Simultaneous hydrodynamic cavitation and nanosecond pulse discharge plasma enhanced by oxygen injection

Highlights

• •A novel Hydrodynamic Cavitation-Assisted Oxygen Plasma (HCAOP) process, which employs a venturi tube and oxygen injection, has been developed for enhancing the production and utilization of hydroxyl radicals (·OH) in the degradation of organic pollutants. The introduction of an appropriate amount of oxygen was found to increase the ·OH energy efficiency by sixfold to 119.8 nmol/J.
• •The impact of fluid characteristics on discharge behavior and hydroxyl radical production has been revealed.
• •Numerical simulations have provided insights into the electric field distribution of the two-phase fluid in the Venturi tube under different flow regimes.
• •HCAOP treatment of model pollutant Indigo Carmine achieved 99% degradation of 20 mg/L E132 in 5 L of water within 2 min, with an EE0 of only 0.26 kWh/m³/order.

Abstract

A novel Hydrodynamic Cavitation-Assisted Oxygen Plasma (HCAOP) process, which employs a venturi tube and oxygen injection, has been developed for enhancing the production and utilization of hydroxyl radicals (·OH) in the degradation of organic pollutants. This study has systematically investigated the fluid characteristics and discharge properties of the gas–liquid two-phase body in the venturi tube. The hydraulic cavitation two-phase body discharge is initiated by the bridging of the cavitation cloud between the electrodes. The discharge mode transitions from diffuse to spark to corona as the oxygen flow rate increases. The spark discharge has the highest current and discharge energy. Excessive oxygen results in the change of the flow from bubbly to annular and a subsequent decrease in discharge energy. The effects of cavitation intensity, oxygen flow rate, and power polarity on discharge characteristics and ·OH production were evaluated using terephthalic acid as a fluorescent probe. It was found that injecting 3 standard liter per minute (SLPM) of oxygen increased the ·OH yield by 6 times with only 1.2 times increase in power, whereas<0.5 SLPM of oxygen did not improve the ·OH yield due to lower breakdown voltage. Negative polarity voltage increased the breakdown voltage and ·OH yield due to asymmetric density and pressure distribution in the throat tube. This polarity effect was explained by numerical simulation. Using indigo carmine (E132) as a model pollutant, the HCAOP process degraded 20 mg/L of dye in 5 L water within 2 min following a first-order reaction. The lowest electric energy per order (E_EO) was 0.26 (kWh/m³/order). The HCAOP process is a highly efficient flow-type advanced oxidation process with potential industrial applications.

1. Introduction

In recent years, there has been a growing concern about the detection of organic pollutants in water environments [1], [2], [3]. Traditional water treatment methods such as coagulation, sedimentation, ozonation and chlorine disinfection are often ineffective for treating pollutants like antibiotics and perfluorooctanoic acid which have demonstrated resistance [4], [5]. Advanced oxidation processes (AOPs) that use high concentrations of hydroxyl radicals (∙OH) are becoming an attractive method for degrading these pollutants [6], [7], [8]. One of the advantages of AOPs is that they can mineralize organic pollutants to CO₂ and H₂O through chain reactions, without creating toxic by-products such as halogenated hydrocarbons, which may be produced by chlorine disinfection [9], [10].

Among AOP technologies, discharge plasma has gained considerable attention because it does not require chemicals or catalysts [11]. Various types of gas discharges, such as corona discharge [12], [13], [14], [15], dielectric barrier discharge (DBD) [16], [17], [18], gliding arc discharge [19], [20], and pulsed spar discharge [21], [22], [23], have been applied in plasma advanced oxidation processes. Although liquid-phase discharge can generate high concentrations of hydroxyl radicals, it requires extremely high breakdown voltages, resulting in significant equipment costs. Therefore, researchers have attempted to use gas-phase discharge above the water surface to be treated. However, this approach has its limitations, as the active species generated by gas discharge have a very short lifetime, which restricts their ability to penetrate the gas–liquid interface and react with the target molecules within a limited depth in the liquid phase [24], [25].

In recent years, researchers have sought to increase the efficiency of plasma activation by exploring the potential of bubble-enhanced cold plasma activation. This approach leverages the advantages of small bubbles, such as high mass transfer efficiency, high surface-to-volume ratio, long residence time in liquid, and high internal pressure, to improve oxidation activation [26], [27], [28], [29]. Additionally, active species like O¹D and O³P generated by the discharge can produce hydroxyl radicals through subsequent reactions. One particularly promising option is the use of gas–liquid two-phase discharge with pure oxygen injection [30], [31], [32]. This approach has the potential to further enhance hydroxyl radical production and improve the degradation of organic pollutants in water.

