Transformation of ZnS microspheres to ZnO, their computational (DFT) validation and dye-sensitized solar cells application Open Access

Dye-sensitized solar cells (DSSCs) have been paid widespread attention from the last few years as a facile renewable energy source, owing to depletion of non-renewable energy sources. Moreover, DSSCs are proposed as a low-cost alternative to the conventional photovoltaics.^1–4 DSSCs are fabricated by using wide bandgap semiconductor materials as photoanodes, such as TiO₂, ZnO, SnO₂, and Nb₂O₅, for advanced outcome of DSSCs.^5–8 Among these, though TiO₂ is still in demand, ZnO due to its excellent electron mobility and wide range of morphological variation shows the promising candidature as photoanode materials in DSSCs. ZnO is a unique material, which has semiconducting as well as piezoelectric properties.^9–11 It is a versatile oxide semiconductor possessing the bandgap (direct bandgap: 3.37 eV) with 60 meV of high free extinction binding energy at room temperature.^12,13 Various methods, such as sol–gel, hydrothermal, solvothermal, chemical vapor deposition (CVD), electrophoretic deposition, vapor–liquid–solid (VLS) method, and thermal decomposition, are reported for synthesis of ZnO nanostructures to obtain different morphologies for enhancing surface properties and surface area.^14–25 The solid state oxidation reaction is a very simple but least explored method for the synthesis of ZnO nanostructures.

Here, we synthesized nanostructured ZnO microspheres by solid state oxidation of ZnS microspheres, which were synthesized by using the hydrothermal method in our previous work on photocatalytic properties of ZnS microspheres.²⁶ So far, conversion of ZnS to ZnO has been reported very few times. Mahmood et al.²⁷ converted cubic ZnS to hexagonal ZnO under a high temperature oxidation process in which ZnS powder was thermally evaporated and deposited on a Si (001) substrate. The developed ZnS films were oxidized at different temperatures from 700 to 1000 °C.²⁷ Wahab et al.²⁸ reported the change of ZnS to ZnO nanoparticles as a function of oxidation temperature (300–900 °C). In their reaction, zinc acetate dihydrate and thiourea precursors in water were refluxed at 90 °C for over 12 h to synthesize ZnS.²⁸ After refluxing, the as-received samples were then subjected to heat treatment at 300, 500, 700, and 900 °C in atmospheric air for 1 h to obtain ZnO nanoparticles. In accordance with these very few available reports, we oxidized (in an ambient atmosphere) previously prepared highly porous ZnS microspheres for the present investigation and such synthesis protocols have never been used for fabrication of photoanodes in DSSCs. Accordingly, we have synthesized ZnO microspheres in order to achieve better surface roughness, high electrical conductivity, light scattering, and porosity for the enhanced performance in DSSC fabrication. Hexagonal ZnO is highly efficient as a photoanode material in DSSCs compared to the other crystalline form (zinc blend and cubic). At annealing temperatures above 500 °C, the cubic structure transforms to a hexagonal structure, and further annealing improves its crystalline nature. Owing to this, ZnS microspheres were heated at various temperatures of 600 and 700 °C to transform into hexagonal ZnO microspheres with high crystallinity so as to prepare highly efficient photoanodes for DSSCs. Therefore, a comprehensive study has been reported here for both batches of ZnO microspheres (synthesized at 600 and 700 °C) in order to study the outcome of high temperature oxidation on the structural and optical properties [further validated by the density functional theory (DFT) studies] as well as on the photovoltaic performance.

The analytical grade precursors and reagents were used as such without any further purification or processing for the reactions.

Zinc acetate (ZnC₄H₆O), thiourea (CH₄N₂S), and ammonia (25% NH₃ solution) received from S.D. Fine, India, were used without further purification. The conductive glass plate made up of fluorine tin oxide (FTO) TEC8 (8 Ω/□), N719 dye, cell electrolyte EL-HPE (high performance electrolyte), and platinum paste (PT1) used for the counter electrode were purchased from Dyesol. The packing of sandwich type assembly of both electrodes was ended by a sealing agent (SX1170-60, 50 µm) obtained from Solaronix.

In our previous work, ZnS microspheres were synthesized by hydrothermal techniques with different times of reaction as 6, 12, and 24 h and these powders were labeled GHZ1, GHZ2, and GHZ3, respectively.²⁶ The obtained end product set of three ZnS microsphere powders were used as the precursor in the present work. Typically, ZnS microsphere powders 200 mg from each set as the first batch were heated at 600 °C for 1 h in a muffle furnace in atmospheric air with the heating rate of 4 °C/min. A similar heat treatment was also carried out at 700 °C to the other batch of powder from the same set. Accordingly, the GHZ1, GHZ2, and GHZ3 samples were further named as GHZ61, GHZ62, and GHZ63 and GHZ71, GHZ72, and GHZ73 based on heat treatment at 600 and 700 °C, respectively. The obtained ZnO powders were collected as such without any further processing for physico-chemical characterization as well as for DSSC photoanode fabrication.

