Improved Efficiency and Stability of Organic Solar Cells by Interface Modification Using Atomic Layer Deposition of Ultrathin Aluminum Oxide

The interfacial contacts between the electron transporting layers (ETLs) and the photoactive layers are crucial to device performance and stability for OSCs with inverted architecture. Herein, atomic layer deposition (ALD) fabricated ultrathin Al2O3 layers are applied to modify the ETLs/active blends (PM6:BTP‐BO‐4F) interfaces of OSCs, thus improving device performance. The ALD‐Al2O3 thin layers on ZnO significantly improved its surface morphology, which led to the decreased work function of ZnO and reduced recombination losses in devices. The simultaneous increase in open‐circuit voltage ( VOC ), short‐circuit current density ( JSC ) and fill factor (FF) were achieved for the OSCs incorporated with ALD‐Al2O3 interlayers of a certain thickness, which produced a maximum PCE of 16.61%. Moreover, the ALD‐Al2O3 interlayers had significantly enhanced device stability by suppressing degradation of the photoactive layers induced by the photocatalytic activity of ZnO and passivating surface defects of ZnO that may play the role of active sites for the adsorption of oxygen and moisture.


Introduction
Organic solar cells (OSCs) have gained intensive research attention due to the priorities of lightweight, flexibility and solution processability, thus showing great potential in commercialization. [1][4] However, materials stability and device lifetime have received far less attention by far, [5] which has been a crucial issue to be addressed so that OSCs could be applicable in the marketplace. [6]The conventional OSCs normally employ poly (3,4ethylenedioxythiophene): poly(styrenesulfonate) (PEDOT:PSS) as the hole transporting layer (HTL), which usually present lower stability due to the inherent hygroscopicity and acidity of PEDOT:PSS, that can corrode the indiumtin oxide (ITO) substrates, thus leading to diffusion of indium and tin into the HTL. [7,8]OSCs with an inverted structure were then developed, which can facilitate the devices with better stability than the conventional OSCs, thus are more favorable to the practical application of OSCs. [9,10]13] However, solution-processed ZnO-incorporated OSCs face quite a few challenges in achieving good performance.[16] Secondly, ZnO usually presents high photocatalytic activity under UV light irradiation, as the photoinduced charge carriers in the ZnO films would oxidate the dangling hydroxyl groups of NFA molecules and form hydroxyl radicals, which would lead to decomposition of NFA molecules. [17,18]One of the most common mechanisms is photo-oxidation, which involves the reaction between the photoactive materials and oxygen molecules in the presence of light.The excited state of the photoactive material can transfer energy to the oxygen molecule, resulting in the formation of reactive oxygen species (ROS), such as singlet oxygen or superoxide.These ROS can then react with the photoactive material, leading to the formation of oxidation products and ultimately the degradation of the material.For instance, the C=C bonds can be broken by the photogenerated hydroxyl radicals, thus causing the decomposition of IT-4F on the ZnO surface, which would lead to poor stability of IT-4F based OSCs. [19,20]Nowadays, PM6:Y6 or Y-series is the most widely used combination in the OSCs The interfacial contacts between the electron transporting layers (ETLs) and the photoactive layers are crucial to device performance and stability for OSCs with inverted architecture.Herein, atomic layer deposition (ALD) fabricated ultrathin Al 2 O 3 layers are applied to modify the ETLs/active blends (PM6:BTP-BO-4F) interfaces of OSCs, thus improving device performance.The ALD-Al 2 O 3 thin layers on ZnO significantly improved its surface morphology, which led to the decreased work function of ZnO and reduced recombination losses in devices.The simultaneous increase in opencircuit voltage (V OC ), short-circuit current density (J SC ) and fill factor (FF) were achieved for the OSCs incorporated with ALD-Al 2 O 3 interlayers of a certain thickness, which produced a maximum PCE of 16.61%.Moreover, the ALD-Al 2 O 3 interlayers had significantly enhanced device stability by suppressing degradation of the photoactive layers induced by the photocatalytic activity of ZnO and passivating surface defects of ZnO that may play the role of active sites for the adsorption of oxygen and moisture.
