Characterization and comparability study of a series of miniaturized neon plasma jets

During the development of new therapeutic devices, comprehensive experimental investigations are necessary in all phases of the process. This requires the provision of device prototypes with reproducible and comparable operating parameters. In the current study, such miniature neon plasma jet prototypes designated for medical applications have been manufactured, characterized, and compared. The multi-parametric characterization included measurements of energy, power, temperature, leakage current, effluent length, and relative as well as absolute radiation. The dissipated electrical power and the optical emission in the UV range were identified as parameters indicating definable tolerances to sort out a device with inadequate species output. A liquid phase model was used to investigate reactive species deposition into simple matrices. Based on these investigations, a quality control procedure for manufacturing new device series is proposed. In conclusion, our findings suggest a test concept of achieving reproducible and comparable plasma device characterization as a putative quality control measure for lab-scale plasma source production.


Introduction
In recent years, several plasma jets for medical applications were developed, and even a few were certified for medical use Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.[1][2][3][4].Commonly, the application is aimed at chronic wounds as well as pathogen-based skin diseases.In addition, some studies investigated the effect of a plasma jet on the mucosa for dental application [5][6][7][8].Using plasma jets for intra-corporeal applications is more challenging because very small dimensions, enhanced flexibility, and the ability to operate in hollow bodies are required.Still, there are already several promising approaches to meet all these requirements.Notable examples are the plasma gun [9,10], a transferred atmospheric pressure plasma jet [11,12], and a miniaturized jet initially developed for endoscopic applications [13,14].This miniaturized plasma jet, optimized for operation in the instrument channel of a medical endoscope, has been chosen as the object of the present study.As it has the advantage of being operable in cavities and is miniaturized, it is potentially suitable for application in the upper respiratory tract for, e.g.antiviral and antimicrobial treatments [15].
Riedel et al [16] investigated four devices of the 'COST reference microplasma jet' to establish a reference source for scientific research with reproducibility as a crucial condition over different labs and countries.They compared the electrical characteristics with optical emission spectra, absolute densities of atomic oxygen, ozone and temperature, and bactericidal activity.In another study, a set of six plasma devices of the same type for dental application were compared to verify safe application and comparability within a multicentric clinical trial [8].
In the present study, six identically in-house manufactured plasma jets were investigated and evaluated regarding their reproducibility and variance to verify comparable operation of plasma sources for biomedical investigation.In addition, a quality control parameter for serial production was aimed to be established.The intended application area for the herepresented system is the upper respiratory tract.The electrical energy dissipated in the plasma is converted into various forms of energy, for instance, the electric excitation of atoms and molecules, thermal, radiative, and chemical energy.Therefore, electrical parameters, optical characteristics, and reactive species output in the gas and liquid phase were analyzed to investigate whether the six devices in the produced series provide comparable performance.Moreover, the appearance of correlations between different operating parameters has been investigated.Finally, a quality control procedure for manufacturing new series of devices is proposed based on the obtained data to validate comparability.

Plasma source
The plasma jet consists of an inner polytetrafluoroethylene (PTFE) tube around which a copper wire (100 µm in diameter) is coiled.This inner tube has an inner and outer diameter of 0.9 mm and 1.16 mm, respectively.
At the end of it, a ceramic nozzle is placed.This nozzle has one smaller end with a diameter of 0.9 mm, which is pressed into the end of the inner tube.The stability of this connection is enhanced by the application of a small amount of epoxy.The opening of the nozzle has a diameter of 0.4 mm.In the nozzle, grooves enable the shielding gas to pass through.The outer PTFE tube (inner diameter 1.5 mm and outer diameter 1.8 mm) is placed around the thicker part of the nozzle (1.7 mm in diameter).The whole ceramic nozzle has a length of 8 mm in total.Figure 1 depicts the setup of the device.
The copper wire acts as a high-voltage electrode and is narrowly wired, as described by Winter et al [13], to avoid plasma ignition inside the tubing.The last windings are fixed with epoxy and end about 1 mm in front of the thicker part of the nozzle.The high-voltage electrode is connected to a system-specific, self-tuning high-voltage generator that generates a sinusoidal signal with a frequency of around 415 kHz.Depending on the load, the frequency changes slightly and the individual frequency was extracted from the individual aquired data set.This sinusoidal signal is clocked with a frequency of 1 kHz and a duty cycle of 32%.
As feed gas, 300 sccm neon (Ne N40, Air Liquide) with an admixture of 0.3% (0.9 sccm) molecular oxygen (O 2 N48, Air Liquide) has been used.Carbon dioxide (CO 2 N45, Air Liquide) was used as shielding gas with a flow rate of 300 sccm.Given the purity levels of the gases under use, contributions of up to 6 ppmv of molecular nitrogen and 2 ppmv of water are considered withing the main working gas channel.In addition up to 25 ppmv of molecular nitrogen and 5 ppmv of water can remain within the supplied carbon dioxide.Further impurities from tubing lines has to be considered.Here, the usage of full teflon tubing is stated to minimize these effect in correspondence with previous studies [17].Gas flows were controlled using microflow controllers (Bronkhorst), two 500 sccm and one 50 sccm.They have an uncertainty of ± 1% of the measured value.
Essential performance parameters of the device type are presented in [14].Noticable features include a leakage current below 10 µA, a concentration of gaseous ozone (O 3 ) and nitrogen dioxide (NO 2 ) in the farfield of below 3 ppm O 3 and 0.3 ppm NO 2 , and a considerable microbial inactivation depending on the treatment time.

