Design, Synthesis, Characterization of Novel Ruthenium(II) Catalysts: Highly Efficient and Selective Hydrogenation of Cinnamaldehyde to (E)-3-Phenylprop-2-en-1-ol

In this contribution, two novel supported and non-supported ruthenium(II) complexes of type [RuCl2(dppme)(NN)] where [dppme is H2C=C(CH2PPh2)2 and NN is N1-(3-(trimethoxysilyl)propyl)ethane-1,2-diamine] were prepared. The NN co-ligand caused release of one of the dppme ligands from [RuCl2(dppme)2] precursor to yield complex 1. The process of substitution of dppme by NN was monitored by 31P{1H}-NMR. Taking advantage of the presence of trimethoxysilane group in the backbone of complex 1, polysiloxane xerogel counterpart, X1, was prepared via sol-gel immobilization using tetraethoxysilane as cross-linker. Both complexes 1 and X1 have been characterized via elemental analysis, CV and a number of spectroscopic techniques including FT-IR, 1H-, 13C-, and 31P-NMR, and mass spectrometry. Importantly, carbonyl selective hydrogenation was successfully accomplished under mild conditions using complex 1 as a homogenous catalyst and X1 as a heterogeneous catalyst, respectively.


Synthetic Investigation of Ruthenium(II) Complex 1 and Xerogel X1
Reaction of [RuCl 2 (dppme) 2 ] with an equivalent amount of N 1 -(3-(trimethoxysilyl)propyl)ethane-1,2diamine in dichloromethane afforded complex 1 as depicted in Scheme 1. Both the change in color of the reaction mixture from brown to light yellow and 31 P{ 1 H}-NMR chemical shifts confirmed that one of the dppme ligands was exchanged equivalently by a diamine ligand. The structure of the complex 1 was confirmed by elemental analysis and by means of various spectroscopic techniques, including IR, TG/DTA 1 H-, 13 C{ 1 H}-and 31 P{ 1 H}-NMR spectroscopy, and FAB-mass spectrometry. For this compound, the 1 H-, 13 C-, and 31 P-NMR spectral features are all in excellent agreement with the suggested structure. Complex 1 has been used to prepare the xerogel X1 through a simple sol-gel polymerization process at room temperature in the presence of 10 equivalents of Si(OEt) 4 as cross linker using methanol/THF/water, as shown in Scheme 1. Due to the poor solubility of X1 in common solvents, it was subjected to solid state NMR measurements.

Spectral Data
The 31 P{ 1 H} signals in the 31 P-NMR spectrum of complex 1 show a splitting of the 31 P{ 1 H} signals; this is due to the asymmetric nature of diamine in N 1 -(3-(trimethoxysilyl)propyl)ethane-1,2-diamine co-ligand without a C2 axis which will lead to AX resonance patterns for complex 1, as shown in Figure 1. The phosphorous chemical shifts and the 31 P-31 P coupling constants (Jpp = 35.8 Hz) suggested that the phosphine ligand was positioned trans to the diamine, with trans-dichloro atoms, to form the kinetically favored trans-Cl 2 Ru(II) isomer [7][8][9][10]. Formation of complex 1 was monitored by 31 P{ 1 H}-NMR spectroscopy. [RuCl 2 (dppme) 2 ] complex has a 31 P{ 1 H}-NMR signal corresponding to complex Cl 2 Ru(dppme) 2 at  p = −1.9 ppm. Addition of N 1 -(3-(trimethoxysilyl)propyl)ethane-1,2-diamine resulted in a fast substitution of this ligand with one molecule of dppme ligand. The substitution reaction was confirmed by a decrease in the intensity and a 48 ppm downfield shift, in addition to the appearance of two new signals, one belonging to free dppme at  p = −17.2 ppm and the other an AX pattern 31 P{ 1 H}signal at  p = 33.6, 46.3 ppm which arises from the desired complex 1 (Figure 1b). [RuCl 2 (dppme) 2 ] was totally converted to complex 1 within 30 min. (Figure 1c). The 31 P{ 1 H}-NMR spectrum also confirmed the absence of any side products in this ligands substitution reaction.

Elemental Analysis and FAB-Mass Spectrum of Complex 1
The elemental analysis of complex 1 is consistent with the proposed molecular formula (Calcd. for

IR Spectrum of Complex 1
The IR spectrum of complex 1 shows the absorption bands of the functional groups present as displayed in Figure 3. Absorption bands in the 3,390-3,280 cm −1 region can be assigned to NH stretching vibrations. An absorption band observed at 3,180 cm −1 is attributed to the stretching vibration of the aromatic C-H bonds, while bands at 2,980-2,740 cm −1 are due to C-H stretching vibrations. Other characteristic bands due to other functional groups are also present in the expected regions.

