CdS–graphene and CdS–CNT nanocomposites as visible-light photocatalysts for hydrogen evolution and organic dye degradation

Abstract

CdS–graphene (GR) and CdS–carbon nanotube (CNT) nanocomposites prepared by a hydrothermal method were studied as photocatalysts for the evolution of hydrogen and the degradation of methyl orange (MO) under visible-light irradiation. The incorporation of GR or CNT into CdS significantly enhanced the photocatalytic activities for both reactions. The photocatalytic activities depended on the mass ratio of CdS : GR or CdS : CNT in the nanocomposites. Under optimized mass ratios, the CdS–GR was more efficient than the CdS–CNT. Our characterizations suggested that CdS nanoparticles of ∼35 nm in size were dispersed on the graphene sheets or CNT surfaces in the nanocomposites. Significant band-gap narrowing was observed due to the incorporation of GR or CNT into CdS, indicating the strong interactions between CdS and GR or CNT. The transient photocurrent response studies suggested a more efficient separation of photogenerated electrons and holes. The stronger interaction or larger contact interface between CdS and GR was proposed to accelerate the transfer of photogenerated electrons from CdS to GR more efficiently, resulting in higher photocatalytic activities of the CdS–GR composite.

---

1. Introduction

Semiconductor photocatalysts have attracted much attention in recent years because of their potential in solving global energy and environmental problems. Many photocatalysts have been reported for the splitting of water in the presence or absence of sacrificial agents, producing H₂ as a clean energy source, and for the degradation of organic pollutants.^1–14 However, most of these catalysts require ultraviolet (UV) light and their efficiencies are far from practical application requirements. The development of efficient visible-light-responsive photocatalysts is a highly challenging but desirable goal for the effective utilization of solar energy because visible light in the solar spectrum is much more abundant (∼43%) than UV light (∼4%).¹¹

CdS, which possesses a bandgap of ∼2.4 eV, is a simple and probably the most extensively studied visible-light-driven photocatalyst.^15,16 However, the efficiency of CdS is very low because of the rapid recombination of the photogenerated electrons and holes. Most of the studies concerning CdS are focused on the enhancement of photocatalytic efficiencies by preparing nanocomposites containing CdS and other semiconductors, such as TiO₂, ZnO, MoS₂, PdS or layered titanate nanosheets, to suppress the fast recombination of electrons and holes.^17–22

On the other hand, recent studies have revealed that carbon nanomaterials, particularly carbon nanotubes (CNTs), are promising co-catalysts in photocatalysis. The enhancement effects of CNTs on the photocatalytic performance of TiO₂, a UV-light-driven photocatalyst, have been investigated by several research groups.^23–26 It is proposed that CNTs are capable of accepting, transporting and storing electrons, and thus retarding or hindering the recombination of the electrons with the holes remaining on the semiconductor particles.^23,25 However, the studies on the enhancement effects of carbon nanomaterials on photocatalytic behaviors of CdS are not numerous. Kamat and co-workers²⁷ once reported that the deposition of CdS onto CNTs could lead to a fast electron transfer from photo-excited CdS into CNTs. Recently, several studies disclosed that the incorporation of CNTs into CdS could promote its photocatalytic performance.^28,29

Graphene (GR), which is a two-dimensional sp²-hybridized carbon nanosheet, possesses many unique properties such as a very high theoretical specific surface area, high mobility of charge carriers and good mechanical strength.^30–34 The use of GR as a co-catalyst in photocatalysis has drawn increasing attention recently.^13,35 Some studies have found that the incorporation of GR into TiO₂ can enhance the photocatalytic performance for the degradation of organic dyes and the evolution of H₂ from aqueous solutions containing sacrificial agents, such as methanol, under UV-light irradiation.^13,36–39

Very few studies have been devoted to investigating the photocatalytic properties of CdS–GR composites.^40,41 Yu and co-workers⁴⁰ recently reported an interesting study on the use of CdS–GR composites, prepared by a solvothermal method⁴² at 453 K, in which graphene oxide (GO), Cd(Ac)₂ and dimethyl sulfoxide (DMSO) were used as precursors for H₂ evolution from a 10% lactic acid aqueous solution under visible-light irradiation. Compared to CdS alone, the combination of GR into CdS in the presence of a Pt co-catalyst afforded a H₂ formation rate ∼4.9 times higher. The use of DMSO as the precursor for sulfide and the solvent in the preparation would result in large amounts of sulfur-containing organic waste and is not environmentally friendly. Xu and co-workers⁴¹ prepared nitrogen-doped graphene (N-GR) by annealing GO in an NH₃ gas flow at 773 K, and a CdS–N-GR composite by precipitating CdS in the suspension containing N-GR. Photocatalytic studies revealed that the CdS–N-GR composite showed a rate of H₂ evolution from an aqueous solution of Na₂S and Na₂SO₃ ∼5 times higher than CdS alone. The CdS–GR composite, which was prepared by the same method using GR instead of N-GR, exhibited a rate of H₂ evolution only twice as high as CdS. As N-GR and GR are both hydrophobic, the interactions between CdS and these carbon materials may be not strong enough. In fact, a large fraction of N-GR or GR could not be well dispersed in aqueous solution.⁴¹

Therefore, it is still necessary to develop an environmentally benign method for the preparation of CdS–GR composites with strong interactions between CdS and GR. Moreover, although both GR^40,41 and CNT^28,29 have been demonstrated to enhance the photocatalytic performances of CdS, there has been no study devoted to comparing the promotional effects of GR and CNT in composites with CdS. Xu and co-workers³⁸ once proposed that the role of GR was in essence the same as that of CNT in promoting the photocatalytic performance of TiO₂ in the degradation of volatile aromatic pollutants. To clarify the advantages of GR or CNT as a co-catalyst in photocatalysis is helpful for the rational design of efficient carbon material-promoted photocatalysts.⁴³

Herein, we report the preparation of CdS–GR composites by a new hydrothermal method using Cd(Ac)₂, Na₂S and GO as precursors. This method is environmentally benign because no organic sulfur-containing compounds are formed. Moreover, the precursor, i.e., GO, which contains a high concentration of oxygen-containing functional groups and is hydrophilic, can be well dispersed in aqueous solutions.³⁴ Stronger interactions between CdS and GR can be expected by using hydrophilic GO instead of hydrophobic GR as the precursor. The photocatalytic behaviors of the prepared CdS–GR composites, for the evolution of H₂ from an aqueous solution containing Na₂S and Na₂SO₃ and the degradation of methyl orange (MO) under visible-light irradiation, are both studied to gain insight into the effect of GR on the photocatalytic properties of CdS. The present paper also contributes by comparing systematically the catalytic behaviours of the CdS–GR and CdS–CNT composites.

