Christian G. Parigger,1,* James O. Hornkohl,1 and László Nemes2
1The University of Tennessee Space Institute, 411 B.H. Goethert Parkway, Tullahoma, Tennessee 373882Laboratory for Laser Spectroscopy, Chemical Research Center of the Hungarian Academy of Sciences, Pusztaszeri ut
The study of laser-induced aluminum plasma is ofinterest in material science with applications for ex-ample in micromachining. Pulsed Q-switched laser ra-diation has found numerous machining applications.Precise micromachining is typically achieved by use ofUV radiation from excimer lasers and�or more re-cently by use of radiation from picosecond or femtosec-ond lasers. Associated with laser material processingis the laser-generated plasma. Here we explore the useof aluminum emission spectroscopy for diagnostic pur-poses of the machining applications.
Laser-induced optical breakdown of gases and sol-ids generates high-temperature microplasma thatshows rich emission spectra. Early in the plasmadecay, after several tens of nanoseconds, spectro-scopic signatures include atomic species embedded infree-electron background. Later in the plasma decay,after several hundreds to thousands of nanoseconds,the typical signatures of highly excited molecular re-combination spectra can be recorded. Laser-inducedbreakdown spectroscopy (LIBS) studies typically uti-lize nominal nanosecond laser radiation for genera-tion of the microplasma [1,2].
In the work presented here, we apply atomic andmolecular emission spectroscopy methods to evaluatethe microplasma at and near the surface of metallicaluminum. The microplasma is induced by nomi-nal nanosecond, Q-switched Nd:YAG laser radiation.Atomic spectroscopy techniques [3,4] can be employedto investigate the temporal evolution of electron num-ber density. Subsequently, molecular emissions [5–8]can be used to characterize postbreakdown phenom-ena. Of interest is the use of hydrogen Balmer serieslines, in particular H� and H�, to infer the plasmastate generated from focusing laser radiation ontosolid aluminum.
The dynamics of the laser–plasma interaction in-cludes high-irradiance interaction with metal vaporduring the laser ablation of aluminum [9]. Thehigh-density vapor shows temperatures exceeding10,000 K ��1 eV� and electron number densities of5 � 1018 cm�3. Pronounced plasma absorption occursvia photoionization typically of the order of milli-meters from the surface, when using an irradianceof 1–2 GW�cm2 for visible copper vapor laser orfrequency-doubled Nd:YAG laser radiation. However,as the irradiance of the laser is increased by approxi-mately 6 orders of magnitude to 900,000 GW�cm2,laser-surface experiments [4] using a high-dispersionx-ray spectrometer to measure the hydrogenic alumi-num Lyman series, show electron temperatures of
the order of 1 keV. The electron number densitiesamount to Ne � 2 � 1022 cm�3 close to the surface.Absorption by inverse bremsstrahlung is the domi-nant mechanism for high Ne. Here we explore theapplication of the hydrogen Balmer series lines inthe characterization of microplasma near aluminumsurfaces. Hydrogen Balmer series lines allow one toinfer number densities up to �1 � 1018 cm�3 from H�
and up to �1 � 1019 cm�3 from H� [10,11].
2. Experimental Details and Results
Optical breakdown is generated using radiation fromthe continuum Q-switched Nd:YAG laser at a 10 Hzrepetition rate. Here we use 12 ns pulses for the1064 nm fundamental mode that are focused with a10 cm focal length lens onto an aluminum block. Thefocal spot diameter is typically 30 �m, yielding anirradiance of approximately 300 GW�cm2. This irra-diance level is approximately a factor of 3 above theoptical breakdown threshold for IR radiation in lab-oratory air and several orders of magnitude above theoptical breakdown threshold for plasma generationat and near the aluminum block.
For the experimental studies we used the OphirModel Wavestar-U to record a spatially and tempo-rally integrated spectrum of the plasma emission. Wealso used a 0.5 m monochromator Model Acton Spec-traPro with an intensified linear diode array and anoptical multichannel analyzer to store spatially aver-aged but time-resolved emission spectra of selectedatomic lines and molecular bands. An optical fiberwas used to collect the emission from the breakdownplasma, and this fiber was directly coupled to theWavestar or Acton monochromator.
