Can anti-soiling coatings cut cleaning water? What the field trials show

Yes, in the field trials we have, but how much depends on the product and on the cleaning method. In the peer-reviewed Green Energy Park trial in Morocco, cleaned by hand, Solar Wash Protect, a cleaner that leaves an antistatic layer, used half the water per string; the two hydrophobic coatings used as much water as plain water cleaning on the same day. On a 330 MW plant in Vietnam cleaned by robot, D-Solar Defendor cut water use by 15–25%. On a tractor-cleaned plant in Chile, the coating showed up as faster passes, and the report gives no water figure.

Chemitek says the same about its own headline number: its figure of about 50% less water is measured against traditional manual cleaning, and automated systems save less (Chemitek).

1. Green Energy Park, Benguerir, Morocco

   +3.22%Solar Wash Protect, cumulative to the end of the dry season. D-Solar Defendor +2.01%, Industrial Glass Protect +1.89%.

   Set-up

   Five strings of seven modules, August 2021 to April 2022

   Cleaning

   Manual: pressure sprayer and brush

   Products

   Solar Wash Protect (SWP), D-Solar Defendor (DSD) and Industrial Glass Protect (IGP)

   Water

   SWP: 2.25 L per string on the day-30 clean, half the 4.5 L per string of the water-only clean at day 15. DSD and IGP: 4 L, the same as the water-cleaned string that day.

2. Ninh Thuan, Vietnam

   +2.64%In the dry season; +0.99% in the rainy season.

   Set-up

   330 MW plant; three treated and three water-cleaned strings of 285 panels; one year

   Cleaning

   Cleaning robot; about four cleans in the dry season, none in the rainy season

   Product

   DSD in the robot’s wash water

   Water

   15–25% less, in line with the time saved, as reported by the O&M team

3. Santiago, Chile

   +2.41%Over six months.

   Set-up

   115 MWc single-axis tracker plant; three treated and three water-cleaned strings; October 2020 to March 2021

   Cleaning

   Tractor-mounted brush; one wet clean to every two dry ones

   Product

   DSD in the wet clean

   Water

   Not reported: the two cleans after coating were dry. Tractor speed went from 1.1 to 1.35 km/h.

Sources: Elamim et al. (2023) and Green Energy Park’s field-test report; Chemitek case studies at the Vietnam plant, with BIM AC Renewables, and at Santiago Solar Plant.

Two details in the Morocco numbers matter. On the same day-30 clean, the SWP string used 2.25 L against 4 L for the water-cleaned string, which is 44% less (our arithmetic). And the two coatings, DSD and IGP, saved time on that clean (11 minutes against 12) but used the same 4 L as the water-cleaned string. Read the Morocco case study and the Vietnam case study.

Cleaning equipment runs on a continuous flow of water, so extra time spent cleaning means extra water (Chemitek). A coating saves water mainly by making each clean quicker. In Vietnam, the O&M team reported the water saving in line with the time saved, and the robot could switch to softer brushes.

With a manual crew, the water a clean takes depends on how long the crew sprays and rinses, so an easier clean can use less. In Morocco, that showed up as less water for Solar Wash Protect and as slightly shorter cleans for the two coatings. A robot or a tractor moves at the pace the machine and the ground allow. Chemitek’s note on its water figure says exactly this: with automated systems such as tractor-mounted brushes the percentage is lower, because terrain limits how much faster the machines can go. In Chile, the tractor gained about a quarter of a kilometre an hour.

For water-short regions, Chemitek recommends a hybrid of wet and dry cleaning. The Chile plant already ran one: the coating went on in a wet clean, and the dry cleans that followed ran faster. At Kalkbult, near De Aar, dry cleaning performed about as well as cleaning with distilled water on loose dust (du Plessis 2017).

Coatings earn most between cleans, when no rain helps. In Vietnam, the DSD strings gained 2.64% in the dry season and 0.99% in the rainy season, when the plant was not cleaned at all. In Morocco, SWP led through the dry months, and IGP, a hydrophobic coating, edged ahead in the rainiest months of March and April (Elamim et al. 2023).

A coating can also cut the number of washes, which is a second water saving on top of the first. Modelling by Ilse et al. (2019) shows that halving the soiling rate cuts the cost-optimal number of cleans by 29%. In Vietnam, the DSD strings went through three more cleans without the coating being reapplied. Chemitek states at least 12 months of protection for its coatings, and about six months for Solar Wash Protect. See how the interval is set in our cleaning frequency guide.

These are well-documented results with limits:

• Small samples. Seven-module strings in Morocco, and three treated strings in Vietnam and in Chile. Water was monitored on two cleans in Morocco, and reported as a range by the O&M team in Vietnam.
• Vendor involvement. The Vietnam and Chile studies are Chemitek case studies, and three of the eight authors of the Morocco paper work for Chemitek, which developed and applied the products.
• Mixed effects. Chemitek says DSD increases light transmission by 1.8% in its testing, and that this effect can’t be separated from the anti-soiling effect in the energy results.
• Sites differ. In a Northern Cape trial, a hydrophobic coating increased dust soiling, opening a performance-ratio gap of up to 5.5% (du Plessis et al. 2020). IEA PVPS (2022) notes there is no agreement on whether hydrophilic or hydrophobic coatings are effective, and that the benefit has to be assessed site by site. Long-term field results are typically 20–50% less soiling, well below the 80% or more seen in short tests (Ilse et al. 2019).

A wet wash uses about 1 L/m² in US literature and about 3 L/m² in a South African dissertation (NREL 2018; Naicker 2018), or roughly 50–150 m³ for one full wash of a 10 MWp plant. As an illustration only, a 50% saving on such a wash would be 25–75 m³, and a 15–25% saving about 8–38 m³. Your number depends on your method, crew, ground and water, so measure it the way the trials did:

1. Choose treated and control strings, matched for dust exposure.
2. Log litres and minutes per string, on both, at every clean.
3. Read energy against the control strings, and fit soiling rates with a sensor.
4. Run it through a dry season.

Match the product to the method. Antistatic Solar Armor 2.0 and D-Solar Defendor are dosed at 1 kg per 1,000 L into the water of rotary brushes, robots and tractors (DSD in soft water), and Solar Wash Protect also suits manual crews. On utility-scale and large C&I plants, our free pilot follows this protocol, and the soiling loss calculator puts the energy side in rand. See also low-water cleaning and equipment selection.