2. Experimental

2.1. Preparation of catalysts

GO was synthesized from graphite powder by the method reported by Hummers and Offeman,⁴⁴ and the detailed synthetic procedures were described in our previous paper.³⁹ Multi-walled CNTs with outer diameters of 20–60 nm and inner diameters of 3–5 nm were synthesized by an established method via CH₄ decomposition over a Ni-MgO catalyst.⁴⁵ The as-synthesized CNTs were pre-treated with concentrated HNO₃ at 413 K under refluxing conditions to remove amorphous carbon and the remaining Ni catalyst. The functional groups, such as carboxylic and hydroxyl groups, were also generated and the hydrophilicity of CNTs was increased after the pre-treatment.⁴⁶ For the preparation of CdS nanoparticles, the aqueous solution of Na₂S was first added dropwise into an aqueous solution of Cd(Ac)₂ under vigorous stirring. After being stirred for 24 h, the obtained suspension was transferred into a Teflon-lined autoclave and was subjected to hydrothermal treatment at 453 K for 40 h. The solid CdS powders were then collected by centrifugation, washed repeatedly with deionized water, and dried in vacuum at 333 K.

For the preparation of CdS–GR nanocomposites, GO was first dissolved in H₂O by ultrasonic treatment for 1 h to give a yellow-brown solution. Then, Cd(Ac)₂ was dissolved in the aqueous solution containing GO. Subsequently, the aqueous solution of Na₂S was added dropwise to the solution containing GO and Cd(Ac)₂ under vigorous stirring, and the suspension was further stirred for 24 h. The mixture was transferred into a Teflon-lined autoclave and was subjected to hydrothermal treatment at 453 K for 40 h. It was expected that both the reduction of GO to GR and the loading of CdS could be achieved during the hydrothermal process.^36,38,39 The resultant solid was recovered by centrifugation, washed thoroughly with deionized water, and dried in vacuum at 353 K. The same procedure was also adopted for the preparation of CdS–CNT nanocomposites except for employing CNT to replace GO. For comparison, a CdS–TiO₂ composite was prepared by a similar procedure using TiO₂ (P25) instead of GR or CNT.

2.2. Catalytic reactions

The photocatalytic H₂ evolution was carried out in a Pyrex reaction cell connected to a closed gas circulation and evacuation system. The light source was a 200 W Xenon arc lamp (PLS-SXE300, Beijing Trusttech Co., Ltd.) with a UV cut-off filter (λ ≥ 420 nm). The focused intensity on the reaction cell was ∼90 mW cm⁻². Typically, the photocatalyst (100 mg) was suspended in an aqueous solution (100 mL) containing 0.1 M Na₂S and 0.05 M Na₂SO₃, which were used as sacrificial agents. The amount of H₂ evolved was determined with an online gas chromatograph (Techcomp; GC-7890II) equipped with a molecular 5 A column and TCD detector. N₂ was used as the carrier gas.

The photodegradation of MO was performed in a quartz tube under the irradiation of visible light. Typically, 0.050 g of the photocatalyst was suspended in 50 mL of 0.010 g L⁻¹ dye solution. The light source was a 200 W Xenon lamp with a UV cut-off filter (λ ≥ 420 nm). The suspension was first magnetically stirred in the dark in order to establish an adsorption and desorption equilibrium of dyes on the catalyst surface. Then, the reaction vessel was exposed to the visible-light irradiation. The reaction was performed at room temperature under continuous stirring. At specific time intervals, the reaction solution was analyzed by a UV-vis spectrophotometer. The degree of degradation was expressed by C/C₀, which is the ratio of the temporal MO concentration to the initial MO concentration after absorption equilibrium in the dark. The C/C₀ was proportional to the A/A₀, which is the normalized maximum absorbance occurring at 465 nm.

2.3. Characterisation of catalysts

Powder X-ray diffraction (XRD) patterns were recorded on a Panalytical X'pert Pro diffractometer using Cu-Kα radiation (40 kV, 30 mA). X-ray photoelectron spectroscopy (XPS) measurements were performed on a Quantum 2000 Scanning ESCA Microprobe (Physical Electronics) using Al-K_α radiation (1846.6 eV) as the X-ray source. Diffusion reflectance UV-visible (UV-vis) spectra were recorded on a Varian-Cary 5000 spectrometer equipped with a diffuse reflectance accessory. The spectra were collected with BaSO₄ as a reference. Transmission electron microscopy (TEM) measurements were performed on a JEM-2100 electron microscope operated at an acceleration voltage of 200 kV.

The total organic carbon (TOC) values of the degradation of MO were measured with a TOC-5000A instrument (Shimadzu, Japan). The terephthalic acid fluorescence probe was employed to detect hydroxyl (OH) radicals possibly generated during the degradation of MO.^47,48 It is known that the OH radical reacts with terephthalic acid (TA) forming TAOH, which fluoresces at ∼426 nm on the excitation of its own 312 nm absorption band.⁴⁷ The fluorescence spectra were recorded on a fluorescence spectrophotometer (Hitachi F-7000, Japan). The detailed experimental procedures were similar to those for the photocatalytic activity measurements except that the aqueous dye solution was replaced by a 5 × 10⁻³ M TA aqueous solution containing 0.01 M NaOH.

The photoelectrochemical measurements were carried out using an Ivium CompactStat (Holand), using a standard three-electrode cell with a working electrode, a platinum wire as the counter electrode, and an SCE electrode as the reference electrode. A 0.5 M solution of Na₂SO₄ was used as the electrolyte. The working electrode was prepared by cleaning an F-doped SnO₂-coated glass (FTO glass, 1 cm × 2 cm). The photocatalyst was dispersed in ethanol, and the suspension was added dropwise directly onto the FTO by microsyringe with a gentle stream of air to speed drying. The film was dried under 393 K for 1 h, and the typical surface density of the photocatalyst was 1 mg cm⁻².

