Symposium EN04—Silicon for Photovoltaics

All major future energy scenarios forecast a key role for photovoltaic solar energy. The crystalline silicon PV technology will dominate the future residential and large-scale utility PV applications. The learning objective of this comprehensive tutorial are 1) understanding the fundamental principles of photovoltaic energy conversions; 2) understanding of the physics and chemistry behind materials and interface engineering for crystalline silicon PV technology, 3) understanding of the design principles of the various silicon PV device architectures and module concepts and 4) the understanding of the role of a circular value chain of crystalline silicon photovoltaics in the energy transition towards a 100% renewable energy future.

The Fundamental Principles of Photovoltaic Energy Conversion

The key factor in getting more efficient and cheaper solar energy panels is the advance in the development of photovoltaic cells. In part 1 you will learn how photovoltaic cells convert solar energy into useable electricity. You will also discover how to tackle potential loss mechanisms in solar cells. By understanding the semiconductor physics and optics involved, you will develop in-depth knowledge of how a photovoltaic cell works under different conditions.

Materials and Interface Engineering for Crystalline Silicon PV Technology

The nature of the materials and interfaces for crystalline silicon PV will be presented. In Part 2 you will gain the latest insights how impurities, doping, crystal orientation, interstitials, and defects affect the optical and electrical performance of the novel materials and interfaces. A special focus will be on new materials, processing methods and innovative interface engineering (chemical and electrical passivation) for passivating contacts.

Introduction in to Silicon PV Device Architectures and Module Concepts

In part 3 you will be introduced to the recent advances in the development of the various crystalline silicon device architectures. You will use the design rules, to reduce the optical and electrical losses as discussed in Part 1, to understand the underlying principles of Al Back Surface Field solar cells, Partial Back Contact (PERC) solar cells, heterojunction junction solar cells (HJ/HIT), interdigitated back contact (IBC), tunnel oxide passivated contact (TOPCon), bifacial solar cells and hybrid multi-junction PV concepts. In addition, you will be introduced to the design rules to manufacture PV modules using PV cells. The various modules architectures and the recent advances will be reviewed.

The Role of Circular Value Chain of Crystalline Silicon Photovoltaics in The Energy Transition Towards a 100% Renewable Energy Future

All major future energy scenarios forecast a key role for photovoltaic solar energy. In part 4 the important role of crystalline silicon in future residential and large-scale utility PV applications will be discussed. You will be introduced to new metrics -beyond that of module efficiencies- that will be important for the successful upscaling of the PV technologies in the coming decades.

In this talk, I will share our recent findings regarding silicon module durability arising from tools we have developed to study potential induced degradation (PID) and the effects of moisture. PID has re-emerged as an alarming reliability issue in silicon PV. Sodium is known to be correlated with PID of the shunting type, but a quantitative description of its kinetics in PV modules has been lacking. I will describe a finite element method model of PID we have developed that relies only on parameterization of Na kinetics within the module stack. This parameterization can be directly correlated with the bill of materials. We use a series of bottom-up and top-down experimental approaches to validate the model, including a new method to measure the kinetics of mobile ion migration through dielectric layers. Ultimately, we provide a sensitivity analysis of PID rates on the bill of materials that can serve as a guide toward robust modules.

The second part of the talk will focus on moisture impacts. Moisture is implicated in many modes of performance degradation in silicon PV, but it has proved difficult to assess the water content within modules to determine its direct role in degradation reactions. We have developed a short-wave IR optical Water Reflectometry Detection (WaRD) technique to quantify moisture levels within the module. We combine this moisture measurement with correlative luminescence imaging to track water content and cell performance at millimeter resolution over 2500 hours of accelerated testing at multiple levels of humidity and temperature. The approach reveals dominant modes of power loss with distinct and rather counterintuitive moisture sensitivities in glass-glass vs. glass-backsheet packages. Metallization interrupts in fingers occur earliest in dry conditions and are delayed in backsheet modules at higher humidity, suggestive of plasticization of the backsheet. Background series resistance sharply increases in backsheet modules at high humidity, by a mechanism not seen in glass-glass modules. Overall, we dissect the losses over time to attribute the marginal power loss from each mode as a function of temperature, humidity, and package type.

These new tools for durability science should aid in understanding the progression and severity of module degradation modes, the development of targeted accelerated tests, and ultimately new technologies that extend silicon durability.

Solar photovoltaic (PV) technologies are at the forefront of the clean energy transition, with at least 1 TW of PV installations projected by 2050. There will be a cumulative demand for 56 million metric tonnes of virgin material by 2050, considering future PV technology evolution and current end-of-life (EoL) management practices. 17 million metric tonnes of waste will need to be managed as PV modules have a lifetime of around 30 years. To mitigate negative impacts and ensure the increased solar energy deployment is sustainable, a circular economy (CE) for PV must be implemented. A quantitative analysis of waste-management interventions is needed to identify which actions should be taken to promote a CE for crystalline-Silicon (c-Si) PV while integrating future technology improvements and social factors. Here, we use a novel, dynamic, Python-based tool called PV ICE (PV in the Circular Economy), created by researchers at the National Renewable Energy Laboratory (NREL), to quantify the future mass flow impacts of ten waste-management interventions. The interventions and corresponding EoL pathway rates come from Walzberg et al.’s agent-based model (ABM). This ABM considered social factors to predict human behavior and decision-making surrounding PV EoL. Progressive simulations were run using PV ICE to explore the impact of improving circular EoL pathways, recycling efficiencies, and module reliability. The results suggest that a certain intervention, Landfill ban, can drastically reduce cumulative PV EoL waste to only 2 metric tonnes by 2050. Improving module reliability (Improved lifetime) is shown to have a powerful impact on maintaining installed capacity, increasing from 0.77 TW to 0.95 TW by 2050. When seeking to maintain a target capacity, lower quality modules can demand up to 16% more virgin materials and produce 200% more waste, pointing to the importance of module reliability for circularity. Further research should widen the scope of the intervention analysis to include energy and cost considerations.

Bifacial photovoltaics are expected to gain significant market share over the next decade due to the reduced levelized cost of energy. However, the thermal performance of bifacial modules is not yet well understood, nor have thermal management strategies been developed for this module configuration. Elevated operating temperatures in Si photovoltaics led to both diminished efficiency and accelerated module degradation.
A significant percentage of this waste heat is generated by sub-bandgap parasitic absorption within the module. Because of this, spectrally selective mirrors that act as above bandgap antireflection coatings and sub-bandgap light rejectors are a promising photon management strategy for reducing the operating temperature of monofacial photovoltaics. However, it remains an open question whether bifacial photovoltaics that possess partial sub-bandgap transparency thermally benefit from this light management strategy. Here we use rigorous finite-element simulations to study the performance benefits of spectrally-selective mirrors on bifacial modules.
The optical properties of test modules containing commercial, 20.6 % efficient bifacial PERC cells were modeled using a combination of ray-tracing, Beer’s law, and transfer matrix calculations. This optical description was combined with a view-factor irradiance model with SMARTS2 integration and a finite-element model to solve for the module’s temperature and power output over a typical meteorological year.
We first studied idealized mirrors, with unity above-bandgap transmittance and unity sub-bandgap reflectance, to determine the limits to performance. To mimic laboratory tests, AM1.5G irradiance was normally incident on the module’s front face while irradiance on the rear face was varied. When there is no rear irradiance, applying mirrors at either the front only or the front and rear provided an approximate 4 % optical benefit, corresponding to the reflectance of an air/glass interface, as well as a 1.1 % thermal benefit, equating to 2.7 °C of module cooling. These results show that partial sub-bandgap transparency does not prohibit spectrally selective mirrors from being a promising thermal management strategy. As the rear irradiance increased, the configuration with the mirror only on the front exhibited decreasing optical and thermal benefits, due to a decreasing percentage of above-bandgap light incident on the front face and sub-bandgap light trapping of rear side incident light, respectively. With mirrors on both interfaces, the optical benefit was maintained as rear irradiance increased, and the thermal benefit increased linearly due to the linear increase in sub-bandgap irradiance rejection.
We then developed 4- and 6-layer photonic mirrors using real materials and compared performance to a traditional antireflection coating (ARC). The mirror designs were classified as thermally aware ARCs, providing 0.2 and 0.6 % more above bandgap antireflection than a traditional ARC, without the associated -0.13 % thermal penalty. Adding a rear spectrally-selective mirror improved the optical benefit when placed over high albedo ground, but also led to an increase in sub-bandgap light trapping, as observed by a small, 0.02 % thermal penalty.
In addition to the studies described above, the idealized and realistic mirrors' performance were tested in single-axis tracking arrays at varying row spacings, array height, and at 12 different US cities. Each study provided insight into how varying angle of incidence distributions, rear irradiance fractions, total irradiance, and varying climate affect the optical and thermal benefits the mirrors can provide.
Our results indicate that spectrally-selective mirrors enhance the energy yield of bifacial photovoltaics through a combination of improved anti-reflection and reduced operating temperature.

Millions of years before people began to generate functional nanostructures, biological systems were using nanometer-scale architectures to produce unique functionalities. Some nocturnal moths use hexagonal arrays of subwavelength nipples as antireflection coatings to reduce reflection from their compound eyes. Similar periodic arrays of nanopillars have also been observed on the wings of cicada to render superhydrophobic surfaces for self-cleaning functionality. Inspired by these natural nanostructures, we have developed three scalable nanofabrication technologies based on colloidal self-assembly, maskless reactive ion etching, and metal-assisted chemical etching for producing wafer-scale, self-cleaning, broadband antireflection coatings on a large variety of PV-relevant substrates, such as single-crystalline silicon, multicrystalline silicon, GaAs, GaSb, and glass. These techniques combine the simplicity and cost benefits of bottom-up self-assembly with the scalability and compatibility of standard top-down microfabrication. The resulting subwavelength-structured antireflection coatings can greatly suppress light reflection from both crystalline silicon and encapsulating glass surfaces, enhancing light absorption and ultimate conversion efficiency of the crystalline silicon solar panels. Importantly, optimal moth-eye antireflection nanostructures with tapered geometries can be easily passivated by conventional passivation technologies, resolving an outstanding challenge faced by many other black silicon approaches.

Concentrator solar cells have served as the absolute efficiency record holder for more than 25 years. Contrary to non-concentrator photovoltaics (PV), where silicon has been the flagship absorber material, concentrator PV research has been dominated by high efficiency III-V absorber layers. The main reason for this is that high Auger losses impose a comparably lower efficiency limit onto Si solar cells under concentration. Beyond the higher efficiencies of concentrator solar cells due to a logarithmic gain in open-circuit voltage, the amount of material needed per unit area scales inversely with increasing concentration. While Auger recombination might be an overall limit for efficiencies of Si concentrator solar cells, the option of saving material per unit area is becoming increasingly attractive. Considering projections of worldwide shortages for elements such as Ag, Cu, In etc, partially caused by the quickly rising demand of the Si PV industry itself, the time for concentrator Si PV might have come.

In this work we introduce a patterned and transparent polymer structure that attempts to address two problems that do not only apply to Si concentrator PV (CPV) but to CPV in general. First of all we want to solve the unsatisfying trade-off between shading losses that arise from high front metal coverage and Joule losses that increase with lower front metal coverage. Secondly, the same structure will actually allow for a decrease of the total area required to obtain a certain short-circuit current density (J_SC), due to lower shading losses.

The mentioned structure is a transparent polymer layer with V-shaped grooves, imprinted above the cell’s metal fingers. These V-grooves redirect light away from the metal grid and onto the absorber material. This renders the contacts effectively transparent and breaks the trade-off between shading losses and Joule losses from the metal grid, thus allowing for higher concentration ratios and higher efficiency. We will demonstrate how performance of a Si solar cell with 25% front metal grid coverage and an initial short-circuit current density (J_SC) of 29.95 mA/cm² can be improved to a J_SC of 39.12 mA/cm² after adding the patterned polymer, mitigating shading losses almost completely. Optical microscope images as well as reflection data from integrating sphere measurements confirm that the contacts have become effectively transparent, showing little to no metal-grid-related features in the reflection spectrum. Furthermore, we will show that the angular performance of the V-grooves allows for shading losses to remain absent for concentrations beyond 1000 suns. Lastly, we present experimental IV-data of solar cells with V-grooves under concentration.

To conclude, the polymeric V-groove design allows for complete mitigation of shading losses, even with high metal coverage. In a concentrator system, this would allow for higher efficiencies, and especially higher power output per unit solar cell area compared to non-concentrating PV. Furthermore, lower shading losses also mean less area usage per concentrator system, as less concentration is needed to achieve the same J_SC. This could be beneficial as it might allow for cheaper and simpler concentration optics.

