Enhancing the indoor performance of organic photovoltaic devices: interface engineering with an aminobenzoic-acid-based self-assembled monolayer

Significant advances in the performance of organic photovoltaic (OPV) devices can facilitate their use in internet of things applications. However, achieving excellent photostability and high efficiency using stable, efficient OPV devices in indoor settings is considerably difficult. To address this issue, a zinc oxide (ZnO) electron transport layer (ETL) was modified with a self-assembled monolayer of 4-aminobenzoic acid (ABA) in the present study, and the impact of this modification was correlated with the indoor performance of an OPV device with the PM6:L8-BO photoactive layer. The ABA-treated ZnO ETL exhibited a significant reduction in the work function (from 4.51 to 4.04 eV), surface roughness (from 0.201 to 0.177 nm), and hydrophilicity of an indium-tin-oxide electrode; this aided in selectively extracting charge carriers from the device and minimizing trap-assisted recombination losses. Additionally, the ABA treatment of the ZnO ETL considerably enhanced the electron mobility and recombination resistance. It reduced the trap density, thereby enabling the ZnO/ABA-based device to achieve improved performance. Consequently, the ZnO/ABA-based device exhibited a noteworthy 14.68% higher maximum power output than that of the device without any ZnO surface modification under 1000 lx halogen (HLG) illumination (P out, max = 354.48 and 309 µA cm−2, respectively). Moreover, under thermal illumination conditions (1000 lx HLG lighting), the ZnO/ABA-based device sustained ∼74% of its initial power conversion efficiency over 120 h, significantly higher than its ABA-free equivalent (∼55%).


Introduction
The proliferation of low-power indoor internet of things (IoT) devices is anticipated to result in the deployment of millions of wireless sensors in the next decade [1], with approximately half being installed inside buildings.However, emerging IoT devices require a small amount of power for operation, face battery-related issues, and have a limited lifespan.Organic photovoltaic (OPV) systems-which exhibit an impressive output power density (>100 µW cm −2 ) [2], high efficiency, and notable stability-have drawn attention as a compelling photovoltaic technological solution to these IoT-related issues.Conventional indoor lighting systems (including light-emitting diodes (LEDs), fluorescent (FL) lamps, and halogen (HLG) lamps), which primarily drive IoT devices, emit confined spectra and a substantially lower irradiance intensity than that of outdoor solar conditions (AM1.5G, 1-sun) [3].Interestingly, OPV systems allow the tailoring of absorption spectra through modification of the active layer material, thereby enabling energy-level refinement and the realization of absorption coefficients sufficiently high for indoor light energy capture [4].Additionally, OPV systems offer aesthetic advantages including color versatility and flexibility [5], thereby catering to remarkably diverse IoT products [6].
Numerous studies have been conducted to improve OPV device performance, with approaches such as innovative material design and device structure engineering being targeted, specifically those involving interface optimization [7][8][9][10][11].These concerted efforts have yielded remarkable outcomes, with non-fullerene-acceptor-based OPV devices having achieved power conversion efficiencies (PCEs) exceeding 33% in controlled indoor settings [12].In terms of OPV device operation, a pivotal factor affecting the device performance is the judicious selection of an appropriate electron transport layer (ETL) material [8,13].This choice profoundly influences the energy alignment, optoelectronic, and surface characteristics of the device.Among the varied ETL materials investigated to date, including zinc oxide (ZnO), zinc tin oxide, and titanium suboxides, ZnO has emerged as the predominant choice owing to its desirable attributes including suitable energy levels, high electron mobility, transparency, and cost-effectiveness [14].Recently, inverted OPV cells with ZnO as the ETL exhibited an impressive indoor efficiency of 17.42% [15].However, ZnO has certain limitations such as instability in humid environments, given its tendency to undergo water-induced oxygen substitution [16], and sensitivity to ultraviolet (UV) radiation [17].Consequently, the practical utility of pristine ZnO as an ETL is questionable, necessitating the exploration of surface treatments to enhance the stability of OPV devices under humid conditions and thermal irradiation.
Strategies such as thermal annealing and UV-ozone treatment have been employed to manipulate the characteristics of each OPV device layer [18,19].In particular, self-assembled monolayer (SAM) construction shows promise as an effective, convenient technique in this regard.SAMs are formed by the chemical adsorption of organic molecule head-groups onto solid surfaces, followed by their spontaneous arrangement into monolayers.SAM-compliant organic molecules must comprise three essential components: anchor, spacer, and terminal groups.The anchor group considerably influences the work function (WF) of the substrate and contact resistance, whereas the length of the spacer group is crucial for electronically isolating one contact from another.The terminal group is responsible for the air-overlayer interface chemistry [20].Notably, SAMs produced using organic molecules with carboxylic-acid-containing anchor groups can enhance the stability of ZnO by enabling the Zn 2+ from ZnO to link with the H − from the carboxylic acid, thereby facilitating molecular adsorption [21].SAMs can serve as a protective barrier for the ZnO ETL, preventing its interactions with H 2 O molecules, thereby augmenting the moisture resistance of ZnO.Moreover, the barrier reinforces the structural integrity of ZnO, thereby helping enhance its photostability [22].Additionally, SAMs can modify the energy levels of ZnO, facilitating efficient electron movement and altering its optical properties including absorption and transmission.Furthermore, SAM-based surface treatment can help smoothen the ZnO-active layer interface, leading to a reduced recombination rate and, consequently, a high charge collection efficiency [23].Therefore, to enhance the stability and performance of OPV devices, surface tuning with SAMs can be adopted as a sensible alternative approach to using ZnO as a standalone ETL.
In the present study, the surface of the ZnO ETL was strategically altered through a 4-aminobenzoic acid (ABA)-based SAM treatment to boost the OPV device performance.ABA structures, characterized by carboxyl and amino groups (serving as anchor and terminal groups, respectively) connected at the para-positions of a benzene ring, have been documented in perovskite solar cells [24].Notably, the modified ETL (ZnO/ABA) demonstrated a substantially low WF, reduced surface roughness, and improved hydrophilicity with an indium tin oxide (ITO) electrode.The alteration significantly facilitated the selective extraction of charge carriers from the device owing to the concurrent energy-level alignment and mitigation of losses from trap-assisted recombination.Consequently, the ZnO/ABA-based device exhibited a noteworthy enhancement in efficacy, as exemplified by its maximum power output (P out, max ) under 1000 lx HLG lighting conditions being 14.68% higher than that of the device with the unmodified ZnO layer (354.48 and 309 µA cm −2 , respectively).