Gas-liquid two-phase flow in a Venturi tube is different from conventional bubble-type gas–liquid two-phase flow. It is generated by cavitation, which creates low gas pressure, depending on the saturated vapor pressure, and is ideal for the generation of discharge plasma. In a hydrodynamic cavitation device based on the Venturi water jet principle, the water volume decreases at the throat section, and the flow velocity and pressure increase sharply [33], [34], [35]. As a result, a large number of tiny cavitation bubbles are produced in the water body when the pressure drops below the saturated vapor pressure. These bubbles are in the micrometer or sub-micrometer range, which is difficult to achieve using traditional bubbling technology. When the bubble cloud reaches the downstream pressure recovery zone, the liquid pressure rises sharply, compressing the gas inside the bubbles and causing their gas pressure and temperature to rise. Upon collapse of the cavitation bubbles, a transient high-temperature (10³ ∼ 10⁴K) and high-pressure (0.1 ∼ 10 MPa) environment is formed, accompanied by strong shock waves and microjets inside the cavitation bubbles [36], [37]. The cavitation effect accelerates the transfer of active substances from the gas phase to the liquid phase and provides extreme physical conditions for the generation of hydroxyl radicals and the oxidative degradation of organic pollutants [36], [38], [39], [23]. In recent years, different research groups have developed three types of gas–liquid two-phase discharge devices based on Venturi tube hydrodynamic cavitation, and these devices have been experimentally verified as a new advanced oxidation technology with potential applications [40], [41], [42], [43]. However, because ·OH in hydrodynamic cavitation is mainly produced by dissociation, ionization, and other reactions of gaseous water molecules, ·OH yield is low. To improve ·OH yield, introducing oxygen is a feasible measure, as the Venturi tube itself has negative pressure suction, which can draw in working gas without installing a gas pump [36].

In this study, we present a design for a hydrodynamic cavitation-assisted oxygen plasma (HCAOP) water treatment process based on a venturi tube. Our investigation demonstrates that, unlike conventional plasma technologies, the fluid characteristics of hydraulic cavitation within the venturi tube are the determining factors for discharge properties and the production rate of hydroxyl radicals (·OH) [39], [43]. By examining the chain process in which fluid characteristics alter discharge properties, subsequently affecting the ·OH production rate and energy consumption, we have identified optimal operating conditions and universal principles for the HCAOP process.

Under the optimal conditions of our system, a high ·OH production per unit energy consumption of up to 119.8 nmol/J was achieved. Furthermore, efficient decolorization was observed in the treatment of indigo carmine red, a model organic pollutant commonly used to evaluate plasma water treatment technologies [43], [44]. This highlights the potential of the HCAOP process in effectively addressing organic contaminants and advancing the field of water treatment.

2. Materials and methods

2.1. Experimental procedures

2.1.1. Experimental rig for water treatment

The experimental setup shown in Fig. 1 consists of two tungsten needle electrodes, each with a 3 mm diameter, inserted into the gas–liquid mixture through inlet and outlet ports, with a fixed distance of 20 mm between them. Oxygen was injected into the gas–liquid mixture at the venturi's throat through a mass flow controller (MFC), and the pressure of the injected oxygen was measured by pressure gauge 3 (YB80A2, −0.5 to 0.5 MPa, YBPCM, China).

Fig. 1.

Experimental arrangement.

To generate discharge plasma, we used a Solid-State (SOS) nanosecond pulsed power supply with a rapid rise time (30 ns) and narrow pulse width (150 ns), resulting in high energy efficiency. The peak output voltage of the SOS power supply in this experiment is 50 kV, and the pulse repetition frequency is adjustable from 1 to 1000 Hz.

2.1.2. Design and fabrication of venturi tube

In this study, a venturi tube was designed using Solidworks software, as shown in Fig. 2, with inlet and outlet ports as well as a gas inlet port. Polymethyl methacrylate (PMMA) was chosen as the material for construction due to its transparency to visible light. However, PMMA does not allow the transmission of UV-light with wavelengths below 420 nm, whereas the emission spectrum of hydroxyl radicals (·OH) is at 309 nm.

Fig. 2.

Schematic diagram of the proposed Venturi channel.

2.1.3. Electrical measurement

The discharge energy of a single pulse can be calculated by integrating its instantaneous power. Since the pulsed discharge is random, we collected 200 pulse waveforms for each experimental condition and calculated the average single-pulse energy (E_mean) by taking their average value. The discharge power can be determined by multiplying the average single-pulse energy by the pulse repetition frequency f, as shown in the following equation:

$$\text{P}=f\overline{\left({\int }_{t_0}^{t_0+T}u\left(t\right)i\left(t\right)dt\right)}$$ (2)

2.1.4. Visual observations of cavitation and plasma

The experimental setup for tracking the cavitation and plasma state in the venturi tube involved the use of a SONY SLT-A57 camera. To visualize the cavitation state, an LED backlight with a diffuser was positioned to illuminate the bubble clouds generated at different flow rates. In contrast, plasma imaging was performed in complete darkness.

2.1.5. Plasma emission spectrum

The optical emission spectroscopy (OES) system consisted of a spectrometer (HR4D1943, Optics Inc., USA) equipped with an optical fiber that was connected to a computer, capable of operating in both ultraviolet and visible light from 200 to 900 nm. To capture more emitted light from the plasma nozzle, an optical condenser was installed at the center of the nozzle with a distance of 5 cm to one tip of the optical fiber.

2.1.6. Fluorescent detectors for hydroxyl radical

The hydroxylation product of phthalic acid, hydroxy phthalic acid, exhibits excitation at a wavelength of 315 nm and emission at 425 nm. To measure the concentration of the sample, we rely on the fluorescence intensity and construct a calibration curve with known concentrations of hydroxy phthalic acid. For fluorescence detection, we employ an enzyme-linked immunosorbent assay (ELISA) instrument (E300). Prior to discharge treatment, phthalic acid must be dissolved in NaOH and prepared as a 2 mM solution.

