Selective Hydrodeoxygenation of Esters to Unsymmetrical Ethers over a Zirconium Oxide-Supported Pt-Mo Catalyst.

The catalytic hydrodeoxygenation (HDO) of carbonyl oxygen in esters using H2 is an attractive method for synthesizing unsymmetrical ethers because water is theoretically the sole coproduct. Herein, we report a heterogeneous catalytic system for the selective HDO of esters to unsymmetrical ethers over a zirconium oxide-supported platinum-molybdenum catalyst (Pt-Mo/ZrO2). A wide range of esters were transformed into the corresponding unsymmetrical ethers under mild reaction conditions (0.5 MPa H2 at 100 °C). The Pt-Mo/ZrO2 catalyst was also successfully applied to the conversion of a biomass-derived triglyceride into the corresponding triether. Physicochemical characterization and control experiments revealed that cooperative catalysis between Pt nanoparticles and neighboring molybdenum oxide species on the ZrO2 surface plays a key role in the highly selective HDO of esters. This Pt-Mo/ZrO2 catalyst system offers a highly efficient strategy for synthesizing unsymmetrical ethers and broadens the scope of sustainable reaction processes.

Introduction

The synthetic strategy for the controlled removal of oxygen from highly oxygenated compounds has attracted both academic and industrial interest. In this context, catalytic hydrodeoxygenation (HDO) using H2 as a reductant is a promising method because water is the sole coproduct of this reaction.[1−6] Although there have been many reports on the HDO of oxygenated compounds via the cleavage of carbon–oxygen bonds (C–O and C=O), the selective HDO of acyl carbonyl oxygen in carboxylic acid derivatives has rarely been achieved.[7,8] In particular, the selective HDO of esters to ethers is significantly challenging because alcohols and/or hydrocarbons are predominantly formed over the corresponding ethers.

As an important class of bulk and fine chemicals, ethers are prevalent in solvents, fuels, fragrances, and pharmaceuticals.[9] Conventionally, ethers are obtained by the acid-catalyzed dehydration of alcohols. However, with this method, the selective synthesis of unsymmetrical ethers can only be achieved using benzylic and propargylic alcohols [Scheme (I-a)].[10−12] A wider variety of unsymmetrical ethers can be synthesized via the Williamson reaction [Scheme (I-b)] or the Ullmann reaction [for aryl ethers; Scheme (I-c)].[13] Nevertheless, these methods have serious drawbacks associated with the large amounts of salt waste resulting from the use of inorganic bases, organic halides, and alkoxide substrates. In comparison, the deoxygenation of esters is considered a straightforward method for the preparation of unsymmetrical ethers because of the ubiquity of esters in natural and synthetic organic compounds. Such reactions often use metal hydride reagents such as LiAlH4 or NaBH4 in the presence of BF3 etherate,[14,15] which produces a large quantity of metal waste [Scheme (II-a)]. As an alternative to the above stoichiometric methods, catalytic hydrosilylation is an area of growing interest [Scheme (II-b)]. These reactions are commonly performed with homogeneous metal complex catalysts under mild reaction conditions, but the formation of siloxane waste results in low atom efficiency.[16−18] From the viewpoints of environmental protection and atom efficiency, the catalytic transformation of esters to ethers using H2 as a reductant is more attractive because water is the only coproduct. Recently, the Beller group and others reported the catalytic etherification of esters using binary catalyst systems consisting of ruthenium/triphos complexes and acid additives such as Al(OTf)3 in the presence of alcohols [Scheme (II-c)].[19−21] However, these homogeneous catalysts still require high H2 pressures (4–6 MPa) and temperatures (140–160 °C). Consequently, to establish green and sustainable processes, it is necessary to develop more efficient catalysts for the production of unsymmetrical ethers from esters using H2 under mild and additive-free conditions.

