Spectrochemical Analysis of Trace Contaminants in Helium (Helium- Fluorine) Pulsed Discharge Plasmas* Irni Journal of Analytical Atomic Spectrometry I I ALEKSEI B. TRESHCHALOV ANDRE1 S. CHIZHIK AND ARNOLD A. VILL Institute of Physics Riia 142 EE2400 Tartu Estonia Time-resolved spontaneous emission and absorption spectra in the 115-860 nm spectral range were studied for high pressure He and He-F pulsed discharge plasmas the aim being the spectroscopic identification and on-line monitoring of transient and stable gaseous contaminants accumulated in an actual F laser gas mixture. Various radicals (CN CH C OH SiF and CF,) excited atoms (0 C Si N and H) molecules (CO 0 and N2) and molecular ions (N2+ and CO') were identified as contaminants. The accumulation and self- transformation kinetics due to plasmo-chemical reactions between these species were investigated.It is suggested that carbon oxygen and carbon monoxide species are desorbed as primary impurities from the metal electrodes during the sputtering of the electrodes from cathode (anode) hot spots. Secondary impurities such as transient radicals and more complex compounds are formed by plasmo-chemical reactions between primary and other trace impurities (H N and 0,) and fluorine. These reactions might be strongly catalytically activated on the surface of pure nickel particles sputtered from the electrodes. The sensitivity of the spontaneous emission methods was estimated to be better than 1 ppm for the detection of stable CO and N impurities in the He discharge plasma. Under static conditions N2 and 0 impurities accumulate at a rate of about 0.1 mbar d-' in He and He-F gas mixtures due to continuous outgassing of the laser chamber materials. Under F laser running conditions a marked increase in the rate of accumulation of 0 contaminants ( z 0.1 mbar per 5 x 105 shots) was observed.Oxygen is the main contaminant responsible for F laser output energy degradation because of the intracavity absorption at 157.6 nm in the oxygen dissociation absorption band. The strong absorption in the far VUV range (130-115 nm) observed in an aged He-F gas mixture was attributed to stable fluorine-containing compounds (tentatively identified as COF or HF molecules). In addition to the well known molecular fluorine emission band at 157 nm caused by the transition from the ionically bonded D' 3112g state two spontaneous emission bands (structural at 255 nm and continuous at 280 nm) were observed in the discharge of an He-F gas mixture.These new bands were attributed to transitions bet ween covalent 1 y bonded excited electronic states of the fluorine molecule. Keywords Fluorine laser; discharge; gas mixture degradation; vacuum ul traviole t-visi bl e emission-a bsorption spectroscopy ; plasmo-chemical reactions; radicals Molecular discharge lasers (excimers F2) are nowadays the most powerful directly pumped commercially available sources of coherent UV and VUV radiation. It is well known however that the critical component of these lasers is their gas mixture. Laser output power usually declines after the laser chamber has been filled with fresh gas due to highly reactive halogen *Presented at the 1996 Winter Conference on Plasma Spectrochemistry Fort Lauderdale FL USA January 8-1 3 1996.donor depletion and formation of contaminant impurities in the gas mixture caused by discharge-induced plasmo-chemical reactions.'-' Even trace amounts ( z 10 ppm) of some contami- nants manifest themselves in a loss of output energy and cause spatial spectral and temporal degradation of the laser beam due to development of discharge instabilities.8-'' Output par- ameters of VUV gas lasers are particularly sensitive to gaseous impurities which can lead to intracavity absorption of VUV radiation or pre-ionization photons." These contaminants are accumulated continuously and removed from the laser chamber only by refilling with fresh gas.The development of a new generation of 'sealed-off' high pressure gas lasers with a quantitative improvement in gas lifetime'3.14 prompted us to initiate a spectrochemical investigation of the contaminants formed within pulsed discharge-pumped lasers under real operating conditions. Sophisticated gas chromatographic mass spectrometric and Fourier-transform infrared spectrometric methods are usually used for the measurement of specific gaseous imp~rities.~-~ However all these techniques are invasive and they only allow the investigation of stable final plasmo-chemical products accumulated in gas mixtures. Nowadays the most sensitive and more importantly non-intrusive discharge plasma diag- nostic methods are based on spontaneous emission and laser- induced fluorescence techniques characterized by high tem- poral and spatial resolution and species selectivity.It is well known that in a non-stationary He discharge plasma high energy electrons which most efficiently excite and ionize molecular additives are formed mainly at the initial breakdown stage of the discharge. Another mechanism of efficient exci- tation of impurities is connected with fast energy exchange chemical reactions between minor additives and excited helium atoms He* excimer He2* molecules and molecular He2 + ion^.'^*'^ For the formation and investigation of short-lived excited species a high pressure fast pulsed discharge is neces- sary because the formation of transients has to be faster than the possible quenching and radiative decay processes. This paper presents spontaneous emission and absorption spectral data in the range 115-860 nm for high pressure He (He-F,) glow discharge plasmas under fast pulsed excita- tion.The aim was the spectroscopic identification and on-line monitoring of transient and stable gaseous contaminants accumulated in a real VUV fluorine laser gas mixture. EXPERIMENTAL The experimental arrangement for measurement of time-resolved spontaneous emission and absorption spectra in a discharge plasma is shown schematically in Fig. 1. The object of investi- gation is the discharge plasma in a commerci- ally available miniature excimer laser (PSX-100) (MPB Technologies Dorval Canada). A traditional thyratron- switched charge-transfer circuit with a storage capacitor of 12 nF and a charging voltage of 8 kV is used in this laser to generate the pumping pulse for the transverse electrical dis- charge.The length of the active medium is 15 cm; the solid Journal of Analytical Atomic Spectrometry September 1996 Vol. 1 I (649-659) 649Fig. 1 Schematic diagram of the experimental apparatus nickel electrode spacing is 0.3 cm and the discharge width 0.2 cm. Automatic homogeneous UV pre-ionization is performed by a surface discharge on two BaTiO ceramic plates that are placed laterally on both sides of the cathode electrode. Pre- ionization discharge is initiated on the growing front of the main discharge voltage. These ceramic plates also serve as extremely low-inductive peaking capacitors with a total capaci- tance of 12nF for the rapid pulsed homogeneous discharge excitation of the high pressure (up to 10 bar of He) gas mixtures. It was not possible to make reliable and direct measurements of discharge voltage and current temporal behaviour for the very compact design of the pumping circuit used here.From electrotechnical considerations