Removal of Pb(II), Cd(II) and Ni(II) Ions from Groundwater by Nonthermal Plasma

The removal of Pb(II), Cd(II) and Ni(II) ions from aqueous solutions by means of nonthermal plasma with a dielectric barrier discharge is investigated. Aqueous solutions with metal ion concentrations from 10 to 100 mg/dm3 in spring water were used. In the first stage, the optimization of the solution flow rate, generator modulation frequency and duty cycle was made in terms of the removal efficiency of the considered metals. The removal was then investigated as a function of the number of passes of the solution through the cold plasma reactor. The effect of the initial concentration of ions in the solution was studied. Techniques such as composite central design, least squares method and Fourier transform infrared spectroscopy were used. The physical and chemical parameters of the solutions, such as electrical conductivity, pH, temperature, concentration of metal ions and the content of other substances (e.g., total organic carbon), were measured, and the presence of microorganisms was also examined. It was found that each pass of the solution through the cold plasma reactor causes a decrease in the concentration of Cd(II) and Ni(II); the concentration of Pb(II) drops rapidly after one pass, but further passes do not improve its removal. The removal percentage was 88% for Cd(II) after six passes and 72% for Pb(II) after one pass, whereas 19% for Ni(II). The purification mechanism corresponds to the precipitation of metal ions due to the increasing pH of the solution after exposure to cold plasma.


Introduction
Contamination of surface water and groundwater with trace elements is one of a worldwide challenges due to their high toxicity, persistence and high level of bioaccumulation in the environment [1,2]. The quality of surface waters has been adversely affected by continuously progressing industrialization, urbanization and the mismanagement of water resources [3]. Among the substances polluting natural waters, heavy metal ions belong to most frequently released, together with sewage, waste and dust. The main sources of pollution of the aquatic environment with these metals are emissions from industrial plants, metal smelters, atmospheric precipitation, leachate from landfills and industrial and municipal sewage from urbanized areas. Source of lead can also include gasoline, household paints, batteries and zinc-lead ores [4,5]. Once introduced into the environment, they remain there and interact for a long period of time due to the negligible degree of degradation [6]. They are non-degradable pollutants that are not subject to significant chemical and microbiological transformations.
The harmfulness of heavy metals polluting the environment is largely due to their biochemical and biological properties. Heavy metal ions, such as lead, nickel and cadmium, in them). Therefore, plasma technologies should be further explored to determine their properties and optimize treatment processes [22].
The aim of the study is to investigate whether Pb(II), Cd(II) and Ni(II) in concentrations comparable with those in industrial wastewater can be removed from aqueous solutions via nonthermal plasma with a dielectric barrier discharge. An attempt was also made to determine the effect of NTP on the quality of the tested water as well as the effect of initial concentration of heavy metals and the pH of the solution on removing the pollutants from the water. The optimization of cold plasma generator settings was also carried out in terms of the effectiveness of metal ion removal.

Lab Stand
The research was carried out using an experimental device with a NTP reactor with DBDs. The schematic diagram is shown in Figure 1, and the lab stand is presented in Figure 2. On a quartz tube, a coil is wound; one of its terminals is connected to AC voltage of 7.5 kV and a frequency of 75 kHz. This voltage is excited with a selected modulation frequency of 100-400 Hz. Inside the quartz tube, a conductive tube is placed. The water to be treated as well as air, which is mixed with the water, is supplied through the lower part of the quartz tube. The aerated water flows between the two pipes and is then directed into a destination tank. From an electrical point of view, the arrangement of the coil, quartz tube, water and metal tube is an LC circuit. When operating in a resonant state, it can generate a high electric field strength in the working area between the coil turns and the metal tube. This field causes electric discharges inside air bubbles and the formation of UV radiation, whose photons have enough energy to ionize or excite water molecules, oxygen and other compounds present in the polluted water. The device has the option of setting the flow rate through the reactor, modulation frequency and duty ratio. ganic compounds in them). Therefore, plasma technologies should be further determine their properties and optimize treatment processes [22].
The aim of the study is to investigate whether Pb(II), Cd(II) and Ni(II) in tions comparable with those in industrial wastewater can be removed from a lutions via nonthermal plasma with a dielectric barrier discharge. An attemp made to determine the effect of NTP on the quality of the tested water as well a of initial concentration of heavy metals and the pH of the solution on removin tants from the water. The optimization of cold plasma generator settings was a out in terms of the effectiveness of metal ion removal.

