Introduction

Chemical lasers have been extensively developed since the mid 1960’s. Almost all established chemical lasers for research and application areas, work in the infrared (IR) spectral range. The invention of a short-wavelength chemical laser would be recognized by the research community as a big step forward. Currently, lasers pumped by a nearly resonant energy transfer from chemically generated singlet oxygen to atomic iodine, chemical oxygen-iodine laser (COIL), remain as the only routinely working and wide-spread scheme in this direction [1]. This laser neither enters the visible (Vis) or ultraviolet (UV) range, working only in the near IR at 1315.2 nm.

Numerous approaches towards a UV–Vis chemical laser have been followed and various schemes tested. They were all either complicated or based on exotic species. Benard [2] investigated chemiluminescence emitted by a mixture of He, FN3 and Bi(CH3)3 thermally initiated in a shock tube reactor, where a very weak lasing at 470 nm was achieved, assigned to the BiF(A → X, Δv = 3) transition. A red emission was registered [3] from an externally chemically generated singlet oxygen stream flowing over various heated metals. However, no lasing has been achieved in this scheme. Yoshida et al. [4] reported a weak lasing at 700 nm assigned to dimolecular oxygen emission in an optimized chemical generator of singlet oxygen. Herbelin [5] proposed a hybrid system of a supersonic oxygen-iodine laser optically coupled to a nitrogen fluoride DF-like supersonic flow to achieve a lasing from NF(b1 Σ) at 528.8 nm. Kudriavtsev et al. [6] investigated the possibility of Vis lasing from IF(B → X, 0 → 4) at 603 nm, generated by I2 + F2. The resonance energy transfer from excited NF to IF was considered to occur in a gas dynamic setup with supersonic flow mixing based on a shock tube. Gavrikov et al. [7] reviews these earlier studies. During the last 25 years, numerous approaches have been carried out toward this task, however only small progress has been achieved.

It seems that more than five decades of chemical laser research has not provided any simple and efficient chemical laser device working in the Vis and UV spectral range. Therefore, the search for an intense (plasma) chemical UV–Vis emission from a simple gaseous system still has a strong motivation. Recently, we found a promising UV emission in a practical plasma-chemical system. In this article, we report on strong UV luminescence emitted in the afterglow far from a radiofrequency (RF) powered discharge in a He–C2F4–NO mixture.

Our original idea was, to allow singlet-oxygen react in the gas phase with C2F4. The expected intermediate tetrafluoro-dioxetane would decompose, due to the excess energy to two F2CO molecules; one of them would be electronically excited and would emit in the UV, similar to the observation of Bogan et al. [8]. Singlet oxygen is produced in an RF discharge, in which minor quantities of NO are added to scavenge O atoms and ozone. The C2F4 is produced by depolymerization of polytetrafluoroethylene (PTFE), by which a separate He gas-discharge chamber is cladded. The two gas flows merge in a supersonic beam, where UV emission is observed. Surprisingly, the emission is from the trace component NO and not from F2CO.

Experimental

The layout and technical parameters of our setup are determined by the fact that this device was originally designed and operated as a testing tool for the development of a supersonic discharge oxygen-iodine laser [9,10,11]. The presented work uses the same device with slightly different parameters. The basic scheme is given in Fig. 1. It consists of two main subsystems, the discharge singlet oxygen generator (DSOG) and the discharge tetrafluoroethylene generator (DTFEG). Together both parts form a single low-pressure sub/supersonic gas flow system for plasma-chemical luminescence studies.

Fig. 1
figure 1

Simplified scheme of the experimental device, consisting of two main subsystems—the discharge singlet oxygen generator (DSOG) and the discharge tetrafluoroethylene generator (DTFEG). Both subsystems are connected in a single low-pressure sub/supersonic gas flow system. He* denotes an electronically excited state of helium

Full size image

Our recent DSOG concept is based on fast mixing of a rare gas RF plasma jet with a neutral oxygen stream. It implies an energy transfer from the RF rare gas plasma jet to the neutral O2 stream. This technique has the potential to compensate for some disadvantages of RF plasma jets containing O2, i.e. high production of O and O3 (both are strong quenchers of singlet oxygen) and a faster erosion of the electrode adjacent to the highly excited and highly reactive oxygen plasma, resulting in increased contamination of the gas flow by the electrode material. Some remaining O and O3 have to be, however, scavenged by NO titration. A survey of DSOG related papers can be found elsewhere [12, 13].

