COMMENTS ON PAPER BY NORBERT NEUROTH

In response to questions from the floor, a number of points of experimental procedure in connection with this investigation were clarified. It was established that the pumped samples were protected from the ultraviolet radiation of the flash lamp by enclosing the sample in a supremax tube. During damage testing which was carried out on a single shot basis, air cooling was employed; while water cooling was employed during rapid pulse sequences. The output of the test laser was not subject to transverse mode control. In the tests the full output of the laser was used. Two methods were used to measure the peak power density in the laser output. One was the observation of the threshold for the formation of an air spark at the focal point of a lens. The other was a technique whereby an iris was placed in the beam of the laser and the aperture diameter at which half of the total energy was transmitted was determined. From this measurement the peak power density could be estimated from the measurement of total energy and pulse duration. The two methods agreed to within 30%. The definition employed for damage threshold was the following: the damage threshold was defined as the lowest power density at which visible damage of the sample was observed with the unaided eye. Some spread in the results was observed. A much greater variation in the observed damage threshold was seen in filter glasses than in laser glass. This was attributed to the much higher degree of reproducibility required in melting laser glass. For a given sample of laser glass the observed damage threshold was seen to depend on focal length of the lens employed in the test and for a given focal length and given sample, damage thresholds were observed to be reproducible to within about 35%. The difference in the damage threshold observed in active tests and passive tests was attributed in part to thermal stresses induced by the pump light in the active measurements, and in part to the self-focusing of the incident laser beam in pumped samples, which in turn raised the local power density. The lower extinction coefficient quoted for the new Schott laser glasses was attributed primarily to a reduction in the iron content of the glass but also, possibly, to a decrease in scattering losses. Two results for damage with picosecond pulses were quoted by the speaker. Jean Davit at C.G.E. has reported a damage threshold in a picosecond regime of 20 joules per square centimeter for filamentary damage. This result was somewhat dependent on sample size and was reported in the Proceedings of the 1970 Symposium on Damage in Laser Materials. Recently, damage well above threshold has been observed at the Max Planck Institute in Munich at power levels of 50 gigawatts per square centimeter in 10 picosecond pulses.

Laser Glass Damage Threshold Studies At Owens-Illinois

N. L. Boling and R. W. Beck

Owens-Illinois Inc. 1700 N. Westwood Ave. Toledo, Ohio

Owens-Illinois is continuing its efforts toward an understanding of damage
induced in laser glass materials exposed to high intensity laser pulses. While
previous efforts have been directed toward a characterization of the magnitudes of
the threshold damage due to particulate inclusions, surface interactions, and self-
focusing, current work concerns itself primarily with an investigation of the
mechanisms responsible for surface damage. To this end, a high intensity, TEM 00
mode, neodymium glass, oscillator-amplifier system is used to induce surface damage
in damples of laser glass, while a TEM mode ruby laser is used to record high
speed holograms of the damage interactions. This technique not only allows an in-
vestigation of the plasma produced at the sample surface but also detects the
creation of shock waves within the sample. With various time delays between the
damaging and the recording laser pulses it is possible to evaluate the temporal
development of both the plasma and the shock wave. By using multiple exposure tech-
niques further information is obtained from the recorded refractive index variations.

A brief discussion of the glass and ruby lasers as well as the holographic technique will be presented with primary emphasis placed upon the interpretation of the recorded plasmas and shock waves and their significance to laser glass surface damage.

Key Words:

Holography, inclusion damage, plasma and shock waves, and self

focusing surface damage.

1. Introduction

The continuing effort at Owens-Illinois to further increase the damage threshold of laser glass has been divided into two areas, inclusion damage and surface damage. The problem of inclusion damage, having previously been characterized in this and other laboratories, is being approached at the glass melting stage, the goal being to completely eliminate all damaging inclusions from the finished laser glass. Surface damage is being studied from two complementary aspects, investigations of mechanisms responsible for surface damage and studies of the efficacy of various treatments of the glass which might increase the damage threshold. The problem of self trapping or bulk damage is under investigation only to the extent that it is related to surface damage and/or the experimental technique used for studying surface damage lends itself to this end. High speed holography is the technique employed to study damage. The holograms obtained yield information about acoustic disturbances in the glass and plasmas on the surface immediately after the passage of a damaging pulse through a glass sample.

