A Novel Multi-Needle-to-Cylinder Dielectric Barrier Discharge Reactor with Deflector Rings for Energy-Efficient Removal of Sulfides and Ammonia from Odor Gases

Abstract

Non-thermal plasma is a promising technology for odor abatement from agricultural and domestic waste. However, its widespread application is often limited by the inherent trade-off between energy efficiency and processing capacity in conventional reactors. To address this challenge, we propose a novel multi-needle-to-cylinder dielectric barrier discharge reactor integrated with a deflector ring. By synergistically optimizing the electrode topology and modulating the flow field, this reactor achieves enhanced removal of complex ammonia–sulfur odor mixtures. The underlying mechanisms were elucidated through computational fluid dynamics (CFD) simulations coupled with systematic performance evaluation. Experimental results demonstrate that an 8-needle electrode configuration provides the optimal balance between discharge density and energy efficiency. CFD simulations further reveal that the deflector ring effectively suppresses gas bypass and promotes recirculation vortices downstream, thereby extending the residence time significantly. Mechanistic studies indicate that the removal of recalcitrant inorganic sulfides (e.g., CS₂ and H₂S) follows a synergistic mass-transfer–reaction controlled process, which is markedly improved by flow field optimization. In contrast, organic sulfides are governed primarily by chemical kinetics and show little dependence on flow variations. Under an extremely short residence time of 0.57 s (corresponding to a flow rate of 2.0 m³/h) and an ultra-low specific energy input of 6.26 J/L, the system achieved nearly complete removal of organic sulfides. Even for challenging inorganic sulfides, removal efficiencies reached 80.9% for H₂S and 45.3% for CS₂, while O₃/NO_x/SO₂ byproducts were quantified. For industrial deployment, these byproducts can be managed by standard downstream polishing. By effectively coordinating discharge characteristics with flow dynamics, this study provides both theoretical insight and technical support for the development of next-generation, energy-efficient, high-throughput industrial odor control systems.

1. Introduction

Odor control in agricultural and domestic waste management has become a pressing environmental issue [1]. Beyond causing severe olfactory discomfort, such emissions intensify public resistance through the “Not In My Backyard” (NIMBY) effect and undercut the prospects for sustainable industry growth [2,3]. The gas streams typically contain a complex mix of ammonia [4], sulfides that cause strong olfactory detectability even at low concentrations [5], volatile organic compounds (VOCs) [6], and terpenes [7]. Among these, ammonia and sulfides are the most frequent pollutants, with concentrations ranging from tens of ppb to hundreds of ppm [8], making their removal a priority. While conventional methods like biofiltration [9] and chemical scrubbing [10] are widely used, they face inherent limitations. Chemical scrubbers typically offer a compact footprint; however, similar to biological systems, their efficiency is heavily constrained by the water solubility of the target pollutants, limiting mass transfer and conversion kinetics for hydrophobic compounds. Consequently, there is an urgent industrial demand for compact, agile, and energy-efficient alternatives.

Non-thermal plasma (NTP) technology stands out for its rapid reaction kinetics and compact design [11], making it ideal for treating high-volume, low-concentration volatile organic compounds (VOCs) [12]. However, a persistent trade-off limits its wider application: the conflict between energy efficiency and gas throughput. Unlike general VOC abatement, effective odor control requires near-total removal efficiency, which invariably demands high energy consumption [13]. For example, while wire-plate pulse corona plasma can remove 93.1% of ethyl mercaptan and 100% of ammonia, it requires a specific input energy (SIE) of 65.1 J/L [14]. For mixed H₂S and NH₃, planar DBD systems typically require ~202 J/L to achieve ~67% removal [15]. Similarly, double dielectric barrier discharge (DBD) systems may achieve high removal rates for carbon disulfide (86.32–100%) but often operate at an excessive SIE of 960–3222 J/L [16]. Developing plasma technologies that balance high efficiency with low energy consumption remains a pressing goal.

To address this challenge, improving the controllability of barrier discharges has become a focal point of recent research. For instance, discharge homogeneity and parameters can be effectively tuned by tailoring applied voltage waveforms, or by manipulating the dielectric microstructure (e.g., via porous dielectrics) to modify local electric field distributions. Complementing these electrical and material-based approaches, optimizing reactor geometry remains a proven pathway to enhance performance. Specific electrode topologies, such as multi-needle-to-plate [17] or coaxial arrays [18], can stabilize discharges at lower voltages to save energy. Furthermore, modulating the internal flow field—for instance, through flow-field induced surface DBD [19]—can significantly improve treatment efficacy. We propose that synergistically combining electrode topology with flow field modulation is the key to efficient, low-energy pollutant removal. Our design strategy focuses on creating inhomogeneous electric fields using multi-needle arrays to induce dense micro-discharges, while actively reconstructing the gas flow path. This forces pollutants into active discharge zones, minimizing the bypass effect.