The scheme of our DSOG is shown in Fig. 1. The RF plasma jet is produced in an Al nozzle with a cylindrical hole, using a gas mixture of Ar + He. Al has a low rate of O2(1Δ) surface quenching. The diameter/length of the nozzle is 5/10 mm. The nozzle (hollow electrode) coupled RF discharge forms the plasma jet and a non-equilibrium plasma is created. The neutral stream of the O2 + He + NO gas mixture is injected radially at the nozzle exit plane perpendicular to the axis of the plasma jet. The injection slit, surrounding the nozzle exit, has a diameter/width of 5.3/0.5 mm. Stability of the plasma jet is achieved by an enhanced vortex flow upstream of the RF nozzle and by an optimized gas mixture flow rate of Ar:He = ~ 10:5 mmol/s. All the metal parts in the RF plasmatron are made of aluminium (99.5% Al, A1050) and are water cooled (water temperature 12 °C). The visible length of the plasma jet is only ~ 5 mm due to the presence of electronegative O2.

The generation of O2(1Δ) is achieved by a laterally symmetric injection of a neutral mixture O2:He:NO = 1.5:2.1:0.11 mmol/s into the RF plasma jet of Ar:He = 10.5:4.8 mmol/s. The RF absorbed power (evaluated by subtraction of the reflected power from the forward power) is 498 W and the frequency is 13.56 MHz. An L-type impedance-matching circuit enables perfect matching when the reflected power is below 1% of the forward power throughout all experiments. The temperature of the gases at the discharge chamber is 73 °C and at the outlet is 26 °C. This temperature is measured by thermocouples electrically shielded (for suppression of electromagnetic noise) and covered by fused quartz (to avoid deactivation of ions and excited species on a metal surface). The yield of O2(1Δ) is ~ 5%, whose determination is described elsewhere [10]. The upstream/downstream nozzle pressure is 16.4/1.2 kPa. The residence time of gases in the nozzle exceeds 10 μs.

Our method of C2F4 (TFE) generation (DTFEG) is based on depolymerization of PTFE by a rare-gas RF plasma (electrode coupled) at a low pressure and at room temperature. Such a C2F4 production method is, for our case, technically the simplest and cheapest: our DTFEG is a simple modification of our hardware for producing atomic I for an oxygen-iodine laser by RF plasma dissociation of iodine compounds [11]. The feasibility of an inductively coupled RF discharge PTFE depolymerization (TFE yield 94% at a temperature of 600 °C and a pressure of 5 kPa) has been demonstrated experimentally [14]. Alternative ways of TFE production (e.g. direct synthesis of TFE [15, 16], laser ablation of PTFE [17] or vacuum pyrolysis of PTFE [18, 19]) are not suitable in our apparatus. Successful reports on PTFE depolymerization highlight the fact that a low pressure (namely 5 kPa [14]/0.67 kPa [19]) is necessary to achieve a high yield of TFE (namely 94% [14]/100% [19]). This pressure condition is fulfilled in our system. The temperature ~ 600 °C is not necessary in our system as the RF plasma, in our arrangement, excludes any reverse polymerization.

