Dye-sensitized solar cell utilizing silver doped reduced graphene oxide films counter electrode: Influence of annealing temperature on its performance

1 Introduction

The work on free-platinum film as counter electrode in DSSC with much cheaper material has extensively been carried out by several research groups (Yun et al., 2014; Huang et al., 2007; Zhu et al., 2008; Nam et al., 2010). The use of carbon film as counter electrode of the device has been attempted due its high electronic conductivity, corrosion resistance and electrochemical reactivity (Zhu et al., 2008; Nam et al., 2010; Roy-Mayhew et al., 2010). Graphene can also replace platinum as counter electrode in DSSC as it is much cheaper than platinum and possesses high conductivity. Several researchers have utilized graphene films as counter electrode in DSSC (Roy-Mayhew et al., 2010; Zhang et al., 2011; Wang et al., 2013; Kavan et al., 2016). It was found that the efficiency of the DSSC utilizing graphene films counter electrode was comparable with that of the device using platinum films counter electrode (Roy-Mayhew et al., 2010; Zhang et al., 2011; Wang et al., 2013; Kavan et al., 2016). In our previous work, we have employed reduced graphene oxide (rGO) as counter electrode in DSSC and achieved the efficiency of 0.09% (Rahman et al., 2017). Then, we utilized gold doped reduced graphene oxide (Au-rGO) as cathode in DSSC and demonstrated the efficiency of 0.175% (Rahman et al., 2018). Nitrogen doped graphene has been used as counter electrode in DSSC (Xue et al., 2012; Ju et al., 2013). Phosphorous doped reduced graphene oxide (P-rGO) has been employed as counter electrode of DSSC (Wang et al., 2014).

Since the efficiency is low, the electronic conductivity of rGO should be improved by doping it with non-metallic or metallic material. This causes the bulk resistance of the device to be reduced and consequently improves the efficiency of the device. Xue and his group utilized nitrogen doped graphene foams as counter electrode in DSSC and demonstrated the efficiency of 7.07% at 7.0% nitrogen. This efficiency is slightly lower than that of the device utilizing Pt counter electrode that was 7.44% (Xue et al., 2012). Ju and his co-workers used nitrogen doped graphene nanoplatelet as superior metal free counter electrode for DSSC and achieved the efficiency of 9.05%, higher than that of the device with Pt counter electrode that was 8.43% (Ju et al., 2013). Wang and his co-researchers utilized phosphorous doped reduced graphene oxide (rGO) as an electrocatalyst counter electrode in DSSC and yielded the efficiency of 6.04% at 1.27 at% phosphorous (Wang et al., 2014). The efficiency of the device utilizing Pt counter electrode was 6.80% (Wang et al., 2014).

In this work, we have synthesized silver doped rGO (Ag-rGO) and utilized it as a counter electrode in DSSC. The Ag-rGO samples underwent annealing treatment at various temperatures in order to improve its electronic conductivity and catalytic activity when employed in DSSC. The originality of this work is the use of annealed Ag-rGO films as counter electrode of DSSC. The objective of this work is to investigate the effect of annealing temperature of Ag-rGO on its properties and also on the performance parameters of the DSSC.

2 Experimental

Reduced graphene oxide (rGO) was synthesized via Hummer’s method. 2.0 g graphite flakes and 1.0 g sodium nitrate were dissolved into 200 ml of 98.0% sulfuric acid (H₂SO₄) in right sequence and stirred with 600 rpm at room temperature for 30 min. Subsequently, 6.0 g potassium permanganate was slowly added to the mixture solution over 30 min at the same temperature. The solution was then allowed to stir for 3 h 30 min. 300 ml deionized water was slowly added to the dark brown slurry, whereby white smoke was released. The slurry was allowed to stir for another 2 h. 160 ml of 5% hydrogen peroxide (H₂O₂) was added to the slurry and stirred for 15 min. A spontaneous change in colour from brown to yellow was observed, indicating the successful oxidation of graphite oxide. The yellow slurry was then washed with 5% of H₂SO₄, 5% of H₂O₂ and 37% of HCl to remove the excess metal ions. Finally, it was washed with deionized water 3 times. The brown dispersion was then dried in an oven at 80 °C for 24 h.

