Dehydrogenative Oxidation of Alcohols Catalyzed by Highly Dispersed Ruthenium Incorporated Titanium Oxide

Abstract

Ruthenium incorporated titanium oxides (Ru_xTiO₂) were prepared by a one-step hydrothermal method using Ti(SO₄)₂ and RuCl₃ as the precursor of Ti and Ru, respectively. X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), transmission electron microscope (TEM), Energy-dispersive X-ray spectroscopy (EDS) mapping, and BET were applied for the analyses of catalysts. Ruthenium atoms are well dispersed in the anatase phase of TiO₂ and the crystallite size of Ru_xTiO₂ (≈17 nm) is smaller than that of pure TiO₂ (≈45 nm). In particular, we found that our homemade pure TiO₂ exhibits a strong Lewis acid property. Therefore, the cooperation of ruthenium atoms playing a role in the hydride elimination and the Lewis acid site of TiO₂ can efficiently transfer primary alcohols into corresponding aldehydes in an oxidant-free condition.

1. Introduction

The oxidation of alcohols into their carbonyl compounds by a cost effective and environmentally benign means is important in synthetic chemistry. Particularly, many researchers have made concentrated efforts on the development of catalytic oxidation reactions using molecular oxygen (O₂) as an oxidant instead of using stoichiometric amounts of expensive and harmful oxidizing reagents [1,2,3]. Moreover, there have been lots of studies seeking the novel oxidation catalytic systems of alcohols in which alcohol oxidation is accomplished by a dehydrogenation pathway under oxidant-free conditions. In homogeneous catalytic systems, iridium complexes, osmium, or ruthenium pincer-type complexes, and Grubbs’ catalyst were commonly applied for the dehydrogenative oxidation [4,5,6,7]. Recently, nickel complexes were also reported [8]. In heterogeneous dehydrogenative catalytic systems, mainly noble metals catalysts were reported: alumina-supported silver cluster, hydrotalcite supported silver nanoparticles, ceria supported gold nanoparticles, alumina-supported Pt nanoparticles, alumina-supported Re nanoparticles [9,10,11,12,13]. Particularly, high reactivity was reported for the oxidation of aliphatic alcohols in anaerobic condition [12]. While aerobic oxidation generates water as byproduct, anaerobic conditions produce hydrogen gas. Therefore, the dehydrogenative oxidation prevents the deactivation of catalysts by water and is safely applied to substrates containing water sensitive functional groups.

Nanocrystalline TiO₂ has been intensively investigated in a wide range research fields due to its practical applications, such as photocatalysts, solar cells, materials for water or air purification, and so on [14,15,16,17]. It is the chemical, physical, and photo-stability of TiO₂ that all applications mentioned above make possible. Therefore, besides the catalytic properties of TiO₂, it can be a suitable material for supporting catalytic active centers in heterogeneous catalysts. Particularly, TiO₂ supported ruthenium catalysts were utilized for diverse oxidation reactions including aerobic oxidation [18,19], catalytic wet air oxidation of various chemicals [20,21,22], and electro-oxidation of alcohols [23]. However, to the best of our knowledge, ruthenium incorporated TiO₂ has not been applied to the oxidation of alcohols under oxidant-free conditions. Here we report the dehydrogenative oxidation of benzylic, allylic, and aliphatic alcohols into their corresponding aldehydes using ruthenium incorporated TiO₂ under oxidant-free conditions. We found that benzylic and allylic alcohols exhibited enhanced catalytic activity compared to aliphatic alcohol. Particularly, our homemade pure TiO₂ is a strong Lewis acid catalyst enough to form the Lewis acid-catalyzed Friedel-Craft alkylation products in a benzyl alcohol reaction. Therefore, the dehydrogenative oxidation pathway is accomplished by the synergistic catalysis of the Lewis acidic TiO₂ surface and the highly dispersed ruthenium species acting as the hydride elimination of α-carbon.

