1 Introduction

There are only several powerful sources of VUV emission with quantum energy of more than 6 eV, providing high-power radiation at deep-UV wavelengths which are widely used in various photochemical processes and technologies such as photolithography, precision micromachining, and ablation of surfaces [1,2,3]. At present, the most powerful sources of the VUV radiation at wavelengths shorter 200 nm are discharge-pumped lasers based on H2, F2 and exciplex ArF* molecules. Therefore, study of the processes occurring in the active medium of these lasers, development of new excitation methods, and the improvement of lasing parameters are highly relevant from scientific and practical points of view.

As a rule, transverse volume discharges with preliminary ionization of a gas mixture are used for excitation of the VUV lasers. However, the features of the kinetics of H2 laser on B1Σ+u → X1Σ+g Lyman band [4] and short lifetime of the volume discharge in fluorine containing gas mixtures [5] make it necessary to use short (~ 10 ns) high-power high-voltage pulses for excitation of the VUV lasers [6,7,8].

It is known that quite uniform diffuse plasma can be produced without additional preionization in gaps formed by electrodes with a small radius of curvature [9,10,11,12]. When high-voltage pulse with short rise-time is applied to electrodes providing sharply non-uniform electric field in a gap, a diffuse discharge is formed in pressured gases due to breakdown initiation by run-away electrons and X-ray radiation. Therewith the diffuse plasma can effectively emit VUV spontaneous and stimulated radiation [11,12,13].

The aim of this work is to study VUV lasing on hydrogen, fluorine and ArF* molecules in a diffuse discharge between two extended electrodes with a small radius of curvature and to reveal the features of a diffuse discharge formation in the non-uniform electric field, as well.

2 Experimental equipment and measurement techniques

The installation drawing is shown in Fig. 1. A laser chamber with two stainless steel electrodes 30 cm long placed at a distance of 1.8 cm and connected to a RADAN-220 pulse generator described in detail in [14] was used in our experiments. The maximum energy E stored in the generator forming line with a capacity of C = 50 pF [15] was 2.1 J. This value is determined by the breakdown voltage U of a P-49 serial spark gap-sharpener, which in our experiments was measured to be 280 ± 10 kV.

Fig. 1
figure 1

Schematic diagram of the experimental setup. 1—gas mixture extraction and preparation system, 2—laser chamber with RADAN-220 generator, 3—VM-502 or MDR-12 monochromator, 4—EMI9781B photomultiplier, 5—photocell with Cu cathode, 6—digital camera, 7—FEK-29SPU photocell, 8–personal computer, 9—FEK-22SPU photocell, 10—TDS-3054 oscilloscope

Full size image

The electrodes were made in the form of blades with rounded edges, had a vertex angle of 5°, and a radius of curvature of the sharp edges of 0.05 mm. This provided increased electric field strength near the electrodes and generation of fast and run-away electrons, as well as X-ray radiation, which made it possible to form volume diffuse discharge in various high pressure gas mixtures [9, 13, 14].

There was a window on the side wall of the chamber closed by a CaF2 plate for photographing the discharge and recording the parameters of its UV–visible spontaneous emission pulses. The laser plane-parallel optical cavity was formed by a flat aluminum coated mirror and a plane-parallel MgF2 plate.

The generator produces voltage pulses with duration ~ 2 ns (FWHM) on the matched load. Due to long connection inductor L = 203 nHn the current pulse in the discharge gap has a half-period of ≈ 10 ns in the oscillating regime. Therewith impedance of the excitation system is relatively high and reaches ρ = 67 Ω. However, in spite of the large L voltage pulse with amplitude of about 300 kV and rise-time ~ 1 ns is applied to the gap.

The laser output was determined by an OPHIR type meter (Ophir Optronics LTD, Inc.) with a PE50BB sensor head, which was installed at a distance of a few tenths of a millimeter from the MgF2 plate, while the remaining gap was purged with helium. Radiation energy on FI laser lines from diffuse discharges in gas mixtures with F2 was measured without blowing the gap and then subtracted from the measurement result.

The VUV laser and spontaneous emission spectra of the diffuse discharge plasma and the emission pulse waveforms on separate radiation lines in the wavelength range of 100–545 nm were measured with a VM—502 vacuum monochromator (Acton Research Corp.) with an EMI9781B photomultiplier (PMT). A DA—780 detector assembly with sodium-salicylate-coated window as a wavelength convertor was placed in front of the PMT. The rise time of the signal of this PMT was 3 ns, the fall time was 30 ns. The monochromator was evacuated with an oil-free NORD—100 pump to a residual pressure of 10–6 Torr.

When measuring VUV spectra of the diffuse discharge, the CaF2 plate was removed, and the window was joined to the VM—502 monochromator LiF input plate.

