Analysis of Accelerated Exposure Testing of Thin-Glass Mirror Matrix

Concentrating solar power (CSP) companies have deployed thin-glass mirrors produced by wet silver processes on ∼1-mm-thick, relatively lightweight glass. These mirrors have been bonded to metal substrates in commercial installations. Initial hemispherical reflectance is ∼93% to 96%, and the cost is ∼$16.1/m² to $43.0/m². These mirrors have the confidence of the CSP industry. However, corrosion was observed in mirror elements of operational solar systems deployed outdoors for 2 years. NREL’s advanced optical materials team was assigned to investigate the problem. First, it was noted that this corrosion is very similar to the corrosion bands and spots observed on small (45 mm × 67 mm) thin-glass mirrors laminated to metal substrates with several different types of adhesives and subjected to accelerated exposure testing (AET) at NREL. These samples exhibited corrosion at the unprotected edges and along cracks, and the choice of adhesive affected the performance of weathered thin-glass mirrors. Secondly, two significant changes in mirror manufacture have occurred in the wet-chemistry process because of environmental concerns. The first is the method of forming a copper-free reflective mirror, and the second is the use of lead-free paints. A test matrix of 84 combinations of sample constructions (mirror type / back protective paint / adhesive / substrate) was devised for AET as a designed experiment to identify the most promising mirrors, paints, and adhesives for use with concentrator designs. Two types of accelerated exposure were used: an Atlas Ci5000 WeatherOmeter (WOM) and a damp-heat chamber. Based on an analysis of variance (ANOVA), the various factors and interactions were modeled. These samples now have almost 24 months of accelerated exposure. Analysis of the thin-glass mirror matrix indicated that the Glaverbel mirror with a copperless formulation demonstrates slightly better performance compared to the Naugatuck standard copper-containing mirror and new copperless constructions although most results are within experimental uncertainty. Analysis of the thin-glass mirror matrix indicates commercial (non-mirror) back-protective paint applied after mirror manufacturing is not beneficial. Degradation of the samples exposed to date in the damp-heat chamber is similar, but at a rate 10 times faster than observed for samples in the WOM. We will discuss the results of the continued exposure testing of these mirror samples. Although glass mirrors with copper back-layers and heavily leaded paints have been considered robust for outdoor use, the new copperless back-layer and lead-free paint systems were designed for interior mirror applications and the outdoor durability must be determined.