Electrical measurements
For the general characterization, the miniaturized, flexible plasma jets were mounted in a non-conducting polyoxymethylene (POM) tube with 20 mm outer diameter to enable simple and reproducible handling.

Electrical energy.
To characterize the electrical behavior of the different devices of the plasma jet, a Petri dish with a conducting copper foil stuck on its backside was used as the counter electrode.To measure the dissipated power, the counter electrode was connected to ground with a capacitor in series, as shown in figure 2.
The electrical charge Q = C m • U C was used to calculate the dissipated energy.From the measured driving voltage V In , the capacitance C m and the voltage drop over the capacitor U C the energy per period E t0−t1 was calculated according to equation (1) [18][19][20][21].For a higher measurement accuracy, 16 individual periods were averaged, which were uniformly distributed over the burst This measurement of the dissipated energy has been performed twice per device, the first time at the beginning and the second time at the end of the characterization period.Thus, the value is used as an indicator of the long-term stability of the device's performance.However, since the operation using a dielectrically covered copper foil as a treatment target provides conditions different from the ones during practical application in a biological or medical environment, the obtained values are not necessarily representative for the practical usage of the system.
The same limitation applies for the measured dissipated power P, that has been calculated by integrating over one complete cycle of a burst T according to equation (2) The power P has been measured as another parameter that is easily accessible in a routine measurement to compare the behavior of different units of the system.Figure 3 depicts the voltage waveform and the bounds of integration used for equations (1) and (2).The voltage waveform results from the construction of the system-specific high-voltage generator [14] that is optimized for reliably igniting the jet while maintaining a low temperature and a tolerable patient leakage current (PLC).During the active phases, the signal is sinusoidal and results in the QV plot depicted in figure 4.

PLC.
The PLC was measured using the device Bender Unimed 800ST in both alternating current (AC) and direct current (DC) conditions.Its internal measurement circuit is depicted in figure 5.With each jet device to test, two different target options were used.As the first target, the same Petri dish with the copper foil at the backside that was applied for the electrical measurements was used.As the second option, a 7 mm × 7 mm copper plate was utilized as a target.Both were electrically connected to the Bender device over the internal leakage current circuit in figure 5.The DC values of the leakage current were exemplarily measured at a distance of 1 mm.All measured DC values were at the lower  detection limit of the Bender Unimed device.The measurements were performed according to standard 'IEC 60601-1-11:2015 ′ [22] and the technical rule 'DIN-Spec 91315' [23].

Optical measurements
For optical characterization of the plasma jets, three investigations were performed.The length of the effluent was measured for each plasma jet, the temperature was determined and the overview emission spectra was recorded per device.

