Ruthenium (Ru) Doped Titanium Dioxide (P25) Electrode for Dye Sensitized Solar Cells

In this study, P25-titanium dioxide (TiO2) was doped with ruthenium (Ru) by systematically varying the Ru content at 0.15, 0.30, 0.45 and 0.6 mol%. The synthesized Ru-doped TiO2 nanomaterials have been characterized by X-ray diffraction (XRD), Raman spectroscopy, energy-dispersive X-ray (EDX) analysis, UV-visible (UV–Vis) spectroscopy, and electrochemical impedance (EIS) spectroscopy. The XRD patterns of undoped and Ru-doped TiO2 nanomaterials confirm the presence of mixed anatase and rutile phases of TiO2 while EDX spectrum confirms the presence of Ti, O and Ru. Further, UV-visible absorption spectra of doped TiO2 nanomaterial reveal a slight red shift on Ru-doping. The short circuit current density (JSC) of the cells fabricated using the Ru-doped TiO2 photoanode was found to be dependent on the amount of Ru present in TiO2. Optimized cells with 0.3 mol% Ru-doped TiO2 electrodes showed efficiency which is 20% more than the efficiency of the control cell (η = 5.8%) under stimulated illumination (100 mWcm−2, 1 sun) with AM 1.5 filter. The increase in JSC resulted from the reduced rate of recombination upon doping of Ru and this was confirmed by EIS analysis.


Introduction
Dye-sensitized solar cells (DSSCs) have been studied intensively as an alternative energy source due to their low cost, easy fabrication and more environmentally friendly nature. A DSSC consists of an electron transporting mesoporous metal oxide layer on a transparent conducting oxide coated glass, dye, electrolyte and a counter electrode. Generally, the visible light is converted into electricity in DSSCs through spectral sensitization of wide bandgap semiconductors such as SnO 2 [1], SrTiO 3 [2], Nb 2 O 5 [3], ZnO [4] and TiO 2 [5]. Among the semiconductors explored for DSSCs so far, TiO 2 remains the most promising material [1]. Although natural dyes used as sensitizers in DSSC are cheap compared to synthetic dyes, their reported efficiencies are rather low [6][7][8][9]. The concept of a dye-sensitized solid-state solar cell was first proposed by Tennakone et al. in 1988 [10] and then O'Regan and Grätzel reported an efficiency of 7.1% and a current density greater than 12 mAcm −2 for DSSC in which I − /I − 3 redox couple and TiO 2 were used as liquid electrolyte and ETM, respectively [5]; later, the maximum efficiency of 11.1% was reported by Nazeeruddin et al. [11,12]. However, relaxation of oxidized dye

Preparation of Ru-Doped TiO 2 Nanomaterials
To achieve a reproducible outcome and comparison, a simple doping of Ru in P25-TiO 2 was used to prepare Ru-doped TiO 2 nanomaterials. Then 0.15, 0.3, 0.45 and 0.6 mol% of RuCl 3 ·xH 2 O were added to the TiO 2 nanopowder (P25) and stirred vigorously for one hour at room temperature (step 1, 2) and then it was dried at 100 • C with continuous stirring for 2 hours (step 3). The product was ground well (step 4) and the resulting Ru-TiO 2 powder mixtures were annealed at 500 • C for 3 h (step 5) ( Figure 1) [26,35,37,38].

Fabrication of DSSCs
The fluorine-doped tin oxide (FTO) coated conducting glass (sheet resistance 7.5 Ω/cm 2 ) was used as the current collector. It was cleaned initially with soap water and subsequently with distilled water and ethanol using an ultrasonic bath. Then, the synthesized Ru-doped TiO2 nanomaterial was made into paste by mixing with deionized water, acetylacetone and Triton TM X-100 binder and coated on FTO by the doctor-blade technique using a glass rod with adhesive tapes (3 M Scotch tapes) as spacers and the thickness of the TiO2 film was about 7 μm. The prepared Ru-doped TiO2 films were dried and calcinated at 500 °C for 30 minutes. Then the coated glasses were soaked in 0.3 mM solution of N719 dye in acetonitrile/tert-butyl alcohol (50%v/v) for 12 h. After the dye-sensitization process, the photoanode was washed with acetonitrile to remove the unanchored dye molecules and dried. A platinum-coated FTO glass plate was used as the counter electrode. The dye-coated Ru-doped TiO2 electrode and Pt counter electrode [26] were used to assemble the cell and ⁄ electrolyte was used as redox electrolyte.