3. Results and discussion

3.1. Photocatalytic behaviors of CdS–GR and CdS–CNT for H₂ evolution

Fig. 1 shows the amount of H₂ evolved from the aqueous solution containing a 0.1 M Na₂S and 0.05 M Na₂SO₃ over the CdS–GR composites, with different mass ratios of CdS : GR under visible-light irradiation. The amount of H₂ evolved over each catalyst increased proportionally to the irradiation time. It is noteworthy that GR alone is inactive under our reaction conditions. The incorporation of GR into CdS significantly accelerated the amount of H₂ evolved. All the CdS–GR composites examined here (mass ratios of CdS : GR = 1 : 0.005–1 : 0.1) exhibited higher activities for H₂ evolution than CdS alone. Fig. 1 also suggests that the mass ratio of CdS : GR plays a key role in determining the amount of H₂ evolution. The composite with a CdS : GR mass ratio of 1 : 0.01 showed the highest H₂ evolution activity. Both higher and lower CdS : GR ratios were disadvantageous for H₂ evolution.

Fig. 1 Amount of H₂ evolved from the aqueous solution containing 0.1 M Na₂S and 0.05 M Na₂SO₃ over the CdS–GR composites with different mass ratios of CdS : GR under visible-light irradiation. Reaction conditions: catalyst, 0.10 g; aqueous solution, 100 mL; light source, 200 W Xe lamp (λ ≥ 420 nm).

The photocatalytic properties of the CdS–CNT composites for H₂ evolution are shown in Fig. 2. The combination of CNT with CdS also enhanced the amount of H₂ evolved. The H₂ formation activity increased in the order of CdS < CdS–CNT (1 : 0.005) < CdS–CNT (1 : 0.01) < CdS–CNT (1 : 0.1) < CdS–CNT (1 : 0.03) < CdS–CNT (1 : 0.05). In other words, the H₂ evolution activity increased with an increasing fraction of CNT in the nanocomposites and arrived at a maximum as the CNT : CdS mass ratio reached 0.05 : 1. A further increase in the mass ratio of CNT : CdS to 0.1 : 1 decreased the activity of H₂ evolution.

Fig. 2 Amount of H₂ evolved from the aqueous solution containing 0.1 M Na₂S and 0.05 M Na₂SO₃ over the CdS–CNT composites with different mass ratios of CdS : CNT, under visible-light irradiation. Reaction conditions: catalyst, 0.10 g; aqueous solution, 100 mL; light source, 200 W Xe lamp (λ ≥ 420 nm).

The rates of H₂ evolution over the CdS–GR and the CdS–CNT composites with optimized mass ratios of CdS : GR and CdS : CNT are compared in Fig. 3. CdS alone provided an H₂ evolution rate of 14.5 μmol h⁻¹. Previous studies demonstrated that the combination of CdS with TiO₂ could also enhance the photocatalytic performance of CdS for H₂ evolution, and the CdS–TiO₂ composites with CdS : TiO₂ mass ratios of ≥ 1.5 : 1 exhibited better performances.¹⁷ Our result showed that the CdS–TiO₂ composite with a CdS : TiO₂ mass ratio of 2 : 1 provided an H₂ evolution rate of 34.4 μmol h⁻¹, which was ∼2.4 times that of CdS alone. Fig. 3 reveals that both the CdS–GR and the CdS–CNT composites displayed higher rates of H₂ evolution than CdS alone and the CdS–TiO₂ composite, indicating the higher efficiency of GR or CNT as the co-catalyst of CdS. The CdS–GR composite was more efficient than the CdS–CNT composite. The H₂ evolution rate over the CdS–GR composite reached 70 μmol h⁻¹, which was 4.8 times higher than CdS and 1.3 times higher than the CdS–CNT (52 μmol h⁻¹). GR alone was inactive for the photocatalysis of H₂ evolution under our reaction conditions. Moreover, we have examined the photocatalytic performance of a physical mixture of CdS and GR with a CdS : GR mass ratio of 1 : 0.01. This mixture showed almost the same rate of H₂ evolution as CdS alone, indicating that the physical mixing was ineffective. This implies that the interaction between GR and CdS plays a key role in enhancing the photocatalytic performance.

Fig. 3 Comparison of photocatalytic performances of CdS, GR, CdS–GR composite (CdS : GR = 1 : 0.01), physical mixture of CdS and GR (CdS : GR = 1 : 0.01), CdS–CNT composite (CdS : CNT = 1 : 0.05) and CdS–TiO₂ (CdS : TiO₂ = 2 : 1) composite for the evolution of H₂ from the aqueous solution containing 0.1 M Na₂S and 0.05 M Na₂SO₃. Reaction conditions: catalyst, 0.10 g; aqueous solution, 100 mL; light source, 200 W Xe lamp (λ ≥ 420 nm).

Instability was a problem for CdS-based photocatalysts because of the possible photocorrosion of CdS in aqueous medium.⁴⁹ Thus, we have examined the stability of the CdS–GR composite (CdS : GR = 1 : 0.01). After each run, the CdS–GR catalyst was evacuated for 30 min and was re-used in the next run. The result in Fig. 4 manifests that no deactivation occurs for 5 recycles.

Fig. 4 Changes of photocatalytic performances of CdS–GR composite (CdS : GR = 1 : 0.01) during repeated uses for the evolution of H₂ from the aqueous solution containing 0.1 M Na₂S and 0.05 M Na₂SO₃. Reaction conditions: catalyst, 0.10 g; aqueous solution, 100 mL; light source, 200 W Xe lamp (λ ≥ 420 nm).

3.2. Photocatalytic behaviors of CdS–GR and CdS–CNT for degradation of MO

Photocatalytic degradation is an effective means to eliminate organic dyes in aqueous solutions.^1,5 Here, we have studied the photocatalytic behavior of the CdS–GR and the CdS–CNT composites for the degradation of MO, which is one of the representative organic dyes,⁵ under visible-light irradiation.

