Segmented YAG Laser Rods and Methods of Manufacture

1. Field of the Invention

The invention relates to solid state laser rods and more particularly concerns a segmented laser rod and the method of its manufacture.

2. Brief Description of the Prior Art

Satisfying a growing demand for solid state laser (Light Amplication by Stimulated Emission of Radiation) rods currently requires relatively expensive investment in straight forward capacity expansion. This is due, in part, to the present state-of-the-art methods of manufacture. For example, the commercially-practiced method of manufacturing YAG boules from which the yag laser rods are cored is the so-called "Czochralski" method. In this method a "seed" crystal is dipped into a high temperature molten pool of the doped garnet material and very slowly rotated and withdrawn. A crystal boule of about 1 to 3 inches in diameter is formed, the length of which depends on the mass of starting melt and available process time. It takes about 2-3 weeks of steady, uninterrupted processing to produce a boule of 6-8 inches in length from which usable laser rods can be "cored" along the axial (long) direction of the boule. Any interruption or significant change in process parameters during that time is likely to introduce a flaw in the boule at the position of the boule/melt interface. This is therefore a factor which determines the maximum length and total number of rods (i.e. the "yield") which can be cored from a boule. The mathematically-inclined will readily see that if process interruptions are random events in time, the process yield decreases exponentially with desired boule (rod) length. Operating at constant yield, increases in production can be achieved only by increasing the number of growth stations.

Also, the Czochralski method generally produces a boule whose quality and concentration of active dopant changes along the boule length This is due to the difference in solubility of the active dopant species in the melt as contrasted with the solid crystal boule. This is called segregation. The degree of non-uniformity along the boule depends upon the relative mass of the final boule to the mass of the starting melt.

Because of these features of monolithic boule growth, with YAG, for example, the yield of rods of even a few inches length is generally less than 50% of the geometrically-available boule material, and the overall yield of highest quality rods suitable, for example, for use in cw-pumped, TEM .sub.oo mode laser operation is even lower. The currently-used process for producing YAG rods penalizes length by ever decreasing yield. One-inch long rods are easy, six-inch lengths very difficult and costly and ten-inch rods virtually impossible to obtain. Similar problems occur with other methods of boule growth such as the Vernuil method.

A further problem occurs with long rods, namely, a dynamic effect called "thermal focusing". The longer the rod, the more the rod focuses, and the more difficult it is to maintain the laser resonator's stability and high quality beam.

Dynamically, an optically pumped (excited) solid state laser rod, by virtue of parasitic heat generated within the rod volume, and the removal of this heat by cooling through the lateral surface, shows a radial temperature gradient. This temperature gradient through the stresses created and the photo-optical properties of the material, causes the rod to exhibit the properties of a lens. In the case of YAG, the lens is positive and causes a focusing effect. The focal length F of this dynamically-produced effect is closely represented by the expression ##EQU1## where L is the length of the rod and .alpha. is a parameter that depends on material properties and the level of excitation (pumping).

Since this dynamic lens is included within the laser resonator, indeed, a part of it, there exist values of F for which the resonator becomes unstable and laser action ceases or decreases markedly. This effect is often observed as a decrease in beam output with an increase in excitation (pumping).

The method of the present invention permits one to solve the above-described problems of the prior art. All of the above-described problems are greatly reduced or entirely eliminated by abandonment of the monolithic rod in favor of a composite rod, built up from sections or segments. When only short (.about.1") active segments are required to build a long, composite rod, the boule yield problem virtually vanishes, dopant concentration variation can be eliminated by matching of sections, compensation of dynamic focusing and birefringence can be incorporated and rods of arbitrary length can be fabricated and usefully operated.

The present invention offers a method for assembly of segments, i.e. short rod segments or slabs from boules so as to obtain a laser rod of usable length. It is worthwhile to note by those skilled in the art that several common optical cements and resins were tried but they failed because of damage as a result of the high internal light intensities that are produced in a laser during operation.

