Picosecond-Pulse Damage Studies of Diffraction Gratings

Picosecond-Pulse gratings are frequently used in dye-laser cavities as wavelength-tuning elements. These gratings often limit the ,fiaximumlaser output energy because of their low damage tl~resholds. We have measured the damage characteristics of both ruled and holo- graphically produced gratings, under a variety of conditions. USing the single-shot-per- site mode, the samples were irradiated by 30-ps, 1.064-pm pulses having a spot size of 0.5-nm radius. It was found that holographic gratings have damage thresholds from 1.5 to 5.0 times higher tnan similar ruled gratings. Thresholds for S-polarized light (E parallel to grooves) were higher by factors of 1.5 to 6. For the same type grating, gold coatings yielded higher thresholds than aluminum, although this is wavelength dependent. For holo- graphic gratings, replicas have slightly higher thresholds than masters. Dependence upon groove spacing was weak, Data are presented to show a variety of comparisons between different types of grat- ings, including two different manufacturers and usage at higher orders of incidence,


Diffraction
gratings are frequently used in dye-laser cavities as wavelength-tuning elements.
These gratings often limit the ,fiaximumlaser output energy because of their low damage tl~resholds.
We have measured the damage characteristics of both ruled and holographically produced gratings, under a variety of conditions. USing the single-shot-persite mode, the samples were irradiated by 30-ps, 1.064-pm pulses having a spot size of 0.5-nm radius.
It was found that holographic gratings have damage thresholds from 1.5 to 5.0 times higher tnan similar ruled gratings. Thresholds for S-polarized light (E parallel to grooves) were higher by factors of 1.5 to 6.
For the same type grating, gold coatings yielded higher thresholds than aluminum, although this is wavelength dependent. For holographic gratings, replicas have slightly higher thresholds than masters. Dependence upon groove spacing was weak, Data are presented to show a variety of comparisons between different types of gratings, including two different manufacturers and usage at higher orders of incidence,

Introduction
Diffraction gratings are frequently used in laser systems, e.g., dye-laser c~vities, as wavelength-tuning elements. These gratings usually have the lowest damage threshold of any component in tha system and thus determine or limit the maximum energy output available from thu laser, It is therefore important to understand the damage characteristics of gratings and how they can be improved.
Damage-threshold characteristics have been r arured for both conventionally ruled (replica) gratings and for the newer holographically produceo Iratings. A comparison between aluminum and gold as tile surface coating has been made, and the effuct of the nurnbi!ruf llnes per millimeter on the dam,age threshold has heen determined, Grating damage threshold" were measured for both S-and P-Polarizations, T~m 'Vfect of using ruled gratings in higher orders was investigated. Comparisons were made between ma6ter and replica holographic gratings dnd between replicn ruled gratings from two different manufacturers.
For theso t~sts the gratings were mounted in Littrow condition, first order, as is commot) for thcfr use as tuning elements in dye lasers, Using the single-shot-per-site mode, the samples were irradiated by 30-ps, 1064-nrn pulges having a spot size of O.S-mm radius, It was found that holographic gratfngs have damaqe thresholds from 1,5 to 5 times higher than similar ruled gratinqs.
For the same type qratinq, gold is a !nIlchbetter overcunt than alumlll~lm at 10G4-nm, as would br expected, Thresholds for gold-coated gratings were approximately 1,5 to 10 timgs high~r than the threshold for alumll!um-coatod gratings,

Experiment.al Setup
The laser damage threshold measurements were made using a Nd:YAG oscillator-amplifier conflguratioc. The laser was mode-'ocked, and a single 30-ps pulse was extract~d from the mode-locked tr'in, Care was used throughout the system to maintain a single transverse spatial mode.
This beam then entered the interaction area shown in figure 1. A 2-m focal-length lens focused the beam at a point beyond the sample (grating) so that the spot size radius at the sample was 0.5 m (Gaussian parameter), The spot size was measured cm every shot using a Reticon linear-diode array.
Damage was detected in three ways: (1) visual observation during the shot; (2) phol.omultiplier observation of spark; and (3) comparative observation of the scattering of a He-Ne laser before and after the shot.
The interaction room was completely darkened during the tests and an observer watched the sample through glasses highly shielded for the 1064-nm radiation. The observer saw a spark if substantial damage occurred, and was also able to detect very slight changes in the scattering from a coincident He-Ne beam. The primary damage detection was the signal from a photomultiplier viewing the spark at the interaction point. The photomulti lier was heavily filtered so that it could only see a narrow range of wavelengths centered at 4200 1.
After each shot the sample was translated so that no jite was irradiated more than once. An average of about 40 shots was used to establish the damage threshold for each test.
The gratings were mounted in Lfttrod condition, first order. Therefore gratings having different groove spacings were i!'radiated at different anales of incidence relative to the plane of the gratings.
A half-wave plate was used to change the polarization of the incident liser beam. In order to avoid confusion over the terms S-plane and P-plane, we will explicitly refer to the electric vector parallel to the grooves, E 11' and the electr!c vector perpendicular to the grooves, [ 1"

Metallic Co9tings
Four pairs of gratings were obtained such that, within a pair, the gratings were identical except for coating, One grating of e~ch pair was coated with aluminum and one with gold. The damage threshold results are shown In table 1.
In every case, the gold coated gratings had higher thresholds by a substantial ratio. This is reasonable at 1064 nm, since the refl~ctivity of gold is higher than that of aluminum, as shown in figure 2.
Below 600 nm, however, the situatioĩs reversed and one would expect that aluminum coatings would have higher damage thresholds than gold.
Thresholds for both mat?rials would be lower at 600 nm than those thresholds reported here, due to the increased absorption at shorter wavelengths. Once the laser pulse has turned off, further heat diffusion cannot reduce the (already attained) maximum temperature.
Although they were not tested in this experiment, ruled masters may have some advantage for long pulse lengths in that they do no: have the insulating epoxy layer. The use uf metallic substrates could also help for high average power applications. The substrates used here were glass.  1.0 1.5 2.2