It was reported that the incorporation of GR into TiO₂ could increase the adsorption of organic dyes onto the photocatalysts, contributing to the enhancement in degradation activities.^36,38Fig. 5 shows the remaining concentration fractions of MO and the corresponding sample pictures for the CdS–GR composites with different mass ratios of CdS : GR after the adsorption equilibrium in the dark. The adsorption of MO was enhanced by the presence of GR in the composites, and a larger amount of MO molecules were adsorbed onto the surface of the CdS–GR composite with a higher fraction of GR. The variation of the degree of degradation (C/C₀) with the time of irradiation under visible light for each catalyst is plotted in Fig. 6. The degrees of degradation for the CdS–GR composites were all significantly higher than that of CdS alone. The photodegradation performance increased with an increase in the mass ratio of GR : CdS from 0 to 0.01 : 1 and reached the maximum at 0.01 : 1. Further increases in the fraction of GR decreased the photodegradation activity.

Fig. 5 Sample pictures and bar plots showing the remaining MO in the aqueous solution after reaching the adsorption equilibrium in the dark for the CdS–GR composites with different mass ratios of CdS : RGO. (a) Initial MO solution, (b) CdS, (c) CdS–GR (CdS : GR = 1 : 0.005), (d) CdS–GR (CdS : GR = 1 : 0.01), (e) CdS–GR (CdS) : GR = 1 : 0.03), (f) CdS–GR (CdS : GR = 1 : 0.05), (g) CdS–GR (CdS : GR = 1 : 0.1).

Fig. 6 Photocatalytic degradation of MO for the CdS–GR composites with different mass ratios of CdS : GR under visible-light irradiation. Reaction conditions: catalyst, 0.050 g; aqueous dye solution, 50 mL; light source, 200 W Xe lamp (λ ≥ 420 nm).

Similar studies have been performed for the CdS–CNT composites with different mass ratios of CdS : CNT. As displayed in Fig. 7, the adsorption of MO in the dark was also enhanced in the presence of CNT, and the adsorption amount increased with the fraction of CNT in the composites. However, the comparison between Fig. 5 and Fig. 7 shows that the efficiency of the MO adsorption on the CdS–CNT is lower than that on the CdS–GR. Fig. 8 demonstrates that, similar to the photocatalytic H₂ evolution, there also exists an optimum mass ratio of CNT : CdS (0.03 : 1) for the photodegradation of MO; both lower and higher CNT : CdS mass ratios are unfavorable for the photodegradation reaction.

Fig. 7 Sample pictures and bar plots showing the remaining MO in the aqueous solution, after reaching the adsorption equilibrium in the dark for the CdS–CNT composites with different mass ratios of CdS : CNT. (a) Initial MO solution, (b) CdS, (c) CdS–CNT (CdS : CNT = 1 : 0.005), (d) CdS–CNT (CdS : CNT = 1 : 0.01), (e) CdS–CNT (CdS : CNT = 1 : 0.03), (f) CdS–CNT (CdS : CNT = 1 : 0.05), (g) CdS–CNT (CdS : CNT = 1 : 0.1).

Fig. 8 Photocatalytic degradation of MO for the CdS–CNT composites with different mass ratios of CdS : CNT under visible-light irradiation. Reaction conditions: catalyst, 0.050 g; aqueous dye solution, 50 mL; light source, 200 W Xe lamp (λ ≥ 420 nm).

The comparison of the photodegradation activities between the two series of composites (Fig. 6 and 8) reveals that GR is more efficient for enhancing the catalytic performance of CdS. After 60 min irradiation, the degree of MO degradation for the optimized CdS–GR (1 : 0.01) reached 95%, while that for the optimized CdS–CNT (1 : 0.03) was 88%. To make a better comparison, we have analyzed the kinetics by assuming a pseudo first-order reaction for the degradation of organic dyes.⁵⁰ The data in Fig. 6 and 8 were fitted by using the following equation: ln(C₀/C) = kt, where k was the apparent rate constant. From the lines plotted in Fig. 9 and using the data points at the initial reaction stage, we calculated the rate constant for each photocatalyst. Table 1 demonstrates that the rate constant of the CdS–GR composite with the optimum composition (GR : CdS = 0.01 : 1) is 7.9 times higher than that of CdS alone and is 1.8 times higher than that of the CdS–CNT composite with the optimum composition (1 : 0.03). These results allow us to conclude that, for both the photocatalytic evolution of H₂ and the photodegradation of MO, the CdS–GR and CdS–CNT composites exhibit significantly better performances than CdS alone, and the presence of GR is more efficient for increasing the photocatalytic activity of CdS.

Fig. 9 Pseudo first-order kinetic treatment for the photodegradation of MO with the CdS–GR (A) and CdS–CNT (B) composites under visible-light irradiation. Reaction conditions: catalyst, 0.050 g; aqueous dye solution, 50 mL; light source, 200 W Xe lamp (λ ≥ 420 nm).

Table 1 Apparent rate constants for the photodegradation of MO for the CdS–GR and CdS–CNT nanocomposites

Mass ratio of GR:CdS or CNT:CdS	Apparent rate constant (10^−2 min^−1)
	CdS–GR	CdS–CNT
0 (CdS alone)	0.75	0.75
0.005	2.4	1.0
0.01	5.9	1.6
0.03	4.1	3.3
0.05	2.8	2.7
0.10	2.5	1.5

---

3.3. Possible reaction mechanism for photocatalytic degradation of MO

We have confirmed that, in the absence of our CdS-based photocatalyst (blank reaction), MO does not undergo self-sensitized degradation either with or without the addition of Na₂S (Fig. 10). Thus, it is clear that the photocatalyst plays a crucial role in the degradation of MO.

Fig. 10 Degradation of MO under different conditions. Standard reaction conditions: catalyst CdS–GR (1 : 0.01), 0.050 g; aqueous dye solution, 50 mL; light source, 200 W Xe lamp (λ ≥ 420 nm). Blank: without adding catalyst.

