Production of Upgraded Metallurgical Grade (UMG) silicon for a low-cost high-efficiency and reliable PV technology

B\"ulent Ar{\i}kan; Carlos del Ca\~nizo; David Fuertes Marr\'on; Eduardo Fornies; Fernando Ruiz; Guillermo S\'anchez Plaza; Hasan H\"useyin Canar; Jos\'e Manuel M\'iguez Novoa; Laura Mendez; Luis Jaime Caballero

arxiv: 2604.04095 · v1 · submitted 2026-04-05 · ❄️ cond-mat.mtrl-sci

Production of Upgraded Metallurgical Grade (UMG) silicon for a low-cost high-efficiency and reliable PV technology

Jos\'e Manuel M\'iguez Novoa , Volker Hoffmann , Eduardo Fornies , Laura Mendez , Marta Tojeiro , Fernando Ruiz , Manuel Funes , Carlos del Ca\~nizo

show 7 more authors

David Fuertes Marr\'on Nerea Dasilva Villanueva Luis Jaime Caballero B\"ulent Ar{\i}kan Ra\c{s}it Turan Hasan H\"useyin Canar Guillermo S\'anchez Plaza

This is my paper

Pith reviewed 2026-05-13 17:14 UTC · model grok-4.3

classification ❄️ cond-mat.mtrl-sci

keywords UMG siliconsolar cellsphotovoltaicsmulticrystalline silicongetteringPERCblack siliconlife cycle assessment

0 comments

The pith

Tailored processing turns upgraded metallurgical silicon into a direct substitute for polysilicon in high-efficiency solar cells and modules.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper shows that UMG silicon, after targeted purification to leave moderate boron and phosphorus compensation, can be turned into wafers that match conventional polysilicon performance in multicrystalline solar cells. Adding gallium at crystallization produces a uniform 1 ohm-cm resistivity profile along the ingot. Subsequent steps—optimized phosphorus diffusion gettering, black-silicon texturing that survives gettering and passivation, and standard BSF or PERC fabrication—deliver cell efficiencies in the same range as poly-Si references. Outdoor modules built from these cells show comparable output after several years, and life-cycle assessment indicates lower environmental impact. The work therefore claims UMG-Si as a lower-cost, more sustainable route to reliable high-efficiency PV without new degradation problems once a regeneration step is included.

Core claim

The central claim is that UMG-Si produced via Ferrosolar's value chain, when gallium-doped for uniform resistivity, subjected to optimized phosphorus gettering, and textured with black-silicon processes compatible with later steps, yields industrial BSF and PERC cells whose efficiencies fall within the range achieved on polysilicon substrates, with TOPCon key steps also showing promise; modules exhibit no significant light-induced or temperature-related degradation after regeneration and perform equivalently to poly-Si references after years of field operation.

What carries the argument

The full tailored manufacturing chain that converts moderately compensated UMG material into high-electronic-quality wafers through gallium doping for resistivity control, phosphorus gettering optimization, and black-silicon texturing that preserves compatibility with gettering and passivation.

If this is right

UMG PERC cells reach efficiencies comparable to poly-Si cells under standard industrial processing.
A regeneration step eliminates significant light or temperature degradation in UMG modules.
Modules built on UMG wafers maintain output parity with poly-Si modules after multiple years outdoors.
Life-cycle assessment shows reduced environmental impact for UMG-based PV across manufacturing and installation scenarios.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

If the processing sequence transfers cleanly, UMG-Si could become the default feedstock for cost-sensitive utility-scale multicrystalline arrays.
Further work on TOPCon or other passivated-contact architectures on the same material base could test whether the efficiency ceiling rises beyond current PERC levels.
Longer-term field data beyond a few years would be needed to confirm that any subtle compensated-material effects remain negligible.

Load-bearing premise

The specialized steps of gallium doping, gettering, and black-silicon texturing can be reproduced at industrial scale without introducing defects or added costs that erase the original material-price advantage.

