Electric field and higher harmonics of RF plasma slit jet measured by antennas and VI probes

The cold atmospheric plasma jets change their character when interacting with the different surfaces. Since such interaction is the primary area of plasma jet applications, it is essential to monitor the process. The non-linearity of the RF plasma slit jet (PSJ) was analyzed using the VI probes and a novel method, the non-intrusive antenna measurements. Regardless of the experimental setup and gas mixture (Ar, Ar/O2, Ar/N2), the PSJ frequency spectrum consisted of the following main features: dominant fundamental frequency peak, relatively strong odd harmonics, and significantly weaker even harmonics. The lowest degree of non-linearity was recorded for the Ar PSJ ignited against a grounded target. Admixing a molecular gas increased the discharge non-linearity. It was attributed to the enhancement of secondary electron emission from the dielectric surfaces. In addition to the non-linearity analysis, the antenna spectra were for the first time used to determine the semi-quantitative values of the PSJ-radiated electric field. The electric fields decreased by a factor of 2 after the admixing of nitrogen and oxygen molecular gases. Out of the studied targets, the highest electric fields were observed when plasma impinged on the grounded targets, followed by the floating target (2× lower) and the PSJ ignited in the open space configuration (4× lower than in the grounded target configuration).


Introduction
The atmospheric pressure plasma discharges have been successfully utilized in many different applications, such as the treatment of polymer surfaces for improved adhesion [1][2][3][4][5], cleaning of metal surfaces [6], plasma activation of water [7,8], plasma gas conversion [9,10] or wound healing [11][12][13].This incomprehensive list demonstrates the uses of plasma in varying configurations-open space and ignited against dielectric, metal, or liquid targets.Naturally, the question of if and how the choice of the target affects the fundamental plasma properties (electric field, charge density) has arisen.In recent years, several groups have studied the interaction of cold atmospheric pressure plasma jets (APPJs) with different targets, either through modeling [14][15][16] or experimentally, by monitoring the flow patterns [17,18] and measuring plasma properties such as electric field [19][20][21] and densities of electrons and different plasma-generated species (He * , OH, O, . ..) [22][23][24].However, these studies were usually limited to single-filament He APPJs driven by a kHz power supply, with the information on the Ar or the RF-driven discharges being almost non-existent.
The characteristics of the He APPJ can significantly differ with the electrical properties of the target (conductivity, dielectric permittivity) or with the proximity of the electrical ground.Of the two options, the choice of the target material is the more impactful.Upon contact with a low-permittivity (ϵ r < 10) dielectric, the initial bulk ionization wave (IW) propagating from the powered electrode towards the target transitions into an outwards spreading surface IW (SIW).The speed and area of the SIW spreading depend on the dielectric constant, with the higher speeds and larger areas observed at the targets with the lower permittivity [14,15,25].On the contrary, the SIW does not develop when plasma interacts with a metallic target.Instead, a return stroke (reflected IW) caused by the impedance mismatch between the electrical system with the HV electrode at its end, the conductive channel, and the target is induced [14,21,23,26].The third possible target material choice, a high-permittivity (ϵ r > 10) dielectric, results in the hybrid behavior as both the SIW and the return stroke occur [14,23].The choice of the target material also affects the plasma properties.The speed of the IW propagation, the magnitude of charge density, and the electric field all increased with the increasing relative permittivity of the target [14,16].
The proximity of the electrical ground, i. e. the grounding of targets, does not significantly influence the development of the SIW or the return strokes.However, it affects the plasma parameters with the higher velocity and the intensity of IW corresponding to the grounded target.Overall, the influence of the ground on the plasma parameter is stronger than the influence of the target material, as the IW has higher velocity and intensity when the APPJ impinges on the grounded low-permittivity dielectric target than on the floating metallic target [25].
The electric field is one of the most crucial parameters of plasma discharges because its magnitude determines the production and energy of the charged particles, hence the behavior and gas chemistry of the discharge [27].Despite the importance, measurements of the electric field of the APPJs are infrequent due to the experimental difficulties.The electric field induced second harmonic generation [28,29] and coherent anti-Stokes Raman scattering (CARS) measurements [30,31] both require a ps laser.Furthermore, the CARS method is limited to the molecular species (mainly N 2 ).The methods based on optical emission spectroscopy, although less equipment demanding, are restricted to the bright part of the discharge and a specific plasma gas chemistry.The Stark polarization spectroscopy method is limited to the He-containing discharges [27,32], whereas the determination of electric field from the intensity ratio of N 2 (C-B) and N + 2 (B-X) emissions requires a valid collision-radiative model in addition to the excitation of both molecular bands [33][34][35].The measurement by electrooptic sensors based on the Pockels effect is not species-specific or overly experimentally demanding [20,36,37].Although nonintrusive and nonperturbative measurements are possible [38,39], electro-optic sensors are standardly used in contact with the plasma, which can, as discussed above, affect the plasma parameters.
In this work, we investigated the electromagnetic characteristics of the atmospheric pressure RF plasma slit jet (PSJ) ignited in argon flow to which oxygen and nitrogen gas or water aerosol were admixed.Compared to the other APPJs, the RF PSJ has a unique power coupling principle, a combination of capacitive and inductive coupling.Thus, we first used the current and voltage probe measurements to improve our understanding of the PSJ electrical character and how it is affected by the gas feed composition.The second part of the paper is dedicated to the novel plasma diagnostics method, measurement of the radiated electromagnetic field by the antennas, which was used to study the interaction of RF PSJ with different targets.The analysis of obtained frequency spectra showed that the effect of a target on the PSJ non-linearity differs with the used gas feed (Ar, Ar/O 2 , and Ar/N 2 ).The electric antenna measurements can also be used to determine the semi-quantitative values of macroscopic PSJ electric field.We demonstrate the feasibility of using the antenna measurements as a possible alternative to other experimentally more challenging methods of electric field determination.