Concerning the nature of the active species for the oxidative degradation of MO, the photogenerated hole and the hydroxyl (OH) radical may be possible candidates. Further experiments have been performed to gain insight into the probable active species. We have examined the possibility of the photogenerated hole as the active species by adding sodium oxalate (Na₂C₂O₄), which typically functions as a hole scavenger, into the reaction system. The result showed that the activity of MO degradation for CdS–GR (1 : 0.01) after the addition of Na₂C₂O₄ decreased only slightly (Fig. 10). On the other hand, the addition of tert-butanol (t-BuOH), a scavenger of OH radicals, caused a remarkable decrease in the performance of the photocatalytic degradation of MO (Fig. 10). This suggests that the OH radical may play a more important role than the photogenerated hole toward the degradation of MO.

We have detected further OH radicals by using the terephthalic acid fluorescence probe technique.^47,48 In brief, if OH radicals exist in the reaction system, they can react with terephthalic acid (TA) forming TAOH, which fluoresces at ∼426 nm due to the excitation of its own 312 nm absorption band.⁴⁷ We observed the fluorescence at ∼426 nm during our experiment for the suspension containing photocatalysts and an aqueous solution of TA, confirming the generation of OH radicals. Fig. 11 shows that the CdS–GR (1 : 0.01) composite displays a much higher intensity of the fluorescence than the CdS–CNT composite (1 : 0.03) and CdS alone. This suggests that a higher concentration of OH radicals has been generated in the case of the CdS–GR composite, which exhibits a higher activity in the photodegradation of MO (Table 1). As the photogenerated hole cannot oxidize H₂O or HO⁻ directly into OH radicals, we speculate that the OH radical may be generated from the activation of O₂ by the photogenerated electrons via O⁻₂ [eqn (1)–(3)].^47,51

e⁻_CB + O₂ → O⁻₂ (1)

2H⁺ + e⁻_CB + O⁻₂ → H₂O₂ (2)

e^¬ + H₂O₂ → ˙HO + HO⁻ (3)

We have also measured the total organic carbon (TOC) values for the MO solutions before and after photodegradation. The results summarized in Table 2 show that the TOC values all decrease after the photodegradation, suggesting that part of MO can be mineralized to CO₂ and H₂O. Table 2 further demonstrates that the CdS–GR composite is more active than the CdS–CNT composite and both composites are more effective than CdS alone.

Table 2 TOC values for the MO solutions before and after catalytic degradation, in the presence of different catalysts

Catalyst	TOC values (mg L^−1)
	Initial	After photodegradation
CdS	4.8	4.3
CdS–GR (1 : 0.01)	4.8	2.1
CdS–CNT (1 : 0.03)	4.8	3.4

---

Fig. 11 The changes to the fluorescence intensity of TAOH at 426 nm, at different irradiation times (λ_ex = 312 nm).

3.4. Characterizations of CdS–GR and CdS–CNT nanocomposites

The crystalline structures of the nanocomposites were studied by XRD. The XRD pattern of GO showed a strong and sharp diffraction peak at 2θ of 10.6° (Fig. 12), corresponding to a d-spacing of ∼0.84 nm. This was consistent with the lamellar structure of GO. For the GR obtained after the hydrothermal process without CdS, this strong diffraction peak disappeared and a very broad and very weak peak at 2θ of ∼25°, corresponding to a d-spacing of ∼0.36 nm, was observed. This suggests the reduction of GO to GR sheets during the hydrothermal process.⁵² The CNT displayed an intense peak at ∼26°, corresponding to a d-spacing of ∼0.34 nm for the sp²-carbon layer. CdS alone showed three main diffraction peaks at 2θ of 26.5°, 43.9° and 52.0°. These peaks were assigned to a cubic phase of CdS (zinc blend structure), while the other weak peaks at 2θ of 24.9°, 28.2°, 36.7° and 47.9° could be ascribed to a hexagonal phase of CdS (wurzite structure).²² Thus, the present CdS crystallites were composed of both cubic and hexagonal phases. The CdS–GR and the CdS–CNT composites displayed XRD patterns similar to that of CdS alone. The diffraction peak of CNT at ∼26° was probably overlapped with a main peak of CdS at almost the same position.

Fig. 12 XRD patterns of CdS, CNT, GO, GR, CdS–GR with different mass ratios of CdS : GR, and CdS–CNT composites with different mass ratios of CdS : CNT.

The chemical states of elements on the surfaces of the nanocomposites were investigated by XPS. Fig. 13 shows the C 1s, Cd 3d and S 2p XPS spectra for the CdS–GR and the CdS–CNT composites together with carbon materials and CdS alone. As shown in Fig. 13A, two main peaks were observed for GO; the peak at 284.6 eV was assignable to the sp²-carbon species, while the peak at higher binding energies (286–288 eV) was ascribed to oxygenated carbon species such as hydroxyl and carboxyl groups. For the GR and the CdS–GR (CdS : GR = 1 : 0.01), the peak at a higher binding energy position became significantly weakened or almost disappeared, indicating the reduction of GO to GR sheets in the CdS–GR nanocomposites. The C 1s spectra for CNT and the CdS–CNT (CdS : CNT = 1 : 0.05) are also displayed in Fig. 13A. The peaks of oxygenated carbon species for these samples were weak. The binding energies of Cd 3d_5/2 were found at 406.0–406.2 eV for CdS, CdS–GR and CdS–CNT composites (Fig. 13B), suggesting that cadmium was in the Cd²⁺ state in all these samples.²⁸ The S 2p_3/2 peaks were observed at 161.9–162.0 eV for these samples (Fig. 13C), in agreement with the expectation that sulfur existed as the sulfide species (S²⁻).²⁸ The C 1s, Cd 3d and S 2p XPS spectra for the CdS–GR composite (CdS : GR = 1 : 0.01) after the photocatalytic evolution of H₂ are also shown in Fig. 13. No significant changes could be observed in these spectra, confirming that no photocorrosion of CdS occurred and the catalyst was stable under our reactions conditions.

Fig. 13 XPS spectra. (A) C 1s: (a) GO, (b) RGO, (c) CdS–RGO (CdS : RGO = 1 : 0.01), (d) CdS–RGO (CdS : RGO = 1 : 0.01) after reaction, (e) CNT, (f) CdS–CNT (CdS : CNT = 1 : 0.05). (B) Cd 3d: (a) CdS, (b) CdS–RGO (CdS : RGO = 1 : 0.01), (c) CdS–RGO (CdS : RGO = 1 : 0.01) after reaction, (d) CdS–CNT (CdS : CNT = 1 : 0.05). (C) S 2p: (a) CdS, (b) CdS–RGO (CdS : RGO = 1 : 0.01), (c) CdS–RGO (CdS : RGO = 1 : 0.01) after reaction, (d) CdS–CNT (CdS : CNT = 1 : 0.05).

