Improved interfacial charge transfer and visible light activity of reduced graphene oxide–graphitic carbon nitride photocatalysts

Abstract

Development of efficient visible light-driven photocatalysts is an important approach towards sustainability. Herein, reduced graphene oxide–graphitic carbon nitride (rGO–gCN) photocatalysts were successfully prepared by in situ photocatalytic reduction of graphene oxide (GO) in the presence of gCN as the photocatalyst. The XPS spectra revealed the presence of N-graphene, which confirmed the interaction between the rGO and the gCN. The rGO–gCN with GO loading amount of 0.1 wt% exhibited almost three times higher photocatalytic activity than gCN for degradation of phenol under visible light irradiation. Fluorescence spectroscopy, EIS, and photocurrent response studies indicated that the presence of rGO significantly promoted the electron–hole separation and improved the interfacial charge transfer on the gCN. These factors played a crucial role in enhancing the photocatalytic activity of the gCN under visible light irradiation.

---

Introduction

In recent years, the discovery and development of highly efficient photocatalysts are immensely reported worldwide, in which most studies are devoted to the investigation and employment of sustainable semiconductors which have visible-light active behaviour, owing to issues surrounding the depletion of natural resources. Previously, semiconductors such as metal oxides and metal sulphides were commonly used as photocatalysts, owing to their suitable band structure and high stability.^1,2 However, most of these materials are undesirable as they are costly, not visible-light active, and are not practical for large-scale production. Only a very few of them are visible-light active, such as a recently discovered mixed-valence tin oxide, Sn₃O₄, with narrow band gap energy of 2.4 eV.³ Therefore, there arises a great need to find an alternative semiconductor photocatalyst which is non-toxic, cheap and abundant, as well as visible-light active. One such highly promising material is graphitic carbon nitride (gCN).

gCN is a polymeric semiconductor which has attracted tremendous attention due to its remarkable physicochemical properties, such as small band gap of 2.7 eV, chemical inertness, high stability in both acid and basic media, and thermal stability in air up to 500 °C.^4–9 It was reported that this polymeric semiconductor shows great ability as a photocatalyst in the H₂ generation process in the presence of a sacrificial donor.¹⁰ Recent studies also demonstrated that modification of gCN with Ag₃PO₄ could produce a hybrid photocatalyst with remarkable performance in generating O₂ from water.^11,12 However, the pure gCN suffers from low photocatalytic performance, owing to its fast electron–hole recombination. In order to solve this problem, many attempts were made to enhance its photocatalytic performance through various modifications such as doping,⁴ crystal engineering,¹³ and hybrid coupling.^4,14 More recently, fabrication and modification of gCN with graphene (GR) to produce GR based photocatalysts is of particular interest.

Since two-dimensional (2D) GR was introduced in 2004,¹⁵ GR has been widely explored and used to modify semiconductor photocatalysts, owing to its outstanding properties such as high specific surface area, high electron conductivity, and high thermal stability.^16–22 Among them, the most modified semiconductor photocatalyst is TiO₂ since it has excellent properties such as non-toxicity, high stability and effective photocatalytic performance. The previous studies showed that the GR–TiO₂ photocatalysts showed better photocatalytic performance than bare TiO₂ due to the successful suppression of fast electron–hole recombination on TiO₂ by the GR.^16–22 However, even though the GR–TiO₂ composite showed a sign of slight visible light-responsive behaviour, the performance of GR–TiO₂ photocatalysts in the visible light region is still restricted due to the wide band gap of TiO₂ (∼3.2 eV). Therefore, developing graphene based visible light-responsive semiconductor photocatalysts is still a great challenge nowadays, considering that visible light contributes the most part of the solar spectrum.

On the other hand, graphene oxide (GO) is usually used as a modifier since it is more susceptible than GR, owing to the presence of abundant oxygen functional groups.^23–25 These oxygen functional groups enable the GO to interact with other inorganic and organic compounds to produce hybrid composite photocatalysts, such as in the formation of TiO₂/Ag₃PO₄/GO heterostructures.²⁶ However, GO is an insulating material with poor electron conductivity; it has to be reduced to form reduced graphene oxide (rGO) in order to restore its electroconductive network system.^23,24 Generally, rGO–gCN composites can be produced using an in situ approach, in which GO is mixed with the starting precursor of gCN and further reduced by means of chemical and thermal reductions.^27–30 As for the chemical reduction process, hydrazine was reported to act as a strong reducing agent to produce rGO–gCN photocatalysts.²⁷ However, the use of a toxic reducing agent is unfavorable and the residue of the reductant might be introduced into the rGO dispersions.³¹ Another reported reducing agent is sodium borohydride (NaBH₄). NaBH₄ is not preferred as it could react with water, which is the main solvent for GO dispersion. Moreover, additional alcohols were produced during the reduction process.²⁵ On the other hand, thermal reduction processes would produce rGO with more defects and a less crystalline structure.³² Hence, a mild reaction condition is highly desired to reduce GO to rGO. Of particular interest is the use of photocatalytic reduction method to produce rGO based composites in the presence of a semiconductor photocatalyst, as firstly reported for the preparation of the rGO–TiO₂ composite.³³ This method offers a few advantages, such as mild synthesis conditions, no involvement of toxic reducing agents and high heating temperatures, no impurities and no side reactions with water.