the current rise time can be estimated to be 10ns and the discharge peak current to be 10 kA which gives a maximum discharge current density of ~2 kA an- and a pumping power density of ~ 2 0 MW cm-,. This laser operated with a gas mixture of 6 bar of He and 5 mbar of F gives a pulse energy of M 1 mJ at 157.6 nm (D' 3112g +A' ,112 transition in molecular fluorinelg) with a rep- etition rate up to 100 Hz and a gas lifetime of lo6 pulses.The laser chamber with a volume of ~ 1 5 0 0 cm3 is made from aluminium; Viton O-rings are used for sealing the windows. High voltage insulators mounted inside the laser chamber are made from PTFE. The optical cavity is composed of two uncoated MgF plates. The spontaneous VUV (1 15-220 nm) emission of the dis- charge plasma was detected along the laser axes with a 0.5 m VUV grating monochromator (M-12) (VEMO Tartu Estonia) with a 1/12 relative aperture (resolution of the monochromator 0.05 nm) and an FEU-142 solar blind PMT (time resolution about 20ns). The spontaneous UV-visible (UV-VIS) (220-860 nm) emis- sion was collected with a quartz lens through a diaphragm with a 1/50 relative aperture and focused onto the entrance slit of a 0.5 m grating monochromator (MDR-23) (resolution 0.03 nm) with an FEU-79 PMT (time resolution 10 ns).The signal from the PMT was detected by a BCI-280 box- car integrator interfaced with a computer for data storing and processing. Two different registration regimes were used (1) time-integrated (with an integrator gate of 5 ps) for rec- ording the spectra and (2) time-resolved (with an integrator gate of 10 ns) for detection of the transient kinetics of excited species in the discharge. Relative intensities of measured emis- sion spectra are reported in this paper without corrections for the absolute spectral response of the monochromator and PMT combination. A laser spark was used as a continuum pulsed light source for measurement of low-resolution absorption spectra of stable plasmo-chemical products in the UV-VUV range.20 The light from the laser spark is transmitted through the laser resonator length and compared with the transmission of the evacuated laser chamber.A light pulse with an energy of 2 mJ from a tunable dye-laser (VL-lo) pumped by an excimer XeCl laser was focused inside the gas cell containing Ar at 10 bar and trace amounts of Xe as an impurity. The laser spark is initiated by multiphoton resonance breakdown of the dense gas. The dye-laser wavelength at 440nm is tuned to the four-photon resonance with the 4f states of Xe.21 The emission spectrum of a hot and dense spark plasma consists of an intense black- body continuum and several broad bands (243 220 190 167 160 nm) which presumably belong to excited ion and molecu- lar argon dimers.The VUV cut-off threshold of this source is limited to 112 nm by the transmission of the MgF output window of the gas cell. Particular care was taken to provide a reliable leak-free system for gas filling and pumping of the laser. Stainless-steel and copper tubes with Swagelok fittings were used for He F and other gases. In order to minimize possible impurities such as H,O 02 N2 CO CO and hydrocarbons in the He gas (initial purity 99.99%) additional purification during the gas filling stage was performed by passing the gas through a zeolite trap immersed in liquid nitrogen. Hydrogen which may be present as a contaminant in the He carrier gas at a concentration of up to lOppm is not removed by the cold zeolite trap and is always present in our gas mixture. Other commercially available gases which served as minor additives to our gas mixtures were used without further purification.The laser chamber and gas system were evacuated through a separate cooled zeolite trap to a vacuum of not better than lo-' mbar to avoid 'back-~treaming"~ of oil vapour from the rotary vacuum pump. Flushing of the laser chamber with pure He was usually carried out before refilling it with fresh gas. RESULTS AND DISCUSSION Fresh and Aged Helium Gas Mixtures Spontaneous emission spectra The measurement of the reference spontaneous emission spec- trum of pure He is the first step towards the identification of impurities produced in the discharge plasma. Helium was purified by passing it through a cold zeolite trap and emission spectra were measured immediately after the laser chamber had been filled with the gas at a pressure of 6 bar.A portion of the time-integrated spectrum of pure He in the UV-VIS range is shown in Fig. 2. Most of the observed broad emission 650 Journal of Analytical Atomic Spectrometry September 1996 Vol. 1110 . . I ' " " ' " . ' . fresh gas filling He2 - v) .- 3 - He 6 bar He2 c 300 350 400 450 5 00 Wavelengthhm Fig. 2 Time-integrated spontaneous emission spectrum of the dis- charge in fresh specially purified He. Only molecular He,* bands and trace N2+ CN and HP impurity lines are observed bands belong to transitions between several excited He,* excimer states which are formed after the discharge pulse from the rapid chemical reactions between He+ He* He and electrons under high pressure conditions.All of the observed He2* bands were identified according to the assignments in refs. 23 and 24 and are shown in Table 1. Only trace impurity bands from nitrogen molecular ions N2+* B -+X ,Zg+ and the hydrogen atomic HP line are detected in the initial visible emission spectrum. Some new contaminant bands for example CN* B ,Z+ -+X 2 Z + appear and continuously increase in intensity during the scanning of the spectrum. It was found that after the first several thousand discharge shots significant transformation of the emission spectra (par- ticularly in the VUV range) was observed due to the rapid formation of contaminants in the discharge and accumulation of these contaminants in the aged gas mixture. The spontaneous time-integrated emission spectrum of the same pure He gas mixture after 1.5 x lo6 shots changes markedly due to the accumulation of several impurities in the gas mixture (see Fig.3). Lines originating from atomic or molecular He mostly disappear and very strong emission bands from contaminant radicals (C,* CN* CH* and OH*) molecules (O,*) and molecular ions (N,'*) are observed and can be identified. Positions and assignments of the most intense bands are given in Table 1. For comparison of the fluorescence intensity scales in Figs. 2 and 3 the HP line should be taken as the reference (the intensity of this line does not change during the running of the laser). In the red region of the fluorescence spectrum (see Fig. 4) of an aged He gas mixture at 6 bar after 1.5 x lo6 shots Swan bands of C,* radicals the Ha (656.3 nm) line from atomic hydrogen several lines from He* (587.6 nm is the most intense) several well known red lines from atomic fluorine F* (see Table l) and the 777.5 and 844.6 nm lines from atomic oxygen can be identified. The relative intensities of the oxygen lines should be much higher than is the case in Fig.4; this is due to the poor sensitivity of the PMT cathode and the poor mono- chromator grating efficiency in the infrared region of the spec- trum. Atomic fluorine lines appear in the spectrum of the aged He gas mixture because before these experiments the laser chamber was passivated over a long period by the discharge in the