Effluent length.
By imaging the discharge effluent with a reference scale depicted in figure 6.The image was captured by a Canon PowerShot SX50 HS, with fixed settings like aperture, exposure time and iso.The camera was mounted on the bench to get reproducible images and the zoom was set to the same scale each time after camara off, to get the same image resolution.The image resolution varies from 26.5 to 30.5 pixel mm −1 .The length of the plasma jet's effluent was determined by a self-written python script.The effluent length was determined where the intensity dropped to 30% of the detected maximum.As in the previous investigation, two different temperatures were determined [24,25].The first temperature value was measured in the effluent, approximately 1 mm in front of the target.For this measurement, the same Petri dish with the copper foil at the backside that was applied for the electrical measurements was used as a target.However, the connection of the conducting foil to ground was different.For the temperature measurements, the leakage current measurement circuit shown in figure 5 provided the connection between the foil and ground.
The temperature probe (FOT Lab.Kit) was held from the bottom, and the tip touched the effluent 1 mm in front of the target.The distance between the target and the plasma jet nozzle was varied between 1 mm and 6 mm, and for each millimeter, one measurement point was recorded.
The second temperature value (target temperature) was measured at the backside of a copper plate of 7 mm × 7 mm.The copper plate is equipped with a small tube at the center of its backside.This tube ensures a better heat distribution around the thermal probe and allows a fixation.The probe tip was placed in that tube so that it touched the backside of the copper plate.The copper plate was connected to the leakage current circuit, see figure 5.The distance between the target and plasma jet nozzle was varied between 1 mm and 6 mm, and for each millimeter, one measurement point was recorded.
Temperatures and PLC were acquired with a customengineered python-based acquisition software (PlaDinSpec, Leibniz Institute for Plasma Science and Technology), measuring 50 subsequent values per spatial position, after the temperature was stabilized.In a subsequent post evaluation, mean, minimum, and maximum values were calculated.

OES.
For the optical emission spectroscopy (OES) measurements, a silica window was mounted in front of the effluent at 8.5 mm from the plasma jet's nozzle.An optical fiber was placed on the backside of the window and directed towards the nozzle.No collimator was used.The used spectrometer was a single-channel USB spectrometer (Avantes AvaSpec-3648-USB2), capturing a wavelength range from 200 to 1100 nm.The optical setup was mounted via optical bench system equipment, including micrometer screws for adjustment, on an optical bench.
While the present approach is designed to acquire an overview spectrum with commonly used systems for an easy comparison between different replicates of a plasma system, this system will not be able to address the excimer continuum emission that is emitting the VUV range.Here we refer to previous studies performed in these wavelength ranges for plasma jets [26][27][28].The neon excimer emission would be between 70 nm and 100 nm [29], while the absorption by ambient air diffusion, curtain gas and impurities have to be considered based on absorption coefficients.
The integration time of the software (Avantes AvaSpec8) was set to 100 ms, and an averaging of 20 samples was used for one record.For observation of changes over time in the emission spectra, a record time of at least 120 s was chosen.

Liquid diagnostics
For analyzing the effect of the plasma jet on the production of liquid phase reactive oxygen and nitrogen species, the plasma jet was placed perpendicularly in a holder, which was mounted to a xyz-table.In contrast to the other investigations, the liquid diagnostics were not only performed using neon with an addition of 0.3% O 2 as feed gas, but also plasma operated with pure neon was investigated.
For the treatment of the liquid, the so-called conductive mode was used, which means that the effluent of the plasma jet was in direct contact with the liquid [24].The spacing under which the conductive mode was reached differs for the different devices and was determined in parallel to the liquid treatment using the z-position of the xyz-table.
With this defined distance, the plasma jet was positioned above the liquid surface to treat a volume of 300 µl Dulbecco's phosphate-buffered saline (DPBS, without magnesium and calcium; Biowest) in a 96 flat bottom well plate for 60 s.Directly after plasma treatment, liquid samples were taken and subsequently analyzed by the different assays.

Hydrogen peroxide.
As representative stable reactive oxygen species (ROS), the hydrogen peroxide (H 2 O 2 ) concentration in plasma-treated liquid was analyzed by a commercial assay (Pierce Quantitative Peroxide Assay Kit, Thermo Scientific) based on the oxidation of ferrous (Fe 2+ ) to ferric (Fe 3+ ) ion in the presence of xylenol orange (125 µM) and sorbitol (100 mM).Due to the presence of sorbitol, H 2 O 2 is converted to peroxyl radical.This peroxyl radical oxidizes the Fe 2+ to Fe 3+ .Fe 3+ reacts under acidic conditions with xylenol orange forming a purple compound whose absorbance can be measured at 595 nm by a microplate reader (Tecan Infinite M200 Pro; Tecan).
The assay was performed in a 96 flat bottom well plate including standard dilution series from 300 µM H 2 O 2 to 0 µM in triplicates.Each treatment was performed three times (n = 3), and three samples, 10 µl each, were taken from each treatment.By adding 100 µl of the previous prepared working solution (1:100 ratio of Fe 2+ in sulfuric acid to the sorbitol and xylenol orange-containing solution), the reaction was started.An incubation time of 15 min was needed to complete the reaction.Afterwards, the absorbance was detected by the plate reader.By subtracting the blank and using the standard curve, the H 2 O 2 concentrations in the samples were deduced.Due to the low number of independent samples, the error bars represent the maximal and minimal deviation from the mean value.The assay was performed in a 96 flat bottom well plate including standard dilution series from 35 µM to 0 µM NO 2 − in triplicates.Each treatment was performed three times (n = 3), and three samples of 100 µl each were taken from each treatment.After adding 50 µl of each Griess Reagent 1 and 2, the plate was incubated for at least 10 min before the absorbance measurement was performed.By subtracting the blank and using the standard curve, the NO 2 − concentrations in the samples were deduced.Due to the low number of independent samples, the error bars represent the maximal and minimal deviation from the mean value.