Device fabrication
ITO-coated glass substrates were successively cleaned by ultrasonication with deionized (DI) water + detergent, DI water, acetone, and isopropanol, with each step lasting 20 min.ZnO nanoparticles were spread onto the substrates through a 0.2 µm poly (tetrafluoroethylene) (PTFE) filter and spin-coated at 5000 rpm for 60 s, followed by thermal annealing at 200 • C for 30 min.To construct the SAM in the control device, 10 mM ABA was dissolved in ethyl alcohol and underwent stirring on a hot plate set at 100 • C for 15 min in ambient conditions.To ensure that the solution does not change in concentration during heating, the solution-containing vial was tightly sealed with parafilm [25].The resulting solution was immediately dispersed through a 0.45 µm PTFE filter and spin-coated at 3000 rpm for 30 s, followed by annealing at 100 • C for 10 min.The following additional steps were performed to fully fabricate the device of interest: A photoactive layer solution of 13.5 mg ml −1 PM6:L8-BO (1:1.2) dissolved in chloroform with a 0.5 V% 1-fluronaphthalene additive was stirred at 35 • C for 3 h.The resulting solution was spin-coated at 4000 rpm for 30 s, followed by annealing at 100 • C for 10 min.Finally, 5 nm thick MoO x and 100 nm thick Ag electrodes were sequentially deposited by thermal evaporation under a shadow mask at a high vacuum pressure (∼10 −7 hPa).The active area was estimated to be ∼0.045cm 2 .