Characterization
The structural properties of the synthesized nanomaterials were studied by the X-ray diffraction method using scan range (2θ) between 20° and 95° with step size of 0.02° and scan speed of 1°/min. A Raman spectroscopic study was carried out using a laser confocal Raman microscope (Renishaw, UK, Model: Invia). The optical absorbance spectra were recorded using Shimadzu 1800 Scanning Double Beam UV-visible spectrophotometer. The elemental composition of the synthesized nanomaterials was analyzed by the energy-dispersive X-ray spectroscopy technique. The photovoltaic performance of the cells was studied using Keithley-2400 source measurement unit (SMU) under simulated irradiation of intensity 100 mWcm −2 with AM 1.5 filter (Peccell-PEC-L12, Japan). Current-voltage (I-V) characteristics in the dark were measured before and after the illumination which confirmed no change in device behaviour [39,40]. The effective area of the photoelectrode was 0.25 cm 2 . Electrochemical impedance spectroscopy (EIS) measurements were carried out on the DSSCs using Metrohm Autolab potentiostat/galvanostat (PGSTAT 128N) with a frequency response analyzer (FRA 32M).

X-ray Siffraction and Raman Spectroscopy
The crystal structure of the synthesized nanomaterials was investigated by the X-ray diffraction method (XRD). Figure 2 represents the XRD patterns of undoped TiO2 electrode, 0.15, 0.3, 0.45 and 0.6 mol% Ru-doped TiO2 electrodes and diffraction peaks for 2θ diffraction angles were monitored

Fabrication of DSSCs
The fluorine-doped tin oxide (FTO) coated conducting glass (sheet resistance 7.5 Ω/cm 2 ) was used as the current collector. It was cleaned initially with soap water and subsequently with distilled water and ethanol using an ultrasonic bath. Then, the synthesized Ru-doped TiO 2 nanomaterial was made into paste by mixing with deionized water, acetylacetone and Triton TM X-100 binder and coated on FTO by the doctor-blade technique using a glass rod with adhesive tapes (3 M Scotch tapes) as spacers and the thickness of the TiO 2 film was about 7 µm. The prepared Ru-doped TiO 2 films were dried and calcinated at 500 • C for 30 minutes. Then the coated glasses were soaked in 0.3 mM solution of N719 dye in acetonitrile/tert-butyl alcohol (50%v/v) for 12 h. After the dye-sensitization process, the photoanode was washed with acetonitrile to remove the unanchored dye molecules and dried. A platinum-coated FTO glass plate was used as the counter electrode. The dye-coated Ru-doped TiO 2 electrode and Pt counter electrode [26] were used to assemble the cell and I − /I − 3 electrolyte was used as redox electrolyte.

Characterization
The structural properties of the synthesized nanomaterials were studied by the X-ray diffraction method using scan range (2θ) between 20 • and 95 • with step size of 0.02 • and scan speed of 1 • /min. A Raman spectroscopic study was carried out using a laser confocal Raman microscope (Renishaw, UK, Model: Invia). The optical absorbance spectra were recorded using Shimadzu 1800 Scanning Double Beam UV-visible spectrophotometer. The elemental composition of the synthesized nanomaterials was analyzed by the energy-dispersive X-ray spectroscopy technique. The photovoltaic performance of the cells was studied using Keithley-2400 source measurement unit (SMU) under simulated irradiation of intensity 100 mWcm −2 with AM 1.5 filter (Peccell-PEC-L12, Japan). Current-voltage (I-V) characteristics in the dark were measured before and after the illumination which confirmed no change in device behaviour [39,40]. The effective area of the photoelectrode was 0.25 cm 2 . Electrochemical impedance spectroscopy (EIS) measurements were carried out on the DSSCs using Metrohm Autolab potentiostat/galvanostat (PGSTAT 128N) with a frequency response analyzer (FRA 32M).

