A colloidal paste of TiO₂ NPs was prepared using the method of Syrrokostas et al.^[40] as follows: 3  g of commercial Degussa P25 TiO₂ nanopowders (P25) (Degussa AG, Dusseldorf, Germany) was ground in a porcelain pestle and mortar and mixed with 1  ml of distilled water containing acetyl acetone (10%  v/v). Acetyl acetone was used as a dispersing agent because it prevents the coagulation of TiO₂ NPs and its influence on the porosity of the film. The paste was diluted further by the slow addition of distilled water (4  ml) under continuous grinding. The addition of water controls the viscosity and final concentration of the paste. Finally, a few drops of a detergent (Triton X-100) were added for reducing the surface tension of the paste, which facilitates even spreading and reducing the formation of surface cracks.

To synthesize 0.2  wt% of RGO+ TiO₂ nanocomposite paste, the method of Yu et al.^[39] was followed. Briefly, RGO (thickness = 0.553.74  nm, size = 0.5  μ m– 3.0  μ m, purity > 99%, purchased from Chengdu Organic Chemicals Co. Ltd.) was rinsed with deionized water three times and dried at 60  ° C under ambient conditions. Then 6  mg of RGO was added to the previous TiO₂ NPs paste and dispersed using an ultrasonic horn for 30  min. After stirring for 2  h, using the doctor blade technique, the 0.2  wt% RGO+ TiO₂ nanocomposite paste was coated on a transparent conducting glass substrate made from SnO₂:F (FTO) (purchased from Solaronix, TCO22-7 sheet resistance = 7  Ω /sq and > 80% transmittance in the visible region). Then, the films were sintered at 450  ° C for 1  h. After cooling to room temperature, CdS QDs were deposited on the RGO+ TiO₂ nanocomposite photoanode layer by the SILAR deposition method. Two separated solutions of 0.5-M Cd(NO₃)₂ in ethanol and 0.5-M Na₂S in methanol were prepared. The RGO+ TiO₂ nanocomposite photoanodes were dipped into 0.5-M Cd(NO₃)₂ solution for 30  s and rinsed with ethanol and then, dipped into 0.5-M Na₂S for 30  s and rinsed again with methanol. The two-step dipping procedure was considered to be one cycle. This sequential coating was repeated for several cycles to form seven working photoanodes marked from a to g. After solvent evaporation, the final nanocomposite film thickness was 8 ± 1  μ m. The photoanodes were then kept at room temperature. The counter electrodes were FTO substrate sheets coated with Pt.

The working photoanode and Pt counter electrode were assembled into a sandwich-type cell using clamps. Both electrodes were sealed using a hot-melt polymer sheet (Solaronix, SX1170-25PF) that was 25-μ m thick to avoid the evaporation of the electrolyte. The active area of the assembled cells was 1  cm × 1  cm. Finally, an iodide electrolyte solution was prepared by dissolving 0.127  g of 0.05-M iodine (I₂) in 10  ml of water-free ethylene glycol and then adding 0.83  g of 0.5-M potassium iodide (KI). The electrolyte was inserted into the cell with a syringe, filling the space between the two electrodes.

The absorption spectra of the working photoanodes were recorded using an ultraviolet-visible (UV-Vis) spectrophotometer (JASCO V-670). The current density– voltage (J– V) characteristics were recorded with a Keithley 2400 voltage source/ammeter using Green Mountain IV-Sat 3.1 software, while the CdS QDSSCs were subjected to a solar simulator (ABET technologies, Sun 2000 Solar Simulators, USA) operating at 100  mW/cm² (AM1.5). The intensity of the incident solar illumination was adjusted to 1 sun condition using a Leybold certified silicon reference solar cell (Model: (57863) Solar cell 2  V/0.3 A STE 4/100). The photo-voltage decay rates of the assembled CdS QDSSCs were measured using a Tektronix 500  MHz oscilloscope. All experiments were carried out under ambient conditions.

The UV-Vis absorption spectra of CdS QDs sensitized RGO+ TiO₂ nanocomposite photoanodes for several SILAR deposition cycles (0, 2, 3, 4, 6, and 8) are shown in Fig.  2. It is clearly seen that the absorption is enhanced as the number of SILAR cycles increases, due to a greater amount of CdS QD loading. Furthermore, the absorption edges of CdS QDs sensitized (RGO+ TiO₂) nanocomposite photoanodes are redshifted as the number of SILAR cycles increases also. The red-shifts of the absorption shoulder and onset position with increasing SILAR cycles imply a slight growth of the CdS QDs. Such red shifts are in reasonable agreement with the results in Refs.  [18] and [41].

Fig.  2.

	Figure Option ViewDownloadNew Window
	Fig.  2. UV-Vis absorption spectra of CdS QDs sensitized RGO+ TiO2 nanocomposite photoanodes for different SILAR deposition cycles.

