^H (Figure 2) [14], in which the central P-donor is bound to the three nitrogen atoms of the 2-(3-methylindolyl)diphenylphosphine moiety, was also used in these studies.

2. Results

Ligands L1^H and L2^H were prepared in two steps from 3-methylindole as previously reported by our group [14,15]. The new ligands L1^CF3 and L1^OMe were prepared in a similar way to L1^H, by reacting tris-2-(3-methylindolyl)phosphine [17] with the corresponding chlorodiarylphosphine (Figure 1).

The reaction of L1^H with [Ru(η⁶-benzene)Cl(μ-Cl)]₂ in refluxing THF for 16 h led to the formation of the octahedral complex RuCl₂L1^H in quantitative yield (Figure 2). Surprisingly, the other ligands did not react with this ruthenium precursor in refluxing THF. Even after 36 h, only trace amounts of the target complexes were observed using ex-situ ³¹P NMR spectroscopy, and mainly signals of the unreacted ligand were present. Reactions in refluxing THF/toluene (1:3) mixture at 120 °C for a prolonged time of 64 h did lead to the formation of complexes RuCl₂L1^CF3, RuCl₂L1^OMe, and RuCl₂L2^H in >50% isolated yields.

All three L1-based RuCl₂L1 complexes show similar ³¹P NMR spectra: three signals in a ratio 1(td):2(t):1(td) (Figure 3). This splitting pattern indicates the formation of a symmetric complex. Ruthenium(II) complexes tend to form octahedral 18-valence electron complexes, and X-ray analysis and NMR spectroscopy confirmed that complex RuCl₂L1^H indeed features an octahedral geometry with the two chlorides in mutual cis position and all four phosphines bound to the metal center [15]. The above factors and the similarity of the ³¹P NMR spectra of the other L1-based complexes point to the same octahedral geometry around the metal center.

Layering of a dichloromethane solution of RuCl₂L1^CF3 with pentane resulted in the formation of crystals suitable for single-crystal X-ray diffraction. The crystal structure (Figure 4) of RuCl₂L1^CF3 indeed shows an overall octahedral geometry with the chlorido ligands in cis position, in accordance with the observed ³¹P NMR splitting pattern. The P1–Ru–P2 angle in RuCl₂L1^CF3 (157.78(6)°) is comparable to the one previously reported for RuCl₂L1^H (160.04(3)°) (

Table 1. Selected bond lengths (Å) and angles (°) for RuL complexes.

	RuCl2L1H [a]	RuCl2L1CF3	RuCl2L2H	RuN2L1H	RuN2L1CF3
Ru1–P1	2.3189(9)	2.3097(19)	2.3295(8)	2.2747(12)	2.2613(10)
Ru1–P2	2.3727(9)	2.3445(19)	2.4079(8)	2.2752(11)	2.2702(10)
Ru1–P3	2.2671(9)	2.2383(17)	2.2913(8)	2.2774(11)	2.2554(10)
Ru1–P4	2.1932(9)	2.2023(17)	2.1363(8)	2.2133(11)	2.2193(9)
Ru1–Cl1	2.4869(9)	2.4869(16)	2.4829(7)
Ru1–Cl2	2.4471(9)	2.4487(16)	2.4705(7)
Ru1–N1				2.011(4)	2.066(3)
N1–N2				1.085(5)	1.064(5)
P1–Ru1–P2	160.04(3)	157.78(6)	159.24(3)	122.85(4)	119.79(4)
P1–Ru1–P3	102.23(3)	101.39(7)	101.30(3)	115.80(4)	119.28(4)
P2–Ru1–P3	94.27(3)	97.42(7)	93.60(3)	118.33(4)	118.16(4)
P1–Ru1–P4	87.06(3)	86.13(6)	86.13(3)	84.14(4)	84.54(4)
P2–Ru1–P4	82.67(3)	83.12(6)	80.63(3)	83.91(4)	84.13(4)
P3–Ru1–P4	86.85(3)	87.02(6)	85.62(3)	84.63(4)	84.73(3)

^[a] Taken from reference [16].

). The P1–Ru, P2–Ru and P3–Ru bond lengths are all slightly smaller in RuCl₂L1^CF3 (2.3097(19) Å, 2.3445(19) Å and 2.2383(17) Å, respectively) compared with RuCl₂L1^H (2.3189(9) Å, 2.3727(9) Å and 2.2671(9) Å). The –CF₃ groups clearly point away from the metal center, so steric effects do not play a significant role in the coordination chemistry. The C_s symmetry observed in solution by ³¹P NMR spectroscopy for both RuCl₂L1^H and RuCl₂L1^CF3 is not present in the solid state, as the unit cell contains two rotamers. This becomes evident when the symmetry-equivalent molecules in the crystal structure of RuCl₂L1^CF3 are viewed from the bottom: the indolyl moieties either all point clockwise, or counter-clockwise (Figure 5, showing the bottom view of RuCl₂L1^CF3 and a schematic representation). In solution, these two rotamers seem to rapidly interchange on the NMR timescale, resulting in one signal for the two mutually trans phosphines. Likely, the methyl groups of the indolyl moiety do not create enough steric repulsion to suppress this process. Despite several attempts, no crystals suitable for X-ray diffraction of the RuCl₂L1^OMe complex were obtained.

