Back contact passivation with Ga-grading in narrow bandgap (Ag,Cu)(In,Ga)Se2 bifacial solar cells on In2O3:Sn back contact

Back contact passivation with Ga-grading in narrow bandgap (Ag,Cu)(In,Ga)Se2 bifacial solar cells on In2O3:Sn back contact Elizaveta Yakovleva1,2, André F. Violas3,4, Olivier Donzel-Gargand1, Klara Kiselman1, Patrick Pearson1, Pedro M. P. Salomé3,4, Lars Stolt1, Marika Edoff1,2 1Division of Solar Cell Technology, Department of Materials Science and Engineering, Uppsala University, 752 37 Uppsala, Sweden 2Wallenberg Initiative Materials Science for Sustainability, Department of Materials Science and Engineering, Uppsala University, 752 37 Uppsala, Sweden 3International Iberian Nanotechnology Laboratory (INL), 4715-330 Braga, Portugal 4Departamento de Física da Universidade de Aveiro, Campus Universitário de Santiago, 3810-193 Aveiro, Portugal Abstract — This study investigates the effect of Ga-grading in narrow bandgap (Ag,Cu)(In,Ga)Se2 (ACIGS) bifacial solar cells on In2O3:Sn (ITO) transparent substrates. It is found that the performance of cells on ITO equals or supersedes the performance of reference cells made on Mo. Up to 60% bifaciality was found for the cells on ITO. At the ACIGS/ITO interface a layer of 3 nm GaOx is identified with STEM and is believed to have a passivating effect on the back contact. An increase in the steepness of the Ga-grading results in better overall performance for front and rear side measurements, which likely indicates an added rear-contact passivation effect. I. INTRODUCTION (Ag,Cu)(In,Ga)Se2 (ACIGS) solar cells grown on transparent conducting oxides (TCOs) have reached good values of light to power conversion efficiency and have, thanks to their bifaciality, a potential to reach higher yearly yield than monofacial solar cells [1]. One of the limitations often found when going from Mo to TCO back contacts is a loss in fill factor (FF). It has been shown that for co-evaporated ACIGS, the deterioration of cells with TCOs as back contacts is related to an excess of Na enhancing GaOx growth [2]. The GaOx growth can be limited by lower deposition temperature, shorter deposition time, and a reduction of the Na amount [2]. High-efficiency ACIGS solar cells are achieved with >2 μm thick, graded absorbers, where the grading is confined to the back contact region [3]. When the absorber thickness is reduced significantly, to between 1000 and 500 nm, this type of grading of the absorber becomes harder to achieve [4] due to interdiffusion of Ga and In during the deposition. At the same time, the thinner absorbers make the cells more sensitive to back surface recombination, which makes the passivation of the back contact all the more important. Architectures for passivation and light-management have been explored in several research projects to increase the performance of ultrathin solar cells, including using dielectric passivation layers, e.g. Al2O3 or SiOx with openings for current transport [4]-[5]. In this work we present devices utilizing thin, 550-650 nm, graded ACIGS absorbers on In2O3:Sn (ITO) reaching values of efficiency of 15.2% and 9.7% under front and rear illumination, respectively. We achieve a graded bandgap by depositing a high Ga-concentration at the back contact. By using a low deposition temperature and a relatively short deposition time, the interdiffusion of Ga and In, as well as the GaOx growth, is limited. By using Ag, good crystallinity is achieved despite lower deposition temperatures [1]. II. METHODS ACIGS cells were fabricated on ITO substrates as well as on Mo for reference. The ITO substrate was the MSE PRO 1.1 mm ITO coated soda-lime glass (SLG) substrate with ITO thickness of 350 nm and a sheet resistivity of 3~5 Ohm/Sq. The Mo substrates had 300 nm Mo on 3 mm standard SLG. Both types of substrates were treated with NaF prior to absorber deposition. The ITO substrates were coated with 6 nm NaF, the Mo substrates with 12 nm NaF (when comparing cells with 6 nm and 12 nm NaF on Mo substrates, there was no difference in performance). ACIGS absorbers were co-evaporated according to the threestage process described in [6] but with a reduced deposition time and hence a reduced thickness. The target composition ratios for the bulk absorber were [Ga]/([Ga]+[In]) = 0.20 (GGI), ([Ag]+[Cu])/([Ga]+[In]) = 0.80 (ACGI), [Ag]/([Ag]+[Cu]) = 0.10 (AAC), measured in the deposited material with X-ray fluorescence (XRF). Two sets of cells were made using two profiles, one with a