Efficient Hydrogen Peroxide Generation Utilizing Photocatalytic Oxygen Reduction at a Triphase Interface.

Photocatalytic oxygen reduction has garnered attention as an emerging alternative to traditional anthraquinone oxidation process to synthesize H2O2. However, despite great efforts to optimize photocatalyst activity, the formation rate has been largely limited by the deficient accessibility of the photocatalysts to sufficient O2 in water. Here we boost the reaction by reporting an air-liquid-solid triphase photocatalytic system for efficient H2O2 generation. The triphase system allows reactant O2 to reach the reaction interface directly from the ambient atmosphere, greatly increasing the interface O2 concentration, which in turn simultaneously enhanced the kinetics of formation constant and suppressed the unwanted electron-hole recombination and the kinetics of H2O2 decomposition reaction. Compared with a conventional liquid-solid diphase reaction system, the triphase system enables an increase in H2O2 formation by a factor of 44. The triphase system is generally applicable to fundamentally understand and maximize the kinetics of semiconductor-based photocatalytic oxygen reduction for H2O2 generation.

Introduction

Hydrogen peroxide (H2O2) is a valuable chemical with rapidly growing demand in a wide variety of industrial areas, including fuel cells, chemical oxidation, environment protection, and paper and textile industries (Campos-Martin et al., 2006). The global H2O2 market demand is expected to reach 6,000 kilotons in 2024 (www.gminsights.com/pressrelease/hydrogen-peroxide-market). Currently, industrial processes for H2O2 synthesis involve the multistep anthraquinone oxidation, which requires complex large-scale infrastructure and large amounts of energy. Thus, developing efficient and cost-effective alternative routes for H2O2 generation is of ongoing importance (Edwards et al., 2009, Freakley et al., 2016, Jung et al., 2018, Lu et al., 2018, Siahrostami et al., 2013).

The photocatalytic reduction of oxygen to H2O2 has received great attention as it requires only light, water, and O2 (Baur and Neuweiler, 1927, Cooper and Zika, 1983, Kaynan et al., 2014, Kofuji et al., 2016, Kormann et al., 1988, Liu et al., 2014, Moon et al., 2014, Nakata and Fujishima, 2012, Shiraishi et al., 2014, Sorcar et al., 2018, Teranishi et al., 2010, Teranishi et al., 2016, Wang et al., 2015). During the reaction photogenerated conduction band (CB) electrons reduce O2 to produce H2O2; O2 + 2e−CB + 2H+aq→ H2O2 [(O2/H2O2) = 0.695 V versus normal hydrogen electrode (NHE)]. However, to date resultant product concentrations have been quite limited. The low production rate can be ascribed to the following aspects, which may not be strictly independent of one another: first, the low concentration and slow diffusion rate of O2 in liquid phase results in deficient accessibility of the photocatalysts to reactant; second, the recombination of electrons and holes limits the electron utilization efficiency, and such limitation becomes more serious in the presence of higher charge carrier concentrations associated with greater light intensities; third, the degradation of H2O2 by photogenerated charge carriers also reduces the product yield.

The performance of interfacial catalytic reactions is generally governed by the interface environment. Herein, we simultaneously address these limitations by demonstrating a reaction system possessing an air-liquid-solid triphase reaction interface as illustrated in Figure 1A, where the nanostructured semiconductors are deposited on the top surface of a porous superhydrophobic substrate. Learning from nature, based on the cooperative effect between the low surface energy and rough surface structure, superhydrophobic substrates have been fabricated and used in a wide variety of fields (Aebisher et al., 2013, Deng et al., 2012, Feng et al., 2002, Feng and Jiang, 2006, Hong et al., 2007, Lafuma and Quéré, 2003, Lei et al., 2016, Su et al., 2016, Wooh et al., 2017, Wu et al., 2014, Yohe et al., 2012). When immersed in water the superhydrophobic substrate traps air within atmosphere-linked air pockets, resulting in an interface where solid, liquid, and air three phases coexist (Feng et al., 2002, Lafuma and Quéré, 2003). The triphase system allows reactant O2 to diffuse directly from the air phase to the reaction interface, rather than by slow diffusion through the liquid. Benefiting from this interface architecture the accessibility of the photocatalyst to O2 is greatly increased, which in turn (1) enhances the reaction rate between O2 and photogenerated electrons, (2) suppresses the electron-hole recombination and increases the charge utilization efficiency, and (3) reduces the degradation reaction between H2O2 and photogenerated electrons, thus leading to much enhanced rates of H2O2 production.

