Selective Catalytic Transfer Hydrogenolysis of Glycerol to 2-Isopropoxy-Propan-1-Ol over Noble Metal Ion-Exchanged Mordenite Zeolite

Abstract

This study investigated the selective conversion of glycerol to 2-isopropoxy-propan-1-ol over noble metal ion-exchanged mordenite zeolites (RuMOR, RhMOR, and PdMOR) as heterogeneous catalysts via catalytic transfer hydrogenolysis (CTH) using propan-2-ol as the solvent, hydrogen supplier, and reactive coupling reagent with glycerol. The catalytic reactions were performed at 140 °C under inert conditions with a 0.5 MPa initial pressure of N₂. A single product, 2-isopropoxy-propan-1-ol, was catalytically generated without any appreciable by-products. The catalytic results were reproducible, and the catalysts exhibited good recyclability.

1. Introduction

The use of renewable biomass resources for various chemical production processes provides an avenue for chemical industries to reduce their sole dependence on waning fossil-based resources and to alleviate CO₂ emissions [1]. The production of sustainable fuel and chemicals from biomass feedstocks is an attractive approach compared to the use of conventional petroleum-based fuel and chemicals [2]. Biodiesel production through transesterification of vegetable oils cogenerates glycerol on a large scale (~10%). The global production of glycerol is expected to reach 6 million tons by 2025 [3]. Glycerol is currently considered as one of the most promising platform chemicals because of its non-toxic and biodegradable properties [4]. The high demand for and production of biodiesel has resulted in an oversupply of glycerol, consequently reducing the average price in the current chemical market [5]. The efficient conversion of glycerol to several value-added chemicals and fuels has been achieved via various highly applicable catalytic reactions, including oxidation, hydrogenolysis, dehydration, acetalization, and esterification [6,7,8].

Among the aforementioned reactions, hydrogenolysis of glycerol is a useful route for the production of important chemicals such as propane-1,3-diol, propane-1,2-diol, and ethylene glycol, which are commonly utilized in the synthesis of various valuable products, including pharmaceuticals, cosmetics, flavors, fragrances, and antifreezes [9,10]. Various noble and non-noble metals, including Pt, Rh, Ru, Pd, Cu, and Ni, supported on catalytic materials have been widely applied in common hydrogenolysis reactions as they are highly efficient for cleaving the various bonds in the glycerol molecule [11,12,13,14]. The use of bimetallic catalysts over a wide range of reaction temperatures and pressures has also been reported [15,16]. To ensure experimental safety and tempered energy consumption, employing low pressure and avoiding the use of external hydrogen are recommended [17,18]. Previous studies also proved that during hydrogenolysis of glycerol, the solvent itself (e.g., ethanol or propan-2-ol) could produce the necessary hydrogen for the reaction. This methodology is termed catalytic transfer hydrogenolysis (CTH), which provides a facile approach for the hydrogenolysis of glycerol to various value-added products [19].

Herein, we report a new approach for the efficient conversion of glycerol to 2-isopropoxy-propan-1-ol over various noble metal ion-exchanged mordenite (MOR) zeolite catalysts (RuMOR, PdMOR, and RhMOR) in the absence of external H₂ via the CTH reaction, where propan-2-ol is used not only as the solvent but also as the hydrogen supplier and glycerol coupling reagent for the production of 2-isopropoxy-propan-1-ol (Figure S1; Supplementary Materials). MOR zeolite is synthesized via a hydrothermal method, and the noble metals are introduced into the zeolite framework by ion-exchange with the chloride salts of Ru, Rh, and Pd. These catalysts are characterized by powder X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS), thermogravimetric analysis (TGA), solid-state magic angle spinning (MAS) ²⁷Al nuclear magnetic resonance (NMR), N₂ adsorption isotherm, and inductively coupled plasma-optical emission spectrometry (ICP-OES).

