Highly Stable and Enhanced Performance of p–i–n Perovskite Solar Cells via Cuprous Oxide Hole-Transport Layers

Solar light is a renewable source of energy that can be used and transformed into electricity using clean energy technology. In this study, we used direct current magnetron sputtering (DCMS) to sputter p-type cuprous oxide (Cu2O) films with different oxygen flow rates (fO2) as hole-transport layers (HTLs) for perovskite solar cells (PSCs). The PSC device with the structure of ITO/Cu2O/perovskite/[6,6]-phenyl-C61-butyric acid methyl ester (PC61BM)/bathocuproine (BCP)/Ag showed a power conversion efficiency (PCE) of 7.91%. Subsequently, a high-power impulse magnetron sputtering (HiPIMS) Cu2O film was embedded and promoted the device performance to 10.29%. As HiPIMS has a high ionization rate, it can create higher density films with low surface roughness, which passivates surface/interface defects and reduces the leakage current of PSCs. We further applied the superimposed high-power impulse magnetron sputtering (superimposed HiPIMS) derived Cu2O as the HTL, and we observed PCEs of 15.20% under one sun (AM1.5G, 1000 Wm−2) and 25.09% under indoor illumination (TL-84, 1000 lux). In addition, this PSC device outperformed by demonstrating remarkable long-term stability via retaining 97.6% (dark, Ar) of its performance for over 2000 h.


Materials and Methods
Cu 2 O films (10 nm) were deposited by DCMS, HiPIMS [49], and superimposed HiP-IMS [48] techniques, subsequently, on an ITO substrate. Before spin-coating the perovskite precursor solution, we used plasma to modify the Cu 2 O surface for 30 s and coat the precursor solution smoothly. Details of the device fabrication and characterization of the materials are shown in the Supplementary Materials.

Results
We fabricated the PSCs with a p-i-n structure of Glass/ITO/Cu 2 O/MAPbI 3 /PC 61 BM/ BCP/Ag. Figure 1 displays the X-ray diffraction patterns of Cu 2 O from DCMS and superimposed HiPIMS at different oxygen flow rates (f O2 = O 2 /(O 2 + Ar) × 100%, Ar: Argon O 2 : Oxygen). For the DCMS (Figure 1a), the increase in oxygen flow rate from f O2 = 10% started to form the Cu 2 O phase and became pure Cu 2 O phase at f O2 = 17.5%. Beyond this, it started to form a mixed Cu 2 O phase. For the normal HiPIMS sample, we observed the Cu 2 O phase at f O2 = 20%, which is consistent with our previously published study [49]. For the XRD of the superimposed HiPIMS process (Figure 1b), the increase in oxygen flow rate from f O2 = 17.5% started to form the Cu 2 O phase and became pure Cu 2 O phase at f O2 = 30%, and later, with a further increase in f O2 , it became a mixed-Cu 2 O phase. Accordingly, we obtained the optimal parameters for the Cu 2 O phase and used them as the HTLs of PSCs for further study. HiPIMS sample, we observed the Cu2O phase at fO2 = 20%, which is consistent with our previously published study [49]. For the XRD of the superimposed HiPIMS process (Figure 1b), the increase in oxygen flow rate from fO2 = 17.5% started to form the Cu2O phase and became pure Cu2O phase at fO2 = 30%, and later, with a further increase in fO2, it became a mixed-Cu2O phase. Accordingly, we obtained the optimal parameters for the Cu2O phase and used them as the HTLs of PSCs for further study. We determined the UV-Vis transmittance of these Cu2O-deposited ITO samples using air as a background (Figure 2a). The transmittance of the HiPIMS sample was slightly more than 80%, while in contrast, that of the DCMS and superimposed HiPIMS samples were each a bit less than 80% (<3%), in a wavelength range (500-900 nm). Due to the high ionization rate and high-energy plasma of the HiPIMS process, it offers dense and highly crystalline Cu2O films. Therefore, we observed higher transmittance of respective Cu2O in comparison with the DCMS sample (i.e., the structure was less dense). Although the superimposed HiPIMS process has the same advantages as the HiPIMS process, the MF power applied during the off-time bombarded more copper that became deposited as film, possibly contributing to the decrease in transmittance. We determined the UV-Vis transmittance of these Cu 2 O-deposited ITO samples using air as a background (Figure 2a). The transmittance of the HiPIMS sample was slightly more than 80%, while in contrast, that of the DCMS and superimposed HiPIMS samples were each a bit less than 80% (<3%), in a wavelength range (500-900 nm). Due to the high ionization rate and high-energy plasma of the HiPIMS process, it offers dense and highly crystalline Cu 2 O films. Therefore, we observed higher transmittance of respective Cu 2 O in comparison with the DCMS sample (i.e., the