2. EXPERIMENTAL SETUP

Investigation of the luminescence spectra of nuclear-excited gas mixtures was conducted with radioisotope pumping. A cylinder of diameter 25 mm and length 70 mm, carrying on its surface 18 sources containing ²¹⁰Po, was installed in a stainless steel chamber. Under normal conditions, the maximum range in He of α particles of energy 5 MeV is 183 mm. Before installation of the sources, the chamber was heated and degassed in a vacuum of about 10⁻³ Pa. After installation in the chamber, the α sources were pumped without heating for two to three weeks until well-reproducible (up to 3–7% of intensity for different gases) luminescence spectra were obtained. The gas pressure was measured by a vacuum manometer and a VDG-1 vacuum meter. The purity of the spent gases was Ne 99.996%, He 99.99%, Ar 99.992%, and Kr 99.999%. Hydrogen and deuterium (enrichment by D₂ was 99%, nitrogen impurities about 0.1% and oxygen about 0.05%) were purified by passage through silica gel and active copper. The emission spectrum was analyzed by an SPM-2 monochromator with quartz prism and a FEU-106 photomultiplier operating in photon-counting mode. The activity of the α sources was 9.6 GBq, which corresponds to an average energy deposition under 2 atm of helium of W ≈ 3 × 10⁻⁵ Wcm⁻³ and a mean ionization rate according to gas volume of S ≈ 4 × 10¹² cm⁻³s⁻¹. The apparatus for study of luminescence spectra of gas mixtures in the active zone of a nuclear reactor is described in Khasenov (Reference Khasenov2004; Reference Khasenov2014).

3. NEON

Lasing on three neon lines in the visible region (Fig. 1) with wavelengths of 585.3 nm (3p'[1/2]₀-3s'[1/2]⁰₁ transition), 703.2 nm (3p[1/2]₁-3s[3/2]⁰₂), and 724.5 nm (3p[1/2]₁-3s[3/2]⁰₁) was achieved under pumping mixtures with neon by products of nuclear reactions (Voinov et al., Reference Voinov, Krivonosov, Mel'nikov, Pavlovskii and Sinyanskii1990; Hebner, Reference Hebner1993). Processes in the active media of lasers at 3p-3s transitions of NeI are considered to be well studied (Shon et al., Reference Shon, Rhodes, Verdeyen and Kushner1993a; Karelin & Yakovlenko, Reference Karelin and Yakovlenko1995): population of the upper laser level is due mainly to dissociative recombination of Ne₂⁺ and HeNe⁺ molecular ions. Under comparatively weak pumping, the HeNe⁺ ions also form Ne₂⁺ ions in the substitution reaction

$${\rm He}{{\rm Ne}^{\rm + }}\, + \, {\rm Ne}\, \to \, {\rm N}{{\rm e}_2^{\rm + }} \, + \, {\rm He}{\rm .}$$

Assuming that the dependence of the luminescence intensity of the 585 nm line is determined by competition between Ne₂⁺ charge exchange with the quenching additive and recombination of electrons with Ne₂⁺, we obtain the additive concentration P at which the intensity is halved. In this case, the rate of recombination of molecular ions Ne₂⁺ is equal to the rate of ions recharge on additive:

$${\rm \beta} [{\rm N}{{\rm e}_2^ + }] n\, = \, k[{\rm N}{{\rm e}_2^ + }] P\comma \; \,$$

where k is the coefficient of Ne₂⁺ charge exchange with the additive, β is the coefficient of Ne₂⁺ recombination with electrons and n is density of electrons. Taking the values of the coefficients of recombination of basic molecular ions in the plasma approximately equal to the β value for Ne₂⁺ ion, we obtain:

(1)$$\eqalign{& n \approx \sqrt {S/{\rm \beta} } \comma \; \cr & P \approx \sqrt {{\rm \beta} S} /k.}$$

