Multi-faceted investigation of ZnO nanoparticles with 1-ethyl-3-methylimidazolium iodide for enhancing DSSC performance

Abstract

This study presents a comprehensive experimental and computational investigation of ZnO nanoparticles with 1-ethyl-3-methylimidazolium iodide (EmimI) to enhance the performance of dye-sensitized solar cells (DSSCs). Two distinct ZnO morphologies, star-like (ZnO-S) and plate-like (ZnO-P), were synthesized and characterized. ZnO-P exhibited a higher specific surface area (6.06 m² g⁻¹) and a wider bandgap (3.31 eV) compared to ZnO-S (1.86 m² g⁻¹, 3.19 eV). When integrated into DSSCs with EmimI-based electrolytes, ZnO-P achieved a maximum efficiency of 9.86% after 3 h of dye immersion, while ZnO-S reached 8.38% after 1 h. Molecular dynamics simulations revealed that EmimI enhances iodide ion mobility and suppresses charge recombination by competitively blocking triiodide adsorption sites at the ZnO interface. The simulations also demonstrated strong interactions between Emim⁺ cations and the ZnO surface, optimizing electrolyte penetration and charge transport. These findings provide mechanistic insights into the role of ZnO morphology and EmimI in improving DSSC efficiency, offering a pathway for the design of high-performance solar energy devices.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-18470-4.

Introduction

Today, our main need is energy, and fossil fuels provide the majority of that energy. However, these fuels are non-renewable and pose significant environmental risks, such as air pollution and the accumulation of carbon dioxide in the atmosphere^1,2. There has been an increasing global demand for clean, affordable, and renewable energy sources. During the recent years, notable progress has been achieved in the utilization of abundant resources like sunlight, ocean waves, and wind energy for power generation. These developments have helped to partially address the growing need for sustainable energy production processes that mitigate environmental impacts^3,4. In 1991, a groundbreaking experiment was conducted by O’regan and Grätzel, introducing dye-sensitized solar cells (DSSC)⁵. These solar cells marked the first instance of combining organic semiconductor materials with inorganic semiconductors to harness solar energy efficiently. This pioneering work opened up new possibilities for renewable energy technology and paved the way for advancements in photovoltaic devices. DSSCs have since gained significant attention as a promising alternative to traditional silicon-based solar cells⁶, due to their cost-effectiveness and potential for flexible applications⁷.

To enhance electron transport and minimize charge carrier recombination, researchers have focused on optimizing the morphology of nanostructured electrodes. Various hierarchical nanostructures, including nanorods and nanowires, have been extensively studied in this promising approach for nanodevices and nanoelectronics^8–10. Ideally, these nanostructures should exhibit uniform shape, size, perfect crystalline structure, and be free from morphological defects. However, a significant challenge arises due to the heterogeneous interface between the substrate and the nanostructures. This interface can lead to electron scattering events that hinder efficient charge transport. Moreover, the diminished dimensions and smaller sizes of these nanostructures can jeopardize their thermal and chemical stability¹⁰. As previously proposed by Grätzel, structures based on high surface area nanoparticles are considered optimal for DSSCs^11,12. While several metal oxide materials with different nanostructures have been explored—including zinc oxide (ZnO)¹³, tin IV oxide (SnO₂)¹⁴, niobium pentoxide (Nb₂O₅)¹⁵, zinc stannate oxide (Zn₂SnO₄)¹⁶, none have demonstrated performance comparable to titanium dioxide (TiO₂)¹⁷.

In recent years, the hexagonal wurtzite structure of ZnO semiconductors has attracted significant attention for their application as photoanodes in various fields. ZnO offers several advantages in solar cell technology due to its wide band gap, cost-effectiveness, high optical transparency, and ability to emit light in the near-UV and visible ranges. Its conduction band edge level makes it suitable for anti-reflective coatings and transparent conductive materials in solar cells. The high exciton binding energy of ZnO allows for efficient excitonic transitions even at room temperature, resulting in improved radiative recombination efficiency for spontaneous emission and a lower threshold voltage for laser emission^13,18–20. When used as the anode component in DSSCs utilizing natural or organic dyes, ZnO, with its porous structure, facilitates dye adsorption and enhances charge collection from incident light energy¹³. In addition, the synthesis of ZnO can be achieved through various methods, including precipitation²¹, sol-gel²², and hydrothermal techniques²³. Each method yields different structures and morphologies such as nanowires²⁴, nanotubes²⁵, nanorods²⁶, nanosheets²⁷, and tetrapods²⁸. This versatility allows for the optimization of the photoanode’s morphology to enhance electron diffusion length. By carefully selecting the synthesis technique and controlling parameters such as temperature, time, precursor concentration, and pH conditions, researchers can tailor the morphology of ZnO to meet specific requirements regarding surface area, porosity, and crystal orientation²⁹. These optimized morphologies improve performance in applications related to photocatalysis or energy conversion devices.

Several solution-based methods have been developed for ZnO electrode fabrication, with chemical bath deposition (CBD), hydrothermal deposition, and electrodeposition (ED) being the most prominent. While CBD offers substrate versatility and ED provides faster deposition rates (despite requiring conductive substrates) along with low-temperature processing advantages for flexible electronics, hydrothermal synthesis excels in producing highly crystalline ZnO nanostructures with controlled morphologies through precise regulation of temperature, pressure, and precursor chemistry^30,31. Significant advances in ED techniques have been demonstrated by Goux et al.³², who systematically investigated ZnO thin film growth from zinc (II) ion solutions using various oxygen precursors (O₂, NO₃⁻, H₂O₂), while highlighting the critical influence of eosin Y dye concentration on film morphology and properties. Further developments were reported by Izaki and colleagues³³, who successfully fabricated Cl-doped ZnO/Cu₂O heterostructures through sequential ED processes. Complementarily, hydrothermal methods have enabled tailored ZnO architectures by optimizing growth parameters such as pH, mineralizer concentration, and reaction time, advantages particularly valuable for photoelectrochemical applications where crystallinity and surface area are paramount^34–36.

In optoelectronic device applications, ZnO nanostructures have emerged as cost-effective alternatives to TiO₂, exhibiting efficient electron transformation leading to reduced charge recombination rates through increased electron diffusion length. However, a limitation of ZnO-based DSSCs employing natural dyes is higher levels of free electron recombination with oxidized dye molecules than dyes based on ruthenium (Ru) complexes due to electron transfer constraints within ZnO photoanodes. Addressing these challenges associated with electron transfer limitations within ZnO photoanodes holds great potential to enhance the efficiency of ZnO-based DSSCs and unlock their full capability for solar energy conversion^37–39.

Efforts to enhance the efficiency of DSSCs have extensively explored the combination of traditional organic solvents such as acetonitrile, propylene carbonate, valeronitrile, dimethyl formamide, and dimethyl sulfoxide with iodide/tri-iodide redox couple sources. These endeavors have resulted in remarkable achievements, surpassing conversion efficiencies exceeding > 13%^40–42.

Concurrently, scientists have suggested the utilization of ionic liquids (ILs) as a component in ionic electrolytes. Several ILs, including imidazolium tricyanomethanide, 1-ethyl-3-methylimidazolium dicyanamide, triethylammonium perfluorocarboxylic and, imidazolium selenocyanate have been suggested as potential solvents for these electrolytes^43–46.

ILs exhibit a range of beneficial properties that make them well-suited for various applications in electrochemistry, such as lithium batteries, supercapacitors, fuel cells, and photovoltaics like DSSCs. These properties include nonflammability, negligible vapor pressure, environmental friendliness, high thermal stability, variable conductivity and viscosity, wide electrochemical windows, and miscibility with different solvents^43,47,48.

In the context of using ILs as electrolytes in DSSCs specifically, the physicochemical properties of these mixtures have been thoroughly examined. Critical parameters such as viscosity and ionic conductivity have been investigated to evaluate their potential effectiveness in enhancing DSSC performance⁴⁹.

IL-based electrolytes for DSSCs often involves reducing viscosity, which can be achieved by blending ILs with organic solvents or selecting inherently low-viscosity ILs. Studies confirm that this approach significantly enhances DSSC efficiency^50,51. However, high-viscosity ILs typically hinder the transport of redox species, leading to inefficient dye regeneration^52,53. To address this, low-viscosity ILs are preferred for DSSC applications⁵⁴, such as 1-ethyl-3-methylimidazolium dicyanamide (EMIDCA) with a viscosity of 21 mPa s at room temperature^55,56, 1-ethyl-3-methylimidazolium thiocyanate (EMISCN) with a viscosity of 21 mPa s at 25 °C ^46,57, 1-(3-hexenyl)-3-methylimidazolium iodide (HeMII) with a viscosity of 104 mPa s at 27 °C⁵¹, and 1-(3-butenyl)-3-methylimidazolium iodide (BeMII) with a viscosity of 80 mPa s at 27 °C⁵¹. Additionally, electrolytes combining 1-ethyl-3-methylimidazolium iodide (EmimI) with deep eutectic solvents and acetonitrile have demonstrated viscosities below 20 mPa s at 25 °C^43,47,58, making them particularly suitable for DSSCs due to their balanced ionic conductivity and fluidity.

