Ruthenium(II) Phosphine/Picolylamine Dichloride Complexes Hydrogenation and DFT Calculations

Abstract

Treating [Ru(PPh₃)₃Cl₂] with the amine 2-picolylamine (Picam) ligand in a 1:1 molar ratio, the Ru(II) complex trans-Ru(PPh₃)₂(Picam)Cl₂ (1) is obtained in methylene chloride and can be isolated as a pure solid compound. The single-crystal structure of 1 was determined by X-ray crystallography. The geometry at the Ru metal center is a distorted octahedral environment with a trans arrangement of the two chlorides. A trans effect of the bond lengths was observed within the structure. Similarly, treating [Ru(PPh₃)₃Cl₂] with 1:1:1 molar ratios of 2-picolylamine (Picam) and 1,1′-bis(diphenylphosphine)ferrocene (DPPF) ligands yielded the Ru(II) complex trans-Ru(DPPF)(Picam)Cl₂ (2). In identical conditions, the homogeneous hydrogen transfer catalytic reactivity of complexes 1 and 2 has been tested in a basic 2-propanol solution and they indicate different catalytic activity. It was discovered that monodentate and bidentate phosphine ligands of Ru(II) complexes, as well as cis- and trans-chloro configuration display different catalytic properties from our experimental data, in agreement with literature data. Based on DFT calculations, the relative molecular catalytic reactivity of all available experimental data is understood from the relative calculated molecular energy.

1. Introduction

The use of catalysts to accelerate chemical reactions is a very efficient route in chemical synthesis and industrial manufacturing [1]. Most transition metals and lanthanides show various catalytic activities due to unoccupied d and f orbitals [2,3]. Ruthenium, in the second transition metal group, has been studied extensively due to its excellent catalytic properties for either photocatalysis or hydrogenation catalysis in various reactions [4]. J. Halpern and B. James were the first researchers to explore the hydrogenation catalytic property of ruthenium metal chlorides RuCl₃ in the early 1960s [5,6]. Thereafter, different Ru metal chemicals and complexes were used as either heterogeneous [7,8] or homogenous catalysts [9,10].

The synthesis routes that transfer hydrogenation through ruthenium(II) catalyzing become popular [11,12]. Ruthenium(II) complexes containing mixed nitrogen and phosphino ligands are efficient catalysts and a large number of derivatives with different combinations of P and N donor ligands has been synthesized [13,14]. Noyori’s research group has developed a highly active catalytic system with both chiral phosphine and amine (P–P and H₂N–NH₂) ligands, e.g., trans-RuCl₂(diphosphine)(1,2-diamine) [15,16]; it was determined that the Ru–H/–NH₂ motif plays a key role in generating highly catalytic hydrogenation reactivities as well as catalyzing the transfer hydrogenation of ketones in a 2-propanol solution [17,18].

We have previously synthesized a series of Ru(II) complexes and studied their hydrogen-transfer reactivity by the selection of 1,1′-bis(diphenylphosphino)ferrocene (DPPF) as a ‘constant’ bidentate P–P ligand and varied the bidentate N–N ligand using different diamine and diimine ligands [18]. We found that bidentate diimines have comparable catalytic ability as do bidentate diamines. Recently we also demonstrated that monodentate or bidentate amine/imine Ru(II) complexes show very different catalytic reactivities [19]. The Ru(II) with a monodentate imine pyridine ligand (Py) demonstrated less activity compared to that for the bidentate diimine 2,2′-bipyridine (Bipy) or 1,10-phenanthroline (Phen) [18] and even worse compared to that for the 2-aminopyridine ligand [19]. Only when both bidentate ligands of phosphine and diamine/diimine are present can it provide the stable square plane structure (Scheme 1(1)) during the reaction period and through this stable square plane can conduct the hydride transfer to the substrate. This leads us to believe that the stable square plane catalytic transitional structure is the main contributing factor to a fast transfer for carrying forward the hydrogenation reaction. Herein we have designed two Ru(II) complexes with one side fixed using a mixed amine and imine ligand of 2-picolylamine, meanwhile varying the other side with monodentate triphenylphosphine and bidentate DPPF. We will compare catalytic activities of the two complexes of RuCl₂(PPh₃)₂(Picam) (1) and RuCl₂(DPPF)(Picam) (2) under the same conditions to determine how these different two types of phosphines, affect the reactivity (monodentate (PPh₃) and bidentate (DPPF)) and to understand which factors control the catalytic capability of the synthesis complexes. This will in turn contribute to our understanding of the catalytic mechanism. So far advanced theoretical simulations by DFT calculations to assist to understand the molecular interaction have been addressed previously [20,21]. To fully understand this system in comparison to related literature data [14,22]. Gaussian DFT calculations for two types of phosphine ligands (monodentate and bidentate) in different geometries were systematically conducted, and these calculations helped us to explain the experimental results.