The UV–visible radiation pulses were measured with a FEK-22SPU vacuum photocell. The use of a coaxial design of the photocell provides a rise time of the electrical signal on the level 0.1–0.9 no more than 0.5 ns.

Waveforms of radiation pulses on red FI lines were recorded with a MDR—23 monochromator equipped with a 1200 lines mm–1 grating and FEK—29SPU coaxial vacuum photocell with a sensitivity range from 400 to 1100 nm.

To measure the VUV pulses, the photo-cathode in the standard FEK—22 coaxial vacuum photocell was replaced with a polished copper plate. The photocell was vacuum-tightly attached to the MgF2 plate and pumped out to a residual pressure of 5 × 10–4 Torr. The work function of the copper cathode is 4.53–5.10 eV, which corresponds to the photoelectric threshold of λ = 274–245 nm [16]. The photocell with copper cathode was tested by laser pulses at λ = 351–353 nm (XeF* laser), λ = 337.1 nm (N2 laser), and λ = 248 nm (KrF* laser). The laser pulse shapes at λ = 248 nm, recorded by FEK—22 photocell and the modified one, coincided, while the radiation pulses at λ = 337 and λ = 351–353 nm was not detected by the copper-cathode photocell.

It should be noted that the techniques for measuring VUV pulses of plasma radiation by devices with metal photocathodes are well developed, provide high accuracy, are quite widespread [17, 18] and are also used to register laser pulses [19]. To weaken the influence of the integrated spontaneous emission from the diffuse discharge on the VUV laser signal, the Cu photocathode was installed at a distance of 30 cm from the output MgF2 plate.

Images of the diffuse discharge plasma glow were captured with a Sony A100 digital camera Due to short duration of the diffuse discharge glow [9,10,11,12,13,14], the pictures were taken in one pulse. The exposure time was 1 s to synchronize the triggering of the camera and the RADAN generator.

To measure the discharge current pulses, an ohmic shunt assembled from low-inductive microchip resistors was used. Electrical signals were recorded using a TDS 3054 digital oscilloscope (0.5 GHz, 5 GS s−1). The oscilloscope channels were synchronized using a four-channel BNC generator, the signals from which were simultaneously fed to the channels via high-frequency cables. The distance from the Cu photocathode to the MgF2 plate was taking into account, when determining the laser pulse onset.

3 VUV lasers pumped by a diffuse discharge

3.1 Hydrogen laser

The possibility of obtaining stimulated emission in the VUV region on transitions of molecular hydrogen at a high excitation power was substantiated in 1965 [20]. A few years later, a theoretical model of the VUV hydrogen laser was created [21]. First H2 lasers based on the Werner and Lyman bands pumped by a self-sustained discharge were launched in 1970 [22, 23]. According to the theoretical calculations [21, 24], the upper laser level (ULL) of an H2 laser, similarly to a nitrogen laser at 337.1 nm [25], is populated by direct electron impact. In this case, the electron energy must be at least 12.5 eV. Since the upper laser level lifetime is short (≈ 1 ns), for its optimal excitation, according to estimates [4], it is necessary to provide across the laser gap the value of the parameter E/p ≈ 500 V cm−1 Torr−1, where E is electric field strength, p is gas pressure. As in a nitrogen laser, after breakdown of the laser gap due to fast electron multiplication in the discharge plasma, the value of the E/p parameter and the average electron temperature sharply decreases. Accordingly, to quickly populate the ULL and reach the lasing threshold on numerous transitions of B1Σ+u → X1Σ+g Lyman band, it is necessary to apply high—voltage pulses (≈ 100 kV) with a short rise-time and provide at the stage of voltage drop very high discharge current density (~ 100 kA cm−2). Such high-power excitation pulses can be obtained using low-inductance generators based on strip lines and short discharge gaps of small width [4, 22, 23, 26,27,28], as well as electrodes in the form of points [22] or thin foils [4, 26,27,28]. As a rule, the gap length is varied in the range 0.1–2 cm and the discharge width is limited to a few mm and optimal H2 pressure was usually 10–100 Torr.

Thus, short high-voltage pulses were applied to the gaps providing a sharply inhomogeneous electric field near electrodes. With a high degree of probability, it can be concluded that in the above references run-away electrons could appear and similarly to [9,10,11,12,13,14] provided preionization and diffuse discharge formation.

It should be noted that in the conditions of high current densities, quenching of B1Σ+u state in collisions with electrons becomes noticeable, which significantly reduces the ULL lifetime and the efficiency of the high-current excitation mode [29]. Probably for this reason, in the references, cited above, duration of the H2 laser pulses was no more than 1–3 ns, and the output did not exceed 0.5 mJ. The efficiency of H2 lasers (~ 0.01%) was an order of magnitude lower than that of a UV nitrogen laser obtained using this excitation technique.