RF PSJ and choice of substrates
The RF PSJ construction (figure 1(a)) is unique in comparison with most of the cold APPJs [40].In the PSJ, the plasma is ignited inside a mica composite slit (151 × 3 mm) placed in a specially designed coil serving as the periodic deceleration structure.This co-joined component is situated inside the tunable cavity composed of the fixed metal cover and movable conductive plates used as the resonance matching circuit.Therefore, adjusting the matching plate positions changed the discharge impedance, allowing us to match it to the RF generator 50 Ω output impedance without requiring a separate matching unit.The PSJ is designed to work at the frequency of 13.56 MHz in the width of 150 mm.The applied power of 500 or 600 W was delivered by the CESAR 136 generator (Advanced Energy).High voltage, required to sustain the discharge, is induced on the last three turns of the coil serving as a system of high-voltage (HV) high-frequency electrodes.The bottom part of the PSJ metal cover/shielding (see figure 1(a)) serves as the grounded electrodes.
The gases (and water aerosol) are injected into the PSJ through a tube with evenly distributed holes facing the topmost metal cover.This configuration results in the formation of random turbulences that ensure the intermixing of discrete flows leaving the outlets.As the gas flows downwards through the slit, it homogenizes, leading to the uniform laminar flow at the slit exit.Four working gas feeds were studied: Ar, Ar/O 2 , Ar/N 2 , and Ar/H 2 O aerosol.The Ar flow rate was set to either 67 or 100 slm, the O 2 flow rate was kept constant at 1 slm, and N 2 flow rate was varied between 0.5 and 3.5 slm.Water aerosol was produced from demineralized water (50 ml) by an ultrasonic nebulizer working at the 1.7 MHz frequency.It was gathered by a separate Ar flow (flow rates 1-3 slm) and intermixed with the main Ar gas stream before entering the PSJ body.
The measurements were carried out with the jet freely expanding to open air or impinging on five different targets placed 10 mm away from the slit exit.Two targets, the tables (0.22 m in height) with differently designed tops, were on a floating potential: • The target denoted as the dielectric table had the top made of a 6 mm thick mica composite plate (ϵ r = 6).
• In the table with glass window, plasma impinged on a 1 mm thick sheet of quartz glass (ϵ r = 3.8) covering a slightly larger 3 mm thick mica composite plate with a rectangular hole (95 × 35 mm) cut in its center.The PSJ was aligned with the longer principal axis of the rectangular cut-out.
Three remaining targets consisted of a mica composite plate in contact with plasma filaments and a ground beneath it: • In the table with a mesh, a grounded metallic mesh was held in between two 3 mm thick mica composite plates.• In the conveyor belt, the plasma impinged on a 3 mm thick mica composite plate positioned above a 2 mm thick aluminum plate inserted into the grounded conveyor belt cover.shift, and the delivered power.The impedance was also monitored, though, it was ∼50 Ω in all the measurements with the ignited plasma as the discharge was matched to this value.Due to the insufficient time resolution (0.1 s), the Octiv VI sensor could not follow the VI waveforms (the wave period of 13.56 MHz signal is 73.7 ns).Therefore, we measured the voltage and current waveforms with high-voltage (Cal Test CT4028) and current (Pearson 6585) probes connected to the oscilloscope (Agilent DSO9054A).The current was recorded in two parts of the circuit-directly after the RF generator and at the grounded electrode, where the measured signal corresponded to the discharge current.The voltage was measured only at the RF generator output because the highvoltage probe was not connectible to the high-voltage RF electrodes inside the PSJ body where the discharge voltage is induced.
The power dissipated directly in the plasma is standardly determined from the VI characteristics measured after the matching box.In our experimental setup, the matching unit is not a separate box but an integral part of the PSJ body (figure 1).Therefore, we would need to connect the highvoltage probe to the high-voltage RF electrodes inside the PSJ body, which, as explained above, was not viable.Instead, we used an indirect power measurement method proposed by Hofmann et al [41] using the VI data measured directly after the generator before the matching.Presuming the power dissipated in the matching unit with and without the discharge was the same for the same RMS current, the power dissipated in the plasma is equal to the difference of power delivered by the RF generator with and without the ignited plasma (figure 2).