Herein, we report the synthesis of rGO–gCN photocatalysts by a photocatalytic reduction method using the bulk gCN as the photocatalyst. The use of photocatalytic reduction method to produce rGO–gCN photocatalysts has never been addressed yet. Moreover, there is no preliminary study on using bulk gCN as the semiconductor to perform photocatalytic reduction of GO to produce rGO–gCN photocatalysts. The photocatalytic performance of the prepared rGO–gCN photocatalysts was evaluated for degradation of phenol under visible light irradiation. Notably, with as low as 0.1 wt% of GO, we could obtain almost three times better photocatalytic efficiency as compared to the bare bulk gCN. The excellent photocatalytic performance of the rGO–gCN was due to the presence of rGO in promoting charge separation of electron–hole as well as improving interfacial charge transfer between gCN and rGO, as proven by EIS and transient photocurrent studies.

Experimental

Synthesis of GO

GO was synthesized from graphite flakes (Merck) via an improved Hummers’ method similar to the reported literature.³⁴ A mixture of concentrated H₂SO₄ (>95%, Fischer Scientific) and H₃PO₄ (85%, Merck) with a ratio of 9 : 1 (135 and 15 mL, respectively) was added to a mixture of graphite flakes (1.125 g) and KMnO₄ (6.75 g). The reaction was then heated to 50 °C and stirred using a water bath for 24 h. The reaction was cooled to room temperature and poured onto ice. Double distilled water (200 mL) was added, producing an exothermic reaction. H₂O₂ (30%, Fischer Scientific, 3 mL) was then added to reduce residual permanganate. The mixture was centrifuged at 3500 rpm for 10 minutes. The solid obtained was washed with HCl (30%, Fischer Scientific, 200 mL) twice, followed by washing with double distilled water until pH 6. For each washing, the mixture was centrifuged at 4000 rpm for 10 minutes. The solid was then dispersed in methanol and ultrasonicated for 60 minutes. Lastly, the solid was vacuum-dried overnight at room temperature.

Synthesis of gCN

gCN was synthesized by using a thermal polymerization approach according to the reported literature with cyanamide (99%, Sigma-Aldrich) as a starting precursor.^35,36 Briefly, a certain amount of cyanamide (50 g) was added into a crucible and calcined in a muffle furnace at 550 °C for 4 h with a ramp rate of 2.3 °C min⁻¹. The resulting yellow solid was then ground into powder.

Synthesis of rGO(x)–gCN

The rGO–gCN samples were synthesized via a photocatalytic reduction method in a similar way to the reported literature,³³ except that the gCN was used as the photocatalyst to reduce GO. A certain amount of the GO was added to gCN (1 g) that was dispersed in methanol (99.99%, Fischer Scientific, 70 mL) and ultrasonicated for 30 min to produce a good dispersion of the GO sheets with the gCN. The mixture was then irradiated under a UV lamp (8 W, I = 0.40 mW cm⁻²) for 24 h. The mixture was filtered and washed with double distilled water and ethanol, followed by drying overnight in an oven at 80 °C. The samples were denoted as rGO(x)–gCN, where x represented the ratio of the GO to the gCN (wt%) with x varying from 0.05 to 0.3 wt%. The reproducibility of the synthesis method up to three times was confirmed by synthesis of rGO(0.1)–gCN samples. For comparison purposes, a mixture sample consisting of GO and gCN with GO loading of 0.1 wt% was prepared by physical mixing using a mortar. In addition, the chemical reduction method using NaBH₄ as the reducing agent was also employed to prepare the rGO(0.1)–gCN composite in a similar way to the reported literature.³⁷ Typically, a mixture of the gCN (0.5 g) and GO was dispersed in methanol (99.99%, Fischer Scientific, 25 mL), followed by ultrasonication for 30 min. The NaBH₄ (0.284 g) was added into the dispersed mixture and the mixture was refluxed at 70 °C for 2 h. The mixture was then filtered under vacuum and washed with methanol and water until the pH of the filtrate was neutral. The obtained solid sample was dried in an oven at 30 °C for 5 days.

Characterizations

The crystal structure of prepared samples was investigated by an X-ray diffractometer (XRD) using a Bruker Advance D8 diffractometer with Cu Kα radiation (λ = 1.5406 Å) at a scan rate of 0.05° s⁻¹. The accelerating voltage and the applied current were 40 kV and 40 mA. The morphology of some selected samples was studied by a transmission electron microscope (TEM) using a JEOL JEM-2100 with an accelerating voltage of 200 kV. The surface chemical composition of the prepared samples was investigated by an X-ray photoelectron spectrometer (XPS) using a Shimadzu Axis Ultra DLD with monochromator Al Kα radiation source. Raman spectra of the samples were measured at room temperature on an XploRA Plus Raman Microscope HORIBA at wavelength of 532 nm. The functional groups of the prepared samples were confirmed by a Fourier transform infrared (FTIR) spectroscopy on a Thermo Scientific Nicolet iS50 with KBr as a reference sample. The optical properties of the samples were analysed by a diffuse reflectance (DR) UV-Vis spectroscopy using a Shimadzu UV-2600 spectrophotometer with BaSO₄ as a reflectance standard material. The fluorescence spectra were measured by a fluorescence spectroscopy (JASCO, FP-8500) at room temperature in the range of 200–800 nm. The excitation wavelengths were 278 and 369 nm, while the excitation and emission bandwidths were fixed at 5 nm.

