Rapid production of 38 mM H₂O₂ in an alcoholic suspension of a WO₃ photocatalyst under visible light

Abstract

Specific approaches to H₂O₂ production over a WO₃ photocatalyst under visible light irradiation, i.e., (1) no use of a cocatalyst, (2) reaction under an O₂ atmosphere, (3) reaction in a high concentration of alcohol, and (4) reaction at 60 °C, were found to be effective for rapid production of 38 mM H₂O₂.

---

Hydrogen peroxide (H₂O₂) is a versatile chemical compound that is used in various applications including applications in pharmaceuticals, organic synthesis, and semiconductor manufacturing processes.¹ It is also gaining attention for its potential to reduce costs and promote environmental conservation in chemical production post-processing. The reaction of H₂O₂ produces only water and oxygen as by-products. However, the anthraquinone method, which is currently the mainstream industrial method for producing H₂O₂, is a multi-step process that consumes a significant amount of energy, introduces hydrogen gas, and requires toxic organic solvents. Therefore, a more efficient and cleaner method for synthesizing H₂O₂ is necessary.

One method for producing H₂O₂ is through photocatalysis using semiconductors.² When a semiconductor is irradiated by light of which the energy is larger than the band gap, photogenerated electrons in the conduction band (CB) and positive holes in the valence band cause reduction and oxidation, respectively. It was reported that organic and inorganic semiconductor photocatalysts such as titanium(IV) oxide (TiO₂),^3,4 carbon nitride (C₃N₄),^5,6 cadmium sulfide⁷ and metal organic frameworks (MOFs)⁸ generate H₂O₂ by two-electron reduction of O₂ (O₂ + 2e⁻ + 2H⁺→ H₂O₂, E° = +0.69 V vs. NHE)⁹ in the presence of an appropriate hole scavenger (Fig. 1). However, since the CB of these photocatalysts is rather negative, oxygen radicals such as ˙O₂⁻ and ˙HO₂ can also be formed by one-electron reduction of O₂ (O₂ + e⁻ + H⁺ → ˙O₂H, E° = −0.05 V vs. NHE), resulting in less efficiency (or selectivity) for H₂O₂ production. In addition, since H₂O₂ decomposes with irradiation of UV light,¹⁰ TiO₂ might be eliminated from candidates for an efficient H₂O₂-producing photocatalyst. We reached the idea that tungsten(VI) oxide (WO₃) would be an ideal photocatalyst for H₂O₂ production for two reasons: (1) tungsten(VI) oxide works under irradiation of visible light (band gap: 2.8 eV), which reduces the possibility of H₂O₂ decomposition by UV light, and (2) the bottom of the CB of WO₃ (+0.50 V vs. NHE) is located under the potential for one-electron reduction of O₂ but over the potential for two-electron reduction of O₂, which eliminates the possibility of H₂O₂ decomposition induced by active oxygen species formed by one-electron reduction of O₂ (Fig. 1).

Fig. 1 Energy potentials of O₂ reductions and the conduction band bottom of WO₃.

The concentration of H₂O₂ ([H₂O₂]) is expressed as a function of reaction time (t)³ (eqn (1)), assuming that the photocatalytic H₂O₂ production kinetics is in the zero order and photocatalytic H₂O₂ consumption is in the first order of its concentration,

The equilibrium (saturated) concentration of H₂O₂ ([H₂O₂]_∞) is calculated from eqn (2),

and is determined to be

From the kinetic consideration, the strategy for rapid production of H₂O₂ with a high concentration is to simultaneously satisfy a large k_f and a small k_d toward the WO₃ photocatalyst. The tactic for achieving a small k_d is not to use a cocatalyst because a cocatalyst induces the decomposition of H₂O₂ as well as acceleration of O₂ reduction.

