Symposium EN01—Silicon for Photovoltaics

Solar cells made from crystalline silicon not only dominate the emerging photovoltaic market for many years. They also undergo a technology transformation towards higher efficiency concepts, both, in industrial production as well as in laboratory scale prototypes. All these concepts rely on contacting schemes that rather put the contacts on the silicon wafer instead of the in-diffusion of classical p-n or high-low junctions. This paradigm shift has also enforced a renewed discussion on the working principles of solar cells with a special focus on the so-called selectivity of contacts to solar cells in general. The contribution will start with the practical discussion of the transparency, conductivity, and passivation properties of the solar cell surfaces with view on the minimization of optical, resistive, and recombination losses, respectively. It will be shown that efficiency gains mostly result from optimizations along physically conflicting lines. For instance different concepts for contact selectivity in solar cells predict maximum efficiencies as a compromise between the recombination and resistive losses. Diving deeper in this topic, it will be shown that all present concepts for selectivity up to now are incomplete as they either focus on the electrical picture or on the chemical potential losses. The contribution will propose a more general thermodynamic concept for charge carrier collection and photovoltage generation in solar cells. This concept embraces the chemical and electrostatic potential in all functional layers of the solar cell, thus enabling a full thermodynamic gain/loss analysis. Finally, recent technological achievements and experimental findings will be discussed in the frame of the new concepts.

Transparent conducting electrodes (TCEs) are essential in many optoelectronic devices including solar cells, LEDs, sensors and displays. The most common approach for TCEs is the use of transparent metal oxide layers, in particular Indium tin oxide (ITO). However, ITO comes with numerous drawbacks such as the relatively high cost, fragility, rarity of Indium, strong UV absorption and relatively high sheet resistance. An exciting TCE alternative is the use of interconnected metallic nanowires, where their sub-wavelength cross-sections makes them transparent to visible light. Especially Ag nanowires are a good candidate because of their high conductivity. However, the high material cost of Ag limits the usage in large-scale applications and therefore Cu is used as an alternative.
In this work, we show a promising cost-effective method to fabricate metal nanowire structures using selective-area electrochemical deposition (SAEC). SAEC is based on electrochemical deposition through an insulating template onto a conductive substrate. Next to the low fabrication cost, electrochemical depositions is a bottom-up method with numerous advantages such as scalability, ambient operation conditions, control over nucleation and high purity. While submicron sized features have been already demonstrated by SAEC, obtaining homogeneous void free filling of features below 100 nm using direct plating is a major challenge because of the terminal effect and the poor control of nucleation of the seed layer.
Here, we have explored the SAEC fabrication of Ag and Cu nanowire grids over a large area by using soft-conformal imprint lithography as means to fabricate the template on ITO. Finally, we compare the optical and electrical properties of the SAEC-fabricated metal grids with those fabricated using conventional thermal or e-beam evaporation followed by lift-off. The SAEC- fabricated metal grids show high transmissivity (~90%) and low sheet resistance (< 50 Ω/sq), whose can be easily tuned by adjusting the thickness of the wires by controlling the growth time. This work represents a low-cost and scalable strategy to fabricate sub 100 nm predefined metal structures, enabling the commercial viable fabrication of nanowire based transparent conductive electrodes for optoelectronic devices.

Photon upconversion enables the conversion of low-energy photons into high-energy photons. It has potential applications in many fields, such as solar cells, bioimaging and photodynamic therapy.^[1] Among photon upconverting systems, silicon quantum dot (Si QD) -molecular structures emerged recently due to non-toxic and bio-compatible properties. A high photon upconversion efficiency of 7% has been reported for the Si QD - 9-ethylanthracene (Si:9EA) system that is potentially biocompatibility.^[2] This initial demonstration of triplet energy transfer between Si QDs and molecular transmitters motivated us to investigate various physical parameters to improve the photon upconversion efficiency. In this work, we explore factors like Si QD size, molecular transmitter loading, the length of surface-passivating ligand and investigate how they affect triplet energy transfer and photon upconversion. In addition, we designed a new molecular transmitter (9-vinylanthracene) with a double-bond bridge. This conjugated carbon bridge introduces strong electronic coupling between the Si QD and anthracene transmitter. We found a much shorter triplet transfer time constant (3ns) for the Si:9VA system compared with the Si:9EA system (15ns) that has less electronic coupling. A high photon upconversion efficiency exceeding 10% was achieved for Si:9VA system. Overall, this work demonstrates that the triplet energy transfer in silicon quantum dot-molecular hybrid system could be mediated and improved by promoting electronic coupling. This is a new strategy for controlling energy or charge transfer at nanoscale interfaces.

[1] Yanai, N.; Kimizuka, N. New triplet sensitization routes for photon upconversion: thermally activated delayed fluorescence molecules, inorganic nanocrystals, and singlet-to-triplet absorption. Acc. Chem. Res. 2017, 50, 2487–2495.
[2] Xia, P., Raulerson, E.K., Coleman, D., Gerke, C.S., Mangolini, L., Tang, M.L., Roberts, S.T. Achieving spin-triplet exciton transfer between silicon and molecular acceptors for photon upconversion. Nature Chem. 2020, 12, 137–144.

Inorganic solar cells on the market today fall short of absorbing all of the energy available in sunlight. Inorganic quantum dots functionalized with triplet exciton accepting surface ligands are a promising platform for the upconversion to improve the efficiency of these photovoltaic (PV) cells and overcome the Shockley–Queisser limit. Here, we functionalize silicon quantum dots (Si QDs) with triplet exciton-accepting perylene ligands which capture photons with wavelengths as long as 730 nm to deliver spin-triplet excitons to their surrounding environment that drive upconversion via triplet fusion. Using transient absorption spectroscopy, we have characterized the efficiency of key steps in this energy transfer chain. We find spin-triplet excitons produced in perylene-functionalized Si QDs establish an equilibrium wherein they readily move between the Si QD core and surface-bound perylene molecules. Adjusting the electronic structure of the surface-bound exciton acceptor can shift this equilibrium, making nearly all excitons available for extraction. Importantly, our work not only identifies ways to improve the function of Si QD-based photon upconversion systems, but also decisively demonstrates molecule-to-silicon spin-triplet exciton transfer is possible. This finding can enable the design of light-harvesting systems and quantum information devices that make use of singlet fission, triplet fusion’s inverse process, to achieve unique functionality.

Broadband absorption of solar photons is a perquisite for any application concerning harvesting of solar energy. The current study shows excellent broadband absorption of the solar spectrum with arrays composed of light funnel arrays integrated with top nanolenses. Light funnel arrays provide high broadband absorption governed by low transmission levels. However, light funnel arrays suffer from relatively high reflection. Herein we show how the broadband absorption is governed by the inherent low transmission levels induced by the light funnel geometry coupled with efficient broadband reflection decrement induced by the presence of top nanolenses. The integration of nanopillar arrays and light funnel arrays with nanolenses is examined. The light funnel geometry provides a more favorable optical coupling with nanolenses with a broadband absorption enhancement of ∼12% under normal illumination. The enhancement in the broadband absorption of light funnel arrays due to nanolenses is also examined and confirmed for oblique illumination. Photovoltaic cells based on light funnel arrays with nanolenses are numerically realized and an increase of ∼10% in short-circuit current is demonstrated. Finally, the associated light trapping mechanisms are qualitatively discussed.

Improving c-Si solar cell contacts is becoming progressively important as efficiencies creep towards the limit of 29.4%. An efficient contacting scheme should exhibit a combination of functionalities: It extracts majority charge carriers with a low contact resistivity and suppresses the recombination of minority carriers at the contact. The contact should moreover exhibit a high transparency as well as provide sufficient lateral conductivity towards the metal grid. The goal of simultaneously satisfying these requirements to the maximum extent is the main challenge for contact design.

In this work, we show how SiO₂/ZnO/Al₂O₃ stacks are an all-round excellent contact to n-type surfaces, checking all the boxes above. Recently we already showed these stacks yield state-of-the-art passivation of n-type c-Si surfaces.² While conductive ZnO has historically not been known for its passivating properties, the innovation that enabled excellent passivation is the Al₂O₃ capping layer: During an activation anneal at 400-500 ^oC the Al₂O₃ prevents hydrogen effusion from the ZnO, such that the H can passivate the defects in the tunnel SiO₂.

Interestingly, the distinguishing aspect of this novel passivation stack is the very conductive nature of the passivating ZnO:Al, potentially enabling electrical contact to the c-Si. Note that contacting the ZnO:Al by metal is possible since the Al₂O₃ can be removed after annealing with a wet etch without impairing the ZnO passivation. In this work we show how such an approach leads to excellent contacting properties on both diffused c-Si(n⁺) and poly-Si(n) surfaces, as typically found in PERC and TOPCon solar cells.³

1) Passivation: The stacks yield excellent passivation of textured c-Si(n) surfaces (iV_oc 728 mV), n⁺-diffused c-Si surfaces (R_sheet 260 Ω/sq, iV_oc 699 mV) as well as hydrogenation of poly-Si(n) contacts (iV_oc 735 mV). Note that the result on the n⁺-diffused surface was obtained on an industrial PERC half-fabricate cell and is on par with conventional SiN_x.

2) Contact resistivity: Doping of both the ZnO:Al and the c-Si is crucial for obtaining a low contact resistivity. Excellent values down to 15, 23 and 42 mΩcm² were obtained on 130 Ω/sq n⁺-diffused Si, 260 Ω/sq n⁺-diffused Si and poly-Si(n) surfaces, respectively.

3) Conductivity and transparency: The activation anneal also strongly enhances the carrier mobility of the ZnO:Al, enabling a low resistivity (0.5 mΩcm) and low free-carrier absorption that is typically only attainable with indium-based TCOs.

4) Thermal stability: The stacks are thermally stable up to 600 ^oC. This is not firing-stable yet significantly higher in comparison to a-Si:H and most oxide-based contacts.