The present invention describes an optical cement for joining the segments or slabs which cement does not fail under dynamic condition and has the necessary mechanical strength, ability to survive the intense light flux from both the laser beam and the pumping-lamp and a coefficient of thermal expansion and an index of refraction, very near that of the host solid-state material.

In accordance with a specific embodiment, a glass binder is selected so that its properties match the refractive index and thermal expansion properties of the host solid-state material. The latter is required so that upon the application of heat to melt the glass which is interposed between two rod segments and thereafter cooled down, the glass will join the rods together at the interface without fracturing or introducing stresses. The former is required so that the interface, in effect, disappears optically and light is not significantly reflected or scattered because the change in index of refraction, R, at the interface is nearly zero; that is, the optical index of refraction, of the host material and the glass are nearly matched or are the same. The power reflection coefficient, R, of light incident on a boundary between two regions of different index of refraction is given by ##EQU2## where n.sub.g is the index of the bonding cement and n.sub.h that of the solid-state laser host. A reflection coefficient of up to 0.5% is generally acceptable for good laser performance which gives, in the case of YAG (n.sub.h =1.82), (n.sub.g -n.sub.h)=.DELTA.n.apprxeq..+-.0.26, a relatively wide range. In some specific lasers either larger or smaller reflections are acceptable so other index deviations may be acceptable. Accordingly, the optical properties of the segmented rod produced by the present invention are essentially the same as a monolithic rod.

Thus, long sections of YAG rods are produced more easily by this invention. It is now possible to produce yag rods of indefinite length. This latter was practically impossible heretofore by conventional crystal growth technology.

FIG. 1 shows a diagrammatic cross-sectional, side elevation of a portion of a segmented YAG laser rod of the invention.

Yttrium-aluminum-garnet (YAG) boule manufacturing techniques are well known, including those which provide doped boules, i.e., doped with, for example, neodymium, erbium or other rare-earth elements. In the method of the present invention, relatively short YAG and doped YAG rods or slabs cut from boules are assembled to form a segmented laser rod or "boule" of relatively greater length.

Referring now to the accompanying drawings there is seen in FIG. 1 an enlarged, cross-sectional side elevation of an embodiment of a segmented laser rod 10 of the invention. The rod 10 is made up of a plurality of segments or slabs 12 of YAG or doped YAG fashioned so that when assembled together and joined by an optical glass cement 14, a segmented laser rod or boule is obtained. The glass cement 14 is an optical glass composition which satisfies the following criteria:

(a) refractive index match.+-.0.5 (i.e.=index of laser host material.+-.0.5) (Note: matching does not mean precisely equal).

(b) expansion coefficient match to YAG.+-.apprx. 5%, the exact value depending on process temperatures.

(c) non-solarizing (i.e. no optical degradation in high ambient pump light.

(d) optically linear and non-absorbing at laser wavelength (i.e. does not behave non-linearly or degrade under laser dynamic conditions).

(e) strength comparable to host.

(f) compatible with host material, i.e. no undesirable chemical reaction.

(g) water insoluability where water cooling of rod is used.

An example of a glass which satisfies these criteria is one having the following approximate stoichiometry:

This glass cement has an index of refraction of n=1.82, a thermal expansion coefficient of 71.times.10.sup.-7 /.degree.C., a bond strength of 22,000 PSI; and a laser damage threshold of 72 MW/cm.sup.2. It is also non-solarizing and water insoluble.

Other glass compositions of different refractive index and/or expansion coefficient are readily formulated knowing the desired index of refraction and thermal expansion coefficient by superimposing the graphs appearing in the article "Properties of Low-Temperature Solder Glass" by John Gallup and A. G. F. Dingwall appearing in Ceramic Bulletin, Volume 36, No. 2 (1957). NOTE: The example stoichiometry given above differs from those in Gallup and Dingwall in that the example includes a few percent of Al.sub.2 O.sub.3 while Gallup and Dingwall omit this constituent. The Al.sub.2 O.sub.3 in the given example is included to stabilize and inhibit phase separation in the glass.

