**Emission Spectroscopic Characterization of a Helium Atmospheric Pressure Plasma Jet with Various Mixtures of Argon Gas in the Presence and the Absence of De-Ionized Water as a Target**

#### **Nima Bolouki 1,\* , Jang-Hsing Hsieh 1,2, Chuan Li <sup>3</sup> and Yi-Zheng Yang <sup>2</sup>**


Received: 25 April 2019; Accepted: 3 July 2019; Published: 4 July 2019

**Abstract:** A helium-based atmospheric pressure plasma jet (APPJ) with various flow rates of argon gas as a variable working gas was characterized by utilizing optical emission spectroscopy (OES) alongside the plasma jet. The spectroscopic characterization was performed through plasma exposure in direct and indirect interaction with and without de-ionized (DI) water. The electron density and electron temperature, which were estimated by Stark broadening of atomic hydrogen (486.1 nm) and the Boltzmann plot, were investigated as a function of the flow rate of argon gas. The spectra obtained by OES indicate that the hydroxyl concentrations reached a maximum value in the case of direct interaction with DI water as well as upstream of the plasma jet for all cases. The relative intensities of hydroxyl were optimized by changing the flow rate of argon gas.

**Keywords:** atmospheric pressure plasma jet; plasma characterization; optical emission spectroscopy

#### **1. Introduction**

Cold atmospheric plasma devices, mainly based on the atmospheric pressure plasma jet (APPJ) [1], have emerged over recent decades. Such devices offer the possibility of direct and indirect (remote) methods for bacteria inactivation [2], biofilm control [3], cancer cell treatment [4], water purification [5], plasma activated water [6], and so on. The APPJ with a tube-shaped configuration fed by helium gas was presented by Laroussi et al. [7] for the first time. Then, the APPJ was developed and extended with different electrode configurations and working gases.

Generally, an APPJ with various mixtures of feed gases provides numerous types of reactive species with different concentrations [8,9]. One of the species is a hydroxyl radical, which provides excellent benefits for biological and medical applications although it has a short lifetime [10,11]. Some articles reported the characterization of the argon-based APPJ with a mixture of water to enhance the hydroxyl species [12–14]. Argon as a working gas might be considered as an alternative candidate since scarce and costly helium is consumed in a large volume in the APPJ. The argon APPJ has better energy transfer efficiency compared to the helium jet; however, it releases considerable heat [15]. The conditions of electrical discharges in a pure argon gas might be unstable as a result of changing the parameters related to experimental conditions. For example, an additive gas such as oxygen is able to extinguish the argon discharge and shrink the range of discharge stabilities due to low electron temperature compared to the helium discharge [15,16]. Therefore, a plan of mixing argon and helium gases is proposed not only to reduce helium consumption but also to enhance the concentration of reactive species, specifically hydroxyl.

Based on the specific applications, the APPJ-based devices are designed and constructed with different power sources, geometries, and working gases, which results in changing the electrical conditions of the APPJ as well as the produced reactive species. Moreover, interactions of the plasma jet with liquid targets, which are most of the biological cases, influence the plasma conditions and the produced reactive species. Therefore, it is undoubtedly necessary to measure the plasma parameters to characterize a homemade APPJ in interaction with liquid.

Optical emission spectroscopy (OES) [17], as an affordable and a non-intrusive diagnostics method with an easy experimental setup, is used to identify the plasma parameters and reactive species produced by an APPJ. In this case, the radiations emitted by excited atoms, molecules, and reactive species in the plasma source are collected and analyzed to determine the plasma parameters, such as electron density, electron temperature, gas temperature, and so on. The Boltzmann plot [18] is an established method to estimate the electron temperature by assuming that the plasma condition is in a state of partial local thermodynamic equilibrium due to high collision frequency between particles in atmospheric pressure [19]. The upper levels of the atomic transitions follow the Saha–Boltzmann distribution. Hence, the excitation and electron temperatures are assumed to be the same, although this is an inappropriate assumption in mid and low pressures [20,21]. Stark broadening of the Balmer series lines of atomic hydrogen (Hα, Hβ, Hγ) [18], which are broadenings or shifts of the spectral lines due to the presence of the electric fields of charged particles, allows us to calculate the electron density in atmospheric pressure. Electron density is almost independent of electron temperature for a given broadening of the H<sup>β</sup> line, while in the cases of H<sup>α</sup> and Hγ, this dependency is obvious specifically for H<sup>α</sup> [22]. Therefore, the H<sup>β</sup> broadening (486.1 nm) is utilized to diagnose the electron density. The Stark broadening can be considered as a reliable method for the values of electron density higher than 10<sup>19</sup> m-3. Otherwise, less than this value, the contributions of Doppler and Van der Waals would be dominant, leading to a high error range in the estimation of electron density [23].

In addition to measurements of electron density and electron temperature, the neutral gas temperature plays a vital role in plasma characterization and processes. For instance, in biological applications, the high gas temperature of the APPJ is capable of damaging biological cells and tissues. Generally, there is no specific threshold temperature for heating damage, which depends on the type of biological cells and tissues; however, in biological testing, the temperature should be less than 42 ◦C [24]. Therefore, to confirm whether the APPJ might be suitable for the desired application, measurements of the gas temperature would be necessary. Also, to estimate the electron density by the Stark effect, the contributions of Van der Waals and Doppler broadenings, which depend on the gas temperature, should be considered. Hence, knowing the gas temperature is needed. In atmospheric pressure, it is assumed that the rotational temperature of the second positive system of nitrogen gas is equal to the neutral gas temperature [25].

In this study, to characterize and optimize the APPJ, the relative intensities of reactive excited species have been measured using OES. The electron temperature has been estimated by the Boltzmann plot alongside the plasma jet. Since the Balmer series line of atomic hydrogen has been too weak, the measurements of Stark broadening have been carried out at the bottom of the plasma jet. While the flow rate of helium gas is held constant, the influences of argon gas with various flow rates on relative intensities of species and plasma parameters in the presence and the absence of de-ionized (DI) water as a target have been investigated upstream, midstream, and downstream of the plasma jet. The rotational temperature of the second positive system of nitrogen gas has been measured to estimate the neutral gas temperature downstream of the plasma jet in direct interaction with DI water.

#### **2. Materials and Methods**

#### *2.1. Experimental Setup of the Atmospheric Pressure Plasma Jet*

The experimental setup of the APPJ is shown in Figure 1. The setup consists of a quartz tube with two copper electrodes in a cylindrical shape with a gap distance of 15 mm. A high voltage DC pulsed power supply delivers a monopolar pulsed voltage with a square wave-form pulse. The on-time and off-time of the pulsed voltage were adjusted to be 25 μs, while the rise and fall time was adjusted to be 3 μs. A voltage probe (Rigol-RP1018H) and a current probe (Cybertek-CP8030B) were used to measure the applied voltage and the plasma current. The voltage and current waveform were recorded using an oscilloscope (Rigol DS1054z, 50 MHz, 1 GS/s). During the experiment, helium as a working gas with the flow rate of 5 slm (standard liter/min) remained unchanged, while the argon flow rate as a variable parameter of the working gas was controlled and adjusted to be 0–2000 sccm (standard cubic centimeter/min) by mass flow controller (MFC). The experiments based on the target situation were carried out in three cases. The first case represents the APPJ without de-ionized (DI) water. The second and third cases refer to the presence of DI water as a target exposed by the APPJ directly and indirectly. The maximum distance between the downstream position of the APPJ and the surface of DI water in the case of the indirect plasma exposure was selected to be 4 mm. The APPJ was mounted on an adjustable stage and, thus, the distance of 4 mm was adjusted by the stage.

**Figure 1.** Experimental setup of the helium APPJ with various flow rates of argon gas. The spectroscopic characterization was performed alongside the plasma jet in direct and indirect interaction with and without DI water.

#### *2.2. Electrical Measurements*

Figure 2 shows the voltage and current characteristics of the discharge. The maximum values of applied pulsed voltage and frequency were adjusted to be 8.5 kV and 17.8 kHz, respectively. The flow rate of argon gas of the recorded voltage and current was 1600 sccm. It should be noted that the voltage–current waveform was examined with different flow rates of argon gas. The waveform of the discharge did not change significantly.

**Figure 2.** Voltage–current characteristic of the helium-based APPJ with the flow rate of 1600 sccm of argon gas.

#### *2.3. Spectroscopic Measurements*

The emission spectra of the APPJ was measured using a spectrometer (Avantes, AvaSpec-2048L, focal length of 75 mm, grating with line density of 300 mm<sup>−</sup>1, entrance slit of 25 μm, 2048-pixel CCD detector) with a spectral range of 200–1100 nm and a spectral resolution of 1.4 nm. A fiber optics cable including a collecting lens was used to capture the light emitted from the APPJ. The emission spectra were recorded for 5 accumulations with an exposure time of 100 ms. The measuring points of the electron temperature were selected to be upstream, midstream, and downstream of the APPJ as shown in Figure 1. Since the hydrogen Balmer line in the wavelength of 486.1 nm (Hβ) was too weak to be detected in the plasma jet, the spectrometer was placed at the bottom of the plasma jet to enhance the signal intensity of Stark broadening related to the hydrogen Balmer line for measuring the electron density.

#### **3. Results and Discussion**

#### *3.1. Measurements of Relative Intensities of Species*

Figure 3 presents the spectra of optical emission of the plasma jet obtained by OES in the cases of non-contact (free of DI water—Figure 3a,b), indirect contact (Figure 3c,d), and direct contact (Figure 3e,f) with DI water. The spectra were measured with and without argon feed gas (flow rate of 1600 sccm) downstream and upstream of the jet stream. Hydroxyl (309 nm), helium (706 nm), and argon (750 nm) species, as well as the second positive system of nitrogen gas, were identified in the spectra. As mentioned before and shown in the figure, in all cases, the hydrogen Balmer line (Hβ) is too weak to estimate the electron density.

**Figure 3.** The emission spectra of the plasma jet with and without argon gas in the cases of (**a**) and (**b**) non-contact, (**c**) and (**d**) indirect contact with DI water, and (**e**) and (**f**) direct contact with DI water. The measurements have been performed at the downstream and the upstream of the plasma jet.

Figure 4 shows the spatial profile of relative intensities of hydroxyl (309 nm), helium (706 nm), and argon (750 nm) species of the APPJ obtained by OES with different flow rates of argon gas in the cases of free of DI water (Figure 4a), and indirect contact (Figure 4b) and direct contact with DI water (Figure 4c). The figure shows that relative intensities of hydroxyl are higher upstream of the jet compared to the other measured positions for all cases. In the case of the indirect contact with DI

water, shown in Figure 4b, the relative intensity of hydroxyl increases more. The enhancement of the hydroxyl concentration is boosted in the case of direct contact with DI water alongside the plasma jet. The intensity value of hydroxyl species rises more than two times compared to the case of free of DI water by considering the values shown in Figure 4a,c. Moreover, the relative intensities of hydroxyl species reach the maximum amount at the flow rate of 1600 sccm specifically at the upstream of the plasma jet; then, the intensities of hydroxyl decrease for all cases. In the case of direct contact, at the flow rate of 2000 sccm, the spectrum was not available to measure the intensities of the species as the length of the plasma jet decreased to half. Since the flow rate of helium gas was held constant during the experiment, the intensities of helium species remained nearly unchanged, as shown in Figure 4, for all cases.