He-F gas mixture. The metal fluoride film which is produced on the inner surface of the laser chamber and on the discharge electrodes provides a continuous source of trace amounts of fluorine during the running of the laser with the He gas mixture.It must be emphasized that all of the vibrational series for both C2* and CN* emission bands are well developed and the hot bands are unusually intense (such as the 1-1 2-2 3-3 2-3 3-4 and 4-5 transitions for CN* and the 1-2 3-4 4-5 5-6 and 6-7 transitions for C2* bands). It is obvious that in He gas at a pressure of 6 bar these radicals do not thermal- ize rapidly after plasmo-chemical reactions which populate selectively highly excited vibronic levels. It is difficult to measure the spontaneous emission spectrum of the He discharge in the VUV region after filling the laser chamber with fresh gas because of the rapid rate of accumu- lation of carbon impurities during the measurement of the spectrum.Fig. 5 shows the time-integrated VUV emission spectrum of the discharge in aged He after 1.5 x lo6 shots. The strong contaminant emission lines (see Table 1) belong to atomic carbon CO* A 'I'I-+X 'Z+ molecular bands and CO+ B 'Z+ +X 'C+ molecular ion bands. It is interesting that fairly strong hot emission lines from (v= 1,2) vibronic levels are observed in the fluorescence spectra of CO* and CO+*. This shows that these electronically excited species are created in highly excited vibronic levels and that at a buffer gas pressure of 6 bar vibrational relaxation is not complete during the radiative lifetime of CO* (7% 10 ns3') and CO+* (7x53 ns3') species. Rotational relaxation is however virtually complete.From the measured rotational contour of the individual CO+ (B-+X) 0-0 emission line the rotational temperature of CO+* ions is estimated to be 380f40 K. Emission lines from atomic nitrogen N* (120.0 and 149.5 nm) oxygen O* (130Snrn) and hydrogen H* (121.5nm) are observed. The relative intensities of the N* and H* lines at wavelengths around or shorter than 120nm should be much higher than they appear in Fig. 5; this discrepancy might be due to the absorption of some of the VUV emission by the thick (8 mm) MgF laser window and to the poor sensitivity of the PMT in the VUV region (PMT cut-off threshold 112 nm). The emission from two weak molecular bands (a structural band at 255 nm and a broad continuous band at 280 nm) is very difficult to attribute to any impurity molecules or radicals. These bands are tentatively assigned to new transitions between excited states of molecular fluorine The nature of these electronic states will be discussed later in this paper.Accumulation kinetics of impurities in the discharge The accumulation and self-transformation kinetics of some identified contaminants in pure He as a function of the number of discharge shots are shown in Fig. 6(a) and (b). On-line monitoring of impurities is controlled by the time-integrated spontaneous emission intensity of their specific spectral lines and bands (see Table 1) C* (193.1 nm) CO* (171.2 nm) CO+* (219.0 nm) O* (130.5 nm) N* (120.0 nm) F2* (255.5 nm) CH* (431.4 nm) C2* (516.5 nm) CN* (387.1 nm) N2+* (388.4 nm) HP* (486.1 nm) and 02* (351.6 nm).The most rapid accumulation rate is observed for C* impurities which quickly reach a maximum after x 6 x lo4 shots and do not increase during further running of the laser. The accumulation kinetics for CO* CO+ CH* and C2* impurities are fairly similar to each other. Emission signals from atomic and molecular nitrogen decrease continuously during the running of the laser. Nitrogen impurities which cause the initial level of the signal are present in He as trace additives. No kinetics were observed for the emission intensity of 02* and H* impurities. Only a slow increase in F,* emission was observed during the running of the laser in a He gas mixture. The emission from CN* radicals increases rapidly during the first 5 x lo4 shots which correlates with the initial rapid increase in the emission from C* species and then decreases slowly which correlates with the decrease in nitrogen impurities during the running of the laser.It is important to note that the formation of impurities is observed only under the discharge conditions. If the running of Journal of Analytical Atomic Spectrometry September 1996 Vol. 1 1 651Table 1 of the most intense characteristic lines are underlined Species Wavelengt h/nm Assignment Species Wavelength/nm Assignment He* 25 388.8 3p3P+2s3S j ~ ~ ~ * 23.24 335.6 5pn3 n +a3 C + 587.6 3d3D + 2p3P0 367.6 4pn3ng+a3Zu+ 706.5 3s3S+2p3PO 399.7 m3Cu+ 4 b3n ~ * 2 5 - 120.0 3s4P+2p3 4s0 445.7 j3Cu+ +b3n 149.5 3s2P+2p3 'Do 464.8 e3n,+a3xuf o* 25 130.5 3s3so+2p4 3P 513.3 E'lI,+A'Z,,+ 777.5 3 p5P + 3 s5s0 573.3 f3A,+b3n 3p3P+3s3SO 639.8 d3 C,+ + b3 l7 844.6 Identified atomic (ionic) emission lines and molecular bands observed for the discharge in He and He-F gas mixtures.The wavelengths ~ * 2 5 121.5 2pZPO-+ lS2S ~ 0 * 2 4 . 2 7 154.4 0-0 A'n+X'EC+ 486.1 4dZD + 2p2P0 157.7 2-2 656.2 3d2D+2p2Po 159.7 0-1 c * 25 156.1 2p3 3D0+2p2 3P 165.3 0-2 247.8 3s'P0 +2p2 'S 174.4 0-4 165.7 3s3P0 +2p2 3P 171.2 0-3 193.1 3s'P0+2p2 'D 173.0 1-4 F* 25 623.9 3p4so +3s4P 179.2 1-5 634.8 3p4so +3s4P 181.1 2-6 685.6 3p4D0 +3s4P 187.8 2-7 3p4D0 +3s4P 189.7 3-8 703.7 3pZPO +3s2P " - 219.0 0-0 B2C+ +X2Z+ 712.8 3pZPO +3s2P 211.2 1-0 739.8 3p4PO +3s4P 213.7 2-1 748.2 3p4PO +3S4P 230.0 0-1 755.2 3p4PO +3s4P 232.5 1-2 22 1.7 3p3 3D0 +3p2 3P 397.3 ,-o+ * 24,27 690.2 757.3 3p4PO 43s4P 235.2 2-3 0-0 Q1B2C+ +A211j 4s3P0 +3p2 'D 395.3 0-0 Qz 251.4 4s3P0 +3p2 3P (z2* 2427 436.3 4-2 d3lIg+a3& 250.6 251.9 4s3P0+3p2 'D 437.1 3-1 252.4 4s3P0 +3p2 3P 466.8 6-5 252.8 4s3P0 +3p2 3P 467.8 5-4 253.2 5S'PO +3p2 'S 468.4 4-3 288.2 4s'Po+3p2 'D 464.7 3-2 217.4 471.5 2-1 225.5 473.7 1-0 226.6 509.7 2-2 228.9 5 12.9 1-1 230.0 516.5 0-0 231.7 301.2 3 10.2 338.1 341.5 3 52.4 361.9 - 385.5 386.2 387.1 388.3 358.4 358.6 359.0 41 5.2 415.8 416.8 418.1 419.7 421.6 CH* 24 431.4 OH * 24.27 287.5 294.5 - - 307.8 312.1 3 18.4 3-3 B2C+ +X2C+ (I2* 24,27 2-2 1-1 0-0 2- 1 1-0 5-6 4-5 3-4 2-3 1-2 $ ; i ~ * 24 0- 1 0-0 A ~ A +x2n 2-1 A2Z'+X211 3-2 0-0 1-1 2-2 3-2 (:F2* 28,29 lq2 + * 2427 543.3 6-7 544.7 5-6 547.0 4-5 550.1 3-4 554.0 2-3 558.5 563.5 - 337.0 351.6 367.3 384.1 284.1 289.4 295.0 300.9 306.9 313.5 320.0 433.4 436.8 440.0 443.0 - 446.2 449.6 358.2 39 1.4 427.8 1-2 0-1 0-14 B3C,- +X3C,- 0-15 0-16 0-17 000-030 A'B1 +XIAl 000-040 000-050 000-060 000-070 000-080 000-090 3-2 A2C+-+X2n 0-0 0-0 1-1 2-2 3-3 1-0 B2C,+ +X2C,+ 0-0 0-1 the laser is stopped and then restarted after several minutes or hours the signals from all the impurities formed remain at the same levels that were achieved immediately before the interrup- and excited by the subsequent discharge shots and give emission signals from specific transient fragmentation products.tion of the running of the laser. This shows that all highly reactive species such as atoms and radicals are transformed Calibrated additions of impurities during the time between the discharge shots to stable gaseous compounds.These accumulated contaminants are destroyed For interpretation of these complicated accumulation kinetics and to reveal the nature of the stable contaminants the 652 Journal of Analytical Atomic