In this study, to compensate for the loss of the cocatalyst effect and to further increase the rate of H₂O₂ production by O₂ reduction over a cocatalyst-free WO₃ photocatalyst under visible light irradiation, several methods for achievement of a larger k_f and a high concentration of H₂O₂ were investigated. We found that rapid production of 38 mM H₂O₂ was achieved by a very simple reaction system, i.e., cocatalyst-free WO₃ in an alcohol–water solvent with a high alcohol content under O₂ at a slightly elevated temperature.

Fig. 2a shows the time courses of photocatalytic H₂O₂ production by O₂ reduction in a 2-propanol–water (9 : 1) suspension of WO₃ at various temperatures under visible light irradiation. At 27 °C, the H₂O₂ concentration reached 4.8 mM at 4 h, which is much larger than that reported for cocatalyst-free WO₃ (8.55 × 10⁻³ mM at 5 h) and still larger than that of Au–WO₃ (1.2 mM at 5 h) in a 4% ethanol aqueous solution,¹¹ although the reaction conditions such as the kind of WO₃, alcohol concentration, and light intensity were different. The large difference between our result and the reported results suggests that the rate of H₂O₂ production and H₂O₂ concentration can be further increased by controlling the reaction conditions. No H₂O₂ was produced in the absence of WO₃.

Fig. 2 (a) Time courses of photocatalytic H₂O₂ production by O₂ reduction in a 2-propanol–water (9 : 1) suspension of WO₃ at various temperatures under irradiation of visible light and (b) Arrhenius plot using k_f values determined at various temperatures.

Fig. 2a shows that a slight increase in reaction temperature significantly increased both the rate of H₂O₂ production and concentration of H₂O₂. After 2 h at 60 °C, the concentration of H₂O₂ surprisingly reached 28 mM and thereafter remained largely unchanged. Kao et al.¹² provide a summary of the concentrations of H₂O₂ produced in various photocatalytic systems, along with the reaction conditions, in the supporting information of their paper. The rate and concentration of H₂O₂ production in our reaction system exceed those previously reported. For example, Teranishi et al.¹³ reported that the concentration of H₂O₂ reached 17 mM after 23 h in a 4% ethanol–water suspension of Au/TiO₂ under UV light irradiation. The concentration of H₂O₂ increased to 11 mM after 8 h over modified Au/TiO₂ in an aqueous solution containing 3.2 vol% formic acid.¹² In previous studies, Au was mainly used as the sole cocatalyst for H₂O₂ production because of its effectiveness in catalyzing O₂ reduction and low activity for H₂O₂ decomposition. All of the studies were conducted at room temperature without considering the reaction temperature. Our results indicate that a significant amount of H₂O₂ can be produced rapidly by raising the reaction temperature by only about 30 °C, without the need for an Au cocatalyst. To clarify the effect of an Au cocatalyst, 0.3 wt% Au/WO₃ was prepared using the photodeposition method and used for H₂O₂ production under the same conditions as those for cocatalyst-free WO₃. The reaction was carried out in a 2-propanol–water (9 : 1) solution under O₂ at 60 °C with visible light irradiation (Fig. S1, ESI†). The concentration of H₂O₂ after 2 h was 12 mM, which was less than half of that produced by Au-free WO₃, although the reaction rate and concentration were much larger than those previously reported.