The conductive glass substrates (FTO: fluorine doped tin oxide) were sequentially cleaned with de-ionized (DI) water, ethanol, and acetone and then dried out with nitrogen gas. ZnO powders, α-terpinol, ethanol, and ethyl cellulose were used to prepare ZnO paste. The ZnO paste was then used to prepare ZnO thin films on FTO by using doctor’s blade procedure and then annealed at 450 °C for 60 min at the rate of 4 °C/min in air. The thickness of each film was obtained to be ∼8 µm. The annealed electrodes were instantly dipped in dye solution (0.3 mM N719 dye in acetonitrile:tert-butyl alcohol) for 12 h followed by washing off non-adsorbed dye in ethanol. Concurrently, the platinum paste was used to prepare the counter electrode by screen printing it on another FTO plate and annealed in a tubular furnace at 400 °C for 30 min in air with 4 °C/min rate. The DSSC device was made up by adjusting both electrodes placed (the photoanode and the counter electrode) in a sandwich way by confronting each other leaving the space for contact to the external load. This sandwich device was sealed by a sealant from all side apart from one side through which the electrolyte solution was injected. The electrical contacts on both electrodes were made by copper wires with the help of silver paste and Araldite.

The XRD used to determine the crystalline structure of synthesized ZnO powders was done using a Bruker AXS model D-8 equipped with a Ni-filtered Cu-K_α radiation (λ = 1.54 Å) at 2ϴ varying from 20° to 80° with a scan speed of 4° min⁻¹ and a monochromator. A Field Emission Scanning Electron Microscope (FESEM) (Hitachi S-4800) was used to determine the surface morphology. A Shimadzu UV–Vis–NIR spectrophotometer (Model: UV-3600) was used to obtain the absorption spectrum of the synthesized samples. The photovoltaic performance: open-circuit voltage (V_oc, V), short circuit current (J_sc, mA/cm²), fill factor (FF), and conversion efficiency (ɳ) were calculated with the recorded J–V curve by using a digital Keithley 236SMU using a 1000 W/HS xenon arc lamp with a light intensity of 100 mW/cm² and 1.5 AM. A Xe lamp (300 W) equipped with a monochromator (Oriel) was used to calculate the incident photon-to-electron conversion efficiency (IPCE). The Si photodiode was used as a reference for calibration. The impedance spectroscopy [electrochemical impedance spectroscopy (EIS)] study was carried out in light and dark by applying a forward bias at −0.70 V using a potentiostat (VersaSTAT3) with a frequency range of 10⁻²–10⁵ Hz at room temperature at 10 mV alternating voltage amplitude at open-circuit potential (OCP) configuration.

The density functional theory (DFT) with generalized gradient approximation (GGA) was executed for electronic structure calculations^29,30 using Perdew–Burke–Ernzerhof (PBE) as exchange-correlation functional³¹ as executed in Quantum Espresso (QE).³² A plane wave basis set with an energy cutoff of 680 eV and charge density cutoff of 6122 eV is used. A 7 × 7 × 7 k-point mesh is sampled for the Brillouin zone with Monkhorst–Pack. The convergence criteria for force during unconstrained unit cell optimization was set to 25 × 10⁻³ eV/Å, and that for electronic self-consistency was 13 × 10⁻⁹ eV. It is well known that bandgaps obtained using PBE can be severely underestimated. Therefore, we computed the band structure using the Heyd–Scuseria–Ernzerhof (HSE)³³ functional using 20 k-points each on each symmetry lines in the Brillouin zone.

The electronic band structure changes with temperature mainly due to electron–phonon interactions. We modeled these interactions with the help of Zero Point Motion renormalization (ZPR)^34,35 using Allen–Heine–Cardona (AHC) theory^36–38 as implemented in ABINIT.^39,40 We used the PBE exchange-correlation functional and with 8 × 8 × 8 Monkhorst–Pack k-point grid with the kinetic energy cutoff to be 625 eV. The obtained corrections were then applied to the bandgap obtained using HSE calculations as noted before. The corrected renormalized bandgap at the Γ point was 2.56 eV at zero temperature. The DFT generated computational data were then correlated with the obtained experimental results.

ZnO powder synthesized by heating ZnS at 600 and 700 °C was subjected to x-ray diffraction to ascertain the conversion of ZnS into ZnO (Fig. 1). The indexed peaks for all the six samples prepared using both heat treatment cases (600 and 700 °C) exhibited wurtzite hexagonal nature of powder with (110), (002), (101), (102), (110), (103), (200), (112), (201), (004), and (202) planes, which matched with JCPDS file No.80-0074.