2. Results and Discussion

2.1. Synthesis of Complexes

The reaction of trans-RuCl₂(PPh₃)₃ with a 1:1 molar ratio of 2-picolylamine (Picam) in methylene chloride complexes 1 yielded RuCl₂(PPh₃)₂(Picam) (see Figure 1). From complex 1 via consecutive substitution reactions by addition of one equiv. of 1,1′-bis(diphenylphosphino)ferrocene (DPPF) to a solution of (RuCl₂(PPh₃)₃+2-picolylamine) under aerobic conditions resulted in a rapid color change from blackish-purple to red, and straightforward work-up of the solutions gave high yields of high-purity yellow products of complex 2. In addition, trans-RuCl₂(PPh₃)₂(Picam) complex 1 was crystallized from the solution.

2.2. Crystal Structure of trans-[RuCl₂(PPh₃)₂(Picam)] (1)

Unfortunately, no crystal structure data for complex 1 is available from the literature. By our effort, compound 1 was crystallized from a dissolved powder sample and the crystal structure was solved in a single crystal by X-ray crystallography diffraction in the solid state [23,24]. Selected bond lengths and bond angles are listed in the caption of Figure 2. The geometry at the Ru metal for complex 1 is an octahedral environment (Figure 2) with a trans arrangement of the two chloride ligands. The bond lengths of Ru–N are Ru–N(1)=2.1482(19) and Ru–N(2)=2.155(2) Å for Ru–imine and Ru–amine nitrogen atoms, respectively. Bond lengths of Ru–P(1)=2.3346(6) and Ru–P(2)=2.3190(6) Å as well as Ru–Cl(1)=2.4295(6) and Ru–Cl(2)=2.4321(6) Å are also determined. It is obvious that the Ru–N bond length for the imine (Py) N atom is slightly shorter than that for the amine (–NH₂) N atom. The trans effect is observed in the structure, such as a relatively shorter bond length for Ru–N(1) (2.1482(19) Å) and a relatively longer Ru–P(1) (2.3346(6) Å) bond. In contrast, a longer Ru–amine N bond length (2.155(2) Å) with a shorter Ru–P(2) (2.3190(6) Å bond are seen, which are also the same trans effect to Ru–Cl(1) and Ru–Cl(2) bond lengths, whereas the Ru–P(1) bond trans to the imine–N displays a slightly longer distance (0.0156 Å) than Ru–P(2) trans to the amine–N. The angle N(1)–Ru–N(2) for Picam is 76.26(8)°, whereas the Cl(1)–Ru–Cl(2) and Cl(1)–Ru–N(1) angles are 164.89(2)° and 80.64°, respectively, leading to a more distorted octahedral geometry than those obtained with symmetric diamine ligands (Figure 2) [18]. The Ru–N(1)–C(1)–N(2) torsion angle is less than that for N(2)–N(1)–P(2)–P(1), which indicates that Ru metal with a bidentate mixed amine/imine ligand is more planar.

Selected Bond Lengths (Å) and Angles (°): Ru–N(1)=2.1482(19); Ru–N(2)=2.155(2); Ru–Cl(1)=2.4295(6); Ru–Cl(2)=2.4321(6); Ru–P(1)=2.3346(6); Ru–P(2)=2.3190(6). Cl(1–Ru–Cl(2)=164.89(2); Cl(1)–Ru–P(1)=99.29(2); Cl(1)–Ru–N(1)=80.65(6); P(1)–Ru–P(2)=98.75(2); Cl(1)–Ru–P(2)=104.91(2); P(1)–Ru–N(1)=165.19(5).

2.3. Comparison of Reactivities of Catalytic Hydrogen Transfer of a Ketone

Herein, the activity of homogeneous catalytic transfer hydrogenation of ketones (C=O bond) for the synthesized pre-catalysts of Ru(II) complexes 1 and 2 was tested in basic isopropanol at 80 °C under an inert nitrogen atmosphere (Figure 3) [25,26]. All of the tested complexes performed the homogeneous catalytic hydrogen transfer reactivity effectively under the given conditions (

Table 1. Catalyzed transfer hydrogenation reduction of acetophenone.

