Enhanced Photocatalytic H2O2 Production over Inverse Opal ZnO@Polydopamine S-Scheme Heterojunctions

Gaowei Han , Feiyan Xu , Bei Cheng , Youji Li , Jiaguo Yu , Liuyang Zhang ,＊

1 State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology,Wuhan 430070, China.

2 Laboratory of Solar Fuel, Faculty of Materials Science and Chemistry, China University of Geosciences, Wuhan 430074, China.

3 College of chemistry and Chemical engineering, Jishou University, Jishou 416000, Hunan Province, China.

Abstract: Photocatalytic H2O2 production is a sustainable and inexpensive process that requires water and gaseous O2 as raw materials and sunlight as the energy source. However, the slow kinetics of current photocatalysts limits its practical application. ZnO is commonly used as a photocatalytic material in the solar-to-chemical conversion, owing to its high electron mobility, nontoxicity, and relatively low cost. The adsorption capacity of H2O2 on the ZnO surface is low, which leads to the continuous production of H2O2. However, its photoresponse is limited to the ultraviolet(UV) region due to its wide bandgap (3.2 eV). Polydopamine (PDA) has emerged as an effective surface functionalization material in the field of photocatalysis due to its abundant functional groups. PDA can be strongly anchored onto the surface of a semiconducting photocatalyst through covalent and noncovalent bonds. The superior properties of PDA served as a motivation for this study. Herein, we prepare an inverse opal-structured porous PDA-modified ZnO (ZnO@PDA) photocatalyst by in situ self-polymerization of dopamine hydrochloride. The crystal structure, morphology, valency, stability, and energy band structure of photocatalysts are characterized by X-ray diffraction (XRD), Fourier-transform infrared spectroscopy (FTIR), transmission electron microscopy(TEM), field-emission scanning electron microscopy (FE-SEM), X-ray photoelectron spectroscopy (XPS), UV-visible diffuse reflectance spectroscopy (UV-Vis DRS), electrochemical impedance spectroscopy (EIS), Mott-Schottky curve (MS), and electron paramagnetic resonance (EPR). The experimental results showed that electrons in PDA are transferred to ZnO upon contact, which results in an electric field at their interface in the direction from PDA to ZnO. The photoexcited electrons in the ZnO conduction bands flow into PDA, driven by the electric field and bent bands, and are recombined with the holes of the highest occupied molecular orbital of PDA, thereby exhibiting an S-scheme charge transfer. This unique S-scheme mechanism ensures effective electron/hole separation and preserves the strong redox ability of used photocarriers. In addition, the inverse opal structure of ZnO@PDA promotes light-harvesting due to the supposed “slow photon” effect, as well as Bragg diffraction and scattering. Moreover, the enhanced surface area provides a high adsorption capacity and increased active sites for photocatalytic reactions. Therefore, the resulting ZnO@PDA (0.03% (atomic fraction) PDA)exhibits the optimal H2O2 production performance (1011.4 μmol·L-1·h-1), which is 4.4 and 8.9 times higher than pristine ZnO and PDA, respectively. The enhanced performance is ascribed to the improved light absorption, efficient charge separation, and strong redox capability of photocarriers in the S-scheme heterojunction. Therefore, this study provides a novel strategy for the design of inorganic/organic S-scheme heterojunctions for efficient photocatalytic H2O2 production.

Key Words: Step-scheme heterojunction; Polydopamine; Inverse opal ZnO; Photocatalytic H2O2 production

1 Introduction

Hydrogen peroxide (H2O2), a strong oxidizing agent, has been widely used in medical, chemical and environmental fields because of its effectiveness and eco-friendliness1-4. It is also recognized as a promising solar fuel due to its high energy density (3.0 MJ·L-1) and easy storability5,6. The global energy market demands three million tons of H2O2per year7. Currently,H2O2is producedviaanthraquinone (AQ) process8-10, direct hydrogen-oxygen reaction11,12and oxidation of alcohols13,14.These methods either demand high energy input or cause environmental pollution, which significantly hinder their largescale application15-17. Thus, it is urgent to develop an economic,efficient and eco-friendly method.