Figure 1 shows a typical overview spectrum fromlaser-induced optical breakdown events. The occur-rence of the aluminum lines is noteworthy, and herewe marked the two neutral lines measured at 394.5and 396.2 nm and the singly ionized line at 466.4 nm.The wavelength resolution of the Wavestar mono-chromator is 0.5 nm with indicated peak wave-lengths to 0.1 nm accuracy. For this experimentalstudy we operated asynchronously the Wavestarmonochromator and the Nd:YAG laser. An openingtime of 500 ms for the Wavestar shutter was selected,
which resulted in the capture of 4 to 5 plasma emis-sions induced by 10 Hz repetition rate laser radia-tion.
Selected aluminum atomic lines were studied us-ing time-resolved techniques. Figure 2 shows therecorded, single-shot spectra for several time delaysfrom optical breakdown. A gate width of 0.1 �s wasselected to record at the indicated time delays of 0.45,0.75, 1, and 1.5 �s. The resolution of these time-resolved spectra amounts to approximately 0.4 nmfor the selected slit width of the 0.5 m monochroma-tor with a 1200 groove�mm grating. The early Allines are intense, Stark-broadened, and noticeablyStark redshifted.
Molecular recombination radiation can be observedlater in the plasma decay. Aluminum monoxide spec-tra were recorded at a time delay of 50 �s, using agate width of 10 �s. Figure 3 shows the experimentalresults combined with the synthetic fit of the molecularspectrum. The spectral resolution amounted to 0.4 nmfor the emission spectra collected from 300 subse-quent laser-induced breakdown events. An equilib-rium AIO temperature of T � 4220 K was inferredfrom the fitted, synthetic AIO spectrum.
For the initial time-resolved experiments [11] weused 150 mJ, 8 ns laser pulses with a peak irradianceof 1000 GW�cm2. This peak irradiance correspondsto an electric field strength of E � 1�100 � 2.8GV�cm, or 1�100 of the field strength that holds thehydrogen atom together. Hydrogen Balmer serieslines are used to infer the number density of alumi-
Fig. 1. (Color online) Spatially and temporally integrated emis-sion spectrum following nanosecond Nd:YAG laser-generation-induced optical breakdown at and near the aluminum block.
Fig. 2. (Color online) Time-resolved, single-shot aluminum 394.4and 396.2 nm atomic line emissions for selected time delays fromthe optical breakdown pulse.
Fig. 3. (Color online) Measured and synthetic molecular spectraof aluminum monoxide 50 �s after nanosecond Nd:YAG laser op-tical breakdown.
num expanding into a cell filled with 125 Torr ofhydrogen and mixtures of residual air and argon atan estimated pressure of an additional 50 Torr.
For the work reported below, an energy of 45 mJ isavailable for 12 ns laser pulses, resulting in a peakirradiance of 300 GW�cm2. Meticulous attention wasgiven to evacuating the cell and filling it to a level of100 Torr hydrogen gas, without the presence of re-sidual air or argon. The spectra were recorded usingfiber imaging and coupling to a 0.5 m Acton spec-trometer with a 2400 groove�mm grating, and for theselected slit width a spectral resolution of 0.34 nm.The individual points were measured using an in-tensified (6 ns gate) optical multichannel analyzer.The data were wavelength calibrated by use of
hollow-cathode discharge lamps and were linear di-ode array-sensitivity corrected by use of a tungstenlamp. Figure 4 illustrates the 21 point, second-orderSavitzky–Golay smoothed results for time delays of25, 50, 100, and 200 ns. For the earlier time delay�25 ns� from the 12 ns laser pulse, note the relativelylarge background contribution from free-electron ra-diation. Previous studies [11] with higher laser en-ergy of 150 mJ per pulse, for slightly higher hydrogenpressure and residual laboratory air, show the occur-rence of singly ionized Al II 390.0675 nm and a bunchof lines near 399 nm: Al II 399.6182, Al II 399.6143,Al II 399.6143, Al II 399.6075, and Al 399.5838 nm.In the current 100 Torr hydrogen experiments, and
Fig. 4. (Color online) Recorded Stark widths and shifts for Al I 394.40 nm and Al I 396.15 nm lines from aluminum plasma in 100 Torrhydrogen.