Antenna measurements
The electromagnetic field generated by the PSJ was measured using electric and magnetic antennas placed in several different positions (figure 1(b)) relative to the PSJ.The electric field was assessed by two differently sized dipole antennas.The smaller custom-made dipole antenna (further in the text denoted simply as the dipole antenna) had a diameter of 10 mm and was used in two different positions, below and above the target.In the below-target position, the dipole antenna was situated directly beneath the plasma at a 1 mm distance between the topmost antenna part and the bottom target surface.In the above-target position, the dipole antenna was placed on the longer PSJ side, most often 20 mm away from the PSJ and 10 mm above the target (i.e. the center of the antenna was at the level of the PSJ slit outlet).The exception was the setup with the cryo-table, where the dipole antenna was positioned 45 mm above the substrate and 100 mm away from the slit exit to prevent its malfunction.The larger dipole antenna was the commercial biconical antenna BicoLOG 20300 (biconical antenna), 350 mm in a diameter.Thanks to its much larger diameter, it could be placed further away from the PSJ without losing the sensitivity even to weak overtone signals.In the utilized experimental configuration, the biconical antenna was placed 460 mm away from the shorter side of PSJ with the antenna center (maximum gain) leveled with the slit exit.
The magnetic field (results can be found in S2 section of supplementary information) was probed by Rigol antennas NFP-3-P1 (further referred to as the ring magnetic antenna) and NFP-3-P4 (small magnetic antenna).Both antennas were utilized in the same below-target configuration as the dipole antenna.The ring magnetic antenna was used oriented perpendicular to the plasma.The PSJ electromagnetic radiation (detected by the antennas) was recorded as power frequency spectra using the frequency analyzer ZVL-6 (Rohde & Schwarz) set to the maximum peak regime.The presented data were measured over frequency range 10-700 MHz using 50 accumulations (figure 3).The total acquisition time was 35.5 s.
The y-axis (intensity) of recorded power frequency spectra was calibrated using frequency-dependent antenna factors.The commercial antennas (biconical, ring magnetic, small magnetic) antenna factors were provided by the manufacturers.The antenna factors of the custom-built dipole antenna were measured in the transversal electromagnetic mode cell using the standard field method without a reference electric field probe.The dipole was placed between the plates of the parallel plate waveguide with a defined impedance.The antenna factor of the dipole antenna was obtained as the ratio of the theoretically calculated electric field induced between the waveguide plates (for the known applied power) and the effective voltage measured at the dipole antenna connector.
The biconical antenna measurements were calibrated using two different antenna factors depending on the presence of ground beneath the target.The original manufacturer-supplied antenna factors were used to calibrate the data obtained with the PSJ in open space configuration and ignited against the floating targets.In the remaining cases, the presence of the ground beneath a dielectric target distorted the relative intensity of the first and second harmonics compared to the current probe measurement.Since the biconical antenna is further away from the PSJ, the distortion is presumably not the result of a capacitive coupling between the discharge and the antenna.Therefore, the antenna factors were corrected by 15 dB for the first and second harmonics.
The influence of antenna directivity was neglected, as all the dipole antennas had relatively high radiation pattern widths.Besides, the measurements were carried out in a room with metal walls where the real antenna directivity is likely considerably different from the theoretical values, and its determination would require complex simulations far out of this study's scope.The laboratory room also acted as a resonance cavity.However, there was no overlap between the frequencies of higher harmonics generated by the discharge and the room resonance frequencies, and no harmonics data needed to be discarded.
The intensity of each higher harmonic was determined from the calibrated frequency spectrum by integrating its peak area.The average standard deviation of antenna measurements was 3.5 dB, with the maximum errors reaching up to 10 dB.The error bars are not shown in any graphs to maintain clarity and ease of orientation.

Determination of electric field intensity
The antenna factor AF used to calibrate the measured antenna signal is defined as the ratio of electric field intensity amplitude E to the effective voltage induced at the antenna connector U by the said electric field The influence of antenna directivity is again neglected.As a logarithm, the equation ( 1) takes the form of Thus, the calibrated antenna signal is directly proportional to the strength of the measured electric field, i. e. to the radiated electric field of the PSJ system (a mix of applied, plasma and surface-charge-induced electric fields).
The spectrum analyzer recorded (and displayed) the signal induced by the electric field in the form of power level P [dBm] expressed in decibels with reference to one milliwatt.The effective voltage U found in the equations ( 1) and ( 2) was derived from the power level using a basic electrical power equation where Z 0 = 50 Ω is the spectrum analyzer input connector impedance.Substituting the Z 0 value into the formula and using a logarithm, the power level expression becomes and the formula used to calculate the amplitude of electric field intensity measured by the antenna at its position can be rewritten using only the known quantities All the antenna measurements were carried out in the reactive near-field region (outer boundary of the region at the 13.56 MHz is ∼3.5 m) where the phase shift is frequencydependent.Therefore, the values of the electric field derived from the harmonics (subharmonics and ultraharmonics) peak could not be summed up, and the total amplitude of radiated electric field intensity was approximated by a value derived from the most intense contributor, the fundamental frequency peak.
An antenna measures the amplitude of electric field intensity E a in its position.In the reactive near field region, the decrease of electric field intensity with the distance from the source can be approximated by the inverse-cube law [42], the same as for the dipole field.Therefore, the antenna position (i.e. the distance between the antenna and the plasma) has to be factored in the calculation.The electric field recorded by the antenna consists of three contributions: the applied electric field, the electric field induced by the space charges (plasma), and the electric field induced by the surface charges deposited on the dielectric target surface by the discharge.Of the three contributions, the applied electric field is decidedly the weakest (the antenna signal measured when the RF generator delivers the power, but the discharge is not yet switched on is one to two orders of magnitude lower than with the ignited discharge) and thus was neglected.The electric field induced by the surface charges is also excluded from the calculation because it differs based on the target setup (dielectric type, ground proximity), and its estimation would require advanced numerical simulations.Thus, only the plasma and plasma sheath contributions are considered further on.
Since plasma is conductive and the potential gradient is concentrated in the thin plasma sheath, for an approximate estimate, plasma can also be disregarded, considering only the sheath of a constant height.Such approximation is possible because the electric field was measured after the discharge had stabilized (at least five minutes passed between the PSJ ignition and the spectrum recording), and the measurement duration (38 s) was much longer than the RF cycle (74 ns).The field within the sheath was modeled as a spatially extended dipole (similar to the field inside a capacitor), and so was the field outside.The recalculation ratio from electric field E a at the antenna position to E p inside the sheath was then obtained as the ratio of electric fields inside the extended dipole and at the antenna position.The effective geometry of the dipole was obtained using previous results of filament fast camera video analysis [43].Because of rapid horizontal filament movement and many repetitions in the antenna measurements, a simple rectangular cuboid geometry was used to represent the mean of all possible filament states.The cuboid length was the slit length (150 mm), its height was the sheath thickness (0.3 mm [44]), and its width was computed to match the estimated average area of the plasma sheath (1.2 mm for Ar).The electric field calculation was based on the biconical antenna data.The remaining valid setup, the dipole antenna in the below-target position, was too complex for the approximation as it does not account for the dampening of the electromagnetic radiation by the target.