At the conference, detailed mechanistic insights behind these findings will be presented. Moreover, we will address the high potential of this stack as a building block for c-Si solar cells: as passivating front contact in PERC-type solar cells, as interface contact in PERC-perovskite tandem cells, as well as conductive hydrogenation source for TOPCon structures.

This work is part of the TKI PERCSpective project and we gratefully acknowledge all the co-workers within this consortium.

1. T. G. Allen, J. Bullock, X. Yang, A. Javey, and S. De Wolf: Passivating contacts for crystalline silicon solar cells. Nat. Energy 2 (2019).
2. B. W. H. Van De Loo, B. Macco, J. Melskens, W. Beyer, and W. M. M. Kessels: Silicon surface passivation by transparent conductive zinc oxide. J. Appl. Phys. 125(10) (2019).
3. B. Macco, B. W. H. van de Loo, M. Dielen, D. G. J. A. Loeffen, B. B. van Pelt, N. Phung, J. Melskens, M. A. Verheijen, and W. M. M. Kessels: Atomic-layer-deposited Al-doped zinc oxide as a passivating conductive contacting layer for n+-doped surfaces in silicon solar cells. Sol. Energy Mater. Sol. Cells 233(September), 111386 (2021).

Hafnium oxide (HfO₂) is a wide bandgap electrical insulator, frequently used in the electronics industry due to its dielectric properties. With its high dielectric constant, high refractive index, inherent mechanical robustness, and potential for charge storage, HfO₂ makes an excellent candidate for silicon surface passivation. Importantly, HfO₂ can easily be deposited using established industrial processes, such as atomic layer deposition (ALD), which is now widely used in the silicon PV industry to deposit Al₂O₃ films. Importantly, in contrast to Al₂O₃, HfO₂ generally provides a positively charged film¹, giving an additional degree of freedom for field effect passivation in certain solar cell designs.

We have investigated the effects of various post-deposition annealing conditions on the carrier lifetime produced by ALD-grown HfO₂ films, focusing on optimising the temperature, time and ambient. Prior research on this topic used float-zone wafers^1–3, whose lifetime is has been found to degrade at the temperatures at which the passivation is activated^4,5. In contrast, we use Czochralski silicon (the material of choice for high efficiency solar cells), in conjunction with room temperature superacid passivation to isolate surface and bulk effects which may have influenced previous work.

We deposit our hafnium oxide using plasma-enhanced ALD at 200 °C, and subsequently anneal the sample to activate the surface passivation. We find that effective carrier lifetime initially increases with increasing temperature, up to around 450-500^oC, and at higher temperatures is begins to degrade. This temperature is significantly higher than past reports which all used float zone silicon, which we believe could have been influenced by bulk carrier lifetime degradation effects not related to surface passivation^4,5. Through this optimisation process we have been able to achieve surface recombination velocities <5 cm/s.

To support our lifetime results we conduct a series of superacid re-passivation experiments, whereby the HfO₂ films are etched away and the sample is subsequently re-passivated at room temperature using a superacid-based method^6,7. This approach, which has not been applied in this context previously, enables us to separate competing surface and bulk effects. Using this method ensures that the measured effects are solely attributed to the HfO₂ films themselves, and not influenced by bulk degradation due to the thermally induced recombination centres. Initial experiments show no obvious degradation to the bulk of the Czochralski silicon at these temperatures, indicating the beneficial effects of higher temperature activation in the context of photovoltaic cells.

With the aim of understanding the origin of the effects observed, we use high resolution grazing incidence X-ray diffraction to characterise the HfO₂ films as a function of activation temperature. Results suggest the crystallinity of the film is important in achieving good passivation but there are other factors involved, including changes to field effect and chemical passivation. We study these competing effects by analysing the injection dependence of the effective lifetime and also through Kelvin probe measurements. This work clarifies the role both HfO₂ crystallinity and silicon bulk effects play within the passivation process, and furthers the development of HfO₂ as the main candidate for a positively charged ALD-grown passivation layer for solar cells.

1 J. Cui, et al., Appl. Phys. Lett., 2017, 110, 021602.
2 A. B. Gougam, et al. Mater. Sci. Semicond. Process., 2019, 95, 42–47.
3 S. Tomer, et al., IEEE J. Photovoltaics, 2020, 10, 1614–1623.
4 N. E. Grant, et al., Phys. Status Solidi - Rapid Res. Lett., 2016, 10, 443–447.
5 N. E. Grant, et al., Phys. Status Solidi Appl. Mater. Sci., 2016, 213, 2844–2849.
6 N. E. Grant, et al., IEEE J. Photovoltaics, 2017, 7, 1574–1583.
7 A. I. Pointon, et al., Sol. Energy Mater. Sol. Cells, 2018, 183, 164–172.

Excellent passivation of silicon surfaces is essential to realize high-efficiency c-Si solar cells¹. As energy conversion efficiencies get closer to the theoretical limit, requirements on passivation schemes become more stringent. Current industrial solar cells often use doping to achieve carrier selectivity, which can be enhanced by applying a passivation scheme with appropriate fixed charge. For p⁺-doped Si surfaces, Al₂O₃ is the gold standard passivation scheme, because it combines a very low interface defect density (D_it) and high negative fixed charge density (Q_f) on Si.² Recently, phosphorous oxide capped by aluminum oxide (PO_x/Al₂O₃) has been found to be able to reach similarly very low D_it values combined with similarly high, but positive fixed charge density. These stacks show great promise for the passivation of n⁺-doped Si surfaces.

In this work, material and interface properties of PO_x/Al₂O₃ stacks were investigated and more insight into the passivation mechanism was gained. Additionally, state-of-the-art surface passivation was demonstrated on textured n⁺-doped Si.

It was found that deposition temperature mainly impacts D_it, which improved at lower deposition temperature. The high positive Q_f was found to be present directly after deposition of the PO_x/Al₂O₃ stack and was not altered significantly by deposition temperature or post-deposition annealing. Annealing did however lead to a significant decrease in D_it by almost three orders of magnitude (~10¹³ to 10¹⁰ eV^-1 cm^-2). For optimal conditions, a D_it as low as 5 × 10¹⁰ eV^-1 cm^-2 in combination with a high Q_f of +5 × 10¹² cm^-2 was obtained, leading to a surface recombination velocity of 2 cm/s.

From a combination of infrared spectroscopy, SIMS, and hydrogen effusion measurements, it was found that the chemical passivation mechanism of the PO_x/Al₂O₃ stacks is related to the passivation of Si dangling bonds provided by hydrogen from the PO_x/Al₂O₃ stack and mixing of aluminum into the PO_x layer. The latter leads to the formation of AlPO₄ upon annealing.

The PO_x/Al₂O₃ stacks were used to passivate textured n⁺ Si surfaces. A recombination parameter (J₀) value of 12 fA cm^-2 is reached for a relatively light surface diffusion (R_sheet = 260 Ω sq^-1). For a stronger surface diffusion (R_sheet = 137 Ω sq^-1), a value of J₀ = 30 fA cm^-2 is obtained. These values are on par with state-of-the-art passivation schemes such as alnealed SiO_x³ and SiO₂/SiN_x/SiO₂ stacks⁴ clearly demonstrating the promise of PO_x/Al₂O₃ for the passivation of n-type c-Si surfaces, in particular for n⁺-doped regions used in industrial solar cells.

1. J. Schmidt, R. Peibst, R. Brendel, Sol. Energy Mater Sol. Cells, 2018, 187, 39-54.
2. B. Hoex, J. Schmidt, R. Bock, P. P. Altermatt, M. C. M. Van De Sanden, W. M. M. Kessels, Appl. Phys. Lett., 2007, 91, 112107.
3. M. J. Kerr, J. Schmidt, A. Cuevas, J. H. Bultman, J. Appl. Phys., 2001, 89, 3821-3826.
4. T. C. Kho, K. C. Fong, M. Stocks, K. McIntosh, E. Franklin, S. P. Phang, A. Blakers, Prog Photovolt, 2020, 28, 1034-1044.

How well can you improve a cell production line? This talk will focus on the use of electrical and material characterization methods to improve the balance between cell and module efficiency and cost by applying a complete IV power loss analysis to individual cell measurements and large-scale production data sets. The main aim of any production facility or researcher in the silicon PV industry is to drive efficiency gains while balancing cell or module reliability and pricing aspects through design and production. For this reason, having the ability to improve cell manufacturing depends on gathering the information to decide where changes in line or cell design need to be made. However, simple characterizations of efficiency, V_oc, I_sc, etc. do not speak well to where these efficiency gains can be made. By computing a complete IV curve power loss breakdown based on minority-carrier lifetime, substrate doping, thickness, J_o, and series resistance, folks managing cell lines and aiming for improved efficiency gains can easily access the information they need to reach target efficiencies or surpass them. For those in research, a complete power loss analysis allows for the evaluation of potential sources of cell process improvement. The question here is what information can production line managers use to improve their efficiencies, and how can they use it without spending hours collecting or analyzing data?
To that end, the biggest existing opportunity to easily measure cell efficiency contributions is at the Cell IV measurement step. At this point, the IV parameters in addition to the analysis parameters can be used to construct an easily interpretable map for production managers or anyone else to use in improving cell and module line efficiencies. Because the data comes in real-time with cell measurements, line managers can monitor results in real-time to track efficiency losses in production. The loss analysis works from the ideal limit of cell efficiency, the total current with only intrinsic recombination present, and steps back in loss mechanism to the true Pmp of the cell as measured by the IV curve. This result provides the power loss as a percentage of the total potential power for that device. For example, a cell with a maximum power of 5.91 W could have a total power loss as a percentage of the Auger limited power potential of 11.85%. This is broken down in the analysis to a loss due to R_s of 4.179%, R_sh of 0.007805%, and a Recombination of 7.668%. The recombination can then be further broken into its constituent parts of 2.572% due to J_o and 5.095% due to bulk SRH effects.
This talk will focus on the applications for the power loss analysis using cell line data and suggest methods of monitoring and improving manufacturing lines. Working with specific cell line data examples the presentation will detail where the most efficiency gains can be made. We will also detail opportunities for catching issues upstream in the production chain to avoid the post-mortem effect that is measuring power loss at the cell IV test step. By using the power loss analysis, based on physical models, optimizations can be applied to the production equipment for metallization, surface passivation, doping, thickness, and raw substrate.

