Generation of High-Density Pulsed Gas–Liquid Discharge Plasma Using Floating Electrode Configuration at Atmospheric Pressure

Abstract

In this paper, a high-density gas–liquid discharge plasma is obtained combined with nanosecond pulse voltage and a floating electrode. The discharge images, the waveforms of pulse voltage and discharge current, and the optical emission spectra are recorded. Gas temperature and electron density are calculated by the optical emission spectra of N₂ (C³Π_u → B³Π_g) and the Stark broadening of H_α, respectively. The emission intensities of N₂ (C³Π_u → B³Π_g), ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π), OH (A²Σ → X²Π), O (3p⁵P → 3s⁵S⁰), He (3d³D → 3p³${\mathrm{P}}_2^0$), gas temperature, and electron density are acquired by optical emission spectra to discuss plasma characteristics varying with spatial distribution, discharge gap, and gas flow rate. The spatial distributions of discharge characteristics, including gas temperature, electron density, and emission intensities of N₂ (C³Π_u → B³Π_g), ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π), OH (A²Σ → X²Π), O (3p⁵P → 3s⁵S⁰), and He (3d³D → 3p³${\mathrm{P}}_2^0$), are presented. It is found that a high-density discharge plasma with the electron density of 2.2 × 10¹⁵ cm⁻³ and low gas temperature close to room temperature is generated. While setting the discharge gap distance at 10 mm, the discharge area over liquid surface has the largest diameter of 20 mm; under the same conditions, electron density is in the order of 10¹⁵ cm⁻³, and gas temperature is approximately 330 K. In addition, the discharge plasma characteristics are not kept consistent in the axial direction, in which the emission intensities of ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π), N₂ (C³Π_u → B³Π_g), OH (A²Σ → X²Π), and gas temperature increased near the liquid surface. As the discharge gap is enlarged, the gas temperature increases, whereas the electron density remains almost constant. Moreover, as the gas flow rate was turned up, the electron density increased and the gas temperature was kept constant at 320 K.

1. Introduction

Atmospheric pressure (AP) gas–liquid discharge plasma has been actively investigated for its promising application prospects in water purification [1,2,3], biomedicine [4,5,6], and materials processing [7,8,9]. However, random micro discharge channels occur easily which can lead to non-directed energy distribution leading to serious issues and adversely affecting industrial applications. Hence a stable gas–liquid discharge at AP is expected. Over the last two decades, various factors have been studied for improving discharge stability, including the electrode structures [10,11], the driving power supply [12,13], the working gases characteristic [14,15], and many other experimental parameters. In addition, the plasma characteristic could be different even under the same conditions [16], because it is influenced by electron transport. Thus, it is necessary to discuss the mechanism that leads to the change of the discharge mode, such as the glow to arc transition (GAT), which is a common result of thermal and electronic instability. Both of those instabilities are intensified by positive feedback via Joule heating [17]; more specifically, a slight increase in temperature can lead to a significant increase in vibrational–translational relaxation rate, which improves gas temperature further via their exponential relationship [18]. In addition, the heat release happens in a small volume that allows thermal instability to develop rapidly.

Nanosecond pulse voltage can be used to generate stable non-thermal plasma, where the fast-rising time can prevent the GAT effectively. Since the rising time of nanosecond pulse voltage is shorter than the character time of gas heating (in microsecond scale) [19], the increase in gas temperature is not obvious. When the pulse voltage is applied, the rapidly rising voltage pulse induces a high electric field, which promotes the occurrence of non-equilibrium plasma [20], and the electrical energy delivered in the plasma discharge is mainly deposited in the energetic electrons instead of heating the heavy particles. Therefore, the nanosecond pulse voltage can be employed to generate a stable gas–liquid discharge with high electron density and low gas temperature [21,22]. Herrmann et al. [23] studied the influences of peak voltage and discharge gap distance on pulsed discharge behavior over the liquid surface. Gromov et al. [24] reported the physical process at the gas–liquid interface in a single pulse mode and high-frequency burst. Xu et al. [25] compared the pulsed discharge plasma and sinusoidal discharge plasma, the results showing that nanosecond pulsed discharge plasma has lower gas temperature and higher efficiency of bacterial inactivation in liquid than that of the sinusoidal discharge plasma. Wang et al. [26] used bipolar nanosecond pulse power to obtain a gas–liquid discharge plasma with electron density in the order of 10¹⁵ cm⁻³ and gas temperature of 350 K. They studied it further and reported in [27] that the electron density can be elevated to the order of 10¹⁷ cm⁻³ for the transient to spark discharge, with a gas temperature of 380 K.

