Non‐Thermal Plasma in Contact with Water: The Origin of Species

Abstract Non‐thermal atmospheric pressure plasma has attracted considerable attention in recent years due to its potential for biomedical applications. Determining the mechanism of the formation of reactive species in liquid treated with plasma is thus of paramount importance for both fundamental and applied research. In this work, the origin of reactive species in plasma‐treated aqueous solutions was investigated by using spin‐trapping, hydrogen and oxygen isotopic labelling and electron paramagnetic resonance (EPR) spectroscopy. The species originating from molecules in the liquid phase and those introduced with the feed gas were differentiated by EPR and 1H NMR analysis of liquid samples. The effects of water vapour and oxygen admixtures in the feed gas were investigated. All the reactive species detected in the liquid samples were shown to be formed largely in the plasma gas phase. It is suggested that hydrogen peroxide (determined by UV/Vis analysis) is formed primarily in the plasma tube, whereas the radical species ⋅OOH, ⋅OH and ⋅H are proposed to originate from the region between the plasma nozzle and the liquid sample.


Table of contents
Experimental details 2 Figure S1. Voltage and return current waveforms with different feed gas compositions 4 Table S1. The results of the OES analysis of the gas phase plasma inside the quartz tube between the electrodes 4 Figure S2. Calibration curves for the analyses 5 Table S2. Error assessment 6 Figure S3. A schematic representation of the experimental setup 6 Table S3. H 2 O concentration in the feed gas with helium passing through a Drechsel flask 6 Table S4. H 2 O content in the liquid H 2 17 O sample after plasma exposure 7 Figure S4. EPR spectra of DMPO and PBN spin adducts in experiments with magnesium fluoride window 7 Figure S5. H 2 O amount in a D 2 O liquid sample delivered by helium flow 7 Figure S6. H 2 O 2 concentration in the liquid as a function of O 2 content of the feed gas 8 Figure S7. DMPO-H adduct concentration in the liquid sample 8 Figure S8. DMPO-OH adduct concentration in the liquid 8 Table S5. Experimental investigation of primary kinetic isotope effect 9 Table S6. Concentrations and relative amounts of PBN-H and PBN-D radical adducts 10 Table S7. Experimental and calculated relative amount of PBN adducts 11 Figure S9. DEPMPO-OOH and DEPMPO-OH radical adduct concentration 11 Figure S10. The concentration of the DEPMPO-H adduct 12 Figure S11. The concentration of the DEPMPO adduct of a carbon-centred radical 12 Figure S12. DEPMPO-OOH and DEPMPO-OH adduct concentrations 13 Figure S13. Concentration of the formed TEMPO 14 Figure S14. Concentration of the formed TEMPO with added sodium azide 14

Experimental Methods
The plasma was ignited in a quartz tube (4 mm ID and 6 mm OD, 100 mm length) surrounded by copper electrodes (10 mm width) separated by 20 mm. A PVM500 Plasma Resonant and Dielectric Barrier Corona Driver power supply (Information Unlimited) was used to sustain the plasma. A high voltage probe (Tektronix P6015A) and current probe (Ion Physics Corporation CM-100-L) were used with a Teledyne LeCroy WaveJet 354A oscilloscope to measure time resolved current and voltage. Voltage and frequency were kept constant throughout all experiments at 18.3 ± 0.2 kV (peak-to-peak) and 24.9 kHz, respectively. The return current values were between ca. 4 and 7 mA. The voltage and current waveforms at some of the experimental conditions are shown in Fig. S1. OES measurements of the plasma between the electrodes were performed with Ocean Optics HR-4000CG-UV-NIR spectrophotometer (Table S1).
The plasma was operated with a feed gas of helium with oxygen and water admixtures controlled by mass flow controllers (MFCs) (Brooks Instruments and Brooks Instruments 0254 microcomputer controller). All experiments were carried out with a total flow of feed gas of 2 L/min. The percentage of the O 2 admixture is shown in vol%, and the concentration of water vapour is quoted in percent of the saturation and mol%.
The experimental setup was positioned inside a large Faraday cage with the mesh size of 22 mm.