Integrated photograph of the glow in the gap is shown in Fig. 2, while spontaneous emission spectra of the diffuse discharges in pure hydrogen and its mixtures with some rare gases are shown in Fig. 3. The discharge similarly to [13, 14] has the form of weakly luminous in the visible spectral range diffuse channels starting from bright spots on the edge of the blade electrodes and then quickly expanding towards the middle of the gap, forming a glow without signs of arc channels. The number of spots increases while width of the jets reduces with the gas pressure. Bright arc channels appear against the background of diffuse glow in gas mixtures with argon. As a result, intensity of UV and visible emission in mixture with Ar becomes more noticeable. The intensity of the UV and visible glow near the electrodes and in the discharge balk slightly increased with He additions, while the intensity of visible radiation becomes more noticeable in mixtures with neon.

Fig. 2
figure 2

Diffuse discharge view in He: H2 = 2 atm: 100 Torr gas mixture. The discharge glow brightness has been increased

Full size image
Fig. 3
figure 3

Spectra of the diffuse discharge in pure H2 (a) and in mixtures of 100 Torr H2 with Ar, He (b) and Ne (c)

Full size image

It is seen from Fig. 3 that the Lyman band of H2 in the wavelength range 120–170 nm is most intense in the VUV spectra of the diffuse discharge plasma in pure H2 and its mixtures with He and Ne. Intensities of the atomic hydrogen lines at 100–120 nm were low compared to that of the Lyman band. Emission peak at ≈ 143 nm followed by a drop in radiation power observed in various gas discharges can be related to the spectrum structure of the Lyman band of H2, see, for example, [30, 31]. Similar complex spectrum of the microwave discharge in H2 was observed in [32, 33], while increase of H2 pressure and additions of helium and argon noticeably reduced the intensity of the Lyman band. Therewith changes in the emission intensity on some Lyman band transitions are possible by selective population and depopulation of vibrational H2 levels in collision with the diffuse discharge plasma components which can lead to the appearance and disappearance of individual peaks of the VUV radiation.

The radiation intensity of the discharge plasma in the UV and visible ranges was found to be relatively low in accordance with the data shown in Figs. 2 and 3.

The Lyman band intensity increases with decreasing hydrogen pressure, and its maximum is observed at a pressure of 15–35 Torr. This is due to an increase of the overvoltage across the gap during the diffuse discharge formation by nanosecond high-voltages pulses when H2 pressure is decreased [34, 35]. Consequently, the E/p parameter across the discharge gap increases which improves population rate of B1Σ+u level of hydrogen molecule. Increase in H2 pressure results in the fast temporal fall-off of the average electron energy and strongly decreases the overall efficiency of the upper laser level population [36].

Additions of helium and argon reduced the intensity of the VUV band. Similar effect was observed in [37] where intensity of the H2 laser lines decreased markedly when He and Ar were added to H2 gas.

The increase in the argon buffer gas pressure results in decrease of efficiency of the VUV molecular hydrogen laser. While maintaining uniformity of diffuse discharge, it can be associated with quenching of the upper laser level by argon, with a decrease in the reduced electric field, and with appearing bands of the “third continuum”, which radiates in the region of 200–400 nm over time in tens of nanoseconds. One can see the radiation in this region of the spectrum in Fig. 3b. Note that the formation of the “third” continuum has been studied in many works, see, for example, [38, 39]. Moreover, as shown in the theoretical work [40], discharge in argon at high pressures, in addition to the well-known 1st and 2nd continuum [41], emits several continuums in different spectral regions, which are sometimes combined into one third continuum.

Besides, the VUV emission intensity in mixtures with argon decreased due to development of diffuse plasma inhomogeneities. The discharge contraction led to the appearance of continuum radiation in the visible region (380–520 nm), see Fig. 3b associated with the recombination radiation of the spark leaders and channels [10].

It was found that neon sufficiently improved the intensity of Lyman band, see Fig. 3c. Since the intensity of spontaneous emission P(t) is proportional to the population of the upper level n of B1Σ+u–X1Σg+ transition and the probability of spontaneous emission A [42]:

$$ P\left( t \right) = n \, \left[ {B^1 S^+_u } \right]A, $$
(1)

it follows from Fig. 3 that He and Ar decrease, while Ne increase population of the upper laser level of the Lyman hydrogen band.

Spectrum of the hydrogen laser radiation in our experiments was similar to [8, 22, 23, 27, 37, 43] and consisted of 10–15 VUV lines in the range 148–161 nm.