Computational methods, including molecular dynamic (MD) simulations, have become crucial in the study of DSSCs. These simulations offer a powerful approach to gaining molecular-level insights and comprehending the underlying physics of various systems, including those involving ILs. For instance, we utilized MD techniques to investigate the adsorption modes of N719 dye in ethanol solvation. By employing MD simulations, we were able to determine how N719 dye interacts with TiO₂ and Bi₂WO₆ photoanodes⁵⁹. Mandal et al. conducted a purely computational investigation focused on exploring the adsorption behavior of organic azo dyes on TiO₂ and ZnO surfaces within DSSCs. These studies exemplify how MD simulations are pivotal in understanding essential aspects such as dye adsorption mechanisms and solvation effects within DSSC systems⁶⁰. Numerous computational studies have been conducted to investigate the impact of dye on the surfaces of the photoanode and cathode, using molecular dynamics simulations. Additionally, extensive research has been carried out on electrolytes based on ILs. For instance, Heydari et al. performed atomistic molecular dynamics simulations to study mixture of choline chloride and ethylene glycol as a deep eutectic solvent in combination with MeCN, lithium iodide, EmimI, and iodine as DSSC electrolytes⁴⁷. In addition, Wang et al. conducted a study in which they performed MD simulations to investigate the wetting behavior and regulating mechanism of Li⁺-doped IL droplets on the TiO₂-B (100) surface⁶¹.

In this work, we synthesize two different morphologies of zinc oxide nanoparticles. We aim to compare these morphologies regarding their crystal structure, surface area, surface charge, energy band gap, and recombination of electron–hole pairs. To achieve this, we employ various analytical techniques, including scanning electron microscopy, X-ray diffraction, N₂ adsorption–desorption isotherms, zeta potential measurements, diffuse reflectance spectroscopy and photoluminescence spectroscopy. These synthesized zinc oxide nanoparticles were utilized as photoanodes in DSSCs using EmimI electrolyte. This study presents a systematic investigation of ZnO photoanodes with distinct star-like and plate-like morphologies for DSSC applications, introducing several key innovations. First, we establish an optimized dye immersion protocol that reveals morphology-dependent performance characteristics. Second, our work provides the first comprehensive molecular dynamics analysis of EmimI-based electrolytes in acetonitrile, comparing systems with and without IL additives to elucidate interfacial interactions at both the ZnO photoanode and platinum cathode. The simulations reveal previously unreported mechanisms of competitive site blocking by Emim⁺ cations that suppress charge recombination. Third, we demonstrate how the unique physicochemical properties of our low-viscosity EmimI electrolyte enhance iodide ion mobility while maintaining optimal interfacial contact. These findings offer new molecular-level insights into structure-performance relationships in DSSCs, advancing both the fundamental understanding and practical design of ZnO-based photovoltaic systems.

Methods

Materials required

The required materials for this study were sourced from reputable suppliers. Platinum paste and conductive glass plates, specifically fluorine-doped tin oxide (FTO) glass with a sheet resistance of 15 Ω cm⁻² and standard iodine-based electrolyte, were purchased from Sharif Solar. Ethyl cellulose (EC), terpineol, ethanol, zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O), sodium hydroxide (NaOH), acetonitrile (ACN), lithium iodide (LiI), iodine (I₂), ethyl iodide (C₂H₅I) and N719 dye powder were purchased from Sigma-Aldrich (US Research Nanomaterial, Lnc, USA), a well-known supplier in the industry. Additionally, zinc acetate dihydrate (Zn(CH₃COO)₂·2H₂O, 1-methylimidazole (C₄H₆N₂), ethyl acetate (C₄H₈O₂) and diethylamine (DEA) were acquired from Merck. All materials used in the experiments were of analytical grade and did not require further purification before use.

Synthesis of ZnO plates (ZnO-P)

The synthesis of ZnO plates was carried out using (Zn(CH₃COO)₂·2H₂O as the precursor. Two solutions, A and B, were prepared for the synthesis. In solution A, 3.15 g of (Zn(CH₃COO)₂·2H₂O were completely dissolved in 30 mL of distilled water. In solution B, 3.82 g of NaOH was dissolved in 30 mL of fully distilled water. To begin the synthesis, solution B was added drop by drop to solution A, and the resulting mixture was stirred at 500 rpm for 40 min. This led to the formation of mixture C, which contained Zn(OH)₂. Next, mixture C was transferred to a 80 mL capacity Teflon-lined stainless-steel autoclave and placed in an oven for 7 h at a temperature of 200 °C. After this time, the autoclave was removed from the oven and cooled. The solid powder formed in the solution was centrifuged and washed three times with distilled water. Finally, the powder was completely dried in an oven at 180 °C for 4 h. This resulted in the synthesis of ZnO nanoplates using the hydrothermal method.

Synthesis of ZnO star-like (ZnO-S)

Hydrothermal method was applied to synthesize ZnO star-likes. We used Zn(NO₃)₂·6H₂O as the source of zinc. Specifically, we dissolved 2.97 g of Zn(NO₃)₂·6H₂O and 4 g of NaOH in distilled water to create a 10 mL solution with a molar ratio 1:10 of Zn²⁺ to OH⁻. We mixed 3 mL of this solution with 5 mL of distilled water and added 30 mL of ethanol, stirring the mixture for 10 min. Next, we added 10 mL of DEA to the solution while stirring, adjusting the pH to 13 using a pH meter. We placed the solution in an ultrasonic device for 30 min. The resulting solution was transferred to a Teflon-lined stainless-steel autoclave and heated in an oven at 180 °C for 2 h. After removing the solution from the autoclave, a white precipitate formed, which we centrifuged, filtered, and washed with distilled water and ethanol. The purified powder was completely dried in an oven at 70 °C for 10 h.

Preparation of EmimI IL

The EmimI IL was prepared using a standard procedure⁶², in which 1-methylimidazole was reacted with an excess of ethyl iodide and stirred at 70 °C for 72 h. The resulting salt was then carefully purified by repeated washing with ethyl acetate. To remove the ethyl acetate, the IL was first subjected to rotary evaporation for over 90 min at 70 °C. The residual ethyl acetate was then removed by drying the IL overnight at 70 °C under reduced pressure.

Fabrication of the DSSCs

To prepare a paste of ZnO nanoparticles, the following steps were taken, in the first step, 0.03 g of EC was added to a glass vial containing 2 mL of ethanol. This mixture was placed on a magnetic stirrer at 70 °C and stirred at 500 rpm until the ethyl cellulose was completely dissolved. Then, 500 μL of terpineol was added to the solution using a micropipette. The mixture was put back on the magnetic stirrer and stirred at 500 rpm until the terpineol was completely dissolved. In the second step, 0.03 g of ZnO (ZnO-S or ZnO-P) nanoparticles were added to the prepared solution. The mixture was then placed in an ultrasonic bath for 35 min to disperse the nanoparticles in the solution completely. The solution was then placed in a paraffin oil bath at a temperature of 70 °C and stirred on a magnetic stirrer for 24 h until the complete ethanol evaporated and a homogeneous nanoparticles paste was formed. In the preparation of the ZnO photoelectrode, FTO substrates with dimensions of 2 × 1.5 cm and a resistance of 15 Ω cm⁻² were first washed and cleaned with a detergent solution, followed by ethanol, acetone, and distilled water for 15 min each. After the FTO surface was cleaned, the paste of ZnO nanoparticles was coated onto the FTO using the Doctor Blade method. The coated FTO was then placed in a furnace at a temperature of 450 °C for 1 h to prepare the zinc oxide photoelectrode which the thickness of the ZnO-S and ZnO-P samples measured approximately 3.93 μm and 3.09 μm, respectively (Fig. 1). Next, a commercial Pt paste (Sharif solar) in the form of a viscous paste was coated onto the FTO using the Doctor Blade technique. The as-prepared cathode was then annealed in the furnace for 30 min at a temperature of 400 °C. In the dye solution (0.4 mM) preparation section, 0.012 g of N719 powder was poured into a 25 mL volumetric flask, and ethanol was added to reach a volume of 25 mL. The solution was then stirred on a magnetic stirrer at 500 rpm for 24 h. In a completely dark environment, the prepared zinc oxide photoanode was placed in the dye solution at different time intervals to investigate the absorption of the dye on the surface of zinc oxide nanoparticles. The electrolyte was prepared using ACN solvent, where 0.025 g of iodine, 0.2 g of lithium iodide, and 0.357 g of EmimI were dissolved in 5 mL of solvent. To prevent short-circuiting between the two electrodes, a thin layer of hot-melt spacer made of Surlyn was applied. The electrolyte was infused into the cell via a vacuum backfilling process through a hole drilled in the counter-electrode. In the assembly of DSSCs, alligator clamps were utilized to establish the necessary electrical connection between the anode and cathode electrodes, ensuring proper functionality of the system.

Fig. 1.

SEM images of ZnO-S (A,B) and ZnO-P (D,E) synthesized under hydrothermal conditions and cross-section SEM images of ZnO-S (C) and ZnO-P (F) deposited on FTO.

Instrumentation and measurements

To calibrate a solar simulator, the Abet Technologies reference solar cell (4 cm² single crystal silicon) is used to measure radiation levels. The electrochemical characterization involved using an Autolab PGSTAT302N electrochemical station with GPES software to measure the photocurrent density–voltage (J–V) characteristics of each ZnO sample. These J-V responses were captured under simulated solar light conditions (AM 1.5G, 100 mW/cm²). The solar simulator employed is the Onsol-60A, designed to adhere to the spatial distribution and thermal stability criteria outlined in the IEC-60904-9 standard. This equipment is dedicated to assessing and regulating the performance of solar cells. It features a unique light source that combines halogen and LED lamps to ensure precise illumination conditions. J–V measurements were performed four times on each sample.

Electrochemical impedance spectroscopy (EIS) was performed using the Autolab system integrated with a frequency response analyzer (FRA). The frequency range adjustable from 1 mHz to 1 MHz was applied while DSSCs were under the open-circuit voltage bias with constant light illumination (AM 1.5G, 100 mW/cm²). The impedance spectra obtained were analyzed and fitted using ZView software. Furthermore, the incident photon to current conversion efficiency (IPCE) was evaluated using an IPCE-020 model from sharif solar.