Spectrometry September 1996 Vol. 1 1300 350 400 450 500 Wavelengthhm Fig. 3 Spontaneous emission spectrum of aged He (after 1.5 x lo6 shots). Lines originating from atomic or molecular He have largely disappeared. Strong emission bands are observed for CN* C2* CH* CO' N OH* and 02* 550 600 650 700 750 Wavelengthlnm Fig.4 Spontaneous emission spectrum in the red spectral region of aged He (after 1.5 x lo6 shots) I50 200 250 300 Wavelengthhm Fig.5 Spontaneous VUV emission spectrum of aged He (after 1.5 x lo6 shots).The strong emission lines belonbg to C* O* and N* atoms CO* molecules and CO' molecular ions. Two new weak molecular bands (structural at 255 nm and broad continuous at 280 nm) belong to transitions from a covalently bonded fluorine molecular state populated in a neutral energy exchange reaction uiz. F* + F +F,* + F influence of calibrated additions of impurities (02 CO CO N2 and H,) on the emission spectrum of pure He was studied. Fig.7 shows the dependence of the emission intensity from some specific impurity bands as a function of calibrated additions of N to pure He at 6 bar. The emission from N2* molecules N2+* ions and CN* radicals increases linearly with added N up to a pressure of 1 mbar of N2 whereas the 0 150 3 00 450 Number of shots/l000 Fig.6 Accumulation and self-transformation kinetics of some ident- ified contaminants in the He discharge as a function of number of shots I He 6 bar '.OO 0.20 0.40 0.60 0.80 N2 additive/mbar Fig. 7 Dependence of the emission intensity of some specific contami- nants on the calibrated addition of N to pure He emission from C2* and CH* radicals and C* 0* H* CO* and CO+* species remains independent of N addition. Fig. 8(a) and (b) shows the influence of calibrated additions of CO on the emission intensity of some specific species in a discharge of pure He. The emission from CO* CO+* and C2* species increases linearly with CO addition. Some saturation of CO* and CO+* emission signals is observed at a partial pressure of CO higher than 0.5 mbar. The emission from C* 0* CN* and CH* species increases linearly and then levels off for C* and O* at a partial pressure of CO higher than 0.5mbar and for CH* and CN* at a partial pressure of CO higher than 1.3 mbar.The emission intensity for H* N* and N2+* species is independent of CO addition. The influence of CO addition is very similar to that of CO. The addition of O2 to pure He up to a pressure of 0.5 mbar does not have a significant influence on the emission intensity from any of the above-mentioned impurity species except for the O* and 02* emission lines which increase linearly with O2 addition. Journal of Analytical Atomic Spectrometry September 1996 Vol. 11 6532 - C C .- .- m 'Z w 1 10 1 I ! I I (a) He.6 bar 8 - 0.00 0.25 0.50 0.75 I .oo CO additive/mbar Fig.8 nants on the calibrated addition of CO to pure He Dependence of the emission intensity of some specific contami- The addition of H up to a pressure of 1 mbar produces a linear increase in the emission from CH* OH* and H* species and a marked decrease in the emission from C2* CN* CO* CO+* N2* N2+* C* O* and N* impurities. It is interesting that the spontaneous emission spectrum of pure He with 0.5 mbar of added CO is very similar to that obtained for aged He gas after prolonged (1.5 x lo6 shots) running of the laser (see Figs. 3-5). It can therefore be postulated that the CO molecule is one of the main stable contaminants that is accumulated in a He gas mixture after prolonged running of the laser. From these experiments the sensitivity of the spontaneous emission method can be estimated bearing in mind that the geometry conditions used here have not been optimized.A detection limit better than 1 ppm (which is equivalent to 0.006mbar of impurities in our gas mixture) is obtained for the direct monitoring of stable trace impurities such as N or CO molecules. As regards transient species with strong emis- sion bands for example CN* and C,* radicals the sensitivity of their detection by the spontaneous emission method might be even higher; however no attempts were made to measure the absolute density of these transients in the plasma discharge. Time-resolved emission of excited species In order to investigate the mechanisms by which specific contaminant species are produced the time-resolved emission following a short discharge excitation was monitored for several specific impurities accumulated in aged He after pro- longed (1.5 x lo6 shots) running of the laser [see Fig.9(a) and (b)]. For our gas mixture which contains many trace contami- nants and where several competitive chemical reactions take place simultaneously after the initiation of the discharge it is very difficult to reveal all the precursors responsible for the formation of specific species. We will therefore only discuss some of the more probable reactions. The most rapid spontaneous emission kinetics are observed for H* N2* and He* transients which indicates direct high- energy electron-impact excitation of these species at the initial stage of the discharge. The next short-lived components are 10 8 6 4 1 2 2. E3 h c m .- C 5 0 a x v Y l .l . l l . I I I 0 100 200 300 400 500 .- C . 8 - E w .- 6 - 4 - 2 - 0 - & 0 200 400 600 800 1000 1200 Time/ns Fig.9 Time dependence of the spontaneous emission from some identified contaminants in an aged He gas mixture CO+* and N2+* molecular ions which are formed by rapid charge-transfer reactions of He2+ ions (the lowest X 2Cu+ state of the He,' ion has an energy of 22.45 eV) with CO and N r n o l e ~ u l e s . ' ~ ~ ~ ~ The rate of decay of CO+* emission kinetics is attributed to the decay of the CO+* excited electronic state B ,Z+ (radiative lifetime z x 5 3 ns3'). The emission decay time of N2+* ions is significantly shorter than the radiative lifetime of the N2+* excited electronic state B 'XUf (z x 60 ns3') which shows that rapid plasmo-chemical reactions or collisional quenching are important for N2+* ions. An alternative exci- tation mechanism for the emission of CO+* and N,+* ions could be Penning ionization of CO and N2 molecules by 3S1 and 'So metastable states of He (the corresponding energies are 19.82 and 20.62 eV).The emission from 0," impurities has a rapid rise time while the decay time constant of about 120 ns is fairly long. It seems that 0 molecules are directly excited by electron impact in the discharge; however the collisional quenching of the excited electronic state B 3Zu- is much faster than its radiative lifetime (z= 5 x s3'). As regards the emission kinetics of N* O* and C* excited atomic components it should be noted that a certain detectable delay of 15-20 ns is observed for the first emission maximum in comparison with the very rapid emission maximum for H* atoms.It seems that N* O* and C* atoms are produced by direct electron-impact excitation and consequent fragmen- tation of stable O N and CO molecules. Fast electron recombination processes such as CO+ +e+C* + 0 (0" +C)3' and N,++e-'N*+N could also be one of the mechanisms for the production of N* O* and C* atoms. The spontaneous emission decay profiles for these atomic species exhibit a fairly pronounced tailing. A strong second maximum in the emission kinetics of C* atoms and a small second maximum in the emission kinetics