Scheme 1

Methods for the Synthesis of Unsymmetrical Ethers: (I) Conventional Methods;[10−13] (II) Ether Synthesis from Esters Using Hydride Reagents or Homogeneous Catalysts;[14−21] and (III) This Work

Herein, we achieved the HDO of various esters, including a biomass-derived glyceride, to the corresponding unsymmetrical ethers without additives over a zirconium oxide-supported Pt–Mo catalyst (Pt–Mo/ZrO2). This catalyst exhibited high activity for direct HDO under mild reaction conditions [0.5 MPa H2 and 100 °C; Scheme (III)]. Cooperative catalysis using Pt nanoparticles and neighboring molybdenum oxide clusters on the ZrO2 support was found to be critical for the highly selective direct HDO of esters to unsymmetrical ethers.

Results and Discussion

Catalytic Performance

Various supported bimetallic catalysts were prepared by a sequential impregnation method (Table S1). The effects of the catalyst on the HDO of cyclohexyl acetate (1a) to cyclohexyl ethyl ether (2a) were examined using n-hexane as a solvent at 100 °C under 0.5 MPa H2 for 4 h (Table ). Among the supported bimetallic catalysts evaluated for the HDO of 1a, Pt–Mo/ZrO2 showed high catalytic activity and selectivity, affording 2a in 70% yield with cyclohexane (4a) formed in 23% yield as a byproduct (entry 1; see Figure S1 and Scheme S1 for details). Other ZrO2-supported bimetallic M1–Mo catalysts (M1 = Ru, Rh, or Pd) did not show significant HDO activity (entries 4–6). Furthermore, bimetallic catalysts in which the Mo species was replaced with Re, W, or V (Pt–M2/ZrO2) were ineffective for the HDO of 1a (entries 7–9). The support material of the catalyst also strongly affected its HDO activity and selectivity for 2a. For instance, although Pt–Mo/TiO2 exhibited high catalytic activity, its selectivity for 2a was lower than that of Pt–Mo/ZrO2 (entry 10). Catalysts with other supports (Pt–Mo/hydroxyapatite; entry 11) were less effective than those with TiO2 or ZrO2 supports; some of which hardly promoted the HDO of 1a (Pt–Mo/CeO2 and Pt–Mo/MgO; entries 12 and 13, respectively). To clarify the effects of the catalyst metal components, Pt/MoO3, Pt/ZrO2, and Mo/ZrO2 catalysts were used for the HDO of 1a (entries 14–16, respectively). Unlike Pt–Mo/ZrO2, these catalysts provided only trace amounts of 2a. We also found that the physical mixture of Pt/ZrO2 and Mo/ZrO2 afforded a lower yield of 2a than that in the case of Pt–Mo/ZrO2 (entry 17), thus clearly demonstrating that the simultaneous presence of both Pt and Mo on the ZrO2 support is essential for the efficient HDO of 1a to 2a.

Table 1

HDO of Cyclohexyl Acetate (1a) to Cyclohexyl Ethyl Ether (2a) over Various Catalystsa

			yield [%]b
entry	catalyst	conv. [%]b	2a	3a	4a
1c	Pt–Mo/ZrO2	97	70	0	23
2d	Pt–Mo/ZrO2 (0.1 MPa H2)	100	57	0	26
3e	Pt–Mo/ZrO2 (reuse)	97	67	0	25
4	Ru–Mo/ZrO2	2	0	2	0
5	Rh–Mo/ZrO2	2	0	2	0
6	Pd–Mo/ZrO2	2	0	2	0
7	Pt–Re/ZrO2	10	6	4	1
8	Pt–W/ZrO2	9	1	6	1
9	Pt–V/ZrO2	5	0	5	0
10	Pt–Mo/TiO2	87	52	<1	28
11	Pt–Mo/hydroxyapatite	30	19	1	10
12	Pt–Mo/CeO2	4	1	2	1
13	Pt–Mo/MgO	3	0	3	0
14	Pt/MoO3	16	3	3	10
15	Pt/ZrO2	8	2	4	2
16	Mo/ZrO2	2	0	2	0
17	Pt/ZrO2 + Mo/ZrO2	30	21	5	4

Reaction conditions: Pt–Mo/ZrO2 (0.15 g, 2 mol % Pt and 0.375 mol % Mo), 1a (1 mmol), n-hexane (3 mL), H2 (0.5 MPa), 100 °C, 4 h. For entries 4–9 and 15–17, M1–Mo/ZrO2 (2 mol % M1 and 0.375 mol % Mo) and Pt–M2/ZrO2 (2 mol % Pt and 0.375 mol % M2) were used.

