Birefringence of tempered glass
The tempering process strongly increases the resistance of glass to breakage by creating mechanical stress within the material, and is therefore widely used for meeting safety standards. It consists in heating the glass to a temperature where it becomes soft, then quickly cooling it down so that the outer shell solidifies before the core. This lets the outer shell in a very tough compressive state, making the glass harder to break, while the core is in an unstable tensile state, causing the glass to split into many harmless fragments, if nevertheless the glass breaks.
Tempered glass anisotropy is a dreaded phenomenon which may look very unpleasant on a glazed facade. The mechanical stress created by the tempering process slightly modifies optical properties of the glass : its refractive index is slightly modified, and depends on light direction. Such a material is called a birefringent material, and has strong effects on polarized light, such as blue sky or light reflected on the glass surface under angle.
The phenomenon can have a very significant impact on the visual appearance of a building, mostly in clear sky conditions:
Simulating the effect
A new material model was developed in Ocean for reproducing the phenomenon. The anisotropic refractive index of the glass material is physically described using birefringence coefficients, and the model also allows coatings to be applied on both sides of the glass sheet.
In order to reproduce the non-uniform nature of the effect, the birefringence coefficients may depend on position, using any function. In this work, the birefringence coefficients were modulated according to random vertical stripe functions, corresponding to glass traveling under quenching nozzles in the tempering furnace. However, the defined shapes are only qualitatively similar to real tempered anisotropy patterns.
Simulations were performed:
- Under clear sky, using the Ocean polarized sky model
- Under overcast sky, using an environment map
Tests were made with and without a low-E coating. The results are presented below
No coating, cloudy sky
The anisotropy is already visible on this picture, but is quite low. Overcast skies are almost non-polarized, so the polarization comes only from reflection and refraction under angle on the glass.
No coating, clear sky
The effect is now very strong, and unpleasant. This confirms usual observations, where glass anisotropy is mostly seen by reflection of blue sky.
Low-E coating, cloudy sky
The patterns are now slightly colored, but not more visible than without coating.
Low-E coating, clear sky
The anisotropy effect is now colored : the stripes are not only lighter and darker, but vary between purple and cyan. This is similar to what was observed in the real picture.
Discussion and future work
In the last simulated image, we can observe some differences with the real picture. This can be due to multiple factors:
- The amount of birefringence : we have set it to 5.10-4 arbitrarily. This value strongly varies between tempering processes, in order to make a quantitative validation, it should be measured on the glass sample to reproduce.
- The sky conditions : the effect is stronger with very clear skies, and low haze. In a future article, we will present the results of using a real polarized sky capture, instead of a model.
- The angle of observation, and orientation of the glass in respect to the sky
- The coating : an arbitrarily simulated low-E coating was used in the simulation. As the nature of the coating has an influence on reflection colors, it should be defined from real measurements of a sample identical to the glass used in the real building.
To our knowledge, this is currently the only simulation of birefringence on architectural glass. This could be used for assessing optimal tolerance on birefringence, depending on glass configuration, coatings, and environment, so that the visual appearance of a building is not affected.