To better understand the production of 28 mM H₂O₂ at 2 h in a 2-propanol–water (9 : 1) suspension of metal-free WO₃ at various temperatures, we analyzed the kinetics. Table 1 shows the rate constants of H₂O₂ production and consumption (k_f and k_d), which were determined from curve fitting of experimental results (Fig. 2a) using eqn (1). Simulation curves using k_f and k_d were well-fitted to the results. The value of k_f at 27 °C (2.0 mM h⁻¹) was much larger than that reported for Au–WO₃ (0.36 mM h⁻¹).¹¹ This indicates that the amount of alcohol and O₂ pressure greatly affect H₂O₂ production by a WO₃ photocatalyst, regardless of the use of an Au cocatalyst. As the reaction temperature was increased, the value of k_f greatly increased and the value at 60 °C (20 mM h⁻¹) was ten-times larger than that of cocatalyst-free WO₃ at 27 °C. These results suggest that increasing the reaction temperature is a simple yet highly efficient approach to rapidly produce H₂O₂ with a high concentration using cocatalyst-free WO₃ photocatalysis under visible light irradiation. The kinetics of Au–WO₃ at 60 °C were also analyzed and its rate constants (k_f and k_d) are also shown in Table 1. The value of k_f of Au–WO₃ at 60 °C was much smaller than that of Au-free WO₃ at the same temperature. As predicted, the value of k_d of Au–WO₃ at 60 °C was larger than that of Au-free WO₃ at the same temperature. A comparison of the kinetic parameters between Au-free WO₃ and Au–WO₃ at 60 °C suggests that our approach of operating the reaction in an alcohol–water solvent with a high alcohol content under O₂ at a slightly elevated temperature without using a cocatalyst is appropriate for achieving rapid production and a high concentration of H₂O₂.

Table 1 Kinetic parameters of photocatalytic H₂O₂ production by O₂ reduction in a 2-propanol–water (9 : 1) suspension of WO₃ at various temperatures under irradiation of visible light

Temp./°C	k f/mM h^−1	k d/h^−1	k f/kd/mM
a Calculated from the initial 2 hours of irradiation. b Data for 0.3 wt% Au–WO3.
27	2.0	0.34	5.9
40	4.7	0.21	22
50	9.4	0.30	31
60	20^a	0.39^a	51
60^a	9.9^b	0.54^b	15^a

---

The thermal effects on H₂O₂ production over metal-free WO₃ were analyzed, and Fig. 2b shows an Arrhenius plot. A linear correlation was observed between the logarithm of the rate constant for H₂O₂ production (ln k_f) and reciprocal temperature (T⁻¹). The apparent activation energy (E_a) and the frequency factor (A) were determined to be 62 kJ mol⁻¹ and 9.8 × 10¹⁰ mM h⁻¹, respectively, under the present reaction conditions. These kinetic parameters clarify the characteristics of photocatalytic H₂O₂ formation over cocatalyst-free WO₃ under the present conditions. The large activation energy suppresses a rate-determining process at around room temperature, either O₂ reduction or 2-propanol oxidation on the surface of WO₃. This difficulty can be overcome by a slight increase in the reaction temperature. Interestingly, the effect of reaction temperature on k_d of WO₃ was considerably less than the effect on k_f (refer to Table 1). This may be a characteristic of cocatalyst-free WO₃. At elevated temperatures, the large increase in k_f value and the slight increase in k_d value result in a larger k_f/k_d value (Table 1), leading to rapid H₂O₂ production and large H₂O₂ concentration. A thermal effect on H₂O₂ production was observed only under light irradiation (Fig. S2, ESI†). In the dark at 60 °C, H₂O₂ was consumed, while the amount of acetone increased, indicating that oxidation of 2-propanol by H₂O₂ occurred over WO₃.

Fig. 3 shows an absorption spectrum of WO₃ and an action spectrum of H₂O₂ production in a 2-propanol–water (9 : 1) suspension of WO₃ obtained at 60 °C. The change in AQE against wavelength of light was similar to that in light absorption of WO₃, suggesting that H₂O₂ production was induced by photoabsorption of WO₃. It should be noted that the value of AQE at 450 ± 10 nm (55%) was quite large, which means that H₂O₂ production by O₂ reduction proceeds very efficiently over WO₃ under the present conditions. The value of AQE was almost the same in contrast to the further increase in the photoabsorption at 400 ± 10 nm. Since inevitable consumption of H₂O₂ by 2-propanol oxidation occurs, H₂O₂ production may be almost saturated at 450 nm under the present conditions.

Fig. 3 Absorption spectrum of WO₃ and action spectrum of H₂O₂ production in a 2-propanol–water (9 : 1) suspension of WO₃ obtained at 60 °C.

