1. Manuscript Received Investigation of the Effect of Surface Finish D.W. Mordaunt and D.E. Maguire Hughes Aircraft Company Electro-Optical and Data Systems Group El Segundo, CA 90245-0902 1-13-89 This paper reports on the effect of varying surface polish and fabrication techniques on the laser damage threshold of uncoated Nd: Cr:GSGG slabs. A series of GSGG samples were prepared with a uniformly high quality surface polish. Chamfers were applied to these samples using a standard 220 grit grind, a window polish and a laser quality polish to finish the surface of the chamfer. The chamfers produced by each technique were analyzed for the degree of subsurface damage produced during fabrication. Samples of each method were damage tested in the chamfer area and on the polished surface. The test results are presented and compared to typical results for antireflection coatings on GSGG laser rods. Implications for the use of Nd: Cr:GSGG slabs in high power laser systems are presented. Key words: Introduction laser damage; Nd:Cr:GSGG damage thresholds; chamfer damage. In recent years, Nd: Cr:GSGG has matured as a laser material to the point where it can be considered as a substitute for Nd:YAG in production laser systems. The higher efficiency of GSGG compared to YAG makes this material very attractive for applications where size, weight or operational lifetime are important. Hughes currently has an internal R&D program on solid state laser technology aimed at developing new materials and scaling of solid state laser technology to higher energies. During this research program, we developed a GSGG oscillator/amplifier configuration using rectangular slabs of GSGG for the amplifiers. For this configuration, the oscillator output is expanded and overfills the amplifier slabs. The oscillator beam is amplified once on the first pass, reflected and fully amplified on a second pass through the amplifier slabs. At the end of the second pass the energy in the laser beam is at a maximum and is still filling the entire aperture of the amplifier slabs. During experiments designed to scale this configuration to energies in the 5 to 10 J range, we observed damage to the beveled edges or chamfers of the amplifier slabs. This damage occurred because the GSGG amplifier slabs have a nonuniform gain profile with the highest gain occurring at the edges of the slab closest to the flashlamps. This result was not particularly surprising when the high absorption of this material is considered. When combined with diffraction effects, this nonuniform gain causes the highest fluence in the entire laser system to occur at the edges of the amplifier slabs as the amplified beam completes its second pass through the amplifier. This high fluence resulted in damage to the chamfer, accompanied by ablation of material off of This ablated material can then deposit on the antireflection (AR) coated end faces of the amplifier slabs and cause catastrophic surface damage and a resulting failure of the laser system. In order to scale this configuration up to higher energies, the chamfer damage threshold needs to be increased to the point where it is at least as durable as the AR coatings on the end faces of the slabs. This paper reports on a series of damage test experiments performed in order to determine how the fabrication method and degree of surface polish applied to the chamfers affects the damage threshold. We report on two sets of experiments. The first series of tests were performed on samples provided by Litton Airtron of Morris Plains, NJ. These samples were used to provide a baseline for typical performance of coated and uncoated Nd: Cr:GSGG substrates. Two of these samples were produced with a normal ground chamfer, two with a polished chamfer and two with acid etched chamfers. Because the results of these experiments were not conclusive, an additional series of GSGG substrates were fabricated under known and well controlled conditions. These samples consisted of uncoated or AR coated substrates with chamfers which were fabricated with either a normal 220 grit grind, a low quality window polish or a high quality surface polish similar to that which would be applied to the face of a laser rod. The results of all of these experiments are presented in the remainder of this section. A layout of the damage test facility used for these experiments is given in figure 1. The test station consists of a Nd:YAG based laser system which uses an oscillator/amplifier configuration and a phase conjugate mirror to produce a high quality output at a wavelength of 1.06 μm. The beam quality is quite good and consists of a 9 mm diameter circular beam with approximately 250 urad beam divergence. All of our damage test measurements were performed using a 1.06 μm beam and typical test conditions consisted of a 15 nsec pulsewidth (FWHM), a pulse repetition rate of 1 Hz and a 1 mm diameter spot size in the target plane. All tests were performed as unconditioned tests in which each test site on the surface is exposed to a single fluence level. Each site was exposed to at least 20 shots and damage was defined as any permanent change in surface scatter when the test site is illuminated by a He-Ne laser and viewed online under 20x magnification. Preliminary testing was performed on samples provided by Litton Airtron of Morris Plains, NJ. Two coated samples with different coating designs were provided as well as two uncoated GSGG pieces which had the chamfers fabricated by three different techniques. The damage test results are presented in table 1 and are summarized here. The coated samples and the polished surfaces of the uncoated pieces had damage thresholds in the 20 J/cm2 range. The chamfers of the two uncoated pieces were fabricated using three different manufacturing techniques: a 220 grit grind, a polished chamfer and a polished chamfer which was then acid etched. The chamfer test results were not conclusive and showed little difference between the chamfer damage thresholds for the three different fabrication techniques. All of the chamfers had damage thresholds between 5 and 8 J/cm2, although one might expect that the polished chamfers would exhibit behavior similar to the polished faces of the samples. Two main factors tend to cast some doubt on these test results. First, we did not have any control and were not able to adequately specify the surface finish of the chamfers and the possibility is quite strong that they were not