of N* atoms are observed at 100-13011s. This stage is a very late phase of the pumping discharge when the plasma has mostly recombined. This maximum can be explained by a chemical reaction between stable CO and N 654 Journal of Analytical Atomic Spectrometry September 1996 Vol.1 1molecules and long-lived highly energetic neutral species for example He2* excimer molecules. These molecules are produced from He* atoms by the following reaction He* + 2He+He2* +He (rate constant 2.5 x cm6 s-' 32). The long-lived triplet He2* electronic states have energies between 18.26 and 20.74eV,24 which is sufficient for the dissociation of CO and N2 molecules and for the resonant electronic excitation of atomic carbon and nitrogen fragments. The spontaneous emission kinetics for CO* impurities [see Fig. 9(a)] show a fast rise time immediately after the discharge excitation pulse and a long decay tail with a small second maximum at about 130 ns.This maximum in the CO* kinetics correlates with a strong second maximum in the emission kinetics of C* atoms. The formation of CO* (A 'II) molecules in the initial stage of the discharge is caused by the direct electron-impact excitation of CO molecules accumulated in the gas mixture from the previous discharge shots. This mech- anism however does not explain the long decay tail in the CO* emission kinetics [see Fig. 9(a)]. The emission A 'II-+X'Z+ transition of CO* molecules has a fairly short radiative lifetime (z M 10 ns3'); therefore the long decay kinetics reflect the destruction of long-lived precursors of the plasmo- chemical reaction for the formation of CO* A 'II molecules. He,* or some other electronically excited neutral molecule with a stored energy higher than 8eV could be such a precursor.It is possible that CO* molecules could also be produced in any fast reactions with the participation of C* atoms for example C* + O2 +CO*(A 'II) + 0. The analogous reaction where C atoms and CO molecules are in the ground electronic state with a fairly high rate constant k = 5 x lo-'' cm3 s-' has been studied by Fairbai~m.~~ In our reaction with the partici- pation of excited C atoms the rate constant might be even higher. The emissions from CH* and OH* radicals [Fig. 9(a) and (b)] have fairly short rise times and exponential decay tails with time constants of 160 and 75 ns respectively. These decay times are much faster than the radiative lifetimes of the investigated transitions (530 ns for the A 2A state of CH* 30 and 700 ns for the A 'C+ state of OH* radicals3').Therefore collisional quenching is very strong for these species. It appears highly unlikely that these radicals are formed from fast plasmo-chemical reactions with the participation of primary discharge-excited atomic species (C* H* 0* . . .). The spontaneous emission kinetics of the C,* and CN* radicals [Fig. 9(b)] have a specific 'incubation' delay time and a very long non-exponential decay tail. Since the radiative lifetimes of the corresponding transitions are fairly short (65 ns for the CN* B2C+ state3' and 12011s for the C2* d 3 n state3') these kinetics do not reflect the radiative decay of excited electronic states but rather the destruction of some long-lived ( M 1 ps) precursors.C,* radicals in the excited d 317 state could be formed from the association of two free carbon atoms by the following reaction C + C + He+C,* + He.34935 According to the potential energy level d i a g ~ a m ~ ? ~ ~ C2* radicals in the d 3 n state could be produced from carbon atoms in either the ground state 3P or in the excited 'D state. CN* radicals in the excited B 2Z+ electronic state could be produced by the reaction of long-lived metastable N2* mol- ecules in the A 3&,+ state with carbon atoms or some carbon- containing contaminant molecule. The highly reactive A &+ state of N is the lowest excited state with an energy of 6.1 eV and a radiative lifetime of about 2 s.~' The influence of meta- stable N as a possible long-lived transient contaminant on the performance of an XeCl laser has been discussed by Gabzdyl et al." This could explain the linear dependence of CN* fluorescence on N addition (see Fig.7) and the connec- tion between the accumulation kinetics of CN* N2+* and C* transients [see Fig. 6(a) and (b)]. If it is assumed that the concentration of carbon-containing impurities is proportional to the emission intensity of the emitted C* species and that the concentration of metastable N is linearly related to the emission of N2+* impurities in the discharge the CN* radicals should be a product of these precursors which is in fact observed in Fig. 6(b). The rise and decay times of the spontaneous emission kinetics of the molecular F2* band at 255 nm depend strongly on the density of the fluorine impurities.For trace concen- trations of fluorine in an aged He gas mixture these kinetics are fairly slow [see Fig. 9(a)]. However for He with fluorine added at a pressure of 4.5 mbar they become very fast (see Fig. 13). A possible reaction for F2* formation will be discussed in the next section. The following explanation for the production in the dis- charge and accumulation in an aged He gas mixture of carbon- containing contaminants is proposed. Carbon 0 and CO impurities which are always present in metals are extensively desorbed as primary contaminants from the main nickel elec- trodes and pre-ionization aluminium plates during the sputter- ing of the metal from cathode (anode) hot spots created under the very high discharge current density. Electron (ion)-bom- bardment desorption of CO from the electrodes36 could also be occurring under our discharge conditions.Hence the purity of the material of the main and pre-ionization electrodes is a very important parameter for a long operational lifetime of the gas mixtures of discharge lasers. Secondary contaminants such as CN C2 CH and OH radicals are formed by plasmo- chemical reactions with other trace impurities (N2 0 and H2). These reactions might be catalytically activated on the surface of pure metal particles sputtered from the electrode^.^^ In an aged He gas mixture the kinetics of the accumulation of contaminants will tend to reach saturation [see Fig.6(a) and (b)]. We believe that after about 1 million shots i.e. for a highly contaminated gas mixture the laser chamber walls serve as both a source and a sink for highly reactive species in the discharge.Surface recombination and pseudo-polymer layer formation reactions on the chamber wall could be an effective channel for the loss of long-lived chemically active radicals such as CN CH C and OH.38 The pseudo-polymer layer itself could act as a source of gaseous contaminants; hence under conditions where there is a dynamic equilibrium between the loss and creation processes the accumulation kinetics will also be saturated. Thus in order to obtain reproducible results in kinetics the purity of the laser chamber walls and their history are very important. Contaminants in an Aged He-F Gas Mixture Spontaneous emission spectra A mixture of He at a pressure of 6 bar with F added at a pressure of 4.5 mbar as investigated in this work is a standard gas mixture usually used in discharge-pumped molecular fluor- ine lasers.After filling the laser chamber with a fresh gas mixture a strong VUV lasing at 157 nm causes a very high level of stray light in the VUV monochromator and it is difficult to measure any spontaneous VUV emission spectrum. The time-integrated spontaneous VUV emission spectrum from an aged He-F gas mixture after 2.5 x lo6 shots when the VUV lasing has disappeared is shown in Fig. 10. Together with the well known molecular fluorine transition D' 3112g -+A' 3112u (157 nm) two new molecular fluorine emission bands at 255 and 280 nm are also observed. The intensity of these two bands is much higher than in the VUV spectrum of aged He containing trace amounts of fluorine impurities (see Fig.5). From the spontaneous VUV emission spectrum in Fig. 10 it is possible to identify also O* and C* atomic lines and CO* molecular bands which are very similar to the impurity lines Journal of Analytical Atomic Spectrometry September 1996 Vol. 1 1 655c) .d C c .