Table 1. Linewidth (FWHM) and Redshift for � � 394.4 nm Line of Al I
Delay[ns]
Width[nm]
Shift[nm]
Ne
[1018 cm�3]
25 0.59 � 0.08 0.28 � 0.03 1.2 � 0.5a
30 0.52 � 0.08 0.25 � 0.03 1.0 � 0.5a
50 0.25 � 0.05 0.16 � 0.03 0.9 � 0.4a
100 0.17 � 0.05 0.095 � 0.03 0.3 � 0.05b
150 0.15 � 0.05 0.042 � 0.02 0.2 � 0.02b
200 0.084 � 0.04 0.036 � 0.02 0.1 � 0.01b
400 0.002 � 0.04 0.027 � 0.02 0.05 � 0.01b
aThe electron number density is inferred using FWHM of H�.bThe electron number density is inferred using FWHM of H�
and H�.
Table 2. Linewidth (FWHM) and Redshift for � � 396.15 nm Line ofAl I
Delay[ns]
Width[nm]
Shift[nm]
Ne
[1018 cm�3]
25 0.72 � 0.08 0.31 � 0.03 1.2 � 0.5a
30 0.61 � 0.08 0.29 � 0.03 1.0 � 0.5a
50 0.42 � 0.08 0.17 � 0.05 0.9 � 0.4a
100 0.26 � 0.08 0.11 � 0.05 0.3 � 0.05b
150 0.22 � 0.05 0.053 � 0.03 0.2 � 0.02b
200 0.087 � 0.04 0.047 � 0.03 0.1 � 0.01b
400 0.003 � 0.04 0.038 � 0.03 0.05 � 0.01b
aThe electron number density is inferred using FWHM of H�.bThe electron number density is inferred using FWHM of H�
using approximately 50 mJ�pulse, 12 ns Nd:YAG IRlaser radiation, these lines are not observed.
Tables 1 and 2 show the aluminum Stark widthand Stark shifts for selected time delays from opticalbreakdown. The listed values are deconvolved fromthe spectrometer resolution of 0.34 nm. The typicalvalues for errors are indicated. Note the relative largeerrors for longer time delays, owing to the experimen-tal design to measure larger Stark widths and shifts.The large absolute errors for short time delays arecaused by difficulties in extracting accurate widths atearly time delays. The electron number densities, in-ferred by comparison with hydrogen Balmer serieslines, are indicated. For the first three time delays of25, 30, and 50 ns, only the FWHM of H� was used toinfer Ne.
Figure 5 illustrates selected time-resolved spectraof the H� Balmer series line for delay times of 25, 50,100, and 200 ns, corresponding to the selected delaysfor the aluminum spectra. The experimental spectralresolution amounted to 0.27 nm for the Acton spec-trometer optical multichannel analyzer (OMA) con-figuration using a 1200 groove�mm grating.
Table 3 shows the H� Balmer series Stark widthsand Stark shifts for selected time delays from opti-cal breakdown. Also indicated are the number den-sities that are inferred from width and redshift,using the Parigger et al. [10] Tables 3 and 4 for H�.The number densities for H� are extracted by use ofresults from applying Oks’s convergent theory. Theindicated error ranges for early time delays reflect
the difficulties in extracting widths of atomic linesembedded in large free-electron background radia-tion.
Figure 6 illustrates selected time-resolved H� spec-tra for time delays of 75, 100, 150, and 200 ns. Thespectral resolution for the H� experiments amountedto 0.29 nm for the Acton spectrometer�OMA config-uration using a 1200 groove�mm grating. Obviously,H� FWHM determination appears complex, consider-ing the significant free-electron background radia-tion. Yet the red�blue separation of peaks may beused as an additional diagnostic for the determina-tion of high electron number density using H�.