In addition, the floating electrode (FE), as a kind of independent structure between electrodes, shows the potential to improve the discharge plasma characteristics [28,29]. As Hu et al. [30] reported, the FE can influence the place of initial current density increase. Lee et al. [31] compared electric field intensity distributions in a plasma jet with and without the inner-FE. The results show that the removal of FE improves plasma reactivity and lowers gas breakdown voltage. Liu et al. [32] developed a diffuse-like discharge and studied the spatial electric field distribution when the FE is set. The report shows that the spatial electric field near the FE is advantageous for generating uniform-density plasma. But there are few studies on the gas–liquid discharge when FE is employed.

In this paper, with the employment of bipolar nanosecond pulse voltage, a kind of new electrode structure is used to attempt to obtain a stable discharge over the liquid surface. The electron structure is composed of a floating electrode and a hollow high voltage electron. A series of studies on the floating electrode gas–liquid discharge (FEGLD) is presented in this paper, including discharge images, electrical characteristics, optical emission spectra, plasma gas temperatures, and electron density. The spatial distribution characteristics of FEGLD are also carried out. Moreover, the interaction area between high density and low gas temperature at AP is expected to be enlarged by adjusting the discharge operation conditions. The influence on FEGLD by the discharge gap distance and gas flow rate is described in detail.

2. Experiment Setup

Figure 1a illustrates the experiment setup of FEGLD, comprising a discharge reactor, a nanosecond pulse power supply, the electrical characterization measurement system, and the optical emission spectra (OES) detection system. He (99.999%) is used as work gas. Figure 1b shows the diagram of the discharge reactor in which a stainless-steel hollow electrode inserts into PTFE. The floating electrode is fixed 2 mm below the end of the high voltage electrode. The bipolar nanosecond pulse power supply (DGM-40, 100 MHz, Dalian Power Supply Technology Co., Ltd., Dalian, China) provides pulse peak voltage which is adjustable in the range of 0–60 kV. A high voltage probe (Tektronix P6015A 1000×, 3.0 pF, 100 MΩ) and a current probe (Tektronix TCP312, bandwidth 100 MHz) are used to measure pulse voltage and discharge current, which are displayed on the oscilloscope (Tektronix, TDS5054B, 500 MHz). These compose the electrical characterization measurement system. The optical fiber and grating monochromator (Andor SR750i) are employed as the OES detection system.

3. Result and Discussion

3.1. Visualization Characteristics of FEGLD Plasma

Since the plasma and liquid surface interaction area could be varied by changing the discharge gap, Figure 2 shows the discharge images of FEGLD at the discharge gap of 10 mm, 13 mm, and 15 mm, captured by a Canon 550D digital camera with 500 ms exposure time, under the measurement conditions of 3.5 L/min gas flow rate and 30 kV pulse peak voltage. As Figure 2 shows, when the discharge gap is set at 10 mm, there is a discharge area over the liquid surface with the diameter larger than 20 mm, and no flames or unstable discharge model transition can be observed. In the middle of this, a mean discharge area has the most intense emission intensity. While enlarging the discharge gap distance from 10 mm to 15 mm, as shown in the figure, there is an obvious contraction in the discharge area over the liquid surface.

3.2. Spectral Characteristics of FEGLD

The OES emitted from the FEGLD are illustrated in Figure 3, under the experimental conditions of 30 kV pulse peak voltage, 12 mm discharge gap, and 3.5 L/min gas flow rate. The OES of FEGLD is mainly composed of the second positive system of N₂ (C³Π_u→B³Π_g), the first negative system ${\mathrm{N}}_2^{+}$ (B²${\mathsf{\Sigma }}_{\mathrm{u}}^{+}$→X²${\mathsf{\Sigma }}_{\mathrm{g}}^{+}$), the bands of OH (A²Σ→X²Π), and the lines of He, H_α and O. The FEGLD is operated into open air filled with N₂, O₂, and H₂O originating in liquid surface. In He discharge, the metastable helium is easily quenched by those molecules. These processes are shown as R(1)–R(3) in

Table 1. Related reactions for active species in FEGLD.