To prepare Ag doped rGO, 0.1 g GO was first ultrasonically dispersed in 10 ml DI water for 1 h. Silver nitrate (AgNO₃) with the concentration of 0.20 M was dissolved in 5 ml DI water which was then added to GO solution. The solution was sonicated for another 30 min and followed by stirring for 3 h. The solution was spin coated on glass substrate at 1500 rpm for 30 s to obtain Ag-rGO film. The sample was then annealed at 350 °C under nitrogen gas for one hour. These procedures were repeated for preparing Ag doped rGO annealed at other temperatures, namely, 360, 370, 380, 390 and 400 °C. The un-annealed sample was also prepared as the control sample. The phase structure of the un-annealed and annealed samples was studied by XRD. The morphology of the samples was characterized by FESEM. The optical transmission was characterized by UV–VIS.

The DSSC was fabricated by utilizing TiO₂ coated N719 dye as photoanode, Ag doped rGO counter electrode and iodide/triiodide electrolyte solution. The TiO₂ films were prepared by liquid phase deposition. 0.1 M ammonium hexafluorotitanate was mixed with 0.2 M boric acid and heated in an oven at 50 °C for 5 h. The sample was then annealed at 400 °C for 1 h. The photoanode was sensitized by immersing it into an ethanol solution containing 0.5 mM  N719 dye for 15 h at room temperature. The area of TiO₂ photoanode and Ag-rGO counter electrode was 0.23 cm². The device was fabricated by assembling Ag doped rGO and TiO₂ coated N719 dye separated by parafilm. The current-voltage measurement under illumination of 100 mW cm⁻² light from tungsten halogen lamp was performed to analyze photovoltaic performance of the devices. The electrochemical impedance spectroscopy (EIS) was also performed under illumination of 100 mW cm⁻² light to investigate the charge transfer resistance and carrier lifetime. The performance stability has been performed by investigating the efficiency of the best device within the period of 8 days. The best device is the device utilizing silver doped rGO annealed at 350 °C. The efficiency was calculated on day 1, 4, 5, 6, and 7. The EIS data such as the charge transfer resistance of the interface Ag-rGO/electrolyte was also determined during these days to support the data of efficiency. The cyclic voltammetry (CV) measurement was performed by using electrochemical cell with three-electrodes in acetonitrile solution consisting of 10 mM LiI, 1 mM I₂ and 0.1 M LiClO₄. The Ag-rGO samples were used as working electrode, Pt wire as counter electrode and Ag/AgCl as reference electrode, respectively.

3 Results and discussion

Fig. 1 represents the XRD spectra of Ag-rGO sample prepared at 25 °C and Ag-rGO samples with different annealing temperature; namely, 350 °C, 360 °C, 370 °C, 380 °C, 390 °C and 400 °C. The GO peak of 25 °C sample appears at 11.42° with interlayer d-spacing of 0.88 nm. Upon thermal annealing, the GO peak completely disappears and a broad peak belonging to rGO appears at around 2θ = 24.0° for all annealed samples with interlayer d-spacing of 0.37 nm, a typical pattern of amorphous carbon structure, revealing the formation of rGO. A decrease in the interlayer d-spacing without and with thermal annealing is ascribed to the removal of the oxygen functional groups in GO during the thermal annealing process (Soo et al., 2016). It is interesting to note that the broad peak at 2θ = 24.0° is shifted to the higher diffraction angle as the annealing temperature increases. Upon thermal annealing, the interlayer distance is decreased further to 0.36 nm as a result of more removal of functional groups at higher temperature (Huh et al., 2010; Umar et al., 2013).

Close

Fig. 1 XRD patterns of silver-doped rGO film with various annealing temperatures.