2.1.7. Spectra of E132

The UV–Vis spectrum of E132 was obtained using an AUCY UV1901PC Spectrophotometer, which detected wavelengths ranging from 200 to 800 nm. We calculated the dye degradation efficiency (η) and degradation rate by the following equations:

$$\eta =\frac{C_0-C_t}{C_0}$$ (3)

where C_t and C₀ are the dye concentration at treatment time t and initial time 0. The degradation of E132 was determined based on the absorption intensity at 610 nm. The initial E132 solution had a concentration of 20 mg/L, pH of 5.7, and conductivity of 2.1 μS/cm.

2.2. Numerical simulation

The Multiphase-Mixture method available in the fluent module of ANSYS Workbench software (version 19.2.0) was employed to perform numerical simulations of the fluid in the Venturi tube. The Schnerr-Sauer cavitation model was utilized to calculate the phase transition between the water-liquid phase and the vapor phase. The realizable k-ε model, which is commonly used to model turbulence, was employed to treat the axisymmetric viscous flow in this study.

While the k-ε model is not capable of capturing nonstationary fluctuations in the cavitation phenomena, a quasi-steady behavior of the cavitation flow can be obtained after a transient fluctuation of the cavity length. The quasi-steady flow was employed to approximately determine the distributions of density and pressure in the gas–liquid mixture.

The geometric parameters used in the axisymmetric model were identical to those of the actual device, including the electrodes at the axis, as illustrated in Fig. 4. The electrodes at the water inlet and outlet were designated as electrodein and electrodout, respectively. The boundary conditions at the water inlet and outlet were set as the walls of pressure that determined the water flow. If oxygen injection was considered, the oxygen flow was set at the oxygen inlet. The study focused on the hydrodynamic cavitation of the yellow region in Fig. 4, where the discharge gap is located.

Fig. 4.

Cavitating flow region without gas suction.

To determine the distribution of the electric field at breakdown, it was first necessary to obtain the electric field distribution in the gas–liquid mixture at breakdown. For this purpose, a numerical simulation method was used, which involved converting the local relative permittivity (ε_eq) of the mixture to the relative permittivity (ε_l, ε_g) and volume fraction (V_l, V_g) of both the gas and liquid phases using the following formula [45].

$${\epsilon }_{\mathit{eq}}(r,z)=\frac{V_l(r,z)}{V_m(r,z)}{\epsilon }_l+\frac{V_g(r,z)}{V_m(r,z)}{\epsilon }_g=\alpha (r,z){\epsilon }_l+\left(1-\alpha (r,z)\right){\epsilon }_g$$ (4)

where α was defined as the volume fraction of liquid in the mixture. Since the density of gas is much lower than that of liquid, the volume fraction of liquid in the mixture can be expressed as the ratio of the mixture density to the density of the liquid.

$$\alpha =\frac{V_l}{V_m}=\frac{m_l/{\rho }_l}{m_m/{\rho }_m}\approx \frac{{\rho }_m}{{\rho }_l}$$ (5)

The electric field distribution in the gas–liquid mixture prior to breakdown at a specified applied pulse voltage is determined by the distribution of the dielectric constant ε_eq. To determine the electric field distribution, the known ε_eq distribution is used as input into the COMSOL Multiphysics 5.6 software. We employed the static electric currents module in COMSOL Multiphysics to calculate the distribution of electric field in the gas–liquid mixture before breakdown, without considering the influence of discharge channels on the electric field, as these channels had not yet formed at this stage.

3. Result and discussion

3.1. Cavitating flow without O₂ suction

Fig. 3 illustrates the relationship between the Venturi tube's flow rate and suction pressure in response to variations in inlet water pressure under constant outlet pressure and absence of gas suction. The findings indicate that rising inlet water pressure enhances flow rate; however, cavitation-induced flow blockage leads to a saturation trend upon further flow rate increase.

Fig. 3.

Variation of water flow and suction port pressure with water inlet pressure.

The impact of inlet pressure on the formation of cavitating regions in cavitating flow, in the absence of gas suction, is illustrated in Fig. 4 below. The white region in Fig. 4 represents the cavitating region generated at each pressure ratio, with the direction of water flow being from left to right. The findings indicate that the length of the cavitating region increases as the water flow increases. When the water flow rate is as low as 3.1 L/min, the cavitating region disappears entirely.

The cloud cavity's instability leads to random connection or disconnection between electrodes when water flow lies between 7.2 and 8.4 L/min. However, above 9.5 L/min, the gas–liquid two-phase body produced by cloud cavitation stably connects the electrodes. Breakdown characteristics of the two-phase body are significantly influenced by whether cloud cavitation bridges the electrodes, as the primary breakdown occurs through a series of bubbles via the small bridge effect in the discharge gap.

3.2. Discharge characteristic without O₂ suction

In the absence of gas suction, a 50 kV peak pulsed voltage with a 150 ns half-width was applied between two tungsten electrodes. For each constant water flow, 200 independent electrical pulses were administered, and the breakdown events in the two-phase medium were counted. Fig. 5 depicts the breakdown probability as a function of water flow at a maintained 50 kV voltage amplitude. Breakdown initiates at 7.4 L/min water flow, and its probability surges with increasing flow, nearly saturating at 7.8 L/min and reaching 100% at 9.5 L/min.

Fig. 5.