NO.	Reaction	Rate Constant	Ref.
1	${\mathrm{He}}^{*}+{\mathrm{N}}_2\to {\mathrm{N}}_2^{+}+\mathrm{He}+\mathrm{e}$	7.0 × 10−11 cm3 s−1	[35]
2	${\mathrm{He}}^{*}+{\mathrm{O}}_2\to {\mathrm{O}}_2^{+}+\mathrm{He}+\mathrm{e}$	2.6 × 10−16 cm3 s−1	[36]
3	${\mathrm{He}}^{*}+{\mathrm{H}}_2\mathrm{O}\to {\mathrm{He}+\mathrm{OH}}^{+}+{\mathrm{H}}^{+}+\mathrm{e}$	1.5 × 10−10 cm3 s−1	[37]
4	${\mathrm{e}+\mathrm{N}}_2(\mathrm{X})\to {\mathrm{e}+\mathrm{N}}_2(\mathrm{A})$	1.1 × 10−10 cm3 s−1	[33]
5	${\mathrm{N}}_2{(\mathrm{A})+\mathrm{N}}_2(\mathrm{A})\to {\mathrm{N}}_2{(\mathrm{C})+\mathrm{N}}_2(\mathrm{X})$	8.0 × 10−11 cm3 s−1	[33]
6	${\mathrm{e}+\mathrm{N}}_2(\mathrm{A})\to {2\mathrm{e}+\mathrm{N}}_2^{+}$	2.4 × 10−12 cm3 s−1	[34]
7	${\mathrm{e}+\mathrm{O}}_2\to \mathrm{e}+\mathrm{O}+\mathrm{O}$	3.2 × 10−11 cm3 s−1	[34]
8	${\mathrm{e}+\mathrm{H}}_2\mathrm{O}\to \mathrm{e}+\mathrm{H}+\mathrm{OH}$	2.6 × 10−12 cm3 s−1	[34]
9	${\mathrm{OH}(\mathrm{A})+\mathrm{O}}_2(\mathrm{X})\to {\mathrm{HO}}_2(\mathrm{X})+\mathrm{O}({}_{}{}^3\mathrm{P})$	——	[38]

* Metastable state.

. While the pulse voltage is employed, the production of electrons via Penning ionization sustains the plasma between two discharge pulses [14]. The N₂ (C) state could be produced by the ground state of nitrogen molecules N₂ (X), as shown in R(4)–R(5). It should be mentioned that R(5) could release additional energy [18] contributing to temperature improvement as the gas fast heating mechanism. The first negative system ${\mathrm{N}}_2^{+}$ (B²${\mathsf{\Sigma }}_{\mathrm{u}}^{+}$→X²${\mathsf{\Sigma }}_{\mathrm{g}}^{+}$) in FEGLD exhibited high emission intensity, as shown in Figure 3, the physicochemical process being shown in R6. Besides, the non-elastic collisions of the energetic electrons with N₂, O₂, and H₂O molecules will induce its dissociation, excitation, and ionization [33]. The process of energetic electrons created by the discharge pulse colliding with O₂ [34] is shown in R7. The physicochemical process of OH generation is shown in R8. Meanwhile, when the surrounding gas contains O₂ and H₂O, the collisional quenching of OH (A) by O₂ is an important process, as shown in R9.

3.3. Electrical Characteristics of FEGLD

The distinction in discharge characteristics can be observed at pulse voltage duration times of positive and negative. Hence, the waveforms of the discharge current and topical voltage in positive pulse voltage and negative pulse voltage are presented in Figure 4a,b, under the experimental condition of 12 mm discharge gap, 30 kV pulse peak voltage, and 3.5 L/min gas flow rate. Figure 4a shows that the discharges mainly occur at the rising edge of positive pulse voltage, which has the rising time of about 250 ns. When the gas is broken down, one discharge current peak with 1.5 μs pulse width can be observed. In addition, three burr-like current peaks appear at the rising edge of the main discharge current peak. Each of those burr-like current peaks has the width of 50 ns. The characteristic of discharge current waveforms in negative pulse voltage is similar to that in positive pulse voltage, except their opposite polarity. In addition, there are a number of oscillations and a long plateau for voltage in each pulse. It is due to the power supply and capacitive loads in FEGLD that the charge accumulation at the plasma–liquid interface reduced the space charge loss, and sustained discharge [39].

