Can Higher Rear Irradiance Always Deliver Higher Bifacial Module Power

No. More rear irradiance usually helps a bifacial module produce more power, but the increase is rarely one-to-one. A rear sensor may show a much higher W/m² value while the actual power gain stays modest. The reason is simple: the cells also have to deal with temperature, uneven light, shadows, resistance and changes on the front side.

If front irradiance and cell temperature stay about the same, extra rear light should raise current and usually Pmax. That is the cleanest case. Outdoor systems are rarely that neat.

Start With the Numbers

Take a module with 1,000 W/m² on the front and 80% Isc bifaciality.

At 200 W/m² rear irradiance:

1,000 + (200 × 0.80) = 1,160 W/m²

At 300 W/m²:

1,000 + (300 × 0.80) = 1,240 W/m²

Those 1,160 and 1,240 W/m² figures are adjusted irradiance values. They are not a promise that Pmax will rise by exactly 16% or 24%.

IEC TS 60904-1-2:2024+AMD1:2026 uses Isc bifaciality when equivalent irradiance is calculated for bifacial I-V testing.^[1]

Rear Irradiance and Bifaciality

Rear irradiance tells you how much light reaches the back of the module. Bifaciality tells you how strongly the cell responds to rear light compared with front light.

The two numbers have to be read together.

With the same 200 W/m² on the rear, moving from 70% to 90% bifaciality changes the simple rear contribution from 140 to 180 W/m².

So 200 W/m² of rear irradiance is not the same thing as a 20% power gain.

The datasheet for Tongwei's TNC-G12R bifacial cell lists front and rear Pmpp, Impp, Voc, Isc and fill factor separately. That is useful because rear performance cannot be described by one percentage alone.

Tongwei also lists a mass-production bifaciality of 90±5% for the TWMNH-66QD module. This is a device specification. The actual energy gain still depends on how much rear light reaches the module after installation.

Current Changes More Than Voltage

Extra light mainly raises current.

Tongwei's TNC-G12R light-intensity data make that easy to see:

When irradiance falls from 1,000 to 400 W/m², normalized Isc falls to 0.403. Voc is still 0.962. Current follows irradiance much more closely than voltage does.

Module power is commonly described by:

Pmax = Voc × Isc × Fill Factor

Pmax therefore depends on more than current. Temperature, resistance and mismatch can change voltage and fill factor at the same time.

Resistance becomes more noticeable as current rises because:

Resistive loss ∝ I²R

If current goes from 10 A to 11 A, the I² term goes from 100 to 121. Current rose 10%, while the I² term increased 21%.

Average Rear Irradiance Can Hide a Problem

A single average can look fine even when some cells receive far less light than others.

Case B has the higher average, but its brightest area is 320 W/m² and its darkest is only 60 W/m². The gap is 260 W/m², and the brightest value is about 5.3 times the darkest.

Cells in the same series path have to carry the same current. A weakly lit section can therefore hold back cells that are receiving much more rear light.

In NREL's tested rooftop layouts, a very low mounting height of 0.15 m produced annual mismatch losses of roughly 1.5% to 2%, while higher-clearance cases were generally below about 0.5%. Those numbers apply to the tested layouts, not every bifacial project.^[2]

Sandia measurements also found rear-irradiance differences of more than 50 W/m² between cell positions on a sunny day with a test module mounted about 0.6 m above the ground.^[3]

Similar average rear readings can therefore come with different currents. Tongwei discusses the same issue in its article on similar rear irradiance but different rear-side current.

Cell Layout Changes the Loss

Where a shadow falls matters.

Two modules may each have 5% of the rear surface shaded. On one module, the shade is spread across several cell paths. On the other, one narrow rail shadow cuts across a single substring. The shaded area is the same, but the power loss can be different.

The result depends on:

• full-cell or half-cell design;
• cell orientation;
• series and parallel connections;
• substring layout;
• bypass-diode position.

Bypass diodes do not remove normal mismatch. Pmax can fall before a bypass diode turns on.

A 20%, 30% or 50% difference in rear irradiance is also not a fixed bypass threshold. Front irradiance and the current of the weakest connected cells still matter. This is covered in Tongwei's discussion of uneven rear irradiance and bypass activation.

Racking Shadows

Torque tubes, rails, tracker bearings, braces, junction boxes and cable bundles can all block rear light.

In this example, the 35 W/m² point behind the rail is about 85% below the 230 W/m² open area.

Sandia models rear irradiance across different cell positions because nearby structures and array geometry can create sharp local differences.^[4]

Rails, junction boxes and row spacing can all contribute to rear-side power losses, especially when a shadow falls across one electrical substring.

Mounting Height and Row Spacing

A module close to the ground sees a smaller area below and around it. Some of that ground may already be shaded by the module or the next row.

Raising the module can give the rear side a wider view of lit ground and can reduce sharp light differences across the cells.

Sandia's RADIANCE work found that cell-spacing effects were stronger at a mounting height of 0.5 m and became much smaller at heights of 1 m or more in one modeled case.^[5]

One meter is not the right height for every system. Tilt, pitch, tracker shape and ground conditions still change the result.

More row spacing can also reduce row-to-row shade and let more light reach the ground. The trade-off is simple: wider spacing uses more land per module.

Ground Albedo

Albedo is the share of incoming solar radiation that a surface reflects.

NREL measurements show how different real surfaces can be:

These are site-specific measurements, not universal values for every surface.^[6]

At Changshu, white-painted concrete measured 0.533 versus 0.236 for ordinary concrete. That is more than twice the reflected fraction.

