Plasma Monitoring of Laser Beam Welds

Light and sound intensity measurements

can be used for the monitoring of plasma initiation and propagation

during laser beam welding

BY G. K. LEWIS AND R. D. DIXON

ABSTRACT. Experimental and theoretical depth of joint penetration for any one set with industrial CO2 lasers. studies using high power density laser of laser conditions, particularly on high Although the same plasma theory pulses (greater than 109 W/cm2) at pulse reflectivity materials. The data, along with explained in the text applies to the higher lengths less than 1 microsecond have the appreciation that more must be multikilowatt power levels, the CO2 proven the existence of laser supported known about the process, led to experi­ lasers produce different plasma waves absorption waves. In our ND:YAG laser ments using acoustic emission as a moni­ with different characteristics than the LSC beam welding power and pulse regimes toring technique. These studies showed waves experienced in the lower power (400 watts maximum average power with significant acoustic energy generation but regime of the Nd:YAG The type of pulse durations of 0.5-8 ms), high speed no apparent correlation with weld mor­ plasma wave produced depends on the photography, microphone, and light phology. The affect of the plasma was power density of the laser beam. Hence, intensity measurements show that multi­ not considered in these studies. As a the laser-plasma interaction may change ple laser supported waves form during a result, a literature review and experimen­ as power is increased. Other effects such single pulse and produce enhanced cou­ tal study were performed to determine as radiation trapping during deep joint pling with the target. the laser light-plasma-material interaction penetration keyhole welding are not sig­ process and the affect on welding. nificant at the power levels discussed here. Introduction Theoretical and experimental studies, primarily at power densities much greater The use of lasers for welding and 9 2 than 10 W/cm and pulse lengths of less Laser-Plasma-Material Interaction materials processing has become an than 1 microsecond (ps) duration, have important manufacturing process. The attempted to characterize the laser-plas­ Absorption of laser energy depends Nd:YAG (1.06 pm)* and C02 (10.6 pm)* ma-target interaction (Refs. 1-25). These initially on the intrinsic absorptivity of the lasers have evolved into primary tools for studies have shown that the interaction target material. Absorptivity generally laser beam welding operations. Typical involves many variables such as the type increases with decreasing wavelength; it Nd:YAG systems are pulsed with average of plasma formed, laser light intensity, may increase because of a change in powers up to 400 watts, while CO2 wavelength, interaction time, energy spa­ metal temperature, base metal oxidation systems may be pulsed or continuous tial and temporal distribution, environ­ state or other surface reaction. wave with average powers up to 20 ment above the target, surface condi­ Figure 1 shows the spectral reflectance kilowatts. These systems are used on a tions, target composition, and material (reflectance equals 1-absorptivity) for wide range of metals for low heat input, physical properties, among others. various elements at room temperature. high precision welds. Although far removed from the laser At the Nd:YAG and CO2 wavelengths 4 9 At the Los Alamos National Laboratory, beam welding power regime of 10 — 10 (1.06 and 10.6 pm, respectively), less than 2 a significant effort is in progress to devel­ W/cm and pulse durations of several 10% of the light is absorbed by Al and less op techniques for controlling and certify­ milliseconds to continuous wave, these than 5% by Cu, Ag and Au. Absorption of ing the laser beam welding process. Initial studies have provided information appli­ the laser radiation typically does not weld morphology vs. weld variable cable to the plasma effects occurring account for the amount of material studies indicated a large variation in during laser beam welding. melted during the welding process, par­ This paper deals with the laser-plasma- ticularly for highly reflective materials. material interaction as experienced with a Instead, an increase in coupling efficiency * Ifim (micrometer) = 0.00004 in. Raytheon Nd:YAG laser in the pulsed is noted as laser intensities are increased mode at average power levels up to 400 above plasma initiation threshold intensi­ ties. This phenomenon is referred to as Based on a paper sponsored by the DOE watts. In this power regime, we are con­ "enhanced coupling" (Ref. 18) and is Interagency Group for presentation at the 64th cerned primarily with low joint penetra­ Annual A WS Convention held in Philadelphia, tion welds on the order of 1.5-3.0 mm related to the plasma formation and sub­ Pennsylvania, during April 24-29, 1983. (0.06-0.12 in.) penetration. The results of sequent radiation heating of the base metal by the plasma. G. K. LEWIS and R. D. DIXON are with the our studies are applicable to this welding Materials Science and Technology Division, regime, and we have not extended the The significance of the radiant heat flux Los Alamos National Laboratory, Los Alamos, same experimental tests to the multikilo- from the plasma to the base metal is New Mexico. watt deep joint penetration welds made shown in Fig. 2 and equation (1). As the

