Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
CHAPTER I A REVIEW 0F TECHNICAL CHARACTERISTICS 0F LASERS J. A. Carruthers* and Martin S. Litwin** INTR0DUCTI0N Since Maiman first succeeded in obtaining laser action in a synthetic ruby rod in I960l, the discovery of new laser transitions has proceeded at a very rapid pace, so that today one would scarcely try to count the number of wavelengths at which laser action has been observed. A few of the more promising lasers have received intensive development effort, which has led to some rather remarkable achievements in terms of power, coherence, lifetime, and other characteristics. Concurrent with these improvements there has been a steady increase in the numbers and types of lasers in research, education, industry, and military applications. There is every indication that the use of lasers will continue to expand and that human contact with lasers in research, teaching, and other activities will increase enormously in the next decade. To aid in the detailed discussions which follow in the later chapters on laser eye effects, a review is given of the characteristics of lasers and of the radiation emitted by them. Certain properties and principles of operation are common to all lasers, but others are best treated by con- sidering the three main laser classifications, solid state, gaseous, and semiconductor. A brief statement of the present capabilities of lasers is in order. Solid-state giant pulse (Q-switched) lasers, such as ruby and neodymium, produce pulses that are on the order of 5 to 50 nanoseconds (I nanosecond equals 10~9 seconds) in length with peak powers of the order of 10 to 1000 megawatts and pulse energies, typically, from 0.1 to 10 joules. Alter- natively, the same lasers operated in the normal or "long-pulse" mode generate pulses on the order of a millisecond in length, peak powers up to hundreds of kilowatts, and pulse energies from a few joules to a few kilojoules. Such "long-pulse" operation is typically about an order of magnitude more efficient than Q-switched operation, although neither is particularly efficient in converting electrical input energy into laser energy. Most work requiring high power has been done in the red (chromium- in-sapphire or ruby, bS^tmy) or near infrared (neodymium-in-glass or crys- tals, 1060 rnn). The most common of the continuous (CW) lasers is the helium-neon gas system which has its principle lines at 633 mu, 1150 mu and 3390 mjj. The output is usually a few milliwatts but can be as high as a few tenths of a watt at 633 <n>J. The argon-ion laser, which is presently capable of several watts CW in the green (principally 488 and 51^.5 mu), ^Department of Electrical Engineering, University of Minnesota, Minneapolis, Minnesota **Department of Surgery, Tulane University School of Medicine, New 0rleans, Louisiana
is now becoming increasingly common as technical problems are overcome. Still higher continuous powers are available at lO60 mu using nd3+ in yttrium-aluminum-garnet (YAG), and an output of over 200 watts has been reported. Perhaps the simplest of the CW lasers is also capable of the highest average power; the C02 molecular laser oscillates at 10.b microns and can produce about 50 watts in a very elementary design, and kilowatts or more in larger or more complex systems. Furthermore, because of the relatively long lifetimes in the lasing levels of the C02 laser, it can be Q-switched to produce peak powers many orders of magnitude greater than the CW power from the same laser. A summary of the present capabilities of the more common lasers ob- viously cannot include reference to most of the specialized systems in use in research or various special projects. However, it is important to note that higher peak powers can be obtained from some of the CW lasers by pulsing the pump power or by "..-switching the laser cavity, or both. 0ther lasers can only operate in the pulsed mode since the lasing inver- sions are created by transient conditions in the pump cycle, such as the oxygen and nitrogen lasers in the violet and ultraviolet. High-efficiency frequency doubling or even quadrupling is a practical method for obtaining radiation of shorter wave lengths. Further, many lasers can be operated in the multimode phase-locked condition, which causes the output to con- sist of sharp, regular spikes whose peak power may be orders of magnitude larger than the average power. //hen discussing future trends during the next decade, it appears important to consider those systems which are most widely used as well as the state-of-art lasers which will probably be more restricted in use and confined to research laboratories and special projects. The more common laser oscillators, e.g., ruby (69^ my), nd-glass (1060 mjj), he-ne (633 mjj) and argon (k8'6 and 51^.5 m^j) , have already undergone intensive