Interestingly, complex RuCl₂L2^H, which is based on the linkage isomer of L1^H, ligand L2^H, displays a different ³¹P NMR spectrum (Figure 6). The four signals, two triplets of doublets and two doublets of doublets of doublets with an AX pattern, each integrating one phosphine, indicate the formation of a complex with a different symmetry. Crystals obtained by layering of a DCM solution of RuCl₂L2^H with pentane at 5 °C were of sufficient quality for X-ray crystallography. The corresponding structure of complex RuCl₂L2^H (Figure 7) helps to understand the origin of the AX pattern, as the bottom view displays much more steric hindrance around the central phosphorus atom P4 (vide infra).

However, different than the L1-based complexes, where the methyl groups are pointing towards the central phosphine (Figure 5), in complex RuCl₂L2^H, the bulkier phenyl rings of the indolyl moiety point towards the central phosphine (Figure 8). From the crystal structure, it appears that these phenyl rings should encounter more hindrance toward inversion and it therefore seems likely that this process is associated with a high energy barrier in solution. As a result, the two phosphine atoms (P1 and P2, Figure 7) in trans position become inequivalent, leading to the observed AX pattern in the ³¹P NMR spectrum as a result of the P1–P2 coupling.

The crystal structure of RuCl₂L2^H further reveals minor differences in the coordination around the ruthenium center compared with RuCl₂L1^H (Table 1). The P1–Ru, P2–Ru, and P3–Ru bond lengths become slightly larger (2.3295(8) Å, 2.4079(8) and 2.2913(8) Å), but the P4–Ru bond (2.1363(8) Å) shortens, which is in line with a more π-acidic character of the tris(amido)phosphine P4 atom.

Two-electron reduction of RuCl₂L1^H under an N₂ atmosphere led to the formation of the RuN₂L1^H complex, which features a characteristic infrared band corresponding to the coordinated N₂ at 2125 cm⁻¹ (

Table 2. IR frequency of the selected Ru⁰(N₂)L complexes.

Complex	νN2 (cm−1)
RuN2L1CF3	2136
RuN2L1H	2125
RuN2L1OMe	2113
RuN2L2H	2136

) [15]. To explore if the electronic properties of the ligands L1^CF3, L1^OMe, and L2^H have an influence on the binding of N₂ to Ru⁰, the corresponding N₂ complexes were generated in situ by reacting the ruthenium(+II) complexes with two equivalents of KC₈ (Figure 9). The extent of activation of the N₂ bond was quantified by measuring the N≡N stretch frequency with infrared spectroscopy, as this is a direct measure for the ligand-dependent π-back-donation. Table 2 shows the N₂ stretch frequencies of the Ru⁰N₂L complexes.

As expected, the complex coordinated with the less electron-rich, CF₃ substituted ligand shows a band with a higher wavenumber in the infrared spectrum (2136 cm⁻¹), as a result of weaker metal-to-N₂ π-back-donation. The complex coordinated to the more electron-rich OMe substituted ligand shows a band at 2113 cm⁻¹, thus revealing stronger π-back-donation to N₂. Clearly, the peripheral phosphines induce measurable electronic effects on the d-orbitals of the ruthenium center; the phosphines cis to the N₂ ligand become stronger σ-donors and weaker π-acceptors in L1^OMe and vice versa for L1^CF3, which changes the binding properties of the N₂ ligand. Going from RuN₂L1^H (2125 cm⁻¹) to the structural isomer RuN₂L2^H (2136 cm⁻¹), the N₂ stretch frequency increases by 11 cm⁻¹. This is likely a result of the more π-acidic pivotal P-atom (atom P4, Figure 10, top) of the tris(amido)phosphine donor ligand, which competes with the N₂ ligand for π-back-donation from the same metal d-orbitals. Thus, substitution of the phosphine trans to the N₂ ligand for a more π-acidic group leads to weaker π-back-donation of Ru to N₂. These results show that the electronic properties of the dinitrogen complexes can be tuned by ligand design, either by manipulating the substitutions on the donor atoms coordinated cis to N₂ or by changing the π-acidity of the atom bound trans to N₂.

Layering of a THF solution of RuN₂L1^CF3 with pentane resulted in the formation of single crystals suitable for X-ray diffraction (Figure 10). Similar to the RuN₂L1^H complex reported previously [15], RuN₂L1^CF3 also shows a trigonal bipyramidal geometry. The N≡N bond is slightly shorter in RuN₂L1^CF3 (1.064(5)°) than in RuN₂L1^H (1.085(5)°) (Table 1), which is in accordance with the observed trend in the activation of the N₂ stretch frequency (Table 2). In solution, these complexes remain trigonal bipyramidal, as judged from the occurrence of a doublet and a quartet in all ³¹P NMR spectra.