WELDING RESEARCH SUPPLEMENT 149-s A :: ::: 0.8 - ' 4r^^^^^ ^^ ^ 0.6 " // ' A c/> 0.4 ~ 2^r^ ,000 0.2

i i i i i i II 0.2 0.4 0.6 LO 2.0 4.0 6.0 10.0 5000 10,000 5,000 20.000& Wavelength , pm \— Fig. 1 — Reflectance as a function of wavelength Fig. 2 — Blackbody energy flux emitted as a function of wave­ length and temperature temperature of the plasma increases, the 1). This is a significant improvement target leaving a shocked region for the maximum energy flux emitted increases over the 1-10% absorbed without the absorption wave to propagate into. This by the fourth power of temperature and plasma. process is shown schematically in Fig. 3 shifts towards shorter wavelengths. The Because the plasma formation signifi­ (from Ref. 1) with the regions for the short wavelengths are more readily cantly affects the welding process, initia­ relevant plasma variables: pressure (P), absorbed than the infrared laser radiation tion and propagation mechanisms must density (p), particle velocity (U), tempera­ as shown in Fig. 1. The total energy be understood. Incident laser light partial­ ture (T), enthalpy (h) with spatial coordi­ emitted, assuming the plasma radiates as ly absorbed by the target and surround­ nates X and L. The subscript S refers to a black body, is given by the Stefan- ing medium causes heating, vaporization the shock region and the subscript co Boltzmann law: of asperites at high energy density sites, refers to ambient conditions. Note the Wp- = o-T4 (1) electron emission and ion emission (Ref. laser is incident from the left and that the where rr = 5669 X 10"5 ers cm"2 s"1 k~4 20-25). The electron, ion, and neutral absorption wave has been identified as a and T = absolute temperature. atoms constitute a plasma that ignites and Laser Supported Combustion (LSC) wave with velocity (v ). Laser plasmas typically reach tempera­ absorbs the incident laser energy when w tures of 5000-20,000 K (Ref. 1) and thus the plasma temperature and density Laser-target interactions produce two can contribute a significant amount of become high enough. Additional mecha­ distinct types of absorption waves: the heat to the part. However, calculation of nisms contribute to the plasma formation previously mentioned LSC wave and a the heat flux from plasma to target must for low power laser systems. The most laser supported detonation (LSD) wave. account for changing plasma tempera­ dominant are: thermoelectron emission, Both waves contribute to enhanced cou­ ture due to plasma density, velocity, photo-electron emission and inverse pling, but the LSC wave is optimal. Pirri, volume and other factors that are difficult "bremstrallung" in the medium above Kemp, Root and Wu (Ref. 1), among to measure. An upper limit of such a the target. other investigators, have modeled the calculation assumes that the laser light is The ignition is signalled by the creation plasma wave formations and characteris­ totally absorbed by the plasma and yields of a laser supported absorption (LSA) tics. Some of their results are discussed an absorption of energy by the material wave and is preceded by a pressure below, and their models are shown in of 50% of the incident laser energy (Ref. pulse that propagates away from the Figs. 3-5.

