CSP mirror reflectivity, measured

A concentrating collector only uses light reflected accurately onto the receiver, and dust scatters light away from it.

Mirrors soil fast. At the Plataforma Solar de Almería, mirror soiling averaged 0.52% a day, with a maximum of 8.9% a day, and cleanliness never fell below 82% with fortnightly cleaning (Wolfertstetter et al. 2018). Near a ferromanganese smelter at Emalahleni, test mirrors lost on average 32.6% of their reflectance per 14 days in the dry season. The best site was 13.1% better than the worst, and the first rains helped (Swart et al. 2023).

South Africa had 600 MW of CSP at 30 June 2025, all in the Northern Cape: KaXu, Khi, Bokpoort, Xina, Ilanga-1, Kathu and Redstone (CSIR 2025; van Zyl et al. 2025). We found no published cleaning data for South African CSP plants, so each plant’s own measurements are the only guide.

Water is part of the sum. South African CSP plants hold water-use licences of 1,397–17,900 m³ a year per MW. Actual raw-water use was 0.35–0.56 m³/MWh at four plants, and 1.9 and 4.5 m³/MWh at two others (van Zyl et al. 2025). A wash that didn’t need doing spends water from that allowance; a late wash spends energy. Reflectance data lets you avoid both. See also low-water cleaning.

The SolarPACES Reflectance Guideline (v3.1, 2020) sets out how to measure and report the reflectance of solar mirrors:

• Specular reflectance counts the light reflected within a small acceptance angle around the mirror direction: the light a receiver can collect.
• Hemispherical reflectance counts all reflected light, scattered or not.
• Cleanliness factor is soiled reflectance divided by clean reflectance for the same mirror. A factor of 0.95 means the mirror delivers 95% of what it would clean.

Field readings use a portable specular reflectometer. The D&S 15R, for example, measures at 660 nm and 15° incidence, with 15, 25 and 46 mrad apertures, and repeats to ±0.002 (Devices & Services). We record the wavelength, incidence angle and aperture with every reading, so results compare over time.

1. Sampling plan. Choose sample mirrors across the field: near roads and dust sources, at the upwind and downwind edges, and in the centre.
2. Soiled and clean readings. Measure each sample as found, clean a small patch, and measure again.
3. Cleanliness map. Calculate the cleanliness factor by area, and track it over time to fit soiling rates by season.
4. Wash triggers. Set the cleanliness level at which washing pays. A modelled optimum at Almería averaged 0.965 ± 0.013, with 0.97–0.98 cited as optimal (Wolfertstetter et al. 2018). Yours depends on your soiling, water and tariff.
5. Treatment trial. Test wash methods and anti-soiling treatment on sample loops, and measure the difference.

The best published evidence for a mirror treatment comes from the same kind of reading this service takes. At the SOLABEN CSP plant in Extremadura, Spain, reflectivity was tracked on identified collectors, treated with Chemitek Mirtek and untreated, over 10 to 17 days (Case study: Chemitek (Portugal), with Atlantica and Rio Glass Services).

1. Jetwash, October 2021Loss 6× slower with Mirtek
2. With Mirtek, 10 days0.40%
3. Untreated, 11 days2.44%
4. Brush, July 2022Loss 13× slower with Mirtek
5. With Mirtek, 10 days0.16%
6. Untreated, 10 days2.17%

SOLABEN CSP plant, Extremadura, Spain: identified collectors, read over 10 to 17 days. Lower is better (Chemitek with Atlantica and Rio Glass Services).

After brush application, reflectivity recovered to its initial level in several cases, and Chemitek recommends the brush. Read the SOLABEN case. Mirtek is dosed into distilled or demineralised wash water. Buy it from us for your wash crews, or have us measure the effect on sample mirrors first.