Panels degrade, and the first year does not resemble the rest

Two curves, not one

Solar panels lose output as they age, which everyone knows, and they lose it in two distinguishable phases, which fewer people do. The first is an initial drop that happens within the first hours to weeks of sunlight exposure and then stops. The second is a slow decline that continues for the rest of the panel’s working life.

The initial drop has a physical cause tied to defects that form in the silicon when it is first illuminated, and it is a well-characterised effect that manufacturers account for in their ratings. It is modest in size, it settles, and it is not a sign that anything is wrong. Cell technologies differ in how susceptible they are, and the industry has spent considerable effort reducing it.

After that the panel enters the long, boring part of its life, which is the part that actually determines whether it was worth installing.

What drives the slow decline

A panel is a laminate of glass, encapsulant, cells, interconnects and a backsheet, sitting outdoors and cycling through temperature swings every single day for decades. Nothing about that is chemically or mechanically gentle, and the degradation mechanisms are all consequences of it.

Ultraviolet light gradually yellows and embrittles the encapsulant, reducing the light reaching the cells. Thermal cycling flexes the solder joints between cells, and metal that flexes repeatedly eventually fatigues and cracks. Moisture works slowly past the edge seal and the backsheet, corroding metallisation and raising resistance where connections should be clean. On some system architectures, voltage between the cells and a grounded frame drives a leakage effect that degrades output more quickly than any of the above.

None of these produces a sudden failure. They produce a panel that makes slightly less each year than it did the year before, at a rate that is usually well under one per cent annually for a good panel and roughly linear once the initial settling is done.

A performance warranty describes a floor

Panels come with two warranties and they are frequently confused. The product warranty covers manufacturing defects and physical failure for a defined period. The performance warranty is a separate promise: that output will still be above some stated percentage of the original rating after a stated number of years.

That warranted line is a floor, not a forecast. It is set conservatively, because a manufacturer honouring claims at scale would be an expensive business, so the guaranteed curve sits well below what a decent panel actually does. Reading the warranty as a prediction of your yield is pessimistic in a way that will make a system look worse than it will be.

It is also worth reading who honours it. A performance warranty is a claim against a company, over a horizon long enough that companies merge, restructure or disappear. The warranty is only as durable as the entity behind it, which is not an argument against buying, only against weighting a twenty-five-year promise too heavily in a purchase decision.

Degradation is not the same thing as failure

Degradation is graceful. Failure is not, and the two follow completely different statistics. Panel failures cluster at the beginning of life, where a manufacturing defect shows itself early, and then again much later as materials reach the end of their durability. The middle stretch is remarkably quiet.

The failures that do occur are mostly mechanical or environmental rather than a matter of the cells wearing out: a cracked junction box, delamination where moisture has broken the bond between layers, a diode that has failed short, glass broken by hail or by someone walking on the array. Microcracks from careless handling during installation are a particular offender, because they may not affect output at all at first and then open up under thermal cycling years later.

The practical point is that panels are the reliable component in a solar installation, not the fragile one. They are the part with no moving pieces, no electrolytic capacitors and no firmware.

What this means when you weigh the numbers

Because degradation is slow and the early years produce the most, a modest annual decline matters far less to the economics than people assume. The first years of production carry most of the weight in any sensible financial comparison, and those are the years least affected. A percentage point of degradation is not what decides whether an array pays for itself; the tariff you buy and sell electricity at decides that, and that varies enormously by country.

What you can influence is small but real. Keep the panels ventilated, since heat accelerates most of the ageing mechanisms as well as reducing instantaneous output. Insist that installers handle modules carefully and never walk on them. In a climate with regular rain and reasonable tilt, cleaning is rarely worth paying for; in a dusty or arid one it may be.

And keep the expectation in proportion. A panel array that has quietly lost a tenth of its output over twenty years is not a disappointment, it is a component performing as designed. The inverter will almost certainly have been replaced by then, and that is the part worth thinking about.