A homo-tandem organic solar cell was fabricated using the ITO/ZnO NPs/PEIE/PTF5:Y6-BO/MoOx/Au/ZnO NPs/PEIE/PTF5:Y6-BO/MoOx/Ag device architecture [31]. The solutions of PTF5:Y6-BO photoactive materials (dissolved in o-xylene solvent) were stirred overnight at 70°C before being subjected to filtration using a 5 mm PTFE filter. Following the filtration, the solutions were heated on a hot stirring plate at 40°C before being spin-coated onto precleaned, glass ITO substrates within a N₂-filled glove box. Before the spin-coating, the ITO substrates underwent thorough cleaning with deionized water containing detergent, followed by cleaning with acetone and isopropyl alcohol. This cleaning cycle was repeated three times. Subsequently, the substrates were exposed to ultra violet ozone (UVO) treatment for 15 minutes. Next, a solution containing ZnO nanoparticles (2.5 wt% in 1-butanol) was spin-coated onto the ITO substrate at 3000 rpm for 60 seconds, which was followed by thermal annealing at 100°C for 10 minutes. A thin layer of ethoxylated polyethyleneimine (PEIE; 0.4 wt% polyethyleneimine ethoxylated solution in 2-methoxyethanol) was then spin-coated at 5,000 rpm for 30 seconds onto the ZnO electron transport layer, after which it was annealed at 100°C for 10 minutes. Subsequently, the substrates were transferred to the N₂-filled glove box, where photoactive solutions were spin-coated onto them at various speeds to achieve the desired thicknesses. After the photoactive films were dried for 2 hours inside the N₂-filled glove box, they were transferred to a thermal evaporation chamber. In the evaporation chamber, a 100 nm layer of MoOx and a 1 nm layer of Au were thermally evaporated onto the substrates. Next, another layer of ZnO NPs was spin-coated onto the substrates and annealed at 100°C for 10 minutes. The PTF5:Y6-BO active layers were then fabricated using the same process mentioned above. Finally, sequential evaporation under a high vacuum was carried out to deposit a 100 nm layer of MoOx followed by a 300 nm layer of Ag. When PEDOT:PSS was used as an interconnecting layer (ICL), the device fabrication proceeded according to the aforementioned description, with the exception that PEDOT:PSS was used as the connecting layer. The PEDOT:PSS was formulated using isopropyl alcohol (3:1) as the solvent. The tandem solar cell device had an active area of 81 mm². Tandem devices underwent characterization through J–V curve measurements using a Keithley 2400 source meter and a K201 LAB55 solar simulator (McScience Inc., Republic of Korea). The characterization took place under simulated irradiation of 100 mW cm^–2 from an arc Xe lamp equipped with an AM 1.5G filter. Light intensity was calibrated with an NREL-certified silicon diode combined with a KGI optical filter, and the simulator’s irradiance was measured using a calibrated spectrometer. External quantum efficiencies (EQEs) were determined using a spectral measurement instrument (K3100 IQX, McScience Inc., Republic of Korea), which utilized monochromatic light from a 100 W Xe arc lamp, filtered through an optical chopper and a monochromator. The cross-sectional structure of a tandem device was observed using a crossbeam 550 focused ion beam (Carl Zeiss, Germany), and transmission electron microscopy (TEM) images were obtained on a FEI Tecnai TF20 (Philips, Netherlands) instrument.