Fig. 4 shows the effect of 2-propanol content in 2-propanol–water solvents on H₂O₂ production for 1 h in a suspension of WO₃ at 60 °C. The H₂O₂ yield increased with an increase in the 2-propanol content and the yield reached the maximum at 90% of 2-propanol. The yield decreased when the reaction was carried out in neat 2-propanol. One of the reasons for the larger yield in solvents having larger 2-propanol contents is that positive holes were effectively trapped by 2-propanol adsorbed on the surface of WO₃, resulting in efficient reduction of O₂ with photogenerated electrons. Another reason is the larger solubility of O₂ in solvents having larger 2-propanol contents,^14–16 which directly increases the rate of O₂ reduction (electron trapping by O₂). The decrease in the yield in neat 2-propanol suggests that the presence of H₂O is necessary as the hydrogen source and/or stabilization of H₂O₂ in the solvent.

Fig. 4 Effect of the 2-propanol content in a 2-propanol–water solvent on H₂O₂ production for 1 h in a suspension of WO₃ at 60 °C.

After reaching 28 mM at 2 h, the H₂O₂ concentration did not increase in a 2-propanol–water (9 : 1) suspension of WO₃ at 60 °C, although the equilibrium concentration was expected to be 51 mM from the value of k_f/k_d (eqn (1) and Table 1). Table 2 shows the amounts of H₂O₂ and acetone in the liquid phase and the amount of O₂ in the gas phase at various times at 60 °C. If the ideal reaction (O₂ + 2-propanol → H₂O₂ + acetone) occurs, the ratio of [H₂O₂]/[acetone] should be unity. However, the observed values were less than unity, as shown in Table 2. This suggests that a subsequent reaction consuming H₂O₂ (H₂O₂ + 2-propanol → 2H₂O + acetone) occurs in part and inevitably. We also noted that O₂ was greatly consumed along with H₂O₂ production (approximately half being consumed after 2 h), as shown in Table 2. The decrease in the partial pressure of O₂ in the gas phase also reduces the O₂ concentration in the liquid phase, resulting in a decrease in the rate of H₂O₂ production. Another reason for the saturation of the H₂O₂ concentration may be the deactivation of WO₃, which will be discussed later. To increase the H₂O₂ concentration, two methods, (1) connection with an O₂ balloon and (2) re-bubbling with O₂, were investigated. When an O₂ balloon was connected to the reactor, the H₂O₂ concentration increased after 2 h and finally reached 34 mM after 4 h (Fig. 5a). In the case of re-bubbling with O₂ at 2 h, the H₂O₂ concentration reached 38 mM after an additional 1 h (total of 3 h) light irradiation (Fig. 5b). These results indicate that O₂ pressure is one of the important factors for achieving a high concentration of H₂O₂.

Table 2 Concentrations of H₂O₂ and acetone produced in the 2-propanol–water (9 : 1) suspension of WO₃ and amount of O₂ in the gas phase at various times at 60 °C under visible light irradiation

Time/h	[H2O2]/mM	[Acetone]/mM	O2/μmol
0	0	0	1400	—
1	17	55	1100	0.31
2	28	116	716	0.24
3	29	246	204	0.12

---

Fig. 5 (a) Time courses of photocatalytic H₂O₂ production by O₂ reduction in a 2-propanol–water (9 : 1) suspension of WO₃ at 60 °C under visible light irradiation: (a) connected with an O₂ balloon and (b) re-bubbled with O₂.

Fig. 5b indicates that WO₃ was deactivated after 3 h. We performed various analyses including XRD, SEM, FT-IR, UV-vis spectrum, and XPS analyses of WO₃ samples before and after the reaction (Fig. S3, ESI†). However, no clear differences were observed between the two samples, suggesting that the saturation of H₂O₂ concentration is due not to an irreversible deactivation but a temporal deactivation of WO₃. We found that calcination of WO₃ in a box furnace at 300 °C for 1 h under air resulted in complete recovery of the initial performance of WO₃ (Fig. S4, ESI†). Since the specific surface area of WO₃ was small, a small amount of organic (peroxide) moieties deposited on the surface may decrease the O₂ adsorption. The organic moieties are probably removed by thermal treatment at 300 °C, resulting in the recovery of the original performance of WO₃.