as active layers, [21][22][23] which also suffer from the same problem. [24]ome strategies were utilized to address this issue, such as surface treatment with polar solvents, [25] Lewis acid or gas plasma, [26,27] introducing dopants, that is, aluminum (Al), [28] zirconium (Zr), [29] indium and lithium cations (Li + ), [30,31] into ZnO layers, surface defects passivation of ZnO via insertion of appropriate interlayers, that is, selfassembled monolayers (SAMs), [32] functionalized fullerenes, [33] ionic liquid and high-molar-mass polymers, [34,35] which have been demonstrated effective to improve the stability and efficiency of OSCs.Amongst them, the insertion of interlayers could also reduce device leakage current, as they could improve charge extraction. [36][40] Recently, there have been some reports on the utilization of the ALD technique for OSCs.Duan et al. [41] inserted an ultrathin ALD-TiO x interlayer in OSCs to improve the efficiency and stability of the devices.Polydorou et al. [18] applied an ultrathin ALD-HfO 2 film as a passivation interlayer on the ZnO ETL to improve device performance.Vasilopoulou et al. [42] introduced ultrathin ALD-Al 2 O 3 and ZrO 2 layers to passivate surface defects of TiO 2 ETL for performance improvement of the inverted OSCs.These works have proved that ALD fabricated interlayers have been successfull in further improving the performance of OSCs.Whilst most of them focused on the passivation of defects on the metal-oxide surface, the effects of the interlayers on the photocatalytic activity of ZnO were rarely reported.
In this work, we report on the utilization of an ultrathin ALD-Al 2 O 3 film to form an interlayer in an inverted PM6:BTP-BO-4F based OSCs for improving device performance.The best device performance was achieved from OSCs introduced with three deposition cycles of ALD-Al 2 O 3 interlayer.Accompanying with synergistically increased open-circuit voltage (V OC ), short circuit current density (J SC ) and fill factor (FF), the PCE realized an absolute 1% enhancement compared to the control devices with pristine ZnO ETL, which was also comparable with that of the conventional devices we fabricated previously. [43]We found that synergistic effects of 1) ZnO surface defects passivation, 2) improved active layer nanomorphology, 3) decreased work function (WF) of ZnO, 4) reduced recombination loss, and 5) more balanced charge transport induced by the ALD-Al 2 O 3 interlayer led to further improved efficiency.Moreover, the devices modified with ALD interlayers exhibited significantly improved stability, resulting from the reduced photocatalytic activity of ZnO and the passivation of surface defects on ZnO.

Results and Discussion
To verify the feasibility of the strategy, the passivation effect of Al 2 O 3 on the ZnO x (0002) surface was first evaluated with density functional theory calculation.In our model, oxygen vacancies were incorporated due to their lowest formation energy amongst all of the ZnO surface defects.In the beginning, we created and optimized the ZnO x (0002) surface, where x = 0.87. Figure 1a (left) shows a quarter of oxygen atoms on the surface were removed from the slab, resulting in an O-vacancy coverage density of ρ = 2.73 × 10 14 cm −2 .The Al 2 O 3 /ZnO x interfaces were then created and optimized (Figure 1a [right]).This process would release an energy of 8.05 eV per Al 2 O 3 adsorbent unit according to our calculation results.Figure 1b exhibited the density of states (DOS) of the O-terminated ZnO x (0002) surface.The states located in the range of −1.5 to −0.27 eV under the Fermi levels were contributed from O states.The calculated DOS of ZnO x /Al 2 O 3 interfaces is illustrated in Figure 1c, which demonstrates the suppression of surface oxygen states around the Fermi levels compared to the pristine ZnO x (0002) (Figure 1b).Bader charge analysis indicates that there existed charge transfer from Al 2 O 3 to ZnO x (0002) surface oxygen atoms at a maximum of ≈1.39 |e|.These calculation results suggest that passivation of the ZnO x surface may be realized by charge transfer from Al 2 O 3 to oxygen atoms.