RULED REPLICA GR.~WNG HOLOGflAPHiC GRAllNG
One possible explanation of the difference is the sharpness of the respective groove shapes, as demonstrated in figure 3. Ruled gratings are actually cut into the metal layer with a diamond tool, producing sharp corner. and whiskers of metal. Electric-field enhancement at these sharl] corners could account for lower thresholds, The ruled gratings used in tnese tests were actually replicas rather than masters, but it fs presumed that the sharp features repl~cate faithfully.

Holographic gratings are made by irradiating a photosensitive surface with an optical interference pattern.
The pattern etched into the surface is some truncated form of a sinusoid and is, therefore, smoother in sh~pe than a ruled pattern. Figure 4 shows photomicroyraphs of two different holographic grntings. The reflective metal layer is deposited on top of the exposed and developi'd photorcslst mater{al.
All of the gratinqs i,l tablu 2 were tested in first order except the ruled 600 groove/mm !]rating, which was tested in second order.

Figure 4. Photomicrographs of two holographic gratings Illustrating groove profiles.
Groove spacing on the left is 1800 k?hm and on the right is 1200 2/mm.

Higher Orders
One possible advantage of ruled gratings is their ability to go to higher orders efficiently. Frnq a laser damage threshold viewpoint, however, using higher orders is not advantageous, as can be seen in table 3.
Here, a holographic grating in first order is compared to ruled gratings in second and fourth orders.
The groove spacing and order number are such that the Llttrow angle is constant 'or the three cases.
Each of the ruled gratings was blazed for vse in the order lfsted in 0.07 ----

Polarization of the incident, beam had a substantial effect on the damage threshold of a gratinq, as seen in table 4.
For all but one of the gratings tested, the threshold was higher fur th? electric vector E parallel to the grooves than fur E perpendicular to the grooves. It is well known that the efficiency of diffraction gratfngs Is diffrrent for the two different polarizations, and that the curves of efficiency versus wavelength vary from one grating to another [2], This would imply that the detafls of the electric field distribution at the surface of the groCing differ, perhaps account-tn~for the different damage thresholds.
Typical g:'atlng efficiency curves are shown in figure 5. A simpli~tic approach of~orr~lating the efficiency of a grating with its damage threshold for the two polarizations does not work at all. We have not attempted a detailed analysis of this area, but it can be seen from figure 6 that simple metal absorption theory would predict higher damage thresholds for S polarization (E parallel to yt-ooves), in qualitative agreement with the data. Speclfylng the angle of incidence is more difficult, however, because of the angles Involved in the grooves themselves.
It is apparent thtit the higher efficiency configuration (E perpendicular to grooves) has the Icwer damage threshold.  The dependence of damaqe threshold on arnove spacina is noc cleal', In table 5, for holoaraohic gratings and E paraliel to the grooves, the~hreshold inc~eases with the number of grooves per-ml'll!meterq However, for the same gratings and E perpendicular to the grooves, threshold increases from 600 to 120G grooves/mm, then decreases from 1200 to 1800 grooves,lmm. For the ruled gratings shown, the two different po';arizations have opposite dependence on grocve spacing, One can only concludp that other factors appear to be more Important than groove spacing,  300 0.32 0.05

Master versus Replica Gratings
Two master-replica pairs of holographic gratings were tested for relative damage thresholds. For both aluminum and gold coatings and for both polarizations the replica gratings had higher thresholds than the master gratings, as shown in table 6. One possible explanation is the different types of material layers used under the reflective coating, as discussed in Section 3. Holographic masters use a phot.oresist material SIJ that the initial pattern can be "recorded." Replicas have no such requirement since they are merely taking the phys{cal shape of the master by impression in an epoxy layer.
Another possible explanation is that sharp points and corners may not replicate faithfully, but may come out more rounded.
It should be nated tt,at these replicas are actually secondgeneration or "positive" replicas so as to have the same phase as the master.
Similar tests on ruled gratings were not performed due to the considerable cost of ruled masters.  Table 7 gives the damage thresholds for two rul~d gratings mdde by two different manufacturers but atherwise identical.
Since the relative positions of the gratings changed between the two polar-fzat{onr and the thresholds are clcse to onc another, neither grating appears to be better than the other, This test was nat performed on holographic gratings.

Conclusions
The thresholds for laser-induced damage to holographic and ruled gratings have been measured and compared for 30->s pulses of 1.064 pm radiation.
For these conditions, holograph',c gratings have higher aamage thresholds than ruled (replica) gratings.
Ruled master gratings were not tested. For both holographic and ruled gratings, gold coatings have higher damage thresholds than aluminum coatings, as would be expected at this wavelength.
In general, the damage tnreshold was h~ghest for the electric vector parallel to the grooves, by a substantial factor. Unfortunately, this orientation normallv oroducus lower efficiency. ODeratino in lower orders oives the best damaae resistance, where ti$ai is !Jractical. Rulina '~ensit'v dces~ot appear to The holographic replica grating; tested-in this series master gratings. This is not fully understood. No s replicas from two different vendors.

11.
Referer,-,,~h ave~clear effect on dam~ge threshold had I igher damage thresholds than holographic gnif cant difference was found between ruled