High H₂O₂ yield in the direct oxidation of H₂ with O₂ on mono dispersed Pd–Au nano colloid under pressurized conditions

Abstract

Pd–Au (75 : 25) nano colloid was prepared by using various reductants and it was found that the mono dispersed Pd–Au nano colloid with 4.7 nm in diameter was prepared by reduction with oxalic acid and this Pd–Au nano colloid is highly active for the synthesis of H₂O₂ by direct oxidation of H₂ with O₂. The yield of H₂O₂ increased with increasing reaction pressure and at 1 MPa, H₂O₂ yield achieved a value of 30%. Under these conditions, H₂ conversion is as high as 80%. H₂O₂ yield was further improved by increasing H₂ partial pressure, and the optimum H₂ partial pressure was 10% at 1 MPa because of safety (out of the explosion limit concentration in reactor) and the formation rate of H₂O₂. Because H₂O₂ decomposition is prevented, H₂O₂ yield also increased with decreasing reaction temperature and at 268 K, H₂O₂ yield and production rate are achieved as high as 46% and 25 mmol h⁻¹, respectively, in aqueous solution. H₂O₂ concentration reaches a value of 5.8% in aqueous solution after 20 h at 268 K.

---

Introduction

Demands for hydrogen peroxide (H₂O₂) are currently increasing, with the increasing importance of green chemistry. Therefore, the market for H₂O₂ is expanding in various fields, like breaching. In particular, H₂O₂ is expected to become an important oxidant for use in catalytic reactions,^1–5 such as the epoxidation of olefins. At present, H₂O₂ is synthesized by the so-called anthraquinone method, which consists of anthraquinone hydrogenation followed by the auto-thermal oxidation of hydroanthraquinone.^6,7 This process is useful and efficient for production of H₂O₂ only at a large scale. On the other hand, the direct synthesis of H₂O₂ from gaseous hydrogen and oxygen is attracting much interest as a simple H₂O₂ production method, in particular, for small scale production, because of the high production rate and simple process. Studies of various catalysts, particularly Pd- and Pt-based catalysts, have been conducted.^8–24 Lunsford et al. reported the high activity of colloidal Pd,^8–11 and Edwards et al. reported that the addition of Au to Pd improves the yield and that high productivity is achieved using TiO₂ (P-25), which is a mixed phase of rutile and anatase, as a support.^12–14 We have also reported that a rutile-type TiO₂ support leads to a high H₂O₂ yield in the direct oxidation of H₂ on a Pd–Au bimetallic catalyst under atmospheric pressure. In addition, we found that Pd–Au nano colloid at Pd : Au = 85 : 15 exhibits high activity for H₂O₂ formation rate¹⁵ since a sharp particle size distribution and a large metal surface area can be achieved.²⁵ In this study, the effects of the type of reductant for Pd–Au nano colloid preparation on H₂O₂ synthesis were investigated. Optimization of reaction conditions is also performed. Although high H₂O₂ yield is reported by using methanol for catalyst suspension solvent,^26,27 separation of H₂O₂ from solvent containing methanol is highly difficult and so high yield of H₂O₂ formation in aqueous solution is highly important and so we used water for solvent in spite of low H₂ and O₂ dissolved amount.

Experiment

Pd–Au bimetallic colloid was prepared by the chemical reduction of HAuCl₄ and PdCl₂ in a mixture of 1 ml of 5.6 M HCl solution, 25 ml of H₂O, 25 ml of C₂H₅OH and 1 g of oxalic acid or ca. 0.25 mmol of hydrazine (namely 1 ml of 0.8 vol% hydrazine). The total amount of Pd and Au was maintained at 50 μmol and the Pd : Au ratio was maintained at 75 : 25 mol% for the catalyst. In order to mono dispersion of the colloid, 0.43 g of polyvinylpyrrolidone (PVP, Kishida Pure Chem. Co. Ltd.) was added for preventing aggregation of nano colloid. Since the amount of PVP as dispersion reagents is insufficient, aggregation of Pd–Au nano colloid occurs easily and the catalytic activity is significantly decreased. However, it is confirmed that the H₂O₂ formation rate is hardly dependent on the amount of PVP when the amount of PVP is larger than 0.4 g. The aqueous solution was refluxed at 353 K for 60 min under atmospheric conditions with a magnetic stirring mixer and the solution gradually changed to a brownish red, which indicated the formation of nano colloid particles. The nano colloidal catalyst was kept in liquid phase and when reaction activity was tested, the liquid phase catalyst was put into the reaction solution to be 17.8 mg in total amount of Pd–Au nano colloid with colloidal solution, unless otherwise noted.