To fully characterize the SZB and to ensure its voltage compatibility with the tandem solar cells, the SZB was fabricated and pretested, prior to the construction of the integrated energy device, in the following manner: the Ag cathode of the OPV was the thin film described in Section 2.1. To produce a structure as close to this as possible, a 300 nm thick layer of Ag was deposited on stainless steel foil (25 μm thick, type 304, Alfa Aesar, U.K.) using an e-beam evaporator (FC2000, Temescal, USA) as the cathode, and Zn foil (250 μm thick, 99.98% pure, Alfa Aesar, U.K.) was used as the anode. The gel electrolyte was prepared as follows: 5 g of deionized water and 0.35 g of polyvinyl alcohol (PVA, C₂H₄O, M.W. 130,000, 87%–89% hydrolyzed, Sigma Aldrich Co., USA) were placed in a glass vial and stirred at 80°C at 500 rpm for 3 h to dissolve and obtain a clear and homogeneous PVA solution. Then, 1–4 g of zinc chloride (ZnCl₂, anhydrous, ≥99.8% pure, Alfa Aesar, U.K.), 0–1 g of glycerol (C₃H₈O₃, ≥99% pure, Alfa Aesar, U.K.), 0.1–0.5 g of glutaraldehyde (GA, C₅H₈O₂, 50 wt% in H₂O, Sigma Aldrich Co, USA) were added to the prepared PVA solution. A gel precursor was stirred at 300 rpm for 1 hour at room temperature to obtain a homogeneous solution. The prepared gel precursor was then poured into a petri dish and dried for 24 hours in a vacuum desiccator at room temperature. The thickness of the resulting gel electrolyte was about 3 mm. The preparation of the gel electrolyte is presented schematically in Fig. S1.

The rate characteristics, internal resistance, and cycling characteristics of the SZB were measured as follows. The SZB underwent two charge/discharge cycles in the range of 0.5 to 1.1 V at a rate of 5 mA cm⁻², during which the battery was reversibly charged and discharged. At this point, the charge was terminated when only 12% (10 μAh cm⁻²) of the total mass of the Ag thin film on the cathode was reacted to ensure the stable operation of the solar cell upon integration. This corresponds to applying a charge current for approximately 7.2 seconds and reaching a charge voltage of approximately 1.1 V. The discharge was terminated when the voltage reached 0.5 V. The rate was characterized at only high charging rates, focusing on the charging situation of a battery for an integrated energy device. The charge currents ranged from 1 to 10 mA cm⁻², and the charge voltages reached were between 1.01 and 1.20 V when only 12% of the total mass of Ag was reacted. The discharge current of 1 mA cm⁻² was applied, regardless of the charge currents, and the lower cut-off voltage during the discharge was 0.4 V. All tests were repeated three times with the same charge current. The difference between the operating voltage at the completion of charging and the open-circuit voltage after 10 seconds of rest time after the completion of charging was divided by the applied charging current density to calculate the internal resistance (DC-IR) of the cell. The cycling characteristics were evaluated at currents of 10 mA cm⁻² (charge) and 0.1 mA cm⁻² (discharge) in the range of 0.5–1.2 V. The charging cut-off method was the same as previously described. All the electrochemical experiments were performed using a potentiostat (BCS 810, Bio-Logic, France) at a temperature of 25°C.

A VIED was fabricated by combining an OPV and SZB as follows (Fig. S2). First, copper foil was used to extend both the anode and cathode of the OPV, respectively, and silicone tape was used to seal the entire area except the surface of the Ag electrode to prevent the other components from being exposed to the air and battery electrolyte applied later. Next, the battery gel electrolyte fabricated in Section 2.2 was placed on the exposed Ag surface. The Zn foil was placed on top of the gel electrolyte. The Zn foil was connected to a stainless steel foil (10 μm thick, type 444, Woori Science, Republic of Korea) by welding it with a spot welder (HL-113, Hanil Electric). The fabrication of the VIED was completed by encapsulating the whole component in silicone tape. The prepared VIED is a complete, three-electrode structure that comprises the anode of the OPV (the ITO electrode), the anode of the SZB (Zn electrode), and a shared electrode (an Ag electrode), which serves as the cathode for both the OPV and the SZB. The photo-charging and electrical discharging of the VIED were performed as follows. The ITO electrode and Zn electrode were electrically connected for the photo-charging. A switch was inserted between the two electrodes, and it was turned on for photo-charging and off for the electrical discharging. Photo-charging was performed by illuminating the ITO electrode of the VIED with 1 sun (A.M. 1.5, 100 mW cm⁻²) using a solar simulator (Class A, PEC-L01, Peccell Technologies, Japan). In this case, the light was irradiated for 3 seconds to prevent causing OPV failure by completely reacting the Ag. This restricts the charge capacity to approximately 9%–13% of the total Ag used during the SZB charging process. The charge curves of the SZB in the VIED were measured using a voltmeter, and the charge current values were calculated by dividing the voltage difference across the switch terminals by its resistance (0.2 Ω). Subsequently, the Zn electrode was disconnected from the ITO electrode, and the discharging of the SZB was performed at a rate of 0.1 mA cm⁻². All the electrochemical tests were conducted using a potentiostat (VMP3, Bio-Logic, France). The morphology of the Ag electrode after photo-charging was observed via scanning electron microscopy (SEM, MIRA3; TESCAN, Czech Republic), and the electrode’s chemical composition was analyzed by energy dispersive X-ray spectroscopy (Octane Elect Plus; AMETEK, USA).