The dinitrogen ligand in all of these ruthenium complexes is weakly coordinating as judged by the relatively small IR-shifts compared to that of free N₂ (ν = 2359 cm⁻¹) (Table 2). Preliminary experiments show no dinitrogen reduction to ammonia when complex RuCl₂L1^H was subjected to excess reductant (KC₈) in the presence of acid ({H(OEt₂)₂}{BArF₂₄}). Therefore, we set out to explore the reactivity of the tripodal tetraphosphine ruthenium(II) complexes for dinitrogen reduction with chlorosilanes (Figure 11), which follows a different mechanism [24].

A mixture of the RuCl₂L complexes with 100 equiv of KC₈ and 100 equiv of chlorotrimethylsilane in 10 mL THF at room temperature under an atmosphere of dinitrogen was stirred for 1 day, whereafter an aliquot of the reaction mixture was analyzed by GC. Modest yields of ~1.5 equivalents of (Me₃Si)₃N relative to Ru were detected, regardless of the complex used (

Table 3. Formation of (Me₃Si)₃N with RuCl₂L complexes.

Complex	Equiv (Me3Si)3N [a]
RuCl2L1CF3	0.92 [b]
RuCl2L1H	1.83 [b]
RuCl2L1OMe	1.74 [b]
RuCl2L2H	1.40 [b]
[Ru(η6-benzene)Cl(μ-Cl)]2	1.29 [c]
-	trace

^[a] Conditions: ~33 μmol of the complex was suspended together with KC₈ (100 equiv) in 10 mL THF whereafter the chlorotrimethylsilane (100 equiv) was added. The suspension was stirred for 23 h at room temperature; ^[b] Average of two runs; ^[c] Equivalents per ruthenium atom.

). The use of metallic sodium as reductant or longer reaction times did not increase the overall yield. In the absence of ruthenium only traces of (Me₃Si)₃N could be detected.

To gain more insight on the N₂ activation using these novel ruthenium complexes, the reduction was followed in time for complexes RuCl₂L1^H and RuCl₂L2^H, analyzing aliquots of the reaction mixture every hour. The reaction profile shows an incubation period of 1 h before product formation is observed (Figure 12). Such an incubation period was also reported for various other iron and cobalt systems that were investigated for N₂ reduction under comparable conditions [24,25,26], and which all showed similar activities for the formation of tris(trimethylsilyl)amine, regardless of the initial catalyst structure. It was speculated that the catalytically active species are generated after ligand dissociation from these iron and cobalt pre-catalysts. The exact nature of the reported catalysts could not be elucidated, as no dinitrogen-containing species were detected in these previous studies; however, test experiments pointed to a molecular nature of the catalyst [24,25].

Given that our complexes also show an incubation period and that electronic effects induced by ligand modification do not affect the catalytic activity, we speculate that our Ru-complexes are all pre-catalysts. To further test this hypothesis, we performed a catalytic reaction with [Ru(η⁶-benzene)Cl(μ-Cl)]₂ in the absence of any tripodal indolyl-phosphine ligand under the same reaction conditions. Comparable amounts (1.29 equiv relative to Ru) of silylamine were formed in this control experiment, which indeed suggests ligand dissociation to activate the catalyst. Thus, these ligands cannot prevent metal leaching under the strongly reducing conditions used. Ex-situ ³¹P NMR spectroscopy of the RuCl₂L1^H catalytic system showed the absence of the signals of the ruthenium complex pointing to decomposition of the pre-catalyst under the applied conditions.

Next, we investigated the potential of these Ru-species in the catalytic dehydrogenation of formic acid (Figure 13), given our interest in the reversible storage of H₂ into liquid fuels [27,28,29,30,31,32,33]. For Ru-catalysts, formic acid dehydrogenation is accelerated when more electron-rich ligands are employed, as was shown by Himeda et al., who used a series of bipyridine ligands with various substituents at the para position (–OH, –OMe, –Me, –CO₂H, and –H) [34]. With the various substituted tripodal tetraphosphine ligands L1 and L2 in hand, we anticipated to observe similar effects for the corresponding complexes in the formic acid dehydrogenation. The RuCl₂L complexes are coordinatively saturated, which necessitates chloride abstraction using AgBF₄ to allow for substrate activation. Addition of 1.2 mmol of formic acid under nitrogen to a refluxing solution of a mixture of Ru-precatalyst (8.5 µmol) and AgBF₄ (17 µmol) in 3 mL THF pre-activated for 2 h (initial concentration of formic acid: 0.4 M, substrate to catalyst ratio: 140) led to gas production within 1–5 min. The reactions were monitored volumetrically with a gas burette for one hour (see experimental section for a schematic representation). A typical time profile of the dehydrogenation reaction with catalyst RuCl₂L1^OMe is displayed in Figure 14. The overall results of the different catalysts used in the dehydrogenation of formic acid are presented in

Table 4. Activities of RuCl₂L catalysts in the formic acid decomposition reaction.