**Figure 4.** *Cont.*

**Figure 4.** The relative intensity of hydroxyl (309 nm), helium (706 nm), and argon (750 nm) obtained by OES as a function of the flow rate of argon gas in the cases of (**a**) non-contact with DI water; (**b**) indirect contact with DI water; and (**c**) direct contact of DI water.

#### *3.2. Measurements of Gas Temperature*

Although the second positive system of nitrogen bands were obvious in all cases (Figure 3), the intensity of the second positive system was strong without argon gas feed in the case of direct contact with DI water as shown in Figure 3e,f. As mentioned above, the nitrogen band allows us to measure the temperature of the neutral particle based on the rotational temperature of nitrogen gas. By fitting the emission line spectra of the second positive system (C <sup>3</sup>Π<sup>u</sup> <sup>→</sup> <sup>B</sup> <sup>3</sup>Π<sup>g</sup> <sup>=</sup> 375–381 nm) [26], the gas temperature was measured to be 0.03 eV, equal to 75 ◦C. Thus, the APPJ is in the non-thermal state and regarded as a cold plasma. In addition, the ambient temperature was measured by a mercury thermometer at a distance of around 10 mm from the tip of the plasma jet for 2 minutes. The temperature was recorded to be 33.2 ◦C.

#### *3.3. Measurements of Electron Density and Electron Temperature*

Figure 5 illustrates the spatial profile of the electron temperature estimated by the Boltzmann plot upstream, midstream, and downstream of the APPJ in the cases of free of DI water (Figure 5a), and indirect (Figure 5b) and direct contact with DI water (Figure 5c). The measured parameters are plotted with different flow rates of argon gas. The estimated electron density based on Stark broadening in the case of free of DI water (no target) is shown in Figure 5a. At the flow rate of 400 sccm, the electron density and the electron temperature were estimated to be 1.2 <sup>×</sup> 1022 m−<sup>3</sup> and 0.25–0.45 eV alongside the plasma jet. The electron density reached the maximum value of 2.3 <sup>×</sup> 1022 m−<sup>3</sup> at the flow rate of 1600 sccm, then decreased to the flow rate of 2000 sccm. However, the trend of electron temperature was downward between the flow rates of 400 to 2000 sccm, and finally, reached the values of 0.15–0.21 eV at the flow rate of 2000 sccm. Injecting the massive species to the plasma discharge leads to a change in the energy transfer between electrons and the species due to the collision frequency that causes a reduction in electron temperature. The electron temperature converges more by increasing the flow rate of argon gas in the case of direct contact; nevertheless, the trend of electron temperature is almost the same for all cases. In the case of direct contact, at the flow rate of 2000 sccm, the spectrum was not available to estimate the electron temperature as the length of the plasma jet decreased to half.

**Figure 5.** Spatial profile measurements of the electron temperature upstream, midstream, and downstream of the APPJ in the cases of (**a**) non-contact with DI water; (**b**) indirect contact with DI water; and (**c**) direct contact of DI water. Measurement of electron density was undertaken at the bottom of the jet stream.

The Boltzmann plot equation is assumed to be ln(*I*λ/*gA*) = −*E*/*kT*<sup>e</sup> + constant, where *I* is the measured intensity obtained by OES, λ is the selected wavelength associated with helium and argon gases, *g* is the statistical weight, *A* is the transition probability, *E* is the excitation energy corresponding to the selected wavelengths, *k* is Boltzmann constant, and *T*<sup>e</sup> is electron temperature. The selected wavelengths of atomic emission lines related to argon and helium gases were obtained from references [12,27] to estimate the electron temperature based on the Boltzmann plot. The spectroscopic data related to the wavelengths of excited atoms and ions were confirmed via the NIST atomic database [28]. By adding argon gas, the contributed wavelengths of helium [27] for the estimation of electron temperature weakened; therefore, the wavelengths of argon [12] were used at the flow rate of 400 to 2000 sccm.

Regarding Stark broadening, although hydrogen gas was not used as a working gas in the APPJ, it initially might result from the impurity of helium and argon working gases or the ambient air. The relationship between electron density and Stark broadening is assumed to be *<sup>n</sup>*<sup>e</sup> <sup>=</sup> (ΔλFWHM/(2 <sup>×</sup> <sup>10</sup><sup>−</sup>11))3/2, where <sup>Δ</sup>λFWHM is Stark broadening and *n*<sup>e</sup> is electron density in terms of per cubic centimeter. The broadening of the spectral line is a convolution of instrumental, Doppler, and Van der Waals broadenings [29]. The Voigt fitting of the measured Stark bordering was performed using Origin Pro 5. The instrumental broadening was obtained by replacing the APPJ with a mercury lamp (wavelength of 546.1 nm) at low pressure. This value was obtained to be 0.06 nm. Finally, the Voigt fitting was excluded from the broadenings mentioned above to estimate the pure Stark broadening at full width at half maximum (FWHM) for measurements of electron density.

The presence of massive particles such as argon species in the helium plasma changes the plasma chemistry and the degree of ionization. Based on the results of relative intensities of excited species obtained by OES, the hydroxyl concentration is highest upstream of the plasma jet compared to the other regions. The production of hydroxyl might be explained by two possible different mechanisms of reactions [30]. The first mechanism corresponds to the direct dissociation of the water molecule by energetic electrons. Based on the results, by increasing the flow rate, the electron temperature drops. So, the energy of the electrons is too low (less than 1 eV) to dissociate the water molecule. The second refers to the possibility of participation of metastable argon to produce H2O<sup>+</sup>. In this case, the dissociative recombination of H2O<sup>+</sup> leads to hydroxyl generation. Therefore, the metastable atoms may play an important role to generate hydroxyls [14,31]. That is the reason the concentration of hydroxyl is highest upstream of the plasma jet for all cases, as the concentration of the metastable species of argon is highest in the same space. In addition, as the plasma jet approaches the water surface, the humidity around the plasma jet enhances due to the water evaporation that causes a boost of hydroxyl generation in the presence of water. Yang et al. [30] reported similar results by using the helium-based APPJ with the fixed flow rate of helium gas, while in our case, the flow rate of the APPJ changed by argon gas.

#### **4. Conclusions**

We characterized and optimized a helium APPJ with a mixture of argon gas in the presence and the absence of DI water. DI water as a target was exposed by the APPJ directly and indirectly. Electron temperature and electron density as a function of the flow rate of the gas mixtures were investigated. In a direct interaction of the jet stream and surface of DI water, the gas temperature was estimated based on the rotational temperature to confirm the plasma condition is in the non-thermal regime. The OES results show that approaching the APPJ so that the jet stream would have an indirect interaction with the water surface leads to more hydroxyl concentration with respect to the free plasma case. Hydroxyl production could be boosted more in direct contact with DI water. Considering positive points and drawbacks of using pure helium or argon gases to produce the APPJ, mixing these two gases might be the right solution for biological and environmental applications.

**Author Contributions:** Formal analysis, N.B.; Investigation, Y.-Z.Y.; Supervision, J.-H.H. and C.L.; Writing–original draft, N.B.; Writing–review & editing, N.B.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Experimental Investigation on the Influence of Target Physical Properties on an Impinging Plasma Jet**

**Emanuele Simoncelli <sup>1</sup> , Augusto Stancampiano 1,2 , Marco Boselli <sup>1</sup> , Matteo Gherardi 1,3 and Vittorio Colombo 1,3,\***


Received: 15 July 2019; Accepted: 11 September 2019; Published: 16 September 2019

**Abstract:** The present work aims to investigate the interaction between a plasma jet and targets with different physical properties. Electrical, morphological and fluid-dynamic characterizations were performed on a plasma jet impinging on metal, dielectric and liquid substrates by means of Intensified Charge-Coupled Device (ICCD) and high-speed Schlieren imaging techniques. The results highlight how the light emission of the discharge, its time behavior and morphology, and the plasma-induced turbulence in the flow are affected by the nature of the target. Surprisingly, the liquid target induces the formation of turbulent fronts in the gas flow similar to the metal target, although the dissipated power in the former case is lower than in the latter. On the other hand, the propagation velocity of the turbulent front is independent of the target nature and it is affected only by the working gas flow rate.

**Keywords:** impinging jet; metal/dielectric/liquid substrate; ICCD imaging; high-speed Schlieren imaging; plasma discharge morphology; turbulence

#### **1. Introduction**

The physical and chemical properties of a cold atmospheric pressure plasma (CAP) jet are not uniquely dependent on the plasma source configuration and operational parameters, but also on the target characteristics. The complex mutual interaction between the plasma and the target has recently been the subject of an increasing number of papers, investigating how the targets significantly affect the plasma properties, such as fluid-dynamics [1–3], electromagnetic field [4,5], reactive and excited species production and distribution [6], and ionization front velocity and propagation [1]. Especially, electrical properties, such as conductivity and potential, play a major role [4,7]. During direct plasma treatment, the target becomes part of the transient electrical circuit, connecting the power supply, the plasma source, the plasma, the target and the return to ground [8,9]. This means that the target's electrical parameters (conductivity, potential, etc.) play a major role in determining treatment conditions. Darny et al. revealed how, with conductive substrates like metal, the electric field adjacent to the substrate is enhanced, allowing for a restrike discharge that in turn can greatly enhance the production of reactive species that play an important role in biological applications [10]. Modelling studies by Norberg et al. also predicted similar modifications of the electric field based on substrate conductivity [11]. As reported by Yuanfu et al. [7], the production of OH radicals over a metal substrate can be ten times higher than in the case of a dielectric substrate. In light of the wide range of industrial, biomedical and agricultural applications, the interaction of atmospheric plasma jets with liquids can be of crucial interest to the scientific community. Liquid substrates, having both

capacitive and conductive components, show a hybrid behavior between the purely dielectric and the purely conductive materials [1,7]. This behavior is modulated by the conductivity of liquid solutions that can greatly vary in a range of several order of magnitude (10−7−10−<sup>2</sup> <sup>S</sup>/cm).

In this frame, the present work provides a direct comparison of the behavior of an atmospheric pressure plasma jet when interacting with a conductive liquid solution, or with dielectric and metal substrates, using ICCD and high-speed Schlieren as imaging techniques. The aim is to gather new insight into the fluid-dynamic behavior of this configuration; an aspect which has not been deeply investigated in the literature, despite being of great influence in plasma assisted processes.

#### **2. Materials and Methods**

#### *2.1. Plasma Source*

The plasma source adopted in this work was a single electrode plasma jet developed at the University of Bologna, Italy, and already described, characterized, and applied in previous works by Colombo et al. [1,2,12]. The plasma source was driven by a commercial nanosecond pulse generator (FPG 20-1NMK, FID GmbH). The electric conditions used for all experiments were 15 kV as the peak voltage (PV) and 125Hz as the pulse repetition frequency (PRF). The main voltage pulse supplied by the generator lasted around 35 ns with 10 ns as rise time. A residual damping signal was recorded up to 100 ns, thus the discharge event lasted less than 100 ns. The ignition of the discharge was repeated every 8 ms as a consequence of imposed PRF.

The high-voltage pin electrode (a stainless steel needle; Ø 0.3 mm) was centered inside a dielectric channel, and a flow rate of 3 slpm of helium gas (99.999% pure) was injected through a 12 hole (Ø 0.3 mm) diffuser. The plasma was ejected from the source into the surrounding atmosphere through an orifice with a diameter of 1 mm, producing a visible plasma plume and possibly interacting with a substrate. All experiments were performed in controlled ambient air at 30 ◦C. To estimate the electrical power dissipated by the plasma source, voltage and current waveforms were recorded on the high-voltage connection powering the plasma source by means of a high-voltage probe (Tektronix P6015A, sensitivity of 0.018%) and a current probe (Pearson 6585, sensitivity of 1%) connected to an oscilloscope (Tektronix DPO40034). The power density was evaluated using the following formula:

$$SPD = PRF \cdot \int V \cdot I dt\tag{1}$$

where *PRF* is the pulse repetition frequency, *V* is the applied voltage, and *I* is the current (the voltage and current signal are related to the main positive peak).