A vertical arrangement of batteries and solar cells proves to be more effective than its planar counterparts for maximizing energy density per unit area/volume. However, such a vertical integration of tandem solar cells and secondary batteries poses several implementation challenges that must be addressed.

1) Voltage matching: When a conventional Li-ion battery is charged by an external device, its open-circuit voltage must be close to 4 V. However, the voltage achievable in a solar tandem configuration is around 2 V at best. Considering the need to set a solar cell’s maximum power voltage higher than the battery’s charging voltage to increase the charging efficiency, using secondary batteries that are capable of low-voltage charging is imperative in implementing vertical integration.

2) Current matching: To achieve high-efficiency, the current density generated by solar cells needs to match the appropriate C-rate of the battery. However, high-efficiency solar cells may generate excessively high current densities that exceed the requirements of the battery and impede PSE, especially challenging in vertical structures where only limited adjustments for current matching can be made to the solar cell area.

3) Process matching: Ensuring no performance degradation in subsequently fabricated components occurs due to the sequence of fabrication processes is crucial for vertically connecting two components. Although directly using metals like Ag and Au for upper electrodes, common in tandem solar cells, may be challenging in secondary batteries, substituting other metals could lead to problems with efficiency. Moreover, the presence of batteries could potentially affect the efficiency of underlying solar cells due to exposure to heat and solvents during the battery manufacturing processes.

To address these problems, this study employs the SZB. These batteries offer a charging voltage range of 0.6 to ~2.0 V, which enables easy charging-discharging in tandem configurations, where the maximum power voltage of the solar cell exceeds the charging voltage by 0.2 V or more and facilitates voltage matching. Additionally, the SZB enables stable charging across a wide range of C-rate and eliminates the need to consider solar cell current density for current matching in vertical structures. The use of SZB also simplifies manufacturing processes by allowing the top Ag electrode in organic solar cells to be used as the battery’s active material, which alleviates the need for high-temperature, high-pressure processes, thereby ensuring convenient coupling processes that do not affect the performance of underlying solar cells during integration.

To achieve a highly efficient VIED, we focus on the principle of the power generation of a photovoltaic cell combined with an energy storage device. Because these two devices are directly connected without a charge controller, they affect each other’s electrical characteristics. Fig. 1a illustrates the principle that the power density of the photovoltaic cell (P_PV) is determined by the secondary battery in a VIED. When solar irradiance is constant, that is, in a situation where we have a single plot of the current density (J_PV) vs. operation voltage (V_PV), the P_PV is determined by both the equilibrium voltage (V_B) and the overvoltage (η = R_B × J_OP) of the secondary battery during the operation of the VIED. Furthermore, it can be observed that the operation voltage of the secondary battery (V_OP) determined by V_B and η (or R_B) should match the maximum power point of the photovoltaic cell (V_MPP) to maximize the PSE of the VIED by charging the secondary battery at the V_MPP. To achieve this, two design strategies are needed for the secondary battery. First, the V_B should be as closely aligned as possible to the V_MPP while V_B < V_MPP is ensured, and second, the cell internal resistance of the secondary battery (R_B) should be minimized, thereby minimizing η, while V_B + η ≈ V_MPP is simultaneously ensured. Accordingly, we designed an electrolyte for an SZB to minimize R_B and to align V_B +η with V_MPP, And, as the third strategy for the photovoltaic cell, a photo active material for the OPV to increase its power conversion efficiency (PCE) and fill factor (FF) is needed in order to realize a high-efficiency VIED. The above is illustrated in Fig. 1b.

