What is module degradation
Module degradation is the gradual decline in solar PV module electrical output over its operating life, driven by physical, electrochemical and environmental mechanisms. Some degradation is intrinsic to crystalline silicon physics (light-induced defect formation), some is mechanical (micro-cracks from thermal cycling), some is electrochemical (PID under voltage stress), some is material ageing (encapsulant yellowing, back sheet embrittlement).
For modern mono PERC modules, total degradation is approximately 1 to 2 percent in the first 100 hours (LID), then linear 0.45 to 0.55 percent per year, giving 25-year retention of 84 to 87 percent of original nameplate. TOPCon and HJT achieve marginally better numbers.
Degradation determines lifecycle yield, warranty terms and replacement planning. Bankability of solar projects depends on degradation assumptions matching field reality.
Degradation explained simply
A new solar panel makes its full rated output on day one. The first month it loses about 1.5 percent (LID). After that it loses about half a percent every year. After 25 years it makes about 85 percent of original output. After 30 years, about 82 percent.
These losses come from various physical processes: photons creating defects in silicon, voltage stress driving ion migration, heat ageing the plastic encapsulant, micro-cracks from thermal expansion.
Why degradation matters
Lifecycle yield. Sum of yearly output depends on the degradation profile, which is why credible state-level solar yield benchmarks build in a degradation curve rather than a flat first-year output assumption.
Financing. Lenders model degradation conservatively, requiring assumptions backed by manufacturer warranty, an input that flows directly into C&I LCOE calculations.
Warranty enforcement. Tier-1 warranty replaces modules that fall below the warranty curve.
Replacement planning. Knowing when capacity drops below threshold guides replacement schedule.
Technology comparison. Degradation rates differentiate PERC, TOPCon and HJT.
Degradation mechanisms
LID (Light-Induced Degradation)
| Aspect | Detail |
|---|---|
| Mechanism | Boron-oxygen defect formation in p-type silicon under illumination |
| Magnitude | 1 to 2% in first 100 hours |
| Affects | All p-type silicon (PERC) |
| Mitigation | Gallium doping, n-type silicon |
PID (Potential-Induced Degradation)
| Aspect | Detail |
|---|---|
| Mechanism | Sodium ion migration from glass through encapsulant under high voltage stress |
| Magnitude | Up to 30% in extreme cases over 5 to 10 years |
| Affects | Modules under high system voltage |
| Mitigation | PID-resistant encapsulant (POE), glass priming, system grounding |
LeTID (Light and Elevated Temperature Induced Degradation)
| Aspect | Detail |
|---|---|
| Mechanism | Hydrogen-related defect under combined light and heat |
| Magnitude | 1 to 5% over years |
| Affects | PERC, especially first generations |
| Mitigation | Process control, n-type silicon (TOPCon, HJT) |
Encapsulant degradation
| Aspect | Detail |
|---|---|
| Mechanism | EVA yellowing under UV |
| Magnitude | Few percent over decades |
| Mitigation | UV-stabilised encapsulant, POE alternatives |
Mechanical degradation
| Aspect | Detail |
|---|---|
| Mechanism | Cell micro-cracks from thermal cycling, hail, handling |
| Magnitude | Variable, can be significant in poor-quality modules |
| Mitigation | Robust manufacturing, careful handling |
Back sheet degradation
| Aspect | Detail |
|---|---|
| Mechanism | Polymer back sheet cracking under UV and humidity |
| Magnitude | Cosmetic to electrical risk |
| Mitigation | Glass-glass bifacial construction, higher quality polymers |
Technical deep dive
Total degradation timeline
| Stage | Loss | Cumulative loss |
|---|---|---|
| LID (first 100 hours) | 1 to 2% | 1 to 2% |
| Year 1 to 5 | 0.5 to 0.7%/yr | 3.5 to 5.5% |
| Year 5 to 15 | 0.4 to 0.5%/yr | 7.5 to 10.5% |
| Year 15 to 25 | 0.4 to 0.5%/yr | 11.5 to 15.5% |
| Year 25 to 30 | 0.4 to 0.5%/yr | 13.5 to 18% |
Cell technology comparison
| Technology | First year (incl. LID) | Year 2 onward | 25-year retention |
|---|---|---|---|
| Mono PERC | 2 to 2.5% | 0.45 to 0.55%/yr | 84 to 87% |
| Mono TOPCon | 1 to 1.5% | 0.40 to 0.50%/yr | 85 to 88% |
| Mono HJT | 0.5 to 1% | 0.30 to 0.40%/yr | 88 to 92% |
| Polycrystalline | 2 to 2.5% | 0.60 to 0.80%/yr | 80 to 83% |
Warranty curves
Tier-1 manufacturers typically warrant 0.45 to 0.55 percent annual against 0.55 percent allowable.