2. Results and Discussion

2.1. Catalyst Characterization

Figure 1 shows the crystal phases of ruthenium incorporated titanium oxides (Ru_xTiO₂, x represents the molar ratio of Ru with respect to Ti) obtained by hydrolysis of Ti(SO₄)₂ in aqueous solution with the different amount of RuCl₃. The pure TiO₂ and Ru_xTiO₂ represent the anatase phase of TiO₂ (JCPDS data file No. 00-21-1272). It was reported that TiO₂ prepared by a hydrothermal method using Ti(SO₄)₂ as a precursor exhibited the anatase phase [24]. The reflection peaks originated from RuO₂ are not presented for all the samples obtained from a hydrothermal condition (160 °C). After annealing the Ru_0.07TiO₂ sample (A-Ru_0.07TiO₂) at 700 °C for six hours, we find the appearance of Bragg’s reflection of RuO₂ (JCPDS file No. 00-040-1290) and phase transformation of TiO₂ from anatase to rutile. This result suggests that ruthenium atoms are randomly incorporated into the crystallite of TiO₂ before annealing. However, the phase segregation of TiO₂ (rutile) and RuO₂ takes place at an elevated temperature. The crystallite sizes of samples in

Table 1. The physical properties of the samples.

Samples	Size (nm)		Surface Area	Ru Content
	XRD	TEM	(m2/g)	(μmol/g)
Pure TiO2	26.4	45.7 ± 5.3	48	-
Ru0.01TiO2	13.2	16.7 ± 0.9	104	9
Ru0.03TiO2	13.3	16.9 ± 1.0	111	24
Ru0.05TiO2	13.4	17.0 ± 1.4	125	37
Ru0.07TiO2	14.7	17.3 ± 1.1	126	51

was calculated from X-ray line broadening by the Scherrer equation (D = 0.89λ/(βcos θ)), where D, λ, β, and θ represent the average crystal size, the Cu Kα wavelength (0.15406 nm), the full-width at half-maximum, and the diffraction angle, respectively. The calculated size of the pure TiO₂ is 26.4 nm which is bigger than that of Ru_xTiO₂ (13–15 nm). The surface area of the pure TiO₂ obtained by the BET method is approximately two-times smaller than that of Ru_xTiO₂ (Table 1).

The elemental surface analyses of samples were done by X-ray photoelectron spectroscopy (XPS). A survey spectrum verify the presence of titanium, oxygen, ruthenium, and carbon contaminant without other species (Figure S1). The Ru 3d XPS spectra of Ru_0.07TiO₂ is shown in Figure 2. It is mainly composed of two parts which corresponds to the Ru 3d_5/2 (280.7 eV) and Ru 3d_3/2 (284.6 eV) band originated from spin-orbital coupling. Ru_0.07TiO₂ was synthesized under highly acidic aqueous conditions because no base was added in the hydrolysis of Ti(SO₄)₂. Therefore, we attributed the binding energy of 3d_5/2 (280.7 eV) to the Ru species with +4 oxidation state considering the above mentioned synthetic condition and previous literature data [25,26,27]. It was known that the Ru 3d_3/2 band overlaps with carbon 1s band and therefore, the Ru 3d_5/2 band is used for quantitative analysis [23,28]. The amount of Ru in Ru_0.07TiO₂ is 0.58 atomic percent and the ratio of Ru to Ti is 0.026 which is smaller than the added molar ratio (0.07) (In quantitative analysis, the relative sensitivity factor (RSF) of Ti 2p and Ru 3d_5/2 is 2.077 and 2.705, respectively.). The XPS peak intensities of 3d Ru of other Ru_xHAP are too weak to analyze quantitatively.

The morphology of samples was characterized by transmission electron microscopy (Figure 3). The pure TiO₂ is similar to corn-shape (Figure 3a). The average size of pure TiO₂ is 45.7 ± 5.3 nm, standard deviation. The shape of Ru_xTiO₂ is similar to pure TiO₂, however, the average size is smaller, approximately 16–27 nm (Figure 3b,e). This result agrees with the crystallite size calculated by Scherrer’s equation based on the full width at half maximum (FWHM) of the (101) facet of anatase. It was reported that the addition of metal ions—such as Na⁺, Zn²⁺, and Fe³⁺—influence the crystal phase resulting in a decrease of the crystal size of TiO₂ [29,30,31,32]. We believe that incorporation of Ru in the TiO₂ lattice causes lattice deformation which might inhibit the crystals growth of TiO₂. After annealing the Ru_0.07TiO₂ at 700 °C, several crystallites aggregate with each other to be the size of a few hundred nanometers (Figure 3f and Figure S2) with rutile phase. Particularly, dark spots (red circle in Figure 3f) are embedded in the crystallite. These dark spots seem to be RuO₂ crystallites observed in XRD data.