High albedo only helps when the surface receives sunlight and the reflected light has a clear path to the rear cells. Low clearance, row shadows or a torque tube can block part of that gain.

Albedo also changes with rain, dust, vegetation, aging and snow conditions. One bright-day measurement should not be treated as a fixed yearly value.

Front and Rear Can Move in Opposite Directions

Rear irradiance can rise while module power falls if front irradiance drops by more.

Position A:

950 + (100 × 0.80) = 1,030 W/m²

Position B:

850 + (180 × 0.80) = 994 W/m²

Rear irradiance rose 80%, from 100 to 180 W/m². Adjusted total irradiance still fell from 1,030 to 994 W/m², a drop of about 3.5%.

The rear side improved. The 100 W/m² loss on the front was simply larger than the usable rear gain.

Tracker angle changes can cause exactly this kind of front/rear trade-off. A change in tracker backtracking can alter front irradiance, rear irradiance and row shading at the same time.

Temperature Reduces Voltage

More irradiance can make the cells hotter. Higher cell temperature usually lowers voltage in crystalline-silicon PV devices.

Sandia's module-temperature model uses irradiance, ambient temperature, wind speed, module construction and mounting because all of them affect cell temperature.^[7]

Using a Pmax temperature coefficient of −0.30%/°C:

A module running 5°C hotter would lose about 1.5% of Pmax compared with the same irradiance condition at the lower temperature.

Tongwei's TNC-G12R lists TkPower at −0.305±0.025%/K, close to the −0.30%/°C value used above.

Spectral Effects

Two surfaces can both produce a rear reading of 200 W/m² and still send a different mix of wavelengths toward the module.

Grass, concrete, soil, roof materials and snow do not reflect the same wavelengths equally. Silicon cells also respond differently across the solar spectrum.

NREL found measurable spectral differences between ground materials and between sensor readings and the effective irradiance seen by bifacial PV devices.^[8]

During troubleshooting, check large shadows and rear-light mismatch before chasing small spectral effects. Spectrum becomes more useful when comparing small differences between different ground surfaces.

Light Angle

Rear light can hit the module almost straight on or arrive at a shallow angle.

At a shallow angle, more light can reflect off the glass instead of reaching the cells. Sandia describes this with an angle-of-incidence modifier.^[9]

Sensor Position

A rear irradiance sensor only measures the light where it sits.

The center sensor looks normal at 185 W/m². The rail-shadow point is about 86.5% lower.

Check sensor position, sensor angle, structural shadows, cleaning, calibration and timestamps before using one rear reading to explain module power.

Time alignment matters when clouds are moving. A rear reading taken a few minutes before or after the current measurement may describe a different light condition.

A change from 180 to 185 W/m² is only 2.8%. A one-off difference this small is hard to judge without checking sensor uncertainty, repeatability and front irradiance at the same time.

Check Current First

Rear light usually shows up first as extra current.

These are illustrative values, not field measurements.

Rear irradiance rose 66.7%, current rose 6.9%, and DC power rose 4.5%. The percentages are clearly not interchangeable.

If rear irradiance goes up but current barely moves, check front irradiance, sensor position, timestamps and rear shadows first. A current plateau under strong rear irradiance can also come from measurement or system limits rather than the cells simply stopping their response.

String-Level Differences

Modules in the same string can receive different rear irradiance.

An edge module may see more open ground than a module deep inside the array. Access roads, row ends, snow lines and changes in ground material can do the same thing.

One rear sensor behind one module cannot represent a long row when the rear-light conditions change from module to module.

DC Power and AC Output

Higher module DC power does not always appear as higher AC output.

Available DC power rises by about 5.6%, but the inverter is already limited to 1.00 MW AC.

Sandia defines clipping as the condition where available DC power exceeds the inverter's usable input or conversion capability.^[10]

The 5.6% DC increase is clipped, so it does not appear in instantaneous AC output.

Snow Can Pull in Both Directions

Snow can raise rear irradiance while cutting front-side output at the same time.

• clean snow can reflect a lot of light toward the rear;
• snow on the front glass blocks direct light;
• cold cells usually hold voltage better;
• partial snow cover can create uneven current.

A rear sensor can therefore look excellent while total module power remains low because part of the front is covered.

Fresh snow, dirty snow and melting snow also reflect different amounts of light.

What to Check

Compare front irradiance, rear irradiance, temperature, current and voltage over the same time interval. One isolated rear W/m² number is not enough to explain bifacial performance.

FAQ

Can rear light push module current above the front-only value?
Yes. Rear light adds current. Inverter inputs, cables and connectors must be sized for the expected bifacial current, not just the front-only condition.

Can two modules with the same front-side wattage have different rear gains?
Yes. Bifaciality, cell layout, resistance and temperature coefficients can differ even when the front-side nameplate power is similar.

How many rear sensors are needed?
One well-placed sensor can work when rear light is fairly even. If rails, torque tubes, terrain or ground surfaces create large differences, more than one measurement point gives a better picture.

What standard covers bifacial I-V measurement?
IEC TS 60904-1-2:2024+AMD1:2026 covers current-voltage measurement requirements for bifacial photovoltaic devices.^[11]

Conclusion

Higher rear irradiance should first show up as more current. If it does not, check the sensor, front irradiance and rear shadows. If current rises but Pmax does not rise as much as expected, look at temperature, voltage and how evenly the rear light reaches the cells. A module with 1,000 W/m² on the front, 200 W/m² on the rear and 80% Isc bifaciality has a simple equivalent irradiance of 1,160 W/m², but that figure is not a guaranteed 16% power gain. The rear W/m² number is useful only when it is read together with the electrical data.