#### *2.2. Substrates*

The plasma source was positioned vertically at 10 mm (fixed gap) above the substrate surface. Different targets characterized by different conductivities were selected for this study, ranging from metal and liquid to dielectric substrates. A stainless steel plate (7.6 cm × 7.6 cm × 1 cm) was chosen as the target with almost infinite conductivity (~3.7 <sup>×</sup> 10<sup>11</sup> <sup>μ</sup>S/cm). On the other hand, to simulate a non-conductive substrate, a 7.6 cm × 7.6 cm × 1 cm PVC plate was used as the dielectric substrate (~10−<sup>17</sup> μS/cm). As far as the liquid substrate was concerned, since the electrical conductivity of the target affects the plasma characteristics [13,14], and in turn the plasma treatment may alter the electrical conductivity of the treated liquid solutions [15,16], the liquid target was prepared as a phosphate buffer solution (made by dissolving sodium phosphate dibasic (Na2HPO4) and potassium phosphate monobasic (KH2PO4) in distilled water). A solution characterized by an electrical conductivity of 119 μS/cm with a pH of 7.2 was realized. The physical properties (conductivity and pH) were monitored and remained unaltered during the experiments. The solution volume was 120 mL contained in a vessel (7.6 cm × 7.6 cm × 2 cm) with quartz sidewalls and an aluminum bottom. Since the metal substrate and the aluminum bottom of the liquid substrate vessel were connected to ground through

a low impedance electrical connection, both could be considered at ground potential. The dielectric substrate was instead positioned on a grounded metal plate to control and fix the associated capacitance. According to the literature [9,17], the equivalent electrical circuits associated with the three substrates result in a single resistance for the metal substrate, a single capacitance for the dielectric substrate, and a resistance and capacitance in parallel for the liquid substrate. Presenting very different electrical characteristics, the three substrates are a good representation of a wide range of possible targets.

#### *2.3. Diagnostic Techniques*

The morphology and the time evolution of the plasma discharge interacting with each substrate were investigated by means of an ICCD camera (Princeton Instruments PIMAX3, spectral response 180–900 nm) equipped with a conventional macro lens (Sigma Dg-Ex-APO-If 180 Mm/F3.5, spectral response 380–900 nm). To synchronize the image acquisition with the discharge event, the synchronization of the camera gating, driven by the voltage pulse, was performed by employing a delay generator (BNC 575 digital pulse/delay generator) and taking into account all possible signal transmission delays, as already described in the literature [12]. An overview of the iCCD imaging setup is showed in Figure 1.

**Figure 1.** ICCD imaging setup.

The overall emission intensity, produced by the plasma discharge during the main voltage pulse, was acquired by imposing both the camera gate exposure (35 ns) and the duration of the voltage pulse. The images were accumulated 30 times with a gain factor set at 50. The time evolution of the plasma discharge during the whole voltage pulse was investigated through the capture of sequential 10 ns camera gate exposures (each frame resulting from 30 accumulations). The first ICCD gate opening (0 ns) was imposed, so as to center on the start of the rising front of the voltage pulse (see Figure 3, top). The ICCD camera gates were superimposed on the excitation voltage waveforms. The reproducibility of the discharge was first verified with single shot (no accumulations) acquisitions (data not shown).

The characterization of the fluid-dynamic of the impinging jet was performed using the Schlieren imaging technique [18]. The Schlieren setup is shown in Figure 2 and was composed of a 450 W ozone free xenon lamp (Newport-Oriel 66355 Simplicity Arc Source) as a light source, a slit and an iris diaphragm, two parabolic mirrors with a focal length of 1 m, a knife edge positioned vertically because the highest gradients of the refractive index around the axis of the jet were horizontal, and a high-speed camera (Memrecam GX-3 NAC image technology). The camera was operated at 8000 fps with 1/200,000 s shutter time. The plasma source was positioned halfway between the two parabolic mirrors.

**Figure 2.** Schlieren imaging setup.

#### **3. Results**

#### *3.1. Electrical and Time-Resolved ICCD Characterization of the Plasma Jet Impinging on Di*ff*erent Substrates*

Table 1 shows the measured values of the electrical power, dissipated by the plasma source interacting with metal, dielectric and liquid substrates, respectively. Although the input operating conditions were the same for all investigated cases, the plasma jet impinging on a metal target dissipated the highest power (0.434 W) in comparison to the other targets.

**Table 1.** Electrical power dissipated by the plasma source for metal, dielectric and liquid substrates.


In this section, the results for the ICCD imaging of the plasma jet plume impinging on different substrates are presented.

Figure 3 shows, for each condition, a single accumulated image of the discharge obtained by capturing the light intensity emitted during the whole main voltage pulse (duration of ~35 ns). The acquisitions highlight how the plasma, and therefore its Visible–Near InfraRed (Vis-NIR) light emission, was highly influenced by the nature of the substrate. The strongest intensity and largest width of the plasma columns were recorded for the case of a metal target (fourth image from the left in Figure 3). On the other hand, the lowest intensity was observed when the plasma jet impinged on the liquid substrate (second image from the left in Figure 3). Furthermore, a spreading of the plasma across the target surface, known as surface ionization wave (SIW), was clearly observed only in the case of the dielectric substrate. For the liquid substrate, the plasma discharge appeared to be focused on a small area, corresponding to the dimple created in the liquid by the jet effluent (first image on the left in Figure 3).

**Figure 3.** ICCD images of the discharge emission related to the dielectric, liquid and metal substrates. (The first image on the left, taken with a longer exposure time, is presented to show the dimple formation on the liquid substrate).

In Figure 4, the sequences of ICCD images (10 ns exposure) for both dielectric and metal cases are shown.

The acquisitions (Figure 4) highlight how the temporal evolution of the plasma Vis-NIR light emission was influenced by the nature of the target. The propagation velocity of the ionization front was estimated to be <sup>≥</sup> <sup>2</sup> <sup>×</sup> <sup>10</sup><sup>8</sup> cm/s, as the crossing of the gap took place in the first two acquisitions.

For both substrates, the peak of the emission intensity was achieved in the acquisition at 20 ns, corresponding to the reach of the maximum applied voltage (Figure 4, third acquisition from the left for both cases). In the case of the dielectric target, the images show how the Vis-NIR light emission remained approximately constant during the whole voltage pulse. In contrast, on the metal substrate, the plasma emission significantly changed with time; therefore, the higher the applied voltage, the more intense the light emission.

**Figure 4.** Time-resolved ICCD imaging of the plasma jet plume impinging on the dielectric and metal substrates.

The ICCD images in Figure 4 reveal how, in the case of a metal substrate, the plasma plume did not show a clear SIW formation, but the discharge was quite focused on a point on the surface. On the other hand, in Figure 3 the image related to the metal substrate shows a weak SIW upon the surface, as a consequence of a time integration of 35 ns corresponding to the imposed exposure time. As shown by the frames 5 and 20 ns in Figure 4, (metal substrate) photons emission covered a wider area compared to the frames of 40 and 70 ns.

In the ICCD time-resolved images shown in Figure 4, the SIW formation and development were clearly enhanced in the case of the plasma jet interacting with the dielectric target, due to its higher capacitance. The dielectric surface was charged more than the metal one, favoring a higher spreading of the plume over it.

#### *3.2. Time-Resolved Schlieren Characterization of the Plasma Jet Impinging on Di*ff*erent Substrates*

Figure 5 presents results for the high-speed Schlieren imaging of the plasma jet plume impinging on the dielectric, liquid and metal substrates. The frames were selected with the aim of emphasizing the most important steps of spatial evolution of the turbulent front induced by the discharge, as follows: the discharge event, the turbulent front exiting the nozzle, the turbulent front approaching the substrate, its impact with the substrate surface, and its expansion upon the surface.

The duration of the plasma discharge emission (around 100 ns, as described in Section 2.1) was entirely captured in the first frame (0 ms), since the exposure of the high-speed camera was set at 0.005 ms. The time values reported on each frame were indicative of the time lapse between the acquisition of the first frame and those that followed.

In the case of the metal target, a turbulent front was observed at the outlet of the plasma jet in the first image, while for all the other cases, in the first frame (0 ms) the He gas flow seemed completely laminar, similar to the case of He gas flow without plasma ignition (data not presented). For dielectric and liquid substrates, the acquisitions at 0 ms showed how during the plasma discharge, the He gas flow was laminar with no flow modifications visible. Nonetheless, a significant perturbation of the He gas flow was clearly visible in the following frames, several tens of microseconds after the plasma discharge and the imposed voltage pulse ended. The turbulent phenomena is thus ignited by the discharge event, and its dynamic behavior is directly affected by the nature of the target.

At 0.25 ms after the end of the plasma discharge, in all investigated cases, a transient turbulent structure appeared as a consequence of an induced alteration of the helium gas characteristics, such as temperature and density, inside the plasma source during the discharge. The turbulent front coming out of the nozzle propagated downstream with a velocity close to that of the gas flow (~60 m/s). Once this transient turbulent structure reached the target surface, eddies were generated and propagated above the surface, away from the column axis. In all investigated cases, a laminar flow was finally re-established (data not shown) in the He column before the next discharge was ignited (8 ms period for PRF 125 Hz).

The strongest turbulent front was observed in the case of the metal substrate. The turbulent structures remained visible more than 5 ms after the discharge event (data not shown). In the case of the liquid substrate, the turbulent phenomenon that was induced was well recognizable. On the other hand, when the plasma jet impinged on a dielectric substrate, the induced turbulence was weaker, and the turbulent structures propagated over the surface were rapidly dissipated by interaction with the surrounding air. While the amplitude and the intensity of the turbulence were dependent on the target nature, the velocity of propagation of the turbulent front appeared independent of the target nature and governed by the gas flow rate only.

Finally, the gas impinging on the liquid substrate caused the formation of an axisymmetric dimple on its surface. The effects of impingement of a gas jet on a liquid surface have been studied in detail in different scientific and industrial fields [19–21]. As described in previous work [1], the impact of the turbulent front upon the liquid surface causes a variation of the dimension of the dimple in time.

#### **4. Discussion**

According to the results, the influence of the electrical properties of the material on the discharge characteristics of an impinging plasma jet is clearly visible. Materials with very low conductivity, such as the dielectric PVC plate (~10−<sup>17</sup> μS/cm), favor the accumulation of charges on their surface. The total charge accumulated increases with the capacitance of the target [22] and, in our case, was significantly higher for the dielectric case in comparison with the metal and liquid substrates. The charge accumulation in turn induces the development of a radial electric field driving the formation of a SIW [11]. This results in the symmetrical radial expansion over the surface of the discharge, as clearly visible in Figure 4. SIW phenomenon is undetectable in the metal and the liquid cases, due to their much higher conductivity (~3.7 <sup>×</sup> 1011 <sup>μ</sup>S/cm for the metal and 119 <sup>μ</sup>S/cm for the liquid) that prevents the accumulation of charges on the surface.