Here, we introduce a novel Multi-Needle-to-Cylinder DBD reactor equipped with Flow Deflector Rings (DR-MCDBD). This design integrates a multi-needle array to maximize micro-discharge density and uses diversion rings to mechanically redirect the flow, enhancing gas-plasma interaction. This study aims to: (1) systematically optimize the electrode configuration and discharge characteristics for high-flow conditions; (2) evaluate the system’s removal performance and energy efficiency for complex sulfide-ammonia mixtures; and (3) elucidate how diversion rings enhance the degradation of different pollutants by combining Computational Fluid Dynamics (CFD) simulations with reaction kinetics. Ultimately, this work seeks to provide a theoretical foundation and technical route for the next generation of industrial odor control reactors.

2. Materials and Methods

2.1. Experimental System

The experimental system primarily consisted of a gas supply unit (air and target gas blending), a custom-made low-temperature plasma reactor, a high-frequency high-voltage AC power supply, and a gas detection and analysis system, as shown in Figure 1a. The total gas flow rate was controlled at 1, 1.5, and 2 m³/h, respectively. Air was used as the carrier gas, and the target pollutants from gas cylinders were diluted to the following concentrations: 5 ppm H₂S, 2.5 ppm CH₃SH, 10 ppm DMS, 1.25 ppm CS₂, 2.5 ppm DMDS, and 30 ppm NH₃.

The DR-MCDBD reactor was employed for the removal of these malodorous gases. The peak-to-peak voltage ranged from 7 kV to 9 kV with a discharge frequency of 20 kHz. The voltage was regulated by a single-phase AC voltage regulator (Chint TDGC, Wenzhou, China) and measured using an oscilloscope (Tektronix MDO3024, Shanghai, China). Discharge images were captured using a digital camera (Sony ILCE-7M4, Shanghai, China), and the reactor’s temperature distribution was monitored using a handheld infrared thermal imager (InfiRay M620, Yantai, China).

The average discharge power was determined using the Q-V Lissajous figure method. The SIE and Energy Efficiency (EE) were calculated using Equations (1) and (2) [15,20], where P denotes average discharge power, Q is the gas flow rate, ${\eta }_i$ is the removal efficiency, $C_{inlet,i}$ is the inlet concentration, and $M_i$ is the molecular weight of the pollutant. The coefficient 3.6 represents the conversion factor between Joules and kilowatt-hours. For industrial feasibility assessment, the SIE value can be converted to the standard industrial unit using the relationship: 3.6 J/L $=$ 1 kWh/1000 m³.

$$\mathrm{S}IE\left({\displaystyle \frac{\mathrm{J}}{\mathrm{L}}}\right)={\displaystyle \frac{3600\left(\frac{\mathrm{s}}{\mathrm{h}}\right)\times P\left(\mathrm{W}\right)}{1000(\mathrm{L}/{\mathrm{m}}^3)\times Q({\mathrm{m}}^3/\mathrm{h})}}$$

(1)

$$\mathrm{E}\mathrm{E}\left({\displaystyle \frac{\mathrm{g}}{\mathrm{k}\mathrm{W}\mathrm{h}}}\right)={\displaystyle \frac{{\eta }_i\times C_{inlet,i}\left(\mathrm{p}\mathrm{p}\mathrm{m}\right)\times M_i(\mathrm{g}/\mathrm{m}\mathrm{o}\mathrm{l})}{SIE(\mathrm{J}/\mathrm{L})}}\times {\displaystyle \frac{3.6}{22.4(\mathrm{L}/\mathrm{m}\mathrm{o}\mathrm{l})}}$$

(2)

2.2. Reactor Configuration

To address the requirements for treating high-flow, low-concentration malodorous gases, this study designed and constructed an innovative DR-MCDBD reactor based on traditional dielectric barrier discharge (DBD) technology. The reactor aims to achieve high removal efficiency with low energy consumption by optimizing the geometric ratio between the discharge gap and the gas channel, as well as by regulating the internal flow field distribution.

The reactor features a single-dielectric coaxial configuration, comprising a quartz glass barrier tube, an external mesh high-voltage electrode, and an internal grounded electrode. A quartz tube (40 mm outer diameter (O.D.), 3 mm wall thickness, 552 mm length) was selected as the dielectric barrier. The high-voltage electrode consists of 500-mesh 304 stainless steel mesh wrapped tightly around the outer wall of the quartz tube, with an effective discharge length of 400 mm. The mesh was mechanically fixed to maintain intimate contact with the quartz surface (i.e., no intentional air gap between the mesh and the dielectric); the minimum clearance from the powered mesh to any grounded metallic supports/frame was ≥60 mm to suppress unintended external discharges. This dense mesh structure helps provide a uniform external potential boundary and facilitates real-time observation of the plasma state.