Figure  3 shows the UV-Vis absorption spectra of a plain TiO₂ NPs photoanode, a RGO+ TiO₂ nanocomposite photoanode, a CdS QDs (6  cycles) sensitized plain TiO₂ NPs photoanode, and a CdS QDs (6  cycles) sensitized RGO+ TiO₂ nanocomposite photoanode. It is clearly evident that the absorption of the CdS QDs (six cycles) sensitized RGO+ TiO₂ nanocomposite photoanode is higher than that of the CdS QDs (6  cycles) sensitized plain TiO₂ NPs photoanode. This behavior indicates that more CdS QDs have been adsorbed on the RGO+ TiO₂ nanocomposite photoanode. Such an increase of CdS QDs in quantity is due to the high roughness and porosity of the RGO+ TiO₂ nanocomposite surface.^{[35, 42]}

Fig.  3.

	Figure Option ViewDownloadNew Window
	Fig.  3. UV-Vis absorption spectra of the plain TiO2 photoanode, RGO+ TiO2 nanocomposite photoanode, CdS QDs (six cycles) sensitized TiO2 NPs photoanode, and CdS QDs (six cycles) sensitized RGO+ TiO2 nanocomposite photoanode.

The photovoltaic performance of CdS QDSSCs based on RGO+ TiO₂ nanocomposite photoanodes for several SILAR deposition cycles is investigated. Figure  4 shows the J– V characteristic curves which are measured under a simulated sunlight with an intensity of 100  mW/cm² (AM1.5). The photovoltaic characteristic parameters (V_oc, J_sc, FF, and η ) for CdS QDSSCs are given in Table  1. Up to six SILAR cycles, it is clearly observed that both J_sc and η increase, peaking at 1.83  mA/cm² and 0.37% respectively. These values lower again to 1.31  mA/cm² and 0.26% for seven SILAR cycles and to 0.92  mA/cm² and 0.17% for eight SILAR cycles. These results could be explained in terms of, as the number of SILAR cycles increases up to six cycles, an additional quantity of CdS QDs are loaded because of the high roughness and porous surface of the RGO+ TiO₂ nanocomposite^{[35, 42]} which, in turn, leads to increasing the absorption of the incident solar power (as shown in Fig.  2). Moreover, increasing the number of SILAR cycles causes a significant red shift of the absorption shoulder (as also shown in Fig.  2) and thus causes the relatively high absorption of the incident photons from the visible solar spectrum. While, as the number of SILAR cycles increases to more than six cycles, the extra quantity of CdS QDs leads to blocking the nanopores of the RGO+ TiO₂ nanocomposite layer, causing both J_sc and η to decrease. These results are in good agreement with our previous studies.^{[10, 43]} We have made measurements on J_sc and η of CdSe QDs-sensitized TiO₂ NPs using the DA technique. We find the values of J_sc and η for the 6-h dipping time are better than those of higher dipping times.

Fig.  4.

	Figure Option ViewDownloadNew Window
	Fig.  4. J– V characteristics curves of CdS QDSSCs using RGO+ TiO2 nanocomposite photoanode for several SILAR cycles (curve a to curve g refers to two to eight cycles).

Table 1.

Table 1.

Table 1. J– V characteristic parameters of CdS QDSSCs using RGO+ TiO2 nanocomposite photoanode for several SILAR cycles (curve a to curve g refers to two to eight cycles). No. of SILAR cycles Voc /V ± 0.01 Jsc /(mA/cm2)± 0.01 FF η /(± 0.01)% 2 0.65 0.22 0.34 0.07 3 0.66 0.39 0.34 0.10 4 0.67 0.63 0.33 0.16 5 0.65 1.58 0.30 0.29 6 0.66 1.83 0.30 0.37 7 0.66 1.31 0.28 0.26 8 0.66 0.92 0.27 0.17

Table 1. J– V characteristic parameters of CdS QDSSCs using RGO+ TiO2 nanocomposite photoanode for several SILAR cycles (curve a to curve g refers to two to eight cycles).