On the other hand, the conductivity of the substrate also limits the power deposited by the plasma discharge. After the impinging of the ionization front on the substrate surface, a highly conductive channel connects the high-voltage electrode inside the jet source to the grounded substrate [10]. Once this connection occurs, the current passing through the plasma channel is limited by the conductivity of the target and the capacitance associated with it. As demonstrated by the power measurements (Table 1), the metal substrate, with a conductivity several order of magnitude higher than the other samples, recorded the highest dissipated power, albeit the values of measured power were in the same order of magnitude for all investigated cases. Moreover, although the liquid, with a higher conductivity, presented the lowest power, we hypothesize that other phenomena and factors, such as the target capacitance (higher for the dielectric), could have played a key role in the frame of ns-discharge interacting with a target. As was visible in the dielectric case, the plasma discharge propagated over the target surface and the emission intensity increased to t = 20 ns when the voltage pulse reached its peak. This suggests that the dielectric substrate, due to its capacitance, was still charging and, therefore, only partially limiting the current flow in the conductive channel. The liquid substrate on the other hand, characterized by a conductivity much lower than the metal (nearly 11 orders of magnitude) and a capacitance much lower than the dielectric, presented a singular behavior; no SIW was formed and there was a limitation of the deposited power. It must be noted that the liquid substrate is also the only substrate among those investigated to induce a partial modification of the surrounding gas composition due to evaporation. As reported by Ji et al. [23] in a similar configuration, the presence of a liquid film on a target presents a reduced impact on the characteristics of the plasma discharge propagation and intensity, while on the other hand, it can greatly enhance the production of OH radicals in close proximity to the target surface.

Concerning the plasma-induced alteration of the gas flow, the turbulent front could be clearly seen in all considered cases, but with different characteristics. In the case of the metal substrate, a turbulent alteration of the gas flow was visible from the discharge event (frame 0 ms in Figure 5). Later, a turbulent front propagated downstream with a velocity comparable with the gas flow's, as already observed in similar conditions in previous work [1]. Since the total expansion of the turbulent front was sufficiently fast enough to observe a re-established laminar He flow before the ignition of the next discharge event, we can confirm that the turbulent front phenomenon is directly associated with a combination of plasma-induced pressure waves and local electrode heating [1,24,25]. Moreover, because the Schlieren images (Figure 5) clearly showed that the magnitude of the turbulent structures varied between the investigated cases, we can hypothesize that turbulent front formation and its propagation is affected by the target characteristics. It may be speculated that there is a relation between turbulences and the target modification of the electric field distribution between the high-voltage electrode inside the plasma source and the targets [23]. In virtue of their conductivities and connections to ground, the metal and liquid targets greatly affected the electric field distribution, while the dielectric PVC target had a relative permittivity comparable to that of air (εr−PVC = 3), thus having a limited impact.

#### **5. Conclusion**

The present study highlights, through an ICCD and Schlieren imaging analysis, how the plasma plume changes its morphology and its light intensity as a consequence of the physical properties of the target, and how the fluid-dynamic of the plasma-induced turbulent front is affected by the substrate's nature. The proposed comparison between metal, dielectric and water substrates highlights significant differences between the associated plasma discharges. The electrical characteristics of the substrates may influence not only the plasma discharge propagation and ionization degree, but also the gas flow dynamics. Nanosecond pulsed plasma, interacting with a liquid substrate, presents a singular behavior that is not easily located midway between highly conductive and dielectric materials. Thus, future analysis should be directed to the investigation of substrate materials covering a wider range of conductivities, especially in the range of semiconductive materials. This would certainly help to better identify possible trends and provide better explanations. Concerning the transient turbulence generated in the gas flow by the plasma discharge, it has highlighted a dependence on the target nature. This aspect should be taken into account during plasma treatments, since it may affect through mixing of the gas composition in the plasma region, and may induce pressure fluctuations on the target surface. These aspects may hinder the uniformity of treatments, especially on complex and non-homogeneous surfaces like biological tissues.

The results of this work aim to be considered as one step closer to a full understanding of the complex interaction of non-equilibrium plasma with a substrate.

**Author Contributions:** Conceptualization, E.S., A.S., M.G. and V.C.; Data curation, E.S. and A.S.; Funding acquisition, V.C.; Investigation, E.S., A.S. and M.B.; Methodology, E.S., A.S. and M.B.; Project administration, V.C.; Resources, V.C.; Supervision, M.G. and V.C.; Writing—original draft, E.S., A.S. and V.C.; Writing—review & editing, E.S., A.S., M.G. and V.C.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors would like to acknowledge Romolo Laurita for his contribution in liquid solution preparation and Eng. Alina Bisag for 3D rendering of the experimental setups.

**Conflicts of Interest:** The authors declare no conflicts of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Hydrogen Peroxide Interference in Chemical Oxygen Demand Assessments of Plasma Treated Waters**

**Joseph Groele 1,\* and John Foster <sup>2</sup>**


Received: 16 May 2019; Accepted: 2 July 2019; Published: 5 July 2019

**Abstract:** Plasma-driven advanced oxidation represents a potential technology to safely re-use waters polluted with recalcitrant contaminants by mineralizing organics via reactions with hydroxyl radicals, thus relieving freshwater stress. The process results in some residual hydrogen peroxide, which can interfere with the standard method for assessing contaminant removal. In this work, methylene blue is used as a model contaminant to present a case in which this interference can impact the measured chemical oxygen demand of samples. Next, the magnitude of this interference is investigated by dosing de-ionized water with hydrogen peroxide via dielectric barrier discharge plasma jet and by solution. The chemical oxygen demand increases with increasing concentration of residual hydrogen peroxide. The interference factor should be considered when assessing the effectiveness of plasma to treat various wastewaters.

**Keywords:** non-equilibrium plasma applications; water treatment; advanced oxidation

#### **1. Introduction**

Demand for freshwater is rapidly rising due to expanding agriculture and industrialization; meanwhile, population growth, changing climate, accidental spills, and deteriorating infrastructure further exacerbates the problem of freshwater scarcity [1]. One possible approach to managing this impeding crisis is to re-use wastewater by removing contaminants; however, wastewaters contain contaminants of emerging concern (CECs) which are not readily removed by traditional water treatment processes. These CECs include pharmaceuticals, industrial chemicals such as per- and polyfluoroalkyl substances (PFAS), pesticides for agriculture, and microcystins [2–6]. Fortunately, these CECs can be removed from wastewaters using advanced oxidation processes [7].

Advanced oxidation processes (AOPs) are a category of chemical treatment methods for removing persistent organic pollutants from waters and wastewaters via reactions with highly reactive oxidizing agents, namely hydroxyl radicals. Traditional AOPs typically leverage combinations of ozone, hydrogen peroxide, and ultraviolet light to facilitate in-situ generation of hydroxyl radicals for non-selective decomposition of organic contaminants [8,9]. Through a series of chain reactions known as mineralization, hydroxyl radicals react with organics to ultimately form carbon dioxide, water, and mineral ions, with aldehydes and carboxylic acids often serving as intermediate breakdown products [10].

Plasma-liquid interaction can generate hydrogen peroxide, ozone, and other reactive species along with ultraviolet radiation and short-lived highly reactive hydroxyl radicals, thus representing a novel AOP for the removal of recalcitrant organic compounds from waters and wastewaters [11–13]. The formation of these reactive species is initiated primarily through energetic electrons from the plasma region colliding with atoms or molecules, either in the gas phase or at the liquid surface. Subsequent production of reactive species can also be facilitated by recombination of radicals and de-excitation of metastables [14].

Plasma discharges in contact with water can generate multitudes of reactive oxygen species, including superoxide, hydroperoxyl, and atomic oxygen. If the discharge is in air, then reactive nitrogen species will also form, including nitric oxide, nitrite, nitrate, and peroxynitrate [15]. While all of these reactive oxygen and nitrogen species may contribute to degradation of contaminants in wastewaters, the most important species for advanced oxidation are hydroxyl radicals, hydrogen peroxide, and ozone. In particular, hydroxyl radicals react non-selectively with organics, including contaminant intermediate products such as short chain carboxylic acids, to allow for complete mineralization of most organic pollutants [16].

The production of hydroxyl radicals in plasmas in or in contact with water can occur through a number of different pathways, with the relative importance of these pathways depending on the plasma parameters and the gas composition [17]. Some of these hydroxyl radicals produced in the discharge diffuse to the liquid surface and are transported across the gas-water interface to either react with pollutants in the water or scavenge themselves to form longer lived hydrogen peroxide [12,18].

In the bulk liquid, hydrogen peroxide can react with any dissolved ozone from the discharge to produce more hydroxyl radicals (i.e., peroxone process [19]) for in-situ oxidation of organics, even after the plasma has been turned off. However, not all of the hydrogen peroxide is consumed through reactions with ozone, and a portion of the post-discharge hydroxyl radicals produced by the peroxone process will dimerize back into hydrogen peroxide. In this way, plasma-driven advanced oxidation can leave residual hydrogen peroxide in treated waters that may persist for days, as can also happen with traditional AOPs [20].

The goal of plasma-driven advanced oxidation, and AOPs in general, is to chemically remove contaminants from waters. The five most prominent methods for assessing contaminant removal are liquid chromatography tandem mass spectrometry (LC-MS), spectrophotometry, total organic carbon (TOC), biochemical oxygen demand (BOD), and chemical oxygen demand (COD) [21–24]. Spectrophotometry and LC-MS allow for the determination of specific species concentrations, whereas TOC, BOD, and COD are non-specific water quality parameters that provide a measurement of the overall pollution potential of a water sample. When initially investigating the performance of AOPs for removal of contaminants from wastewaters, LC-MS should be used to validate results from other methods; however, for general monitoring of effluent quality in water and wastewater treatment processes, TOC, BOD, and COD are more commonly used due to the simplicity of the test procedure and easily interpreted results.

As the name suggests, TOC is a measure of the total organic carbon contained in a water sample. This measurement involves two stages: total carbon (TC) and inorganic carbon (IC) measurements, with the difference providing the TOC of the sample. First, combustion catalytic oxidation of the sample converts the organic carbon to carbon dioxide, which is then cooled and humidified before being detected by an infrared gas analyzer to measure the TC content of the sample. The second step involves acidifying the sample to pH < 4, making bicarbonate and carbonate unstable [21,25]. Thus, the IC in the sample is converted to carbon dioxide, either free CO2(aq) or in the form of carbonic acid, that can be purged with a CO2-free gas to isolate the carbon dioxide, which is subsequently cooled, humidified, and detected by the gas analyzer as the IC measurement.

While TOC measurements focus on the carbon content of a sample directly, BOD and COD are measurements of the amount of oxygen consumed in the complete oxidation of organics to carbon dioxide, water, and mineral salts. As such, BOD and COD are particularly well suited for monitoring water treatment processes involving oxidation of organics and for informing the design of the water treatment process (e.g., how much oxidant must be used to remove the contaminants present in the waste stream).

More specifically, BOD is a measure of the oxygen consumed by bacteria in the oxidation of organics and is representative of contaminant decomposition in the natural environment, making this method most suitable for predicting the effects of the organic contaminants on the dissolved oxygen levels of receiving waters. In contrast, COD is a measure of the oxygen consumed in the chemical oxidation of organics [23]. The BOD test typically takes five days to get results, meanwhile the COD test can provide results in less than three hours.