Error estimation
The used micro flow controllers have an uncertainty of ±1% from the measured value.
Due to the HV generator, the systematic error for the determined frequency was ±2.5%, for the U rms the systematic error was ±5% as the sum of the uncertainty of the high voltage probe (±3%) and issues due to the cable (e.g.delay time) and connection.The error for the averaged energy per period was the minimal and maximal deviation of the mean value averaged for 16 periods.The error for the averaged power was ±8%, as sum of the error for the determination of U rms and the uncertainty of the measurement capacitor, which was checked and determined for 350 kHz and found to be 3%.Furthermore, no drift was obtained for this high frequency.
The PLC measurement error was, according to the manufacturer, ± 3 µA + 5% from the measured value.
The estimated error for the effluent length was 0.25 mm.For the temperature measurements, the mean value as well as the minimal and maximal deviation of 50 measurements were calculated and given in the results.
The used spectrometer for the OES measurements was wavelength calibrated and had an uncertainty of 0.5 nm.One recorded spectrum represents the average of 20 spectra.A background spectrum was recorded prior a measurement set and subtracted before evaluation to avoid stray light caused errors.
For the liquid analysis, the mean value, the minimal and the maximal deviation of in total nine samples (n = 3, with three samples taken from each repetition) were calculated.

Results and discussion
By comparing the electrical and optical parameters as well as by determining the concentration of reactive species in plasma-treated solution, the variation of crucial operating parameters between different devices of the plasma jet was studied.In figure 7(A), the frequency achieved with the system-specific self-tuning high-voltage generator is shown for six devices of the plasma jet under investigation at the beginning and at the end of the measurement procedure described in this paper.The whole measurement took about 30 min.Since in most cases a treatment lasts from 30 s to a maximum of 10 min, a continuous operation of 30 min can be used to evaluate the longevity stability.Different frequencies indicate different impedance and capacity values of the jet devices.The plasma jets 03, 04, 05, and 06 achieved similar values (between 413 kHz and 425 kHz).Compared to jets 01 and 02 380 kHz), they were operated with higher frequency.For all devices investigated, the measured operating frequency did not change with increased operating time.The root mean square (RMS) voltage (figure 7(B)) comparison leads to similar results.Plasma jets 01 and 02 gave slightly lower values than jets 03, 04, 05, and 06, with good stability over the whole experiment, as shown by the values for 'begin' and 'end' of the measurement.Also, the determined dissipated power (figure 7(D)) of the six plasma jets calculated according to equation (2) has a similar distribution.Jets 01 and 02 had the lowest power, whereas the values of jets 03-06 were again relatively close.Only the energy calculated according to equation (1) (figure 7(C)) was slightly different; jet 02 showed the lowest value, and 03-06 were almost equal.Here, only jet 01 behaved differently compared to the three other investigated electrical parameters.At the beginning of the experiment, the energy value of jet 01 was closer to that of jet 02 but at the end it was in the range of the group of jets 03-06, rather than jet 02.
A remarkable observation that can be made when looking at the complete picture provided by all the measurements shown in figure 7 is that, after the operation time, jet 02 can be clearly distinguished from the other devices, because it has a lower energy turnover per period and a lower average power dissipation than all the other individual jets.When including the measurements at the beginning of the measurement series, it  can be noticed that also jet 01 behaves differently: it has lower frequency, energy, and dissipated power values in comparison to jets 03-06.The RMS voltage at the beginning is even lower than the one reached with jet 02.The RMS voltage and the energy values at the end of the measurement of jet 01 increase, but they are still lower than the values measured with the other devices.Hence, also jet 01 did not behave similarly to jets 03, 04, 05, and 06.