Precursor Catalysts	Reaction Time (h)	Conversion (%)	TOF (h−1)
RuCl2(PPh3)2 (Picam) (1)	24	62.2	6600
RuCl2(DPPF) (Picam) (2)	24	90.0	10,320

). Under identical conditions, activities of transfer hydrogenation of acetophenone in a basic 2-propanol for complexes 1 and 2 are very different (Table 1). The time dependence of the conversion rate versus the reaction time is shown in Figure 4 for complexes 1 and 2 under the experimental conditions (see Table 1). The conversion data indicate that complex 2 is a more efficient hydrogen-transfer catalyst than complex 1 under these conditions. The only structural difference between these two complexes is the monodentate phosphine ligand in 1 and bidentate phosphine ligand in 2.

Experimental condition: in 2-propanol, under N₂, 80 °C, KOH Cat/Base/Substrate = 1/20/1000. TOF calculated at reaction time at 5 min, the conversion percentage did not much change between 5 min and 24 h.

We prepared different structural Ru(II) complexes to explore their catalytic hydrogenation transfer reactivities. Initially, we selected the bidentate P–P ligand of 1,1′-bis(diphenylphosphino)ferrocene (DPPF) and the bidentate N–N ligands of diamine or diimine, yielding ruthenium six-coordinate complexes with trans-chloride and equatorial cis-P and cis-N yielding a square-planar structure. The catalysis tests demonstrated that their catalytic activities are similar, with the diamine slightly more favorable than the diamine [18]. Very recently we have tuned the nitrogen ligands of monodentate or bidentate amine/imine Ru(II) complexes showing that Ru(II) with a monodentate imine ligand of pyridine (Py) has less activity compared that for the bidentate diimine 2,2′-bipyridine (Bipy) or 1,10-phenanthroline (Phen) [18,19]. Herein we demonstrated that the catalytic reactivities of monodentate phosphine (P) ligands are less efficient than bidentate phosphine (P–P). Only with bidentate ligands of both phosphine (P–P) and nitrogen (N–N) can it provide the stable square-planar structure during the reaction period, indicating this equatorial square-planar structure is critical to maintaining the highly catalytic hydrogen transfer reactivity.

In addition, coordinated chlorides were considered being replaced by hydrides in the hydrogen transfer reactivities. In all Ru(II) complex pre-catalysts, the chlorides in trans-configuration showed lower catalytic activities than that for the cis-chlorides configuration. It has been reported in the literature that monodentate phosphine (P) is almost twice as catalytic in cis-chlorides than in trans-chlorides. Meanwhile, bidentate phosphine (P–P) is almost 10 times as reactive in cis-chlorides than in trans-chlorides [27]. The cis-hydrides are much more efficient in transferring hydrides from the Ru metal center to the ketone substrates. In summary of our tested results and literature data, the relative activities of Ru(II) complexes in different structures and configurations as hydrogen transfer catalysts are shown in Figure 5. The reactivities are closely controlled by the coordination type of ligands and by the chloride configuration.

2.4. DFT Analysis and Mechanism of Hydrogen Transfer

To understand the correlation between structural geometry and catalytic activities, geometry optimizations and energy calculations for eight different structures for both chlorides (Figure 5) and hydride (chloride position was replaced by hydride in Figure 5) were performed. DFT calculations were performed for eight structures with monodentate and bidentate phosphine ligands, along with both chlorides and hydrides in cis- and trans- geometric configurations. To simplify without changing the geometry the monodentate and bidentate of phosphine were replaced by PH₃ and H₂P–(CH₂)₂–PH₂, respectively. In total, two groups with eight geometric structures were selected to conduct the DFT optimization. Four Ru(II) complex structures of monodentate (PH₃) with two chlorides and two hydrides in trans- and cis- geometry configurations were optimized by DFT calculations. The relative energies of the four structures are plotted and shown in Figure 6a. From an energy perspective, trans-chloro Ru monodentate phosphine complex ([RuCl₂(H₃P)₂)(Picam)]) is the most stable geometry, with a calculated energy 82.2 kJ/mol lower than the cis-chloro complex ([RuCl₂(H₃P)₂(Picam)]). Meanwhile, the relationship is exactly opposite for the Ru-hydride structures, with the cis-hydride complex ([RuH₂(H₃P)₂(Picam)]) being more stabilized than the trans-hydride complex with a relatively lower energy of 15.0 kJ/mol (Figure 6b). For the bidentate phosphine ligand system, trans-chloro Ru complex ([RuCl₂(H₂P(CH₂)₂PH₂)(Picam)]) is also the most stabilized geometry, 28.1 kJ/mol lower than cis-chloro complex ([RuCl₂(H₂P(CH₂)₂PH₂)(Picam)]), as shown in Figure 6a. Cis- and trans-hydrides of ([RuH₂(H₂P(CH₂)₂PH₂)(Picam)]) complexes show the same trend with an energy difference of 4.2 kJ/mol (as shown in Figure 6a). The HOMO and LUMO molecular orbitals (MO) for the bidentate ligand Ru(II) complexes in different geometric arrangements are shown in Figure 7, and those for the monodentate ligands are shown in Figure 8. The HOMO MO indicates that most electron density occurs around the hydride and chloride, suggesting these hydrides are very active in contributing electrons to form chemical bonds with other atoms. In principle theory, the high density electron atom can easily overlap with other atoms to form chemical bonds.