Conversion and yield were determined using GC–MS with an internal standard.

Ethane was detected in the gas phase using GC-thermal conductivity detector and MS.

Pt–Mo/ZrO2 (0.0375 g, 2 mol % Pt and 0.375 mol % Mo), 1a (0.25 mmol), n-hexane (0.8 mL), H2 (0.1 MPa), 100 °C, 8 h.

Reuse of the catalyst from entry 1.

Furthermore, the catalytic performance of Pt–Mo/ZrO2 was strongly affected by the solvent. Among the solvents examined, nonpolar n-hexane and dodecane gave high yields of 2a, whereas oxygen-containing polar solvents (i.e., ethanol, 2-propanol, water, and tetrahydrofuran) were ineffective (Table S2). Notably, the HDO of 1a over Pt–Mo/ZrO2 also proceeded under an atmospheric pressure of H2, affording 2a in 57% yield at 100% conversion of 1a (Table , entry 2). Furthermore, the Pt–Mo/ZrO2 catalyst was easily recovered from the reaction mixture by centrifugation and could be reused without an appreciable loss of activity or selectivity (entry 3).

Substrate Scope

The catalytic performance of Pt–Mo/ZrO2 for the HDO of various esters was investigated (Table and Scheme S2 for substrate limitations). Notably, both unsymmetrical and symmetric aliphatic ethers were obtained in high yields. Acetate esters bearing linear aliphatic chains such as octyl (1b), hexyl (1c), and butyl (1d) gave the corresponding alkyl ethyl ethers in approximately 60% yield at 100% conversion of the esters (entries 2–4). Higher selectivity for ethers was observed in the reactions of branched butyl acetates. For example, isobutyl acetate (1e) and sec-butyl acetate (1f) afforded 2e in 84% yield and 2f in 64% yield, respectively (entries 5 and 6). Sterically bulkier esters afforded even higher yields of ethers; for instance, isopropyl cyclohexanecarboxylate (1g) and isopropyl 2-methyl butyrate (1h) were efficiently converted to 2g (95% yield) and 2h (86% yield), respectively (entries 7 and 8). However, the HDO of vinyl pivalate and hexyl formate did not proceed (Scheme S2). A series of unsymmetrical aliphatic ethers were obtained in high yields from 1i–1k, 1m, and 1n (entries 9–11, 13, and 14). A symmetric ether (2l) was produced in good yield (entry 12), and alkyl benzoates (1o–1q) were transformed into the corresponding ethers in high yields, but aromatic rings were not tolerated under the reaction conditions (entries 15–17). Macrocyclic lactones such as exaltolide (1r) and cyclohexadecanolide (1s) smoothly underwent HDO to provide the corresponding cyclic ethers 2r and 2s in 68% yield and 60% yield, respectively (entries 18 and 19).

Table 2

HDO of Various Esters over Pt–Mo/ZrO2a

Reaction conditions: Pt–Mo/ZrO2 (0.15 g, 2 mol % Pt and 0.375 mol % Mo), ester (1 mmol), n-hexane (3 mL), 100 °C, H2 (0.5 MPa).

Conversion and yield were determined using GC–MS with an internal standard.

Conversion and yield were determined using 1H NMR spectroscopy with an internal standard.

140 °C, H2 (5 MPa).