PLC.
In figure 8, the PLC determined using a Cuplate is given.Interestingly, the value was the highest for jet 01 and jet 02, which were always the jets with the lowest values in all other electrical measurements.As in the previous electrical measurements, jets 03, 04, 05, and 06 behaved similarly over the investigated distance range.The determined PLC was below the upper limit for medical devices of 100 µA for AC and 10 µA for DC [1,30].Moreover, the obtained values for AC of the jets 04 and 06 were at the detection limit at a distance of 4 mm.All others, except for jet 01, reached the lower detection limit threshold at a distance of 5 mm in front of the plasma jet's nozzle.The overall slopes of all jets were comparable with other plasma jet measurements, yet at a different amplitude [1].The measured values for the DC portion of the PLC were at or below the lower detection limit of the Bender device.

Optical measurements 3.2.1. Effluent length.
Table 1 gives the measured effluent length for the different plasma jets.The shortest effluent length was obtained for jet 04 with 3.7 mm, followed by jet 05 and 06 with 4.6 mm.Interestingly, the plasma jets with the lowest energy and dissipated power, jet 01, had the longest visible effluents with 6.2 mm, respectively.Unfortunately, as the length of the effluent did not correlate to one of the other investigated parameters, it cannot be used as a 'quick-andeasy' observation parameter to analyze the plasma jets regarding their similar behavior.Yet, it has to be considered that the precise jet effluent length measurement was determined in open (i.e.non-conductive mode) operation (table 1, first row) while the other parameters were acquired with placing a target in front (i.e.conductive mode).In the conductive mode, the effluent length cannot be determined.Instead, the spacing between the plasma jet nozzle and the liquid surface was determined under which the jets operated in the conductive mode.These values are given in the second row of table 1.

Temperature.
The temperature of the investigated plasma jets was detected at two different positions.For one setting, the temperature was evaluated on a copper plate, mimicking a conductive target.For this setting, the target was moved away from the plasma jets' nozzle, starting the first measurement at 1 mm distance to the nozzle.Similar to the electrical parameters shown in the previous subsection (3.1.1),the treatment with plasma jets 01 and 02 started at the lowest temperature of the copper plate.For a distance up to 4 mm, these temperature values were relatively stable around 33 • C-34 • C. For jet 01, this behavior stays until 6 mm, whereas the temperature for jet 02 decreased by almost 10 • C. The other plasma jets showed a more dynamic behavior over the distance from 1 to 6 mm; Starting with a temperature between 38 to 40 • C at 1 mm, it dropped to 34 • C-36 • C at 2 mm, became stable between 3 to 5 mm at 33 • C-35 • C and decreased again for 6 mm distance to the nozzle of the plasma jet.At 6 mm the variation between the different devices was increasing.
In the effluent of the plasma jets, the obtained temperatures were different (see figure 9(B)).Again, the plasma jet 01 and 02 had the lowest temperatures (between 26 to 28 • C) As the most experiments in this study were performed in the conductive mode.The temperature measurements using the Cu-plate represented this treatment mode.In contrast to it, the effluent temperature was measured without a conductive target and therefore it can be used to represents the non-conductive mode.As one would expect the temperature in the conductive mode directly in front of the plasma jet nozzle was higher as the heat transfer was better compared to a non-conductive material.