The internal electrode (grounded electrode) is primarily a 12 mm diameter aluminum rod, along which multiple groups (6, 8, or 10 groups) of discharge pin assemblies are arranged axially. Each assembly comprises a silicone fixing sleeve (12 mm inner diameter (I.D.), 20 mm length, 2 mm thickness) and four 304 stainless steel discharge pins (1 mm diameter, 10 mm length). The discharge pins are distributed in pairs (180° symmetry), with adjacent pin groups staggered orthogonally (90° offset), as shown in Figure 1b and illustrated in the cross-sectional view in Figure 1c. This design creates a wide annular gas channel (11 mm width)

Between the inner electrode and the dielectric barrier, effectively reducing flow resistance; at a flow rate of 2.0 m³/h, the gas residence time is approximately 0.57 s. Simultaneously, a 1 mm micro-discharge gap is formed between the pin tips and the inner wall of the quartz tube (Figure 1c), providing the physical basis for inducing stable fast-pulse micro-discharges at relatively lower voltages. Furthermore, to optimize the flow field distribution within the tube, a polylactic acid (PLA) flow deflector ring (5 mm base thickness, 26 mm diameter) is installed on the upstream side of each pin assembly.

2.3. Analytical System

The inlet and outlet concentrations of sulfides were determined using a gas pre-concentrator (Acrichi APC-9LN, Beijing, China) coupled with a gas chromatograph (Shimadzu Nexis GC-2030, Shanghai, Japan) equipped with a Flame Photometric Detector (FPD). The FPD utilized an SH-1 capillary column (60 m × 0.32 mm × 3 μm) [21]. Ammonia concentrations were measured using a portable multi-gas detector (MultiRAE PGM-82 × 8, Shanghai, China) [22]. Ozone concentration was monitored using a handheld ozone detector (Aeroqual Series 200, Shanghai, China) [23], while nitrogen oxides (NO_x) and sulfur dioxide (SO₂) were analyzed using a portable infrared flue gas analyzer (MRU MGA6 plus, Shanghai, China). Additionally, to evaluate the thermal characteristics of the reaction system, a high-precision infrared thermal imager (InfiRay M620, Yantai, China) was used to monitor the thermal stability and surface temperature distribution of the reactor in real-time. The removal efficiency (${\eta }_i$) for sulfides and ammonia was calculated according to Equation (3), where $C_{inlet,i}$ and $C_{outlet,i}$ represent the inlet and outlet concentrations, respectively.

$${\eta }_i={\displaystyle \frac{C_{inlet,i}\left(\mathrm{p}\mathrm{p}\mathrm{m}\right)-C_{outlet,i}\left(\mathrm{p}\mathrm{p}\mathrm{m}\right)}{C_{inlet,i}\left(\mathrm{p}\mathrm{p}\mathrm{m}\right)}}\times 100\%$$

(3)

2.4. Numerical Simulation Setup

To elucidate the influence of the deflector ring (DR) on the reactor cold-flow field (bypass suppression and recirculation enhancement), 3D CFD simulations were performed in ANSYS Fluent (2022 R1). The flow was modeled as steady, incompressible, single-phase air at 25 °C ($\rho$ = 1.2 kg/m³, $\mu$ = 1.835 × 10⁻⁵ Pa·s). Turbulence was closed using the SST k-ω model. Plasma-induced body forces, gas heating, and species transport/reactions were not considered.

A mass-flow inlet corresponding to 2.0 m³/h (equivalent bulk inlet velocity U_in ≈ 0.70 m/s), a pressure outlet (0 Pa gauge), and no-slip walls were applied. Unless otherwise stated, the CFD results reported in this work correspond to 2.0 m³/h, as this operating condition is associated with the target performance. A grid-independence test was conducted at $U_{in}$ ≈ 0.35 m/s (=1.0 m³/h, within the experimental range), and the mesh with 1,070,886 cells was adopted for the simulations reported in this study. Detailed descriptions of the governing equations, inlet turbulence parameters, and the grid independence test are provided in the Supplementary Materials (Section S1).

3. Results and Discussion

3.1. Discharge Characteristics

3.1.1. Discharge Morphology and Electrical Waveforms

As illustrated in Figure 2a,b, the discharge within the reactor exhibits significant spatial inhomogeneity. As illustrated in Figure 2a,b, the discharge within the reactor exhibits significant spatial inhomogeneity. Owing to the microscopic radius of curvature at the needle tips, the electric field is strongly enhanced and confined to the vicinity of the grounded pin electrodes that are directly exposed to the process gas. In contrast, the high-voltage mesh electrode is located outside the quartz dielectric and is not in contact with the gas. Consequently, the plasma is composed primarily of dense luminous spots anchored at the needle-tip regions, while no macroscopic discharge is observed on the inner electrode surface away from the pin tips. These luminous spots correspond to numerous discrete filamentary micro-discharges (streamer-like events), resulting in a filamentary, spot-like macroscopic emission pattern.

This physical characteristic is further corroborated by the electrical waveforms. Since the power supply operates in Discontinuous Conduction Mode (DCM) with inductive compensation, both the applied voltage and the displacement current deviate from standard sinusoidal forms [24]. Superimposed on the current waveform are dense nanosecond-scale discharge spikes, confirming that the reactor operates in a typical filamentary (streamer-like) DBD mode. Notably, the current exhibits a significant polarity effect: the spikes are higher in amplitude and denser during the negative half-cycle (Figure 2a), which is typical for asymmetric single-dielectric DBD configurations and reflects polarity-dependent discharge development in the pin–dielectric micro-gap.