Photocatalytic activity measurement

The photocatalytic activities of the samples were evaluated for the degradation of phenol under 3 h of visible light irradiation. A solution of phenol (50 mL, 50 ppm) which was prepared by using acetonitrile (99.9%, Merck) as the solvent, was placed in a 100 mL round-bottom flask, and 0.05 g of the sample was added. A halogen illuminator (150 W, λ > 400 nm) was used as the visible light source for all the photocatalytic tests. Prior to irradiation, the suspension was magnetically stirred in the dark for 1 h for adsorption–desorption equilibrium between the photocatalyst and phenol. All photocatalytic reactions were conducted for 3 h under visible light irradiation in a closed system. An Agilent Technologies 7820A Gas Chromatography (GC) system equipped with a flame ionization detector (FID) was used to determine the percentage degradation of phenol. The percentage degradation of phenol was calculated from the ratio of degraded phenol concentration to its initial concentration. To determine the amount of evolved CO₂, 5 mL of gas sample was taken from the reactor and analysed by a Shimadzu GC-2014 GC system equipped with a thermal conductivity detector (TCD). Reusability tests were carried out on the rGO(0.1)–gCN composite by reusing the photocatalyst for three times. In addition, the activity of the rGO(0.1)–gCN composite was evaluated for a 15 h-reaction under different light wavelengths by employing cut-off filters at 320, 380, 400 and 420 nm. Quantum efficiency was determined from the percentage ratio between the amount of converted phenol after a 24 h-reaction and the number of absorbed photons at 365 nm.

Electrochemical measurement

Electrode preparation. The electrode used for electrochemical measurement was a screen-printed electrode (SPE, DS 110) with carbon as the working and counter electrode, and silver as the reference electrode. The electrode preparation was as follows: the as-obtained sample (0.01 g) was added into a mixture of deionized water (1 mL) and Nafion® 117 solution (Sigma Aldrich, 10 μL). The mixture was then ultrasonicated for 10 min to produce a good dispersion of the catalyst. The supernatant of the sample (20 μL) was dropped onto the SPE and dried with a dryer.

EIS measurement. EIS data were recorded using a Gamry Interphase 1000 and all measurements were conducted under similar conditions with the same amount of sample loading. The electrolyte (6 mL) used was a mixture of Na₂SO₄ (0.1 M) and K₃[Fe(CN)₆] (2.5 mM) aqueous solutions. For Nyquist plots, the amplitude applied was 10 mV with a frequency range from 0.1 to 1 MHz.

Photoelectrochemical measurement

Electrode preparation. The measurements were carried out using a solar-cell sandwich concept, in which the sample was placed in between two pieces of ITO glass. The as-obtained sample (0.05 g) was added into absolute ethanol (0.5 mL) and ultrasonicated for 5 min to give well-dispersed slurry. The slurry (20 μL) was spread evenly onto ITO glass (ca. 1.5 × 1.5 cm) with three sides of the ITO being protected using Scotch tape. The deposited ITO was dried in air and Na₂SO₄ aqueous solution (0.5 M, 20 μL) was dropped onto it. Then, another ITO glass was placed onto the deposited ITO to produce a sandwich-like cell.

Transient photocurrent measurement. The working electrode cable was connected with the working sense electrode cable, whereas the counter, counter sense and reference electrode cables were connected together. The as-prepared ITO glass was irradiated with an 8 W UV lamp (λ = 254 nm). The data were recorded using a Gamry Interphase 1000 using chronoamperometry. All measurements were conducted under similar conditions with the same amount of sample and electrolyte loading.

Results and discussion

Characterizations of materials

The phase structure of the GO, the gCN and the rGO(x)–gCN composite samples were studied using an X-ray diffractometer (XRD). Fig. 1 shows the diffraction peaks of the as-obtained samples recorded at the diffraction angles from 5 to 60°. GO showed a distinct diffraction peak at 9.50° (d = 0.93 nm) for the (001) diffraction plane, corresponding to an interlayer distance of graphitic GO, as shown in Fig. 1(a). The value was found to be in good agreement with the value of previously reported literature, indicating the successful oxidation process of graphite flakes to GO via the improved Hummers’ method.³⁴ On the other hand, Fig. 1(b) shows that the bulk gCN exhibited two diffraction peaks at 12.95° (d = 0.683 nm) and 27.45° (d = 0.325 nm) for the (100) and (002) diffraction planes, which corresponded to in-planar structural packing of s-triazine units and interplanar stacking of the gCN, respectively. These two characteristic peaks of the gCN were consistent with the reported literature.^35,36

Fig. 1 XRD patterns of (a) GO, (b) gCN, (c) rGO(0.05)–gCN, (d) rGO(0.1)–gCN, (e) rGO(0.2)–gCN, and (f) rGO(0.3)–gCN samples.

XRD patterns of the rGO(x)–gCN samples are also shown in Fig. 1. All the samples exhibited the characteristic peaks of gCN, but with reduced diffraction intensity at the (002) plane (Fig. 1(c)–(f)). The decrease of the (002) plane was more obvious in the samples with a higher loading of GO. This result shows that the presence of rGO might affect the ordered graphitic structure of the gCN. On the other hand, the presence of rGO did not much interrupt the in-planar structural packing of the s-triazine units. All samples did not show any diffraction peaks of the rGO. This might be due to the low amount of added GO.

The morphology and microstructure of the GO, the gCN and the rGO(0.1)–gCN samples were investigated by a transmission electron microscopy (TEM). As can be seen from Fig. 2(a), the as-obtained GO appeared as thin-layered sheets with some formation of wrinkles. Such morphology indicated the GO was well exfoliated with enriched phenolic and epoxy functional groups on the basal plane.^23,34 On the other hand, the gCN showed a rough layered-stacking surface with an aggregated worm-like structure as shown in Fig. 2(b). As for the rGO(0.1)–gCN sample, due to the low amount of added GO, it was difficult to observe the wrinkle sheets of the rGO. However, the worm-like aggregation of the gCN was obviously seen in Fig. 2(c). This result showed that the morphology and microstructure of the gCN were not influenced by the rGO.