Figure 2 shows the measured dependence of the luminescence intensity of the 585 nm line in a He-Ne mixture on the pressure of quenching additives under radioisotope pumping. Dependence of the intensity at λ = 703 nm on the H₂ or Kr pressure in the He-Ne-H₂(Kr) mixtures is similar to the dependence for 585 nm. For hydrogen, k = 1.3 × 10⁻¹⁰ cm³s⁻¹ (Mayhev, Reference Mayhev1992); then, at the mean ionization rate, P ≈ 8 × 10¹²cm⁻³ ≈ 3 × 10⁻² Pa. The measured values of P correspond to 370 Pa for Kr, 505 Pa for Ar, and 1070–1200 Pa for H₂ and D₂ in a mixture with 200 kPa of He and 6.7 kPa of Ne (see Fig. 2). For comparison, the P value measured for nitrogen was 465 Pa, and the k values for nitrogen in the literature are 8.2 × 10⁻¹⁰ cm³s⁻¹ (Mayhev, Reference Mayhev1992) and 8.6 × 10⁻¹⁰ cm³s⁻¹ (Collins & Lee, Reference Collins and Lee1980).

Fig. 1. (Color online) Scheme of population of Ne(3p) levels, cascade transitions (blue), which are populated 3p levels of neon are indicated. Lasing lines (red) and resonance lines of Ne are also shown. Wavelengths are in nm and µm.

Fig. 2. Dependence of luminescence intensity at λ = 585 nm on pressure of additives to the He(200 kPa)-Ne(6.7 kPa) mixture. I ₀ is the intensity in the He-Ne mixture without additives.

Population of the 3p′[1/2]₀ level of NeI on excitation by heavy particles seems to be absent during the dissociative recombination of molecular ions. The addition to the He(200 kPa)-Ne(6.7 kPa) mixture of up to 13.3 kPa of nitrogen with 2% O₂ impurity led to the same drop in intensity as for pure Ar and Kr (see Fig. 2), implying that the process of electron attachment to an electronegative impurity has no effect on population of the 3p′[1/2]₀ level of Ne. A similar result, also under radioisotope pumping, was obtained by Mavlyutov et al. (Reference Mavlyutov, Mis'kevich and Salamakha1993): the line intensity at 585 nm in a Ne(100 kPa)-O₂(0.3 kPa) mixture was only half that in Ne at a pressure of 100 kPa, and the intensities of the lines at 703 and 725 nm were one-quarter. In comparison, under nuclear pumping of mercury-containing mixtures, population of HgI levels occurs during dissociative recombination of Hg₂⁺ ions (Batyrbekov et al., Reference Batyrbekov, Batyrbekov, Dolgikh, Khasenov, Rudoi, Soroka and Tleuzhanov1988). Adding 13 Pa of O₂ to the ³He-Hg mixture leads to an approximately 500-fold weakening of triplet lines and the resonance line of Hg (Khasenov & Smirnova, Reference Khasenov and Smirnova2008), which is connected to electron attachment to O₂ molecules. Apparently, the process of He₂⁺ recharge on O₂, which competes with charge exchange on Hg atoms, is not sufficient in this case, since the O₂⁺ ions will also be recharged on the Hg atoms.

It was concluded by Poletaev et al. (Reference Poletaev, Dorofeev, D'yachenko, Kopay Gora, Mavlyutov, Mis'kevich and Salamakha1992) that population of the Ne levels occurs under direct excitation by nuclear particles and secondary delta electrons. In He-Ne mixtures, population also occurs through excitation transfer from He metastable:

$${\rm H}{{\rm e}^{\rm m}} + \, {\rm Ne}\, + \, {\rm He}\, \to \, {\rm Ne\lpar 3}p\rpar \, + \, 2{\rm He}. $$

These conclusions were based on a study of the spectral-temporal characteristics of pure Ne and He-Ne mixtures under pumping by heavy charged particles.

From our point of view, more-probable channels for 3p-level population in the mixture with He are cascade transitions from the 4s levels:

$$\eqalign{& {\rm H}{{\rm e}^{\rm m}}\, + \, {\rm Ne} \, \rightarrow \, {\rm Ne}(4s) \, + \, {\rm He\semicolon \; } \cr & {\rm Ne }(4{\rm s}) \rightarrow {\rm Ne}(3p) + h{\rm \nu} }.$$