Precursor Surface - Surface Shock x* L Laser Intensity y *s I, LSC Wave r .... / a) fc. Shock > JI /_Wave VelocitVeto y Equals Particle \ ^ Shocked |" 's Velocity Behind Shock Air | II C- •IB piasma ^Surface Loser I > ^ Laser Intensity I >I, Weak LSD Wove Flux l( 1 1 <> 2 Shock ! 1 b) •> KP | s | • '/ Wave Velocity > Particle Velocity HI 1> Poo u ; i Surface K " x> Laser Intensity shoc|( •~-;J >ck 13**2 LSD Wove X c) 1\ Vw = Abso lute LSC Wave Velocity -Exp. Fan Fig. 3 —Schematic representation of a LSC wave (Ref. 1) Fig. 4 — Transition from LSC to LSD wave (Ref. 1)

50-s | FEBRUARY 1985 Laser Beam Laser Beam

Shock Wave

Thin Laser •LSD Wave Absorption Zone

±*Hot , High Pressure Air T Blast Wave Expansion FansV ! i: fr Radiation To Surface

vyy^yys^iyyyyy/ V7pfrs^77y Conduction Boundary Conduction Layer Fig. 5— Two-dimensional LSC and LSD waves (Ref. 1)

After plasma initiation the absorption waves; it shows some of the more salient without an aperture in the beam path. wave will propagate into shocked space. features such as the surface boundary The power was measured for each pow­ The absorption wave velocity and the region and the blast wave within expan­ er condition using a Coherent Model 213 conditions within the shocked space sion fans seen in the LSD waves. power meter. The laser pulse shape was behind the precursor determine the type Calculations from one dimensional LSC temporarily monitored on an oscillo­ of absorption wave. If the precursor wave theory (Ref. 1) show that, after scope using a detector on the back of the wave imparts sufficient energy to the air 5 -2 ignition by a 1 X 10 watt-cm C02 laser cavity coupled with a 0.01% trans­ (environment) to cause absorption of the laser, a LSC wave will travel 2 cm (0.79 mitting rear cavity mirror. incident laser energy, then the absorption in.) in less than 100 ps. Two dimensional Plasma shock waves were monitored wave created travels at the shock wave calculations (Ref. 16) for 1 X 106 watt- by a Buehl and Kjaer microphone with a velocity causing considerable mass flow cm-2 CO2 laser radiation show that a 200 Hz to 20 kHz range. The microphone through the wave. Such a wave is de­ high temperature isotherm closes upon was placed approximately 12.5 mm (ap­ noted as a LSD wave. itself in the 2-3 cm (0.79-1.18 in.) region prox. V2 in.) from the focal spot. Visible If the shock is not energetic enough to in less than 100 ps, suggesting separation plasma light was monitored with a photo- allow the shocked gas to absorb the laser of the plasma wave from the target multiplier tube (PMT) having a S-20 energy, then the absorption wave fol­ surface. Most laser welding occurs at response. The tube was coupled to the lows the shock wave at a lower velocity times greater than 100 ps; this implies that laser spot by a fiber optic light guide and the driving energy is from the plasma a series of LSC waves will continually placed within 12.5 mm (0.49 in.) of the radiation produced by absorption of laser initiate, propagate and decay during the focal spot of the laser beam. The fiber light in the plasma. This wave is known as welding pulse. had a 30 deg included angle cone of a LSC wave and is characterized with acceptance. The PMT was not sensitive subsonic wave velocity and less mass to 1.06 pm light from the laser. Experimental Description transport compared to the LSD wave. High speed movies were taken at 4000 Because the plasma temperature de­ The models for LSA waves suggest frames per second with black and white pends upon the input energy rate and the several diagnostics to correlate the plas­ negative film. A light inside the film cavity mass flow rate, the LSD wave will con­ ma initiation, growth and regeneration to turned on with the event trigger pulse tribute less radiant heat to the base metal the laser pulse and resultant target and exposed one edge of the film for a than the LSC wave. affects. Each LSA wave formation is char­ starting time reference. Timing marks The conditions for transition from a acterized by a precursor shock wave were placed on the film for speed calcu­ combustion to detonation wave are followed by a visible plasma wave that lations. shown in Fig. 4. Figure 4A represents a propagates away from the target. The A four channel digital oscilloscope was LSC wave propagating into shocked gas plasma process can be monitored by used to record the diagnostic signals at a at a velocity less than or equal to the using a microphone to detect the shock rate of 5 ps/point. The scope was trig­ shocked gas particle velocity. As the laser wave and a photomultiplier to detect gered by a logic circuit that output a intensity is increased, an absorption wave visible light from the plasma. The signals trigger pulse when the laser shutter was is created that has a velocity greater than are correlated to the laser pulse charac­ open and the next laser pulse occurred. the shocked particle velocity. Such a teristics and high speed photography The trigger pulse started the sweep on wave is not quite a LSC or LSD wave and shown schematically in Fig. 6. the scope and turned the light on inside is denoted as a weak LSD wave as shown A pulsed Nd:YAG (1.06 micron wave­ the movie camera. in Fig. 4B. When the laser intensity is very length) laser (Raytheon Model SS500) All targets were 2.22 cm (0.874 in.) high, the shocked region absorbs the capable of a maximum average power of diameter by 0.15 cm (0.059 in.) thick laser energy and a LSD wave is gener­ 400 watts was used for this study. The discs. The discs were mounted on a ated; the latter condition is shown in Fig. pulse length could be varied from 0.5-8.0 rotary base to index the target on a 4C. Figure 5 gives a two-dimensional ms. Single pulse welds were made at common axis. The target materials were representation of the LSC and LSD visual sharp focus on base metal surfaces Type 304 stainless steel, 5052 aluminum