Fig. 2 TEM images of (a) GO, (b) gCN, and (c) rGO(0.1)–gCN samples.

XPS measurement was performed to analyse the surface chemical composition of the GO, the gCN and the rGO(0.1)–gCN composite sample, and also to prove the interaction between the rGO and the gCN. For the GO, as illustrated in Fig. 3(a), the regional C1s spectra of the GO could be deconvoluted into six peaks which appeared at 283.8, 284.8, 285.7, 286.4, 287.8 and 288.7 eV. These peaks were attributed to sp² hybridized carbon (C=C) of the graphene, adventitious carbon impurities adsorbed on the surface and/or sp³ hybridized carbon, phenolic group (C–OH), epoxy group (C–O–C), carbonyl group (C=O), and carboxyl group (COOH), respectively.^38,39 The presence of abundant oxygen functional groups further confirmed the successful oxidation of the graphite into GO. On the other hand, as can be observed in Fig. 3(b), four peaks were curve-fitted for the regional C1s spectra of the gCN, with binding energy values of 283.6, 284.8, 286.9 and 287.9 eV. The peaks at 283.6 and 284.8 eV were due to the adventitious carbon impurities adsorbed on the surface.⁴⁰ The peaks which appeared at 286.9 and 287.9 were ascribable to the C–N=C and C–(N)₃ groups of typical gCN, suggesting the successful thermal polymerization of the gCN from cyanamide.⁴¹ As for the rGO(0.1)–gCN composite sample, as shown in Fig. 3(c), it can be observed clearly that the C–N=C component was detectable after the introduction of the GO, indicating the undisturbed s-triazine structure of the gCN. The C=C component of the graphene and some unreduced oxygen functional groups (C–OH, C=O) were still observed in the spectra, further suggesting that the UV-assisted photoreduction method successfully provided mild reduction conditions as the GO was not fully reduced.

Fig. 3 XPS C1s spectra of (a) GO, (b) gCN, and (c) rGO(0.1)–gCN with their deconvolutions. The fitted curves are shown with open circle symbols.

Fig. 4 shows the regional XPS N1s spectra of the gCN and rGO(0.1)–gCN composite samples. Upon deconvolution, three peaks were obtained for the gCN with binding energy values of 397.3, 398.2, and 399.6 eV, respectively. The peak at 397.3 eV was ascribed to pyridinic nitrogen (C–N=C), confirming the formation of s-triazine of the CN.⁴¹ On the other hand, the peak at 398.2 eV was assigned to pyrrolic nitrogen (N–(C)₃), showing the existence of bridging N atoms in the CN structure, while the peak at 399.6 eV was attributed to the presence of surface amino (C–N–H) functional groups.⁴¹ For the rGO(0.1)–gCN composite sample, the N1s spectrum was deconvoluted into four peaks, which were 396.4, 397.3, 398.2 and 399.6 eV, respectively. Notably, a new shoulder peak was formed at 396.4, which would be caused by the new nitrogen–graphene interaction. A similar observation was also obtained and reported when gCN was doped with cobalt and supported on graphene.²⁹ The presence of this new peak at 396.4 eV indicated that the gCN could interact strongly with the rGO, which was favourable in the interfacial charge transfer.

Fig. 4 XPS N1s spectra of (a) gCN, and (b) rGO(0.1)–gCN with their deconvolutions. The fitted curves are shown with open circle symbols.

The chemical structure of the samples was further confirmed by Raman and Fourier transform infrared (FTIR) spectroscopies. Raman spectra of the samples are shown in Fig. S1.† The graphene exhibited one sharp G peak, while GO exhibited two peaks which were attributed to the D and G bands. However, no prominent peaks could be detected from the gCN and rGO–gCN composite, owing to the strong fluorescence background of the gCN. The FTIR spectrum of the GO is shown in Fig. 5(a). The peaks at 1732 and 1053 cm⁻¹ were attributed to the C=O stretching of COOH and C–O–C functional groups, respectively.^27–30 The presence of these peaks showed the detection of oxygen functional groups on the GO, indicating that the GO was successfully prepared by the improved Hummers’ method. Meanwhile, the peak at 1623 cm⁻¹ was ascribed to the deformation vibration of intercalated water.^27–30 Fig. 5(b) shows the FTIR spectrum of the gCN. Peaks in the region of 1200–1650 cm⁻¹ were originating from the stretching modes of CN heterocycles and the peak found at 806 cm⁻¹ was corresponding to the s-triazine units of the gCN.^{27–30,35,36,42,43} The broad peaks observed between 3500 and 3000 cm⁻¹ were assigned to the N–H stretches, resulting from hydrogenation on nitrogen atoms during thermal polymerization or the presence of terminal-NHx (x = 1, 2) at the defect regions of the gCN.⁴² The detection of characteristic peaks of the gCN indicated that the gCN was successfully formed via thermal polymerization. Despite the difference in the GO loading amount, all the rGO(x)–gCN samples showed similar FTIR spectra to that of the gCN, as shown in Fig. 5(c)–(f). This result clearly showed that the rGO did not interrupt the chemical structure of the gCN. It can be suggested that the photocatalytic reduction method is a mild and promising method to prepare the rGO–gCN without influencing the structure of the basic building blocks in the gCN.

Fig. 5 FTIR spectra of (a) GO, (b) gCN, (c) rGO(0.05)–gCN, (d) rGO(0.1)–gCN, (e) rGO(0.2)–gCN, and (f) rGO(0.3)–gCN samples.