X-ray Siffraction and Raman Spectroscopy
The crystal structure of the synthesized nanomaterials was investigated by the X-ray diffraction method (XRD). Figure 2 represents the XRD patterns of undoped TiO 2 electrode, 0.15, 0.3, 0.45 and 0.6 mol% Ru-doped TiO 2 electrodes and diffraction peaks for 2θ diffraction angles were monitored where d is the crystallite size, k is a dimensionless shape factor which has a typical value of about 0.89, λ is the X-ray wavelength of Cu (0.5406 nm), θ is the Bragg angle corresponding to the anatase (101) peak, and B is the line broadening at half the maximum intensity (FWHM and 74.89° and they correspond to the reflection planes of (101), (004), (200), (105), (211), (204), (116), (220) and (215) which confirms the presence of well-crystallized pure anatase TiO2 phase and peaks at 27.39°, 36.07° and 41.2° correspond to the reflection planes of (110), (101) and (111) for minor rutile TiO2 phase. This indicates that the anatase and rutile crystal structures are retained even after the TiO2 being doped with Ru. The average crystallite size was calculated by the Scherrer Equation [15] d= Ɵ where d is the crystallite size, k is a dimensionless shape factor which has a typical value of about 0.89, λ is the X-ray wavelength of Cu (0.5406 nm), Ɵ is the Bragg angle corresponding to the anatase (101) peak, and B is the line broadening at half the maximum intensity (FWHM  [35,42]. Figure 3a shows the Raman spectra obtained for the undoped TiO2 and Ru-doped TiO2 nanomaterials. The Raman spectrum of the undoped TiO2 nanoparticles was dominated with five bands corresponding to the six Raman active modes. Well resolved TiO2 Raman peaks with the D4h space group at about 170, 216, and 657 cm −1 (Eg), 539 cm −1 (A1g + B1g), and 420 cm −1 (B1g) were observed which correspond to the anatase phase of TiO2 [43]. Figure 3b shows that the peak intensity became weak and also the peak became broader with the increase in the percentage of Ru-doping. In general, the Raman line shape, intensity and position are strongly influenced by lattice strain, defects, and the crystallite size and shape. There was no peak for the Ru and the same behaviour is observed in the XRD studies. This suggests that Ru ions have been successfully incorporated in the TiO2 lattice.   Figure 3a shows the Raman spectra obtained for the undoped TiO 2 and Ru-doped TiO 2 nanomaterials. The Raman spectrum of the undoped TiO 2 nanoparticles was dominated with five bands corresponding to the six Raman active modes. Well resolved TiO 2 Raman peaks with the D 4h space group at about 170, 216, and 657 cm −1 (E g ), 539 cm −1 (A 1g + B 1g ), and 420 cm −1 (B 1g ) were observed which correspond to the anatase phase of TiO 2 [43]. Figure 3b shows that the peak intensity became weak and also the peak became broader with the increase in the percentage of Ru-doping. In general, the Raman line shape, intensity and position are strongly influenced by lattice strain, defects, Energies 2020, 13, 1532 5 of 12 and the crystallite size and shape. There was no peak for the Ru and the same behaviour is observed in the XRD studies. This suggests that Ru ions have been successfully incorporated in the TiO 2 lattice.

UV-visible Absorption Spectroscopy
UV-visible spectroscopy was employed to study the optical properties of the prepared nanomaterials. Five milligrams of the nanomaterial was dispersed in 100 ml of the deionized water in an ultrasonic bath and 1 ml of the dispersed sample was transferred to a standard quartz cuvette