Lab Stand
The research was carried out using an experimental device with a NTP r DBDs. The schematic diagram is shown in Figure 1, and the lab stand is presen ure 2. On a quartz tube, a coil is wound; one of its terminals is connected to AC 7.5 kV and a frequency of 75 kHz. This voltage is excited with a selected mod quency of 100-400 Hz. Inside the quartz tube, a conductive tube is placed. Th be treated as well as air, which is mixed with the water, is supplied through the of the quartz tube. The aerated water flows between the two pipes and is th into a destination tank. From an electrical point of view, the arrangement quartz tube, water and metal tube is an LC circuit. When operating in a reson can generate a high electric field strength in the working area between the coi the metal tube. This field causes electric discharges inside air bubbles and the of UV radiation, whose photons have enough energy to ionize or excite water oxygen and other compounds present in the polluted water. The device has th setting the flow rate through the reactor, modulation frequency and duty ratio

Physical, Chemical and Biological Analyses
As reported in [46], electrical discharges in demineralized water are weak; therefore, water taken from a spring located in the southern part of Poland was used for the research. Spring water contains a certain number of cations, e.g., sodium, calcium or potassium, and anions, e.g., bicarbonate and sulphate. The presence of ions in the tested water facilitates the formation of • OH radicals and OH − ions during the dielectric barrier discharge in the cold plasma [23,26].

Physical, Chemical and Biological Analyses
As reported in [46], electrical discharges in demineralized water are weak; theref water taken from a spring located in the southern part of Poland was used for the resea Spring water contains a certain number of cations, e.g., sodium, calcium or potassi and anions, e.g., bicarbonate and sulphate. The presence of ions in the tested water fa tates the formation of • OH radicals and OH − ions during the dielectric barrier discha in the cold plasma [23,26].
The tests were carried out on the day of water intake. In the tested water, pH, tur ity, electrolytic conductivity, total hardness, oxidizability, total organic carbon (TOC), monium ions, nitrates, nitrites, chlorides, sodium, calcium, sodium and trace eleme such as total Fe, Mn, Mg, Cd, Pb, Cu, Ni and total Cr, were found. The analyses of individual physical and chemical properties of the tested water were determined us the reference methods indicated in [47]. Each determination was performed five times each measurement series was carried out in the same way. The pH measurements w carried out with an ELMETRON pH/mv CP-401 pH meter with an accuracy of ±0. Turbidity (NTU) was measured with a Hach 2100N turbidimeter. Total organic car was determined on an N/C 3100 multi analyzer (Analytik Jena, Jena, Germany) equip with a TC + TIC module. The specific electrolytic conductivity was determined usin G1409-L01 conductometer by Greisinger according to the standard [48]; total water ha ness according to [49]; permanganate index according to [50]; ammonium ions accord to [51]; nitrates according to [52]; nitrites according to [53]; calcium, magnesium and c rides according to [54]; and sodium according to [55]. Nitrogen and sodium compou were determined with a UV-VIS DR 5000 spectrophotometer by Hach-Lange. The con of heavy metals (total Fe, Mn, Mg, Cd, Pb, Cu, Ni and total Cr) was determined using inductively coupled plasma atomic emission spectrometer (ICP-AES) with Spectro A system (SPECTRO Analytical Instruments, Kleve, Germany).
The microbiological tests included the determination of the total number of micr ganisms, coliform bacteria, Escherichia coli and Enterococcus. The total number of micr ganisms (also referred to as the number of heterotrophic bacteria, total number of bact or number of colonies) was determined according to the standard of [56] by the deep p The tests were carried out on the day of water intake. In the tested water, pH, turbidity, electrolytic conductivity, total hardness, oxidizability, total organic carbon (TOC), ammonium ions, nitrates, nitrites, chlorides, sodium, calcium, sodium and trace elements, such as total Fe, Mn, Mg, Cd, Pb, Cu, Ni and total Cr, were found. The analyses of the individual physical and chemical properties of the tested water were determined using the reference methods indicated in [47]. Each determination was performed five times and each measurement series was carried out in the same way. The pH measurements were carried out with an ELMETRON pH/mv CP-401 pH meter with an accuracy of ±0.002. Turbidity (NTU) was measured with a Hach 2100N turbidimeter. Total organic carbon was determined on an N/C 3100 multi analyzer (Analytik Jena, Jena, Germany) equipped with a TC + TIC module. The specific electrolytic conductivity was determined using a G1409-L01 conductometer by Greisinger according to the standard [48]; total water hardness according to [49]; permanganate index according to [50]; ammonium ions according to [51]; nitrates according to [52]; nitrites according to [53]; calcium, magnesium and chlorides according to [54]; and sodium according to [55]. Nitrogen and sodium compounds were determined with a UV-VIS DR 5000 spectrophotometer by Hach-Lange. The content of heavy metals (total Fe, Mn, Mg, Cd, Pb, Cu, Ni and total Cr) was determined using an inductively coupled plasma atomic emission spectrometer (ICP-AES) with Spectro Arcos system (SPECTRO Analytical Instruments, Kleve, Germany).
The microbiological tests included the determination of the total number of microorganisms, coliform bacteria, Escherichia coli and Enterococcus. The total number of microorganisms (also referred to as the number of heterotrophic bacteria, total number of bacteria or number of colonies) was determined according to the standard of [56] by the deep plating method on agar with a yeast extract at a temperature of 22 ± 2 • C during 72 h of incubation. Enterococcus bacteria were determined according to [57], while bacteria of the coli group and Escherichia coli according to [58]. The basis of the method is cultivation in a liquid medium, and the value of the most probable number (MPN) of the organisms sought is read from the relevant MPN tables.

