Low-Light Performance Showdown: TOPCon, BC, and HJT Backed by Real-World Data
Introduction
Nameplate power is a rated value; low-light response is real-world performance. Across most regions of the world, irradiance stays below 1000 W/m² for over 90% of the time. Only two or three hours around solar noon come close to STC conditions. Sunrise, sunset, overcast skies, rain—cells spend most of their working life under low light. A high rated efficiency does not guarantee high real-world output. Today we break down low-light response: who wins on physics, who proves stronger in the field, and how to judge a cell's low-light quality right on the production line.
The Physics of Low-Light Response: Who Leaks and Recombines Less
From the diode equivalent circuit, the root cause of efficiency drop under low light is simple: the photogenerated current shrinks, but leakage and recombination do not shrink proportionally, so their relative share grows.
The most critical factor: shunt resistance Rsh
Under low light the photogenerated current drops sharply, but the leakage current stays roughly constant (it depends on voltage and Rsh). A larger share of leakage current pulls Voc down, which drags FF down, which lowers efficiency.
The higher the Rsh (the smaller the leakage), the better the low-light response. This is the core physical factor.
| Cell Type | Rsh Characteristics | Low-Light Performance |
|---|---|---|
| HJT | i-a-Si:H passivation layer with excellent insulation, extremely low interface recombination | Best |
| TOPCon | Positive and negative poles split across front and back, few edge isolation zones, controllable leakage paths | Good |
| BC | Rear interdigitated structure, many P⁺/N⁺ isolation trenches, increased edge leakage risk | Weaker |
Secondary factor: ideality factor n
The ideality factor reflects the recombination mechanism: n=1 for ideal diffusion current, n=2 when depletion-region recombination dominates. The larger n is, the heavier the recombination loss under low light. TOPCon's passivated contact structure gives n≈1.1-1.2, BC's rear interdigitated PN junction has more interface recombination channels at n≈1.2-1.4, and HJT's amorphous-silicon passivation excels at n≈1.0-1.1.
Series resistance Rs matters less here. Power loss across Rs is I²R; under low light the current is small, so its relative impact weakens.
Why BC Is Weaker Under Low Light: A Structural Reason
BC places both positive and negative electrodes on the rear, requiring numerous isolation trenches between the P⁺ and N⁺ regions to achieve electrical separation. These trenches bring two problems:
Edge leakage risk: Trench etching can damage the silicon substrate and form leakage paths. A single BC rear surface holds hundreds of isolation trenches, each a potential leakage route.
Interface recombination: The P⁺/N⁺ interface area of the rear interdigitated structure grows larger, adding recombination centers and pushing the ideality factor n higher.
This is an inherent structural challenge, not a question of "who did it badly." Process optimization (controlling trench morphology, improving passivation layers) can help, but the structure puts BC at a natural disadvantage on this point.
The reason HJT performs best under low light is the opposite: the intrinsic amorphous-silicon i-a-Si:H passivation layer delivers outstanding surface passivation, low interface state density, the highest Rsh, and the smallest ideality factor.
Field Evidence: TOPCon Beats BC in Per-Watt Output Under Low Light
The field data from several test institutes points in a consistent direction:
| Test Institute | Location | Scenario | TOPCon vs BC Low-Light Gain |
|---|---|---|---|
| CPVT | Yinchuan, Ningxia | Morning/evening low-light periods | Overcast +3.89%, sunny +2.33% |
| CPVT | Yinchuan, Ningxia | Extreme low irradiance (0-100 W/m²) | +4.38% |
| TÜV Nord | Kagoshima, Japan | <400 W/m² | +10.79% |
| TÜV Rheinland | Chengdu | 90% overcast/rainy days | +2.37%, morning/evening peak +7.18% |
| CGC | Hainan | 127 days including 76 rainy days | +7.83% |
| State Grid | Zhangbei | 200 W/m² | +2.6% |
Under low-light conditions, TOPCon's per-watt output exceeds that of BC, and the lower the irradiance, the wider the gap.
But variation within the same technology route is also large. Multi-supplier comparison testing by Carbon Search Evaluation Lab shows BC products losing 2.78% to 6.57% at 200 W/m² low irradiance, while TOPCon ranges from 2.14% to 4.72%. The gap between the "best products" of the three technologies is smaller than the gap between "good products vs. poor products" within the same route.
Production takeaway: when selecting, a manufacturer's process level matters as much as the choice of technology route.
Don't Confuse Temperature Coefficient With Low-Light Response
Temperature coefficient and low-light response are two independent parameters, but they are easily mixed up.
| Parameter | Relevant Scenario | HJT | TOPCon | BC |
|---|---|---|---|---|
| Temperature coefficient | High-temperature scenarios (module >50°C) | -0.24%/℃ | -0.29%/℃ | -0.26%/℃ |
| Low-light response | Low-irradiance scenarios (<400 W/m²) | Best | Good | Weaker |
On a hot, overcast summer day, high temperature and low light stack together, and HJT leads on both, compounding its advantage. On a cold, overcast winter day, low temperature reduces the influence of the temperature coefficient, and low-light response takes the lead. Don't use the temperature coefficient to explain low-light performance, and don't infer the temperature coefficient from low-light performance—they are two distinct physical quantities.
Low-light optimization and UVID resistance are not inherently physically mutually exclusive either. Low light depends on electrical loss mechanisms (Rsh, n), while UVID depends on material stability (passivation-layer chemical bonds, encapsulant film). The two can be improved separately through independent optimization.
How to Judge a Cell's Low-Light Quality on the Production Line
The most direct indicator: shunt resistance Rsh.
In I-V testing, the higher a cell's Rsh, the more likely it performs well under low light. If a batch shows a wide Rsh distribution with a high proportion of low-Rsh cells, low-light output will surely suffer.
Special note for BC lines: cells showing abnormal bright spots in the isolation-trench regions on EL images are likely to have low Rsh. This corresponds to the "trench edge leakage" mentioned earlier—a problem the structure is naturally prone to.
TOPCon lines: Rsh above 1000 Ω·cm² is generally normal; below 500 calls for investigating edge isolation or pinholes in the passivation layer. Cells with excellent low-light behavior usually show Rsh above 3000.
HJT lines: Rsh is naturally high, and above 5000 is common. But a low Rsh on an HJT cell usually means something has gone wrong at the TCO and a-Si:H interface.
Summary
The physics ledger of low-light response: HJT is best, TOPCon is good, BC faces structural challenges. The field ledger: under low light, TOPCon's per-watt output really does exceed BC's, and the lower the irradiance, the wider the gap. But don't judge by technology route alone—the gap between good and poor products on the same route is even larger than the gap between routes.
Data sources: CPVT Yinchuan field test (2025), TÜV Nord Kagoshima field test, TÜV Rheinland Chengdu field test, CGC Hainan field test, State Grid Zhangbei field test, Carbon Search Evaluation Lab multi-supplier comparison testing (2025).
Ooitech's view: Real-world low-light output, not nameplate efficiency, is the true measure of a solar cell, and shunt resistance is the single factor that decides it most.