Characterization methods

Scanning-electron-microscopy (SEM) analysis was utilized to examine the surface morphology and cross-sectional structure of ZnO nanoparticles and pastes by using Tescan Vega 3. X-ray diffraction analysis (XRD) with a Cu Kα radiation source (λ = 1.54060 Å) was used to check the phase and crystallinity of two zinc oxide samples by Philips-D6792 diffractometer. The samples were analyzed under room temperature in the range of 2θ = 10 − 80 with a scan rate of 0.1° 2θ s⁻¹. The synthesized zinc oxide powders were analyzed using diffuse reflectance spectroscopy (DRS) recorded on a Shimadzu UV-2550. This technique was employed to assess the band gap and optical properties of ZnO. BaSO₄ pellets served as a reference sample during the analysis. Barret−Joyner−Halenda (BJH) and Brunauer−Emmett−Teller (BET) analyses were performed based on nitrogen adsorption−desorption assessments at 77 K to determine the pore volume, surface area, and pore size distribution of the synthesized nanoparticles, in this analysis ZnO samples were degassed at 120 °C for 2 h to eliminate pollutants and moisture. The Zeta potentials of the samples were measured using the Zeta Potential analyzer, Particle Metrix. The powder samples were dispersed in water (Milli-Q water used as the dispersant) for surface charge measurement at room temperature. Room-temperature photoluminescence spectroscopy was conducted using a Perkin-Elmer luminescence spectrometer (LS 55, USA) with an excitation wavelength of 320 nm. The proton nuclear magnetic resonance (¹HNMR) spectra were acquired on a Bruker Avance 400 MHz spectrometer at ambient temperature. Chemical shifts (δ) are reported in parts per million (ppm) relative to tetramethylsilane (TMS, δ 0.00) as the internal standard, with downfield shifts denoted as positive values. Viscosity was measured using an Anton Paar MCR-302 rheometer with a 25-mm parallel-plate geometry (0.5-mm gap). Tests were performed at 25 °C (Peltier-controlled) under controlled shear rate or frequency. Each sample underwent three repetitions of zeta potential and DRS analyses, with subsequent reporting of the average data alongside their respective standard deviations.

Computational details

In the theoretical study investigating the effect of electrolyte solutions on the zinc oxide surface, ACN molecules were optimized using a combined density functional method (B3LYP) in the Gaussian program, and the H, O, N, and C atoms were described using the 6-311++G(d,p) basis set⁶³. The calculated structures were then checked for the state of minimum energy on the potential energy surface by controlling the vibrational frequencies. The partial charges were analyzed using the natural bond orbital (NBO) method and obtained for each set of molecules^64,65. The charges from electrostatic potential (ESP) were simulated using a grid-based method (ESP-based charges)⁶⁶. However, no significant difference in the structural properties of the system was observed due to the atomic charges, and therefore, the simulation continued with the NBO charges. To optimize the initial structures of the IL (EmimI) component, we utilized the Gaussian package and performed calculations at the B3LYP/6-311++G(d,p) level of theory for C, N, and H atoms⁶⁷. For iodide, we used LANL2DZ as the basis set⁶⁸. To determine partial atomic charges that would yield accurate results for ILs and effectively reproduce properties of both pure ILs and aqueous mixtures of ILs, we employed the CHELPG procedure to fit the electrostatic potential surface⁶⁹. The simulations were performed using the GROMACS 5.1.5 program⁷⁰. The Lennard–Jones parameters used to describe the interactions of LiI, ACN, and I₂ molecules were obtained from the OPLS-AA parameter set. For modeling the interactions within EmimI IL, modified OPLS-AA parameters as proposed by Canongia Lopes et al.⁷¹, to represent both intermolecular and intramolecular interactions of zinc (Zn) and oxygen (O) atoms in the ZnO slab, LJ spheres with σZn = 0.171 nm and εZn = 1.254 kcal mol⁻¹ for Zn atoms, as well as σO = 0.212 nm and εO = 0.418 kcal mol⁻¹ for O atoms were employed⁷². The atomic charges assigned to the Zn and O atoms in ZnO were + 1.026e and − 1.026e, respectively⁷³. Regarding the platinum wall, platinum (Pt) atoms were modeled using εPt = 7.80 kcal mol⁻¹and σPt = 0.285 nm parameters⁷⁴. These parameter choices enable accurate representation of interatomic forces within the simulated system. The force field parameters, including Lennard–Jones σ (nm) and ε (kcal/mol) for EmimI, ACN, and LiI constituents, are compiled in Table S1. The objective of the MD simulation was to investigate the behavior of ZnO nanostructures as photoanodes in DSSCs, focusing on their interaction with the electrolyte solution. To assess the performance of the photoanode, we simulated a system where the electrolyte solution was positioned between ZnO and platinum walls. In this study, we employed a model system consisting of 120 ion pairs of LiI and 1200 ion pairs of EmimI dissolved in ACN solvent (24,000 molecules). The simulation was conducted for 20 ns under isothermal-isobaric (NPT) ensemble conditions to mimic bulk phase behavior in electrolyte solutions. For simulations involving solid/liquid interfaces, we confined an equilibrated bulk ensemble of the electrolyte solution within nanosized slit pores (10 nm) formed by ZnO/Pt slabs along the z-direction. The equations of motion were solved using Verlet-Leapfrog integration algorithm, under the periodic boundary conditions. Auxiliary walls composed of graphene layers were implemented at both ends of the slit pores to prevent diffusion into vacuum regions. The MD simulation began by utilizing the resulting structures as starting points, followed by initial energy minimizations. Subsequently, equilibration was performed using a canonical ensemble (NVT) with a velocity-rescaling thermostat set to a time constant of 0.1 ps to maintain the system temperature at 300 K. The final simulation runs were carried out for a duration of 30 ns under NVT ensemble conditions, employing time steps of 1 fs. In each simulation, cutoff distances of 12 Å were chosen for nonbonding interactions. For long-range Coulombic interactions, the particle-mesh Ewald method was employed with a cutoff distance of 15 Å. These parameters and techniques helped ensure accurate and efficient simulations while maintaining consistency with experimental conditions. By employing these simulation setups and conditions, we aimed to gain insights into how ZnO nanostructures interact with and influence the behavior of electrolyte solutions within DSSC systems.

Results and discussions

Characterization

¹H NMR

The ¹H NMR characterization of the IL sample was performed, and a representative spectrum of EmimI is provided in Fig. S1. The assignment of the signals to the respective protons of the Emim⁺ cation was conducted in accordance with the work of D’Agostino⁷⁵. ¹H NMR (400 MHz, DMSO-d₆) δ 0.44–0.47 (d, J = 1.5 Hz, 2H), 2.92–2.93 (d, J = 1.5 Hz, 2H), 3.26–3.28 (m, 1H), 6.78–6.81 (q, J = 1.7 Hz, 1H), 6.87–6.90 (q, J = 1.8 Hz, 1H), 8.26–8.28 (d, J = 1.9 Hz, 1H).

SEM analysis

In Fig. 1A,B, SEM images displayed ZnO particles that were synthesized via a hydrothermal method at a temperature of 180 °C using zinc nitrate as the precursor. These images clearly exhibited the growth of star-shaped structures formed by nanorod aggregation around a central point. This unique dispersion pattern was referred to as “star-like morphology.” Further examination in Fig. 1D,E showcased additional SEM images highlighting the morphological characteristics of ZnO nanoparticles synthesized under hydrothermal conditions at 200 °C using zinc acetate as the base material. In contrast to the previous synthesis process, these images revealed plate-like structures. The observed variations in composition and synthesis parameters significantly impacted the resulting morphologies in both cases. Consequently, based on their respective morphologies, we designated the star-shaped structure as ZnO-S and labeled the plate-shaped structure as ZnO-P. In Fig. 1C,F, the thickness of the ZnO layers in the ZnO-S and ZnO-P samples measured approximately 3.93 μm and 3.09 μm, respectively. The thickness of ZnO plays a critical role in determining the level of resistance necessary to inhibit the charge recombination process or facilitate optimal electron transfer⁷⁶.

XRD pattern

In Fig. 2A, the XRD patterns of ZnO-P and ZnO-S were displayed. Interestingly, these patterns exhibited identical diffraction peaks at specific angles: 31.83° (100), 34.52° (002), 36.29° (101), 47.58° (102), 62.89° (103), 56.64° (110), 62.89° (103), 66.41° (200) 67.93° (112), 69.18° (201) and 77.02° (202)⁷⁷. Sharp and well-defined peaks indicated that the ZnO products had a high degree of crystallization. The XRD analysis further confirmed that both samples exhibited a hexagonal wurtzite phase, with lattice constants of (a = b = 0.3251 nm) and (c = 0.5212 nm), in accordance with standard JCPSD card no. 01–007-2551⁷⁸, which was consistent with their respective morphologies. It was worth noting that these findings from the XRD patterns provide valuable insights into the crystal structure and quality of the synthesized ZnO materials for both configurations. Detailed calculations for crystallite size, lattice strain, and dislocation density are provided in the Supporting Information.

Fig. 2.

Characterization of ZnO-P and ZnO-S nanoparticles: XRD pattern (A), DRS (B), and nitrogen adsorption−desorption isotherms and BJH plot for ZnO-S (C) and ZnO-P (D).

Table 1 demonstrates that the properties of S differ depending on the morphology and method of synthesis, indicating that lattice imperfections may be dependent on particle size. Furthermore, the micro-strain observed in the samples varies with morphology, likely reflecting changes in their microstructure, size, shape, and the presence of defects in the particles. Defects in ZnO materials play a significant role in the transport and collection of electrons in the performance of DSSCs⁷⁹. By lowering the density of surface defects, we can reduce the number of trapped electrons at the semiconductor surface, which in turn helps to mitigate electron recombination^80–83. According to our observations, the value of S for ZnO-P is lower than that for ZnO-S, as shown in Table 1. This suggests that the electron recombination for this nanoparticle can be considered to be less significant.