3.4. Gas Temperature of FEGLD

As a quantitative indicator, the gas temperature of FEGLD is acquired and discussed. Since the small energy gap between the rotational levels of the molecules, the translational–rotational equilibrium is readily achieved by frequent collisions between the heavy particles at AP [40]. The gas temperature can be considered approximately equal to the rotational temperature, and it can be acquired by analyzing the rotational spectra of excited molecular species [41,42]. In this work, the N₂ (C³Π_u → B³Π_g, Δν = −2) is chosen to measure rotational temperature, because there is no overlap with OES emitting from He and other trace impurities originating from air, such as H and OH [43]. In this work, the Specair code [44] can be used to obtain the rotational temperature of N₂ (C³Π_u → B³Π_g, Δν = −2) by comparing the simulated spectra and experimental spectra. As shown in Figure 5, under the experimental conditions of 3.5 L/min gas flow rate, 30 kV pulse peak voltage, and 12 mm discharge gap, the rotational temperature of N₂ (C³Π_u → B³Π_g, Δν = −2) is 320 K, that is, close to the room temperature.

3.5. Electron Density of FEGLD

The hydrogen line broadening effects of electron density on Stark broadening is one of the most widely used techniques for electron density measurement [45]. Both H_α and H_β could be used for the calculation. In this work, the H_α is preferred because of its high line shape sensitivity and adequate emission intensity. The emission line of H_α is the convolution of the Gaussian and Lorentzian lines. The Gaussian profile was determined by both the Doppler and instrument broadening, in which the fullwidth at half maximum (FWHM) of the Gaussian is given by:

$$\Delta {\mathsf{\lambda }}_{\mathrm{G}}=(\Delta {\mathsf{\lambda }}_{\mathrm{D}}^2+\Delta {\mathsf{\lambda }}_{\mathrm{I}}^2{)}^{\frac{1}{2}}$$

(1)

where Δλ_D and Δλ_I are the FWHM of the Doppler broadening and instrumental broadening, respectively.

Doppler broadening mainly depends on the temperature of the emitter [46]. The FWHM of the Doppler broadening is given by the expression [47]:

$$\Delta {\mathsf{\lambda }}_{\mathrm{D}}={7.16\times 10}^{-7}\mathsf{\lambda }(\frac{{\mathrm{T}}_{\mathrm{g}}}{\mathrm{M}}{)}^{\frac{1}{2}}$$

(2)

where λ is the emission wavelength, T_g is the gas temperature of plasma in K and M is the atomic weight of hydrogen atoms in g·mol⁻¹.

The broadening induced by the detection system is called apparatus function or instrumental broadening [41]. The instrument broadening in this work is determined by measuring the FWHM of the mercury lamp characteristic spectrum at 435.8 nm. Its value is given as:

$$\Delta {\mathsf{\lambda }}_{\mathrm{I}}=0.078\mathrm{nm}$$

(3)

while grating 1200 lines·mm⁻¹ is used.

Lorentzian profiles are mainly determined by the Stark, van der Waals, resonance, and natural broadening, and the expression could be given by:

$$\Delta {\mathsf{\lambda }}_{\mathrm{L}}=\Delta {\mathsf{\lambda }}_{\mathrm{V}}+\Delta {\mathsf{\lambda }}_{\mathrm{S}}+\Delta {\mathsf{\lambda }}_{\mathrm{R}}+\Delta {\mathsf{\lambda }}_{\mathrm{N}}$$

(4)

Natural broadening is caused by the uncertainty principle of atomic stable state energy, and the expression could be given by:

$$\Delta \mathrm{E}=\frac{\mathrm{h}}{\Delta {\mathsf{\lambda }}_{\mathrm{N}}}=\frac{\mathrm{h}}{2\mathsf{\pi }\mathsf{\tau }}$$

(5)

where Δ$\mathrm{E}$ is uncertainty of the energy level, h is Planck constant, τ is energy level lifetime in the order of 10⁻⁹. The FWHM of natural broadening could be expressed as:

$$\Delta {\mathsf{\lambda }}_{\mathrm{N}}=2\mathsf{\pi }\mathsf{\tau }$$

(6)

The value of natural line broadening is very small by calculation, in the orders of magnitude of 10⁻⁵ nm, which can be negligible in this study [48].

