Figure 8. Reflectivity loss of a TiO2/SiO2/HfO2 coating as a function of 248- and 351-nm uv exposure. 1. Thermal Imaging Studies Of Laser Irradiated Coated Optical Surfaces Alan F. Stewart, Adam Rusek and Arthur H. Guenther* Air Force Weapons Laboratory The detection of a localized temperature excursion induced by a laser beam incident on a coated optical surface has been previously employed to measure the optical absorption and the thermal properties of the coating. The application of thermal imaging to map coating absorption as a nondestructive diagnostic is reported in this study. Near angle light scatter was also used to generate maps of the same areas on these coatings. The data shows a limited correlation between coating absorption anomalies as determined with irradiation at 1064 nm and scattering defects identified at 633 nm. Calibration of absorption data obtained using thermal imaging was obtained from laser calorimetry measurements. Key words: absorption; calorimetry; infrared imaging; scattering; thermal conductivity Introduction A growing awareness of the potential importance of the thermal conductivity and diffusivity in the overall performance of thin film coatings for high power, high energy applications has spurred the development of theoretical models and several new measurement techniques. That a relationship should exist between the thermal properties of a coating and the laser damage threshold under continuous power loading is intuitively clear. For pulsed laser testing, recent analysis of existing and new experimental data comparing short pulse laser damage data to thin film thermal properties has demonstrated a surprising degree of correlation. [1] Lange, McIver and Guenther assumed that thermal transport in films was equal to that of the bulk material. This has been the implicit or explicit assumption made in virtually all thin film models since only bulk material properties were available. However, experimental data on thin films provided independently by Decker, Jacobs, Swimm and Ristau appears to invalidate that assumption with indications that film thermal properties can be lower than the bulk by factors up to several hundred. [2-5] With the results of these recent investigations in mind, the development of measurement techniques which could determine the thermal properties of thin films with increasing accuracy becomes even more important. The higher the level of absorption of light in thin films, the greater the need to transport heat efficiently to preclude or obviate damage. The laser calorimeter developed in the early 1970's provided the sensitivity required to measure the absorption of optical coatings with great precision. [6] In 1979, a new technique was developed for the measurement of optical absorption which was called infrared photothermal radiometry (IPTR) [7]. This technique relies on the detection of infrared thermal radiation emitted from a sample illuminated by a laser or other light source. Draggoo et al., applied IPTR to the measurement of the absorption of coated optical surfaces using a thermal imaging system and a high average power laser. [8] The technique was simultaneously applied by Ristau et al., who demonstrated the capability to measure film absorption as well as thermal conductivity. [5] *Current addresses are: (A.F.S.) Battelle Pacific Northwest Laboratories, P.O. Box 999, Richland, WA 99352. (A.R.) Dept. of Physics, Univ. of New Mexico, Albuquerque, NM 87131 (A.H.G.) Los Alamos National Laboratory, MS A110, Los Alamos, NM 87545. Using this method, they were able to show a good correlation in the data between the thermal conductivity, absorption, melting point and damage threshold. As in the earlier studies, Ristau et al., obtained values for most film thermal conductivities which were considerably lower than for the bulk material. We report on the application of the thermal imaging technique to the determination of optical absorption in thin film coatings. Thermal imaging as used here refers to the detection of infrared radiation to measure the temperature of a surface illuminated by a Our results have been directly compared to laser calorimetry data for calibration of the technique and to determine if the empirical models developed for calibration are adequate. In the course of these individual measurements, it became apparent that the local temperatures induced on these coatings varied considerably with location on the surface. Thus, thermal imaging was used to "map" coated surfaces allowing us to study the occurrence of "thermal anomalies" or "hot spots" regions of high absorption or low thermal conductivity. These surface "maps" from the thermal imaging data have been compared to scattered light intensities. The limited correlation between thermal anomalies and light scatter we have observed may confirm earlier work which emphasized the importance of testing coatings at the design wavelength. [9,10] The component layout used for thermal imaging measurements in this study is shown in figure 1. A continuous wave Nd:YAG laser system operational at either 1064 nm or 1319 nm was used to illuminate the sample. [11] The output beam was focussed onto the sample surface using a 20 cm focal length lens. The use of 2x or 4x beam expansion optics allowed us to obtain different beam diameters at the sample surface. At 1064 nm, the beam spatial profile at the sample plane was measured using a photodiode array. A series of transmission measurements through apertures of different sizes at 1064 and 1319 nm confirmed the photodiode array output. The focussed multimode beam spatial profile was circularly symmetric and nearly Gaussian. For measurements at 1319 nm, the beam FWHM with a 4x beam expander in place was 0.6 mm in diameter at the 1/e2 points in the intensity. The maximum beam power level of 75 Watts at 1319 nm was used. Surface mapping experiments at 1064 nm were performed at the same beam power level but this was obtained with different resonator mirrors and at much lower cavity pump powers. At 1064 nm, no beam expander was used and the focussed beam was 1.01 mm in diameter at the 1/e2 points in intensity. The focussed laser beam was incident on the surface of the optic under test. Individual