The warranty says 0.25% a year. The desert measured 8.7% in three.
The heterojunction-versus-TOPCon story is usually told as an architecture race: efficiency records, the 2030 production mix, the tandem endgame. We've mapped that elsewhere. This is the narrower finding sitting underneath it, and it's the one that should change how a specifier reads a datasheet: the independent field and lab record shows both n-type families degrading in ways their warranties don't capture. The variable that predicts it is bill-of-materials and manufacturing discipline, not the passivation chemistry the marketing argues about.
A desert deployment found heterojunction modules degrading an order of magnitude faster than their warranty curve.
A three-year field study at Hamad Bin Khalifa University in Doha, peer-reviewed in Solar Energy (Kivambe et al., 2025), put PERC, TOPCon and heterojunction (HJT) bifacial modules on the same rack in one of the harshest module environments on earth. One HJT module lost 8.73% of its output over three years; a second lost 6.50%. The datasheet these modules ship against warrants degradation of roughly 0.25% a year, about 0.75% across the same three years. The best TOPCon module in the study lost 0.14%, the lowest of anything tested; another TOPCon module was among the worst. The failure mode on the HJT units was edge-seal delamination letting moisture reach the amorphous-silicon and transparent-conductive-oxide layers: a packaging and bill-of-materials failure, not a cell-physics one.
An operator field report from China's State Power Investment Corporation at its Daqing site points the same direction, HJT degrading faster than TOPCon, but that one is a self-published corporate press release, not peer-reviewed, so it corroborates rather than proves. The peer-reviewed desert data is the load-bearing evidence, and it says the same thing: the warranty curve and the field curve are not the same curve.
The lab agrees, on both familiesTOPCon's turn: UV and contamination the linear warranty never modelled.
This is not a heterojunction problem with a TOPCon halo. A Fraunhofer ISE comparative study in Progress in Photovoltaics (2025) stress-tested twenty commercial TOPCon module types and found that while TOPCon genuinely resists the light-induced degradation that plagued PERC, 40% of them lost more than 5% of their power after a UV dose equivalent to about one year in a moderate climate. A 30-year linear warranty simply doesn't describe that mode. Separately, a UNSW study in Solar Energy Materials and Solar Cells (2025) found both HJT and TOPCon (but not PERC) vulnerable to trace contaminants left on the cell during handling, producing 10–16% power loss in damp-heat testing: the HJT cells failed through voltage and recombination loss, the TOPCon cells through a 41% rise in series resistance as the contacts corroded. Two chemistries, two failure paths, one shared cause: what happened on the line and in the lamination, not what the band diagram predicts.
Why the bill of materials is about to move againA silver shock is rewriting the cost case, and it favours the low-temperature architecture.
The reliability story is a manufacturing story, and manufacturing is being reshaped right now by silver. Silver paste is up to 30% of a TOPCon cell's production cost and 16–17% of total module cost, and the silver price ran from about US$28/oz on average in 2024 to a record above US$120/oz in January 2026. That turned de-silverisation from a roadmap line into a survival one. The two architectures diverge here because of the same low-versus-high-temperature split that governs everything else about them. HJT's low-temperature process is compatible with silver-coated copper paste and direct copper electroplating (the copper route Australia's SunDrive is scaling); TOPCon's high-temperature architecture makes copper harder to use, because it oxidises at those temperatures. Per CPIA's 2025 data, HJT has already cut cell silver to 75 mg, the lowest of the mainstream n-type technologies, against TOPCon's 86 mg and back-contact's 135 mg.
None of this makes heterojunction "the winner." TOPCon's manufacturing-economics lead is real and compounding, and we've said so. It means the bill of materials this whole reliability question turns on is itself in motion, and the direction of that motion, for once, runs slightly in HJT's favour.
What the warranties actually promiseThe trade-off buyers are really choosing between.
Set against field data like Doha's, the warranty structures are worth reading as what they are: promises with different shapes, not measurements of what a module will do.
| Product | Performance term | Annual degradation | End-of-term output |
|---|---|---|---|
| REC Alpha Pure-RX (HJT) | 25 years | 0.25%/yr | 92% at year 25 |
| Trina Vertex S+ (TOPCon) | 30 years | 0.4%/yr | 87.4% at year 30 |
| JinkoSolar Tiger Neo (TOPCon) | 30 years | 0.4%/yr | 87.4% at year 30 |
Manufacturer datasheets, 2025–26 revisions. REC's HJT warranty is the more generous on the annual rate and the year-25 floor; the TOPCon makers extend the term further at a lower guaranteed end-of-term output.
Dietreich's readBuy the bill of materials, not the acronym.
I've spent enough years watching panels come off roofs to distrust any argument that turns on cell physics alone, and this is a clean example of why. The passivation chemistry is real engineering, and heterojunction's structural edges in bifaciality and temperature coefficient are genuine. But the two most rigorous independent field studies available found HJT degrading faster than TOPCon in actual deployment, in direct tension with both technologies' warranty claims; the lab work found TOPCon carrying UV and corrosion modes its 30-year linear warranty never described; and the failures, on both sides, traced to encapsulants, edge seals, contact metallisation and handling cleanliness: the bill of materials, not the band diagram. My read is that "HJT versus TOPCon" is the wrong question for a buyer. The right one is narrower and harder to market: who built this specific module, to what bill of materials, and does the warranty behind it belong to a company that will still be standing to honour it? That last clause is the one the datasheet never prints, and it's the one this benchmark keeps in view.
The heterojunction and TOPCon architecture race is mapped in our recombination-chain and 2030-mix pieces.
Sources11 references
- Kivambe et al. (HBKU): Comprehensive assessment of performance and reliability of PERC, TOPCon and SHJ modules in desert climates, Solar Energy (2025)
- pv magazine: Three-year field test shows TOPCon, HJT module reliability in desert areas (May 2025)
- PV Tech: Fraunhofer ISE study (Progress in Photovoltaics) uncovers critical degradation across 20 commercial TOPCon module types
- UNSW SPREE: Hidden traces: how solar cell handling drives damp-heat failures in HJT and TOPCon modules, Solar Energy Materials and Solar Cells (2025)
- Kiwa PVEL: 2025 PV Module Reliability Scorecard insights
- pv magazine / OPIS: CPIA 2025 per-cell silver consumption; module prices lagging soaring silver costs (Jan 2026)
- Investing News Network: silver's all-time high, US$121.62/oz (29 Jan 2026)
- SPIC / pv magazine press release: Daqing demonstration-base field data (operator-published, uncorroborated)
- REC: Alpha Pure-RX datasheet
- Trina Solar: Vertex S+ NEG9R.25 datasheet
- JinkoSolar: Tiger Neo 72HL4-BDV datasheet
Read our ownership, funding and editorial disclosures.
Stay across the benchmark.
New scorecards, score changes and market signals, published as they happen.