Complex	TOF (h−1) [a]
RuCl2L1CF3	124
RuCl2L1H	76
RuCl2L1OMe	121
RuCl2L2H	33
–	– [b]
[{RuCl2(p-cymene)}2]	1540 [c]

^[a] Conditions: ~8.5 µmol preactivated catalyst (2 equiv AgBF₄ RuCl₂L. 2 h), 3 mL THF, 45 μL (1.2 mmol) HCOOH (catalyst to substrate ratio: 1:140), reflux temperature, 1 h reaction time, Turnover frequency (TOF, defined as number of equivalents of H₂ produced by the catalyst in one hour) is the average of two catalytic runs. The gas was collected volumetrically with a burette; ^[b] Conditions: 3 mL THF, 45 μL (1.2 mmol) HCOOH, reflux temperature, 30 min reaction time. ^[c] Conditions: ~6.5 µmol catalyst, 7 mmol ionic liquid, 17.6 mmol HCOOH (catalyst to substrate ratio: 1:2700), 80 °C, 7 h reaction time. Data taken from ref. [35].

. The gas outlet was analyzed using gas chromatography, which confirmed presence of both CO₂ and H₂, while no traces of CO were detected (detection limit of the GC for carbon monoxide was δ = 10 ppm). Under the described conditions, all the substrate is used within less than 2 h. Subsequent addition of aliquots of formic acid allowed for continuation of the reaction without an observed decrease in activity for at least 6 h.

Complexes RuCl₂L1^OMe, RuCl₂L1^H, and RuCl₂L2^H follow the same trend as noticed by Himeda [34], with the most electron-rich complex RuCl₂L1^OMe giving the highest activity (Turnover frequency (TOF) = 121 h⁻¹) followed by RuCl₂L1^H (TOF = 76 h⁻¹) and the least electron-rich RuCl₂L2^H (TOF = 33 h⁻¹). Interestingly, the RuCl₂L1^CF3 complex (TOF = 124 h⁻¹) shows a comparable, even slightly higher activity to the electron-rich RuCl₂L1^OMe complex. The reason for this unusual observation is currently not clear. The TOFs found for the RuCl₂L complexes in the formic acid decomposition reaction are relatively low compared to reported systems based on Ir and Ru. For instance TOF of 487,500 h⁻¹ was obtained with [Cp*Ir(L)Cl]Cl (L = 2,2’-bis-2-imidazoline) in neat HCOOH at 90 °C [36], TOF of 24,000 h⁻¹ in boiling HCOOH/dioxane using [K(dme)₂][RuH(trop₂dad)] (trop₂dad = 1,4-bis(5H-dibenzo[a,d]cyclohepten-5-yl)-1,4-diazabuta-1,3-diene [37], TOF of 4556 h⁻¹ with the more related Ru(H)₂(P4) complexes (P4 = meso-1,1,4,7,10,10-hexaphenyl-1,4,7,10-tetraphosphadecane) in HCOOH/N,N-dimethyloctylamine/propylene carbonate mixtures at 80 °C [38]. However, these initial results show that the RuCl₂L complexes are active in the dehydrogenation of formic acid and that the reactivity can be steered by ligand modification. Further reactivity studies with these complexes should be performed to elucidate the mechanism of the reaction and to find the maximum capacities (stability, temperature, solvent) of these catalysts.

3. Materials and Methods

All reactions were carried out under an atmosphere of nitrogen or argon using standard Schlenk techniques or in the glovebox. Reagents were purchased from commercial suppliers and used without further purification. THF, pentane, hexane, and Et₂O were distilled from sodium benzophenone ketyl. These solvents were degassed using the freeze–pump–thaw method (three cycles) and stored under dinitrogen atmosphere. CH₂Cl₂ was distilled from CaH₂ under dinitrogen. NMR spectra (¹H, ³¹P, and ¹³C{¹H, ³¹P}) were measured on a Bruker DRX 500, Bruker AV 400, Bruker DRX 300, or on a Bruker AV 300 spectrometer (Bruker, Billerica, MA, USA). IR spectra (ATR mode) were recorded with a Bruker Alpha-p FT-IR spectrometer (Bruker, Billerica, MA, USA). High resolution mass spectra were recorded on a JEOL AccuTOF LC, JMS-T100LP mass spectrometer (JEOL, Tokyo, Japan) using cold electron-spray ionization (CSI) at −40 °C. GC measurements were performed on a Shimadzu GC-17A Gas Chromatograph (Shimadzu Corporation, Kyoto, Japan) with a Supelco SPB-1 fused silica capillary column. Tris-2-(3-methylindolyl)phosphine [14], L1^H [14], L2^H [14] and potassium graphite (KC₈) [39] were prepared according to literature procedures. Chlorobis[4-(trifluoromethyl)phenyl]phosphine (Alfa Aesar, Haverhill, MA, USA) and chlorobis[4-methoxyphenyl]phosphine (Sigma Aldrich, St. Louis, MO, USA) are commercially available chemicals and were used as received.