We used non-fullerene, organic homo-tandem solar cells to enable the efficient charging of an alkali metalion battery within an integrated solar-battery hybrid energy device. These solar cells consist of two blends of the wide-band-gap donor polymer PTF5 and the nonfullerene acceptor Y6-BO, connected in series via a MoO_x/Au recombination layer. This material combination was chosen because of its balanced charge carrier mobility, its beneficial properties in low charge recombination dynamics, and its minimal reduction in efficiency with an increase in area, all of which make it advantageous in large area energy production/storage devices [31]. The device architecture, along with the chemical structure of the donor and acceptor materials, is depicted in Fig. 2a. Fig. 2b presents the energy diagrams of the tandem solar cell and the HOMO and LUMO energies of the individual components. Additionally, Fig. 2c presents the focused ion beam TEM image of the corresponding organic homo-tandem solar cell.

To increase the energy capacity of a solar cell-battery integrated device, it is necessary to adjust the area of the components, and for this purpose, the characteristics of the ICL of the solar cell are crucial. In the case of commonly used PEDOT/PSS interlayer materials, it is difficult to ensure the uniformity of the coating due to the difference between the surface energy of the interlayer and that of the active layer. This is especially true when the coating is applied over a large area, which leads to difficulty in realizing high-efficiency devices due to the pinholes. For the PEDOT/PSS ICL, the best performing device exhibited a PCE of 9.49%, with a short circuit current density (J_SC) of 9.07 mA cm⁻², an open-circuit voltage (V_OC) of 1.45 V, and an FF of 71.94% with an optimized active layer thickness of 70 nm/130 nm. These results are detailed in Table S1, and their corresponding J–V curves are depicted in Fig. S3.

To further enhance the performance, we switched from using PEDOT:PSS to MoOx/Au as the ICLs. This strategic shift was done to address the limitations imposed by PEDOT:PSS and to achieve better performance with the homotandem solar cells. Au can be uniformly and effectively deposited on the MoO₃ hole transport layer (HTL) without pinholes, which enables high reproducibility of the tandem cell implementation, even over large areas (Fig. S4).

With an optimized active layer thickness of 70 nm/130 nm, the top performing device achieved a PCE of 13.22%. This was accompanied by J_SC = 10.79 mA cm⁻², V_OC = 1.67 V, and FF = 73.13%, which may have been due to an improvement in interfacial contact. The J–V vs. power density characteristics of the best tandem solar cells are displayed in Fig. 2d, and the photovoltaic performance parameters are summarized in Table 1. The optimized J–V curves for different active layer thicknesses are presented in Fig. S5. The maximum EQE for this device surpassed 35%, which indicates a highly efficient photon to electron conversion. Even with larger electrode areas, high V_OC values and reasonable PCE values above 13% were achieved, which affirms the viability in this photovoltaic system for the implementation of an integrated hybrid energy system (Fig. 2e,f) [32,33]. In this study, a tandem cell based on a 13.22% efficiency was implemented; however, considering that the PTF5 material demonstrated efficiencies above 18% in single cells, it is expected that with further optimization of the tandem cell process, integrated energy devices having even higher efficiency can be produced.

The primary commercial SZB is known to have outstanding rate capability owing to the low charge transfer resistance of its Ag cathode and the high ionic conductivity of its liquid electrolyte. In addition, we have shown that SZB can also be used as a secondary battery to achieve a superb charging rate [34], making SZB ideally suited for high-efficiency integrated energy devices. Unfortunately, however, we have performed dozens of repeated experiments that have shown that when the liquid electrolyte used in conventional batteries is placed over the Ag electrode of a solar cell, the liquid electrolyte penetrates into the solar cell within a relatively short time, even if the Ag electrode is thick (~3 μm) and dense, and it damages the photoactive layer [38], which results in the solar cell losing its functionality. To solve this problem and to ensure the reliability and durability of the VIED, it is necessary to use a solid electrolyte, or at least a gel-type, quasi-solid-state electrolyte, that has low fluidity and almost no ability to penetrate into the solar cell. However, an SZB with a solid (or gel) electrolyte inevitably has high electrolyte resistance and electrode/electrolyte interfacial resistance relative to those with a liquid electrolyte. To ensure the stability of OPVs under these circumstances, while also ensuring that the SZB has as small resistance as possible, we design a gel electrolyte for the SZB by considering the following three factors: electrolyte fluidity, electrolyte resistance, and electrode/electrolyte interface resistance.