The following examples describe the manner and process of making and using the invention and set forth the best mode contemplated by the inventors but are not to be construed as limiting the invention.

The oxide components of the glass cement are weighed and mixed in proportions described above to provide a 1 kg batch. The mixed batch is first melted in an 0.5 liter fused silica crucible at 1000.degree. C. and held for 45 minutes before being transferred to a precleaned 0.5 liter platinum crucible. After transfer, the glass is refined at 1400.degree. C. for two hours with constant stirring at 45 rpm. After completion of refining, the temperature is decreased to 1100.degree. C. and the stirring speed reduced to 30 rpm. These conditions are maintained for one half hour for clarification and as a precaution against striae. The glass is cast at 1100.degree. C. into a 50.times.150.times.30 mm.sup.3 graphite-coated stainless steel mold and cooled at 20.degree. C./hr. from 1100.degree. C. to room temperature. The glass thus formed is of sufficient optical quality to be used as a YAG rod cement.

In this example, two 40 mm long, 4 mm diameter Nd doped YAG rod segments are to be joined to produce one 80 mm long laser rod.

The faces to be joined are lapped flat (but not necessarily optically flat) using 5 micron Al.sub.2 O.sub.3, then washed with detergent and blown dry with dry nitrogen. Perpendicularity of the faces to be joined with respect to the segment axes is maintained to about 5 minutes of arc. A coin-shaped piece of the glass cement previously described is cut and ground and lapped down to.apprxeq.0.005 inch thickness. Its traverse dimensions (diameter) slightly exceed those of the YAG segments above.

One of the rod segments and the glass preform are mounted in a quartz fixture such that they are stacked with the rod segment axis vertical, preform resting on the upper face. The fixture carrying the rod segment and preform are placed into a fused silica tube and the tube with the fixture is lowered into a muffle furnace which has been preheated to 1100.degree. C. After approximately 8 minutes the silica tube and quartz fixture are withdrawn and allowed to cool sufficiently to allow the final assembly step. At this point, the glass preform will be seen to have melted and wetted the rod face. Furthermore, because of surface tension forces, the surface of the glass will be seen to be spherical, much like the shape of a water droplet on a flat surface. (In the case of large diameter rod segments, the glass surface may show a dimple near its center and it may be necessary to invert the fixture, reheat to melting and cool again).

The final step consists in placing the second rod segment into the fixture, aligned, with its end surface resting on the spherical glass surface. A weight of 100 grams is placed atop the second rod segment which is also free to slide axially a small amount. The fixture is replaced into the silica tube and is reheated as above. The force provided by the weight of the second segment plus the 100 gram weight forces the segment faces into intimate contact and squeezes the excess glass out of the bond zone.

After cooling, the now-bonded rod is given a light centerless grind to remove the excess glass which was squeezed from the joint.

The length of the composite rod compared to the combined lengths of the original segments will show less than 0.001 inches for the bond section thickness. The bond section should not be visible to the naked eye when looking through the rod, indicating good index of refraction match between the glass cement and YAG. Images should be transmitted undistorted through the composite rod indicating a high level of homogenity of the index of refraction of the glass cement. Polariscope inspection of the rod should show no extra birefrigence due to the bonding process, indicating a very good match in thermal expansion coefficients between the glass cement and YAG crystal.

After satisfying all of the above passive tests the rod ends are optically polished and antireflection coated as with a monolithic rod so that active laser tests can be performed. The rod is installed inside a standard water-cooled laser pump cavity. A DC Krypton arc lamp is used to provide pump radiation at approximately a 4.4 kW input level, a standard condition for this rod size. Two flat mirrors, one 99.96% reflective, the other 88% reflective, are used to establish laser oscillation when placed about 50 cm apart with the rod located centrally between them. The CW multimode output power of this laser oscillator will measure about 50 W.

Thus the several aforenoted objects and advantages are most effectively attained. Although several somewhat preferred embodiments have been disclosed and described in detail herein, it should be understood that this invention is in no sense limited thereby and its scope is to be determined by that of the appended claims.