Tris-2-(3-methyl-N-di[4-(trifluoromethyl)phenyl]phosphinoindolyl)phosphine (L1^CF3) was prepared in the analogous way to L1^H using chlorobis[4-(trifluoromethyl)phenyl]phosphine: To a solution of tris-2-(3-methylindolyl)phosphine (1.9 g, 4.53 mmol) in THF (50 mL) was added n-BuLi (2.5 M in hexanes, 5.7 mL, 14.3 mmol) at −78 °C. The resulting solution was stirred for 1 h and chlorobis[4-(trifluoromethyl)phenyl]phosphine (3.5 mL, 14.0 mmol) was added. The reaction mixture was stirred for 16 h and allowed to warm slowly to room temperature. The resulting suspension was concentrated in vacuo and dissolved in CH₂Cl₂ (total amount of 50 mL including washing of the pads). The suspension was filtered through a pad of basic alumina and subsequently through a pad of SiO₂. The solvent was removed under reduced pressure and resulted in a yellow foam. This yellow foam was purified by SiO₂ chromatography using a gradient from pure hexane to 2% Et₂O in hexane. Yield: 2.66 g (42%). ¹H NMR (300 MHz CDCl₃) δ: 7.44 (d, J = 8.6 Hz, 15H), 7.33–7.22 (m, 6H), 7.21–7.14 (m, 6H), 7.11 (t, J = 7.6 Hz, 3H), 6.93 (t, J = 7.8 Hz, 3H), 6.59 (d, J = 8.4 Hz, 3H), 2.14 (s, 9H). ¹⁹F NMR (282 MHz, CDCl₃) δ = −60.78 ppm. ³¹P NMR (121 MHz, CDCl₃) δ: 32.63 (d, J = 151.2 Hz, 3P), −76.09 (q, J = 151.6 Hz, 1P). ¹³C NMR (75 MHz, CDCl₃) δ: 1140.28, 140.13, 139.77, 139.53, 139.36, 133.67, 131.91, 131.71, 131.62, 131.43, 129.27, 125.60, 125.53, 125.42, 125.33, 123.72, 122.00, 121.21, 119.50, 113.97, 9.94 ppm.

Tris-2-(3-methyl-N-di[4-methoxyphenyl]phosphinoindolyl)phosphine (L1^OMe) was prepared in the analogous way to L1^H using chlorobis[4-methoxyphenyl]phosphine: To a solution of tris-2-(3-methylindolyl)phosphine (2.20 g, 5.22 mmol) in THF (50 mL) was added n-BuLi (2.5 M in hexanes, 6.6mL, 16.44 mmol) at −78 °C. The resulting solution was stirred for 1 h and chlorobis[4-methoxyphenyl]phosphine (4.83 g, 17.23 mmol) dissolved in 10 mL THF was added. The reaction mixture was stirred for 3 days and allowed to warm to room temperature. The resulting suspension was concentrated in vacuo, dissolved in CH₂Cl₂ (total amount of 50 mL including washing of the pads) and filtered through a pad of basic alumina. Evaporation of the solvent and trituration with Et₂O yielded a white solid which was recrystallized from vapor diffusion evaporation of Et₂O to a concentrated THF solution. Washing with Et₂O and drying under vacuum yielded the ligand in pure form. Yield: 4.03 g (67%). ¹H NMR (300 MHz, (CD₃)₂CO): δ 7.46 (d, J = 7.9, 1.0 Hz, 3H), 7.39–7.27 (m, 6H), 7.20–7.07 (m, 6H), 7.05–6.95 (t, 3H), 6.94–6.81 (m, 12H), 6.77 (d, 6H), 3.74 (s, 18H), 2.01 (s, 9H) ppm. ³¹P NMR (122 MHz, (CD₃)₂CO): δ 36.83 (d, J = 166.0 Hz, 3P), −76.21 (q, J = 166.7, 1P) ppm. ¹³C NMR (75 MHz, (CD₃)₂CO) δ: 161.42, 141.33, 141.16, 134.63, 134.33, 134.06, 134.03, 127.95, 127.78, 127.58, 127.38, 124.22, 123.07, 120.83, 119.56, 115.06, 114.91, 114.86, 114.81, 114.77, 55.52, 9.87 ppm.