Variation of cavitation number and breakdown probability with water flow (No gas suction, U_source = 50 kV, breakdown probability is a statistical value at 200 single pulses).

Fig. 5 demonstrates that increasing water flow rate led to a gradual decrease in the cavitation number, signifying intensified cavitation. This trend is evident in Fig. 4. When the cavitation number reached 0.1 and could not decrease further, a cavitation cloud occupied the entire discharge gap, resulting in stable discharge breakdown between the two-phase bodies under adequate applied voltage. This observation suggests that sufficient cavitation intensity is necessary for stable breakdown in gas–liquid two-phase bodies within this Venturi tube.

3.3. Generation of ·OH without oxygen suction

The hydroxyl radical (·OH) production rate was assessed by measuring the fluorescence intensity of 2-hydroxyphthalic acid (2-HPA). For each water flow rate, a 5 L solution containing 2 mM TPA was treated for 6 min with a 50 kV pulse and a 1 kHz repetition frequency. Discharge power at varying water flow rates was calculated. As shown in Fig. 5, Fig. 6, increasing water flow rate led to higher breakdown probability, discharge power, and ·OH production rate. However, beyond 9.5 L/min, the breakdown probability reached 100%, and further flow rate increase did not significantly enhance the ·OH production rate, due to the primary generation of hydroxyl radicals through discharge plasma, and the plateauing of discharge power after stable breakdown. At 7.2 L/min, no breakdown occurred, but minor ·OH production resulted from local corona discharge. Therefore, ensuring breakdown discharge is crucial for optimizing ·OH production rate and energy efficiency.

Fig. 6.

Hydroxyl radical yield at different water flow rates.

3.4. Cavitating flow with O₂ suction

3.4.1. Different flow patterns with O₂ injection

As previously shown, attaining a high yield of hydroxyl radicals via in-situ nanosecond pulsed discharge in a Venturi tube necessitates maintaining a substantial water flow rate, ensuring the cavitation cloud spans the electrode gap. Consequently, water flow rate was consistently set at 9.5 L/min in subsequent experiments. With adequate water flow, gas flow below 3 SLPM did not alter the two-phase flow's morphology, maintaining a uniform bubbly flow. However, surpassing 4 SLPM gas flow rate led to a transition to annular flow, as illustrated in Fig. 7(c), causing a distinct separation of gas and liquid phases and deviating from a homogeneous mixture. This flow pattern exhibited a sandwich-like liquid–gas-liquid layer arrangement from the axis to the wall, aligning with the pattern derived from numerical simulations in later sections.

Fig. 7.

Flow patterns under different O₂ flow rates: (a) Plate-like cavitation and bubbly flow under 0 SLPM O₂, (b) Ventilated cavitation and bubbly flow under 0.5 SLPM O₂, (c) Annular flow under 4 SLPM O₂ (water flow rate is 9.5 L/min, and the water flow direction is from bottom to top).

3.4.2. Pressure change with O₂ suction

It was observed that with a gas injection rate below 3 SLPM, the gas–liquid two-phase flow in the throat retained a bubbly flow pattern (Fig. 8). However, the gas injection rate influenced the pressure of the gas–liquid mixture. The results revealed a linear decrease in inlet pressure as the gas injection rate increased while maintaining a water flow rate of 9.5 L/min, as depicted in Fig. 8. In pure gas discharge, increasing gas pressure initially reduces and then raises the breakdown voltage, known as the Paschen curve. Pressure fluctuations in the Venturi tube's gas–liquid mixture may induce similar discharge characteristics.

Fig. 8.

Absolute pressure at the gas inlet under a water flow rate of 9.5 L/min with gas suction.

3.5. Discharge characteristic with O₂ suction

3.5.1. Discharge characteristics under different flow patterns

At a low peak voltage of 20 kV, injecting oxygen can lead to different discharge modes. Fig. 9 displays the various discharge forms under different ventilation rates while keeping the water flow rate constant at 9.5 L/min. When there is no gas injection, the discharge takes on a diffuse form resembling a glow discharge, with a peak current of 5.3 A. However, when the injected gas rate is 0.5 SLPM, the bubbly flow pattern of the two-phase flow persists, but the higher gas pressure reduces the breakdown voltage, leading to spark discharge with noticeable discharge channels and a high peak current of 27.5 A. Spark discharge provides more energy, which enhances the yield of hydroxyl radicals. When the oxygen flow reaches 4 SLPM, the two-phase flow pattern transforms into an annular flow. The high gas pressure in the discharge gap increases the breakdown voltage, resulting in a corona discharge restricted to the inlet electrode head, with a peak current of only 1.4 A.

Fig. 9.

Discharge types and voltage-current waveforms under different oxygen injection rates: (a) no gas injection, bubbly flow generated cavitation with diffuse glow discharge; (b) 0.5 SLPM oxygen injection, bubbly flow with spark discharge exhibiting clear discharge channels; (c) 4 SLPM oxygen injection, annular flow with corona discharge confined to the electrode head at the inlet. (Water flow rate of 9.5 L/min, water flows from bottom to top, power output voltage maintained at 20 kV).