Silicon photovoltaics (PVs) continue to improve through generational advancements, yet two main challenges remain. Photons with energies below the bandgap of the PV are not absorbed, leading to overheating and a decrease in device performance. Photons with energies above the bandgap of the PV undergo thermalization, damaging the device as well. To address these issues, we present a photovoltaic/thermal (PV/T) system which utilizes an infrared-absorbing thermal fluid containing Stokes shifting dyes. Together, the fluid and dyes allow for increased energy extraction from photons outside the bandgap of the PV. Through the use of a 3D curved architecture, we have shown that under solar simulation, when compared to a planar Si PV of the same performance, we are able to collect 98% more electrical energy over the course of a day. In addition to the increased electrical performance there is 35 W/m² of thermal energy generated which otherwise would damages the quality of the PV over time, a common issue with planar traditional PVs. Now, we plan to capture the thermalization of the high energy photons using Stokes shifting dyes and use it to increase overall device performance. We use multiple dyes, including Coumarin 6 and Coumarin 153, at varied concentrations under AM 1.5g standard conditions to study the angular response and thermal performance of the PV/T system. The data is supported by an optical ray tracing analysis using COMSOL software. Through preliminary results we see no major change in PV performance while there is a ~15% increase in thermal performance. With the combined collection of PV and thermal power, an effective efficiency of performance can be significantly above that of planar Si devices, while also being cost-effective and protecting the Si from damaging UV and overheating.

Metallization pastes have played a key role in determining the performance as well as the cost of solar cells and there is continued research activity to develop newer pastes to adapt to the evolving solar cell technology. There are a number of constituents that comprise the metallization pastes. Development of these pastes with improved characteristics requires a thorough understanding of the role of various constituents in the paste and the mechanism of electrical contact formation. Glass, which is one of the key constituents in silver front contact paste, not only acts as an inorganic binder to provide adhesion strength between the silver electrode and silicon substrate, but it also plays a crucial role in contact formation. Glass is used for etching of silicon nitride layer and precipitation of silver nanocrystals at the interface of silver and silicon for contact formation [1] [2]. This mechanism is not yet well understood. It is important to study the interaction of glass with silver and silicon nitride layer to understand the current conduction mechanism at the interface of solar cell. Here we report a study on the effect of composition and T_g of the glass on contact formation by closely examining the microstructural evolution at the interface. Two different pastes A and B composed of lower and higher T_g glass frits respectively were printed on 5cm x 5cm wafers (65Ω/sq) for solar cell fabrication. Electrical properties are correlated with the flow behaviour of glass during firing [3]. Current-Voltage measurement and Fill Factor (FF) analysis suggest that the loss in FF of Cell B is due to high series resistance. Electroluminescence (EL) imaging shows the quality of electrical contacts over the cell area [4]. Si-silver paste interface investigation is done using the Focused Ion Beam-Scanning Electron Microscope (FIB-SEM) technique. A thin interfacial glass layer decorated with silver nanocrystals is found in case of Cell A. Higher T_g glass is not flowable during firing and causes increased thickness of residual glass among Ag bulk. So, silicon nitride layer remains intact even after firing, which is evident from the high series resistance of Cell B. Sequential removal of Ag bulk finger and interfacial glass layer removal is also done. Results are correlated by analyzing microstructure development at the tip of textured silicon pyramid. It is expected that the understanding gained may help in the development of glass compositions/pastes that may help achieve the desired contact formation at lower firing temperatures.
Keywords: Metallization Paste, Solar Cell, Glass Transition Temperature, Focused Ion Beam, Electroluminescence Imaging
References:
[1] F. H. P. F. Gunnar Schubert, "Physical understanding of printed thick-film front contacts of crystalline Si solar cells—Review of existing models and recent developments," Solar Energy Materials & Solar Cells, vol. 90, p. 3399–3406, 2006.
[2] F.-W. lM.EbersteinH, "Sintering and Contact Formation of Glass Containing Silver Pastes," Energy Procedia, vol. 27, pp. 512-530, 2012.
[3] S. S. S. H. a. H. K. Dongsun Kim, "Electrical Properties of Screen-Printed Silicon Solar Cell Dependent upon Thermophysical Behavior of Glass Frits in Ag Pastes," Japanese Journal of Applied Physics, 2009.
[4] K.-M. D. S. M. L. Drabczyk, "Electroluminescence imaging for determining the influence of metallization parameters for solar cell metal contacts," Solar Energy, vol. 126, pp. 14-21, March 2016.

Neutral-colored transparent crystalline silicon photovoltaics (t-Si PV) was reported as remarkable building-integrated photovoltaics (BIPV) due to its high efficiency of 12.2% with excellent transmittance (20% in the whole absorption range). Deep reactive ion etching (DRIE) is required to fabricate this photovoltaics where the microhole-shaped light transmission windows were placed on bare crystalline silicon (c-Si, ca. 200 μm thickness) to transmit all visible wavelengths, which causes both high fabrication cost and the plasma damage on Si surface. A prominent method replaceable DRIE is metal-assisted chemical etching (MACE) that occurs through the redox reaction at the Si/metal catalyst (Au, Ag are representative)/etchant solution interfaces because it features simplicity, versatility, and cost-effectiveness. However, the metal catalysts that grow via electroless deposition are deposited as particle-form and act individually during MACE (i.e., uneven MACE), thus yielding a lot of defects on the resultant Si.
Here, the porous monolith Ag is successfully deposited on c-Si by regulating the interaction between Ag cation and Si surface. The epitaxial growth of Ag is manipulated preferentially to occur along the lateral by introducing the acetonitrile that coordinates with Ag cation on the c-Si surface and acts as a surfactant between c-Si and Ag precursor solution. The random movement of Ag catalysts is inhibited owing to the cooperative motion of its porous monolith morphology without etching rate loss compared to its particle-form, which reduces the uneven etching of c-Si. Besides, the chemical stability of this Ag catalyst during MACE is improved through the Au passivation, thus, also preventing the indiscriminate defect generation. As a result, t-Si PV fabricated through MACE based on the porous monolith Ag/Au shows a high performance of 12.1% with 20% neutral-colored transparency, which is comparable to t-Si PV through DRIE and would contribute to the t-Si PV commercialization.

Luminescent solar concentrators (LSCs) are large, flat waveguides doped or coated with fluorophores that reemit absorbed solar radiation at longer wavelengths; the reemitted luminescence is then guided to peripheral or surface-mounted solar cells. LSCs are promising for greenhouse roof panels that produce electricity while transmitting sunlight. However, although there have been many reports of high-efficient LSCs, only a few studies investigated the integration of LSCs into greenhouses. Here, we study the impact of LSCs integration on greenhouses and their potential to meet the greenhouse energy demand. We consider Si quantum dots (QDs) LSC roof panels with Si photovoltaic (PV) cells attached in a square pattern. Silicon QDs are attractive for LSC applications due to being environmentally benign, cost-effective, and having desirable optical properties. Using dynamic energy balance models for Si LSCs greenhouses and analytical models for Si LSC roof panels, we show that the solar power generated can satisfy the energy demand for greenhouses deploying Si LSC panels, and thus enable net-zero energy (NZE) greenhouses. While the deployment of Si LSC panels reduces solar radiation entering the greenhouse, solar resource and transmission calculations indicate that sunlight over the photosynthetically active radiation (PAR) spectrum (400-700 nm) is sufficient for plant growth. Furthermore, the solar radiation drop is found to decrease water consumption for evaporative cooling in greenhouses. In summary, the results of the current research demonstrate the strong potential of integrating Si LSCs into greenhouses to save energy and water, and provide a guideline for LSCs design and assessment in agriculture application.
The authors also acknowledge partial support from the Minnesota Environment and Natural Resources Trust Fund (M.L. 2018, Chp. 214, Art. 4, Sec. 02, Subd. 07a) and the University of Minnesota through the Ronald L. and Janet A. Christenson Chair in Renewable Energy.