The Effects of Al 2 O 3 ALD Layers Coated on ZnO
X-ray photoelectron spectroscopy (XPS) was used to analyze the elemental details to confirm the existence of ALD-Al 2 O 3 layers on ZnO.Only the elements zinc (Zn), oxygen (O) and aluminum (Al) were detected.Shirley background method was adopted to fit the spectra in Avantage software.As shown in Figure Appendix S1, Supporting Information, the Zn 2p 3/2 and Zn 2p 1/2 appeared at a binding energy of 1020.9 and 1044.1 eV, respectively. [41]No variation of the binding energy at the Zn 2p peaks has been observed before and after ALD-Al 2 O 3 deposition.Only the peak intensities decreased significantly along with the increased number of ALD-Al 2 O 3 cycles.The O 1s spectra were also recorded from samples with ALD-Al 2 O 3 layers of different thicknesses.Three peaks were observed from deconvoluted O 1s XPS spectra (Figure 2a-c).The peak at the lowest binding energy (530.0AE 0.1 eV) was correlated with the Zn-O bonds, [44,45] the peak at a higher binding energy (531.1 AE 0.1 eV) originated from Al-O bonds, [46] while the peak at around 532 eV was associated with the surface oxygen in the form of hydroxyl groups, which were formed from the reaction of oxygen with environmental air. [47]Furthermore, the atomic percentage of the deconvoluted O 1s peak associated with a different bond is presented in Table Appendix S1, Supporting Information.We observed that both the Zn-O and Zn-OH peak intensities decreased while the Al-O peak intensity increased along with increased ALD-Al 2 O 3 thickness, indicating that Al had consumed oxygen on the ZnO layer and strong Al-O bonds had formed.As a result, the reduction of −OH concentration can effectively suppress the recombination centers and realize passivation at the ZnO/Al 2 O 3 interface. [41,48]The Al 2p peaks at a binding energy of 73.9 AE 0.01 eV were also observed (Figure 2e,f ), and the peak intensity was enhanced along with more ALD-Al 2 O 3 deposition cycles, indicating the formation of Al-O bonds.Accordingly, we conclude that more ALD-Al 2 O 3 cycles would lead to better passivation.
The surface morphology of the ETL influences the device performance of OSCs significantly.We used scanning electron microscope (SEM) and atomic force microscopy (AFM) to investigate the surface morphology of ZnO modified with ALD-Al 2 O 3 of various thicknesses.Fiber-like surface nanomorphology patterns can be observed from the pristine ZnO films (Figure 3a).The patterns became blurring after being modified with 3 cycles of ALD-Al 2 O 3 , and almost disappeared when 10 cycles ALD-Al 2 O 3 were deposited, which indicates that ALD-Al 2 O 3 layers can smooth the surface morphology of ZnO effectively, and that may passivate surface defects of ZnO successfully.Moreover, the pristine ZnO layer delivered a root-mean-square (RMS) roughness of 3.97 nm (Figure 3b), while for ZnO modified with 3 and 10 cycles of ALD-Al 2 O 3 , the RMS decreased to 2.32 and 1.58 nm, respectively, which means that ALD-Al 2 O 3 treatment had effectively reduced surface roughness of ZnO.We then conclude that a more conformal interface between the ZnO ETL and photoactive layer could be created by several cycles of ALD-Al 2 O 3 .AFM and transmission electron microscopy (TEM) were also utilized to explore the surface topography of the active layers PM6:BTP-BO-4F on ALD-Al 2 O 3 modified ZnO substrates.As shown in Figure S2a, Supporting Information, the PM6: BTP-BO-4F films showed smoother surface, while the RMS of the active layers also reduced with increased ALD-Al 2 O 3 thickness.These results have demonstrated that ALD-Al 2 O 3 modified ETLs also improved the surface morphology of the photoactive layers.
The work function (WF) of ZnO is also essential to the charge extraction process, which affects the overall performance of the devices.We performed Kelvin Probe Force Microscopy (KPFM) measurements to estimate the WF of ZnO layers deposited with ALD-Al 2 O 3 .As presented in Figure S3, Supporting Information, the bare ZnO films presented a WF of 4.64 eV, whilst it reduced when ALD-Al 2 O 3 was performed.The ETL ZnO modified with three cycles of ALD-Al 2 O 3 showed the lowest WF of 4.55 eV.Consequently, the reduced WF would lead to lower electron extraction barrier, thus facilitating effective electron extraction from the active layer to ZnO films, as well as allowing better ohmic contact between ZnO films and BTP-BO-4F with the lowest unoccupied molecular orbital (LUMO) levels of −4.39 eV. [41,49]