The solar industry has seen rapid growth over the last few decades with installations over 600 GW contributing to over 2.2% of the global electricity¹. Silicon (Si) solar cells contribute to the bulk of the PV market, having more than 95% of the total production. The trend is for advanced high-efficiency device structures on n-type monocrystalline silicon to dominate by 2025%². Current examples having commercial-scale production include tunnel-oxide passivated contact (TOPCon), Interdigitated back contact (IBC), and Heterojunction (HJ), which employs advanced improved interface passivation. In particular, Si HJ structures could be favorable due to their ease of fabrication, compatibility with thin wafers (< 100 µm), low thermal budget (< 300°C), and potential to reduce the cost per watt. The performance of these solar cells critically depends on the interface passivation quality of the amorphous silicon (a-Si:H) and crystalline silicon (c-Si) hetero-interface. It has been shown that extrinsic methods like hydrogen plasma treatment (HPT) and thermal annealing treatment (AT) strongly enhance the passivation of dangling bonds at the a-Si:H/c-Si interface³. In this work, we have investigated the underlying surface passivation mechanism of HPT and AT by analyzing the hydrogen bonding using Fourier transform Infrared spectroscopy. The hydrogen bonding environment in a-Si:H samples undergoing different treatments has been corroborated with the passivation quality in terms of effective minority carrier lifetime and implied open-circuit voltage (iV_OC) measured by quasi-steady-state photoconductance. Our initial results show that just AT reduces the Si-H₂ dihydrides, whereas a combination of HPT with AT improves monohydride concentration, in addition to the reduction of the dihydride complexes, and resulted in higher minority carrier lifetime & implied V_OC. Temperature-dependent infrared spectroscopic analysis helps to understand the kinetics of hydrogen redistribution and effusion from the a-Si:H films due to annealing and will probe the activation energies of these processes. This will provide a complete understanding of the passivation improvement mechanism under the post-deposition extrinsic treatments. Additional characterization of the HJ structures was performed using Raman and UV-VIS-NIR spectroscopy to validate the quality of the a-Si:H films and interfaces to be used for making Silicon Heterojunction solar cells.

References:
(1) Solar PV - Analysis - IEA https://www.iea.org/reports/solar-pv.
(2) Fraunhofer Institute for Solar Energy Systems. Photovoltaics Report; 2020.
(3) Soman, A.; Nsofor, U.; Das, U.; Gu, T.; Hegedus, S. Correlation between in Situ Diagnostics of the Hydrogen Plasma and the Interface Passivation Quality of Hydrogen Plasma Post-Treated a-Si:H in Silicon Heterojunction Solar Cells. ACS Appl. Mater. Interfaces 11, 16181–16190.

Black silicon (Si) is a nanostructured type of Si capable of near-total absorption of light. Therefore, the use of black Si can lead to highly efficient solar cells. The performance of Si solar cells is dependent on the active dopant carrier distribution in emitters. The carrier distribution in black Si solar cells has been difficult to measure, and thus to optimize, due to the irregular, nano-scaled structure of the emitter.[1]-[5]
Thus, obtaining more information regarding active dopant density distribution in black Si solar cell requires measurements of carrier distributions on a nanoscopic scale.
One of the authors have previously succeeded in quantitatively analyzing such distributions in monocrystalline Si solar cells using scanning nonlinear dielectric microscopy (SNDM) [6]-[9].
In the present study, we investigated quantitatively the carrier distribution in a phosphorous (P) diffused black Si solar cell using SNDM. As a reference, we also measured the carrier distribution on a flat Si sample fabricated under the same P diffusion conditions. The precise carrier distributions in the emitter were visualized, which revealed the feature of carrier distribution in the emitter of black Si solar cell. Super-higher-order-SNDM was also employed to perform a quantitative analysis of the depletion layer distribution. It was found that the carrier density profile and the depletion layer thickness is less regular in the black Si than in the flat emitter, suggesting that this fluctuation may affect the power conversion efficiency of black Si solar cell.
This work demonstrates that SNDM is a very useful means of investigating carrier distributions in the nano-scaled emitter of black Si solar cell.

References:
[1] Y. Liu, T. Lai, H. Li, Y. Wang, Z Mei, H. Liang, Z. Li, F. Zhang, W. Wang, A.Y. Kuznetsov, X. Du, Small, 8, 1392 (2012).
[2]K. Peng, Y. Xu, Y. Wu, Y. Yan, S.-T. Lee, J. Zhu, Small, 1, 1062 (2005).
[3] K. Peng, X. Wang, S.T. Lee, Appl. Phys. Lett., 92, 163103 (2008).
[4] J Z Shen, B. Liu, Y. Xia, J. Liu, J. Liu, S. Zhong, C. Li, Scr. Mater., 68, 199 (2013).
[5] C.-H. Hsu, J.-R. Wu, Y.-T. Lu, D. J. Flood, A. R. Barron, L.-C. Chen, Materials Science in Semiconductor Processing, 25, 2 (2014).
[6] K. Hirose, K. Tanahashi, H. Takato, and Y. Cho, Appl. Phys. Lett. 111, 032101 (2017).
[7] Y. Cho, S. Jonai, and A. Masuda, Appl. Phys. Lett., 116, 182107 (2020).
[8] Y. Cho, A. Kirihara, and T. Saeki, Rev. Sci. Instrum. 67, 2297 (1996).
[9] Y. Cho, “Scanning Nonlinear Dielectric Microscopy”, Elsevier, ISBN 9780128172469 (2020).

The desired direction for modern c-Si solar cells is to make them both thinner and higher efficiency, which makes passivation of the c-Si surface to reduce interface recombination a top priority. Traditional passivation techniques such as thermal oxidation and a-Si can cause issues in processing due to various temperature limitations [1]–[3]. Passivation of the c-Si surface using H₂S gas has been demonstrated to produce high quality passivation and has resulted in high (>2000 µs) minority carrier lifetimes [4]. It has been demonstrated that by H₂S reaction, c-Si surface dangling bonds can be reduced by bonding with S [4].

This work presents an investigation of the surface passivation mechanism by S by capacitance-voltage-frequency (C-V-f) measurements and will be compared with the surface passivation process by SiN_x, and a-Si. Surface passivation quality is quantified by effective minority carrier lifetime (τ_eff) measured by quasi-steady-state photoconductance and the interface trap density (D_it) and fixed charge (Q_fix) are obtained from C-V-f measurements. The metal-insulator-semiconductor (MIS) structures were fabricated with the Si surface passivated by different materials for C-V-f measurements.

Lifetime and CV data are used together to characterize the passivation quality of the different treatments on two types of c-Si wafers. Both the a-Si and H₂S treatments ( τ_eff ≈ 1000 μs and 1700 μs for a-Si and H₂S treatment, respectively) show better passivation than the SiN_x,( τ_eff ≈ 14 μs) with the H₂S giving the highest lifetime. Additionally, the data shows a reduction in Q_fix when the H₂S passivation process is used compared to the SiN_x sample, with Q_fix = 7.5×10¹¹ cm^-2 and 2.1×10¹³ cm^-2, for H₂S and SiN_x respectively.

The change in lifetime is explained by Q_fix and D_it which are functions of passivation quality. CV data allows calculation of Q_fix and D_it, which are used to fit the lifetime data using the method of Shu et al. [7][8], and further allowing for extraction of electron and hole capture cross sections (σ_n, σ_p).

In summary, two types of c-Si wafer with three different passivation treatments will be used to prepare MIS devices for CV measurement. Fixed charge calculated based on CV data shows a reduction in Q_fix for H₂S samples compared to the SiN_x passivated sample. Additionally, qualitative analysis of the CV curves as well as fitting of lifetime data shows a reduction in D_it for the H₂S samples, indicating high-quality passivation of the c-Si surface is due to low D_it ≈ 4.0×10¹⁰cm^-2eV^-1 with a moderate positive Q_fix.

A detailed comparative study of interface passivation mechanisms by SiN_x, a-Si, and H₂S treatment will be provided in the final conference submission. Additionally, low frequency capacitance measurements, and conductance vs. AC frequency measurements will be included for all samples, which will allow for an accurate and quantitative determination of D_it for all samples. This in turn will yield a more robust lifetime fit to extract values for capture cross sections.

[1] A. G. Aberle, et al., Progress in Photovoltaics: Research and Applications, Vol. 2, pp. 265-273, 1994.
[2] U. K. Das, et al., Apl. Phys. Lett., Vol. 92, 063504, 2008.
[3] W. Liu, et al., ACS Appl. Mater. Interfaces, Vol. 7, pp. 26522-26529, 2015.
[4] H. S. Liu, et al., Langmuir, Vol. 33, pp. 14580-14585, 2017.
[5] P. A. G. O’Hare, Pure Appl. Chem.,Vol. 71, pp. 1243-1248, 1999.
[6] A. Haas, Angew. Chem. Int. Ed. Engl., Vol. 4, pp. 1014-1023, 1965.
[7] Z. Shu, et al., Prog. in Photovoltaics: Research and Applications, Vol. 23, pp.78-93, 2013.
[8] U. Das, et al., Proc. 7th WCPEC, 2018.
[9] D. K. Schroder, John Wiley & Sons, 2006.

The desired direction for modern c-Si solar cells is to make them both thinner and higher efficiency, which makes passivation of the c-Si surface to reduce interface recombination a top priority. Traditional passivation techniques such as thermal oxidation and a-Si can cause issues in processing due to various temperature limitations [1]–[3]. Passivation of the c-Si surface using H₂S gas has been demonstrated to produce high quality passivation and has resulted in high (>2000 µs) minority carrier lifetimes [4]. It has been demonstrated that by H₂S reaction, c-Si surface dangling bonds can be reduced by bonding with S [4].

This work presents an investigation of the surface passivation mechanism by S by capacitance-voltage-frequency (C-V-f) measurements and will be compared with the surface passivation process by SiN_x, and a-Si. Surface passivation quality is quantified by effective minority carrier lifetime (τ_eff) measured by quasi-steady-state photoconductance and the interface trap density (D_it) and fixed charge (Q_fix) are obtained from C-V-f measurements. The metal-insulator-semiconductor (MIS) structures were fabricated with the Si surface passivated by different materials for C-V-f measurements.

Lifetime and CV data are used together to characterize the passivation quality of the different treatments on two types of c-Si wafers. Both the a-Si and H₂S treatments ( τ_eff ≈ 1000 μs and 1700 μs for a-Si and H₂S treatment, respectively) show better passivation than the SiN_x,( τ_eff ≈ 14 μs) with the H₂S giving the highest lifetime. Additionally, the data shows a reduction in Q_fix when the H₂S passivation process is used compared to the SiN_x sample, with Q_fix = 7.5×10¹¹ cm^-2 and 2.1×10¹³ cm^-2, for H₂S and SiN_x respectively.

The change in lifetime is explained by Q_fix and D_it which are functions of passivation quality. CV data allows calculation of Q_fix and D_it, which are used to fit the lifetime data using the method of Shu et al. [7][8], and further allowing for extraction of electron and hole capture cross sections (σ_n, σ_p).

In summary, two types of c-Si wafer with three different passivation treatments will be used to prepare MIS devices for CV measurement. Fixed charge calculated based on CV data shows a reduction in Q_fix for H₂S samples compared to the SiN_x passivated sample. Additionally, qualitative analysis of the CV curves as well as fitting of lifetime data shows a reduction in D_it for the H₂S samples, indicating high-quality passivation of the c-Si surface is due to low D_it ≈ 4.0×10¹⁰cm^-2eV^-1 with a moderate positive Q_fix.

A detailed comparative study of interface passivation mechanisms by SiN_x, a-Si, and H₂S treatment will be provided in the final conference submission. Additionally, low frequency capacitance measurements, and conductance vs. AC frequency measurements will be included for all samples, which will allow for an accurate and quantitative determination of D_it for all samples. This in turn will yield a more robust lifetime fit to extract values for capture cross sections.

[1] A. G. Aberle, et al., Progress in Photovoltaics: Research and Applications, Vol. 2, pp. 265-273, 1994.
[2] U. K. Das, et al., Apl. Phys. Lett., Vol. 92, 063504, 2008.
[3] W. Liu, et al., ACS Appl. Mater. Interfaces, Vol. 7, pp. 26522-26529, 2015.
[4] H. S. Liu, et al., Langmuir, Vol. 33, pp. 14580-14585, 2017.
[5] P. A. G. O’Hare, Pure Appl. Chem.,Vol. 71, pp. 1243-1248, 1999.
[6] A. Haas, Angew. Chem. Int. Ed. Engl., Vol. 4, pp. 1014-1023, 1965.
[7] Z. Shu, et al., Prog. in Photovoltaics: Research and Applications, Vol. 23, pp.78-93, 2013.
[8] U. Das, et al., Proc. 7th WCPEC, 2018.
[9] D. K. Schroder, John Wiley & Sons, 2006.

Black silicon (Si) is a nanostructured type of Si capable of near-total absorption of light. Therefore, the use of black Si can lead to highly efficient solar cells. The performance of Si solar cells is dependent on the active dopant carrier distribution in emitters. The carrier distribution in black Si solar cells has been difficult to measure, and thus to optimize, due to the irregular, nano-scaled structure of the emitter.[1]-[5]
Thus, obtaining more information regarding active dopant density distribution in black Si solar cell requires measurements of carrier distributions on a nanoscopic scale.
One of the authors have previously succeeded in quantitatively analyzing such distributions in monocrystalline Si solar cells using scanning nonlinear dielectric microscopy (SNDM) [6]-[9].
In the present study, we investigated quantitatively the carrier distribution in a phosphorous (P) diffused black Si solar cell using SNDM. As a reference, we also measured the carrier distribution on a flat Si sample fabricated under the same P diffusion conditions. The precise carrier distributions in the emitter were visualized, which revealed the feature of carrier distribution in the emitter of black Si solar cell. Super-higher-order-SNDM was also employed to perform a quantitative analysis of the depletion layer distribution. It was found that the carrier density profile and the depletion layer thickness is less regular in the black Si than in the flat emitter, suggesting that this fluctuation may affect the power conversion efficiency of black Si solar cell.
This work demonstrates that SNDM is a very useful means of investigating carrier distributions in the nano-scaled emitter of black Si solar cell.