Due to the short analysis time and ability to correlate with BOD, the COD test has become the industry standard for rapid and frequent monitoring of water treatment process efficiency and effluent quality. The most common method is the closed reflux, colorimetric method with potassium dichromate oxidizing agent [22]. Pre-formulated commercial reagent mixtures are available for rapid COD analysis and provide consistent results between different laboratories. The problem with COD assessments of plasma treated wastewaters is that residual peroxide interferes with the measurement. Indeed, this interference is an issue for any AOP involving residual peroxide [20]. This paper first discusses a case in which this interference can lead to misinterpretation of plasma treatment results, and then investigates the magnitude of the interference in COD assessments of waters containing residual hydrogen peroxide.

#### **2. Materials and Methods**

#### *2.1. Underwater Dielectric Barrier Discharge Plasma Jet*

The discharge configuration used to bring non-equilibrium plasma in contact with water in this investigation is an underwater dielectric barrier discharge (DBD) plasma jet. The set-up is shown in Figure 1 and is based on the design by Foster et al. [26]. The plasma applicator consists of an 18 gauge tin-plated copper high-voltage wire centered coaxially within a quartz tube with 4 mm ID, 6.35 mm OD, and 150 mm length, and a tin-plated copper wire coil wrapped around the outside of the quartz tube to serve as the grounded electrode. In this case, the quartz acts as the dielectric barrier preventing a thermal arc from forming.

**Figure 1.** (**a**) Schematic of the DBD plasma jet experimental set-up used in this work. (**b**) Schematic and image of the DBD plasma jet apparatus operating in steam-mode discharge in de-ionized water.

The central high-voltage electrode is excited with a 5 kHz sinusoidal voltage at 4 kVp-p, provided by an Elgar model 501SL power supply (AMETEK, Inc., Berwyn, PA, USA) with a 50:1 step-up transformer (SP-225 Plasma Technics, Inc., Racine, WI, USA). Voltage was measured using a high-voltage probe (P6015A, Tektronix, Beaverton, OR, USA), and the discharge current was measured with a Pearson coil

current monitor (6585, Pearson Electronics, Inc., Palo Alto, CA, USA). Current and voltage waveforms were recorded using a 2 GHz oscilloscope (Wavepro 7200a, Teledyne LeCroy, Chestnut Ridge, NY, USA). Typical current-voltage data is shown in Figure 2. Data from previous Lissajous experiments by Garcia et al. with the DBD plasma jet operating in steam-mode at these power supply conditions indicate that about 56 W of power are deposited in the discharge, with peak steam temperatures around 2800 K, as determined from comparing theoretical simulation of OH(A-X) spectra with experimental optical emission spectra [27].

**Figure 2.** Discharge voltage and current signals measured during operation of the underwater DBD plasma jet operating in steam-mode at 5 kHz.

This DBD plasma jet can be operated with or without gas flow. In this work, the plasma jet operates without any input gas. This so-called "steam mode" discharge relies on local evaporation near the tip of the high-voltage electrode to form a vapor bubble that acts as the low-density region for plasma formation. Although a portion of the deposited power goes into heating and evaporating water, operating in steam-mode requires no consumables and minimizes the production of NOx, as evidenced from optical emission spectroscopy [28,29], which can interfere with the iodometric titration method for quantifying hydrogen peroxide concentration, discussed in Section 2.3.

#### *2.2. Sample Preparation*

Samples of de-ionized water (Milli-Q EMD Millipore, Burlington, MA, USA, electrical conductivity ≈ 3 μS/cm) are dosed with hydrogen peroxide either by solution (30%, Fisher Chemical, Hampton, NH, USA) using a variable volume sampler pipette (Thermo Fisher Scientific, Waltham, MA, USA, resolution: 1 μL) or by underwater DBD plasma jet. When dosing with the DBD plasma jet, the opening of the quartz tube is placed approximately in the center of a 50 mL sample of de-ionized water and power is provided to the central high-voltage electrode. Approximately one second is required for the vapor bubble to form at the electrode tip in which the plasma discharge subsequently develops.

Plasma is applied to the sample via the DBD plasma jet for 30 s at a time, and the increase in sample temperature throughout the treatment duration is measured using a Fluke 87V true RMS multimeter with a type-K thermocouple (Fluke Corporation, Everett, WA, USA, resolution: 0.1 ◦C, accuracy: 0.05%). After 30 s of treatment, the sample is placed in an ice bath to cool down to room temperature before further treatment. The objective was to keep the bulk liquid sample below 60 ◦C. After treatment, the samples are stored in a dark cabinet for approximately 20 h before the hydrogen peroxide concentration and COD are measured to allow post-discharge reactions to take place while mitigating hydrogen peroxide photo-dissociation.

Samples of methylene blue (MB) are prepared by dissolving solid MB powder (M291-25 Fisher Chemical) in de-ionized water (Milli-Q EMD Millipore). A low concentration sample of 5 mg/L and a high concentration sample of 1000 mg/L are prepared and treated with the underwater DBD plasma jet to demonstrate the hydrogen peroxide interference problem, discussed in Section 3. Plasma treatment of the MB samples follows the same procedure as described for dosing de-ionized water samples with hydrogen peroxide by DBD plasma jet, including monitoring the temperature to keep the sample below 60 ◦C. Again, treated MB samples were stored in a dark cabinet for 20 h before assessing the COD.

#### *2.3. Hydrogen Peroxide Measurement*

The hydrogen peroxide concentration in the sample is quantified using the iodometric titration method (Hach test kit model HYP-1, Hach Company, Loveland, CO, USA, resolution: 1 mg/L H2O2) immediately prior to COD assessment. The titration method involves the oxidation of iodide to iodine (Equation (1)) in the presence of a molybdate catalyst and a starch indicator. The starch indicator forms a dark blue complex with iodine. Subsequent titration with thiosulfate under acidic conditions (approximate pH of 4) converts the iodine back to iodide (Equation (2)), and the color change from blue to transparent indicates the titration endpoint.

$$\rm H\_2O\_2 + 2KI + H\_2SO\_4 \rightarrow I\_2 + K\_2SO\_4 + 2H\_2O \tag{1}$$

$$\text{Na}\_2 + 2\text{Na}\_2\text{S}\_2\text{O}\_3 \to \text{Na}\_2\text{S}\_4\text{O}\_6 + 2\text{NaI} \tag{2}$$

$$2\text{I}^- + 2\text{NO}\_2^- + 4\text{H}^+ \rightarrow \text{I}\_2 + 2\text{NO} + 2\text{H}\_2\text{O} \tag{3}$$

This iodometric titration method is subject to interference from nitrite ions (Equation (3)); thus, nitrogen species from air plasma discharges will interfere with this measurement. To mitigate the generation of reactive nitrogen species that could result in aqueous nitrite, the DBD plasma jet is operated in steam mode, as discussed in Section 2.1. For future studies, colorimetric assay by titanyl sulfate reagent with azide for elimination of nitrite interference is a preferred method of hydrogen peroxide quantification due to superior selectivity, making it suitable for hydrogen peroxide concentration measurements in waters treated with air plasmas [15,30].

#### *2.4. Chemical Oxygen Demand Measurement*

The assessment of COD in this work was performed using the USEPA 4.10.4 approved method, which is the closed reflux, colorimetric method with potassium dichromate oxidizer. Pre-formulated reagent was purchased from Hanna Instruments (HI9375A-25 COD reagent low range: 0 to 150 mg/L as O2, accuracy: ±5 mg/L or ±4% of reading at 25 ◦C, resolution: 1 mg/L) along with a Hanna Instruments photometer (HI83399, Hanna Instruments, Woonsocket, RI, USA) for measuring sample absorbance using an LED light source with narrow band interference filter at 420 nm.

#### **3. Results and Discussion**

To illustrate a picture of the problem, a 100 mL sample of 5 mg/L MB is prepared and the COD is assessed before and after treatment. The results are shown in Table 1. In general, the COD is expected to decrease with plasma treatment time as the MB molecules are mineralized. Indeed, the color of the dye disappears after 15 min of treatment indicating the destruction of MB, as seen in Figure 3a; yet, the measured COD actually increases with treatment time. No additional organics are being added to the solution by the plasma; thus, one of the plasma-derived species being transported to the liquid during treatment must be contributing to the measured COD of the sample.


**Table 1.** Measured COD values for untreated and treated samples of MB. MB-A samples (left) contain initial MB concentration of 5 mg/L while MB-B samples (right) start with 1000 mg/L MB.

**Figure 3.** (**a**) Images of samples starting with 5 mg/L MB showing decrease in color indicating destruction of the MB dye. (**b**) Image of untreated sample with 1000 mg/L MB and sample after 40 min of plasma treatment showing MB precipitates and mostly clear supernatant.

The two primary long-lived plasma-derived species that could be present in the sample after plasma treatment are hydrogen peroxide and ozone, since nitrates and nitrites should not be generated under steam-mode operation. Of these, hydrogen peroxide has been found to interfere with the standard COD assessment by reducing the potassium dichromate oxidizing agent according to the overall oxidation reaction given in Equation (4) [31]. From this equation, the theoretical hydrogen peroxide interference is calculated to be 470.6 mg of COD as O2 per 1000 mg H2O2.

$$\rm K\rm \rm Cr\_2O + 3H\_2O\_2 + 4H\_2SO\_4 \to K\_2SO\_4 + Cr\_2(SO\_4)\_3 + 7H\_2O + 3O\_2 \tag{4}$$

Researchers investigating plasma treatment of wastewaters must be particularly careful when interpreting COD results because the interference from residual hydrogen peroxide may not be evident. For example, two 100 mL samples of 1000 mg/L MB were prepared and treated with the underwater DBD plasma jet. After treatment, the COD decreased from the initial value of 1680 ± 75 mg/L, as shown in Table 1, and the supernatant becomes transparent as oxidized MB precipitates out of the solution, as shown in Figure 3b. This case of high initial COD relative to the residual peroxide concentration masks the peroxide interference effect. Still, any residual peroxide left in the sample after treatment will contribute to the COD. Hence, the COD contribution from organics should be less than the 99 mg/L after 40 min of treatment indicated in Table 1, since there is some contribution from residual peroxide.

In this case of high initial dye concentration (MB-B Sample, Table 1), the residual peroxide contribution to the COD may have been up 99 mg/L, but since the COD decreases from 1680 mg/L to 99 mg/L, it appears that the process is working as intended (COD decreases as MB is oxidized). In this case, the COD of the residual peroxide is at least 17 times less than the COD of the organics present in the sample before treatment, thus the interference effect is not immediately apparent. In reality, the treatment process is likely working better (faster organic removal rate) than would be indicated by the COD test because of the interference from residual peroxide. However, in the case of low initial dye concentration (MB-A Sample, Table 1), the measured COD increases with plasma treatment time

because the COD contribution from residual peroxide after treatment is greater than the COD of the organics present in the sample before treatment. By simply looking at the measured COD before and after treatment, it would appear as if the process did not work (i.e., organics were not removed), because the peroxide interference has not been corrected for.

To investigate the magnitude of hydrogen peroxide interference in COD assessments, de-ionized water was dosed with hydrogen peroxide both by solution and by DBD plasma jet according to the procedure described in Section 2.2. The hydrogen peroxide concentration is measured immediately before COD assessment. In both cases, the COD increases with increasing hydrogen peroxide concentration, as shown in Figure 4, even though there were no organics present in the samples. This proportional relationship was also found by Lee et al. for conventional advanced oxidation processes, although the interference magnitude varied depending on the water quality and treatment process [20].

**Figure 4.** (**a**) COD vs. H2O2 concentration, where H2O2 is added to samples by a solution of 30% H2O2. (**b**) COD vs. H2O2 concentration, where H2O2 is added to samples by exposure to the underwater DBD plasma jet.