The optical properties of the GO, the gCN and the rGO(x)–gCN samples were investigated by diffuse reflectance ultraviolet-visible (DR UV-Vis) spectroscopy. As shown in Fig. 6, the GO exhibited a broad peak centered at 345 nm, suggesting the presence of C=O functional groups in the GO with electron transition from n to π*.³⁴ The gCN exhibited two distinct peaks at 291 and 371 nm, which corresponded to the electron transitions from π to π* at the C=N bonds and from n to π* at the terminal C–N groups, respectively.^{35,36,42–44} All the rGO(x)–gCN samples gave similar absorption spectra to that of the gCN. The similar absorption spectra indicated that the formed rGO did not have a significant influence on the optical properties of the gCN. As shown in the inset of Fig. 6, the background absorption in the visible light region of the gCN was only slightly improved by the presence of rGO. Moreover, all the rGO(x)–gCN samples exhibited similar band edge absorption from the gCN, which was up to 460 nm. This result clearly suggested that the band gap of the gCN was not altered by the rGO, which was in good agreement with the rGO–gCN composites prepared by chemical and thermal reduction methods.^27,28

Fig. 6 DR UV-Vis spectra of the GO, the gCN and the rGO(x)–gCN samples.

The effect of rGO on the emission properties of the gCN and the interactions between the rGO and the gCN were investigated by fluorescence spectroscopy. Two excitation wavelengths, which were 278 and 369 nm, were used to monitor the emission spectra as shown in Fig. 7. Analogous to the peak assignment for the DR UV-Vis spectra, the excitation peak at 278 nm was ascribed to π–π* electronic transition, which originated from C=N aromatic rings in the gCN. On the other hand, the excitation peak at 369 nm was attributed to n–π* electronic transition, which originated from the lone pair electrons at the terminal N–C groups of the gCN. Only one emission peak at 460 nm was detected when the gCN was excited at either 278 or 369 nm. These values were in good agreement with the previous reported literature.^35,36

Fig. 7 Emission spectra of the gCN and the rGO(x)–gCN samples monitored at the excitation wavelength of (a) 278 and (b) 369 nm, respectively.

Under an excitation wavelength of 278 nm, all the rGO(x)–gCN samples displayed similar emission peaks to that observed on the gCN. The rGO(0.05)–gCN sample exhibited a slightly more quenched emission intensity than that of the gCN, whereas the other samples with higher loading of GO (0.1–0.3 wt%) showed more significantly reduced emission intensities. This result suggested that the rGO could interact with the gCN via π–π stacking, owing to the presence of aromatic networks in both gCN and rGO.^28,44 When the emission spectra were monitored under an excitation wavelength of 369 nm, a similar emission trend was observed, which indicated that the rGO could also interact with the gCN via terminal N–C groups. These results were in good agreement with the XPS results, showing that there were certain interactions between the N of the gCN and the rGO. Good interactions between the rGO and the gCN could be associated with the good electronic charge transfer between them. The lower emission intensity observed on the rGO(x)–gCN samples also suggested lower electron–hole recombination on these samples than on the gCN. In other words, it was demonstrated that the presence of rGO successfully suppressed electron–hole recombination on the gCN.

In order to determine the favourable interaction site between the rGO and the gCN, the Stern–Volmer equation was used, as shown in eqn (1).^35,36,45

where I_o is the initial emission intensity of the gCN, I is the quenched emission intensity in the presence of the rGO, K_SV is the Stern–Volmer quenching constant, and Q is the concentration of GO (wt%). The K_SV values were determined as the slope gradient of the Stern–Volmer equation, in which a higher K_SV value is ascribed to a more favourable interaction site. The Stern–Volmer plots for the excitation wavelengths of 278 and 369 nm are shown in Fig. 8. Upon calculation using the Stern–Volmer equation, the K_SV values for the excitation wavelengths of 278 and 369 nm were determined to be 0.41 and 0.33 wt%⁻¹, respectively. The K_SV value at an excitation wavelength of 278 nm was only slightly higher than that at 369 nm, implying that the rGO could interact almost equally with both emission sites of the gCN, which were N=C and N-terminal groups. These interactions could contribute to efficient interfacial charge transfer, which subsequently might lead to the improved photocatalytic performance of the gCN.

Fig. 8 Stern–Volmer plots between the relative emission ratio and the weight percentages of GO at excitation wavelengths of 278 and 369 nm.

Photocatalytic tests

The photocatalytic performances of the gCN and rGO(x)–gCN samples were assessed for the degradation of phenol under visible light irradiation. Dark adsorption of phenol was conducted for 1 h to ensure adsorption–desorption equilibrium. It was confirmed that there was no remarkable adsorption of phenol that could be detected in the dark after stirring for 1 h. The photocatalytic activity of the gCN and rGO(x)–gCN samples are shown in Fig. 9. The gCN showed photocatalytic activity for degradation of phenol with percentage degradation of 6%. It was clear that all rGO(x)–gCN composite samples exhibited significant enhancements in the photocatalytic efficiency for the degradation of phenol as compared to the gCN, even though the amount of GO loaded was extremely small. For only 0.05 wt% of GO loading, the activity of the composite sample was increased to 14% as compared to the gCN with only 6% activity. The activity of the gCN was further increased by almost three times, which was 17% phenol degradation, when the loading of GO was increased to 0.1 wt%. Although higher loadings of GO wt% did not further improve the photocatalytic degradation of phenol, the rGO(0.2)–gCN and rGO(0.3)–gCN composite samples still showed better performance than the gCN.

Fig. 9 Photocatalytic degradation of phenol by the gCN and the rGO(x)–gCN photocatalysts under 3 h of visible light irradiation.