WELDING RESEARCH SUPPLEMENT 151-s Nd-YAG 304 Stainless Steel 320 Watts Laser Rod Turning Mirror ^r Si Detector

Laser Beam

4-Channel Amplifier Storage Oscilloscope 304 Stainless Steel Focus Lens- ^ 320 Watts

C^-^Micicrophonr e Plasma Target I z5 FiDer °Ptic

Photomultiplier

Rotary Base

Fig. 6 - Laser plasma diagnostics experimental set up

alloy, and pure gold. These materials ma initiation and propagation past the were chosen to be representative of fiber optic detector. both high reflectivity and low reflectivity The microphone and photomultiplier materials. Single spot welds were made signals were correlated in time by com­ on each material at 320 watts and 70 paring peak times for both signals. These watts average power at a pulse length of times correlated well in many cases, par­ Fig. 7 —Signals for Type 304 stainless steel at 8.0 ms. Each target was swabbed with ticularly in the beginning stages of plasma 320 W average power and 8 ms pulse: A — alcohol just prior to making the spot initiation. In each microphone signal laser; B —microphone; C—photomultiplier weld. shown, an intense negative spike indi­ cates the start of multiple plasma regen­ eration and may be correlated to peaks Results and Discussion or slope changes that are slightly later in ending at t = 1.0-1.25 ms where the Typical signals recorded for 320 watts time on the photomultiplier signals. These plasma propagates out of the picture. average power laser pulses on Type 304 correlations are masked somewhat by Similar correlations may be made for stainless steel, 5052 aluminum alloy and microphone ringdown and by the inte­ the photographs in Figs. 11 and 12 with coated gold are shown in Figs. 7-9. The gration of the plasma wave light intensity the data in Figs. 8 and 9, respectively, for relative signal intensities are not signifi­ signals caused by the wide cone of 5052 aluminum alloy and coated gold. cant, because the gains and detector acceptance of the fiber optic detector. The photographs shown correspond to distances from the source were varied Photographs of the plasma formations signal changes for the first parts of the depending on material. It was observed for the data in Fig. 7 are shown in Fig. 10. laser pulse. At times greater than those that the signals recorded from the Type Each frame represents an integrated shown in Figs. 10-12, many plasmas were 304 stainless steel plasmas were of much event time of 250 ps. The first small formed and propagated higher than the greater intensity than the signals from negative microphone signal was at height of the picture until the laser pulse the 5052 aluminum alloy and pure gold t = 0.38 ms corresponding to the first ended. plasmas. visible plasma formation in the frame In comparing the plasma data for the The most prominent features of these representing t = 0.25 ms to t = 0.50 ms. three materials studied, the Type 304 signals are the multiple peaks shown in The photomultiplier data in Fig. 7B show stainless steel plasmas initiated earlier dur­ the microphone and photomultiplier the first rise in signal intensity initiating at ing the laser pulse and regenerated other curves. These peaks indicate that many 0.38 ms and continuing to rise until plasmas more rapidly than in the tests on plasma waves form sequentially during a t = 0.57 ms where the intensity levels off. 5052 aluminum alloy and coated gold. In single 8 ms laser pulse. Each major micro­ The large negative peak in the micro­ tests at 70 watts average power, plasmas phone peak and associated ring down phone signal starts at t = 0.73 ms and were initiated to a lesser extent for Type represents a precursor shock wave, and continues to t = 0.90 ms corresponding 304 stainless steel and 5052 aluminum each slope change and inflection point in to the steep increase in the photomultipli­ alloy and not at all for pure gold without the photomultiplier signal represent plas­ er tube signal starting at t = 0.73 ms and a coating. For pure gold without a coating