2. Results and Discussion

2.1. Characterization of the Catalysts

2.1.1. XRD and SEM Analysis

Figure 1a shows the wide-angle powder XRD patterns of MOR zeolite and the series of noble metal ion-exchanged zeolites. The XRD patterns of as-synthesized NaMOR zeolite (i.e., NaMOR(as-syn)) and the corresponding calcined sample (i.e., NaMOR(cal)) showed the XRD reflections of the fully crystalline MOR framework (JCPDS no. 29-1257) [20]. The XRD patterns of all the noble metal ion-exchanged MOR zeolites (i.e., XMOR(ion-ex) where X is Ru, Pd, or Rh) and the corresponding calcined samples showed that the pristine structure of the MOR framework remained intact after the introduction of various noble metals and subsequent calcination. No new XRD peaks originating from the noble metals were observed, proposing the very small particle size of the noble metals introduced into the MOR zeolites. Figure 1b–e present the SEM images of the calcined MOR zeolite samples (i.e., NaMOR(cal), RuMOR(cal), PdMOR(cal), and RhMOR(cal)), illustrating the micron-scale bulk crystalline morphology. There were no remarkable differences in the SEM images of the samples before and after ion-exchange with the noble metals and subsequent calcination. In addition, no appreciable amorphous phase was observed.

2.1.2. TEM and EDX Analysis

Figure 2(a–d)_i present the TEM images of the calcined samples and the corresponding selected area electron diffraction (SAED) patterns (inset), confirming the bulk crystalline morphology of various MOR zeolite materials. Figure 2(a–d)_ii present the high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images. Energy-dispersive X-ray spectroscopy (EDX) elemental mapping was used to identify the location (occupancy) of the elements (Si, Al, O, Na, Ru, Pd, and Ru) in the zeolite samples. Figure 2(a–d)_iii, (a–d)_iv, and (a–d)_v show the elemental mapping of Si, O, and Al, respectively, confirming the uniform distribution of all the zeolite framework elements throughout the samples. Similarly, the elemental mapping images of Na, Ru, Pd, and Rh are shown in Figure 2(a–d)_vi. The ambiguous elemental mapping in Figure 2(a–d)_vi is possibly due to the low content of these elements in the catalysts. Some amount of the noble metals was localized at the periphery of the zeolite crystals. However, highly dispersed noble metal atoms were also observed. Isothermal N₂ adsorption analysis was used to assess the Brunauer-Emmett-Teller (BET) surface area (S_BET) of the zeolite samples (Table S1; Supplementary Materials).

2.1.3. XPS Analysis

XPS analysis was performed to determine the chemical and electronic state of Ru, Pd, and Rh species in the MOR zeolite samples. Figure 3a shows the full scan survey spectra of RuMOR(cal), PdMOR(cal), and RhMOR(cal), confirming the existence of Ru, Pd, and Rh in the respective RuMOR (~280.0–286.0 eV), PdMOR (~334.0–345.0 eV), and RhMOR (~307.0–31.05 eV) catalysts [21]. Figure 3b–m present the deconvoluted and fitted Ru3d, Pd3d, and Rh3d XPS spectra, along with the O1s, Si2p, and Al2p XPS spectra. The fitted Ru3d spectra are shown in Figure 3b. Deconvolution of the Ru3d curve of RuMOR revealed three binding energy peaks at 280.5 eV (assigned to Ru3d_5/2), 282.1 eV (assigned to Ru3d_5/2), and 284.8 eV (assigned to Ru3d_3/2), which were related to the presence of Ru(0), RuO₂, and Ru(0), species, respectively [22,23]. The XPS data indicated that a few Ru species in RuMOR(cal) were partially reduced to form Ru(0) nanoparticles supported on MOR. Moreover, the Ru3d_3/2 line was expected to fall in the binding energy range of 284.1 to 284.2 eV, which was very close to the previous observation (at 284.6 eV) [24]. The peak at 282.1 eV was assigned to the RuO₂ species corresponding to the Ru(VI) oxidation state, which was formed by oxidation upon exposure of RuMOR(cal) to air [25]. The Pd3d spectra of PdMOR(cal) are shown in Figure 3c.