Analysis
Electron paramagnetic resonance (EPR) measurements were carried out on a Bruker EMX Micro EPR spectrometer. The EPR analysis parameters were as follows: frequency 9.83 GHz, power 3.17 mW, modulation frequency 100 kHz, modulation amplitude 1 G, time constant 40.96 msec, number of scans 5, sweep width 100 G (DMPO and PBN adducts, TEMPO) or 170 G (DEPMPO addcuts). For the measurements, all samples were contained in glass capillary tubes (80 x 1 mm) purchased from Marienfeld Laboratory Glassware. EPR calibration was performed using aqueous solutions of a stable radical (TEMPO) in a range of concentrations 2-200 M (Fig. S2). After each plasma exposure experiment, the samples were immediately placed in a capillary tube. The overall time between the exposure and recording the spectrum was 2 minutes.
Concentration of H 2 O 2 in the samples was determined by UV-Vis measurements performed on a UV-1800 Shimadzu UV-Vis Spectrophotometer with Optical Glass High Precision Cells (10 mm light path) provided by Hellma Analytics. UV-Vis calibration was done using 500 L titanium(IV) reagent with added 300 L aqueous hydrogen peroxide solutions in a range of concentrations 0.0979-4.895 mM (Fig. S2). Titanium(IV) reagent was prepared by dissolving 3.54 g of potassium bis(oxalato)oxotitanate(IV) dihydrate in a mixture of 27.2 mL of sulphuric acid and 30 mL of H 2 O, and diluting the resulting solution to a total volume of 100 mL.
UV-Vis spectra of samples were recorded by adding a mixture of 65 L of plasma-exposed sample (taken immediately after plasma exposure) with 235 L of H 2 O to 500 L of titanium(IV) reagent. The resulting solutions were incubated for 1 min before analysis. The H 2 O 2 concentration was determined from the UV-Vis intensity of the peak at 400 nm.
For 1 H NMR analysis, 50 L of plasma-exposed sample was added to 500-600 L of 0.5 M sodium tosylate solution in D 2 O in a Young NMR tube and kept under argon. The NMR spectra were recorded on a JEOL ECS400 spectrometer.
The composition of H 2 17 O samples was analysed as follows. In a typical experiment, 10 L of a 1.5 M solution of cinnamoyl chloride in acetonitrile was added to 10-15 uL of samples containing H 2 17 O in a small vial filled with argon. The mixture was heated to 60 o C for 2 min to allow full hydrolysis. The samples were cooled down and diluted to ca. 100  concentrations with 1:1 water:acetonitirile mixture. The cinnamic acid formed as a result of the cinnamoyl chloride hydrolysis was analysed using high-resolution MS spectrometry. Mass spectra were acquired using a Bruker 9.4T solariX XR Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometer (Bremen, Germany). The samples were ionized in positive ion mode using the ESI ion source. Spectra were measured with a transient length of 0.84 s resulting in a resolving power of 400000 at m/z 150. The instrument was externally calibrated using sodium formate clusters.  (Fig. S2).
The concentrations of all reactive species in the liquid samples are quoted after correction for the material lost through solvent evaporation.

Error assessment
It must be acknowledged that we have found that the results of the plasma exposure of the samples (e.g., the absolute values of concentration of DMPO-OH) were largely affected by small changes in the configuration of the jet, such as the electrodes contact with the quartz tube, the depth of the tube protrusion inside the reactor, and the vertical alignment of the tube. However, while the numerical values changed, the observed trends remained persistent. For example, the concentration of DMPO-OH increased with the initial introduction of H 2 O to He feed gas and decreased with higher H 2 O content, the concentration of DMPO-OH was lower at 4 mm distance than 10 mm, etc.
Thus, the error assessment was performed within a set configuration of the jet for several conditions (Table S2). We found that conditions of less uniform plasma nature (i.e., in the presence of large amounts of admixtures in the feed gas) generally lead to an increase in standard deviation of the concentration values. The maximum deviation from the mean was found to be ca. 12%.

Plasma exposure experiments
In a typical experiment, 100 L of liquid sample was placed in a well on top of a glass stand inside the reactor. The distance from the nozzle to the sample was 10 mm unless stated otherwise. In experiments when the samples were at the 4 mm distance from the sample to the nozzle, the distance between the live electrode and the sample was maintained at 20 mm (Fig. S3). Thus, the plasma length from the core plasma remained the same throughout all experiments, and the ratio of its quartz surroundings changed (we acknowledge that the plasma jet in contact with quartz will propagate slightly differently than in contact with surrounding gas; nevertheless, this still provided insight into the interaction dynamics with the surrounding atmosphere). The distance between the electrodes was 20 mm in all experiments. The reactor was flushed with the feed gas for 20 s and then exposed to plasma for 60 s.
In The experiments involving different feed gas humidity were performed by using split helium flow (i.e., by mixing dry helium with water-saturated helium in desired proportions). Water-saturated helium was made by bubbling dry helium through a water-filled Drechsel flask at 20 °C. The relative humidity was determined by weighing the flask before and after the experiment and comparing the data with the available literature values (Table S3) Figure S4. EPR spectra of DMPO (left) and PBN (right) spin adducts obtained in experiments with direct plasma exposure of the sample and through the magnesium fluoride window.   H and PBN-D). *** Calculated as shown in Eq. S1.
Primary kinetic isotope effect factor was calculated using the following equation: Here, the KIE calculations were performed under assumption that H and D radicals diffuse into the liquid from the gas phase with the same rate. Realistically, in a hypothetical situation when same amounts of H and D are formed in the gas phase, (2/1)^0.5 = 1.4 times less D atoms will reach the liquid sample surface in the same period of time.