Table 1.

XRD analysis showing lattice parameters, size, lattice strain, and dislocation density of ZnO samples.

Sample	Peak Position	FWHM	a = b (Å)	c (Å)	D (nm)		lattice Strain	S (10–3)
	(101)				Scherrer	W–H	W–H
ZnO-P	36.37	0.24	3.25	5.21	34.21	38.68	0.0016	0.85
ZnO-S	36.23	0.26	3.25	5.21	32.14	35.24	0.002	0.96

DRS

In Fig. 2B, the DRS spectrums of the synthesized samples are presented. Notably, it was observed that ZnO-P exhibits higher light absorption in the UV region compared to ZnO-S. This disparity suggests that ZnO-P has a greater potential for UV photovoltaic activity. The band gap energies of pure ZnO-S and ZnO-P were estimated using the Tauc/David-Matt model based on the UV–vis absorption spectrum^59,84,85. The analysis revealed that the band gap energy for pure ZnO-S was determined to be 3.19 eV, while for pure ZnO-P it was found to be 3.31 eV. These findings indicated that ZnO-P possessed a wider bandgap compared to ZnO-S.

The band edge potentials of the valence band (VB) and conduction band (CB) for ZnO particles were determined using specific equations. The conduction band edge potential (E_CB) is defined by the equation E_CB = E_VB − E_g⁸⁶, whereas the valence band edge potential (E_VB) is calculated using E_VB = X − E_e + 0.5E_g^59,85. In these expressions, X represents the absolute electronegativity of the semiconductor, E_e indicates the energy of free electrons on the hydrogen scale (approximately 4.5 eV)⁵⁹, and E_g refers to the band gap of the ZnO photoanodes. For ZnO, the electronegativity (X) is approximately 5.79 eV⁸⁷. From these calculations, the CB and VB edge potentials for ZnO-P were evaluated at − 0.36 eV and 2.94 eV, respectively. In comparison, for ZnO-S, the determined CB and VB edge potentials are − 0.30 eV and 2.88 eV, respectively. The schematic representation of the DSSC (see Scheme 1) depicts the electrical band diagrams for ZnO-P and ZnO-S. The downward shift in the CB position enhances the kinetic driving force for electron injection from photoexcited dyes into the photoanode⁸⁸. Meanwhile, ZnO-P’s wider bandgap provides thermodynamic stability to the electrons while suppressing recombination losses in DSSCs^13,89,90. This improvement stems from ZnO-P’s superior crystallinity, which reduces charge carrier trapping through fewer crystal defects and enables more efficient electron transport. Experimental studies indicate that photoanodes with a wider bandgap, such as ZnO-P, exhibit significantly reduced charge recombination at the anode-electrolyte interface^91,92.

Scheme 1.

The band diagram of the ZnO-P and ZnO-S, the sensitizer N719 dyes, and the redox couple of I^-/I^-₃ in the electrolyte.

ZETA potential

Zeta potential measurements were performed to assess the surface charge of ZnO nanoparticles. Prior to measurement, the powders were dispersed in water. The zeta potential values were recorded for different morphologies of nanoparticles. Figure S3A illustrates that the zeta potential of ZnO-S was − 13.55  mV, indicating a more positive charge compared to ZnO-P (− 31.11  mV). Zeta potential, a metric of the effective electric charge at nanoparticle surfaces, indicates that ZnO-S, with a comparatively less negative surface charge than ZnO-P, has the potential to enhance the adsorption of anionic dye N719, which carries a negatively charged⁵⁹.

N₂ adsorption−desorption isotherms

The N₂ adsorption–desorption isotherm of ZnO, as illustrated in Fig. 2C,D, provides valuable insights into the porosity and surface characteristics of the material. In this graphical representation, the amount of N₂ adsorbed (V_a, measured in cm³ g⁻¹) is plotted against the relative equilibrium pressure (P/P₀), where P₀ represents the vapor pressure of the bulk liquid nitrogen and P denotes the equilibrium pressure of desorption at the liquid nitrogen temperature (~ 77 K).

The BET surface area, a key parameter for assessing the available surface area for adsorption, was determined to be 1.86 m² g⁻¹ for ZnO-S and 6.06 m² g⁻¹ for ZnO-P. This difference in surface area values indicates variations in the surface properties of the two ZnO samples. Additionally, the pore volume, representing the total volume of pores within the material, was found to be 0.006 cm³ g⁻¹ for both ZnO samples. By utilizing the BJH model, which is commonly employed to analyze pore size distributions, the investigation of porosity revealed a mean pore diameter of 1.66 nm for ZnO-S and 1.22 nm for ZnO-P. These pore size measurements provide crucial information about the pore structures present in the ZnO samples, influencing their adsorption and catalytic properties. The increased specific surface area of ZnO-P nanoparticles provides a greater number of surface sites for the adsorption of dye molecules and ions, potentially leading to enhanced DSSC efficiency. In accordance with the BET classification, the nanostructures of ZnO exhibited microporosity, indicating the presence of small pores. Furthermore, the hysteresis loop observed for the photocatalysts was classified as type III, following the IUPAC classification^93,94.

Photoluminescence

Photoluminescence (PL) spectroscopy serves as a powerful diagnostic tool for evaluating the performance of nanoparticle-based DSSCs. This technique provides critical insights into: (1) the optical properties of photoanode materials, (2) charge transfer dynamics at nanoparticle-dye interfaces, and (3) electron recombination processes that fundamentally govern cell efficiency^95,96. PL studies demonstrate that reduced photoluminescence intensity correlates strongly with enhanced DSSC performance, as evidenced by improved charge carrier lifetimes and minimized energy losses^96,97. The comparative PL analysis (Figure S3) reveals that ZnO-P exhibits lower emission intensity than ZnO-S. This PL quenching phenomenon can be attributed to three key factors: First, the decrease in PL could be related to higher electron mobility which reduces electron–hole recombination probability. Second, optimized defect engineering minimizes non-radiative recombination centers that compete with dye regeneration. Third, controlled nanoparticle film morphology ensures unimpeded charge transport while maintaining sufficient surface area for dye loading^39,88,97,98. Notably, materials exhibiting lower PL intensities demonstrate improved environmental stability, suggesting their potential for prolonged operational lifetimes in practical DSSC applications^88,99,100.

Photovoltaic performance of the DSSCs

We synthesized ZnO-S and ZnO-P nanoparticle powders, which were then coated onto the FTO substrate. The coated substrates were immersed in N719 dye solutions for different durations of 1, 3, 6, and 12 h. To assess the performance and efficiency of the fabricated DSSCs, we conducted J-V measurements utilizing a solar simulator. From these measurements, we determined crucial parameters, including current density (J_sc), open circuit voltage (V_oc), fill factor (FF), and conversion efficiency () for each solar cell. In the provided Fig. 3A and Table 1, it is evident that the duration of immersion in dye solutions for ZnO-S photoanodes (1, 3, 6, and 12 h) correlates with an increase in V_oc and a decrease in J_sc, , and FF. In Fig. 3A, an increase in the V_oc was observed from 0.64 V to 0.75 V after 1 and 12 h, respectively. Conversely, the J_sc of this cell decreased from 17.37 mA cm⁻² (for an immersion time of 1 h) to 2.24 mA cm⁻² (after an immersion time of 12 h), as depicted in Table 2. These findings indicate that with an increase in the photoanode’s immersion time in the dye solution, the overall efficiency of the cell decreases. As shown in Fig. 3B, the best efficiencies were obtained for different periods: for durations of 1, 3, 6, and 12 h, they were recorded as being approximately equal to; 8.38%, 6.90%, 5.35%, and finally dropped down to only about; 0.72% respectively (see Table 2). Figure 3C presents the efficiency of ZnO-P in dye solution immersion for various time durations: 1, 3, 6, and 12 h. The recorded efficiencies were 6.92%, 9.86%, 5.68%, and 3.10%, respectively. Based on Fig. 3D, the highest solar energy conversion efficiency was achieved after three hours with a J_sc of 20.28 mA cm⁻², a V_oc of 0.7 V, and FF of 0.69. Furthermore, in Table 2, it was observed that the maximum J_sc values at different immersion times were as follows: at one hour, 15.53 mA cm⁻²; six hours, 12.50 mA cm⁻²; and twelve hours, 6.91 mA cm⁻². The photovoltaic performance parameters (efficiency, J_sc, V_oc, and FF) exhibited distinct correlations with dye adsorption time for ZnO-S and ZnO-P photoanodes. ZnO-S reached peak efficiency (8.38%) after just 1 h of dye immersion, with performance declining at longer times due to likely dye aggregation and pore blocking^101,102, despite continued dye uptake. In contrast, ZnO-P demonstrated superior performance with maximum efficiency (9.86%) at 3 h, benefiting from its larger pore volume and more negative surface charge that enabled enhanced and prolonged dye adsorption. Moreover, ZnO-P’s thinner layer thickness (3.09 μm vs. 3.93 μm for ZnO-S) further contributed to its higher efficiency by reducing electron transport distance and interfacial resistance, minimizing recombination losses¹⁰³. The negative surface charge of ZnO-P, allowing optimal loading without immediate pore clogging. J_sc initially increased with dye loading but plateaued or decreased at excessive loadings due to recombination and reduced light penetration, while V_oc trends suggested morphology-dependent recombination kinetics. The presence of EmimI further improved performance through enhanced dye regeneration and reduced recombination, as evidenced by higher FF and J_sc values. The decrease in V_oc but improvement in FF for ZnO-P with longer dye immersion suggests two competing effects: (1) Increased dye loading introduces more recombination sites, lowering V_oc due to faster electron loss to the electrolyte, while (2) enhanced dye-electrolyte interfacial contact and better pore filling (facilitated by EmimI) improve charge transport, boosting FF. This trade-off highlights that optimal immersion balances dye coverage (for light absorption) with minimal recombination (for voltage retention). The plate-like ZnO-P structure likely supports more efficient charge collection than ZnO-S, even at higher dye loads.