VI waveforms and power dissipated in PSJ
The RF PSJ is a filamentary discharge (figure 4).The appearance, number, and movement of PSJ filaments differ depending on the working conditions, especially on the gas feed.The constricted filaments of the Ar PSJ were the most numerous and shuffled along the length of the slit (except for stationary filaments situated at both ends of the slit).Admixing of oxygen (Ar/O 2 PSJ) did not affect the constricted nature of the filaments, although their number decreased, and they became almost stationary.In the Ar/N 2 gas mixture, the filaments were composed of short constricted centers surrounded by longer diffuse plasma double plumes.Their number and the degree of movement were between the Ar and Ar/O 2 PSJ.In addition to the motion, filament formation and annihilation events happened.Most crucially, the PSJ filaments formed self-organized patterns characterized by inter-filament distance.The behavior resembles the self-organizing patterns of a quasi-1D dielectric barrier discharge (DBD) [43,45].
Electrical characteristics provide valuable information about the discharge conditions.For example, the VI waveforms are often used to distinguish between the filamentary and homogeneous modes in DBDs, as the filamentary mode is easily recognizable by sharp current peaks, each corresponding to a newly ignited filament superimposed on the discharge current waveform [46,47].Despite the similarities between the PSJ and a DBD, this was not the case for the PSJ, whose discharge current waveforms were smooth with no sharp peaks (figure 5), although the plasma was decidedly filamentary.The RF PSJ operated at three orders of magnitude higher frequency than a typical DBD.Therefore, it cannot be understood as a series of consecutive discharges with reversed electrode polarity.At the RF frequency, the breakdown occurred only once.The filaments were not extinguished during the individual half cycles because the fast polarity switching prevented the charge build-ups on dielectrics.Hence, the current waveforms do not have sharp peaks in the RF PSJ or the standard RF-driven filamentary DBDs [48][49][50].Admixing oxygen or nitrogen gas into the Ar PSJ led to a decrease in discharge current amplitude and a higher distortion of the applied sinusoidal waveform (figure 5).In other words, the PSJ load behaved more non-linearly, and a higher number of more intense harmonics of the fundamental frequency 13.56 MHz was generated.
The phase shift was probed on the coaxial cable after the RF generator before the PSJ.Prior to the breakdown, current lagged the voltage by circa −83 • regardless of the matching conductive plates position (figure 1), meaning the load of PSJ body (and the coaxial cable) was inductive.With the plasma on, the load of the whole system, i. e. plasma, PSJ body, and coaxial cable, became mostly resistive, with phase shifts in the range of −22 • to 18 • (table 1).Approaching the zero phase shift for the switched-on plasma was expected because the voltage and current waveforms should be in phase in the matched circuit.
Increasing the applied power and admixing oxygen and nitrogen gases had the opposite effect on the phase shift.The phase shift had a more positive value (table 1) when the applied power was changed from 500 to 600 W. On the other hand, admixing of molecular gas into the Ar working gas flow moved the phase shift to more negative values, with nitrogen admixing having a more pronounced influence.Changes in plasma conductivity can explain the influence of both the applied power and gas feed.Increasing the applied power from 500 to 600 W increased the power dissipated in the plasma, resulting in a higher ionization rate and, hence, higher electron density, directly proportional to plasma conductivity.Molecular gas admixing had the opposite effect due to a loss of energy in dissociation and excitation of numerous rotational and vibrational levels and a loss of electrons due to oxygen electronegativity.The power dissipated in the PSJ discharge P diss was more or less the same for all the studied gas feeds, 450-460 W at 500 W, and 530-550 W at 600 W applied power (table 1).The corresponding power transfer efficiencies, calculated as the ratio of power dissipated in plasma and the applied power measured by the VI probe at the RF generator output, were 0.92-0.95,with the Ar PSJ having the highest efficiency and the Ar/N 2 PSJ the lowest.The measured P diss and thus the power transfer efficiency values are likely somewhat overestimated compared to reality because the used indirect measurement method expects the same heat dissipation (i.e. the same power losses induced by Joule heating) with and without plasma.This assumption was not fulfilled in the PSJ, as the PSJ body (mainly the coil) was heated up more when the discharge was on.Furthermore, the temperature of the PSJ body with the ignited plasma depended on the position of the aluminum plates used for the impedance matching.The heating was notably stronger when the upper part of the coil was situated between the matching plates (the pure Ar discharge).The influence of plasma gas chemistry (e. g. vibrational excitation transfer in the Ar/N 2 PSJ) on the PSJ body heating can be disregarded as the rotational temperature used to approximate the plasma gas temperature was the same in all the gas feeds (T rot ≈ 750 K).Relative intensities of higher harmonics present in the antenna spectra taken for PSJ ignited against the conveyor belt using different gas feeds, flow rates, and power.