low Ga-grading and the other with a high Ga-grading. The Gagrading was achieved by increasing the GGI in the beginning of the deposition, while keeping a constant ACGI, before decreasing the GGI to a constant value for the rest of the deposition. The low Ga-grading profile aimed at doubling the GGI at the back contact compared to the bulk GGI, the high Gagrading profile aimed at tripling the GGI at the back. In Fig. 1, a sketch of the deposition profile, with the variation of Ga and In for the two profiles (profile 1 for low Ga-grading, 979-8-3315-3444-8/25/$31.00 ©2025 IEEE 0229 2025 IEEE 53rd Photovoltaic Specialists Conference (PVSC) | 979-8-3315-3444-8/25/$31.00 ©2025 IEEE | DOI: 10.1109/PVSC59419.2025.11133411 Authorized licensed use limited to: Uppsala Universitetsbibliotek. Downloaded on October 24,2025 at 13:40:13 UTC from IEEE Xplore. Restrictions apply. profile 2 for high Ga-grading), is shown. The substrate reached a maximum temperature of 450°C during deposition. The active deposition time was less than 8 minutes (460s) for both runs. The resulting absorber thickness for the cells made with profile 1 was on average 550 nm, and for the cells with profile 2, 650 nm, measured with Scanning Transmission Electron Microscopy (STEM). No post-deposition alkali treatment was done. Fig. 1. Deposition profile with rates of the co-evaporated elements. The low Ga profile has the Ga1 and In1 rates in the beginning of the deposition, while the high Ga profile has the Ga2 and In2 rates. The second profile had slightly higher rates for all elements to achieve a thicker absorber. After ACIGS deposition, a 50 nm layer of CdS was deposited in a CBD process also described in [6]. The i-ZnO/ZnO:Al (AZO) window layers were deposited through sputtering. Cells with an area of 0.1 cm2 were defined with photolithography and chemically etched with acetic acid. Measurements of current density-voltage (JV) were performed at AM1.5G with a Pico™ Small Area LED Solar Simulator. External Quantum Efficiency (EQE) was measured with a home-built system using filtered light from a Xe lamp. The composition profile of the samples was analyzed by Glow-discharge optical emission spectroscopy (GDOES). The samples were further prepared by Ga-based Focused Ion Beam (FIB), then analyzed with STEM and Energy-dispersive X-ray Spectroscopy (STEM-EDS) at 200 keV. III. RESULTS A. Performance of Ga-graded ACIGS on ITO To compare the resulting sets of cells, made with the two profiles and on two different substrates, a short name for each sample set is introduced in Table I. The solar cell characteristics for the best cell from each set of samples are also summarized in Table I. The efficiencies were calculated with the JSC integrated from EQE measurements. The JV-parameters for front and rear illumination show a bifaciality (Effrear/Efffront) close to 60% for the cells on ITO. The EQE spectra measured for the cells listed in Table I are presented in Fig. 2. In the EQE spectra for the cells illuminated from the rear we find a similar reduction in the near infrared (NIR) wavelengths (800-1200 nm) for both thicknesses and Gagradings. This loss can presumably be explained by high freecarrier absorption in the commercial ITO and will be investigated further. Fig. 2. EQE for the best cells for Mo-lowG, Mo-highG, ITO-lowG and ITO-highG (front and rear illumination). An improvement in JSC and FF can be seen for the cells with higher Ga-grading. Both JSC and FF increase the efficiency, which can be a contribution of the increased absorber thickness as well as the passivating effects of the Ga-grading. Further experiments are needed to distinguish the role of the thickness from the back passivation effects, however the comparison between ITO and Mo is valid for each type of Ga gradient and temperature. As such, for both profiles, the cells made on ITO substrates showed a higher performance than the reference cells on Mo. Most notably, FF values are improved when going from Mo to ITO. Such a result is a preliminary proof-of-concept that ITO can be a good substitute for Mo as a back contact material. B. Material characterization The GGI profile calculated from the GDOES measurements is presented in Fig. 3. The ITO samples have an almost flat GGI profile compared to the Mo samples, where even Mo-lowG has a significant slope in the GGI profile. Diffusion of In from the ITO and formation of GaOx at the back contact lower the GGI on ITO samples. The deposition profile with higher GGI ratio