The laser employed for inducing damage is a high power, Q-switched, glass laser. It is currently undergoing modification to yield an output in the TEM o mode.

2. Inclusion Studies

Several types of inclusions have been hypothesized to cause damage to laser glass but the most likely suspect in glass melted in platinum crucibles is platinum inclusions. Consequently, in attempting to prevent the occurrence of inclusions it is proper to address the question of possible mechanisms by which platinum can enter the glass during the melting process. Many such mechanisms have been suggested but the most probable culprit is oxygen in the melting environment. Obviously, then, the thing to do is to eliminate oxygen from the melting environment. However, this can be done only to a degree with glass melted in a platinum crucible. When the partial pressure of oxygen is lowered too

tar, the platinum of the crucible is attacked by components of the glass, silicon being the first to react with platinum as the oxygen partial pressure is lowered. The trick is to lower the oxygen content far enough to prevent formation of platinum inclusions in the glass but not so far as to cause crucible attack.

During the past year we have theoretically examined the thermodynamics of the melting of laser glass in a platinum crucible under a reduced partial pressure of oxygen. These studies have indicated that it is indeed possible to strike the happy medium between introduction of platinum into the glass and crucible attack. The desired partial pressure of oxygen can be attained and maintained through the use of CO-CO2 as a purging gas during melting. This acts as a buffer which controls the amount of oxygen within relatively narrow limits.

In the near future, melts will be made utilizing CO-CO2 as a means of oxygen control. from these melts will be analyzed for platinum and damage tested by a high power laser beam.

The glass

3. TEMoo Mode Laser

In the past many studies of surface and self trapping damage have been conducted by various groups. Some of these groups have used low energy output lasers operating in the TEM o mode. The laser beam is focused in order to obtain the power densities required to cause damage. Other investigators have used high energy output lasers operating multimode. Several objections can be raised to either of these procedures. The validity of these objections is suggested by the fact that wide discrepancies exist among reported damage thresholds.

Because of these various objections raised to focused or multimodal systems we have decided to modify the glass laser system which we have been using for damage studies at Owens-Illinois. Until recently this oscillator-amplifier system was operated multimode and was capable of delivering more

than 100 joules in a 40 ns pulse. The modification, which is still in process, will result in a system which operates in the TEMoo mode with an output of 40-50 joules in 30-50 nanoseconds. The oscillator and two amplifiers of this system have been assembled. A diagram of the oscillator is shown in Figure 1. The Brewster-Brewster rod is of 3/4" diameter x 12" length, with 9" pumped by two close wrapped helical flashlamps. The energy to each lamp is nominally 3500 joules. The TEM。。 mode is achieved through the use of a 2.0 mm aperture in the cavity. The output of the oscillator is approximately 250 mj.

A plano-plano mirror configuration is utilized in the oscillator, although it is expected that improved operation could be obtained with a plano-concave arrangement. The lack of difficulty in obtaining the TEMoo mode and the fact that we were anxious to proceed with other work caused us to forego experimenting with curved mirrors. We plan to do this soon.

Figure 2 shows the oscillator-amplifier system as completed up to the present time. The two amplifiers are similar in construction to the oscillator lamp and rod configuration described above. One more similar amplifier will be added soon. The output of the system in its present form is about 8 joules in a 5.5 mm diameter beam. The divergence of the beam from the 5.5 mm aperture varies from 140 microradians to 230 microradians depending on the energy output.

Although the system is not complete and further modifications even of the oscillator and two amplifiers described will probably be made, some comments concerning some of the problems encountered in constructing the system are perhaps in order.

In the first attempt to obtain the TEMoo mode no particular emphasis was placed on the quality of the oscillator rod used. With the same configuration as that shown in Figure 1 and a rod with one wave full aperture stress in it, we were unable to obtain any single spatial mode, even with an aperture as small as 0.5 mm diameter. Replacement of this rod with a carefully selected rod exhibiting less than 1/5 wave full aperture optical distortion resulted in easy attainment of the TEM。o mode, even with an aperture as large as 3.5 mm.