Antenna measurement of radiated electric field
The electric antenna measurements provided information about the distortion of applied sinusoidal waveforms by the plasma and the amplitude of the PSJ system's radiated electric field.Focusing on the analysis of the PSJ system non-linearity, the advantage of the antenna measurement over the standard VI probes lies in the non-intrusivity of the former method.An antenna monitors the distortion without being connected to the system, whereas VI probes must be included in the electrical circuit, which is not always possible.Additionally, the antennas usually have higher efficiency at the higher frequencies (starting at ∼100-150 MHz), enabling a precise intensity measurement of very weak higher harmonics (seventh and higher in our setup).
At the fundamental frequency 13.56 MHz (wavelength λ = 22.12 m), all the antenna measurements were carried out in the reactive near field region (outer boundary ∼3.5 m).In this region, adjacent conductive objects can absorb and re-emit the radiation, distorting the initial electromagnetic wave radiated by the PSJ.More crucially to our measurements, placing antennas too close to the plasma may affect its parameters, thus the emitted spectrum.Therefore, a procedure based on comparing the relative intensities of frequency spectra obtained by different antennas and from the discharge current recorded by the current probe was used to validate the measurements, see the S1 section of supplementary information.It uncovered capacitive coupling between the PSJ plasma and the dipole antenna situated above a target, whose data were therefore discarded.As for the validity of the obtained electric field values, the electric antennas measured the radiated electric field of the PSJ system, which consisted of applied, plasma, and surface-charge-induced contributions.Due to the limitation of the experimental setup and several approximations used in the calculation (section 2.4), the obtained values are only semiquantitative.However, despite the uncertainty in absolute values, the obtained trends are reliable as they are reproducible by all the antennas.

Generation of higher harmonics, subharmonics, and ultraharmonics.
The gas feed induced non-linearity was studied in the configuration with the PSJ ignited against a 3 mm thick mica composite plate placed on the grounded conveyor belt, i. e. in the standard setup for plasma treatment experiments.In all the tested working conditions, the recorded frequency spectra had the same main features (figure 6) -an intense fundamental frequency and significantly weaker higher harmonics.Moreover, the odd harmonics were markedly stronger than the even ones.In the Ar/N 2 PSJ, this trend persisted over the whole frequency range, while in the Ar and Ar/O 2 PSJ, their intensities became comparable starting from the sixth harmonic.The ratio of spectral weights of higher harmonics and the fundamental frequency 13.56 MHz was used to describe a distortion of generator-supplied sinusoidal signal by the plasma.The results obtained by the antennas correspond well with the discharge current measurements, with the lowest ratio observed in Ar (0.005-0.03), followed by Ar/O 2 (0.05-0.10), and the Ar/N 2 (0.08-0.16) gas mixtures.In the Ar discharge, changing the power and Ar flow rate did not affect the relative intensities of harmonics.Similar to the Ar PSJ, the frequency spectra of Ar/O 2 PSJ were within the experimental errors the same for all the tested combinations of O 2 :Ar flow rate ratio and applied power.
The non-linearity of the PSJ load induced by nitrogen admixing changed with the N 2 :Ar flow rate ratio.Overall, the N 2 flow rate did not affect the relative intensities of odd harmonics within the experimental errors (figure 7).The same does not apply to the even harmonics whose relative intensities increased with the higher N 2 flow rate.
In addition to higher harmonics, subharmonics (partial multiples of the fundamental frequency) and ultraharmonics (integer multiples of subharmonics) were generated by the PSJ ignited in nonstandard working conditions, i. e. against an atypically rough mica composite target with and without admixing of water aerosol into Ar gas flow.The increase in target roughness led to the excitation of weak half-harmonic and its ultraharmonics.Contrary to the higher harmonics, changing the working parameters (gas feed, flow rates, power) did not affect the subharmonic and ultraharmonics relative intensities, except for the water aerosol admixing that increased their strength by 10-20 dB probably due to the droplets deposited on the surface increasing its roughness.The addition of the water aerosol also generated the fifth subharmonic and its multiples (1/5, 2/5, 3/5, 4/5, 6/5, . ..) once a threshold concentration of aerosol in the plasma was crossed (2 slm of Ar carrier gas in our PSJ setup).
We hypothesize that the fifth sub-and ultraharmonics originated from the oscillation asymmetric sheaths formed around the water droplets [51].The hypothesis is based on the analogy with a generation of sub and ultraharmonics by the insonated microbubbles in the liquid ultrasound contrast agents that have been attributed to the oscillation of microbubbles diameters [52,53].As for the effect of roughness, two possibilities exist.The generation of the half subharmonic and its multiples can be caused by the inhomogeneity of plasma density (plasma density is higher at the apex of roughness modulations).Alternatively, the generation could be a result of the SIW detaching from the surface and transitioning into a new bulk IW launched from the apex of the surface curvature towards the next across the valley or, in the droplet case, towards the target surface [54].

Electric field intensity.
The amplitudes of the electric field intensity E changed substantially with the different applied power, gas feed, flow rate, or configuration (figures 8 and 9).Generally, the semi-quantitative E values ranged between 10 to 70 kV cm −1 except for the pure Ar discharge impinging on the conveyor belt target.In this setup, the measured electric fields increased to (140 ± 50) kV cm −1 when the Ar PSJ was ignited in more extreme working conditions (100 slm Ar and 600 W).Overall, the E values increased with the applied power and decreased with the admixing of molecular gas.The effect of the Ar flow rate differed with the gas feed.
The measured PSJ electric field was a sum of the three contributions: the applied electric field, the electric field induced by the space charges (plasma), and the electric field induced by the surface charges deposited on the dielectric surface by the discharge.All three grow with applied power, resulting in higher E values (figure 8).The influence of molecular gases (O 2 or N 2 ) admixing is also straightforward.In the Ar/N 2 and Ar/O 2 gas mixtures, both the space and surface charge density were presumably lower because electrons lost their energy by dissociation and excitation of molecules rather than in the ionization reactions, hence the lower electron density and lower electric field.This explanation was corroborated by the decrease in the amplitude of electric field intensity with the increasing N 2 :Ar flow rate ratio while the other working parameters were kept constant (figure 9).In the Ar/O 2 PSJ, electrons could also be lost in the attachment processes with the electronegative O − 2 ions [55].A change in the Ar flow rate affected the electric field intensity only in the Ar PSJ, where a stronger electric field was generated at the higher Ar flow rate (100 slm of Ar in comparison to 67 slm) and the same applied power of 500 W (figure 8).