polished to the same degree as the faces of the GSGG pieces. Although they were not characterized quantitatively, we feel that the surface quality of the polished chamfers was not up to the quality of the polished faces, which had a typical laser rod type finish applied to them. Second, the chamfers were quite small in size with a diameter of only about 0.15 to 0.20 mm while the laser damage test beam had a diameter of approximately 1 mm. As a result, the laser test beam overlapped the chamfer on the polished surface and also on the outside edges of the chamfers. Because the first round of testing did not produce conclusive results, a more carefully controlled series of experiments was designed. For these tests, GSGG substrates were supplied to Lightning Optical Corp. of Tarpon Springs, FL, where they were polished and coated in a prescribed and well controlled manner. Two coated samples were provided with normal ground chamfers and a 1.06 μm AR coating on one face. In addition, nine uncoated pieces were provided with polished faces and three different chamfer fabrication techniques. For these nine samples, a large chamfer with a width of approximately 2.5 mm was applied to provide a chamfer which was wider than the laser spot size used for the test. This allowed reliable positioning of the beam on the chamfer without overlapping either edge. Three samples were provided with a normal 220 grit ground chamfer, three with a low quality "window polish" and three with a high quality surface polish similar to that which would normally be applied to the polished end face of a laser rod. The results of the second round of damage testing are presented in table 2. This data follows much more of an intuitive trend than the previously tested samples. The coated parts were consistent with the previous test samples and had damage thresholds around 20 J/cm2. The uncoated polished faces of the samples were consistently around 17 J/cm2. The chamfers fabricated with a 220 grit grind were also consistent with previous results and averaged around 5 J/cm2. With these samples a large improvement in the chamfer damage threshold was obtained by polishing the chamfer. Both the window polish and the high quality laser rod type polish showed substantial increases in the damage threshold, with the good polish exhibiting the largest increase. The window polished chamfers averaged around 14 J/cm2 while the highly polished chamfers averaged around 16 J/cm2, which is essentially equivalent to the values measured for the polished faces of the GSGG samples. The lower thresholds measured for the window polished chamfers were due to a large amount of small surface scratches and subsurface damage which were not removed during the polishing process. Both the window polish and the highly polished chamfers looked very clean when observed visually under a bright light without magnification. In both cases no visible scratches or pits were observed. However, a large difference between the two fabrication methods was observed under 100x magnification using a Nomarski phase contrast microscope. The difference between the two surfaces is presented in figure 2. Under these observation conditions, the highly polished chamfers still appeared virtually featureless while the window polish showed a grid of fine scratches which had not been fully polished out. We believe that these scratches were the reason for the lower damage threshold for the window polished chamfers. A similar phenomenon may have caused the inconclusive results for the samples tested during the first round. The polished chamfers from these parts were also examined at 100X under the Nomarski microscope and showed a grid of very fine scratches similar to that observed in figure 2b. 3. Conclusions In conclusion, we have shown that the damage threshold observed for the chamfers of polished GSGG substrates is dependent on the fabrication conditions. In the context of addressing an engineering problem in the development of a medium energy Nd: Cr:GSGG laser source, we encountered severe problems with damage to the GSGG amplifier slabs. By controlling the chamfer fabrication conditions and applying a high quality laser rod type finish to the chamfers, we were able to increase their damage thresholds by greater than a factor of 3. The polished chamfers exhibited damage thresholds equivalent to the uncoated polished faces of the GSGG pieces at a level of about 17 J/cm2. These results provide a means of controlling the chamfer damage in real laser hardware systems and show the importance of carefully controlling the fabrication conditions of laser materials when laser induced damage is a critical issue. Table 1. Results of the preliminary damage testing.on the samples provided by Table 2. Results of the damage testing on the samples provided by Lightning Figure 1. The laser damage test facility consists of a Nd:YAG oscillator/amplifier configuration. The abbreviations are defined as follows: OSC = oscillator; AMP = amplifier; EXP = beam expanding telescope; TFP = thin film polarizer; 1/4 = quarter wave plate; 1/2 = half wave plate; and DET energy detector. Figure 2. The two photomicrographs shown below compare the surface quality of the (a) highly polished chamfers and (b) the chamfers with the window polish. Both pictures were taken at 100X magnification using a Nomarski phase contrast microscope. COMMENTS Question: Answer: Question: Answer: Question: Answer: Question: Answer: Did you try to characterize the sub-surface damage and if so, how did you do it? What's the difference between the damage threshold of the polished area and of On the coated surfaces, we cut some pieces off the undamaged areas and the coatings actually tested around 40 joules per sq. centimeter. We didn't test the chamfers on those, and I don't know the exact fluence that the system was running at So I can't really tell you how that compares with the damage threshold. You still have a line of intersection between the chamfer and the face, what about damage there? saw The damage was pretty much catastrophic on the chamfer, typically what we What I was referring to was if you solved the problem of damage on the chamfer and you have no damage on the face, you still have an edge that's essentially not as high quality and I'm wondering whether fixing chamfer problem really solves the damage problem. I would say it's got to be an improvement, I don't know if it's going to solve all of our problems. |