Photocatalysis is a green and sustainable method to produce H2O2since it is powered by inexhaustible sunlight18. Moreover,it also reduces our energy reliance on unsustainable fossil fuels and alleviates environmental pollution15,19,20. To date, various semiconductors have been developed for photocatalytic H2O2production, including TiO213,16,21, ZnO22-24, g-C3N425-27,CdS28,29, graphene30,31and BiVO432. Among them, ZnO,featured with non-toxicity, stability and earth-abundance, has attracted much interest33, but it still suffers from poor absorption of sunlight and fast recombination of photoinduced charge carriers34.

Three-dimensional inverse opal (3DIO) structured photocatalysts show improved light-harvesting performance35.The 3DIO texture shows a unique slow-photon effect because of its large specific surface area and well-ordered porous architecture36, which increases sunlight utilization efficiency and concurrently weakens photo-damage towards the photocatalyst37,38. Polydopamine (PDA) is a widely-explored photosensitizer which can promote solar light utilization of semiconductor materials39-42. Because of enriched functional groups on PDA, e.g., primary amine, tertiary amino and catechol, it can be firmly anchored on the surface of ZnO and stabilize throughout the reaction34,42.

Additionally, an efficient step-scheme (S-scheme)heterojunction can be constructed by PDA-modified ZnO43. In S-scheme photocatalyst, when an oxidation photocatalyst (OP)is coupled with a reduction photocatalyst (RP), an internal electric field (IEF) that directing from RP to OP is formed42,44,45.Then, when illuminated, the photoinduced holes in the RP recombine with the photoinduced electrons in the OP under the IEF46-49. Finally, the photogenerated carriers with strong redox ability are reserved, which explains the enhancement of the performance for photocatalysts perfectly50-55. Considering the energy band positions of PDA and ZnO, it is anticipated that an S-scheme heterojunction can be constructed between them.

Herein, the PDA-modified 3DIO ZnO was prepared by a simplein situself-polymerization. 3DIO ZnO possesses periodically ordered structures and voids that can expose more active sites. The experimental results show that PDA and ZnO are able to form S-scheme heterojunction. Compared with pristine ZnO, PDA-wrapped counterpart promotes the transfer of photogenerated carriers and reserves the strong reduction ability of the photoinduced electrons of ZnO, resulting in higher H2O2production performance.

2 Experimental

2.1 Chemicals

Zn(NO3)2·6H2O (99%, AR), sodium dodecyl sulfate (SDS,86%, CP), potassium persulfate (KPS, 99.5%, AR), styrene(99.5%, AR) and ethanol (99.7%, AR) were obtained from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China).Dopamine hydrochloride (98%, AR) and 5,5-dimethyl-1-pyrroline N-oxide (DMPO, 97%, AR) were purchased from Macklin. All reagents were analytical grade and used as received.

2.2 Synthesis of polystyrene (PS) spheres

The preparation of PS spheres refers to the method in the literature with minor changes36. Typically, 60 mg of SDS was added into 135 mL of deionized water under stirring and heated at 80 °C for 10 min in an oil bath to form a clear solution. Then 15 g of styrene was added and stirred for another 10 min. Finally,150 mg of KPS was added into the solution and stirred for 12 h.The suspension was cooled naturally to room temperature after the reaction was finished.

2.3 Synthesis of 3DIO ZnO

FTO glasses were ultrasonically cleaned for 20 min with water, ethanol and acetone, respectively, followed by UV-ozone treatment for 15 min to remove any residual organics and increase hydrophilicity. The prepared PS suspension was then dropped evenly on the glasses and then heated in an oven at 80 °C to evaporate the solvent. The PS template was soaked with ZnO precursor, which was prepared by dissolving 9.0 g of Zn(NO3)2into the mixture of deionized water (32 mL) and ethanol (28 mL)56. As-treated FTO glasses were dried naturally and calcined at 450 °C for 2 h in the air at a heating rate of 1 °C·min-1to remove PS spheres. After cooling to room temperature, 3DIO ZnO was obtained and labeled as Z.