Solar Cell Performance Analysis
As addressed, the effects of ALD coatings on passivating ZnO layers had been demonstrated, and we then utilized the passivated ZnO films to fabricate OSCs with an inverted architecture of indium tin oxide (ITO)/ZnO/ALD-Al 2 O 3 /active layer/MoO 3 /Ag (Figure 4a).4c, and the corresponding photovoltaic parameters are listed in Table 1 and exhibited as a box-plot diagram in Figure S4, Supporting Information.We found that the devices incorporated with ALD-Al 2 O 3 layers exhibited considerable enhancement in PCE compared with the control devices.Amongst them, the devices modified with 3 cycles of ALD-Al 2 O 3 yielded the maximum PCE of 16.61%, along with a V OC of 0.853 V, a J SC of 25.94 mA cm −2 and an FF of 75.08%.When 10 cycles of ALD-Al 2 O 3 deposition were completed, the device performance started to deteriorate, due to the insulating property of alumina that may start hindering electron transport when it reaches a certain thickness.In addition, the devices modified with three cycles of ALD-Al 2 O 3 showed significantly higher V OC than the control devices, attributing to the reduced WF of ZnO, thus a lower energy offset between ZnO and the photoactive layer. [50]Besides this, the decline in WF lowered the electron extraction barrier and elevated the built-in voltage, which facilitated electron extraction and transport, thus resulting in higher J SC of the three cycles ALD-Al 2 O 3 modified devices finally. [49]Figure S5, Supporting Information shows the dark J-V curves for the devices.The current density at −1.0 V is defined as leakage current density. [51]Obviously, the dark currents in devices with three cycles of ALD-Al 2 O 3 exhibited lower reverse leakage currents compared with the control devices, which suggested effective passivation in the devices.Figure 4d shows the external quantum efficiency (EQE) measurement results.All of the integrated photocurrents obtained from the EQE spectra agreed well with the results achieved from the J-V curves (with <5% mismatch).We also found the incorporation of ALD-Al 2 O 3 layers had not affected the spectral shape of EQE while improving the device efficiency.