- lA Y; ._ f w F He / F (6 bar / 4.5 mbar) F,? after 2.5 lo6 shots C V 150 200 250 300 Wavelengthhm Fig. 10 Spontaneous VUV emission spectrum of an aged He-F gas mixture. Atomic 0* C* and Ni* and molecular CO* contaminant lines are observed. Together with the well known molecular fluorine D 311,g+A' 3112u band at 157 nm two new molecular fluorine bands at 255 and 280 nm are detected in an aged He gas mixture (see Fig. 5). The intensity of these lines remains fairly low until a strong F2* fluorescence band at 157 nm is observed.For a very old He-F gas mixture (after 5 x lo6 shots) when the fluorescence from the 157 nm band is totally absent (but where there are still sufficient F molecules in the aged gas mixture) the emission from C* and CO* contaminants becomes much more pronounced than in Fig. 10. However the emission from the 255 and 280nm bands does not change significantly which shows that the chemical reac- tions responsible for the population of these specific excited electronic states of the F molecule are different. In order to explain our experimental results a schematic potential energy diagram of some F electronic states relevant to the discussion in this paper is proposed in Fig.11. To the best of our knowledge the emission spectra of F2 molecules have not been extensively studied and the determination of the energies of the excited electronic states involved in the emission transitions has still to be ~ompleted.~'-~~ 0 1 2 3 4 5 6 RIA Fig. 11 Schematic potential energy diagram of some F electronic states relevant to the discussion in this paper. Two known absorption transitions at 100 and 290 nm are shown. The upper electronic state for the well known emission band at 157 nm belongs to the ionically bonded D 3112g term. The new molecular fluorine emission bands at 255 and 280 nm originate from the covalently bonded B state It is generally considered that the main process for the formation of the F upper electronic state D 3112g for an F2 laser is a neutral energy exchange reaction viz.F* t F +F2* + F.32943 The rate constant of this reaction is 5.1 x lo-'' cm3 s - ' ; ~ ~ hence the characteristic time constant for the formation of F2* molecules is about 12 ns for our gas mixture. F* denotes both doublet 2P and quartet 4P low-lying excited states of atomic fluorine between 12.7 and 13.0 eV.45 We propose that this reaction can only populate the excited covalent states of the F molecule and not the ionically bonded states (see Fig. 11). Ionic fluorine states in particular the D' 3112g state are populated through the very fast ion-ion recom- bination reaction F+ +F- +He+F2* + He where F+ should be in the 3P2 3P0 or 3P1 lowest electronic states with an energy of about 17.42eV. The rate constant for this reaction in He buffer gas at a pressure of 6 bar is 1.4 x cm3 s - ' .~ ~ Therefore for the efficient pumping of an F laser a significant fraction of the F2 molecules in the He-F gas mixture need to be ionized by the discharge. The main contribution of the ion-ion recombination channel to the pumping of a VUV F2 laser was recently confirmed by Takahashi et aZ.46 In an aged He-F gas mixture with the accumulation of easily ionizable contaminants the ionization of F2 and as a consequence the emission band at 157 nm are depressed strongly. However the pumping of the red atomic fluorine laser lines requires only excited (an energy of about 14.5-14.7 eV) and not ionized fluorine atoms. A reservoir of excited F* atoms with an energy of 12.7-13.0 eV2' is created after the discharge excitation and fast electron-collisional energy-exchange cascades.We believe that this reservoir serves as a precursor for the production of F molecules in the covalently bonded electronic state denoted tentatively as B'. The population of this excited electronic state is responsible for the 255 and 280 nm F emission bands. These bands are not as sensitive to the presence of contaminants in the He-F gas mixture as is the ionic band at 157 nm. The vibronic structure of the 255nm band with a 0-0 transition at 222nm seems to be composed of several over- lapping progressions. The measured sequential distance of 380_+20cm-1 in the short wavelength wing reflects the vib- ronic quantum of the lowest electronic state tentatively assigned as the A3111 state.The vibronic structure of the long-wavelength wing of the 255 nm band is more complicated because of the overlap with the strong carbon (247.8 nm) and several silicon (250.6 251.4 251.9 252.4 252.8 and 253.2 nm) emission lines. We suggest by analogy with the discussion in ref. 19 that more than one near-lying covalent B' elec- tronic state is involved in the 255 nm band transition due to spin-orbit splitting of the F and F* states. Since no vibronic structure is observed for the 280nm emission band the lower electronic state tentatively assigned as B 311u must be a repulsive state producing two ,P fluorine atoms. The fact that there are no observed absorption bands corresponding to transitions to the B and B' states from the X 'Zg+ ground state in indicates that these trans- itions are forbidden by the selection rules.Further evaluation of the molecular F emission spectrum with better resolution is necessary to confirm these suggestions. These investigations are currently in progress. The VUV emission spectrum of an aged He-F gas mixture (Fig. 10) also contains many Ni* atomic emission lines. These lines appear in the spectrum only in the presence of added F in the He gas mixture. This shows that in comparison with pure He the addition of F2 stimulates markedly the sputtering of nickel from the cathode (anode) hot spots created under a very high discharge current density. The most interesting feature in the VUV emission spectrum in Fig. 10 is a strong continuous absorption observed in the far VUV region revealed by the depression of the O* (130.5 nm) impurity emission line in the aged He-F gas mixture.It is 656 Journal of Analytical Atomic Spectrometry September 1996 Vol. 1 1well known that F gas is transparent in the UV-VUV region up to 105 nm (the beginning of absorption in the C 'Xu+ +-X 'Zg+ band4') except for a weak broad continuous A 'nu +-X ' Zg+ absorption of 290 nm1.24 We assign the absorp- tion at a wavelength of less than 140nm in the aged He-F gas mixture to stable fluorine-containing contaminants. The nature of these compounds will be discussed