Role of the configuration (open space, different targets).
Previous parts discussed variations of PSJ interacting with mica composite on the grounded electrode (conveyor belt) due to the applied power, gas feed composition, and flow rates.The PSJ electrical properties also depended on the experimental setup, specifically, whether the plasma extended into an open space or impinged on a floating or dielectrically covered grounded target (figure 10).
The frequency spectra of the PSJ ignited in Ar and Ar/N 2 gas feeds reacted to the presence of a grounded target differently.In the Ar PSJ, the relative intensities of odd harmonics decreased after the target addition (figure 10).In contrast, the frequency spectra of the Ar/N 2 PSJ were identical for both the open space and the grounded target configurations.The arrangement of the grounded targets (table with mesh, conveyor belt, and cryo-table all covered by mica composite) did not have any influence, either due to the ground proximity or because the plasma was in contact with the same dielectric (mica composite).The difference in the Ar and Ar/N 2 PSJ behavior stemmed from the gas chemistry.In both cases, the presence of a grounded target enhanced the intensity of the PSJ electric field (figure 9), which is consistent with the results reported for the He APPJs interacting with different surfaces [15,25].In these studies, an increase in the ionization rate and the production of neutral species (O, OH, N * 2 ) was observed, together with a higher acceleration of charged particles towards the target surface.Assuming the same for the PSJ, the presence of a grounded target increased the plasma density and the velocity of charged particles.In the Ar PSJ, it led to a decrease in non-linearity of the discharge, presumably because the number of volumetric ionization processes increased significantly compared to the localized electron production by the long-lived neutral reactive species.In the Ar/N 2 PSJ, admixing of N 2 into the Ar working gas prevented one process of electron production from dominating, and thus, the relative intensities of higher harmonics were the same in both the open space and grounded target configurations.
The floating target influence was monitored using the Ar/N 2 discharge.Compared to the grounded target and open space configurations, the relative intensities of both odd and even harmonics notably decreased when plasma impinged on the floating targets (figure 10).Additionally, a slight difference in the frequency spectra of the PSJ ignited against the different dielectric targets was observed.Without the grounded electrode situated beneath a non-conductive target, charge deposited by the plasma accumulated on the target surface, inducing a potential that inhibited the discharge, presumably resulting in lower density of both charged and neutral reactive species [15,20].Thus, the relative intensities of higher harmonics also decreased.The amount of charge deposited on the target surface depended on its capacitance-the lower the capacitance (relative permittivity), the stronger the discharge inhibition, and the lower the relative intensities of higher harmonics, as seen in the differences between the spectra measured for plasma interacting with the dielectric table (ϵ r = 6) and the table with glass window (ϵ r = 3.8).

Role of secondary electrons
Secondary electrons have significant impact on the discharge response as seen by the example of the RF atmospheric pressure glow discharge (APGD), transitioning from the linear α-mode (the discharge is sustained by volumetric ionization processes) to the non-linear γ-mode (the localized secondary electron emission appears and sustains the discharge mode) [56][57][58].In the studied PSJ, we also observed a transition from linear to non-linear loads after admixing oxygen and most importantly nitrogen into the working argon gas feed (section 3.2.1 and figure 6).We propose that the transition appears due to the secondary electron emission from dielectric surfaces, similar to the APGD.Khamphan et al [59] concluded that secondary emission induced by metastable nitrogen species is significant for understanding the formation of a DBD in nitrogen at atmospheric pressure.
In our experiments, increasing the N 2 :Ar ratio resulted in the highest non-linearity of the discharge (figure 7).Nitrogen N 2 (A) metastables are specifically efficient at producing secondary electrons, as demonstrated by studies of homogeneous DBDs [46,60].Increasing the N 2 :Ar ratio generates a higher amount of nitrogen species, including nitrogen metastables enhancing the emission of secondary electrons.Thus, the load non-linearity increases.
The homogeneous mode of DBD in N 2 was previously attributed to the memory effect (i. e. production of electrons between the pulses) driven by secondary emission induced by N 2 (A) metastable impacts [46,60].The visual appearance of the PSJ discharges also confirms the importance of nitrogen metastables creating low-energy secondary electrons emitted from the dielectric surface slit by nitrogen metastable impacts.Compared to the Ar and Ar/O 2 thin well-defined filaments (section 3.1 and figure 4), the Ar/N 2 PSJ filaments were composed of a diffuse region surrounding the short constricted center.At the higher N 2 :Ar ratio, the central filament completely disappeared inside the slit, leaving only the diffuse plasma reaching outside [43].
The dynamics of RF sheaths is strongly non-linear but if the effects of two symmetrical sheaths are combined, as in symmetrical capacitively coupled RF discharges, only odd higher harmonics remain in the self-consistent analytical solution  considering collisionless RF sheaths [61].Thus, the even harmonics react sensitively to the discharge asymmetry.Electronneutral collisions damp current oscillations, yet higher harmonic frequencies were theoretically predicted even for atmospheric pressure discharges [62] and a weak third harmonics was observed experimentally in quite linear α-mode of RF APGD [63].In the PSJ, nitrogen addition to Ar led not only to a significant increase of an overall non-linearity response but also to a relative increase of even harmonics compared to the odd ones.We attribute it to the transition from filamentary to more diffuse discharge mode discussed above.