To produce a low fluidity electrolyte that does not easily penetrate the OPV, we prepared a gel electrolyte by varying the concentration of the hardener glutaraldehyde (GA). GA reduces the electrolyte fluidity by increasing its concentration, as it chemically combines with the alkyl groups of the gelling agent, polyvinyl alcohol (PVA). This significantly suppresses the penetration of electrolyte into the OPV; however, as the concentration of GA increases, the ductility of the gel electrolyte decreases, which causes the electrolyte to become brittle during the fabrication of the SZB. After checking the brittleness of the gel electrolyte by varying the concentration of GA, with the concentration of salt (ZnCl₂) fixed and no plasticizer (glycerol) added, the ratio of ZnCl₂:GA was found to be 2:1 (the amount of hardener was 0.5 g) for the gel electrolyte to not crumble when making SZBs. When a gel electrolyte prepared under these conditions was placed on the Ag electrode of the OPV, the OPV remained fully functional for a relatively long time (~40 hours) (Fig. S6), which confirmed that the designed gel electrolyte could be suitable for VIEDs in terms of its stability and durability. However, as shown in Fig. S7, the electrolyte resistance and the electrode/electrolyte interface resistance increased to 293% and 558%, respectively, relative to the original liquid electrolyte. Therefore, the next task was to adjust the composition of the salt and the plasticizer to reduce the electrolyte resistance and interfacial resistance without significantly compromising the stability and durability.

We first observed a change in the electrolyte resistance and in the electrode/electrolyte interface resistance related to salt concentration, and we found that both values decreased with an increase in salt concentration (Fig. S8). We then confirmed that if a gel electrolyte containing more than a certain amount of salt (ZnCl₂:GA = 8:1, where the amount of ZnCl₂ was 4 g) was applied to the Ag electrode of the OPV, the solar cell stopped working within a few minutes. Through repeated experiments, we additionally confirmed that, to maintain the functionality of the OPV for a long time (>20 hours), the maximum weight of salt that could be used was 2 g (where ZnCl₂:GA = 4:1). When the gel electrolyte was prepared at this ratio, the electrolyte and interfacial resistances significantly decreased (~43.5% and ~46.8%, respectively), but they were still too high for the battery to become highly efficient. In particular, relative to the initial increase in interfacial resistance when the gel electrolyte was applied, the decrease in interfacial resistance was quite small, which was thought to be due to the still inferior interfacial contact between the electrode and the electrolyte. In general, an external force can enhance the physical contact properties, but in our VIEDs, external forces can destroy the Au and MoO₃ layers of the solar cell and cause device failure, so improving the interfacial contact properties by design alone without external forces is necessary. Thus, with the amounts of hardener and salt fixed at the values that we had determined, we sought to enhance the contact properties of the electrode/electrolyte interface by improving the adhesion of the electrolyte by adjusting the plasticizer [39].

Fig. S9 shows the changes in electrolyte resistance and electrode/electrolyte interface resistance according to the plasticizer concentration. The effect of plasticizers on resistance is not yet well understood, but phenomenologically, it appears that the electrolyte resistance tends to increase monotonically with an increase in plasticizer, where the electrode/electrolyte interface resistance decreases and then increases with greater plasticizer concentration. The electrolyte and interface resistances of all the designed electrolytes are summarized in Fig. 4a,b, respectively. The expected operating voltages and efficiencies obtained when these electrolytes are used in the SZB were calculated, assuming that the sum of the two resistances was equal to the total resistance of the battery (Fig. 4c,d). Based on these experiments and analyses, the optimal gel electrolyte component ratio for high-efficiency and low-resistance SZBs for VIEDs was determined to be ZnCl₂:GA:glycerol = 2:0.5:0.1 (Table 2, condition 6). We confirmed that the gel electrolyte prepared under this condition was stable (>10 hours) when placed on the Ag electrode of the OPV, so this electrolyte was temporarily adopted as the gel electrolyte of the SZB for VIEDs to determine whether it satisfied the second design strategy of the secondary battery (V_B + η ≈ V_MPP).