later in this section. The emission spectrum of an aged He-F gas mixture in the UV-VIS region (see Fig. 12) contains together with two fluor- ine emission bands at 255 and 280 nm several specific impurity emission bands.A long progression in the 665 cm-' v,' bending vibrational mode of the CF radical in the ground electronic state is observed for the A 'B +X 'A CF,* emission band." The most intense is the 000+050 vibronic transition at 295.0 nm in accordance with the assignment in ref. 28. Another specific impurity band is connected with the SiF* A 'Z+ +X 'I3 fluorescence band (the radiative lifetime of this transition is z 230 ns 30). According to assignments in ref. 24 the most intense is the 0-0 (440.0 nm) transition. No emission from the well known CF* A 'I=+ +X 2rl[47 fluorescence band with a maximum at 247.6 nm and the NiF* fluorescence band with a maximum at 506.5 nm4* was observed. The molecular bands from 02* CH* and Cz* impurity species are very similar to those observed in the emission spectrum of an aged He gas mixture (see Fig.3). As regards the excited atomic species several very strong Si* emission lines are evident in Fig. 12 and are listed in Table 1. We believe that these silicon contaminants in our laser arise from the sputtering of the aluminium pre-ionization electrodes in the discharge of the He-F gas mixture (silicon is the main impurity in aluminium). A number of Ni* lines the most intense line being at 352.4 nm; the C* line at 247.8 nm together with He* (587.5nm) H/? (486.1 nm) and Ha (656.2 nm) lines several strong red F* lines and the 0" (777.7 nm) line were observed in the emission spectrum of an aged He-F gas mixture. No emission lines from N* N2* Nz+* and CO+* species with an excitation energy higher than that of the added F2 were observed for the aged He-F gas mixture. Time-resolved emission of excited species The time dependence of the specific emission lines (bands) in an aged He-F gas mixture is shown in Fig.13. The atomic fluorine emission lines exhibit fast excitation-decay kinetics which virtually coincide with the kinetics of He*. The same results have been reported by Takahashi et and Peet and Tre~hchalov.~~ It seems that F* atoms are produced 10 3 8 h c) .- v) g .d r 3 6 v x - .- v) c) 5 4 z 2 'Z .- C .d W 0 250 300 350 400 450 500 Wavelengthhm Fig. 12 Spontaneous UV-VIS emission spectrum of an aged He-F gas mixture. The emission bands from C* 02* C2* CH* and HB* Contaminants are similar to those in aged He (see Fig. 3). Lines from Si* SiF* CF2* and F2* appear in the presence of F h v) r~ 1 I I I M h ."FF2 after 5 1 O6 shots He I F (6 bar / 4.5 mbar) 100 150 200 250 0 50 Timehs Fig.13 Time dependence of some specific emission bands in the discharge of an aged He-F gas mixture from very fast energy transfer reactions such as He*( He,*) + F + He-+F* + F + 2He. The rate constant for this reaction is about 2 x s-' (Ref. 43) for a He pressure of 6 bar. The emission from the 255 and 280 nm molecular fluorine bands exhibits fairly fast kinetics similar to the atomic fluorine band but delayed by 15 ns. The emission kinetics of Si* and Ni* atoms exhibit a fast rise time and a long decay time with a time constant of about loons which shows the long exci- tation of these species in the cathode hot spots.SiF* molecules and CF,* radicals exhibit typical kinetics which possibly reflect their production from plasmo-chemical reactions of carbon and silicon atoms with F,. Identification of stable contaminants from absorption spectra In order to identify stable contaminants which accumulate in an aged He-F gas mixture and deteriorate the output energy of an F laser the emission and VUV absorption spectra of the gas mixture were measured during the running of the laser. After filling the laser chamber with fresh gas (6 bar of pure He or an He-F mixture with 4.5mbar of F added) there was no detectable absorption in the spectral range 190-1 15 nm for the gas in a laser chamber with an optical absorption length of 25 cm. Under static conditions (without laser running) N and 0 molecules were detected as the main impurities that are slowly accumulated inside the laser chamber. The absolute concen- tration of N impurities was detected from the fluorescence of N2+* ions.Oxygen molecules were monitored from the strong 0 dissociation continuum absorption band with a maximum at 145 nm.50 The accumulation rate for both N and 0 gases is about 0.1 mbar d-l. The rate of gas ageing is almost identical for pure He and an He-F gas mixture under static conditions but it depends on the quality of the laser chamber passivation. It seems that continuous out-gassing of 0 and N from the laser chamber materials is the main mechanism for the formation of these impurities. Oxygen contaminants accumulate much faster with the He-F discharge running than under static conditions.Fig. 14 shows the absorption spectrum of an aged He-F gas mixture in the VUV spectral region. From the absorption of the oxygen dissociation continuum band the accumulation rate for 0 impurities was estimated to be 0.1 mbar per 5 x lo5 pulses. Running the laser with discharges in pure He does not lead to a considerable increase in the accumulation rate for 0 and N impurities compared with static conditions. It seems that an extensive fluoridation (passivation) process viz. Me0 + F*+MeF + O converts the metal oxide film inside the laser chamber into a stable metal fluorine film and removes 0 from the metal surface into the gas mixture. Highly reactive species such as fluorine atoms are necessary for the passivation process because it is necessary not only to remove oxides Journal of Analytical Atomic Spectrometry September 1996 Vol.I I 657I00 He I F (6 bar / 4.5 mbar) after 5 10’ shots 80 - h 5 - c 60 5 . $ 4 0 - F laser d - .& D 20 0 180 170 160 150 140 130 120 I I . . l l l l . l I . . l l l . _ Wavelengt h/nm Fig. 14 Absorption spectrum of an He-F gas mixture after 5 x lo5 shots. The broad absorption band at 145nm belongs to oxygen contaminants. The F laser line position (157.6 nm) is denoted by an arrow. The strong absorption at 130-120 nm is tentatively assigned to COF or HF contaminants hydrocarbons and water from the surface of the laser chamber but also to produce a corrosion-resistant metal fluorine film. We believe that the strong intracavity absorption at the F2 laser wavelength of 157.6 nm caused by the molecular 0 impurities is the main reason for the degradation of the F laser output.In our laser 0.1 mbar of 0 gives about 40% intracavity absorption at the lasing wavelength. The spectral dependence of the 0 absorption band explains the results obtained by Kakehata e t al.,l2 who investigated the absorp- tion of the active medium of an F discharge-pumped laser. In contrast to the relatively low absorption measured at 168.6nm,12 the energy output of the F laser seems to be limited by the 5-fold greater intracavity absorption. This observation is consistent with the results presented in Fig. 14 where the absorption of oxygen contaminants at 168.6 nm is about five times less than at 157.6 nm. In addition to the 0 absorption band a strong continuum absorption band appears in the range 135-120nm with an extremely large increase in absorption towards the far VUV region. This continuum band only appears when the laser is run with an He-F gas mixture.The identity of the contami- nant responsible for this band is not known. Presumably it could be a stable fluorine-containing compound. We believe that COF or HF molecules could be the contaminant. It is known that CO burns in an atmosphere of F uiz. CO + F + COF2. Hydrogen fluoride is produced by burning trace amounts of hydrocarbons or water in an atmosphere of fluorine. This very effective reaction needs fluorine atoms and is easily initiated by the discharge. Both COF and HF molecules have been identified from Fourier-transform infrared absorption spectra7 and observed in an aged KrF-excimer laser gas mixture (after 2.5 x lo6 shots) at concentrations of about 60ppm for COF and 1OOOppm for HF.The lowest excitation dissocation pathway of COF by analogy with the photolysis of phosgene (COC12),’1 could give after the dis- charge excitation CO CO* and F molecules and FCO radicals.’ The self-transformation reactions between F CO F FCO and COF are not yet fully understood and are cur- rently being investigated in connection with the stratospheric photochemistry of fluorocarbon radical^.'^ In discharge-pumped lasers with fluorine-containing gas mixtures the accumulated COF and HF contaminants have an adverse effect not only because they act as a means of F fuel removal but also because they strongly absorb pre- ionization photons in the far VUV region which seriously deteriorates the spatial homogeneity of the discharge.Indeed the discharge in an He-F mixture after prolonged running of the laser contains visually more pronounced sparks than in a fresh gas mixture. According to the data for the VUV absorp- tion of COF molecules a strong continuous absorption is observed near 131 nm.54 At shorter wavelengths the absorption increases up to the limit of observation of 121.5nm. HF molecules also have a strong X ‘Z+ -+A ‘II photoabsorption band with a maximum at 120nm.55 Both these absorption phenomena are very similar to those shown in Fig. 14. In order to determine whether COF and/or HF is the contam- inant additional experimental studies are necessary. These investigations are currently in progress.CONCLUSIONS This work represents the results of a comprehensive investi- gation of the production and accumulation of transient and stable gaseous contaminant impurities in high pressure He and He-F gas discharge plasmas. Various radicals (CN CH C OH SiF and CF,) excited atoms (0 C Si N and H) molecules (CO 0 and N,) and molecular ions (N,’ and CO + ) were identified as contaminants. The accumulation and self-transformation kinetics due to plasmo-chemical reactions between these species were investigated. Possible mechanisms for the production of the contaminant impurities are proposed. This work was supported in part by the Estonian Science Foundation under Grant No. 2276. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Tennant R.Laser Focus 17 65. Gower M. C. Kearsley A. J. and Webb C. E. J. Quantum Electron. 1980 16 231. Mandl A. E. and Hyman H. Appl. Phys. Lett. 1986 49 841. Karwacki E. J. Jr. and Hanton S . D. Laser Focus World 1992 28 81. Hwang H. H. James K. Hui R. and Kushner M. J. J. Appl. Phys. 1991 69 7419. Stevens R. E. Kung C. Y. Kittrell C. and Kinsey J. L. Rev. Sci. Instrum. 1994 65 2464. Jursich G. Rufin D. Von Drasek W. Reid J. and Znotins T. Laser Focus World 1989 25 56. Brannon J. H. J. Quantum Electron. 1982 18 1302. Kimura W. D. and Seamans J. F. J. Quantum Electron. 1988 24 2124. Gabzdyl J. Cleaver K. Stevenson A. OKey M. A. and Osborne M. R. J. Appl. Phys. 1994,75 1213. Shimauchi M. Miura T. and Takuma H. Jpn. J. Appl. Phys. 1994,33,4628.Kakehata M. Ueno Y. Tamura K. and Kannari F. J. Appl. Phys. 1994 75 1304. Endert H. Patzel R. Rebhan U. Powell M. and Basting D. Jpn. J. Appl. Phys. 1995 34,4050. Nikolaus B. Endert H. Rebhan U. Patzel R. Vob F. and Basting D. in Lambda Highlights ed. Brinkmann U. Lambda Physik 1994 44 1. Treshchalov A. B. in Proceedings of the International Symposium on High-Power Lasers and Laser Applications V Vienna Austria April 5-8 1994 314/SPIE 2206 pp. 314-322. Chizhik A. Vill A. and Treshchalov A. in Proceedings of I I International Conference on Laser Physics & Spectroscopy ICLPS’95 Grodno Belarus September 25-27 1995 pp. 96-97. Collins C. B. and Robertson W. W. J. Chem. Phys. 1964,40,701. Richardson W. C. and Setser D. W. J. Chem. Phys. 1973 58 1809. Diegelmann M.Hohla K. Rebentrost F. and Kompa K. L. J. Chem. Phys. 1982 76 1233. Hemici M. Mottin S. Bon M. Roncin J. Y. and Laporte P. J. Phys. III Pr. 1991 1 2061. Peet V. E. Phys. Rev. A 1995 51 3982. Heppell T. A. Vacuum 1987 37 593. Neeser S. Tietz R. Schulz M. and Langhoff H. Z . Phys. D 1994 31 61. Rosen B. Spectroscopic Data Relative to Diatomic Molecules Pergamon Press Oxford 1970. Striganov A. R. and Sventitskij N. S. Tables of Spectrum Lines Atomizdat Moscow 1966. Zaidel A. N. Prokof’ev V. K. and Raiskii S . M. Tables of Spectrum Lines Pergamon Press New York 1961. 658 Journal of Analytical Atomic Spectrometry September 15196 Vol. 1127 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Pearse R. W. B. and Gaydon A. G. The Identification of Molecular Spectra Chapman and Hall London 1965. Mathews C.W. Can. J . Phys. 1967 45 2355. Mann D. E. and Thrush B. A. J. Chem. Phys. 1960 33 1732. Huber K. P. and Herzberg G. Constants of Diatomic Molecules Van Nostrand Reinhold New York 1979. Wan B. N. and Langhoff H. J. Chem. Phys. 1992 97 8137. Kim Y.-P. Obara M. and Suzuki T. J. Appl. Phys. 1986 59 1815. Fairbairn A. R. Proc. R. SOC. London Ser. A 1969,312,207. Bokor J. Zavelovich J. and Rhodes C. K. J. Chem. Phys. 1980 72 965. Little C. E. and Browne P. G. Chem. Phys. Lett. 1987 134 560. Sandstrom D. R. Leck J. H. and Donaldson E. E. J. Appl. Phys. 1967 38 2851. Zhang J. Wu N. Wang J. Zhao J. and Xu Y. Opt. Laser Technol. 1994 26 355. O’Neill J. A. and Singh J. J. Appl. Phys. 1995 77 497. Porter T. L. J. Chem. Phys. 1968 48 2071. Colbourn E. A. Dagenais M. Douglas A. E. and Raymonda J. W. Can. J . Phys. 1976 54 1343. Levelt P. F. Eikema K. S. E. Stolte S. Hogervorst W. and Ubachs W. Chem. Phys. Lett. 1993 210 307. Bressler C. Lawrence W. Chergui M. and Schwentner N. J. Lumin. 1994 608~61 570. 43 44 45 46 47 48 49 50 51 52 53 54 55 Ohwa M. and Obara M. Appl. Phys. Lett. 1987 51 958. Huestis D. L. Hill R. M. Nakano H. H. and Lorents D. C. J . Chem. Phys. 1978,69 5133. Sumida S. Obara M. and Fujioka T. J. Appl. Phys. 1979 50 3884. Takahashi M. Maeda K. Kitamura T. Takasaki M. and Horiguchi S. Opt. Commun. 1995 116 269. Porter T. L. Mann D. E. and Acquista N. J. Mol. Spectrosc. 1965 16 228. Bouddou A. J. Mol. Spectrosc. 1994 168 477. Peet V. E. and Treshchalov A. B. Laser Phys. 1993 3 88. Okabe H. Photochemistry of Small Molecules Wiley New York 1978 p. 214. Okabe H. Laufer A. H. and Ball J. J. J . Chem. Phys. 1971 55 373. Maricq M. M. and Szente J . J. J. Chem. Phys. 1994 100 8673. Wallington T. J. Ellermann T. Nielsen 0. J. and Sehested J. J. Phys. Chem. 1994 98 2346. Workman G. L. and Duncan A. B. F. J. Chem. Phys. 1970 52 3204. Cacelli I. Chem. Phys. Lett. 1996 249 149. Paper 61003 95 H Received January 18 1996 Accepted May 13 1996 Journal of Analytical Atomic Spectrometry September 1996 Vol. 1 1 659