Advances in materials engineering of CdS coupled with dual cocatalysts of graphene and MoS₂ for photocatalytic hydrogen evolution

The effect of CdS morphology

Considering the role of CdS as photoactive component to produce electron-hole pairs, the optimization of CdS component is mainly aimed at improving its capability for more efficiently generating charge carriers to enhance the photoactivity of the composites. It has been demonstrated that tuning the morphology of CdS is able to boost the performance of CdS-based photocatalysts for H₂ production from water [58], [60], [61], [62]. For example, Liu et al. have synthesized MoS₂-graphene/CdS composites through a hydrothermal process using as-prepared CdS nanoparticles (NPs), MoS₂ microspheres and graphene oxide (GO) as the precursors [58]. As shown in Fig. 5a, in the ternary composite obtained at 180°C for 12 h, CdS exhibits a mixed nanoparticles (NPs)/nanorods (NRs) morphology; accordingly, the composite is denoted as M-G/CdS(NP-NR). It has been demonstrated that the interactions between the functional groups on GO and the surface of CdS NPs during the preparation might facilitate the growth of CdS NRs in the c-axis [001] direction. As for the sample prepared at 160°C for 12 h, only CdS NPs can be observed and such sample is denoted as denoted as M-G/CdS(NP). Notably, M-G/CdS(NR) composites cannot be prepared by the similar hydrothermal treatment when using CdS NPs as the precursor. Therefore, for comparison, M-G/CdS(NR) sample has been prepared using CdS NRs as the starting materials.

Fig. 5:

(a) TEM image of M-G/CdS(NP-NR); (b) photocatalytic H₂ evolution over different photocatalysts under visible light irradiation (λ>400 nm); (c) transient photocurrent responses of the M-G/CdS(NP-NR), M-G/CdS(NR), M-G/CdS(NP), M/CdS, G/CdS, GO, MoS₂ and CdS electrodes in 0.1 M Na₂SO₄ electrolyte; (d) schematic illustration for the charge transfer in the M-G/CdS(NP-NR) system. Reproduced from Ref. [58] with permission of Royal Society of Chemistry.

The photoactivity of these samples as well as CdS, MoS₂-CdS (M/CdS), graphene-CdS (G/CdS) as comparison counterparts has been evaluated by H₂ evolution reaction (HER) under visible light irradiation (λ>400 nm) with lactic acid as a sacrificial agent. It can be seen from Fig. 5b that bare CdS shows a relatively low photocatalytic activity with a H₂ evolution rate of 0.18 mmol h⁻¹ due to its rapid recombination of electrons and holes. The rates of H₂ evolution for the M/CdS and G/CdS are increased to 1.34 mmol h⁻¹ and 0.27 mmol h⁻¹, respectively. These results indicate that MoS₂ and graphene can act as co-catalyst to improve the HER performances of CdS. In the presence of both MoS₂ and graphene with their weight ratio of 19:1, the H₂ evolution rate of M-G/CdS(NP) is 1.87 mmol h⁻¹, which is higher than both of M/CdS and G/CdS. Notably, M-G/CdS(NP-NR) exhibits the highest H₂ evolution rate of 2.32 mmol h⁻¹. The quantum yield (QY) for H₂ evolution over M-G/CdS(NP-NR) was 65.8% at 420 nm, which is comparable to those reported over Pt/CdS [63], [64]. The enhanced photocatalytic performance of M-G/CdS(NP-NR) has been ascribed to the synergistic effects of three aspects: (i) effective electron-hole separation and transfer of CdS with mixed NPs/NRs morphology [as manifested by the highest photocurrent density of M-G/CdS(NP-NR) among the samples in Fig. 5c], (ii) excellent electron conductivity and large surface area of graphene, and (iii) the increased surface area and HER reactive sites [58]. Based on the above results, a tentative mechanism for HER over M-G/CdS has been proposed, as illustrated in Fig. 5d. Under visible light illumination, the electrons are excited to the conduction band (CB) of CdS NPs and NRs, creating holes in the valence band (VB) simultaneously. The CB electrons of CdS can be injected into the graphene sheets because the graphene/graphene˙⁻ redox potential is lower than the CB of CdS [65]. The MoS₂ nanosheets can accept electrons (either directly from CdS or through the graphene electron transfer mediator) and act as reactive sites for H₂ evolution. Such two electron transfer ways can effectively suppress the charge recombination and prolong the lifetime of the charge carriers, which in turn results in enhanced photocatalytic activity for H₂ evolution. The holes in the system are trapped by the sacrificial agent of lactic acid. In addition, the M-G/CdS(NP-NR) composite photocatalyst exhibits no obvious deactivation during the HER process.