3.1.2. Power Consumption Characteristics

Figure 3a illustrates the interactive effects of the number of needle electrode sets (6, 8, and 10 sets) and the applied voltage on the average discharge power. Generally, the average discharge power exhibits a significant non-linear positive correlation with the applied voltage. As the voltage increases from 7.0 to 9.0 kV, the intensification of the inter-electrode electric field enhances both the frequency and intensity of micro-discharges, thereby improving the energy injection efficiency.

At a fixed voltage level, a distinct step-wise increase in average discharge power is observed with an increasing number of needle sets. Specifically, at the peak voltage of 9.0 kV, the average power consumption for the 6, 8, and 10-needle configurations reaches 2.9 W, 3.9 W, and 5.1 W, respectively. This enhancement is primarily attributed to the increase in effective discharge sites, which induces a greater number of discharge channels, thereby elevating the total transport current and power deposition. Notably, the power curves exhibit a distinct inflection point in the 7.5–8.0 kV range, where the growth slope is maximized. This surge signifies the reactor’s transition from an initial localized corona to a fully developed filamentary discharge regime. As the voltage rises further (8.0–9.0 kV), the average discharge power demonstrates a steady quasi-linear increase. In this region, the gas gap is in a stable breakdown state, the linear dependence of power on voltage observed here aligns well with the classical theory of ozonator discharge proposed by Manley [25]. This indicates that the power enhancement is governed by the theoretical scaling law, driven primarily by the increase in transferred charge per discharge cycle rather than an explosive expansion of the discharge area.

Figure 3b further investigates the influence of gas flow rate (0.5–2.5 m³/h) on the average discharge power of the 8-needle reactor. The results indicate that the applied voltage is the dominant governing factor for power consumption, while the effect of gas flow rate is negligible. Even with a five-fold increase in flow rate, the average discharge power remains relatively constant at a fixed voltage (e.g., 8.0 kV), showing only marginal fluctuations. This is because the development of micro-discharges occurs on the nanosecond (ns) scale, which is orders of magnitude faster than the millisecond (ms) scale residence time of the gas flow. Consequently, aerodynamic disturbances are insufficient to alter the charge accumulation and breakdown mechanisms within the discharge gap. This flow-independent characteristic demonstrates the engineering robustness of the system, providing strong support for subsequent high-flow industrial applications.

3.2. Optimization of Electrode Configuration

To determine the optimal reactor configuration capable of sustaining high-throughput conditions, the degradation characteristics were systematically evaluated under arrangements of 6, 8, and 10 needle sets. Initially, the flow rate adaptability of the 6- and 8-set arrays was examined within the range of 1.0–2.0 m³/h (Figure 4 and Figure 5). The results indicate that the 6-set configuration exhibits pronounced sensitivity to flow rate variations. As the flow rate increased, the removal efficiency of inorganic sulfides decayed significantly without exhibiting a stable plateau: the removal rates for H₂S and CS₂ plummeted from peaks of 82.4% and 57.2% to 39.5% and 38.9%, respectively. Conversely, the decline for organic sulfides was minimal (<2%).

This discrepancy is attributed to the differences in degradation kinetics between the pollutants. For inorganic sulfides, the inherently sparse discharge distribution of the 6-set configuration failed to provide sufficient contact time for active species under high gas velocities (shortened residence time). In contrast, organic sulfides exhibit much faster reaction rates due to their rapid oxidation mechanisms, rendering them less sensitive to variations in residence time. Detailed reaction pathways and kinetic mechanisms for these species will be thoroughly discussed in Section 3.7.

Comparatively, the 8-set array demonstrated superior flow rate adaptability (Figure 5). Although the overall removal rate of CS₂ decreased (from 68.2% to 48.2%, a drop of approximately 20 percentage points), the efficiency stabilized within the industrially relevant range of 1.5–2.0 m³/h, with the decline controlled within 6%, forming a distinct “performance plateau.” Meanwhile, H₂S maintained a high removal level (86.0% to 81.6%). This confirms that the denser discharge network effectively increased the density of the active zone, compensating for the limitations imposed by the shortened residence time at high flow rates.

To further optimize the electrode configuration, a lateral comparison was conducted under typical conditions of a 2.0 m³/h gas flow rate and an applied voltage of 8.5 kV (note that 9.0 kV was excluded because it yielded <4% efficiency gain but caused >13% increase in SIE) (Figure 6). Increasing the number of needle sets from 6 to 8 resulted in a substantial improvement in system performance, with H₂S removal rising from 37.9% to 80.9%. This improvement is attributed to the expansion of the effective plasma-active region, which increases the probability of interactions between pollutants and energetic electrons/reactive species.

