Atoms and Nuclei

Zeitschrift for Physik A

Z. Phys. A - Atomsand Nuclei 304, 197-199 (1982)

9 Springer-Verlag 1982

Laser Excitation of an Atomic Beam of Tungsten for Lifetime Measurements in the Configuration ( 5 d + 6 s ) 5 6p M. Kwiatkowski, G. Micali, K. Werner, M. Schmidt, and P. Zimmermann Institut fiir Strahlungs- und Kernphysik der Technischen Universit~it Berlin, Germany Received October 5, 198i The controlled evaporation of a heated tungsten wire was used to produce an atomic beam of tungsten. The atoms were excited selectively by the radiation of a pulsed dye laser. Time-resolved observation of the reemitted fluorescence with single-photon-counting technique yielded the lifetimes of 13 levels in the configuration (5d + 6s) s 6p.

Impurities in hot plasmas can influence the radiative transfer and the plasma diagnosis. Therefore, accurate oscillator strengths are needed for those elements which are commonly used in plasmaphysics. Tungsten with its very high melting point at 3,653 K is very important as a material for high temperature applications. There exists, however, only a small number of spectroscopic data on WI. One reason may be seen in the complexity of the spectrum which - as a common feature of the elements in the platinum group - is caused by the large number of levels in configurations of unfilled 5d shells. In the NBS-compilation of atomic energy levels [i], for example, only 20 .~o of the 355 listed W I levels are tentatively assigned by terms and configurations. In later experimental investigations by Laun and Corliss [2] and Corliss [3] 101 new levels were published without assignments by terms or configurations. Shadmi and Caspi [4] have calculated the levels of the configuration (5d+6s) 6 and have fitted 57 observed levels with a mean error of 100cm -1. They have shown that due to strong spin-orbit and interconfiguration interaction the assignment by LSvalues and configurations is meaningless especially for the higher levels. The determination of absolute oscillator strengths is based on the accurate knowledge of the physical conditions of the emitting or absorbing vapour. Therefore, in many cases only relative values are measured [5, 63. Lifetime measurements can then be used to establish an absolute scale for the oscillator strengths. This was done recently by Obbarius and Kock [10]. They determined a relative set of oscit-

lator strengths in W I by a combination of hook and emission measurements which was converted to an absolute scale by our values for the lifetimes of the levels z 7D3 and z VP2. With the experimental values of atomic lifetimes one can also test the quality of theoretical calculations, as the transition probabilities are sensitive to changes in the coupling properties within a given configuration or to admixtures caused by interconfiguration interaction.

Experiments Apart from the complexity of the W I spectrum, the main difficulty in the experiments with tungsten is the production of free atoms of this highly refractory metal. One possibility is the dissociation of compounds like WF6 in a hot plasma. This method was applied for the determination of oscillator strengths in WI by Clawson and Miller E5] in a gas-driven shock tube, by Shukhtin et al. [6] in a pulsed discharge, and by Obbarius and Kock [10] in a wallstabilized arc. Another possibility is the sputtering technique. Lifetime measurements of refractory elements were reported by Ramanujam [7] who used heavy-ion-induced sputtering or by Hannaford and Lowe [83 who used cathodic sputtering in a low-pressure raregas discharge. These methods are based on relatively high local particle densities which can cause certain problems with density-dependent effects like radiative trapping, pressure broadening, Stark mixing, etc. We have therefore decided to use the direct

0340-2193/82/0304/0197/$01.00

I98

M. Kwiatkowski et al.: Lifetime Measurements in the Configuration ( f d + 6s) s 6p

thermal evaporation from the metal for the production of an atomic beam of tungsten. A disadvantage of this method is, however, that one has to deal with temperatures of about 3,600 K in order to obtain a sufficiently high vapour pressure. For these temperatures the blackbody radiation is on the order of 1 kW/cm 2 and the total emissivity of tungsten is e=0.35. Therefore, one is forced to use only a small area of the metal from which the atoms evaporize. This can be done if one concentrates an electron beam on the tip of a tungsten wire and produces a droplet of molten material. We found out, however, that it is easier to use a tungsten wire with a diameter of 0.5 mm which is directly heated by a current of about 25 A. The concentration on a small emitting area of about I mm 2 was achieved by slightly reducing the diameter of the wire at one point which was placed on the axis of the atomic beam. With this set-up we could maintain an atomic beam of tungsten for 2-3 h. The amount of evaporized material was about 3 mg/h. Consequently, the particle density at the place of the laser excitation in a distance of about 10cm from the oven can be estimated to be on the order of 107 atoms/cm 3. Figure 1 shows the investigated levels of the configuration (5d+ 6s) 5 6p together with the levels a 5Dj and a 7S 3 of the configuration (5d+ 6s) 6 from which the laser excitation took place. The LS-designation is taken from Ref. 1 although it is obvious from the considerable overlapping of the terms and from the strengths of transitions with A S = I or A L > I that there are large deviations from LS-coupling. One can also expect strong interconfiguration interaction for the odd levels as it was shown for the even levels by the calculation of Shadmi and Caspi [41. That is why we prefer the designation of the configuration in the form (5d+6s) x. Due to the high temperatures which are necessary for the evaporation of tungsten, all the levels a 5Dj and a 7S 3 of the configuration (5d+6s) 6 were sufficiently populated. The wavelengths of the transitions to the investigated levels of the configuration (5d+6s) 5 6p are found between 360 and 430nm and are easily accessible for a dye laser. The lifetimes were measured with the single-photon-counting technique after pulsed laser excitation. The laser system consisted of an amplified dye laser which was pumped by a nitrogen laser. The peak power of the nitrogen laser was about t MW, the pulse length 3 ns, and the repetition rate up to 250 Hz. The path of the laser radiation inside the vacum chamber was carefully focussed through light baffles in order to avoid stray light. A monochromator was used to select the corresponding transitions for the observation of the reemitted fluorescence. In most cases