The deconvoluted Pd3d spectra of PdMOR(cal) showed two peaks at 336.7 eV (assigned to Pd3d_5/2) and 342.1 eV (assigned to Pd3d_3/2), both of which were correlated to the PdO (Pd²⁺) state [26]. The Rh3d signal for RhMOR(cal) (Figure 3d) was fitted to two discernible doublets with different intensities. The doublet observed at ~313.0 eV (assigned to Rh3d_3/2) and ~309.0 eV (assigned to Rh3d_5/2) was attributed to Rh₂O₃ [27]. The additional binding energy peak at 310.8 eV (assigned to Rh3d_3/2) was related to the presence of the Rh(0) NPs in the RhMOR sample [28]. The O1s spectra of all three samples (RuMOR, PdMOR, and RhMOR) are presented in Figure 3e–g). The major peak at 532.8 eV and a smaller peak at ~531.4 eV can be assigned to Si–O–Si and Si–O–Al linkages, respectively. The shoulder peak at 529.3 eV was assigned to the chemisorbed oxygen (O–Ru) in the RuMOR catalyst [29]. Figure 3h–j present the Si2p spectra for all catalysts. The major binding energy peak at 103.5 eV and the relatively smaller peak at 104.2 eV indicated the presence of the SiO₂ (Si⁴⁺ state). The peak at 102.8 eV corresponded to the Si₂O₃ species. The major peak at 74.8 eV in the Al2p spectrum confirmed the presence of the Al₂O₃ species (Figure 3k–m).

2.1.4. TGA and Solid-State MAS ²⁷Al NMR Spectroscopy

Figure 4a shows the TGA data for the NaMOR(cal), RuMOR(cal), PdMOR(cal), and RhMOR(cal) catalysts. A total weight loss of ~9% was observed over different temperature ranges, i.e., from 50 to 300 °C and thereafter up from 300 to 600 °C and beyond. The rapid weight loss in the temperature range of 50 to 300 °C was attributed to the desorption of water molecules present in the MOR zeolites. The weight loss in the temperature range of 300 to 600 °C with respect to RuMOR(cal), RhMOR(cal), and PdMOR(cal) could be ascribed in terms of the amount of Ru, Rh, and Pd present in these catalysts. It has been proven that the considerable weight loss in the metal-exchanged MOR zeolites (in this temperature range) increases with the amount of metal exchanged [30]. Our study firmly corroborates similar trends. Table S1 (ICP-OES data) presents the quantitative analysis (chemical composition) of the noble metal ion-exchanged MOR zeolites. The amount of Ru, Rh, and Pd in the ion-exchanged MOR is in 1:2.1:3.7 ratio. Therefore, the weight-loss trend (Figure 4a) obtained was PdMOR > RhMOR > RuMOR, which was attributed to the dehydroxylation from the hydroxylated noble metal species. These results indicate that all the materials are quite stable at high temperatures and that the stability of NaMOR(as-syn) was sustained after the addition of the noble metals. To understand the coordination of Al species within the MOR framework, all the samples (NaMOR(cal), RuMOR(cal), PdMOR(cal), and RhMOR(cal)) were characterized by solid-state MAS ²⁷Al NMR spectroscopy.

Figure 4b presents the solid-state MAS ²⁷Al NMR spectra of all the samples. The most intense peak at ~59 ppm confirms the presence of tetrahedrally coordinated framework aluminum atoms (F_Al). A very low-intensity peak was also observed at ~0 ppm for the RuMOR(cal), PdMOR(cal), and RhMOR(cal) samples, which was attributed to the octahedrally coordinated extra-framework Al atoms (EF_Al) [31]. This analysis indicates that Al atoms in the MOR frameworks of NaMOR(cal), RuMOR(cal), PdMOR(cal), and RhMOR(cal) are mainly arranged as tetrahedrally coordinated F_Al, which further corroborates synthesis of the highly stable MOR catalysts.