3.5.2. Discharge characteristics in bubbly flow with O₂ suction

Bubbly flow, in comparison to annular flow, exhibits a higher mass transfer efficiency and is more conducive to spark discharge with higher discharge power, as shown in Fig. 9B. To generate sufficient energy required for producing hydroxyl radicals through spark discharge, it is imperative to maximize the discharge voltage. However, even if the power supply's peak output voltage is increased to 50 kV, attaining a breakdown voltage exceeding 45 kV remains challenging. This is because the discharge gap completes the breakdown process before the voltage reaches 50 kV. Upon breakdown, the resistance of the discharge gap drops sharply and the current increases rapidly, causing the voltage of the discharge gap to drop sharply, resulting in the inability to reach the breakdown voltage of the power supply's output voltage.

The injection of oxygen into the mixture not only affects the void fraction but also alters the gas pressure distribution in the discharge gap, both of which influence the breakdown voltage. We defined the positive polarity when the high voltage positive electrode was placed near the water inlet, while negative polarity was defined when the high voltage was applied to the electrode near the water outlet. We carried out 200 individual pulse discharges separately for different oxygen flow rates and analyzed the breakdown voltage statistically. The median of the discharge voltage, U₅₀, was defined as the 50% discharge voltage. The power supply output voltage was kept constant at 50 kV throughout the measurement process. Fig. 10 illustrates the distribution of discharge voltage under 200 independent pulses when 3 SLPM oxygen was injected. It is evident that, even when the power supply output voltage was constant, the breakdown voltage still exhibited significant randomness and followed a normal distribution pattern.

Fig. 10.

U_break distribution of 200 discharge pulses under an oxygen flow rate of 3 SLPM. (Water flow rate of 9.5 L/min, U_source = 50 kV).

Fig. 11 illustrates the distribution of U₅₀ under different oxygen flow rates (ranging from 0 to 3 SLPM) and voltage polarities, and highlights two interesting observations. First, U₅₀ initially decreases and then increases with the increase of oxygen injection rate, reaching a minimum at an oxygen flow rate of 0.3 SLPM, resembling the Paschen curve in gas discharge. As the oxygen flow rate increases, the pressure of the gas–liquid mixture increases linearly, as shown in Fig. 8. Following Paschen's law, at low pressure, there are few particles colliding, and increasing pressure facilitates the avalanche process of discharge. At high pressure, however, there are too many collisions, making the mean free path of molecules too short and thus, reducing the ionization rate. Continuing to increase pressure makes the discharge difficult to develop. Second, U₅₀ exhibits a significant polarity effect, with negative polarity discharge voltage higher than positive polarity. This observation contrasts with the rule observed in gas discharge. The needle-needle electrode used in this study is symmetrically structured in terms of spatial arrangement. In gas breakdown, the uniform distribution of pressure and dielectric eliminates the polarity effect. In this experiment, however, the cavitation effect leads to non-uniform density and pressure distribution of the gas–liquid mixture, where the density and pressure of the mixture at the electrode-in head are significantly lower than those at the electrode-out head. As a result, the distribution of dielectric constant ε is uneven, and the electric field strength at the electrode-in head is higher, leading to easier flow injection and lower discharge voltage for positive polarity. This is the cause of polarity effect in this experiment.

Fig. 11.

U-shaped curve of U₅₀ variation with oxygen injection rate under positive and negative polarities for 200 pulses at a water flow rate of 9.5 L/min (U_source = 50 kV).

The breakdown voltage plays a crucial role in the production of hydroxyl radicals due to two primary effects. Firstly, it directly influences the discharge power, which, in turn, impacts the yield of hydroxyl radicals (as illustrated in Fig. 12). Secondly, the breakdown voltage affects the electron energy distribution of the plasma, which significantly influences the discharge products. It is important to note that different discharge products exhibit different efficiencies in generating hydroxyl radicals, particularly when oxygen is involved (Fig. 13).

$$\begin{array}{cc}{\text{O(}}^3\text{P)}+H_{2}O\to 2·OH& K=2.36\ast {10}^{-12}c{m}^3/\end{array}(mol·s)$$ (6)

$$\begin{array}{cc}{\text{O(}}^1\text{D)}+H_{2}O\to 2·OH& K=2.3\ast {10}^{-10}c{m}^3/\end{array}(mol·s)$$ (7)

Fig. 12.

U-shaped curve of the variation of E_mean with oxygen injection rate under positive and negative polarities for 200 pulses at a water flow rate of 9.5 L/min (U_source = 50 kV).

Fig. 13.

Production of 2-HPA at different polarities and oxygen injection rates (U_source = 50 kV, repetition frequency f = 1 kHz, total water volume of 5 L, water flow rate of 9.5 L/min, treatment time of 6 min).

The production rate of ·OH by O¹D is significantly higher than that by O³P, by two orders of magnitude, as demonstrated in equations (7), (8). However, generating O¹D requires more electron energy, as depicted in equations (10), (11) below.

$$O_2+e(>6.1eV)\to e+O_2\left(A^3\mathrm{\Sigma }{u}^{+}\right)\to e+{\text{O(}}^3\text{P)}+{\text{O(}}^3\text{P)}$$ (8)

$$O_2+e(>8.4eV)\to e+O_2\left({\text{B}}^3\mathrm{\Sigma }{u}^{-}\right)\to e+{\text{O(}}^3\text{P)}+{\text{O(}}^1\text{D)}$$ (9)

To summarize, the generation of hydroxyl radicals is influenced by two main factors: the discharge power and the breakdown voltage. As seen in equations (7), (8), the rate of ·OH production by O¹D is much higher than that by O³P, but producing O¹D requires more electron energy, as shown in equations (8), (9). Thus, increasing the discharge power and breakdown voltage can enhance the generation of hydroxyl radicals by promoting the production of O¹D and providing the necessary electron energy for its formation. Therefore, these factors should be carefully controlled to optimize the production of hydroxyl radicals in the plasma system.