development, and the power capabilities, reproducibi 1ity, reliability, and efficiency have been improved enormously from earlier models. It would appear that further advance will encounter the law of diminishing returns and that an additional factor of ten in the peak power capabilities will represent an upper limit. High-power amplifiers, often using the same material as the laser oscillator, are used to obtain higher pulse energies or powers. The present limits are determined by the ability of the optical and laser materials themselves to withstand permanent damage. The molecular lasers, such as the C02~N2 system, are relatively new and major advances can be expected, although the wavelengths will probably be restricted to the several micron region and longer. Because the power and efficiencies are high and the systems are comparatively inexpensive, it is expected that extensive use will be made of this type of laser. The vast majority of transitions which are suitable for lasers are in the infrared region of the spectrum, and this condition is not likely to change. However, there is need for practical laser sources throughout the whole of the visible spectrum, and it is expected that considerable effort will be directed toward meeting this need. The widescale use of holography and photographic plates insures an increasing market for visible
lasers. Also, it is more convenient to work with lasers whose radiation is visible. The he-ne 633 mu laser will probably be a popular model for many years because of its relatively low cost, dependable operation, and long life. The output power of the majority of units will be between I and 100 milliwatts, although less than a milliwatt is adequate for many applications. 0ther CW lasers in the visible spectrum can be expected to become available, probably producing about the same power or somewhat more. The argon-ion laser will partially fill this requirement, and commercial units will probably reach the ten watt level. Frequency doubling, and even quadrupling, from 1060 mjj (nd'+ in yttrium-aluminum-garnet) or other infrared lasers may be a practical means for providing visible coherent sources at other wavelengths if direct lasing action throughout the visible remains difficult to achieve. Also, tunable pulsed and CW parametric oscillators and Raman lasers will almost certainly be developed. »3 Tun- able CW parametric fluroescence has also been demonstrated, and this may become an important research tool, although the power will probably be less than a microwatt. The ruby (69^ mp) and nd-glass (1060 nyu) lasers are expected to continue to dominate the market for high-power pulsed lasers. Frequency doubling is already a practical method of obtaining shorter wavelengths from these lasers, and since the doubling efficiency is high (up to 25%), there is not a compelling need to develop new systems. The principal laser lines are summarized in Tables 1, 2, and 3. The recent book by Smith and Sorokin14 is also an excellent reference for details on current lasers and lasing transitions. THE LASER PRINCIPLE Conventional optical sources can be grouped into one of three broad classifications according to the nature of the radiation, incandescent, and emission, or line emission. Lasers have something in common with the last type of source since the active material must normally have an extremely narrow line. This line represents a transition between two energy levels in the atom, ion, or molecule. An atom which is in an excited, or upper level, tends to revert to a lower level, and the excess energy may be radiated spontaneously. This process is referred to as spontaneous emission, or fluorescence. It should be emphasized that for fluorescence to occur, it is necessary only that some of the atoms be in the upper state, and it is not required that more be in the upper than the lower level. The frequency, I/, of the emitted radiation is related to the energy difference, E, between the two levels by the Einstein relation E = h i/where h is Planck's constant.
Table 1 Principal Solid State Laser Elements Currently in use Active Element (and Valence) Europium (3+) Chromium (3+) Samarium (2+) Ytterbium (3+) Praseodymium (3+) Neodymium (3+) Host Lattice Ma t e r i a 1 Yttrium oxide plastic chelate in alcohol Alumi num oxi de Fluorides of calcium and strontium Glass Calcium tungstate Vari ous fluori des, molybdates, glass Principal Output Wavelength .bl p .70 p .71 p 1.02 p 1.05 p I.0b u Thu 1 i urn (2+) Ca lei um f luor i de 1.12 u Erbi urn (3+) Calcium tungstate 1.61 u Thul i urn (3+) Calcium tungstate, strontium fluoride K91 p Holmi um (3+) Calcium f luor i de, 2.05 p calcium tungstate, glass Dyspros ium (2+) Ca lei um f luor i de 2.36 p Urani um (3+) Various fluorides 2.k - 2.6 p Note. â The laser element may be doped into several different lattice materials in certain cases.