RuCl₂L1^CF3: L1^CF3 (1.3732 g, 0.99 mmol) and [Ru(η⁶-benzene)Cl(μCl)]₂ (249.3 mg, 0.50 mmol) were suspended in THF (4 mL) and toluene (8 mL) and stirred at 60 °C for 64 h. After cooling, the yellow precipitated complex was filtered, washed with toluene (1 × 2 mL) and pentane (3 × 3 mL), and dried overnight in the vacuum oven at 40 °C. Yield: 0.9914 g (64%) of a yellow solid. Recrystallization of the complex by layering a DCM solution with pentane at 5 °C gave crystals suitable for X-ray diffraction analysis. ¹H NMR (300 MHz, CDCl₃): δ 7.72 (dd, J = 10.8, 8.0 Hz, 3H), 7.59 (dt, J = 10.5, 5.3 Hz, 4H), 7.53–7.31 (m, 8H), 7.24 (t, J = 7.6 Hz, 4H), 7.18–7.01 (m, 7H), 6.99–6.81 (m, 7H), 6.11 (d, J = 8.5 Hz, 2H), 5.86 (d, J = 8.5 Hz, 1H), 2.93 (s, 6H), 2.69 (s, 3H) ppm. ³¹P NMR (122 MHz, CDCl₃): δ 99.83 (dt, J = 31.5, 26.4 Hz, 1P), 78.90 (t, J = 26.3 Hz, 2P), 49.00 (dt, J = 30.3, 26.7 Hz, 1P) ppm. ¹³C NMR (75 MHz, CDCl₃): δ 134.83, 134.73, 133.95, 133.82, 132.74, 132.63, 132.31, 132.18, 131.43, 131.31, 129.54, 128.96, 128.73, 128.49, 128.18, 128.05, 127.88, 127.74, 127.68, 127.58, 127.44, 126.91, 126.76, 126.44, 126.30, 125.15, 124.71, 124.64, 122.90, 122.45, 122.32, 120.76, 120.22, 117.80, 115.44, 115.00, 13.64, 12.27, 11.99. ppm. Mass Analysis (CSI) [C₆₉H₄₅Cl₁F₁₈N₃P₄Ru]⁺: calc: 1518.1024 found 1518.1067.

RuCl₂L1^OMe: L1^OMe (1.02 g, 0.88 mmol) and [Ru(η⁶-benzene)Cl(μCl)]₂ (227.9 mg, 0.45 mmol) were suspended in THF (4 mL) and toluene (8 mL) and stirred at 120 °C for 3 days leaving a pale brown solution. Upon cooling of the reaction mixture, an off-white solid precipitated. This solid was filtered and washed with hexane (3 × 3 mL) and dried overnight in the vacuum oven at 40 °C, resulting in the product as a brown-white solid. Yield: 1.15 g of an off-white solid (97.8%). ¹H NMR (300 MHz, CDCl₃): δ 7.66 (d, J = 8.0 Hz, 2H), 7.58 (d, J = 7.9 Hz, 1H), 7.44–7.32 (m, 4H), 7.32–7.24 (m, 1H), 7.24–7.02 (m, 10H), 6.89 (q, J = 7.8, 6.9 Hz, 4H), 6.77 (t, J = 7.8 Hz, 1H), 6.68 (d, J = 8.5 Hz, 4H), 6.28 (d, J = 8.5 Hz, 2H), 6.24 (d, J = 7.9 Hz, 2H), 6.12 (d, J = 8.5 Hz, 4H), 5.90 (d, J = 8.5 Hz, 1H), 3.70 (d, J = 1.8 Hz, 12H), 3.62 (s, 6H), 2.84 (s, 6H), 2.60 (s, 3H) ppm. ³¹P NMR (121 MHz, CDCl₃): δ 107.24 (dt, J = 32.0, 26.0 Hz, 1P), 83.78 (t, J = 26.6 Hz, 2P), 54.51 (dt, J = 30.9, 26.9 Hz, 1P) ppm. ¹³C{¹H,³¹P} NMR (75 MHz, CDCl₃): δ = 159.95, 159.46, 139.85, 139.70, 135.39, 135.35, 135.27, 133.02, 132.98, 132.86, 123.72, 123.69, 121.47, 121.27, 120.20, 117.09, 113.38, 113.29, 112.88, 112.72, 112.03, 111.96, 111.89, 77.43, 77.00, 76.58, 54.90, 54.82, 54.79, 11.92, 11.68 ppm. Mass Analysis (CSI) [C₆₉H₆₃ClN₃O₆P₄Ru]⁺: calc: 1290.2399 found: 1290.2400.