Activity comparison was made using the sample prepared by physical mixing between GO (0.1 wt%) and gCN. The activity obtained on this sample was only 7%, suggesting the importance of the rGO species in the initial reaction. GO itself did not give any activity under the current reaction conditions. On the other hand, the composite sample prepared via the chemical reduction method only showed a slight improvement in photocatalytic activity, which was 10%. These results further supported that the photocatalytic reduction method is a good method to prepare rGO(x)–gCN composite photocatalysts with efficient photocatalytic performance. Moreover, it is worth noting here that the present synthesis method showed good reproducibility from batch to batch. From three batches, the prepared rGO(0.1)–gCN samples gave very similar activity performance, which was ca. 17% phenol degradation. It was confirmed that CO₂ gas was evolved as the final product of the reaction as shown in Fig. S2 and Table S1.† When most of the phenol was totally degraded, the mol ratio of experimentally detected CO₂ to the theoretical expected CO₂ was close to one, suggesting that the photocatalyst was able to fully degrade the phenol to CO₂ and H₂O.

The photostability and reusability of the rGO(0.1)–gCN photocatalyst was evaluated for three successive cycles under the same experimental conditions. As illustrated in Fig. S3,† no significant deterioration was observed in the photocatalytic activity after three cycles. Moreover, as can be seen in Fig. S4,† the structure of the composite still remained intact and unchanged after three cycles, suggesting that the rGO(0.1)–gCN did not suffer from self-decomposition under irradiation. These results indicated that the rGO(0.1)–gCN composite exhibited good photostability and reusability for the degradation of phenol.

In order to demonstrate that the rGO(0.1)–gCN composite is an effective photocatalyst, an action spectrum was investigated by analysing its photocatalytic performance under different light wavelengths. As shown in Fig. S5,† the trend of photocatalytic activity well-resembled the absorption spectrum of the rGO(0.1)–gCN, indicating that the degradation of phenol was driven by the light absorption of the photocatalyst. The quantum efficiency of the rGO(0.1)–gCN photocatalyst was determined to be 2.92%.

In this study, as observed earlier in the DR UV-Vis spectra, the presence of the rGO only slightly improved the background absorption of the gCN in the visible-light region. And yet, this slight improvement of background absorption might not be the influential factor in enhancing the photocatalytic performance of the gCN photocatalyst. On the other hand, fluorescence studies showed that the rGO interacted with the gCN and it successfully reduced the electron–hole recombination on the gCN. However, the optimum loading of GO was important to give the optimum photocatalytic activity as shown in Fig. 9. The rGO(x)–gCN samples prepared via the photocatalytic reduction method gave remarkable photocatalytic performance with optimum loading of GO of 0.1 wt%. A high loading of GO might cause a masking effect, in which the active sites of the gCN photocatalyst might be blocked by the formed rGO or the added GO. Moreover, the excess amount of GO might have affected the interfacial charge transfer as GO has poor electron conductivity. Therefore, the excellent photocatalytic performance of the composite samples should be attributed to the optimal interactions formed between the rGO and gCN, which led to the suppression of electron–hole recombination and improved interfacial charge transfer. In order to reveal the cause and mechanism of why the optimum loading promoted the photocatalysis, further investigations on the electrochemical behaviour of the gCN and rGO(x)–gCN composite samples were carried out via EIS and photocurrent response studies.

Interfacial charge transfer

The efficiency of the charge separation on the gCN, rGO(0.1)–gCN and rGO(0.2)–gCN was investigated by using EIS and is shown in Fig. 10. The Nyquist plot showed that the semicircle in the high frequency region was attributed to the charge transfer resistance (R_ct) and constant phase element (CPE) at the electrode/electrolyte interface.⁴⁶ A smaller arc radius of the semicircle indicates a smaller R_ct between the working electrode and electrolyte. With a smaller R_ct, the material would demonstrate better electron conductivity and charge transfer capability. Whereas, the straight line inclined at approximately 45° to the real axis in the low frequency region corresponded to the diffusion of ions at the electrode/electrolyte interface and is known as Warburg impedance (W_d).⁴⁶ In this study, an equivalent circuit model (CPE with diffusion) was employed and fitted by using a simplex model program in Gamry Echem Analyst.

Fig. 10 EIS investigation on the gCN, rGO(0.1)–gCN and rGO(0.2)–gCN photocatalysts using 0.1 M Na₂SO₄ and 2.5 mM K₃[Fe(CN)₆] electrolyte.

As shown in Fig. 10, it was observed that both the rGO(0.1)–gCN and rGO(0.2)–gCN composite samples showed smaller semicircles compared to the gCN, implying a smaller charge resistance than the bare gCN. The obtained R_ct values for the gCN, rGO(0.1)–gCN and GO(0.2)–gCN were 7.3, 2.7, and 5.2 kΩ, respectively. These results indicated that the presence of the rGO on both composite samples could accelerate the electron transfer and enhanced the electron conductivity of the gCN. However, the rGO(0.2)–gCN composite sample showed larger R_ct than the rGO(0.1)–gCN. This might be due to the presence of excess GO that could have impeded the charge transfer as GO has poor electron conductivity. This would be one of the reasons for the decreased photocatalytic performance in the degradation of phenol at higher loadings of GO.

The heterogeneous electron-transfer rate constant (k) can be calculated using eqn (2), where R is the gas constant, T is temperature (K), F is the Faraday constant, A is the electrode area (cm²), R_ct is the charge transfer resistance, C⁰ is the concentration of redox couples in the bulk solution, and n is the number of transferred electrons per molecule of the redox probe.