The synthesis of H₂O₂ from a gaseous mixture of H₂ and O₂ (1 : 9 ratio in general) was tested in a stainless steel autoclave reactor with magnetic stirring at 283 K, 1 MPa. This reactor was charged with an aqueous solution containing a catalyst comprising 17.8 mg in total concentration of Pd and Au, 84 mM of NaCl and 0.368 M of H₂SO₄ for experiments performed under all conditions used. In addition, a gaseous mixture of H₂–O₂–N₂ (2.5 : 19.5 : 78 volume ratio, assuming the H₂ diluted with air) was mainly fed through a porous glass filter (pore size, 10 μm in diameter) and the catalyst suspension was stirred mechanically at 800 rpm using a motor. It is also noted that the hydrogen partial pressure used (2.5%) is smaller than that of the explosion limit (4.5%). A back pressure valve (TESCOM type 2500) was used for pressurizing the reactor and a thermal flow controller were used for gas flow rate control. The reactor (Tiatsu Glass Co. Ltd.) was immersed in temperature-controlled water (283 K) using a cryostat (Taitec CR-80R). Since H₂O₂ yield decreased with reaction time, we used the reaction data after 2 h reaction for comparison unless otherwise mentioned.

The amount of H₂O₂ formed was analyzed by the UV absorption method, in which TiO(SO₄)₂ was used as the pigment. The amounts of gaseous H₂ and O₂ were measured with a TCD gas chromatograph (Shimadzu GC 8A). The selectivity of H₂ to H₂O₂ was defined as the H₂O₂ formation rate divided by the H₂ consumption rate. Samples for XPS measurement were prepared by drying the colloidal solution on a clean glass plate at 333 K and the commercial equipment (Shimadzu type 165). It was confirmed no change in surface composition with XPS and no elution of Pd and Au during the reaction over 5 h with ICP measurements. TEM observation was performed by using FEI Tecnai type G2 with 200 kV accelerated voltage and around 200 particles of Pd–Au colloid were used for estimating the particle size distribution.

Results and discussion

H₂O₂ synthesis at ambient pressure

Fig. 1 shows TEM images of the Pd–Au nano colloid obtained by using oxalic acid (a), hydrazine (b), and NaBH₄ (c) as reductant. As shown in Fig. 1(a), Pd–Au colloid with a particle size of 1–7 nm was obtained when oxalic acid was used for the reductant. However, finer (1–4 nm) metal particles were obtained using hydrazine as the reductant (Fig. 1(b)). In contrast, slightly larger but similar size particles to those of oxalic acid preparation are obtained by using NaBH₄ as reductant (Fig. 1(c)). The particles obtained by hydrazine reductant were uniform in the fine size area (1–4 nm), but some of particles are aggregated and so a bimodal distribution is observed. Fig. 1(d) shows the lattice image of the nano colloid obtained with oxalic acid reductant. Considering the lattice constant of Pd–Au particles (0.228 nm), which is intermediate between Pd and Au, Pd–Au alloy seems to form in the Pd–Au nano colloids prepared using both of these reductants. The surface composition of the obtained Pd–Au nano colloids was also analyzed quantitatively by XPS using a mixture of Pd and Au powders as a standard and the results are shown in Table 1. Note that the surface of the Pd–Au nano colloids is slightly rich in Pd compared with that of the bulk. Considering the clear lattice image is observed, it seems that Pd–Au nano colloid is a functionally graded structure. Such enrichment of Pd has also been reported for a Pd–Au bimetallic supported on TiO₂ or Al₂O₃.²⁸ Compared with the surface composition of the colloid obtained using the oxalic acid as reductant, that of the Pd–Au colloid obtained by the hydrazine reductant is slightly rich in Au. In contrast, in the case of NaBH₄ reductant, the surface is enriched with Au. According to the quantum calculations,^29,30 a surface composition of Pd : Au = 3 : 1 is ideal for H₂O₂ synthesis by direct oxidation of H₂. This is because dissociation of the O–O bond is prevented and desorption of H₂O₂ formed is accelerated with Au. Therefore, an excessively enriched composition with Au or Pd is not suitable for H₂O₂ synthesis.

Fig. 1 TEM images and particle distribution of Pd–Au nano colloids obtained using oxalic acid (a), hydrazine (b) and NaBH₄ (c) as reductants. (d) A high resolution image of nano colloid obtained by oxalic acid.

Table 1 Effects of type of reductant for preparation of Pd–Au (75 : 25) nano colloids on results of H₂O₂ synthesis by H₂ oxidation and surface composition

Reductant	H2 Conversion/%	H2O2 Formation				Surface Pd amount/mol (%)
		Rate/mmol h^−1	Productivity mmol/(h g-cat)	Selectivity^b/%	Yield/%
Pressure, atmospheric pressure. Glass reactor, flow rates: H2, 50 ml min^−1; O2, 50 ml min^−1; H2SO4, 0.368 M; NaCl, 84 mM; amount of Pd–Au catalyst, 23.9 mg; volume of water, 100 ml. Temperature: 283 K.a Pd–Au/rutile TiO2: Pd : Au = 82 : 18 molar ratio (Pd–Au 25.5mg).b Hydrogen base value.
C2H2O4	6.1	5.96	249	99.6	6.1	0.84
N2H4	7.7	7.58	317	78.5	6.1	0.79
NaBH4	8.6	4.04	169	37.9	3.3	0.68
Rutile-TiO2 supported^a	2.9	3.60	141	99.7	2.9	—

