Characterization of an Atmospheric-Pressure Argon Plasma Generated by 915 MHz Microwaves Using Optical Emission Spectroscopy

The paper presents the investigations of an atmospheric-pressure argon plasma generated at 915MHzmicrowaves using the optical emission spectroscopy (OES). The 915MHz microwave plasma was inducted and sustained in a waveguide-supplied coaxial-linebased nozzleless microwave plasma source. The aim of presented investigations was to estimate parameters of the generated plasma, that is, excitation temperature of electrons Texc, temperature of plasma gas Tg, and concentration of electrons ne. Assuming that excited levels of argon atoms are in local thermodynamic equilibrium, Boltzmann method allowed in determining the Texc temperature in the range of 8100–11000K. The temperature of plasma gas Tg was estimated by comparing the simulated spectra of the OH radical to the measured one in LIFBASE program. The obtained Tg temperature ranged in 1200–2800K. Using a method based on Stark broadening of the Hβ line, the concentration of electrons ne was determined in the range from 1.4× 10 to 1.7× 10 cm, depending on the power absorbed by the microwave plasma.

Since in process of the gas treatment by the plasma, the temperature of plasma gas and concentration of electrons play an important role; therefore, the knowledge of these basic parameters is crucial for understanding the chemical kinetics and its optimization.

Experiment
In Figure 1(a), a photo of the MPS is shown, whereas a draft of an experimental setup is presented in Figure 1(b).The MPS is supplied via a standard waveguide WR 975 and ended by a movable plunger which allows for an effective transfer of the microwave power from the electric field to the plasma.
Inside the used MPS, a quartz tube is placed.In the tube, two gas flows are formed.The first one is the axial flow (processed gas), and the second one is the swirl flow (cooling gas).In the axial flow, the processed gas is introduced into the quartz tube through the inner electrode.In the swirl flow, the cooling gas is delivered into the MPS by four inlets located tangentially to the quartz tube wall [1].This resulted in a vortex (swirl) flow inside the quartz tube.In this type of the MPS, plasma in a form of flame occurs inside the quartz tube at the tip of the inner electrode.The additional swirl flow stabilized the discharge in the center of the quartz tube and protects the quartz wall from overheating [1].Below the MPS waveguide, the quartz tube is surrounded by a metal cylinder with a vertical slit for the observation of the generated discharge.
Symmetrical double convex quartz lens (50 mm in diameter, focal length 75 mm) was used to focus light emitted by the microwave plasma.Additionally to collimate the emitted light, two diaphragms with pinholes of 1 mm in diameter were placed.In these investigations, the spectrum of the microwave plasma was measured by McPherson model 209 spectrometer, equipped with double-pass scanning monochromator.The used spectrograph is equipped with sensitivity-calibrated iCCD camera and diffraction grating 1200 grooves/mm.
Using Hg I lines λ = 365.02nm, 435.84 nm, and 546.07 nm emitted from low-pressure calibration Hg-Ne lamp, the instrumental line broadening Δλ I has been determined.The obtained values of Δλ I was about 0.07 nm.

