Copper doped manganese dioxide as counter electrode for dye-sensitized solar cells

1 Introduction

Photovoltaic is truly a sustainable and environmentally friendly method of producing energy [Kiran et al., 2016; Zulkifli et al., 2015]. Dye-sensitized solar cells (DSSCs) are alternatives to silicon solar cells with the advantages of higher efficiency in diffused light, better cost-effectiveness, and easy fabrication [Chiba et al., 2006]. DSSC has four major parts: photoanode, sensitizer, redox electrolyte, and photocathode. The typical choice of photocathode or counter electrode (CE) material is platinum (Pt) due to its high catalytic activity and excellent stability toward the iodide redox species [Kim et al., 2006]. Since it is a costly noble metal and retreating supply, other materials have been under search to substitute Pt [Thomas et al., 2014]. Conducting polymers, carbonaceous materials, and transition metal compounds are some of the alternative materials introduced for the fabrication of DSSCs [Thomas et al., 2014].

It should have an energy level that matches the electrolyte potential and possess enhanced electrocatalytic activity, good reflectivity, and high surface area. Besides transparency, good reflectivity of CE is needed for illumination on the photoanode side. The electrochemically active surface of the CE should also preferentially be porous, have an optimum thickness, have chemical and electrochemical stability, be resistant to corrosion, have good adhesivity with the substrate, be low-cost, and be environmentally friendly. Atli and Yildiz (2022) reported a power conversion efficiency of 5.06% even with opaque Pt film assigned as a CE compared to transparent Pt CE (4.31%). This is because the high coverage (homogeneity) occurs in the case of transparent Pt CE due to the smaller particle size leading to an increase in the possibility of light interactions and reflections. [Atli and Yildiz, 2021; Atilgan and Yildiz, 2022].

Noticed, wide studies have been reported towards conducting polymers (poly(3,4-propylenedioxythiophene), polypyrrole (PPy), porous poly(3,4-propylenedioxythiophene), polythiophene (PTs), etc.,) as cathodes in DSSCs due to their good conductivity, and environmental stability [Jeon et al., 2011; Huang et al., 2012; Kim et al., 2011]. Likewise, carbon families (activated carbon, carbon nanotubes, and graphene) show exceptional catalytic activity towards the redox species upon being used as CEs in DSSCs [Murakami et al., 2008; Nam et al., 2010].