3.6. Generation of ·OH with O₂ suction

Similar to the breakdown voltage, the production of ·OH exhibits a polarity effect. That is, under the same oxygen flow rate, the ·OH production rate for negative polarity is higher than that for positive polarity, due to the higher breakdown voltage increasing the average electron energy (Fig. 13).

When the oxygen flow rate is increased, the production rate of ·OH does not follow the same trend as the change in breakdown voltage, which initially decreases and then increases. Instead, at low oxygen flow rates, the production rate of ·OH remains almost constant (Fig. 13). This is because, at low oxygen flow rates (<0.5 L/min), the oxygen input leads to a decrease in breakdown voltage and reduced electric field strength, resulting in minimal change in the ·OH production rate.

When the oxygen flow rate exceeds 0.5 L/min, the breakdown voltage no longer decreases but instead increases with the increasing oxygen flow rate, and the ·OH production rate also sharply rises (Fig. 13). Tsung-Rong Lin and colleagues employed plasma jet treatment on a 12 ml water sample, yielding a peak hydroxyl radical (·OH) concentration of about 4.5 μM post a 1-minute treatment [46]. However, HCAOP technology outperforms plasma jet methods, delivering larger discharge power and exceptionally high field strength, thus enhancing the ·OH production rate. Under negative polarity, the ·OH production rate at an oxygen flow rate of 3 L/min is 53.2 μM, which is 6.04 times the production rate without gas input.

However, if the oxygen flow rate is further increased to 4 L/min, the production rate of ·OH experiences a steep decline (Fig. 13). At this point, the gas–liquid mixture has transitioned from bubble flow to stratified annular flow. Under the annular flow regime, the internal gas pressure of the flow is high, and both electrode surfaces are covered by a water film, making it difficult for the discharge gap to be stably broken down, resulting in corona discharge most of the time. Interestingly, under the annular flow pattern, the ·OH production rate for positive polarity is higher than that for negative polarity. This is precisely because positive polarity has a lower breakdown voltage distribution, increasing its probability of breakdown.

3.7. Plasma emission spectrum

We focused our analysis on the emission spectra between 425 and 870 nm. Among the various spectral lines observed, two were deemed most significant, namely the Hα line at 656 nm and the excited state oxygen atom emission line at 777 nm. The former is a result of hydrogen atom dissociation from water molecules, while the latter originates from the emission of excited state oxygen atoms. The intensity of the spectral line at 656 nm serves as a measure of the degree of dissociation of water vapor in the system.

$$H_{2}O+e\to ·H+·OH\left(A^2{\mathrm{\Sigma }}^{+}\right)+e$$ (10)

The experimental results in Fig. 14 provide insights into the reaction pathways leading to the production of hydrogen atoms and ·OH radicals in the plasma system. As illustrated in Equation (11), the electron collision dissociation of water molecules leads to the production of hydrogen atoms and ·OH radicals, with one free electron producing one ·OH. On the other hand, excited state oxygen atoms generated through the reaction of oxygen molecules with free electrons can also react with water molecules to produce ·OH radicals, as shown in Equation (8) in the preceding section. Importantly, one free electron can produce two ·OH via this pathway, thereby increasing the overall ·OH yield.

Fig. 14.

Spectrum of the plasma with oxygen injection (a) 0 SLPM O₂ (b) 0 SLPM O₂ (c) 0 SLPM O₂ (wavelength range 425–870 nm, exposure time 100 ms).

Fig. 14 presents the emission spectra of discharge plasma under different oxygen flow rates. Notably, when no oxygen was introduced into the system, both the Hα and excited state oxygen atom emission lines were detected (Fig. 14.a). This indicates that the sources of hydrogen and oxygen atoms were both the dissociation of water molecules via electron collision dissociation. As the oxygen flow rate increased from 0.5 SLPM to 3 SLPM, the emission line intensity at 656 nm (corresponding to the Hα line) decreased, while that at 777 nm (corresponding to the excited state oxygen atom emission line) increased sharply. These observations suggest that the plasma system underwent less water dissociation and more oxygen dissociation excitation. This finding helps explain the previously observed increase in ·OH yield by 6 times when the oxygen flow rate increased from 0.5 SLPM to 3 SLPM, while the discharge power increased by only 1.2 times (as discussed in Section 3.7).

3.8. Effect of initial pH and conductivity

Fig. 15 illustrates the effects of varying initial pH values on the concentrations of TRO, H₂O₂, and ·OH, with the pH range in the solution spanning from 4 to 9. As the pH value increases, the TRO concentration gradually decreases from 19.1 mg/L to 8.34 mg/L·H₂O₂ is undetectable at pH = 4, but its terminal concentration increases to 274.3 μM as the pH continues to rise. The accumulated concentration of ·OH increases from 24.8 μM to 59.7 μM with the increase in pH. These results suggest that under alkaline conditions, this process can more rapidly consume TRO and generate a higher amount of ·OH.

Fig. 15.