2.4 Synthesis of 3DIO ZnO@PDA

Typically, 60 mg of 3DIO ZnO and a certain amount of dopamine hydrochloride were dispersed in 30 mL of deionized water and stirred for 3 h. The pH of the aqueous suspension was adjusted to 8.5 with ammonium hydroxide42. The obtained gray product was centrifuged, washed and dried at 80 °C overnight.The molar ratios of polydopamine in ZP1, ZP2 and ZP3 samples 0.05%, 0.3% and 0.8% respectively. Pristine PDA was also synthesized using similar procedure in the absence of 3DIO ZnO and labeled as P. The synthesis process is illustrated as Fig. 1.

2.5 Characterization

Detailed information for material characterizations and photocatalytic experiments are presented in Supporting Information (SI).

3 Result and discussion

The X-ray diffraction (XRD) patterns of pure ZnO were indexed to zincite-type ZnO (JCPDS No. 36-1451) without any detectable impurities (Fig. 2a)38. Upon PDA deposition, the ZnO@PDA showed no changes in phase except for a slight decrease in the diffraction intensity, which indicated that PDA did not change the structure of ZnO lattice.

Fourier transform infrared (FT-IR) spectrometer was used to investigate the functional groups of ZnO@PDA samples, and the obtained spectra were shown in Fig. 2b. The ―OH stretching vibrations at 3400 and 1639 cm-1in ZnO and ZP2 corresponded to the surface-absorbed water molecules57. The band around 450 cm-1was ascribed to the Zn―O stretching vibration. The bands at 1280, 1438, 1507 and 1576 cm-1were ascribed to the C―O,C―H, aromatic C―C and C ＝C vibrations of pure PDA,respectively40,58. Compared with the spectrum of ZnO, due to the shielding effect of PDA, the characteristic absorbance of ZnO in ZP3 showed a slight decrease.

Fig. 1 Schematic illustration of the formation process of the 3DIO ZnO@PDA inverse opal.

Fig. 2 (a) XRD patterns of Z, P, ZP1, ZP2 and ZP3 samples. (b) FTIR spectra of Z, P and ZP2 samples.

The morphology of the samples was observed by field emission scanning electron microscopy (FESEM) and transmission electron microscopy (TEM). As shown in Fig. S1a,PS spheres (～180 nm in diameters) were closely and uniformly distributed. After the removal of PS spheres, 3DIO ZnO emerged with periodically ordered structures and consecutive voids (Fig.S1b)37,38. Such unique periodic porous structures were maintained even after depositing a layer of PDA (Fig. 3a,b). The polymerized dopamine was attached to the framework of the skeleton and increased the surface roughness of the sample. The TEM and high-resolution TEM images of ZP2 (Fig. 3c,d)showed that the PDA was wrapped on ZnO, in agreement with the above FESEM observations. A lattice fringe spacing of 0.19 nm was observed in the HRTEM image (Fig. 3d), corresponding to the ZnO (002) plane. In addition, energy-dispersive X-ray(EDX) spectroscopy mappings (Fig. 3e-i) confirmed the presence of Zn, O, C and N elements in ZP2, affirming the successful formation of ZnO@PDA nanohybrids.

The photocatalytic performance of ZnO@PDA was measured under the illumination of a 300 W Xenon lamp (28 mW·cm-2).Pure PDA and ZnO showed poor H2O2production of 113.3 and 230.3 μmol·L-1·h-1, respectively, as shown in Fig. 4a. This result was due to the rapid charge recombination in unitary photocatalysts. After coupling PDA with ZnO, the yield of H2O2for ZP2 increased to 1011.4 μmol·L-1·h-1, which was 4.4 and 8.9 times higher than that of pure ZnO and PDA, respectively. The apparent quantum yield of ZP2 was 13.96% under the irradiation of a 365 nm light source. Further increasing the PDA amount could deteriorate the photocatalytic H2O2production performance (e.g., 656.1 μmol·L-1·h-1for ZP3) because of the light-shielding effect13,41. The recyclability and stability of ZnO@PDA were also investigated. After four photocatalytic cycles, ZP2 almost maintained its initial photocatalytic performance (decreased 10.5%) (Fig. 4b). The phase, surface groups and chemical states of the spent ZP2 were unchanged as shown in the XRD pattern (Fig. S2), FTIR spectrum (Fig. S3)and XPS spectra (Fig. S4), revealing its acceptable stability.