Transition metal carbides (MoC, VC, WC, NbC, TiC, Cr₃C_2, and Ta₄C₃) may act as smart CE materials because of superior catalytic activity, high electrical conductivity, and good thermal stability properties [Wu et al., 2012; Yun et al., 2013; Paranthaman et al., 2016]. Cai et al., 2017 fabricated DSSCs using copper-doped iron carbide (0.75%) as CEs resulting in the highest energy conversion efficiency of 5.68%. Such efficiency may be due to the excellent conductive properties of uniformly dispersed copper over iron carbide, which facilitates high electron transfer. Noble metal-like behavior of transition metal nitrides (TiN, MoN, WN, VN, NbN, CrN, NiN, Mo₂N, W₂N, Fe₂N, and Ta₄N₅) due to the electronic structure similarity renders them as CE materials in DSSCs [Furimsky 2003; Kang et al., 2015; Zhang et al., 2012; Jiang et al., 2009]. Also, metal chalcogenides of various compositions were introduced as CEs show enhanced photovoltaic performance than ubiquitous platinized cathodes [Olsen et al., 2000]. Wu et al. fabricated DSSCs with MoS₂ and WS₂ as CEs that results in photoconversion efficiency of 7.59 % and 7.73 %, respectively [Wu et al., 2011]. Yildiz et al. reported Iron Phosphide (FeP) as CEs generates an efficiency of 3.96% which is comparate to the Pt CE. Besides, DSSCs are fabricated with transition metal oxides such as WO₂ and NbO₂ as CE generates photoconversion efficiency of 7.25% and 7.88% found close to DSSC with Pt CE (7.57%) [Zhou et al., 2013; Lin et al., 2011]. Ahmad et al., 2017 have synthesized α-MnO₂ nanorods by facile hydrothermal method and utilized them as CE in DSSCs resulting in an open-circuit voltage of 0.75 V and fill factor (0.38). Likewise, MnO₂-coated carbon nanofiber was used as CE which resulted in higher open-circuit voltage (0.78 V), fill factor (0.68), and PCE (8.86%). In continuation, various researchers coupled MnO₂ with NiO and TiO₂ layer-by-layer architecture and utilized them as CEs for DSSCs which shows higher efficiency compared to the pristine MnO₂ layer [Kakroo et al. 2020, Datta et al. 2020]. Copper having excellent conductive properties upon doped on MnO₂ results in enhanced electrochemical behaviour [Li et al., 2013; Ding 2010; Davis et al., 2014; Gao et al., 2018; Julien et al., 2017; Wei et al., 2011; Kakroo et al., 2020]. Considering the various choices of materials, the search for finding a low-resistance and low-cost alternative to Pt still needs more research [Thomas et al., 2014]. The promising results from doped metal oxides have motivated us to further study MnO₂ and Cu doped MnO₂ as CEs in DSSCs, particularly focusing on the effect of doping level on the performance of the MnO₂ electrode as a CE for DSSCs. This is because the transition metal oxide MnO₂ has certain advantages such as low-cost, high natural abundance, high theoretical capacity (1370 Fg⁻¹), high voltage window, etc. Since its capacitance and conductivity are low, cationic doping has been proved to be effective to improve the conductivity and of which copper cations are found to improve electrochemical performance far better compared to other cations [An et al. 2019]. Further, the conductivity is based on electron concentration and mobility of electrons and hence we preferred to choose low dopant ratios (1, 5, 10 %wt). Hence in this study, MnO₂ and copper doped MnO₂ nanoparticles were prepared by hydrothermal approach and investigated as CEs in DSSCs.

2 Experimental methods

135 ml deionized water (DI) was added to potassium permanganate (KMnO₄) (1185.25 mg) slowly as the solution was stirred until it was dissolved completely. Then, different weight percentages (1, 5, 10 %) of copper acetate were added until it was fully dissolved. Afterward, the mixture was acidified by adding 2.5 ml of HCl. The whole mixture was transferred to a Teflon-lined autoclave and kept at 140 °C for 14 h. Finally, the product (Cu doped MnO₂) was regenerated by washing with DI water, and ethanol and dried at 80 °C for 24 hrs. A similar procedure was followed in the absence of copper acetate to generate pristine MnO₂ [Luo et al., 2008].

Refer to supporting information regarding a list of equipment used to characterize the prepared nanomaterials and for the experimental procedure followed towards the fabrication of solar cell devices.

3 Results and discussions

As shown in Fig. 1(A), the XRD pattern of the synthesized pristine MnO2 and 1, 5, & 10 wt% copper doped MnO₂ nanoparticles by the hydrothermal method exhibited a well-defined diffraction pattern which is in good harmony with the testified crystalline α-MnO2 (JCPDS data card no. 44–0141). [Ahmad et al., 2017; Yang et al., 2019] The diffraction peaks of both the samples appear at (12.6°), (18.03°), (28.5°), (37.4°), (41.5°), (49.8°), (55.9°), (59.8°), and (65.3°), are corresponding to the (1 1 0), (2 0 0), (3 1 0), (2 1 1), (3 0 1), (4 1 1), (6 0 0), (5 2 1), and (0 0 2) planes of α-MnO2 respectively [Ahmad et al., 2017]. XRD patterns of Cu doped MnO₂ samples were also studied and results without any alterations in peaks belonging to α-MnO₂ were observed. Since the copper species are highly homogeneous and well dispersed; hence no characteristic peak of Cu or CuO has been seen even for the 10 wt% Cu doped MnO₂ sample [Yang et al., 2019]. The calculated crystallite size (using Scherrer formula) of 10 wt% Cu doped MnO₂ is found smaller (23 nm) compared to the pristine MnO₂ sample (27 nm).