Mainstream solar cell production now preferentially uses gallium doped silicon wafers. Gallium doped substrates offer inherently better carrier lifetime stability than boron doped ones, without a need for post-cell production stabilisation processes. The transition to gallium occurred very rapidly, as no changes to cell lines were required. Although the benefits are clear, the properties of the material are relatively poorly understood, and a set of specific research issues specific to gallium doped silicon solar cells has started to emerge.
The first issue concerns potential performance limitations specific to gallium doped silicon. In terms of carrier lifetime – the most important property for solar cells – the highest reported historical values were typically ~1 ms. We have evaluated the performance limit of modern gallium doped silicon materials with state-of-the-art surface passivation [1]. We find this to result in considerably higher lifetimes than usually reported, with lifetimes > 9 ms achieved in some cases. The lifetimes we measure in ungettered gallium doped silicon are similar to those for boron doped silicon which had undergone both gettering and boron-oxygen stabilisation processes. As such, equivalent high performance can be achieved with gallium doped silicon substrates without the need for stabilisation processing.
A second issue is the stability of gallium doped solar cells, particularly those of the market-dominating passivated emitter and rear cell (PERC) variety. We use a photoluminescence imaging proxy method to study the performance of commercially produced cells under illumination for very long times (> 3000 h in some cases). Our published reports have shown gallium doped PERC cells undergo a degradation in performance, which then recovers, reminiscent of light and elevated temperature induced degradation (LeTID) [1, 2]. The magnitude of the degradation with gallium doping is less than in equivalent boron doped PERC cells. Stabilisation processes of the kind used to mitigate boron-oxygen light induced degradation do not appear to influence behaviour in our gallium doped cells. A dark pre-anneal (up to 300 °C) can influence the LeTID-like behaviour, although the behaviour has been found to be inconsistent between batches, highlighting the need for further controlled studies. Ongoing work aims to address the dependence of degradation and recovery on cell substrate resistivity, with factors such as unintentional boron doping also being taken into account.
The final issue we address is carrier lifetime changes which occur in illuminated gallium doped silicon wafers. Historically, lifetimes in gallium doped silicon wafers are assumed to be stable, however it is now clear that this is not always the case. Kwapil et al. have recently reported LeTID behaviour in gallium doped wafers [3] and interestingly the degradation has a larger magnitude when the samples are exposed to relatively low levels of illumination (0.1 Suns). We will report new data which support these findings, using samples with different levels of gallium.
In summary, our work generally confirms the positive effects of a transition from B to Ga doping for solar cells. However, given the recent switch, understanding and mitigating these newly found degradation processes is one of the most pressing issues in silicon photovoltaic research at present.
[1] N. E. Grant, P. P. Altermatt, T. Niewelt, R. Post, W. Kwapil, M. C. Schubert, J. D. Murphy, Solar RRL, 5 2000754 (2021), doi: 10.1002/solr.202000754.
[2] N. E. Grant, J. R. Scowcroft, A. I. Pointon, M. Al-Amin, P. P. Altermatt, J. D. Murphy, Solar Energy Materials & Solar Cells, 206 110299 (2020), doi: 10.1016/j.solmat.2019.110299.
[3] W. Kwapil, J. Dalke, R. Post, T. Niewelt, Solar RRL, 5 2100147 (2021), doi: 10.1002/solr.202100147.

Light- and elevated-temperature-induced degradation (LeTID) is a decrease in the minority carrier lifetime in p-type Si which occurs under simultaneous heat-exposure and carrier injection. It is well accepted that the defect structure and degradation mechanism are tied to hydrogen that is injected into the Si bulk during the firing of the Si paste to make contact to the solar cell. There are three characteristic stages of LeTID: an annealed state where defects are metastable and non-recombination-active, a degraded state where defects are recombination active, and a stabilized state where defects are permanently non-recombination-active. Even though there have been many publications on LeTID over the past 10 years, the fundamental defect structure and degradation mechanism of LeTID is largely speculative. Empirical studies relate H to the defect structure and degradation mechanism, but atomistic evidence is lacking. If the exact structure of LeTID would be discovered, mitigation strategies that are industrially feasible could be quickly studied, discovered, and implemented, allowing for increased reliability of conventional solar cells.
In this work we utilize electron paramagnetic resonance (EPR) spectroscopy and minority carrier lifetime spectroscopy to study the structure of the LeTID defect and the degradation mechanism. We report on the EPR spectra of Ga-doped Cz Si in different stages of LeTID and correlate these results to lifetime spectroscopy measurements. Firstly, we show that the light- and elevated-temperature fully degraded shows a strong Si-dangling bond EPR signature at g = 2.0055, compared to samples in the annealed state, where the Si-dangling bond EPR signature has a decreased intensity. When we regenerate the samples, the Si-dangling bond EPR signature decreases again, this time to a minimum intensity. Thus, the Si-dangling bond EPR signature correlates to the stage of LeTID the sample is in. Additionally, in the fully degraded state of LeTID, we observe hyperfine splitting EPR signatures that we attribute to H that is injected into the Si bulk during fast firing and which has been long suspected to play an integral role in the degradation and defect structure of LeTID. We confirm that the hyperfine splitting EPR signatures originate from H with isotope experiments. We observe a direct correlation between the minority carrier lifetime and the peak-to-peak intensity of the Si-dangling bond EPR signature further supporting the involvement of a Si-dangling bond in the LeTID defect.
Finally, in control samples that did not undergo a hydrogenation step, we do not observe an increase in the Si-dangling bond EPR signature as LeTID progresses, hyperfine splitting EPR signatures due to H, or a decrease in the minority carrier lifetime upon simultaneous heat and light exposure. Thus, for the first time, EPR has been used to show manipulation of the LeTID defect during simultaneous heat and light exposure and confirm at the atomistic level that H is involved in the degradation mechanism and defect structure of LeTID.

Auger recombination is an intrinsic, non-radiative recombination mechanism involving three carriers – either two electrons and a hole (eeh) or two holes and an electron (hhe). Despite silicon’s overwhelming importance as a semiconductor, the microscopic mechanisms of Auger recombination in silicon remain poorly understood. In this work, we use first principles methods to probe both direct Auger, where momentum is strictly conserved by the recombining electrons and holes, as well as indirect (phonon-assisted) Auger, which is enabled by the additional momentum provided via electron-phonon coupling. In addition to assessing the relative strength of the different Auger mechanisms, we also investigate the contribution of different electronic valleys and phonon modes to the overall Auger recombination rate. We also examine the effects of temperature and carrier density in the high-density regime on the Auger rates. The result of these efforts is a hitherto inaccessible understanding of Auger recombination at an atomic scale.

We demonstrate that phonon-assisted Auger is the dominant mechanism for both the eeh and hhe processes. Our results are in excellent agreement with experimental measurements. Furthermore, our analysis of the valley contributions to Auger reveals that that for eeh Auger, the strongest recombination occurs from electrons occupying valleys that are perpendicular to one another. This finding indicates that it may be possible to tune the Auger recombination rate in silicon via strain engineering – a boon given that Auger is an intrinsic recombination process. Furthermore, we have found that the hhe process is dominated by short wavelength phonons, while phonons across the wavelength spectrum all contribute to the eeh process. Ultimately, our computational characterization paves the way for a clearer understanding of the microscopic Auger recombination mechanisms in silicon, and points to engineering solutions that may improve the efficiency of silicon solar cells.

This work was supported as part of the Computational Materials Sciences Program funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, under Award No. DE-SC0020129. Computational resources were provided by the National Energy Research Scientific Computing (NERSC) Center, a DOE Office of Science User Facility supported under Contract No. DE-AC02–05CH11231. K.B. acknowledges the support of the U.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing Research, Department of Energy Computational Science Graduate Fellowship under Award Number DE-SC0020347.

We relate the firing temperature of hydrogen-containing dielectric films on polycrystalline silicon on silicon oxide (poly-Si/SiO_x) passivating contacts to the final passivation of solar cell test structures. By subjecting passivating contact structures with different passivating dielectric films and film stacks to different thermal treatments, we can develop improved processes for firing of high-efficiency Si solar cells. Through isotopic studies using quadrupole mass spectrometry (QMS), Fourier transform infrared spectroscopy (FTIR), and Raman spectroscopy, we determine how hydrogen content and stability within each type of film relates to final passivation quality of solar cell test structures. Because H can be a very difficult element to observe and deuterium is chemically identical to hydrogen within these systems, we will test films containing D in place of H for such studies. D gives different signals from H in FTIR and Raman spectroscopy as well as in QMS, so isotopic substitution can be used as an excellent tool to probe the H behavior within films. Thermal treatments will determine at what temperatures H evolves from the dielectric films and serve as a guide to determine the optimal firing conditions. Following firing of the passivating films, lifetime and saturation current density (J₀) values of poly-Si/SiO_x passivated contact test structures and bare Si are measured using quasi-steady-state photoconductance decay on a Sinton lifetime tester to evaluate overall performance. Si solar cells using passivating contacts are at the forefront of Si solar cell research and emerging as top performers within industrial production. Performance of passivating contact Si solar cell test structures is largely determined by a parameter known as the implied open-circuit voltage iV_oc, which directly relates to material quality within the bulk of the device and at surfaces. High iV_oc is achieved when defects within the bulk crystalline silicon (c-Si) and at interfaces are passivated, preventing them from acting as charge carrier recombination centers. One of the most common ways of passivating defects within Si solar cells is hydrogenation, injecting large amounts of H to satisfy dangling bonds in the bulk and at interfaces. At elevated temperatures, the hydrogen becomes mobile enough to find and satisfy dangling bonds. Industrial silicon solar cells are often metallized through co-firing of silver pastes at high temperatures through SiN_x. This firing results in release of H out of the SiN_x film and into the rest of the device. Though some amount of H is beneficial to passivate defects in the bulk Si and at interfaces, too much H can lead to H-induced degradation or light and elevated temperature induced degradation (LeTID) [1]. It has been shown that SiN_x, Al₂O₃, and a-Si:H passivate the interfaces of poly-Si passivating contacts through different mechanisms, and that combinations of these films can lead to differing performance [2]. Though Al₂O₃ is a stoichiometric dielectric material, SiN_x can have many different values of x depending on deposition conditions. We observe FTIR and Raman spectra over several x values to determine the bonding within them and correlate the relative Si, N, and H concentrations to the stability of H within SiN_x and the passivation of each film. Investigations into the performance of these passivating films and film stacks will lead to greater understanding of dielectrics in semiconductor devices, further improvements in passivated contact design, and greater proliferation of solar energy worldwide.

[1] H. C. Sio et al., "Light and elevated temperature induced degradation in p-type and n-type cast-grown multicrystalline and mono-like silicon," Solar Energy Materials and Solar Cells, vol. 182, pp. 98-104, 2018.
[2] T. N. Truong et al., "Hydrogenation mechanisms of poly-Si/SiOx passivating contacts by different capping layers," Solar RRL, vol. 4, no. 3, p. 1900476, 2020.