Removing Metal Ions
The experiments were carried out on synthetic solutions prepared on the basis of chemical compounds of metals (analytical grade) containing the single metal ions cadmium, nickel and lead in initial concentrations comparable to those in industrial wastewater. The tested water was contaminated with heavy metals in concentrations from 10 to 100 mg/dm 3 . The metals of high mobility in the environment, i.e., Cd(II), Ni(II) and Pb(II), were selected for the study, because they often occur locally in surface and underground waters [59]. A total of 5 L of water contaminated with the specified metal were passed through the plasma reactor each time. The research included: the optimization of the device settings, examination of the effect of the number of cleaning cycles and examination of the effect of the initial concentration of metal ions on the removal efficiency.
In the first stage, the focus was set on selecting the generator settings. Their influence can be roughly estimated from the following mathematical model. Let the flow rate be p [dm 3 /min], the modulation frequency f m [Hz] and the duty cycle d. Let the volume of the working space (between the quartz tube and the metal tube) be V [dm 3 ]. The time needed for the displacement of this volume of water through the working space is V/p [min]. However, the duration of the discharge is actually shorter due to the duty cycle d, and equals T eff = dV/p. Therefore, the effective time during which the water is affected to discharges per unit volume of water is t eff = T eff /V = d/p [min/dm 3 ]. The greater the effective time of interaction per unit volume, the greater the expected effect. Therefore, the duty cycle should be as high as possible and the flow rate as low as possible. The above analysis shows that the modulation frequency is not an important factor; however, this analysis does not take into account other factors, such as air flow and the frequency of discharges. Therefore, the measurement results may differ from the predictions. As a result, it was decided to determine the optimal flow rate, modulation frequency and duty cycle. The influence of these settings on the removal of Pb(II), Cd(II) and Ni(II) ions with an initial concentration of 10 mg/dm 3 was investigated. The central composite design method in the inscribed version (CCI) was used for three variables with the method parameter α = 23/4 ≈ 1.68. The flow rate was tested in the range of 0.5-2.5 dm 3 /min, modulation frequency was in the range of 100-400 Hz and duty cycle was 10-90%. These variables were normalized according to the CCI rules and denoted x 1 , x 2 and x 3 , respectively, as in Table 1. For each of the three tested metals Pb(II), Cd(II) and Ni(II), aqueous solutions with a concentration of 10 mg/dm 3 were prepared. Spring water with an initial pH of 7.57 and a temperature of 22 • C was used. After mixing and allowing it to equilibrate, the pH and the metal concentration were measured. The solution was then subjected to the NTP reactor (one pass) for various settings of flow rate, modulation frequency and duty cycle according to the 20 points CCI scheme for three variables. The pH and the concentration of metal ions were measured. Then, the pH change and the percentage of ion removal were calculated as follows: where C initial = 10 mg/dm 3 and C final is the final concentration of metal ions. The determined values of ∆pH and Removal were then used to find the best fit to a function of the following form: The unknown coefficients a 0 , a 1 , . . . , a 9 were found with the least squares method, taking the adjusted R 2 as a measure of the fitting, which penalizes the number of degrees of freedom. The adjustment was performed for all possible combinations of base functions, and the best match was assumed that one that had the highest value of adjusted R 2 . This allowed to eliminate the terms that had only a very insignificant influence on the values of the explained variable. Further tests were carried out for a solution flow of 0.5 dm 3 /min, a modulation frequency of 300 Hz and 90% duty cycle-these parameters were selected as a result of the optimization of the settings from the above-described test stage.
In the next stage, solutions of Pb(II), Cd(II) and Ni(II) ions with an initial concentration of 10 mg/dm 3 were prepared, and studies on the removal of Pb(II), Cd(II) and Ni(II) ions were carried out depending on the number of passes of contaminated water through the plasma reactor. The change in pH and the percentage removal of Pb(II), Cd(II) and Ni(II) ions were investigated. In addition, the temperature of the solution was measured after each pass as well as the energy intake was traced. In addition, the FTIR analysis of the samples obtained after 6 passes was performed.
Apart from the above, the effect of the concentration of Pb(II), Cd(II) and Ni(II) ions was investigated. The concentrations varied in the range of 10-100 mg/dm 3 , and the solutions were passed once through the reactor. The change in pH and the percentage removal were determined, as well as temperature and energy intake being measured.
The energetical efficacy of metal ions removal was calculated as follows: where V is the solution volume and W is the energy intake.

Characteristics of the Used Water
The overall characteristics of the water used in the tests along with the values of the permissible values for pollutants discharged into water and soil as well as industrial wastewater are presented in Tables 2 and 3. The conducted research shows that the parameters of the tested water do not exceed the permissible values for surface waters. * Reference value provided in [47]. ** Need not be determined for water production lower than 10,000 m 3 per day. *** Permissible concentration of magnesium at sulphate concentrations lower than 250 mg/dm 3 . Table 3. Microbiological characteristics of the tested water. 100 CFU/1 mL in the water entering the water supply network 200 CFU/1 mL at the consumer's tap * MPN-most probable number. ** CFU-colony-forming unit. *** Reference value specified in [60].