It is known that the Ne(4s) levels are close to the He(2³S ₁) level, and the operation of the He-Ne laser at 1.15 µm is based on excitation transfer to Ne atoms from He(2³S ₁). The absence of 4s-3p transition lines in high-pressure mixtures (Poletaev et al., Reference Poletaev, Dorofeev, D'yachenko, Kopay Gora, Mavlyutov, Mis'kevich and Salamakha1992) is connected with the location of this transition in the IR region of the spectrum, beyond the sensitivity limit of photomultipliers. In the work of Abramov et al. (Reference Abramov, Gorbunov, Melnikov, Mukhamatullin, Pikulev, Sinitsyn, Sinyanskii and Tsvetkov2006), where measurements were made up to 1100 nm, in the excitation of Ne and He-Ne mixture by uranium fission fragments, there is a line at 966.5 nm corresponding to the 4s[3/2]₂-3p[1/2]₁ transition. Abramov et al. (Reference Abramov, Gorbunov, Melnikov, Mukhamatullin, Pikulev, Sinitsyn, Sinyanskii and Tsvetkov2006) also identified more than 10 lines of the 3d-3p transitions. In these transitions, it is mainly the lowest three 3p levels that are populated.

The intensity changes at 585 nm are connected not only with quenching of the 3p′[1/2]₀ level by additives. The quenching rate constant is 4.6 × 10⁻¹¹ cm³s⁻¹ for H₂ and 5.3 × 10⁻¹¹ cm³s⁻¹ for Ar (Burstein et al., Reference Burstein, Komarovskii, Fedorov and Yurgenson1991), and, considering the lifetime of the level (14.3 ns), we can obtain the value of the H₂ or Ar pressure (about 5 kPa) at which the quenching rate is comparable to the rate of spontaneous decay of the level. The decrease in luminescence intensity with increasing partial pressure of quenching additives appears to be connected mainly with the Penning process of He metastable on the additive.

Considering that the dependence of the intensity at 585 nm in He(200 kPa)-Ne(6.7 kPa)-H₂ (Ar, Kr, D₂, or N₂) is determined by competition between non-resonance excitation transfer from He(2³S ₁) to Ne and the Penning process of He(2³S ₁) on the quenching additive, we may estimate the value of the quenching additive pressure at which the intensity is halved (Table 1). A satisfactory fit of the estimated and measured P values leads to the conclusion that under ionizing pumping of He-Ne (with a substantial He content), 3p-level population occurs as a result of excitation transfer from He(2³S ₁) to Ne and at subsequent cascade 4s-3p transitions.

Table 1. Processes of He(2³S ₁) quenching. Estimates of P for a He-Ne(6.7 kPa)-M mixture (M = H₂, D₂, Ar, Kr, or N₂) with the rate constants K of the processes

Adding 6.7 kPa of Ar to Ne at a pressure of 100 kPa as well as at 300 kPa leads to a two-fold decrease in intensity at λ = 585 nm and a three-fold decrease at 703 nm, which correspond approximately to the 3p-level quenching by Ar. The presence of intense radiation in the 3p-3s transitions on Ne excitation by α particles (Batyrbekov et al., Reference Batyrbekov, Batyrbekov, Danilychev and Khasenov1990) and by spontaneous fission fragments of ²⁵²Cf (W ≈ 10⁻⁸ Wcm⁻³) (Poletaev et al., Reference Poletaev, Dorofeev, D'yachenko, Kopay Gora, Mavlyutov, Mis'kevich and Salamakha1992) leads to the conclusion that in pure Ne, 3p-level population occurs through processes not connected with dissociative recombination of Ne₂⁺ ions. At low degrees of ionization, Ne₂⁺ ions are preferably recharged by impurities in the gas. Under pumping by uranium fission fragments (Gorbunov et al., Reference Gorbunov, Grigor'ev, Dovbysh, Mel'nikov, Sinitsyn, Sinjanskii and Tsetkov2004), the luminescence intensity of Ne(64 kPa)-NF₃(200 Pa) is approximately equal to the emission intensity in the 3p-3s transitions in Ne with the addition of 1.06 kPa of Ar, Kr, or Xe. As with the He-Ne mixture, the presence of an electronegative impurity in Ne does not lead to a rapid drop in radiation intensity in the 3p-3s transitions.

