Bifacial Solar Panels on Commercial Roofs: When the Rear Side Pays
Updated 23 October 2026 · SEO Dons Editorial
What does the bifaciality factor on a datasheet mean?
Bifaciality is the rear side's output as a share of the front side's, each measured separately. Canadian Solar defines it on its CS6.3-66TB sheet as rear-side maximum power divided by front-side maximum power, both under standard test conditions. The three sheets reviewed quote 80%, 85% and 85%, each with a ±5% tolerance. IEA PVPS Task 13 describes bifacial cells with rear-side efficiency from above 60% to over 90% of the front side's.
That is a ceiling, not an uplift. The back would deliver 85% of the front's power only if it received the same light, and a roof gives it a fraction. Canadian Solar's gain table lists a 645 W module at 677 W, 710 W and 774 W for 5%, 10% and 20% gain, and says gain depends on mounting structure, height, tilt angle and ground albedo. It is a menu of scenarios, not a forecast for your roof. The other datasheet lines are explained on the commercial PV panels page.
Which watts is the quote selling?
Trina's TSM-NEG19RC.20 sheet lists a 625 W module at standard test conditions and 692 W under BNPI, which adds 135 W/m² of rear light to 1,000 W/m² on the front. That is 10.7% more (692 divided by 625). MCS 005 issue 4.0 says bifacial power should be determined at BNPI conditions, and compliance has been mandatory for new MCS product certification since 1 August 2026. A quote built on BNPI watts is not comparable with one built on standard watts, so ask which was used. The kWp, kW and kWh guide explains the unit.
How much does the roof surface add?
Albedo is the share of light a surface reflects. The IEA report calls it one of the most important site characteristics for bifacial performance, and says gain rises with albedo, diffuse fraction, array height, row spacing and the space between modules.
Berkeley Lab's Heat Island Group puts a clean white roof at 80% reflection of sunlight and a gray one at 20%. Sandia's modelling collaborative lists dry asphalt at 0.09 to 0.15, concrete at 0.25 to 0.35, and galvanised steel at 0.35 when new but 0.08 when very dirty. So ask for a membrane's aged reflectance, not its new value.
For scale, the IEA reports gains below 10% for single-axis trackers over natural ground with an albedo of 0.2 to 0.3. That is a ground-mount benchmark, not a roof.
Why do height and spacing matter as much as the surface?
Rear light has to arrive. The IEA report describes a close-mount rooftop as having little to no rear-side irradiance, which is the position of panels clipped to a metal roof. Raising the array helps, but rear light is uneven, causing mismatch within modules and strings. The US laboratory formerly called NREL, renamed the National Laboratory of the Rockies on 1 December 2025, modelled rooftop systems over high-albedo reflective roofs. It found mismatch losses of up to 2% a year at 0.15 m mounting height, against under 0.5% for high-clearance systems.
On a roof of fixed size, compare energy per square metre of roof, not gain per module. In the IEA's area-constrained example, raising tilt increases row spacing and lowers the module count. The flat roof mounting guide covers spacing and ballast.
Three further cautions apply. Bifacial arrays run at higher DC current, so wire, fuse and inverter sizing may need adjusting. Edge effects dominate small research rigs, so their gains should not be expected on larger arrays. And models disagree: in an IEA comparison with 13 research and commercial entities, predicted gains varied by as much as a factor of two. Ask for any rear-gain claim to be run at the roof's measured albedo and again at a low case such as 0.2.
What did the UK vertical bifacial study find?
The UK study most often cited is Badran and Dhimish's Scientific Reports paper (August 2024). University of York researchers mounted 36 series-connected heterojunction units (1.5 kW) vertically on the Physics Tower roof, over white gravel with 50 cm between rows. They monitored them from February to December 2023 against a vertical monofacial array and a monofacial polycrystalline array tilted at 45 degrees, each also rated at 1.5 kW.
It reports 26.91% (early morning) and 22.88% (late afternoon) more power than the tilted array, and seasonal average gains over it from 14.77% in summer to 24.52% in winter. Against the vertical monofacial array the seasonal averages ran from 8.13% to 12.45%. Read it as research, not a quote for a warehouse roof:
- It is one site, one year and a 1.5 kW rig.