52-s | FEBRUARY 1985 mu Fig. 10-Photographs at 4000 fps correspond­ ing to data in Fig. 7. From left to right, pictures represent 0.25 ms intervals starting at event time = 0.25 ms. X5.6 (reduced 40% on repro­ duction)

5052 Al 1.0 320 Wotts

0 I V m*.j^.r\.r\f\t^~^- ii Fig. 12-Photographs at 4000 fps correspond­ -1.0 H ing to data in Fig. 9. From left to right, pictures represent 0.25 ms intervals starting at event times as follows: A—0.60 ms; B — 2.60 ms. -2.0 " ® X5.6 with reductions as follows on reproduc­ tion: A-27%; B-50%

5052 Al tion energy threshold must be crossed to 2.0 320 Watts start the plasma regeneration process. -^ ft Once that energy level is exceeded, a \ l\ /\/\ A LSA wave is formed and propagates up S 1-5 - j\ \ II 1 »1 the beam. The LSA wave allows some k 1 1 n \ A 111 \ fraction of the incident laser fluence to = 1.0 1 VV \ reach the target. If that amount of energy plus the plasma radiant energy is less than £ 0.5 - j the plasma initiation threshold, a second plasma cannot form. As the wave propa­ 1 1 1 1 gates up the focused beam, it may decrease in density because of increased Fig. 8 —Signals for 5052 aluminum alloy at 320 turbulence or lower laser power density W average power and 8 ms pulse: A—laser; and allow a larger fraction of incident B — microphone; C—photomultiplier uf\1 laser light to reach the target. At some time during the life of the wave, the fraction of incident light allowed through at 320 watts, a very small plasma was will exceed the plasma threshold energy, Fig. 11 — Photographs at 4000 fps correspond­ barely visible on the 4000 fps film. The and another LSA wave can form and ing to data in Fig. 8. From left to right, pictures propagate away from the surface. addition of an ink coating using the same represent 0.25 ms intervals starting at event laser parameters produced multiple plas­ times as follows: A—0.45 ms; B — 2.45 ms; mas during the pure gold test. In each C—3.45ms. X5.6. With reductions as follows Conclusion case where plasma formation and regen­ on reproduction: A—28%; B—38%, C—40% eration were not detected, there was no Plasma initiation and propagation dur­ visible melting of material. As the number of molten material increased. ing laser welding can be monitored by of LSA waves increased, either by The change in the number of plasmas measuring the light intensity and sound increasing laser power or by application formed with power increase or surface intensity that is characteristic of a laser of a coating in the case of Au, the amount modification suggests that a plasma initia- supported absorption wave. The number

Au-Coated 320 Watts

Fig. 9 —Signals for gold (coated) at 320 W and 8 ms pulse: A—laser; B—microphone; C—photomultiplier