Another point of interest is the effect of an inclusion in an amplifier rod. When the rod of the first amplifier was utilized in the previous multimode system, an inclusion present in the rod did not manifest itself in any apparent way in the output beam. However, when the TEM o mode beam was amplified by this rod the presence of the inclusion became very apparent in the near field burn pattern. Figure 3 shows this burn pattern. The relatively large hole probably stems from the heating of the inclusion by the laser pulse, and the fact that it shows up in single mode operation while it does not in multimode operation can be attributed to the spatial coherence across the entire single mode beam.

The use of apertures in the system has been found necessary in order to obtain a "clean" output. However, the positioning of these apertures is critical. The slightest misalignment results in the loss of the circular symmetry in the output beam.

Finally, it should be mentioned that the system described here is the result of a modification of an already existing system. The system would be different in several respects if it had been designed from scratch.

Several investigators have hypothesized various mechanisms to be responsible for surface damage of glass subjected to a high power laser pulse. Some of these are as follows:

(1) The electrostrictive interaction between the laser beam and the glass causes a radial constriction of the glass. The Poisson effect resulting from the squeeze causes a compression wave to be propagated along the laser beam. When this compression reaches the surface the unloading effect causes rupture of the surface.

(2)

(3)

Stimulated Brilloiun scattering initiates and amplifies an acoustic wave which propagates along the forward direction of the beam. This wave ruptures the surface upon incidence.

The surface plasma, which invariably accompanies surface damage and which forms in the
first few nanoseconds of the laser pulse, rapidly expands upon partial absorption of the
tail end of the pulse. This expansion creates a shock wave which damages the surface.
The plasma originally stems from desorption of impurities from the surface.

(4) The surface plasma of (3) bombards the surface with thermally energetic ions. This results in thermal erosion of the surface.

Self trapping could play a role in any of these mechanisms. It would be particularly effective in cases (1) and (2) where the magnitude of the internal acoustic disturbances would be much greater in a trapped portion of the beam.

4.2. Holographic Studies of Damage

We have been using high speed holography to investigate the mechanisms responsible for damage. Figure 4 is a diagram of the essentials of the experimental arrangement. A 40 ns damaging pulse from the glass laser is passed through a one inch cube of ED-2 laser glass. (Some of the samples in this paper were 3/4" on one side.) The damaging beam was passed through a lens prior to incidence on the sample.

This was either a 10 or 25 cm focal length lens. In the 10 cm case the surface to be studied was placed near the focal point, this inspite of the objections alluded to above. The reason for this was that only one amplifier was in the glass laser system at the time and such operation was necessary to easily achieve damaging power densities. When the 25 cm lens was used the focal point was several cm from the exit surface of the sample.

At a selected time after the damage pulse, a hologram is made of the sample through the use of a ruby laser which emits a TEM。o pulse of 20 ns duration. The time interval between the damage pulse and the hologram pulse can be varied from 0 to several microseconds. Up to 500 ns the interval can be controlled to within 10 ns.

Two types of holograms are made with this apparatus, single exposure and double exposure. The single exposure, in which the ruby laser is fired only once, results in a shadowgram superimposed on the hologram of the sample. The shadowgram shows regions in which the optical density changes rapidly in space. The double exposure technique, which is of course quite well known, requires that a hologram first be made of the sample without the damaging pulse. Next, the sample is subjected to the damaging pulse and another hologram is made after the selected time interval. The result is a hologram of the sample with fringes due to the difference in its state between the two holograms. A shadowgram is also present in the resulting hologram.

The damaging

Figure 5 is a photo of the virtual image of a hologram made by a single exposure. beam was passed through a 10 cm lens focused just beyond the exit surface. (The exit face is in the right hand side of the photo in this and all other photos in this paper.) The delay, that is the time between the damaging pulse and the ruby pulse, is 190 ns. The damage to this sample took the form of a small pit, which is characteristic of exit damage. Two acoustic waves can be seen clearly in the photo. Note that the origin of these waves is obviously the damage site in the surface and that no other disturbances are seen inside the sample.