A statistical study of the relationship between surface quality and laser induced damage

Laser induced damage of optical components is a concern in many applications in the commercial, scientific and military market sectors. Numerous component manufacturers supply “high laser damage threshold” (HLDT) optics to meet the needs of this market, and consumers pay a premium price for these products. While there’s no question that HLDT optics are manufactured to more rigorous standards (and are therefore inherently more expensive) than conventional products, it is not clear how this added expense translates directly into better performance. This is because the standard methods for evaluating laser damage, and the underlying assumptions about the validity of traditional laser damage testing, are flawed. In particular, the surface and coating defects that generally lead to laser damage (in many laserparameter regimes of interest) are widely distributed over the component surface with large spaces in between them. As a result, laser damage testing typically doesn’t include enough of these defects to achieve the sample sizes necessary to make its results statistically meaningful. The result is a poor correlation between defect characteristics and damage events. This paper establishes specifically why this is the case, and provides some indication of what might be done to remedy the problem.