Table 2 Principal Gases Known to Exhibit Laser Action Gas Emission Di stri but ion Emission Frequency Range Argon 86 (strong 1 i nes 1 i nes from 35H-5145A) 2753A - 26.95 u Bromi ne k 1 ines regi on of 8446 A Carbon 9 1 i nes kbkj A - 5.5956 u Ces i um 2 1 ines 3.204 u and 8.1821 p Chlorine 11 1 i nes 4781 A - 2.2060 u He 1 i um 2 1 ines 1 .9543 u and 2.0603 /J lodi ne 8 1 i nes 5407.4 A - 3.431 H Krypton 59 1 i nes 3050A - 7.0565 y Mercury 25 1 i nes 4797 A - 1.813 v Neon 155 lines (line at 3.39 u has gain 40dB/meter) 2678. 6 A - 132.8 jt Ni trogen 9 1 i nes 3478.7 A - 1.4547 H (Table continued on next page)
Table 2 - concluded Gas 0xygen Sulfur Xenon Carbon monoxide Carbon dioxide Deuterium 0xide Ni trogen Ammon i a (NH2, NH3, Water vapor Ni trous oxi de Hydrogen Cvanide Emission Distribution 13 lines 2 1i nes 64 1i nes (line at 3.507 u highest gai n 60 dB/meter) three visible bands two bands Emission Frequency Range 2984.6 A - 8446.37 A 1.0455 u and 1.0636 u 2983.8 A - 18.5 u 5590.6 A - 6613.5 p 9 u and 10 u 16 unidentified transitions from 33.9 u - 107.7 4 bands 7 wavelengths 32 unidentified transitions one band 8683.5 A - 1.2347 2.47 u - 31.94 M region of 10.8 region of 337 microns Note. -- With the exception of cesium vapor, all are stimulated by passage of an electrical current.
Table 3 Principal Diode Materials Currently in use in Semiconductor Lasers Laser D i ode Mater ia 1 Gal 1i urn arsenide Indi urn phosphide Indi um arseni de Gallium indium-arsenide Indi um antimoni de Gallium arsenide-phosphide Em i s s i on F requency Range 0.840 p - 0.932 p 0.900 ^i - 0.919 M 3.112 u 0.840 p - 3.100 y 5.200 y 0.610 - 0.840 Note. â Efficiency of operation is high approaching 40%; however, total power density that can be achieved is low. 10
If a material shows a strong emission line in fluorescence, it is also found to be a strong absorber of radiation of the same wavelength. When absorption occurs the atom is raised from the lower to the upper state, or it may be said that the atom is stimulated to a higher level. A corresponding process in emission also occurs; atoms in the upper state can be stimulated to emit radiation and, in the process, drop to the lower level. It should be emphasized that the stimulation process is basically different from the spontaneous process, since the former occurs in either direction (absorption or emission) while the latter occurs only in emission. Absorption and stimulated emission are funda- mentally the same process, and both occur simultaneously whenever radi- ation of the resonant wavelength passes through the material. Which process is dominant depends on the relative number of atoms in the two levels: if more are in the higher level, there is net emission; while if more are in the lower state, there is net absorption. Normally the lower state has the larger population and absorption occurs. When the upper state can be made to have the larger population, the material has gain and is a possible source for laser action. This condition is comparatively difficult to achieve, and the material is said to have an inverted population. To further illustrate the laser principle, consider a flame to which a small amount of salt is added in order to produce the yellow lines of sodium in fluorescence. If a narrow beam of radiation at the same wavelength as one of the sodium lines is passed through the flame, the radiation is slightly absorbed. In order to have laser action, it would be necessary for the yellow flame to cause an increase in the intensity of the narrow beam as it passed through. To produce laser action, it is required that the population of the two levels be inverted so that the material amplifies its own spontaneous emission. To achieve an inverted population, the upper level must have a relatively long lifetime, and it is necessary to pump or excite the atom into this state. The commonest means of pumping a solid state laser is by means of intense light from a flashlamp. For gas lasers, a discharge is set up by means of an rf or dc excitation. In semicon- ductor lasers, a large dc current is passed through the junction. A laser material amplifies by means of stimulated emission, but does not constitute an oscillator unless a resonator is provided. The simplest form of resonator consists of two flat mirrors which are made parallel and positioned one on either side of the active material so as to intercept as much of the stimulated emission as possible. The same conditions hold for an optical frequency oscillator as for one at lower frequency; if the gain of the laser material is sufficient to over- come losses, the system will oscillate at one of the natural frequencies of the system. The amplitude of the oscillation increases until satura- tion effects cause the gain to become equal to loss. Multiple-layer dielectric coatings can be produced with losses as low as a few tenths of a percent, so that even low gain transitions can be made to oscillate. Although flat mirrors are still used in most solid state lasers, concave spherical reflectors are more common in gas lasers since the tilt of the mirrors does not have to be adjusted as accurately. 11