The decomposition of H2O2is a common side reaction for photocatalytic H2O2production, which should be inhibited largely. As shown in Fig. 4c, all the composite samples showed a lower H2O2decomposition rate than pure ZnO. The zero-order kinetics of H2O2production and first-order kinetics of H2O2decomposition were employed to investigate the apparent kinetics of H2O2production:

whereKfandKdare the production rate constants (μmol·L-1·h-1)and decomposition rate constant (min-1) of H2O2, respectively59,60.The calculated values ofKfandKdfor different samples were illustrated in Fig. 4d. The production of H2O2increased with thePDA content in the composites; while the decomposition rate of H2O2decreased significantly compared with pure ZnO. The results show that ZP2 has highestKfand lowerKdvalues, both of which contribute to a higher H2O2productivity. Compared with pure PDA and ZnO, the optimized 3DIO ZP2 exhibits a high performance in H2O2production because (i) the 3DIO structure offers a large surface area and boosts the visible light utilization; (ii) the heteroepitaxial growth of PDA shell boosts the visible light absorption.

Fig. 4 (a) Photocatalytic H2O2 production of different samples under visible light irradiation. (b) Stability test results of ZP2 for photocatalytic H2O2 production. (c) Photocatalytic degradation of H2O2 in O2-free condition under visible light irradiation.(d) Production and decomposition rates of H2O2 (Kf, Kd) of all the samples.

The XPS spectra of Z, P and ZP2 were recorded to confirm the chemical states of constituent elements. The Zn 2pXPS spectra of pure ZnO and ZP2 (Fig. 5a) showed two peaks at 1021.9 and 1045 eV, which were assigned to Zn 2p3/2and Zn 2p1/2, respectively. The O 1sspectrum (Fig. 5b) confirmed the presence of lattice Zn－O (530.5 eV) and surface ―OH (532.2 eV) in ZnO41,57. The C 1sspectra (Fig. 5c) suggested the presence of graphiticsp3-hybridized carbon (284.8 eV), C－N(286.4 eV) andsp2-hybridized carbon (288.1 eV) in the Ncontaining aromatic ring. The N 1sXPS spectra revealed the presence of－N＝ species of the indole group in PDA (Fig.5d)42. Note that the N 1sbinding energy (BE) of ZP2 shifted positively compared with that of PDA; while the BEs of Zn 2pand O 1sof ZP2 showed negative shifts relative to those of pure ZnO, which suggested electron transfer from PDA to ZnO upon their hybridization. The electron transfer created a built-in electric field (BEF) at interfaces with direction pointing from PDA to ZnO and concurrently bent their energy bands at interfaces52.

In situirradiated XPS was also conducted to explore the transfer of photoinduced charge samples under irradiation. The BEs of Zn 2pand O 1sof ZP2 shifted positively under ultraviolet(UV) light irradiation, compared with those collected in dark. On the contrary, the N 1sBE of ZP2 shifted negatively from 399.9 eV (in dark) to 399.7 eV (under irradiation). The BE shifts proved that the photoinduced electrons were transferred from the ZnO conduction band (CB) to the DPA HOMO under UV light irradiation, exhibiting an apparent S-scheme charge pathway driven by the BEF and bent energy bands at the interfaces of ZP241,42.