Close

Fig. 1

(A) XRD patterns and (B) FTIR spectra of MnO₂ and 1, 5, & 10 wt% Cu doped MnO₂ nanoparticles.

Export to PPT

Similar to XRD, both MnO₂ and 1, 5, & 10 wt% Cu doped MnO₂ show the same IR stretching peaks upon the analysis in the Fourier Transform-Infrared (FTIR) spectrometer (Fig. 1(B)). The stretching and bending vibrations of the hydroxyl groups and the hydrogen-bonded surface water molecules are found at 3438 cm⁻¹ and 1629 cm⁻¹ [Yang et al., 2019; Mondal et al., 2018; Mondal et al., 2019] The Mn-O vibration band is seen near 700 cm⁻¹ which further supports the formation of MnO₂. Such observed results are due to two factors: (i) the amount of doping is a trace (even increased to 10%) and (ii) due to the same crystal size Cu doping can substitute manganese ions (results peak broadening). [Mondal et al., 2018].

To determine the optical energy bandgap of MnO₂ and Cu doped MnO₂ nanoparticles, the UV–Vis absorption spectrum was recorded. All the samples show a broad absorption band starting from 800 nm onwards which can be recognized due to d-d transitions of manganese ions (Figure not provided). The photoexcited electron from the valence band to the conduction band takes place due to Mn 3d energy level splitting [Mondal et al., 2018; Mondal et al., 2019]. Converting the photoabsorption data to the Tauc plots in Fig. 2, the optical energy bandgap (E_g) was predicted indirectly from the plots by using the following formula: $$\mathit{Eg}=\frac{1240}{\mathrm{W}\mathrm{a}\mathrm{v}\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{n}\mathrm{g}\mathrm{t}\mathrm{h}}$$

Close

Fig. 2

Tauc plot of MnO₂ and 1, 5, & 10 wt% Cu doped MnO₂ nanoparticles.

Export to PPT

Pristine MnO₂ shows a bandgap value of about 1.67 eV and upon doping with copper (1, 5, 10 wt%) noticed bandgap get lowers to 1.6, 1.49, and 1.43 eV which clearly illustrates that light absorption got enhanced with Cu dopant (Fig. 2). In addition, such copper dopants may suppress the recombination rates at the electrode/electrolyte/dye interfaces [Atilgan and Yildiz, 2022].

Both pure MnO₂ and Cu doped MnO₂ looked semispherical with coarse surface morphology under FE-SEM analysis (Fig. 3 a-d). Compared with pristine MnO_2, Cu doped MnO₂ has a slight difference in morphology and not noticed any lump which proved no agglomeration of particles. Such a larger particle size increases the scattering of light and thus allows the way light travels inside the device [Atli and Yildiz, 2022; Atilgan and Yildiz, 2022]. Besides it favors the penetration of the electrolyte and hence occurs quick electron transfer between the mediators (iodide/triiodide) in the electrolyte and the CE [Yang et al., 2019; Vijayakumar et al., 2015]. The EDX elemental analysis confirms the existence of individual metals only in the prepared pure MnO₂ (Mn-67.21; O-32.78) and 10 wt% Cu doped MnO₂ (Cu-9.12; Mn-55.22; O-35.66) samples (Fig. 3 e,f). It is found that the observed EDX results are almost matching with the 10 wt% Copper dopant amount taken for the experiment. In TEM analysis (Fig. 3 g,h,j,k), also, samples show similar spherical morphology as seen in the SEM images. The selected area electron diffraction patterns also illustrate the samples were well crystallized in nature. (Fig. 3 i,l).

Close

Fig. 3

FE-SEM micrographs of (a-b) MnO_2, (c-d) 10 wt% Cu- MnO₂ and (e,f) EDS images of MnO₂ and 10 wt% Cu- MnO₂ nanoparticles. TEM micrographs of (g,h) MnO_2, (j,k) 10 wt% Cu- MnO₂ and (i,l) SAED images of MnO₂ and 10 wt% Cu- MnO₂ nanoparticles.