NMR Spectra of Complex 1 and Xerogel X1
There is a good agreement between the 1 H-NMR spectrum of the prepared complex 1 and its assigned structure. Displayed in Figures 5a,b are the 1 H-NMR spectra of Cl 2 Ru(dppme) recorded in

Abs.
CDCl 3 before addition of the diamine ( Figure 5a) and after (Figure 5b) to prepare complex 1. The spectrum of 1 (Figure 5b) revealed signals of aromatic and aliphatic protons of dppme and [3-(2-aminoethyl)aminopropyl]trimethoxysilane ligands that appear as complex multiplets in the region~6.0-8.0 and 0.5-4.5 ppm, respectively. Integration of 1 H signals confirms that the dppme to diamine ratio is in agreement with the structural composition of 1.
In the 13 C-NMR spectra for complex 1 and X1 recorded in CDCl 3 , signals associated with the different types of carbons in the dppme ligand as well as [3-(2-aminoethyl)aminopropyl]trimethoxysilane diamine co-ligand were observed. Their assignment was achieved by free ligand 13 C-NMR and 135 DEPT studies; DEPT experiments were employed to differentiate secondary and quaternary carbons from primary and tertiary carbons. Several sets of aliphatic and aromatic carbons related to the phosphine and diamine were assigned to their positions. The 13 C-NMR spectra of the free [3-(2-aminoethyl)aminopropyl]trimethoxysilane, Cl 2 Ru(dppme) 2 complex, complex 1; reaction mixture of Cl 2 Ru(dppme) 2 and [3-(2-aminoethyl)aminopropyl]trimethoxysilane] ligand and solid state 13 C-CP-MAS-NMR of xerogel X1 are shown in Figure 6. Examination of the 13 C-CP-MAS-NMR spectrum of the modified solids along with the solution phase spectrum of the corresponding molecular precursor led to the conclusion that the organic fragments in complex 1 and xerogel X1 remained intact during the grafting and subsequent workup without noticeable decomposition ( Figure 5). The absence of the CH 3 O signal at  C = 36.8 which belongs to (CH 3 O) 3 Si in [3-(2-aminoethyl)-aminopropyl]trimethoxysilane co-ligand after sol-gel reaction of complex 1 to form xerogel X1 was the major difference recorded in spectra; this is in agreement with the immobilization of the desired hybrid Ru(II) complexes. In addition, the total disappearance of groups in X1 (Figure 6b) compared with complex 1 (Figure 6a) provides good confirmation that the sol-gel process has proceeded to completion [22][23][24][25][26][27][28][29][30][31][32][33][34].
Solid-state 29 Si-NMR provided further information about the silicon environment and the degree of functionalization [22][23][24][25][26]. In all cases, the organometallic/organic fragment of the precursor molecule was covalently grafted onto the solid, and the precursors were, in general, attached to the surface of the polysiloxane by multiple siloxane bridges. The presence of T m sites in case of xerogel 1 and 2 in the spectral region of T 2 at  Si = −55.8 ppm and T 3 at  Si = −66.1 ppm was confirmed. Additionally, Q silicon sites due to Si(OEt) 4 condensation agent were also recorded to Q 3 at  Si = −101.2 and Q 4 at  Si = −109.5 ppm silicon sites of the silica framework [22][23][24][25][26][32][33][34].

Thermal Studies
A typical thermal TG/DTA curve of complex 1 is given in Figure 7. The thermal decomposition study of the complex was investigated in the 25-900 °C temperature range under open atmosphere at a heating rate of 10 °C/min. There is no weight loss in the range 25-290 °C which indicates the absence of coordinated or uncoordinated water molecules. As Figure 7 reveals, the complex undergoes a one-step decomposition with 88% weight loss due to the loss of coordinated chlorides, diamine, dppme ligands from the complex between 293 and 385 °C with an exothermic DTA peak at 340.7 °C. The final residue remaining after the complex was heated to 900 °C was analyzed by IR spectroscopy and identified as ruthenium oxide.

Electrochemistry
The electron-transfer behavior of the complex 1 was examined before sol-gel by cyclic voltammetry and the corresponding results are represented in Figure 8. One-electron oxidation reversible wave (∆E = 100 mV) was observed around 0.0-0.20 V vs. Cp 2 Fe 0/+ which was assigned to Ru(II/III) oxidation reduction couple reaction. The half-wave potential, E ½~0 .15 v, was calculated from the average of the anodic and cathodic wave's potential. This is in agreement with the observed for the trans-[RuCl 2 (dpme)(diamine)] species [7][8][9].