The scheme of our DTFEG is shown in Fig. 1. A central component made of pure Al combines a supersonic nozzle, an RF discharge chamber and a TFE injector. Its outer shape is designed to form, together with the cavity walls, a supersonic double-slit nozzle. The critical section of each half of the nozzle is 2.9 mm in height and 50 mm in width. The discharge is sustained in a central cylindrical space with a 9 mm inner diameter. The tungsten electrode has a free space length/outer diameter of 101/2 mm. The left and right channels of the 5 mm inner diameter serve for cooling water flow. The whole body is grounded (the second electrode). The RF power/frequency is kept at ~ 200 W/40 MHz. Also this RF discharge uses the L-type impedance-matching circuit for perfect matching, i.e. reflected power ≤ 1% of the forward power, throughout all the experiments. C2F4 (TFE) vapour, diluted in a He carrier gas (flow rate 1.22 mmol/s), is generated by depolymerization of a PTFE sheet inserted into the discharge volume of the injector to cover its inner cylindrical wall. The TFE flow rate is evaluated from the weight loss of the PTFE sheet and the operation time interval, assuming a 100% yield of TFE at low pressure < 1 kPa. A single piece of PTFE sheet (50 × 22 × 1 mm, 2.4 g) can supply the experiment for ~ 20 min with an almost steady state TFE flow rate of 11 μmol/s. The gas velocity inside the injector exceeds 50 m/s with a residence time of < 2 ms at a plasma pressure of 836 Pa. The gas leaving the discharge is mixed with the effluent of the DSOG (≤ 19 mmol/s), which flows around the injector. The discharge products are injected into the DSOG flow by two rows of circular apertures (2 × 15) each having a diameter/mutual distance of 2.2/3.1 mm. The aperture rows are placed 3.4 mm downstream of the nozzle throat. The flow far downstream is determined by the pumping velocity and a cavity pressure of 80 Pa, enabling supersonic flow. This gas flow system is pumped by a combination of vacuum pumps (rotary + Roots) with a pumping velocity of 3000 m3/h. The mixture of the injected gas with the effluent from the DSOG is accelerated in the nozzle to supersonic velocities (M ~ 2). The pressure in the detection cavity is measured 144 mm downstream of the injection point. The pressure in the discharge chamber is measured at the He inlet point. The following gases (purity) were used in this flow system: Ar (99.998%), He (99.996%), O2 (99.9992%), and NO (99.9%), respectively (all supplied by Air Products, Czech Republic).

The supersonically expanding gases produce luminescence, which can be detected by a spectrograph or a camera and also by the naked eye. Spectrograph scans are taken at a distance of 55 mm from the injection point. This distance is the shortest technically achievable in our setup. The side walls of the supersonic expansion duct are wedged windows (angle 2°) made of high quality optical fused quartz, enabling spectral scans free of interference patterns over a wide range of the spectrum from the mid UV > 190 nm to the IR. The spectral measurements region (see Fig. 1) is determined by the geometry of the optical coupling at the entrance aperture of the spectrograph. The spectrometer (Andor Shamrock SR 303i-A) is equipped by a camera (Newton EM DU97OP-UVB). The spectrometer itself has a Czerny-Turner arrangement with imaging toroidal optics (focal length 303 mm, aperture f/4, grating 150 l/mm, blaze 500 nm and span ~ 520 nm). The camera uses an electron-multiplying charge coupled device (EMCCD). The whole spectrograph is calibrated by means of a mercury calibration lamp using the Hg 435.8 nm line. The spectrograph control, spectral data acquisition and data processing are carried out by a PC. Optical noise is reduced by accumulation of 100 subsequent scans and their superposition for each final spectrum. Wide range spectra (190–710 nm) are composed of two sets of spectral scans (190–350 nm without an optical glass filter and 350–710 nm with an optical glass filter, in order to avoid the second diffraction order of the strong UV bands overlaying the visible part of the spectrum). The visible part of the gas luminescence can also be photographed by a standard CCD compact camera with 3.5 megapixel resolution. The camera is located 123 mm downstream of the mixing point at a distance of ~ 300 mm from the wedged window on the opposite side to the spectrograph.

Results and discussion

Spectral lines of strong UV luminescence were observed during the experiments as shown in Fig. 2. Assignments of the observed UV bands to corresponding transitions and their comparison with previous experiments [20] are listed in Table 1.

Fig. 2
figure 2

Experimental observed spectral lines. The majority are due to NO(A2Σ+) UV emission

Full size image
Table 1 Assignment of the UV spectral lines
Full size table

According to Table 1, most of the bands originate from the second excited state of NO(A2Σ+), with and without vibrational excitation (v′ = 0 to 3). The expected sequence bands (v′ > 0 to v′′ > 0) were also observed but are not detailed in the table. A higher-resolution spectrum also shows the two spin components (j = 1/2 and 3/2) in the correct ratio. A numerical integration of the spectrum shows that this emitter generates 68% of the monitored total radiation power over the spectral range of 190–710 nm. The weak C → B emission of nitrogen is probably from an impurity, whose origin is not known at present. For helium, see below.

The upper state A2Σ+ of the emitting NO cannot originate from the discharge in the DSOG, as its radiative lifetime is only 0.20 μs [21] and therefore it would not survive the time (tens of milliseconds) taken to travel the 1.27 m distance (Fig. 1) to the observation region. Also, it cannot originate from a collision with an excited molecule near the merging region of the two gas flows, as the time taken to travel over 55 mm is still too long (> 100 μs). Hence this state must be produced by collisions with a longer-lived excited species.