Based on eqn (2), the heterogeneous electron-transfer rate constants for the gCN, rGO(0.1)–gCN and rGO(0.2)–gCN were determined to be 1.16 × 10⁻⁴, 3.14 × 10⁻⁴ and 1.63 × 10⁻⁴ cm s⁻¹, respectively. These results showed that the electron transfer occurred faster on the composite samples, where the rGO(0.1)–gCN gave faster electron transfer than the rGO(0.2)–gCN.

Since W_d was related to the diffusion of ions at the electrode/electrolyte interface, a smaller W_d value indicates that the material has better diffusion as less resistance is present in the flow of ions at the interface. The obtained W_d values for the gCN, rGO(0.1)–gCN and rGO(0.2)–gCN were 110 × 10⁻⁶, 80 × 10⁻⁶, and 83 × 10⁻⁶ (Ss^1/2), respectively. Herein, it was demonstrated that both composite samples exhibited better diffusion than the bare gCN, with the rGO(0.1)–gCN showing slightly better diffusion than the rGO(0.2)–gCN. All these findings indicated that the rGO can act as a good electron shuttle in transferring electrons, in which better electron transfer can lead to better photocatalytic performance.

There have been many studies reporting that rGO has good electron conductivity, owing to its two-dimensional and π-conjugated network system.^16–24 Thus, in the rGO(0.1)–gCN composite sample, rGO would act as a good electron transport, which would successfully shuttle the photosensitive charge carriers and suppress the fast electron–hole recombination in the gCN. Likewise, a significantly enhanced charge transfer could occur from gCN to rGO due to the good distribution of charge density prompted by the good mounting of rGO on gCN, as investigated by the hybrid DFT method.⁴⁴ All these findings suggested that the enhanced charge separations and improved interfacial charge transfers are the key factors that contributed to the better photocatalytic performance of rGO(0.1)–gCN in the degradation of phenol.

To give further evidence to confirm the better charge separation and interfacial charge transfer on the rGO–gCN than the gCN, transient photocurrent studies of gCN and rGO(0.1)–gCN were investigated for several cycles under intermittent irradiation. As shown in Fig. 11, it can be clearly seen that the photocurrent density value increased rapidly until it reached a constant value when the light was turned on, whereas the value decreased quickly to zero when the light was turned off. These dramatic changes suggested that both the gCN and the rGO(0.1)–gCN are light-responsive materials. For the gCN sample, it was observed that the photocurrent density was slightly decayed after each cycle. In contrast, the rGO(0.1)–gCN showed relatively constant photocurrent density, indicating its better stability than gCN. Moreover, the photocurrent density of the rGO(0.1)–gCN was much higher than the gCN, suggesting that the rGO(0.1)–gCN has better electron conductivity than gCN. The rGO(0.1)–gCN sample gave off a photocurrent spike at the first second of each cycle of irradiation and the current spike decreased with the time until a constant current was attained. This might be due to a certain degree of photorecombination. However, there was no photocurrent decay observed for each intermittent irradiation, showing that the rGO could suppress electron–hole recombination and also facilitate the interfacial charge transfer.

Fig. 11 Transient photocurrent responses of the gCN and the rGO(0.1)–gCN in 0.1 M Na₂SO₄ electrolyte under 8 W UV irradiation at 0 V.

Conclusions

In conclusion, the mild synthesis condition of the photocatalytic reduction method is a good approach to fabricate rGO–gCN photocatalysts. The presence of rGO did not interrupt the in-planar structural packing of gCN but inhibited the recombination of photogenerated charge carriers and further enhanced the interfacial charge transfer. The hybrid photocatalyst synthesized under this mild synthesis condition showed remarkable photocatalytic performance in the degradation of phenol. With GO loading as small as 0.1 wt%, the fast electron–hole recombination in the gCN was suppressed and the interface charge transfer was improved, leading to almost three times better performance than that of gCN.

Acknowledgements

This work was financially supported by the Ministry of Higher Education (MOHE) and Universiti Teknologi Malaysia via Research University Grant (Tier 1, cost center code: Q.J130000.2526.03H85). Financial support from MyBrain PhD Scholarship is also greatly acknowledged (PT).