---

The H₂O₂ formation rates, H₂O₂ concentrations for 2 h accumulation, H₂O₂ selectivity, and H₂ conversion rates are summarized in Table 1 for the Pd–Au nano colloids prepared using three different reductants. For this experiment, we used a glass reactor and slightly larger amount of the calayst (23.9 mg) was used. In Table 1, the H₂O₂ synthesis results on rutile-TiO₂-supported Pd–Au are also listed for comparison. Although the optimized composition of Pd–Au for H₂O₂ synthesis differed between the nano colloid and the TiO₂-supported one, all Pd–Au nano colloid catalysts showed much higher H₂O₂ formation rates and also higher H₂ conversion rates than the TiO₂-supported one. Therefore, it is clear that a high H₂O₂ formation rate can be obtained using Pd–Au nano colloids. Moreover, H₂O₂ formation activity clearly varies markedly depending on the type of reductant used for the colloid preparation, despite the use of the same type and ratio of starting materials. H₂O₂ formation rate increases in the following order of reductants used: hydrazine > oxalic acid > NaBH₄. Although the high selectivity is exhibited on Pd–Au nano colloid obtained by oxalic acid, H₂O₂ formation rate, namely, H₂O₂ yield, is important considering the difficulty in recycling the reactant when air is used for H₂O₂ synthesis. There are two possible reasons for the difference by reductant for the preparation of the colloid: one is a difference in particle size resulting in a difference in surface area and the other is a difference in surface composition. As shown in Fig. 1, smaller particles tend to be obtained by using hydrazine than oxalic acid. In addition, a slightly higher concentration of Au on the surface is also effective for suppressing H₂O₂ decomposition. However, in the case of NaBH₄, Au concentration becomes excessively high, thereby decreasing the H₂O₂ formation rate. This is because Au has a low activity for H₂O₂ formation. Therefore, among those reductants examined, Pd–Au nano colloid prepared by using hydrazine has a high activity for H₂O₂ formation by H₂ direct oxidation under the high H₂ concentration condition, which is a problem for safety.

Although PVP was used as a dispersion agent for the Pd–Au nano colloid, it is expected that PVP is also effective for preventing H₂O₂ decomposition. Effects of PVP on H₂O₂ formation rate were further studied (Fig. 2) by using the nano colloid prepared by oxalic acid. Evidently, H₂O₂ formation rate as well as selectivity increased with increasing PVP concentration and saturated at a PVP concentration higher than 40 μM. Under these conditions, H₂O₂ selectivity is close to 100%. This suggests that surface PVP may also have a function of preventing the decomposition of formed H₂O₂. This is because the H₂O₂ synthesis catalyst is also an active H₂O₂ decomposition catalyst. In addition, a slight excess amount of PVP is also effective for preventing the aggregation of the nano colloid during the reaction. Since the saturated value was achieved and the mono dispersion of Pd–Au nano colloid is stably sustained under reaction conditions, a PVP concentration of 40 μM is used in the following part of this study.

Fig. 2 H₂O₂ formation rate as a function of polyvinylpyrrolidone (PVP) dispersion concentration. Reaction conditions: 283 K, 0.1 MPa, H₂: 50 vol%, O₂:5 0 vol%, total flow rate: 100 ml min⁻¹.

Fig. 3 shows H₂O₂ concentration as a function of reaction time for the Pd–Au nano colloid prepared by using the hydrazine reductant under various H₂ and O₂ concentrations at atmospheric pressure. A low concentration of H₂O₂ is the most significant drawbacks of the direct synthesis of H₂O₂ from H₂ oxidation. In the most reports, H₂O₂ concentration is smaller than 0.5% when water is used as the solvent for catalyst suspension, while much higher H₂O₂ concentration is required for application of breaching etc., at least 5 wt%. The H₂O₂ formation rate increases with increasing H₂ concentration, and the H₂O₂ accumulated amount is also strongly dependent on the H₂/O₂ ratio. H₂O₂ concentration tends to increase with decreasing H₂/O₂ ratio. At H₂ : O₂ = 1 : 5, H₂O₂ concentration reaches a value as high as 2 wt.% in spite of atmospheric pressure and water for solvent. Since the H₂O₂ decomposition rate increases with increasing H₂O₂ concentration, accumulation of a high concentration of H₂O₂ means a low H₂O₂ decomposition rate. In reply to the improved decomposition rate of H₂O₂, the selectivity of H₂O₂ is high (ca. 80%) in the initial 2 h but decreased drastically with reaction period.

Fig. 3 H₂O₂ concentration as a function of reaction time for the Pd–Au nano colloid prepared by using the hydrazine reductant under various H₂ and O₂ concentrations. Reaction condition: 283 K, 0.1 MPa, total gas flow: 100 ml min⁻¹ and no dilute gas.