OES.
In figure 10, the optical emission spectrum of jet 08 normalized to the highest peak at 703 nm (Ne emission line) is given as a representative example for all investigated plasma jets.In the ultraviolet range, molecular bands of the γsystem of the nitric monoxide radical ( • NO, A 2 Σ + ↔ X 2 Π), molecular bands of the Å-system of the hydroxyl radical ( • OH, A 2 Σ + -X 2 Π) as well as molecular bands of the second positive system of molecular nitrogen (N 2 , C 3 Π → B 3 Π g ) were apparent [31].From the visible (VIS) range to the nearinfrared (NIR) region, atomic peaks of Ne (predominantly 3p → 3s) [32] were dominant.In addition, also peaks of atomic oxygen ( • O) occurred in the NIR range at 777 nm (3p 5 P → 3s 5 S) and 844 nm (3p 3 P → 3s 3 S) [33].Beside the active working gases of neon, molecular oxygen and carbon dioxide contributing to the emitting species, even weak impurities within the working gas and remaining in the Teflon tubing show contributions to the emission spectra.In figure 11, the overview spectra of all investigated plasma jets are given.Each spectrum is normalized to the Ne peak at 703 nm as the dominant one in each spectrum.In the VIS/NIR range, recorded spectra appear overall similar, except for atomic oxygen peaks being more intense in some jets over others.The most prominent differences by comparison of these spectra were found in the UV range.
A closer look at the UV spectral range was chosen to evaluate additional details of the plasma jets' emission profiles (see figure 12).Two plasma jets, jet 01 and jet 03, showed the most intense emission of molecular bands.Jets 04, 05, and 06 also emitted • OH and N 2 molecular bands, whereas jet 02 had by far the lowest emission in the UV range.The • NO band was only well evaluable in the spectra of jets 01 and 03.In the spectra of jets 04, 05, and 06, the molecular band system of • NO was mainly identifiable due to the identification in the spectra of jets 01 and 03, but the overall intensities were relatively low.
These three molecular species ( • OH, • NO, N 2 ), together with • O, are highly important precursors for forming reactive species in the gas and liquid phase.• OH and N 2 are highly  relevant as precursor for the subsequent production of reactive oxygen and reactive nitrogen species in the liquid [34,35].Therefore, the behavior of five representative peaks for the four species ( • OH, • NO, N 2 , and • O) is shown in figure 13 for the different plasma jets.The given intensities were again normalized to the dominant peak of Ne (703 nm).The • O peaks, 844 nm and 777 nm, were for jets 01, 04, 05, and 06 almost identical in their intensity and only for jet 02 and jet 03 lower compared to the remaining four.By having a closer look at N 2 , • OH, and • NO, jet 03 had instead the strongest emission, whereas jet 02 was still the less intense one.The observed low intensity of the two • O peaks was not surprising since • OH, • NO, and O 2 are present.These are known to react with • O to form nitrogen dioxide ( • NO 2 ), ozone (O 3 , via a third body), or O 2 , including by-products such as atomic hydrogen ( • H) or atomic nitrogen ( • N) [36,37].
By comparing the species behavior for the different plasma jets with the frequency, the dissipated power, and the energy shown in figure 7(C), the trend is reproduced.As the depicted OES spectra give an impression of 20 averaged spectra, the values at the end of the measurements were used for comparison with the electrical parameter.For jet 03, the highest energy value (1.38 µJ) was determined.Jets 01, 04, 05, and 06 had values between 1.17-1.3µJ.Only jet 02 gave lower results (0.8 µJ).Concerning OES, jet 03 resulted in slightly more intense emission in the UV range than jets 01, 04, 05, and 06, whereas jet 02 emitted a lower intensity in that range.The combination of the electrical measurements of these three parameters-frequency, dissipated power, and energy-with OES, especially in the UV range, indicated partially pronounced differences in custom-built plasma jets.

Liquid diagnostics
To infer whether the jets' electrical and spectral differences translate to altered reactive species transport to a target, plasma-treated liquids were investigated.In figure 14, the generated H 2 O 2 concentration (figure 14(A)) and the NO 2 − concentration (figure 14(B)) in a buffered saline solution are given for the different plasma jets.
For NO 2 − , jets 01 and 02 differed substantially from all other plasma jets, which, in turn, showed similar concentrations (mean: 54.57µM).In the case of H 2 O 2 , only jet 02 produced lower (4.84 µM) concentrations than the other four plasma jets (mean: 130.23 µM).By comparison of the trend of the • OH peak for the different jets with the observed behavior for the H 2 O 2 concentration in the liquid for the jets, as shown in figure 15, similarities became visible.Besides jet 02, all other jets resulted in comparable normalized intensities of the • OH emission band as well as H 2 O 2 concentration.If the feed gas is humidified, the fact that H 2 O 2 in the plasma-treated liquid mainly originates from the gas phase is also known for the argon-driven atmospheric pressure plasma jet kINPen from a previous study [38].In the here presented study, the feed gas was not humidified.Hence, the H 2 O 2 was likely generated via dissociation of the treated water-based saline solution, DPBS.By comparison of the trend for the average energy per period (figure 15(A)) with the trends for • OH (figure 15(B)) and H 2 O 2 (figure 15(C)), a similarity is recognizable.This make sense, not only for the H 2 O 2 but as well for gaseous • OH, since it will be generated also by the dissociation of water molecules, but only in the gas phase [39].Instead, the observed behavior of the • NO and N 2 peaks (figure 13) compared with the NO 2 − concentration (figure 14(B)) for the different jets showed no relationship.Jet 01 showed similar normalized intensities of N 2 and • NO compared to jets 04 and 06 but resulted in lower NO 2 − concentrations being generated in the treated liquid.Furthermore, jet 03, which had the strongest emission for these two nitrogencontaining molecules, did not produce the highest NO 2 − concentration in the liquid phase.In another study [35], also investigating the kINPen, it was found that aqueous • NO was not created in the gas phase and, therefore, not transferred by solvation.As • NO is the main precursor of NO 2 − , the here obtained behavior for a different plasma jet pointed in the same direction.The previously discussed comparable behavior of energy and emission in the UV range was also reflected in the liquid chemistry, especially by the NO 2 − concentration.Although jet 05 was not included in the liquid investigations presented here, as it was already in a collaborator's laboratory at the time of measurement, it was tested with pure neon as feed gas and compared with the other five plasma jets (data not shown).The behavior of all jets under this condition was similar to the data given here, with jet 05 fitting into the series of jets 03, 04 and 06.The absolute values were slightly higher due to the absence of O 2 in the feed gas.
It should be considered, however, that in conductive conditions (i.e. the plasma jet being in direct contact with the treatment target), RNS reactive species generation and deposition dramatically increase, as recently observed for the kIN-Pen plasma jet [24].
H 2 O 2 and NO 2 − are known to be important in the regulation of biological systems [40].These two species can be measured relatively easily by colorimetric assays, allowing them to be determined in almost any laboratory.Therefore, they are ideally suited for regular determination as quality control parameters.Furthermore, H 2 O 2 and NO 2 − can be used as first hints for the more reactive, and therefore, more difficult to determine, ROS and RNS such as radicals.