To help identify this species, we systematically switched on and off the conceivable sources, either the electrical or gas supplies, as summarized in Fig. 3, which shows the strength of the highest-intensity line (236.2 nm) for all combinations of switching on only one or both discharges and blocking them (or not) from some of their feed material (PTFE, O2 and NO). Other parameters (Ar and He flow rates and both RF powers) are kept constant.

Fig. 3
figure 3

NO(A2Σ+, 0–1) emission intensity at 236.2 nm under the conditions of all combinations of inputs (PTFE, O2, and NO) and sources of gas excitation (DSOG, DTFEG and DSOG + DTFEG)

Full size image

It seems obvious that there is no 236 nm emission in the absence of NO (4 columns “NO off” in Fig. 3). However, this can be considered as an independent test that the assignment of the emission spectrum is correct. It is very interesting that the emission is not observed when only the DSOG is switched on (line “DSOG” in Fig. 3). Hence the long-lived excited species must come predominantly from the DTFEG discharge. A clue on its identity is given by the helium emission (Table 1), which is indeed only observed, if the DTFEG is switched on. A suitable long-lived species of helium is in its lowest metastable state He(2 3S1) (energy 19.82 eV, radiative lifetime 7900 s [21]). It is certainly also generated in the DSOG discharge, but then obviously nearly completely quenched on the long way (1272 mm, Fig. 1) from the discharge to the observation region. Therefore, NO is probably excited to its A2Σ+ state (energy 5.48 eV, radiative lifetime 0.20 μs [21]) by single collisions with He(2 3S1). Another spin-allowed product of such a collision is probably the lowest excited state NO(a4Πi) (energy 4.71 eV, radiative lifetime 160 ms [21]). However, its emission is too weak to be directly observed, but by another collision with metastable helium it could be excited to the emitting A state. (Note that multiple collisions can still take place over the 55 mm distance between the C2F4 injection and observation region; otherwise the singlet oxygen pumped iodine laser, which uses a similar setup and conditions of pressure and flow, would not work. Also the short-lived upper state of the He emission (see Table 1) must be produced locally by a collision of metastable He with another excited species.)

It is worth mentioning that we did not detect emission from higher states of NO. In view of the large energy difference it is surprising that triplet He produces only the lowest excited states of NO. Collisions of He(2 3S1) with NO were previously observed to give rise to A–X emission of the cation NO+ [22]. This emission is in the vacuum UV, a region not accessible with our spectrometer.

The role of gaseous O2 is surprisingly harmful for UV emission intensity. The peak intensity at 236.2 nm reaches a value ~ 3.7 times higher when the oxygen flow is switched off (Fig. 3). For interpretation, one should note that NO is always a minor component of the gas flow, designed to scavenge O atoms and ozone formed as side products of the discharge in oxygen. Hence a large part of NO is simply consumed when the O2 supply is switched on.

Solid state PTFE serves as a source of gaseous C2F4 (TFE). Depolymerization is brought about by collisions with metastable excited species such as triplet He(2 3S1). According to our interpretation above, an excess of this species collides near and after the mixing zone in the molecular beam with NO, causing emission from its A2Σ+ state. As also stated, such a collision must also produce NO in its metastable a4Πi state. Figure 3 (columns 3 and 7) shows that the NO luminescence is enhanced by a factor of 1.6–1.7, when PTFE is inserted into the DTFEG generator, no matter whether the other discharge (DSOG) is switched on or off. Obviously, this enhancement cannot be caused by any species formed from singlet oxygen (such as the dioxetane discussed in the introduction) and therefore must be due to a long-lived fluorocarbon species. The most plausible candidate is CF2 in its lowest triplet state. It is a by-product of C2F4 photodissociation [23, 24] at ≤ 193 nm (≥ 6.4 eV). It must also be expected from dissociation by collision with triplet He (19.82 eV). It may even be a by-product from the collisional depolymerization. This long-lived species (triplet CF2 (energy 2.46 eV) [25]) could provide its energy to excite the metastable quartet NO(a4Πi) in a collision to its luminescing A state (energy difference 0.77 eV).

