Skip to content
WattCurious

How to Clean Solar Panels Without Damaging Them

Updated 8 October 20267 min readSolar Energy

A dirty module loses output because light never reaches the cell. A badly cleaned module loses output permanently, because the anti-reflective coating, the glass or the cells themselves have been damaged. The gap between those two outcomes is entirely a matter of method — water quality, surface temperature, pressure and timing — and none of it is guesswork.

Key takeaways

  • Uniform dust costs a little output everywhere; one opaque patch costs far more, because cells in series are limited by the worst of them.
  • Rain cleans a tilted module reasonably well and a nearly flat one badly, because runoff is what carries the dust away.
  • The anti-reflective coating on the glass is microns thick and is the first thing an abrasive cloth or a gritty brush destroys.
  • Cold water on hot glass is a thermal shock the module was never designed for; clean early, late or on an overcast day.
  • Hard water dries into a mineral film that scatters light as effectively as the dirt it replaced.
On this page
  1. What dirt actually does to a module
  2. Uniform dust and one opaque patch are not the same problem
  3. When rain does the job, and when it does not
  4. What actually damages a module during cleaning
  5. A method that respects all of that
  6. Deciding whether cleaning is worth doing

What dirt actually does to a module

Nothing about soiling is mysterious. Light has to cross the glass to reach the cell, and anything sitting on that glass either absorbs it or scatters it somewhere else.

Cross-section through the top of a solar module. Sunlight arrives at the glass surface. On the clean half, light passes through the anti-reflective coating and glass into the cell. On the soiled half, a layer of dust absorbs part of the light, scatters part of it back to the sky, and lets a reduced amount through to the cell, with the scattered fraction shown spreading in several directions.Cross-section through the top of a solar module. Sunlight arrives at the glass surface. On the clean half, light passes through the anti-reflective coating and glass into the cell. On the soiled half, a layer of dust absorbs part of the light, scatters part of it back to the sky, and lets a reduced amount through to the cell, with the scattered fraction shown spreading in several directions.
Absorbed, scattered away, or through to the cell. A dust layer converts the first two into heat and loss before the cell has any say in it.

Two properties of the layer decide how much is lost. Coverage is the fraction of the surface obscured. Opacity is how completely each particle blocks light — a fine mineral dust scatters much of what it intercepts, while a leaf or a bird dropping is effectively a hole punched in the aperture.

Soiling is also wavelength-dependent and angle-dependent, which is why a panel can look hazy and still perform acceptably, or look almost clean and have lost real output. The eye is a poor instrument here. The only reliable measure is the electricity.

Uniform dust and one opaque patch are not the same problem

This is where soiling stops being simple arithmetic. Cells within a module are wired in series, so the same current has to pass through all of them, and a shaded cell limits that current for the whole string it belongs to.

Two modules compared. The first carries an even film of dust across every cell and produces slightly reduced current everywhere. The second is clean except for one opaque bird dropping covering a single cell; the string containing that cell is current-limited, its bypass diode conducts and switches the whole sub-string out, and the covered cell dissipates power as heat, shown as a hotspot.Two modules compared. The first carries an even film of dust across every cell and produces slightly reduced current everywhere. The second is clean except for one opaque bird dropping covering a single cell; the string containing that cell is current-limited, its bypass diode conducts and switches the whole sub-string out, and the covered cell dissipates power as heat, shown as a hotspot.
An even film costs a few percent. One opaque patch can switch out a third of the module and heat the covered cell — a disproportionate penalty from a very small amount of dirt.

An even film reduces output roughly in proportion to the light it blocks. One opaque patch does something worse: it throttles its whole series string, and the bypass diode protecting that string conducts to route current around it — taking a third of a typical module out of service to save a few square centimetres of cell. The covered cell, meanwhile, is reverse-biased and dissipates rather than generates, which is the heating mechanism behind hotspots. The same physics governs partial shade from objects, and it is why a single dropping deserves attention that a season of dust does not.