Correlation of the results
To verify that the electrical, optical, and liquid data were correlated, a Spearman correlation test was performed using the correlation coefficient calculator in OriginPro 2022b.The calculated ρ values are shown in table 2. The values highlighted in green have a ρ value above |ρ| > 0.7, and the bold values have significance at the 0.05 level.
The parameter with the most significant correlations was the PLC, which showed a significant correlation with all other electrical parameters as well as with the two temperature values.
Unfortunately, the OES measurements did not correlate significantly with any of the other measurements performed.
The liquid analysis showed several correlations with the electrical measurements performed, although none of the correlations were significant.The strongest correlation value for NO 2 − , with ρ = −0.87,was found for PLC, while for H 2 O 2 a strong correlation with PLC was also found, with ρ = −0.87, in addition a perfect correlation, with ρ = 1, could be found for f end , E end and P end .
From the correlation evaluation it can be concluded that the electrical parameters, especially the PLC, in combination with the liquid analysis of at least one of the stable species can be recommended to evaluate the reproducibility and could serve as a quality control parameter set.
Furthermore, the use of OES spectra for reproducibility and variance control proved to be insufficient in the present design.Considering that OES describes purely the plasma neglecting the electrical system components, it is highly suited to characterized the discharge.However, for a comparison of multiple devices, the interplay of the plasma discharge as for instance, a resistor and a capacitor with the electrical circuit components can thus not be detected via OES, while the PLC detects the current passed through the electrical system in total.This might be an indication, why the electrical measurements outscore the optical plasma spectroscopy.However, more detailed application of OES could yield better result and could be considered for further investigation.

Summary and conclusion
By comparison of the electrical data, namely frequency, voltage, power dissipation, and leakage current, a remarkable variation between the individual devices of the jet was observed.In contrast, the species output was in the same order of magnitude for most devices, apart from one more and one less apparent exception.Consequently, the studied plasma jets show no remarkable dependency between electrical quantities and species output.By comparing the liquid phase reactive species output, all jets but the two exceptions delivered similar concentrations of H 2 O 2 and NO 2 − .These results suggest that the investigated plasma source type is suitable for a distribution to different laboratories.Although the variation of operational parameters in this type of in-house manufactured plasma jet is not negligible, the comparability and reproducibility of experiments performed with different devices of the plasma jet are ensured regarding the reactive species output.
Two devices with deviating characteristics regarding the species output required a closer examination.Distinctive features of both jets were significantly lower for, e.g.frequency, dissipated power, and energy per period.Furthermore, the one device with obvious deviating performance, jet 02, could also be recognized from the optical emission spectra, as the emission bands of ˙OH and N 2 in the ultraviolet range showed a remarkably lower intensity than the other jets.Unfortunately, this is not valid for jet 01, another device behaving differently.Therefore, the OES spectra alone are not an appropriate tool for quality control.This output was further verified by the correlation test, where the OES spectra showed no relevant correlation to one of the other investigated parameters.
Consequently, by screening the electrical parameters, especially the dissipated power, a first hint of unwanted differences can be obtained, which need further verifications by an additional measurement.In the field of plasma life science, for instance liquid chemistry analysis or biological experiments can be a useful addition.
In general, it can be stated that an indicator from one phase is not enough to identify faulty plasma sources, but in a combination of electrical, optical, and liquid characterization, the unwanted hidden differences can become visible.For the here investigated type of plasma jet, lower threshold values for the frequency, the dissipated power, and the energy can be set, for instance, to 400 kHz, 100 mW, and 1 µJ, respectively.
One reason for the differences in the plasma devices' performance could be an 'aging' of the sources, as jet 01 was used the most compared to the other five devices.Perhaps the used materials or the combination of them, for instance, the epoxy used to glue the last copper wire windings at the nozzle weaken over time.This would explain why jet 01 performed worse than the other, but it did not explain why jet 02 behaved worse as well.In the conductive mode, the jet plasma is a combination of a guided streamer and a glow like discharge.Therefore, it could be that for plasma jets 01 and 02, the glow like discharge occurred more frequently and therefore the RMS power would be lower due to the more frequent voltage drops.This would also explain why the PLC was higher for these two jets, having a more frequently established conductive channel to the surface and resulting in a higher current being transported to the target.These hypotheses need further investigation to identify critical and non-critical steps for better reproducibility in plasma jet construction and performance.Since the PLC seems to be one of the parameters suitable for assessing reproducibility, it is highly relevant whether measurements and treatments were performed in conductive or non-conductive mode.Moreover, it is already known that the different modes have a great influence on the species yield as well as on the biological effect of a treatment [24,41], underscoring the importance of these modes and even more relevant for localized, comparable treatments [15].
Finally, in the present study, control parameters such as the PLC and the analysis of relatively stable reactive species in the aqueous liquid were identified to verify the reproducibility and variability of custom-built sets of plasma jets.