Charge Separation, Transport and Recombination
To evaluate exciton dissociation and charge generation efficiency, the relation of photocurrent density (J ph ) with respect to effective voltage (V eff ) of the devices incorporated with ALD-Al 2 O 3 layers was also analyzed.As shown in Figure 5a, J ph can be estimated as J L ÀJ D , where J L and J D represent the current densities measured separately under light illumination at AM 1.5G and in dark. [52,53]V eff is determined by V 0 ÀV bias , where V 0 is the voltage when J ph ¼ 0 mA cm −2 and V bias is the applied voltage.The saturation current density (J sat ) is defined as the photocurrent when all of the photogenerated excitons are dissociated into free charges and collected by the electrodes at a high V eff , that is, when V eff ≥ 2V.The exciton dissociation possibility (P diss ) can be calculated as J Ã ph =J sat , where J Ã ph is J ph under short-circuit condition.Table S2, Supporting Information summarizes the detailed exciton dissociation and charge collection efficiencies of the OSCs.Notably, the devices with three cycles of ALD-Al 2 O 3 realized the highest exciton dissociation efficiency of 96.7%.They also achieved higher J sat than that of the control devices.However, the devices with 10 cycles ALD-Al 2 O 3 showed a reduced J sat , which indicates that Al 2 O 3 layers may impede charge transport when they reach a certain thickness.Light-dependent Energy Environ.Mater.2024, 7, e12620 V OC and J SC were also measured to gain further insights into the charge separation, transport and recombination properties of the devices.The relationship of V OC versus P light can be expressed as: where k denotes Boltzmann's constant, T is the absolute temperature and q is the elementary charge. [54]If n is close to 2, it means severe trap-assisted recombination would have been induced by morphology defects.As presented in Figure 5b, the fitted n were 1.29, 1.16, 1.38 for devices with pristine ZnO, 3 cycles and 10 cycles ALD-Al 2 O 3 treated ZnO, respectively.The 3 cycles ALD-Al 2 O 3 incorporated devices presented the best performance in reducing trapassisted recombination.The relationship between J SC and P light is defined as: J SC / P α light , [55] in which α represents the exponential factor and can be obtained from the slopes (s) of fitting curves.An α of 0.959 was obtained for the control devices, which increased to 0.978 for three cycles ALD-Al 2 O 3 inserted devices (Figure 5c), indicating the biomolecular recombination in three cycles ALD-Al 2 O 3 passivated devices had been effectively suppressed.To clarify the influence of ALD-Al 2 O 3 interlayers on charge transport properties of the devices, the hole (μ h ) and electron (μ e ) mobilities of devices with Al 2 O 3 layers were evaluated using the space charge limited current (SCLC) method.As summarized in Figure 5d,e and Table S3, Supporting Information, the hole and electron mobilities of the control devices were 4.44 × 10 −4 mA 0.5 cm −1 and 2.36 × 10 −4 mA 0.5 cm −1 , respectively.The mobilities marginally reduced after treatment with three cycles ALD-Al 2 O 3 , and continued to decrease along with the increased thickness of ALD-Al 2 O 3 .The μ h =μ e ratio was utilized to evaluate the charge transport balance in the OSCs. [56]A μ h =μ e ratio of 1.8 was obtained for the devices with three cycles ALD-Al 2 O 3 , which was closer to 1 compared with that of 1.89 for the control devices.These results manifest that the insulating property of alumina may affect charge mobility, whereas an ALD-Al 2 O 3 interlayer of appropriate thickness can lead to more balanced charge transport in devices, which agrees well with the photovoltaic measurement results.
In order to further explore charge separation and recombination dynamics at the ZnO/PM6 interfaces, steady state photoluminescence (PL) measurements were carried out on neat PM6 films on top of ZnO layers with and without ALD-Al 2 O 3 modification (Figure S6a, Supporting Information).All of the films were spin coated at a same spin speed The standard deviations of PCEs were calculated from those of 12 independent cells.with 5 mg mL −1 PM6 solution in order to obtain constant film thickness.Enhanced PL intensity was observed from PM6 films on ZnO/ALD-Al 2 O 3 (three cycles) surfaces, compared with that from PM6 films on pristine ZnO, which was attributed to effective passivation of traps on ZnO surfaces by the ALD-Al 2 O 3 layers. [49]Next, time-resolved photoluminescence (TRPL) measurements were performed in order to investigate interfacial charge transfer from PM6 to ZnO (Figure S6b, Supporting Information).The fluorescence lifetime (τ) was calculated by fitting the TRPL semilog decay curves using a two-exponential function, and the extracted parameters are presented in Table S4, Supporting Information.As a result, the exciton lifetime of PM6 films on top of ZnO deposited with three cycles ALD-Al 2 O 3 achieved an apparent increase.The increased exciton lifetime resulted from the effective passivation of deep surface defect states on ZnO, which served as recombination centers for holes, thereby reducing hole trapping. [42]eferring to the above discussions, we concluded that effective passivation of traps on ZnO had been realized by introducing ALD-