Theoretical characterization of materials using first-principles tools can provide useful information to explain experimental observations and to discover new materials. However, it is challenging to explain the silicon PV characteristics from a purely theoretical perspective because the indirect band gap of silicon requires the consideration of processes beyond direct electron-photon interactions, including phonon-assisted, charged-impurity assisted, and conductivity-induced optical absorption. In our work, we developed the first theoretical characterization tools to quantify both the direct and the various indirect processes of optical absorption in both pure and doped silicon from first principles. Our model is based on density functional theory, many-body perturbation theory and the maximally localized Wannier function approach, and has been implemented in the open-source Electron-Phonon Wannier (EPW) code.
Our model allows us to quantify the impact of free carrier absorption (FCA) as well as of different crystal polytypes on silicon-based PV. We first show that our calculated indirect absorption coefficients of cubic silicon agree well with experimental measurements and that our calculated FCA coefficients agree with experimental measurements over a broad range of carrier densities and photon wavelengths. Our theoretical characterization explains the different mechanisms dominating FCA at different photon wavelengths. Further, we show that our calculated absorption coefficients for the cubic and the recently synthesized 4H silicon predict a higher absorption coefficient for the 4H polytype in the visible and IR region.
Taking a step forward, and combined with models to calculate absorbance in textured films, we calculate the electron-hole pair induced current density, and we show that we can quantify the influence of free carriers and polytypes on the PV performance. We show that the practical impact of FCA on silicon PV is small due to the thin n-type emitter and the low carrier density in p-type base. However, the same analysis shows that the 4H silicon polytype can potentially allow up to five times thinner film compared to conventional cubic silicon and is a promising polytype for silicon-based PV applications.
This work was supported as part of the Computational Materials Sciences Program funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, under Award No. DE-SC0020129. It used resources of the National Energy Research Scientific Computing (NERSC) Center, a DOE Office of Science User Facility supported under Contract No. DE-AC02–05CH11231.

Nitrogen hyperdoping has been proposed to generate intermediate band (IB) in silicon. This enables two photon absorption of low energy photons. The enhanced IR absorption adds photogenerated carriers to the carriers generated through photon absorption via valence band-conduction band transitions. In this work we include the influence of the ubiquitous oxygen in silicon on N hyperdoping. To evaluate that effect Density Functional Theory (DFT) calculations have been performed, where the exchange and correlation functional MGGA,[1] Pseudo-Dojo pseudopotential,[2] a moderately accurate basis set, and 7x7x7 k-points for sampling the Brillouin zone have been used. To reduce the defect perturbation of the Si supercell, the latter was chosen large (3x3x3 Si unit cells, that is 216 Si atoms), which led to extensive computation. This heavy computation was unavoidable in order to have adequate atomic system that matches the concentrations of N and O species in hyperdoped Cz silicon. The symmetry reduced the k-points, for instance to 172 in the case of the substitutional nitrogen (Ns). This makes one complex per 216 Si atoms, that is a N concentration of 2.3E20/cm3, which is within the hyperdoping concentration range. The DFT calculations have shown that high concentration of N and O impurities in silicon bonded to vacancies (V), or located in interstitial sites as well as clusters of such species significantly transform Si energy band structure.
VN complexes in silicon are N atoms in substitutional sites; by occupying tetravalent vacancy sites, they are negatively charged (Ns-). The band structure has shown a higher density of bands due to degeneracy removal by the perturbation caused by Ns-. These results are analyzed in twelve VxNyOz complexes for odd numbers of N atoms (y=1,3,5,7). The negatively charged complexes drastically effect, and may totally control, the band structure of the IB material. The complexes with odd numbers of N atoms in silicon have been discussed as deep levels by Voronkov et al [3].
The DFT calculations have shown also that added oxygen atom does not fundamentally change the energy band structure. This result has been verified for a wide range of N-related complexes detected by high resolution FTIR mapping along the depth of the specimen. We used bevel polished samples to enable the depth profiling of the N-O complexes. The samples are obtained from heat-treated N doped Cz and N doped FZ silicon wafers. The identification of N-O complexes combined with the DFT calculations demonstrated that O does neither hinder the hyperdoping of silicon with N nor the formation of an IB in silicon. The fact that annealed N free Cz Si has not produced IB, and never been reported to do so, is a sign of the oxygen ineptitude to making an intermediate band in silicon. We propose that the reason for the little influence of O on the band structure is the nature of the Si-O-Si bridging bonds, which are well relaxed.
[1] K. Lejaeghere, G. Bihlmayer, T. Björkman, P. Blaha, S. Blügel, V. Blum, Reproducibility in density functional theory calculations of solids, Science 351 (6280), aad3000 (2016) doi: 10.1126/science.aad3000
[2] http://www.pseudo-dojo.org/paperlist.html
[3] G. I. Voronkova, A.V. Batunina, V.V. Voronkov, R. Falster, V.N. Golovina, A.S. Guliaeva, N. B. Tiurina, Electrical properties of nitrogen doped float-zoned silicon annealed in the range of 200 to 900C, Thin Solid Films, 2350 (2010).

The relevance of solar photovoltaic (PV) energy in the share of electricity generation is nowadays unquestionable and all the most relevant outlooks report PV as the key player in the next decades for the transition towards a zero-fossil-fuel society. When transferred into the manufacturing and deployment phase, the ambitious targets meet, among others, the challenge of accurate and reliable power rating of the PV products, for which differences of a fraction of percent mean significant investment gains or losses for the manufacturers, investors, and buyers. Thereafter it is of ever-growing importance the availability of measurement equipment (i.e. solar simulators) with the highest possible quality and rating. Among the factors that mainly affect the power rating accuracy in accredited testing and calibration laboratories are: Spatial uniformity of the total irradiance; and Spectral mismatch (between the simulator’s and AM1.5 spectrum and between the reference standard and the testing module). More recently, with the increasing wafer size and module efficiency, the effect of capacitance has also represented a significant source of uncertainty in measurements and significant effort is ongoing to detect and solve it. In addition, when addressing the accuracy of power rating in the manufacturing phase, uncertainty related to thermal effects become more relevant than in testing and calibration laboratories, requiring an accurate evaluation of the temperature coefficients with the minimum uncertainty, and accurate methods to correct for temperature deviations. In any case, the stability of the reference standard to set the total irradiance of solar simulators (i.e. “golden” or “silver” modules) is of utmost importance in the production line.
Most of the factors that cause instability of current, voltage, and fill factor (hence, of maximum power) in PV modules are well-known. Silicon modules are usually more stable than thin-film modules, which are subject to metastable behaviours. However, light, the combination of light and temperature, humidity permeation, deterioration of the contacts, and other mechanical stresses may cause an unexpected degradation in the electrical parameters of reference devices, even on silicon ones. In the testing and manufacturing practice, reference modules are generally stabilised via light soaking before use and then periodically recalibrated to ensure their long-term stability.
In the metrological practice, repeated measurements with the same procedure, same location, and same testing device, over an extended period of time, are referred to as intermediate precision conditions of measurement: this, in terms of maximum achievable precision, falls in between repeatability (which is the same as intermediate precision, but over the shortest period of time) and reproducibility (which may include different locations, operators, and measuring systems). In this work we present the results of our year-long study of stability in intermediate precision conditions of measurements of a variety of commercial PV modules, including p-type Si, n-type mono-Si, Silicon heterojunction (HJT), amorphous silicon, CIGS and CdTe. The results are analysed through significant tests (Student’s t-test and F-test), and allow to determine the conditions of stability and metastability of the testing samples, providing a benchmark of values for intermediate precision conditions that are typical of an accredited testing laboratory. The results also illustrate the typical time rate at which metastabilities occur in thin-film modules; Show that metastability is not observed in HJT; And compare the intermediate precision stability of (homojunction) crystalline silicon with HJT and with the metastabilities of thin-films, indicating at which extent it is possible to identify significant drifts in the electrical parameters of the reference modules under examination.

Silicon solar cells are the leading candidate for the bottom cell in hybrid tandem photovoltaics, which combine dissimilar materials to achieve high efficiencies at a reasonable cost. A key feature that can enable high tandem PV efficiencies is the interdigitated back contact design, which can be used to provide two back contacts to a tandem stack. If the interconnect between the top and bottom cell is conductive (as in a two-terminal tandem), this provides an additional degree of freedom, creating a three-terminal tandem (3TT). 3TTs do not require current matching at the device level, meaning that they are a very flexible design to accommodate different top cell band gaps and thicknesses without efficiency trade-offs. They also do not require lateral transport layers between the cells, as four terminal devices would. Thus, 3TTs may provide the highest device efficiency for realistic material combinations in tandems. There are also many different possible device and operational conditions for 3TTs, offering different currents and voltage outputs and different device constraints, such as designs that do not require tunnel junctions. However, the 3TTs are complex since the two (or more) cells are electrically and optically coupled, meaning that characterization and operation requires simultaneous measurements of both circuits. Additionally, stringing cells requires matching of voltages based on groups of cells, which must be designed optimally for a given device configuration. We present cell results for III-V/Si prototype 3TTs fabricated using transparent conductive adhesives, and we use III-V 3TTs to demonstrate robust cell characterization methods as well as integration of multiple cells into strings. These lessons apply to a broad set of device and material configurations, including emerging tandem cells such as perovskite/Si 3TTs.

Energy harvesting with solar cells have been the most effective way to generate portable energy in rural regions. While silicon solar cells have the potential to generate significant currents based on the solar radiation intensity, voltages generated by a single solar cell is limited. On the other hand, thermoelectric modules that convert heat into electricity generates high voltages, but the current is limited by the intrinsic properties of the thermoelectric materials. In a new approach to produce portable energy for rural Africa, silicon solar cells have been combined with thermoelectric modules to enhance power generation. The new concept was tested in a rural tribal village in Botswana under an NSF project. Communities in rural Africa burn wood for their cooking needs. Most of the heat generated is wasted to the environment. Under the new hybrid concept, thermoelectric modules generated high voltages from waste-heat while silicon solar cells generated high currents. The hybrid of the combination in a unique configuration has resulted in power outputs that are much higher than the added outputs of the individual devices. The IV characteristics and the power generation efficiencies of this hybrid devices will be presented.