Fig. 4 shows the H₂O₂ formation rate, H₂O₂ yield and H₂ conversion for the Pd–Au colloid prepared using the hydrazine reductant as a function of H₂ concentration. Since the H₂ concentration in Table 1 is high and an explosion problem is anticipated, a high H₂O₂ formation rate at diluted H₂ concentration is strongly requested from safety issues. H₂O₂ formation rate monotonously decreased with decreasing H₂ concentration (C_H2) from 50 to 0.5 vol%, suggesting that reaction is mass transfer limited one. In contrast, H₂ conversion increased with decreasing H₂ concentration. Hence, the H₂O₂ formation rate and the H₂ conversion rate showed the opposite dependency on hydrogen concentration (C_H2). Since H₂ conversion increased, H₂O₂ yield also increased with decreasing C_H2. At a C_H2 of 0.5 vol%, H₂ conversion and H₂O₂ yield as high as 32% and 30% were achieved, even at atmospheric pressure. It is also noted that selectivity of H₂O₂ is as high as 93% at this C_H2 and is always higher than 90%. Since the amount of hydrogen dissolved in water decreases with decreasing C_H2, an increase in H₂ conversion rate with decreasing C_H2 is opposite to the expected result. In fact, in the case of Pd–Au/rutile TiO₂, H₂ conversion the H₂O₂ formation rate decreased with decreasing C_H2. At present, the reason Pd–Au nano colloids exhibit a small dependency on hydrogen partial pressure (r_H2O2 ∝ P_H2^0.85) remains unclear comparing with that of TiO₂ one (r_H2O2 ∝ P_H2^1.35); however, it seems that the high surface activity of Pd–Au nano colloids is related to such observations and that a high activity for hydrogen activation is sustained under a low hydrogen concentration.

Fig. 4 H₂O₂ formation rate, H₂O₂ yield and H₂ conversion for the Pd–Au colloid prepared using the hydrazine reductant as a function of H₂ concentration. Reaction condition: 283 K, 0.1 MPa, H₂: X vol%, O₂: 50 vol%, N₂: 50 − X vol%. Total flow rate 100 ml min⁻¹.

H₂O₂ synthesis under pressurized conditions

Table 2 summarizes the results of H₂O₂ synthesis on a Pd–Au nano colloid catalyst prepared using the hydrazine or oxalic acid as reductant under 1 MPa pressurized conditions. For pressurization, the reactor was changed from a glass reactor to a stainless steel autoclave. In Table 2, comparison of H₂O₂ formation rate in a glass reactor and stainless reactor was also shown under atmospheric pressure. H₂O₂ formation rate and H₂O₂ selectivity decreased markedly at 0.1 MPa with the change of the reactor because of H₂O₂ decomposition on the stainless steel parts in the reactor, such as the paddle, and the formation of a much larger bubble of the reactant gas mixture. In fact, we confirmed that H₂O₂ decomposition occurred rapidly in an autoclave reactor under H₂ flow. However, as shown in Table 2, an increase in reaction pressure is highly effective for improving the H₂O₂ formation rate because of the increased amount of hydrogen dissolved in water, and H₂ conversion becomes higher than 95% at 283 K. In contrast, H₂O₂ selectivity decreased markedly with the change of the reactor (from 95 to 38%) and slightly with pressurization (from 38 to 33%). Since the oxidation of hydrogen by H₂O₂ occurs with increasing hydrogen concentration, decrease in H₂O₂ selectivity at 1 MPa could be explained by the improved H₂O₂ decomposition rate as a result of the increased amount of hydrogen in water. Because the H₂ conversion is as high as 95% at 1 MPa on Pd–Au nano colloid, it is expected that a high H₂O₂ yield will be achieved by the suppression of H₂O₂ decomposition under pressurized conditions. On the other hand, a slightly higher selectivity is achieved on Pd–Au nano colloid obtained using the oxalic acid reductant at 1 MPa because of the low surface activity for H₂O₂ decomposition; thus a H₂O₂ yield higher than 40% is obtained at 1 MPa. At present, the reason why Pd–Au prepared with oxalic acid reductant shows slightly higher selectivity to H₂O₂ at 1 MPa is not clear, however, mono dispersion of metal particles seems to effectively work for synthesis of H₂O₂ selectively at high pressure. H₂O₂ synthesis on Pd–Au nano colloid obtained by oxalic acid under pressurized conditions was studied in detail in the following part of this study.