The main features of the hydrogen laser operation pumped by the diffuse discharge formed by runaway electrons are shown in Figs. 4, 5 and 6.

Fig. 4
figure 4

Characteristic waveforms of UV–Visible spontaneous emission (Psp), laser radiation (Plas) and diffuse discharge current (Id) in pure H2 at 15 Torr (a) and in He: H2 = 4 atm: 35 Torr gas mixture (b)

Full size image
Fig. 5
figure 5

Waveforms of laser pulses on Lyman molecular band (Plas): a in pure H2 at 35 Torr (1) and in mixtures of 35 Torr H2 with He at 100 (2), 400 (3), 750 (4); b in pure H2 at 60 (1), 100 (2), 150 (3), 250, 350, 450 (4–6) Torr; c in pure H2 at 150 Torr (1) and in mixtures of 150 Torr H2 with Ne at 150 (2), 500 (3), and 1150 (4) and 2250 (5) Torr

Full size image
Fig. 6
figure 6

a Laser power (1, 2) and energy (3) as function of He (1, 3), H2 content is 35 Torr and H2 pressure (2) and b laser energy versus Ne pressure, H2 content is 35 (1), 75 (2) and 150 (3) Torr

Full size image

The characteristic waveforms of intensity of diffuse discharge UV and visible spontaneous emission Psp and H2 laser pulse Plas both with the discharge current Id. are shown in Fig. 4. Therewith two modes of discharge operation were observed depending on the pressure and composition of the active mixture. Shot pulse of discharge current and powerful laser emission were observed at low H2 pressure, see Fig. 4a. This occurs due to the peculiarities of a diffuse discharge formation. It was found that when high overvoltage is applied to the blade electrodes, ionization waves (streamers) starts from edges of electrodes to the gap center [44, 45]. The velocity of the streamer usually increased with a decrease in the pressure and can reach several cm ns−1. A sharp increase of the electric field on the front of the streamers [9, 12, 13, 45, 46] and the sharp edges results in formation of fast and run-away electrons which provide preionizaton of a gas volume. As a result, the diffuse discharge is formed in the form of wide diffuse plasma channels. The discharge looks like a spherical ball or a pumpkin, while its width sharply decreases with increasing H2 pressure [12]. At low gas pressure the discharge width can reach several cm and the discharge current closes on the walls of the chamber, and the signal from the shunt disappears within ≈ 1 ns.

However in this case, during the gap closing by the streamers, a narrow region with increased electric field strength and high current density is first formed between the electrodes, where the upper laser level is effectively populated and then the discharge expansion occurs. An increase in the E/p parameter in the gap with sharply non-uniform electric field with a decrease in the H2 pressure is evidenced by the results obtained in [34]. In this work, an increase in the current of runaway electrons both with the duration of the voltage pulse across the gap and a slowdown in the rate of the voltage decay with decreasing pressure were observed.

As a result, the maximum energy and radiation power of the H2 laser of 0.12 mJ and 50 kW, respectively, were achieved in the diffuse discharge at hydrogen pressure of 15–35 Torr. It seems that in the conditions of low pressure operation, the authors of [4, 22, 23, 26,27,28] were forced to limit the discharge width for improving the laser parameters. Therewith maximal laser output was estimated to be 0.5 mJ [43], but laser pulse duration was as short as 1 ns due to very high current density. In our experiments, total laser pulse duration at optimal H2 pressure was as long as 5 ns and 2.5 ns at half-maximum.

When pressure of H2 or He–H2 gas mixture was increased, the discharge width is noticeably reduced, and current oscillations with a half-period of about 10 ns were measured in the gap, see Fig. 4b. The spontaneous emission pulse has short peak and then only weak UV–visible emission is observed during several periods of the discharge current. As in the low pressure operation mode, lasing on the Lyman band starts ≈ 1 ns after the discharge formation and was as long as 5 ns, but the laser power reduces to 12 kW due to a decrease in the E/p parameter in the gap when the pressure is increased. Note that, lasing was observed in pure H2 at pressure up to 3 atm, but the radiation power and energy still fell by 5 times.

Effect of foreign gases on the H2 laser pulse waveform is shown in Fig. 5. No VUV signal was observed in mixtures of H2 with nitrogen and carbon monoxide, therefore data for these mixtures are not shown in Fig. 5. Increase of H2 pressure and (or) additions of 0.15–1 atm of helium significantly changed the laser pulse waveform. The laser pulse duration was doubled in certain pressure ranges of H2 or He–H2 mixture. The effect of additives had a threshold character, see Fig. 5a, b. When the gas pressure is increased, the second laser peak first appears, but the laser power in the first peak decreased. Then, the pause between the peaks shortened and then a continuous pulse is evident with a total duration of 10 ns which approximately corresponds to the duration of the half-cycle of the discharge current. Finally, the laser pulse begins to shorten as the pressure is further increased.