It has been shown that a scheme in which Ne 3p-3s transitions are excited by dissociative recombination of Ne molecular ions with electrons is not consistent with experimental data on weak pumping by ionizing radiation: (1) The effective luminescence of 3p-3s transitions observed with admixtures at pressures of hundreds of Pascal's, at recombination mechanism of the levels populating the radiation intensity at W ~ 3 × 10⁻⁵ Wcm⁻³ would rapidly decrease at H₂ or N₂ pressures of about 0.01–0.1 Pa. (2) Processes of electron attachment to electronegative impurities (O₂ and NF₃) have no effect on population of the Ne 3p′[1/2]₀ level.

Excitation transfer to Ne atoms from metastable He atoms and direct excitation of Ne by nuclear particles and secondary electrons are assumed to be the most likely channels of Ne(3p) population (see Fig. 1). In Ne, the 3d, 4s, and 5s levels are excited by electrons, and population of Ne(3p) occurs in cascade transitions from these levels. The 3d, 4s, and 5s levels optically connected with the ground state are effectively excited by electron impact. The absence of lines representing transitions from the 5s levels (Batyrbekov et al., Reference Batyrbekov, Batyrbekov, Danilychev and Khasenov1990; Poletaev et al., Reference Poletaev, Dorofeev, D'yachenko, Kopay Gora, Mavlyutov, Mis'kevich and Salamakha1992) is explained by the fact that the probabilities of the 5s-3p transitions are far less than those of the 5s-4p transitions in the IR region of the spectrum. Cascade transitions 5s-4p, 4p-4s, and 4s-3p may also contribute to population of the 3p levels of NeI. In He-Ne mixtures, population of the 3p levels occurs also in the processes of non-resonance excitation transfer from He(2³S ₁) and He (2¹S ₀) metastable atoms to the 4s and 5s levels of Ne and in the subsequent cascade transitions.

Lasing of the He-Ne-Ar mixtures pumped by powerful beam of electrons with duration of about 5 ns was investigated by Shon et al. (Reference Shon, Rhodes, Verdeyen and Kushner1993a). Laser radiation at 585 nm occurred with a delay of 20–40 ns, and continued for about 200 ns after the pump pulse. According to Mel'nikov et al. (Reference Mel'nikov, Sizov and Sinyanskii2008), these results indicate the recombination mechanism of 3p levels population. In our opinion, the delay of laser pulse and its long duration should be associated with cascaded population of 3p levels of neon from higher levels, as well as transfer of energy from exited helium atoms to neon.

4. ARGON, KRYPTON AND XENON

The main mechanisms for population of the nd levels that have been discussed previously (Mel'nikov et al., Reference Mel'nikov, Sizov and Sinyanskii2008) are as follows (B = Xe, Ar, Kr, and A = buffer gas atom): (1) impact-radiation recombination: B⁺ + e + e(M) → B(nd) + e(M) (M = third particle); (2) recombination of heteronuclear ion molecules: AB⁺ + e → B(nd) + A (probably in cascade transitions from the upper p and s levels); (3) electron–ion recombination: B₂⁺ + e → B(nd) + е; (4) excitation transfer in inelastic collisions: A* + B → B(nd) + A; (5) step-by-step excitation of Xe(5d): Xe(6s, 6s′) + e → Xe(5d) + e.

The most common hypothesis is the second of these processes, namely level population through dissociative recombination of heteronuclear ionic molecules with electrons (Ohwa et al., Reference Ohwa, Moratz and Kushner1989; Shon et al., Reference Shon, Kushner, Hebner and Hays1993b; Shon & Kushner, Reference Shon and Kushner1994). Heteronuclear ionic molecules are effectively destroyed in the collision process, so their recombination at the relatively low density of electrons in nuclear-induced plasma cannot make a significant contribution to the formation of excited atoms (Mel'nikov et al., Reference Mel'nikov, Sizov and Sinyanskii2008). Apruzese et al.(Reference Apruzese, Giuliani, Wolford, Sethian, Petrov, Hinshelwood, Myers, Ponce, Hegeler and Petrova2006; Reference Apruzese, Giuliani, Wolford, Sethian, Petrov, Hinshelwood, Myers, Dasgupta, Hegeler and Petrova2008) and Mel'nikov et al. (Reference Mel'nikov, Sizov and Sinyanskii2008) considered electron-ion recombination of B₂⁺ ions (mechanism 3) to be the main channel for population of the nd levels. In most kinetic models, dissociative recombination of B₂⁺ molecular ions is considered to be the channel for losses to populate the lower laser р levels and (n + 1)s levels.