RuCl₂L2^H: L2^H (587.3 mg, 0.60 mmol) and [Ru(η⁶-benzene)Cl(μ-Cl)]₂ (150.1 mg, 0.30 mmol) were suspended in THF (3 mL) and toluene (6 mL) and stirred at 120 °C for 64 h. After cooling, 20 mL of hexane was added, which resulted in the precipitation of a yellow solid. The solid was filtered, washed with hexane (3 × 3 mL), and dried overnight in the vacuum oven at 60 °C. Yield: 477.5 mg (69.1%) of a yellow solid. Recrystallization of the complex by layering a DCM solution with pentane at 5 °C gave crystals suitable for X-ray diffraction analysis. ¹H NMR (300 MHz, CDCl₃): δ 7.90 (dd, J = 11.6, 7.7 Hz, 2H), 7.84–7.65 (m, 6H), 7.60 (d, J = 7.8 Hz, 1H), 7.53 (t, J = 9.3 Hz, 3H), 7.45 (d, J = 7.8 Hz, 1H), 7.37 (t, J = 7.6 Hz, 1H), 7.32–7.12 (m, 9H), 7.11–6.92 (m, 6H), 6.90–6.75 (m, 7H), 6.70 (d, J = 8.4 Hz, 1H), 6.36 (t, J = 8.4 Hz, 3H), 6.23 (dd, J = 7.5 Hz, 2H), 1.97 (s, 3H), 1.72 (s, 3H), 1.40 (s, 3H) ppm. ³¹P NMR (121 MHz, CDCl₃): δ 144.43 (dt, J = 38.9, 37.8 Hz, 1P), 39.22 (dt, J = 45.0, 23.6 Hz, 1P), 24.70 (ddd, J = 328.6, 31.0, 21.1 Hz, 1P), 16.62 (ddd, J = 331.4, 39.7, 25.6 Hz, 1P) ppm. ¹³C NMR (75 MHz, CDCl₃): δ 134.83, 134.73, 133.95, 133.82, 132.74, 132.63, 132.31, 132.18, 131.43, 131.31, 129.54, 128.73, 128.49, 128.18, 128.05, 127.88, 127.74, 127.68, 127.58, 127.44, 126.91, 126.76, 126.44, 126.31, 125.15, 124.71, 124.64, 122.90, 122.45, 122.32, 120.76, 120.22, 117.80, 115.44, 115.00, 13.64, 12.27, 11.99 ppm. Mass Analysis (CSI) [C₆₃H₅₁ClN₃P₄Ru]⁺: calc: 1110.1789; found: 1110.1779.

3.1. Reduction of RuCl₂L to RuN₂L

RuCl₂L (±17 µmol) and KC₈ (2–3 equiv) were transferred to a flame-dried Schlenk flask in the glovebox. THF (2 mL) was added and the suspension was stirred for two to three hours; after filtration, a red solution was obtained. Part of this solution was used for in-situ infrared spectrometry, part of the solution was used for in-situ ³¹P NMR analysis, and the rest of the solution was set for crystallization via slow diffusion evaporation of pentane. Applying this procedure to RuCl₂L1^OMe led to formation of a purple solution after filtration, which did not show an infrared band after two hours. However, after one day, the solution turned more red and the corresponding infrared band could be detected. The infrared analysis of the thus obtained solutions yielded the corresponding infrared bands as given in Table 2. The ³¹P-NMR analysis showed the presence of one doublet and one quartet, indicative of a symmetric species, which corresponds with the proposed trigonal bipyramidal geometry.

3.2. Catalysis: Dinitrogen Reduction

The catalyst (33 µmol) and KC₈ (100 equiv) were transferred to a 100 mL flame-dried Schlenk flask equipped with a glass stirring bar in the glovebox. 10 mL of dry THF was added, followed by the addition of the chlorotrimethylsilane (100 equiv). The suspension was stirred overnight. A quantity of 10 µL of n-decane was added as internal standard, whereafter a filtered aliquot of the reaction mixture was measured on the Gas-GC.

3.3. Catalysis in Time: Dinitrogen Reduction

The same procedure as for the normal nitrogen reduction was used, adding the 10 µL of n-decane right away and taking a filtered aliquot of the reaction mixture every hour, which was subsequently measured on the Gas-GC.

3.4. Catalysis: Formic Acid Dehydrogenation

Hydrogen evolution was initiated by the addition of formic acid (45 μL, 1.2 mmol) to a solution of catalyst (±8.5 μmol) in refluxing THF (3 mL) in a 10 mL reaction flask equipped with a condenser. The gas was cooled via a condenser, with the outlet connected to a burette filled with water (Figure 15). The displacement of the water level in the burette was measured in time. The second burette was also filled with water and used to compensate for pressure buildup, keeping the water levels at equal height. The results are shown in

Table 5. Duplo measurements for formic acid dehydrogenation by the depicted catalysts.

Complex	TOF (h−1) #1	TOF (h−1) #2	TOF (h−1) Average
RuCl2L1CF3	120.8	127.2	124
RuCl2L1H	79.3	71.9	75.6
RuCl2L1OMe	123.5	117.4	120.5
RuCl2L2H	22.4	42.7	32.6

.

TOF_H2 was calculated using Equation (1) [40].

$$TO{F}_{H2}=\frac{V_{obs}·0.5}{V_m·n_{cat}}$$

(1)

V_obs: measured gas volume displacement cylinder [mL]; V_m: molar gas volume: 24.49 [mL/mmol]; n_cat: amount of catalyst [mmol].

3.5. X-ray Diffraction Analysis

Crystallographic data was obtained using a Bruker D8 Quest Eco diffractometer equipped with a Triumph monochromator. The intensities were integrated with the SAINT software package [41]. Multiscan absorption correction and scaling was performed with SADABS [42]. The structure was solved with Intrinsic Phasing Methods using SHELXT [43]. Least-squares refinement was performed with SHELXL 2013 [44] against F² of all reflections. Non-hydrogen atoms were refined freely with anisotropic displacement parameters. All hydrogen atoms were located in difference Fourier maps and refined with a riding model. All structures have solvent-accessible voids filled with disordered solvent. Their contribution to the structure factors in the refinement was taken into account with the PLATON/SQUEEZE approach [45] (

Table 6. Crystallographic data for RuCl₂L1^CF3, RuCl₂L2^H and RuN₂L1^H.