However, a further increase from 8 to 10 sets produced only marginal gains (<3% for all pollutants), indicating diminishing marginal returns despite the higher power input. This saturation is likely governed by a combination of kinetic and transport-related limitations rather than discharge suppression. For readily oxidized pollutants such as H₂S, once high removal is achieved in the upstream stages, the residual concentration in the downstream region becomes low; consequently, the additional stages (sets 9–10) contribute only limited incremental removal (reactant-depletion-driven kinetic limitation). For more recalcitrant species such as CS₂, adding more discharge stages does not necessarily translate into proportional removal enhancement under high-flow conditions, because the overall performance can become constrained by non-ideal gas–plasma contacting and transport/hydrodynamic effects associated with a denser internal structure (e.g., reduced axial spacing between adjacent internal elements), which may intensify local flow disturbance/wake interaction and reduce the effective residence time in the active discharge zones. The flow-rate-dependent decomposition characteristics of the 10-needle configuration are provided in the Supplementary Materials (Section S2, Figure S3). Notably, spatially resolved measurements of active species (e.g., OH) and axial concentration profiles were not performed in this work; therefore, the above mechanisms are proposed as plausible interpretations based on the observed performance trends.

Consequently, the 8-set configuration, which strikes a balance between high-throughput processing capacity and energy efficiency (SIE maintained at 1.94–7.04 J/L), was selected as the optimal design for this system.

3.3. Energy Efficiency Analysis

The energy efficiency (EE, g/kWh) of the optimized 8-group needle configuration was characterized under varying voltages and flow rates (Figure 7). Across all flow rates, EE for ammonia and sulfides exhibited a non-monotonic trend, consistently peaking at 7.5 kV. Below this threshold, discharge intensity was insufficient for rapid degradation. Conversely, voltages exceeding 7.5 kV caused a sharp EE decline (e.g., NH₃ EE dropped from ~35 to ~12 g/kWh at 9.0 kV). This drop occurs because power consumption rises quadratically while removal rates exhibit diminishing returns, identifying 7.5 kV as the point of maximum energy economy.

Additionally, higher flow rates progressively enhanced EE, with peak NH₃ efficiency rising from ~19 to ~35 g/kWh. This is attributed to improved active species utilization; low flow rates result in excessive residence time, leading to non-productive quenching or byproduct over-oxidation. In contrast, higher flows ensure pollutant molecules effectively capture the generated plasma energy.

Regarding pollutant specificity, NH₃ and DMS achieved significantly higher EE (>30 and >20 g/kWh) compared to inorganic sulfides (<5 g/kWh). This aligns with the lower bond dissociation energies and faster reaction pathways of amino and methyl groups compared to recalcitrant inorganic structures.

Ultimately, parameter selection requires balancing economy and efficacy. Although 7.5 kV offers peak energy efficiency, removal efficiency continued to rise significantly up to 8.5 kV. Prioritizing regulatory discharge standards, 8.5 kV is recommended as the optimal operating voltage to ensure robust pollutant abatement with acceptable energy consumption.

3.4. Effect of Flow Deflector Ring (DR)

3.4.1. Hydrodynamic Simulation Analysis

Figure 8a illustrates the axial velocity distribution of the cold flow field without the Deflector Ring (DR) at a flow rate of 2 m³/h. The flow field exhibits a relatively smooth distribution of streamlines. Due to the cross-sectional constriction caused by the 2-mm-thick needle holder, the maximum axial velocity reaches 1.19 m/s near the needle tips, which is approximately 1.7 times the inlet velocity. However, a low-velocity bypass region is observed near the root of the discharge needles and the inner electrode surface. Fluid passing through this region fails to penetrate the core discharge zone, potentially limiting the interaction between the gas and active species.

The flow field with the inclusion of the DR is presented in Figure 8b. The DR significantly accelerates the flow, with the maximum axial velocity peaking at 2.28 m/s at the throat of the guide plate (3.26 times the inlet velocity). The DR modifies the hydrodynamics in two beneficial ways: First, the gas forms a high-speed flow layer downstream of the throat. Given that the intense electric field is primarily concentrated near the tips of the discharge needles, this flow pattern forces the gas to traverse the region enriched with high-energy electrons and active radicals (e.g., ·OH), thereby enhancing the gas-species interaction. Second, a distinct recirculation zone is formed downstream of the DR and the holder, with a maximum negative axial velocity of 0.8 m/s. This vortex structure promotes turbulent mixing and extends the residence time of the gas within the discharge region, further contributing to the degradation efficiency.

3.4.2. Enhancement of Degradation Efficiency

To validate the theoretical benefits of the modified flow field observed in the CFD simulation, comparative experiments were conducted. Figure 9a and b present a comparative analysis of the removal efficiencies for inorganic sulfides (H₂S, CS₂) and organic sulfides (DMS, DMDS, CH₃SH) under a 6-needle electrode configuration at a flow rate of 1.5 m³/h, with and without the implementation of a Deflector Ring (DR).

As shown in Figure 9a, the introduction of the DR resulted in a pronounced enhancement in the degradation of inorganic species. For H₂S, the removal efficiency with the DR (H₂S-DR) increased linearly with voltage, reaching approximately 62% at 9.0 kV, significantly outperforming the configuration without the DR (~36%). A similar, albeit slightly less dramatic, improvement was observed for CS₂, where the DR increased removal efficiency by roughly 10–15 percentage points across the voltage range. This substantial divergence indicates that the degradation of these inorganic species is highly sensitive to hydrodynamic conditions, hinting at a mass-transfer limitation in the absence of the DR.