Fig. 14 shows the TEM images of the CdS–GR (CdS : GR = 1 : 0.01) and CdS–CNT (CdS : CNT = 1 : 0.05) composites together with GO, GR and CdS alone. The sheet-like structure was clearly observed for GO and GR, while CdS prepared in this work had a nanoparticle morphology. In the CdS–CNT and the CdS–GR composites, CdS nanoparticles were dispersed on the CNT surfaces or the GR sheets. The high-resolution TEM image for the CdS–GR composite (Fig. 14f) reveals that a crystallized CdS nanoparticle with a lattice spacing of 0.34 nm, corresponding to the (111) plane of cubic CdS, is directly connected with the GR sheets. The CdS particle size distributions are also shown in Fig. 14. The average CdS particle size for CdS alone was 33 nm. The average CdS particle sizes became slightly larger in the CdS–CNT (34 nm) and the CdS–GR (36 nm) composites. The phenomenon that the presence of GO or GR promoted the crystallization of CdS during the synthetic procedure was also observed in a previous study.⁴⁰ This may also suggest that there exists interactions between CdS and GR in the composites.

Fig. 14 TEM micrographs and CdS particle size distributions. TEM micrographs: (a) GO, (b) GR, (c) CdS, (d) CdS–CNT (CdS : CNT = 1 : 0.05), (e) CdS–GR (CdS : GR = 1 : 0.01), (f) CdS–GR (CdS : GR = 1 : 0.01) high resolution. CdS particle size distributions: (g) CdS, (h) CdS–CNT (CdS : CNT = 1 : 0.05), (i) CdS–GR (CdS : GR = 1 : 0.01).

The CdS–GR and the CdS–CNT composites have further been characterized by diffuse reflectance UV-vis spectroscopy. The results for the two series of composites are shown in Fig. 15 and 16. The background in the visible light region (≥ 550 nm) was increased in the composites containing GR or CNT. This corresponded to the change in the sample color from orange to olive due to the combination of black GR or CNT with CdS. From the plots of the modified Kubelka-Munk function, i.e., [F(R_∞)hυ]^1/2, versus the energy of exciting light (hυ), the band-gap energy for a semiconductor can be evaluated.⁵³ The plots of [F(R_∞)hυ]^1/2 against hυ for the CdS–GR and the CdS–CNT composites with different compositions are shown in Fig. 15B and 16B. The plot for CdS alone afforded a band-gap energy of 2.27 eV. The increase in the mass ratio of GR : CdS in the CdS–GR composites decreased the band-gap energy; the band-gap energies for the CdS–GR composites with GR : CdS mass ratios of 0.005, 0.01, 0.03 and 0.05 declined to 2.22, 2.21, 2.18 and 2.15 eV, respectively. This band-gap narrowing was also observed for TiO₂ in the TiO₂–GR composites, which was suggested to correspond to the formation of a Ti–O–C bond between TiO₂ and GR.^36,38,39 We propose that the band-gap narrowing observed for the CdS–GR composite indicates the strong interactions between CdS and GR.

Fig. 15 (A) Diffuse reflectance UV–vis spectra for CdS and CdS–GR composites with different mass ratios of CdS : GR. (B) Corresponding plot of transformed Kubelka-Munk function versus the energy of the light.

Fig. 16 (A) Diffuse reflectance UV–vis spectra for CdS and CdS–CNT composites with different mass ratios of CdS : CNT. (B) Corresponding plot of transformed Kubelka-Munk function versus the energy of the light.

Similarly, the increase in the mass ratio of CNT : CdS in the CdS–CNT composites also decreased the band-gap energy as shown in Fig. 16B; the band-gap energies for the CdS–CNT composites with CNT : CdS mass ratios of 0.005, 0.01, 0.03 and 0.05 were 2.25, 2.23, 2.21 and 2.18 eV, respectively. The comparison between the two series of composites demonstrates that, for the sample with the same mass ratio of carbon : CdS, the band-gap narrowing is relatively larger for the CdS–GR composite. Our previous studies indicated that the degree of the band-gap narrowing may reflect the degree of the interactions or the contact interfaces between TiO₂ and carbon materials.³⁹ Thus, a larger band-gap narrowing observed for the CdS–GR composite may suggest a stronger interaction or a larger contact interface between CdS and RGO than that between CdS and CNT.

The transient photocurrent response has been demonstrated to be a useful technique for investigating the efficiency of the separation of photogenerated electron-hole pairs.^26,54–56Fig. 17 shows the curves of transient photocurrent density versus time with several on-off cycles of intermittent visible-light irradiation for CdS, the CdS–GR (CdS : GR = 1 : 0.01) and CdS–CNT (CdS : CNT = 1 : 0.05) composites. As expected, in each case, the photocurrent density decreased rapidly to zero as soon as the light irradiation was turned off, and it increased again to almost the same value as the light was turned on. Fig. 17 revealed that the CdS–GR composite exhibited a higher photocurrent than the CdS–CNT composite, and the photocurrents for both composites were significantly higher than that for CdS alone. These results demonstrate that the efficiency of the separation of photogenerated electron-hole pairs increases in the order of CdS < CdS–CNT < CdS–GR.

Fig. 17 Transient photocurrent responses for CdS, CdS–GR (1 : 0.01) and CdS–CNT (1 : 0.05).

3.5. Structure–performance correlation

We have demonstrated that the incorporation of both GR and CNT into CdS to form nanocomposites can enhance the photocatalytic performances of CdS for the evolution of H₂ and the degradation of MO. There exists optimum mass ratios of GR : CdS and CNT : CdS for both types of photocatalytic reactions. For the CdS–GR composites, the maximum photocatalytic performances have been achieved at a relatively lower mass ratio of GR : CdS (0.01 : 1), whereas the optimum CNT : CdS mass ratios are 0.05 : 1 and 0.03 : 1 for the H₂ evolution and the MO degradation, respectively. The catalytic performances of the optimized CdS–GR composite have been found to be better than those of the optimized CdS–CNT composite for both reactions. The characterizations for the two series of nanocomposites suggest that CdS nanoparticles with mean sizes of ∼35 nm are dispersed on the GR sheets or on the CNT outer surfaces. As suggested by the larger band-gap narrowing for the CdS–GR composite (Fig. 15), there exists stronger interactions or larger contact interfaces between CdS and GR than those between CdS and CNT.