2.2. Catalytic Transfer Hydrogenolysis of Glycerol

All four materials (NaMOR(cal), RuMOR(cal), PdMOR(cal), and RhMOR(cal)) were applied in the CTH conversion of glycerol to 2-isopropoxy-propan-1-ol, a highly applicable solvent for resin, dyes, and surface coatings [32,33]. A series of experiments were conducted under different conditions with variation in reaction time (24, 48, 72, and 96 h) and temperature (120, 140, and 160 °C). Prior to the reactions, the system was purged with N₂ gas and 0.5 MPa was used as the initial reactor pressure for all experiments (the reactor pressure was 1.4 MPa under the optimized reaction conditions (at 140 °C and 500 rpm for 72 h)). The experimental parameters had a major impact on the reaction efficiency, and most importantly, on the product selectivity. The products were identified by gas chromatography mass spectrometry (GCMS) and proton NMR analysis (Figures S2 and S3; Supplementary Materials). The strong peak at m/z 117 (M-1 peak) corroborated the conversion of glycerol to 2-isopropoxy-propan-1-ol (Figure S2), and the ¹H NMR spectra after CTH conversion of glycerol in the presence of propan-2-ol showed the chemical shifts of 2-isopropoxy-propan-1-ol. The analyzed compound also contained residual glycerol and propan-2-ol, and, therefore, the signature peaks of propan-1-ol were merged with the product peaks (peaks a and b); the peaks of unutilized glycerol may also overlap with peaks c, d, and e (Figure S3). In order to confirm the product formation, we also analyzed the ¹H NMR spectra of the starting reaction mixture (glycerol and propan-2-ol in deuterated methanol, shown in Figure S4). The results clearly show differences in the intensity of the proton peaks at the corresponding chemical shifts.

A comparative analysis of the results obtained with the fresh (first run) and recycled (second run) catalysts for CTH conversion of glycerol is presented in Figure 5. The data represent the optimized results for the conversion of glycerol over all the catalysts. The catalytic efficiency (in terms of % conversion) in the first and second reaction runs with the catalysts is represented by red and blue lines, respectively. Interestingly, this reaction resulted in 100% product selectivity (2-isopropoxy-propan-1-ol), as only one additional peak in the GC and GCMS profiles of the solution after reaction was observed. The catalytic efficiency of fresh NaMOR(cal), RuMOR(cal), PdMOR(cal), and RhMOR(cal) was 22.3%, 69.7%, 26.9%, and 50.9%, respectively, for the reaction at 140 °C for 72 h (see the results for the first run catalysts). The catalytic performance changed slightly for all the recycled catalysts (see the results for the second run catalysts). The glycerol conversion achieved with the spent catalysts was 20.1%, 62.9%, 22.1%, and 46.7%, respectively, for NaMOR(cal), RuMOR(cal), PdMOR(cal), and RhMOR(cal). The enhanced catalytic performance of the RuMOR(cal) and RhMOR(cal) catalysts relative to that of PdMOR(cal) was attributed to the strong coordination of Ru and Rh with MOR (due to the +3 oxidation state of Ru and Rh, whereas Pd remains in the divalent state in PdMOR(cal)). The highest catalytic efficiency exhibited by the RuMOR(cal) catalyst can be correlated to the XPS data, which indicates the presence of chemisorbed oxygen (O–Ru) with a binding energy of 529.3 eV [29]. The chemisorbed oxygen would lead to strong interaction between MOR and Ru metal, thereby augmenting the stability of the catalyst. Moreover, the impact of the larger surface area could also be a key factor contributing to the superior catalytic performance of RuMOR.

3. Materials and Methods

3.1. Materials

All chemicals used for the study were of analytical reagent grade (from Sigma-Aldrich (MO, USA), TCI (Tokyo, Japan), and Daejung chemicals (Gyeonggi-do, Korea), and were used without further purification. NaAlO₂, Ludox silica (30 wt.%), RuCl₃, PdCl₂, and glycerol were purchased from Sigma-Aldrich (MO, USA). NaOH, isopropyl alcohol, and RhCl₃·_XH₂O were obtained from Daejung chemicals (Gyeonggi-do, Korea), and dodecane was purchased from TCI chemicals (Tokyo, Japan). Deionized water with a specific resistivity of 16.82 MΩ cm at 25 °C was prepared with a commercial deionizer (Seoul, Korea).