Figure 1

Schematic Illustration of the Triphase Photocatalytic Reaction System

(A) Photocatalysts are immobilized on the porous superhydrophobic substrate.

(B) Enlarged view of the solid-liquid-air triphase reaction zone. Reactant O2 is rapidly delivered from the air to the reaction interface resulting in a significantly enhanced rate of H2O2 production.

Results and Discussion

As a proof of concept, we constructed a triphase photocatalytic interface architecture by immobilizing Au-decorated TiO2 nanoparticles (Au-TiO2 NPs) (Teranishi et al., 2010) onto a polytetrafluoroethylene-treated superhydrophobic porous membrane composed of carbon fiber (see the “Methods”) as shown in Figures 1A and S1 (Supplemental Information). Oxygen can diffuse perpendicularly through the membrane, via air phase, to the reaction interface. Upon UV light illumination the photogenerated electrons transfer from TiO2 to the Au co-catalyst to react with O2 via two-electron reaction, in turn producing H2O2 (Figure 1B). A structural analysis of the Au-TiO2/triphase system is shown in Figure 2; the anatase TiO2 NPs (Figure S2) have an average size of about 200 nm. The hydrophobic carbon fiber substrate has a water contact angle (CA) of 148° (Figure 2A), whereas after photocatalyst deposition (see top of Figure 2A) the surface becomes hydrophilic with a CA of 47°. In such a case, water can wet the hydrophilic photocatalysts but cannot penetrate into the porous hydrophobic substrate, leading to the formation of a triphase reaction interface microenvironment. Imaging by transmission electron microscopy (TEM, high-resolution TEM), see Figure 2C, indicates that the Au NPs are uniformly distributed upon the TiO2 surface. Fringe spacing of 0.204 and 0.352 nm, respectively, corresponding to the d-spacing of Au (200) and TiO2 (101) planes, can be observed in Figures 2D and 2E. Figure 2F shows that the average size of the Au NPs is about 5.2 nm.

Figure 2

Surface Morphologies and Microstructure Characterizations of the Triphase Reaction System

(A) Scanning electron microscopic (SEM) image of polytetrafluoroethylene-treated superhydrophobic carbon fiber substrate immobilized with Au-TiO2 nanoparticles; inserts are photographs of water droplets placed on the substrate (bottom) and Au-TiO2/substrate (top).

(B) SEM image of the carbon fiber and Au-TiO2 nanoparticles.

(C) TEM image of one individual Au-decorated TiO2 nanoparticle.

(D and E) High-resolution TEM images of (E) TiO2 and (D) Au nanoparticles.

(F) Size distribution of the Au nanoparticles.

Photocatalytic synthesis of H2O2 was carried out under UV light with a wavelength of 367 ± 5 nm (Figure S3). Control experiments based on a diphase photocatalytic system where the same amount of photocatalyst was dispersed in 1.5 mL water was also conducted. The photocatalytic performances of diphase and triphase reaction systems were first evaluated under UV light of different intensities. Figure S4 shows the H2O2 concentration ([H2O2]) after 1-h reaction under UV light illumination using the triphase reaction system. The reactions initially demonstrate zero-order kinetics, with [H2O2] increasing linearly with time. The formation and degradation of H2O2 are generally considered to follow, respectively, zero- and first-order kinetics (Kaynan et al., 2014, Kormann et al., 1988, Teranishi et al., 2010). Thus it can be assumed there is negligible H2O2 degradation within the first 1 h, allowing the production rate to be calculated as d[H2O2]/dt. Figure 3A shows the relationship between the calculated H2O2 formation rate and light intensities varying between 1 and 120 mW cm−2. Using the triphase system (Figure 3A, red line), the rate of production increased with light intensity up to about 60 mW cm−2, whereas in the diphase reaction system (Figure 3A, dark line), the production rate saturates at a light intensity of 3 mW cm−2. The rate of production reaction of the triphase system at a light intensity of 60 mW cm−2 was about 18 times faster than that of the diphase reaction system.