Complex	RuCl2L1CF3	RuCl2L2H	RuN2L1CF3
Empirical formula	C69H45Cl2F18N3P4Ru + solvent	C63H51Cl2N3P4Ru, 2(CH2Cl2) + solvent	C69H45F18N5P4Ru + solvent
FW	1553.93 a	1315.77 a	1511.05 a
Temperature [K]	150	150	150
Radiation	Mo Kα	Mo Kα	Mo Kα
Wavelength [Å]	0.71073	0.71073	0.71073
Cryst syst.	monoclinic	monoclinic	monoclinic
Space group	C 2/c	P 21/n	P 21/c
a [Å]	21.594(3)	19.1879(9)	14.2851(6)
b [Å]	27.753(4)	18.1604(8)	18.9090(8)
c [Å]	23.938(3)	19.7879(9)	29.8721(12)
α [deg.]	90	90	90
β [deg.]	93.631(3)	91.012(2)	97.011(2)
γ [deg.]	90	90	90
Volume [Å3]	14317(3)	6894.2(5)	8008.6(6)
Z	8	4	4
Color	pale yellow	yellow	dark red
θ-max	25.135	25.030	26.450
Density [kg·m−3]	1.442 a	1.268 a	1.253 a
Absorp. Coeff. [mm−1]	0.472 a	0.591 a	0.356 a
F(000)	6240 a	2688.0	3040 a
R1/wR2/S	0.0715/0.1975/1.053	0.0412/0.1435/1.134	0.0666/0.2002/1.436

^a Excluding the disordered solvent contribution.

).

RuCl₂L1^CF3: The structure contains voids (2544 ų per unit cell) filled with disordered solvent molecules. Their contribution to the structure factors was secured by back-Fourier transformation using the SQUEEZE routine of the PLATON package [45], resulting in 671 electrons per unit cell which corresponds to 16 disordered molecules of CH₂Cl₂ per unit cell. Alerts A and most Alerts B generated by the IUCr checkCIF program resulted from large displacement parameters of F atoms implying intense rotation of the CF₃ groups of the L1^CF3 ligand. Thirteen FCF reflections below θ-min were obscured by the beam-stop which generated Alert B.

RuCl₂L2^H: The structure contains voids (1518 ų per unit cell) filled with disordered solvent molecules. Their contribution to the structure factors was secured by back-Fourier transformation using the SQUEEZE routine of the PLATON package [45], resulting in 503 electrons per unit cell which corresponds to 12 disordered molecules of CH₂Cl₂ per unit cell. Eleven FCF reflections below θ-min were obscured by the beam-stop which generated Alert B by the IUCr checkCIF program.

RuN₂L1^CF3: The structure contains voids (2224 ų per unit cell) filled with disordered solvent molecules. Their contribution to the structure factors was secured by back-Fourier transformation using the SQUEEZE routine of the PLATON package [45], resulting in 514 electrons per unit cell which corresponds to 12 disordered molecules of pentane per unit cell. Alerts A generated by the IUCr checkCIF program resulted from large displacement parameters of F atoms implying intense rotation of the CF₃ groups of the L1^CF3 ligand. Thirteen FCF reflections below θ-min were obscured by the beam-stop which generated Alert B.

CCDC 1574140, 1574141, 1574142 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge via http://www.ccdc.cam.ac.uk/conts/retrieving.html (or from the CCDC, 12 Union Road, Cambridge CB2 1EZ, UK; Fax: +44-1223-336033; E-mail: [email protected]).

4. Conclusions

In conclusion, we have shown the formation of new ruthenium complexes based on a tripodal 3-methylindolephosphine scaffold. The ligand has two structural (“linkage”) isomers, and, for one of those isomers, several analogues with electron-withdrawing and -donating groups were prepared. These ligands were successfully coordinated to [Ru(η⁶-benzene)Cl(μ-Cl)]₂, resulting in the formation of octahedral complexes with the two chlorides in cis position. The stoichiometric reduction of these complexes with KC₈ yielded the corresponding dinitrogen complexes. Electronic effects in the ligands translated to a change in the N≡N stretch frequency of the coordinated dinitrogen. The complexes were studied in the N₂ reduction with chlorosilanes and KC₈, yielding stoichiometric amounts of the silylamines. When the reaction was followed in time, an incubation period was observed. This incubation period, and the near-identical activities found for all complexes tested, suggests that these species are mere pre-catalysts. The complexes were also employed for formic acid dehydrogenation. Turnover frequencies between 33 and 124 h⁻¹ were reached, depending on the ligand used, clearly showing that the electronic properties of the ligand do have an effect in this reaction. Further studies with these complexes have to be performed to elucidate the reaction mechanism and stability of the catalysts.