The excellent HDO performance of the Pt–Mo/ZrO2 catalyst was further demonstrated for the valorization of biomass-derived compounds. Although fatty acid methyl esters are widely used as biodiesel, the instability of their ester moieties against hydrolysis is a drawback.[22] Notably, the HDO of methyl stearate (1t) and ethyl stearate (1u) proceeded to afford methyl octadecyl ether (2t) in 76% yield and ethyl octadecyl ether (2u) in 87% yield (entries 20 and 21), which could be used as biodiesel with improved hydrolytic stability. The Pt–Mo/ZrO2 catalyst was successfully applied to HDO on the preparative scale. Using 3 g of 1u, 2u was obtained in 65% isolated yield alongside octadecane (4u) in 27% yield; the turnover number (TON) for 2u (based on surface Pt atoms) reached 260 (Scheme a and eq S1). Consequently, this is a new, green synthetic pathway to access 2t and 2u. Furthermore, an abundant natural glyceride, trilauryl glyceride 1v, which is an ingredient of coconut oil, was converted to tridodecyl glyceryl ether (2v) in 51% isolated yield (Scheme b). These results clearly demonstrate the versatility of Pt–Mo/ZrO2 for the HDO of esters to unsymmetrical ethers.

Scheme 2

HDO of Biomass-Derived Esters; (a) Preparative-Scale HDO Reaction of 1u; and (b) HDO of a Triglyceride

Heterogeneous Nature of the Pt–Mo/ZrO2 Catalyst

As mentioned previously, Pt–Mo/ZrO2 was easily recovered from the reaction mixture and could be reused (Table , entry 3). To investigate the leaching of metal species during the catalytic reaction, a hot filtration test was carried out. After the Pt–Mo/ZrO2 catalyst was removed by filtration at 60% conversion of 1a, the filtrate was subjected to the same reaction conditions for a further 3 h (Figure S2). No additional product was formed in the filtrate, indicating that the HDO of esters proceeded on the Pt–Mo/ZrO2 surface. Inductively coupled plasma atomic emission spectroscopy (ICP–AES) analysis revealed that the Pt and Mo loading amounts were the same in the fresh and used Pt–Mo/ZrO2 catalysts, which confirmed that Pt and Mo leaching was negligible during the HDO reaction (Table S3). These results demonstrate that the Pt and Mo species are strongly immobilized on the ZrO2 support and that the HDO of esters involves heterogeneous catalysis.

Reaction Pathway

To elucidate the origin of the high selectivity for the desired unsymmetrical ethers over the Pt–Mo/ZrO2 catalyst, we proposed three reaction pathways: (a) hydrogenolysis/alcohol condensation, (b) keto–enol tautomerization/hydrogenation/hydrogenolysis, and (c) hydrogenation/hydrogenolysis (Scheme ). The feasibilities of these pathways were evaluated based on various control experiments. The treatment of 3a with ethanol under HDO conditions did not produce any ether products, including unsymmetrical ether 2a and symmetrical ethers (Scheme a). This result indicated that ether formation via the hydrogenolysis/alcohol condensation pathway (Scheme a), which includes dehydrative alcohol condensation,[21] is negligible. Furthermore, Pt–Mo/ZrO2 catalyzed the HDO of cyclohexyl acetate-d3 (5a), affording cyclohexyl ethyl-d3 ether in 67% yield (Scheme b). As no H–D scrambling product was observed in this experiment, the HDO reaction is unlikely to proceed via the keto–enol tautomerization/hydrogenation/hydrogenolysis pathway shown in Scheme b. Hence, among the proposed pathways in Scheme , the hydrogenation/hydrogenolysis pathway without CO–R bond cleavage (Scheme c) is the most plausible reaction pathway.

Scheme 3

Possible Reaction Pathways for the HDO of 1a to 2a

Scheme 4

Control Experiments: (a) Condensation of 3a and Ethanol and (b) HDO of 5a; Yields Were Determined Using GC–MS with an Internal Standard