~[103cm-13

40.

zfP 30.

zfD 3-2"

3

zfF 5--

zTP z7D

3

1

4

3

2

2--3

(fd,6s) 5 6p 20

10.

,, ,,,,,,

3

0

aTS

a5D 4

2 t

WI 3

o

(5d.6s)6

Fig. l. Part of the Grotrian diagram of W I showing the investigated levels

O(:3 C4n3

W I z7P2 =71.6 ns

0

1 O0

200

900

400

Fig. 2. Decay curve of the level z 7Pz (26,229 cm-t) of W I

the fluorescence could be observed at a wavelength which differed from the exciting laser light. In this way we could eliminate the residual stray light of the laser pulse. The time-resolved observation of the fluorescence was achieved by a time-to-amplitude converter (TAC) which was started by the laser pulse and stopped by the first observed photon. For the correction of a possible pile-up effect which is caused by the fact that the TAC only registers one photon after each laser pulse, the mean counting rate was simultaneously monitored and adjusted to about 0.1 photon per laser pulse. A decay curve for the level z 7P2 (26,229 cm -1) is shown in Fig. 2. The decay curves were fitted by single exponentials to deduce the lifetimes of the corresponding levels. The results are listed in Table 1. The error limits are

M. Kwiatkowski et al.: Lifetime Measurements in the Configuration (5d+6s) 5 6p Table 1. Radiative lifetimes of 13 levels in the configuration (5d +6s) s 6p of W I Level

Lifetimes (ns)

Energy (cm -1)

Designation [1]

This work

CB ~

26,189.20 28,797.24 27,662,52 29,139.11 31,432.91 33,370.03 26,229.77 27,488.11 28,198.90 29,393.40 30,586.64 29,195.84 29,912.85

z 7D 3 z 7D 4 z 5F2 z SF3 z 5F, z 5F5 z 7P2 z 7P3 z 5P1 z 5P2 z SP3 z 5D 2 z 5D 3

161(8) 182(7) 188(7) 254(20) 449(25) 420(35) 72(4) 85(4) 126(5) 73(4) 49(3) 289(25) 170(9)

153 150 137 136 210 166 78 65 70 51 42 206 107

Calculated from the experimental transition probabilities of Corliss and Bozman [9]

a

due to the statistical deviations of the different measurements and to the uncertainty in the time calibration of the TAC. For comparison we have also listed the lifetimes which can be calculated from the experimental transition probabilities of Corliss and Bozman [9]. With the exception of the longer lifetimes the agreement is satisfactory. Therefore, the corresponding gf-values given in [9] should be reliable for the stronger transitions with an uncertainty of 30-50~o. As mentioned above, Obbarius and Kock [10] recently have determined oscillator strengths of 43 WI lines. These values were put to

199

an absolute scale by our results for the lifetimes of the levels z 7D3 and z 7P2. They have also checked their f-values with the other lifetimes of Table 1 and have found agreement within an uncertainty of 15 ~o. This work was supported by the Deutsche gemeinschaft, Sonderforschungsbereich 161.

Forschungs-

References 1. Moore, C.E.: Atomic Energy Levels, NBS Circular No. 467, Vol. 3 (Washington, DC: US Government Printing Off., 1971) 2. Laun, D.D., Corliss, C.M.: J. Res. Natl. Bur. Stand. (U.S.) A72, 609 (1968) 3. Corliss, C.M.: J. Res. Natl. Bur. Stand. (U.S.) A73, 277 (1969) 4. Shadmi, Y., Caspi, E.: J. Res. Natl. Bur. Stand. (U.S.) A72, 757 (1968) 5. Clawson, J.E., Miller, M.H.: J. Opt. Soc. Am. 63, 1598 (1973) 6. Shukhtin, A.M., Plekhotkin, G.A., Mishakov, V.G.: Opt. Spectrosk. 45, 438 (1978) 7. Ramanujam, RS.: Phys. Rev. Lett. 39, 1192 (1977) 8. Hannaford, P., Lowe, R.M.: J. Phys. B: At. Mol. Phys. 14, L5 (1981) 9. Corliss, C.M., Bozman, W.R.: Experimental Transition Probabilities, NBS Monograph No. 53. Washington, DC: U.S. Government Printing Office 1962 i0. Obbarius, H.U., Kock, M.: J. Phys. B: At. Mol. Phys. (to be published) M. Kwiatkowski G. Micali K. Werner M. Schmidt P. Zimmermann Institut f'tir Strahlungs- und Kernphysik Technische Universit~it Berlin Rondellstrasse 5 D-1000 Berlin 37 Germany