Laser Gain Measurements in a Xenon-Krypton Discharge H. Brunet

Centre de Recherches de la Compagnie Généale d'EIectricité, Département Recherches Physiques de Base, Marcoussis (Seine-et-Oise), France. Received 15 February 1965.

It has been previously reported1 that observed laser oscil­lations may result from inversions enhanced by means of col­lisions of the second kind between krypton ls5 metastable atoms and xenon atoms; an experiment has been reported by Grosof and Targ1 on xenon 2p states where enhancement of s-p transi­tions was observed. (In what follows we shall use Racah no­tation. )

In a similar manner, it would seem to be possible to enhance inversions of 5d xenon states via collisions of the second kind with krypton 53 metastable atoms. In this laboratory we found enhancement of the 5d-6p transitions arising from this mech­anism which gives rise to strong laser oscillations at 3.6798 μ.

Transfer of energy by collisions of the second kind have a high probability between levels whose energies have a coincidence within room temperature, kT (about 200 cm -1). The prob­ability of such energy transfer is, generally, much greater when spin conservation (the Wigner rule) is obeyed than when it is violated, but no conclusion with respect to this rule may be drawn from our experiments.

Consider the 5d xenon states which lie closest in energy to the 5s krypton metastable level, the 5d[½]0

1, 5d[½]00, 5d[½]0

4 and 5d[3/2]0

2 levels. Enhancement of these levels is very probable, especially the 5d[½]0

1 level. The 5d~6p transitions thus can be expected to be stronger in Kr-Xe discharges than they are in pure xenon discharges since in the latter case, the 5d states are pop­ulated solely by direct electron impacts.

Two Brewster-angle type lasers in an oscillator-amplifier configuration were employed in our experiment.2 The oscillator tube, as previously described,3 was dc excited, had a length of 120 cm, and was enclosed by metallic mirrors of 150 cm sep­aration. A prism was inserted in the cavity and allowed to isolate a single line from the oscillator before sending it through the amplifier. The amplifier tube was dc excited and had a 75 cm discharge length. Both tubes employed calcium fluoride windows.

Attenuators consisting of thin glass slides were inserted be­tween the oscillator and the amplifier to avoid saturation effects in the amplifier. Our experiment was limited to detection of oscillations in the range between 3 μ and 6 μ. In Fig. 1, we have

Fig. 1. Gain/unit length at 5.57 μ vs input signal level.

1354 APPLIED OPTICS / Vol. 4, No. 10 / October 1965

Table I. Gain Measurementsa

a ΔE denotes energy difference between the xenon upper level and the krypton metastable level. The filling pressures are 0.02 torr for pure xenon, and for Xe-Kr mixture impurities of Xe (about 5 · 10 - 3 torr) in 0.05 torr of krypton.

plotted the intensity gain per unit length we measured vs input signal intensity for the strong laser oscillations at 5.5754 μ which was previously reported by Faust and co-workers.4

The beam emanating from the laser was chopped before it entered the amplifier tube. After amplification, detection was accomplished using an infrared sensitive device. Since the spontaneous emission arising in the amplifier tube was not chopped, it was not detected and thus did not affect our gain measurements.

Experimental data are tabulated in Table I for several lines for which the gain was measured.

Our results indicate that only the 5d [½]01 level of xenon is

significantly enhanced by collisions with krypton metastable atoms. This shows that direct energy transfer can be important in a Kr-Xe mixture when the energy difference is very small with regard to kT.

The large gain measured for the 3.6798 μ line allows us to call it a strong transition. This result has its obvious interpretation in an enhanced inversion population arising from collisions with krypton metastable atoms and by the relatively large value of the matrix element for the transition.5

The transition probability for the 3.6798 μ line is similar to the large transition probabilities encountered in the strongest 5d-6p transitions. It is in the ratio of 5:27 to the 5.5754 μ line and 1:4 to the 3.5080 μ line.5 We note that the 3.6798 μ transition obeys Δk = ΔJ = 0.

References 1. G. Grosof and R. Targ, Appl. Opt. 2, 299 (1963). 2. W. B. Bridges, Appl. Phys. Letters 3, 45 (1963). 3. H. Brunet and P. Laurès, Phys. Letters 12, 106 (1964). 4. W. L. Faust, R. A. MacFarlane, C. K. N. Patel, and C. G. B.

Garrett, Appl. Phys. Letters 1, 85 (1962). 5. H. Statz, C. L. Tang, and G. F. Koster, J. Appl. Phys. 34,

2625 (1963).