These results seem to confirm that the measured COD results from residual peroxide reacting with the potassium dichromate oxidizing agent during the COD assessment. The magnitude of the interference found in this experiment is approximately equal to the theoretical value from Equation (4), but this value can vary in practice by up to 16% percent depending on the specific organic compounds and oxidizing species that are present in the sample being assessed, particularly in real wastewaters [20]. When AOPs are used to remove trace levels of contaminants from relatively low COD waters, as can occur at final disinfection stages of drinking water treatment plants, the uncertainty in the actual interference magnitude can lead to misinterpretation of results. While reagents may be used to quench residual hydrogen peroxide in attempt to remove this interference, these reagents often introduce their own interferences with the COD assessment [31]. More research is needed to investigate how the residual oxidants and other long-lived species from plasma treatment affect the COD measurement in various synthetic and real wastewaters in order to develop correction methods for appropriate interpretation of results.

#### **4. Conclusions**

Wastewater treatment plants use COD to measure treatment process efficiency and quality of treated water. Plasma-water interaction produces residual peroxide which interferes with the COD measurement. As a result, COD assessments of plasma treated waters will show incomplete removal of oxygen demand and may be viewed as a disadvantage of plasma-based advanced oxidation for municipal wastewater treatment plants and companies looking to treat their wastewater streams. Indeed, when researching new wastewater treatment technologies, LC-MS should be used to determine residual organic content before applying standard water quality diagnostics. This paper therefore describes precautions necessary for plasma practitioners to take into account so as to yield credible contaminant decomposition measurements. Plasma in contact with water produces a host of reactive oxygen species, and so caution is necessary when interpreting the results.

**Author Contributions:** Investigation, analysis, and writing, J.G.; conceptualization, resources, and supervision, J.F.

**Funding:** This research was funded by the National Science Foundation (NSF 1700848) and the U.S. Department of Energy (DOE DE-SC0001939).

**Conflicts of Interest:** The authors declare no conflicts of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Ignition of a Plasma Discharge Inside an Electrodeless Chamber: Methods and Characteristics**

#### **Mounir Laroussi**

Electrical & Computer Engineering Department, Old Dominion University, Norfolk, VA 23529, USA; mlarouss@odu.edu; Tel.: +1-757-683-6369

Received: 31 July 2019; Accepted: 8 October 2019; Published: 14 October 2019

**Abstract:** In this paper the generation and diagnostics of a reduced pressure (300 mTorr to 3 Torr) plasma generated inside an electrodeless containment vessel/chamber are presented. The plasma is ignited by a guided ionization wave emitted by a low temperature pulsed plasma jet. The diagnostics techniques include Intensified Charge Coupled Device (ICCD) imaging, emission spectroscopy, and Langmuir probe. The reduced-pressure discharge parameters measured are the magnitude of the electric field, the plasma electron number density and temperature, and discharge expansion speed.

**Keywords:** plasma jet; electric field; plasma bullet; diffuse plasma; ionization wave

#### **1. Introduction**

Low temperature plasma jets exhibit large electric fields at the tip of their plasma plumes. The magnitude of this field was measured by several investigators and was found to be in the 10–30 kV/cm range [1–4]. Figure 1 shows measurements of the electric field reported by Begum et al. [1] while Figure 2 shows simulation results reported by Naidis [5]. When impinging on a dielectric material, the plasma plume causes charge build up on the surface of the dielectric surface. Therefore, via capacitive coupling, the electric field can be effectively transmitted to the area behind the dielectric barrier [6–8]. Under reduced pressure, the transmitted field can be large enough to start a discharge, which quickly propagates and grows in volume [9].

**Figure 1.** Mean magnitude of the electric field along the axis for a plasma plume emitted by a pulsed plasma jet [1].

**Figure 2.** Electric field axial profile at different times for a helium plasma jet emerging into surrounding air. The jet radius is 0.25 cm. Figure adapted from Figure 1 of [5].

In this paper the characteristics of the plasma ignited inside an electrodeless dielectric chamber by an external plasma jet are reported. The chamber has no direct physical or electric connections to the externally applied plasma jet. The diagnostics techniques used, which include emission spectroscopy, fast imaging, and Langmuir probe, allow for the measurement of the magnitude of the transmitted electric field inside the chamber, the plasma electron number density and temperature, and discharge propagation speed.

#### **2. Materials and Methods**

Low temperature plasma jets launched in the ambient environment are enabled by guided ionization waves [10]. The electric field at the front of these waves can be quite large and can be transmitted across a dielectric barrier. If the dielectric barrier constitutes the wall of a chamber where the pressure can be controlled, then a diffuse plasma can be generated inside the chamber below a certain pressure threshold [9]. This principle was used here to generate a reduced pressure plasma inside a Pyrex cross-shaped tube that has no electrodes or electrical connections. Figure 3 shows the experimental setup and a photo of the reduced pressure plasma inside the Pyrex tubing.

**Figure 3.** *Cont.*

**Figure 3.** Schematic of the experimental setup (top) and photo of the large volume plasma inside the tubing/chamber (bottom). The plasma jet is applied externally on the far side of the chamber. The chamber is a Pyrex glass cross with an arm 18 inches long (inner diameter is 6 inches) and a second arm 16 inches long (inner diameter is 4 inches). For this photo, the pressure inside the chamber was 3 Torr and the gas was air.

The tubing/chamber is made of Pyrex glass, where the pressure can be reduced gradually down to the mTorr range. The plasma jet is physically independent of the chamber and is applied externally. It is driven by repetitive narrow pulses (ns–μs) with magnitudes in the kV range. The tip of the plasma plume is brought against the external wall of the chamber to ignite a plasma at reduced pressure.

#### **3. Results and Discussion**

For all the experiments described below, the pulsed power supply and its associated circuitry were all placed inside a Faraday cage located in a separate location, away from the chamber and the diagnostics equipment, and all cables were properly shielded. All spectroscopic and Langmuir probe measurements presented here reflect average values. No time- and space-resolved measurements are reported. In addition, in order to have a ground reference for the diagnostics circuitry, a thin copper ring was wrapped around one arm of the Pyrex chamber (externally) and connected to a hard ground. For all experiments, the jet was driven by the following parameters, unless otherwise mentioned: Voltage was 7 kV, pulse width was 1 μs, repetition rate was 7 kHz, and helium flow rate was 7 slm.

#### *3.1. Measurement of the Deposited Charge*

The electric field generated behind a dielectric barrier is a function of the amount of charges deposited on the outer/opposite surface of the dielectric. To measure the charges deposited on the surface of a dielectric surface by the plume of a plasma jet, the following circuit was used (Figure 4) [11]. The charge was calculated by integrating the current flowing through the readout resistor. The current through the resistor flowed via capacitive coupling, with the capacitance having the dielectric constant of the Pyrex glass, and the area was that of the copper plate electrode placed against the dielectric.

Figure 5 is a plot of the deposited charges versus time when the plasma jet is powered by 1 μs wide pulses of magnitude of 7 kV.

**Figure 4.** Circuit used to measure the amount of charge deposited on the surface of a dielectric. Copper plate was 1.5 inches × 1.5 inches thick. Dielectric plate was 2 inches × 2 inches and 0.6 inches thick. Resistance was 1 kΩ.

**Figure 5.** Deposited charge as a function of time for pulse 1 μs wide, magnitude of 7 kV, pulse frequency of 7 kHz, and helium flow rate of 7 slm.

The magnitude of the electric field generated by such charge accumulation was estimated to be in the 10–15 kV/cm range for a dielectric thickness of about 1.5 cm. This is in relative agreement with the spectroscopic measurements presented in the next section.

#### *3.2. Measurement of the Transmitted Electric Field*

The electric field was measured using the Stark splitting and shifting of the helium visible lines and their forbidden lines. The displacement of the Stark sublevels of He I 447.1 nm and its forbidden component was calculated. The wavelength separation (Δλallowed-forbidden) of the π components of allowed and forbidden lines (mupper = 0 to mlower = 0) was measured and its relationship to the electric field strength was used to calculate E using a polynomial relation [12,13] (see Equation (1)). Using this method, electric field strength up to 18 kV/cm was calculated [11]. Figure 6 shows the experimental setup while Figure 7 shows a result of such a measurement.

$$
\Delta\lambda\_{\text{allowed-forbridden}} = -1.6 \times 10^{-5} \times \text{E}^3 + 5.95 \times 10^{-4} \times \text{E}^2 + 2.5 \times 10^{-4} \times \text{E} + 0.15,\tag{1}
$$

where E is expressed in kV/cm.

**Figure 6.** Sketch of the experimental setup to measure the electric field in the chamber. The electric field in the bulk of the plasma is assumed to be weak so the radiation collected to measure the axial electric field in the chamber gives a measure of the field close to the chamber wall (opposite side to where the jet impinges on the wall), where the field is high. The spectrometer was a 0.5 m Spectra-Pro-500i imaging spectrometer (Acton Research). The grating was 3600 g/mm, the slit width was 300 μm, and the spectral resolution was set at 0.02 nm. ICCD was a Dicam-Pro. The pressure inside the chamber was 1.5 Torr and the gas was helium. For each wavelength, more than a million ICCD images were integrated over a collection time of 100 ms.

**Figure 7.** Electric field (EF) measurement using Stark splitting. An EF average strength of about 18 kV/cm was measured.

#### *3.3. Measurement of the Electrons Density and Temperature*

A Langmuir probe was used to measure the electrons density and temperature of the plasma discharge inside the chamber. It is important to note again that only time-averaged values were measured. The measurements presented here are preliminary and only meant to give an approximate idea on the order of magnitude of the density and temperature. The probe was made of a shielded 5 mm long metal needle with a surface area of 2.36 mm2. A variable DC power supply was used to bias the probe. For an air plasma at a pressure of 400 mTorr, the average electron density was found to be around 1.6 <sup>×</sup> <sup>10</sup><sup>10</sup> cm−<sup>3</sup> and the average electron temperature was 2.2 eV. For a helium plasma, with a small admixture of air, at a pressure of 300 mTorr, the electron density was around 6.4 <sup>×</sup> 10<sup>10</sup> cm−<sup>3</sup> and the average temperature was 2.7 eV.

#### *3.4. Dynamics of the Plasma Ignition Inside the Chamber*

ICCD images of the expansion of the diffuse plasma inside the chamber, taken 20 nanoseconds apart, are shown in Figure 8 (using a Dicam Pro ICCD). The images show that, as the ionization wave front (i.e., plasma bullet) impinged on the outer surface of the chamber, a discharge was ignited behind the Pyrex glass wall. A glowing and expanding plasma advanced radially in all directions inside the chamber. The speed of expansion was estimated to be in the 106–107 cm/s range, consistent with that of an ionization wave in gases at low pressure.

**Figure 8.** ICCD images of the plasma bullet (left panel) and of the expanding discharge inside the Pyrex tubing/chamber (right panel). The air pressure in the chamber was around 1 Torr. The contour of the chamber wall (grey circle) is added for illustrative purposes. The vertical line is added as a reference to better visualize the expansion of reduced pressure plasma. The images on the left panel show the plasma bullet arriving and spreading over the outer surface of the Pyrex glass wall.