Characterization of Pt–Mo/ZrO2 after the HDO Reaction

X-ray absorption fine structure (XAFS) analysis was used to investigate the atomic-scale structures of the Pt and Mo species in Pt–Mo/ZrO2. Pt L-edge X-ray absorption near-edge structure (XANES) spectra of the fresh and used Pt–Mo/ZrO2 catalysts are depicted in Figure a. The white line of the Pt L3-edge XANES spectrum of the used Pt–Mo/ZrO2 catalyst showed an intensity lower than that of the fresh Pt–Mo/ZrO2 catalyst and similar to that of the Pt foil, which reveals the formation of metallic Pt species. In the Mo K-edge XANES spectra, the absorption edge of used Pt–Mo/ZrO2 was shifted toward a lower energy relative to that of fresh Pt–Mo/ZrO2, indicating that the Mo6+ species were reduced to Mo4+ (Figure b). Interestingly, the pre-edge shoulder peak remained for used Pt–Mo/ZrO2, which differs from the spectral features of Mo4+ in MoO2. It has been reported that the pre-edge peak for the Mo6+ species in MoO3 is derived from a distorted MoO6 octahedral structure.[23] Therefore, the Mo species in used Pt–Mo/ZrO2 are present in the 4+ valence state but retain the pristine distorted octahedral structure of MoO3, resulting in the presence of oxygen vacancy (Ov) sites on the MoO3 surface (HMoO).[24,25] Overall, these XAFS results suggest that Pt(0) species and Ov sites on HMoO are generated under the employed reaction conditions, which may play key roles in the direct HDO of esters.

Figure 1

(a) Pt L3-edge XANES spectra of fresh Pt–Mo/ZrO2, PtO2, used Pt–Mo/ZrO2, and Pt foil. (b) Mo K-edge XANES spectra of fresh Pt–Mo/ZrO2, used Pt–Mo/ZrO2, MoO2, and (NH4)6Mo7O24. (c) TEM image (scale bar: 10 nm) and size distribution histogram (inset) of used Pt–Mo/ZrO2. (d) FT-IR spectra of ethyl butyrate vapor and ethyl butyrate adsorbed on Pt–Mo/ZrO2, Pt/ZrO2, and ZrO2.

Furthermore, structural analyses of the Pt–Mo/ZrO2 catalyst were carried out. The X-ray diffraction pattern of used Pt–Mo/ZrO2 was similar to those of fresh Pt–Mo/ZrO2 and pristine ZrO2. No diffraction peaks attributable to crystalline Pt or Mo species were observed (Figure S3), suggesting that the Pt and Mo species were highly dispersed. Transmission electron microscopy (TEM) images of used Pt–Mo/ZrO2 revealed the presence of nanoparticles with a mean diameter of 2.4 nm (Figures c and S4 for other supports). A high-resolution TEM measurement of the nanoparticle showed that the measured d-spacing value of the lattice fringe was approximately 0.199 nm, which corresponds to the (200) plane of face-centered cubic Pt nanoparticles (Figure S5). CO pulse chemisorption analysis revealed that H2-pretreated Pt–Mo/ZrO2 and Pt/ZrO2 adsorbed similar amounts of CO (Table S4), indicating that the Mo species in Pt–Mo/ZrO2 did not decorate the surface of the Pt nanoparticles.[26] Hence, these results suggest that the Pt nanoparticles and HxMoOy species are highly dispersed on the ZrO2 surface.

Proposed Catalytic Cycle

To investigate the interaction between Pt–Mo/ZrO2 and esters, Fourier transform infrared (FT-IR) analysis was conducted using ethyl butyrate as a probe. When ethyl butyrate vapor was adsorbed on H2-pretreated Pt–Mo/ZrO2 at 100 °C, an intense band appeared at 1716 cm–1 (Figure d), which was attributed to the C=O stretching vibration of the ester functionality. This band was shifted to a much lower wavenumber than that observed for ethyl butyrate vapor alone (1755 cm–1). In sharp contrast, no bands were observed in the spectra of Pt/ZrO2 and ZrO2. The Ov sites of HMoO have been reported to act as Lewis acid sites for the activation of acyl carbonyl groups.[27,28] Thus, the FT-IR spectroscopy and XAFS results suggest that the carbonyl moiety of the ester is directly adsorbed and activated on the Ov sites of HMoO.