Export to PPT

The cyclic voltammograms (CVs) of MnO₂ and copper doped MnO₂ thin layer electrodes (Fig. 4) look sigmoidal and symmetrical in shape due to the slow ion transport that may stimulate the fabricated DSSCs towards the effect of sunlight and as well as avoid back electron transfer. As can be observed from CV loops, the device having the Pt shows the highest I_PC and E_PP value (4.80 mAcm⁻² and 1000 mV) which indirectly indicates the good electrocatalytic activity of Pt CEs toward the liquid electrolyte [Yildiz et. al. 2021]. Again, the higher I_PC of Pt suggests a better electrocatalytic activity which may be reflected in the device’s performance. No certain oxidation and reduction peaks for Cu doped MnO₂ samples, hence it is not possible to determine I_PC and E_PP values. Since the prepared Pt-coated electrode showed a symmetric response with the two distinct peaks due to fast electron transfer from the counter electrode to the electrolyte (reduction of iodine/iodide electrolyte: I₃^- + 2e^- → 3I^-) [Vijaya et al., 2021].

Close

Fig. 4

CV curves of MnO₂ and 1, 5, & 10 wt% Cu doped MnO₂ Nanoparticles. (Inset shows CV for Pt electrode for comparison).

Export to PPT

Cu doped MnO₂ coated FTO glass plates were sandwiched one above the other for symmetrical dummy cell fabrication towards electrochemical impedance spectroscopy (EIS) measurements. Parafilm was used to avoid contact with each other. The electrolyte (an iodide/triiodide (I^-/I₃^-) solution) consisting of LiI/I₂ (0.05 M/0.01 M), and LiClO₄ (0.5 M) dissolved in 3-methoxy-propionitrile was injected in between the two glass plates. The electrochemical impedance spectroscopy (EIS) studies for such thin layer symmetrical electrodes were carried out under the following experimental conditions (frequency − 0.1 to 10⁵ Hz; AC voltage − 10 mV). The Nyquist plot (Fig. 5) is projected against real impedance (Z') vs imaginary impedance (-Z“) gives information about the resistance data at TiO₂/dye/electrolyte and electrolyte/CE. Also included equivalent circuit model in the inset of Fig. 5. The electron-hole charge transfer, diffusion of ions, and electron recombination at the TiO₂/dye/electrolyte interface are related to resistance. The series/bulk resistance (R_S) represents the FTO plates resistance along with the resistance of the counter electrodes can be investigated from the origin of the semicircle (the high-frequency region). The diameter of the semicircle of the mid-frequency region represents the charge transfer resistance (R_CT) indicates the I^-/I₃^- reduction catalytic activity [Wu et. al. 2017]. The response at low frequency region is attributed to Warburg resistance, Z_w of I₃^-, tri-iodide diffusion [Wang et al. (2015)]. Further, the EIS parameters extracted from the Nyquist plot (Fig. 5) are listed in Table 1. From this table, the R_S values of the respective CEs follow the order: 1 wt% Cu- MnO₂ CE (38.62 Ω cm²) > 10 wt% Cu- MnO₂ CE (35.55 Ω cm²) > Pt CE (31.97 Ω cm²) > MnO₂ CE (30.29 Ω cm²) > 5 wt% Cu- MnO₂ CE (29.87 Ω cm²). The R_S values of the Pt (31.97 Ω cm⁻²) is less than 10 wt% Cu doped MnO₂ (35.55 Ω cm⁻²) confirming the superior electrical conductivity of the Pt electrode. Further, the 1 wt% Cu doped MnO₂ shows the lowest R_CT value of 157.95 Ω cm⁻² compared to Pt (191.53 Ω cm⁻²) and 10 wt% Cu doped MnO₂ (204.9 Ω cm⁻²). The better electrocatalytic activity of Cu doped MnO₂ toward the liquid iodide/triiodide electrolyte is confirmed by these EIS results.