at the beginning of the deposition (profile 2 in Fig. 1) resulted in a more pronounced GGI profile on both substrates. The thickness of the GaOx layer was estimated with STEM to be 3 nm for both sample ITO-lowG and ITO-highG, which could indicate that the GaOx layer thickness is not necessarily 979-8-3315-3444-8/25/$31.00 ©2025 IEEE 0230 Authorized licensed use limited to: Uppsala Universitetsbibliotek. Downloaded on October 24,2025 at 13:40:13 UTC from IEEE Xplore. Restrictions apply. increased with a higher Ga-content. Despite the GaOx layer, which consumes Ga and lowers the GGI profile as a result, the cells on ITO still performed better than on Mo. This indicates that the GaOx layer might act as a back contact passivation layer. Fig. 3. GGI ratios from GDOES for the four types of samples MolowG, Mo-highG, ITO-lowG and ITO-highG IV. CONCLUSION Two different Ga-grading profiles were evaluated for ACIGS cells on ITO. Compared to reference cells deposited on Mo, the cells on ITO reached higher efficiencies, and most notably gained in FF. With the higher Ga-grading and slightly thicker absorber of this study, the cells reached a maximum efficiency of 15.2%. The cells performed well under rear illumination, reaching a bifaciality of 60%. However, no increase in the EQE for NIR-wavelengths is seen under rear illumination, which would have been expected unless there is considerable parasitic absorption in the ITO. GaOx was identified in the ITO devices and its effects in these samples appear to be beneficial, likely due to passivation of the back contact. Further studies are needed to identify the contribution of the Ga-grading passivating effects to the high efficiencies seen under front and rear illumination in this study. ACKNOWLEDGEMENT This work was partially supported by the Wallenberg Initiative Materials Science for Sustainability (WISE) funded by the Knut and Alice Wallenberg Foundation. Additionally Fundação para a Ciência e Tecnologia (FCT) grant with number 2020.07073.BD is acknowledged. This project was conducted in collaboration with the European project Hi-BITS, grant agreement 101122203, funded by the European Union. REFERENCES [1] S. Yang, T. Lin, M. Ochoa, H. Lai, R. Kothandaraman, F. Fu, A. N. Tiwari, R. Carron, "Efficiency boost of bifacial Cu(In,Ga)Se2 thin-film solar cells for flexible and tandem applications with silver-assisted low-temperature process," Nature Energy, vol. 8, pp. 40-51, 2023. [2] J. Keller, L. Stolt, O. Donzel-Gargand, T. Kubart, M. Edoff, "Wide-Gap Chalcopyrite Solar Cells with Indium Oxide–Based Transparent Back Contacts," Solar PRL, vol. 6, 2022. [3] J. Keller, K. Kiselman, O. Donzel-Gargand, N. M. Martin, M. Babucci, O. Lundberg, E. Wallin, L. Stolt, M. Edoff, "Highconcentration silver alloying and steep back-contact gallium grading enabling copper indium gallium selenide solar cell with 23.6% efficiency," Nature Energy, vol. 9, pp. 467–478, 2024. [4] A. F. Violas, A. J. N. Oliveira, E. Yakovleva, B. L. Sieira, F. Pinto, B. Rocha, E. J. Ribeiro, X. L. Pinheiro, O. Bondarchuk, J. Capitão, A. Medes, P. A. Fernandes, J. P. Teixeira, M. Edoff, P. M. P. Salomé, "One-step Lithography Nanostructured Au Encapsulation for Light Management in Ultrathin ACIGS Solar Cells," Solar PRL, 10 January 2025. [5] B. Vermang, V. Fjällström, J. Pettersson, P. Salomé, M. Edoff, " Development of rear surface passivated Cu(In,Ga)Se2 thin film solar cells with nano-sized local rear point contacts," Solar Energy Materials and Solar Cells, 2013. [6] M. Edoff, T. Jarmar, N. S. Nilsson, E. Wallin, D. Högström, O. Stolt, O. Lundberg, W. Shafarman, L. Stolt, "High Voc in (Cu,Ag)(In,Ga)Se2 Solar Cells," IEEE Journal of photovoltaics, vol. 7, no. 6, 2017. TABLE I SAMPLE NAMING DESCRIPTION AND SUMMARY OF IV PARAMETERS MEASURED UNDER FRONT AND REAR ILLUMINATION Sample name Substrate Sample description Light IV parameters (best cells) Profile Avg. thick. Eff JSC (EQE) VOC FF (nm) (%) (mA/cm2) (mV) (%) ITO-lowG ITO Low Ga-grading 550 13.8 28.8 637 75.5 Mo-lowG Mo Low Ga-grading 550 13.2 29.9 608 72.6 ITO-highG ITO High Ga-grading 650 15.2 31.3 637 76.4 Mo-highG Mo High Ga-grading 650 15.0 31.4 633 75.5 ITO-lowG (rear illumination) ITO Low Ga-grading 550 8.1 16.8 630 76.3 ITO-highG (rear illumination) ITO High Ga-grading 650 9.7 20.0 631 76.8 979-8-3315-3444-8/25/$31.00 ©2025 IEEE 0231 Authorized licensed use limited to: Uppsala Universitetsbibliotek. Downloaded on October 24,2025 at 13:40:13 UTC from IEEE Xplore. Restrictions apply.