DFT calculation results show that for both monodentate and bidentate phosphine ligands, the cis-chloro to cis-hydride transformation is more energy favorable. This explains why the cis-chloro configuration shows highly catalytic reactivity. Comparing the monodentate with the bidentate, the latter ligand requires less energy than the monodentate ligand (Figure 6). This suggests that the monodentate phosphine ruthenium complex consumes more energy to form a hydride Ru-complex, which is the activated catalyst species. This is also consistent with the experimental test results for this system from the data presented here and in the literature [14,22].

Regarding these types of catalysts, one of the proposed mechanisms for reduction of ketones is through hydride bifunctional group H-bonding interactions between a ‘P₂RuH–NH’ unit and the ketone (Scheme 1(2)), finally the ketones are reduced and form alcohols [27,28]. Perhaps this is because the monodentate ligand for either nitrogen N(Py) or phosphine (PPh₃) bonding to Ru(II) is not as stable as the bidentate ligands and is easily dissociated in solution. Consequently, due to the loss of the square plane, the foundation for activation of the ‘P₂RuH–NH’ unit is no longer available, so the formation of a transition activation structure between ‘RuH–NH’ and substrate ketone cannot occur [19]. Our DFT calculation results also indicate that bidentate phosphine and amine/imine bidentate ligands are more energy favorable than monodentate phosphine ligands probably because these two types of bidentate ligands can maintain a stable square equatorial plane. This is the critical condition to ensure greater catalyst efficiency through the proposed bifunctional intermediate mechanism to reduce the ketones. All the DFT and experiment results indicate that the cis-chloro bidentate phosphine ruthenium complex is the best and fastest pre-catalyst.

3. Materials and Methods

The phosphines (PPh₃ and DPPF), 2-picolylamine (Picam) and the ketones were used as received from Aldrich. RuCl₃·3H₂O was donated by Colonial Metals, Inc.; RuCl₂(PPh₃)₃, was synthesized by the literature methods [29]. All the solvents were dried before use.

Standard Schlenk techniques were adopted to conducted experiments and manipulations under an inert atmosphere. The solvents, ruthenium precursors of [RuCl₂(PPh₃)₃], NMR spectra and Chromatography (GC) analyses were prepared and operated exactly as previously described [18,19]. Elements were analysed by The Carlo Erba EA 1108 elemental analyzer. IR, GC and EA were conducted in The Analytical and Instrumentation Laboratory, Chemistry Department, University of Alberta.

3.1. Synthesis of trans-RuCl₂(PPh₃)₂(Picam) (1)

Synthesis of trans-RuCl₂(PPh₃)₂(Picam) (1): RuCl₂(PPh₃)₃ (0.80 g, 0.84 mmol) was loaded in a round-bottom flask suspended in 10 mL of methylene chloride, then 2-picolylamine (Picam) (90 μL, 0.88 mmol) was added. The mixture was stirred at room temperature for 2 h and concentrated under reduced pressure; addition of pentane afforded a yellow powder precipitate. After filtration the product was washed twice with diethyl ether (Et₂O) and dried under vacuum. Yield: 520 mg (77.0%). Anal. Calcd for C₄₂H₃₈Cl₂N₂P₂Ru: C, 62.69; H, 4.76; N, 3.48. Found: C, 62.81; H, 4.78; N, 3.51. ¹H NMR (CDCl₃, 20 °C): 8.60–6.50 (m, 34H; phenyl and C₅H₄N protons), 4.45 (s; br, 2H; CH₂), 3.28 (s; br, 2H; NH₂). ¹³C{¹H} NMR (CDCl₃, 20 °C): 162.7 (s; NCCH₂), 157.6 (s; NCH of C₅H₄N), 136.6–120.1 (m; aromatic carbons), 50.8 (s; CH₂). ³¹P{¹H} NMR (CDCl₃, 20 °C): 44.0 (d, ²J_PP = 32.6 Hz), 40.1 (d, ²J_PP = 32.6 Hz).