Export to PPT

The peaks at 31, 36, 51 and 61° belong to ITO substrate (Roza et al., 2012). The silver peaks appear at 2θ = 38, 44, 64, and 77°, representing the crystallographic planes of (1 1 1), (2 0 0), (2 2 0), and (3 1 1) of the face-centered cubic of Ag, respectively, according to the International Centre for Diffraction Data (ICDD). The peak intensity of silver at 2θ = 38° increases with the annealing temperature up to 390 °C. At the temperature of 400 °C, the peak intensity of Ag decreases. In general, the intensity of silver peaks of the annealed samples is higher than compared with that of un-annealed sample prepared at the diffraction angle of 38°. The crystallite size of Ag is estimated at the diffraction angle of 38° and illustrated in Table 1. From the table, it is found that the crystallite size is in the range of 19–32 nm corresponding with the annealing temperature of 350–400 °C. Therefore, the crystallite size of Ag increases as the annealing temperature increases.

Fig. 2 depicts the FESEM images of Ag-rGO prepared at various annealing temperatures with 5000 and 30,000× magnifications, respectively. The images shows the dark region and white strips for all samples. It is observed the white silver particles reside on the surface of rGO for all samples. The silver particles that do not occupy the site of rGO lattice and reside on the surface is known as interstitial dopant (Mustaffa et al., 2018). It is also observed that at 350 °C, the white strips just start to appear. Once the temperature is increased to 370 °C, the size of white strips become bigger. Once the temperature is further increased to 400 °C, the white strips size decreases. However, the number of white strip is higher than that at 370 °C.

Close

Fig. 2 FESEM images of Ag-rGO prepared at various temperatures; (1) 350 °C, (2) 370 °C, and (3) 400 °C at (a) 5000 and (b) 30,000× magnifications.

Export to PPT

Fig. 3 illustrates the optical transmission spectra of the Ag-rGO samples annealed at various temperatures in ultraviolet and visible region. As observed from the figure, there is no increasing or decreasing trend with the annealing temperature of Ag-rGO. In other words, the optical transmission neither increase nor decrease with the annealing temperature. The sample annealed at 350 °C possesses the lowest transmission while that with the annealing temperature of 400 °C has the highest transmission around 9.8% in the visible region. By comparing with the unannealed Ag-rGO, its is found that the samples underwent annealing treatment possesses much lower transmission with that of the unannealed samples that is 30% (Mustaffa et al., 2018).

Close

Fig. 3 UV–Vis optical transmission of the Ag-rGO samples with various annealing temperatures.

Export to PPT

Fig. 4 shows the I-V curves in dark of the DSSC utilizing silver doped rGO films counter electrodes with various annealing temperatures of Ag-rGO. It is noticed that the devices do not show rectification property since the dark current in reverse bias is slightly larger than that in forward bias. The difference in the dark current is quite significant signifying that the annealing temperature affects the dark current. The device utilizing the sample annealed at 400 °C possesses the lowest leak current and the one utilizing the sample annealed at 360 °C possesses the highest leak current. Leak current is the current in reverse bias.

Close

Fig. 4 I-V curves of the devices in dark utilizing the samples with various annealing temperatures of Ag-rGO.

Export to PPT

Fig. 5 shows the J-V curves of the device utilizing silver doped rGO counter electrode annealed at various temperatures. The device with the sample annealed at 350 °C has the highest output power leading to the highest power conversion efficiency as illustrated in Table 2. The device utilizing the un-annealed sample generates the lowest output power, resulting in the lowest efficiency. The curves for all devices have high slope, producing low FF as also illustrated in Table 1. The other photovoltaic parameters such as J_SC and V_OC are also presented in Table 1.

Close

Fig. 5 J-V curves of the devices with various annealing temperatures under illumination of 100 mW cm⁻².