In stark contrast, Figure 9b reveals a negligible impact of the DR on the removal of organic sulfides. Both the DR-equipped and non-DR configurations exhibited nearly identical performance curves. The removal efficiencies for DMDS, DMS, and CH₃SH surged sharply between 7.0 and 8.0 kV, achieving near 100% removal at voltages above 8.0 kV regardless of the presence of the DR.

This overlap suggests that the removal of organic sulfides is primarily governed by energy input (voltage) rather than gas flow patterns, characteristic of a kinetic-controlled regime. Further analysis regarding the underlying reaction mechanisms and bond dissociation energies will be discussed in Section 3.7.

3.5. By-Product Analysis

3.5.1. Byproduct Formation Characteristics

Figure 10 illustrates the variations in the concentrations of secondary pollutants (O₃, NO_x, and SO₂) with applied voltage (7.0–9.0 kV) in the optimized 8-needle reactor at a flow rate of 2 m³/h. The results indicate a monotonic increase in the generation of all three by-products as the discharge voltage rises. Specifically, ozone (O₃) concentration exhibited the most significant growth, rising sharply from less than 0.5 ppm at 7.0 kV to approximately 5.2 ppm at 9.0 kV. This is primarily attributed to the enhanced electric field strength at higher voltages, which intensifies the electron-impact dissociation of oxygen molecules in the background gas (${\mathrm{O}}_2+{\mathrm{e}}^{-\mathrm{*}}\to 2{\mathrm{O}}^{*}+{\mathrm{e}}^{-}$), thereby promoting ozone synthesis. Concurrently, NO_x concentration increased linearly, reaching about 2.6 ppm at 9.0 kV. This increase is linked to the enhanced oxidation of nitrogen species, originating from both the background air and the decomposition of ammonia. In contrast, SO₂ generation remained at a low level (<1.5 ppm) with a moderate growth trend, likely originating from the oxidation of sulfur-containing intermediates. Notably, at the recommended operating voltage of 8.5 kV, the ozone concentration was maintained below 4.6 ppm. This suggests that while increasing voltage enhances active species density, it also heightens the risk of secondary pollution.

3.5.2. Byproduct Mitigation and Environmental Considerations

Although O₃, NO_x, and SO₂ were detected at the ppm level, continuous operation at high gas flow rates may still lead to non-negligible mass emissions; therefore, end-of-pipe polishing is required to avoid secondary pollution and ensure compliant discharge. The proposed DR-MCDBD reactor can be readily integrated with mature post-treatment units. Specifically, O₃ can be efficiently removed using a compact catalytic ozone destruction unit installed downstream of the reactor [26,27]. SO₂ can be effectively removed by mature alkaline wet scrubbing (wet flue gas desulfurization (FGD)), most commonly limestone/lime-based systems (and, in specific cases, caustic scrubbing) [28]. NO_x control can follow established industrial routes when required by the emission limit (e.g., selective catalytic reduction (SCR)) or site-specific absorption/oxidation strategies depending on the gas composition and operating window [29]. These measures are proposed as practical engineering options; their optimization and verification will be addressed in future work.

3.6. Performance Comparison with Other Studies

Table 1 benchmarks the DR-MCDBD reactor against representative NTP systems reported in the recent literature. Many conventional direct-treatment NTP reactors exhibit a performance–energy trade-off and often require relatively high SIE to achieve high removal. Under optimal conditions (8-needle, 8.5 kV), the DR-MCDBD achieves robust removal of reduced-sulfur odorants at a single-digit SIE (6.26 J/L). Specifically, for a multi-component feed (H₂S 5 ppm, CS₂ 1.25 ppm, DMDS 2.5 ppm, CH₃SH 2.5 ppm, DMS 10 ppm, and NH₃ 30 ppm) at Q = 2 m³/h, the removal efficiencies reach 80.9% (H₂S), 45.3% (CS₂), 96.2% (DMDS), and 100% for CH₃SH/DMS/NH₃, with energy efficiency of 1.11–16 g/kWh depending on the species.

We note that advanced DBD variants can also achieve low-SIE/high-removal performance. For example, ref. [11] reports 95% NH₃ and 54% H₂S removal at 2 J/L with 21.6 ppm residual O₃, and ref. [30] reports ~100% removal at 3.37 J/L for single-digit ppm odorants using a wire-mesh DBD, while showing NO₂ formation (1.2–16 ppm). These results indicate that inlet concentration, residence time/flow conditions, and by-product formation strongly affect the apparent energy efficiency. In our case, the DR-MCDBD shows low outlet by-products (4.54 ppm O₃ and 2.06 ppm NO_x, with 1.43 ppm SO₂) while treating a more complex, ammonia-rich odor mixture, benefiting from direct micro-gap discharges and a flow-deflector design that enhances residence time and active-species utilization.

