Volt-ampere (V-A) characteristics of the pin-APPJ system were investigated and power deposition was calculated for different input conditions. The characterization was done in the identical conditions as in treatments, with the AB25 aqueous sample as a grounded target and Ar flow of 1 slm.

The results of electrical characterization are given in Figures 4, 5. Figure 4A represents the V-A at the jet, while Figure 4B shows the V-A of the plasma in contact with the sample. In both plots, RMS (root mean square) values of the voltage and appropriate current are presented. The linear part of the dependence represents the conditions when plasma is not ignited. At around 1.6 kV_RMS voltage the plasma was ignited resulting in a considerable reduction in voltage. One needs to keep in mind that here we did not measured the breakdown voltage but the operating voltage. With a further increase of power given by the RF power supply after plasma ignition, operating voltage did not change significantly and remained at about 1.1 kV_RMS. However, after ignition, the values of I_RMS at ground, i.e., in the plasma, raise more than 50%. In comparison to the I_RMS in the source, I_RMS in the part of the circuit that includes the sample were observed to be more than two fold lower.

The most important parameter for plasma treatments is power deposited into the plasma system. The variation in the power deposition at the source and sample are illustrated in Figures 5A,B, respectively. For the lowest powers applied, when the plasma was not ignited, it was observed that there was a small power deposition of 1–2 W at the source even though the system is mainly capacitive. This is due to the small resistive impedance of the electrical connections and components. When the applied voltage increased above 1.6 kV_RMS the breakdown occurs and with the inception of plasma the power deposition is increased. We can see that the increase of the power deposited in the system occurred while the voltage stayed almost constant or slightly decreasing indicating that the power increment is significantly due to the higher current in the source and through the ground i.e., in contact with the sample. Of particular importance was to calculate the power at the ground, i.e., deposited in the part of the electrical circuit in contact with the sample. It can be observed in Figure 5B that when the plasma was ignited the power delivered to the sample was in the range of about 5–11 W.

OES was used for the detection of excited species in argon plasma in the air interacting with the liquid. A typical OES spectrum during AB25 treatment in the range of 200–1,100 nm can be seen in Figure 6. Excited species such as hydroxyl radical (HO), atomic hydrogen (H), atomic oxygen (O), atomic argon, nitrogen second positive system (SPS), nitrogen first positive system (FPS) and nitrogen first negative system (FNS) were observed.

A-X band of HO and H_α line were detected with maximum emission of at 309 and 656.2 nm, respectively. In the discharge HO emission could have originated due to dissociative excitation of water molecules in collisions of electron or argon metastables [57–59]. The main source of water molecules in the gaseous phase or discharge region mostly comes from the evaporation of the target sample due to the effect of purging caused by gas flow or could be via local heating caused by the impact of heavy ions in APPJ [57, 60]. Since the plasma was ignited in the mixture of argon with the surrounding air, excited argon lines, molecular nitrogen bands and atomic oxygen lines were present. The emission of O lines were at 777.5 and 844.6 nm, excited molecular nitrogen SPS between 300 and 405.8 nm with intensive band-head at 337.7 nm. For the molecular ions, the nitrogen molecular FNS between 391.4 and 428 nm and a weak intensity of FPS are also present. Excited argon lines were observed in the spectral region between 696.5 and 965.7 nm and the strongest emission occurred at 773.5 and 811.5 nm, respectively. The formation of O and molecular nitrogen can be attributed to the reactions of energetic electrons or argon metastables with ambient air [57]. Many authors have reported that HO and O radicals are considered as one of the main short-lived ROS in the plasma discharge, and they can directly react with organic pollutants which lead to degradation [14, 19, 24, 61].