Results
During the experiment, the nitrogen was used as a cooling gas with constant flow rate of 50 NL/min.The investigations were performed with argon as processed gas with flow rate equal to 50 NL/min.The power absorbed by the microwave plasma P A was calculated as a difference between power incident P I and power reflected P R in the MPS [1].The P I and P R were directly measured using a directional coupler.In these investigations, the power P A was changed from 2 to 4 kW.
The spectra were recorded 15 mm below the tip of the inner electrode.In our measurements, we focused on the range of emission spectra from 300-600 nm.An example of the recorded spectra is shown in Figure 2. To detect emission spectral lines of the H or the OH radicals, a small addition of H 2 O vapour (H 2 O-0.1 kg/h, H 2 O vapour temperature was equal to 400 °C) was added to the process gas flow.
In performed investigations, we assumed that microwave plasma at atmospheric pressure is generally in partial local thermodynamic equilibrium [13,14,[18][19][20].This assumption allowed us to use the Boltzmann plot method to determine the T exc [14,15].Five transition lines of argon (see Figure 2) were selected to determine the T exc .Selected argon lines with the parameters for the Boltzmann plot method were presented in Table 1.An example of Boltzmann plot is shown in Figure 3. Conformity of a straight line with experimental points indicates balance in the excited states of argon atoms.The obtained T exc was in the range of 8100-11000 K, as shown in Figure 4.The estimated T exc temperature increased with increasing the absorbed microwave power P A .
It is widely accepted that in the microwave discharges, the rotational temperature of the OH radical T rot corresponds to the translational temperature of heavy particles in the plasma (temperature of plasma gas T g ) [16,17].
To obtain the molecule rotational temperature, the OH band A 2 Σ + → X 2 Π was used.This band is very sensitive against the changes of the rotational temperature [16].After measuring the OH spectrum, we simulated this band in the LIFBASE program [21].This program allows calculating the emitted spectrum of plasma radiation of various gases at individually given rotational and vibrational temperatures.In this program, the simulated OH band was fitted to the experimental one (Figure 5).A good agreement has been found.Spectrum simulations were performed for Gaussian line shapes with a FWHM value equal to  Journal of Spectroscopy 0.07 nm.The obtained gas temperature T g ranged from 1200 to 2800 K (Figure 6).By using the method based on Stark broadening of the hydrogen H β line, the concentrations of electrons n e in the plasma were determined [13][14][15]17].The introduction of water vapour caused the emergence of emission lines of hydrogen H β , H δ , and H γ .In our work, we focus only on the H β (486.13 nm) line.The H δ and H γ lines were hardly noticeable or partially overlapped by the argon lines.Therefore, these two lines were not used to determine the concentration of electrons n e .
The shape of the recorded H β line is affected by several different mechanisms of broadening (instrumental Δλ I , Van der Waals Δλ W , Stark Δλ S , resonance Δλ R , Doppler Δλ D , and natural Δλ N ) which result to a Voigt profile [13][14][15].In order to obtain the FWHM of H β line in investigations to the measured profile, the Voigt function was fitted.The fitting was performed using the Origin software [22].
The Doppler broadening Δλ D is a result of atoms' random motions in the plasma.This effect can be calculated from [15,23] Δλ D = 7 17 × 10 −7 λ 0 T M , 1    where λ 0 is the wavelength, T is the temperature of the emitter in Kelvins, M is the mass of the emitter in a.m.u.In this work, T was assumed equal to the temperature of plasma gas T g .The Van der Waals broadening Δλ W is the effects of dipolar interaction between excited the atoms and the neutral ground state atom [14].The Δλ W broadening can be estimated from [23] Δλ W = 6 48 × 10 −22 p kT 0 7 g , 2 where p is the pressure and k is the Boltzmann constant.Determination of plasma gas temperature T g allows to estimate values of Van der Waals and Doppler broadening effect.Using the above formulas, the values of Δλ D and Δλ W broadening of the H β line were calculated.The obtained value of Δλ D was equal to 0.003 nm while Δλ W = 0.02 nm, respectively.In the tested range of the absorbed microwave power P A , determined values were constant.In the presented work, resonance and natural broadening have been omitted due to low FWHM values in comparison to the other effects [13][14][15].Taking into account the estimated values of Δλ I , Δλ D , and Δλ W and obtained value FWHM of H β line, the Stark broadening Δλ S was calculated [15,23].In the experiment, we observed a linear relationship between the estimated value of Stark broadening Δλ S of H β line and the absorbed microwave power P A by the plasma.In calculation of the n e , a Gig-Card theory [24] was used.The measured concentration of electrons n e ranged from 1.4 × 10 15 to 1.7 × 10 15 cm −3 (Figure 7).The values of the electron concentration indicate that the balance between electrons and heavy particles as a result of collisions cannot be achieved.Thus, the plasma cannot be described by a single temperature.
Adopting a classical ideal gas model and using the measured concentration of electrons, we estimated that the ionization degree in plasma was about ~10 −4 .This indicates that ionization degree is too low to thermalize the electron energy distribution function.Therefore, there may be a lack of balance between the basic state and the excited states of argon atoms in plasma.This cause that the measured temperatures could be overestimated.In measurements, we record the radiation from the excited states, which are the result of collisions, while the basic states remain neutral.

Conclusions
The investigations of an atmospheric-pressure argon plasma generated at 915 MHz microwaves using optical emission  4 Journal of Spectroscopy spectroscopy (OES) are presented in this work.These investigations yielded the excitation temperature of electrons T exc , the gas temperature T g , and the concentration of electrons n e in the generated argon plasma.In the tested range of the absorbed microwave power P A by the plasma, we observed an increase in the excitation temperature T exc , the gas temperature T g , and the concentration of electrons n e .These results indicate that appropriate selection of the gases and the operating parameters of the MPS (central and the additional flow rate, absorbed microwave power) enables in obtaining the plasma with desired parameters.It should also be mentioned that the investigated MPS works very stable with various processing gases (argon, nitrogen, air, and carbon dioxide) at high flow rates and absorbed microwave power by the plasma can be changed in a wide range.Thus, the above properties make the presented MPS an attractive tool for different gas processing at high flow rates.
Gig-Card theory

Figure 1 :
Figure 1: (a) Photo of the microwave plasma source and (b) the experimental setup used for the spectroscopic investigations.

Figure 2 :
Figure 2: Measured emission spectra of argon plasma (with a small amount of H 2 O vapour) with selected argon 4p-4s and 5p-4s transition lines for Boltzmann plot method.Absorbed microwave power P A = 3 kW and argon flow rate Q = 50 l/min.

Figure 3 :
Figure 3: Example of Boltzmann plot for determination of the T exc .I nm -intensity of the recorded emission line from transition n → m, microwave power absorbed P A = 3 kW, and argon flow rate Q = 50 l/min.

Figure 4 :Figure 5 :Figure 6 :
Figure 4: Electron excitation temperature as a function of absorbed microwave power P A .

Figure 7 :
Figure 7: The concentration of electrons as a function of absorbed microwave power P A .

Table 1 :
Parameters of selected argon emission lines used to determine the excitation temperature of the electrons T exc .n/m:energy levels upper/lower, respectively; λ nm : wavelength of transition n → m, A nm : Einstein coefficient for transition n → m, g n : statistical weight of the upper level n.E n is the energy of the upper level n.l.p. λ nm (nm) Transition A nm (10 7 s −1 ) g n E n (cm −1 ) nm λ nm /g n A nm )