The frequency separation between the primary resonator modes is approximately c/2I_ where L is the length of the cavity and c is the velocity of light. For a one-meter resonator, the separation is 1.5 X 10° cycles per second, or 0.005 cm~ . The linewidth of most laser materials is considerably greater than this, so that lasers tend to have several modes oscillating simultaneously. The presence of several independent modes is not of major significance in most applications except that the amplitude noise is high. However, it is possible for the modes to become coupled together so that they are phase-locked to one another and produce intense spikes in the output.5»fa»7i° Even when many modes are present, the radiation is highly monochromatic by most standards. For he-ne (633 my) the linewidth is about 1.5 X 10^ cycles per second, which can be expressed as 0.05 cm~' or as 0.02 Angstroms. CHARACTERISTICS 0F LASER RADIATI0N The most important characteristic of laser radiation is its ex- cellent coherence, both in time and space. That is, the emission is highly monochromatic and can be focused to an extremely small beam area. The linewidth of a single mode he-ne laser is determined by mirror vibration and is of the order of 105 cycles/second, or 10~*> Angstroms, under good laboratory conditions. However, most lasers have many modes present, and the total frequency spread is primarily determined by the width of the fluorescent line of the active material, as discussed briefly in the previous section. Except for nd-glass, the linewidths of most of the common lasers are still very narrow even when oscillating multimode. For example, the use of a resonant reflector in a ruby laser will limit the frequency spread to about 109 cycles per second, or 0.02 Angstroms. The excellent spatial coherence of most lasers results from the use of a resonator with highly precise mirror surfaces. The wavefront of the beam radiated from the output mirror matches the shape of the mirror surface and is, therefore, a plane-wave if the mirror is flat, and a spherical wave if the mirror is curved. In some instances, the active material introduces significant wavefront distortion because of optical inhomogeneities, and the presence of off-axial modes is a further com- plication. However, it has been reported that in the best of ruby lasers under ideal conditions the beam divergence can approach the diffraction 1imi t. The practical effect of good spatial coherence is that when a positive lens is placed in the beam, the size of the focal spot, or image, is extremely small, and for some lasers the diameter of the spot focused by a perfect lens is limited only by diffraction. As long as lens aberrations do not become significant, the larger the diameter of the parallel beam as it enters the lens, the smaller the size of the focal spot, so that even from medium power CW laser, it is possible to obtain extremely high power density over this very small area. The giant 12
pulse lasers have intensities of the order of 10.9 watts per cm^ even when unfocused, so that it is not surprising that a focused beam can shatter crystals or generate a spark. It is interesting to compare the intensity of the sun's image at the focal point of a lens with that of a medium power he-ne visible laser. Assume an f/k lens, with a focal length of the 10 cm and a diameter of 2.5 cm. The sun subtends an angle of about 0.01 radians, so that the diameter of the image is one millimeter. The intensity of the sun's radiation is about 100 milliwatts per car, so that at the focal point the power density is approximately 50 watts per cm^. If the laser beam is 3 millimeters in diameter and the laser is generating only longitudinal modes with no transverse modes, the angular width of the central lobe of the diffraction pattern is approximately given by X /d radians, when X is the wavelength, and d is the diameter of the beam. The diameter of the focal spot is approximately 10 X /d, since the focal length is 10 cm. This gives a diameter of about 20 microns. If the output power of the laser is 50 milliwatts, the intensity at the focus is approximately 1.6 x 10^ watts per cm . S0LID STATE LASERS The laser rod or crystal is situated in a reflecting chamber along with a pump source (usually a flash lamp). Electrical power stored in a capacitor bank is suddenly discharged through the flash lamp. Some of the light is absorbed in the laser rod and excites the atoms or ions (cr3+ ions in the case of ruby) from the ground state to an excited level. In ruby, there is then a very rapid transition to the upper laser level, and laser action occurs to the ground state. Ruby is an example of a three-level system. The laser resonator may consist of the rod itself, in which case one end is coated so as to reflect most of the energy (99%) and the other end is partially transmitting. In other cases, the rod is placed between separate mi rrors. 