Energy Loss Analysis
To further understand the origins for V OC improvement of the ALD-Al 2 O 3 treated devices, we analyzed energy loss (E loss ) in both control devices and ALD-Al 2 O 3 modified devices based on Marcus theory. [57]he total energy loss of an OSC includes three parts: radiative recombination loss ΔE 1 ð Þ, charge generation ΔE 2 ð Þ, and non-radiative recombination ΔE 3 ð Þ, which can be described as the following equation: [58,59] where E g is the optical bandgap of the device, E CT is the charge transfer (CT) energy state.Firstly, we estimated E g of the two kinds of devices from the crossing points of the normalized electroluminescence (EL) and EQE spectra (Figure S7, Supporting Information). [60]E g were 1.403 eV for the control devices and 1.400 eV for the devices with three cycles ALD-Al 2 O 3 .Next, E CT can be estimated by simultaneously fitting the lower energy regions of highly sensitive EQE (s-EQE) and EL spectra. [61]ΔE 2 can then be determined by the formula ΔE 2 ¼ E g ÀE CT , which is related to the energetic difference between singlet excited states and charge transfer states. [62]As seen, the devices modified with 3 cycles ALD-Al 2 O 3 exhibited similar E CT with the control devices, resulting in comparable ΔE 2 (Figures 6a,b).ΔE 1 is induced by radiative recombination, which is unavoidable for all kinds of OSCs.ΔE 3 is the energy loss originating from non radiative resomibanition, which is the main cause for the large energy loss of OSCs, it can be quantified by the equation: where k is Boltzmann constant, T is temperature and EQE EL is external quantum efficiency of EL.As displayed in Figure 6c, EQE EL of the control devices and ALD-Al 2 O 3 treated devices are 0.61 × 10 −4 , 1.12 × 10 −4 , respectively.Correspondingly, ΔE 3 are 0.251 eV for the control devices and 0.235 eV for the three cycles ALD-Al 2 O 3 introduced devices.The suppressed ΔE 3 of ALD-Al 2 O 3 treated devices indicates that the ALD-Al 2 O 3 passivation interlayer can inhibit charge recombination loss in devices, which coincides with the results from V OC and J SC versus light intensity measurements (Figure 5b,c).Table 2 summarizes the detailed parameters extracted from the measurements, the energy losses for the devices are compared in Figure 6d.

Device Stability
Besides device efficiency, stability is also a crucial parameter for OSCs.As known, photodegradation of the photoactive materials mainly causes deterioration of device performance, which is a complex process.To evaluate the ambient stability of our devices, we measured 12 h evolution of J-V curves for unencapsulated devices with and without ALD-Al 2 O 3 under UV light in ambient   S5, Supporting Information, the PCE of 15.63% for the control devices decreased to 5.94% after 12 h of constant UV illumination, only retaining 38% of the initial PCE.However, for the devices with 3 cycles ALD-Al 2 O 3 , the PCE decreased from 16.61% to 14.05%, which retained ~85% of the initial PCE.These results manifested that ALD-Al 2 O 3 interlayers can effectively lower the photocatalytic activity of ZnO, thus reducing the hydroxyl radicals that could react with NFA, thereby preventing the photoactive molecules from degradation. [20]Moreover, we investigated the influence of ALD-Al 2 O 3 interlayer on the long time device stability.We stored the unencapsulated devices in a nitrogenfilled glove box and monitored the steady power output (SPO) stability for a certain period.Figure 7c shows the operational SPO stability performance of the corresponding devices under persistent one sun irradiation.We found that the devices with three cycles ALD-Al 2 O 3 retained 95% of the initial PCE.The performance degradation was negligible as the light intensity decreased from 1 sun to 0.98 sun during 75 h, while the PCE of the control devices dropped to 79.5% of the initial PCE.The significant improvement on the stability of the ALD-Al 2 O 3 treated devices can be explained by the fact that ALD-Al 2 O 3 interlayers had effectively passivated surface defects of ZnO that may act as active sites for the adsorption of oxygen and water molecules.Additionally, during the ALD process, the initially adsorbed water molecules on the ZnO surface participate in reactions with trimethylaluminum (TMA) to form the first monolayer of Al 2 O 3 , which means they are unable to act as corrosion agents. [49]he stability of devices stored in air (RH ~40%) was also investigated (Figure 7d).Inverted devices are inherently more stable than normal devices due to the acidic environment created by PEDOT:PSS.When Al 2 O 3 interface layers were introduced, the inverted devices' stability was further improved, maintaining over 88% of its initial efficiency after 500 hours of degradation.