DSOG is not only the primary source of reactive oxygen species (ROS) O2(1Δ), O and O3, but is also the additional secondary source of gas pre-excitation, namely the long lived metastable species NO(a4Πi), which cannot survive over long distances, and He(2 3S1), which can survive over long distances. Long lived He contributes to the enhancement of the UV emission by a factor of 1.16–1.22 (see the columns 3 and 7 in Fig. 3). This generator alone is not able to trigger the plasma-chemical reaction chain (see Fig. 3). Simultaneous excitation of gases without oxygen by DSOG and DTFEG gives the best results. Processes and data important for the best sensitized mid UV luminescence are summarized in Table 2.

Table 2 Processes and data important for the best sensitized mid UV luminescence
Full size table

The visible part of the gas luminescence in the supersonic expansion duct is shown in Fig. 4. This visualization confirms the supersonic gas flow. Compared to the UV part, the visible luminescence is much weaker (see the spectrum in Fig. 2). We have not identified the emitting species, as characteristic spectral features are absent in the visible range.

Fig. 4
figure 4

Photograph of the visible part of gas luminescence in the supersonic expansion duct as viewed from the side opposite to the spectrograph. Gases leaving the supersonic nozzle (left) are flowing to the vacuum pump system (right). The geometry of the supersonic rectangular duct is following: the length 160 mm, the height in the middle 26 mm, the width in the middle 56 mm and the full angle of vertical and horizontal expansion 4°

Full size image

To date, we cannot present quantitative emission intensities or number densities of the excited NO. However, as the longer wavelength emission is clearly visible by the eye, the UV intensities must be substantial. According to the emission spectrum in Fig. 2, the power per unit bandwidth in the strongest NO bands (with background subtracted) is by a factor of ~ 40 above the other Vis radiation. (Note that the “68% in the UV emission” mentioned above was the spectrally integrated power.) Although the experimental setup currently produces supersonic flows, we expect that subsonic conditions would likewise give rise to this emission.

The strong sensitized UV luminescence of NO molecules, which is experimentally observed in this work, leads to the question of NO laser feasibility. There are three successful methods for NO lasing—(1) NO laser radiating in the UV–Vis part of the spectrum [26,27,28,29], (2) NO laser radiating in the near IR part of the spectrum [30], and (3) NO laser radiating in the mid IR part of the spectrum [31, 32]. In the UV–Vis NO laser [26,27,28,29] NO is optically pumped to a higher state by an F2 laser at 158 nm using intensities of up to 20 MW/cm2. The absorption spectrum of NO is modified by an external magnetic field. The laser emits in pulsed mode from vibronic states at the lines 163, 211, 218, 226, 234, and 589 nm. The near IR NO laser [30] uses flash photolysis of NO by the vacuum UV > 165 nm. The laser emits from the transitions D(2Σ+) → A(2Σ+) (1,1) and C(2Π) → A(2Σ+) (0,0) at the lines 1106.9 and 1223.7 nm, respectively. The mid IR NO laser [31, 32] uses pulsed photodissociation of IBr at 532 nm (a frequency-doubled Nd:YAG pump laser) to provide a high yield of spin–orbit excited Br. Near resonant electronic-to-vibrational energy transfer from Br(2P1/2) to NO(v = 2) enables an NO(v = 2 → 1) laser operating at 5.4 μm. In our scheme, the NO(A2Σ+) state is pumped by energy transfer from a long lived species, probably metastable He and excited CF2. In our plasma-chemical system population inversion of NO(A2Σ+) versus high vibrational states (e.g. v ≥ 4) of the electronic ground state NO(X2Σ+) is conceivable, if the X state is vibrationally not much hotter than the A state. However, it depends on (presently unknown) absolute densities of the excited NO, whether a laser threshold could be overcome, so that one could observe lasing at some of the strongest lines shown in Fig. 2. For lasing, the spectral intensities are the important quantities. Their UV/Vis ratio is a factor of ~ 40 in Fig. 2.

Summary and conclusions

A newly observed phenomenon of unexpectedly strong UV emission is observed from an RF excited supersonically flowing discharge afterglow in a gas mixture of He–C2F4–NO. The dominant emitter is NO(A2Σ+, v′-X, v′′), which emits 68% of the radiation over the range of 190–300 nm of the whole investigated spectral range of 190–710 nm. This emitter gives ~ 40 times higher intensities than any other emitter in the monitored spectral range. The NO molecules are excited by energy transfer from electronically excited He and an excited fluorocarbon species, probably triplet CF2. Any presence of O2 proves to be harmful for UV emission intensity. This strong UV emission has some potential for use for a new UV gas-flow laser.