Export to PPT

According to Table 2, it is found that the device with the Ag-rGO annealed at 350 °C demonstrates the highest V_oc and η. The highest η is due to the smallest R_ct1 that this device owns as illustrated in Table 3. The device with the un-annealed sample performs the lowest η due to the shortest τ according to Table 3. It is also found that the efficiency of the device utilizing the annealed samples is higher that of the device with un-annealed sample. Annealing treatment on Ag-rGO has improved the interfacial contact between Ag-rGO electrode/electrolyte containing the redox mediator of iodide/triiodide. It consequently decreases the charge transfer resistance at the interface Ag-rGO/electrolyte. The lower the resistance at cathode/electrolyte interface, the faster triiodide is reduced to iodide to in order to improve the efficiency of the DSSC. Also, from the table, the V_OC does not change quite significantly with the annealing temperature. The FF is low since the area of maximum power rectangles is much smaller than that of the J-V curves shown in Fig. 4.

The highest η obtained in this work, that is 1.30%, is smaller compared with that of the DSSC utilizing phosphorous doped rGO counter electrode that was 6.04% (Wang et al., 2014). This might be caused the internal resistance of the DSSC fabricated in this work is higher than that reported in (Wang et al., 2014). However, it is higher than that of the device utilizing gold doped rGO counter electrode which was 0.175% (Rahman et al., 2018). This is because the R_ct of the DSSC utilizing silver doped rGO counter electrode in this work is smaller than that of the DSSC utilizing gold doped rGO which was 69 Ω (Rahman et al., 2018). The bulk resistance of the DSSC with silver doped rGO is smaller than that of gold doped rGO due to the electronic conductivity of silver is higher than that of gold.

Fig. 6 depicts the nyquist curves of the devices utilizing Ag-rGO counter annealed at various temperatures. The curves display two semi-circles for which the first semi-circle represents the charge transfer resistance at the interface of Ag-rGO/electrolyte (R_ct1) for reduction of triiodide to iodide and the second one denotes the charge transfer resistance at the interface of electrolyte/TiO₂/N719 dye (R_ct2) for oxidation of iodide to triiodide. R_ct1 and R_ct2 are also referred to semicircle in the high frequency and mid-frequency region, respectively. The bulk resistance (R_s) is from the origin to the starting point of the first semi-circle. The inset of Fig. 6 presents the equivalent circuit used to determine the bulk resistance (R_s), R_ct1 and R_ct2. However, the R_b is not presented in Table 3 since its value is about the same that is around 56 Ω. The capacitance, C₁ and C₂ which are parallel to R_ct1 and R_ct2 are also not determined and presented in Table 3. The other two resistances, R_ct1 and R_ct2 are estimated from the curves and illustrated in Table 3. The R_ct1 is found to increase with the annealing temperature until 380 °C and then drops. However, this trend is not observed for R_ct2. This resistance is found to decrease with the annealing temperature until 370 °C and then increases with the annealing temperature. Nevertheless, this increasing trend terminates at 390 °C and then drops. Also, according to the table, the device with the sample annealed at 350 °C possesses the lowest R_ct1 while the one utilizing the sample annealed at 380 °C has the highest R_ct1. For the case of R_ct2, the device employing the sample annealed at 360 °C has the smallest R_ct2, while that utilizing the sample annealed at 390 °C possesses the biggest R_ct2.

Close

Fig. 6 EIS curves of the devices with various annealing temperatures. Inset: Equivalent circuit model.

Export to PPT

Fig. 7 shows the bode curves of the devices utilizing Ag-rGO counter annealed at various temperatures. Each curve is found to be symmetry about the resonant frequency which is required to compute the carrier lifetime (τ) for each device. The carrier lifetime is then illustrated in Table 3. It is observed neither increasing nor decreasing trend of carrier lifetime with the annealing temperature is observed. The device employing the Ag-rGO annealed at 390 °C exhibits the longest lifetime whereas the one that applies the unannealed sample possesses the lowest lifetime. Also, according to Table 3, there is no clear correlation between R_ct1 and R_ct2 with τ is observed. In other words, the R_ct1 neither increases nor decreases with the annealing temperature. This behaviour is also not observed for R_ct2.

Close

Fig. 7 Bode plots of the devices with various annealing temperatures.