References:
[1] Y. Liu, T. Lai, H. Li, Y. Wang, Z Mei, H. Liang, Z. Li, F. Zhang, W. Wang, A.Y. Kuznetsov, X. Du, Small, 8, 1392 (2012).
[2]K. Peng, Y. Xu, Y. Wu, Y. Yan, S.-T. Lee, J. Zhu, Small, 1, 1062 (2005).
[3] K. Peng, X. Wang, S.T. Lee, Appl. Phys. Lett., 92, 163103 (2008).
[4] J Z Shen, B. Liu, Y. Xia, J. Liu, J. Liu, S. Zhong, C. Li, Scr. Mater., 68, 199 (2013).
[5] C.-H. Hsu, J.-R. Wu, Y.-T. Lu, D. J. Flood, A. R. Barron, L.-C. Chen, Materials Science in Semiconductor Processing, 25, 2 (2014).
[6] K. Hirose, K. Tanahashi, H. Takato, and Y. Cho, Appl. Phys. Lett. 111, 032101 (2017).
[7] Y. Cho, S. Jonai, and A. Masuda, Appl. Phys. Lett., 116, 182107 (2020).
[8] Y. Cho, A. Kirihara, and T. Saeki, Rev. Sci. Instrum. 67, 2297 (1996).
[9] Y. Cho, “Scanning Nonlinear Dielectric Microscopy”, Elsevier, ISBN 9780128172469 (2020).

The global power generation is in a transition toward renewable energy forms in a visible speed. Photovolotaic (PV) technology using crystalline silicon solar panels is playing a more important role, with more than 500GW global installation and relevant solutions becoming more mature. Continuous solar cell efficiency improvement is a dominating factor that accelerates PV industry in recent years, where the mainstream technology Passivated Emitter and Rear Contact (PERC) allows an average solar cell efficiency from 21% to 23%. However, further efficiency increase is limited by the significant carrier recombination loss at the metal-silicon interface regions. Hence, the passivating contacts have attracted great attention in both the research and the industry society, where polysilicon (poly-Si) tunnel junction technology using a thin silicon oxide tunneling layer and doped polysilicon carrier selective layer combination. It allows the contact interface region to achieve a significantly reduced recombination rate and increased open circuit voltage reaching 720mV. The laboratory world records for homo-junction cells, either front-rear contacted or back-contacted, utilized such passivating contact structure: the 25.8% efficiency n-type TOPCon cell from ISE Freiburg and the POLO-IBC cell from ISFH. Strong interests from the PV industrial players promote a rapid development of passivated contact technology in the manufacturing level. This work presents JinkoSolar’s latest attempt of achieving over 25% efficiency for passivated contact structure solar cells using production size n-type silicon wafers.

Solar cell performance
This work presents the overall performance of solar cells from several production batches, where the average efficiency is over 24.8%. Among the batch results, the champion cell achieve a third-party certified efficiency of 25.25%, with an open circuit voltage (Voc) of 715 mV and a fill factor of 84.9%. The distribution of batch cells’ electrical performance indicates a strong alignment between the cell efficiency and the fill factor, where the fill factor parameters vary in a relatively large range. It well suggests that the contact quality of the passivated contact solar cells using screen printing technique is crucial, which is determined by various factors such as the screen paste composition, contact region surface morphology, poly junction thickness and doping concentration, quality of the tunneling oxide layer, etc. This work presents some specific characterization resulting aiming to explain the contacting mechanism of a high efficiency poly-Si passivated contact solar cell. In addition, several loss analysis have been performed for the champion cell.

Module performance and the field performance
Manufacturing produced passivated contact solar cells, particularly those commonly referred as n-type TOPCon cell by the PV industry, are relatively new to the market, with only limited field performance record. Thus the module reliability research for passivated contact technology is a hot research topic, major evaluation methods include a serious of IEC standard tests such as temperature circle, dump heat test, potential induced degradation test, and so on. This work will present some testing and result details. In addition, this work will also present some field power generation results of passivated contact technology modules, including the analysis of performance improvement comparing to conventional PV modules.

Polycrystalline silicon (poly-Si) has proven to be a game-changing material in the field of high thermal budget carrier-selective passivating contacts (CSPCs) for c-Si solar cells beyond PERC architecture [1]-[4]. However, doped poly-Si exhibits a very high free carrier absorption (FCA), which has turned the attention of researchers towards wide bandgap materials, such as poly-SiO_x [5][6]. In these materials, the opto-electronic properties depend on oxygen [6][7] alloying. Such CSPCs consist of heavily doped poly-Si alloyed with oxygen and deposited on an ultra-thin SiO_x layer, prepared by thermal oxidation [8], wet-chemical process [6], UV/O₃ process [9], or low temperature plasma oxide [10]. Solar cells fabricated with such CSPCs on wet-chemical tunnelling SiO_x have exhibited illuminated efficiency of around 21% in front/back contacted (FBC) architecture [11]. In this work, building on our single-side poly-SiO_x processing, an ultra-thin SiO_x layer prepared by thermal oxidation was used as tunnelling oxide. The advantages of using this type of oxide are better uniformity on textured interfaces and better control over the oxidation rate by changing the oxygen flow rate, temperature, and time.
First, we optimized the passivation of n-type and p-type doped poly-SiO_x CSPCs symmetric samples. Ultra-thin thermal oxide of around 1.5 nm was grown at 675 °C for 6 minutes. The standard thickness of the poly-SiO_x layer is around 30 nm for both polarities. An in-situ, single side, doping approach using PH₃ or B₂H₆ gas flow was used to dope the poly-SiO_x CSPCs to n-type or p-type, respectively. Double side textured n-type and double side flat p-type doped poly-SiO_x symmetric samples give an iV_oc of 690 mV after annealing and iV_oc of 710mV after hydrogenation.
Single side textured FBC c-Si solar cells passivated with poly-SiO_x were fabricated on n-type FZ wafers and using the optimized n-type and p-type doped poly-SiO_x CSPCs. These were realized by using n-type poly-SiO_x layer on the sunny side and p-type poly-SiO_x layer on the rear flat surface. All solar cells were annealed at 950 °C for 10 minutes to crystallize the poly-SiO_x layers and drive in the dopants. In this high temperature process, hydrogen effuses from c-Si bulk/SiO_x interface. So, these cells were hydrogenated by forming gas annealing (FGA) at 400 °C for 1 hour after being preliminarily capped with 100-nm thick PECVD SiN_x layer. Upon the removal of SiN_x, indium tin oxide (ITO) layers were sputtered (75-nm and 150-nm thick at front and back side, respectively) to ensure efficient lateral carrier transport of charge carriers. The front ITO layer also acts as anti-reflective coating. Screen printing was used to deposit front and rear metallic contacts. The best solar cell gave an illuminated efficiency of 21.44% (V_oc = 697 mV, J_sc = 38.52 mA/cm², FF = 79.86%, metallization faction ~3%, illuminated area = 3.92 cm²). Moving from wet-chemical tunnelling SiO_x and evaporated metallic contacts [11] as well as further optimizing our doped poly-SiO_x layers, we could improve both the V_oc and the FF of our solar cells based on poly-SiO_x by an absolute 0.7%. Currently we are analysing the integration of this poly-SiO_x passivated c-Si solar cell as bottom cell in 2T perovskite/c-Si tandem configuration. The results will be presented at the conference.

[1] G. Yang, et al., Sol. Energy Mater. Sol. Cells, 2016.
[2] F. Haase, et al., Sol. Energy Mater. Sol. Cells, 2018.
[3] F. Feldmann, et al., Sol. Energy Mater. Sol. Cells, 2017.
[4] A. Richter, et al., Nat. Energy, 2021.
[5] J. Stückelberger, et al., Sol. Energy Mater. Sol. Cells, 2016.
[6] G. Yang, et al., Appl. Phys. Lett., 2018.
[7] M. Singh, et al., Sol. Energy Mater. Sol. Cells, 2020.
[8] F. Haase, et al. SiliconPV, 2018.
[9] R. van der Vossen, et al., 7^th SiliconPV, 2017.
[10] W. Lerch, et al., ECS Transactions, 2012.
[11] G. Yang, et al., Prog. Photovolt. Res. Appl., 2021 (under review).

All current state-of-the-art silicon photovoltaic (PV) devices employ some flavor of passivating contact structure to minimize detrimental recombination while providing good conductivity. These properties are most often provided by a dielectric layer (SiN_x, Al₂O₃, or SiO_x) that is in contact with the crystalline silicon (c-Si) substrate. The effectiveness of these layers is heavily dependent on the nanoscale structure of both the dielectric layers and the crystalline silicon interface. As an example, the tunneling probability for charge carriers across thin SiO_x layers, a process that is critical to maintain high conductivity, is extremely sensitive to the SiO_x thickness. SiO_x with a thickness below 2 nm readily allows charge carrier tunneling, whereas it is impeded for greater thicknesses. In the latter case it has been shown that modifications to the thermal processing schedule can induce disruptions, or pinholes, in thick SiO_x layers that allow transport of charge carriers while maintaining excellent passivation. However, this requires higher temperatures processes that may inhibit widespread adoption by industry. Recently, another option has been presented that may circumvent the necessity for high temperature processing to produce pinholes, namely metal assisted chemical etching (MACE) of the SiO_x layer. MACE relies on deposition of Ag nanoparticles onto the SiO_x layers with subsequent electroless etching to form mesopores in the SiO_x layers that are analogous to the pinholes created with high temperature processing. The size and density can be controlled based on the Ag nanoparticle deposition conditions. The resulting contact structure is known as polysilicon on locally etched oxide (PLEO) contacts. The nanoscale structure of the mesopores in PLEO contacts define the device level observables such as recombination current (J_o) and junction resistance. In this work we present atomic resolution transmission electron microscopy (TEM) analysis of mesopores in PLEO contacts formed with different processing conditions to provide insight into how the nanoscale structure of the SiO_x layer influences PV device properties. Additionally, we have previously used electron beam induced current (EBIC) to probe local transport properties in Si PV devices with a SiO_x layer containing pinholes formed with high temperature processing. Using EBIC we were able to directly show the enhance charge carrier transport through the pinholes. Here we also employ EBIC to study non-uniformities in charge carrier recombination and transport associated with mesopores in PLEO contacts and compare these results with our previous work on devices with pinholes in the SiO_x formed through high temperature processes. The products of this work provide critical information that is required to both further optimize performance of Si PV devices with PLEO contacts and to drive future adoption of this technology by industry.

Black silicon (b-Si) surface texture that consists of nanoscale structures is well known to reduce the optical losses both in multi- and monocrystalline silicon solar cells. The enhanced surface recombination related to the increased surface area of nanostructures is no longer an issue as conformal atomic layer deposited thin films provide excellent surface passivation. [1] Furthermore, pn-junction formation and related control of dopant atoms inside nanostructures has resulted recently in totally recombination-free emitters yielding external quantum efficiencies even above unity in front-junction Si solar cells. [2] These developments have made the b-Si technology attractive for the PV industry as well.

So far only one black silicon fabrication technology has emerged as cost competitive technology to conventional texturization, that is, metal assisted chemical etching (MACE). [3] However, there are other b-Si fabrication technologies that have their benefits over MACE (e.g. extended infrared absorption) and which could find interest among PV industry, too. These include dry-etching as well as laser-based methods. All these technologies have seen lately great progress, so in this contribution we give an updated overview of b-Si fabrication methods and related achievements. Furthermore, in order to make a direct comparison of the methods easier, we present a systematic study on the properties of b-Si wafers fabricated in the same facility on sister wafers using three different fabrication methods: i) MACE ii) reactive ion etching and iii) femtosecond-laser. A special emphasis is placed on optical properties, recombination and surface passivation quality, surface morphology, silicon consumption, throughput, and patternability. Finally, the pros and cons of each method are discussed and the future outlook is given for each technology.

[1] P. Repo et al. IEEE Journal of Photovoltaics 3 (90-94) 2013.
[2] K. Chen et al. APL Materials (submitted 2021).
[3] K. Chen et al. Solar Energy Materials and Solar Cells 191 (1-8) 2019.

In this talk I will introduce low-cost and scalable light management approaches for crystalline silicon (x-Si) photovoltaic (PV) cells utilizing, (i) lithography-free inverted pyramids, (ii) electroless plated metal nanoparticles (NPs), and (iii) 3D printed light trapping films, that my team has been developing recently.

Industrial-scale passivated emitter rear contact (PERC) solar cells typically utilize upright pyramids fabricated utilizing potassium hydroxide (KOH) based alkaline etching of c-Si. ISFH’s champion cell also utilizes the upright pyramid texture. While high efficiency lab-scale PERC solar cells utilize inverted pyramids because of higher antireflection (AR) relative to upright pyramids; these inverted pyramids are fabricated utilizing photolithography and an alkaline etch. Photolithography required to fabricate inverted pyramids is a slow and high cost process that is not compatible with low cost solar cell manufacturing needed for competitive solar energy generation. Hence, to-date, inverted pyramid texture has not been utilized in industrial-scale c-Si PERC cells. I will present results on a vacuum- and lithography-free wet chemical etching process that results in inverted pyramid textures on c-Si with extremely low spectrum weighted surface reflectivity (Rave) of 4.4% and 10% relative improvement in power conversion efficiency (PCE) as compared to c-Si solar cells with no surface texture.