Comparison of antenna measurements with previous electric field measurements
The published measurements or simulations of the electric field usually cover the capillary He APPJs ignited in different configurations (electrode setup, driving voltage, targets) [14,15,20,21,25,27,32,39,[64][65][66][67].Depending on the experimental method, the electric field of the IW (Stark polarization spectroscopy) [21,27,32,65] or the electric field on a dielectric surface induced by the impinging APPJ (electrooptic sensors based on Pockel's effect) [20,39,66] were measured, both methods yielding similar strengths with the maximum electric fields ranging between the 10-45 kV cm −1 .The semi-quantitative values of PSJ system electric fields obtained by the antenna measurements are generally in the same order of magnitude as in the He APPJs, 10-60 kV cm −1 , although in certain conditions (Ar PSJ ignited at high flow rates or powers) they reached up to 140 kV cm −1 .The higher intensities of the PSJ electric fields arise partly from using Ar as the working gas [36,37].However, the primary origin of the higher electric fields values obtained by antennas is the measurement principle and approximations used during the calculation.
The effect of most PSJ working parameters on the measured electric field corresponds with the results obtained in He APPJs.The electric field increases with the applied power and decreases when oxygen and nitrogen gases are admixed into the working Ar gas flow.The exceptions are the changes in the PSJ electric field strengths induced by the configuration and changing the Ar gas flow rate.
Comparing the electric field strengths obtained with the different targets (figure 9), the proximity of the ground was the main factor affecting the measured E values, as the electric field of the PSJ impinging on the grounded target was on average 2× higher than in the floating target configuration, and 3-4× higher than in the open space.These results somewhat correspond with the data published for the He APPJs interacting with different targets [14,21,25], the exception being the open space configuration as in the He APPJs the same electric field strengths have been obtained in both the open space and floating target setups [21].The PSJ in the open space configuration having the lowest electric field strengths is presumably caused by the coarse approximation used in its calculation, i. e. the whole distance between the end of the filament and the target surface is approximated as the plasma sheath.The permittivity of a floating target did not influence the measured electric field, probably because its influence was counterbalanced by the different target thicknesses [67].
Increasing the Ar flow rate while keeping the same applied power led to a generation of a stronger electric field in the pure Ar PSJ (figure 8).The observed change does not match the results presented in the papers on the He APPJs driven by kHz generators.In the open space configuration, the maximum electric field of the He APPJ was the same for all the flow rates [27,32], whereas when the He APPJ was ignited against a target, the electric field strength decreased with the increasing He flow rate [32].However, in all these publications, Stark polarization spectroscopy was used, measuring only the strength of the electric field inside the plasma jet.Using electric antennas, we cannot monitor the changes in electric field contributions separately.Thus, one of the explanations for the electric field increasing with the Ar flow rate is a higher amount of charged particles deposited on the dielectric surface by the discharge.Correspondingly, the length and intensity of the filaments in contact with the dielectric slightly increased with the higher flow rate.
The measured electric field strength might have also been markedly affected by the amount of air entrained in the Ar flow.The above-cited studies of the He APPJs [27,32] agree that the amount of air entrained into the He gas flow is a crucial parameter governing the results, with the maximum electric field corresponding to a certain air mole fraction (e. g. 0.014 for the He APPJ studied by Sobota et al [27]).In the capillary He APPJ, the ideal air mole fraction was achieved at low flow rates, hence the observed decrease of the electric field with the increasing working gas flow rate [32].In the PSJ, the situation might have been reversed.At the lower flow rate of 67 slm, the intermixing of air with the main Ar flow is not high enough, and the optimal mole fraction is only reached once the Ar flow rate increases.Therefore, the higher electric field strength was observed using 100 slm Ar.In the Ar/O 2 (and Ar/N 2 ) discharge, the effects of molecular gas admixing presumably dominated over the influence of Ar flow rate, and we observed no changes in the measured electric field intensity.
Compared with Stark polarization spectroscopy and Pockels technique, antennas do not measure a specified electric field component in a defined discharge region.They record an overall mix of all radiated electric fields regardless of their source: the applied electric field and electric fields induced by space and surface charges.Thus, it is reasonable to presume that the highest electric field of the PSJ system, the field in the plasma sheath, will be the main contributor to the signal measured by the antennas.Numerical simulations of He/O 2 APPJs have shown that the electric fields up to 100-200 kV cm −1 can be induced in the plasma sheath [14,15], which are comparable with our highest obtained values.However, it must be highlighted again that even though there is a nice overlap with the He APPJs, the antenna data are only semi-quantitative.Given the large intensity scale of the measured data (ten orders of magnitude), the reproducibility of the absolute intensities is also lower than in other techniques, which is reflected by the large error bars, about 30% of the measured value.Nonetheless, the observed trends were consistently reproduced in all the different configurations, demonstrating sensitivity of the antenna measurements to changes in the PSJ system.In summary, even though the antenna measurements do not yield accurate absolute electric field values in a well-defined plasma region, they appear wellsuited for tracking the macroscopic changes, which can still provide valuable insights into the plasma processes.

