Laser damage resistance of optical coatings in the sub-picosecond regime
When developing high-power ultrashort-pulse lasers and their related applications, the greatest concern is how to achieve resistance to laser damage in the optical components.1 Such laser systems require multilayer optical coatings that limit pulse lengthening and spectral distortion, while also exhibiting high laser-induced damage thresholds (LIDTs).
In a transparent optical material, damage occurs when there is energy transfer from the laser beam to the material. In the ultrashort regime (<10ps), this energy deposition is triggered by non-linear ionization processes—photo and impact ionization—that lead to an efficient coupling of the laser energy in the material. Damage to the material takes place when the deposited energy is larger than the critical energy needed to make material modifications (such as melting or vaporization). The main material property that drives laser damage is the optical bandgap. Therefore, the greater the bandgap, the higher the LIDT.2
Localized defects (macroscopic impurities at the surface or embedded in the film) are a major source of laser-induced damage, particularly in the nanosecond regime. Such damage is generally stochastic in nature and is related to the defects and their initiation of the damage process. As the pulse duration shortens, the damage mechanism becomes dominated by electron dynamics, and is therefore considered deterministic. However, recent studies suggest that isolated defects can also induce damage in the ultrashort regime. Isolated damage events related to defects can occur for fluences significantly lower than the ‘intrinsic’ damage threshold.3 This process can take place through strong absorption in the defect or through strong light intensification in the material surrounding the defect and intrinsic mechanisms. Commonly, the process results from a combination of both phenomena.
To explore LIDT, we characterized limiting defects by their density and the fluence at which they caused damage. To obtain these metrics, we used specific damage test procedures such as raster scanning.4 Using such techniques in the sub-picosecond regime,4 we observed tens of damage sites per square centimeter on dielectric mirrors at half the intrinsic damage threshold.
Once damage is initiated in an optic, it can grow under successive pulses, causing an obscuration level in the optic that can make the component unusable (see Figure 1). However, because small damage sites that do not grow can be benign, knowledge of damage growth behavior is a key issue for developers of laser systems. The density of growing damage sites is in fact the main parameter that must be evaluated to qualify an optical component.5 The bandgap of the material ultimately limits the damage threshold (and bandgap and refractive index are interrelated). Thus, there is also a clear correlation between the refractive index and the LIDT6 (see Figure 2).
Silica is widely used as a low-index material in many optical interference coatings. Theoretically, it would be possible to reach higher bandgaps with fluoride materials, but thin films of these materials with bulk-like properties have not been demonstrated for laser applications. Hafnium(IV) oxide exhibits good qualities as a high-index material and is used for high-power laser applications. However, the bandgap of 5.5eV is limited. By mixing hafnium(IV) oxide with silica, it is possible to achieve a significant increase in the laser damage threshold.7 We may also use aluminum(III) oxide, which has a lower refractive index than hafnium(IV) oxide but has a higher damage threshold (although in an increased number of layers). Mixtures of scandium sesquioxide and silica thin films have also demonstrated relatively high damage thresholds.8
Designing optical components involves reaching the best compromise between optical performance, LIDT, and fabrication constraints. To enhance laser damage resistance, the development of high-power optics is usually based on the optimization of the field strength distribution in the stack. Based on knowledge of the LIDT and the refractive index of optical thin-film materials deposited in particular conditions, it is possible (in most cases) to estimate the LIDT of a multilayer stack with some accuracy.10 As an alternative to quarter-wave optical thickness (QWOT) stacks (which enable optical layers to be deposited in quarter-wave optical thicknesses), non-QWOT stacks offer a solution to reject the maximum electric field in the low-index material. More recent concepts, such as a refractive index step-down from the substrate to the surface (which use mixtures of materials) can significantly improve the damage threshold.11
Unfortunately, a good design with high bandgap material is not enough to achieve high laser-damage resistance. The key to achieving increased short-pulse damage thresholds is to reduce in size and number the density of the defects (such as nodules), which are introduced by the manufacturing process. The complete elimination of these remains a challenge.
In summary, the laser-resistance performance of thin-film optical materials can be ranked based on their bandgap values with a linear evolution of the LIDT. Theoretically, high laser damage threshold optics can therefore be designed based on available materials and optimization of the electric field distributions in the multilayer stack. Overall, however, the user must consider incubation or fatigue when an optic is submitted to multiple subpicosecond irradiations.12 Finally, LIDT is strongly dependent not only on the material of choice but also on its fabrication, since defects that can trigger damage growth are a main limitation of laser damage resistance. The physical process for laser damage initiation and growth of defects is still unclear, and will be the subject of our future investigations.
The author is grateful for the support of the French Atomic Agency and REOSC, and acknowledges the assistance of partners at the Laser Research Center, Vilnius University, and the Laser Zentrum Hannover.
Laurent Gallais' research is focused on high-power laser/materials interactions. He has worked on laser-induced damage of optical components since 1999, particularly on laser damage of optical interference coatings in the femto- to nanosecond range, and the mitigation of laser damage growth for fusion lasers.