Notes and references

1. H. Wang, L. Zhang, Z. Chen, J. Hu, S. Li, H. Wang, J. Liu and X. Wang, Chem. Soc. Rev., 2014, 43, 5234–5244.
2. K. Zhang and L. Guo, Catal. Sci. Technol., 2013, 3, 1672–1690.
3. M. Manikandan, T. Tanabe, P. Li, S. Ueda, G. V. Ramesh, R. Kodiyath, J. Wang, T. Hara, A. Dakshanamoorthy, S. Ishihara, K. Ariga, J. Ye, N. Umezawa and H. Abe, ACS Appl. Mater. Interfaces, 2014, 6, 3790–3793.
4. Z. Zhao, Y. Sun and F. Do, Nanoscale, 2015, 7, 15–37.
5. Y. Zhang, J. Liu, J. Liang, M. Jaroniec and S. Z. Qiao, Energy Environ. Sci., 2012, 5, 6717–7673.
6. X. Wang, S. Blechert and M. Antonietti, ACS Catal., 2012, 2, 1596–1606.
7. Y. Wang, X. Wang and M. Antonietti, Angew. Chem., Int. Ed., 2012, 51, 68–89.
8. Y. Wang, T. Mori and J. Ye, Sci. Adv. Mater., 2012, 4, 282–291.
9. S. Cao, J. Low, J. Yu and M. Jaroniec, Adv. Mater., 2015, 27, 2150–2176.
10. X. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson, K. Domen and M. Antonietti, Nat. Mater., 2009, 8, 76–80.
11. X. Yang, H. Tang, J. Xu, M. Antonietti and M. Shalom, ChemSusChem, 2015, 8, 1350–1358.
12. X. Yang, Z. Chen, J. Xu, H. Tang, K. Chen and Y. Jiang, ACS Appl. Mater. Interfaces, 2015, 7, 15285–15293.
13. G. Algara-Siller, N. Severin, S. Y. Chong, T. Bjçrkman, R. G. Palgrave, A. Laybourn, M. Antonietti, Y. Z. Khimyak, A. V. Krasheninnikov, J. P. Rabe, U. Kaiser, A. I. Cooper, A. Thomas and M. J. Bojdys, Angew. Chem., Int. Ed., 2014, 53, 7450–7455.
14. J. Zhang, M. Zhang, R.-Q. Sun and X. Wang, Angew. Chem., Int. Ed., 2012, 51, 10145–10149.
15. K. S. Novoselov and A. K. Geim, Science, 2004, 306, 666–669.
16. X. An and J. C. Yu, RSC Adv., 2011, 1, 1426–1434.
17. M.-T. Sergio and P.-M. M. Luisa, Environ. Sci. Pollut. Res., 2012, 19, 3676–3687.
18. L.-L. Tan, S.-P. Chai and A. R. Mohamed, ChemSusChem, 2012, 5, 1868–1882.
19. Q. Xiang, J. Yu and M. Jaroniec, Chem. Soc. Rev., 2012, 41, 782–796.
20. N. Zhang, Y. Zhang and Y.-J. Xu, Nanoscale, 2012, 4, 5792–5813.
21. Q. Xiang and J. Yu, J. Phys. Chem. Lett., 2013, 4, 753–759.
22. W. Tu, Y. Zhou and Z. Zou, Adv. Funct. Mater., 2013, 23, 4996–5008.
23. S. Park and R. S. Ruoff, Nat. Nanotechnol., 2009, 58, 1–8.
24. H.-C. Hsu, I. Shown, H.-Y. Wei, Y.-C. Chang, H.-Y. Du, Y.-G. Lin, C.-A. Tseng, C.-H. Wang, L.-C. Chen, Y.-C. Lin and K.-H. Chen, Nanoscale, 2013, 5, 262–268.
25. D. R. Dreyer, S. Park, C. W. Bielawski and R. S. Ruoff, Chem. Soc. Rev., 2010, 39, 228–240.
26. X. Yang, J. Qin, Y. Jiang, K. Chen, X. Yan, D. Zhang, R. Li and H. Tang, Appl. Catal., B, 2015, 166, 231–240.
27. Q. Xiang, J. Yu and M. Jaroniec, J. Phys. Chem. C, 2011, 115, 7355–7363.
28. Y. Zhang, T. Mori, L. Niu and J. Ye, Energy Environ. Sci., 2011, 4, 4517–4521.
29. Q. Liu and J. Zhang, Langmuir, 2013, 29, 3821–3828.
30. J. Fu, C. Hu, X. Wu and X. Wang, Nanoscale, 2014, 6, 12555–12564.
31. J. Gao, F. Liu, Z. Wang and X. Zhang, Chem. Mater., 2010, 22, 2213–2218.
32. C. H. A. Wong, A. Ambrosi and M. Pumera, Nanoscale, 2012, 4, 4972–4977.
33. G. Williams, B. Seger and P. V. Kamat, ACS Nano, 2008, 2, 1487–1491.
34. D. C. Marcano, D. V. Kosynkin, J. M. Berlin, A. Sinitskii, Z. Sun, A. Slesarev, L. B. Alemany, W. Lu and J. M. Tour, ACS Nano, 2010, 4, 4806–4814.
35. M. S. Sam, H. O. Lintang, S. L. Lee and L. Yuliati, Spectrochim. Acta, Part A, 2014, 124, 357–364.
36. M. S. Sam, P. Tiong, H. O. Lintang, S. L. Lee and L. Yuliati, RSC Adv., 2015, 5, 44578–44586.
37. C. K. Chua and M. Pumera, J. Mater. Chem. A, 2013, 1, 1892–1898.
38. L. Q. Xu, Y. L. Liu, K.-G. Neoh, E.-T. Kang and G. D. Fu, Macromol. Rapid Commun., 2011, 32, 684–688.
39. M. Xing, X. Li and J. Zhang, Sci. Rep., 2014, 4(5493), 1–7.
40. R. C. Pawar, V. Khareb and C. S. Lee, Dalton Trans., 2014, 43, 12514–12527.
41. B. Chai, T. Peng, J. Mao, K. Lia and L. Zan, Phys. Chem. Chem. Phys., 2012, 14, 16745–16752.
42. Y. Li, J. Zhang, Q. Wang, Y. Jin, D. Huang, Q. Cui and G. Zou, J. Phys. Chem. B, 2010, 114, 9429–9434.
43. D. R. Miller, J. Wang and E. G. Gillan, J. Mater. Chem., 2002, 12, 2463–2469.
44. A. Du, S. Sanvito, Z. Li, D. Wang, Y. Jiao, T. Liao, Q. Sun, Y. H. Ng, R. Amal and S. C. Smith, J. Am. Chem. Soc., 2012, 134, 4393–4397.
45. E. Blatt, R. C. Chatelier and W. H. Sawyer, Biophys. J., 1986, 50, 349–356.
46. J. Li, X. Liu, L. Pan, W. Qin, T. Chena and Z. Suna, RSC Adv., 2014, 4, 9647–9651.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra17967j

---