Indicator
From a bacteriological point of view, no pathogenic microorganism from the coliform bacteria, including Escherichia coli and Enterococcus, were found in the tested water. On the other hand, the total number of microorganisms was over 300 CFU and exceeded twice the limit values for tap water [60]. Most of the bacteria that are hazardous to human health belong to the mesophilic group, for which the human body is an excellent incubator due to the optimal temperature for development. Therefore, an increase in the total number of bacteria is always a warning signal of a deterioration in water quality and a health hazard [61].

Optimization of the NTP Generator Settings
The results of the research regarding the selection of the optimal settings of the cold plasma generator, solution flow rate through the reactor, modulation frequency and duty cycle are presented in Tables 4 and 5. Table 4. Measurement results for pH and metal ion concentration at points in the CCI measurement scheme. No.
x 1    Table 5 shows that the variance of the examined indicators (∆pH and Removal) is not very well explained by the assumed approximation (adjusted R 2 in the range of 0.40-0.76). This is probably due to the high content of the random component and the measurement uncertainty. However, it can be seen that after one pass of the water through the reactor, the pH increased, with the average increase being the greatest for Ni(II) and the smallest for Cd(II). Nevertheless, it is difficult to say which of the explanatory variables has the dominant contribution. It is noticeable that an increase in x 1 (flow rate) contributes negatively to the increase in pH, which is in line with the above-mentioned prediction, as the higher the flow rate, the shorter the time of plasma impact on the water. The effect of the remaining variables (modulation frequency and duty cycle) is not unequivocal, although it would be expected that a high duty cycle should result in a stronger impact on water and a greater increase in pH. Moreover, the influence of the tested generator settings on the removal of metal ions is not obvious, although a decrease in concentration is observed in each case, i.e., a positive effect on removal. Regarding the removal of Pb(II), it should be noted that the most of the lead had precipitated out of the solution before exposure to cold plasma, and the subsequent changes in Pb(II) concentration were insignificant. The greatest changes were observed in the case of Cd(II) and the smallest in the case of Ni(II), but in no case was there a clear effect of the settings on the removal of metal ions. Table 6 shows the Pearson correlation coefficients between flow rate, modulation frequency, duty cycle, ∆pH and Removal. Their absolute values are mostly lower than 0.5 and do not indicate a strong correlation between the generator settings and ∆pH and Removal. There was no significant correlation coefficient between the flow rate and other variables. The modulation frequency has a positive effect on the increase in pH in the case of Pb(II) solution (correlation coefficient 0.56), and the duty cycle on Ni(II) removal (0.59). Figure 3 shows the measurement results in (∆pH, Removal) coordinates. The highest correlation coefficient between ∆pH and Removal was observed for Cd(II) (0.77), quite a clear correlation occurred for Pb(II) (0.51), while no correlation was found between these variables for Ni(II). This may be due to the fact that Ni(II) precipitates very poorly in the considered pH range. of Pb(II) solution (correlation coefficient 0.56), and the duty cycle on Ni(II) removal (0.59). Figure 3 shows the measurement results in (ΔpH, Removal) coordinates. The highest correlation coefficient between ΔpH and Removal was observed for Cd(II) (0.77), quite a clear correlation occurred for Pb(II) (0.51), while no correlation was found between these variables for Ni(II). This may be due to the fact that Ni(II) precipitates very poorly in the considered pH range.  Based on the above results, it can be concluded that a minimum flow rate is desirable. The research did not show a clear effect of the modulation frequency and duty cycle. This may indicate that long living reactive species are crucial in the treatment mechanism. The theoretical considerations indicate that the modulation frequency itself is not of great importance, but the duty cycle is important, as it directly affects the reactor's impact time on the water. Therefore, in further tests, the flow rate was set to 0.5 dm 3 /min (lowest possible for the device), the modulation frequency was arbitrarily set to 300 Hz and the duty cycle to 90% (maximum possible for the device).

Physical, Chemical and Microbiological Parameters of the Tested Water
Physical, chemical and microbiological parameters of water samples before and after treatment with cold plasma with DBD discharges are presented in Table 7. It can be seen that the cold plasma increased the electrolytic conductivity by about 2-2.5%. Similar results were obtained in [62]. The increase in conductivity indicates the presence of new ionic forms in the solution resulting from the electric discharge and ionic forms from the destruction of contaminating compounds contained in the water (e.g., H3O + , NO2 − and NO3 − ). The electrolytic conductivity of liquids is one of the important parameters affecting the intensity of UV radiation (it increases in solutions with higher conductivity) as well as the generation of a shock wave or thermal effects [29]. As reported by [63], an exceedingly Based on the above results, it can be concluded that a minimum flow rate is desirable. The research did not show a clear effect of the modulation frequency and duty cycle. This may indicate that long living reactive species are crucial in the treatment mechanism. The theoretical considerations indicate that the modulation frequency itself is not of great importance, but the duty cycle is important, as it directly affects the reactor's impact time on the water. Therefore, in further tests, the flow rate was set to 0.5 dm 3 /min (lowest possible for the device), the modulation frequency was arbitrarily set to 300 Hz and the duty cycle to 90% (maximum possible for the device).