Resonance broadening occurs by electric dipole of the upper- or the lower-level transit to the ground state. The formula for the FWHM of resonance broadening could be written as follows [47,49]:

$$\Delta {\mathsf{\lambda }}_{\mathrm{R}}={8.61\times 10}^{-30}{\mathsf{\lambda }}^2{\mathsf{\lambda }}_{\mathrm{r}}{\mathrm{f}}_{\mathrm{r}}{\mathrm{N}}_{\mathrm{i}}{(\frac{{\mathrm{g}}_{\mathrm{i}}}{{\mathrm{g}}_{\mathrm{k}}})}^{\frac{1}{2}}$$

(7)

where λ is the wavelength of the spectral line, f_r is the absorption oscillator strength, which is 0.6407 when the transition is 2s–3p for H_α line [50], λ_r is the wavelength of the resonance line; g_k and g_i are the statistical weights of its upper and lower levels, which are 8 and 18, respectively [49]. N_i is the ground state number density. Since FEGLD operated in the open air with low H₂O concentration, the discharge happened on the liquid surface with low temperature, which is not enough to boil the water to improve the H₂O concentration highly. The resonance broadening should be out of consideration in this work.

Van der Waals broadening is generated by collisions of excited H atoms with another kind of species, which is the helium atom in this work. An approximate formula for the FWHM of it is [51]:

$$\Delta {\mathsf{\lambda }}_{\mathrm{V}}={8.18\times 10}^{-26}{\mathsf{\lambda }}_0^2{({\overline{\mathrm{R}}}^2)}^{\frac{2}{5}}{\mathrm{T}}_{\mathrm{g}}^{\frac{3}{10}}\mathrm{N}{{\displaystyle \sum }}_{\mathrm{i}}(\frac{{\mathsf{\alpha }}_{\mathrm{i}}^{\frac{2}{5}}{\mathsf{\chi }}_{\mathrm{i}}}{{\mathsf{\mu }}_{\mathrm{i}}^{\frac{3}{10}}})$$

(8)

where λ₀ is wavelength, μ is the reduced mass, N is the neutral particle density, i is He and χ the fraction of the perturber. Hofmann et al. [51] calculated and reported the experiential expressions of the van der Waals broadening for different gas compositions at atmospheric pressure, and the formula can be shown as follows, where the C can be given when discharge plasma operates with different work gases.

$$\Delta {\mathsf{\lambda }}_{\mathrm{V}}=\frac{\mathrm{C}}{{\mathrm{T}}_{\mathrm{g}}^{\frac{7}{10}}}$$

(9)

In this work, the value of C is 2.42, which considers the plasma operating in air with helium as the work gas.

Stark broadening has high sensitivity to electron density. The relationship between the FWHM of the hydrogen alpha line and electron density is shown as follows [51]:

$$\Delta {\mathsf{\lambda }}_{\mathrm{S}}=1.78\times (\frac{{\mathrm{n}}_{\mathrm{e}}}{{10}^{23}{\mathrm{m}}^{-3}}{)}^{\frac{2}{3}}$$

(10)

The line shapes of experimental and simulated H_α broadened by the above broadenings are shown in Figure 6. The values of FWHM for all the broadening mechanisms are presented in

Table 2. Broadening of FEGLD plasma in H_α.

Broadening Effect	Value (nm)
Doppler broadening	0.008
Instrument broadening	0.078
Van der Waals broadening	0.044
Natural broadening	10−5
Stark broadening	0.156

. The electron density is calculated to be about 2.2 × 10¹⁵ cm⁻³ under the conditions of 30 kV pulse peak voltage, 12 mm discharge gap, and 3.5 L/min gas flow rate.

3.6. Spatial Distributions Characteristic of the FEGLD

For understanding the characteristic variation of discharge plasma in space, the spatially resolved emission intensity distributions of N₂ (C³Π_u → B³Π_g), ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π), OH (A²Σ → X²Π), O (3p⁵P → 3s⁵S⁰), and He (3s³S → 2p³P⁰) of FEGLD are shown in Figure 7, under the experimental conditions of 12 mm discharge gap, 30 kV pulse peak voltage, and 3.5 L/min helium flow rate. In this part, the OES emitted from the different parts of FEGLD is obtained by optical fiber which is placed on the adjustable micrometer screw lifting platform.