The effect of MoS₂ morphology and crystallinity

It is known that the MoS₂ as an efficient cocatalyst for HER can accept the photogenerated electrons from semiconductor CdS to reduce H⁺ for H₂ production with low overpotential. There have been some advances achieved on how to improve the efficiency of MoS₂ in enhancing the photocatalytic HER performance of semiconductors [47], [48]. It has demonstrated that the cocatalytic activities of MoS₂ are derived from the uncoordinated sulfur edge sites while its basal planes remain catalytically inert [66], [67], [68]. Therefore, the efficiency of MoS₂ for HER can be improved by optimizing the structure and morphology of MoS₂ to increase its catalytic reactive sites [69], [70], [71].

For instance, our group has prepared two series of graphene (GR)-CdS-MoS₂ composites through two-step wet-chemistry strategies, which are denoted as GR-CdS-MoS₂ (photodeposition, PD) and GR-CdS-MoS₂ (hydrothermal, HT), respectively, as illustrated in Fig. 3 [48]. Their photoactivities for HER have been investigated under visible light irradiation (λ>420 nm). The performance of bare CdS has also been evaluated for comparison. It can be seen from Fig. 6a that the rate of H₂ evolution over CdS is relatively low (19 μmol h⁻¹), which should be ascribed to the rapid recombination of photogenerated electron-hole pairs and the lack of reactive sites for proton reduction. After coupling with MoS₂ or both of MoS₂ and graphene (GR), its photoactivities are much enhanced. With the photodeposition of 1%MoS₂, the composite of CdS-1%MoS₂ displays the H₂ evolution activity of 305 μmol h⁻¹, which is even higher than that of CdS-1%Pt. The modification of CdS with two cocatalysts of GR and MoS₂ through photodeposition method leads to the further improved photoactivity toward H₂ evolution. As shown in Fig. 6a, the (1%GR-CdS)-1%MoS₂ (PD) with optimal ratios exhibits the highest H₂ production rate of 513 μmol h⁻¹, corresponding to an apparent quantum efficiency (AQE) of 26.8% at 420 nm. However, with the same optimal ratios, the (1%GR-CdS)-1%MoS₂ (HT) shows much lower photoactivity than (1%GR-CdS)-1%MoS₂ (PD). These results imply that the photodeposition of MoS₂ is more beneficial than the hydrothermal loading of MoS₂ onto GR-CdS for improving its photocatalytic performance for H₂ production from water. The characterization results indicate that tiny MoS₂ with homogeneous dispersion is obtained in the GR-CdS-MoS₂ (PD) (Fig. 6b), while in the GR-CdS-MoS₂ (HT), MoS₂ displays typical layer structure with the thickness of 4–8 nm (Fig. 6c). Obviously, the homogeneously photodeposited MoS₂ with a tiny size can provide more catalytic reactive sites than the hydrothermal synthesized MoS₂ with stacked layer structure, which is able to facilitate the separation and transfer of photoexcited electron-hole pairs generated from CdS more efficiently, as reflected by the lowered overpotential (Fig. 6d), enhanced photocurrent density (Fig. 6e), and decreased interfacial impedance (Fig. 6f) of (1%GR-CdS)-1%MoS₂ (PD) as compared to other samples. In addition, (1%GR-CdS)-1%MoS₂ (PD) exhibits enhanced anti-photocorrosion performance and thus improved photostability as compared to (1%GR-CdS)-1%MoS₂ (HT).

Fig. 6:

(a) Photocatalytic H₂ evolution activity over blank CdS, CdS-1%MoS₂ (PD), CdS-1%Pt, (1%GR-CdS)-1%MoS₂ (PD), and (1%GR-CdS)-1%MoS₂ (HT) under visible light irradiation (λ>420 nm); HRTEM images of (b) (1%GR-CdS)-2%MoS₂ (PD) and (c) (1%GR-CdS)-1%MoS₂ (HT) composites; (d) polarization curves, (e) transient photocurrent responses, and (f) electrochemical impedance spectroscopy Nyquist plots of the electrodes. Reproduced from Ref. [48] with permission of American Chemistry Society.

