Femtosecond vs. nanosecond laser damage threshold

As laser technology continues to evolve, optical components must also advance to meet the demanding specifications required by high-precision applications. Medical procedures, micromachining, fundamental scientific research, and many other fields have all been revolutionized by the power of ultrafast laser technologies. For industries and applications previously dominated by nanosecond lasers, adopting ultrafast lasers presents challenges, including remarkably different laser damage thresholds for optical components. To ensure efficiency and longevity of your laser system, it is essential to understand how and why this specification differs across nanosecond and femtosecond pulse durations. 

Laser damage threshold (LDT), sometimes referred to as laser-induced damage threshold (LIDT), is a critical parameter to evaluate when selecting optical components for any laser system. ISO 21254 defines LDT as the “highest quantity of laser radiation incident upon the optical component for which the extrapolated probability of damage is zero.”¹ While this definition may seem straightforward, an actual LDT value is dependent on a variety of factors beyond the nature of the optic itself. In particular, an optic’s LDT may vary by several orders of magnitude when evaluated at nanosecond (10^-9 s) vs. femtosecond (10^-15 s) pulse durations. This wide variance is due to the drastically different mechanisms for laser damage that occur at these different time scales (see Fig. 1). 

Nanosecond laser damage mechanisms

Compared to femtosecond pulses, the long pulses of nanosecond lasers induce damage in optical components primarily by thermal mechanisms. The laser deposits a significant amount of energy into the material of the optical component, which drives local heating within the laser spot. This heating may trigger melting directly or it can induce some structural changes through thermal expansion and resultant mechanical stress. This stress may go on to cause cracking or even cause the coating to completely detach from the substrate through a process called delamination.³

Beyond direct heating of the coating material, optics under nanosecond laser irradiation are particularly sensitive to defects within the coating. These defects act like small lightning rods within the optical coating because they are prone to much higher absorption than their surroundings. As a result, these defect areas heat up more rapidly and, in cases of catastrophic laser damage, explode out of the coating. This violent damage mechanism often leaves behind craters on the surface, as well as additional particulate that redeposits immediately following the damage event (see Fig. 2).

Because these defect sites act as initiators for laser damage, a higher presence of defects is commonly associated with a lower LDT for a particular optic. For this reason, major emphasis is placed on the surface quality of optical components used with nanosecond lasers. In addition, LDT testing for nanosecond timescales is a highly statistical process. The probability of damage at any given site on the optical surface is due to a number of related factors, including the size of the incident beam, distribution and density of defect sites, as well as intrinsic material properties. Such a variety of contributing factors also explains why nanosecond LDT values may vary significantly from batch to batch of the same coating. The LDT may be affected by inconsistencies in substrate polishing and preparation, fluctuations during the actual coating deposition, and even changes in the storage conditions after coating. Such a wide variety of contributing factors to nanosecond LDTs contrasts with the dominant mechanisms that drive femtosecond laser damage, which are primarily intrinsic to the applied coating materials.³

Femtosecond laser damage mechanisms

The ultrafast pulses of femtosecond lasers induce damage through different mechanisms due, in part, to the incredibly high peak powers they generate. Even if a nanosecond and a femtosecond laser have the same pulse energy, the femtosecond laser pulse will have approximately a million times higher peak power due solely to the shorter pulse duration. These high-intensity laser pulses are capable of directly exciting electrons from the valence band to the conduction band. Even if the photons of the incident laser pulse are lower in energy than this transition—the so-called band gap of the material—the peak fluence of ultrafast laser pulses is so high that the electrons may absorb more than one photon at a time. This nonlinear mechanism is referred to as multiphoton ionization and is a common damage pathway in ultrafast laser optics.

Tunneling ionization may also present as a pathway to damage under femtosecond laser irradiation. This phenomenon occurs under the very strong electric fields generated by ultrafast laser pulses—so strong that the incident electric field actually warps the energy of the conduction band, which allows an electron to tunnel through from the valence band. Once enough electrons have been excited to the conduction band, the incident radiation begins coupling energy directly into that sea of free electrons, which then induces breakdown of the coating material.³

As a result of these damage pathways, femtosecond LDTs are much more deterministic than their nanosecond counterparts. Laser damage essentially “turns on” at a certain input fluence from a femtosecond laser, which scales with the band gap of the applied dielectric coating material. This starkly contrasts with the much more probabilistic nature of nanosecond laser damage (see Fig. 3).

In further contrast to nanosecond laser damage pathways, it is important to note that thermal effects do not govern an optic’s LDT on the femtosecond timescale. This is because the duration of ultrafast laser pulses is actually faster than the timescale of thermal diffusion within the material structure. As a result, femtosecond pulses do not deposit energy into the coating material as heat, so there is no thermal expansion and mechanical stress as with nanosecond laser pulses. For these exact reasons, ultrafast lasers are extremely advantageous in many applications that require high-precision cutting and marking,⁵ such as cardiovascular stent manufacturing.⁶

Be wary of scaling

Just like their pulse durations, typical LDT values for nanosecond and femtosecond pulses can differ by orders of magnitude. A common laser mirror may yield an LDT value around 0.2 J/cm² when measured with a 100-fs pulse, but that same optic could have an LDT closer to 10 J/cm² when measured with a 5-ns pulse. These different values may first be alarming, but they are nevertheless indicative of the vastly different damage mechanisms at play on these timescales. For the same reasons, be wary of using LDT scaling calculators across large timescales. In general, LDTs grow larger as the pulse duration grows longer, but scaling an LDT value from a femtosecond pulse to a nanosecond pulse and vice versa will most likely result in a damaged optical component. The best course of action is to select an optic with a suitable LDT rating obtained under conditions—including wavelength, repetition rate, and pulse duration—as close as possible to those in your application. 

Selecting the right optic

Laser technologies will continue to develop to meet the demand for higher precision. As these new technologies take shape, understanding the differences in laser damage mechanisms—and which are dominant on a particular timescale—will become increasingly critical for selecting the right optic for the application at hand. Keeping these differences in mind will not only promote efficiency and longevity of your laser system today, but will also allow for seamless adaptation to the more advanced laser systems of tomorrow. 

REFERENCES

1. See www.iso.org/standard/43001.html.

2. S. S. Mao et al., Appl. Phys. A, 79, 1695 (2004).

3. D. Ristau et al., Thin Solid Films, 518, 1607 (2009).

4. N. Carlie and K. Firestone, Laser Focus World (2019); www.laserfocusworld.com/14035451.

5. S. Lei et al., J. Manuf. Sci. Eng., 142, 1 (2020).

6. A. G. Demir and B. Previtali, Biointerphases, 9, 029004 (2014).