Close

Fig. 5

Nyquist plots of symmetrical dummy cells fabricated with Pt, MnO₂, 1 wt% Cu-MnO₂, 5 wt% Cu-MnO_2, and 10 wt% Cu-MnO₂ CEs. Inset see equivalent circuit of CEs.

Export to PPT

Also, Bode phase plots for thin layer electrodes were measured to calculate the electron lifetimes (Fig. 6) from the equation. $$\tau =\frac{1}{2\mathrm{\Pi }\mathrm{f}}$$ where τ - electron lifetime and f - frequency [Sarker et al., 2014]. The calculated electron lifetime for Pt, MnO₂, 1 wt% Cu-MnO₂, 5 wt% Cu-MnO_2, and 10 wt% Cu-MnO₂ CEs dummy cells are found to be 99, 32, 43, 57, and 75 ms. The peak noticed in the high-frequency region is correlated to the charge transfer that occurs at the electrode/electrolyte interfaces [Kim et al., 2012]. Copper-doped MnO₂ samples are found to have a phase shift lower than undoped samples indicating lower resistance for pristine MnO₂ samples.

Close

Fig. 6

Bode phase plots of dummy cells fabricated with Pt, MnO₂, 1 wt% Cu-MnO₂, 5 wt% Cu-MnO_2, and 10 wt% Cu-MnO₂ CEs.

Export to PPT

To determine exchange current density (current at zero overpotential), Tafel polarization experiments were conducted for thin layer electrodes. Tafel polarization of the dummy cells has been plotted to further analyze the electrocatalytic behaviour of CEs. For that, dummy cells have been fabricated using two identical electrodes coated with Pt and Cu doped MnO₂ samples by plotting logarithmic current density as a function of voltage. The higher exchange current density (J₀) has been observed (Fig. 7, Table 1) for the Pt CE (1.17 mAcm⁻²) which is well supported by the lower R_CT value compared to that of the 10 wt% Cu doped MnO₂ CE (0.055 mAcm⁻²) indicating better electrocatalysts for good reaction kinetics.

Close

Fig. 7

Comparison of Tafel polarization curve of Pt with respect to MnO₂, 1 wt% Cu-MnO₂, 5 wt% Cu-MnO_2, and 10 wt% Cu-MnO₂ CEs.

Export to PPT

The fabricated solar cells of type FTO/TiO₂/N3dye/liquid electrolyte/copper doped MnO₂/FTO were analyzed under 0.85 Sun (85 mW cm⁻², AM 1.5 G solar simulator; Refer to supporting information for experimental procedure toward the fabrication of solar cells). The open-circuit voltage (V_oc), short-circuit current (I_sc), fill factor (FF), and PCE (η) was calculated from the measured J-V curves (Fig. 8) using the AUTOLAB12/FRA2 PGSTAT302N model.

Close

Fig. 8

I-V characteristics of DSSCs fabricated with Pt, MnO₂, 1 wt% Cu-MnO₂, 5 wt% Cu-MnO_2, and 10 wt% Cu-MnO₂ CEs.

Export to PPT

The device fabricated with pristine MnO₂ CE showed V_OC of 724 mV, I_SC of 2.71 mA/cm^2, FF of 0.53, and a power conversion efficiency of 1.21%. 1 wt% Cu doped MnO₂ showed enhanced % PCE of 1.4 compared with MnO₂ CE with V_OC of 726 mV, I_SC of 3.05 mA/cm^2, and improved FF of 0.54. 5 wt% Cu doped MnO₂ CE showed V_OC of 754 mV, I_SC of 3.17 mA/cm², FF of 0.52, and % PCE of 1.47. 10 wt% Cu doped MnO₂ CE showed the best V_OC of 781 mV, I_SC of 3.69 mA/cm², FF of 0.50, and % PCE of 1.7. The DSSC fabricated with Pt CE showed V_OC of 780 mV, I_SC of 6.8 mA/cm², FF of 0.43, and % PCE of 5.83. These results highlight that Cu doped MnO₂ is far better catalytic activity compared to pristine MnO₂. (Table 2).