Table 4 shows the H� Balmer series Stark widthsand Stark shifts for selected time delays from opticalbreakdown. Also indicated are the number densities
Fig. 5. (Color online) Recorded Stark widths and shifts for hydrogen-alpha Balmer series line for the indicated time delays and in 100Torr hydrogen.
Table 3. Linewidth (FWHM) and Redshift for � � 656.28 nm H� BalmerSeries Line
that are inferred from Stark widths, using the Pari-gger et al. [10] Table 2 for H�. Table 4 also shows theextracted separation of the two H� peaks. Previousexperimental studies by Parigger et al. [10] showedthat this separation amounts to approximately, 1�4of the FWHM Stark width early in the plasma decay;see Parigger et al. [10] Table 1 for H� for this corre-lation with the Stark width. The red and blue peakseparation is �red � �blue � a��1�2, with a value for theconstant a of approximately a � 0.25.
For even earlier time delays, the determination ofStark width and shift becomes increasingly diffi-cult, owing to the presence of large free-electronbackground and the occurrence of the ions’ spectro-scopic signatures. Previous experimental studiesreported by Parigger et al. [10] utilized gas samplesof hydrogen, leading to measurements of numberdensities of almost 1019 cm�3 a few nanoseconds afterthe generation of optical breakdown. The breakdownof solid aluminum causes the generation of typically
1022 cm�3 free-electron density near the surface [4].Measurements of densities of the order of 1019 cm�3
are expected away from the target [4] and for timedelays up to several tens of nanoseconds.
3. Conclusions
The exploratory studies presented here show thatatomic and molecular emission spectroscopy can beused for diagnostic purposes of aluminum laser ma-terials processing. The recorded Stark-broadenedand shifted emissions in air indicate electron numberdensities of approximately Ne � 1017�cm3 at a timedelay of 1 �s. This value for the electron numberdensity can be inferred from the increase in differ-ence for Stark width and shift when comparing 1.5 �swith 1 �s records (see Fig. 2) and using Table 1 ofFleurier et al. [3] and by comparing the listed electronnumber densities for hydrogen for similar time de-lays [10]. The measurement with 100 Torr hydrogenbuffer gas serves as a calibration for the Stark-broadened Al lines to infer electron temperature andelectron number density several tens to hundreds ofnanoseconds after breakdown. Molecular AlO emis-sion is not seen in the atomic-emission calibrationmeasurement. Molecular-emission diagnostics can beapplied to infer spectroscopic temperature severaltens of microseconds after optical breakdown. Therecorded molecular AlO emissions suggest a temper-ature of above 4000 K of the AlO molecule at a timedelay of 50 �s from optical breakdown. For these timedelays, there is no indication of the H� atomic line in
Fig. 6. (Color online) Recorded Stark widths and shifts for hydrogen-beta Balmer series line for the indicated time delays and in 100 Torrhydrogen.
Table 4. Linewidth (FWHM) and Peak Separation for � � 486.14 nm H�
the recorded AlO spectra. The presence of H� emis-sion in air from aluminum impurities has not beeninvestigated in this study.
The detailed measurements of aluminum Starkprofiles compare nicely with the aluminum Stark pro-files measured previously by Fleurier et al. [3]. Awelding-type arc at atmospheric pressure was usedcombined with 95% argon, 5% hydrogen, and AlCl3vapor flow. Using 15 kW electrical power, the Al I394.4 and 396.15 nm lines showed a width of0.054 nm and a shift of 0.031 nm. This resulted in anelectron number density of Ne � 1.28 � 1017�cm3 anda temperature of T � 1.32 � 104 K. Our results for100 Torr hydrogen background pressure are consis-tent with these values. A comparison with experi-ments using approximately 3 times larger energy perpulse, approximately 2�3 shorter pulse widths, andhigher (125 Torr) hydrogen pressure [11], shows thatthe values for Ne inferred from the aluminum Starkwidths should be higher by about a factor of 3 early inthe plasma decay. Here, measurements of H� and H�
show lower electron number densities. We attributethe factor of 3 discrepancy to the smaller size of theplasma. Increased ambiguities and errors result whenassociating hydrogen and aluminum atomic emis-sions recorded in separate experimental runs. Spa-tially resolving the plasma emissions is desirable tohelp decrease this factor of 3 discrepancy.