This operation mode of a discharge H2 laser, apparently, has not been observed previously. It should be noted that an increase in the H2 laser pulse duration from 1.5 to 3 ns was observed in [47] within H2 pressure range 0.4–0.7 atm, but the laser output was only several µJ.

The increase of the laser pulse duration is associated with transition to a four-level scheme of the active medium operation, predicted in [4, 36], when the lower laser level X1Σg+ is quenched in collisions with components of the diffuse discharge plasma in processes of dissociation [29] and VT relaxation at a gas mixture pressure in the order of one atm. As a result, the lifetime of the population inversion increases both with the laser pulse duration.

Additions of neon in hydrogen have a little effect on the laser pulse “tail”, however, the peak power of the laser radiation gradually increases with Ne pressure and eventually exceeds this parameter in pure hydrogen, see Fig. 5c. Therewith neon additions do not change the shape and duration of the laser pulse in mixtures with optimal H2 pressure. However, in this case, the peak of the laser pulse power is reached earlier (see the black line in Fig. 5c), which indicates an increase in the gain in the active medium.

Effect of He and Ne on the laser peak power and output is shown in Fig. 6. In mixtures with He the peak radiation power noticeably falls at pressure lower than 1 atm, and then somewhat increases. Similar dependence is also observed for the radiation energy, but its growth is observed at lower He pressure due to longer duration of the laser pulse.

Impact of the neon addition on the laser parameters depends on hydrogen content, see Fig. 6b. The addition of neon to hydrogen at a pressure of 35 Torr first leads to a sharp drop in the radiation power and energy, and then the laser parameters increase linearly with Ne pressure and eventually reach maximum values obtained at optimal pressure of pure hydrogen. In mixtures with higher hydrogen content, only a linear increase of laser power and energy with Ne pressure is observed.

Apparently, in gas mixture with low Ne and H2 pressure decrease of breakdown voltage both with the E/p parameter across the gap can occur which should result in low population rate of B1Σu+ state. This assumption is confirmed by the results obtained in [48]. In this work, no runaway electron beam was observed in a diffuse discharge in low pressure neon, which was associated with breaking down the gap without forming an ionization waves with a high electric field at their front and, accordingly, by the relatively uniform distribution of the electric field along the gap during the discharge current growth. Increase of H2 pressure results in appearance of ionization waves, and Ne addition has already little effect on the diffuse discharge formation.

Lasing pulses on several lines of the Lyman hydrogen band are shown in the Fig. 7. It follows from Fig. 7, as in the case of the integral laser pulse, that the rise time of the stimulated emission pulses in mixtures with neon is noticeably shorter than in mixtures with helium. The rise time in neon based mixture is about 2 ns which corresponds to the photomultiplier resolution time. This indicates a higher gain in the neon-based active gas media, which ultimately depends on the population of the upper laser level. Increase in concentration of H2 molecules in B1Σu+ state in mixtures with Ne follows from the data shown in Fig. 3, as well. As a result, the radiation energy of the hydrogen laser in mixtures with Ne increased and reaches maximum values. The linear increase in the radiation energy with neon pressure allows us to hope to obtain a further improvement of the laser parameters at Ne pressure above 5 atm.

Fig. 7
figure 7

Laser pulse waveforms on several lines of the hydrogen Lyman band in He: H2 = 1.5 atm: 35 Torr (a) and Ne: H2 = 3.5 atm: 35 Torr (b) gas mixtures

Full size image

Thus, the experiments carried out unambiguously indicate an increase of the upper laser level population in mixtures of hydrogen with neon. The buffer gas effect can be explained in several ways. A volume discharge in neon–hydrogen mixtures had been widely used to pump a laser on Ne transition at λ = 585.3 nm [49, 50]. It is believed that in the kinetics of this laser, hydrogen is used to clean the lower laser level in the Penning ionization reaction:

$$ {\text{Ne}}* + {\text{H}}_2 \to {\text{H}}_2^+ + {\text{Ne}} + {\text{e}} $$
(2)

However, another reaction channel (2) was considered in [51]. Significant fraction of the input electric energy is spent on the formation of excited Ne* atoms and Ne*2 excimer molecules in a volume discharges in Ne–H2 gas mixtures at high Ne pressure. Then, in collisions with excited Ne atoms and molecules H*2 in higher lying states can be produced and population of the B1Σu+ hydrogen state is possible by cascades from the excited H*2 states [52, 53].