Next I will present our work on a light management structure on c-Si utilizing electroless plating of silver nanoparticles (AgNPs). With the help of a MATLAB-based analytical model on Mie theory, the size distribution of AgNPs for desired optical properties was determined, and experimental results confirmed the reduction in reflection by up to 24.8% at a wavelength of 371 nm. An atmospheric degradation study of the AgNPs was also conducted, which demonstrated that the LSPR response of unprotected AgNPs is markedly impaired after 14 days, while the LSPR response of aluminum oxide (Al2O3) protected AgNPs is unchanged even after 90 days. AgNPs covered with an Al2O3 layer on a c-Si solar cell exhibit the highest absolute gain of 19.2% in external quantum efficiency (EQE) at 700 nm and an overall 20% relative increase in PCE compared to the reference c-Si solar cell without NPs.

Finally, I will discuss our recent work on 3D-printed solar concentrators using acrylonitrile butadiene styrene (ABS) printing filament and designed using a CAD software. Post designing and printing, the concentrators underwent a layer of varnish to both smoothen the 3D-printed surface and improve the overall strength. Varnished 3D-printed structures were then spray coated with a layer of metal. The metals chosen were aluminum (Al) and chromium (Cr), due to both having low cost and high reflectivity in the visible spectrum. We obtained around 4x light concentration with the 3D-printed concentrators while reducing the weight of the concentrators to that of milled metal concentrators by two times and reducing the production cost compared to binary-optics by up to 500 times.

Si has been widely used in nearly all aspects of semiconductor technology. However, Si in the bulk, diamond-structured form (d-Si) has an indirect bandgap and weak absorption. Exotic forms of crystalline silicon (silicon allotropes) that have been stabilized at ambient temperature and pressure include caged, open channel, and layered structures. One of the most well-known Si allotropes is Si clathrate, a cage-like, crystalline Si inclusion compound. Si clathrate has been synthesized to form either type I (Na₈Si₄₆) or type II Na_xSi₁₃₆ (0 < x < 1) phases in the presence of alkali atoms, such as Na, which sits inside the cage acting as an interstitial guest. If the Na content is low enough, which can only be achieved in type II (x << 1 in Na_xSi₁₃₆), Si clathrate can become a semiconductor with Na as a shallow, n-type, dopant. Theoretical studies have predicted type II Si clathrate to have a direct or nearly direct bandgap, near 2.0 eV, making it an exciting alterative to d-Si in optoelectronic applications. Most studies of Si clathrate have focused on powder synthesis. Compared with clathrate powder, thin-film Si clathrate is more interesting due to its potential for practical applications. Si clathrate films can be formed using a two-step process adapted from powder synthesis: formation of a Zintl precursor NaSi film and decomposition of the precursor into the clathrate phase. However, this thermal decomposition technique is very sensitive to growth parameters and the local environment, limiting controllable synthesis and complicating the characterization of film properties.
Here, we explore systematically the two-step synthesis technique in more detail by studying the key factors leading to thin, low-Na content, and high-quality films. Films have been synthesized with Na concentrations, varying from near the metal insulator transition (x ~10) values, comparable with the lowest Na content reported for powder samples. However, X-ray diffraction, Raman scattering spectroscopy, and Electron Paramagnetic Resonance measurements all provide evidence for a disordered or amorphous silicon-like phase that can dominate the surface and grain boundaries of the film. Time-of-Flight Secondary Ion Mass Spectrometry also demonstrates inhomogeneity of the ion profiles, showing regions of high Na content at the film surface and near boundaries surrounding low Na content regions, where high-quality clathrate material is located. We find that the presence of disordered Si and the inhomogeneity can significantly affect the film properties, complicating determination of the intrinsic properties of the clathrate.
Several growth and post growth techniques are studied to improve the film crystallinity and isolate the desired type II clathrate phase with low Na content. By exfoliating the film surface from the Si substrate, the high-quality clathrate structure is exposed and detected at the buried interface. Wet and dry acid etching are also applied, of which SF₆reactive ion etching (RIE) is shown to selectively remove the disordered phase, while also decreasing the Na concentration in the films. These techniques provide promising methods of isolating high-quality clathrate material for investigation and application.
Using these techniques, band edge characteristics were probed on high-quality films and surfaces using a range of measurements including absorption, photoluminescence (PL), and ellipsometry. A bandgap near 1.7 eV and an absorption coefficient about two orders of magnitude higher than diamond Si were observed. Room temperature PL is also detected in the near infrared region. An analysis of the temperature dependent PL emission confirms that Na in type II Si clathrate films can be a shallow donor with an activation energy of about 19 meV. This makes type II Si clathrate a n-type Si allotrope with exciting potential for Si based solar cells, LEDs, and sensors. This work was supported by National Science Foundation Awards 1810463.

Light trapping and the broadband absorption of the solar radiation are significant to a plethora of absorption-based photonic devices. Specifically, efficient broadband absorption was recently demonstrated with arrays of subwavelength structures. The current study examines both numerically and experimentally light trapping driven by deep sidewall subwavelength structures (DSSS) in silicon nanopillar (NP) arrays (DSSS arrays). Particularly, the focus is on DSSS geometries that are an inherent outcome of the top-down dry etch approach used in arrays of high aspect ratio NPs due to the periodical operationality of the Bosch dry etch method. ~10% enhancement in the broadband absorption of DSSS arrays compared with NP arrays is demonstrated numerically, as well as the generation of near-IR absorptivity peaks of ~25% for DSSS arrays. Importantly, it is shown that the introduction of DSSS systematically blue-shifts the absorptivity peaks of the NP arrays and in this manner a deterministic light trapping is possible. Finally, decrements of ~40% in direct reflectivity and ~7% in diffused reflectivity in DSSS arrays realized on silicon wafers is demonstrated experimentally.

Increasing the power conversion efficiency of silicon (Si) photovoltaics is a key enabler for continued reductions in the cost of solar electricity. A two-terminal perovskite/Si tandem design that increases the Si cell’s output in the simplest possible manner: by placing a perovskite cell directly on top of the Si bottom cell. The advantageous omission of a conventional interlayer eliminates both optical losses and processing steps and is enabled by the low contact resistivity attainable between n-type TiO₂ and Si, established here using atomic layer deposition. The fabricated proof-of-concept perovskite/Si tandems on both homojunction and passivating contact heterojunction Si cells to demonstrate the broad applicability of the interlayer-free concept. Stabilized efficiencies of 22.9 and >25% were obtained for the homojunction and passivating contact heterojunction tandems, respectively.
Based on this, a clear route is provided for minimizing optical losses aided by optical simulations which result in tandem current densities of ≈20 mAcm⁻² with front-side texture. In addition, electrical modeling on the Si-subcell has also been conducted in order to understand the efficiency potential of this cell under filtered light in a tandem configuration. The possibility of increasing the Si subcell efficiency by 1% absolute is offered through joint improvements to the bulk lifetime, which exceeds 4 ms, and improves surface passivation quality to saturation current densities below 10 fA cm⁻². A combination of optical modeling of the complete tandem structure alongside electrical modeling of the Si-subcell, both with state-of-the-art modeling tools, provides the first complete picture of the practical efficiency potential of over 30% for perovskite/Si tandems.

Monolithic perovskite/silicon tandem solar cells are of interest thanks to their high power conversion efficiency (PCE) potential at affordable cost. Also, the possibility to fabricate silicon heterojunction bottom cells either in the n-i-p or p-i-n polarity provides considerable flexibility to choose the device architecture. The initial perovskite/silicon tandems were in the n-i-p configuration. However, as light in this architecture enters from the p-side, applying the typical electron and hole selective contacts of single-junction devices to the tandem configuration resulted in devices with a poor blue response, mainly due to parasitic absorption in the front-contact stack. Therefore, global tandem research refocused on the inverted structure (p-i-n configuration, with electron collection at the front). So far, several efficient devices have been reported in this configuration. Increasingly the PCE of p-i-n tandems further is constrained by parasitic absorption and recombination-active defects in the front SnO_x/C₆₀ ESL stack, prompting us to revisit the n-i-p structure to benefit from the expansive library of efficient hole selective layers in single-junction n-i-p perovskite solar cells.

In this study, we overcome the challenges of the n-i-p perovskite/silicon tandem solar cell platform and unveil the true potential of this architecture. We do so by developing novel electron and hole selective contacts which combine high broadband transparency with efficient charge extraction. First, we needed to develop efficient electron selective layers exhibiting high conformality on macroscale textured silicon substrates: the texture is needed to minimize light reflection. We implemented this concept by developing room temperature sputtered amorphous niobium oxide (a-NbO_x) films, combined with a C₆₀ self-assembled molecular monolayer. On the sunward side, we needed to overcome the high parasitic absorption and fast degradation challenges of spiro-OMeTAD. We developed conformal, transparent, and efficient hole selective layers (molecularly doped evaporated spiro-TTB). We combined this contact with a low-temperature atomic layer deposited vanadium oxide (VO_x) layer, which prevented sputtering damage and did not introduce significant optical losses. To enhance the stability of this contact stack, we introduced ultrathin TPBI layers between the spiro-TTB and VO_x.
Combining these advancements with a two-dimensional perovskite passivation layer on the perovskite absorber, we achieved a power conversion efficiency exceeding 27%. This represents a significant advance over the best prior report in the n-i-p configuration of only 21.8% on similar bottom cells. The new device platform allows for translating the achievements of n-i-p single-junction solar cells into the context of tandems. The rich library of broadband transparent small-molecule HSLs and the possibility to utilize two-dimensional perovskite passivation strategies can now be easily implemented in tandems. From this, we expect rapid progress to be achieved in the field of perovskite/silicon tandems in the n-i-p configuration. We expect that our research will also propel the perovskite/perovskite tandem solar cells research which has been suffering from a lack of efficient vacuum-deposited hole selective contacts, where we offer a unique solution here.

Silicon heterojunction (SHJ) solar cells have exhibited efficiencies well above 25% [1]. To further boost the efficiencies of c-Si-based solar cells, high-bandgap perovskite cells are stacked on top, achieving a world record efficiency of 29.52% [2]. However, as most of the high-quality perovskite films are solution-processed [3], the front surface of the bottom device should be flat. Therefore, in this work, we optimized SHJ bottom c-Si cells featuring front-side-flat and read-side-textured morphology, which delivers high V_OC [4] together with excellent near-infrared response [5].
We firstly optimized plasma-enhanced chemical-vapor-deposition (PECVD) deposition conditions of (i)a-Si:H for both symmetrical flat <100> and textured <111> n-type c-Si surfaces. We observed that the optimized (i)a-Si:H monolayer (~ 10-nm-thick) on flat <100> surface using pure SiH₄ plasma results in limited performances (τ_eff = 1.2 ms, i-V_OC = 701 mV). While an optimized hydrogen dilution ratio (DR = f[H2]/f[SiH4]) of 3 was found for textured <111> surface (τ_eff = 4.3 ms, i-V_OC = 730 mV). This may be ascribed to an easier defective epitaxial growth for <100> as compared to <111> as reported previously [6]. To understand better the relations between passivation qualities and the microstructure properties (i)a-Si:H on flat <100> surface, we characterized layers via Fourier-transform infrared spectroscopy (FTIR) and spectroscopic ellipsometry (SE). We conclude that the monolayer that contains more H, a higher fraction of monohydrides and fewer defects (high refractive index) is beneficial for achieving a better passivation quality.
To improve further the passivation quality of monolayer (i)a-Si:H on flat <100> surface, we investigate approaches aiming at incorporating more H without promoting detrimental epitaxial growth. Firstly, we tested hydrogen plasma treatment (HPT) [7]. With HPT, we could only slightly improve the i-V_OC from 701 mV to 705 mV. While with a bilayer approach [8] that features firstly a less H-containing (i)a-Si:H to prevent epitaxial growth and then covered by a H-rich (i)a-Si:H, we could double the τ_eff to 2.4 ms with an i-V_OC of 720 mV. Eventually, by combining the bilayer approach with an intermediate HPT, the best sample showed a τ_eff up to 8.3 ms and an i-V_OC of 734 mV for a total thickness of 10-nm (i)a-Si:H on the flat <100> surface.
For the two-terminal tandem solar cells, bottom cells with (n)-contact on top are preferred due to the optical advantage of the perovskite top cells with the p-i-n configuration. Therefore, we symmetrically deposited (n)-type hydrogenated nanocrystalline silicon oxide (nc-SiO_x:H) on (i)a-Si:H coated flat <100> surface. Surprisingly, a drop of i-V_OC to 689 mV was observed after depositing (n)nc-SiO_x:H on (i)a-Si:H obtained via bilayer + HPT approach. We ascribe the deteriorated passivation qualities to possible excessive H incorporation during the (n)nc-SiO_x:H depositions. Therefore, we chose bilayer (i)a-Si:H, which has lower H-content as compared to the combined approach, and we obtained improved τ_eff (5.8 ms) and i-V_OC (724 mV) in combination with doped layers.
With this method, we successfully implemented various (n)-contacts with either (n)a-Si:H, (n)nc-Si:H, (n)nc-SiO_x:H or (n)nc-Si:H/(p)nc-SiO_x:H/(p)nc-Si:H into solar cells, which delivered V_OCs range from 703 to 710 mV and FFs range from 78.8% to 81.0%. Therefore, we have building blocks ready for tandem cell fabrications. Besides, advanced optical simulations aiming at improving the matched tandem current are also being conducted.
[1] X. Ru, et al., Sol. Energy Mater. Sol. Cells, 2020
[2] Oxford PV. oxfordpv.com (Accessed 21^st of June, 2021)
[3] J. Jeong, et al., Nature, 2021
[4] M. Taguchi, et al., Prog. Photovolt. Res. Appl., 2005
[5] ZC Holman, et al., IEEE J. Photovolt., 2013
[6] B. Demaurex, et al. J. Appl. Phys., 2014
[7] M. Mews, et al., Appl. Phys. Lett., 2013
[8] H. Sai, et al., Appl. Phys. Lett., 2018