Physical, Chemical and Microbiological Parameters of the Tested Water
Physical, chemical and microbiological parameters of water samples before and after treatment with cold plasma with DBD discharges are presented in Table 7. It can be seen that the cold plasma increased the electrolytic conductivity by about 2-2.5%. Similar results were obtained in [62]. The increase in conductivity indicates the presence of new ionic forms in the solution resulting from the electric discharge and ionic forms from the destruction of contaminating compounds contained in the water (e.g., H 3 O + , NO 2 − and NO 3 − ). The electrolytic conductivity of liquids is one of the important parameters affecting the intensity of UV radiation (it increases in solutions with higher conductivity) as well as the generation of a shock wave or thermal effects [29]. As reported by [63], an exceedingly high water conductivity is not desirable due to the lower production of • OH radicals and H 2 O 2 in cold plasma. The turbidity increased significantly after the first pass, which is due to the turbulent nature of the flow through the cold plasma reactor. After the next passes, the turbidity decreased due to the precipitation of impurities under the influence of various types of reducing forms. There was a noticeable increase in TOC, while the content of inorganic compounds slightly decreased. It is likely that the carbon present in the water, e.g., carbon dioxide dissolved in water, has been partially converted into simple organic substances.
The action of plasma, due to the UV light photons, ozone, radicals and atomic oxygen it contains, also leads to the degradation of microorganisms [62,64]. The synergistic action of these factors effectively reduced the amount of the total number of microorganisms in the tested water after just one pass through the plasma reactor (87%). Further passes through the reactor, i.e., three and six times, destroyed around 95% and 98% of bacteria, respectively, which proves a very good inactivation of microorganisms by the NTP reactor.

