Question: Answer: Question: Answer: Question: Answer: Question: COMMENTS You show damage of polycrystalline calcium fluoride, mag fluoride, and lithium fluoride. It's hard to get smooth ten angstrom type surfaces of crystal and lithium fluoride. You also stated that high defect densities on lithium fluoride and barium fluoride led to multi-photon processes rather than an avalanche. Do you know what the surface quality, in terms of roughness was of the lithium fluoride as compared to the calcium and barium fluoride? These are commercially optically high grade pieces. More, I don't know. But, I would like to make the point that the barium fluoride and the lithium fluoride have 102 more solubility in water compared to magnesium fluoride and calcium fluoride. This fact is what made me suggest doing the talk that we believe that lithium fluoride and barium fluoride have a large number of defect states on which water can adsorb. However, I am aware of the fact that polished lithium fluoride has a lot of defects introduced by the polishing process. I still am not sure that I understand completely this idea that you need a large Yes, but you have to take into account the high density, you see. The electrons are predominantly in a given spin. And even if you have defect states, they are comparatively very low concentration compared to the density of states in the given spin. So once you have some states available that can promote the multi-photon process, the high density of electrons in the given spin state would favor that process. Am I making myself clear? We'll talk about it some more later. Because if you have occupied defect states on the surface, their density will be in any case very much lower than those electrons in the given spin state. I also have difficulty of being convinced of multi-photon process here in lithium fluoride. A number of years ago, we measured the damage threshold of bulk lithium fluoride and found out that it was loaded with bulk defects which we concluded were non-stoichiometric sites. I would assume that those sites would exist on the surface as well, and that they have their own absorption spectrum and one photon could easily cause damage. And so, if you have those, you don't have to invoke multi-photon absorption to damage non-stoichiometric material. The point is that if you are talking only about 1-photon process, you are most likely only making heat. Now I did not show those morphology pictures, but from those, it's clear that there's no melting or so going on. We surely will have some heat generated, but we think it's a small amount that goes into that channel compared to the other one. Now you questioned the multi-photon process as well. There are several indications here, for example, when you measure by the fluorescent light from these breakdowns, it's a 5-photon dependence. When we measure the desorption, it's a 5-photon dependence. And of course, the first idea here was not to see the multi-photon process, in fact, we were surprised to see slopes like 5, which would sort of bridge the band gap, but I don't see any other interpretation and what we will be following now is this transition from multi-photon to the low defect dielectric breakdown. All I can say is, come up with another interpretation about the slope, and we can discuss it. First I want to compliment you on having the foresight to look at cleave surfaces in comparison with polished surfaces. But, I'd like to comment that I know from lots of work, that if you look at the dislocation content of crystals, particularly lithium fluoride, and you can do this easily by making etch pits Answer: counts, you find that the cleave surfaces have very few dislocations. The polished surfaces have a large variation in dislocation density, depending on the person and the technique used to polish them and that the roughness crystals such as you would get by using sand paper have extremely high surface dislocations. Have you looked at this, and if you have, what do you make of this dislocation possibility? We have not looked at that yet in the systematic manner and I'd like to emphasize again that these are qualitative data. I'm very much aware of the fact that we have to quantify the number of defects on the surface in order to make this case a little bit better. But, give us a year or two and we might be able to do this. What we have done up to now is only a qualitative measurement. But, I thank you for the suggestion. Manuscript Received Laser-Induced Surface Ablation and Optical Damage of ZnS Crystals H. F. Arlinghaus," W. F. Calaway, D. M. Gruen Materials Science, Chemistry, and Chemical Technology Divisions L. L. Chase Lawrence Livermore National Laboratory P.O. Box 5508 Livermore, CA 94550 Velocity distributions and yields of neutral Zn atoms emitted from laser-irradiated ZnS crystals at power densities far below the single-pulse damage threshold have been analyzed by high-resolution two-photon laser-induced fluorescence spectroscopy and also by electron impact ionization mass spectrometry. Large differences in the measured kinetic temperatures have been found between the single- and multiple-pulse laser irradiation experiments. The high-kinetic temperatures, obtained in multiple-pulse experiments, may be caused by cumulative surface modification, such as thermallyinduced cracking, leading to a reduction of the thermal conductivity compared to the bulk value. Optical damage was related to the interaction of a plasma formed at a critical combination of particle density and laser intensity, with the surface. Key words: Laser damage, laser fluorescence spectroscopy, particle emission; temperature, threshold, time-of-flight, vaporization, zinc sulfide. 