structure was less dense). Although the superimposed HiPIMS process has the same advantages as the HiPIMS process, the MF power applied during the off-time bombarded more copper that became deposited as film, possibly contributing to the decrease in transmittance. The values of Ra (the arithmetic average of the roughness profile) of the samples were determined by tapping-mode atomic force microscopy (AFM) to understand the surface morphology of ITO/Cu2O affecting the growth of perovskite; Figure 2b represents surface roughness through AFM images of the samples. We obtained values of Ra for the DCMS (3.06 nm), HiPIMS (2.94 nm), and superimposed HiPIMS (2.93 nm) samples. We speculated that HiPIMS would result in higher density and higher energy plasma than DCMS during deposition, which would result in a smoother surface. The superimposed HiPIMS sample had the same characteristics as the HiPIMS sample, leading to similar Ra values. Figure 3 shows the contact angles of the ITO/Cu2O surfaces when we dropped the perovskite precursor solution on top. As indicated in Figure 3a, the perovskite precursor formed a sphere on the HiPIMS Cu2O samples (i.e., a high angle of 85.13°), spreading poorly on the Cu2O surface and causing failure growth of the perovskite layer. The DCMS and superimposed HiPIMS Cu2O samples also had high angles, of 89.95° and 69.31°, respectively, as shown in Figure S1, which also led to poor spread of the perovskite precursor. To resolve this issue, we applied a plasma treatment to induce an ultraviolet The values of Ra (the arithmetic average of the roughness profile) of the samples were determined by tapping-mode atomic force microscopy (AFM) to understand the surface morphology of ITO/Cu 2 O affecting the growth of perovskite; Figure 2b represents surface roughness through AFM images of the samples. We obtained values of Ra for the DCMS (3.06 nm), HiPIMS (2.94 nm), and superimposed HiPIMS (2.93 nm) samples. We speculated that HiPIMS would result in higher density and higher energy plasma than DCMS during deposition, which would result in a smoother surface. The superimposed HiPIMS sample had the same characteristics as the HiPIMS sample, leading to similar Ra values. Figure 3 shows the contact angles of the ITO/Cu 2 O surfaces when we dropped the perovskite precursor solution on top. As indicated in Figure 3a, the perovskite precursor formed a sphere on the HiPIMS Cu 2 O samples (i.e., a high angle of 85.13 • ), spreading poorly on the Cu 2 O surface and causing failure growth of the perovskite layer. The DCMS and superimposed HiPIMS Cu 2 O samples also had high angles, of 89.95 • and 69.31 • , respectively, as shown in Figure S1, which also led to poor spread of the perovskite precursor. To resolve this issue, we applied a plasma treatment to induce an ultraviolet effect that modified the surface of the Cu 2 O films. As indicated in Figure 3b, the Cu 2 O film became hydrophilic with a low angle of 9.99 • , leading to the preferred growth of the perovskite layer. Nanomaterials 2023, 13, x FOR PEER REVIEW 5 of 12 effect that modified the surface of the Cu2O films. As indicated in Figure 3b, the Cu2O film became hydrophilic with a low angle of 9.99°, leading to the preferred growth of the perovskite layer.  Table 1 and Figure 4 summarize the device performances for various Cu2O films. Table 1 displays the standard deviations of the PSCs calculated from a minimum of three samples deposited using the same technique. The perovskite deposition was from a solution, which may cause slight variations during each deposition. Due to the variability in optoelectronic and surface properties caused by the deposition process of HTLs, the PSCs produced using various conditions of HTLs exhibited different device performances. DCMS-derived devices provided a PCE of 7.12 ± 0.64%, with a short-circuit current density (Jsc) of 13.74 ± 0.97 mAcm −2 , a value of open-circuit voltage (Voc) of 0.89 ± 0.02 V, and a fill factor (FF) of 58.1 ± 3.93%. The HiPIMS-based devices showed slightly higher efficiency, with a PCE of 9.42 ± 0.92%, a value of Jsc of 17.36 ± 0.94 mAcm −2 , a value of Voc of 0.90 ± 0.00 V, and an FF of 61.1 ± 6.95%. The superimposed HiPIMS devices showed the highest performance, with a PCE of 13.04 ± 1.61%, a value of Jsc of 19.27 ± 1.20 mAcm −2 , a value of Voc of 0.98 ± 0.00 V, and an FF of 69.5 ± 4.98%. The improvement in the PCE of the overlaid HiPIMS PSCs was brought on by substantial increases in the Jsc, Voc, and FFs. From the device's J-V curves (Figure 4), we computed the series (Rs) and shunt (Rsh) resistances. The resistances of the electrodes, the interfacial resistance, the hole/electron