What would settle it

Large-scale production runs in which UMG-based PERC cells fall more than 0.5 percentage points below matched polysilicon runs, or fielded modules show faster power loss than poly-Si references after the regeneration step.

read the original abstract

UMG-Si has the potential to reduce the cost of PV technology and to improve its environmental profile. In this contribution, we summarize the extensive work made in the research and development of UMG technology for PV, which has led to the demonstration of UMG-Si as a competitive alternative to polysilicon for the production of high-efficiency multicrystalline solar cells and modules. The tailoring of the processing steps along the complete Ferrosolar's UMG-Si manufacturing value chain has been addressed, commencing with the purification stage that results in a moderately compensated material due to the presence of phosphorous and boron. Gallium is added as a dopant at the crystallization stage to obtain a uniform resistivity profile 1 Ohm*cm along the ingot height. Defect engineering techniques based on phosphorus diffusion gettering have been optimized to improve the bulk electronic quality of UMG-Si wafers. Black silicon texturing, compatible with subsequent gettering and surface passivation, has been successfully implemented. Industrial-type BSF and PERC solar cells have been fabricated, achieving cell efficiencies in the range of those obtained with conventional polysilicon substrates. TOPCon solar cell processing key steps have also been tested to further evaluate the potential of the material in advanced device architectures beyond PERC. Degradation mechanisms related to light exposure and operation temperature have been shown not to be significant in UMG PERC solar cells when a regeneration step is implemented, and PV modules with several years of outdoor operation have demonstrated similar performance to reference ones based on poly-Si. LCA has been carried out to evaluate the environmental impact of UMG-based PV technology when compared to the poly-Si-based one, considering different scenarios both for the manufacturing sites and the PV installations.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

UMG-Si can reach cell efficiencies and module stability close to polysilicon after Ga doping, gettering, and black-Si steps, with a clearer LCA edge, but the paper is mostly a summary and the industrial-scale data remain thin.

read the letter

The main point is that this work shows UMG silicon, after compensation control and targeted processing, can deliver multicrystalline cells whose efficiencies sit in the same range as polysilicon references, with no major light-induced degradation once a regeneration step is added and with modules that have performed comparably in the field for several years. The LCA section also indicates lower embodied energy and emissions under the scenarios they model. That combination is useful because it ties together purification, crystallization, wafering, cell fabrication, and module testing in one narrative. The practical steps—gallium addition for flat resistivity, optimized phosphorus gettering, and black-silicon texturing that survives later processing—are described clearly enough that a process engineer could see what was changed from standard flows. The outdoor module data and the TOPCon trial runs add some weight beyond just lab cells. The soft spots are straightforward. The abstract and summary give no numerical efficiency values, no standard deviations, and no side-by-side tables, so the claim that results are “in the range” cannot be checked without the full figures. The stress-test concern about scale-up is fair: the paper does not report minority-carrier lifetimes or yields from ingots much larger than lab scale, nor does it quantify any added cost or throughput penalty from the extra gettering or texturing steps. Because the manuscript largely consolidates the group’s own prior R&D rather than presenting a single new mechanism or dataset, its novelty is incremental. This paper is aimed at PV manufacturers and materials groups already looking at alternative silicon feeds. It is coherent on its own terms and deserves a serious referee who can ask for the missing quantitative tables and a clearer discussion of how the processes translate to 200-plus kg ingots. I would send it to review rather than desk-reject, with the expectation that the authors supply the detailed metrics the abstract lacks.

Referee Report

3 major / 2 minor

Summary. The manuscript summarizes R&D on Ferrosolar's UMG-Si value chain for PV. It covers purification yielding compensated B+P material, Ga doping at crystallization to achieve uniform 1 Ω·cm resistivity along the ingot, optimized P-diffusion gettering for bulk quality, black-Si texturing compatible with gettering/passivation, fabrication of industrial BSF/PERC/TOPCon cells, light/temperature degradation studies with regeneration, multi-year outdoor module data, and LCA comparisons versus poly-Si.