Emission intensities are determined with various input powers. Figure 7 illustrates that the emission intensities of HO, H, nitrogen SPS, excited argon as a function of deposited power to the sample (ground). It can be observed that the spectral emission intensity of the species increases almost linearly with the increase in deposited power. The largest change in the intensity occurred with species such as O (844.6 nm), nitrogen SPS and excited argon, whereas the intensity of HO, H, and O (777.5 nm) only slightly increases with input powers. The observed influence of input power correlates with results presented by other researchers [57]. Changes in emission intensities with power are related to higher energy delivered to the plasma. Higher power probably could lead to an increase in the electron temperature as well as electron density which influences the emission intensity of excited species [62]. From the viewpoint of treatments, higher intensities and concentrations of reactive species are the essential factors in order to degrade organic pollutants [11, 44]. The OES is a simple and, more important, non-intrusive diagnostic technique that can give information about the excited atomic, molecular and ion species. Its simplicity makes it readily available and the obtained results can be easily interpreted. However, they do not give the whole picture about the plasma chemistry (especially in the gas phase). This can be done by using the Fourier transform infrared (FTIR) spectroscopy, which was not readily available in our case. In the literature, FTIR analysis was carried out to analyze long-lived reactive species in the gas phase. For example, in the study of [11], FTIR analysis was performed to characterize the plasma-liquid interaction, where a plasma jet was formed over the liquid surface in ambient air. It was found that FTIR confirmed the presence of various reactive species (e.g., O₃, NO₂, HNO_3, HNO₂, etc.) during plasma–liquid interaction in ambient air. [63] have demonstrated the effect of a water surface on the production of reactive species and FTIR spectroscopy was applied to identify O₃ and NO₂ in the far-field of the cold atmospheric plasma jet. [64] investigated the synthesis of nanoparticles in liquid with cold atmospheric plasma and they have used FTIR for surface characterization of the TiO₂. Another diagnostic technique that can give information about plasma chemistry (both radicals, atoms, molecules, and ions) is mass spectrometry at atmospheric pressure. We have performed characterization of various plasma sources that operate at atmospheric pressure by using HIDEN HPR60 and obtained mass spectra of neutral radicals (atoms/molecules) and ions (results are presented here) [65–68].

In the laboratory-scale experiments, two initial concentrations of AB25 were used. The degradation of dye in the solutions as a function of treatment time is given in Figure 8A. With the increase of treatment time, the concentration of AB25 was decreased exponentially. In both cases, the degradation was significantly more pronounced in the first 5 min. The removal efficiency values were about 87% for 25 mg/L and 73% for 50 mg/L within 5 min. In case of lower initial concentration, the destruction of AB25 for the longest treatment time (10 min) was about 100%, whereas for the two times greater initial concentration the obtained removal efficiency was around 93%. It can be expected that for a longer time the higher initial concentration of AB25 would be completely removed.

Visually, the change in the coloring of the sample can be observed in Figure 9. For lower AB25 concentration and within 10 min of treatment time, dye almost disappeared. On the other hand, for the higher concentration, the solution was not completely colorless.

The kinetic study was performed in order to understand the degradation rate of AB25. It was found that the degradation of AB25 for both initial concentrations followed the first-order kinetics during investigated treatment duration (Figure 8B) with good regression coefficient values (R ² ≥ 0.98).

First-order kinetics is defined by the following Eq. 7, where k is rate constant and t is treatment time. C C o =   e − k t (7)

The half-life (t_1/2) AB25 decomposition time was determined by the given Eq. 8 and shown in Table 1. t 1 / 2 =   0.693   k (8)

It can be inferred from Table 1 that the lower initial concentration of AB25 promotes the higher degradation rate constant. The half-life time for 25 mg/L was almost 2 times lower than 50 mg/L, which attributed to a faster degradation rate for the lower AB25 initial concentration. The higher concentration of AB25 means a higher number of dye molecules in the solution, whereas the formation of plasma species remains almost the same. Moreover, at the higher concentration, intermediate products may become more relevant and not all introduced plasma species continue reacting with the original AB25 molecules (competition with intermediates). Yet, concerning that equation for first-order kinetics (7) precisely fits the data, the effect involving intermediate products is not pronounced. Almost a similar behavior was reported in the previous research on the degradation of AB25, where DC glow discharge was formed in ambient air over the liquid surface [9].