PV-powered vehicle application are very attractive for reducing CO₂ emission and creation of new market. Development of high-efficiency (> 30%) and low-cost solar cell modules is necessary [1]. According to our analysis [2], the 2-junction and 3-junction tandem solar cells have potential efficiencies of more than 36% and 42%, respectively. The perovskite/Si tandem solar cells are thought to be one of the most promising PV devices because of high-efficiency and low-cost potential. However, efficiencies of perovkite/Si tandem solar cells with an efficiency of 29.15% are lower compared to 37.9% with III-V 3-junction tandem solar cells and 35.9% with III-V/Si 3-junction tandem solar cells. Therefore, it is necessary to clarify and reduce several losses of perovskite/Si tandem solar cells. This paper presents high efficiency potential of perovskite/Si tandem solar cells analyzed by using our analytical procedure [3] and discusses about non-radiative recombination, optical and resistance losses in those tandem solar cells.
One of problems to attain the higher efficiency perovskite/Si tandem solar cells is to reduce non-radiative recombination loss. The open-circuit voltage V_oc drop compared to bandgap energy (E_g/q - V_oc) is dependent upon non-radiative voltage loss (V_{oc, nrad}) that is expressed by external radiative efficiency (ERE). Open-circuit voltage is expressed by
V_oc = V_{oc, rad} + (kT/q)log(ERE), (1)
where the second term shows non-radiative voltage loss, is radiative open-circuit voltage and 0.28V was used as △V_{oc, rad} (= E_g/q - V_{oc, rad} ) for perovskite and III-V compounds and 0.26V was used for Si in this study. The practical limiting efficiencies of perovskite/Si tandem solar cells were calculated by assuming ERE = 30%, optical loss of 5% and resistance loss of 2%.
The perovskite/Si 2-junction tandem solar cells is shown to have efficiency potential of 37.4% as a result of non-radiative recombination loss of 2.3%, optical loss of 2.7% and resistance loss of 3.1%. The perovskite/Si 3-junction tandem solar cells are thought to be very attractive because the III-V and III-V/Si 3-junction tandem solar cells have potential efficiencies of 45.9% and 42.9%, respectively. Although they still have non-radiative recombination loss of 1.9% and 1.5%, optical loss of 2.9% and 2.3% and resistance loss of 3.2% and 2.7%, respectively.
Several non-radiative recombination losses such as bulk recombination, interface recombination, resistance loss including inter-connection and so forth in tandem solar cells are discussed in this paper. In addition, practical limiting efficiency (29.4%) and several losses of crystalline Si single-junction solar cells are discussed in order to compare with those of tandem solar cells in this study.

References
[1] M. Yamaguchi et al., Prog. Photovolt. (2020) https://doi.org/10.1002/pip.3343.
[2] M. Yamaguchi et al., J. Phys. D. Appl. Phys. 51, 133002 (2018).
[3] M. Yamaguchi et al., J. Mater. Res. 32, 3445 (2017).

Tandem solar cells with perovskite top cell and Si bottom cell have attracted a lot of attention for their potential to achieve energy conversion efficiency beyond the Shockley-Queisser limit. In the n-i-p perovskite/c-Si variant of this tandem solar cell, recombination at the interface of c-Si and the electron transporting layer of perovskite top cell greatly affect the performance of the device. In this study, we propose the utilization of Scanning Zone Annealing (SZA), a selective heat treatment with high degree of control, to control carrier recombination at Si/TiO₂ interface for perovskite/Si tandem solar cell application. Here, we found that recombination velocity at Si/TiO₂ interface can be reduced down to 40 cm/s using SZA. We also found that the amorphous silica interlayer affected the recombination dynamics at Si/TiO₂ greatly. To understand this phenomenon, theoretical approaches to model triple layer interfaces of Si/SiO_x/TiO₂ were utilized. A combination of molecular dynamics and density functional theory calculations was employed to study the effect of amorphous silica stoichiometry and layer thickness on the electronic properties of the interfaces including band alignment, charge transport, and carrier lifetime.

In the past few years, perovskite/silicon (Si) tandem solar cells have emerged as one of the most promising candidates for next generation photovoltaics with the potential to beat the theoretical efficiency limit of single junction solar cells (∼33%). Lately, a record power conversion efficiency (PCE) > 29 % has been demonstrated for two-terminal (2T) perovskite/Si tandem solar cells, which already surpasses the record PCE of single junction Si solar cells.

Yet the vast majority of high efficiency 2T perovskite/Si tandem solar cells reported to date, have been demonstrated using front-side polished Si wafers. While this approach is compatible with laboratory scale processes, it is not practical as industrial standard for Si photovoltaics due to the high costs. Industrial textured Si wafers are essential to achieve efficient light incoupling (i.e., to reduce the reflection losses) and light trapping near the band gap and typically exhibit pyramids that partly exceed 3 µm in size. However, such large textures are not compatible with the fabrication of conventional solution-processed perovskite thin films (thickness ∼500 nm) on top.

In response to this challenge, we report on industrial-like textured Si wafers with a reduced pyramid size of ∼1-2 µm that demonstrates (1) very low reflectance, similar to the conventional industrial textured Si wafers and (2) compatibility with thick solution-processed perovskite solar cells (PSCs) on top. To investigate the light management in PSCs fabricated over textured surfaces in depth, we replicate the Si texture by nanoimprint lithography on glass substrates and fabricate highly efficient p-i-n PSCs over the replicated texture.

A thick PSC absorber layer is necessary to cover all pyramid tips and prevent shunting paths. Therefore, we first developed highly-efficient p-i-n PSCs on a planar surface (PCE > 19 %) with a 1 µm thick double cation perovskite absorber layer (Cs_0.17FA_0.83PbI_2.75Br_0.25). Efficient charge carrier extraction is achieved by using urea additive to enhance the charge carrier diffusion length. Then, we optimized the spin-coating parameters for perovskite thin films processed over the replicated textures such that all pyramids are fully covered. In comparison to the planar PSCs, the PSCs processed over the textured substrates exhibit a voltage loss of ∼50 mV. To explain the underlying reasons of the voltage loss, we investigated the perovskite crystallization dynamics over the industrial-like textured substrates by performing a full characterization scheme using XRD, EDX, EBIC, cross sectional SEM as well as the standard electrical and optical characterization means. Thereby, we present a guideline on how to minimize these voltage losses [1].

Finally, to demonstrate the optical gains that can be obtained by processing PSCs over textured surfaces, we study in detail the optical properties of the optimized thick PSCs processed over the replicated texture as well as the planar surface. While the textured stack exhibits a considerable broadband antireflection behavior, the planar stack suffers from higher reflection losses due to the thin film interference pattern (2.5 times higher reflection than the textured stack at around 570 nm). The textured stack also exhibits very efficient light incoupling even up to 70^ο angle of incidence as compared to the planar reference. The efficient light harvesting in our devices results in a remarkable improvement in the broadband spectral response, which leads to efficient J_sc amplification with one of the highest reported J_sc /J_SQ ratios [2,3].

In summary, our contribution shows that, by means of optimized pyramidal surface textures and adapted thick solution-processed PSC, we can solve a key hurdle in the fabrication of textured perovskite/Si tandem photovoltaics.


[1] A. Farag, et al. (in preparation)
[2] A. Farag, et al. (submitted)
[3] F. Gota, et al. (submitted)

Till date, the photovoltaic industry is dominated by crystalline silicon (c-Si) solar cells because of their reliability, rapid cost reduction and improved efficiency. However, the thermalization losses occurring in the device limits its performance. Most common approach to surpass this limit is to use solar cells in tandem configuration with the top sub-cell having wide band gap absorber and bottom sub-cell possessing silicon as the absorber layer. Perovskite solar cells have proven to be a promising candidate to be used as a top sub-cell in tandem configuration because of their band gap tuning. Monolithic integration of silicon/perovskite devices reduces the wiring complexity and the additional fabrication steps. Most of the monolithic silicon/perovskite solar cells have been obtained either by homojunction or heterojunction silicon bottom cells. However, a tremendous amount of work shows an improved efficiency of transparent metal oxide (TMO) based carrier selective contact (CSC) silicon solar cells. The doped a-Si:H layers in the conventional SHJ solar cells are replaced by TMOs such as V₂O₅, MoO_x, TiO₂, and NiO_x . MoO_x has been proven to be an efficient substitute for p-doped aSi:H layer. The high work function (about 6.7eV) in MoO_x and higher band gap (about 3.2 eV) increases the fraction of light reaching the CSC device. The larger work function also causes an inversion in the c-Si thereby facilitating the transfer of holes from c-Si to the front end electrode. The CSC solar cell incorporating MoO_x as a hole transport layer has obtained an experimental efficiency of 22.5%. The rear side p-doped a-Si:H is also being substituted by TiO_x which introduces higher conductivity for electrons and lower conductivity for holes in CSC configuration. The highest efficiency reported for TiO_x based CSC devices is 21.6%. The monolithic tandem configuration requires optical coupling and minimized electrical losses which is obtained by adding a tunnel layer in between the top and bottom solar cells.
In this work, the performance parameters of tunnel junction (TJ) formed by TMO layers are investigated on the top of a carrier selective contact (CSC) based silicon solar cell in Sentaurus. We have employed a tunnel junction by adding TiO_x on the top of a hole transport layer, MoO_x from the silicon bottom sub-cell. The band alignment between TiO_x and MoO_x would allow transport of charge carriers by tunneling thus substituting the conventional tunnel layers. The simulations give a profound picture of the transport mechanism occurring in the device. The variations in electron affinity values and defect density of tunneling layers shows their effect on the hole collection ability of MoO_x.

High quality stable surface passivation is an essential requirement for high efficiency silicon solar cells, and there are many ways of achieving this (e.g. Al₂O₃, SiN_x, their stacks and more recently TOPCON). Good surface passivation is also required for the accurate diagnosis of defect-related degradation phenomena in cells and wafers, although the stability requirement is less critical in this context. Established passivation strategies for solar cells and test structures often complicate material and degradation studies, as they can modify the bulk properties of the sample under investigation. This is because they include elevated temperature (>300 °C) steps during which recombination-active defects can form, be annihilated or even be gettered. Changes in the hydrogenation state of the material also need to be considered.

Over the past five years, we have developed a range of temporary surface passivation schemes for measuring very long carrier lifetimes in silicon. Thin films are formed on wafer surfaces by treatment with bis(trifluoromethanesulfonyl)-based solutions, including superacidic and non-acidic chemicals, at room temperature. The films provide excellent surface passivation (surface recombination velocity < 1 cm/s), without modifying the bulk lifetime under investigation. Effective lifetimes in excess of 75 ms have been measured, and bulk lifetimes exceeding the currently accepted Auger limit have been extracted in some cases. As part of the talk’s introduction, the latest developments in our passivation methods will be presented, building on studies we have published previously [1-4].

The main part of the talk will include new insights gained from using these temporary passivation schemes on the latest silicon materials and photovoltaic cells. One key feature of the superacid-derived approach is that it can be applied after samples have been stripped of dielectric passivation, diffusions or even metallisation. We can therefore use it to measure the true impact of a process on the material’s lifetime without the modification that occurs in other passivation processes. New results to be presented will show how the activation annealing widely used for Al₂O₃ surface passivation can modify bulk lifetime, as well as chemical and field effect passivation. Our data show that the optimal activation temperature of around 460 °C occurs because recombination centres form in the bulk at higher processing temperatures. Our approach is also valuable in the study of light and elevated temperature induced degradation (LeTID), and data for some of the latest commercial Ga doped passivated emitter and rear cell (PERC) devices will be presented.

The final part of the talk will focus on the stability of superacid-derived passivation. Our standard approach is stable on the timescale of hours, which is sufficient for diagnostics but not for devices. Our standard approach involves an HF treatment to remove any native oxide layer on the sample surface, as native oxides provide very little surface passivation. Importantly, we have found that treating a thin (<1 nm) native oxide layer on the sample’s surface with a superacid-based solution at room temperature can however give rise to good levels of surface passivation (surface recombination velocity < 20 cm/s), which can last for several months. Using X-ray photoelectron spectroscopy (XPS) we show that fragments of the superacid molecule (particularly fluorine) become incorporated into the treated native oxide layer. Room temperature chemically enhanced thin oxides could therefore be a promising concept for silicon-based solar cell devices.