Real-world applications
Lifecycle solar yield modelling.
Performance warranty enforcement.
Replacement timing decisions.
Module technology comparison.
Site-specific degradation under Indian climate.
Advantages of low-degradation modules
Higher lifecycle yield.
Better LCOE.
Longer warranty terms.
Reduced replacement risk.
Limitations
Degradation is irreversible. Cannot be “fixed” through cleaning or maintenance.
Manufacturer claims vary; field validation takes years.
Warranty enforcement requires documentation discipline.
Common problems
PID in legacy modules without PID-resistant materials. 10 to 30 percent loss over 5 years.
Back sheet cracking in tropical climates.
Encapsulant browning visible after 10 to 15 years.
Hot spot induced local degradation.
Snail trail cosmetic marks (often benign).
Junction box bypass diode failure.
Best practices
Specify PID test certification.
Use modules with linear performance warranty of 0.55 percent or better.
Document module electrical parameters at commissioning.
Schedule annual IV curve audits during O&M.
Photograph module condition every 3 years.
Maintain warranty correspondence files.
Track yield against warranty curve.
Comparison: degradation drivers
| Driver | Effect | Mitigation |
|---|---|---|
| LID | First 100 hours | n-type silicon |
| PID | Multi-year drift | POE encapsulant, grounding |
| LeTID | Gradual under heat | Process control |
| UV | Encapsulant ageing | Stable encapsulant |
| Thermal cycling | Micro-cracks | Quality manufacturing |
| Soiling | Apparent loss (reversible) | Cleaning |
Indian market context
ALMM-listed Indian modules, available through authorized solar equipment suppliers, typically warrant linear degradation of 0.45 to 0.55 percent per year with 84 to 87 percent retention at year 25.
PID test certification under IEC 62804 is increasingly demanded by Indian project developers.
LeTID has affected some PERC installations from the late 2010s. Modern PERC processes control LeTID.
Climate variation across India (cold Himalayas to hot Rajasthan to humid coast) creates region-specific degradation profiles.
PM Surya Ghar requires modules with linear performance warranty.
Standards and certifications
| Standard | Scope |
|---|---|
| IEC 61215 | Module qualification |
| IEC 61730 | Module safety |
| IEC 62804 | PID test |
| IEC TS 63209 | Extended thermal cycling |
| IS 14286 | Indian Standard for crystalline modules |
| MNRE ALMM | Listing for subsidies |
Common mistakes
Ignoring LID in first-year yield modelling. Causes “underperformance” panic.
Choosing modules without PID test certification.
Not auditing module IV curves periodically.
Failing to document degradation correspondence for warranty.
Assuming uniform degradation across years (real curves are slightly non-linear).
Skipping initial commissioning baseline measurement.
Key takeaways
Module degradation is the gradual decline in solar output over time. Modern modules degrade 0.30 to 0.55 percent per year after a 1 to 2 percent initial LID. 25-year retention is 84 to 92 percent depending on cell technology. PID, LeTID and encapsulant ageing are the main controllable mechanisms. Tier-1 ALMM-listed modules with linear performance warranty are the safe choice for Indian solar projects.
Further reading
For how Module Degradation plays out in real projects, these guides go deeper:
- How Does a Solar Power System Work in India?
- How Shadow on One Solar Panel Kills Your Generation (And How to Fix It)
- Solar Inverter Low Output, 12 Causes and Fixes India
- Best Solar System Under ₹1 Lakh in India 2026
Sources
- IEC 61215, Module qualification and type approval.
- IEC 61730, Module safety.
- IEC 62804, PID test method.
- IEC TS 63209, Extended thermal cycling.
- IS 14286, Indian Standard for crystalline silicon PV modules.
- MNRE Approved List of Models and Manufacturers (ALMM).
- NREL Technical Report TP-5J00-78211, PV Module Degradation Rates.
- ITRPV Roadmap, latest edition, VDMA.
- IEA PVPS Task 13 Report on PV performance and reliability.
- Jordan D. C., Kurtz S. R., Photovoltaic Degradation Rates: An Analytical Review, Progress in Photovoltaics, 2013.
Frequently Asked Questions
What is module degradation in simple words?
How much do solar panels degrade per year?
What is the first-year degradation?
What is the 25-year retention?
What causes degradation?
What is LID?
What is PID?
What is LeTID?
Are warranty terms based on degradation?
Does temperature affect degradation?
Can degraded modules be replaced under warranty?
Does cleaning affect degradation?
Are bifacial modules more degradation-prone?
How is degradation measured?
What is the warranty curve?
From the Blog
Co-founder & CEO driving Qbits Energy's product strategy and growth across India.