transporting layers, and the perovskite all contributed to the value of Rs. The highest value of Rs was for the DCMS device (9.47 Ω cm 2 ), and the lowest value of Rs was for the superimposed HiPIMS device (1.42 Ω cm 2 ). Further, we performed a four-point probe experiment to understand the electrical properties of Cu2O. This analysis showed the highest value of electrical conductivity (σ) for the superimposed HiPIMS Cu2O thin film (3.33 S cm −1 ) and the lowest for the DCMS Cu2O thin film (0.11 S cm −1 ). In addition, through XRD analysis ( Figure S2), the crystallinity of the Cu2O improved from the DCMS to superimposed HiPIMS deposition techniques. According to previous studies, an increase in film's crystallinity helps to increase the charge-carrier mobility (µ) [50]. In addition, varying the exposure times of the sputtering and plasma treatments impacted the optoelectronic properties of the thin films. In an attempt to reduce the exposure time during sputtering, we observed that the quality of the Cu2O film deteriorated significantly, potentially due to the film becoming discontinuous. Conversely, increasing the exposure time led to a well-formed Cu2O film. However, it resulted in unsatisfactory formation of the perovskite layer, and we are currently investigating the underlying reasons for this. As for the plasma treatment, we have not yet examined the impact of different exposure times on the optoelectronic properties of Cu2O film, as our primary goal was to improve its wettability. However, we speculate that excessively long exposure times during plasma treatment may damage Cu2O film and decrease its photoelectric properties. Lower values   Figure 4 summarize the device performances for various Cu 2 O films. Table 1 displays the standard deviations of the PSCs calculated from a minimum of three samples deposited using the same technique. The perovskite deposition was from a solution, which may cause slight variations during each deposition. Due to the variability in optoelectronic and surface properties caused by the deposition process of HTLs, the PSCs produced using various conditions of HTLs exhibited different device performances. DCMS-derived devices provided a PCE of 7.12 ± 0.64%, with a short-circuit current density (J sc ) of 13.74 ± 0.97 mAcm −2 , a value of open-circuit voltage (V oc ) of 0.89 ± 0.02 V, and a fill factor (FF) of 58.1 ± 3.93%. The HiPIMS-based devices showed slightly higher efficiency, with a PCE of 9.42 ± 0.92%, a value of J sc of 17.36 ± 0.94 mAcm −2 , a value of V oc of 0.90 ± 0.00 V, and an FF of 61.1 ± 6.95%. The superimposed HiPIMS devices showed the highest performance, with a PCE of 13.04 ± 1.61%, a value of J sc of 19.27 ± 1.20 mAcm −2 , a value of V oc of 0.98 ± 0.00 V, and an FF of 69.5 ± 4.98%. The improvement in the PCE of the overlaid HiPIMS PSCs was brought on by substantial increases in the J sc , V oc , and FFs. From the device's J-V curves (Figure 4), we computed the series (R s ) and shunt (R sh ) resistances. The resistances of the electrodes, the interfacial resistance, the hole/electron transporting layers, and the perovskite all contributed to the value of R s . The highest value of R s was for the DCMS device (9.47 Ω cm 2 ), and the lowest value of R s was for the superimposed HiPIMS device (1.42 Ω cm 2 ). Further, we performed a four-point probe experiment to understand the electrical properties of Cu 2 O. This analysis showed the highest value of electrical conductivity (σ) for the superimposed HiPIMS Cu 2 O thin film (3.33 S cm −1 ) and the lowest for the DCMS Cu 2 O thin film (0.11 S cm −1 ). In addition, through XRD analysis ( Figure S2), the crystallinity of the Cu 2 O improved from the DCMS to superimposed HiPIMS deposition techniques. According to previous studies, an increase in film's crystallinity helps to increase the charge-carrier mobility (µ) [50]. In addition, varying the exposure times of the sputtering and plasma treatments impacted the optoelectronic properties of the thin films. In an attempt to reduce the exposure time during sputtering, we observed that the quality of the Cu 2 O film deteriorated significantly, potentially due to the film becoming discontinuous. Conversely, increasing the exposure time led to a well-formed Cu 2 O film. However, it resulted in unsatisfactory formation of the perovskite layer, and we are currently investigating the underlying reasons for this. As for the plasma treatment, we have not yet examined the impact of different exposure times on the optoelectronic properties of Cu 2 O film, as our primary goal was to improve its wettability. However, we speculate that excessively long exposure