Table 1. Comparison of removal performance and energy efficiency between the DR-MCDBD reactor (under optimal conditions) and other NTP systems reported in the literature.

Plasma Parameters	Concentration	Energy Supply	Removal Efficiency	EE	By-Products	Ref.
DBD	75 ppm of NH3	113 J/L, 19 kV	80%	/	38 ppm of O3, 32 ppm of NOx	[31]
DBD, RH = 20%	20 ppm of NH3 and H2S	2 J/L, 9 kV	95% and 54%	/	21.6 ppm of O3	[11]
Wire and tube plasma reactor, Q = 6 L/min	80 ppm of NH3 and H2S	235 J/L, 16 kV, 200 Hz	73.3% and 84.5%	0.75 and 2 g/kWh	66.7 ppm of SO2, 266.5 ppm of NO2, 55.4 ppm of O3	[32]
Wire-mesh DBD, residence time = 1 s	4.5 ppm of H2S, 6.7 ppm of CH3SH and 3 ppm of NH3	3.37 J/L	99.8%, 100% and 100%	6.7,14.1 and 2.3 g/kWh	1.2–16 ppm of NO2	[30]
DBD, Q = 140 L/min	125 mg/m3 of NH3	78.9 J/L, 9 kV	77.9%	4.5 g/kWh	0.53 mg/L of ${\mathrm{N}\mathrm{O}}_2$ and 2.79 mg/L of NO3	[12]
Wire-plate pulse corona reactor, Q = 13 m3/h	105 mg/m3 of CH3CH2SH and 40 mg/m3 of NH3	65.1 J/L, 200 Hz, 54 kV	93.1% and 100%	/	NOx, O3, SO2, CO2 and CO	[14]
	200 mg/m3 of CH3CH2SH and H2S	65.1–81.4 J/L	95.6% and 100%	/	NOx, O3, SO2, CO2 and CO
DDBD, Q = 1 L/min	100 ppm of CS2	960–3222 J/L, 6.52–7.04 kV	86.32% and 100%	0.35 and 1.01 g/kWh	SO2 and CO2	[16]
gliding arc discharge plasma reactor	300 mg/m3 of NH3 and 200 mg/m3 of H2S	1262 J/L, 11 kV	77% and 100%	0.13 g/kWh	22 mg/L of HCN and 27.8 mg/L of SO2	[33]
wire and tube low temperature plasma reactor	240 mg/m3 of NH3 and 120 mg/m3 of H2S	202.8 J/L, 15 kV, 200 Hz	66.1% and 67.3%	/	O3, SOx and NOx	[15]
DBD, Q = 0.25 m3/h	22 mg/m3 of DMDS	22.5 J/L, 20 kV	61%	/	SO2 selectivity 33.86%	[34]
	60 mg/m3 of DMDS	22.5 J/L, 20 kV	22%	/	SO2 selectivity 7.75%
	22 mg/m3 of DMDS	10 J/L, 20 kV	32.68%	/	SO2 selectivity 10%
DBD, Q = 3.75 L/min	75 mg/m3 of H2S	235.2 JL, 20 kV	22%	0.25 g/kWh	O3 and SOx	[35]
	75 mg/m3 of H2S	382.4 JL, 27.5 kV	47%	0.32 g/kWh	O3 and SOx
DR-MCDBD, Q = 2 m3/h	H2S (5 ppm), CS2 (1.25 ppm), DMDS (2.5 ppm), CH3SH (2.5 ppm), DMS (10 ppm) and NH3 (30 ppm)	6.26 J/L, 8.5 kV	80.9%, 45.3%, 96.2%, and 100%	1.11–16 g/ kWh	4.54 ppm of O3, 2.06 ppm of NOx, 1.43 ppm of SO2	This work

Table 1. Comparison of removal performance and energy efficiency between the DR-MCDBD reactor (under optimal conditions) and other NTP systems reported in the literature.