The elemental mapping of Ru_0.07TiO₂ was investigated by energy dispersive X-ray spectroscopy (Figure 4). Individual mapping of Ti, O, and Ru is shown Figure 4b–d, respectively. Figure 4e represents the elemental mapping with Ti, O, and Ru overlaid. We found that Ru is homogeneously dispersed within TiO₂. This result is in good accordance with the previous XRD data of samples in which the reflection peak from RuO₂ does not observed before annealing Ru_0.07TiO₂. In addition, high resolution TEM image also showed no aggregation of Ru species (Figure 4f). Therefore, combining XRD and TEM data, we conclude that ruthenium species are highly dispersed in TiO₂ in monomeric form.

2.2. Catalytic Tests

Catalytic properties were tested by the oxidation reaction of alcohols (benzyl alcohol, cinnamyl alcohol, and 1-octanol) under oxidant-free condition (

Table 2. Catalytic properties of alcohol oxidation under oxidant-free conditions ¹.

Entry	Catalysts	Substrates	Products	Conversion (%)	Yield to Aldehyde (%)	Selectivity (%)
1	None			0	-	-
2	Degussa, P25			2	1	60
3	pure TiO2			49	1	2
4	Ru0.01TiO2			41	34	86
5	Ru0.03TiO2			79	72	93
6	Ru0.05TiO2			88	82	93
7	Ru0.07TiO2			>99	92	93
8	Reuse 7			99	92	93
9	Reuse 8			98	90	92
10	A-Ru0.07TiO2			2	2	99
11	None			0	-	-
12	Degussa, P25			3	3	99
13	pure TiO2			44	43	98
14	Ru0.01TiO2			78	78	>99
15	Ru0.03TiO2			85	84	>99
16	Ru0.05TiO2			96	94	98
17	Ru0.07TiO2			>99	98	99
18	None			0	-	-
19	Degussa, P25			3	2	93
20	pure TiO2			4	4	98
21	Ru0.01TiO2			12	12	>99
22	Ru0.03TiO2			23	22	97
23	Ru0.05TiO2			23	22	>99
24	Ru0.07TiO2			24	23	98

¹ Reaction conditions: substrate amount: 0.04 mmol, catalyst amount: 5 mg, reaction time: 3 h, Conversion is calculated by an internal standard method. Yield to aldehyde is calculated by calibration curves (Figure S4). Selectivity is calculated by the peak area of aldehyde divided by the total peak area of product.

). The entries 1–10 showed the oxidative catalytic activities of benzyl alcohol. Compared to the short reaction period (30 min) under O₂ atmosphere (

Table 3. Catalytic properties of alcohol oxidation using O₂ as an oxidant.

Entry	Catalysts	Substrates	Products	Conversion (%)	Selectivity (%)
1	None			0	-
2	Degussa, P25			3	>99
3	pure TiO2			3	>99
4	Ru0.01TiO2			38	>99
5	Ru0.03TiO2			77	>99
6	Ru0.05TiO2			94	>99
7	Ru0.07TiO2			>99	>99

), reaction time was adjusted to 3 h in N₂ environment. The conversion was less than 2% in the absence of catalyst or in the presence of Degussa, P25 (entries 1 and 2). However, the conversion of benzyl alcohol with homemade pure TiO₂ (entry 3) was 49% which is high compared to the commercial Degussa P25 (Degussa, Frankfurt, Germany) (entry 2). Importantly, the conversion into benzaldehyde is only 2%. Instead, o, m, p-benzyltoluenes (the molar ratio of ortho/meta/para = 43:6:51) and dibenzyl ether were produced as the major products. These o, m, p-benzyltoluenes were produced by the acid catalyzed Friedel-Crafts alkylation of toluene (solvent) with benzyl alcohol (Figure S3). Dibenzyl ether is produced by the homo-coupling of benzyl alcohol. Both reactions are proceeded by the same intermediate that is benzylic carbocation as indicated in the Scheme 1. Therefore, the surface of homemade pure TiO₂ acts as a Lewis acid catalyst. It was reported that TiO₂ and sulfated TiO₂ are applied to Friedel-Crafts acylation or Beckman rearrangement in which these solid catalysts act as Lewis acid catalysts [33,34]. In the case of Ru_xTiO₂, the major product is benzaldehyde, however, still approximately 7% of benzyl toluene and benzyl ether are remained (entries 4–9). As the Ru content increases, the catalytic activity increases (entries 4–7 in Table 2). For example, Ru_0.07TiO₂ exhibits 98% conversion with 93% selectivity within 3 h (entry 7). In addition, Ru_0.07TiO₂ exhibited the catalytic activity to be retained without significant loss for up to three cycles (entries 7–9). However, after annealing Ru_0.07TiO₂ at 700 °C, the catalytic activity decreased dramatically (entry 10), and the Lewis acid catalyzed products (o, m, p-benzyltoluenes and dibenzyl ether) were not produced by A-Ru_0.07TiO₂ in the same reaction conditions. Based on the previous mechanistic studies using the single crystal of TiO₂ (typically rutile (110) surface) in vacuum conditions, the reactivity of alcohol on TiO₂ is strongly dependent on the surface defects of TiO₂, such as oxygen vacant sites [35,36,37]. Besides the reduction of surface area and the aggregation of Ru into RuO₂ during annealing process, the decrease of the surface defects after annealing at 700 °C might result in the low reactivity of A-Ru_0.07TiO₂.