times during plasma treatment may damage Cu 2 O film and decrease its photoelectric properties. Lower values of R s appear to be associated with higher σ and µ of Cu 2 O thin films and increased J sc values of the device. We speculate that the increase in J sc could have been due to a reduction in the carrier recombination rate at the p-i interface. This explanation is supported by other studies reporting that increasing hole conductivity can lead to higher J sc [51]. We acknowledge that we do not have the necessary facilities to perform spectral-response characterization experiments, but our findings align with the existing literature. Overall, the use of superimposed HiPIMS Cu 2 O as HTLs resulted in higher FF, J sc , and PCE values compared with the devices based on DCMS and HiPIMS. As indicated in Figure 4, 15.20% was the best performance of the superimposed HiPIMS devices. Table 2 compares Cu 2 O films as the HTLs of PSCs prepared by different processes; V oc , J sc , FF, and PCE are the best performance values of the device. The Cu 2 O films were prepared by various methods such as spin coating [52,53], successive ionic layer adsorption and reaction (SILAR) [54], sputtering [55,56], the thermal oxidation method [57,58], electrodeposition [59,60], chemical vapor deposition (CVD) [61], etc. Although it is known from simulation results that, theoretically, the PCE of Cu 2 O-based PSCs can be as high as 25.2% [62], it was found that the actual device PCE was mostly in the range of 6 to 13%. Based on the above method, we observed that the improvement in PCE was due to the low series and high shunt resistance of the respective devices. The high conductivity of the Cu 2 O thin film resulted in low recombination loss and smaller bulk resistance of the device. These factors contributed to a reduction in the energy loss within the device and increased the efficiency of charge extraction and transfer. This phenomenon is consistent with findings in previous reports [54,60]. of Rs appear to be associated with higher σ and µ of Cu2O thin films and increased Jsc values of the device. We speculate that the increase in Jsc could have been due to a reduction in the carrier recombination rate at the p-i interface. This explanation is supported by other studies reporting that increasing hole conductivity can lead to higher Jsc [51]. We acknowledge that we do not have the necessary facilities to perform spectralresponse characterization experiments, but our findings align with the existing literature.
Overall, the use of superimposed HiPIMS Cu2O as HTLs resulted in higher FF, Jsc, and PCE values compared with the devices based on DCMS and HiPIMS. As indicated in Figure 4, 15.20% was the best performance of the superimposed HiPIMS devices. Table 2 compares Cu2O films as the HTLs of PSCs prepared by different processes; Voc, Jsc, FF, and PCE are the best performance values of the device. The Cu2O films were prepared by various methods such as spin coating [52,53], successive ionic layer adsorption and reaction (SILAR) [54], sputtering [55,56], the thermal oxidation method [57,58], electrodeposition [59,60], chemical vapor deposition (CVD) [61], etc. Although it is known from simulation results that, theoretically, the PCE of Cu2O-based PSCs can be as high as 25.2% [62], it was found that the actual device PCE was mostly in the range of 6 to 13%. Based on the above method, we observed that the improvement in PCE was due to the low series and high shunt resistance of the respective devices. The high conductivity of the Cu2O thin film resulted in low recombination loss and smaller bulk resistance of the device. These factors contributed to a reduction in the energy loss within the device and increased the efficiency of charge extraction and transfer. This phenomenon is consistent with findings in previous reports [54,60].    In this study, we achieved a competitive result of 15.21% via embedding Cu 2 O thin films using superimposed HiPIMS. The MF power-supply sputtering in the superimposed HiPIMS system allowed us to retain a high ionization rate as the advantage of the HiPIMS. Compared with DCMS processes, HiPIMS and superimposed HiPIMS can bombard sputtering ions with higher energy, which helps to form the preferred stoichiometry of Cu 2 O films, resulting in a denser coating with lower surface roughness. It helped to reduce the defects in the film and led to a smoother interface, resulting in a decrease in the leakage current and an improvement in the PCE. Our EPMA analysis (Table S1) showed that the Cu 2 O chemical composition ratio as Cu/O was 1.9, which is close to the form of perfect stoichiometric Cu 2 O.