3.2. Methods

3.2.1. Synthesis of Mordenite (MOR)

In a typical synthesis, 0.40 g NaOH was placed into 14 g of deionized water and stirred for 15 min to form a transparent solution. To this solution, 0.2 g of NaAlO₂ was added, stirred for another 15 min, and 6 g Ludox silica (30 wt.% aqueous solution) was further added to the solution. The resultant gel comprising 29.96 SiO₂:5.97 Na₂O:1.22 Al₂O₃:777.78 H₂O was stirred for 2 h at ambient temperature. The final gel was transferred to a Teflon-lined stainless-steel autoclave and kept in an oven for crystallization at 180 °C for 48 h at a tumbling rate of 25 rpm. After the synthesis, the autoclave was cooled in water, and the product was filtered, washed with deionized water, and dried at 100 °C overnight. The obtained powder was further calcined in air at 450 °C for 4 h. The synthesized mordenite powder was denoted as NaMOR(as-syn) and NaMOR(cal) before and after calcination, respectively.

3.2.2. Synthesis of Noble Metal Ion-Exchanged Mordenites (RuMOR, RhMOR, and PdMOR)

RuMOR, RhMOR, and PdMOR catalysts were synthesized via ion-exchange using RuCl₃ (Ru content: 45%–55% by weight), RhCl₃·xH₂O (Rh content: ~40% by weight), and PdCl₂ (Pd content: 60% by weight) as noble metal precursors. A calculated amount of noble metal precursor was dispersed in a 250 mL round-bottom flask containing 2 g calcined NaMOR zeolite dispersed in 150 mL deionized water (with a final ratio of ~3 wt.% noble metals to MOR). The resultant mixture was stirred overnight (~12 h) under ambient conditions. Subsequently, the solution mixture was filtered/centrifuged, washed with deionized water, dried at 100 °C overnight, and calcined in air at 450 °C for 4 h. A similar procedure was adopted for the synthesis of the RuMOR, RhMOR, and PdMOR catalysts.

3.2.3. Characterization of Catalysts

Powder XRD patterns were recorded on a PANalytical Empyrean X-ray diffractometer (Almelo, Netherlands) equipped with Cu-Kα radiation (40 kV, 30 mA) at a rate of 7.5° min⁻¹ over the 2θ range of 5 to 55°. SEM images were recorded on an Analytical UHR Schottky emission scanning electron Hitachi SU-70 microscope (Tokyo, Japan). The SEM samples (in powder form) were prepared on carbon tape, and later the platinum spin coating of each sample was done for 100 s. The SEM analysis for all the samples was carried out at the accelerating voltage of 15.0 kV. TEM images of the catalysts were obtained by using a Tecnai F-20 (Philips) instrument (OR, USA) operating at 200 kV (lattice resolution: 0.19 nm). Images were recorded with a charged coupled device (CCD) camera (CA, USA) at high magnification under low-dose conditions. All the TEM samples were prepared by dispersing the catalyst powder in ethyl alcohol under ultrasonication. A few droplets of suspension were kept on a carbon-coated copper grid and allowed to dry under ambient conditions. Elemental mapping was carried out by energy dispersive spectroscopy (EDAX STD-S PV7747/67, MA, USA) with an energy resolution of 129 eV. X-ray photoelectron spectroscopy was performed on a high-performance Thermo Scientific (HP-XPS) K-ALPHA⁺ instrument (MA, USA) with an ultimate vacuum of 5 × 10⁻⁹ mbar pressure (pass energy of 200 eV for survey spectra and 50 eV for narrow spectra, take-off angle of 60°) equipped with a monochromated Al-Kα source connected to a 128-channel detector. Thermogravimetric analysis (TGA) of the catalysts was performed in the temperature range of 50 to 900 °C at the ramping rate of 10 °C min⁻¹ under N₂ atmosphere (20 mL min⁻¹) on a Perkin Elmer STA 6000 (simultaneous thermal analyser, MA, USA) instrument. Solid-state MAS ²⁷Al NMR data were obtained at a spinning rate of 10 kHz on a JEOL ECZ400R instrument (Tokyo, Japan). For each analysis, a total 2048 number of scans were performed (with the relaxation delay of 1 s) using a standard of 1.1 M Al(NO₃)₃ in D₂O. The surface area of all the materials was determined by applying the BET equation to the isothermal N₂ adsorption data acquired at 77 K using a BEL/Besorp mini II instrument (Tokyo, Japan). Prior to the measurements, all the samples were degassed at 400 °C for 2 h. Elemental analysis was carried out with a Thermo Scientific iCAP 6300 DuO ICP-OES instrument equipped with a charge injection device (CID) detector (MA, USA) working in the wavelength range of 166 to 841 nm. The data were obtained at wavelengths of 396.1, 212.4, 340.4, 240.2, and 343.4 nm for Al, Si, Pd, Ru, and Rh, respectively.