Figure 3

Performance of the Triphase Photocatalytic System for H2O2 Generation

(A) Rate of H2O2 formation under different light intensities using the triphase (red line) and diphase (dark line) systems.

(B) H2O2 formation rates based on different operational conditions using these two systems under illumination of 60 mW cm−2. The bottom panel presents a schematic illustration of the four operational conditions.

(C) IAQYs for the two systems under 1 mW cm−2, 9 mW cm−2, and 60 mW cm−2 illumination. Data of IAQYs for 1, 9, and 60 mW cm−2 illumination are 17.29 ± 0.44, 15.23 ± 0.57, and 5.49 ± 0.21, respectively, in the triphase system and 2.45 ± 0.64, 0.55 ± 0.08, and 0.08 ± 0.009, respectively, in the diphase system.

The difference in the reaction rates of the diphase and triphase systems can be attributed to their fundamentally different reaction interfacial architectures. With the diphase system O2 is delivered to the photocatalyst surface through the liquid phase with a slow rate of diffusion. Even with air being continuously fed into the solution (Figure S5) the enhancement in H2O2 production with the diphase system is limited (see Figures 3B-I and 3B-II). In contrast, the triphase architecture enables sufficient O2 to be delivered directly from air to the reaction interface. Because the diffusion coefficient of O2 in air (2.0 × 10−1 cm2 s−1) is approximately four orders of magnitude higher than that in water (2.1× 10−5 cm2 s−1) (Cussler, 1997), O2 consumed at the photocatalyst surface is rapidly resupplied, with O2 concentration in the liquid phase having little impact on the reaction kinetics. As seen in Figure 3B, when O2 levels in water are decreased by feeding the solution with nitrogen (Figure 3B-I) and increased by feeding the solution with air (Figure 3B-iii), the rate of H2O2 formation is essentially unchanged (Figure 3B-ii), indicating that the triphase photocatalytic reaction kinetics is air phase dependent. To confirm that the activity enhancement was due to the enhanced access of the photocatalyst to O2, O2 in the air phase was replaced with N2 as illustrated in Figure S6. The significantly enhanced production rate indicates that rapid mass transport of O2 from air phase to the triphase interface plays a key role in enhancing oxidase kinetics. Increasing interface O2 partial pressure can further increase the rate of H2O2 production; as an example, Figure 3B-iv indicates the effect of replacing air by (pure) oxygen.

The higher O2 levels at the triphase interface significantly enhance the reaction kinetics between O2 and electrons, whereas suppressing electron-hole recombination leads, in turn, to higher charge utilization efficiency and quantum yields. The apparent quantum yield (AQY) is defined as the number of electrons used to produce H2O2 molecules per unit time to the number of incident photons (Kato et al., 2013). The calculated initial AQY (IAQY) based on triphase system is much higher than that of diphase system over the whole range of light intensities. As shown in Figures 3C and S7, under light intensities of 1 mW cm−2, 9 mW cm−2, and 60 mW cm−2, the IAQYs of the triphase system are, respectively, approximately 7, 27, and 66 times higher than those of the diphase system. From Figure 3C it can also be seen that the IAQY of the diphase system decreases rapidly from 2.45% to 0.55%, a factor of four, as the light intensity is increased from 1 to 9 mW cm−2. The rapid decrease in the IAQY suggests that electron-hole recombination is the dominant process even at modest light intensities. With the triphase system the IAQY decreases only from 17.29% to 15.23% as the light intensity is increased from 1 to 9 mW cm−2, whereas even at 60 mW cm−2 the IAQY of the triphase system is still much higher than that of the diphase system at 1 mW cm−2. The effective suppression of electron-hole recombination enables the triphase system to be operated at high light intensities with the large amount of photogenerated electrons efficiently utilized for H2O2 production.