Studies on the Removal of Pb(II), Cd(II) and Ni(II)
The measurement results are summarized in Table 8 and presented in Figures 4 and 5. As shown in Figure 4, the pH increased after each pass through the cold plasma reactor. After the six passes, the increase was approximately 0.65. Growth weakened in subsequent cycles. Moreover, the percentage of metal removal tended to increase with the number of cycles ( Figure 5). It was observed that Pb(II) concentration decreased after the first pass, then increased slightly after the third pass, to decrease again after the sixth pass. These changes, however, were not large and may result from the turbulent nature of the solution flow and the related fluctuations in the local values of Pb(II) concentration. It is also possible that some components of the already precipitated phases re-ionized at a certain pH and entered the dissolved phase.
The percentage removal of the metal ions increased along with the number of passes ( Figure 5). It reached 72% after one pass for Pb(II) and 88% after six passes for Cd(II), whereas only around 19% for Ni(II). Figure 6 shows the results of the measurements in the coordinate system (∆pH, Removal). The values of the correlation coefficients between these two variables are also shown. Their values range from 0.83 to 0.99.    The percentage removal of the metal ions increased along with the nu ( Figure 5). It reached 72% after one pass for Pb(II) and 88% after six pas whereas only around 19% for Ni(II). Figure 6 shows the results of the me the coordinate system (ΔpH, Removal). The values of the correlation coeffi    The percentage removal of the metal ions increased along with the num ( Figure 5). It reached 72% after one pass for Pb(II) and 88% after six pas whereas only around 19% for Ni(II). Figure 6 shows the results of the me the coordinate system (ΔpH, Removal). The values of the correlation coeffic these two variables are also shown. Their values range from 0.83 to 0.99. Since pH has a great impact in the precipitation of the considered me that pH values were too low for the effective precipitation of Ni(II) compou removal of Ni(II) than Cd(II) is rather strange, because most references Ni(OH)2 precipitates at pH slightly lower than that for Cd(OH)2. For exam retical solubility diagram indicates that Ni(OH)2 precipitates at pH around that for Cd(OH)2 [65], but in [66] the difference is around 0.5. According t cipitation of Pb(OH) + occurs in the pH range of 4-11, and Pb(OH)2 starts at a 8, whereas the precipitation of Cd(OH) + occurs in the pH range of 7-12, starts at a pH of around 8. As for Ni(OH) + and Ni(OH)2, the starting pH is 8, respectively [68]. According to [69], the precipitation of Cd(II), Ni(II) and ides occurs for pH 7-14, 8-14 and 7-8, respectively. Thus, the differences b Ni(II) and Cd(II) are not great, and the presence of other factors, e.g., sp other forms due to cold plasma, could change the pH values as well as inf mation of other compounds [68]. This requires further research.
In addition, an increase in water temperature is visible in the subsequ ble 8). After the first pass, the temperature rise was 2-2.5 °C. In the subsequ increase was lower due to the increasing losses related to the transfer of solution to the device components. After six passes, the temperature incre 8-8.5 °C. The increase in water temperature is a visible symptom of ener water, although from the point of view of water purification, it is not a favo enon-a higher temperature usually increases the solubility of the compou poses their precipitation. Figures 7-9 show the Fourier transform infrared (FTIR) spectra of the t The samples Cd-0, Ni-0 and Pb-0 relate to the prepared solutions of Cd Pb(II) with a concentration of 10 mg/dm 3 before treatment with NTP. Al look very similar. There are three peaks characteristic for water and corres H bond stretching (around 3300 cm −1 ), H2O bending (1640 cm −1 ) and O-H cm −1 ) [70]. The lack of peaks characteristic for bonds with Cd(II), Ni(II) or P Since pH has a great impact in the precipitation of the considered metals, it follows that pH values were too low for the effective precipitation of Ni(II) compounds. The lower removal of Ni(II) than Cd(II) is rather strange, because most references indicate that Ni(OH) 2 precipitates at pH slightly lower than that for Cd(OH) 2 . For example, the theoretical solubility diagram indicates that Ni(OH) 2 precipitates at pH around 1 lower than that for Cd(OH) 2 [65], but in [66] the difference is around 0.5. According to [67], the precipitation of Pb(OH) + occurs in the pH range of 4-11, and Pb(OH) 2 starts at a pH of around 8, whereas the precipitation of Cd(OH) + occurs in the pH range of 7-12, and Cd(OH) 2 starts at a pH of around 8. As for Ni(OH) + and Ni(OH) 2 , the starting pH is around 7 and 8, respectively [68]. According to [69], the precipitation of Cd(II), Ni(II) and Pb(II) hydroxides occurs for pH 7-14, 8-14 and 7-8, respectively. Thus, the differences between pH for Ni(II) and Cd(II) are not great, and the presence of other factors, e.g., specific ions and other forms due to cold plasma, could change the pH values as well as influence the formation of other compounds [68]. This requires further research.
In addition, an increase in water temperature is visible in the subsequent passes (Table 8). After the first pass, the temperature rise was 2-2.5 • C. In the subsequent passes, the increase was lower due to the increasing losses related to the transfer of heat from the solution to the device components. After six passes, the temperature increased by about 8-8.5 • C. The increase in water temperature is a visible symptom of energy transfer to water, although from the point of view of water purification, it is not a favorable phenomenon-a higher temperature usually increases the solubility of the compounds, which opposes their precipitation.          The samples Cd-6, Ni-6 and Pb-6 relate to solutions after six passes through the cold plasma reactor. Compared to the Cd-0, Ni-0 and Pb-0 spectra, additional peaks are visible, which indicate the formation of various types of bonds. The peaks are summarized in Table 9. They are not unequivocally characteristic for the metals Cd(II), Ni(II) and Pb(II), which makes it difficult to identify the bonds clearly. According to [71,72], peaks 400-510 cm −1 in Cd-6 can be attributed to CdO. Peaks in the same region in Ni-6 spectrum are assigned to NiO and Ni-OH bonds [73,74]. No characteristic peaks related with Pb(II) are observed for Pb-6, which confirms Pb(II) was removed from the solution. Other peaks in the spectra are present in all samples of Cd-6, Ni-6 and Pb-6, which indicates the formation of bonds between elements naturally present in the tested water. Apart from the peaks related with water (~3300 cm −1 , 1640 cm −1 and 620 cm −1 ), there are other peaks indicating vibrations of OH (660 cm −1 -out-of-plane bending; 1236 cm −1 -in-plane bending). The large peak in region 1300-1500 cm −1 can be decomposed into two peaks: around 1327-1361 cm −1 and 1397-1456 cm −1 . The first one can be attributed to CO 3 2− [73] or CH 2 wagging [75], whereas the second can may indicate C-H bending or O-H bending [70,75]. The peaks in region 800-900 cm −1 are assigned to C-O-O vibrations, and those in region 1000-1150 cm −1 -to C-O bond [70,71,75]. Small peaks in region 2800-3000 cm −1 are related to CH 2 stretching [75]. This suggests that some organic substances are created during the plasma reactor operation, which agrees with the observation of the increased value of TOC (see Table 7).  [70,76] The samples marked Cd-6s, Ni-6s and Pb-6s were taken from the sludge after six passes. The spectra of Cd-6s and Ni-6s are similar to those for Cd-6 and Ni-6, respectively. This is because the collected sediment was very diluted due to its small amount. Therefore, control sediments were made (NaOH was added to the prepared Cd(II) and Ni(II) solutions not subjected to plasma treatment to precipitate Cd(II) and Ni(II))-these samples are marked Cd-cs and Ni-cs. Apart from the peaks related with CdO and NiO in region 400-510 cm −1 , there are also very sharp peaks 3600 cm −1 and 3639 cm −1 , which can be attributed to Cd(OH) 2 and Ni(OH) 2 , respectively [70,76]. Comparing the spectra of Cd-6s with Cd-cs and Ni-6s with Ni-cs, it can be seen that some of the peaks appear in both spectra, which indicates the presence of Cd(II) and Ni(II) in Ni-6s and Cd-6s samples in the form of various types of complexes.
In the case of Pb-6s, the sediment was clearly visible, and it has a white tint. The FTIR spectrum of Pb-6s contains a distinct wide peak around 1400 cm −1 and several narrow small peaks: 1733 cm −1 (attributed to C=O bond stretching [71]), 1051 cm −1 (attributed to C-O bond [70]), 839 cm −1 (attributed to C-O-O vibrations [70]) and 678 cm −1 (out-of-plane bending of OH [70]). This suggest that at least part of Pb(II) could be formed in lead white (2PbCO 3 ·Pb(OH) 2 ).
The results of the research on the effect of metal ions concentration are presented in Table 10. Based on them, the percentage removal (R) was calculated and presented in Figure 10. It follows that, in the analyzed range of concentrations, the percentage removal depended insignificantly on the initial concentration. 6s with Cd-cs and Ni-6s with Ni-cs, it can be seen that some of the peaks app spectra, which indicates the presence of Cd(II) and Ni(II) in Ni-6s and Cd-6s the form of various types of complexes.
In the case of Pb-6s, the sediment was clearly visible, and it has a white tin spectrum of Pb-6s contains a distinct wide peak around 1400 cm −1 and seve small peaks: 1733 cm −1 (attributed to C=O bond stretching [71]), 1051 cm −1 (a C-O bond [70]), 839 cm −1 (attributed to C-O-O vibrations [70]) and 678 cm −1 (o bending of OH [70]). This suggest that at least part of Pb(II) could be formed in (2PbCO3⋅Pb(OH)2).
The results of the research on the effect of metal ions concentration are p Table 10. Based on them, the percentage removal ( ) was calculated and prese ure 10. It follows that, in the analyzed range of concentrations, the percenta depended insignificantly on the initial concentration.   Figure 11 shows the change in the pH of the tested solutions after one p the cold plasma reactor. In each case, an increase in pH was observed af through the reactor. For higher metal concentrations, there was also an overall trend in ΔpH (Table 10). This could be explained by the precipitation mechan of OHions were removed from the solution in the form of metal hydroxides.  Figure 11 shows the change in the pH of the tested solutions after one pass through the cold plasma reactor. In each case, an increase in pH was observed after passing through the reactor. For higher metal concentrations, there was also an overall downward trend in ∆pH (Table 10). This could be explained by the precipitation mechanism-a part of OH − ions were removed from the solution in the form of metal hydroxides.
In order to estimate the energy used for water ionization, the energy consumed by the reactor was measured. In one cycle, it was on average 0.095 kWh (340 ± 30 kJ). A part of this energy was used to heat the water-the average increase in temperature was around ∆T = 2.3 ± 0.2 • C. As the volume of the solution was 5 L, the energy for heating the water in one cycle was approximately 5 kg × 4.2 kJ/(kg • C) × 2.3 • C ≈ 48 kJ. The rest of the energy-about 86%-was mainly used to ionize the water. In order to estimate the energy used for water ionization, the energy the reactor was measured. In one cycle, it was on average 0.095 kWh (340 of this energy was used to heat the water-the average increase in tem around Δ = 2.3 ± 0.2 °C. As the volume of the solution was 5 L, the ene the water in one cycle was approximately 5 kg × 4.2 kJ/(kg °C) × 2.3 °C ≈ 48 the energy-about 86%-was mainly used to ionize the water.
The energetical efficacy of the metal ions removal after one pass is pr ure 12. In the considered range of the concentrations of metal ions, the effi imately proportional to the initial concentration.