Device characterization
Current-voltage (J-V) characteristics of the OPV devices were evaluated using a source meter (2401, Keithley Instruments, Cleveland, OH, USA) and the control program K730 (McScience, ROK).Four light sources were used to investigate the device characteristics under different light illumination conditions: an AM1.5 G solar simulator (McScience, ROK); an LED lamp (K3000 (LED 100), 5600 K, McScience, ROK); an FL lamp (OSRAM DULUXSTAR STIC 11 W; OSRAM, Munich, Germany); and an HLG lamp (OSRAM ECO PRO CLASSIC 64 543 A 240 V 46 W E27; OSRAM, Munich, Germany).The illumination light intensities were calibrated using a Si reference cell (K801S-K058, McScience).The external quantum efficiency (EQE) spectra of the OPV with the ZnO/ABA bilayer ETL were collected using a spectral solar-cell quantum-efficiency measurement system (K3100, McScience).The thin-film transmittance was measured using a UV-vis near-infrared spectrophotometer (Cary 5000, Agilent).The surface morphologies of the coated photoactive layers were examined by atomic force microscopy (AFM; XE-100, Park Systems, ROK), whereas the root-mean-square (RMS) roughness values of the photoactive layers were estimated using park systems software.Maxwell's equations-based finite-difference time-domain (FDTD) simulations (Lumerical) were performed using a standard plane-wave source.Onto the X-and Y-axes were imposed periodic boundary conditions whereas on the Z-axis were applied perfect matching-layer boundary conditions.The photoactive layer's absorption specifications were measured using a 3D frequency-domain power monitor with utilization of the refractive indices and extinction coefficients of the materials acquired using an ellipsometer (J. A. Woollam) [26].