In summary, it has been believed that only Au serves as an efficient cocatalyst for the photocatalytic production of H₂O₂ through O₂ reduction. However, the production rate and the concentration of H₂O₂ have remained low. In this study, a WO₃ photocatalyst was used to achieve rapid production and a high concentration of H₂O₂. Tungsten(VI) oxide has the advantage of working under visible light irradiation, and its CB bottom is located between the potentials for one-electron and two-electron reduction of O₂. The strategy employed in this study involves avoiding the use of a cocatalyst, which induces H₂O₂ decomposition, even though this may result in a decrease in the formation rate. To compensate for the loss of the cocatalyst effect in the cocatalyst-free WO₃ photocatalyst, other factors were investigated. Finally, rapid production of 38 mM H₂O₂ was achieved in a high alcohol concentration (90%) under an O₂ atmosphere at 60 °C with additional O₂ supply.

The data supporting this article have been included as part of the ESI.†

This work was partly supported by JSPS KAKENHI Grants 23K17964 and 23H01767 and by a fund from Nippon Sheet Glass Foundation for Materials Science and Engineering. A. T. is grateful for financial support from Japan Petroleum Energy Center (JPEC).

Conflicts of interest

There are no conflicts to declare.

Notes and references

1. J. M. Campos-Martin, G. Blanco-Brieva and J. L. G. Fierro, Angew. Chem., Int. Ed., 2006, 45, 6962.
2. H. Hou, X. Zeng and X. Zhang, Angew. Chem., Int. Ed., 2020, 59, 17356.
3. M. Teranishi, S. Naya and H. Tada, J. Am. Chem. Soc., 2010, 132, 7850.
4. Y. Shiraishi, S. Kanazawa, D. Tsukamoto, A. Shiro, Y. Sugano and T. Hirai, ACS Catal., 2013, 3, 2222.
5. Z. Zhu, H. Pan, M. Murugananthan, J. Gong and Y. Zhang, Appl. Catal., B, 2018, 232, 19.
6. Z. Wei, M. Liu, Z. Zhang, W. Yao, H. Tan and Y. Zhu, Energy Environ. Sci., 2018, 11, 2581.
7. S. Thakur, T. Kshetri, N. H. Kim and J. H. Lee, J. Catal., 2017, 345, 78.
8. Y. Isaka, Y. Kondou, Y. Kawase, Y. Kuwahara, K. Mori and H. Yamashita, Chem. Commun., 2018, 54, 9270.
9. Y. Nosaka and A. Y. Nosaka, Chem. Rev., 2017, 117, 11302.
10. P. Krystynik, P. Kluson, S. Hejda, D. Buzek, P. Masin and D. N. Tito, Appl. Catal., B, 2014, 160–161, 506.
11. Y. Wang, Y. Wang, J. Zhao, M. Chen, X. Huang and Y. Xu, Appl. Catal., B, 2021, 284, 119691.
12. K.-C. Kao, S.-J. Huang, Y.-F. Hsia, J.-H. Huang and C.-Y. Mou, ACS Appl. Nano Mater., 2024, 7, 218.
13. M. Teranishi, S. Naya and H. Tada, J. Phys. Chem. C, 2016, 120, 1083.
14. C. B. Kretschmer, J. Nowakowska and R. Wiebe, Ind. Eng. Chem., 1946, 38, 506.
15. R. Battino, T. R. Rettich and T. Tominaga, J. Phys. Chem. Ref. Data, 1983, 12, 163.
16. M. Quaranta, M. Murkovic and I. Klimant, Analyst, 2013, 138, 6243.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental procedure, Tables S1 and S2, and Fig. S1–S4. See DOI: https://doi.org/10.1039/d4cc01402b

---