It is expected that the stronger interactions or the larger contact interfaces between CdS and GR may result in a more efficient electron transfer from the CdS nanoparticles to GR sheets under visible-light irradiation and a more efficient separation of photogenerated electrons and holes. The increase in the efficiency of the separation of photogenerated electron-hole pairs has been confirmed by the transient photocurrent response studies (Fig. 17). We have also demonstrated that the physical mixture of CdS and GR did not show any enhancing effect of GR (Fig. 3). Thus, we propose that the stronger interactions or the larger contact interfaces is the main reason for the higher photocatalytic activities of the CdS–GR composite for both the H₂ evolution and the organic dye degradation. The unique two dimensional sheet-like structure, the larger surface area, and the higher concentration of oxygen-containing functional groups of GO may favor the direct contact between CdS and the GR sheet during the preparation stage.

It is noteworthy that the degree of band-gap narrowing or the contact interfaces between CdS and CNT could be increased by increasing the mass ratio of CNT : CdS in the CdS–CNT nanocomposites. We speculate that this is why a higher fraction of carbon materials is required for obtaining the maximum photocatalytic performances over the CdS–CNT composites. Here, the next question may arise. Why are the maximum photocatalytic performances of the CdS–CNT lower than those of the CdS–GR if similar contact interfaces can be obtained? It is expected that there are two main contradictory roles of carbon materials in influencing the photocatalytic property of a semiconductor catalyst. The positive role is to accelerate the electron-hole separation as described above. On the other hand, the black carbon materials in the nanocomposites may exert a negative “shielding effect”.^26,40 In other words, CNT or GR in the composites may shield the light absorption of the semiconductor photocatalyst. This negative role has been observed in several studies for the TiO₂–CNT and TiO₂–GR composites, particularly at higher fractions of carbon materials.^26,36–40 We believe that this is the main reason for the decrease in the photocatalytic performances at higher GR : CdS or CNT : CdS mass ratios. Therefore, the better photocatalytic performance of the optimized CdS–GR as compared to the optimized CdS–CNT can be interpreted by the fact that the CdS–GR composites can provide stronger interactions or larger contact interfaces between CdS and carbon materials at a lower ratio of carbon:CdS.

4. Conclusions

We succeeded in preparing the nanocomposites of CdS–GR and CdS–CNT by an environmentally benign hydrothermal method. Hydrophilic GO was used as the precursor with an aim to obtaining a composite with strong interactions between CdS and GR. The characterizations using different techniques uncovered that GO was reduced to GR sheets in the nanocomposites. Characterization results also suggested that CdS nanoparticles with a mean diameter of ∼35 nm were dispersed on the GR sheets or on the CNT outer surfaces in the nanocomposites. As compared to CdS alone, the incorporation of GR or CNT led to band-gap narrowing, indicating the strong interactions between CdS and GR or CdS and CNT in the composites. There exists stronger interactions or larger contact interfaces between CdS and GR than those between CdS and CNT in the composite with the same mass ratio of carbon : CdS. We studied the photocatalytic properties of the CdS–GR and the CdS–CNT nanocomposites for the evolution of H₂ from the aqueous solution containing Na₂S and Na₂SO₃ and the degradation of MO under visible-light irradiation. The CdS–GR and the CdS–CNT composites both exhibited significantly better performances for the two photocatalytic reactions than CdS alone. The photocatalytic performances of the two series of composites depended on the mass ratios of CdS : GR or CdS : CNT. The rate of H₂ evolution for the CdS–GR nanocomposite with an optimum mass ratio of CdS : GR (1 : 0.01) was 4.8 times higher than that over CdS alone, whereas the physical mixture of CdS and GR showed a similar photocatalytic activity with CdS. The rate of H₂ evolution for this CdS–GR composite was higher than that of the CdS–CNT composite with an optimum mass ratio of CdS : CNT (1 : 0.05). The CdS–GR composite was stable and could be used repeatedly for H₂ evolution. The rate of photodegradation of MO was also the highest for the CdS–GR composite with a mass ratio of CdS : GR of 1 : 0.01. The apparent rate constant of the pseudo first-order degradation of MO for this catalyst was 7.9 times higher than that of CdS alone, and 1.8 times higher than that of the CdS–CNT composite with an optimum mass ratio of CdS : CNT (1 : 0.03). The stronger interactions or the larger contact interfaces favor the transfer of photogenerated electrons from CdS to GR, leading to a higher efficiency in the separation of photogenerated electron-hole pairs and a higher photocatalytic performance of the CdS–GR composite.

Acknowledgements

This work was supported by the National Basic Research Program of China (No. 2010CB732303), the NSF of China (Nos. 21173172, 21173174 and 21161130522), the Program for Changjiang Scholars and Innovative Research Team in University (No. IRT1036) and the Research Fund for the Doctoral Program of Higher Education (No. 20090121110007).