Ru_xTiO₂ can be successfully applied oxidation of allylic alcohol (cinnamyl alcohol) into corresponding a α,β-unsaturated carbonyl compound (cinnamaldehyde) with high selectivity (>98%). In the case of cinnamyl alcohol, the by-products from the electrophilic aromatic substitution or acid catalyzed homo-coupling were not produced (entries 14–17 in Table 2). Similar to benzyl alcohol oxidation, homemade pure TiO₂ showed the high conversion compared to Degussa, P25 (entries 12 and 13 in Table 2). In addition, pure TiO₂ exhibits high selectivity in transformation into the cinnamaldehyde (entry 13 in Table 2). We want to point out that the conversion of alcohol into aldehyde in our catalytic system is not a common phenomenon because the dehydration of various alcohols into alkenes is a more general reaction on TiO₂ surface instead of dehydrogenation reaction [35,38]. Unfortunately, the kinetics of primary aliphatic alcohol (1-octanol) were not as efficient as that of benzyl alcohol and cinnamyl alcohol in the given conditions (entries 19–24 in Table 2). In the case of the oxidation of secondary alcohol (1-phenyl ethanol), the major product is styrene instead of acetophenone, which indicates that secondary benzylic alcohols are prone to elimination under the same experimental conditions (Table S1).

In our initial experiment set-up, the role of TiO₂ is catalytic supports to incorporate ruthenium atoms that are catalytic active centers. However, our homemade TiO₂ itself significantly involves in the reaction (entries 3 and 13 in Table 2). For the conversion of alcohol into aldehyde, two hydrogen atoms (one is attached to oxygen, the other is attached to α-carbon) are detached. A previous high resolution STM study verified that adsorption process of alcohol on TiO₂ surface induces the detachment of the hydroxyl hydrogen in which the regular surface Ti sites of TiO₂ enable alcohols to be adsorbed molecularly and dissociatively (deprotonated alkoxide) at room temperature [37]. Particularly, our homemade highly acidic surface of pure TiO₂ facilitates alcohol adsorption on the surface to form deprotonated alkoxide species. The hydrogen elimination of the α-carbon seems to take place on exclusively ruthenium species (Figure 5). This interpretation is supported by a dramatic change of major products from o, m, p-benzyltoluenes to benzaldehyde when pure TiO₂ is replaced Ru_0.01TiO₂ (entries 3 and 4 in Table 2). Similarly, Yamaguchi et al. proposed that in an alcohol oxidation reaction on Ru_xTiO₂ using molecular oxygen as oxidant, the reaction was accomplished by the ruthenium alcoholate formation followed by hydride elimination and the rate determining step is the hydride abstraction step [18].

The turnover frequency (TOF) based on Ru for benzyl alcohol oxidation using Ru_0.01TiO₂ amounts to 130 h⁻¹ where the total amount of ruthenium obtained from Inductively coupled plasma mass spectrometry (ICP) data are used for TOF calculation. In comparison with other heterogeneous catalytic systems in oxidant-free conditions (Au/CeO₂, (TOF = 125 h⁻¹) [7], Ag/Al₂O₃ (>2 h⁻¹) [9], Ru/AlO(OH) (7 h⁻¹) [39], Cu/hydrotalcite (1.5 h⁻¹) [40], Cu/Al₂O₃ (8.3 h⁻¹) [41]), our value is higher except for the silver nanoparticles on the hydrotalcite system (590 h⁻¹) [11].