Ne2* excimer emission on the second continuum which corresponds to transitions from 3Σu excimer state to the repulsive ground state at wavelength range 76–88 nm with peak at 84 nm and first excimer continua due to the radiative decay of vibrationally excited levels of the 1Σu excimer state, centered between 73 and 75 nm along with Ne* resonance lines at 73.59 and 74.37 nm are dominated in discharge plasma spectra in mixtures with neon [54].

The radiation absorption cross section by hydrogen molecules in this spectral range falls within (1–3) × 10–17 cm2 [55, 56]. Under the conditions of our experiment, the radiation of neon molecules and atoms will be effectively absorbed by H2 molecules in the ground state forming highly excited H2*, and the process described above can again be repeated.

The upper laser level can also be populated in the processes of recombination of H2+ molecular ions and collisions of H2 molecules with excited H* atoms [57, 58], which are produced in the reaction [59]:

$$ {\text{Ne}}_2^* + {\text{H}}_2 \to 2{\text{Ne}} + {\text{H}} + {\text{H}}^{*} $$
(3)

To determine the physical reasons for the increase in the output parameters of the H2 laser in mixtures with neon, additional calculations and studies are required.

3.2 Molecular fluorine laser

3.2.1 Lasing on atomic fluorine lines

The kinetic processes in mixtures of helium with fluorine that form inversion at atomic fluorine transitions are part of the kinetics of VUV lasers on F2 molecules. One of the channels for populating the upper laser level of the VUV fluorine transition D′(3П2g) → A′(3П2u) is the energy transfer reaction:

$$ {\text{F}}^* + {\text{F}}_2 \to {\text{F}}_2^* + {\text{F}}, $$
(4)

where F* is excited fluorine atom in metastable 3s4Pj, j = 5/2, 3/2, 1/2 states, which are the lower laser levels for a laser on fluorine atomic lines [6]. Thus study of lasing on FI lines is important for a better understanding the F2 laser kinetics.

Radiation spectrum of the FI laser pumped by diffuse discharge contains five doublet and quartet lines in the range of 634–755 nm. The letters Q and D on denote quartets (transitions from 3p4S03/2 and 3p4P05/2,3/2,1/2 levels) and doublets (transitions from the 3p2S01/2 и 3p2P01/2,3/2 levels), respectively, see Fig. 8. Measurement of the spectrum over the width of the laser domain showed that at low gas mixture.

Fig. 8
figure 8

Spectra of FI laser as functions of helium pressure p(He), partial pressure of F2 is 5 Torr. Letters Q and D designates the quartets and doublets

Full size image

pressure the laser spot periphery is dominated by radiation with λ = 634 nm, while lines with λ > 700 nm are the most intensive in the laser spot centre. This looks like a drop in the radiation intensity, since the eye is unable to see radiation at wavelength longer ≈ 700 nm.

A sharp increase in the lasing power with the buffer gas pressure is observed on the 3p4P05/2–3s4P5/2,3/2 quartet transitions at λ = 739.8 and 755.2 nm, while the emission intensity on the other lines practically disappears. It can be assumed that these dependences are due to the rapid relaxation of the upper levels of other four transitions in collisions of F2 molecules and helium atoms.

A distinctive feature of the FI laser is delay time of the laser pulse relative to onset of the discharge formation. The laser pulse is evident within 1 ns within after onset of the discharge current. Total pulse duration is about 5 ns. However, the lasing pulse widens from 5 to 15 ns with an increase in the Q factor of the laser cavity, see Fig. 9.

Fig. 9
figure 9

Waveforms of UV and visible spontaneous (Psp) emission, diffuse discharge current (Id), integral FI laser pulses (Plas) and laser pulses at 739 and 755 nm. Reflectance of the output mirror is 70% (a) and 7% (b). Gas mixture He: F2 = 5 atm: 5 Torr is used

Full size image

Another feature is a wide lasing spot width which can reach 1 cm at pressure 1–2 atm. This is due to low lasing threshold on fluorine lines and wide current flow region in the initial stage of the diffuse discharge formation [60]. However, at He pressure of more than 5 atm the FI laser spot narrows to 3–4 mm, and the visible intensity dip in the center is not evident.

The VUV radiation domain does not exceed 3–4 mm showing that the region with perceptible discharge current density sharply narrows because high current density is required to reach the lasing threshold at 157 nm.

The total radiation energy on the FI lines increases linearly with the mixture pressure, and exceeds 0.35 mJ at p = 10 atm. This confirms the diffuse discharge stability in mixtures with fluorine, since the lasing energy on the FI lines drops sharply [61] in a non-uniform discharge.