Production of ∙OH, TRO and H₂O₂ at different pH (50 kV negative output voltage, 9.5 L/min water flow, 3 SLPM O₂, treatment time of 6 min).

This can be attributed to the fact that under high pH conditions, OH– acts as the initiator for the decomposition of O₃ [47]. According to the SBH mechanism, chain initiation reactions involve the transfer of ·O⁺ from O₃ to form ·O₂– by interacting with OH^–. As the rate of ·OH production is determined by the chain initiation reaction, which serves as the controlling step, higher concentrations of ·OH are observed under alkaline conditions [48]. Simultaneously, a high concentration of ·OH can react with O₃ to generate ·O₂^– or the protonated form of ·O, the perhydroxyl radical ·OH₂. Both ·OH₂ and ·O₂^– can further combine with O₃ to produce more free radicals. Therefore, under high pH conditions, the concentrations of ·OH are higher while those of TRO and H₂O₂ are lower [36].

Fig. 16 presents the variation of ·OH accumulation concentration as the conductivity increases from 0.52 S/m to 6.76 S/m. When the solution conductivity is no higher than 1.43 S/m, the accumulated concentration of ·OH is essentially unaffected by the conductivity, remaining approximately 26.17 μM. As the conductivity continues to increase, the accumulated concentration of ·OH sharply declines to 4.21 μM at a conductivity of 2.37 S/m. With further increases in conductivity, the accumulated concentration of ·OH gradually decreases until it ceases to form when the conductivity approaches 7 S/m. This trend can be attributed to the change in the discharge characteristics of the two-phase system at high conductivity. When the conductivity exceeds 7 S/m, the voltage is unable to load across the discharge gap, and discharge does not occur.

Fig. 16.

Production of ∙OH at different conductivity of water (50 kV negative output voltage, 9.5 L/min water flow, 3 SLPM O₂, treatment time of 6 min).

Fig. 17 depicts the relationship between the power consumed in the discharge gap and the conductivity. When the conductivity is no higher than 2.37 S/m, the power consumed in the discharge gap dramatically increases with rising conductivity, from 31.2 W to approximately three times that value at 95.3 W, due to the increased power consumption caused by leakage conductivity. However, as the conductivity continues to rise, the power consumed in the gap starts to decrease, as the gap voltage drops and the discharge becomes increasingly weak.

Fig. 17.

Discharge power at different conductivity of water (50 kV negative output voltage, 9.5 L/min water flow, 3 SLPM O₂).

The aforementioned study demonstrates that the hydrodynamic cavitation gas–liquid two-phase system with nanosecond pulsed discharge advanced oxidation process can treat solutions with a certain conductivity (<1.42 S/m). If the solution conductivity is too high, both the energy efficiency and ·OH production rate of this process will decline drastically.

3.9. Evaluation on dye degradation efficiency

In the previous sections, we have determined the optimal operating parameters for the production of hydroxyl radicals in the venturi tube through the combination of hydrodynamic cavitation and nanosecond pulsed discharge. These parameters include a sufficient water flow rate of 9.5 L/min, a larger oxygen flow rate of 3 SLPM under bubbly flow pattern, and negative polarity voltage. The effectiveness of this technique was evaluated by treating 5 L of E132 solution with a concentration of 20 mg/L under these parameters, and the absorption spectra of the solution was recorded (Fig. 18). By observing the change in the absorption spectrum at 610 nm, we observed that the device was able to efficiently degrade 100 mg of E132 in 5 L of water within 2 min. This result confirms the rapid and efficient nature of HCAOP as a technique for the treatment of organic dye-contaminated water.

Fig. 18.

UV–Vis absorption spectra of E132 degradation in HCAOP process for a 160 s treatment (20 mg/L initial concentration, 50 kV negative output voltage, 9.5 L/min water flow, 3SLPM O₂).

In the molecular structure of Indigo Disulfonate Sodium (IDS), the central C C double bond exhibits high chemical reactivity. When IDS is exposed to a discharge plasma environment, the plasma generates reactive species such as ozone (O₃) and hydroxyl radicals (·OH) [49]. These reactive species possess high oxidizing properties, enabling them to react with the C C double bond in IDS molecules. During this process, the molecular structure of IDS is disrupted and ultimately transformed into a colorless product, Indirubin Disulfonate Sodium (IRS). This conversion process can be attributed to the redox reactions between the plasma-generated reactive species, such as ozone and the C C double bond of IDS, which are accompanied by the destruction and reorganization of the IDS molecular structure [50]. This chemical reaction process is illustrated in Fig. 19.

Fig. 19.

Decolorization of sodium indigo disulfonate.

The degradation curves of E132 under HCAOP process with different oxygen flow rates are illustrated in Fig. 20. The figure clearly shows that the degradation process of E132 follows a typical first-order reaction [23], [51].

$$\ln \left(\frac{C_0}{C_t}\right)=kt$$ (11)

$${\text{E}}_{\text{E0}}=\frac{Pt\text{1000}}{V\text{lg}(c_i/c_f)}$$ (12)

Fig. 20.

Degradation efficiency η of E132 in HCAOP process for a 6 min treatment with different O₂ flow (20 mg/L initial concentration, 50 kV negative output voltage, 9.5 L/min water flow).

E_EO is the amount of electrical energy, measured in kWh, needed to degrade a contaminant C by one order of magnitude in 1 m3 of contaminated water [52].