Recharging of molecular ions on admixtures competes with dissociative electron-ion recombination. Atoms of Kr have high rate constants for charge exchange on Ar₂⁺ ions: 7.5 × 10⁻¹⁰ (Collins & Lee, Reference Collins and Lee1979), 6 × 10⁻¹⁰ (Johnsen et al., Reference Johnsen, Macdonald and Biondi1978), and 5.3 × 10⁻¹⁰ cm³s⁻¹ (Shul et al., Reference Shul, Passarella, Upschulte, Keesee and Castleman1987). For Xe atoms, the rate constants for charge exchange on Ar₂⁺ range from 2.3 × 10⁻¹⁰ cm³s⁻¹ (Shul et al., Reference Shul, Passarella, Upschulte, Keesee and Castleman1987) to 12.5 × 10⁻¹⁰ cm³s⁻¹ (Collins & Lee, Reference Collins and Lee1979).

The rate constant for the charge-exchange process

$${\rm K}{{\rm r}}_2^ + + {\rm Xe } \to {\rm X}{{\rm e}^ + } + {\rm }2{\rm Kr}$$

is k ≈ 2 × 10⁻¹⁰ cm³s⁻¹ (Kebarle et al., Reference Kebarle, Haynes and Searles1967). From Eq. (1) at k = 2 × 10⁻¹⁰ cm³s⁻¹, β = 10⁻⁶ cm³s⁻¹ and a gas ionization rate S = 10¹⁹ cm⁻³s⁻¹ we obtain the impurity pressure at which the transition intensity in the case of the recombination mechanism for level population is halved: Р = 1.6 × 10¹⁶ cm⁻³ ≈ 60 Pa.

Abramov et al. (Reference Abramov, Gorbunov, Melnikov, Mukhamatullin, Pikulev, Sinitsyn, Sinyanskii and Tsvetkov2006) and (Gorbunov et al., Reference Gorbunov, Grigor'ev, Dovbysh, Mel'nikov, Sinitsyn, Sinjanskii and Tsetkov2004) measured luminescence spectra of inert gases and their mixtures under pumping by ²³⁵U fission fragments; the pumping power was 40 W cm⁻³, which corresponds to a gas ionization rate of 10¹⁹ cm⁻³s⁻¹. Table 2 shows the intensity values for the 4p-4s transitions of Ar atoms in Ar and in Ar mixtures with Xe and Kr based on data from Abramov et al. (Reference Abramov, Gorbunov, Melnikov, Mukhamatullin, Pikulev, Sinitsyn, Sinyanskii and Tsvetkov2006). As can be seen from the table, on the addition of 1 kPa of Kr, the line intensity decreases 1.5-fold, and on the addition of Xe, it is halved. With a lifetime of the 4p levels of Ar atoms equal to τ ≈ 30 ns, this corresponds to quenching of 4p levels of argon by Kr or Xe atoms with a quenching rate constant of about 10⁻¹⁰ cm³s⁻¹. In the case of a recombination mechanism for 4p-level population, Ar₂⁺ ion charge exchange on additives should lead to a 50- to 100-fold decrease in line intensity. Similar conclusions may be drawn in relation to the 5p-5s transitions of Kr. The intensities of the Kr lines are given in Table 3, which is also based on data from Abramov et al. (Reference Abramov, Gorbunov, Melnikov, Mukhamatullin, Pikulev, Sinitsyn, Sinyanskii and Tsvetkov2006).

Table 2. Emission power in the 4p-4s transitions of Ar atoms in pure Ar and in Ar-Kr and Ar-Xe mixtures under pumping by uranium fission fragments

Table 3. Emission power in the 5p-5s transitions of Kr atoms in Kr and Kr-Xe under pumping by uranium fission fragments

The total intensity of lines of the 4p-4s transitions of the Ar atom in the spectral range 696.5–842.5 nm decreases only 1.5-fold on addition of 200 Pa of NF₃ to 0.45 kPa of Ar (Gorbunov et al., Reference Gorbunov, Grigor'ev, Dovbysh, Mel'nikov, Sinitsyn, Sinjanskii and Tsetkov2004), confirming that recombination of molecular ions with electrons is not the basic mechanism for population of the 4p levels of Ar.

