Preparation of rGO/ZnO photoanodes and their DSSCs performance

Microcrystalline graphite (average particle size 30 μm, concentration >98%), concentrated sulfuric acid (H₂SO₄, concentration 96%–98%), concentrated hydrochloric acid (HCl, concentration 36%), hydrogen peroxide (H₂O₂, concentration 30%), potassium permanganate (KMnO₄), potassium persulfate (K₂S₂O₈), phosphorus pentoxide (P₂O₅), zinc nitrate hexahydrate ((Zn(NO₃)₂·6H₂O), sucrose (C₁₂H₂₂O₁₁), zinc acetate (Zn(Ac)₂·2H₂O), polyethylene glycol-300 (PEG-300), hydrazine hydrate (H₆N₂O, 80%), ethyl cellulose (viscosity:10 cp), ethyl cellulose (viscosity:46 cp), and terpineol and anhydrous ethanol (C₂H₆O) were used. All reagents were of analytical grade and used without further purification.

GO was prepared by the improved Hummers method with the reactive system temperature from high, low to medium, and the schematic diagram of GO preparation process is as shown in Figure 1.

Fig. 1

In the high-temperature oxidation period, 9.7 mL H₂SO₄ was added into a 500 mL beaker accurately, and 1.2 g microcrystalline graphite, 2 g K₂S₂O₈, and 1.2 g P₂O₅ were successively added to a beaker under stirring condition. Then the beaker was placed in a water bath at 80°C with stirring for 4.5 h. In the low-temperature oxidation period, 50 mL H₂SO₄ and 6 g KMnO₄ were added to the above mixed suspension in an ice-water bath at 5°C with stirring for 1 h, and a dark green suspension was obtained. In the medium temperature oxidation period, the water bath temperature was raised to 35°C, and the dark green suspension was stirred for 2 h, and then 100 mL deionized water was added to the suspension with stirring for 2 h. Finally, 280 mL deionized water was added to the above suspension with stirring condition, and then 8 mL H₂O₂ was added until the suspension turned to bright yellow. The bright yellow suspension was filtered and washed using 5% HCl solution and deionized water several times until the pH value of the filtrate was equal to 7. GO powder was obtained by freeze-drying the filter. The freeze-drying process consists of two stages. The first stage is that of pre-freezing, where the temperature, time, and vacuum degree were − 40°C, 5 h, and 1.013 × 10⁵ Pa, respectively; the second stage was the drying which is from − 40°C to 5°C with heating rate of 0.05°C/min under vacuum degree of 30 Pa.

ZnO ultrafine powder was synthesized by the solution combustion method. The precursor solution was obtained by mixing zinc nitrate, sucrose, and deionized water with a certain proportion. According to the principle of propellant chemistry, sucrose-zinc nitrate solution was completely combusted with releasing of N₂, CO₂, and H₂O gas according to Striker and Ruud [23], and the combustion reaction equation was as follows: (1) Zn(NO3)2+5φ24C12H22O11+5(φ−1)2O2→ignitingZnO+55φ24H2O+5φ2CO2+N2 \matrix{ {{\rm{Zn(N}}{{\rm{O}}_3}{)_2} + {{5\varphi } \over {24}}{{\rm{C}}_{12}}{{\rm{H}}_{22}}{{\rm{O}}_{11}} + {{5(\varphi - 1)} \over 2}{{\rm{O}}_2}} \cr {\buildrel {{\rm{igniting}}} \over \longrightarrow {\rm{ZnO}} + {{55\varphi } \over {24}}{{\rm{H}}_2}{\rm{O}} + {{5\varphi } \over 2}{\rm{C}}{{\rm{O}}_2} + {{\rm{N}}_2}} \cr } where φ is the molar ratio of sucrose and nitrate, which is called stoichiometric coefficient. When φ is equal to 1, the system is completely combusted according to the theory. Considering the factors such as heat dissipation of the system, φ equal to 1.5 was chosen (excessive fuel). Raw materials were weighed according to stoichiometric ratio (see Eq. (1)) and dissolved in deionized water. In the precursor solution, Zn²⁺ concentration was 0.2 mol/L and sucrose concentration was determined by using Eq. (1). A certain volume of precursor solution was taken to a refractory crucible and then placed in a furnace at 500°C. Subsequently, combustion reaction was occurred and powder was obtained after completion of the reaction. The detailed preparation process is shown in the relevant literature according to Li and Liu [25].