The oxidant or energy dosage required for advanced oxidation processes (AOPs) is dependent on the volume of water being treated and the desired reduction in contaminant concentration. Therefore, a figure of merit, known as the Electrical Energy per Order (E_EO), has been defined for electrical-driven AOPs [53]. We calculated the rate constants and E_EO under different oxygen flow rates, as shown in Table 1 below. Table 1. Degradation rate constant k of E132 with HCAOP process.

Table 1.

Degradation rate constant k of E132 with HCAOP process.

O2 flow (SLPM)	0	0.5	1	2	3
Degradation rate constant k (min−1)	0.186	0.205	0.60	0.82	1.81
EEO(kWh/m3/order)	2.33	1.99	0.71	0.56	0.26

3.10. Numerical simulation

3.10.1. Hydraulic characteristics

Fig. 21 demonstrates a consistency with the observations from the experiments in Fig. 7, indicating that a small amount of oxygen input (Q_gas = 0.5 L/min) does not alter the flow pattern of the two-phase medium. In Fig. 21a, the water is injected into the throat through the inlet, and due to intense cavitation, the density of the two-phase medium within the entire throat is considerably low. A high-density region is formed only along the central axis of the throat, resulting from the obstruction of the downstream outlet electrode.

Fig. 21.

Density distribution in the Venturi tube at different oxygen flow rates obtained through numerical simulation (a) No gas injection, (b) 0.5 SLPM oxygen, (c) 3 SLPM oxygen, (d) 4 SLPM oxygen (Water flow rate is kept constant at 9.5 L/min).

The high-density region is located at the central axis of the throat, surrounded by a two-phase medium containing a large number of gas bubbles. Therefore, in the experiments shown in Fig. 7, the high-density region along the central axis of the two-phase throat could not be captured, and only a uniform bubble flow in the periphery was observed. Moreover, the distribution of the high-density region is axially asymmetric, exhibiting a general trend of low density near the inlet electrode and high density near the outlet electrode.

This phenomenon is attributed to the fact that the inlet electrode head serves as the backflow surface for the two-phase medium, characterized by low pressure and intensified cavitation. In contrast, the outlet electrode head acts as the flow-facing surface of the two-phase medium, where a rapid pressure increase at the drainage surface head causes the vapor generated by cavitation to quickly re-condense, returning to a high-density state. Such density differences lead to the asymmetry of the electric field distribution in the two-phase medium. Additionally, since the electric field distribution is inversely proportional to the dielectric constant distribution, the strong field region should be located in the low-density area near the inlet head.

3.10.2. Electric field distribution

Fig. 22a-c illustrates that under the bubble flow regime (Q_gas = 0, 0.5, 3 L/min), the asymmetric distribution of high-density regions in the discharge gap (as shown in Fig. 21) leads to the formation of a high-field-strength region at the head of the inlet electrode, resulting in a highly asymmetric electric field distribution. This explains why the 50% breakdown voltage (U₅₀) in Fig. 11 is significantly higher under negative polarity compared to positive polarity. When the flow pattern transitions to the stratified annular flow (Fig. 22d), the high-field-strength region surrounds the inlet electrode head, which is consistent with the observation of weak corona discharge at the head of the inlet electrode in Fig. 9A.

Fig. 22.

Electric field distribution of discharge gap under different oxygen flow rates at the breakdown voltage (a) No gas injection, (b) 0.5 SLPM oxygen, (c) 3 SLPM oxygen, (d) 4 SLPM oxygen (Water flow rate is kept constant at 9.5 L/min).

4. Conclusions

Hydrodynamic cavitation, by modulating the density and pressure of the gas–liquid mixture, governs the discharge characteristics in HCAOP process, substantially influencing the ·OH production rate and energy efficiency. To achieve optimal ·OH production rates and energy efficiency, the following conditions must be met:

1. Sufficient cavitation intensity for stable bridging of gas–liquid mixtures across discharge gaps, resulting in reliable spark breakdown;

2. Appropriate oxygen flow rate, as too low of a rate may reduce breakdown field strength and lead to lower electron energy, while too high of a rate can cause the flow regime to transition into stratified annular flow;

3. Negative polarity external voltage is more favorable than positive polarity, due to the non-symmetrical density and pressure distribution of the two-phase mixture, with higher breakdown field strength under negative polarity conditions.

These principles apply to analogous systems of varying sizes. The introduction of an appropriate amount of oxygen was found to increase the ·OH energy efficiency by sixfold to 119.8 nmol/J. This improvement is attributed to the change in gas-phase composition and reaction system caused by the oxygen, which leads to an increase in oxygen atom concentration. Oxygen atoms can directly react with water to produce hydroxyl radicals, and the ozone generated can rapidly form ·OH through the Peroxone reaction with H₂O₂ produced in the HCAOP process.

CRediT authorship contribution statement

Qiong Wu: Methodology, Investigation, Visualization, Writing – original draft. Haiyun Luo: Conceptualization, Funding acquisition, Supervision. Hao Wang: Investigation, Formal analysis, Validation, Methodology. Zhigang Liu: Investigation, Formal analysis, Validation. Liyang Zhang: Investigation, Formal analysis, Validation. Yutai Li: Investigation, Formal analysis, Validation. Xiaobing Zou: Funding acquisition, Conceptualization, Supervision. Xinxin Wang: Funding acquisition, Conceptualization, Supervision.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.