3.2. trans-RuCl₂(DPPF)(Picam) (2)

Complex 2 was synthesized using the same procedure and solvents as described for the synthesis of 1. Yield: 81 mg (97.0%); characterization details were given in our previous publication [19].

3.3. Crystal Structure Determination

Data collection was performed on a Bruker PLATFORM/SMART 1000 CCD diffractometer. The structures were solved using the Patterson search/structure expansion (DIRDIF-99) and were refined using full-matrix least-squares on F² (SHELXL-2018) [23,24]. The selected crystal data and structural refinement details for 1 are listed in

Table 2. Crystal data and structure refinement details for complex 1.

Empirical formula	C42H38Cl2N2P2Ru·2CH2Cl2 (1)
Formula weight	974.51
Crystal system Crystal Dimensions	Monoclinic 0.42 × 0.23 × 0.14 mm
Space group	P21/c (No. 14)
Unit cell parameters a (Å)	18.1129(11)
b (Å)	9.8700(6)
c (Å)	24.4824(15)
β (°)	102.8701(8)
Volume (Å3)	4266.9(5)
Z	4
Calculated density (g cm−3)	1.517
Temperature, K	193.2(1)
μ (MoKα), (mm−1)	0.853
θ range for data collection (°) Index ranges Independent reflections Observed reflections Data/restraints/parameters Goodness-of-fit on F2	0.3 to 27.48 −23 ≤ h ≤ 23 −12 ≤ k ≤ 12 −31 ≤ l ≤ 31 9773 8046 9773/0/496 1.043
Final R indices [F02 ≥ 2σ(F0)] wR2 [F02 ≥ −3σ(F02)]	R1 = 0.0362 wR2 = 0.0924
Large difference peak and hole	−1.062 and 0.881 e/Å3

.

3.4. General Procedure for the Catalytic Transfer Hydrogenation

Hydrogen-transfer experiments were carried out in standard Schlenk glassware [25,26]. All reactions were set up in a dry glove-box under N₂ using 10⁻⁵ M catalyst in 2-propanol (10 mL), the test calculated amount of acetophenone (PhCOCH₃) was loaded in a final molar ratio of catalyst/KOH/substrate was in 1:20:1000. The temperature was set to 80 °C under refluxed and quickly magnetic stirring. The yields were calculated after the products were isolated and measured using gas chromatography (GC) by comparison with data of known compounds.

3.5. Computational Details

The DFT calculations of geometry and energy were performed with the Gaussian 03 program [30]. Becke’s three-parameter hybrid functional with the Lee-Yang-Parr correlation functional (B3LYP) was employed for all calculations [31]. The LanL2DZ basis set on Ru and the 6-311G basis set on P, C, and H were selected. Once an optimized geometry was finished, imaginary frequencies were checked at the same level by vibration analysis to verify the genuine minimum on the potential energy surface (PES) and to evaluate the zero-point energy (ZPE) correction.

4. Conclusions

Two different types of phosphine complexes, Ru(II) trans-RuCl₂(PPh₃)₂(Picam) (1) and trans-RuCl₂(DPPF)(Picam) (2) as pre-catalysts were studied for their transfer hydrogenation catalytic activities under the same experimental conditions. The transfer hydrogenation activities demonstrated that the bidentate phosphine (DPPF) coordinated to Ru metal has higher catalytic activity than the monodentate phosphine (PPh₃). These results are similar to data previously reported whereby the nitrogen bonding ligands were varied [18,19]. Combining the new results with previous data, we conclude that both bidentate phosphine and nitrogen ligands need to be present to maintain the highly catalytic transfer hydrogenation activity. These results indicate that phosphine and nitrogen ligands bond strongly to Ru(II) and thus guarantee the formation of stabilized planar ‘P₂RuN₂′ units in situ, which is critical as an efficient hydrogenation transfer catalyst. DFT calculations for both monodentate and bidentate phosphine chlorides and hydrides show their different energy requirements, however, both cis-chloro and cis-hydrides are more energy favorable to formation with bidentate ligands compared to monodentate ligands. By reviewing all available data, we achieved a full understanding of the different geometric complexes and their varying catalytic reactivities.