Results and discussion
Two distinct films-each comprising an ITO electrode coupled with the pristine or ABA-modified ZnO ETL-were prepared to investigate the thin-film properties of the ETLs (see figure S1 for comprehensive information on the chemical structures and definitions, and figure 1(a) for the intensity-fermi level plot and figure 1(b) for energy level diagram).The WFs of the ITO/ETL systems were determined by Kelvin probe force microscopy and ultraviolet photoelectron spectroscopy (UPS), whereas the other energy levels were obtained from the literature [27,28].Notably, both the measurement methods yielded consistent results (table 1).The ABA-modified ITO/ETL system (ITO/ZnO/ABA) exhibited a conspicuously lower WF (4.04 eV) than that of its unmodified counterpart (ITO/ZnO; 4.51 eV).This WF adjustment helped align the energy level of the ETL with the lowest unoccupied molecular orbital (LUMO) energy of L8-BO, which enhanced the active electron transmission and suppressed the hole transport, leading to inhibited charge recombination and improved photovoltaic performance.
To optimize the performance of a photovoltaic device, ensuring that abundant incident photons penetrate the photoactive layer is crucial.This is particularly important under indoor lighting conditions, where the photon count is significantly lower than that in outdoor settings.Therefore, the ABA incorporation had to be performed such that it minimally affected the transmittance of the ETL.This was validated by acquiring transmittance spectra of the ZnO and ZnO/ABA films (figure 1(c)), which showed only a marginal difference in transmittance.Moreover, the cumulative photon transmission of the ZnO/ABA film-calculated based on the area under its spectrum-decreased by only 1.6% across the entire investigated wavelength range (300-1000 nm), indicating that the additional ABA layer hardly influenced the photon transmission.Furthermore, this difference in transmittance confirmed the successful attachment of ABA to the ZnO layer.
AFM analysis was conducted thereafter to elucidate the dynamic evolution of the film surface morphology following the deposition of ABA onto ZnO.The height images of the two films and their respective RMS surface roughness values measured over a 1.0 µm 2 area provided crucial insights (figure 1(d)).The striking disparity in surface roughness between the two films (0.201 and 0.177 nm for ZnO and ZnO/ABA, respectively) validated the integration of the ABA layer and affirmed the remarkable  effectiveness of ABA in leveling the surface.Smooth surfaces of OPV device layers are known to yield fewer defect sites at the interfacial junctures during device fabrication.Consequently, the uniform surface morphology of ABA-coated ZnO was anticipated to impart a discernible boost to the photovoltaic performance of the OPV device.Top-view scanning electron microscopy (SEM) analysis of the two films (figure S2) corroborated the AFM results.As mentioned previously, the strategic utilization of a SAM to passivate the ZnO layer can notably enhance its stability, particularly for alleviating the deleterious effects of moisture exposure.Therefore, the potency of ABA as a protective barrier was assessed by analyzing its hydrophobicity through water contact angle analysis (figure 1(e)).The ABA-coated surface demonstrated a notably higher contact angle (47 • ) than that of the relatively hydrophilic pristine ZnO (38 • ).This was likely associated with the reduction in the concentration of hydrophilic hydroxyl groups on the ZnO surface [29].This helped markedly enhance the hydrophobicity, which engendered a pivotal shift in the interfacial dynamics between the ETL and photoactive layer, fostering an environment that enhanced their compatibility and helped optimize the device's performance.
To assess the manner in which the ABA adsorption onto the ZnO ETL influenced the photovoltaic performance of the OPV device, exposure studies were conducted under various illumination conditions, encompassing both outdoor and indoor settings.Irradiation spectra were acquired under standard 1-sun illumination (AM1.5 G), which simulated outdoor conditions, and under LED, HLG, and FL lighting, which replicated indoor conditions (figure 2(a)).Inverted-structure devices comprising an ITO/ZnO unit with or without the ABA/PM6:L8-BO/MoO x /Ag layer system were explored (see the device schematic in figure 2(b) and the experimental section for more details).Cross-sectional SEM images of the two configurations were obtained to acquire deeper insights (figure 2(c)).The imaging results conformed to expectations, revealing a slight increase in the thickness of the ZnO ETL from 35 to 37 nm upon ABA coating.
Initially, the devices were tested under AM 1.5 G (1-sun) irradiation at an input power density (P in ) of 100 mW cm −2 .The obtained J-V curves (figure 3(a)) suggested that both devices performed similarly.Specifically, the ZnO-and ZnO/ABA-based devices exhibited the following photovoltaic metrics: open-circuit voltage (V OC ), 833.74 and 843.53 mV; short-circuit current density (J SC ), 23.68 and 23.83 mA cm −2 ; fill factor (FF), 75.62% and 74.50%; maximum output power density (P out, max ), 14.9 and  14.97 mW cm −2 ; and PCE, 15.21% and 15.24%, respectively.The slightly higher V OC of the ZnO/ABA device was likely related to its suitable energy-level alignment with the LUMO energy of the photoactive layer.EQE measurements were conducted thereafter to elucidate the discrepancy in J SC values between the two devices.The acquired EQE curves (figure 3(b)) indicated that the relatively high J SC exhibited by the ZnO/ABA device was evidently due to its improved EQE at shorter wavelengths (300-470 nm).Consequently, the current density calculated using the EQE values (J EQE ) for the ZnO/ABA-based device was higher than that  Light source ETL V OC (mV) J SC (µA cm −2 ) FF (%) for its ABA-free counterpart (22.92 and 22.57 mA cm −2 , respectively; see table 2 for more details on the performance of the devices under AM 1.5 G illumination, and equation (S1) for details on J EQE ).
HLG lamps are widely used as indoor lighting sources, primarily because of their notable output power and cost-effectiveness.They are frequently preferred over alternative indoor lighting sources, largely because of their superior current density and extended irradiation spectrum (360-2500 nm).Specifically, their irradiation spectra are relatively less distributed at high-energy wavelengths, rendering them particularly appealing for OPV applications, given the inherent vulnerability of state-of-the-art photoactive layer materials to UV irradiation exposure [17,30,31].Moreover, HLG lamps emit radiation at a blackbody temperature of ∼3200 K. To date, research on enhancing the indoor OPV device performance under halogen lighting has been limited.Therefore, in this study, the OPV device performance under harsh thermal conditions was assessed using an HLG light source.The J-V characteristics of the two devices were examined under 1000 lx HLG lighting (figure 3(c)) at a P in value of 4.72 mW cm −2 .According to the results, the attachment of ABA onto the ZnO ETL led to significantly higher V OC , J SC , FF, and P out, max values than those of the ABA-free device (739.41 and 726.99 mV, 614.65 and 543.52 µA cm −2 , 77.87% and 77.76%, and 354.48 and 309.1 µA cm −2 , respectively).The remarkable 14.68% increase in P out, max , predominantly facilitated by the sizeable 13.09% increase in J SC , was evidently due to the manifold improvements in performance induced by ABA integration.Specifically, the substantially high P out, max value of the ABA-modified device reflected its ability to power indoor electronic devices, particularly IoT network components that typically consume low electrical power (1-100 µW) in an idle/standby state.