UV-Visible Absorption Spectroscopy
UV-visible spectroscopy was employed to study the optical properties of the prepared nanomaterials. Five milligrams of the nanomaterial was dispersed in 100 ml of the deionized water in an ultrasonic bath and 1 ml of the dispersed sample was transferred to a standard quartz cuvette for measurement [23]. Figure 4a represents the optical absorption spectra of undoped TiO 2 , 0.15, 0.3, 0.45 and 0.6 mol% Ru-doped TiO 2 nanomaterials. As can be seen from Figure 4a the absorption peak of undoped TiO 2 nanomaterial appears in the UV region whereas there is a slight red shift in the absorption spectrum of Ru-doped TiO 2 and this red shift is found to increase with the increase in Ru-doping indicating the bandgap narrowing due to the introduction of a mid-bandgap or impurity levels located between the valence band and the conduction band of TiO 2 . In addition, the light absorption in the range from 400 to 700 nm is found to increase with increasing Ru content in the Ru-doped TiO 2 , accompanied with a change in the colour from white to reddish black [41]. The bandgap (E g ) of undoped TiO 2 , 0.15, 0.3, 0.45 and 0.6 mol% Ru-doped TiO 2 nanomaterials was calculated using Tauc plots, where the intercept of the tangent to the plot (αhυ) 2 versus hυ gives a good estimation of the direct bandgap for a semiconductor. The optical absorbance coefficient of a semiconductor for direct transition is given by the equation where hυ = photon energy, α = absorbance coefficient, E g = bandgap energy, A = constant and the exponent 'n' depends on the type of transition and it may have values of 1/2, 2, 3/2 and 3, corresponding to the allowed direct, allowed indirect, forbidden direct and forbidden indirect transitions, respectively. As shown in the Figure 4b,  for measurement [23]. Figure 4a represents the optical absorption spectra of undoped TiO2, 0.15, 0.3, 0.45 and 0.6 mol% Ru-doped TiO2 nanomaterials. As can be seen from Figure 4a the absorption peak of undoped TiO2 nanomaterial appears in the UV region whereas there is a slight red shift in the absorption spectrum of Ru-doped TiO2 and this red shift is found to increase with the increase in Ru-doping indicating the bandgap narrowing due to the introduction of a mid-bandgap or impurity levels located between the valence band and the conduction band of TiO2. In addition, the light absorption in the range from 400 to 700 nm is found to increase with increasing Ru content in the Ru-doped TiO2, accompanied with a change in the colour from white to reddish black [41]. The bandgap (Eg) of undoped TiO2, 0.15, 0.3, 0.45 and 0.6 mol% Ru-doped TiO2 nanomaterials was calculated using Tauc plots, where the intercept of the tangent to the plot (αhυ) 2 versus hυ gives a good estimation of the direct bandgap for a semiconductor. The optical absorbance coefficient of a semiconductor for direct transition is given by the equation where hυ = photon energy, α = absorbance coefficient, Eg = bandgap energy, A = constant and the exponent 'n' depends on the type of transition and it may have values of 1/2, 2, 3/2 and 3, corresponding to the allowed direct, allowed indirect, forbidden direct and forbidden indirect transitions, respectively. As shown in the Figure 4b, Figure 5 illustrates the elemental analysis of the undoped and Ru-doped TiO2 samples, studied using energy-dispersive X-ray spectroscopy in the binding energy region of 0.0-20.0 KeV and the results are summarized in Table 1, which reveals the existence of Ti, O and Ru elements in Ru-doped TiO2.  Figure 5 illustrates the elemental analysis of the undoped and Ru-doped TiO 2 samples, studied using energy-dispersive X-ray spectroscopy in the binding energy region of 0.0-20.0 KeV and the results are summarized in Table 1, which reveals the existence of Ti, O and Ru elements in Ru-doped TiO 2 .

Energy-Dispersive X-ray Spectroscopy
Energies 2020, 13 Figure 6a represents the J-V characteristics of DSSCs (in each type four devices were made and readings of the champion cells have been reported) made of photoanodes containing TiO2 electrodes doped with different mol% of Ru and its control under simulated irradiation intensity of 100 mWcm −2 with AM 1.5 filter. The corresponding photovoltaic parameters such as the short circuit current density (JSC), open-circuit photovoltage (VOC), fill factor (FF) and power conversion efficiency (η) of these cells with Ru-doped TiO2 electrodes and undoped TiO2 electrode are summarized in Table 2. Moreover, Figure 6b indicates the influence of different Ru mol% dopant on JSC and η. Ru-doping produces little improvement in VOC and FF. The control cell showed VOC of 0.66 V, which slightly increased to 0.69 V when 0.3 mol% Ru was doped. When the mol% of Ru dopant increases, Jsc shows a slight increase from 12.9 to 14.42 mAcm −2 up to 0.15 mol% Ru and then it attains a maximum of 14.73 mA/cm 2 for 0.3 mol% Ru, subsequently, the JSC values show a downward trend with further increase in Ru mol%. A similar trend was reflected in the η versus Ru mol% plot. The overall efficiency of the cell is mainly influenced by JSC. Similar observations have been reported by So and co-workers (2012) in DSSCs with Ru-doped TiO2 nanotubes [32]. In our study, the cell fabricated with 0.3 mol% Ru-doped TiO2 electrode showed the best η of 7% which is over 20% enhancement relative to undoped TiO2 based DSSC (η = 5.78%).  Table 2. Moreover, Figure 6b indicates the influence of different Ru mol% dopant on J SC and η. Ru-doping produces little improvement in V OC and FF. The control cell showed V OC of 0.66 V, which slightly increased to 0.69 V when 0.3 mol% Ru was doped. When the mol% of Ru dopant increases, Jsc shows a slight increase from 12.9 to 14.42 mAcm −2 up to 0.15 mol% Ru and then it attains a maximum of 14.73 mA/cm 2 for 0.3 mol% Ru, subsequently, the J SC values show a downward trend with further increase in Ru mol%. A similar trend was reflected in the η versus Ru mol% plot. The overall efficiency of the cell is mainly influenced by J SC . Similar observations have been reported by So and co-workers (2012) in DSSCs with Ru-doped TiO 2 nanotubes [32]. In our study, the cell fabricated with 0.3 mol% Ru-doped TiO 2 electrode showed the best η of 7% which is over 20% enhancement relative to undoped TiO 2 based DSSC (η = 5.78%).