Ghodbane et al. [9] have studied the degradation of AB25 with direct exposure to glow discharge, whereas plasma-treated samples were characterized by using HPLC and HPLC-MS. They have reported that the key reactive species such as high energy electrons, O₃, and HO can attack at various positions of AB25 molecules and lead to the formation of low molecular weight residues. For example, initially, the reactive species can attack various chemical bonds in the AB25 molecule, including amino, sulfonic and vinylsulfonyl since they are located far from anthraquinone rings (stable), and then various bonds in naphthaquinone rings can be eliminated. It was found that low molecular weight products such as pyrocatechol, 3,6-diamino-4-methylphthalic acid, benzene, and 2,5-dimethylphenol appeared due to oxidation of AB25. [69] have investigated the degradation of AB25 with AOPs (e.g., UV-H₂O₂ and UV-TiO₂), and they reported the possible degradation pathways for AB25. It was mentioned that degradation can be initiated by attacking at C-N bond to separate the aromatic ring and then hydroxylation in the secondary amine position. In further steps, splitting of quinone ring leading to the formation of 2-amino-4-(2-carboxybenzoyl)-5-hydroxybenzenesulfonate, after that it further oxidizes to phthalic acid and then turns into formic and oxalic acids. Finally, ending up with complete mineralization of the dye. In the study of [70] corona discharge was used for degradation of AB25. Authors reported degradation pathways of AB25, whereas O₃ and H₂O₂ were the major reactive species that are responsible for degradation. The products such as CO₂, H₂O, SO₄ ²⁻, NO₃ ⁻ were found by the end of treatment.

The energy yield was calculated and it depended on the removal percentage of AB25 for both initial concentrations. The energy yield was higher at the beginning of the treatment (shorter treatment times) due to a faster degradation rate. However, for fixed removal percentage the difference in the energy yield was not significant for both initial AB25 concentrations. In this case, the energy yield was about 259 mg/kWh for 50 mg/L and 252 mg/kWh for 25 mg/L at 50% removal, respectively.

The pH and the conductivity of the treated solutions were also measured. The results are presented in Figure 10. For both initial concentrations, the pH changed in the same way and it drastically dropped with the treatment time. After 10 min of treatment time, the final pH was about 2.5. The decrease in the pH can be attributed to the formation of acidic molecules (e.g., HNO₃ and HNO₂) and the higher concentration of hydrogen ions (H⁺) inside the solution [11, 71–73]. Based on measured pH, the concentration of H⁺ ions was determined by Eq. 9 [74]. The concentration of H⁺ ions with treatment time is plotted in Figure 10A, where it increased linearly with treatment time. [ H + ] = 10 − p H (9)

It was observed that the conductivity of the treated solutions increased with the increase in treatment time. The initial conductivity, which was quite small (21–31 μS/cm), was enhanced up to about 850–900 μS/cm after 10 min of treatment. The conductivity was somewhat higher for the higher initial concentration of AB25, but the difference was not significant. The electrical conductivity is determined by the Kohlrausch law taking into account the concentration and nature of ions inside solution, whereby H⁺ ions can contribute strongly because of the Grotthus proton-hopping mechanism leading to an exceptionally high limiting molar ionic conductivity (349.8 S cm²/mol). Figure 10B shows that the H⁺ ions concentration measured against conductivity has a linear dependence.

The influence of deposited power on AB25 degradation was also investigated. The experiment was carried at the initial AB25 concentration of 50 mg/L and treatment time of 5 min. In Figure 11, the effect of the deposited power into the discharge on the change in AB25 concentration and removal percentage is shown. As expected, the higher degradation of AB25 was accomplished when there was more power transferred to the sample. At higher input power, more energy can be delivered to the plasma to support the formation of more reactive species. It was evident from OES result (Figure 7), where emission intensities of reactive species increased with input powers. Therefore, the presence of more reactive species in the gas phase can lead to the degradation of AB25. The observed effect was illustrated in various papers, showing that input power played a beneficial role in the degradation of various organic dyes [75, 76].

The kinetic curves for the removal of AB25 as a function of deposited power are shown in Figure 11. In the first-order reaction (Eq. 7), the time was replaced by input power to determine what is more cost-efficient: to increase the power or to use a longer treatment time. The value of the rate constant revealed that increasing the input power shows slower degradation compared to increasing the treatment time. A modified first-order and half-life equation is described by a given equation. C C o =   e − k ′ P (10)

It takes about 7 W of input power to decompose 50% of AB25 (Table 2), calculated by Eq. 11. P 1 / 2 ′ =   0.693 k ′ (11)

Since the treatment was carried out at different input powers, the energy yield was also determined based on the amount of power consumed to remove 50% of AB25 at a fixed treatment time of 5 min and AB25 concentration of 50 mg/L. The calculated energy yield was about 203 mg/kWh. Comparing to Table 1 for 50% proves that removal varying the power at fixed treatment time was not cost-effective. It is more effective to use lower power and longer treatment times.