WELDING RESEARCH SUPPLEMENT 153-s of plasmas produced during a single 8 ms head to follow the weldment. Plasma 2145. laser pulse depends on substrate materi­ activity could be monitored with statisti­ 8. McKay, J. A., Bleach, R. D., Nagel, D. J., al, surface condition and laser power cal sampling at various points along the and Schriemph, J. T. 1979. Pulsed-C02-Iaser interaction with aluminum in air: thermal density. The amount of melting increases weldment for analysis. response and characteristics, journal of as the number of plasma regenerations The monitoring technique described is Applied Physics 50(5):3231-3240. increases; this suggests that radiant heat­ by no means complete. However, it 9. Walters, C. T., Barnes, R. H„ and Beverly, ing from the plasma is an important would help in identifying the power R. E. 1978. Initiation of laser-supported-deto- heating mechanism during the laser beam regimes where the phenomenon of nation (LSD) waves, journal of Applied Physics welding process. An ink coating used on enhanced coupling is significantly chang­ 49(5):2937-2949. a gold target increased the number of ing the amount of energy coupled to the 10. Krokhin, O. N. 1972. Generation of plasmas formed and increased the target surface. An improved monitor high-temperature vapors and plasmas by laser amount of molten material formed com­ must account for the energy balance for radiation. Laser handbook, eds. F. T. Arecchi pared to gold without a coating. the entire process. Such a monitor and E. O. Schulz-DuBois, pp. 1371-1406. North-Holland Publishing Co. requires spatial and temporal character­ ization of the laser pulse, measurement 11. Pirri, A. N., Root, R. G., and Wu, P. K. S. 1978. Plasma energy transfer to metal surfaces of the absorbed and reflected light at the Process Monitoring of Laser Beam Welds irradiated by pulsed lasers. AIAA lournal material surface and measurement of the 15(12):1296-1304. The experiments described have absorbed, transmitted and reflected light 12. McKay, J. A., and Schriempf, J. T. 1979. attempted to characterize the interaction from the plasma. Work in these areas is Transient surface heating of metals by CO2 between the incident laser light, the plas­ continuing to understand the laser beam laser pulses with air-plasma ignition, journal of ma formation and the target material welding process more completely than is Applied Physics 50(8):5202-52O5. during pulse welding with a Nd:YAG presently known and to develop the 13. Beverly, R. E., and Walters, C. T. 1976. laser. Although the overall laser beam monitoring instrumentation required. Measurements of CCVlaser-induced shocked welding process is complex, the instru­ pressures above and below LSD-wave thresh­ ments used in these tests may lead to a olds, lournal of Applied Physics 47(8):3485- method of real time monitoring, at least in A ckno wledgments 3495. our Nd:YAG welding regime of average 14. Raizer, Q. P. 1965. Heating of a gas by a The authors wish to acknowledge the powerful light pulse. Soviet Physics jEPT powers up to 400 watts. assistance and expertise of the following 21(5):1009-1017. At present, microphones and fiber people: Mike Ward, EG&C, Los Ala­ 15. Basov, N. G, Gribkov, V. A., Korkhin, optic light detectors are used to signal mos—set up the instrumentation and O. N., and Sklizkov, G. V. 1968. High temper­ plasma initiations and propagation. The fiber optics for the plasma light intensity ature effects of intense laser emission focused number of plasmas generated appears to measurements; Tony Rollett and John on a solid target. Soviet Physics jEPT 27 (4):57 5- 582. correlate with the weld pool penetration Buchen, Los Alamos National Laborato­ 16. Jackson, J. P., and Nielsen, P. E. 1974. in a target. Velocities of plasma shock ry—wrote the data reduction computer waves moving away from the target can Ignition transients and two-dimensional effects program and designed the triggering cir­ in the propagation of laser-supported-absorp- be calculated by spacing the micro­ cuit for data acquisition, respectively; tion waves. Laser Digest, report AFWL-TR- phones at a