However, the peak sharpness was slightly increased in the case of ZnO samples obtained at 700 °C, which was obviously due to the increased grain size. Accordingly, the crystallite sizes of representative ZnO powders GHZ62 and GHZ72 were estimated to be 23 and 26 nm, respectively, by using the Scherrer equation. We further validated our data by comparing it with DFT calculations. Our calculations showed the zero-temperature ground state of ZnO is wurtzite hexagonal by the lattice constants a = 3.282 Å and c/a = 1.6133 (inset of Fig. 2). We computed the XRD spectrum for this structure, which is plotted in Fig. 2, and matched well with the obtained x-ray diffractograms.

The UV–Vis spectra in diffused reflectance mode (UV-DRS) of the synthesized ZnO powder are shown in Fig. 3. The absorption peak varied from 410 to 413 nm for ZnO powders heated at 600 °C [Fig. 3(a)]. Figure 3(b) shows the absorption peak of ZnO powder samples at 407, 406, and 397 nm for synthesized GHZ71, GHZ72, and GHZ73 samples, respectively, which were heated at 700 °C. Figure 4 exhibits the Tauc plot of ZnO samples, which were synthesized by heat treatment of ZnS microspheres. In Fig. 4(a), the Tauc plot of ZnO powders obtained at 600 °C revealed the bandgap varying from 3.02 to 3.06 eV for GHZ61 to GHZ63. The bandgap of ZnO obtained at 700 °C varied from 3.04, 3.05, and 3.12 eV for GHZ71, GHZ72, and GHZ73 samples, respectively [Fig. 4(b)]. We compared these bandgap values with those obtained with DFT calculations.

Figure 5 shows the band structures obtained by HSE functionals. A DFT study underestimates the bandgap as it is known behavior. The bandgap obtained by PBE potentials was found to be significantly underestimated (0.44 eV), but HSE calculations (2.54 eV) showed relatively good agreement with experiments.

The variations of bandgap are plotted using ZPR corrections in Fig. 6. At high temperatures of synthesis, i.e., 600–700 K, the bandgap changed from 2.69 to 2.71 eV, that is, by 0.02 eV, which is in good accordance with the experimental results. Rather, lower corrections to the bandgap suggest that the electron–phonon interaction is relatively weak.

Figure 7 shows the FESEM image of ZnO microspheres, which were synthesized from ZnS microspheres heated at 600 and 700 °C for 1 h. FESEM images of the obtained ZnO samples showed formation of spherical microspheres. Occasionally, the microspheres appeared to be fused to each other. These microspheres were made up of bunches of spherical and faceted nanoparticles, which were self-assembled to generate these hierarchical shapes. Figures 7(a-i)–7(f-i), corresponding to GHZ61, GHZ62, and GHZ63 samples (obtained by heating ZnS at 600 °C), revealed spherical microspheres having different sizes, which were produced from bunches of nanoparticles. The surface roughness of microspheres goes on increasing in the order GHZ63 > GHZ62 > GHZ61. However, it is noticed that in the case of sample GHZ62, porosity is more. The change in porosity may affect the dye loading and consequently the efficiency. The average particle sizes of GHZ61, GHZ62, and GHZ63 samples were found to be 62, 80, and 102 nm, respectively. In the case of the ZnO samples obtained at the heat treatment of 700 °C, namely, GHZ71, GHZ72, and GHZ73, the spherical morphology of the microspheres was obtained as shown in Figs. 7(g-ii)–7(l-ii). The surface roughness of microspheres was found to be increased gradually as GHZ73 > GHZ72 > GHZ71. Nevertheless, the increase in porosity did not follow the same trend. Quite interestingly, the porosity of the GHZ72 sample seemed to be better than the other two samples of the same temperature heating series (700 °C). The average particle size of GHZ71, GHZ72, and GHZ73 samples was found to be 65, 97, and 112 nm, respectively. It is interesting to note that the average microsphere size of the samples heated at 700 °C was found to be smaller than the samples heated at 600 °C. Yet, the surface roughness increased. The decrease in the microsphere size might be due to densification and particle growth of microspheres during the higher temperature processing. The surface particle growth increased the surface roughness of the microspheres, which might assist positively in the dye loading. At the same time, the decrease in the porosity due to densification would have the negative impact on the effective dye loading and affects the resultant photovoltaic performance of the solar cell. Based on the FESEM analysis, it was speculated that GHZ62 and GHZ72 may exhibit better solar cell characteristics owing to their improved surface roughness and porosity simultaneously as compared to other samples.