3.2.2 VUV fluorine molecular laser

Typical waveforms of the discharge current, spontaneous emission in the gap at 200–600 nm, and F2* laser pulses are shown in Fig. 10. The VUV lasing threshold was reached at the mixture pressure of about 3 atm. The short current peak about 2 ns long should be highlighted at the beginning of the discharge current pulse, which can be explained by the presence of a dynamic bias current [62]. The bias current arises during motion of the ionization wave front. The ionization waves start from both blade electrodes and meet in the gap, which leads to the appearance of the first peak of the current through the gap. As the ionization waves approach each other near the gap center, an increase in the electric field occurs in the region between them. The field enhancement made it possible to obtain from the central part of the gap UV lasing on the second positive nitrogen system (λ = 337.1 nm) in pressured nitrogen and on the first negative one (λ = 428 nm) in mixture He–N2 using the laser device with blade electrodes [13, 14].

Fig. 10
figure 10

Waveforms of UV and visible spontaneous emission in the gap (Psp), F2* laser radiation (P157) and diffuse discharge current (Id) in He: F2 = 10 atm: 5 Torr (a) and He: F2 = 6 atm: 2 Torr (b) gas mixtures

Full size image

As follows from Fig. 10, the main part of input electric energy is deposited in the helium–fluorine discharge plasma after the gap was closed by ionization waves during the first and subsequent half cycles of the discharge current. The laser pulse at λ = 157 nm starts during the first peak of the discharge current, but its maxima were observed after the maxima of id. The laser pulse had three pronounced vertices corresponding to three successive half cycles of the diffuse discharge current. Total duration of the laser pulse similarly to [63] can be as long as 50 ns in the mixtures with fluorine content 1–1.5 Torr (see Fig. 11). The persistence of the diffuse nature of the discharge during six half-cycles of the discharge current is confirmed by the continuation of the laser pulse.

Fig. 11
figure 11

Waveforms of UV and visible spontaneous (Psp) emission, diffuse discharge current (Id) and laser pulse at 157 nm (P157), gas mixture He: F2 = 4 atm: 1.5 Torr is used

Full size image

It should be noted that there are noticeable differences between the peaks of the discharge current pulses and peaks on the laser pulses in gas mixture with different composition and pressure. Thus, at high helium pressures and fluorine concentrations in the gas mixture, the amplitude of the short peak in the Id pulse could be more than 50% of the maximal amplitude of the discharge current during the first half-cycle. Under these conditions, the laser pulse began to form as early as ≈ 1–2 ns after onset of the current pulse, and first laser peak had the highest amplitude. In the mixtures with low pressure and F2 concentration the amplitude of the first current peak decreases, the delay time of the lasing pulse onset increases, and the second peak in the laser pulse has the maximum intensity.

At high He pressures and F2 concentrations in the mixture, see Fig. 10a, the breakdown voltage increases [64], which leads to an increase in the electric field near the electrodes and in the gap, and also improves the rate of current rise in the gap. In this case, as in [10], most of the incident wave of the voltage pulse is reflected from the gap, and the main discharge is formed after the arrival of the voltage wave reflected from the generator.

It can be seen from Fig. 10a that the pumping power in the first current peak is sufficient to reach the lasing threshold. This ensures a rapid increase in the lasing power during the first half-cycle of the discharge current. Similar laser operation mode was obtained with an inductive energy storage generator (IES) [65]. The IES ensured a short powerful pumping pulse and rapid formation of an inversion population. Under conditions of Fig. 10b the inversion population slows down, and therefore the second laser peak has the maximum power.

Laser output in mixtures with He and Ne buffer gases is shown in Fig. 12. The radiation energy of the molecular fluorine laser similarly to [6, 7, 11] increased linearly with helium pressure. This is due to an increase of the excitation power, since the voltage across the discharge gap increases with He pressure [64], while the discharge current in our experiments is mainly determined by the circuit impedance.

Fig. 12
figure 12

Laser output at 157 nm as functions buffer gas pressure in mixtures of 5 Torr F2 with He and Ne

Full size image

Maximum F2* laser output was as high as Q157 = 3.8 mJ at He pressure of 10 atm. This energy Q157 corresponds to the electrical efficiency of F2 laser of η0 = 0.18% which exceeds the maximum efficiency of the F2* laser pumped by a volume self-sustained discharge with preionization, obtained in [6, 7].

Addition of neon to a mixture with helium or complete replacement of helium by neon led to a noticeable drop in the laser output. This is due to the following reasons. The first one is related to the kinetics of the upper laser level in the fluorine laser. As was mentioned above, one of the channels for populating the D′(3П2g) state of F2* molecule is the energy transfer process (see Eq. 4). Therewith in a diffuse discharge in mixtures with He about 50% of fluorine molecules in the D′(3П2g) state are formed in reaction (4) [11]. In mixtures with Ne, concentration of excited fluorine atoms F* drops noticeably, which leads to lower the gain in the active medium and decreases the laser output [66]. The second reason is related to the insufficient discharge resistance in mixtures with Ne [64], which, under the conditions of our experiment, reduces the pumping power and input electric energy, resulting in low laser output in Ne based mixtures.