To extend the analysis to diverse indoor illumination conditions, the performance of the devices was assessed under LED and FL illumination conditions (figure S3).Under 1000 lx 5200 K LED illumination (P in : 0.254 mW cm −2 ), the device with the ZnO/ABA ETL demonstrated higher V OC , J SC , and FF values than those of the ABA-free system (678.10 and 654.80 mV, 127.23 and 126.37 µA cm −2 , and 75.41% and 74.25%, respectively), culminating in a 6.02% higher power output (P out , 66.58 and 62.8 µW cm −2 , respectively).This enhancement was also observed under 1000 lx FL illumination (P in : 0.300 mW cm −2 ), given that the ZnO/ABA-based device achieved an impressively higher P out, max than that of its ABA-free counterpart (74.67 and 72.9 µW cm −2 , respectively).Integrated J SC (J SC, int ) represents the J SC computed based on the efficacy of the device in transforming the incident photon flux into an output current.Therefore, J SC calculations were performed under LED and FL illumination by examining the photon flux density alongside J SC, int (figures 3(d) and (e)).Under 1000 lx LED illumination, the ZnO/ABA-based device exhibited a higher saturated J SC, int value than that of its ABA-free equivalent (129.82 and 127.02 µA cm −2 , respectively).Comparatively, under FL 1000 lx lighting, the saturation points increased in value to 146.15 and 149.12 µA cm −2 for the ZnO and ZnO/ABA-based devices, respectively.The similarity between these values and those measured confirmed the validity of the measured data and the contribution of ABA to the performance enhancement (see table 3 for results of the performance measurements conducted under various indoor lighting conditions).A greater performance enhancement under indoor lighting than AM1.5 G condition can, to an extent, be attributed to the series and shunt resistances (table S1), as explicable by the OPV device equivalent circuit diagram (figure S4) and the Shockley diode equation [28].More specifically, under the 1-sun condition, the series resistance (R S ) values of the two devices are comparable, whereas, Under low-intensity indoor lighting conditions, which are characterized by a restricted influx of incident photons, the efficiency of OPV devices is significantly compromised owing to the emergence of recombination loss as a pivotal factor.Therefore, space charge-limited current (SCLC) measurements were performed to obtain precise quantitative insights into the electron trap density of the fabricated device [8,32].Two electron-only devices comprising the ITO/ZnO unit with or without the ABA/PM6:L8-BO/PDINO/Al system were fabricated for this analysis (refer to figure S5 for details on PDINO).The J-V curves of the two devices obtained under dark conditions (figure 4(a)) suggested that both devices exhibited a slope (m) of 1 in the low-voltage region, indicating the existence of ohmic conduction zones.Moreover, the shallow-trap SCLC region was absent owing to the trap-free behavior of the devices [33].Notably, the devices attained the trap-free zone beyond the measured trap-filled limit voltage (V TFL ) values of approximately 1.07 and 0.86 V for the devices with ZnO and ZnO/ABA as the ETL, respectively, leading to a current with an m value of 2. This region adheres to the Mott-Gurney law [34], which is expressed as where J is the current density, V is the applied voltage, ε r is the relative permittivity of the ETL material (approximated to 8.5), ε 0 is the permittivity of free space, µ is the mobility of the charge carrier, and t is the thickness of the ETL.Using equation (1), the electron mobilities of the ZnO and ZnO/ABA films were estimated to be 2.20 × 10 −8 cm 2 V −1 s −1 and 6.67 × 10 −8 cm 2 V −1 s −1 , respectively.The electron trap density (N t ) was determined as follows [28,35]: where e is the elementary charge of an electron.The calculated N t value for the ZnO/ABA-based device (5.90 × 10 17 cm −3 ) was lower than that of its ABA-free counterpart (8.21 × 10 17 cm −3 ), suggesting that the integration of ABA with ZnO helped reduce the trap density.To clarify the manner in which the high electron mobility and low electron-trap density of ABA helped suppress the recombination loss, trap-assisted recombination loss was systematically investigated by analyzing the dependence of J SC on the dim light intensity.In essence, the current density was plotted against the light intensity (J LI ) on a logarithmic scale (figure 3(f)), and the results were interpreted based on the power-law relationship between J SC and J LI (J SC = J LI s , where s denotes the bimolecular recombination rate, with deviations in s from 1 signifying variations in the rate).The devices with ZnO or ZnO/ABA as the ETL exhibited different slopes (0.94 and ∼1, respectively).Consequently, the contrast in J SC between the two devices was presumably due to the ability of ABA to impede charge recombination.
Further charge dynamics analysis was conducted by subjecting the two devices to electrochemical impedance spectroscopy (EIS) under dark conditions.Based on the diameters of the Nyquist plot data, the recombination resistances of the devices with ZnO or ZnO/ABA as the ETL were determined to be 53.4 and 108.5 kΩ, respectively (figure 4(b)).The lower recombination resistance of the ZnO-based device is consistent with its previously calculated higher trap density.FDTD simulations were subsequently performed to investigate the differences in the optical properties of the two devices.Consequently, the normalized power absorption data of the devices were acquired in the wavelength range of 300-800 nm (figure S6; see experimental section for details pertaining to the simulation procedure).A noticeable distinction was observed in the wavelength range of 300-350 nm, where the ZnO/ABA-based device exhibited a power absorption peak that was more intense than that of its pristine counterpart.This was likely related to the more uniform surface morphology of the ETL and increased ETL-photoactive layer compatibility.The amplified power absorption of the ABA-incorporated device aided in improving its J SC .
To ascertain the efficacy of ABA in enhancing the photothermal stability of the OPV system, the fabricated devices were subjected to prolonged irradiation under continuous illumination from a 1000 lx HLG lamp (P IN = 4.72 mW cm −2 ) for 120 h in a nitrogen-filled environment.The normalized PCE of the devices decreased progressively over 120 h (figure 4(c)).Notably, after 6 h, the ZnO-based device exhibited a substantial PCE loss of ∼30%, in contrast to the meager reduction of its ZnO/ABA-based counterpart (∼5%), which retained a significant portion of its initial PCE.Notably, the ZnO/ABA-based device showed excellent consistency in stability over time, preserving ∼74% of its initial PCE after 120 h, whereas its ABA-free counterpart exhibited considerable performance degradation, retaining only ∼44% of its original PCE.The use of a ZnO ETL has been shown to result in the degradation of OPV devices owing to its diffusion into the organic photoactive layer [36,37].Consequently, the notably superior stability of the ZnO/ABA-based device could be attributed to the presence of the ABA layer, which effectively suppressed the miscibility between ZnO and the photoactive layer.