The inconsistent spatial distributions of FEGLD reactive species emission intensities along the axial direction can be found in Figure 7. In the range of 0–10 mm, the emission intensities of N₂ (C³Π_u → B³Π_g), ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π), and OH (A²Σ → X²Π) remains constant. But in the range of 10–12 mm which is near to the liquid surface, the emission intensities of ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π), N₂ (C³Π_u → B³Π_g), and OH (A²Σ → X²Π) show obvious increases and attain a peak value at 12 mm. Meanwhile, the emission intensities of He (3s³S → 2p³P⁰) and O (3p⁵P → 3s⁵S⁰) keep decreasing in the range of 0–12 mm. This is because there are differences in the excitation mechanism of radiative states. The formation of He (3s³S → 2p³P⁰) and N₂ (C³Π_u → B³Π_g) is dominated by direct excitation by energetic electrons [52], which relies on the high electric field and the electron density. But near the liquid surface, there is no obvious increase in electron density which is clearly shown in Figure 8. According to the research in [53], a mushroom-like shape discharge area is generated due to the interaction with a liquid surface. The N₂ molecule from ambient air is involved in the discharge area. Thus, the emission intensity of N₂ (C³Π_u → B³Π_g) increased here, and helium diffused here leading to the emission intensity of He (3s³S → 2p³P⁰) decreasing. The significant elevation of ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π) emission intensity comes from the Penning process [52], in which the increase of nitrogen will induce the improvement of ${\mathrm{N}}_2^{+}$ (B²${\mathsf{\Sigma }}_{\mathrm{u}}^{+}$)_u generation. The formation of OH (A) via direct dissociative electron excitation of water or dissociative recombination of H₂O⁺ or H₃O⁺ [54] The increase of OH (A²Σ → X²Π) emission intensity near the liquid surface is also shown in [52]. The H₂O⁺(A) ion could exist in a mushroom-like shape area due to its relatively long radiative lifetime (about 10 µs), so the increase of OH (A²Σ → X²Π) emission intensity can be observed near the liquid surface.

As the important parameters, gas temperature and electron density influence the gas–liquid discharge application. The spatially resolved gas temperature and electron density of FEGLD are shown in Figure 8 when the pulse peak voltage, gas flow rate, and discharge gap are kept at 30 kV, 3.5 L/min, and 12 mm, respectively. It clearly shows that the electron density remains at 2.2 × 10¹⁵ cm⁻³, almost without changes in the axial direction. The gas temperature remains at 320 K in the range of 1–12 mm, and reaches maxima of 340 K at 12 mm which is near the liquid surface. The increase in gas temperature near the liquid is also shown by Verreycken et al. [55] who attributed it to the fast vibration-to-vibration (V–V) process of O₂–H₂O and N₂–H₂O and, at the same time, H₂O’s higher vibration-to-translation (V–T) rate compared to O₂ and N₂, which means that as the H₂O concentration increases, additional gas heating occurs. One et al. [56] found that the temperature will be elevated while water vapor concentration increases. Moreover, since the inelastic collisions between electrons and heavy particles induce the plasma chemistry [57], which could be have potential applications in the treatment of living human and animal tissue [58], for instance, the high electron density plasma with approximately room temperature at AP is expected to meet temperature-sensitive needs.

3.7. The Effects of Discharge Gap on FEGLD

Considering the space compatibility during actual operation, the effect of the discharge gap on active species, the gas temperature, and the electron density are discussed. Figure 9 shows the effects of the discharge gap on the emission intensities of N₂ (C³Π_u → B³Π_g), ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π), OH (A²Σ → X²Π), O (3p⁵P → 3s⁵S⁰), and He (3s³S → 2p³P⁰) of the FEGLD under the measurement condition of 30 kV plus peak voltage and 3.5 L/min gas flow rate. It clearly shows that the emission intensities of N₂ (C³Π_u → B³Π_g), ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π), OH (A²Σ → X²Π), and O (3p⁵P → 3s⁵S⁰) remain almost constant as the discharge gap is enlarged from 10 mm to 15 mm. But in the same conditions, the emission intensity of He (3s³S → 2p³P⁰) decreases while the discharge gap changes from 10 mm to 13 mm. Since the emission spectrum of He (3s³S → 2p³P⁰) is an indicator of energetic electrons with relatively high threshold energy of 22.7 eV [59], the decrease of emission intensity of He (3s³S → 2p³P⁰) reflects the change in electron energy distribution, which is controlled by the applied reduced field E/N [60]. As the discharge gap is enlarged, the increase in the value of pressure multiplied by electrode gap distance led to the fall in the reduced electric field, which induces the energy of electrons to decrease [61], and then the emission intensity of He (3s³S → 2p³P⁰) is obviously decreased here.