Table 2 Effects of pressure and reductant for catalyst preparation on H₂O₂ synthesis on Pd–Au nano colloids

Catalyst	H2 Conversion/%	H2 Conversion/%	H2O2 Formation
			Rate/mmol h^−1	Productivity mmol/(h g-cat)	Selectivity^a/%	Yield/%
Stainless steel autoclave reactor with PTFE inner vessel; Total flow rate: 100 ml min^−1, flow rates of each gas: H2, 1.25 ml min^−1; O2, 9.75 ml min^−1; N2, 15 ml min^−1; Ar, 24 ml min^−1; amount of catalyst: 17.9 mg; volume of water, 75 ml. H2SO4: 0.368 M, NaCl: 84 mM Pd–Au(O): Pd–Au obtained using oxalic acid reductant Pd–Au(H): Pd–Au obtained using hydrazine reductant. Rutile TiO2 supported: Pd–Au (Pd : Au = 82 : 18 molar ratio) supported on rutile TiO2.a H2O2 amount after 2-h reactionb Glass reactor; flow rates: H2, 2.5 ml min^−1; O2, 19.5 ml min^−1; Ar, 78 ml min^−1; volume of water, 100 ml.
Pd–Au(H)^b	0.1	30.7	1.53	92	95.4	29.0
Pd–Au(H)	0.1	60.2	0.67	54	37.5	22.6
Pd–Au(H)	1.0	96.6	0.98	78	33.1	32.0
Pd–Au(O)	1.0	84.4	1.24	92	48.0	32.9
Rutile TiO2 supported	1.0	54.2	0.92	64	56.0	30.3

---

Fig. 5 shows the effects of total flow rate on the formation rate, selectivity, and yield of H₂O₂ and also H₂ conversion on Pd–Au nano colloid prepared by oxalic acid. Evidently, high H₂ conversion which was close to 90% was exhibited on Pd–Au nano colloid and sustained up to the 233 ml min⁻¹ examined. This suggests that the surface of Pd–Au nano colloid is highly active for H₂ activation. With increasing flow rate, yield as well as selectivity was slightly decreased, however, the decrease is quite small. Therefore, H₂O₂ formation rate is monotonously increased with increasing total flow rate. At 233 ml min⁻¹, the H₂O₂ formation rate achieved a value of 4.5 mmol h⁻¹. Therefore, in the following part, we fixed 233 ml min⁻¹ for total gas flow rate, which is the maximum flow rate for the flow controlling system used in this study.

Fig. 5 Effects of total gas flow rate on H₂O₂ formation rate, selectivity, yield and H₂ conversion. Reaction condition; 283 K, H₂ 2.5%, O₂ 19.5%, N₂ 78%, 1.0 MPa, 800 rpm.

The effects of reaction pressure on H₂O₂ synthesis were further studied for increasing H₂O₂ formation rate by using Pd–Au nano colloids prepared with oxalic acid as the reductant. Here it is also noted that in Fig. 6, we used air composition for oxygen, which is also industrially important. Since the concentration of H₂ dissolved in water is small, mass transport of H₂ from the gas to water phase tends to be a rate determining step and an increase in the dissolved amount of H₂ into water seems to be effective for increasing the H₂O₂ formation rate. Fig. 6 shows the H₂ conversion, H₂O₂ formation rate, and H₂O₂ yield as a function of reaction pressure. In this experiment, H₂ partial pressure was fixed at 2.5%, which is lower than that of the explosion limit. Evidently, as expected, the H₂O₂ formation rate as well as H₂ conversion monotonously increased with increasing total pressure and at 1 MPa, the H₂O₂ formation rate achieved a value of ca. 4.5 mmol h⁻¹, which corresponds to 251 mmol h⁻¹ g-cat⁻¹. Comparing with the formation rate on Pd–Au/TiO₂ (160 mmol h⁻¹ g-cat⁻¹ in 66% CH₃OH solution) reported by Edwards et al.,^31,32 H₂O₂ formation rate of 4.5 mmol g⁻¹ achieved is higher even using H₂O solution, however, for industrial application, H₂O₂ yield is still not high enough.

Fig. 6 H₂ conversion, H₂O₂ formation rate, and H₂O₂ yield as a function of reaction pressure. 283 K, H₂ 2.5 vol%, O₂ 19.5 vol%, N₂ 78 vol%, total flow rate: 233 ml min⁻¹.

Fig. 7 shows the H₂O₂ formation rate, H₂O₂ selectivity, and H₂ conversion as a function of H₂ partial pressure in feed gas. As discussed, the dissolved amount of H₂ in water is an important factor for achieving high H₂O₂ yield, and so increase in H₂ partial pressure seems to be positive effects on H₂O₂ formation rate. As expected, H₂ formation rate monotonously increased with increasing H₂ partial pressure and at 10% P_H2, it achieved a value as high as ca. 17 mmol h⁻¹. In contrast, selectivity to H₂O₂ slightly decreased with increasing H₂ partial pressure because of the improved activity to H₂O₂ decomposition with increasing P_H2. However, even at 10% P_H2, selectivity to H₂O₂ is close to 40%, which is reasonably high comparing with that reported for a H₂O₂ direct synthesis catalyst. On the other hand, H₂ conversion is also increased by increasing H₂ partial pressure and at 10% H₂, H₂ conversion was close to 90%. Although the explosion limit of H₂ is around 4.5%, 90% of H₂ fed was converted and so on this Pd–Au nano colloid catalyst, it seems that H₂ partial pressure of 10% can be used within an explosion limit. Under this condition, H₂O₂ production rate and H₂O₂ yield are as high as 531 mmoh h⁻¹ g-cat⁻¹ and 33%, respectively.