The Nyquist plots of fabricated DSSCs with Pt, MnO₂, 1 wt% Cu-MnO₂, 5 wt% Cu-MnO_2, and 10 wt% Cu-MnO₂ CEs were shown in Fig. 9(A) (Table 2). The doped samples were found to have lower R_CT than pure MnO₂ samples indicating that the resistance for charge transfer got reduced. A lower R_CT value illustrates less overpotential is required for the electron transport between the electrolyte and CE indicating a higher tri-iodide reduction (I^-/I₃^-) and hence high fill factor.

Close

Fig. 9

Nyquist (A) and Bode phase (B) plots (under dark) of DSSCs fabricated with Pt, MnO₂, 1 wt% Cu-MnO₂, 5 wt% Cu-MnO_2, and 10 wt% Cu-MnO₂ CEs.

Export to PPT

In DSSCs, after electron injection from dye molecule into TiO₂, there is three possible recombination reactions with the electron acceptors viz. oxidized dye molecule, oxidized electrolyte species, and electron scavenger contaminants in the system. But the recombination reaction of the electron with the oxidized dye molecule is often least considered because this recombination is found to be slower (in the range of 10^-4 s) than the regeneration of the dye molecule (10^-6 s) by the iodide ion [Bisquert et al. (2004)]. Hence, the primary recombination reaction is between TiO₂ and the oxidized ions in the electrolyte solution. Here, the electron lifetime is calculated using the formula τ = 1/2πf, where the frequency is obtained from the Bode plot under illumination conditions at open circuit voltage. The Bode phase plot of the fabricated DSSCs with Pt, MnO₂, 1 wt% Cu-MnO₂, 5 wt% Cu-MnO_2, and 10 wt% Cu-MnO₂ CEs was shown in Fig. 9(B). From the Bode phase plot, the electron lifetime can be calculated which illustrates Cu doped MnO₂ samples are better than pristine MnO₂ (Table 2). The device fabricated with Pt CE showed the lower electron lifetime (0.18 ms) compared to 10 wt% Cu doped MnO₂ CE (0.79 ms) suggests decreased electron transfer at TiO₂ layer makes Pt CE less efficient for dye regeneration. While, the longer electron lifetime of Cu doped MnO₂ CE strongly supports the efficient dye regeneration due to fast electron transfer evidencing the superior electrocatalytic activity. The device fabricated with Pt CE showed the lowest R_CT value of 96.02 Ω cm⁻² suggesting the fast electron transfer process at CE/electrolyte interface.

Further, a comparable table reflection the improvement with a novel CE are presented in Table 3 illustrate that the initial results of copper doped MnO₂ are found satisfactory. This is because in the case of copper doped MnO₂, the added copper ion leads to increase the amount of Mn³⁺ at the surface and in turn improves the catalytic activity. Besides, added copper dopant increased the covalent nature of Mn-O bonds and results in stabilization and faster kinetics rate of electrode. All these factors may responsible for the improvement of efficiency upon compared to pristine MnO₂.

4 Conclusions

MnO₂ and copper doped MnO₂ nanoparticles have been successfully produced by the hydrothermal method. The formation of the tetragonal crystal structure of α-MnO₂ and its copper doped versions was established by XRD analysis. The electrochemical characterization of MnO₂ and copper doped MnO₂ nanoparticles show the electrocatalytic property of the respective electrodes toward the reduction of I₃^- ions. Even though the samples do not show the distinct oxidation and reduction peaks in the CV study, the DSSC devices fabricated with MnO₂ and copper doped MnO₂ CEs behave like diodes under dark (no light is present). Among the fabricated DSSCs 10 wt% Cu doped MnO2 cathodes yield the highest energy conversion efficiency of 1.7% than the pristine MnO2 which had only 1.21% PCE. Further study regarding the choice of a material having electrocatalytic activity as good as Pt may lead to the fabrication of low-cost Pt-free DSSCs.