Figure  5 shows the J– V characteristic curves of CdS QDSSCs based on both the plain TiO₂ NP photoanode and the RGO+ TiO₂ nanocomposite photoanode for six-SILAR deposition cycles. The photovoltaic parameters for both QDSSCs are summarized in Table  2. It is clearly observed that there are notable increases in V_oc (∼ 10%), J_sc (∼ 26%), and η (∼ 33%) in CdS QDSSCs based on the RGO+ TiO₂ nanocomposite as compared with those based on plain TiO₂ NPs under an AM1.5 simulated sunlight. These considerable improvements in the photovoltaic parameters for CdS QDSSCs based on the RGO+ TiO₂ nanocomposite photoanode are mainly attributed to many factors. First, the enhancement in the absorption of the RGO+ TiO₂ nanocomposite photoanode, due to the rough surface structure of RGO+ TiO₂ nanocomosite film, ^[42] which leads to increasing the CdS QD loading on the film. Second, the rapid electron transport through RGO, offering fast interpaths due to the superior electrical conductivity of RGO (∼ 4 × 10³  Ω ^{− 1}· m^{− 1[44]}) compared with that of the TiO₂ NPs (∼ 8.11× 10^{− 9}  Ω ^{− 1}· m^{− 1[45]}). The addition of a small amount of RGO leads to improving the electron mobility of the nanocomposite.^{[38, 39, 42]} The RGO dispersed into the TiO₂ matrix facilitates the faster transport of photoexcited electrons (from the CdS QDs) across the TiO₂ network and the collection at the FTO surface. Our results are in good agreement with the results in Refs.  [38] and [39]. Peining et al.^[38] fabricated DSSCs using TiO₂ coated with CNTs (TiO₂-CNTs). In their work, they concluded that the TiO₂-CNT content (0.2  wt%) cell showed a ∼ 50% increase in η . They attributed this enhancement mainly to the improvement in interconnectivity between TiO₂ particles and TiO₂-CNTs in the porous TiO₂ film. Third, the electron back recombination rate decreases due to the existence of RGO in the nanocomposite photoanode, as discussed below.

Fig.  5.

	Figure Option ViewDownloadNew Window
	Fig.  5. J– V characteristic curves of CdS QDSSCs using (curve a) plain TiO2 NPs photoanode and (curve b) RGO+ TiO2 nanocomposite photoanode for six SILAR cycles.

Table 2.

Table 2.

Table 2. J– V characteristic parameters of CdS QDSSCs with and without RGO working electrode for six SILAR cycles. Working electrode Voc /V ± 0.01 Jsc /(mA/cm2) ± 0.01 FF η /(± 0.01)% Without RGO 0.61 1.45 0.31 0.28 With RGO 0.66 1.83 0.30 0.37

Table 2. J– V characteristic parameters of CdS QDSSCs with and without RGO working electrode for six SILAR cycles.

Figure  6(a) shows V_oc decay rates of CdS QDSSCs based on both plain TiO₂ NPs and RGO+ TiO₂ nanocomposite photoanodes for six SILAR cycles, measured by using a 500-MHz oscilloscope (Tektronix). It is clearly seen that the V_oc decay rate of CdS QDSSCs based on the RGO+ TiO₂ nanocomposite photoanode is much slower than that when using the plain TiO₂ NPs photoanode. This deliberated decay implies that the charge recombination between the injected electrons and the holes is remarkably retarded. RGO facilitates the electron transport from CdS QDs across the TiO₂ network to the FTO surface, since the RGO energy level lies between the CB of TiO₂ and FTO. The energy level diagram of CdS QDSSC using the RGO+ TiO₂ photoanode is shown in Fig.  6(b). The stepwise energy levels of CdS QDs, TiO₂, RGO, and FTO are beneficial for electrons to transport from the conduction band of CdS QDs to the FTO substrate. RGO leads to reducing the electron back recombination rate significantly. Therefore, RGO enhances the photovoltaic performance of CdS QDSSCs.

Fig.  6.

Figure Option ViewDownloadNew Window

Fig.  6. (a) Open circuit voltage (Voc) decay rates of CdS QDSSCs based on both plains. The TiO2 NPs photoanode and RGO+ TiO2 nanocomposite photoanode, (b) the energy level diagram of CdS QDSSC using the RGO+ TiO2 photoanode, where the Fermi levels of TiO2, RGO, and redox (I− /I3− ) are set to be at − 4.21, [48] − 4.42, [48] and − 4.94  eV.[49] respectively, and (c) the variations of electron lifetime τ with voltage, obtained from Eq.  (1) for six SILAR cycles

The electron lifetime constant (τ ) as a function of the decaying V_oc is calculated from the following equation:^{[46, 47]}

The values of lifetime constant (τ ) of CdS QDSSCs based on both plain TiO₂ NPs and RGO+ TiO₂ nanocomposite photoanodes for six SILAR cycles are shown in Fig.  6(c). It is clearly seen that τ for CdS QDSSCs based on the RGO+ TiO₂ nanocomposite photoanode is at least one order of magnitude larger than that based on the plain TiO₂ NPs photoanode. This novel evidence verifies that RGO acts as a blocking layer to reduce the back electron recombination in QDSSCs. Similar results were achieved in Refs.  [46] and [47]. Al-juaid et al.^[46] concluded that the addition of a blocking layer of ZnO to DSSCs plays an important role in reducing the back electron– carrier recombination, which enhances the whole efficiency of DSSCs. In our work, RGO facilitates the electrons transport from CdS QDs across the TiO₂ network to the FTO surface as mentioned before, while the blocking ZnO layer^[46] prevents electrons from recombining due to the energy barrier caused by the ZnO layer.