Effects of Cerium Removal from Glass on Photovoltaic Module Performance and Stability

National Renewable Energy Laboratory, Golden, CO

Schematics and photo of the lap-shear design.

In photovoltaic (PV) modules, polymeric materials provide electrical insulation and protect modules from mechanical damage and environmentally induced corrosion. When used in front of a PV cell, the principal attributes of interest are that the encapsulant transmits photons and maintains adhesion to all surfaces. PV module qualification tests are designed to provide minimum standards for module durability and to demonstrate a degree of safety in the production of electricity.

Photovoltaic modules are exposed to extremely harsh conditions of heat, humidity, high voltage, mechanical stress, thermal cycling, and ultraviolet (UV) radiation. Because of the extreme difficulty of exposing production modules to concentrated light sources, the UV exposure required by these tests corresponds to an equivalent field exposure of several months to 1.5 years rather than the desired service life of 30 years. In the mid-1980s, there were a number of notable problems with ethylene vinyl-acetate (EVA) yellowing. This degradation was partially the result of polymer embrittlement, delamination, and/or discoloration (yellowing). One part of the solution to these issues was to use glass containing Cerium (Ce), which selectively blocks UVB radiation. In recent years, better stabilizer formulations for EVA have been developed, giving more confidence in the long-term stability of PV packaging materials. Because of this, some manufacturers have stopped using Ce-doped glass.

Schematic diagrams of samples used to estimate the useful photon flux for a PV device.

In typical low-Fe Ce-doped glass, the Ce is present in both the Ce³⁺ and Ce⁴⁺ states. It is the Ce³⁺ state that absorbs ultraviolet light, with peak absorption around 314 nm. Upon exposure to UV radiation, the Ce³⁺ will be oxidized to the Ce⁴⁺ state. Because diffusion of oxygen into soda lime glass occurs on geologic time scales, Ce cannot be oxidized without a corresponding reduction reaction. In glass containing iron, Fe³⁺ will be reduced to Fe²⁺ since Fe²⁺ has a broad absorption peak around 1050 nm, which is within the useful range of typical silicon-based PV technologies. Fe3+ has a relatively weak absorption peak in the UV range around 370 nm. Thus, “solarization” results in a small decrease in PV performance upon field exposure of Ce-containing glass.

Recently, photovoltaic panel manufacturers have been using glass that does not contain Cerium. This has the advantage of providing about 1.3-1.8% more photon transmission but potentially at the expense of long-term stability. The additional transmission of light in the 300-340 nm range can cause delamination to occur about 3.8 times faster. Similarly, UV radiation will cause polymeric encapsulants, such as EVA, to turn yellow faster, losing photon transmission. Silicones do not suffer from light-induced degradation as hydrocarbon-based polymers do, therefore if silicone encapsulants are used, a 1.6-1.9% increase in photon transmission can be obtained from removal of Ce from glass, with no tradeoff in long-term stability. Additionally, antimony (Sb) can be added to non-Ce-containing glass to further improve photon transmission (principally in the IR range) by an additional 0.4-0.7%; however, this does not significantly affect UV transmission, so the same UV-induced reliability concerns will still exist with common hydrocarbon-based encapsulants.

1-in. x 1-in. laminated composite EVA specimens after exposure to 42 UV suns.

Experiments have focused on the benefits and problems associated with the use of Ce in low-Fe glasses used in the PV industry. It has been shown that the removal of Ce from PV glass provides improvements to Jsc of about 0.9-2.3%, and that the addition of Sb can also improve light transmission. These benefits arise from improved transmission in the near-IR region because of a decreased concentration of Fe²⁺ ions in the glass. The improved transmittance in the UV range does not significantly improve the PV performance because of low intensity and low quantum efficiency in the UV, and because the encapsulants absorb much of the UV light.

Unfortunately, the removal of Ce from glass can accelerate the delamination of EVA from the front glass by about 3.4-3.8 times. Because most encapsulant materials strongly absorb UV light, Ce removal should not affect the adhesion of the encapsulant to the PV cells. With regard to loss in transmittance through the encapsulants over time, significant problems were seen in the polyvinyl butyral (PVB) sample tested. The EVA and thermoplastic polyurethane (TPU) samples gave equivalent results at 42 UV suns, indicating that with regard to light transmittance, the use of non-Ce glass may be acceptable if a well-formulated material is used. The ionomer tested retained excellent transmission even without UV-blocking Ce.

Photograph of ionomer sample after exposure to 42 UV suns.

Lastly, the polydimethylsiloxane (PDMS) encapsulants showed no signs of UV-induced degradation with respect to either lap-shear strength or optical transmittance. If not for the typically higher prices for using PDMS materials, they would be the clear choice for PV modules. Removal of Ce and addition of Sb to glass has the potential to improve PV module performance, but these improvements may be short-lived because of enhanced UV degradation. One solution would be to use an antireflective coating that also blocks UV light below 350 nm. If one could also develop it to reflect the far-IR, the module temperature could be reduced, creating further improvements in module performance.

This work was done by M.D. Kempe and T. Moricone of the National Renewable Energy Laboratory, and M. Kilkenny of Skyline Solar.