GO powder was accurately weighed and dispersed in 100 mL deionized water to prepare GO suspension at concentrations of 3.3 mg/mL, 9.8 mg/mL, 16.3 mg/mL, 22.8 mg/mL, and 29.3 mg/mL, respectively. After strong ultrasonic treatment for 2 h, the stable GO suspension was obtained. Then, the above GO suspension (1 mL) and ZnO powder (1.3 g) were mixed, respectively, with stirring for 60 min in a water bath at 35°C, and then 30 μL 80% hydrazine hydrate was added and stirred for 48 h to reduce it. Finally, the mixed suspension of rGO and ZnO was filtered and centrifuged several times until the pH value of the filtrate was equal to 7. Then the filter was dried in vacuum at 80°C for 2 h, so rGO/ZnO composite powder with 0%, 0.25%, 0.75%, 1.25%, 1.75%, and 2.25% rGO were obtained, respectively. The schematic diagram of rGO/ZnO composite powder preparation process is shown in Figure 2.

Fig. 2

Phase composition of the samples was analyzed by using XRD (D8-Advance, Brucker, Karlsruhe, Germany) with CuKα radiation (λ = 0.15418 nm) and step scan mode (range: 5°–90° of 2θ, step-time: 0.60 s, and step-size: 0.04°). The microstructure of the samples was characterized by using SEM (JSM-6700F, JEOL, Tokyo, Japan) and TEM (JEM-2010 HR, JEOL, Tokyo, Japan). Chemical structure of the samples was detected by using FT-IR (Nicolet 5700, Massachusetts, America) between 400/cm and 4,000/cm wavenumber with a resolution of 4/cm and Raman spectrometer (in Via, Renishaw, Gloucestershire, England) between 100/cm and 3,200/cm wavenumber with a resolution of 1/cm. The specific surface area of the samples was tested by using automatic physical-chemical adsorption instrument (ASAP-2020, Micromeritics, America) with high-purity nitrogen as adsorbent, and it was calculated by using BET (Brunauer-Emmett-Teller) method. Conductivity of rGO was tested by using a digital four-point probe tester (ST-2258A, Jingle, Suzhou, China). Photoelectric conversion performance of the solar cells was tested by using solar simulated light source (SXDN-150, Nowdata Corporation, Tokyo, Japan) with light intensity of AM1.5 (100 mW/cm²) and the area of the solar cells was 0.16 cm². The outer quantum efficiency of the solar cells was tested by using a photon-electron conversion tester (Solar Cell Scan100, ZOLIX, Beijing, China), and testing wavelength range was from 300 nm to 800 nm. EIS of the solar cells was tested by using electrochemical workstation (Zahner Zennium, Kronach, Germany) with detection frequency range from 10⁻¹ Hz to 10⁵ Hz.

Figure 4 shows XRD patterns of the samples prepared by the improved Hummers method. It can be seen from Figure 4 that there was an obvious characteristic diffraction peak at 2θ = 10°, which is consistent with the characteristic diffraction peak of GO according to Zhu et al. [2], indicating that GO is successfully prepared by the improved Hummers method.

Fig. 4

According to Bragg equation 2dsinθ = nλ, the interlaminar spacing for GO, rGO, and graphite is 0.81 nm, 0.41 nm, and 0.34 nm, respectively. The increase of the interlaminar spacing of GO is due to the insertion of a large number of oxygen-containing functional groups into carbon layer during oxidation process (see “3.2 Infrared and Raman Spectra Analysis” section, and with regard to rGO, due to the loss of the functional groups (see “3.2 Infrared and Raman Spectra Analysis” section), the interlayer spacing decreases, but it is still greater than that of graphite. In addition, the characteristic diffraction peaks of GO tended to widen comparing with the characteristic diffraction peaks of graphite. This is due to covalent bonding between functional groups and carbon atoms, resulting in the bending and folding of GO microstructure (as shown in Figure 4), and causing order degree of GO lamellar is lowered in the vertical direction than that of graphite as reported by Zhu et al. [2].

Figure 5(A) shows XRD pattern of the sample prepared by solution combustion method. It can be seen from Figure 5(A) that these typically peaks can be indexed to the crystal planes (100), (002), (101), (102), (110), (103), (112), and (201) (JCPDS cards No: 36-1451, ZnO), respectively. And there are no other impurity peaks in the pattern, indicating that the as-synthesized ZnO has high purity. Besides, it can also be seen from the characteristic peaks that the as-synthesized ZnO possesses hexagonal structure.