3. Materials and Methods

3.1. Synthesis of TiO₂ and Ruthenium-Incorporated TiO₂ (Ru_xTiO₂)

Ti(SO₄)₂ and RuCl₃ nH₂O were purchased from Kanto and Sigma-Aldrich, respectively. TiO₂ and Ru_xTiO₂ were synthesized by a hydrothermal method using Ti(SO₄)₂ and RuCl₃ nH₂O as the precursor of Ti and Ru, respectively. 15 mL of 24% Ti(SO₄)₂ (Ti: 20 mmol) was added dropwise into the 60 mL of aqueous solution containing the different amount of RuCl₃ (Ru: 0, 0.2, 0.6, 1, 1.4 mmol) under vigorous stirring. The samples are labeled Ru_xTiO₂, in which x represents the mole ratio of ruthenium with respect to titanium. The solution was transferred to a Teflon-lined autoclave reactor and treated at 160 °C for 6 h. After the reaction vessel was cooled down to room temperature, the samples were separated by centrifugation and purified by washing with distilled water three times. Finally, the precipitates were obtained by drying under vacuum at room temperature. A-Ru_0.07TiO₂ was prepared by annealing the Ru_0.07TiO₂ at 700 °C for 6 h.

3.2. Characterization of Catalysts

X-ray diffraction (XRD) patterns were obtained using a Bruker D8 Advance X-ray diffractometer with Cu Kα radiation (Bruker, Karlsruhe, Germany). Diffraction patterns were taken over the 2θ range 20°–75°. The XPS results were obtained on a PHI 5000 Versaprobe instrument (ULVAC-PHI Inc., Chigasaki, Japan) in which monochromated Al Kα radiation (hν = 1486.6 eV) is illuminated as the light source. The X-ray anode was run at 25 W and the voltage was maintained at 15 kV. The pass energy was fixed at 23.5 eV to ensure sufficient resolution to determine peak positions accurately. The binding energies were calibrated by using the Au 4f_7/2 signal at 84.0 eV. The prepared sample morphology was characterized by a TITAN G2 60-300 (FEI, Hillsboro, OR, USA) transmission electron microscopy (TEM) equipped with CETCOR corrector and 4-silicon drift detector system energy dispersive X-ray spectroscopy (Super-X EDS, Bruker) operating at an accelerating voltage of 80 kV. The quantitative EDS area mapping was carried out using a scanning electron probe (Bruker) with a size of 1 nm over the sample. The surface area of samples was measured by Brunauer-Emmett-Teller (BET) method. N₂ adsorption-desorption isotherms of samples at liquid nitrogen temperature (77 K) were obtained using Belsorp Mini II (BEL Japan, Osaka, Japan).

3.3. Catalytic Tests

2 mL of alcohols (benzyl alcohol, 1-octanol, cynnamyl alcohol, 20 mM) dissolved in toluene was added to 16 mL test tubes containing 5 mg of catalysts. To ensure oxygen-free conditions, nitrogen gas was flowed for 20 min in the reaction mixture for degassing, followed by connection of an N₂ balloon that was connected to a syringe needle, which was inserted into the septum covering the test tube. The reaction mixture was heated 105 °C for 3 h. After cooling the reaction mixture to room temperature, the solid catalysts were separated from solution phase by centrifugation. The solution phase was analyzed by gas chromatography (GC, aglient 7890A, Agilent Technologies, Santa Clara, CA, USA) with a flame ionization detector (FID). Dodecane was used as an internal standard for the calculation of conversion. The catalytic property of the commercial Degussa P25 was also tested for comparison.

4. Conclusions

In summary, we prepared the highly dispersed ruthenium incorporated TiO₂ by a one-pot hydrothermal method. Ru_xTiO₂ can transfer the benzylic, allylic, and alkyl alcohols into corresponding carbonyl compounds in oxidant-free conditions. In our catalytic system, particularly, TiO₂ acts as not only the catalytic support but also a Lewis acid catalytic site which improves reactivity by facilitating the deprotonated adsorption of alcohols on the catalysts. Therefore, the synergistic effect of hydride elimination by ruthenium and the Lewis acid site on TiO₂ can efficiently transfer primary alcohols into corresponding aldehydes in oxidant-free conditions.