Based on this analysis, a reaction mechanism was proposed for the conversion of esters to ethers over Pt–Mo/ZrO2 (Scheme ). First, Pt nanoparticles are formed via reduction with H2 (I), followed by H2 dissociation (II). The spillover hydrogen from the Pt nanoparticles generates Ov sites on the HMoO clusters adjacent to the Pt nanoparticles (III). Ester activation occurs following the adsorption of the acyl carbonyl moiety on the Ov sites of HMoO (IV). Then, the HDO of the ester affords the corresponding ether. Overall, the HDO of esters proceeds via a “reverse Mars–van Krevelen”-type mechanism,[29−33] with Pt nanoparticles activating H2 and partly reduced HMoO promoting the removal of carbonyl oxygen through coordination at its Ov sites, which are Lewis acid sites. This cooperative catalysis between Pt nanoparticles and HMoO species successfully promotes the direct HDO of various esters to ethers.

Scheme 5

Proposed Catalytic Cycle for the Direct HDO of Esters to Ethers over Pt–Mo/ZrO2

Conclusions

We developed a highly efficient heterogeneous catalytic system for the direct and selective HDO of esters to ethers over Pt–Mo/ZrO2. A wide range of esters could be converted into the corresponding unsymmetrical and symmetric ethers, including a biomass-derived glyceryl ether, in moderate to excellent yields (51–95%). Pt–Mo/ZrO2 operated well under mild reaction conditions (0.5 MPa H2 at 100 °C) and was applicable to preparative-scale production with a high TON (260). Moreover, Pt–Mo/ZrO2 was easily separated from the reaction mixture and could be reused without sacrificing its high activity, thus offering a simple and clean synthetic methodology for the synthesis of unsymmetrical ethers. XAFS and FT-IR spectroscopy studies revealed that the excellent performance of Pt–Mo/ZrO2 for the HDO of esters was attributable to cooperative catalysis between Pt nanoparticles and the Ov sites of HMoO clusters on the ZrO2 surface via a reverse Mars–van Krevelen-type mechanism. The high activity of Pt–Mo/ZrO2 for unsymmetrical ether synthesis makes a significant contribution toward the development of future sustainable reaction processes.

Experimental Section

Preparation of Metal Oxide-Supported Pt–Mo Catalysts

The Pt–Mo/ZrO2 catalyst was prepared using a sequential impregnation method. An aqueous solution of (NH4)6Mo7O24·4H2O (1 mL, 37.5 mM) and ZrO2 (1.5 g) was added to distilled water (50 mL) at room temperature. After stirring for 12 h in air, water was removed by rotary evaporation under reduced pressure to obtain the solid product. The obtained powder was dried at 110 °C for 5 h. After drying, the product and an aqueous solution of H2PtCl6 (2 mL, 100 mM) were added to distilled water (50 mL) at room temperature and stirred for 12 h. Then, water was removed by rotary evaporation, and the product was dried at 110 °C for 5 h. Finally, the obtained powder was calcined at 500 °C for 3 h under a static air atmosphere to obtain Pt–Mo/ZrO2 as a gray powder. As determined using ICP–AES, the Pt and Mo contents in Pt–Mo/ZrO2 were 2.45 and 0.24 wt %, respectively. The other M1–M2 bimetallic catalysts were prepared in a similar way, with the loading amounts of M1 and M2 adjusted to 6.0 and 1.13 μmol/m2, respectively. The specific surface areas and amounts of each support are summarized in Table S1.

Typical HDO Reaction Procedure (Table , Entry 1)

The reaction with cyclohexyl acetate (1a) was carried out in a 50 mL stainless-steel autoclave equipped with a Teflon vessel. The vessel was charged with 1a (1 mmol), Pt–Mo/ZrO2 (0.15 g), and n-hexane (3 mL), and a Teflon-coated magnetic stirring bar was added. The reactor was sealed, purged five times with 0.5 MPa H2, and then pressurized (0.5 MPa), heated to 100 °C, and stirred at 900 rpm for 4 h. After the reaction, the autoclave was cooled in an ice-water bath, and hydrogen gas was released. The resulting reaction mixture was diluted with ethyl acetate and analyzed using gas chromatography–mass spectrometry (GC–MS).