18 May 2007
Creating optics for petawatt lasers requires careful consideration of the actual-use conditions and an understanding of thin-film performance trade-offs, as David Kemp and Dave Collier of Alpine Research Optics, US, explain.
Laser systems in the petawatt power range can deliver peak intensities as high as 1022 W/cm2, enabling extreme experiments in several disciplines, including relativistic particle acceleration; high harmonic generation; and high pressure and temperature condensed matter studies. Such petawatt systems require unique and high-performance coated optics to relay the beam between the different laser and amplifier stages.
High-damage threshold coatings
Currently, the two ways to reach petawatt power levels are with very high pulse energy using Nd:glass amplification, or with ultrashort (around 30 fs) pulses using a high-bandwidth Ti:sapphire amplifier chain. Some hybrid systems use a mixture of these approaches.
One system funded by the US Department of Energy at the University of Texas at Austin uses a Ti:sapphire oscillator, multiple optical parametric amplifiers (OPAs) and unique glass amplifiers to reach a pulse energy of 200 J with a duration of 150 fs.
“Damage threshold at pulsewidths on the order of 100 fs is our concern,” project manager Mikael Martinez explained. “The damage threshold falls off exponentially as the pulsewidth decreases, but not much is known about damage thresholds much below a few hundred picoseconds. It is short-pulse damage that ultimately drives our maximum attainable power levels.”
For coating manufacturers, the proven way to reach the necessary damage threshold levels is traditional e-beam evaporation in conjunction with high-quality substrates and precision surface-cleaning techniques. But other coating methods for producing denser films, such as ion-assisted deposition, have made progress in recent years.
The Center for Ultrafast Optical Science at the University of Michigan, US, has developed a Ti:sapphire laser with a final output pulse of only 30 fs and 0.1 PW peak power. “30 fs translates into a minimum spectral bandwidth of 100 nm that must be maintained throughout our 810 nm system,” explained researcher Victor Yanovsky. “For any type of coating, higher bandwidth means using more layers and this complicates achieving high-damage thresholds.”
Coating designs invariably involve trade-offs and, sometimes, relaxing a parameter of limited importance results in a significant improvement in a more important area. A case in point is thin-film polarizers, which are used in conjunction with waveplates to make optical isolators and to provide variable beam attenuation.
A thin-film polarizer only works in the spectral band where there is a wide difference in the s and p reflectivities. But, the broadband operation of high-power lasers means that this region must be quite wide with both the s and p curves as flat as possible over the entire target bandwidth.
A typical broadband polarizer coating is designed to operate at Brewster’s angle (56°) where it can deliver up to 3% bandwidth (e.g. 24 nm at 800 nm). But, if the coating is designed to work at a 70° incidence angle, the bandwidth can be extended to 5% (40 nm at 800 nm).
However, even this wider performance is not sufficient for some Ti:sapphire and OPA-based systems. For this reason, Alpine Research Optics (ARO) has developed wider bandwidth polarizers where the extinction ratio is lowered from the typical 100:1 to a value closer to 6:1. These include so-called Tp biased polarizers, which deliver 99% p transmission but only 80% s reflection, and Rs biased optics, which deliver over 97% s reflectivity, but only 80% p transmission.
Film stress and post-coating flatness
When it comes to optics for petawatt systems, another priority is minimizing wavefront distortion. “We need an undistorted wavefront to achieve a diffraction-limited focal spot and the highest focused intensity,” added Yanovsky. “Also, higher levels of wavefront distortion can damage downstream amplifier components.” In the Michigan system, the specification is for λ/10 peak-to-valley or better in the reflected/transmitted wavefront. In reflection this translates into a surface figure of λ/20.
While λ/20 surface flatness optics are not uncommon, several factors combine to make it a challenge in petawatt lasers. First, some of the system mirrors are seven or more inches in diameter, so the surface quality must be maintained over a large clear aperture. Many of these large optics are fabricated using BK7 glass as the substrate, rather than fused silica, which could be very expensive. However, BK7 tends to reflow and undergo surface distortion during polishing. Lastly, the coatings must contain many layers in order to deliver the necessary broadband performance (particularly for highly curved parts with varying incidence angles such as aspheres).
Most optical thin-films are deposited at high temperature and mechanical stress can arise in a coating as it cools. The stress in each coating layer can be either compressive or tensive, depending on the wavelength-specific materials used and the exact deposition parameters. E-beam deposition produces somewhat porous films that can absorb water depending on ambient humidity, causing additional, slight time-varying changes in internal thin-film stress. The net result of all of these effects is that an optic that starts with a λ/20 specification might end up with a surface accuracy of λ/2 in actual use. “For this reason, it is vital to specify the post-coating surface figure,” commented Yanovsky.
There are a number of ways to achieve high post-coating flatness. One approach is prefiguring, which involves deliberately polishing a slight curvature on a substrate and allowing the coating stress to distort the optic back to flatness. Another method, useful for first-surface reflectors, is backside coating. Here a coating is placed on the second surface of the optic in order to null out the overall mechanical stress. But neither of these approaches is deterministic.
To reliably achieve high post-coating flatness, ARO has developed a new approach based on our experience of producing demanding optics for the National Ignition Facility. The starting point should be a substrate with a diameter:thickness ratio no higher than 5:1. In practice however, the polarizers in many pettawatt systems must also control the amount of dispersion as the beam passes through the substrate, meaning a thinner substrate is preferred. This illustrates the care that designers must take to balance the trade-off between optimal thickness and wavefront control.
The fabrication process begins with the production of a substrate with high flatness in terms of both power and irregularity. A thin film is then applied that is specifically optimized to introduce minimal distortion to the part. This is done by carefully selecting a combination of coating materials with both tensive and compressive characteristics.
In addition to minimizing overall stress, the design is also optimized to produce the required damage threshold and mechanical durability. Crucially, ARO has successfully characterized the relationship between temperature/humidity and coating stress. Understanding how coating stress changes with production conditions and the final operating environment has enabled the consistent production of optics that meet specification in actual use.
• This article originally appeared in the May 2007 issue of Optics & Laser Europe magazine.