Fig. 5

Moreover, Figure 5(B) shows XRD pattern of the composite powder prepared by the chemical reduction method. It may also be seen from the pattern that there is a characteristic diffraction peak of rGO at 2θ = 24° according to Park et al. [26], and this indicates that GO is reduced to rGO. Through XRD analysis of the composite powder, it indicates that rGO/ZnO composite powder is synthesized.

Figure 6 shows FT-IR spectra of GO and rGO/ZnO prepared by the improved Hummers and the chemical reduction method, respectively. It can be seen from FT-IR spectrum of GO that there are many characteristic peaks of oxygen-containing functional groups between GO layers, among which absorption peak near 3,420/cm is caused by stretching vibration of −OH group; an obvious absorption peak can be found near 1,730/cm, which is main deformation peak caused by stretching vibration of −COOH and C=O groups at the edge of GO layer; peak near 1,625/cm corresponds to bending vibration of −OH group adsorbed on water molecule, while absorption peak near 1,220/cm is caused by stretching vibration of C–O. Due to oxygen-containing functional groups such as −OH, −COOH, C=O, and C–O, they can enter into graphite interlayer during oxidation reaction, and then destroyed the crystal structure of graphite. As a result, the layered structure of graphite was transformed into a separated lamellar structure, so GO is formed as reported by Shuai et al. [27] and Zaaba et al. [28]. Moreover, it can be seen from FT-IR spectrum of rGO/ZnO that absorption peak of −OH groups disappear, and absorption peaks of −COOH, C=O, and C–O disappear or decrease, and stretching vibration peak of C=C (1,564/cm) appears. In addition, there are many absorption peaks near 500/cm, which can be attributed to the vibrational absorption of Zn–O bond in ZnO. These indicate rGO/ZnO composite powder is synthesized by the chemical reduction method. Moreover, other rGO/ZnO with different contents of rGO have similar FT-IR spectra.

Fig. 6

Raman spectra for GO and rGO are shown in Figure 7. The observed peaks near 1,350/cm and 1,590/cm can be assigned to D (κ-point phonons breathing mode) and G (carbon atoms sp² bonds) bands, respectively. The ratio of the integral area values of D and G peaks (I_D/I_G, i.e., degree of disorder) changes from 1.05 for GO to 1.11 for rGO, which indicates an increase of defective sites by the chemical reduction and conversion of GO to rGO. However, there are still relatively low level of defects in rGO.

Fig. 7

SEM images of GO are shown in Figure 8. It can be seen from Figure 8A that GO has a layered structure, and there are a lot of folds on the surface of the layers, so the as-synthesized GO has a higher specific surface area of 650 m²/g. These indicate that GO has lost surface characteristics of graphite and changed from a flat and smooth surface to a rough and stepped surface. Moreover, it may be further seen from Figure 8(B) that there are different distances between GO layers. This is due to the strong oxidation of oxidant, hydroxyl, carbonyl, and other groups are introduced to the edge and interlayer of GO. At the same time, surface energy of the layers is increased when graphite is oxidized to GO. To maintain the stability of the layered surface, a large number of folds are formed on the surface. These indicate that the as-prepared GO has good microstructure.

Fig. 8

Figure 9 shows TEM images and selected area electron diffraction (SAED) of GO. It can be seen from Figure 9A that the layered GO is stacked disorderly and has good light transmittance near the edge. Figure 9B shows that GO presents transparent characteristics. To decrease the surface energy, the layered surface of GO is uneven and there are a lot of folds, which is consistent with the SEM analysis results. In addition, it can be seen from the inset in Figure 9B that there is a series of halo around the bright spot, which indicates that the as-synthesized GO has a polycrystalline structure.

Fig. 9

Figure 10 shows SEM images of ZnO powder, rGO, and the cross section of rGO/ZnO photoanode. Figure 10A shows that the as-synthesized ZnO shows short rods with particle size about 50 nm, and has good dispersibility, while Figure 10B shows that rGO has a typical layered structure, and the layered structure has a little agglomerate than that of GO. The rGO/ZnO photoanode shows porous and high porosity (see Figure 10C), and it can also be seen from Figure 10D that rGO/ZnO porous layer consists of rod and flake particles. The short rod particles is ZnO while the lamellar with size about 100–200 nm is rGO. In addition, it can be further seen from Figure 10D that ZnO film is about 200 nm between rGO/ZnO layer and FTO glass, which may increase utilization of incident light, and binding ability between rGO/ZnO layer and FTO glass. The photoanode with porous structure is beneficial to increasing adsorption capacity of dye molecules. Moreover, other photoanodes with adding different contents of rGO have similar microstructure.