In addition to the microstructure of MoS₂, its crystallinity has also been demonstrated to be influential on the cocatalytic performance of MoS₂. Ye and coworkers have found that proper annealing treatment of MoS₂/graphene-CdS (denoted as MoS₂/G-CdS) in an Ar atmosphere can improve the visible-light-driven photoactivity of the composites for H₂ production [47]. As shown in Fig. 7a and b, the rate of H₂ evolution over the as-prepared sample obtained at 453 K is about 0.88 mmol h⁻¹. After the annealing process, the photoactivity of the samples firstly increases and then declines with the rise of the annealing temperature. The MoS₂/G-CdS with annealing treatment at 573 K exhibits the optimal performance. Such observations have been ascribed to the effects of annealing process on the crystallinity and morphology of the composites. More specifically, when relatively low synthesis temperature condition is adopted, the crystallization of MoS₂ is poor, which is not advantageous to photogenerated electrons migrating to the surface for reaction. Increasing the annealing temperature to 573 K, the crystallinity of MoS₂ is improved, as reflected by the XRD pattern in Fig. 7c. The HRTEM image (Fig. 7d) of MoS₂/G-CdS composites (the molar ratio of MoS₂ to graphene is 1:2 and the amount of MoS₂/G is 2.0 wt%) after annealing at 573 K for 2 h shows a typical layered MoS₂ with an interlayer distance of 0.62 nm corresponding to the information on the (002) peak of MoS₂. Further increasing the annealing temperature (e.g. 673 K) leads to the agglomeration of CdS particles and uniform dispersion of MoS₂-graphene cocatalysts (Fig. 7e), thus resulting in the decrease of the photoactivity. Therefore, MoS₂/G-CdS composites obtained at the annealing temperature of 573 K exhibit the best photocatalytic H₂ evolution activities.

Fig. 7:

Visible light (λ>420 nm) photocatalytic H₂ production activities of MoS₂/G-CdS with 2.0 wt% cocatalyst and a molar ratio of MoS₂ to graphene of 1:2 (a) annealing at different temperatures and (b) the cycling behavior in 5 h; (c) XRD patterns of as-prepared MoS₂/G, MoS₂/G, and MoS₂ after annealing at 573 K for 2 h in an Ar atmosphere; (d) HRTEM image of the MoS₂/ G-CdS composite after annealing at 573 K for 2 h in an Ar atmosphere; (e) SEM image of MoS₂/G-CdS sample annealed at 673 K for 2 h in an Ar atmosphere. Reproduced from Ref. [47] with permission of American Chemistry Society.

The effect of graphene electrical conductivity

As discussed in the third Section ‘The roles of dual cocatalysts in enhancing the photoactivity of CdS for HER’, in the ternary CdS-graphene-MoS₂ composites, one important role of graphene is to accept the electrons photogenerated from CdS and shuttle to MoS₂ for the reduction of water to H₂. As demonstrated by the results shown in Figs. 5–7, the introduction of graphene is able to further promote the separation and transfer of electron-hole pairs and thus the HER photoactivity enhancement of the ternary composites as compared to the binary CdS-MoS₂ counterparts. Considering such function of graphene in shuttling electrons, the electrical conductivity is a key factor to determine its electron-accepting and transfer capability, thus affecting the photocatalytic HER performance of the composites.

Liu and coworkers have investigated the effect of the electrical conductivity of graphene on the photoactivity of CdS-graphene-MoS₂ composites [72]. They have employed two-step hydrothermal processes to synthesize CdS-graphene-MoS₂, as illustrated in Fig. 8a. Firstly, GO, ammonium molybdate tetrahydrate and thiourea are mixed to undergo hydrothermal reactions. They have found that with the reaction time of 24 h, MoS₂-RGO (MRGO) is obtained; when the reaction time is prolonged to 72 h, nitrogen doping in the RGO occurs, thus leading to the formation of MoS₂/N-doped RGO (MNRGO). Based on MRGO or MNRGO, CdS component is then introduced. Taking MNRGO as an example, it can be seen from Fig. 8b and c that MoS₂ nanosheets grow on the surfaces of NRGO sheets and the CdS nanoparticles (NPs) of hexagonal phase are distributed on both the surfaces and the edges of MoS₂ and NRGO nanosheets. This has been further confirmed by the TEM elemental mappings. As shown in Fig. 8d–i, the elements of Mo, S, and Cd are uniformly distributed across the NGRO formwork consisting of C and N elementals.

Fig. 8:

(a) Schematic illustration of the processing steps to prepare CdS@MNRGO hierarchical nanoarchitecture; (b) HRTEM and (c) the corresponding amplified image of CdS@MNRGO; (d) high-magnification dark-field FESEM–STEM image of CdS@MNRGO; (e–i) elemental mappings of the selected area of white box in (d); (j) visible-light (λ>420 nm) photocatalytic H₂ evolution rates of CdS@MNRGOx and CdS@MRGOx (x=1, 3, 5 and 8). Reproduced from Ref. [72] with permission of John Wiley & Sons, Inc.

The photocatalytic HER performance of the samples has been evaluated under visible light irradiation (λ>420 nm). Figure 8j displays the comparison of the hydrogen production rates of CdS@MNRGOx with CdS@MRGOx (x is the weight percentage of MRGO or MNRGO in the composites, which equals to 1, 3, 5, or 8). It is observed that with the same content of cocatalysts, CdS@MNRGO hybrids exhibit higher H₂ evolution rates than CdS@MRGO. This should be ascribed to that in the N-doped RGO (NRGO), the N atom activates the adjacent C atom by tailoring its electron donor acceptor property and consequently lead to improved electrical conductivity [73], [74], [75]. Therefore, the hierarchical nanoarchitectures of CdS@MNRGO would facilitate more efficient electron separation and transfer, and thus exhibit better photocatalytic performance for H₂ evolution than the non-N-doped counterpart CdS@MRGO [72].