The contact potential differences (CPD) of Z, P and ZP2 were measured under dark conditions to illustrate the charge-transfer pathway during the interfaces of ZnO and PDA (Fig. S5). The work function (W) and Fermi level (Ef) of samples were calculated according to the following equations:

Fig. 5 High-resolution XPS spectra of (a) Zn 2p, (b) O 1s, (c) C 1s and (d) N 1s of ZnO, PDA and ZP2.ZP2 and ZP2-UV denote measurements taken dark and UV-irradiated (λ = 365 nm) conditions, respectively.

whereWandWtipare the work functions of the sample and the gold tip (4.25 eV), e is the charge of an electron, and CPDsampleis the CPD value obtained from the sample27. Based on the above equations, the work functions of Z, ZP2 and P are 4.60,4.50 and 4.41 eV, respectively. Correspondingly, the Fermi levels of Z, ZP2 and P are -4.60, -4.50 and -4.41 eV,respectively. The larger work function of ZnO than that of PDA is accounted for the electron transfer within ZP2, which is in accordance with the above XPS results.

The UV-visible diffuse reflection spectroscopy (UV-Vis DRS)of all the samples were recorded to explore their optical properties and light absorption capability (Fig. 6a). The absorption edge of pure ZnO was located at 380 nm, whereas ZP2 exhibited an enhanced light absorption at 410-1400 nm,favoring the photocatalytic H2O2production. The band gaps of pure ZnO and PDA were calculated to be 3.16 and 1.6 eV,respectively (Fig. S6).

The Mott-Schottky (M-S) plots of PDA and ZnO were recorded under different frequencies. As shown in Fig. 6b,c, the positive slopes suggested then-type nature of ZnO and PDA38,42.The flat band potential (Efb) of ZnO and PDA derived from the M-S plots was -0.16 and -0.98 eV (vsNHE, pH = 0),respectively. TheEfbis close to the CBM or LUMO level ofntype semiconductors or polymers. Correspondingly, the valence band maximum (VBM) of ZnO and the HOMO level of PDA were estimated to be 2.8 eV and 0.52 eV (vs.NHE, pH = 0),respectively (Fig. 6d).

Steady-state photoluminescence (PL) and time-resolved photoluminescence (TRPL) spectra were recorded to investigate the separation efficiency of photogenerated carriers. The ZP2 showed weaker PL intensity than pure ZnO (Fig. 7a), indicating enhanced electron-hole separation efficiency. Fig. 7b displayed the TRPL spectra of ZP2 and ZnO. The average lifetime (τa) of samples were calculated according to the following equation:

whereτ1,τ2andτ3represent lifetime of radiative, non-radiative and energy transfer process, respectively.A1,A2andA3are the preexponential factors of decay curves42.

As shown in Table 1, ZP2 had a longer average lifetime since the radiation and non-radiative recombination were suppressed61,62.The long-life charge carriers assured the composites of better H2O2production performance. We also compared the transient photocurrent responses and electrochemical impedance spectra(EIS) of the samples59,63. The ZP2 composites showed an enhanced photocurrent than pure ZnO or PDA (Fig. 7c),manifesting that the charge recombination in ZP2 was restrained.Additionally, ZP2 had the smallest radius of the Nyquist diagram among the samples (Fig. 7d), which implied that ZP2 had the minimum charge transfer resistance and showed great potential in photocatalytic applications64.

To compare the redox ability of the samples, electron paramagnetic resonance (EPR) was applied to identify the reactive radicals. As shown in Fig. 8a, PDA showed no DMPO-·O2-EPR signal, suggesting that the LUMO electrons readily recombined with HOMO holes in pure PDA. In contrast,ZP2 showed an enhanced DMPO-·O2-signal compared with pure ZnO. This suggests that PDA in ZP2 effectively inhibited the recombination of photogenerated carriers owing to the S-scheme charge transfer in the composites. The survived electrons were located in LUMO of PDA, which had strong reducing ability to reduce O2into ·O2-.

Fig. 6 (a) UV-Vis diffuse reflectance spectra of Z, P, ZP1, ZP2 and ZP3, (b) Mott-Schottky plots of P, (c) Mott-Schottky plots of Z,(d) band structure of Z and P.