0ther types of solid state lasers, such as those containing neodymium, function at four energy levels. In these crystals, the terminal laser level is above that of the ground state far enough so that it is normally nearly empty at the operating temperature of the laser, thus reducing pumping power requirements. In addition, the pump bands absorbed by neo- dymium are much wider than for most similar crystals and pumping can also be accomplished to several upper energy states rather than one or two narrow bands as is the case with ruby. While there are three principal lines for emission from this material, over 98% of the emitted energy lies at the wavelength of 1060 mjj. As energy outputs increase, it may be ex- pected that the total amount of energy available in the lesser lines will also increase proportionately. 13
Solid state lasers most used to date include the following: (a) ruby, which is chromium (3+) substitutionally replacing aluminum in aluminum oxide, (b) neodymium (3+) in crown glass or in various fluorides, calcium tungstate, or yttrium aluminum garnet (YAG), (c) europium (3+) in yttrium oxide, and (d) uranium (3+) in various fluorides. These and others of importance are listed in Table 1. Wavelengths achieved from solid state lasers have varied between 610 mjj for europium and 2600 nju for uranium. Energy outputs using neo- dymium in glass have gone up into the thousands of joules per pulse with repetition rates as high as 5 to 10 per minute. Most biologic work has been done with crystals doped with chromium and neodymium. GAS LASERS The first gas laser was built by Javan, Bennett and Herriott in 196l9. It gave a continuous output of one milliwatt at several wavelengths near one micron with the principal emission at 1150 mp. The gases used were helium and neon. While much work with lasers has been done using these gases, at present more than 500 lines are known to exhibit laser action under either continuous or pulsed excitation. Wavelengths from gas lasers range from the ultraviolet (neon-268 nyj) into the infrared (HCN-337 microns), and they may be pulsed, continuous, or quasi-continuous. The helium-neon combination, just as most of the others, can give coherent laser output at many wavelength including several in the visibile portion of the spectrum. Using certain reflector techniques, various spectral regions can be selected. The strongest visible laser line from the helium-neon mixture is at 633 ">«⢠Recently, considerable work has been done using carbon dioxide and nitrogen, and also ionized rare gas lasers. Relatively high powers have been achieved from these devices with outputs reported in the 20 watt range for ionized argon and over 1,000 watts for carbon dioxide. Table 2 lists those gases which have demonstrated laser action together with the wavelengths achieved. Although most emit a large number of lines, tuning techniques may be used to select the desired wavelength. Gas lasers are usually operated as continuous (CW) sources, and have a beam divergence close to the theoretical minimum. Although most gas lasers have several modes present they can be operated so as to produce only a single mode. Most gas lasers are four-level systems. Population inversion is maintained between two upper excited levels, which can be labelled E3 and E2. The upper level, E3, undergoes stimulated emission to the lower state, E2. Depopulation of E2 occurs by spontaneous emission to an inter- mediate metastable state, Ej, rather than to the ground state, EQ, from whence repumping to E3 can be accomplished. The fact that Ej is metas- table hinders the decay of this state and its return to EQ. Thus, popu- lation inversion density versus pumping rate can saturate or even decrease due not only to resonance trapping at the lower metastable level but also
to electron re-excitation from that level. This so-called metastable "bottleneck" has prevented most gas lasers from achieving higher power outputs or more efficient operating conditions. The ionized gas laser seems to be the first step in overcoming this metastable drawback. In the ion laser, the contained gas is ionized by passage of an electrical current and the active laser material consists of excited ions. Excitation then occurs principally by electron impact with ground state neutral atoms. Although efficiency is low, less than 1%, no distinct saturation limit on pumping rate relative to power output has been observed. Another class of lasers has been proposed called "collision" lasers. In collision lasers the metastable problem can be overcome by using various atomic collision processes against one another so as to cause depopulation of the metastable level to the ground state. Using these devices, 50% efficiency is not an unreasonable figure to consider attainable. More recently energy