The effect of the components spatial distribution

The intimate interfacial contact among the components has proven to be crucial for the efficient charge separation and thus for enhancing the photocatalytic performance of the composites, which has been evidenced by the higher photoactivity of the hybrids prepared through in situ growth approach than the counterparts obtained by mechanical mixing method [50], [72]. It is worth noting that in addition to the interfacial contact, the spatial distribution of the components also has important influence on the performance of the composites. For example, Li et al. have found that MoS₂ can be controllably loaded on the surface of graphene or CdS of the graphene-CdS composites at different pH values (Fig. 9e) [76]. More specifically, reduced graphene oxide (RGO)/CdS composites are firstly prepared by a solvothermal method using dimethyl sulfoxide (DMSO) as a sulfide source and reducing agent. Then MoS₂ is loaded onto the RGO/CdS by a facile photodeposition method using (NH₄)₂MoS₄ as a precursor. The selective deposition of MoS₂ onto RGO or CdS can be achieved by adjusting the pH value of the solution. At low pH value (e.g. 7), MoS₂ preferentially deposits on the surface of the CdS particles of the composite; as for high pH value of 11, it loads on the exposed RGO, as observed in Fig. 9a–d, respectively. Such modulation on the spatial distribution of MoS₂ can be ascribed to the variation of the zeta potentials of CdS and RGO and thus their interaction with MoS₄²⁻ along with the change of pH value [76].

Fig. 9:

TEM and HRTEM images of (a, b) RGO1.5/CdS/MoS₂-7 and (c, d) RGO1.5/CdS/MoS₂-11; (e) schematic illustration of tunable photodeposition of MoS₂ on RGO1.5/CdS; (f) photocatalytic hydrogen evolution under visible light irradiation (λ>420 nm) for 1 h over RGO1.5/CdS/MoS₂ prepared at different pH; schematic illustration of (g) incompatibility between the photoexcited electron transfer from CdS to RGO and the HER at MoS₂ for RGO1.5/CdS/MoS₂-7, and (h) the compatibility for RGO1.5/CdS/MoS₂-11. Reproduced from Ref. [76] with permission of American Chemistry Society.

Figure 9f displays the effect of different spatial distribution of MoS₂ on the photocatalytic H₂ evolution over RGO/CdS/MoS₂ composites under visible light irradiation (λ>420 nm). For the RGO1.5(wt%)/CdS/MoS₂-7 (pH 7) sample, a low activity (23 μmol h⁻¹) can be observed. With the increase of pH value, the activity exhibits a parabolic trend. At the pH value of 11, the sample RGO1.5/CdS/MoS₂-11 (pH 11) reaches the highest photocatalytic HER activity (99 μmol h⁻¹), which is enhanced by a factor of 4.3 as compared to RGO1.5/CdS/MoS₂-7. However, when the depositing pH is 12, the activity decreases to some extent (80 μmol h⁻¹), which may be due to the too high loading ratio of MoS₂ (~77 wt%) to RGO1.5/CdS [76]. The distinct difference in the photoactivity of RGO1.5/CdS/MoS₂-7 and RGO1.5/CdS/MoS₂-11 has been ascribed to the different spatial distribution of the components, which leads to dissimilar charge transfer process. For the sample of RGO1.5/CdS/MoS₂-7 with MoS₂ loading on CdS, the photoexcited electrons of CdS that are injected into RGO could not effectively reduce water into H₂ due to the lack of reactive sites. In this case, RGO would by contraries act as a recombination center of the electron-hole pairs (Fig. 9g), leading to the low HER activity. This indicates that the electron transferring from CdS to RGO is antisynergic with the HER at CdS. Besides, the MoS₂ deposited on the CdS may shield the light absorption of CdS, which is disadvantage to the formation of charge carriers and thus the photoactivity of the composites. For RGO1.5/CdS/MoS₂-11, because of MoS₂ depositing on RGO, the electrons can transfer from the conduction band of CdS to RGO and then to MoS₂, as illustrated in Fig. 9h. Such effective separation of the electron-hole pairs by RGO is compatible with HER at the MoS₂ on RGO, which leads to positive synergic effect among the components for photocatalytic HER [76].

Based on this work, we can acquire that as for multinary composite photocatalysts, besides the intimate interfacial contact among the components, their spatial distribution also needs to be taken into account. The modification of the component spatial distribution is able to tune the types of interfaces formed in the composites, which will affect the charge transfer pathway, the efficacy of the components, and thus the photoactivity of the composites.