Fig. 7 (a) PL spectra and (b) TRPL spectra of Z, P and ZP2; (c) transient photocurrent response and(d) EIS spectra of Z, P and ZP2 under irradiation.

Sample τa/ns(Rel. %)τ1/ns(Rel. %)τ2/ns(Rel. %)τ3/ns(Rel. %)P 6.77 0.38(23.33)2.59(48.15)10.78(28.52)18.05(30.18)ZP2 17.13 0.65(10.62)Z 10.67 0.61(22.94)3.24(46.89)3.42(30.01)28.68(59.37)

Generally, photogenerated electrons and holes can be effectively used in surface redox reactions to form H2O2viaeither single-step two-electron reduction reaction (Eq. 5) or the stepwise single-electron reduction reaction (Eqs. 6 and 7)processes27. The EPR results also reveal that the conversion of O2to H2O2undergoes stepwise one-electron processes (Eqs. 6 and 7), evidenced by the formation and accumulation of typical intermediate ·O2-species13. In addition, no DMPO-·OH signal was observed for pure PDA (Fig. 8b), suggesting that PDA was incapable of oxidizing water into ·OH. However, ZP2 showed a stronger DMPO-·OH signal than pure ZnO. This phenomenon correlated with the holes that survived in ZnO valence band (VB)with stronger oxidation ability owing to S-scheme charge transfer in ZP213,42,46.

Fig. 9 illustrated the formation of ZnO@PDA S-scheme heterojunctions and the mechanism for photocatalytic H2O2production. Before contact, ZnO and PDA show staggered band structures (Fig. 9a). Upon contact, owing to the higher work function of ZnO (4.60vs.4.41 eV), electrons in PDA spontaneously transfer towards ZnO until their Fermi levels reach the same level (Fig. 9b). Thus, positive and negative charges gather on the side of PDA and ZnO, respectively, which bends the energy bands and forms an IEF at interfaces with direction pointing from PDA to ZnO65. Under light irradiation,the electrons in PDA HOMO and ZnO VB are photoexcited and accumulated at the CB of ZnO and LUMO of PDA, respectively.Driven by the bent energy band and IEF, the photogenerated electrons in ZnO CB flow into PDA and recombine with the HOMO holes of PDA, exhibiting S-scheme charge transfer (Fig.9c). As a result, the photogenerated carriers (electrons in PDA LUMO and holes in ZnO VB) with strong redox ability are preserved, thereby improving the lifetime of the carriers and the performance of photocatalytic H2O2production.

Fig. 8 EPR spectra of (a) DMPO-·O2- in methanol and (b) DMPO-·OH species in aqueous solution produced over ZnO, PDA and ZP2.

Fig. 9 Schematic illustration of (a) the relative band energy positions of ZnO and PDA before contact, (b) the formation of the built-in electric field after contact, and (c) the S-scheme charge transfer mechanism between ZnO and PDA under irradiation.

4 Conclusion

In summary, three-dimensional ZnO@PDA with inverse opal structures was prepared byin situself-polymerization of PDA on the periodically-ordered ZnO.Ex-situXPS confirmed the presence of a built-in electric field pointing from PDA to ZnO owing to the electron transfer upon their contact. Driven by the IEF, the photoinduced electrons in ZnO CB recombined with the holes in the HOMO of PDA, as confirmed byin situirradiated XPS, corroborating S-scheme charge transfer. The formation of S-scheme heterojunction enhanced the light absorption,increased the lifetime of photogenerated carriers, and allowed the photogenerated carriers to preserve strong redox ability.Compared with pure PDA and ZnO, the ZnO@PDA showed better performance of photocatalytic H2O2production, owing to the S-scheme charge separation and improved light harvesting.This work demonstrated an important contribution of S-scheme heterojunctions for economical and efficient production of hydrogen peroxide.

Supporting Information:available free of chargeviathe internet at http://www.whxb.pku.edu.cn.