Long-term VUV lasing during several half-cycles of Id indicates high stability of the diffuse discharge formed in a non-uniform electric field in mixtures with high fluorine content. This is due to the conditions for the diffuse discharge formation with high-voltage pulses with amplitude about 300 kV and fast increase of the conduction current density in the gap dj/dt > 50 A cm−1 ns−1. In this pumping regime, the process of direct ionization by electron impact dominates during several nanoseconds at the stage of the current rise [65], which provides high uniformity and stability of the diffuse discharge plasma.

3.3 ArF laser

The operating modes of ArF laser were similar those of F2 laser, see Fig. 13. The laser pulse had three distinct peaks. The maximum power of laser radiation was observed in the first or second stimulated emission peak depending on the composition and pressure of the gas mixture. The first peak was the most intensive in gas mixtures with high pressure and high fluorine content. The laser pulse had continued during two periods of the discharge current. Total duration of the pulse was as long as 35 ns.

Fig. 13
figure 13

Waveforms of UV–Visible spontaneous emission (Psp), ArF laser radiation (P193) and diffuse discharge current (Id) in He: Ne: Ar: F2 = 2: 3 atm: 150: 3 Torr (a) and Ne: He: Ar: F2 = 1: 5 atm: 200: 4.5 Torr (b) gas mixtures

Full size image

The effect of buffer gas on the ArF laser output energy is shown in Fig. 14. Similarly to [67] neon additions into helium based mixture first lead to a sharp increase in the laser output, and then the radiation energy does not change. The experimental and theoretical study of the buffer gas effect on the laser parameters was reported in [68, 69]. It was found that additions of Ne in the gas mixture improve the discharge plasma uniformity, but has small effect on discharge plasma active resistance. Thus, it is the best homogeneity of the discharge plasma in mixtures with small Ne amount that ensures the initial the lasing energy gain. Further increase of neon amount has no longer effect on the discharge uniformity and does not greatly improve impedance matching of the circuit impedance and the discharge resistance.

Fig. 14
figure 14

ArF* laser output versus additions of He in Ne: Ar: F2 = 5 atm: 200:4.5 Torr (1) and Ne in He: Ar: F2 = 5 atm: 200: 4.5 Torr gas mixtures

Full size image

The diffuse discharge plasma resistance increase linearly with He pressure and the efficient energy transfer into the diffuse discharge load occurs. Therefore, the laser output grows linearly with He content in the Ne—based gas mixture since the homogeneity of the plasma is provided initially. Maximal ArF laser output in our study reach 2.4 mJ from active volume of 20 cm3. In Ar containing gas mixtures, diffuse discharge consisted of narrow diffuse channels unevenly distributed along the electrodes, spark channels were often observed in the gap, while the glow of the rest of volume was relatively low. Low output of an ArF laser can be related to the diffuse discharge inhomogeneities in gas mixtures with argon.

4 Conclusion

Parameters of VUV stimulated emission of a diffuse discharge formed in a non-uniform electric field by run-away electrons in hydrogen and H2–He (Ne), He (Ne)–F2, He (Ne)–Ar–F2 gas mixtures were studied. Efficient lasing in the VUV spectral region on the Lyman band of molecular hydrogen, the D′(3П2g) → A′(3П2u) band of molecular fluorine and ArF* excimer molecules has been obtained.

It is shown that the addition of He to hydrogen results in a twofold increase in the duration of the VUV laser pulses on H2 molecules. Similar effect was found in pure H2 in certain range of gas pressure. The laser pulse duration was doubled from 5 to 10 ns in certain pressure ranges of H2 or He–H2 mixture. The increase of the laser pulse duration can be explained by the transition to the four-level laser operation mode, when the lower laser level X1Σg+ is quenched in collisions with components of the active gas mixture.

Addition of large amount of Ne to hydrogen was found to greatly improve the laser output on Lyman band of hydrogen. Therewith the laser energy grows linearly with Ne pressure and further improve of the laser operation can be achieved at buffer gas pressure over 5 atm.

The volume stage of the diffuse discharge in mixtures with F2 was found to maintained for 2–3 periods of the gap current.

Laser output at 157 nm up to 3.8 mJ and the radiation pulse duration up to 55 ns were obtained. The electrical efficiency of the F2 laser of η0 = 0.18% was realized, which exceeds the efficiency of lasers of this type excited by transverse volume discharges with additional preionization, obtained in [6, 7]. Long-pulse operation of discharge ArF laser was demonstrated, as well.