Thus, dissociative recombination of molecular ions of Ar and Kr are not the basic processes responsible for population of the lower p levels of lasers on d-p transitions of atoms. As the p levels are populated intensively as a result of d-p transitions, it appears that similar conclusions may be drawn in relation to the nd levels of Ar and Kr. The absence of any temporary delay of luminescence on excitation of Xe by a nanosecond electron beam proves the absence of a channel for population of the 5d[3/2]₁ level of Xe via formation of the Xe₂⁺ molecular ion (Denezhkin & D'yachenko, Reference Denezhkin and D'yachenko2009).

The results obtained are consistent with the data of Barrios et al. (Reference Barrios, Sheldon, Hardy and Peterson1992) and Ramos et al. (Reference Ramos, Schlamkovitz, Sheldon, Hardy and Peterson1995) concerning the predominant formation of Ar, Kr, and Ne atoms in the (n + 1)s or ground (np) states upon dissociative recombination of Ar₂⁺, Kr₂⁺, and Ne₂⁺ ions. Populating the nd levels of Ar, Kr, and Xe under pumping by hard ionizers apparently occurs mainly as a result of direct excitation by secondary electrons and excitation transfer from atoms of the buffer gases as well.

Different mechanisms of level population can be engaged depending on experimental conditions. When pumped by nanosecond electron beam Ar-Xe mixture lased pulse with delay of tens of nanoseconds, while the duration of lasing pulse at atmospheric pressure reached several microseconds (Peters et al., Reference Peters, Mei and Witterman1989; Scrobol et al., Reference Skrobol, Heindl, Krücken, Morozov, Steinhübl, Wieser and Ulrich2009). Investigation of luminescence at λ = 1.73 microns of He-Ar-Xe mixtures excited by an electron beam of 2 ns duration and 500 A current was reported in (Denezhkin & D'yachenko, Reference Denezhkin and D'yachenko2013). It was shown that there were two channels of 5d[3/2]₁ level population. Characteristic time constants of these channels depended on composition of the mixture and amounted to the order of a hundred nanoseconds for one channel and few microseconds for the second channel. These channels are well separated in He-Ar-Xe mixtures, which results in significant fraction (70–80%) of photons emitted during the first 300 ns. Ar-Xe mixtures exhibited a single population time constant of about one microsecond with delay of radiation pulse of tens of nanoseconds.

Since the fraction of xenon in optimal mixtures (He)-Ar-Xe is small, the main pumping occurs with the transfer of excitation from the argon to xenon atoms (Fig. 3). In our opinion, the long lasted lasing and luminescence emission continued after the pump pulse stems from the stepwise excitation processes:

$$\eqalignno{{\rm Xe}\left({6s\comma \; {\rm }6s^{\prime}} \right){\rm } + e &\to {\rm Xe}\left({5d} \right){\rm } + e\semicolon \;\cr {\rm Xe}\left({6s\comma \; {\rm }6s^{\prime}} \right){\rm } + e &\to {\rm Xe}\left({6{\rm p}} \right){\rm } + e\semicolon \;\cr{\rm Xe}\left({6p} \right){\rm } + e &\to {\rm Xe}\left({5d} \right){\rm } + e.}$$

The formation of xenon atoms in the (6s, 6s') states can be the result of several processes: direct excitation from the ground level during the pump pulse, excitation transfer from argon atoms, cascaded transitions from the upper levels, the recombination of Xe₂⁺ ions with electrons, and possibly other mechanisms. The microsecond fraction of the radiation is less pronounced in He-Ar-Xe then in Ar-Xe mixtures because electrons are faster thermalized in He then in Ar. In pure xenon this channel is negligible due to the high rate of formation of Xe₂* excimers generated from Xe (6s, 6s').

