Formation of Nitrogen Oxides by Nanosecond Pulsed Plasma Discharges in Gas–Liquid Reactors

Abstract

A gas–liquid-film flow reactor with a nanosecond pulsed power supply was utilized to produce nitrogen oxides from Ar/N₂ mixtures (gas phase) and deionized water (liquid phase). Chemical analysis of the stable products found in both the gas and liquid phases was performed and chemical quenching was incorporated for the liquid phase samples in order to eliminate post plasma reactions. Significant amounts of NO and NO₂ in the gas phase and ${\text{NO}}_{2}^{ - }$ and ${\text{NO}}_{3}^{ - }$ in the liquid phase were determined using FTIR spectroscopy and ion chromatography, respectively. The production rate of all nitrogen oxides produced increased significantly with N₂ concentration while H₂O₂ formation decreased slightly. The gas temperature of the plasma was approximately 525 K and was unaffected by N₂ concentration while the electron density ranged from 1 × 10¹⁷ cm⁻³ in pure Ar to 5.5 × 10¹⁷ cm⁻³ in 28% N₂. The role of the ${\cdot {\text{OH}}}$ in the reaction pathway was assessed by adding CO as a gas phase radical scavenger showing that ${\cdot {\text{OH}}}$ is essential for conversion of the gas phase NO and NO₂ into water soluble ${\text{NO}}_{2}^{ - }$ and ${\text{NO}}_{3}^{ - }$. Conversely, atomic oxygen originating from water is likely responsible for NO and NO₂ generation. Experiments with N₂/O₂/Ar mixtures and air showed a significant increase in NO₂ production caused by the additional generation of reactive oxygen species. An overall energy yield for all nitrogen oxides produced in the most efficient case was 50 eV/molecule.

Appendix

Electrical Measurements

Temporal alignment of current and voltage probes is essential for accurate analysis of nanosecond discharges. For temporal alignment of these probes the nanopulser was discharged across a resistor placed inside the Rogowski coil while the two voltage probes measured the electrical potential across the resistor. The deskew of each voltage probe was then adjusted to ensure simultaneous rise in each voltage probe with current.

For analysis of the plasma discharge, the high voltage probes were connected across the stainless-steel inlet and outlet capillaries to measure the electric potential across the plasma discharge region. The reactor was placed inside the Rogowski coil so that the discharge region was centered within the coil to measure current flow through the discharge region only. With placement of the voltage and current probes in this fashion, the measured values give the electrical potential between the electrodes and current flow only within the formed plasma channels. More details of the electrical analysis can be found in our previous work [60, 62].

H₂O₂ Quantification

For H₂O₂ quantification the effluent was discharged directly into a sodium azide solution of sufficient quantity to rapidly remove all ${\text{NO}}_{2}^{ - }$ from the liquid phase. 2 mL of effluent from reactor was discharged into a 1 M sodium azide solution at an equal 1–1 volume ratio (i.e. 2 mL sodium azide+2 mL effluent). The H₂O₂ in the resulting solution was then quantified with the TiSO₄ colorimetric method using a UV–Vis spectrometer detailed in our previous publications as well as others (PerkinElmer, Lambda 35) [58, 61, 87]. The resulting concentration was then multiplied by a factor of 2 in order to account for the dilution with the sodium azide solution.

Nitrate and Nitrite Quantification

For nitrate and nitrite quantification the pH of the effluent must be rapidly increased to slow/stop the reaction of ${\text{NO}}_{2}^{ - }$ with H₂O₂. In addition, the solutions were diluted to approximately 10 ppm total nitrogen for protection of the IC instrument. For this procedure, 3 mL of effluent was directly discharged into a 26.7 mL solution of 1 mM NaOH. The pH of the resulting mixture after effluent addition was approximately 10. Calibration curves for the IC were prepared using various dilutions of NaNO₂ and KNO₃. All samples were analyzed on a Dionex, Aquion ion chromatograph equipped with an IonPac™ AS22 column (4 × 250 mm), an AERS™ 500 carbonate suppressor, and a conductivity detector. The eluent was 4.5 mM sodium carbonate mixed with 1.4 mM sodium bicarbonate and supplied to the IC system at an isocratic flow of 1.2 mL/min. A sample IC spectrum is shown in Fig. 13.

Fig. 13

Sample IC spectrum for 28% N₂ with balance argon as the carrier gas

NO and NO₂ Quantification

An FTIR (Jasco – 6700) equipped with a gas flow cell was utilized to analyze the gas phase effluent. Calibration curves for NO and NO₂ were generated by flowing dilutions of 1000 ppm gas standards generated with a mass flow controller (Brooks, 5850E) and balanced with N₂ at the same total flow rate as the experimental conditions (3.4 L/min). For NO the band at 1917.86 cm⁻¹ was used for quantification while for NO₂ the band at 1637.78 cm⁻¹ was utilized. The following instrument conditions were used for all FTIR measurements: 8 accumulations, 0.25 cm⁻¹ resolution, 32 gain, 1.8 mm aperture, and 2 mm/s scanning speed. Prior to measurements the gas flow cell was purged with N₂ to remove water and air. All measurements were taken after a steady state in the flow cell had been reached by ensuring no change in the intensity of the bands for a period of two minutes. Sample experimental FTIR spectra are shown in Figs. 14 and 15.

Fig. 14

Sample FTIR spectrum for 28% N₂ with balance Ar as the carrier gas

Fig. 15

Sample FTIR spectrum for air as the carrier gas

Optical Emission Spectroscopy

For OES measurements the probe tip was oriented towards the center of the discharge region in as close proximity as possible to the quartz reactor body (approx. 5 mm). Stark broadening of the $H_{\alpha }$ peak was utilized for estimation of the electron density. The $H_{\alpha }$ peak was observed using an optical emission microscope (Avantes, AvaSpec-ULS3848; wavelength range: 600–700 nm) at 656.3 nm. The $H_{\alpha }$ peak was deconvoluted into two Voigt profile. The Voigt profile was then deconvoluted into Gaussian and Lorentz profiles. The FWHM of Gaussian profile was determined by Doppler and instrument broadening where Doppler broadening was obtained using gas temperature, and instrumental broadening was obtained by calculating the FWHM of the spectral line with the optical fiber connected to a mercury-argon lamp (Avalight-CAL). The FWHM of the Stark broadening was calculated from the relationship between the FWHM of the Van der Waal’s broadening and the FWHM of the Lorentzian profile. More details of this method can be found in a previous publication and a sample of the peak fitting can be found in Figs. 16 and 17 [63].

Fig. 16

$H_{\alpha }$ peak fitting in argon discharge by two Voigt profile for 28% N₂ in argon

Fig. 17

Optical emission spectra for 28% nitrogen in argon showing a band at 777 nm which indicates presence of atomic oxygen