Fig. 3.

Photovoltaic performance of DSSCs with (A,B) ZnO-S and (C,D) ZnO-P photoanodes: (A,C) J-V curves with EmimI-based electrolyte; (B, D) efficiency evolution versus dye loading time.

Table 2.

Photovoltaic parameters of the DSSC for ZnO-S photoelectrode with EmimI-based electrolyte.

ZnO-S (Time)	Jsc (mA/cm2)	Voc (V)	FF	(%)
1 h	17.37	0.64	0.75	8.38
3 h	13.81	0.72	0.69	6.90
6 h	10.91	0.74	0.66	5.35
12 h	2.24	0.75	0.42	0.72

For more detailed key parameters regarding DSSCs, Tables 2 and 3 as well as Figs. 3A–D can be referred.

Table 3.

Photovoltaic parameters of the DSSC for ZnO-P photoelectrode with EmimI-based electrolyte.

ZnO-P (Time)	Jsc (mA/cm2)	Voc (V)	FF	(%)
3 h	20.28	0.70	0.69	9.86
1 h	15.53	0.67	0.66	6.92
6 h	12.50	0.66	0.68	5.68
12 h	6.91	0.64	0.70	3.10

Figure 4A illustrates the IPCE of two distinct ZnO structures. These results provide insights into the number of incident photons within the cell and their corresponding efficiencies. An increase in IPCE values signifies not only enhanced optical absorption but also an improvement in the conversion efficiency of photons into electrons. Analysis of the IPCE results reveals that the incident light wavelength predominantly falls within the 300–800 nm range in DSSCs. When comparing the optimal performance of both ZnO structures with EmimI electrolyte (3-h immersion for ZnO-P vs 1-h for ZnO-S), ZnO-P demonstrated a peak IPCE of 91.63% while ZnO-S reached 88.48% at 540 nm. The IPCE values showed strong positive correlation with JSC measurements, confirming consistency with the J–V curve analysis. Notably, the highest IPCE values occurred within the 400–650 nm range, corresponding perfectly with the absorption spectrum of N719 dye. The observed IPCE reduction in ZnO-S compared to ZnO-P may be attributed to less efficient dye regeneration by iodide in the ZnO-S system¹⁰⁴.

Fig. 4.

(A) IPCE spectra and integrated photocurrent of ZnO-P (3h dye immersion) and ZnO-S (1h immersion), (B) EIS Nyquist plots at Voc (inset: equivalent circuit) with EmimI-based electrolyte.

EIS analysis was conducted on DSSCs composed of FTO/ZnO (photoanode)/dye/electrolyte/Pt (counter electrode)/FTO. The Nyquist curves, captured at V_oc (Fig. 4B), exhibit two semicircles across the frequency spectrum, as per the equivalent circuit model shown in the figure’s inset. Key features of the solid/liquid interface, such as charge-transfer resistance at the counter electrode/electrolyte interface (Rct1) (generally observed in high-frequency range), and the charge-transfer resistance at the ZnO/electrolyte interface (Rct2) (in middle-frequency range), and Zw is the electrolyte diffusion resistance (in low-frequency range)  were investigated. EIS measurements were performed over a frequency range of 1 mHz to 1 MHz under V_oc with constant light illumination^47,58. Since an identical counter electrode is applied for all the two types of devices, the difference at the ZnO/dye/electrolyte interface is completely responsible for the resistance values. The EIS plot of Fig. 4B reveals that the resistance at the counter electrode/electrolyte interface (Rct1) is very close and as expected have no impact on the photovoltaic performance of DSSCs (21.2 Ω for ZnO-P and 23.3 Ω for ZnO-S). Notably, both the Jsc (Jsc: 20.28 mA/cm² for ZnO-P vs. 17.37 mA/cm² for ZnO-S) and power conversion efficiency (9.86% for ZnO-P vs. 8.38% for ZnO-S) exhibit a clear correlation with the charge-transfer resistance at the ZnO/electrolyte interface (Rct2: 96.5 Ω for ZnO-P vs. 114.9 Ω for ZnO-S). This performance enhancement corresponds to the dye immersion time (3 h for ZnO-P compared to 1 h for ZnO-S). ZnO-P DSSC has the smallest Rct2 resistance, which implies the reduced electron recombination and enhanced electron transport. This could be further attributed to the ZnO plate-shaped structure, which can provide much shorter pathways for electrons getting to the substrate. Nyquist plots (Fig. 4B) exhibit only one dominant semicircle (SC) in the intermediate frequency region, corresponding to the charge transfer resistance (Rct) at the photoanode/electrolyte interface^52,105. In contrast, similar studies on DSSCs with ionic liquid-based electrolytes often reveal additional SCs in the low-frequency range⁵⁴. These SCs, which reflect charge transport resistance, are typically attributed to the diffusion of I₃⁻ in the electrolyte and are commonly observed in high-viscosity systems^52,106. The absence of these SCs in our case can be explained by the low viscosity of both pure synthesized EmimI (15.11 mPa·s at 298 K) and the prepared electrolyte (0.76 mPa·s), which is significantly lower than that of other ionic liquids, such as the low-viscosity 1-allyl-3-methylimidazolium dicyanamide (AMIM-DCA)⁵⁴. At this stage of analysis, the incorporation of EmimI appears to cause minimal disruption to redox species diffusion. Furthermore, the electrolyte’s low viscosity enhances its penetration into the porous ZnO layer, promoting efficient dye regeneration and redox species mobility. Conversely, excessively high electrolyte viscosity could hinder I₃⁻ diffusion, impede electrolyte permeation into the photoanode, and ultimately compromise electron transport dynamics in the solar cell.

It is worth noting that IPCE and EIS analyses were conducted on the ZnO-S sample immersed in the dye for 1 h and the ZnO-P sample immersed for 3 h, which exhibited the highest efficiency in the J–V analysis.

To provide a thorough and exhaustive comparison of the electrolyte systems employed in this study, two additional electrolytes, namely the standard electrolyte and electrolyte without EmimI (comprising 5 mL of ACN, 0.025 g of I, and 0.2 g of LiI), are considered. This comparative analysis aims to enhance the understanding of the impact of EmimI IL on the efficiency of DSSCs.

The investigation of ZnO-S with the standard electrolyte revealed notable trends. As the immersion time in the dye increased from 1 to 12 h, the J_sc decreased from 14.29 to 3.49 Ma cm⁻², also the FF decreased from 0.72 to 0.47. The optimal performance for ZnO-S with the standard electrolyte was achieved at 1 h of immersion, showcasing a J_sc of 14.29 mA cm⁻², a V_oc of 0.67 V, an FF of 0.72, and an efficiency of 7.02%. For a more comprehensive analysis, detailed insights can be gleaned from the accompanying Table 4 and Figure S4A.

Table 4.

Photovoltaic parameters of the ZnO-S photoelectrode with standard electrolyte.

ZnO-S (Time)	Jsc (mA/cm2)	Voc (V)	FF	(%)
1 h	14.29	0.68	0.72	7.06
3 h	12.39	0.67	0.66	5.59
6 h	8.00	0.67	0.61	3.28
12 h	3.49	0.69	0.47	1.14

Based on the J-V plot depicted in Figure S4C, the investigation of the ZnO-P photoanode with the standard electrolyte highlighted the peak performance of the DSSC after 3 h of immersion in the dye. This configuration demonstrated a J_sc of 15.58 mA cm⁻², a V_oc of 0.68 V, an FF of 0.67, and an efficiency of 7.23%. The efficiency rankings for immersion durations of 1, 6, and 12 h were as follows: 6.05%, 4.02%, and 1.27%, respectively. For a comprehensive overview of the key parameters of the ZnO-P DSSC with the standard electrolyte, please refer to Table 5. The performance loss with prolonged dye immersion time is primarily due to excessive dye loading, which leads to aggregation and pore blocking. This reduces light absorption (lower J_sc) and hinders electrolyte diffusion, increasing charge recombination (lower FF). While V_oc may slightly rise due to reduced back-electron transfer, the overall efficiency drops as dye overloading disrupts optimal charge generation and transport. The effect is more pronounced in ZnO-S (peak at 1 h) than ZnO-P (peak at 3 h) due to its lower surface area and faster pore saturation.

Table 5.

Photovoltaic parameters of the ZnO-P photoelectrode with standard electrolyte.

ZnO-P (Time)	Jsc (mA/cm2)	Voc (V)	FF	(%)
3 h	15.58	0.68	0.67	7.23
1 h	13.68	0.64	0.68	6.05
6 h	9.18	0.65	0.67	4.02
12 h	3.40	0.63	0.59	1.28

The utilization of an electrolyte without EmimI in DSSC performance based on ZnO-S and ZnO-P photoanodes revealed interesting trends, as illustrated in Figure S4B,D and Table 6 and 7. Notably, ZnO-S experienced a decrease in J_sc from 7.29 to 1.68 mA cm⁻² as the immersion time increased from 1 to 12 h. Conversely, the V_oc showed a rising trend from 0.61 to 0.66 V. The high performance of ZnO-S with the EmimI-free electrolyte was achieved at 1 h of immersion, showcasing a J_sc of 7.29 mA cm⁻², a V_oc of 0.61 V, an FF of 0.55, and an efficiency of 2.50%. As shown in Table 6, the J_sc and efficiency exhibited a decline with longer dye immersion time, whereas the V_oc showed a corresponding increase.