Fig. 7 H₂O₂ formation rate, H₂O₂ selectivity, and H₂ conversion as a function of H₂ partial pressure in feed gas. 283 K, H₂: X vol%, O₂: 0.2 × (100 − X) vol%, N₂: 0.8 × (100 − X) vol%, total flow rate: 233 ml min⁻¹.

Fig. 8 shows the H₂O₂ concentration on Pd–Au colloid prepared with oxalic acid as a function of the reaction period. Another important issue for the direct synthesis method for H₂O₂ is the concentration of H₂O₂ accumulated as discussed because the reaction rate for H₂O₂ decomposition was much improved with increasing H₂O₂ concentration. As shown in Fig. 8, H₂O₂ concentration monotonously increased with reaction period at initial 20 h and achieved a value of ca. 2.6 wt%. However, after 20 h, concentration of H₂O₂ hardly increased with time and this suggests that the formation rate and decomposition rate of H₂O₂ was balanced because the decomposition reaction rate increased as the amount of H₂O₂ increases. Here it is noted that H₂O₂ concentration of 2.6 wt% is one of the highest values in water solvent reported in the open literature. Since the H₂O₂ accumulated concentration is determined by formation rate and decomposition rate of the formed H₂O₂. Therefore, by preventing H₂O₂ decomposition, much higher H₂O₂ concentration could be achieved.

Fig. 8 H₂O₂ concentration on Pd–Au colloid prepared with oxalic acid as a function of the reaction period. 283 K, H₂: 2.5 vol%, O₂: 19.5 vol%, N₂: 78 vol%, total flow rate: 233 ml min⁻¹.

Since H₂O₂ decomposition could be suppressed by decreasing reaction temperature, temperature dependence of H₂O₂ formation rate was studied and Fig. 9 shows H₂O₂ formation rate, H₂O₂ selectivity and H₂ conversion as a function of reaction temperature. Because of freezing point depression and exothermic reaction, direct synthesis of H₂O₂ was performed at 268 K without freezing of catalyst suspension. As expected, H₂O₂ formation rate and selectivity to H₂O₂ are monotonously increased with decreasing reaction temperature. At 268 K, H₂O₂ formation rate is achieved at 25 mmol h⁻¹ and the selectivity is also higher than 54%. In addition, H₂ conversion is almost independent of reaction temperature and so H₂O₂ yield achieved a value as high as 46%, which is the highest class value for yield reported considering water used for solvent and the average value over 2 h after reaction started. Under these conditions, H₂O₂ production rate is achieved to a value of 782.5 mmol h⁻¹ g-cat⁻¹, which is 5 times higher than that reported by Edwards et al.^31,32 This high yield of H₂O₂ suggests that the primary product in H₂ oxidation on Pd–Au nano colloid is H₂O₂ and preventing the formed H₂O₂ decomposition is highly important for achieving the high yield of H₂O₂. Because the surface activity of Pd–Au nano colloid for H₂ activation is high, high H₂ conversion e.g. 85% was still exhibited even at 268 K. Therefore, decrease in reaction temperature is highly effective for obtaining the high yield of H₂O₂. Here, it is also noted that one targeted yield for commercialization of the direct H₂O₂ synthesis method is 50% yield with one-pass reaction and so 46% H₂O₂ yield achieved in this study is highly interesting from the viewpoint of an industrial process.

Fig. 9 H₂O₂ formation rate, H₂O₂ selectivity and H₂ conversion as a function of reaction temperature. Reaction condition: H₂: 10 vol%, O₂: 18 vol%, N₂: 72 vol%, total flow rate: 233 ml min⁻¹.

Since a high yield of H₂O₂ is achieved, accumulation of H₂O₂ was further studied at 268 K. H₂O₂ concentration at 268 K was shown in Fig. 10 as a function of reaction period. The H₂O₂ formation rate is much improved and the H₂O₂ concentration also increased greatly, as shown in Fig. 10. After 10 h, the H₂O₂ concentration achieved a value of 5.8% and then it is saturated because the H₂O₂ decomposition rate is balanced with formation rate. It is said that one of the application areas of H₂O₂ is a bleaching process in paper and the required concentration of H₂O₂ is around 5%. Therefore, it is expected that 5.8% H₂O₂ concentration achieved in this study is reasonably high. In addition, we used water as a solvent for catalyst suspension, and so it is expected that aqueous H₂O₂ water can be obtained easily. Comparing the ethanol or methanol solvent, which usually leads to a high yield or concentration of H₂O₂, no separation process of formed H₂O₂ is required for the present reaction system and this is the great advantage for the use of aqueous solution.^26,27 As discussed before, high H₂O₂ concentration is another important issue for direct synthesis of H₂O₂ from H₂ and so Pd–Au nano colloid prepared by oxalic acid is highly useful for the synthesis of H₂O₂ with reasonably high concentration. In any way, Pd–Au nano colloid is highly active to H₂O₂ synthesis and a relatively high concentration of H₂O₂ can be obtained by direct oxidation of H₂ with O₂.

