Manuscript Received Measurement of the Three Photon Absorption Cross Section and Lin Simpson, X. A. Shen, Scott C. Jones and Peter Braunlich Washington State University and Paul Kelly National Research Council, Ottawa, K1A 0R6, Canada The self trapped exciton recombination luminescence (STERL) technique is used to measure the three photon absorption cross section in KI at 532 nm. STERL is used as a thermometric property to obtain the laser induced temperature increase, up to values near the melting point, before sample damage occurs. Computer modelling of the results indicates that the intrinsic damage process is absorption of light by three photon generated free carriers, according to the mechanism reported for KBr by Shen et al.[1]. Key Words: alkali halides; bulk damage mechanism; exciton recombination luminescence; 532 nm laser; free electron heating; potassium iodide; self-trapped exciton; three photon absorption cross section. 1. Introduction The self-trapped exciton recombination luminescence (STERL) technique of measuring multiphoton absorption and the resulting laser induced temperature increase was devised by Shen et al.[1-3] to study intrinsic bulk laser damage in K Br at 532 nm. Using this method it was shown that bulk damage to high purity samples is a result of secondary energy absorption by four-photon generated free electrons, causing lattice melting. The purpose of this work is to validate the carlier results by repeating the experiments and calculations described in references [1-3] in a different material; potassium iodide. Potassium iodide exhibits the self-trapped exciton (STE) phenomena in common with all alkali halides, in particular the two-band emission spectrum and temperature dependent luminous efficiency. However, KI presents a three-photon band gap at 532 nm, thus the free electron generation will be of third order. In order to perform the computational analysis, the three-photon absorption cross-section is required. This information is obtained by the methods described in ref.[4] except the system is calibrated using the third harmonic of Nd:YAG, 355 nm (two 355 nm and three 532 nm photons have equal total energy). The result is (3) = (6±2.8) × 10-81 cm6 sec2, where o(3) is defined in the equation Here ne is the free electron density, F is the photon flux (photons cm density. 2. Induced Temperature Measurements The laser induced temperature measurements are carried out in precisely the same manner as used by Shen et al. described in references [1] and [3]. The description of the experimental apparatus and procedures are contained therein. First the luminous efficiency is obtained by monitoring the relative emission output following stimulation by two-photon (355 nm) absorption. The crystal temperature is ältered while the stimulating beam remains at a constant low (non-heating) intensity. The results are presented in figure 1. The data fits the equation where n is the efficiency, with TV = 5200 and E = 0.069 eV. Thermal quenching begins at T = 70 K. -2 sec (2) -1 Armed with this information, the crystal is then irradiated with 532 nm photons while the cryostat maintains the initial crystal temperature at 50 K. The energy of the 100 ps pulse is increased by small increments, beginning with low intensities, which just yield measurable luminescence, up to damaging levels. One such run is shown in fig. 2. At relatively low levels of photon flux (below 8 x 1028 photons cm peak flux) the luminescence signal follows a third-order dependence on flux (slope of three on the doublelogarithmic plot), indicating three-photon absorption as the carrier generation process. The curve begins to bend over (i.e., the dependence falls below third-order) when the center of the focused Gaussian beam reaches the thermal quenching temperature, T = 70 K, because the luminescence efficiency begins to decrease from a constant level. In this experiment, the center of the irradiated volume thus reaches a temperature of 70 K (AT = 20 K) at F = 8 x 1028 photons cm-2 sec1 (see fig. 2). This yields one calibration point for the luminescence thermometer. -2 -1 Computational modelling [3,4] demonstrates that the temperature rise depends on the photon flux in the third order, and with the fixed point at T = 70 K (AT = 20 K), F = 8 x 1028 photons cm sec-1, the temperature at the center of the focal volume resulting from a given pulse can be calculated (fig. 3). This assumes no new processes become involved in the interaction, e.g., avalanche ionization. Barring this, the peak lattice temperature for the most intense nondamaging pulse we observed was 890 K, 50 degrees below the melting point. This occurs at a peak flux of 2 × 1029 photons cm-2 sec1 (≈ 75GW/cm2). If another carrier generation process were operative, the expectation is that the crystal would heat more rapidly than shown in figure 3, and thus damage would occur at a lower temperature than that inferred in fig. 3. Thus we conclude that no additional processes occur. 3. Conclusion The experimental results given here do validate the mechanism of intrinsic damage of alkali halides by visible light discovered by Shen et al. Free electrons are generated according to eq. 1 with o(3) = (6 ± 2.8) × 10-81 cm3 sec2, and these carriers subsequently absorb light energy according to the theory of Epifanov [5] inducing lattice heating expressed by 6 with small contributions (< 10%) by laser generated lattice defects. In the above, m* is the electron effective mass, k Boltzmann's constant, lac and v, are the electron mean free path and speed of sound respectively, e the electron charge, w the laser frequency and E the electric field amplitude. Again, we see no evidence for the existence of electron avalanche impact ionization in the damage process. This work was supported by the U.S. Air Force Office of Scientific Research under Grant No. AFOSR87-0081. 