Table 6.

Photovoltaic parameters of the ZnO-S photoelectrode without EmimI.

ZnO-S (Time)	Jsc (mA/cm2)	Voc (V)	FF	(%)
1 h	7.29	0.61	0.55	2.50
3 h	6.50	0.64	0.53	2.27
6 h	3.47	0.65	0.45	1.03
12 h	1.68	0.66	0.45	0.51

Table 7.

Photovoltaic parameters of the ZnO-P photoelectrode without EmimI.

ZnO-P (Time)	Jsc (mA/cm2)	Voc (V)	FF	(%)
3 h	9.38	0.65	0.48	2.97
1 h	8.34	0.64	0.41	2.21
6 h	4.74	0.63	0.42	1.26
12 h	2.37	0.61	0.41	0.61

Furthermore, the performance of the DSSC based on ZnO-P with an electrolyte without EmimI was explored in Figure S4D. The observed trend for J_sc over immersion times of 3, 1, 6, and 12 h was recorded at 9.38, 8.34, 4.74, and 2.37 mA cm⁻², respectively. Similarly, the V_oc values during the same time intervals were measured at 0.65, 0.64, 0.63, and 0.61 V, respectively. The optimal performance of this DSSC was achieved during a 3-h immersion period, exhibiting an efficiency of 2.97%. As seen in Table 7, the drastic reduction in J_sc following 12 h of dye immersion resulted in a severe efficiency loss in the cell.

Upon review of the results, it was noted that the general trend of DSSC efficiency, based on the immersion time in the dye solution, remains consistent for the two electrolytes standard and without EmimI, aligning with the electrolyte containing EmimI. The addition of EmimI to the electrolyte of DSSCs plays a crucial role in improving their photovoltaic performance. The positively charged Emim⁺ cation tends to accumulate near the ZnO photoanode surface, where it competes with the triiodide ions (I₃⁻) for adsorption sites. This competitive interaction effectively suppresses the undesirable recombination of photoinjected electrons with I₃⁻, a process known as dark current⁵⁴. As a result, charge collection efficiency is enhanced, leading to a notable increase in the FF of the solar cell. Moreover, the imidazolium ring in EmimI, with its near-planar geometry^107,108, contributes to optimal electrolyte viscosity, while its flexible alkyl chains facilitate penetration into the porous ZnO structure¹⁰⁹. This property facilitates deeper and more uniform infiltration of the electrolyte into the porous ZnO nanostructure, ensuring improved interfacial contact and ion transport. Consequently, the overall photovoltaic performance of the DSSC is significantly enhanced.

The reproducibility of devices fabricated with ZnO-P and ZnO-S electrodes was confirmed by constructing and analyzing four samples for each condition. The performance values presented in Tables 2, 3, 4, 5, 6 and 7 are the averages and standard deviations derived from these replicates.

Molecular dynamic simulation results

To assess the average orientation and spatial distribution of molecules at the interface between solid and liquid phases, density profiles were analyzed. This analysis involved examining the distribution of electrolyte molecules within the electrolyte mixture across slabs. Figure 5A presents the number density profile of electrolyte boxes obtained from all-atom simulations conducted along the Z-direction for ZnO/Pt systems. The results reveal three distinct regions based on their positions along the Z-axis:

1. Region 1: This region corresponds to the interface between the photoanode and electrolyte solution, with a position range of 3 nm < Z < 5 nm.

2. Region 2: This region represents the interface between Pt and electrolyte solution, from 13 nm < Z < 15 nm.

Fig. 5.

The number density profiles of electrolyte solutions confined within nanopores of a ZnO/Pt model at 300 K (A). The snapshot from the MD simulation of ZnO/Pt slabs, iodide and lithium are depicted in orange and violet spheres. Zn, O, Pt, C, Emim⁺ and I-Emim are depicted in silver, red, yellow, pink, tan, and green spheres, respectively. ACN molecules are illustrated as light blue sticks (B).

In the first region, as indicated by the density profile diagram, a significantly higher density of ACN molecules and Emim cations was observed compared to that of lithium, iodide, and I-Emim ions. The aforementioned observations indicate that there was a strong interaction between the majority of ACN solvent atoms and the ZnO surface in the ZnO/Pt model. Additionally, Figure 5B snapshot provides further evidence by demonstrating that the first adsorption layer on the ZnO surface predominantly comprises of ACN solvent. Close to this layer, Emim cation ions were observed to be more abundant compared to other ions. The accumulation of Emim⁺ cations at the ZnO photoanode interface competitively blocks I₃⁻ adsorption sites, effectively suppressing dark current through electron I₃⁻ recombination, as confirmed by our J-V characteristics. ACN molecules are accumulated at the ZnO surface while being interacted by Emim cations. Moving on to region 2 near the Pt slab, we observed significant peaks corresponding to higher lithium ions concentration on the platinum surface compared to I-Emim, iodide ions, and Emim ions. This investigation highlighted a robust interaction between lithium ions and platinum surfaces. Furthermore, Figure 5A,B illustrated that in this region, first adsorption layer primarily consists ACN solvent. Moreover, a higher density of lithium ions compared to other species are observed.

The radial distribution function (RDF) is a statistical tool that characterizes the probability of finding an atom at a specific distance from a reference atom in a system of particles, obtained by averaging over the simulation trajectory. In the study of electrolytes, another helpful approach to probe their distribution is by calculating the RDFs of the electrolyte between parallel slabs. This technique can provide insights into the spatial arrangement and intermolecular interactions of different species in the electrolyte.

In this section, the RDFs of electrolyte atoms about Pt and ZnO slabs are studied. The investigated ACN atoms were labeled as H, C1, C2, and N (see Scheme 2). Hydrogen atoms from the Emim cation were labeled with H12, H13, and H14. The anion of the EmimI complex was depicted with I-Emim label. Additionally, lithium and iodide ions were denoted by Li and I label, respectively (refer to Scheme 2 for corresponding labels).

Scheme 2.

The structure of EmimI and ACN molecules, including atom labeling.

Investigation of interactions between electrolyte solution species

The interactions between imidazolium ring hydrogens with I-Emim and ACN atoms are shown in Fig. 6A–C. The results indicate that sharp first peaks between H12, H13, H14 and I-Emim are at 0.25 nm, 0.29 nm and 0.33 nm, respectively. The first peaks of H12, H13, and H14 with hydrogen atoms located at 0.18 nm, 0.21 nm, and 0.21 nm, respectively. These findings signify a strong correlation between the Emim cation and I-Emim as well as acetonitrile solvent components. In Fig. 6D and Table S2, significant peaks were located for I-Emim when it interacted with H atoms and C1 atoms at distances of 0.33 nm and 0.41 nm, respectively. These findings suggested a strong correlation between I-Emim and the ACN solvent. This analysis highlighted the pronounced affinity of I-Emim to interact and dissolve with ACN. For a comprehensive overview of various interactions between imidazolium hydrogen atoms (H12, H13, H14), I-Emim, LiI and ACN atoms (C1, H, N, C2), refer to Table S2.

Fig. 6.

RDFs between of H12, H13 and H14 with ACN atoms and I-Emim ions denoted as (A–C), respectively. RDFs of I-Emim ions with ACN atoms (D).

Figure S5A and Table S2 provided insightful observations regarding the interactions between iodine ions and Emim. Notably, three robust imidazolium hydrogens interactions with LiI anion, I…H12, I…H13, and I…H14, were identified at distances of 0.23 nm, 0.31 nm, and 0.31 nm, respectively. These findings suggest a strong correlation between the iodine ions of LiI and Emim cation. The analysis of Figure S5B and Table S2 yielded significant findings regarding the interactions between Li ions and various species. We observed high intensity first peaks for Li…H and Li…C1 interactions at 0.18 nm and 0.26 nm, respectively. These findings indicate significant interactions between Li ions and ACN molecule. In contrast, the peaks corresponding to Li interactions with C2 and N atoms of ACN are located at higher distances and lower intensities. Additionally, a relatively lower probability peak was observed for Li…I-Emim at 0.25 nm.

Investigation of electrolyte solution with solid surfaces

Figure 7A,C demonstrates that H12…Zn, H12…O, H14…Zn and H14…O peaks appearing at shorter distances than H12…Pt and H14…Pt. These peaks indicate significant interactions between the Emim cations and the ZnO surface. Furthermore, Fig. 7B revealed that H13 interactions with both slabs are weaker due to the steric hindrance of ethyl group. In Fig. 7D, it was evident that I-Emim peaks are located at a considerable distance from the surface atoms of Zn (0.87 nm), O (0.87 nm), and Pt (0.62 nm). As a result, weaker interactions were observed between I-Emim and the photoanode and cathode surfaces.

Fig. 7.

RDFs of hydrogen atoms within Emim, specifically H12, H13 and H14 were analyzed in relation to surface atoms (Zn, O and Pt) denoted as (A–C) respectively. RDFs of I-Emim ions of EmimI complex with surface atoms (Zn, O and Pt) (D).

In Fig. 8A, we observed a sharp first peak between H…Zn, located at 0.18 nm, as well as H…O peak at 0.21 nm. However, a shoulder indicating a less probable interaction was observed for H…Pt. Furthermore, in Fig. 8B, sharp peaks were observed for N…Pt interactions at a distance of 0.21 nm, while weaker interactions are obvious for N…Zn and N…O. In the investigation of other ACN atoms interacting with surface atoms, we observed significant interactions between C2…Pt, C2…Zn, and C2…O at distances of 0.27 nm, 0.37 nm, and 0.40 nm respectively in Fig. 8C. Additionally, both C1…Zn and C1…O interactions were located at a distance of 0.25 nm for each interaction. However, a weak interaction was observed between C1 and Pt (see Fig. 8D). These results indicate that ACN molecules interact with Pt slab from their N side and oriented toward ZnO slab by their methyl group.