Fig. 10 H₂O₂ concentration at 268 K as a function of reaction period. Reaction condition: 268 K, H₂: 10 vol%, O₂: 18 vol%, N₂: 72 vol%, total flow rate: 233 ml min⁻¹.

For H₂O₂ selective synthesis, suppression of H₂O₂ decomposition is important. A H₂O₂ synthesis catalyst is also active to H₂O₂ decomposition, in particular, under a H₂ coexisting atmosphere. Decomposition activity of H₂O₂ on Pd–Au nano colloid was studied as a function of reaction temperature. Under the conditions used, simple decomposition of H₂O₂ into H₂O is hardly observed and so the Pd–Au under addition of HCl and H₂SO₄ is almost inactive for the simple decomposition of H₂O₂. On the other hand, the H₂O₂ decomposition proceeds quickly when H₂ is fed. Therefore, H₂O₂ decomposition mainly occurs by reaction with H₂ and H₂O₂ (H₂ + H₂O₂ = 2H₂O), which is also reported by several groups.^32,33Fig. 11 shows the H₂O₂ decomposition rate on Pd–Au prepared with oxalic acid as a function of reaction temperature. As expected, the H₂O₂ decomposition rate was monotonously decreased with decreasing reaction temperature and at 276 K, the decomposition rate is almost half of that at 283 K. Therefore, a much improved H₂O₂ yield or formation rate at 268 K can be explained by suppression of H₂O₂ decomposition with H₂ remaining in reactor. Since H₂O₂ selectivity is slightly higher than 50%, direct oxidation of H₂ into H₂O (H₂ + ½O₂ = H₂O) seems also to proceed simultaneously and more significantly with reaction period. Therefore, preventing the simple oxidation of H₂ with O₂ should be requested for achieving further higher H₂O₂ yield.

Fig. 11 H₂O₂ decomposition rate on Pd–Au prepared with oxalic acid as a function of reaction temperature. Reaction condition: H₂O₂: 2 wt% (566.8 mM), H₂: 2 vol%, N₂: 98 vol%, total flow rate 233 ml min⁻¹.

In summary, this study revealed that Pd–Au nano colloid is highly active and selective for H₂O₂ synthesis by direct oxidation of H₂ and that a high H₂O₂ yield of 46% and H₂O₂ concentration of 5.8% was achieved at 10 vol% H₂, 1 MPa, 268 K.

Conclusions

Development of an active catalytic process for H₂O₂ synthesis from H₂ is a highly important subject from green chemistry. Mono dispersed Pd–Au nano colloid was successfully prepared by using oxalic acid as reductant. Under atmospheric pressure, a slightly higher yield of H₂O₂ is obtained on Pd–Au nano colloid prepared by hydrazine as reductant because of finer particles. However, under pressurized conditions, Pd–Au nano colloid with mono dispersion seems to be important and so Pd–Au nano colloid prepared with oxalic acid shows higher H₂O₂ yield. The reaction mechanism for H₂O₂ formation on Pd–Au metal surface has been estimated with quantum chemistry calculations.^29,30 From the theoretical approaches, H₂O₂ formation proceeds through hydrogenation of oxygen molecule and water forms dominantly when dissociation of oxygen molecules occurs. Addition of Au into Pd is effective for preventing O–O bond dissociation and also makes the desorption of H₂O₂ easier from the surface. Quantum calculations suggest that clusters consisting of Pd : Au = 3 : 1 are highly important for H₂O₂ selective synthesis, and larger particles may not be active for this reaction because of the larger portion of Pd domain. Therefore, we think that Pd–Au nano colloid with a small size and mono dispersion is highly important for the synthesis of H₂O₂ selectively. The high activity of Pd–Au nano colloid prepared with oxalic acid may suggest that the highly disperse state of Pd–Au particle is important for a high yield of H₂O₂. Because H₂O₂ decomposition with H₂ was suppressed by decreasing temperature and high H₂ conversion is sustained, a much higher H₂O₂ yield of 46% and formation rate of 25 mmol h⁻¹ were achieved at 268 K. Under these conditions, H₂O₂ concentration can be improved to 5.8 wt%. Cooling the reaction media lower than ambient temperature requires energy resulting in increased cost. Therefore, a reaction temperature slightly higher than room temperature is ideal for an industrial process for direct synthesis of H₂O₂ from H₂. Since H₂O₂ formation rate drastically decreased with increasing reaction temperature, a reaction temperature slightly lower than ambient temperature is requested at present. We think that optimum reaction temperature may relate with added halogen for preventing H₂O₂ decomposition and so, further detailed study on optimizing the halogen compound added may lead to increased reaction temperature and this is now under investigation. The results will be reported in the future. Consequently, this study reveals that Pd–Au nano colloid with Pd : Au = 3 : 1 is highly active for H₂O₂ synthesis by H₂ oxidation.

Acknowledgements

This work was supported in part by an R&D grant for novel interdisciplinary fields based on nanotechnology and materials from MEXT.