1. N.E. Grant et al., IEEE Journal of Photovoltaics 7, 1574-1583 (2017).
2. A.I. Pointon et al., Solar Energy Materials & Solar Cells 183, 164-172 (2018).
3. A.I. Pointon et al., ACS Applied Electronic Materials 1, 1322-1329 (2019).
4. A.I. Pointon et al., Applied Physics Letters 116, 121601 (2020).

Reduction of the surface and interface recombination losses on crystalline silicon (c-Si) solar cells is at present a crucial need for the fabrication of high efficiency devices. At an industrial level, these losses account for at least 0.5%abs efficiency decrease on passivated emitter and rear cells (PERC). In recent years many efforts have been made to reduce these recombination currents for which the concept of a passivated contact cell has aroused and gained a lot of attention. This more advanced concept of a cell, which is often referred to as carrier-selective contact cell, is meant to supply a carrier collection channel that will be able to reduce metal-semiconductor interface recombination while allowing free flow of selected charge carriers. Several approaches for the formation of selective contacts in silicon solar cells have been proposed. Amongst them there is: 1) the application of heavily doped regions under the contacts which reduce minority carrier conductivity as the result of a chemical potential introduction and low minority carrier mobility, 2) reduction of the minority carrier concentration at the surface by the application of external potential sources from materials with large fixed charges (Qf) or a low (/high) work-functions, or 3) the formation of a heterojunction with a wide-bandgap material with the desired conductivity type (n or p). At present, the highest achieved power conversion efficiencies (PCE) on silicon devices have been achieved by the application of amorphous silicon (a-Si) passivated contact and heterojunction technology. This technology benefits from carrier selectivity provided by the combined cases 2 and 3 given by the a-Si properties, and has led to the remarkable PCE = 26.3%, with the potential to reach >27% efficiencies.
A graphite oxide monolayer, also known as graphene oxide (GO), is a carbon-based material
whose properties have been investigated since its discovery and first-time fabrication over a
hundred years ago. This material exhibits some remarkable properties such as high transmittance and fixed surface negative charge which makes it very appealing for optoelectronic applications. Further to this, graphene oxide optical absorption can be tuned by the degree of oxidation of the material. Thorough studies of this process have been shown experimentally, resulting in optical bandgap tuneability from 1.15 eV and up to 3.06 eV. Similar to the a-Si case, these two properties of graphene oxide are likely to provide the carrier selectivity required for a high-quality passivating contact on PV devices, with the advantage of very low temperature and low-cost fabrication.
It has already been shown that graphene oxide films on silicon can act as passivation layers achieving surface recombination velocities as low as 17.4 cm/s. In this work, we study the application of multilayered graphene oxide films as passivating contacts on passivated emitter and rear locally diffused cells (PERL). It is found that graphene oxide interlayers aid carrier collection and reduce rear recombination leading to an increase of open-circuit voltages in the excess of 5mV and short-circuit currents greater than 3.5mA. A discussion of the cell performance and recombination parameters will be discussed during the presentation.

Continuous effort is being put towards device optimization for high efficiency solar energy conversion systems. As these devices are greatly affected by the various loss mechanisms and charge transport properties of the structures and materials, in-depth mapping of loss and charge transport mechanisms can be key for rapid device optimization.
By combining wavelength dependent photoluminescence quantum yield (PLQY) measurements with optical modeling, we extract the depth dependent spatial external luminescence efficiency (SELE), which is defined as the probability of an electron-hole pair that is photo-generated at a certain depth to contribute to photoluminescence from the device. Since the external luminescence efficiency is related to the obtainable photovoltage from the device, extraction of the SELE profile could be used to map the contribution of different regions in the device to the photovoltage. In this contribution, we will discuss the SELE extraction method and analyze the SELE profiles of Silicon and GaAs wafers, serving as model systems.
In addition, the use of finite elements simulations of the measured structures while accounting for charge transport and photon recycling effects, allows computing of the SELE profile and attributing features in the profile to important physical quantities such as the surface recombination velocity, internal luminescence yield and the optical coupling between the device and its ambient. Analysis of SELE profiles for advanced solar cells could assist in mapping loss mechanisms and further promote high efficiency solar energy conversion devices optimization.

Aluminum nitride (AlN) is a versatile material for chemical and field-effect passivation of silicon since fixed charges can be positive or negative, enabling to passivate n-type and p-type c-Si [1]. In this study, hydrogenated AlN layers (AlN:H) were deposited onto p-type c-Si by sputtering at hydrogen fluxes of 1 sccm, 3 sccm and 5 sccm and negative charges were activated on AlN through rapid thermal annealing (RTA) at 850^oC. AlN also can work as a converter layer as it is suitable as a host for rare earth ions such as Terbium Ions for spectral downshift in solar cells [2].
Transient surface photovoltage (SPV) spectroscopy allows to investigate transport, generation and recombination processes as well as their relaxation times. In addition, due to its non-contact characteristic of transient SPV spectroscopy, it is possible to perform measurements where other techniques are not capable.
SPV signals associated with the transfer of electrons and holes at the AlNxOy: H / c-Si (p) interfaces were separated from those related to reducing the surface band bending. The distribution of negative fixed charges was high so that an accumulation regime near the silicon surface was established. At high hydrogen flux, the band bending was low due to an increase in the interface defects that compensate for the negative fixed charge. Complementary measurements with modulated SPV give additional information on defect and tail energies [3].
By measuring SPV spectra at different times, one can obtain information about the charge transfer [4], the tail energies and offset deep defects [3]. Comparing the results obtained from the tail energies and offsets, we found that charge transfer to defects was dominated by a disordered layer at the AlNxOy:H / c-Si interface and by deep defect states for the samples deposited at 1 and 5 sccm of hydrogen flow, respectively.
The three-variable measurements (time, photon energy/wavelength, SPV) allows for a deeper study of dominating charge separation mechanisms.
References
[1] G. Krugel et al., Energy Procedia 55, 792-804 (2014).
[2] K. Tucto et al., MRS Advances, vol. 2, 52, 2989–2995 (2017)
[3] J. Dulanto et al., J. Phys.: Conf. Ser. 1841 012003 (2021)
[4] S. Fengler et al., ACS Appl. Mater. Interfaces 12, 3140-3149 (2020)

A levelized cost of electricity (LCOE) of 3¢/w, for U.S. locations with average solar resource, has been identified as a critical threshold for wide-spread adoption of photovoltaics without subsidies. Achieving this cost goal requires improvements in all facets of the technology, from cell efficiency increases, to component cost decreases, to lower soft costs associated with installation, to reduced system performance degradation, and so forth. Maintaining long-term high performance in modules is often an under-recognized opportunity in this cost reduction because it is not fully quantified at the time of manufacture. While certain standardized tests minimize risk of early failure, information on long-term performance for new modules is largely unknown and thus does not appear in datasheets or quotes.

In this work, we explore several aspects of how light and elevated temperature degradation (LeTID) impacts long-term module performance and thus LCOE. First, we document the behavior of two commercial module types fdeployed at NREL starting in 2016 and 2018. ~7 kW of each module type was installed on-site at NREL, and both exhibited degradation hypothesized to be LeTID. In cooperation with the international testing and standards community, an accelerated test for LeTID was developed. This accelerated test was performed on unfielded spare modules from the degrading NREL arrays, and the presence of LeTID was confirmed. An approach for predicting long-term performance loss due to LeTID, utilizing the accelerated test results and observed defect behavior, is presented. The predictions from this approach are compared with yearly flash test data on fielded modules from the two degrading 7 kW arrays at NREL and other published results. Cost models based on the predictions indicate that LeTID has a significant and climate-dependent impact on LCOE. Areas where predictions could benefit from a deeper understanding of LeTID are identified.

Although there are significant uncertainties in the predictions, the general process demonstrated in this work is the same as that needed to develop reliability predictions more widely. The process starts with observation of degradation in the field or in accelerated tests. Next a physical model – a mathematical relationship between the stresses and the degradation - is hypothesized, and accelerated tests to probe the kinetics are performed. Long-term predictions are compared with fielded behavior and extended experiments with modest acceleration. As more information about the physical mechanism and how to measure it become available, the model may be refined. This iterative approach can supply some information for stakeholders to estimate costs associated with new products, while successively decreasing uncertainty of predictions and addressing apparent disagreement between data sets.

The Department of Energy’s Solar Energy Technologies Office (SETO) supports applied research related to photovoltaics, concentrating solar power, and systems integration with the goal of making solar energy more affordable and facilitating solar energy deployment. Over the past decade, commercial silicon cell developments have targeted efficiency losses under standard test conditions due to surface and bulk defects. The silicon photovoltaic cell industry continues to move towards new architectures to reduce losses from defects and increase efficiency. In the future, addressing performance challenges at the module-level, controlling field performance through better understanding of degradation pathways, such as potential and carrier induced degradation, and improving sustainability will become increasingly important. This talk will highlight the research funded by the SETO in silicon-based photovoltaics, how research areas have evolved as industrial capabilities and technologies change, and a few of the materials-level research problems in silicon photovoltaics that will promote future solar deployment.

Two of the major trends in nowadays energy sector are electrification and digitalization. Large-scale introduction of variable renewable energy sources, energy storage and power-electronics components, all based on direct current (DC), is fundamentally changing the electrical energy system of today, based on alternating current (AC). This trend leads to a complex hybrid AC/DC power system with the extensive deployment of information and communication technologies (ICT) to keep the system stable and reliable. Furthermore, the expanding urban areas are increasingly so energy intensive that the demand for local and optimal generation of clean electricity is growing as well. In this scenario, photovoltaics (PV) technology will play an essential role. However, to take full advantage of PV in the urban environment, PV technology must become intelligent and PV devices must become multi-functional.

Recently, we identified, described, and labelled a new research field that deals with intelligent PV and its application in components with multiple functionalities. We denoted this field photovoltatronics [1]. Photovoltatronics aims at facilitating the large-scale implementation of photovoltaic technology in urban areas by combining it with electronic and photonic devices as well as digital technologies. As a result, photovoltatronics will deliver intelligent devices that we refer to as multi-functional PV-based intelligent energy agents (PV-IEA), by bringing together disciplines of energy and informatics. Since photons and electrons are carriers of both energy and information, photovoltatronics is the field that designs and delivers autonomous devices for electricity generation and information communication. It introduces a pathway from harvesting energy of photons (hν) to creating bits of information (01) through the energy of photo-generated electrons (eV). Photovoltatronics will deliver solutions for electricity generation and information communication in applications such as building environment, e-mobility, sensing, and domotics. This aim is realized by combining different research areas. Five major research areas that lay the foundations for the photovoltatronics research field are: (1) modelling and multi-layer mapping for maximizing energy harvesting from ambient energy sources at a particular location; (2) design of PV-IEAs for delivering optimum electrical energy including reconfigurable electrical topologies of PV generators, (passive) cooling elements, electronics, sensor systems, and control; (3) security and flexibility of electrical energy supply by integrated storage (4) wireless transmission of electricity through a novel design of PV module electrode, and (5) light-based communication (e.g. LiFi) by integrating and controlling embedded light-generating elements such as light-emitting diodes (LEDs) and by leveraging the inherent ability of PV devices to work as receiver in light-based communication systems while still being able to generate energy from the received natural or artificial light.

Soon, these devices can become building blocks for energy and information flows in cities, as every surface in urban areas can play an important role in providing useful energy from solar radiation to urban dwellers. Multi-functional PV-IEAs will thus represent the backbone of the combined energy/information system of the future.

We review photovoltatronics research areas and introduce new directions for each area. In this process, we show that ~10 keV energy is at least needed for transceiving one bit of information in the energy-information chain of the photovoltatronics, while the ultimate efficiency of the chain can reach up to 33.4%. We show that the number of publications related to photovoltatronics is exponentially increasing and the publication rate of combined research areas has been doubled in the present decade and reached 3.4% as a clear sign of its emergence.

[1] H. Ziar, et al., Energy Environ. Sci., 2021.

Over the past several decades, the semiconductor materials landscape has expanded from silicon-group elements to binary compounds, ternary alloys, perovskites, and 2D materials. However, each one of these materials has been created under the constraints of Earth’s gravitational field. What if this force could be negated? Microgravity offers a unique environment to achieve materials innovation in ways that cannot be accomplished on Earth. Changes in fluid behavior such as the lack of buoyancy-driven convection and sedimentation in microgravity lead to beneficial characteristics in materials like semiconductor crystals, 2D materials, and photonic materials. Some demonstrated benefits of the microgravity environment include the ability to isolate and study non-gravitational phenomena, preparing defect-free crystals, diffusion dominated processing, and containerless processing of materials.

While early focus of experiments in space was on fundamental research to enable and improve space exploration, the advent of the new space age has made it possible to bring the benefits of space closer to home. The emergence of commercial space stations like Axiom Station opens new opportunities for academic and industry researchers to perform focused applied research that leads to manufacturing of advanced materials in space for electronics and energy applications. This talk will highlight: (1) the fundamental ways in which microgravity affects materials processes; (2) the use of microgravity environment an innovation platform; and (3) potential future in-space manufacturing applications.