3.2.4. Catalytic Transfer Hydrogenolysis (CTH) of Glycerol to 2-Isopropoxy-Propan-1-Ol

CTH reactions were performed in a 50 mL stainless-steel autoclave high pressure liquid batch reactor (Korea) equipped with a pressure regulator, under N₂ atmosphere. In a typical reaction, 0.4 g of freshly prepared catalyst (NaMOR(cal), RuMOR(cal), PdMOR(cal), or RhMOR(cal)) was charged into an autoclave containing 9 g isopropyl alcohol (propan-2-ol). To this mixture, 0.5 g (5.43 mmol) of glycerol was added. The resultant mixture was thoroughly purged twice with N₂ to remove residual gases and moisture inside the vessel, and an initial N₂ pressure of 0.5 MPa was applied for each reaction. The reactor was then heated to 140 °C and stirred at 500 rpm for 72 h. The reactor was autogenously pressurized to 1.4 MPa under the aforementioned reaction conditions. After a designated time, the reaction was stopped, and the reactor was cooled to ambient temperature. An internal standard (dodecane) was added to the reaction solution. The solid catalyst was filtered out by using a syringe filter (pore size 0.22 µm), and the liquid part of the reaction mixture was analyzed by gas chromatography (6500 GC, Younglin instrument, Gyeonggi-do, Korea) with a DB-5 column and a flame ionization detector (FID, CA, USA).

3.2.5. Product Analysis

The reaction solution and products were analyzed and identified by GC, GC-MS, and ¹H NMR. The compounds were first analysed by GC (6500 GC, YL instrument, Gyeonggi-do, Korea) equipped with a DB-5 column (30 m × 0.32 mm × 1.50 µm) and a flame ionization detector (FID, CA, USA). GCMS analysis was performed on an Agilent Technologies 5975B Inter XL MSD instrument equipped with an Agilent J&W GC column (30 m × 0.32 mm × 0.25 µm) and mass selective detector (MSD, CA, USA). ¹H NMR analysis was carried out with an Agilent ProPulse 500 MHz NMR spectrometer (CA, USA). The NMR samples were prepared in deuterated methanol (CD₃OD). The catalytic performance was calculated in terms of the conversion (%). The reaction between glycerol and propan-2-ol was confirmed from the GC, GCMS, and ¹H NMR data. The % conversion was calculated based on the glycerol concentration in the reaction mixture before and after the reaction. The product selectivity was confirmed from the GC and GCMS data, where only one additional peak (apart from those of the reactants) was observed after the reaction.

$$\% \mathrm{Conversion}=\frac{\mathrm{mole}\mathrm{of}\mathrm{converted}\mathrm{glycerol}}{\mathrm{mole}\mathrm{of}\mathrm{initial}\mathrm{glycerol}}\times 100$$

4. Conclusions

In summary, we have reported a novel approach for catalytic application of mono-metallic noble metal ion-exchanged MOR zeolites for the selective CTH conversion of glycerol to 2-isopropoxy-propan-1-ol in the presence of propan-2-ol under mild reaction conditions, in the absence of external hydrogen. The synthesized noble metal-based catalysts (RuMOR, RhMOR, and PdMOR) were characterized via XRD, SEM, TEM, XPS, TGA, solid-state ²⁷Al MAS NMR, and a BET surface area analyzer, which affirms the synthesis of highly stable materials. This study, for the first time, proposes an alternative use of glycerol to obtain 2-isopropoxy-propan-1-ol by controlling the reaction conditions with 100% product selectivity, and confirms that the ruthenium-based ion-exchanged mordenit (RuMOR) displayed best catalytic performance compared to the other catalysts. As part of future research, the effects of other catalytic parameters and reaction conditions are currently under investigation. The obtained product is highly applicable as an effective solvent for resin, dyes, and use in surface coating.