Discussion on the Mechanism of Metal Ion Removal
Based on the results obtained, it follows that the increase in pH seem for metal ions removal with NTP. The mechanism of the increase is the processes occurring in the solution affected by the cold plasma. A sufficien Figure 11. pH increase of the Pb(II), Ni(II) and Cd(II) aqueous solutions versus the initial concentration before the exposure to cold plasma and after one cycle of exposure.
The energetical efficacy of the metal ions removal after one pass is presented in Figure 12. In the considered range of the concentrations of metal ions, the efficacy is approximately proportional to the initial concentration. Figure 11. pH increase of the Pb(II), Ni(II) and Cd(II) aqueous solutions versus the tration before the exposure to cold plasma and after one cycle of exposure.
In order to estimate the energy used for water ionization, the energy the reactor was measured. In one cycle, it was on average 0.095 kWh (340 ± of this energy was used to heat the water-the average increase in tem around Δ = 2.3 ± 0.2 °C. As the volume of the solution was 5 L, the ener the water in one cycle was approximately 5 kg × 4.2 kJ/(kg °C) × 2.3 °C ≈ 48 the energy-about 86%-was mainly used to ionize the water.
The energetical efficacy of the metal ions removal after one pass is pr ure 12. In the considered range of the concentrations of metal ions, the effic imately proportional to the initial concentration.