- The tilted array used polycrystalline cells and the bifacial array heterojunction cells, so part of the gap may be cell type.
- The arrays match in rated power, not roof area, so the paper cannot say how many kW a roof holds.
- The paper puts commercial peak demand in the late morning and early afternoon, while its largest gains came at the start and end of the day.
The IEA adds that east-west vertical arrays offer a wider generation period that better matches load profiles, and very low soiling rates. Whether that suits a building depends on its half-hourly demand, which the feasibility study guide covers.
Does extra weight or the fire class change the roof calculation?
Most bifacial cells end up in double-glass modules or modules with a transparent polymer backsheet, according to the IEA, and double glass is heavier. The glass-glass versus glass-backsheet comparison sets out weights and failure evidence. RC62 says ballasted arrays can significantly increase roof weight loading and asks for a structural survey by a competent person, so the module is one line in that load. On fire, the Jinko and Canadian Solar bifacial sheets both show IEC fire Class C, while RC62 says research is in process to develop a UK fire test standard for PV modules. A datasheet class is evidence about the module, not a ruling on your roof; the RC62 guide covers the code.
When does the rear side pay on a commercial roof?
Use the table as a list of questions for a designer. None of the sources reviewed gives a measured figure for a UK commercial roof.
| Roof and layout | Rear light | What the sources suggest | Question to ask |
|---|---|---|---|
| Clipped close to a metal roof | Little to none | IEA: close-mount rooftops have little to no rear-side irradiance | What rear gain is assumed, and why? |
| Low-tilt ballasted frames over dark membrane | Low albedo, low clearance | Asphalt 0.09 to 0.15 (Sandia); mismatch up to 2% a year at 0.15 m (NLR) | Which albedo and mismatch loss are in the model? |
| Raised frames, wide pitch, clean white membrane | Highest of the four | Clean white roof 80% (Berkeley Lab); gain rises with height and spacing (IEA) | What is the aged reflectance? |
| East-west vertical rows | Both faces all day | Wider generation period (IEA); one 1.5 kW York site | How many kW fit per square metre of this roof? |
A reader holding a quote with a bifacial line can send an enquiry through this site's form. Cell technology behind the front-side figures is compared in the TOPCon, HJT and PERC guide.
Sources
- IEA PVPS Task 13, Bifacial PV Modules and Systems: executive summary and full report, April 2021, checked 22 October 2026.
- Sandia National Laboratories, Albedo; Berkeley Lab Heat Island Group, Cool Roofs, checked 22 October 2026.
- Badran and Dhimish, Scientific Reports, vertical bifacial PV: a UK case study, 8 August 2024, checked 22 October 2026.
- National Laboratory of the Rockies (formerly NREL), bifacial mismatch loss paper, 2019, and laboratory history, checked 22 October 2026.
- MCS, MCS 005 issue 4.0, 1 August 2024, checked 22 October 2026.
- Datasheets, each for that product only: Canadian Solar CS6.3-66TB v1.11, June 2026; Jinko JKM650-670N-66QL6-BDV; Trina TSM-NEG19RC.20, TSM_EN_2025_D, checked 22 October 2026.
- RISCAuthority and FPA with MCS and Solar Energy UK, RC62, version 2, February 2023, checked 22 October 2026.
Common questions
Do bifacial panels still make sense on a grey or dark roof?
They can, but the rear side adds little. Sandia lists dry asphalt at 0.09 to 0.15, and the IEA reports gains below 10% over ordinary ground. Ask for the monofacial option's yield alongside the bifacial one.
Is the 20% bifacial gain on some datasheets realistic?
It is the top row of a scenario table, not a measured result. Canadian Solar's sheet lists 5%, 10% and 20% and says gain depends on mounting and ground albedo. Treat 20% as an upper case that needs a bright surface and open geometry.
Do bifacial modules need different inverters or cabling?
Not a different type, but the sizing may change. The IEA summary says bifacial arrays operate at higher DC current levels, so wire, fuse and inverter sizing may need adjusting. Ask the designer to show the currents used.
Are vertical bifacial layouts easier to keep clean?
The IEA lists very low soiling rates among the advantages of east-west vertical arrays, and the York paper says vertical systems are less prone to accumulating dirt and debris. Neither is a measurement on a UK commercial roof.