The main objective of crystalline silicon photovoltaic (PV) technology development is to increase the power conversion efficiency and further reduce the production costs, aiming to reduce the levelized cost of electricity (LCOE). As one of the technologies with passivating contacts, silicon heterojunction (SHJ) solar cell technology is considered to expand its share in the PV industry in the coming years due to the high-power conversion efficiency, lean fabrication process and low temperature coefficient.
In this work, we show how to improve the rear side of silicon heterojunction (SHJ) solar cells. The improved rear side increases the cell certified efficiency of 23.55% to 24.51% on a full-size n-type wafer (total area, 244.53 cm²). The route towards the higher efficiency SHJ solar cell is proposed through a combination of experimental work accompanied by device simulation. In this study the gain in in efficiency was achieved primarily by (i) adjusting the properties of the hydrogenated intrinsic amorphous silicon (a-Si:H) i-layer on the rear side of the solar cell and (ii) improving the optical properties of the back reflector of the solar cell.
All SHJ solar cells were fabricated using M2-size n-type Czochralski (CZ) crystalline silicon as-cut wafers from LONGi company textured in alkaline solution and cleaned via an ozone cleaning process. The plasma enhanced chemical vapor deposition (PECVD) process was performed in an AK 1000 system from Meyer Burger for intrinsic and doped a-Si:H layers. Afterwards, indium tin oxide (ITO) layers were sputtered from an 3% Sn-doped In₂O₃ target. Silver grids with a busbar-less design were screen printed and subsequently cured at 170 ^°C for 40 min. The MgF₂ and Ag layers were evaporated. For material characterization we used an ellipsometer (J. A. Woollam M-2000) and the structural characterization was performed by Fourier Transform Infrared Spectroscopy (FTIR) in a Nicolet 5700 system. To evaluate the solar cells performance, current-voltage characteristics, the external quantum efficiency, and reflectance were measured by the integrated solar cell characterization system called LOANA from pv-tools with a Wavelabs Sinus 220 light source.
The properties of the a-Si:H i-layers play multiple roles in SHJ solar cells not only as a surface passivation layer but also as a carrier transport to the electrode. To improve both, intrinsic a-Si:H bi-layers with a porous first layer (i₁) and a dense second layer (i₂) are a viable route. The silane (SiH₄)/hydrogen (H₂) flow ratio and thickness of the second intrinsic a-Si:H (i₂), were specially optimized here to reduce the vertical charge carrier transport loss but still achieving an excellent passivation quality. By using a denser and thinner second part of the intrinsic a-Si:H layer, a reduced vertical rear resistance loss was achieved. This favored vertical carrier transport and led to a higher FF.
To improve the optical properties of the back reflector we reduced the plasmonic absorption at the back reflector by inserting a low-refractive-index magnesium fluoride (MgF₂) before the Ag layer resulting in an increased short circuit current density (J_sc). In total, together with MgF₂ double antireflection coating and other small optimizations during cell fabrication process, ~1% absolute efficiency enhancement is finally obtained. Finally, we were able to improve the efficiency from certified 23.55% mainly due to an improvement in FF and J_sc improvement to 24.51%, with an J_sc of 39.52 mA/cm², an open circuit voltage (V_oc) of 741.8 mV and a FF of 83.61%.

Although carrier-selective contacts are crucial for the operation of any photovoltaic energy converter, it was only the increasing interest in non-traditional contact schemes in c-Si PV that has sparked a debate of how a metric for the ability of a contact to extract desired charge carriers versus blocking undesired ones should be defined and measured.
In the last years, several authors defined metrics for the quantification of the selectivity of contacts to solar cells. They can be classified in two categories: the first one compares the internal versus external voltage of the solar cell at open-circuit operation and the second one compares the transport coefficients for electron and hole transport through the contact. The different approaches are not identical in terms of defining the spatial boundaries of the selective contact and the processes involved in the charge carrier extraction at the contact. In the following, we present the carrier-selective contact from a thermodynamic and kinetic point of view, and compare different figures of merit for carrier-selective contacts. We present experimental investigations to access the different figures of merit by means of a three-terminal photovoltaic device.
First, we briefly summarize the thermodynamic interpretation of a solar cell in terms of a two-step conversion process with two respective conversion efficiencies: After converting sun’s heat radiation into an excited electron-hole gas in the absorber, the contacts convert the chemical energy of the electron-hole gas into electrical energy. The (carrier) extraction efficiency corresponds to the efficiency of the latter and measures the contact’s ability to extract charge carriers selectively. For the kinetic considerations, we discuss the spatial boundaries of the absorber, the carrier-selective membranes and the electrodes of a photovoltaic energy converter and pay special attention to the charge carrier processes occurring in the selective membrane and the electrode. We review the existing definitions of the contact selectivity metric and its measurement from the literature.
In order to examine the selectivity and the extraction efficiency, we use a three-terminal photovoltaic device with almost perfectly electron-selective and hole-selective reference contacts and a TiO_x-based electron-selective contact as model system. The perfectly selective reference contacts allow to measure the electron and hole current densities at the TiO_x-based contact separately. Therefore, it enables us to study the two different categories of selectivity metrics in the literature on a single test device by using two different methods.
For the first method – the three-terminal Suns-V_OC method – our device operates without an external current flow and provides access to the internal and external open-circuit voltage of the device as often used by several author in c-Si, organic and perovskite PV. From a TiO_x-based-contact-limited test device, we determine the extraction efficiency and selectivity coefficient of the TiO_x-based contact as function of the illumination intensity.
For the second method, we use the transport coefficients of electron and hole transport through the TiO_x-based contact to quantify the selectivity coefficient. For this purpose, we estimate the transport coefficient for holes – the recombination pre-factor J_0,C – from J_SC-V_OC characteristic between the electron-selective TiO_x-based contact and the perfectly hole-selective reference contacts. From dark J-V characteristic between the electron-selective TiO_x-based contact and the perfectly electron-selective reference contact of the three-terminal device, we determine the transport coefficient for electrons in the TiO_x-based contact – the contact resistivity. Finally, we compare the results of both methods.

Transparent Conducting Oxides (TCOs) are widely accepted transparent electrodes in a variety of solar cells employing passivating and selective contacts, such as silicon heterojunction (SHJ) solar cells and thin film technologies such as CIGS and perovskite solar cells¹. Among all possible ways to deposit TCOs, sputter deposition is the most employed technology, both in academia and industry. However, the high kinetic energies of the arriving species during sputtering may damage sensitive functional layers beneath, affecting interface formation, passivation and ultimately device performance. This has motivated the search for alternative sputter system designs or deposition techniques, allowing low damage growth of TCOs on the sensitive layers. Here we explore pulsed laser deposition (PLD) as an alternative low-damage physical vapor deposition technique. PLD is operated at high deposition pressures, promoting thermalization of particles, and therefore reducing the kinetic energy of ablated species. Broadband transparent and high mobility TCOs were developed using wafer-scale (4-inch) PLD with deposition rates on par with lab-scale RF sputtering². To investigate the effects of the deposition pressure, we developed ITO with identical sheet resistance (50 Ohm/sq) at deposition pressures three orders of magnitude apart, and applied them as a front TCO in SHJ cells and as rear-contact in semitransparent perovskite cells. The role of the deposition pressure on TCO properties, absorber carrier lifetime and its link to SHJ and perovskite solar cell parameters such as series resistance and fill factor (FF) will be discussed. Strategies to mitigate damage, for each type of cell will be furthermore discussed.

References
¹M.Morales-Masis et al. Adv. Electron. Mater. 3 (5), 1600529, 2017.
https://doi.org/10.1002/aelm.201600529
²Y. Smirnov et al. Adv. Mater. Technol. 6 (2), 2000856, 2021. https://doi.org/10.1002/admt.202000856

Since the first inception of silicon heterojunction (SHJ) solar cells, prominent developments have been realized in both front/back-contacted (FBC) [1] and interdigitated back-contacted (IBC) [2] (world record efficiency of 26.7%) architectures. However, the conventional silicon-based doped layers, which work as carrier-selective transport layers, are not optimally transparent especially when used as window layers.

To circumvent this fundamental limitation, several transition metal oxide (TMO) layers have been investigated because of their good transparency in the wavelength range of interest and appealing electronic properties to efficiently extract charge carriers from the device [3]. MoO_x as the most promising for hole-selective transport layer (HTL) has been applied in SHJ solar cells to replace the conventional doped layer in SHJ FBC devices. Normally, the thickness of HTL is between 6 and10 nm, while a power conversion efficiency (PCE) of 23.5% SHJ solar cell with 4-nm thick MoO_x was recently demonstrated by J. Dreon et al. [4]. The thinner the layer is, the less the optical loss is. However, still 0.5 mA/cm² was found to be lost in MoO_x [5]. According to the work by L. Mazzarella et al. [6], who applied a plasma treatment (PT) on the interface of i-a-Si:H and MoO_x to improve the electric performance, thinner MoO_x layer could be possible for application in solar cells. This is the focus of our contribution.

We used 4-in wide double-side textured n-type FZ wafers as substrate. A layer of i-a-Si:H was deposited on both sides and n-a-Si:H film was deposited on the rear side by plasma-enhanced chemical vapor deposition (PECVD). A PT was applied on the front surface before the deposition of MoO_x. A thickness series of MoO_x HTL was thermally evaporated at the pressure of 5×10^-6 mbar. Tested thickness of MoO_x were 0 nm, 1 nm, 1.7 nm, 2 nm, 2.9 nm, and 4 nm. After that, indium tin oxide (ITO) was sputtered though a mask on the front and rear side with the thickness of 70 nm and 150 nm, respectively, forming five 2×2-cm² wide cells per wafer. From the light facing side, the stack of the layer is therefore ITO/MoO_x/i-a-Si:H/n-Si/(i+n)-a-Si:H/ITO. Initially, front grid and rear contact were screen printed with Ag paste followed by curing (30 min at 170 °C) and two standard SHJ solar cells [7] were fabricated as reference. After measuring our devices, we found that screen printing was rather aggressive on MoO_x-based cells with V_oc dropping by more than 20 mV with respect to the implied V_oc prior metallization. So, to isolate the performance of the thin MoO_x HTL, we used a thin Ag layer as seed for room temperature Cu-plating at the front side (~1.6% metallization fraction) and sputtered a full-cell wide 500-nm thick Ag at the rear side. Overall, thanks to our PT, thinner-than-4nm MoO_x layers in Cu-plated SHJ FBC cells delivered excellent transport (FF > 81%) in combination with good passivation (Voc > 720 mV ). So far, the best device was endowed with 2.9-nm thick MoO_x and exhibited η_active = 22.92% (V_oc = 722 mV, J_{sc_EQE} = 39.04 mA/cm², FF = 81.3%). The realization of other thicknesses of MoO_x in SHJ FBC solar cells is ongoing and will be presented at the conference. From preliminary passivation tests up until front/rear sputtered ITO, we expect even higher performance for thickness of MoO_x < 2 nm, which would positively translate into larger cell J_sc.

[1] X. Ru, et al., Sol. Energy Mater. Sol. Cells, 2020.
[2] K. Yoshikawa, et al., Nat. Energy, 2017.
[3] H. Mehmood, et al., Renew. Energy, 2019.
[4] J. Dréon, et al., Nano Energy, 2020.
[5] J. Dréon, et al., IEEE J. Photovolt., 2021.
[6] L. Mazzarella, et al.,Prog. Photovolt. Res. Appl., 2020.
[7] Y. Zhao, et al., Prog. Photovolt. Res. Appl., 2020.

Silicon heterojunction (SHJ) solar cells have exhibited high efficiencies in both academia and industry [1,2]. Key challenges to be addressed in the upscaling process are the cost of materials and the abundance of the utilized elements. Therefore, consumption of Ag and In, widely used in the electrodes of SHJ devices, must be reduced [3]. Cu has proven to be a good candidate for replacing Ag electrode; however, complexity in the metallization process remains to be tackled [4]. For the widely used electrochemical Cu-plating technique, separate control is generally needed for processing both sides of the wafer, which means that one side of the wafer is protected when plating the other side [5]. This step further increases the cell production costs. On the other hand, to reduce the In consumption, basic strategies include: (i) use of In-free TCOs, such as AZO; (ii) reduction of the TCO thickness (TCO-less option); (iii) development of TCO-free SHJ devices. For (i), challenges may lie in low carrier mobility, interface/material stability issues [6]; For (iii), proof-of-concept TCO-free SHJ cells have been demonstrated recently [7], but efforts have been partially frustrated from the passivation deterioration, and contact problems [8]. Also, it is still an open question whether TCO-free design could give optimal device performance. Moreover, a TCO layer is practically needed to act as a Cu diffusion barrier layer in Cu-plating processes. To circumvent these limitations, we focus on the TCO-less solution in combination with Cu-plating metallization approach.