Fig. 3. (Color online) Scheme of some levels of xenon and argon. The arrows indicate the direction of the laser transitions (red) and 6p-6s transitions (blue), analogous to 4p-4s transitions in argon and 5p-5s transitions in krypton.

It is worth mentioning that in the reported computational models of the Ar-Xe laser (Klopovskii et al., Reference Klopovskii, Lukyanova, Rakhimov and Suetin1989) the major channels population of 5d[3/2]₁ level of xenon are stepwise excitation via 6s metastable level, and the transfer of excitation from the buffer gas.

5. MERCURY

It was shown by Batyrbekov et al. (Reference Batyrbekov, Batyrbekov, Dolgikh, Khasenov, Rudoi, Soroka and Tleuzhanov1987; 1989) that dissociative recombination of molecular ions is the basic process populating the 7³S ₁ level of the Hg atom under ionizing pumping, possibly in cascade transitions from the 7³P levels. The following results point to the recombination mechanism of population: (1) The intensity of the Hg triplet lines in He-Xe-Hg mixtures was practically independent of Hg pressure in the range 0.1–40 Pa in a radioisotope installation and 0.1–13 kPa in an in-core installation, which excludes the possibility of direct excitation of the 7³S₁ level from the ground state by electron impact. (2) In a He-Xe-Hg-D₂ mixture, on increasing the partial pressure of D₂ to 13 kPa, the intensity of the resonance line of Hg (253.7 nm) decreased more than 300-fold, whereas the intensity of triplet lines decreased only 2.5-fold. This indicates that population of the HgI levels cannot be explained by step-by-step excitation through the 6³P ₁ level. (3) On addition of an electronegative impurity (13 Pa of O₂) to a ³He-Hg mixture, the intensity of triplet lines decreased 500-fold (Khasenov & Smirnova, Reference Khasenov and Smirnova2008).

In the work of Bochkov et al. (Reference Bochkov, Kryzhanowskii, Magda, Mukhin, Murzin and Neznakhina1992) and Rhoades and Verdeyen (Reference Rhoades and Verdeyen1992), in which laser action was obtained on the 546.1 nm line of the Hg atom, recombination of Hg molecular ions with electrons was also considered as a basic mechanism for population of the 7³S ₁ level. The 7¹S ₀ level is also populated; notably, although the 407.7 nm line is comparatively weak under our conditions, the probability of the IR transition from this level (λ = 1014 nm) is approximately seven times higher than that for the 7¹S ₀-6³P ₁ transition (Benck et al., Reference Benck, Lawler and Dakin1989). From the lifetime of the 7¹S ₀ level, the probability of the 7¹S ₀-6³P ₁ transition and the intensity of the 407.7 nm line, it is possible to determine the ratio between the number of photons radiated from the 7¹S ₀ level and the number of photons radiated from the 7³S ₁ level (Table 4). Line intensities of mercury were measured at 22 GBq total activities of nuclides sources in the radioisotope installation and at thermal neutron flux of 10¹³ n/cm²s on the reactor facility. Due to the proximity of lines 407.7 nm and 404.7 nm and the known values of the transition probabilities of 7³S ₁ level it was not necessary to calibrate the spectral sensitivity of the apparatus. It appears that Xe (and, to a lesser extent Ar) quenches the 7¹S ₀ level, with level deactivation occurring through the process

$${\rm Hg}\left({{7^1}{S_0}} \right){\rm } + {\rm }2{\rm Xe } \to {\rm HgXe}\ast {\rm } + {\rm Xe}.$$

Table 4. Intensity of transitions (relative units) from 7³S ₁ and 7¹S ₀ levels of Hg atoms under excitation by products of ³He(n,p)³H reaction in the core of a nuclear reactor and by α particles from ²¹⁰Po

Estimation of the corresponding rate constant from the results shown in Table 4 gives a value of about 3 × 10⁻³¹ cm⁶s⁻¹. Khasenov and Smirnova (Reference Khasenov and Smirnova2008) showed the possibility of population of the 7³S ₁ level of the Hg atom by recombination of Hg atomic ions and negative oxygen ions:

$${\rm H}{{\rm g}^ + } + {\rm }{{\rm O}^-} + {\rm M } \to {\rm Hg}\ast {\rm } + {\rm O } + {\rm M}.$$