measurements at 1319 nm were carried out with incidence angles of 5 and 15 degrees. Mapping of surfaces at 1064 nm was performed with the test surface at 6 degrees elevation from the horizontal plane. A video camera and the thermal imaging camera were located in the horizontal plane as close to the sample surface as was practical. The video camera was used to monitor the progress of the mapping experiments and to determine sample registration. thermal imaging camera had a telephoto lens with a 15 cm working distance. The infrared camera utilized a single element cooled HgCdTe detector with front end scanners to produce standard video images. [12] An electronic zoom which reduced the field of view of the scanners was used in conjunction with the telephoto optics. The resulting field of view was 5 x 7.7 mm. The software available on the imaging system allowed image data to be averaged over a small user selectable area within the field of view. The temperature data thus recorded was the average value observed within a 0.5 mm square centered on the incident beam footprint. Sixteen video frames were averaged for each measurement. Throughout these experiments, the infrared camera system operated reproducibly with 0.1 degree Centigrade resolution. Typical temperatures observed in the coatings testing ranged from 0.1 to 30 degrees or more above ambient. A shifting baseline in the measurements was caused by fluctuations in the room temperature. The laboratory air conditioning system was controlled with a 2 degree bandpass and these fluctuations were very clearly seen by the infrared camera system. This was of no consequence for individual measurements because the data was taken very rapidly compared to environmental fluctuations. However, mapping experiments required several hours to complete with as many as ten temperature cycles in evidence. The most serious limitation of the thermal imaging technique for the study of coated optical surfaces is that only coatings deposited on substrates with low thermal conductivity can be studied. For this reason, fused silica and BK-7 substrates were used exclusively throughout these experiments. The peak temperature at the center of the illuminated spot was strongly affected by heat conduction into the surrounding coating and the substrate. 3. Absorption Measurements at 1319 nm Absorption measurements taken at 1319 nm required only a few minutes to complete for each sample. The ambient coating temperature was recorded with the beam blocked. A second reading was recorded with the beam illuminating the surface. Unlike calorimetry which might require several minutes for an equilibrium temperature distribution to become established in a sample, the thermal imaging technique allows the local temperature at the surface to be recorded almost instantly. The onset time for the surface temperature profile could not be resolved using this imaging system but would provide potentially useful data about thermal transport in the coating and the surface. The absorption of high reflectance coatings designed for operation at 1319 nm was measured using the thermal imaging system. It was during the course of these measurements that it became apparent that there were regions on these coatings with anomalously high absorption or low thermal conductivity. Both average surface temperatures and peak surface temperatures were recorded. The samples used in this survey were several different types of coatings and included all-dielectric stacks and enhanced metal designs. Substrates were fused silica in three sizes - 2.54 cm diameter by 0.0254 cm thick wafers, 3.85 cm diameter by 0.95 cm thick, and 15.25 cm diameter by 2.54 cm thick. Coatings of the same type were deposited on identical substrate sets in three different runs. Since absolute measurements of coating absorption were the objective of this test series, the emissivity of these coated samples was measured. Each sample was placed in an oven with an open port in the side for line-of-sight measurements. The measured temperature obtained from the image from the infrared camera was compared to the actual temperature to derive the emissivity. For the all dielectric reflectors, emissivities ranged from 0.83 to 0.89. For enhanced metal reflectors, emissivities ranged from 0.05 to 0.09. Absorption in the thin coated wafers was also measured using a second independent method to assist in a calibration of the thermal imaging technique. Laser calorimetry using the system described in reference [13] was performed at 1319 nm. Measurements were performed only on the wafers due to limitations on sample mass, characteristic of the calorimetric technique. 4. Analysis of Measurements at 1319 nm Data obtained in this survey of high reflectance coating designs was analyzed using the theoretical relationship derived in reference [8]: Absorption = 0.045* t1.11 0.83 / p1.05 where t is the peak temperature difference in Centigrade r is the beam radius in cm p is laser power in Watts (flat topped beam profile) Each data point appearing in figures 2-4 represents the average of measurements at three different positions on the sample surface. The measured emissivity data was used to correct the measured or apparent temperature to obtain the actual surface temperature. For the dielectric coatings tested in figure 2, the inclusion of emissivity amounts to an upward correction of 16% in the thermal imaging data. The same correction factor exists in figure 4. In figure 3, the emissivity is very small for these enhanced metal reflectors and hence the "correction" is large multiplication by a factor of 13 over measured temperature readings. There are several striking features in the data in figure 2. The absorption measured using thermal imaging on parts of very different aspect ratios is practically unchanged. While the coating design is the same for all the samples tested in this figure, the deposition conditions derived for series 3 a-c were used to deposit series 4 a-c and 5 a-c. It is clear that trends in the thermal imaging data are replicated in the calorimetry data although the magnitudes differ by a factor of 2 to 4. The emissivity correction for these coatings is much smaller in comparison to these overall differences. |