Significance. If substantiated with quantitative data, the work would establish UMG-Si as a cost- and environment-competitive feedstock for high-efficiency multicrystalline cells, directly addressing PV cost and sustainability goals. The experimental integration of Ga doping, gettering, and black-Si on compensated material, plus module-level validation, represents a practical advance over prior UMG efforts. However, the absence of numerical efficiency values, spreads, and side-by-side metrics limits the ability to confirm the 'in the range' claim or the scalability of the tailored steps.

major comments (3)

[Abstract; solar-cell fabrication results] Abstract and cell-fabrication section: the central claim that 'cell efficiencies in the range of those obtained with conventional polysilicon substrates' is unsupported by any numerical values, standard deviations, sample sizes, or direct comparison tables. This omission is load-bearing for the competitiveness assertion and prevents quantitative assessment of whether UMG cells truly match poly-Si performance.
[Manufacturing value chain; conclusions] Scalability discussion and results: the competitiveness claim rests on lab-scale tailoring of Ga doping, gettering, and black-Si texturing, yet no data are supplied on minority-carrier lifetime, wafer yield, or added cost at industrial ingot sizes (>200 kg) or throughputs. If compensation-induced defects or extra processing time appear at scale, the UMG cost/environmental advantage disappears; this gap must be addressed with explicit metrics or pilot-line results.
[Degradation mechanisms; outdoor module performance] Degradation and module sections: statements that degradation 'is not significant' and that outdoor modules show 'similar performance' lack quantitative efficiency-loss percentages, time/temperature dependencies, or statistical comparison to poly-Si references. These are required to support the reliability claim.

minor comments (2)

[Abstract; crystallization section] Resistivity is written as '1 Ohm*cm'; adopt standard notation '1 Ω·cm' throughout.
[LCA section] LCA scenarios are mentioned but the specific assumptions (electricity mix, site locations, functional unit) are not tabulated; add a summary table for reproducibility.

Simulated Author's Rebuttal

3 responses · 1 unresolved

We thank the referee for the detailed and constructive review of our manuscript on UMG-Si for PV applications. We have prepared point-by-point responses to the major comments below. Where possible, we will incorporate additional quantitative details in a revised version to strengthen the presentation of our results.

read point-by-point responses

Referee: [Abstract; solar-cell fabrication results] Abstract and cell-fabrication section: the central claim that 'cell efficiencies in the range of those obtained with conventional polysilicon substrates' is unsupported by any numerical values, standard deviations, sample sizes, or direct comparison tables. This omission is load-bearing for the competitiveness assertion and prevents quantitative assessment of whether UMG cells truly match poly-Si performance.

Authors: We agree that explicit numerical values would improve clarity and allow direct quantitative comparison. Although efficiency data appear in the results section and associated figures, the revised manuscript will include a new summary table listing average cell efficiencies (e.g., for BSF, PERC, and TOPCon architectures), standard deviations, sample sizes, and side-by-side values for reference polysilicon cells processed in parallel. This will directly support the competitiveness claim with concrete metrics. revision: yes
Referee: [Manufacturing value chain; conclusions] Scalability discussion and results: the competitiveness claim rests on lab-scale tailoring of Ga doping, gettering, and black-Si texturing, yet no data are supplied on minority-carrier lifetime, wafer yield, or added cost at industrial ingot sizes (>200 kg) or throughputs. If compensation-induced defects or extra processing time appear at scale, the UMG cost/environmental advantage disappears; this gap must be addressed with explicit metrics or pilot-line results.

Authors: The ingots and processes described are from Ferrosolar's industrial production (>200 kg ingots), with gettering and texturing steps already integrated into standard manufacturing flows. Minority-carrier lifetime improvements after gettering are shown via maps and average values in the manuscript. However, detailed wafer yield statistics and precise incremental cost figures at full commercial throughput involve proprietary process data that cannot be disclosed. We will expand the discussion to emphasize the industrial-scale context of the presented results and the minimal added processing burden, but we cannot supply additional pilot-line yield or cost metrics beyond what is already reported. revision: partial
Referee: [Degradation mechanisms; outdoor module performance] Degradation and module sections: statements that degradation 'is not significant' and that outdoor modules show 'similar performance' lack quantitative efficiency-loss percentages, time/temperature dependencies, or statistical comparison to poly-Si references. These are required to support the reliability claim.