Conclusions
The modification of the ZnO ETL using ABA was shown to result in significant enhancements in the performance and stability of an OPV device intended for indoor applications in IoT systems.The ABA-treated ZnO ETL demonstrated remarkable changes in the work function, surface roughness, and hydrophilicity with the ITO electrode, facilitating improved charge-carrier extraction and minimizing trap-assisted recombination losses.The ABA-based modification resulted in enhanced electron mobilities, reduced trap densities, and high recombination resistances, which compelled the ZnO/ABA-based device to exhibit superior performance.Notably, the P out, max value of the ZnO/ABA-based device obtained under 1000 lx HLG illumination (354.48 µA cm −2 ) was 14.68% higher than that of the device with unmodified ZnO (309 µA cm −2 ).Furthermore, the device demonstrated impressive resilience under thermal illumination conditions (1000 lx HLG lighting), maintaining ∼74% of its initial PCE over 120 h, in contrast to the considerable performance degradation of its ZnO-based counterpart (∼55% of the original value).These findings highlight the significant potential of modifying ZnO with ABA for simultaneously enhancing the efficiency and stability of OPV devices for indoor IoT applications, paving the way for their practical implementation in next-generation energy-harvesting systems.

Figure 1 .
Figure 1.(a) Intensity-Fermi level plots for the ZnO and ZnO/ABA films.(b) Energy band diagram of the components of the fabricated devices.(c) Transmittance spectra of ZnO and ZnO/ABA films.(d) AFM profiles of ZnO and ZnO/ABA films, and their respective root-mean-square roughness values.(e) Water contact angles of ZnO and ZnO/ABA films.

Figure 3 .
Figure 3. (a), (c) J-V characteristics of the devices under (a) AM 1.5 G (1-sun) illumination and (c) 500 lx and 1000 lx (HLG) lighting.(b) EQE spectra of the devices.Photon flux densities and integrated JSC spectra of the devices under (d) 1000 lx LED and (e) 1000 lx FL illumination.(f) Light intensity dependence of JSC for the two devices.

Figure 4 .
Figure 4. (a) J-V curves obtained under dark conditions for trap-density analysis.(b) Nyquist plots of electron-only devices constructed using EIS data.(c) Normalized PCE data of the two devices over a 120 h period under thermal illumination (HLG lighting).

Table 2 .
Photovoltaic performances of the devices with ZnO or ZnO/ABA as the ETL under AM 1.5 G (1-sun) illumination (P in : 100 mW cm −2 ).

Table 3 .
Photovoltaic performances of the devices with ZnO or ZnO/ABA as the ETL under different indoor illumination conditions (HLG 1000 lx, LED 1000 lx, and FL 1000 lx).