We developed TCOs with different opto-electrical properties, which are tin-, fluorine-, and tungsten-doped indium oxides, namely, ITO, IFO, and IWO [9][10]. By utilizing a double layer anti-reflection coating formed by TCO/SiO_x layer stacks on both sides of our SHJ device structure, we performed optical simulations with varied front and back TCO/SiO_x thicknesses. Results show that bifacial solar cell architecture provides the most effective way to reduce the TCO use, and the optimal optical response appears at the point where a thin TCO layer is applied, due to a low free carrier absorption and a favorable refractive index enabling light in-coupling. Thus, we applied 25 nm-thick of different TCOs on both sides of the cell precursors fabricated using plasma-enhanced chemical vapor deposition technique. Finally, we developed single step Cu-plating process on lithography patterned openings, which could well control the metal finger growth on both sides of the wafer and largely simplify the simultaneous metallization on both sides. Champions bifacial cell was observed in the devices with IWO use on front and ITO use on the rear side of the device. After applying an additional 110 nm-thick SiO_x layers on both sides of the complete cell, the device parameters were measured to be: V_OC = 721 mV, J_SC = 39.20 mA/cm², FF = 79.57%, η = 22.50% (n side illumination); V_OC = 717 mV, J_SC = 38.72 mA/cm², FF = 79.91%, η = 22.19% (p side illumination). The bifaciality factor is 0.99. Compared to the standard TCO utilization of 75 nm and 150 nm on front and rear side of the monofacial SHJ device, respectively, we achieved comparable cell performance with a 78% reduction in TCO use. TCO thickness-dependent contact properties for both n-contact and p-contact are still under investigation.

[1] K. Yoshikawa, et al., Sol. Energy Mater. Sol. Cells, 173, 37-42(2017)
[2] TaiyangNews, 2nd Heterojunction report, 15-17 (2020)
[3] E. Gervais, et al., Renew. Sustain. Energy Rev., 137, 110589 (2021)
[4] J. Yu, et al., Sol. Energy Mater. Sol. Cells, 224, 110993 (2021).
[5] H. H. Kuhnlein, SiliconPV 2021.
[6] A. B. Morales-Vilches, et al., IEEE J. Photovolt., 9, 34-39(2019)
[7] S. Li, et al., Joule, 5, 1535-1547 (2021)
[8] M. Bivour, et al., SiliconPV 2021.
[9] C. Han, et al., ACS Appl. Mater. Interfaces, 11, 45586 (2019)
[10] C. Han, et al., Sol. Energy Mater. Sol. Cells, 227, 111082 (2021)

The high open circuit voltage (up to 750 mV) of heterojunction solar cellsis provided by the excellent ability of a layer of hydrogenated amorphous silicon to passivate the surface of the silicon wafer. A number of methods for the deposition of a-Si:H layers are well known. To obtain solar cells with high parameters, it is necessary not only to get good passivation of the crystalline silicon surface, but also to create conditions for efficient charge transfer. One of the widely used methods of deposition of a-SiH is the method of PECVD; during the formation of a-Si:H layer, various types of defects arise. Various process steps, including annealing, lead to a decrease in the number of defects.
In this work, we investigated the formation of defects and the properties of the a-Si:H layer after two-stage deposition with intermediate hydrogen plasma treatment, followed by deposition of a transparent conducting oxide, annealing of defects, and the influence of these factors on the characteristics of solar cells.
For the study, we usedn-type monocrystalline silicon wafer with a resistivity of 1-5 Ohm * cm, with a size of 156.75 * 156.75 mm, and 180 microns thick.
After the damaged layer removal, texturing and appropriate cleaning a-Si:H layers were applied to the surface of silicon wafer with the goal to obtain a p-i-i-n structure on one side and n-i-i- n⁺ on the other side. An ITO layer about 75 nm thick was deposited by magnetron sputtering of an In-Sn target in an argon-oxygen atmosphere. Metallization was formed as a final step and the parameters of solar cells were measured.
As a result of the studies carried out, it was found that the use of an intermediate hydrogen plasma treatment of the first layer of amorphous silicon leads to the passivation of defects with hydrogen; an increase in the number of silicon-hydrogen bonds and, as a consequence, an increase in the microstructure factor R from 0.4 to 0.55.This additional step in the PECVD process results in an increase in the lifetime of minority charge carriers from 2 to 5 milliseconds and an increase in the efficiency of the solar cell.

Impurities and defects continue to play an important role in limiting the efficiency of both laboratory and industrial silicon solar cells. In this work, we review recent progress in reducing the impact of some of the most important of these limiting defects through impurity gettering and hydrogenation steps, which occur coincidentally during device fabrication. Since the dominant defects in a material depend very strongly on the crystal growth conditions, we categorise them by the three common silicon crystal growth techniques – Czochralski-grown mono-crystalline silicon, Cast-grown silicon (including both multicrystalline and cast-mono silicon), and Float-Zone silicon.

Czochralski-grown monocrystalline silicon (Cz-Si)
The primary defects in Cz-si, whether p- or n-type, are oxygen-related. In boron-doped p-type silicon, the very well-known Light-Induced Degradation (LID) caused by boron-oxygen defects dominates. Although in principle it is not affected by gettering, hydrogen is widely thought to play a key role in the permanent regeneration of this defect, which is a crucial step in stabilising the high efficiency p-type PERC cells that now dominate the PV industry.
In n-type Cz-Si, which is the preferred substrate for IBC, SHJ and poly-silicon passivating contact solar cells, an important class of defects that sometimes occur during high-temperature processing are the oxygen-related ring defects. These consist of oxygen precipitates, and sometimes their associated dislocations and decorating metallic impurities. Under certain conditions, they can be partially mitigated by impurity gettering and hydrogenation steps. These ring defects can also occur in p-type Cz-Si wafers.
Industrial Cz-Si wafers often also contain traces of dissolved metals such as Fe, which can significantly affect the carrier lifetime, if not removed during processing by gettering. Such gettering can be achieved by phosphorus or boron diffusion, aluminium alloying, doped-poly-silicon layers, and even deposited dielectric films such as silicon nitride and aluminium oxide.

Cast-grown silicon (HP mc-Si and cast-mono Si)
Cast high-performance multicrystalline silicon (HP mc-Si), by design, contains many grain boundaries. Fortunately the recombination activity of these grain boundaries can be significantly reduced after hydrogenation, such as via fired silicon nitride films during cell metallisation. Cast-monocrystalline silicon (cast-mono Si), on the other hand, tends to be dominated by dislocation clusters. These are much more difficult to negate through hydrogenation, which reduces the achievable efficiency and yield for devices made on such wafers. Both HP mc-Si and cast-mono wafers also contain significant quantities of metallic impurities, some of which are precipitated at structural defects, whilst others remain dissolved. Under the right conditions, they can be gettered to a significant extent during phosphorus or boron diffusions, or by deposited films. This provides a convenient method to identify the metallic species present in these materials.

Float-Zone silicon (FZ-Si)
Although not used in industrial solar cells, Float-Zone silicon (FZ-Si) is often used in research laboratories, partly due to its immunity to oxygen-related LID and ring defects, and very low metallic impurity concentrations. However, due to the addition of nitrogen during crystal growth to control dislocations, they are subject to the formation of nitrogen-vacancy complexes. Once formed, these defects are highly recombination active. However they can be partly mitigated by hydrogenation.

Finally, although hydrogen is helpful in passivating some defects in silicon, it can also cause recombination-active defects when present in excess, such as the widely-researched Light and elevated-Temperature Induced Degradation (LeTID). Since these defects are caused by firing hydrogen-containing dielectric films, they can be found in all three types silicon wafer described above, whether n- or p-type.

It has been known for decades that solar cells fabricated from Czochralski-grown silicon (Cz-Si) using Boron as p-dopant display a loss of efficiency in their first few hours of operation. In order for the effect to be observed both Oxygen and Boron need to be present, giving the effects its usual name of Boron Oxygen light induced degradation or BOLID. It has been recently argued that a complex consisting of a substitutional boron atom and two oxygen atoms (B_sO₂) is a center with negative-U properties, and it is responsible for minority carrier trapping effects, persistent photoconductivity and light induced degradation of lifetime (and hence BOLID) in these materials. Depending on the position of the Fermi level, the B_sO₂ defect is either a deep donor or it can be in a few configurations with a shallow acceptor level. The E(-/+) occupancy level of the defect has been found to be at E_v + 0.32 eV using deep level transient spectroscopy (DLTS). Transitions between the different states associated with hole/electron emission or capture by the defect have been monitored with the use of junction capacitance techniques. The presenter will review the evidence that this defect is responsible for this energy loss process in commercial cells from a combination of technqiues, DLTS and optical spectroscopy.
The ability to detect the defect electrically gives scientists a vital tool to detect the presence of this defect, for Boron and, potentially, other alternative p-type dopants. The presenter will review recent work at Manchester on alternative dopants. We have detected a hole emission signal similar to that due to the B_sO₂ defect in Schottky diodes on p-type Si crystals doped with either Al, Ga, or In impurities. It appears that nearly identical hole emission signals with their maxima at about 391-392 K occur in all the DLTS spectra. The hole emission rates have been measured in these diodes in the temperature range from 360 K to 420 K with the use of the high-resolution Laplace DLTS technique. Arrhenius plots of the measured hole emission rates in Si samples doped with either B, Al, Ga, or In impurity atoms show that the emission rates and the derived values of activation energy for hole emission are very close for the differently doped samples. Finally we have carried out measurements of temperature dependencies of equilibrium occupancy and hole capture kinetics for the detected traps in samples doped with the different acceptor impurities. It is found from the analysis of the changes in magnitude of the DLTS signals with temperature that in all the samples the equilibrium occupancy function of the traps are characteristic for a defect with negative-U properties. Further, it is revealed that positions of the E(+/-) occupancy level are very close in different samples, E(+/-) » E_v + 0.32 eV. We suggest that the observed negative-U defects in Cz-Si samples doped with different acceptor impurities are associated with the complexes incorporating a substitutional group-III impurity atom and two oxygen atoms (A_sO₂, where A can be B, Al, Ga, or In). According to the model proposed transformations between the deep donor and shallow acceptor configurations of the B_sO₂ center are related to a transformation of the oxygen dimer in the vicinity of a B_s atom from a “square” to a staggered configuration. It appears that the presence of different group-III atoms in the A_sO₂ complex does not result in significant changes of the O₂ transformation processes. However, the (B_sO₂) to (B_sO₂)⁺ reaction is enhanced compared to that for other Group III atoms probably because of local tensile strain induced in the lattice by a relatively small substitutional boron atom. It is argued that this enhancement is one of possible reasons of much more effective formation of recombination active defects, which are responsible for LID, in p-type material doped with boron compared to that in Si crystals doped with other group-III impurity atoms.

Predicting degradation rates under fielded conditions is important for understanding energy yield, project return on investment, and how to assign higher value to products where the degradation has been eradicated. Some processes, such as light and elevated temperature induced degradation (LeTID), and the regeneration portion of boron-oxygen light induced degradation (BO LID), are accelerated by the presence of excess carriers. It is therefore important to understand how excess carrier concentration changes as a function of conditions and materials properties both in operating solar cells, and in wafers, where these processes are often studied. In this investigation, we simulate excess carrier concentration as a function of wafer thickness and minority carrier lifetime in solar cells and wafers using SCAPS and Quokka. The same results are obtained whether one uses simulation software or analytic expressions based on the physics of p-n junctions. For wafers, there is a simple nearly linear relationship between excess carrier concentration, lifetime, and thickness. For solar cells, as lifetime increases or thickness decreases, excess carrier concentration becomes limited by rear surface recombination. Thus, LeTID may appear to progress more quickly in wafers than in cells, and the degradation rate will have a stronger dependence on minority carrier lifetime in wafers than in cells.

Over the last years, tremendous progress could be observed in improving the efficiency of industrial crystalline Si solar cells. The race towards higher efficiencies brings degradation phenomena occurring under working conditions of the solar cell more and more into the focus as the bulk material quality and surface passivation both play a crucial role for obtaining high efficiencies. In this presentation an overview of the known degradation phenomena on solar cell level will be given. This includes degradation of the bulk quality like the well-known boron-oxygen(BO)-related degradation under injection of charge carriers and the more recently detected LeTID (light and elevated temperature induced degradation) phenomenon as well as degradation of surface passivation quality under injection conditions in conjunction with elevated temperatures.
For BO-related degradation, the presence of H in the Si bulk allows for a regeneration of bulk material quality after degradation, which is stable under further illumination. The LeTID phenomenon seems to be triggered by the presence of H in the Si bulk, too, but here H seems to be the (direct or indirect) cause for degradation. As for BO-related degradation, a regeneration process for LeTID can be observed for longer treatment times, making it harder to cope with in cell processing. For longer treatment times, often changes in the surface passivation quality can be observed as well, at least for some dielectric layer systems. While the detailed microscopic picture of the defect formation is still unclear at least for LeTID and most surface passivation degradation phenomena, H seems to be involved for all of them in a direct or catalytic way.
For BO-related degradation a recently developed generalized model including the interactions of the BO-related defect and H in the c-Si wafer will be discussed. For LeTID, some key findings and experiments will be presented to highlight the state-of -the-art of understanding. For degradation of surface passivation selected examples will be given and discussed. In the last decades, B-doped c-Si had highest PV market share. Currently a trend towards Ga-doped c-Si wafer material can be observed which eliminated the BO-related degradation. But LeTID seems to be observed in Ga-doped c-Si wafers, too, although with a different kinetics. Some recent findings in this field will be presented and discussed.