Notes and references

1. K. Kamata, K. Yonehara, Y. Sumida and K. Yamaguchi, Science, 2003, 300, 964–966.
2. T. M. Kin, K. Markus, B. Santosh, D. Christian, A. Gopinathan, H. Herbert, M. Wolfgang and B. Matthias, Org. Lett., 2005, 7, 987–990.
3. D. Bianchi, R. D’Aloisio, R. Bortolo and M. Ricci, Appl. Catal., A, 2007, 327, 295–299.
4. Y. Sawaki and Y. Ogata, Bull. Chem. Soc. Jpn., 1965, 38, 2103–2106.
5. N. Ullrich, B. Kolbe and N. Bredemeyer, ThyssenKrupp Techforum, 2007, 1, 38–43.
6. G. Goor, W. Kunkel and O. Weiberg, in Ulman’s Encyclopedia of Industrial Chemistry, WILEY-VCH, Weinheim, 1989, vol. A13, pp. 443–466.
7. J. M. Campos-Martin, G. Blanco-Brieva and J. L. G. Fierro, Angew. Chem., Int. Ed., 2006, 45, 6962–6984.
8. D. P. Dissanayake and J. H. Lunsford, J. Catal., 2002, 206, 173–176.
9. Y. F. Han and J. H. Lunsford, J. Catal., 2005, 230, 313–316.
10. Q. Liu and J. H. Lunsford, J. Catal., 2006, 239, 237–243.
11. Q. Liu and J. H. Lunsford, Appl. Catal., A, 2006, 314, 94–100.
12. J. K. Edwards, A. Thomas, B. E. Solsona, P. Landon, A. F. Carley and G. J. Hutchings, Catal. Today, 2007, 122, 397–402.
13. J. K. Edwards, B. E. Solsona, P. Landon, A. F. Carley, A. Herzing, C. J. Kiely and G. J. Hutchings, J. Catal., 2005, 236, 69–79.
14. P. Landon, P. J. Collier, A. F. Carley, D. Chadwick, A. J. Papworth, A. Burrows, C. J. Kiely and G. J. Hutchings, Phys. Chem. Chem. Phys., 2003, 5, 1917–1923.
15. T. Ishihara, Y. Hata, Y. Nomura, K. Kaneko and H. Matsumoto, Chem. Lett., 2007, 36, 878–879.
16. G. B. Brieva, E. C. Serrano, J. M. C. Martin and J. L. G. Fierro, Chem. Commun., 2004, 1184–1185.
17. C. Burato, P. Centomo, M. Rizzoli, A. Biffis, S. Campestrini and B. Corain, Adv. Synth. Catal., 2006, 348, 255–259.
18. K. P. Reis, V. K. Joshi and M. E. Thompson, J. Catal., 1996, 161, 62–67.
19. K. Kamata, K. Yonehara, Y. Sumida and K. Yamaguchi, Science, 2003, 300, 964–966.
20. V. V. Krishnan, A. G. Dokoutchaev and M. E. Thompson, J. Catal., 2000, 196, 366–374.
21. S. E. Park, L. Huang. C. W. Lee and J.-S. Chang, Catal. Today, 2000, 61, 117–122.
22. Q. Chen and E. J. Beckman, Green Chem., 2007, 9, 802–808.
23. A. G. Gaikwad, S. D. Sansare and V. R. Choudhary, J. Mol. Catal. A: Chem., 2002, 181, 143–149.
24. S. Melada, R. Rioda, F. Menegazzo, F. Pinna and G. Strukul, J. Catal., 2006, 239, 422–430.
25. Y. Nomura, H. Matsumoto and T. Ishihara, ChemSusChem, 2008, 1(7), 619–621.
26. S. Park, D. R. Park, J. H. Choi, T. J. Kim, Y. M. Chung, S. H. Oh and I. K. Song, J. Mol. Catal. A: Chem., 2011, 336, 78–86.
27. R. Burch and P. R. Ellis, Appl. Catal., B, 2003, 42, 203–211.
28. A. A. Herzing, A. F. Carley, J. K. Edwards, G. J. Hutchings and C. J. Kiely, Chem. Mater., 2008, 20, 1492–1501.
29. A. Staykov, T. Kamachi, T. Ishihara and K. Yoshizawa, J. Phys. Chem. C, 2008, 112(49), 19501–19505.
30. J. Li, A. Staykov, T. Ishihara and K. Yoshizawa, J. Phys. Chem. C, 2011, 115, 7392–7398.
31. J. K. Edwards, B. Solsona, E. Ntainjua N, A. F. Carley, A. Herzing, C. J. Kiety and G. J. Hutchings, Science, 2009, 323, 1037–1041.
32. M. Piccinini, E. Ntainjuas N, J. F. Edwards, A. F. Carley, J. A. Moulijn and G. J. Hutching, Phys. Chem. Chem. Phys., 2010, 12, 2488–2492.
33. T. Deguchi and M. Iwamoto, Ind. Eng. Chem. Res., 2010, 50, 4351–4358.

---