When rain does the job, and when it does not

Rain is the default cleaning system, and on a well-tilted array it is a good one. The mechanism is runoff: water has to move down the glass carrying the loosened material with it.

Three modules at different tilt angles under rain. At a steep tilt, water runs off quickly and carries dust to the lower edge and away. At a shallow tilt, water pools and evaporates in place, leaving a dried mineral film and a band of accumulated dirt along the lower frame. A fourth panel shows dew wetting a dust layer overnight and drying into a cemented crust that rain no longer removes.Three modules at different tilt angles under rain. At a steep tilt, water runs off quickly and carries dust to the lower edge and away. At a shallow tilt, water pools and evaporates in place, leaving a dried mineral film and a band of accumulated dirt along the lower frame. A fourth panel shows dew wetting a dust layer overnight and drying into a cemented crust that rain no longer removes.
Runoff is the cleaning agent, not the water itself. Shallow tilts leave a dirt line at the lower frame, and repeated wet-dry cycles cement dust into a layer rain cannot lift.

Three things defeat it. A shallow tilt slows runoff so water evaporates where it sits, depositing whatever it dissolved. The lower frame edge interrupts runoff on every module, which is why a persistent band of dirt collects along it. And cementation — repeated cycles of dew or light rain wetting a dust layer and drying it again — binds particles to the glass until only mechanical contact will remove them.

Light rain is the worst case of all: enough to wet and redistribute the dust, not enough to carry it off. A dry spell after that leaves the module dirtier than before it rained.

What actually damages a module during cleaning

The module's top surface is not plain glass. It carries an anti-reflective coating a fraction of a micron thick whose whole purpose is to stop light bouncing off the surface, and it is the first casualty of aggressive cleaning.

Four damage mechanisms shown across one module. Abrasion, where a gritty cloth scratches through the anti-reflective coating leaving a hazed track. Thermal shock, where cold water hits glass heated by the sun and the sudden contraction stresses it. Pressure, where a high-pressure jet drives water past the edge seal and flexes the laminate. Mechanical load, where a person standing on the glass causes microcracks in the cells beneath.Four damage mechanisms shown across one module. Abrasion, where a gritty cloth scratches through the anti-reflective coating leaving a hazed track. Thermal shock, where cold water hits glass heated by the sun and the sudden contraction stresses it. Pressure, where a high-pressure jet drives water past the edge seal and flexes the laminate. Mechanical load, where a person standing on the glass causes microcracks in the cells beneath.
Abrasion, thermal shock, pressure and load. Each produces a module that looks fine and produces less, which is why the damage usually goes unnoticed.
  • Abrasion. A dry wipe drags the grit already on the surface across the coating. The scratch pattern scatters light permanently, and no later cleaning recovers it.
  • Thermal shock. Glass in full sun sits well above air temperature — for the same reasons that make modules lose efficiency when hot. Cold water on hot glass contracts the surface faster than the bulk, and the module was qualified for weather transitions, not for a hose.
  • Pressure and flex. High-pressure water and body weight both bend the laminate. Crystalline silicon cells crack at strains far below what breaks glass, giving exactly the invisible damage that makes hail impacts so hard to diagnose.
  • Chemical residue. Detergents leave films, solvents attack sealants, and hard water dries into a mineral haze — all of them replacing dirt you can remove with dirt you cannot.

A method that respects all of that

Every choice below follows from one of the mechanisms above rather than from habit.