Figure 1 .
Figure 1.Scheme of the miniature neon plasma jet in front of a grounded target.

Figure 2 .
Figure 2. Equivalent circuit to measure the dissipated power.

Figure 3 .
Figure 3. (a) Voltage waveform over one entire clock cycle.(b) One sinusoidal period with the resulting charge signal within the active phase as a detail.

Figure 4 .
Figure 4. QV plot made from the signals in figure 3(b).

Figure 5 .
Figure 5. Equivalent circuit for leakage current measurement.

Figure 6 .
Figure 6.Image of the effluent for length calculation, with intensity distribution of the effluent in the red channel of the image.

Figure 7 .
Figure 7. Frequency ((A), orange diamonds), RMS voltage ((B), purple upwards oriented triangles), average energy per period ((C), brown left oriented triangles), and average power (D), rose right oriented triangles) for the different devices of plasma jets at begin (op open symbols) and at the end (closed symbols) of the investigations.

Figure 8 .
Figure 8. AC-patient leakage current measured over a grounded copper plate placed at different distances.DC-leakage was below the detection limit in all cases (1 µA).

Figure 9 .
Figure 9. Temperature of the target (copper plate; (A)) and in the effluent of the plasma jets (B) in relation to the distance between the device and copper plater; the ambient room temperature for the measurement was 22 • C.

Figure 10 .
Figure 10.Optical emission spectrum of jet 08 with the indication of the observed lines.

Figure 11 .
Figure 11.Comparison of the optical emission spectra as overview of all investigated plasma jets.

Figure 12 .
Figure 12.Detailed view of the ultraviolet range of the optical emission spectra of all investigated plasma jets.

Figure 13 .
Figure 13.Behavior of normalized selected peaks of relevant (reactive) species or their precursors shown for the six plasma jets.Ne 703 nm is used for normalization per spectra.

Figure 14 .
Figure 14.Detected concentrations of H 2 O 2 (A) or NO 2 − (B) in DPBS after 60 s plasma treatment in conductive mode with five different plasma jets.

Figure 15 .
Figure 15.Comparison of the trend for the averaged energy per period (A), the • OH peak around 308 nm (B) and the determined H 2 O 2 concentration in plasma-treated DPBS (C).
2.4.2.Nitrite.Besides H 2 O 2 , also nitrite (NO 2 − ), as stable representative of reactive nitrogen species (RNS), was measured using a commercial assay (Nitrate/Nitrite Colorimetric Assay Kit; Cayman Chemicals).This assay is based on the Griess reaction where NO 2 − is converted by Griess Reagents to a deep purple azo compound.The absorbance of this compound was measured afterwards by a multiplate reader (Tecan Infinite M200 Pro; Tecan) at 540 nm.

Table 1 .
For the non-conductive mode, the effluent length and, for the conductive mode, the spacing between the plasma jet nozzle and the liquid surface, of the different plasma jets is given.

Table 2 .
Correlation evaluation for the examined parameters.Fields highlighted in green indicate a very high correlation (|ρ| > 0.7) according to the Spearman correlation test.Correlation values that are significant at the 0.05 level are shown in bold.