The effects of the discharge gap on gas temperature and electron density of FEGLD are illustrated in Figure 10 under the measurement conditions of 30 kV plus peak voltage and 3.5 L/min gas flow rate. As shown in the figure, the electron density shows almost no change as the discharge gap increases. However, the gas temperature increased from 280 K to 350 K with the discharge gap elevated from 10 mm to 14 mm. Since the optical fiber is aimed at the medium area of FEGLD, it means this increase in gas temperature happened in the middle of the discharge plasma. Considering Figure 2, which shows clearly that the diameter of the discharge area near the liquid surface decreased with the discharge gap enlarging, the mushroom-like shape discharge, which is around the mean middle discharge, is not obvious. Therefore, the cooling effect is limited, which is expressly significant for the helium gas with high thermal conductivity at 1.6 × 10³ J cm⁻¹ s⁻¹ K⁻¹. Hence, in this work, the weakness of the cooling effect around the middle discharge area leads to the increase in gas temperature.

3.8. The Effects of the Helium Flow Rate on FEGLD

Since the gas flow rate influences the time for gas stay, it further affects temperature transition and discharge characteristics. The effect of helium flow rate on the emission intensity of N₂ (C³Π_u → B³Π_g), ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π), OH (A²Σ → X²Π), O (3p⁵P → 3s⁵S⁰), and He (3s³S → 2p³P⁰) are obtained and shown in Figure 11, under the experimental conditions of 12 mm discharge gap and 30 kV pulse peak voltage. It can be seen that as the gas flow rate is turned up, the emission intensities of ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π), He (3s³S → 2p³P⁰), and O (3p⁵P → 3s⁵S⁰) increased gradually. Meanwhile, the reduction in emission intensity of N₂ (C³Π_u → B³Π_g) and OH (A²Σ → X²Π) can be observed. In the study by Yue et al. [62], OH distribution with different gas flow rates is reported. As the gas flow rate is increased, the OH distribution trends to uniform; meanwhile, the peak intensity value of OH-LIF is getting weaker. The influence of gas flow rate on O-LIF is also reported; it is due to the O (3p⁵P) having generated a relatively long lifetime of up to milliseconds [63] that the distribution of atoms could be influenced by gas flow. Similarly, as discussed before, the emission intensity of ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π) is dominated by the Penning process, and the metastable helium with long lifetime is driven by gas flow in the FEGLD, which leads to emission intensities of ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π) increasing.

The effects of gas flow rate on gas temperature and electron density are illustrated in Figure 12, where the pulse peak voltage and discharge gap are kept at 30 kV and 12 mm, respectively. It clearly shows that the gas temperature of FEGLD almost remains at 330 ± 10 K when the gas flow rate varied from 0.5 L/min to 3.5 L/min, suggesting that the FEGLD has good stability. The electron density increases obviously from 2.3 × 10¹⁵ cm⁻³ to 2.7 × 10¹⁵ cm⁻³, while the helium gas flow rate increased from 0.5 L/min to 3.5 L/min, which is the most obvious consistent improvement in this work. Since gas flow could promote collision processes [34], the reactions that could generate electrons are highly promoted by the increase in the gas flow rate. Meanwhile, the increase in high energy metastable helium amount is beneficial to generating the Penning process, which provides seed electrons for discharge. As the results show, increasing the gas flow can avoid gas heating and at the same time improve electron density and the capacity that generates reactive species such as O, which is one of the most important active species in plasma medicine applications [52]. This shows the potential for meeting the needs of such temperature-sensitive applications.

4. Conclusions

In this paper, combined with nanosecond pulse voltage and a floating electrode, a stable gas–liquid discharge with electron density of 2.3 × 10¹⁵ cm⁻³ and low gas temperature of 320 K ± 20 K is obtained at AP. The OES spatial distributions show that gas temperature and emission intensities of N₂ (C³Π_u → B³Π_g), ${\mathrm{N}}_2^{+}$ (B²Σ → X²Π), and OH (A²Σ → X²Π) increase noticeably near the liquid surface. A large discharge area of 20 mm diameter is generated over the liquid surface when the discharge gap is set at 10 mm. As the discharge gap is increased further, obvious contraction is clearly shown in this area, with emission intensities of reactive species and electron density remaining constant. However, the electron density and emission intensity of reactive species such as O can be elevated by turning up the gas flow rate, whereas no obvious gas heating is observed.