Export to PPT

Fig. 8 illustrates the J-V curves of the device utilizing Ag-rGO counter electrode annealed at 350 °C with various days. The photovoltaic parameter such as power conversion efficiency extracted from Fig. 7 and the corresponding charge transfer resistance at the interface of Ag-rGO/electrolyte electrode estimated from Fig. 8 are illustrated in Table 4. It is found from the table that the efficiency of the device degrades quite significantly within one week time. On the other hand, the corresponding charge transfer resistance determined from Fig. 9 has been found to increase with time in term of day. This phenomena is consistent with the theory that the higher resistance results in the lower performance in term of efficiency. Performance stability testing is an important aspect for practical application of counter electrode of DSSC. Some systematical evaluation techniques for counter electrode stability in the device have been reported in recent literatures (Yun et al., 2015; Yun et al., 2014; Li et al., 2018; Gong et al., 2018). This is to evaluate the counter electrode stability for most primary researches which has been given from different angles using various characterization techniques to perform the counter electrode stability assessment.

Close

Fig. 8 J-V curves of the device utilizing Ag-rGO counter electrode annealed at 350 °C with various days.

Export to PPT

Close

Fig. 9 Nyquist plots of the device utilizing Ag-rGO counter electrode annealed at 350 °C with various days.

Export to PPT

The CV curve of Ag-rGO prepared at 25 °C and annealed at 350 °C are depicted in Fig. 10(a). The presence of weak reduction peak in Ag-rGO at 25 °C confirms that the catalytic activity of the sample for the reduction of I₃⁻ to I⁻ is slow (Abdulkarim et al., 2017). The Ag-rGO at 350 °C shows higher cathodic peak current density representing a higher electrocatalytic activity than that of un-annealed Ag-rGO for the reduction of triiodide to iodide (Huang et al., 2017). The peak-to-peak separation (E_pp) and current density of reduction peak (J_red) are two noticeable parameters that can evaluate electrocatalytic activity. Reduction peak current density (J_red) reflects the catalytic activity of electrode. Peak-to-peak separation (E_pp) implies the overpotential and reaction rate in the catalytic reaction process (Elbohy et al., 2018). The E_pp of Ag-rGO annealed at 350 °C is 111 mV which is lower than 270 mV that is for Ag-rGO prepared at 25 °C. The J_red for Ag-rGO at 25 and 350 °C is −6.02 and −6.48 mA cm⁻², respectively. Higher J_red and lower E_pp lead to an enhanced electrocatalytic activity (Wang et al., 2016). Ag-rGO annealed at 350 °C exhibits smaller E_pp and higher J_red than the sample prepared at 25 °C which implies lowered overpotential and improved catalytic reaction rate of triiodide to iodide (Guai et al., 2012).

Close

Fig. 10 (a) CV curves of Ag-rGO at 25 °C and 350 °C at a scan rate of 50 mV s⁻¹ and (b) CV curves of Ag-rGO at 350 °C at different scan rates; inset: the relationship between the square root of scan rate and the resultant redox peak currents.

Export to PPT

Fig. 10(b) shows the CV curves of Ag-rGO annealed at 350 °C at different scan rates, ranging from 50 to 120 mV s⁻¹. The curves imply that the anodic and cathodic peak current gradually increase with the increase of scan rate. A good linear relationship between the redox peak current density and the square root of the scan rate (inset of Fig. 10(b)) is clearly observed. It indicates that the electrochemical redox reaction is controlled by the ionic diffusion transport and the electrode material does not react with the electrolyte (Yuan et al., 2016).

4 Conclusions

Ag-rGO films were successfully synthesized and employed as counter electrode in DSSC. Annealing treatment on Ag-rGO has been found to decrease the charge transfer resistance of the interface Ag-rGO/electrolyte and consequently improves the performance of the DSSC. The DSSC utilizing the sample annealed at 350 °C demonstrates the highest η that is 1.302% due to the smallest charge-transfer resistance at the interface of electrolyte/Ag-rGO that is 14.36 Ω. The performance stability and electrochemical stability results signify that argentum doped rGO has potential to be employed as a free platinum counter electrode of DSSC.

Conflict of interest statements

There is no conflict of interest in this work.