In addition to evaluating the surface morphologies of the Cu 2 O films, we also evaluated the energy level alignment at the Cu 2 O/p-i interface using an incident light energy of 21.2 eV (He(I) emission) via ultraviolet photoelectron spectroscopy (UPS). Figure 5 presents the E Cut-off regions of the DCMS, HiPIMS, and superimposed HiPIMS Cu 2 O samples (WF = 21.2 − E Cut-off ) [63,64]. The DCMS Cu 2 O film had a work function (WF) of -6.40 eV, while the WF of the HiPIMS Cu 2 O film was -6.38 eV, which was 0.02 eV lesser than that of the DCMS Cu 2 O film. The WF of the superimposed HiPIMS sample was -6.07 eV (the negative work function means how much energy is required to be added to the bound electron by the photon it absorbs). Changes in the WF at the p-i interface can significantly impact the energy alignment of Cu 2 O with perovskite, hole transport, and effectiveness of charge extraction as well as transfer. Therefore, a shift in the energy levels of the Cu 2 O may have contributed to the improved performance of the Cu 2 O devices. We believe that the significantly different composition ratio (Table S1) and phase structure ( Figure S2) of the superimposed HiPIMS Cu 2 O film may have affected the WF, resulting in the improved device performance.
We examined the shelf lives of the unencapsulated devices under an argon glove box in the dark to assess the stability of the superimposed HiPIMS Cu 2 O sample for PSC applications (we measured the PCE results at least three times to calculate the error bars). The devices initially offered the best PCE of 15.20%, as shown in Figure 6. After testing for 168 h, the devices retained 99.4% of their initial performance; after testing over 1000 h, the devices retained 98.3% of their initial performance; and after testing over 2000 h, the devices retained 97.6% of their initial performance. Thus, our PSC device suggested high stability. Finally, we determined the performance of the superimposed HiPIMS device under dim light (a TL-84 fluorescent lamp with an illumination of 1000 lux). Table 3 shows the performance of the respective device. The device showed a PCE of 25.09%. Based on these results, we demonstrated that Cu 2 O film has great potential as HTLs of PSCs under one sun and indoor illumination. We examined the shelf lives of the unencapsulated devices under an argon glove box in the dark to assess the stability of the superimposed HiPIMS Cu2O sample for PSC applications (we measured the PCE results at least three times to calculate the error bars). The devices initially offered the best PCE of 15.20%, as shown in Figure 6. After testing for 168 h, the devices retained 99.4% of their initial performance; after testing over 1000 h, the devices retained 98.3% of their initial performance; and after testing over 2000 h, the devices retained 97.6% of their initial performance. Thus, our PSC device suggested high stability. Finally, we determined the performance of the superimposed HiPIMS device under dim light (a TL-84 fluorescent lamp with an illumination of 1000 lux). Table 3 shows the performance of the respective device. The device showed a PCE of 25.09%. Based on these results, we demonstrated that Cu2O film has great potential as HTLs of PSCs under one sun and indoor illumination.   We examined the shelf lives of the unencapsulated devices under an argon glove box in the dark to assess the stability of the superimposed HiPIMS Cu2O sample for PSC applications (we measured the PCE results at least three times to calculate the error bars). The devices initially offered the best PCE of 15.20%, as shown in Figure 6. After testing for 168 h, the devices retained 99.4% of their initial performance; after testing over 1000 h, the devices retained 98.3% of their initial performance; and after testing over 2000 h, the devices retained 97.6% of their initial performance. Thus, our PSC device suggested high stability. Finally, we determined the performance of the superimposed HiPIMS device under dim light (a TL-84 fluorescent lamp with an illumination of 1000 lux). Table 3 shows the performance of the respective device. The device showed a PCE of 25.09%. Based on these results, we demonstrated that Cu2O film has great potential as HTLs of PSCs under one sun and indoor illumination.

Conclusions
In this study, we prepared Cu 2 O films via DCMS, HiPIMS, and superimposed HiPIMS and used them as HTLs for PSC applications. In the comparison of different process technologies for Cu 2 O samples, superimposed HiPIMS enhanced the WF (measured by UPS), surface roughness (measured by AFM), and optoelectronic properties of the Cu 2 O films and, in turn, increased carrier extraction and transport at the perovskite-Cu 2 O Nanomaterials 2023, 13, 1363 9 of 12 interface. Because the optimized Cu 2 O film was obtained from superimposed HiPIMS technology, we observed an increase in the PCE of the PSCs from 7.91 (DCMS) to 15.20%. The extended shelf life of the unencapsulated devices in the glove box attested to their excellent stability and support Cu 2 O as a contender for p-type transporting layers in PSC applications.