#### **4. Conclusions**

Reduced pressure diffuse plasma can be generated inside a chamber having dielectric walls using an externally applied ionization wave emitted by a plasma jet. The chamber had no electrodes or any electrical connection. The dynamics of the diffuse plasma was elucidated using ICCD images, the electric field transmitted inside the chamber was measured using a spectroscopic technique, and the electron density and temperature were measured by a Langmuir probe. The advantage of the generation method discussed in this paper is that the plasma inside the chamber is free of metal contamination (since there are no electrodes). Such plasma can be very useful for material processing. Most low-pressure RF plasmas (capacitively or inductively coupled) require impedance matching when operated at high power. However, the new method described in this paper does not need an impedance matching module since the plasma jet is electrically independent of the reduced pressure plasma chamber. It only serves as the source of the ionization wave. Therefore, the plasma inside the chamber may be characterized as a "remotely" generated plasma.

**Conflicts of Interest:** The author declares no conflict of interest.

#### **References**


© 2019 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Editorial* **Special Issue on Plasma Medicine**

#### **Mounir Laroussi 1,\*, Michael Keidar <sup>2</sup> and Masaru Hori <sup>3</sup>**


Received: 2 October 2018; Accepted: 15 October 2018; Published: 17 October 2018

Research on the applications of atmospheric pressure low temperature plasma (LTP) in biology and medicine started in the mid-1990s with experiments on the inactivation of bacteria on biotic and abiotic surfaces and in liquid media [1,2]. This was soon followed by investigations on the effects of LTP on mammalian cells [3–9]. The encouraging results obtained in these early works led to the consideration of LTP technology for new potential therapies in wound healing, dentistry, and cancer treatment [5–9]. By the end of the first decade of the 2000s, many LTP sources had been approved for medical use. These included the Rhytec Portrait® for use in dermatology (approved in 2008), the Bovie J-Plasma® (Clearwater, FL, USA), the Canady Helios Cold Plasma and Hybrid Plasma™ (Takoma Park, MD, USA) Scalpel, the Adtec MicroPLaSter® (Hounslow, UK, approved for clinical trials in 2008), the kINPen® (developed by INP, Greifswald, Germany, and medically certified as class IIa in 2013), and the PlasmaDerm® device (CINOGY GmbH, Duderstadt, Germany). In addition to potential applications in medicine here on Earth, LTP may prove to be a crucial technology for space medicine. In long-duration manned deep space missions, using LTP for decontamination and wound treatment, for example, would be a more suitable/applicable option than transporting and storing perishable chemical-based medication. In this context. LTP offers energy-based medical options that mostly require the availability of electrical power.

The effects of LTP on biological cells are believed to be mainly mediated by its reactive oxygen species (ROS) and reactive nitrogen species (RNS) [10,11]. These include hydroxyl, OH, atomic oxygen, O, singlet delta oxygen, O2( <sup>1</sup>Δ), superoxide, O2 −, hydrogen peroxide, H2O2, and nitric oxide, NO. These species (radicals and non-radicals) can interact with cells membranes, enter the cells, and increase the intracellular ROS concentrations, which may lead to DNA damage and may compromise the integrity of other organelles and macromolecules [12–16]. ROS and RNS can also trigger cell signaling cascades, which can ultimately lead to cellular death pathways, such as apoptosis. Other plasma-generated agents that may play biological roles are charged particles and photons. In addition, LTP can exhibit large electric fields that are suspected to also play a role, such as in cellular electroporation, allowing large molecules to enter the cells.

This Special Issue contains eight papers discussing the latest results on the application of LTP to various cell lines and tissues. These papers discuss a variety of plasma medicine topics, including the treatment of ovarian cancer, triple-negative breast cancer, malignant solid tumors, new LTP devices, as well as a mini review and a paper describing atomic scale simulations on glucose uptake under LTP treatment.

To conclude, the guest editors would like to thank all the authors for their valuable contributions and the reviewers for their time and efforts.

#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Plasma Medicine: A Brief Introduction**

#### **Mounir Laroussi**

Electrical & Computer Engineering Department, Old Dominion University, Norfolk, VA 23529, USA; mlarouss@odu.edu; Tel.: +757-683-6369

Received: 28 January 2018; Accepted: 17 February 2018; Published: 19 February 2018

**Abstract:** This mini review is to introduce the readers of *Plasma* to the field of plasma medicine. This is a multidisciplinary field of research at the intersection of physics, engineering, biology and medicine. Plasma medicine is only about two decades old, but the research community active in this emerging field has grown tremendously in the last few years. Today, research is being conducted on a number of applications including wound healing and cancer treatment. Although a lot of knowledge has been created and our understanding of the fundamental mechanisms that play important roles in the interaction between low temperature plasma and biological cells and tissues has greatly expanded, much remains to be done to get a thorough and detailed picture of all the physical and biochemical processes that enter into play.

**Keywords:** low temperature plasma; plasma jet; cells; tissue; apoptosis; cancer; wound healing; reactive species

#### **1. Introduction**

In the mid-1990s, experiments were conducted that showed that low temperature atmospheric pressure plasmas (LTP) can be used to inactivate bacteria [1]. Based on these results, the Physics and Electronics Directorate of the US Air Force Office of Scientific Research (AFOSR) funded a proof of principle research program in 1997 and supported such research for a number of years. The results from this research program were widely disseminated in the literature, including in peer-reviewed journals and conference proceedings, therefore attracting the attention of the plasma physics community to new and emerging applications of low temperature plasma in biology and medicine [2–8]. The goals of the AFOSR program were to apply low temperature plasmas (LTP) to treat the wounds of injured soldiers and to sterilize/disinfect both biotic and abiotic surfaces. By the early 2000s, research expanded to include eukaryotic cells when small doses of LTP were found to enhance phagocytosis, accelerate the proliferation of fibroblasts, detach mammalian cells without causing necrosis, and under some conditions, lead to apoptosis [9,10].

The above-described groundbreaking research efforts showed that nonthermal plasma can gently interact with biological cells (prokaryotes and eukaryotes) to induce certain desired outcomes. These early achievements raised great interest and paved the way for many laboratories from around the world to investigate the biomedical applications of LTP and by the end of the first decade of the 2000s, a global scientific community was established around such research activities. The field is today known by the term plasma medicine, and in the last few years a number of extensive reviews and tutorials were published (see Refs [11–18] and references therein) as well as a few books [19–21].

Today, the field of plasma medicine encompasses several applications of low temperature plasmas in biology and medicine [22–53]. These include:


#### *Plasma* **2018**, *1*


In the late 2000s, several LTP sources were approved for cosmetic and medical use. Examples are: in 2008 the US FDA approved the Rhytec Portrait® (plasma jet) for use in dermatology. Also in the US other plasma devices are in use today for various medical applications, such as the Bovie J-Plasma® and the Canady Helios Cold Plasma and Hybrid PlasmaTM Scalpel. In Germany, the medical device certification class IIa was given to the kINPen® (plasma jet) in 2013, and the PlasmaDerm® device (CINOGY GmbH) was also approved. Figure 1 is a timeline graph showing the major milestones in the development of the field of low temperature plasma medicine.

**Figure 1.** Timeline showing some major milestones of the new field of the biomedical applications of low temperature atmospheric pressure plasma. This timeline does not show the case of thermal (hot) plasmas, which were used for many decades in medical applications requiring heat, such as cauterization and blood coagulation.

#### **2. LTP Takes on Hygiene and Medical Challenges**

As can be seen from Figure 1 the biomedical applications of LTP started with experiments on the inactivation of bacteria on biotic and abiotic surfaces and media. Bacterial contamination proved to pose severe challenges for some industries and in the healthcare arena. The industrial challenges are mainly around the problem of food contamination and sterilization of food packaging. Several well-publicized food poisoning incidents (EHEC, *Listeria*, *Salmonella*) pointed out to consumers that the present methods employed by the food industry may not be adequate to insure food safety. The healthcare challenges are linked to nosocomial infections caused by antibiotic resistant strains of bacteria, such as Methicillin-resistant *Staphylococcus aureus* (MRSA) and *Clostridium difficile* (C-diff). Every year in the US, hospital acquired infections (HAI) kill thousands of patients with compromised immune systems. HAI are caused by inadequate sterilization/decontamination of instruments, surfaces, clothing, bedsheets, and personnel (nurses and doctors). In most cases, contamination by strains of bacteria resistant to the best antibiotic medications available today is the cause of HAI. LTP is therefore considered as a novel technology that can be successfully applied to help solve some of the challenges described above.

The most recent application presently receiving much attention is the use of LTP to destroy cancer cells and tumors in a selective manner [38–58]. Starting around the mid-2000s several investigators reported experiments showing that low temperature plasmas (LTP) can destroy cancerous cells in vitro. This was followed by some in vivo work showing that LTP can reduce the size of cancer tumors in animal models. The in vitro work covered a host of cancerous cell lines, which included glioblastoma, melanoma, papilloma, carcinoma, colorectal cancer, ovarian cancer, prostate cancer cells, squamous cell carcinoma, leukemia, and lung cancer. The in vivo (animal model) work can be found in [38,44,45,53].

In addition to direct plasma applications to cancer cells and tissues, investigators reported that plasma-activated media (PAM) can also be used to destroy cancer cells [38,50,54–58]. Plasma-activated medium is produced by exposing a biological liquid medium to LTP for a length of time (minutes). In this case, the plasma-generated reactive species interact with the contents of the medium and generate solvated long-lived reactive species in the liquid, such as hydrogen peroxide, H2O2, nitrite, NO2 −, nitrate, NO3 −, peroxynitrite. ONOO−, and organic radicals. These molecules subsequently react with the cells and tissues causing various biological outcomes.

#### **3. Mechanisms of Biological Action of LTP: Brief Summary**

Investigators reported that the effects of LTP on biological cells (prokaryotes and eukaryotes) are mediated by reactive oxygen and nitrogen species (RONS) [11,12,59–66]. These species include hydroxyl, OH, atomic oxygen, O, singlet delta oxygen, O2( <sup>1</sup>Δ), superoxide, O2 −, hydrogen peroxide, H2O2, and nitric oxide, NO. For example, the hydroxyl radical is known to cause the peroxidation of unsaturated fatty acids, which make up the lipids constituting the cell membrane. The biological effects of hydrogen peroxide are mediated by its strong oxidative properties affecting lipids, proteins, and DNA (single and possibly double-strand breaks). Nitric oxide, which acts as an intracellular messenger and regulator in biological functions, is known to affect the regulation of immune deficiencies, cell proliferation, induction of phagocytosis, regulation of collagen synthesis, and angiogenesis.

In cancer cells, the mechanisms of action of LTP are suspected to be related to an increase of intracellular reactive oxygen species (ROS), which can lead to cell cycle arrest at the S-phase, DNA double-strand breaks, and induction of apoptosis. Research by various groups showed that RONS generated by LTP react with cell membranes and can even penetrate the cells and induce subsequent reactions within the cells that can trigger cell-signaling cascades, which can ultimately lead to apoptosis in cancer cells [56–66]. In addition, investigators have shown that plasma-generated RONS can indeed penetrate biological tissues up to depths of more than 1 mm and therefore interact not only with the cells on the surface but with those underneath [67–72].

LTP delivers not only reactive species but it also can exhibit large enough electric fields [73–77]. The magnitudes of these electric fields are several kV/cm and they are suspected to play a role, such as in cellular electroporation, which may allow large molecules to enter the cells.

#### **4. Two LTP Sources for Biomedical Applications: Brief Description**

The main LTP devices used in plasma medicine research are the dielectric barrier discharge (DBD) and nonequilibrium atmospheric pressure plasma jets (N-APPJ). In fact, the DBD was the device used in the first experiments on the inactivation of bacteria [1]. The DBD uses plate electrodes covered by a dielectric (such as glass). The plasma is generated in the gap separating the electrodes by the application of high sinusoidal voltages in the kHz frequency range. Gases such helium with admixtures of oxygen or air are usually used. For more information on the working of the DBD see references [61,78,79]. Figure 2 shows a schematic of the DBD and a photograph of a diffuse plasma at atmospheric pressure generated by a DBD.