Notes and references

1. M. R. Hoffmann, S. T. Martin, W. Choi and D. W. Bahnemann, Chem. Rev., 1995, 95, 69.
2. A. L. Linsebiger, G. Lu and J. T. Yates Jr., Chem. Rev., 1995, 95, 735.
3. K. Maeda and K. Domen, J. Phys. Chem. C, 2007, 111, 76.
4. F. E. Osterloh, Chem. Mater., 2008, 20, 35.
5. K. Rajeshwar, M. E. Osugi, W. Chanmanee, C. R. Chenthamarakshan, M. V. B. Zanoni, P. Kajitvichyanukul and R. Krishnan-Ayer, J. Photochem. Photobiol., C, 2008, 9, 171.
6. A. Kudo and Y. Miseki, Chem. Soc. Rev., 2009, 38, 253.
7. R. M. N. Yerga, M. C. Á. Galván, F. Del Valle, J. A. V. De la Mano and J. L. G. Fierro, ChemSusChem, 2009, 2, 471.
8. Y. Inoue, Energy Environ. Sci., 2009, 2, 364.
9. M. D. Hernández-Alonso, F. Fresno, S. Suárez and J. M. Coronado, Energy Environ. Sci., 2009, 2, 1231.
10. D. Leung, X. Fu, C. Wang, M. Ni, M. Leung, X. Wang and X. Fu, ChemSusChem, 2010, 3, 681.
11. R. Abe, J. Photochem. Photobiol., C, 2010, 11, 179.
12. X. Chen, S. Shen, J. Guo and S. S. Mao, Chem. Rev., 2010, 110, 6503.
13. Q. Xiang, J. Yu and M. Jaroniec, Chem. Soc. Rev., 2012, 41, 782.
14. Y. J. Li and W. Chen, Catal. Sci. Technol., 2011, 1, 802.
15. M. Ashokkumar, Int. J. Hydrogen Energy, 1998, 23, 427.
16. P. V. Kamat, J. Phys. Chem. C, 2007, 111, 2834.
17. H. Park, W. Choi and M. R. Hoffmann, J. Mater. Chem., 2008, 18, 2379.
18. X. W. Wang, G. Liu, Z. G. Chen, F. Li, L. Wang, G. Q. Lu and H. M. Cheng, Chem. Commun., 2009, 3452.
19. X. Zong, H. Yan, G. Wu, G. Ma, F. Wen, L. Wang and C. Li, J. Am. Chem. Soc., 2008, 130, 7176.
20. H. Yan, J. Yang, G. Ma, G. Wu, X. Zong, Z. Lei, J. Shi and C. Li, J. Catal., 2009, 266, 165.
21. Y. B. Wang, J. C. Wu, J. W. Zheng and R. Xu, Catal. Sci. Technol., 2011, 1, 940.
22. H. N. Kim, T. W. Kim, I. Y. Kim and S. J. Hwang, Adv. Funct. Mater., 2011, 21, 3111.
23. K. Woan, G. Pyrgiotakis and W. Sigmund, Adv. Mater., 2009, 21, 2233.
24. C. G. Silva and J. L. Faria, ChemSusChem, 2010, 3, 609.
25. R. Leary and A. Westwood, Carbon, 2011, 49, 741.
26. J. Yu, T. Ma and S. Liu, Phys. Chem. Chem. Phys., 2011, 13, 3491.
27. I. Robel, B. A. Bunker and P. V. Kamat, Adv. Mater., 2005, 17, 2458.
28. Y. K. Kim and H. Park, Energy Environ. Sci., 2011, 4, 685.
29. T. Peng, P. Zeng, D. Ke, X. Liu and X. Zhang, Energy Fuels, 2011, 25, 2203.
30. A. K. Geim and K. S. Novoselov, Nat. Mater., 2007, 6, 183.
31. A. K. Geim, Science, 2009, 324, 1530.
32. V. N. R. Rao, A. K. Sood, K. S. Subrahmanyam and A. Govindaraj, Angew. Chem., Int. Ed., 2009, 48, 7752.
33. M. J. Allem, V. C. Tung and R. B. Kaner, Chem. Rev., 2010, 110, 132.
34. K. S. Novoselov, Angew. Chem., Int. Ed., 2011, 50, 6986.
35. X. J. Liu, L. K. Pan, T. Lv, T. Lu, G. Zhu, Z. Sun and C. Q. Sun, Catal. Sci. Technol., 2011, 1, 1189.
36. H. Zhang, X. Lv, Y. Li, Y. Wang and J. Li, ACS Nano, 2010, 4, 380.
37. X. Y. Zhang, H. P. Li, X. L. Cui and Y. Lin, J. Mater. Chem., 2010, 20, 2801.
38. Y. Zhang, Z. R. Tang, X. Fu and Y. J. Xu, ACS Nano, 2010, 4, 7303.
39. W. Fan, Q. Lai, Q. Zhang and Y. Wang, J. Phys. Chem. C, 2011, 115, 10694.
40. Q. Li, B. Guo, J. Yu, J. Ran, B. Zhang, H. Yan and J. R. Gong, J. Am. Chem. Soc., 2011, 133, 10878.
41. L. Jia, D. H. Wang, Y. X. Huang, A. W. Xu and H. Q. Yu, J. Phys. Chem. C, 2011, 115, 11466.
42. A. Cao, Z. Liu, S. Chu, M. Wu, Z. Ye, Z. Cai, Y. Chang, S. Wang, Q. Gong and Y. Liu, Adv. Mater., 2010, 22, 103.
43. Y. Zhang, Z. R. Tang, X. Fu and Y. J. Xu, ACS Nano, 2011, 5, 7426.
44. W. S. Hummers and R. J. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
45. P. Chen, H. B. Zhang, G. D. Lin, Q. Hong and K. R. Tsai, Carbon, 1997, 35, 1495.
46. J. Kang, S. Zhang, Q. Zhang and Y. Wang, Angew. Chem., Int. Ed., 2009, 48, 2565.
47. T. Hirakawa and Y. Nosaka, Langmuir, 2002, 18, 3247.
48. Y. Fu and X. Wang, Ind. Eng. Chem. Res., 2011, 50, 7210.
49. H. Zhang and Y. Zhu, J. Phys. Chem. C, 2010, 114, 5822.
50. Y. Liang, H. Wang, H. S. Casalongue, Z. Chen and H. J. Dai, Nano Res., 2010, 3, 701.
51. X. Zhang, X. Quan, S. Chen and H. Yu, Appl. Catal., B, 2011, 105, 237.
52. C. Nethravath and M. Rajamathi, Carbon, 2008, 46, 1994.
53. S. Sakthivel and H. Kisch, Angew. Chem., Int. Ed., 2003, 42, 4908.
54. N. J. Bell, Y. H. Ng, A. Du, H. Coster, S. C. Smith and R. Amal, J. Phys. Chem. C, 2011, 115, 6004.
55. Q. Xiang, J. Yu and M. Jaroniec, Nanoscale, 2011, 3, 3670.
56. Q. Xiang, J. Yu and M. Jaroniec, J. Phys. Chem. C, 2011, 115, 7355.

---