Complexes 1 and X1 as Catalysts in the Hydrogenation of Cinnamaldehyde
[RuCl 2 (diphosphine)(diamine)] has been previously used as a hydrogenation catalyst under a H 2 atmosphere using a strong base as co-catalyst and 2-propanol as solvent for the hydrogenation of α,β-unsaturated ketones and aldehydes [2][3][4][5][6]. In this investigation, the catalytic activity and selectivity of the newly prepared complexes 1 and X1 are studied and compared with our previous complexes. Results of the hydrogenation process using complex 1 [RuCl 2 (dppme)NN] and its polysiloxane xerogel X1 counterpart were compared with those obtained using the complex 1' [RuCl 2 (dppp)NN] and its polysiloxane xerogel counterpart X1' [22] under the same hydrogenation conditions, listed in Table 1. Using cinnamaldehyde as model substrate, three different regioselective hydrogenation products are expected, as shown in Scheme 2. Selective hydrogenation of the carbonyl group affords the corresponding unsaturated alcohol A. Other possible hydrogenation routes (B and C) are undesired. The hydrogenation reactions using complex 1 and xerogel X1 as catalysts were carried out under identical experimental conditions. Scheme 2. Different hydrogenation possibilities of cinnamaldehydes: Selective carbonyl group hydrogenation to produce A, selective C=C hydrogenation to produce B, full hydrogenation path with no selectivity to produce C.
Taken together, the combined results clearly indicate that the newly prepared catalysts were only effective in the presence of excess of hydrogen in 2-propanol as solvent and a strong basic co-catalyst such as KOH and tert-BuOK. Without a strong base or even with a weak base such as K 2 CO 3 , no hydrogenation reaction was observed.
Thus, complexes 1 and 1' are highly active under the abovementioned conditions and resulted in 99% conversion (TOF higher than 1,000) in addition to selective hydrogenation of the C=O group while keeping the C=C intact. Similarly, the xerogels X1 and X1' displayed high conversion ratios and selectivity ~90% in the C=O selective hydrogenation of cinnamaldehyde using strong basic conditions. Decrease in the activity and selectivity were observed upon comparing the homogenous complexes with the heterogeneous xerogels precursors under identical condition. Moreover, complexes 1 and X1 were highly active under identical experimental conditions compared to their corresponding complexes with dppp complex 1' and X1' [RuCl 2 (dppp)(NN)] [22][23][24][25][26]. The presence of double bond in the diphosphine backbone increases the hydrogenation activity due to an increase in the electrophilicity of the ruthenium metal center [7][8][9][10].

Synthesis of Complex 1
Complex 1 is prepared according to the following general procedure: a solution of [3-(2-aminoethyl)aminopropyl]trimethoxysilane (0.10 g, 0.455 mmol, 5% excess) in dichloromethane (10 mL) was added dropwise to a stirred solution of Cl 2 Ru(dppme) 2 (0.22 mmol) in dichloromethane (10 mL) over a 2 min period. The mixture was maintained at room temperature with stirring for ca. 1 h during which the color changed from brown to yellow. The volume of solution was then concentrated to about 2 mL under reduced pressure. Addition of diethyl ether (40 mL) caused the precipitation of a solid which was filtered (P4), dissolved in dichloromethane (40 mL), and concentrated again under vacuum to a volume of 5 mL. Addition of n-hexane (80 mL) caused the precipitation of a solid which was filtered (P4), washed several times with n-hexane, and dried under vacuum. Complex 1 was obtained in analytically pure form with very good yields. m.p. = 340 °C (dec.

General Procedure for the Catalytic Studies
A mixture of 0.02 mmol of the respective complexes, 2.0 mmol of cinnamaldehyde, 0.20 mmol of KOH or tert-BuOK or K 2 CO 3 as co-catalysts and 50 mL of 2-propanol was placed in a 100 mL Schlenk tube. The mixture was sonicated for 5 min to assure that the solids in reaction mixture were completely dissolved. The reaction mixture was vigorously stirred, degassed by two freeze-pump-thaw cycles, and then pressurized with hydrogen gas at 2 bars. The mixture was vigorously stirred at 35 °C for 1 h in case of homogenous or 12 h in heterogeneous. During the hydrogenation process, samples were taken from the reaction mixture after the gas was removed to determine the conversion percentage and hence turnover frequency (TOF). Samples were inserted into a gas chromatograph using a special glass syringe, and the various types of reaction products were compared with authentic samples.

Conclusions
In summary, two novel [3-(2-aminoethyl)aminopropyl]trimethoxysilane/dppme/ruthenium(II) complexes were prepared. Complex 1 was prepared in high yield by a ligand exchange substitution reaction. Sol-gel polymerization of complex 1 afforded xerogel X1 due to the T-silyl functions on the diamine co-ligand backbone. The formation reaction of the desired complexes was monitored by 31 P{ 1 H}-NMR. The structure of complexes 1 and xerogel X1 have been confirmed by elemental analyses, CV, IR, FAB-MS, TG/DTA and 1 H-, 13 C-, and 31 P-NMR spectroscopy. When these complexes were tested as catalysts for the selective hydrogenation of cinnamaldehyde in both homogenous and heterogeneous phases, they revealed high degree of activity and excellent selectivity for carbonyl hydrogenation under mild conditions.