Fig. 8.

RDFs of H, N, C2 and C1 atoms of acetonitrile solvent with surface atoms (Zn, O and Pt) denoted as (A–D) respectively (see the atom labeling of Scheme 2).

The influence of I ions on the surfaces of the photoanode (ZnO) and cathode (Pt) was investigated as well (see Fig. 9A and Table S3). The observed peaks were not well structured and the correlations are rather long range with a first peak at 0.60 nm, 0.59 nm, and 0.54 nm for interaction of iodine ions with Zn, O, and Pt, respectively. These observations indicate correlations between iodine ions and the solid slabs are not substantial, which can be attributed to the substantial interactions between iodide ions with Emim cations (see Figure S5). Therefore, significant role played by ILs when studying electrolytes impact on surface interactions. As shown in Fig. 9B, the interactions of Li ions with Pt, O and Zn are located at 0.19 nm, 0.21 nm and 0.25 nm, respectively. These interactions have relatively sharp peaks characteristic of solid systems. These findings indicate a significant correlation and accumulation of Li ions on the ZnO (photoanode) and Pt (cathode) slabs.

Fig. 9.

RDFs between I and Li atoms with surface atoms (Zn, O and Pt) denoted as (A) and (B), respectively. (see Scheme 2).

Based on the data presented in the tables and graphs of RDFs, notable peaks were identified at short distances, specifically arising from interactions between Emim cation hydrogen atoms and zinc oxide oxygen atoms. The first peak was pinpointed at 0.18 nm, underscoring the robust hydrogen bonding between Emim cation hydrogen and the zinc oxide surface. These findings suggest a compelling proximity of the electrolyte ions to the surface due to this strong interaction. Additionally, the interaction between Li and O, occurring at a short distance of about 0.18 nm, further highlights the significant intermolecular interactions between Li ions and the ZnO surface. The RDFs of Li with a ZnO slab resemble solid systems with low dynamics, while the interactions of Li with Pt show higher dynamics characteristic of a liquid system. In contrast, the RDFs of iodide ions demonstrate weak interactions at long distances.

The RDF plots in Figure S6 illustrate the interactions of C1…Pt, Li…O, I(Emim)…C1, and H14(Emim)…Zn. These specific atomic interactions were investigated over time intervals of 10, 15, 20, 25, 27, and 30 ns to ensure molecular and atomic equilibrium within a 30-ns timeframe. The analysis demonstrated that the positions of the first peak in each RDF remained consistently fixed throughout all the time periods. This consistent positioning of the peaks signifies the achievement of atomic and molecular equilibrium within the simulated structures within the 30-ns duration. This observation indicates that the system attained a stable state where interactions between atoms and molecules were balanced and sustained throughout the simulation period.

Mean-squared displacements (MSDs) measure the average squared distance that a particle moves away from its starting position during a given time interval τ. In this study, simulated trajectories were used to obtain the MSD values.

1

where is the location of the center of mass of the ith particle at time t, and the angular brackets indicate an ensemble average over time origins. The MSD analysis provides insight into the dynamics of the confined electrolyte solution, including the diffusion coefficient and the mobility of the electrolyte ions. By analyzing the MSD curves, it is possible to determine the diffusion behavior of the electrolyte ions within the confined system. This information is essential for understanding the transport properties of electrolyte solutions in a variety of applications, including energy storage devices and biological systems. To examine the component of MSD parallel to the vector, which is specified in the xy-plane (parallel to the ZnO slab) and in the z-direction, the following equation was used:

2

The diffusion coefficients of molecules, D_i, can be determined from the slope of the linear domain of the MSD curves obtained from simulation trajectories. This is related to the Einstein relation, which establishes a relationship between diffusion coefficient and the mobility of particles in a medium. The diffusion coefficient, D_i, was calculated from the MSD plots using the following equation:

3

Additionally, the molar electrical conductivity of the ionic species in the electrolytes was calculated using the Nernst–Einstein Eq. ¹¹⁰:

4

In this equation, and represent the absolute values of the valencies of the cations and anions, respectively. is the Faraday constant, and is the ideal gas constant. It is worth noting that for the calculation of (m²/s) (in which ( +) and ( −) subscripts refer to the cations and anions, respectively) and Λ (S m²/mol), all units of the parameters in the above equations were converted to SI units.

This section focuses on a detailed study of the dynamics of influencing species in the DSSC electrolyte. Hereafter, the in the electrolyte (LiI anion and EmimI anion) will be collectively referred to as IE due to their similar nature.

To study the dynamics of ACN, lithium, IE, and Emim species in the electrolyte mixture, we compared MSD curves of these species during the last 30 ns of MD trajectories. Figure S7A shows that the simulated MSD (in bulk) of ACN solvent is significantly higher than the electrolyte ions. According to Table 8, diffusion coefficient of electrolyte components changes in the order of When the electrolyte solution is confined between solid surfaces, the MSD curves depicted in Figure S7B,D reveal that the dynamics of the acetonitrile solvent show notably higher values in both the xy-plane and the z-direction compared to the other ions within the electrolyte. Additionally, the mobility of IE ions is significantly higher in both directions when contrasted with Emim and Li ions. By analyzing the linear parts of the MSD curves, all diffusion coefficients were computed. To investigate the impact of the electrolyte containing EmimI when confined between the solid surfaces of ZnO/Pt model, the diffusion coefficients of the electrolyte in the xy-plane and the z-direction are depicted in Fig. 10. The findings indicate a significant decrease in the diffusion coefficients of both the solvent and ions when the electrolyte is confined between surfaces. As illustrated in Fig. 10, the diffusion coefficients in the xy-plane for ACN, Li, IE, and Emim are 8.8 × 10^–9 m²/s, 0.58 × 10^–9 m²/s, 1.38 × 10^–9 m²/s, and 0.69 × 10^–9 m²/s, respectively. In the z-direction, the diffusion coefficients are 6.32 × 10^–9 m²/s, 0.19 × 10^–9 m²/s, 0.51 × 10^–9m²/s, and 0.24 × 10^–9 m²/s for ACN, Li, IE, and Emim, respectively. The lower diffusion coefficients observed for ions like Li and Emim suggest strong interactions and correlations of these ions with the ZnO/Pt surfaces. These results are in line with RDF results that was discussed earlier. The diffusion coefficients of lithium and iodine ions have been extensively measured using experimental techniques in diverse organic solvents, such as acetonitrile. Across all experimental datasets, these coefficients consistently fall within the range of 10^–9, indicating a high degree of agreement between the experimental and calculated data^111–113. This consistency is observed in both the simulated models conducted in the presence and absence of the EmimI ionic liquid.

Table 8.

The diffusion coefficients (10^–9 m² s⁻¹) for bulk (electrolyte solution) were calculated by analyzing the MSD-time curves within the NVT ensemble.

System	ACN	Li+	IE	Emim+
ZnO	15.16	0.79	1.88	1.27

Fig. 10.

Diffusion coefficient (10^–9 m² s⁻¹) of electrolyte in the xy-plane and the z-direction.

In Table S4, we examine the ionic conductivity of the EmimI-containing electrolyte solution in three orientations based on simulation results. The findings show that conductivity is higher in the bulk solution and the XY-plane compared to the Z-direction, indicating stronger electrolyte interactions occur along the Z-axis.

Investigation of electrolyte solution without EmimI

In Fig. 11A, the number density profile of the electrolyte without EmimI, derived from all-atom simulations conducted along the z-direction for ZnO/Pt systems, is illustrated. In the density profile diagram, a significantly higher density of ACN molecules and lithium ions was observed at the ZnO surface compared to iodide ions. This observation suggests a strong interaction between the ACN solvent atoms and the ZnO surface within the ZnO/Pt model. Moreover, the snapshot provided in Fig. 11B further supports these findings by demonstrating that the primary adsorption layer on the ZnO and Pt surface predominantly consists of ACN solvent. Notably, in proximity to the Pt slab, there are evident peaks indicating a higher concentration of lithium ions on the platinum surface in contrast to iodide ions. The density results of acetonitrile solvent, iodine, and lithium ions interacting with ZnO and Pt surfaces align with the electrolyte model containing EmimI. One noticeable distinction between the two models is the elevated density of Emim cations in contact with the ZnO surface compared to other ions like lithium and iodine. This disparity can potentially lead to a closer proximity of the electrolyte containing EmimI to the zinc oxide surface. These findings hint at the enhanced effectiveness of the electrolyte with EmimI, suggesting potential improvements in DSSC efficiency.

Fig. 11.

The number density profiles of electrolyte solutions without EmimI of a ZnO/Pt model at 300 K (A). The snapshot from the MD simulation of ZnO/Pt slabs, Zn, O, Pt and C are depicted in silver, red, yellow and pink spheres, respectively. Iodide and lithium are depicted in orange and violet spheres. ACN molecules are illustrated as light blue sticks (B).

As depicted in Figure S8A, notable interactions and distinct peaks were observed for H…Zn and H…O at positions of 0.18 nm and 0.21 nm, respectively. Additionally, a prominent primary peak for N…Pt was detected at a distance of 0.21 nm (refer to Figure S8B). The placement of C2 at a distance of 0.27 nm from the Pt surface is illustrated in Figure S8C. In Figure S8D, sharp peaks for C1…Zn and C1…O were identified at a distance of 0.25 nm. For a comprehensive analysis of peaks related to H…Pt, N…Zn, N…O, C2…Zn, C2…O, and C1…Pt, please consult Table S4.