Authors: We acknowledge the need for quantitative support. The revised manuscript will add explicit degradation metrics, including efficiency loss percentages after light exposure and regeneration (typically <1% relative), temperature-dependent behavior from accelerated testing, and statistical outdoor module performance data (e.g., power output retention over multiple years) with direct comparison to poly-Si reference modules. This will substantiate the reliability statements with concrete numbers and analysis. revision: yes

standing simulated objections not resolved

Detailed wafer yield percentages and added cost at full industrial throughputs (>200 kg ingots), as these involve proprietary information beyond the scope of the current R&D summary.

Circularity Check

0 steps flagged

No circularity: experimental results and LCA rest on direct measurements

full rationale

The paper reports fabrication outcomes, cell efficiencies, degradation tests, outdoor module performance, and LCA comparisons for UMG-Si versus poly-Si. No equations, fitted parameters, predictions derived from inputs, or load-bearing self-citations appear in the provided text. All central claims (e.g., efficiencies 'in the range' of polysilicon, non-significant degradation with regeneration) are grounded in measured data rather than any derivation that reduces to its own definitions or prior author work by construction. This is a standard experimental demonstration paper with self-contained evidence.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard silicon PV processing assumptions and experimental tuning of known steps; no new physical entities or first-principles derivations are introduced.

free parameters (2)

Gallium doping level
Chosen to produce uniform 1 Ohm*cm resistivity profile along ingot height
Gettering process parameters
Optimized specifically for compensated UMG-Si material

axioms (2)

domain assumption Phosphorus diffusion gettering improves bulk electronic quality in compensated silicon
Invoked as the basis for defect engineering step
domain assumption Black silicon texturing remains compatible with subsequent gettering and surface passivation
Stated as successfully implemented without further qualification

pith-pipeline@v0.9.0 · 5697 in / 1341 out tokens · 37049 ms · 2026-05-13T17:14:49.914029+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

2 extracted references · 2 canonical work pages

[1]

with permission from EUPVSEC. The trends for the normalized values in all EF categories for PERC architectures considering ES and CH mixes are summarized in Figure 17, revealing the strong influence of the manufacturing location over all the environmental categories considered. The predominant presence of fossil fuels, specifically coal, in the Chinese mi...

work page doi:10.13039/501100011033/ 2019
[2]

Frischknecht, R

Available at: http://dx.doi.org/10.4229/35thEUPVSEC20182018-2AV.1.5. Frischknecht, R. (2021) Environmental life cycle assessment of electricity from PV systems. Fact sheet. IEA-PVPS. Available at: https://iea-pvps.org/fact-sheets/factsheet-environmental-life- cycle-assessment-of-electricity-from-pv-systems/. Guerra, M.R., De La Parra Laita, I., Solano, M....

work page doi:10.4229/35theupvsec20182018-2av.1.5 2021

[1] [1]

with permission from EUPVSEC. The trends for the normalized values in all EF categories for PERC architectures considering ES and CH mixes are summarized in Figure 17, revealing the strong influence of the manufacturing location over all the environmental categories considered. The predominant presence of fossil fuels, specifically coal, in the Chinese mi...

work page doi:10.13039/501100011033/ 2019

[2] [2]

Frischknecht, R

Available at: http://dx.doi.org/10.4229/35thEUPVSEC20182018-2AV.1.5. Frischknecht, R. (2021) Environmental life cycle assessment of electricity from PV systems. Fact sheet. IEA-PVPS. Available at: https://iea-pvps.org/fact-sheets/factsheet-environmental-life- cycle-assessment-of-electricity-from-pv-systems/. Guerra, M.R., De La Parra Laita, I., Solano, M....

work page doi:10.4229/35theupvsec20182018-2av.1.5 2021