Electrochemical Impedance Spectroscopy
The interfacial charge transport phenomena of the DSSCs can be studied using electrochemical impedance spectroscopy (EIS). Figure 7 shows the Nyquist plots of the electrochemical impedance spectra of DSSCs based on the control and 0.15, 0.3, 0.45 and 0.6 mol% Ru-doped TiO 2 photoanodes, which were measured at frequencies from 10 −2 to 10 6 Hz in the dark, with a bias applied voltage of 10 mV [24]. The small semicircle in the high-frequency range corresponds to the charge transfer resistance (R 1 ), which is related to the charge transfer at the interface of the electrolyte/Pt counter electrode and FTO/TiO 2 interface. A larger semicircle in the low frequency region is mainly related to the charge recombination resistance (R 2 ) across the TiO 2 /electrolyte interface with a partial contribution from electron transport and accumulation in TiO 2 photoanode [44]. R 1 and R 2 values can be estimated from the diameter of the semicircles and resistance related to electron recombination (R 2 ) increases with Ru-doping up to 0.3 mol% and then it starts to decrease with the increase in concentration of Ru. The higher values of resistance related to electron recombination indicate reduced electron recombination in Ru-doped electrodes [45]. This is the reason for the higher Jsc values for the Ru-doped devices in the J-V measurement [32] even though increased recombination resistance and reduced recombination rate improve the Voc of the Ru-doped devices, this can also be attributed to the reduction in the bandgap as well [15]. A similar observation was reported by Wang et al. in perovskite solar cells [29] and also Ismael et al. (2019) reported that Ru-doping could facilitate the separation and migration of photogenerated electron-hole pairs [41].

Electrochemical Impedance Spectroscopy
The interfacial charge transport phenomena of the DSSCs can be studied using electrochemical impedance spectroscopy (EIS). Figure 7 shows the Nyquist plots of the electrochemical impedance spectra of DSSCs based on the control and 0.15, 0.3, 0.45 and 0.6 mol% Ru-doped TiO2 photoanodes, which were measured at frequencies from 10 −2 to 10 6 Hz in the dark, with a bias applied voltage of 10 mV [24]. The small semicircle in the high-frequency range corresponds to the charge transfer resistance (R1), which is related to the charge transfer at the interface of the electrolyte/Pt counter electrode and FTO/TiO2 interface. A larger semicircle in the low frequency region is mainly related to the charge recombination resistance (R2) across the TiO2/electrolyte interface with a partial contribution from electron transport and accumulation in TiO2 photoanode [44]. R1 and R2 values can be estimated from the diameter of the semicircles and resistance related to electron recombination (R2) increases with Ru-doping up to 0.3 mol% and then it starts to decrease with the increase in concentration of Ru. The higher values of resistance related to electron recombination indicate reduced electron recombination in Ru-doped electrodes [45]. This is the reason for the higher Jsc values for the Ru-doped devices in the J-V measurement [32] even though increased recombination resistance and reduced recombination rate improve the Voc of the Ru-doped devices, this can also be attributed to the reduction in the bandgap as well [15]. A similar observation was reported by Wang et al. in perovskite solar cells [29] and also Ismael et al. (2019) reported that Ru-doping could facilitate the separation and migration of photogenerated electron-hole pairs [41].

Conclusions
In the present study, Ru-doped TiO2 electrodes were fabricated by treating TiO2 with RuCl3.XH2O and by systematically varying the Ru content from 0.15% to 0.6%. The XRD pattern of Ru-doped TiO2 nanomaterial confirms the presence of mixed anatase and rutile phases of TiO2.

Conclusions
In the present study, Ru-doped TiO 2 electrodes were fabricated by treating TiO 2 with RuCl 3 ·xH 2 O and by systematically varying the Ru content from 0.15% to 0.6%. The XRD pattern of Ru-doped TiO 2 nanomaterial confirms the presence of mixed anatase and rutile phases of TiO 2 . Optical absorption spectra of pure TiO 2 and Ru-doped TiO 2 reveal a slight red shift in the absorption spectrum upon Ru-doping. Among the DSSCs fabricated, the cell with 0.3 mol% Ru-doped TiO 2 electrode exhibited an optimum efficiency which is over 20% enhancement when compared to the control cell under stimulated AM 1.5 filter (100 mWcm −2 , 1 sun). The EIS analysis of the cells confirms that charge recombination resistance is significantly increased upon Ru-doping that effectively suppresses the charge recombination rate, which results in better electron transport in the cell.