4. References 1. X. A. Shen, Peter Braunlich, Scott C. Jones and Paul Kelly, Proceedings of the 19th Boulder Damage X. A. Shen, Peter Braunlich, Scott C. Jones and Paul Kelly, Phys. Rev. Lett. 59, 1605 (1987). 2. 3. X. A. Shen, Peter Braunlich, Scott C. Jones and Paul Kelly, Phys. Rev. B 38, 3494 (1988). 4. X. A. Shen, Scott C. Jones, Peter Braunlich and Paul Kelly, Phys. Rev. B 36, 2831 (1987). 5. A. S. Epifanov, Zh. Eksp. Teor. Fiz. 67, 1805 (1974) [Sov. Phys. JETP 40, 897 (1975)]. Figure 1. Luminescence efficiency (7) vs temperature (T) in KI. The curve is given by eq.(2). Thermal quenching begins to occur at T 70K. As the temperature increases, the luminescence continues to decrease even though the laser intensity is the same for each point. 1000 Figure 2. Experimentally determined and numerically calculated dependence of luminescence yield on incident peak photon flux. The log-log slope of three in the lower par of the curve is indicative of the three photon process we are studying here. At about 8 x 1028 photons cm3 sec-1 incident flux we observe a deviation from the slope of three. Here the quenching temperature at the interaction volume's center has been reached and thus temperature effects become important above this flux. The curved part is due to a decrease in luminescence efficiency. The theoretical curve was generated using the Epifanov free carrier heating mechanism and the end point at the top of the curve indicates a temperature approximately 50 K lower than the KI melting temperature of 954 K. Manuscript Received 1-9-89 DAMAGE TO SILVER COATINGS FROM HIGH AVERAGE POWER 1-um LASER V. Sanders, L. Jolin, and S. Salazar Los Alamos, NM 87545 Optical coatings are being developed for use in high average power, pseudo continuous wave, 1-μm lasers. In particular, the subject of this study is a silver coating to be applied to a cooled substrate for use in a radio frequency-free electron laser (RF-FEL). Damage threshold measurements and finite element thermal calculations were made that demonstrate the pure thermal nature of heating, and eventual damage, from this RF-FEL pulse format. Key Words: silver coatings, damage thresholds, thermal damage, radio frequency-free electron laser, cooled Introduction The rise in temperature to eventual melt of a silver coating from exposure to a continuous wave, 1-μm wavelength, highaverage power laser is caused by pure thermal equilibrium heating from absorption of light. However, the subject silver coating in this study is being considered for use in a 1-um RF-FEL where the pulse format is referred to as pseudo continuous wave. The purpose of this study is to determine if the train of picosecond pulses from the RF-FEL interacts with the silver coating as if it were a true continuous wave exposure. Therefore, the experiments that were conducted and will be described were fashioned with an emphasis on demonstrating the physical nature of the minimal melt damage threshold. Experiments and Calculations Three experiments were conducted each of which included a controlled variable to be compared with damage threshold measurements. These measurements were either compared with calculations or compared with each other to indicate a scaling dependence on the controlled variable to determine the physical nature of the damage. The damage threshold experiments were conducted using a simulation of the RF-FEL pulse format. The simulator was constructed from a mode-locked Nd:YAG laser directed through a series of four amplifiers, two active pulse-shaping/slicing attenuating optical devices, and two passive Faraday rotation isolation optical devices. The goal was to create the constant amplitude pulse train indicated in figure 1. The results were: a train of 100-ps pulses at 100 MHz with approximately a constant peak-to-peak amplitude (within ±5%) on the 100-ps pulses for 100 μs with negligible energy associated with post-pulse ringing. The beam was plane polarized to better than 300:1. The damage threshold determined, using this pulse format, was one associated with a statistical zero probability of creating damage. Damage is indicated by any minimal microscopic visible indication of disruption of the exposed surface. For the purpose of this study we are assuming minimal surface melt, or near melt, at the surface of the coating as the physical nature of the observed disruption. In the first of the three experiments mentioned above, the controlled variables were the substrate materials and the thickness of the silver coating. The coated substrates were Si, Mo, and SiO2. These three substrates were coated at the same time and in close proximity to each other. This ensures the same thickness and coefficient of absorption for each coating. A damage threshold was determined for each. This was done for three particular silver coating thicknesses: 0.2, 2.0, and 20 μm. The data is graphed in figure 2 in a comparative fashion to each other and compared with a corresponding calculation. Keep in mind that the goal of this study is to determine the physical nature of damage. Therefore, we have plotted ratios of damage threshold measurements on the corresponding coatings. For example, the dot in the upper left portion of the graph indicates that the damage threshold for the 0.2-μm-thick silver coating on Mo substrate was a ratio factor of 2.0 greater than the same coating on Si. Likewise, for this same coating, the damage threshold comparing SiO, and Si as the substrates, the damage threshold for SiO, is only 0.3 that of Si. |