Discussion on the Mechanism of Metal Ion Removal
Based on the results obtained, it follows that the increase in pH seem for metal ions removal with NTP. The mechanism of the increase is the r processes occurring in the solution affected by the cold plasma. A sufficien tric field causes the ionization of gas in the air bubbles in the water. Additi tion is facilitated by the natural presence of a certain number of free elec electrons are accelerated in the electric field and gain energy large enoug excite other molecules. As a result, new electrons are released from the mole radiation quanta are produced, which also enhance the ionization proces reactions leading to the formation of free radicals, ions and excited forms d action of an electron with water molecules are as follows [22]:

Discussion on the Mechanism of Metal Ion Removal
Based on the results obtained, it follows that the increase in pH seems to be crucial for metal ions removal with NTP. The mechanism of the increase is the result of many processes occurring in the solution affected by the cold plasma. A sufficiently strong electric field causes the ionization of gas in the air bubbles in the water. Additionally, ionization is facilitated by the natural presence of a certain number of free electrons [23]. The electrons are accelerated in the electric field and gain energy large enough to ionize or excite other molecules. As a result, new electrons are released from the molecules, and UV radiation quanta are produced, which also enhance the ionization process. Examples of reactions leading to the formation of free radicals, ions and excited forms due to the interaction of an electron with water molecules are as follows [22]: The excited water molecule H 2 O* in interaction with other water molecule can decompose into H • and • OH radicals or into H 2 and O or into 2H • and O. From the point of view of purifying water from metal ions, such as Pb(II), Cd(II) and Ni(II), due to the precipitation of their complexes in the form of hydroxides, the presence of OH − ions in the water is important. Barrier discharges increase their content. These ions can be formed as a result of many different reactions, amongst which the most intensive are as follows: Some of the reactions as well as others are illustrated in Figure 13.
H2O + e − → H2O* + e − The excited water molecule H2O* in interaction with other water molecule can decompose into H • and • OH radicals or into H2 and O or into 2H • and O. From the point of view of purifying water from metal ions, such as Pb(II), Cd(II) and Ni(II), due to the precipitation of their complexes in the form of hydroxides, the presence of OH − ions in the water is important. Barrier discharges increase their content. These ions can be formed as a result of many different reactions, amongst which the most intensive are as follows: Some of the reactions as well as others are illustrated in Figure 13. The presence of oxygen significantly increases the production of OH − ions. This mechanism can be described as a series of the following reactions: O + O2 → O3 (12) Figure 13. Suggested mechanism of metal ion precipitation in the aqueous solution due to nonthermal plasma (selected reactions).
The presence of oxygen significantly increases the production of OH − ions. This mechanism can be described as a series of the following reactions: (13) and then the reaction (8) leading to the formation of the OH − ion. Some of the other oxygen-mediated reactions under the conditions under consideration are [77]: They all increase the pH of the solution directly or indirectly. The increase in pH was observed in all considered cases (see Tables 8 and 10, Figure 11). In such conditions, the metal ions Pb(II), Cd(II) or Ni(II) can more easily combine with OH − ions, e.g., Similar reactions take place with the participation of Cd(II) and Ni(II) ions. It is also possible that other compounds, such as oxides or carbonates, can be formed.

Comparison with Chemical Precipitation
The suggested mechanism is based on pH increase, which is similar to traditional precipitation methods based on adding various agents, such as NaOH and Ca(OH) 2 . Therefore, it is worth comparing both methods (Table 11). It follows that the NTP method offers similar levels of percentage removal for particular metals as chemical precipitation with the use of low-cost precipitants. It is worth emphasizing that the NTP technology does not require the addition of other substances, often expensive (e.g., active carbons, synthesized adsorbents and precipitants, and ion exchange resins). It also does not contain expensive components subject to frequent replacement (e.g., filter membranes). It does not produce secondary waste (such as in adsorption), which is often troublesome in further processing. It is characterized by a fast and effective operation for many types of pollutants [35]. The need to supply electricity may be considered a disadvantage, but in the era of the development of photovoltaics, solutions based on solar energy are not a problem-it even can be considered as an advantage.

Conclusions
Based on the results obtained in the research, the following conclusions can be formulated: • Cold plasma with a dielectric barrier discharge can be used to remove metal ions from aqueous solutions; in the considered case, the percentage removal was 72% for Pb(II), 88% for Cd(II) and 19% for Ni(II).

•
The removal mechanism is based on an increase in pH of the solution affected by the cold plasma. Each pass of the solution through the cold plasma reactor increased pH by around 0.1-0.2; therefore, a significant effect on pH increase requires several passes (cycles) of the same solution through the cold plasma region. This may be changed by lowering the flow rate, enlarging the length of the active area or applying a cascade of several ionizers.

•
The initial concentration of metal ions in the tested ranges had a small impact on the percentage removal of the metal ions-this method could be used for larger metal ions concentrations, because the energetical efficacy is largest.

•
The optimization of the NTP generator settings confirmed that the most important is the flow rate, which should be as low as possible to increase the time of the impact of the cold plasma on the treated solution.

•
The chemical indicators in the tested water mostly changed positively; in some cases, such as TOC, the change was undesired, but remained within the permissible limits.

•
The total number of microorganisms dropped significantly after the cold plasma treatment, which is in agreement with the literature.