Conclusion
In summary, we systematically investigated the application of an ultrathin ALD-Al 2 O 3 interlayer on ZnO as the ETLs in OSCs with inverted structures.It was found that the incorporation of ALD-Al 2 O 3 thin layers led to effective passivation of surface defects on ZnO, which thus resulted in lowered electron extraction barrier and reduced surface recombination.In contrast to the control devices that used pristine ZnO as ETLs, the devices modified with three cycles ALD-Al 2 O 3 showed optimal characteristics with improved V OC , J SC and FF, which delivered an absolute 1% increase in efficiency.Furthermore, the unencapsulated ALD-Al 2 O 3 modified devices exhibited exceptional stability either under ultraviolet light in ambient conditions or under operational conditions.This work has explicitly shown the potential of ALD modification strategy in manipulating ETLs' properties and suppressing defect trap states and also demonstrated the importance of ALD-dominated interface engineering in photovoltaic applications.

Experimental Section
Detailed information related to the synthesis of active electrodes, physicochemical characterization, and electrochemical evaluation of bifunctional electrodes towards UOR and supercapacitor application is provided in Supporting Information.

Figure 1 .
Figure 1.a) The atomic structures of ZnOx (left) and ZnOx/Al 2 O 3 (right) interfaces.b) The density of states of ZnOx and c) ZnOx/ Al 2 O 3 interfaces.
ZnO and MoO 3 films were used as the ETL and HTL, respectively.The chemical structures and the corresponding energy levels of the materials in photoactive layers are shown in Figure 4b.To examine the impact of ALD-Al 2 O 3 thickness on device performance, a series of PM6:BTP-BO-4F based OSCs were fabricated, with the ZnO layers unpassivated or treated with ALD-Al 2 O 3 (Deposition cycles n = 1, 3, 5 and 10).The current density-voltage curves of the devices modified with ALD-Al 2 O 3 are presented in Figure

Figure 4 .of 9 ©
Figure 4. a) The device architecture; b) The energy levels of devices modified with ALD-Al 2 O 3 layers; c) The current-voltage (J-V) measurements; d) The EQE spectra and integrated J SC curves; and e) V OC , J SC and FF of the devices modified with ALD-Al 2 O 3 layers (Deposition cycle: n = 0, 1, 3, 5 and 10).

Figure 6 .
Figure 6.s-EQE and EL spectra of a) devices with pristine ZnO and b) devices with ALD-Al 2 O 3 (Deposition cycle: n = 3).c) EQE EL of the corresponding OSCs.d) Schematic diagram of energy loss.

Table 1 .
Photovoltaic parameters of devices with different cycles of ALD-Al 2 O 3 .

Table 2 .
Energy losses of the control devices and devices modified with three cycles ALD-Al 2 O 3 .