Conclusion
The RF PSJ consisted of about a dozen filaments selforganized into patterns defined by a characteristic interfilament distance, similar to self-organized patterns of DBD filaments.Despite the shared similarities, the PSJ discharge current waveforms were smooth, with no sharp peaks typical of the filamentary DBDs.At the RF frequency, the breakdown occurred only once, as the fast polarity switching prevented the charge from building up on dielectrics; thus, the filaments did not extinguish before the polarity switching.The choice of the gas had no effect on the power dissipated in the PSJ discharge that was more or less the same for all the studied gas feeds, 450-460 W at 500 W, and 530-550 W at 600 W applied power.However, these values are likely somewhat overestimated as we did not consider the different heating dissipation in the coil with and without the ignited plasma.
The non-linearity of the plasma system was studied using the higher harmonics, subharmonics, and ultraharmonics, whose relative intensities were for the first time measured non-intrusively by the antennas.The properties of the PSJ discharge, hence its frequency spectrum, changed with the applied power, gas feed, flow rate, and target.Nevertheless, all the recorded frequency spectra had the same main features: dominant fundamental frequency peak, relatively intense odd harmonics, and significantly weaker even harmonics.The intensity of both higher harmonics increased after the admixing of molecular gas.We propose that the transition appears due to the secondary electron emission from dielectric surfaces, similar to the APGD.In the Ar/N 2 PSJ, a high amount of nitrogen species, including nitrogen metastables enhancing the emission of secondary electrons, were created, resulting in the highest load non-linearity.
The influence of the target on the PSJ non-linearity differed with the gas chemistry.The relative intensities of odd harmonics of the Ar PSJ decreased when the plasma was in contact with the grounded dielectric target, presumably due to the enhanced volumetric ionization.On the other hand, the frequency spectra of the Ar/N 2 PSJ were identical for both the open space and the dielectrically covered grounded target configurations, as the addition of molecular gas limited the number of electron ionization collisions.Without a ground beneath a non-conductive target (floating target), the charge deposited by the plasma accumulated on the target surface, inducing a potential that inhibited the discharge, causing a decrease in the relative intensities of both the odd and even harmonics.
The electrical antenna output is the effective voltage induced at the antenna connector by the external electric field.
In our experimental setup, we record a global mix of three electric fields radiated by the PSJ system: the applied electric field and electric fields induced by space and surface charges.Approximating all these sources by the electric field of the plasma sheath (expected to be the highest contributor), the antenna signal was recalculated to the semi-quantitative values of the macroscopic PSJ system electric fields.Out of the tested gas feeds, the strongest electric field was generated in the Ar PSJ (50-140 kV cm −1 , depending on the applied power and Ar flow rate).The admixing of oxygen and nitrogen gas reduced the measured electric field by 1.5 to 2.5.As for the target influence, the proximity of the ground affected the measured electric field values, while the target setup (permittivity, spacial configuration) did not have any observable influence.The electric field of the PSJ impinging on the grounded target was, on average, 2× higher than in the floating target configuration and 3-4× higher than in the open space configuration.Though the obtained values of the electric field are only semi-quantitative, the antennas exhibited high sensitivity to the changes in the PSJ system, reliably following the abovediscussed trends, thus demonstrating the potential of this novel electric field measurement method.

Figure 1 .
Figure 1.(a) Schematic drawing of the RF plasma slit jet (side view) and (b) of the experimental setups used for the electromagnetic characterisation of the discharge: 1 -combined RF voltage/current sensor or separate voltage and current probes , 2 -coaxial cable, 3 -RF plasma slit jet, 4 -current probe, 5 -dipole antenna positioned above a substrate, 6 -substrate, 7 -biconical antenna, and 8 -dipole or magnetic antennas positioned under a substrate.

Figure 2 .
Figure 2. The determination the power dissipated in the Ar PSJ discharge as the difference of power delivered with and without plasma at the same RMS current.

Figure 3 .
Figure 3.A segment of uncalibrated frequency spectrum (full range 10-700 MHz) measured by the biconical antenna for the Ar/N 2 PSJ (67 slm Ar/1.5 slm N 2 , 500 W) ignited against the conveyor belt covered with mica substrate.The distance between the slit outlet and the surface was 10 mm.

Figure 4 .
Figure 4. Front (left) and side (right) views of the RF plasma slit jet discharge ignited in three different gas mixtures at 500 W: (a) 67 slm Ar, (b) 67 slm Ar/1 slm O 2 , and (c) 67 slm Ar/1.5 slm N 2 .The distance between the slit outlet and the dielectric surface (mica composite, thickness of 6 mm) was kept at 10 mm.The exposure time of the images was 25 µs, corresponding to approximately 340 RF cycles.Reproduced from[43].© IOP Publishing Ltd.All rights reserved.

Figure 5 .
Figure 5. Applied voltage and grounded electrode current waveforms taken at the fixed applied power 500 W and flow rates 67 slm Ar, 67 slm Ar/1.5 slm N 2 and 67 slm Ar/1 slm O 2 .

Figure 6 .
Figure 6.Relative intensities of higher harmonics present in the antenna spectra taken for PSJ ignited against the conveyor belt using different gas feeds, flow rates, and power.

Figure 7 .
Figure 7.The influence of N 2 flow rate on the relative intensities of higher harmonics present in the frequency spectrum of Ar/N 2 PSJ (67 slm Ar/0.5-3 slm N 2 , 500 W) impinging on different targets (table with mesh, dielectric table, and table with glass window).

Figure 8 .
Figure8.The effect of applied power and gas feed composition on the electric field of the PSJ ignited against the conveyor belt.

Figure 9 .
Figure 9.The amplitude of electric field intensity in the PSJ ignited in different configurations-open space, in contact with different grounded and floating targets.The applied power and Ar flow rate were fixed at 500 W and 67 slm, respectively.

Figure 10 .
Figure 10.The relative intensities of higher harmonics present in the frequency spectrum of Ar (67 slm Ar, 500 W) and Ar/N 2 (67 slm Ar/1.5 slm N 2 , 500 W) PSJ ignited in different configurations-open space, in contact with dielectric targets on a floating potential (dielectric table and table with glass window) or placed on a grounded electrode (table with mesh, conveyor belt, and both cryo-table setups).

Table 1 .
Electrical parameters of the PSJ operated in Ar (67 slm Ar), Ar/O 2 (67 slm Ar/1 slm O 2 , and Ar/N 2 (67 slm Ar/1.5 slm N 2 ) gas feeds at two different applied powers (500 and 600 W).The RMS current I RMS , voltage U RMS , and their phase shift were measured on the coaxial cable after the RF generator before the PSJ (and the matching).
table with mesh, dielectric table, and table with glass window).
table with glass window) or placed on a grounded electrode (table with mesh, conveyor belt, and both cryo-table setups).