Tandem solar cells consisting of perovskite (PVK) and silicon (Si) cells have already surpassed the efficiency of each single-junction solar cell, so they are expected to reach the stage of commercialization due to much higher Wp/$ compared to silicon solar cells. Recently, PVK/Si tandem cells with a photovoltaic conversion efficiency of >29% have been reported, but PVK/Si tandem solar cells still cannot absorb long-wavelength infrared rays beyond the bandgap of silicon, so there is still room for efficiency improvement. In this work, we propose a new method, which combines a PVK/Si tandem solar cell and a thermoelectric generator (TEG). Most of the energy loss of solar cells is emitted as heat, which causes considerable efficiency and stability loss of the solar cells. The TEG restrain the temperature rise of the solar cells and generate additional output power by using the temperature difference between the heated solar cells and the cold side. So far, most studies on solar cell-thermoelectric (SC-TE) hybrid systems have been applied to single-junction solar cells, including silicon (Si), copper indium gallium selenide (CIGS), dye-sensitized solar cells (DSSC), and perovskite solar cells. The sunlight in a relatively wide wavelength range is not absorbed by those solar cells and thus can be used as a driving force of the TEG. On the other hand, the tandem solar cell absorbs sunlight in a wide wavelength band and generates relatively little heat, generating insufficient heat to generate TEG. However, we fabricated a tandem SC-TE hybrid system for the first time by introducing a solar absorbing layer that absorbs more than 95% of sunlight in all wavelength bands from UV to far IR between the solar cell and TEG to compensate for insufficient thermal energy. The bottom electrode of the solar cell deposited on the full area was modified in a grid pattern to facilitate the passage of IR, which is possible to increase the temperature of the solar absorbing layer by more than 5 °C. In addition, since current matching between TEG and tandem solar cells is essential to manufacture tandem solar cells, we design an optimal parallel/series structure of TEG and an optimal number of TE legs through simulation, and we realize this design to increase the open-circuit voltage (V_oc) of the tandem solar cell without current loss. As a result, it was observed that the Voc of the tandem SC-TE hybrid system was improved by more than 100mV according to the TEG design without current loss compared to the tandem solar cell (15.4mA/cm2). With the high V_oc, It was successfully applied to water splitting with high solar-to-hydrogen conversion efficiency by utilizing the tandem SC-TE hybrid system without the help of an external power source.

BIPVs (building-integrated photovoltaics) are regarded as next-generation photovoltaic technologies that can generate electricity in urban areas with limited available land while also serving as aesthetic architectural elements. In terms of aesthetics, new functions are additionally required other than power-conversion efficiency, long-term stability, and price competitiveness of conventional solar cells, such as transparency, color tunability, and flexibility. There are two main approaches to achieving a balance between PCE and AVT (average visible transmittance).1, 2 The first approach is usage wavelength-selective active materials that absorb ultraviolet (UV) and near-infrared (NIR) radiation while transmitting light in the visible region. The second approach is control thickness of active material to decrease light absorption. Recently, Lee et. al.3 demonstrated TSC device based on opaque crystalline-Si. The devices demonstrated 6.7% PCE with 50% transparency, but the problem with rigidity and coloring remains unsolved. Kang et. al.4 demonstrated TSC device using Si microwires array. The device showed 50% transparency and flexibility, but efficiency was only 1.8%.
Here, we demonstrate a flexible, color-tunable, and large-area transparent solar cell (TSC) based on a 100 μm thick crystalline-Si wafer with a periodic square hole array embedded with PDMS. The periodic hole array was formed using a dry etching process (Bosch process). The ratio of etching gas (SF6) and deposition gas (C4F8) was optimized to reduce Si wall damage during etching. Due to that, the TSC devices exhibit PCE values of 7.38% and 5.52% at the average visible transparencies (AVT) of 45% and 60%, respectively. It should be noted that, following the formation of the periodic hole array, the flexibility of the device was dramatically enhanced, such that the minimum bending radius decreased from 100 mm to 6 mm; it further decreased to 3 mm after PDMS embedding. To explain the extreme enhancement of flexibility, finite difference time domain (FDTD) simulations were applied. The results of the FDTD simulations showed that the periodic hole array structure uniformly distributes the stress across the entire area. The highest stress appeared near hole edges, while the Si grid experienced relatively low stress. Such unequal distribution of stress helps to stop random cracks and their propagation in TSC devices under bending stress. PDMS encapsulation helped to decrease bending radius to 3 mm due to stress relief at the edges of holes and create a neutral plane. TSC device and PDMS maintained 95% of initial PCE after 1000 repeated bending with a radius of 8 mm. In addition, TSCs of various colors were obtained by dispersing the dye inside the PDMS. Blue, green, yellow semitransparent solar cells were fabricated using organic dyes. Despite parasitic absorption of the dyes, the efficiency of the colored devices was reduced only by 5% of colorless device. The PDMS-embedded TSCs demonstrate extremely high flexibility and long-term stability without significant degradation even after cycles bending deformations up to 1000 cycles and 1500 h of the standard damp heat test. We expect that the silicon-based TSCs with a square hole-array and the features of transparency, color tunability, flexibility, high efficiency, and long-term stability showed in the study will speed up the convenient application of BIPV technology.
1. Du, X. Y.; Li, X. R.; Lin, H.; Zhou, L.; Zheng, C. J.; Tao, S. L., A 2019, 7 (13), 7437-7450.
2. Gholami, H.; Rostvik, H. N.; Muller-Eie, D., Energy and Buildings 2019, 203.
3. Lee, K.; Kim, N.; Kim, K.; Um, H. D.; Jin, W.; Choi, D.; Park, J.; Park, K. J.; Lee, S.; Seo, K.. Joule 2020, 4 (1), 235-246.
4. Kang, S. B.; Kim, J. H.; Jeong, M. H.; Sanger, A.; Kim, C. U.; Kim, C. M.; Choi, K. J., Light-Science & Applications 2019, 8.

Understanding the optics of solar cells was and is an important task to improve the charge carrier generation of solar cells. A variety of solutions were developed, e.g. OPAL1 and OPAL2, RaySim, Sunrays, Wafer ray tracer from PV lighthouse, CROWM, SunSolve, Daidalos, TCAD Sentaurus from Synopsys, OPTOS. Most of them are based on e.g. the transfer matrix method (TMM) and/or raytracing. The latter is most likely as, especially, silicon-based solar cells have a front and/or rear texture. However, to the knowledge of the author, none of the mentioned solutions is freely usable or with flexible input, e.g. dielectric layers with your own optical constants n and k, or working on current computers. Thus, we present a state-of-the-art Matlab-based free software using TMM and raytracing called FRay for optical calculations of typical silicon-based solar cells. It allows the flexible input of dielectric layers in number and optical constants on silicon, the texture to be change between (random) upright/inverted pyramids with flexible angle or flat surfaces. By adding of EVA and glass layers the encapsulated solar cell can also be examined.
The output is a detailed description of wavelength-dependent reflection, absorption (separated by each layer) and transmission. Simulation results are validated in comparison with at least the Wafer ray tracer from PV lighthouse. The hope is to support Quokka 2 simulations by inserting the correct generation with FRay similarly easy and for free.

Crystalline silicon solar cells, which are widely used due to their advantages such as high reliability and eco-friendliness, have problems such as electrode corrosion and reduced efficiency when moisture permeation occurs through the back sheet (BS) at the bottom of the module. An inorganic layer like amorphous Al₂O₃ have been investigated to suppress the moisture permeation through the BS. However, the amorphous Al₂O₃ layer undergoes phase transition to the crystalline Al₂O₃, which rather enhances the moisture permeation, when it is exposed to a condition of high temperature and high humidity. In this study, a multilayered structure of stearic acid-La(OH)₃/Al₂O₃/Silane-based polymer was explored to solve those problems. Surface planarization using Silane-based polymer layers enabled the formation of a 50 nm, dense, and defect-free Al₂O₃ layer through atomic-layer-deposition. Subsequently La(OH)₃ was spray coated with the surface modification by stearic acid treatment to have hydrophobicity. This surface modification resulted in the reduced surface energy and large contact angle over 150 degree. Ca-test under 85 °C/85% RH condition showed much enhanced water vapor transmission rate of these multilayered organic/inorganic hybrid moisture barriers. Based on the results in this study, if a large-area deposition process is established and applied to an actual solar cell module unit, this hybrid multilayer structure is expected to be a new alternative to secure the long-term reliability of various solar cells and various electronic devices.