**Figure 2.** Schematic (**a**) and a photograph (**b**) of an atmospheric pressure diffuse plasma generated by a dielectric barrier discharge (DBD). The discharge in the photo on the right is driven by kHz sinusoidal high voltage and the gas is helium with a small admixture of air. Photo taken at the author's laboratory.

Nonequilibrium atmospheric pressure plasma jets (N-APPJs) produce plasma plumes that propagate away from the confinement of electrodes and into the ambient air. The reactive species generated by the plasma can therefore safely and conveniently be transported to a target at a remote location and away from the main plasma generation area. This characteristic made N-APPJs very attractive tools for applications in biology and medicine [60,80–82]. Various power driving methods that include pulsed DC, RF, and microwave power have been used. In addition, various electrode configurations ranging from single electrode, to two-ring electrodes wrapped around the outside wall of a cylindrical dielectric body, to two-ring electrodes attached to centrally perforated dielectric disks have been used. Figure 3 shows photographs of two N-APPJs, the plasma pencil and the kINPen, which have been used extensively in plasma medicine research.

**Figure 3.** Photographs of two plasma jets that have been used in various biomedical applications. (**a**) is the plasma pencil (ODU, Norfolk, VA, USA), and (**b**) is the kINPen (INP, Greifswald, Germany).

The plasma plumes emitted by N-APPJs turned out to be made of small plasma packets traveling at very high velocities (tens of km/s). These plasma packets came to be known as "plasma bullets" and they were independently first reported in the mid-2000s by Teschke et al and by Lu and Laroussi [83,84]. Lu and Laroussi used nanosecond-pulsed DC power while Teschke et al used RF power. The plasma bullets were subsequently researched extensively, both experimentally and by modeling, by various investigators [85–91]. Today there is agreement that the plasma bullets are guided ionization waves. To learn about these guided ionization waves in greater detail, the reader is referred to [92].

#### **5. Two Biomedical Applications of LTP**

To illustrate the effects of LTP on biological targets, two applications are shown here. The first concerns the bactericidal property of LTP and the second shows the effects of direct plasma exposure as well as plasma activated media on cancerous and healthy epithelial cells. The results presented below are based on the use of the plasma pencil described earlier. The results shown were obtained by the application of the LTP plume generated by the plasma pencil on a bacterial lawn seeded on the surface of a Petri dish (see Figure 4). The bacterium used was *Acinetobacter calcoaceticus*, a gram-negative soil bacterium also found in the tiger mosquito, which is known to be a transmission vector of yellow and dengue fevers. Figure 5 shows zones of inactivation (dark circular areas) around the center of the dish where the plasma plume was applied. The photo to the left is for an initial bacteria concentration of 109/mL, while that on the right is for an initial concentration of 107/mL. It is clear that the killing effects are more extended and pronounced for the lower initial concentration. For more information on the dependence of inactivation on the plasma exposure time and on the type of bacteria, the reader is referred to [93].

**Figure 4.** Experimental setup for the bacterial inactivation experiments.

**Figure 5.** Killing property of LTP: Dependence of the killing efficacy on the initial bacteria concentration. Left picture is for 109/mL and right picture is for 107/mL. Bacterium is *A. calcoaceticus*. LTP source is the plasma pencil operated with helium as a carrier gas [93].

Figure 6 shows the effects of direct application of LTP on suspensions of cancerous cells. The cancer cell line used was a squamous cell carcinoma of the bladder (SCaBER, ATCC HTB-3TM) originally obtained from a human bladder. After LTP exposure and proper incubation process (37 °C under 5% CO2 atmosphere), Trypan-blue exclusion assay was used to count the number of live and dead

cells. For details of the experimental protocol please refer to [40]. The counts immediately after LTP treatment (at 0 h) revealed no dead cells, which suggested there were no immediate physical effects. However, the viability of cells reduced to around 50% at 24 h after a 2-min LTP treatment. As seen in Figure 6, higher plasma exposure times result in more cells killed (5-min plasma treatment results in 75% of loss of viability at 24 h post-treatment) [40]. These results indicate that LTP does not apply immediate brute physical force on the cells, but its effects require longer biological times to show. This is an indication that plasma agents, such as reactive species and electric fields, interact with the cells and induce reactions and/or trigger biochemical pathways that ultimately result in the death of the cancer cells hours later.

**Figure 6.** Viability of SCaBER cells in media treated directly by the LTP plume of the plasma pencil reveal dead (black bars on top) and live (green bars) cells. The viability was monitored at 0, 12, 24 and 48h post-LTP treatment [40].

Figure 7 shows the selective effect of LTP when it comes to destroying cancer cells versus healthy cells in vitro. The viability results shown in the figure below were obtained using plasma activated media (PAM), which was created by exposing biological liquid media to the plasma pencil for certain lengths of time. The cancerous cell line used was SCaBER and the healthy/normal cells were MDCK (Madin-Darby canine kidney) cells from normal epithelial tissue of a dog kidney. The media used to make PAM were MEM (minimum essential media) for SCaBER and Eagle Minimum Essential Media (EMEM) for MDCK. Figure 7 shows the results [57].

**Figure 7.** Viability in percent of SCaBER (cancerous) and MDCK cells (noncancerous) treated by PAM for various lengths of time. Viability was assessed after 12 hours incubation with PAM using MTS assay and Trypan-blue exclusion assay [57].

#### *Plasma* **2018**, *1*

Figure 7 shows that PAM created using longer exposures to LTP has increasing killing effects on SCaBER cancer cells, reducing their viability to below 10% for irradiation times longer than 2 min. However, normal MDCK cells were able to withstand exposure to PAM for 3 min. This illustrates the selectivity of PAM in killing cancer cells while sparing healthy cells. But for PAM created with longer exposures to LTP (6 minutes and more) extensive killing of MDCK cells was obtained. This illustrates that the plasma dose is an important factor to take into consideration for optimal outcomes.

#### **6. Penetration of RONS in Tissues**

One of the key questions in plasma medicine is the following: Do the RONS generated by LTP only interact and affect cells on the surface of a tissue (or tumor) or do they penetrate the tissue and affect cells in deeper layers? Experimental evidence has shown that LTP does indeed affect cells underneath the tissue surface but what remains unclear is how. One possible explanation is what is referred to as the "bystander effect," which implies that there are chemical signals sent by the cells on the surface (in contact with plasma) to cells in the layer below [41]. These signals would trigger reactions similar to those occurring at the cells on the surface, including the onset of apoptosis. However, and to the best of this author's knowledge, there has been no experimental proof this occurs when LTP interacts with tissues. So, the possibility is there, but reliable data that can be replicated needs to emerge first. Therefore, in this section, only experiments that reported qualitatively and/or quantitatively on the penetration of RONS are presented.

In order to qualitatively and quantitatively elucidate RONS penetration into tissues, investigators used various in vitro models. Oh et al. investigated the penetration of RONS using a model made of an agarose film covering a volume of deionized water contained in a quartz cuvette [67]. They found that RONS kept being delivered from the agarose film to deionized water underneath it for up to 25 min after the plasma was removed. To study the delivery of reactive oxygen species (ROS) into cells, Hong et al. used a model comprising phospholipids vesicles encapsulated within a gelatin matrix and equipped with reactive oxygen species (ROS) reporter [68]. They found that ROS were delivered to the cells without rupturing the membranes of the vesicles. To simulate biological tissue, Szili et al. used gelatin gel, a derivative of collagen, and reported on the penetration behavior of H2O2 through a 1.5 mm thickness gelatin film [69]. The same authors also investigated the effects on DNA in synthetic tissue fluids, tissue, and cells [94].

Tissue models are useful and provide preliminary data regarding the penetration of RONS through biological targets. However, to simulate more realistic conditions, Duan et al. used slices of pig muscle tissue of different thicknesses placed on top of a PBS solution [72]. Figure 8 shows the experimental setup. A plasma jet operated with a helium/oxygen mixture was used. To ignite the plasma sinusoidal high voltages at a frequency of 1 kHz were employed. The plasma treatment times were 0, 5, 10, and 15 min.

**Figure 8.** Experimental setup using pig muscle tissue [72]. Reproduced from Duan, J.; Lu, X.; and He, G. *Phys. Plasmas* **2017**, *24*, 073506, with the permission of AIP Publishing.

The concentrations of H2O2, OH, and that of the total of (NO2 − + NO3 −) were measured for different thicknesses of the tissue slice. A comparison of these concentrations when no tissue was used and when a tissue was placed on top of the solution showed that the concentrations of O3, OH, and H2O2 were mostly consumed by the tissue and could not pass through 500-μm or greater tissue thickness. However, more than 80% of the (NO2 − + NO3 −) penetrated a 500-μm-thick tissue slice. Figure 9 shows the measured concentrations of (NO2 − + NO3 −) as a function of tissue thickness and for three plasma treatment times (5, 10, and 15 min).

**Figure 9.** Total nitrite and nitrate concentration versus tissue thickness for three plasma exposure times [72]. Reproduced from Duan, J.; Lu, X.; and He, G. *Phys. Plasmas* **2017**, *24*, 073506, with the permission of AIP Publishing.

Figure 9 shows that the concentrations of the nitrogen reactive species, RNS, decrease with the tissue thickness, but increase with the plasma treatment time. The concentration of (NO2 − + NO3 −) for the 500-μm tissue thickness was comparable to the concentration when no tissue was placed on top of the PBS solution. This means that (RNS) were able to penetrate the tissue slice. This was not the case for ROS, which were absorbed by the tissue, unlike the case when a gelatin model (not real tissue) was used. For that model, ROS were able to penetrate the gelatin film.

The above examples illustrate that RONS do not simply react with the surface of tissues but can indeed penetrate relatively deeply. However, in more realistic conditions using actual tissue, it was shown that not all RONS can cross the same thickness. Some can be absorbed within a few tens of micrometers by the tissue, while others can penetrate up to 1.5 mm below the surface. Of course, the above results may not completely reflect what would happen under in vivo conditions. Such experiments need to be conducted and compared to results obtained for in vitro models and to those obtained under ex vivo conditions [95].

#### **7. Conclusions**

To conclude this brief introduction of the field of plasma medicine, it is safe to say that the biomedical applications of low temperature plasma have opened up an entirely new multidisciplinary field of research requiring close collaboration between physicists, engineers, biologists, biochemists, and medical experts. This multidisciplinary field started in mid-1990s with seminal experiments on the inactivation of bacteria by low temperature atmospheric pressure plasma generated by a dielectric barrier discharge and slowly expanded to include investigations on eukaryotic cells. Applications in dermatology, wound healing, dentistry, and cancer have led to various scientific advances and to the idea that LTP can be a technology upon which various innovative medical therapies can be developed to overcome present healthcare challenges. However, a lot remains to be done in order to fully understand the mechanisms of action of LTP against biological cells and tissues, both in vitro and in vivo. There is strong indication that LTP acts selectively on cancer cells and tumors and can penetrate deep below the surface, but much more work, including extensive clinical trials, is needed

before LTP can be considered a safe technology ready for use in hospitals to treat chronic wounds, cancer lesions and tumors, and other ailments.

**Conflicts of Interest:** The author declares no conflicts of interest.

#### **References**


© 2018 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