Laser-plasma studies by Renner et al. [4] using170 J, 0.4 ns pulses, and a peak irradiance of900 TW�cm2 showed Ne � 2 � 1022�cm3 and a tem-perature of T � 1 keV. A spatially resolved Lymanseries of aluminum was recorded. Electron numberdensities above Ne � 0.5–1.0 � 1018�cm3 are difficultto measure using only H� linewidths. These difficul-ties, as Fig. 6 shows for early time delay, are primar-ily attributable to (i) relatively large Stark widths oflarger than 10 nm and (ii) the relatively large free-electron background radiation resulting in a ratio ofthe order of 0.2 for signal (H� emission) to noise (free-electron radiation). The peak separation for H�,however, may be used to infer N in the 0.5–1.0� 1018�cm3. The peak separation amounts to ap-
proximately 0.25 of the FWHM [10]. Similarly, thefree-electron background radiation poses limits indetermining electron number density by use of theneutral aluminum Al I 394.4 and Al I 396.15 nmlines.
This work was supported in part by the Universityof Tennessee Space Institute (UTSI) and the UTSICenter for Laser Applications.
References1. D. A. Cremers and L. J. Radziemski, eds., Handbook of Laser-
Induced Breakdown Spectroscopy (Wiley, 2006).2. C. G. Parigger, “Laser-induced breakdown in gases: experiments
and simulation,” in Laser-Induced Breakdown Spectroscopy, A.Miziolek, V. Palleschi, and I. Schechter, eds. (Cambridge U.Press, 2006), pp. 171–193 and references therein.
3. C. Fleurier, S. Sahal-Bréchot, and J. Chapelle, “Stark profilesof Al I and Al II lines,” J. Phys. B 10, 3435–3441 (1977).
4. O. Renner, D. Salzmann, P. Sondhauss, A. Djaoui, E. Krouskýand E. Förster, “Experimental evidence of plasma shifts in Ly-man series of aluminum,” J. Phys. B 31, 1379–1390 (1998).
5. I. G. Dors, C. Parigger, and J. W. L. Lewis, “Spectroscopictemperature determination of aluminum monoxide in laserablation with 266 nm radiation,” Opt. Lett. 23, 1778–1780(1998), and references therein.
6. C. Parigger, J. O. Hornkohl, A. M. Keszler, and L. Nemes,“Measurement and analysis of atomic and diatomic carbonspectra from laser ablation of graphite,” Appl. Opt. 42, 6192–6198 (2003), and references therein.
7. J. O. Hornkohl and C. G. Parigger, Boltzmann EquilibriumSpectrum Program (BESP) (The University of TennesseeSpace Institute, 2002), http://view.utsi.edu/besp.
8. J. O. Hornkohl, C. G. Parigger, and L. Nemes, “Diatomic Hönl–London factor computer program,” Appl. Opt. 44, 3686–3695(2005).
9. J. J. Chang and B. E. Warner, “Laser-plasma interaction dur-ing visible-laser ablation of metals,” Appl. Phys. Lett 69, 473–475 (1996).
10. C. G. Parigger, D. H. Plemmons, and E. Oks, “Balmer series H�
measurments in a laser-induced hydrogen plasma,” Appl. Opt.42, 5992–6000 (2003).
11. C. G. Parigger, “Measurements of laser-induced plasma andoptical breakdown spectra of aluminum,” in Proceedings of the18th International Conference on Spectral Line Shapes, E. Oksand M. Pindzola, eds. (AIP Conference Proceedings 874, 2006)pp. 101–111.