Each cleaning decision and the mechanism behind it
ChoiceWhyWhat happens otherwise
Clean early morning, evening, or under cloudGlass is near air temperatureThermal shock on hot glass; water dries before it is wiped
Rinse thoroughly before touching the surfaceFloats grit off so nothing is draggedAbrasion through the anti-reflective coating
Soft brush or microfibre, generous water, light pressureRemoves bonded dust without cuttingScratches, or flexed laminate and cracked cells
Softened, deionised or filtered waterDries without leaving mineralsA mineral haze that scatters as much as the dirt did
Low pressure, from a hose rather than a jetStays inside the module's sealing designWater past the edge seal; long-term moisture ingress
Work top to bottom, one module at a timeDirty water always runs onto uncleaned glassRepeated passes, and more contact with the surface
Stand on the roof or a platform, never on the glassLoad path is the frame, not the laminateInvisible microcracks and permanent output loss

A method, not a product recommendation. Any tool and any water source that satisfies the middle column is acceptable; module manufacturers publish their own limits and those take precedence.

Never scrape a dropping dry, and never use a metal blade on the surface. Soak it, wait, and let the water do the work — patience is free, and the coating is not replaceable.

Deciding whether cleaning is worth doing

The honest answer for most systems in most climates is that it is not, or not often. Soiling loss recovers on its own after a meaningful rain shower, and the accumulation rate at a given site is slow and fairly steady.

Chart of performance ratio over several months. Output drifts downward through a dry period as soiling accumulates, then steps back up sharply after heavy rain, then drifts down again. A cleaning event is marked, showing a similar recovery step. The vertical size of each recovery step is labelled as the amount of output the soiling was actually costing.Chart of performance ratio over several months. Output drifts downward through a dry period as soiling accumulates, then steps back up sharply after heavy rain, then drifts down again. A cleaning event is marked, showing a similar recovery step. The vertical size of each recovery step is labelled as the amount of output the soiling was actually costing.
Illustrative. The recovery step after a rain or a cleaning is the measurement — it tells you what the dirt was costing, which is the only basis for deciding how often to repeat the exercise.

Measure rather than assume. Compare the array's output to its own record under similar irradiance, and watch the size of the recovery step after rain. If a cleaning produces no step, the dirt was not the problem — and the usual alternative explanations are the ones behind a system producing less than expected.

That framing matters because cleaning is not free of risk. Every visit puts a person, a tool and a quantity of water on a sealed electrical product designed to be left alone for decades. Do it when the measurement says it is worth doing, and do it in a way the module can survive.

Frequently asked questions

Do solar panels need cleaning at all?

In most climates with regular rainfall and a reasonable tilt, very little. Soiling losses recover after rain, and the accumulation rate is slow. Cleaning earns its place where rain is rare, where the tilt is shallow, or where a specific local source — pollen, agricultural dust, salt, construction, birds — puts material on the glass faster than weather takes it off.

Can I use a pressure washer?

It is the single most common way to damage a module. High-pressure water can drive past the edge seal, stress the glass-to-frame joint, force water into the junction box area and, at close range, erode the coating on the glass. The pressure that removes baked-on dirt is well above what the laminate was designed to tolerate.

What water should I use?

The softest available. Hard water evaporates and leaves the dissolved minerals behind as a film that scatters light, so the panel can read dirtier after cleaning than before. Deionised or filtered water dries clear and is why professional systems use it; if only hard water is available, dry the glass rather than letting it evaporate in place.

Is it safe to walk on a module to reach the next one?

No. The glass is supported at its frame and is not a floor. Standing on it flexes the laminate and can crack cells invisibly, leaving a module that looks perfect and produces less for the rest of its life — the same mechanism that makes hail damage so hard to spot.

How often should cleaning happen?

There is no universal interval, because soiling rate is a property of the site, not of the technology. The useful approach is to measure: compare output against a clean reference or against the system's own history under similar conditions, and let the recovery achieved by a cleaning tell you when the next one is worth doing.

Sources

Named organisations whose published material underpins this article. Where no link is given, the source is named rather than linked.

Editorial Team

Research, drafting and review

Articles are drafted from primary engineering and physics references with AI-assisted tools, then reviewed and fact-checked line by line by a human editor before publication. We publish explanations, not recommendations: no products, no pricing, no country-specific rules, and no invented author personas.

Last reviewed 8 October 2026. How we research and review