levels of the vibration-rotation spectrum of carbon dioxide have been used to achieve continuous outputs at high powers (1,000 watts has been reported). In the operation of this device, a dc discharge directly into a mixture of carbon dioxide and nitrogen-helium results in the excitation of large numbers of nitrogen molecules to a vibrational energy level almost identical with an energy level for carbon dioxide. A selective transfer of energy to carbon dioxide molecules occurs with population inversion being thereby accomplished. Since the vibrational level of nitrogen has a relatively long life- time, the pumping mechanism is extremely efficient, about 1^%. While other gases have been used in place of carbon dioxide, i.e., nitrous oxide, carbon monoxide, and carbon disulfide, highest power outputs have been achieved using the former. SEMIC0NDUCT0R LASERS In November 1962, successful laser action was first achieved in the junction region of a gallium arsenide semiconductor diode^. '"» H. Since that time, laser action has been achieved in gallium arsenide- phosphide, indium arsenide, indium phosphide, gallium indium-arsenide, and indium antimonide at wavelengths ranging from 0.65 microns to 5.2 microns. The wavelength range has been extended to 22 microns using lead sulfide, lead telluride, and lead selenide. Principal diodes cur- rently in use are listed in Table 3. If an electrical forward bias is applied to a semiconductor diode, electrons enter the junction region and occupy energetic states near or within the conduction band. Simultaneously, electron vacancies in less energetic states near or within the valence band appear in the diode junction. Radiation occurs as electrons pass this junction and make the transition back to the valence band from the conduction band. When the 15
junction current is large enough, there will be more electrons near the edge of the conduction band than at the edge of the valence band and a population inversion may occur. Under these conditions, and utilizing crystals such as gallium arsenide that have been properly made, laser action will occur thoughout the junction region. Useful laser emission is obtained at the semiconductor-air interface at the edge of the junction. Internal ohmic loss in the diode is the principal problem associated with obtaining high peak power pulses from most semiconductor or injection lasers for current pulses of less than 1 microsecond duration. Raising the temperature of a laser diode tends to increase the current density necessary for threshold. At 20°K GaAs diodes have been reported to yield 6 watts of continuous power and at 80°K power in excess of 1 watt has been obtained under CW conditions. Room temperature pulsed operation of GaAs has produced 60 watts of peak power in 100 microsecond pulses at 30 cycles per second. At lower temperatures, several hundred watts of peak pulse power are available from GaAs. Total energy delivered per pulse, however, has been small. At room temperature, laser emission from pure GaAs occurs at a wavelength of about 900 mu. At lower temperatures the output wavelength decreases to about 840 mp. The output can therefore be tuned over a considerable range simply by varying the temperature. Still further change in wavelength can be achieved by using a three-element alloy mixture such as gallium arsenide phosphide. Depending on the ratio of arsenic to phosphorus, such a diode will operate between 610 mp and QkO mu. A three-element mixture of gallium indium arsenide shows promise of operating from 0.8k to 3.1 microns. By adjusting the temperature of operation, the output can be brought close to any wavelength desired. While it is difficult to achieve an accurately reproducible wavelength, diode lasers have the advantages of compactness, simplicity, increased efficiency of operation, a power output that may be adjusted and quickly controlled by simply changing the electrical input into the diode. SPECIAL SYSTEMS Gaint pulse laser Using special techniques it is possible to achieve enormously high peak powers in very short pulses. This method is called gaint pulsing, or Q-switching, and is accomplished by interposing a fast action shutter between one end of the laser rod and the mirror. During most of the pumping cycle, the shutter is closed so that the Q of the resonant cavity is low, I.e., the losses are high, since light from the end of the rod cannot reach the mirror and be reflected back. Therefore, since the gain does not exceed the losses, laser oscillations cannot start even though a large number of atoms are in the excited state and the gain is high. When the laser element has been pumped sufficiently, the shutter is opened quickly. This restores the high Q of the resonant cavity and the oscillations build up very rapidly. Under these conditions, the stored energy is delivered in one giant pulse, and peak powers as high as a few 16