1. Introduction The ability to produce optical materials that can transmit or reflect intense laser radiation with-out suffering surface damage requires the identification of the initial mechanism for deposition of laser energy at the surface and an understanding of the resulting physical processes that precipitate catastrophic material failure. Despite its importance, the nature of this damage is still not well understood. In transparent optical materials, for example, there are several possible interaction mechanisms that could cause absorption of laser energy at the surface. The problem of detecting these surface absorption processes is formidable because they may be very localized and cause small unobservable changes in the transmitted or reflected laser intensities and average surface temperatures [1]. To get more information about the interaction phenomenology, the effects of laser radiation on materials, and the processes that cause laser damage, we investigated sputtered particles originating from laser-irradiated surfaces with fluences far below the accepted damage threshold. These particles are characterized by their yield, kinetic energy and angular distributions, Work performed under the auspices of the U.S. Department of Energy, BES-Materials Sciences, under Contract W-31-109-ENG-38 (ANL) and Contract W-7405-ENG-48 (LLNL). Present address: Atom Sciences, Inc., 114 Ridgeway Center, Oak Ridge, TN 37830 relative abundance, and spatial origin. Chase and Smith, and Arlinghaus et al. [1-6] have found that, in general, most of the particles sputtered at laser fluences far below the damage threshold are neutral atoms or molecules, and have shown that this particle emission is a precursor to observable damage. Last year at this conference Chase and Smith presented time of flight (TOF) data on the emission of Zn from single-pulse irradiated ZnS single crystals using an electron impact ionization mass spectrometer (MS) as a detection system [1]. As shown in figure 1, they observed Zn emission occurring for a time interval of several hundred microseconds. At fluences above 8 mJ/cm2, a peak develops at a delay time of about 250 μs, which is attributed to promptly emitted Zn particles with a kinetic temperature of less than 1000 K. As the damage threshold is approached; however, the characteristic temperature increased to several thousand Kelvin. Similar data has been obtained at 532 nm, 355 nm, and 266 m and for the emission of S and S2 [1]. With decreasing wavelength, the general characteristics of the TOF distributions are unchanged, which suggests that the same processes are involved in the emission at all four laser wavelengths. In this work, we used 308 nm laser irradiation deposited on ZnS to study the above described phenomena in more detail. This photon energy lies between the third and fourth harmonic of the Nd:YAG; thus, direct comparison to the above described single-pulse experiments should be possible. An experimental technique, which has proven to be more sensitive than electron impact ionization MS and is particularly useful in measuring velocity distributions and particle densities of ejected neutral particles, is laser-induced fluorescence spectroscopy (LFS). To determine the velocity distribution for laser-ablated particles from nanosecond laser pulses, TOF measurements have been performed. This is accomplished by detecting the ejected particles at a known distance from the target at various times after the ablating laser strikes the target. The capabilities of this type of experiment have been expanded in our apparatus by incorporating a high-resolution (singlemode) tunable laser into the LFS system in order to obtain Doppler-shifted velocity profiles of ablated atoms as a function of TOF. Using the velocity determined from the Doppler shift and the measured TOF, the time at which an atom was ejected from the surface can be calculated. Thus, it is possible to separate prompt from delayed emission of ablated atoms, as well as to probe possible molecular or cluster fragmentation. This approach of determining velocity distributions, both from Doppler shift and TOF, was thought to be required to investigate delayed emission of atoms as seen for the singlepulse experiments described above. To estimate the amount of material removed with each laser shot and the number of shots needed to remove one monolayer, the absolute yield of Zn atoms was determined by comparison with LFS measurements of Zn atoms sputtered from pure metal targets by Ar+ ions. An important problem in laser-induced optical damage is the dependence of the damage threshold on the number of laser shots striking the surface. Significant reduction in the damage threshold over the single-pulse value appears to occur as a result of multiple-pulse operation [7,8]. This problem was examined by measuring the velocity distribution and the neutral yield from ablated particles as a function of the number of laser shots. 2. Experimental Technique The experimental arrangement is shown in figure 2. It consists of an ion gun, an ablating laser, an LFS detection system, an ultrahigh vacuum (UHV) chamber (typical base pressure 3x10-8 Pa), and computer-controlled electronics. The experimental apparatus was recently described in detail [4,5] so only certain key features will be discuss in this account. The interaction of the ablation laser with the sample causes atoms to be ejected from the surface. After a specific delay time, a narrowband probe laser (85 MHz bandwidth, 1.3 mJ/pulse, 15 ns pulse) is triggered. Its wavelength is set to twice the wavelength of the Zn transition, 4s2 1So - 4d 1D2 (160 nm). The induced fluorescence is detected at 636 nm (4d 1D2 - 4p 1P01) by light collection optics that image the fluorescence from the probed volume, a known distance from the sample through a narrow bandpass interference filter onto a photomultiplier. The frequency of zero velocity for Zn was determined by a two-photon Doppler-free experiment [9]. |