The steady-state concentrations (SSC) of H2O2 produced using the two interfacial architectures were further investigated. As seen in Figure 4A, an SSC of 26.5 mM was achieved with the triphase system, approximately 44-fold higher than that obtained with the diphase counterpart (0.6 mM). Using our triphase system a steady state H2O2 yield of 59 μmol per unit photocatalyst weight (mg) can be achieved, a value much higher than that of other group reports (Table S1). The SSC of H2O2 depends on the kinetics of both formation and degradation reactions. Generally, the reaction kinetics can be analyzed using the following equation (Kim et al., 2016, Kormann et al., 1988, Teranishi et al., 2010):where kf and kd are, respectively, the formation and degradation rate constants for H2O2 and t is the reaction time. The [H2O2] formation and degradation reaction rates follow, respectively, zero- and first-order kinetics, where kf is expressed in mM h−1 and k in h−1.

Figure 4

Formation and Degradation Behavior of H2O2

(A) The steady-state concentration of H2O2 produced using the triphase system (red line) and diphase system (dark line) under illumination of 60 mW cm−2.

(B) Schematic illustration of formation and degradation reactions of H2O2 at the triphase interface.

As clearly shown in Figure 4A, the experimental data are quite accurately modeled using the equation. The calculated kf and kd values of the triphase system are 5.06 mM h−1 and 0.18 h−1 (Table 1), whereas for the diphase system the kf and kd values are 0.26 mM h−1 and 0.43 h−1, respectively. We note that for both systems, under 60 mW cm−2 illumination, the calculated kf values are in good agreement with the results from Figure 3A (kf is equal to the formation rate for zero-order kinetics). With the triphase system not only the kf was greatly increased but also the kd was effectively suppressed. As illustrated in Figure 4B the photocatalytic degradation of H2O2 is initiated by its reaction with CB electrons: H2O2 + e−CB → OH− +·OH [(H2O2/·OH) = 0.71 V versus NHE] (Nakata and Fujishima, 2012, Sheng et al., 2014), which competes with the reaction between O2 and electrons (Zhuang et al., 2015). By providing significantly greater amounts of reactant oxygen to the photocatalytic interface, the triphase system suppresses the degradation reaction and enhances the formation reaction, in turn leading to greater H2O2 production.

Table 1

Calculated kf, kd, and kf/kd Values of Triphase System and Diphase System

	kf (mM h−1)	kd (h−1)	kf/kd (mM)
Triphase System	5.06	0.18	28.11
Diphase System	0.26	0.43	0.60

The stability of the triphase reaction system was further evaluated. We have conducted water breakthrough pressure measurement on Au-TiO2/substrates before and after continuous 24-h UV illumination (60 mM cm−2), in each case achieving a comparable water column height (∼89 cm), suggesting a good substrate photostability during photocatalysis of H2O2, of crucial importance to practical applications. The triphase system demonstrated here is applicable to enhance the performance of other photocatalysts. We have studied the activity of photocatalyst ZnFe2O4 (Su et al., 2012) for H2O2 generation. As shown in Figure S8, an SSC of 3.3 mM was achieved based on the triphase system, which is about seven times higher than that of the diphase system under air mass (AM) 1.5 simulated sunlight. This result indicates that the triphase system provides an exploratory platform, on which different kinds of photocatalysts can be applied for efficient H2O2 generation.

In conclusion, we have constructed a photocatalytic system with a triphase solid-liquid-air reaction interface for efficient H2O2 synthesis. The triphase interface allows reactant O2 to be rapidly delivered to the photocatalyst surface, greatly enhancing the formation reaction and reducing the degradation reaction. The rapid accessibility of O2 to the photocatalyst surface effectively suppresses electron-hole recombination, enabling the triphase system to efficiently utilize the larger amounts of electrons obtained at higher light intensities to, in turn, produce more H2O2. Our results reveal that rational interface microenvironment (wettability and architecture) design is crucial for achieving efficient photocatalytic reaction system for H2O2 generation. The triphase reaction system is general; for practical applications, photocatalysts of much lower cost could presumably be used for efficient synthesis of the desired products.

Limitations of the Study

The Au NPs play a vital role in H2O2 generation. In this article, the size and density of Au NPs was not adjusted. In our future work, we will adjust the amount and the size of Au to further optimize the H2O2 photocatalytic synthesis.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.