Fig. 10

The J-V curves of the assembled solar cells with rGO/ZnO photoanodes are shown in Figure 11, and photoelectric performance parameters of the corresponding solar cells are shown in Table 1. It can be seen from Figure 11 and Table 1 that the effects of the photoanodes by adding different contents of rGO on the performance of the solar cells are very remarkable. PCE of D0 assembled by the photoanode without rGO (the pure ZnO) is 5.20%, and J_sc, V_oc, and FF are 0.63V, 12.73 mA/cm², and 64.70%, respectively. By increasing the content of adding rGO in the photoanodes, PCE, J_sc, and FF values of DSSCs increase initially and then decrease. The V_oc of the corresponding solar cells are maintained between 0.63 V and 0.66 V. When adding content of rGO in the photoanode at 1.25%, PCE of D3 reaches the maximum value of 6.27%, and J_sc, V_oc, and FF are 0.66 V, 13.11 mA/cm², and 72.71%, respectively. With the further increase of rGO content in the photoanodes, PCE of the solar cells decreases gradually. For example, when adding content of rGO in the photoanodes is 2.25%, PCE of D5 is decreased to 5.10%.

Fig. 11

The reasons for the above experimental results can be explained as follows: With pure ZnO as the photoanode, the solar cell has a relatively high efficiency (5.20%). This is because ZnO has a small crystal size (50 nm), relatively larger specific surface area (24.83 m²/g), and perfect crystal structure, but the slow separation rate of electrons and holes, and low transmission rate of electrons from ZnO conduction band to FTO glass lead to very easy recombination of electrons and holes, and thus PCE of the solar cell is lower than that of theoretical value as reported by Babar et al. [12] and Dhamodharan et al. [13]. As the as-synthesized rGO has excellent conductivity (5,530 S/m) and larger specific surface area (450 m²/g), when rGO is added to the photoanodes, it promotes separation of electrons and holes, decreases their recombination as reported by Wu et al. [29], and increases adsorption capacity of dye molecules (see Figure 10C), so PCE of the solar cells is increased. However, when adding content of rGO is too much, the content of ZnO in the photoanodes is relatively reduced. Subsequently, the number of electrons and holes is reduced, which is excited by photoelectric effect, and then PCE of the solar cells are lowered.

Figure 12 shows the external quantum efficiency (EQE) of D0, D1, D2, D3, D4, and D5. Figure 12 shows that with the increase of rGO content in the photoanodes, the EQE of the solar cells have a prominent enhancement trend in 400—700 nm. When adding content of rGO in the photoanodes is 1.25%, D3 has a maximum value of EQE. With further increase of rGO content in the photoanodes, the EQE of the solar cells decreases gradually. This is due to rGO that can improve the absorption ability and utilization efficiency of DSSCs for the light, which increases the number of photogenerated electrons and electrons mobility as reported by Riaza et al. [30] and Kumar et al. [31], so PCE of the solar cells is increased.

Fig. 12

Figure 13 shows Nyquist curves of D0, D1, D2, D3, D4, and D5 under dark conditions. Figure 13 shows that there are two arcs in the high-frequency region and the low-frequency region, in which the high-frequency region represents the charge transfer resistance (R_ct) at the interface between the photoanodes and FTO glass, and the low-frequency region represents the composite resistance (R_rec) at the interface of the photoanodes-dye-electrolyte. In the high-frequency region, the charge transfer resistance R_ct of D3 is lower than those of the other solar cells, while in the low-frequency region, the composite resistance R_rec of D3 is higher than those of the other solar cells, so D3 has the highest PEC. These results indicate that an appropriate amount of rGO added to the photoanodes can promote the transport of electrons and holes between the photoanodes and FTO glass, and reduces recombination of electrons and holes at the interface of the photoanodes-dye-electrolyte according to Omar et al. [32] and Quang et al. [33], and thus improves the performance of the solar cells. This also confirms that rGO may increase PCE of the solar cells from the electrochemical point of view.

Fig. 13