Rapid deep and cheap greenhouse emissions reductions are possible using solar PV and wind. Together, they comprise 3/4 of global net generation capacity additions, because they are cheap and getting cheaper. They must do the heavy lifting to eliminate global emissions well before 2050.
About 4 terawatts of solar and wind must be deployed each year to do the whole job of eliminating fossil fuels, with a value of ~$4 trillion per year (including storage & transmission). Overall, energy costs will fall because solar and wind are so cheap. We will eliminate greenhouse emissions, oil spills, oil related warfare, urban smog, gas fracking, open cut coal mines, power station smoke plumes, ash dumps, and other unpleasant consequences of burning fossil fuels.
The emissions profile of a country typically comprises 70-80% from energy-related activities, 10-20% from the land sector and 5-15% from other activities. Electrification of nearly all energy-related activities entails a doubling of electricity demand and allows solar & wind to eliminate most emissions.
Balancing of 50-100% solar and wind in an electricity network is straightforward using off the shelf technology from vast production runs, namely storage (pumped hydro and batteries), strong inter-regional transmission connections (to smooth out local weather and demand) and demand management in all of its myriad forms.
Decarbonized electricity from solar & wind can be used to eliminate oil from land transport (via electric vehicles) and eliminate gas from heating (via electric heat pumps and electric furnaces). These three steps typically eliminate 60-80% of the greenhouse emissions of a country. Since the required technology is off the shelf from vast production runs, rapid progress is possible.
Elimination of 60-80% of “easy” emissions by the mid-2030s buys time to develop low-cost techniques to eliminate emissions from the harder sectors: aviation, shipping, metals, chemicals and the land sector.
Australia is a global pathfinder for rapid per capita deployment of solar and wind - deploying solar and wind 3-5 times faster per capita then the EU, USA, Japan or China.
Currently, the Australian National Electricity Market derives 36% of its energy from renewables, mostly solar & wind, and is tracking towards 50% in 2025. Australia is the only non-European country in the top 10 in terms of installed solar and wind power (Watts per person). Of course, Australia's solar resource is better than in Europe, and more typical of the 3/4 of the global population who live at low latitudes. Unlike other countries with high levels of solar & wind per capita, Australia has no grid connection to neighbors, which means that it is not possible to balance the grid when there is an excess or deficit of solar & wind generation by trading with neighbors.
Australia has only a small amount of hydro (7% of generation) and is learning to cope with high levels of variable solar & wind. The key lessons so far are that (i) balancing is neither difficult nor expensive and (ii) the more solar & wind that is deployed, the lower is the wholesale spot price in the market.
So, what is required to accelerate the elimination of emissions?
- Continued development of higher efficiency and cheaper photovoltaics is of prime importance, because solar PV will be responsible for the elimination of about 60% of global emissions.
- Improved batteries for electric vehicles are needed; in particular, removal of rare battery metals such as lithium and cobalt.
- Substantial work is required to bring down the cost of electrolyzers for the production of hydrogen for the metals and chemical industry, and for synthetic jet and shipping fuel.
- Post 2050 it may well be necessary to capture carbon dioxide from the atmosphere and sequester it permanently in the deep earth in order to stabilize the climate. This requires vast amounts of solar & wind energy coupled with highly efficient and recyclable carbon capture chemicals and materials.

Vehicle integrated photovoltaics (VIPV) for passenger cars potentially offer increased autonomy, lower operating costs and lower lifetime CO₂ emissions by creating power directly on an alternative fuel vehicle through the integration of solar components in the body. For maximum impact, VIPV must be possible to integrate with multiple vehicle bodies and systems, while still being cheaply manufactured. The key challenge in design of a PV product adaptable for mass integration in different products is to allow for shape and size flexibility while keeping the manufacturing and the electrical system design simple, flexible, and robust. In order to address this market requirement, we have developed a concept of module laminate semi-fabricates building blocks that can be mass manufactured and easily adapted and integrated into various PV enabled products, including vehicles. The technology is based on conductive back contact foil which allows easy inclusion of smart electrical design to optimize performance on 3D shapes and simple, robust interconnection for reliability and stability over time. We will report on the performance of these semi-fabricates before and after final integration, including accelerated lifetime testing. Based on these results, we will evaluate the potential for VIPV to offer competitive business opportunities in the near future.

Silicon photovoltaic (PV) modules in the field typically lose performance over time due to small cumulative changes in their active materials and packaging. Some potential damage leading to underperformance can manifest in the cell at the macro- level including cell cracks, corrosion, and interconnect fatigue. Other cell-level changes can occur at the micro-level such as light-induced degradation (LID), potential-induced degradation (PID) and light & elevated temperature induced degradation (LeTID).
Investigation of large-scale photovoltaic system performance throughout the United States is being conducted within the US Department of Energy’s-sponsored PV Fleet Performance Data Initiative. This collaboration with commercial PV system owners collects and evaluates PV field performance data, and provides reports on aggregated results. Drawing on over 1700 sites across the US and over 19,000 separate PV inverters we have collected in excess of 7.2 gigawatts (GW) of performance data, representing 6-7% of the entire US installed PV capacity. A mixture of utility-scale and large commercial systems are represented, averaging 4.1 megawatts (MW) in size and 5 years in age. The vast majority of the module technologies included are silicon, including legacy Al-BSF, monocrystalline PERC, multicrystalline PERC, N-type interdigitated back contact (IBC) and silicon heterojunction (SHJ).
Initial data analysis leverages NREL’s open-source software toolkit ‘RdTools’. This automated Python-based toolkit conducts degradation rate assessment to identify continuous performance loss rates for each AC power data stream. On-site measured irradiance and temperature data is assessed for data quality prior to use in the analysis. If meteorological station data is insufficiently robust, satellite-based irradiance is substituted, which was required in a minority (<25%) of cases.
Preliminary results indicate that overall performance loss occurs at a median rate of -0.75%/year across the fleet. Some regional and climate dependence is identified which is statistically significant. In particular, warmer climate zones result in slightly faster performance loss than more temperate climates. The module technology type was not found to have a significant impact on performance loss rate in general, with mono-silicon modules degrading at the same rate as multi-silicon modules. Cadmium Telluride (CdTe) modules were also found to lose performance at roughly the same rate as Si modules. One interesting performance difference was found when the degradation analysis was restricted to look only at low-light conditions (< 600 W/m²). In the case of PERC and SHJ module types, the annual performance loss rate under low-light conditions was nearly double that of unfiltered conditions. Conversely, Al-BSF module types showed no difference between low-light and all-light conditions. And N-type IBC modules followed the opposite trend with typical degradation around -0.6 %/yr under all-sky conditions, but an annual loss rate around 0%/yr when the analysis was restricted to below 600 W/m². Further investigation into causes of this performance trend is underway, as well as whether this is a continuous effect or restricted to early-life performance, and whether it is associated with any known mechanisms such as LeTID or degradation of the passivation layer in PERC or SHJ modules.

PV production and installations are growing rapidly, with a record growth of 18% in 2020. Globally, the cumulative installed PV capacity has reached already 700 GW and the first terawatt is expected next year. This is a huge amount and yet it covers “only” 3.7% of the global electricity demand. According to the “broad electrification scenario” about 70 TW installed capacity will be needed by 2050 to cover the demand of all energy sectors including transport, buildings and industry. With 700 GW about 1% of the required capacity is installed – 99% more to go!
Manufacturing volumes in the terawatt range are postulated already for 2030 and the industry is capable to grow at very high rates and even to continuously reduce the cost. The only challenge seen for terawatt levels is the supply of rare materials like silver and indium. For one solar cell only approximately 80 mg silver are consumed (PERC). But this adds up to more than 2 000 tons per year, more than 10% of the entire annual silver supply, for the current global solar cell production of 150 GW. Even with strongly reduced silver consumption per cell, 50 mg for PERC and 100 mg for n-type and bifacial cells (ITRPV projection for 2030), 12 000-21 000 tons of silver will be required for an annual production of 1.5 terawatt. Therefore, replacement of silver by copper for solar cell metallization will be mandatory to avoid material shortage and to ensure fast growing PV production at continuously decreasing cost.
Heterojunction cells are symmetrical and intrinsically bifacial. Beside high efficiency and a simple processing sequence, with all processes at low temperature, the structure offers an important advantage for copper metallization: TCOs are excellent barriers against copper diffusion. This makes heterojunction cells resistant against copper ingress and the perfect structures for copper metallization.
Simply replacing the screen-printed silver paste by copper paste leads to very low cell efficiency because the resistivity of printed copper lines is too high. One approach for plating, already proven in the semiconductor industry, is sputtering a seed layer and use an organic mask for forming the grid by electrodeposition of copper. Using this process sequence, we could demonstrate high efficiency above 24.7% on industrial heterojunction cell precursors.
Several different processing routes have been developed by the industry and by research institutes following the aim for further cost reduction, especially by avoiding PVD equipment and replacing the thick organic mask by cheaper alternatives. We present a simple process sequence comprising screen printing a seed-grid, applying a thin mask and reinforcement of the seed-grid by electrodeposited copper. For forming the seed-grid fine lines are screen-printed with copper paste. The required high line conductivity is provided only by the subsequently electrodeposited layer. For masking either an inorganic dielectric layer, like Al₂O₃, silicon oxide or nitride, or so called self-assembling molecules (SAMs) are used. With this process sequence the same efficiency has been reached as on reference cells with standard screen-printed silver paste but using only copper paste and electrodeposited copper.
The heterojunction cell technology is seen as one of the mainstream future technologies and the recent efficiency records surpassing 26% as well as the particular suitability for copper metallization support this prediction.

The basic function of photovoltaic devices can be distinguished in two fundamental processes. First the electron and holes are generated by the incident photons. In the second step these electrons and holes have to be separated and transferred to either the n- or p-electrode of the cell. This is achieved by a layer or layer system with a high carrier selectivity placed under the p- and n-electrode. The carrier-selective layers of classical crystalline silicon solar cells are created by near-surface n- or p-type doping of the silicon absorber, resulting in diffused pn- and high-low-junctions. Silicon solar cells using this approach have achieved a maximum efficiency of 25%. A major limitation of such cells is the increased Auger-recombination due to the high doping concentration in the diffused regions. The principal idea to overcome this limitation is to decouple the contact layer system from the absorber. For other types of photovoltaic devices, such as organic solar cells, this idea is not new at all. In fact, decoupled selective contacts are required for many PV technologies since the formation of diffused pn-junctions is technically not feasible. For crystalline silicon solar cells, however, it took more than 50 years after the first diffused pn-junctions were created in the Bell labs that this approach gained significant importance. This switch to passivating contacts allowed efficiencies beyond 25%.
In principle we can distinguish four types of passivating contacts for crystalline silicon solar cells:
(i) a-Si/c-Si heterojunctions (SHJ), (ii) transparent layers with high or low work functions as metal oxides, (iii) metal-insulator silicon solar cells (MIS) and (iv) doped polycrystalline silicon on SiO_x (SIS).
In this contribution, we will focus on the last option, i.e. passivating contacts based on poly-Si/SiO_x structures also known as TOPCon (tunnel oxide passivating contacts). They have a great potential to increase the efficiency of crystalline silicon solar cells, leading to efficiencies higher than 26%. Along with a-Si/c-Si heterojunction solar cells, the TOPCon technology is considered by the photovoltaic industry as one of the main routes for the next generation of industrial solar cells. The efficiency potential of TOPCon cells and its compatibility with the state-of-the-art PERC process flow make this technology very interesting for a transfer to industrial production. Indeed, efficiencies higher than 25% have been achieved on large areas. However, it is well known that mass production of solar cells has a lot of stringent requirements. Therefore, for mass production of industrial TOPCon (i-TOPCon) cells, the evaluation of reliable and high-throughput technologies for the cell fabrication is essential. For example, the question of the best deposition technique is still open: LPCVD, PECVD, APCVD, and sputtering are viable technological options with their specific pros and cons.
This contribution will give an overview of the historical development of poly-Si/SiO_x contact structures which have started already in the 1980s and describes the current state-of-the-art in laboratory and industry. In order to demonstrate the great variety of scientific and technological research, four different research topics are addressed in more detail: (i) the superior passivation quality of TOPCon structures made it necessary to re-parametrize intrinsic recombination in silicon, (ii) the control of diffusion of dopants through the intermediate SiO_x layer is essential to optimize passivation and transport properties, (iii) single-sided deposition of the poly-Si layer would reduce process complexity for industrial TOPCon cells, and (iv) silicon-based tunnel junctions for perovskite-silicon tandem cells can be fabricated using the TOPCon technology.