An SZB for high-efficiency, high-stability VIEDs was fabricated according to the electrolyte composition established in the previous sections, and its electrochemical properties were evaluated. Fig. 5a shows the voltage curve for the charging and discharging of the SZB at a current density of 1 mA cm⁻². Under the SZB charging conditions used in this study, AgCl formation occurs at 0.22 V vs. SHE at the Ag cathode (Eq. 4-1), and the Zn reduction reaction takes place at −0.76 V vs. SHE at the Zn anode (Eq. 5), which result in a thermodynamic cell voltage of 0.98 V (Eq. 6-1). The SZB charge curve shows an almost constant anodic overvoltage over the entire charging region, which results in charging under a constant voltage of approximately 1.03 V. This voltage value is close to the V_MPP of the OPVs for VIEDs, which is approximately 1.25 V, and the cell voltage comes closer to 1.25 V as the overvoltage increases with an increase in the charging current. In other words, the combination of SZBs and OPVs proposed in this study is expected to charge the SZBs near the maximum PCE of the OPVs.

To determine the variation in overvoltage and charging voltage with the SZB charging current, we observed the change in the voltage curve when SZB was charged at different rates between 1 and 10 mA cm⁻² (Fig. 5b). The reason that 10 mA cm⁻² was chosen as the maximum charge current is that the OPV used in this study generates a current of approximately 10 mA cm⁻² when generated at the V_MPP (Fig. 2d). The overvoltage was observed to increase linearly with an increasing charge current, so that when charging at 10 mA cm⁻², the V_OP of the SZB was found to be around 1.2 V, which is only 0.05 V different from the V_MPP of 1.25 V for the OPV (Fig. 5c). In other words, as a result of the low-resistance design of the SZB, the fabricated SZB is expected to be charged near the V_MPP of the OPV when used in a VIED, which realizes our second design strategy for the SZB. Fig. 5d presents the change in the discharge capacity retention and in the energy efficiency when the SZB is charged and discharged under currents of 10 mA cm⁻² and 0.1 mA cm⁻², respectively. Although conventional SZBs are known to suffer from rapid capacity degradation during charging and discharging [40], the SZB in this work exhibited a very stable capacity and energy efficiency retention up to 500 cycles when it was charged and discharged under VIED operational conditions (i.e., limited SoC swing).

Fig. 6a schematically illustrates the structure of the VIED proposed in this work. The VIED is a three-electrode structure that contains an Ag electrode (middle) shared between the SZB (bottom) and the OPV (top). The operation of the VIED, i.e., the OPV generation–SZB charge, is shown in Fig. 6b. First, when sunlight is incident on the ITO electrode of the OPV, excitons are excited in the photoactive layer and separated to charge carriers (i.e., electrons, holes). When the switch between the OPV and the SZB is closed, the electrons flow from the photoactive layer through the electrode transport layer and ITO electrode to the Zn electrode, which is the anode of the SZB, whereas holes flow from the photoactive layer through the HTL to the Ag electrode, which is the cathode of the OPV and the cathode of the SZB. Accordingly, the potential of the (negatively charged) Zn electrode decreases and a reduction reaction (Eq. 5) occurs at the Zn/electrolyte interface, whereas the potential of the (positively charged) Ag electrode increases and an oxidation reaction (Eq. 4-1) occurs at the Ag/electrolyte interface, which results in the charging of the SZB (Eq. 6-1). Fig. 6c illustrates the scenario in which the charged SZB in the VIED is electrically discharged through the connected resistor. When the switch between the anode and cathode of the SZB is closed, an oxidation reaction occurs at the Zn/electrolyte interface (Eq. 5), whereas a reduction reaction takes place at the Ag/electrolyte interface (Eq. 4-1), which is opposite to the charging process. The voltage of the SZB drops rapidly and the discharge is terminated when all the AgCl formed during the charging is reduced to Ag.