gigawatts have been achieved. The pulse width is extremely short, usually lasting for only 5 to 30 nanoseconds (I nanosecond = 10~9 second). Devices currently in use for 0_-switching include electro- optical shutters, rapidly rotating mirrors, rotating prisms, bleachable absorbers, and various combinations of these. It should be emphasized that even though very high peak powers are obtained, the pulses are short and the total amount of energy delivered is small, rarely exceeding a few joules. Raman effect When a laser beam is passed into certain liquids such as benzene or nitrobenzene, additional coherent, we 1 1 -col 1 i mated wavelengths are propagated out of the liquid. These new wavelengths result from frequency shifts of the original wavelength by multiples of the various molecular vibrational frequencies of the liquid through which the laser beam is passed. Such a shift is called the "Raman effect". While most of these frequency shifts are downwards (Stokes lines), under certain conditions upward shifts may occur (ant i -Stokes lines). If a chamber containing the liquid being irradiated is placed between two parallel end mirrors so as to form a light resonating cavity, laser light oscillation and amplification will occur at either one or more of the Stokes frequencies, depending on whether or not oscillation threshold for the particular line is reached. Discovery of the stimulated Raman effect using laser beams has made available many new wavelengths of high intensity coherent light. Harmonic generation Because of the intense power density which can be obtained in focused laser beams, and in particular with giant pulse lasers, it is possible to generate appreciable harmonic power in many materials. All substances, presumably, are nonlinear when the intensity of the radiation is suffi- ciently high, but a few crystals are particularly effective in generating harmonics. The most widely used materials for second harmonic generation are potassium dihydrogen phosphate (KDP), ammonium dihydrogen phosphate (ADP) and lithium niobate ( A small amount of harmonic field is generated as the laser beam traverses the crystal. In order for the efficiency of conversion to be significant, it is necessary for the harmonic and fundamental waves to travel with the same velocity so that the harmonic generation is cumula- tive. Because of dispersion in the index of refraction, the wave velocity changes with wavelength, and velocity matching, or phase matching, is normally not obtained. However, phase matching is possible in birefringent crystals. For example, in LiNbO^ the laser can be transmitted as an ordi- nary wave (o-wave) and the second harmonic as an extraordinary wave (e-wave) 17
By rotation of the crystal axes or by changing the temperature of the crystal, the two velocities can be made equal. Lithium niobate is particularly useful for generating second harmonic from a laser which is operating at about 1000 my, while KDP and ADP are used for wavelengths of the fundamental from about 1000 my to 500 mp. REFERENCES 1. T. H. Maitnan, Nature JjJ7, ^93 (I960). 2. J. A. Giordmaine and R. C. Miller, Phys. Rev. Letters J4, 193 (1965) 3. G. 0. Boyd and A. Ashkin, Phys. Rev. Jjt6, 187 (1966). 4. S. E. Harris, M. K. 0shman and R. L. Byers, Phys. Rev. Letters l8, 732 (1967). 5. M. H. Crowell, IEEE J. Quant. Elect. QE-1, 12 (1965). b. T. Uchida, IEEE J. Quant. Elect. QE-3, 7 (1967). 7. T. Uchida and A. Ueki, IEEE J. Quant. Elect. QE-3, 17 (1967). 8. F. R. Nash, IEEE J. Quant. Elect. QE-3, 189 (1967). 9. A. Javan, W. R. Bennet, and 0. R. Herriott, Phys. Rev. 6, l06 (1961) 10. R. N. Hall, G. E. Fenner, J. D. Kingsley, T. J. Soltys, and R. 0 Carlson, Phys. Rev. Letters 9_, November (1962). 11. M. E. Nathan, W. P. Dumke, G. Burns, F. H. Dill, Jr., G. T. Lasker, Phys. Rev. Letters, 9_, November (1962). 12. T. M. Quist, R. H. Rediker, R. J. Keyes, W. E. Drag, B. Lax, A. L. McWhorter, and H. J. Feiger, Appl. Phys. Letters, 9, December (1962) 13. B. A. Lengyel, Introduction to Laser Physics, Wiley (1966). 14. W. V. Smith, and P. P. Sorokin, The Laser, McGraw-Hill (1966). 15. S. Fine and E. Klein, Biological Effects of Laser Radiation Manage- ment Study, Report on Contract AF-29(600)-5136. Submitted to Air Force Missile Development Center, Air Force Systems Command of the United States Air Force, Parts I and II (1965). 16. P. A. Franken and J. F. Ward, Rev. Mod. Phys. 35, 23 (1963). 17. R. H. Kingston, Lasers, in The Encyclopedia of Physics, Edited by Robert M. Besancon, Reinhold Publishing Company, New York (1966). 18
18. M. S. Litwin, K. M. Earle, and D. H. Glew, Biological Effects of Laser Radiation, Washington, D. C., Federation of American Societies for Experimental Biology, 177 (1965). 19. M. S. Litwin and D. H. Glew J.A.M.A. 187. 842 (1964). 20. A. L. Schawlow, Sciet. Amer. 209, 34 (1963). 21. A. L. Schawlow, Science J4£, 13 (1965). 22. K. E. Schuler and W. R. Bennett, Jr. , Chemical Lasers, Washington, D. C., 0ptical Society of America, Supp. 2 to Applied 0ptics, 224 (1965). 19