Currently, the main technology for multi-crystalline silicon (mc-Si) production is directional solidification with uniform small-grain nucleation, the so-called high performance mc-Si (HPM). The presence of recombination-active defects is a key characteristic of mc-Si determining the solar cell quality. The density and location of defects in a silicon ingot depend on many factors: distribution of thermal fields at crystallization and cooling, the type and thickness of Si₃N₄ crucible protective layer, the raw material type.
The paper presents the assembly procedure of three-dimensional images of recombination-active defects in the melted ingot volume by a photoluminescence (PL) method of silicon wafers. The images registration of silicon wafer photoluminescence was made on HE-PL-02 measurement unit.
To achieve this approach the following problems have been solved:
- images sorting of the intensity of silicon wafer photoluminescence according to their location in the ingot;
- images alignment and cropping of the intensity of silicon wafer photoluminescence;
- images rotation of the intensity of silicon wafer photoluminescence in case of their incorrect arrangement in the measurement unit;
- registration of crystalline grain boundaries on the images of the photoluminescence intensity data;
- formation of a three-dimensional model of crystalline grains on the registered grain boundary on the images of the photoluminescence intensity.
The paper provides the algorithms for assembling PL images into an integrated three-dimensional ingot model and demonstrates the opportunities to study the mc-Si production technology.

All-silicon tandem solar cells have the potential to overcome the Shockley-Queisser efficiency limit of single-junction Si solar cells while maintaining sufficiently low cost of fabrication. This is attributed to the fact that the tandem structure is composed of two layers of material with different band gap energy, which allows it to absorb the different wavelengths of the solar spectrum. This results in higher open-circuited voltage and theoretically can improve the efficiency over 30% if the materials are matched properly.

In this research, the porous silicon (PSi) is chosen as the top emitter layer due to its wider band gap energy as a result from the quantum size effect and its absence of the lattice mismatch to the crystalline silicon (c-Si) bottom layer. As the etching process creates dangling bonds, the top surface needs to undergo the electrochemical passivation. The etching also converts the crystalline structure on the top surface into amorphous silicon (a-Si).

The effects of porosity and passivation on the resulting band gap and band alignment between different Si components are important determinants of cell efficiency. We computed these effects of the porosity and different degrees of passivation in relatively large-scale density functional theory (DFT) and density functional tight binding (DFTB) simulations by using the a-Si as the base cell. We found that the structure obtained with less time-consuming molecular dynamics simulations results in a similar bandgap as the DFT-optimized structure. More significant differences from the DFT structures, as well as in the band gaps, were observed with DFTB with different parameterizations. The results are discussed in the context of experimental measurements on surface chemical state as well as the optical and electronic properties.

MXenes are 2D materials that have drawn the attention of researches over the past few years, mainly, due to their outstanding electrical and optical properties. Few studies were made on their usage as contacts for solar cells. In this novel work, we invistage the performance of our zirconia-passivated PERC (Passivated Emitter and Rear Cell) solar cell when Ti3C2Tx MXene is used as a contacts in lieu of silver. The effect of annealing on its structure and the overall efficiency of the solar cell is presented in this work.

As silicon solar cells approach their theoretical maximum efficiency, recent focus is towards increasing photovoltaic (PV) module power output and reliability. Bifacial silicon modules can achieve greater power densities from additional rear side illumination when the opaque backsheet is replaced by a transparent back covering, such as glass or a transparent backsheet. Glass/glass (G/G) PV modules are more rigid due to the symmetry of the structure and less permeable to moisture and oxygen compared to conventional glass/backsheet (G/B) modules. The aim of this work is to understand the distinct degradation pathways in Si PV modules of different structures (G/G and glass/transparent backsheet (G/TB)) and encapsulation type. Historical field data from G/G Si PV modules encapsulated with poly(ethylene-co-vinyl acetate) (EVA) suggest these modules suffer from additional degradation causing more severe power losses. New-generation G/G bifacial modules are emerging with better packaging and robust mounting. Alternative encapsulants to EVA have been proposed to mitigate the degradation of G/G PV modules and maintain long-term field durability. Polyolefin elastomer (POE) is a promising candidate to replace EVA due to the lack of the degradation byproduct acetic acid. Furthermore, POEs with higher volume resistivity and lower water vapor transmission rate (WVTR) are likely to limit ionic and moisture migration through the encapsulant and reduce the risk of potential induced degradation (PID) in Si cells.
In this project, we observed different degradation modes that occurred under stress in bifacial Si p-PERC mini-modules in G/G and G/TB configurations with either EVA or POE encapsulants. Mini-modules of all 4 types were subjected to accelerated stress testing consisting of 85°C and 85% relative humidity and either zero, positive, or negative system voltage of 1000 V for 1000 hours. Module performance testing and spatially-resolved imaging revealed PID of type shunting (PID-s), polarization (PID-p), and corrosion (PID-c) depending on module configuration and applied stress. POE-based modules experienced the lowest degradation at all test conditions, with less than 5% power drop compared to a maximum of 35% for EVA-containing modules. The extent of power degradation in EVA-containing modules depends on the bias polarity, with most degradation being observed in G/G modules under damp heat with positive bias.
The aim of this work is to provide a better understanding of the chemical processes responsible for the observed degradation modes by performing detailed chemical and mechanical analyses of the cell/encapsulant and encapsulant/glass interfaces. The characterization methods include X-ray photoelectron spectroscopy (XPS) that confirms ionic migration from glass to the cell and vice-versa as well as any changes in chemical bonding at the interfaces. Preliminary results indicate presence of Na near the cell gridline in EVA-containing G/G modules, whereas lower and no Na was present in POE-containing G/G modules and pristine encapsulants respectively. The adhesion testing at the interfaces suggests that even though the interfaces weakened considerably in all module configurations, mechanical degradation through loss of adhesion was unlikely related to the power loss observed here, but can cause long-term packaging issues such as delamination. Fourier transform infrared spectroscopy (FTIR) compares chemical changes and degradation of the encapsulants. Greater changes in the polymer structure were observed in the encapsulants from G/TB modules, which can be related to enhanced moisture ingress through the backsheet compared to glass. This study will provide a better understanding of degradation pathways in bifacial Si modules depending on the module configuration and provide guidance on the choice of packaging materials for bifacial modules.

Reducing the Auger recombination in silicon solar cells is the ultimate challenge to push the efficiency towards its fundamental limit. The use of lightly doped wafers provides a way to overcome this challenge and serves as an opportunity to increase the breakdown voltage of solar cells, which can potentially redefine the way we design modules and photovoltaics systems. Commercial silicon solar cells have typically bulk resistivities below 10 Ωcm. This limitation is the result of a compromise between the doping density and the excess carrier concentration at the maximum power point. Recent improvements in the quality of the wafer bulk and the surface passivation shows a path to avert this compromise, enabling the use of high-resistivity wafers for cell and module applications. Therefore, insights into the performance of silicon solar cells and modules using such wafers in relevant field conditions are of significant interest. Studies of solar cells using wafers beyond the commercial resistivity range (>10 Ωcm), are very limited, and surprisingly, almost none of the previous studies determine the performance of cells under realistic operating conditions. In this study we experimentally evaluate how the bulk resistivity (up to >15,000 Ωcm) affects the response of silicon heterojunction solar cells to illumination conditions (0.1-1 suns) and elevated temperatures (25-80^oC), normally experienced by modules in the field. We also examine how the breakdown voltage increases with the bulk resistivity. The results indicate that cells with very high bulk resistivities, over 15,000 Ωcm, can withstand very low illuminations intensities down to 0.1 suns when compared against much lower bulk resistivity cells. The temperature coefficients measured on these cells are also comparable with values previously reported by other groups on cells using wafers with standard resistivities. Currently we are completing our temperature dependent contact resistance study for different bulk resistivities, and we plan to present those results at the conference. The cells with bulk resistivities over 1,000 Ωcm show breakdown voltages larger than -1,000 V, almost two orders of magnitude higher than in typical silicon solar cells. Although modules have bypass-diodes to prevent cells from going into breakdown, higher breakdown voltages can improve the reliability of modules in case of bypass-diode failure and reduce the module cost by easing the number of bypass-diodes required. Our simulation results on modules show that solar cells with larger breakdown voltages, in the absence of bypass diodes, allows the shaded solar cell to operate in forward-bias, even under extreme shading conditions, protecting the integrity of the cell and module. Together, these results highlight the high potential of using high-resistivity wafers to manufacture high efficiency silicon solar cells suitable to operate under relevant field conditions, and potentially enable more robust and cost-effective module designs.