The electrochemical behaviors of the VIED are presented in Fig. 7. The variations in the SZB voltage and charge current over time during VIED operation (OPV generation–SZB charge) are shown in Fig. 7a. The SZB voltage increases from 1.10 to 1.18 V, whereas the charge current decreases from 11.31 to 10.52 mA cm⁻². When the ranges of the SZB voltage and charge current during VIED operation are plotted against the J–V curve of the OPV, the values can be observed to be close to the V_MPP of the OPV and the current value at that time, respectively (Fig. 7b). Remarkably, the current of the OPV (10.9 mA cm⁻², closed red circle in Fig. 7b) calculated from the individual SZB characteristics (V_B = 0.98 V, R_B = 17.5 Ω) is within the range of the OPV current values measured during actual VIED operation. This strongly indicates that the SZB in the VIED has all the characteristics of an individual SZB. However, the surface morphology of the Ag electrode during SZB charging is shown in the inset of Fig. 7a, where it can be observed that the surface becomes slightly rougher as the charge capacity (i.e., formed AgCl amount) increases.

Fig. 7c shows the variations in the PCE of the OPV and the charge capacity of the SZB as a function of VIED operating time, as reproduced from the voltage and current curves of Fig. 7a. As the VIED operation time (i.e., the SZB charging time) increases, the SZB voltage rises and approaches the V_MPP of the OPV. As a result, the actual PCE of the OPV gradually increases, as can be observed in Fig. 7c. In addition, the charge capacity of the SZB increases almost linearly due to the relatively constant charge current during VIED operation, as shown in Fig. 7a. In this case, the average PCE of the OPV is 12.5%, which is 94.6% of the maximum PCE (13.22%). For reference, in conventional integrated energy devices that use supercapacitors as their energy storage system, the PCE of solar cells is approximately 50% of the maximum PCE [17,42,43]. In this study, it was possible to increase this value by 44.6%p compared with previous studies by enhancing the efficiencies of the OPV and SZB and by improving their voltage compatibility. The charge capacity was approximately 9.6 μAh, which means that approximately 12% of the total Ag was used during the charge process (i.e., SoC 12).

Fig. 7d presents the voltage curve measured during VEID operation with the repeated charging of the SZB by the OPV and the constant current discharging of the SZB by the potentiostat, which shows a very stable charging and discharging behavior for approximately 20 cycles. To further analyze the VIED performance, the discharge capacity, discharge energy, energy efficiency, and average discharge voltage of SZB over the charge cycles are shown in Fig. 8a. First, it is noteworthy that the average discharge voltage remains almost constant, regardless of the number of cycles (inset in Fig. 8a), suggesting that there was little change in the SZB resistance during charging and discharging, and furthermore that the integrity of the bulk and interface inside the SZB remained intact. The energy efficiency was quite high, ~90%, despite the device being charged under relatively high current, due to the excellent rate capability characteristics of the SZB. In addition, the average discharge capacity and energy of the SZB were determined to be approximately 9.6 μAh cm⁻² and 9.1 μWh cm⁻², respectively. Notably, when the energy per electrode area was converted to an energy per electrode volume and compared with the corresponding values for integrated energy devices of different structures reported in the literature, the VIED presented in this work was shown to have the highest energy density per volume (Fig. 8b). This is believed to be due to a combination of the high specific capacity of Ag, the unique three-electrode structure with shared substrate and active material, the low-resistance cell design, and voltage compatibility. Furthermore, owing to the application of SZBs with high energy density, the capacity was comparable with or higher than that of conventional powder electrodes [46], despite the use of thin film electrodes.

The PSE, which is one of the main performance indicators of integrated energy devices, refers to the ratio of the battery discharge energy to the irradiated solar energy, and it is calculated as follows [22].

where E_discharge is the discharge energy (Wh), P_input is the irradiation power (W), and t_charge is the irradiation time (hours). Fig. 9a exhibits the variation in the PSE of the VIED according to the SZB cycles. It shows a very high PSE having an average of 10.9% (maximum 11.2%). Fig. 9b compares the PSE of the VIED in this study with values previously reported for OPV-based integrated energy devices [17,28,41–50]. The VIED in this study has the highest PSE of any OPV-based integrated device in the literature.