In Figure S9A, weaker interactions of iodine with the surfaces of zinc oxide and platinum were noted. Contrastingly, sharp peaks for Li…O and Li…Zn were detected at distances of 0.21 nm and 0.25 nm in Figure S9B. Furthermore, sharp peaks were observed for the iodine interactions with hydrogen and carbon (C1) at distances of 0.35 nm and 0.43 nm, respectively (refer to Figure S9C). For a detailed analysis of the peaks related to Li…Pt, I…N, I…C2, Li…C2, Li…N, and Li…N from the RDFs depicted in Figure S9, please refer to Table S5. Moving to Figure S9D, two robust interactions displaying sharp peaks were identified at distances of Li…H and Li…C1, at distances 0.18 nm and 0.26 nm, respectively.

As delineated in Table S6, a significant disparity in the positioning of the RDF peaks is observable between the electrolyte incorporating EmimI and the one devoid of EmimI. The results suggest that in the presence of EmimI, there is a notable reduction in the interatomic distances compared to the system without EmimI. Specifically, the data highlights that within the EmimI-containing system, the distances between iodine ions and the surface of zinc oxide, as well as lithium and the surface of platinum, exhibit a decreased distance relative to the EmimI-absent system.

Analysis of the MSD curves for electrolyte solutions without EmimI in bulk, as well as in the xy-plane and z-direction during a 30 ns simulation, illustrated in Figure S10A–C, indicates a notably higher MSD for the ACN solvent in comparison to the Li and I ions.

Table S7 and Fig. 12 present the diffusion coefficients for the bulk, xy-plane, and z-direction. The coefficients for the bulk are 15.14 × 10^–9 m²/s, 0.84 × 10^–9 m²/s, and 1.83 × 10^–9 m²/s for ACN, Li, and I, respectively, similarly, in the xy-plane and z-direction, they are 10.34 × 10^–9 m²/s, 0.63 × 10^–9 m²/s, 1.33 × 10^–9 m²/s, and 4.81 × 10^–9 m²/s, 0.21 × 10^–9 m²/s, and 0.49 × 10^–9 m²/s, respectively. The dynamics of the electrolyte components (ACN, Li, I) in the three investigated spaces followed the order of bulk > xy-plane > z-direction. The higher diffusion coefficients observed in the xy-plane compared to the z-direction suggest that electrolyte mobility in the surface regions is influenced by xy-plane diffusion. A comprehensive analysis of the MSD and diffusion coefficients in electrolytes containing and excluding EmimI reveals that the iodide ions present in EmimI and lithium iodide (IE) significantly enhance the dynamics and diffusion characteristics of these ions.

Fig. 12.

Diffusion coefficient (10^–9 m² s⁻¹) of electrolyte without EmimI in the xy-plane and the z-direction.

Iodide ions play a crucial role in DSSCs by facilitating the iodide/triiodide redox couple, which is essential for the regeneration of light-adsorbing dyes after they have lost an electron. Their high mobility, coupled with their ability to exist in multiple oxidation states, substantially improves charge transfer mechanisms and electrical conductivity within the electrolyte^114,115. This enhancement leads to a reduction in resistive losses, ultimately elevating the overall efficiency of the DSSCs. Furthermore, iodide ions contribute to the stability of the electrolyte through ion exchange reactions, which help maintain the balance of ionic species. They also enhance electron transfer processes at the photoanode, promoting more efficient operation of the DSSC. Collectively, these factors exert a positive influence on the FF, V_oc, and efficiency of DSSCs, resulting in improved energy conversion rates^114,116. Consequently, the incorporation of EmimI in the electrolyte optimizes these processes and parameters for iodide ions in DSSCs.

The data shown in Table S4 highlight the positive impact of incorporating EmimI into the electrolyte on its conductivity. Specifically, the presence of EmimI facilitates ion transport within the electrolyte, which is crucial for effective charge transfer in DSSCs. The enhanced conductivity is attributed to the near-planar geometry of the imidazolium ring in EmimI⁵⁴, which facilitates π-π stacking interactions with the photoanode surface, improves charge transfer efficiency, reduces steric hindrance for ion mobility, and enables competitive adsorption at the semiconductor-electrolyte interface to suppress charge recombination, while the flexible alkyl side chains (ethyl/methyl groups) maintain optimal electrolyte viscosity for effective pore penetration. In comparison, the electrolyte lacking EmimI displays lower conductivity, which can hinder charge transport and result in decreased efficiency⁵⁸. In summary, the inclusion of EmimI not only boosts the electrolyte’s conductivity but also plays a vital role in optimizing the performance of DSSCs, making them more effective at converting sunlight into usable electrical energy. This finding underscores the importance of electrolyte composition in the design of high-performance solar energy devices.

The force field was validated by comparing simulated shear viscosity values with experimental data. Shear viscosities were computed using the Green–Kubo formalism^117,118, which relates viscosity to the time autocorrelation function of the pressure tensor:

5

where η is the shear viscosity, V is the system volume, is the Boltzmann constant, T is the temperature (298 K), and denotes the αβ component of the pressure tensor. To ensure statistical reliability, ensemble averages were obtained from 25 independent trajectories for bulk systems, each 30 ns in length, simulated in the NVT ensemble⁴⁷. The simulated viscosity of the EmimI-containing electrolyte (0.82 mPa s) shows excellent agreement with experimental data (0.76 mPa s). Similarly, for the EmimI-free system, the calculated viscosity (0.59 mPa s) closely matches the experimental value (0.42 mPa s). These results confirm the accuracy of the force field in capturing the transport properties of the electrolyte systems.

Conclusions

Zinc oxide in two different morphologies (ZnO-S and ZnO-P) were successfully synthesized and characterized. Under the same conditions, it was shown that just 1h duration of immersion time for ZnO-S nanoparticles leads to the efficiency of 8.38%. Also, an immersion time of 3h is needed for ZnO-P nanoparticles to obtain the efficiency of 9.86%. These results indicate that the immersion time in the dye solution has a significant effect on the performance and efficiency of DSSCs. Through a comprehensive analysis of the structural and functional characteristics of these nanoparticles as photoanodes in DSSC, we determined that the superior efficiency and performance of ZnO-P, compared to ZnO-S, can be attributed to several key factors. The higher specific surface area of ZnO-P significantly enhances the adsorption of the N719 dye, which subsequently leads to improved overall efficiency. Additionally, ZnO-P exhibits fewer crystalline defects, which reduces electron trapping on the surface and decreases the likelihood of electron recombination. The observed photoluminescence quenching in ZnO-P could be attributed to improved electron mobility (reducing recombination losses), prolonged charge carrier lifetimes, and lower defect density. It is important to note that while the efficiency outcomes of these two photoanodes in DSSCs are remarkably similar, the minor differences observed can be attributed to the distinct structural and functional variations that have been carefully analyzed. Overall, the enhanced performance of ZnO-P represents a significant advancement in DSSC development. The findings from the IPCE and EIS analyses provide valuable insights that improve the understanding of the performance and efficiency of these two photoanodes.

MD simulations have provided a comprehensive analysis of the behavior of electrolyte ions, revealing that the addition of EmimI significantly enhances the excitability and dynamic properties of iodide ions. This improvement is attributed to the favorable interactions between EmimI and iodide, which facilitate better ion mobility and promote a more efficient charge transfer process. The electrical conductivity and ionic diffusion in electrolytes containing EmimI are superior to those in formulations lacking this ionic liquid. This increase in conductivity and diffusion rates directly contributes to an enhanced electron transfer process within DSSCs, promoting more efficient energy conversion and improved overall performance. The increased availability of iodide ions near the interfaces of the photoanode and cathode plays a crucial role in enhancing the kinetics of the redox reactions, ultimately leading to a greater accumulation of reactive species at these surfaces, thereby optimizing the performance of the DSSC. Furthermore, results revealed a notable contrast, emphasizing the impact of the Emim cation on solid surfaces. The accumulation of Emim⁺ cations at the ZnO photoanode interface serves a dual function: (1) they competitively block I₃⁻ adsorption sites, effectively suppressing dark current by impeding electron-I₃⁻ recombination (as evidenced by our J-V characteristics), while (2) simultaneously modulating the interfacial ionic distribution. Notably, Emim⁺ promotes closer proximity of both I⁻/I₃⁻ redox mediators and Li⁺ ions to the ZnO and Pt surfaces, respectively. This synergistic effect enhances dye regeneration kinetics while simultaneously reducing charge recombination losses, as confirmed by the improved photovoltaic parameters in our DSSC devices¹¹⁹. This configuration lead to the accumulation of triiodide molecules near platinum¹²⁰. Moreover, the presence of small Li⁺ ions on the ZnO surface induced a positive shift in the conduction band’s energy level (quasi-Fermi level) due to Li⁺ strong electropositive nature. This shift widened the band gap between the dye’s excited state and ZnO conduction band, facilitating efficient electron injection from the dye to ZnO¹²¹. This configuration also enhanced the gathering of triiodide molecules near platinum. Conversely, in the EmimI-free model, the increased distance between iodine and lithium interactions with zinc oxide and platinum surfaces slowed down surface electron transfer rates, leading to a reduction in exchange current density at the photoanode-electrolyte interface.

Author contributions

OE: Methodology; Data curation; Analysis; Software; Writing—Original Draft; Investigations, ARZ: Supervision; Conceptualization; Methodology; Writing—Review and Editing; Investigations, HS: Supervision; Conceptualization; Methodology; Writing—Review & Editing; Investigations.

Funding

HS received funding from Iran National Science Foundation (INSF) under project No. 4031115.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.