Based on the above observations, the role of heat energy and atmospheric oxygen on the ZnS sample could be easily understood. As the annealing temperature increased, the sulfur of ZnS reacted by atmospheric oxygen and produced sulfur dioxide gas as follows:

The zinc ions of ZnS then continually reacted by the accessible atmospheric oxygen from the furnace and produced ZnO as follows:

The final reaction equation is then expressed as

ZnO powder obtained by heat treatment in air (oxidation) of ZnS at different temperatures (600 and 700 °C) was exploited as the photoanode for DSSCs. Figure 8 shows the current–potential (J–V) curves of DSSCs under air mass 1.5 global (AM 1.5 G) illumination at 100 mW/cm² light intensity. Among the ZnO samples obtained at 600 °C (GHZ61, GHZ62, and GHZ63), GHZ62 showed better performance as compared to the other two. The open-circuit voltage V_oc of 0.606 V, fill factor (FF) of 0.58, and power conversion efficiency (η) of 3.38% were noted for the GHZ62 sample. Among the ZnO samples obtained at 700 °C, GHZ72 exhibited good performance as compared to GHZ71 and GHZ73 with an open-circuit voltage (V_oc) of 0.600 V, fill factor (FF) of 0.64, and power conversion efficiency (η) of 3.06%. The efficiency of GHZ62 is greater than GHZ72 because of: (i) maximum electron–hole generation due to high surface roughness and porosity resulting in better dye loading and (ii) better electronic conduction pathway. At 700 °C, the increase in the size of nanoparticles forming the microsphere structure of GHZ72 was noticed thus losing the porosity. The amount of dye loading influenced the exciton (electron–hole pair) generation, which consequently affected the cell efficiency of the device. The photovoltaic parameters are summarized in Table I.

The IPCE for the DSSC indicated the ratio of photons that generate electrons in the external circuit to the total incident photons of monochromatic light. The calculation of IPCE was accomplished by noting the values of short-circuit photocurrent (J_sc) as a function of the wavelength of monochromatic light in the range of 400–700 nm. Figure 9 corresponds to the IPCE curves for N719 dye-sensitized ZnO synthesized from single step heat treatment of ZnS at 600 and 700 °C.

The maximum value of IPCE was obtained at 532 nm (visible region), which is in line with the utmost solar radiation absorption (λ_max) value of the N719 dye. The IPCE values are shown in Table I. The maximum IPCE value was obtained for the DSSC fabricated using GHZ62 powder, which is owing to high surface roughness for more dye adsorption and better electronic conduction pathway, which together facilitate less recombination.

EIS elaborates the recombination and charge transfer properties for the interfaces in DSSCs.⁴¹ Figure 10 shows the Nyquist plot with equivalent circuit (inset) for the cell, in which the circuit was given to fitted according to the series resistance (R_s), charge transfer resistance (R_ct), and the equivalent constant phase angle element (CPE) in DSCs.^42,43 In EIS, the high and middle frequency zone semicircles of Nyquist plots are accredited to the oxidation–reduction reaction and electron transfer at the Pt counter electrode and that at the ZnO/dye/electrolyte interfaces, respectively. In the middle frequency region, the large semicircle represented the charge transfer resistance of ZnO to the electrolyte (R_rec) and the resistance decreased with decreasing semicircle diameter for the ZnO/dye/electrolyte interface in light. Figures 10(a) and 10(b) show the Nyquist plot for ZnO synthesized at 600 °C (GHZ61–GHZ63) and 700 °C (GHZ71–GHZ73), respectively. In the presence of light, the radius of the semicircle is in the order of GHZ63 > GHZ61 > GHZ62 and GHZ73 > GHZ71 > GHZ72 for the ZnO at 600 and 700 °C correspondingly, which signified a declining manner of the charge transfer resistance (R_ct) with an increase in the charge transfer rate.⁴⁴ 

In this study, ZnO microspheres were synthesized from heat treatment of ZnS microspheres at different oxidation temperatures of 600 and 700 °C. The crystalline and electronic structures were confirmed by physico-chemical characterization techniques as well as DFT studies. Synthesized ZnO microspheres were used as photoanodes to fabricate the DSSCs, and their photovoltaic parameters (J–V, IPCE) were measured. ZnO microspheres synthesized at 600 °C exhibited better efficiency than the ones synthesized at 700 °C. As temperature increased from 600 to 700 °C, porosity got reduced while the density of particles got increased, which affected the dye loading in addition to the DSSC efficiency. The enhanced photovoltaic parameters were obtained for GHZ62 with a photocurrent value of 9.63 mA/cm² and efficiency (η) of 3.38%. This enhanced performance is the result of high surface roughness and porosity, which assisted in improving the efficiency, dye loading, and better current conduction pathway.

S.W.G. would like to thank the SAKURA SCIENCE Program and PIRC, TUS, for financial support. This work was funded by the Researchers Supporting Project (No. RSP-2021/117), King Saud University, Riyadh, Saudi Arabia.

The authors have no conflicts to disclose.

The data that support the findings of this study are available within the article.