known distance and measur­ Carolina Munoz, University of Texas, El 74-100:212-226. ing the time interval between micro­ Paso — correlated diagnostic signals to 17. Boni, A. A., and Su, F. Y. 1974. Propaga­ phone signals for any single plasma initia­ high speed photographs. In memoriam, tion of laser supported deflagration waves. tion. The wave velocity characterizes the Robb Gordon, Los Alamos National Labo­ The Physics of Fluids 17(2):340-342. wave as LSC (subsonic) or LSD (superson­ ratory—photographed the plasma for­ 18. Marcus, S., Lowder, J. E., and Mooney, ic). The presence of LSC waves indicates mations at 4000 fps. D. L. 1976. Large spot thermal coupling of CO2 that enhanced coupling may dominate laser radiation to metallic surfaces, journal of the process, implying that radiation from Applied Physics 17(7):2966-2968. References the plasma is contributing significantly to 19. Dimitrijevic, M., and Konjevic, N. 1980 the weld pool depth. 1. Pirri, A. N., Kemp, N. H., Root, R. C, and (lune). The importance of the pulse shape for Wu, R. K. S. 1977 (Jan.). Theoretical laser the laser-beam target interaction. Optics and Ideally, plasma temperature data cou­ effects studies. Final report, PSI-TR-89. Physical laser technology, pp. 145-147. pled with the velocity and initiation data Sciences, Inc. 20. Smith, D. C. 1971. Gas breakdown would allow calculation of the radiation dependence on beam size and pulse duration 2. Boni, A. A., Su, F. V., Thomas, P. D., and with 10.6 pm wavelength radiation. Applied flow to the target. Temperature mea­ Musal, H. M. 1977 (April). Theoretical study of Physics Letters 19(10):405-408. surement can be made by measuring laser target fabrication. Final report, SAI-77-567 plasma density with time but is complex. LJ. Science Applications, Inc. 21. Smith, D. C. 1979 (Aug.). Laser induced gas breakdown and plasma shielding. SPIE However, comparison of numbers of 3. Casey, H. 1980 (Sept.). Plasma phenome­ 195:171-181. plasmas and their velocities with depths na during Nd-YAG laser welding. Sixth Interna­ 22. Smith, D. C, and Brown, R. T. 1975. of penetration at various laser settings tional Conference of the Institute of Electrical Aerosol-induced air breakdown With CO2 can be made and a relative evaluation of Engineers/Welding Institute, Edinburgh, UK. laser radiation. Journal of Applied Physics 46 penetration depth established. If the 4. Nichols, D. B., and Hall, R. B. 1978. Threshold conditions for the formation of (3):1146-1154. wave is LSD, other factors such as reflec­ surface plasmas by HF and DF laser radiation. 23. Smith, D. C. and Fowler, M. C. 1973. tion of laser light from the plasma may /. Applied Physics. 49(10):5155-5169. Ignition and maintenance of a CW plasma in become significant and decrease the 5. Godwin, R. P. 1979. Absorption in laser- atmospheric-pressure air with CO2 laser radia­ enhanced coupling affect. The LSD wave produced plasma experiments: a personal tion. Applied Physics Letters 22(10):5OO-5O2. regime is recommended as the subject of view. Applied Optics 18(21):3555-3561. 24. Smith, D. C. 1977. Gas breakdown initiated by laser radiation interaction with future work. 6. laonimagi, P. A., and Richardson, M. C. aerosols and solid surfaces, journal of Applied 1983. Time resolved x-ray photography of the The instrumentation described for such Physics 48(6):2217-2225. expansion of nanosecond C0 laser produced a monitor is practical from the standpoint 2 plasmas. Optics Communications 44(3):180- 25. Fowler, M. C, and Smith, D. C. 1975. of size and non-interference with the 184. Ignition and maintenance of subsonic plasma welding process. Microphones 8-12 mm waves in atmospheric pressure air by CW CO2 7. Maher, W. E., Hall, R. B., and Johnson, laser radiation and their effect on laser beam (0.31-0.47 in.) diameter and fiber optic R. R. 1974. Experimental study of ingnition and propagation, lournal of Applied Physics strands 0.2-1.0 mm (0.008-0.04 in.) can propagation of laser-supported detonation 46(1):138-149. easily be attached to the laser focusing waves, lournal of Applied Physics 45(5):2138-

54-s I FEBRUARY 1985