Plasma Parameters	Concentration	Energy Supply	Removal Efficiency	EE	By-Products	Ref.
DBD	75 ppm of NH3	113 J/L, 19 kV	80%	/	38 ppm of O3, 32 ppm of NOx	[31]
DBD, RH = 20%	20 ppm of NH3 and H2S	2 J/L, 9 kV	95% and 54%	/	21.6 ppm of O3	[11]
Wire and tube plasma reactor, Q = 6 L/min	80 ppm of NH3 and H2S	235 J/L, 16 kV, 200 Hz	73.3% and 84.5%	0.75 and 2 g/kWh	66.7 ppm of SO2, 266.5 ppm of NO2, 55.4 ppm of O3	[32]
Wire-mesh DBD, residence time = 1 s	4.5 ppm of H2S, 6.7 ppm of CH3SH and 3 ppm of NH3	3.37 J/L	99.8%, 100% and 100%	6.7,14.1 and 2.3 g/kWh	1.2–16 ppm of NO2	[30]
DBD, Q = 140 L/min	125 mg/m3 of NH3	78.9 J/L, 9 kV	77.9%	4.5 g/kWh	0.53 mg/L of ${\mathrm{N}\mathrm{O}}_2$ and 2.79 mg/L of NO3	[12]
Wire-plate pulse corona reactor, Q = 13 m3/h	105 mg/m3 of CH3CH2SH and 40 mg/m3 of NH3	65.1 J/L, 200 Hz, 54 kV	93.1% and 100%	/	NOx, O3, SO2, CO2 and CO	[14]
	200 mg/m3 of CH3CH2SH and H2S	65.1–81.4 J/L	95.6% and 100%	/	NOx, O3, SO2, CO2 and CO
DDBD, Q = 1 L/min	100 ppm of CS2	960–3222 J/L, 6.52–7.04 kV	86.32% and 100%	0.35 and 1.01 g/kWh	SO2 and CO2	[16]
gliding arc discharge plasma reactor	300 mg/m3 of NH3 and 200 mg/m3 of H2S	1262 J/L, 11 kV	77% and 100%	0.13 g/kWh	22 mg/L of HCN and 27.8 mg/L of SO2	[33]
wire and tube low temperature plasma reactor	240 mg/m3 of NH3 and 120 mg/m3 of H2S	202.8 J/L, 15 kV, 200 Hz	66.1% and 67.3%	/	O3, SOx and NOx	[15]
DBD, Q = 0.25 m3/h	22 mg/m3 of DMDS	22.5 J/L, 20 kV	61%	/	SO2 selectivity 33.86%	[34]
	60 mg/m3 of DMDS	22.5 J/L, 20 kV	22%	/	SO2 selectivity 7.75%
	22 mg/m3 of DMDS	10 J/L, 20 kV	32.68%	/	SO2 selectivity 10%
DBD, Q = 3.75 L/min	75 mg/m3 of H2S	235.2 JL, 20 kV	22%	0.25 g/kWh	O3 and SOx	[35]
	75 mg/m3 of H2S	382.4 JL, 27.5 kV	47%	0.32 g/kWh	O3 and SOx
DR-MCDBD, Q = 2 m3/h	H2S (5 ppm), CS2 (1.25 ppm), DMDS (2.5 ppm), CH3SH (2.5 ppm), DMS (10 ppm) and NH3 (30 ppm)	6.26 J/L, 8.5 kV	80.9%, 45.3%, 96.2%, and 100%	1.11–16 g/ kWh	4.54 ppm of O3, 2.06 ppm of NOx, 1.43 ppm of SO2	This work

3.7. Mechanism Discussion

The experimental results above highlight a distinct disparity in removal behaviors: organic sulfides remained insensitive to flow field alterations (Figure 9b), while inorganic sulfides showed significant enhancement with the DR (Figure 9a). Based on these observations, we propose a mechanism that couples hydrodynamics with the localized energy field to interpret the degradation pathways, supported by CFD simulations and reaction kinetics analysis.

3.7.1. Molecular Intrinsic Reactivity and Degradation Pathways

The macroscopic degradation performance differs significantly between pollutants (Figure 9). This disparity fundamentally originates from the intrinsic differences in their molecular stability and reaction kinetics. The $\cdot \mathrm{O}\mathrm{H}$-initiated oxidation of a pollutant P is described by the elementary bimolecular rate law r = k[ $\cdot \mathrm{O}\mathrm{H}$][P], which is first-order in [$\cdot \mathrm{O}\mathrm{H}$] and first-order in [P] (second-order overall). Under quasi-steady [$\cdot \mathrm{O}\mathrm{H}$] conditions, the pollutant decay follows pseudo-first-order kinetics with k′ = k[$\cdot \mathrm{O}\mathrm{H}$]_eff, where [$\cdot \mathrm{O}\mathrm{H}$]_eff denotes the effective exposure-averaged radical level experienced by pollutant molecules in plasma-reactive zones.

Quantitatively, the reaction rate constants (k) of organic sulfides (e.g., DMDS, CH₃SH) with hydroxyl radicals ($\cdot \mathrm{O}\mathrm{H}$) are 1–2 orders of magnitude higher than those of inorganic sulfides (e.g., H₂S, CS₂). According to evaluated kinetic sources [36,37], the rate constants follow the order: DMDS > CH₃SH > DMS > H₂S > CS₂.

Specifically, organic sulfides exhibit rapid kinetics. The rate constant for DMDS is approximately $2.0\times 10^{-10}\mathrm{c}{\mathrm{m}}^3\mathrm{m}\mathrm{o}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{u}\mathrm{l}{\mathrm{e}}^{-1}{\mathrm{s}}^{-1}$, and for CH₃SH it is $3.3\times 10^{-11}\mathrm{c}{\mathrm{m}}^3\mathrm{m}\mathrm{o}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{u}\mathrm{l}{\mathrm{e}}^{-1}{\mathrm{s}}^{-1}$. Although its reaction with $\cdot \mathrm{O}\mathrm{H}$ is slower ($k\approx 6.5\times 10^{-12}\mathrm{c}{\mathrm{m}}^3\mathrm{m}\mathrm{o}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{u}\mathrm{l}{\mathrm{e}}^{-1}{\mathrm{s}}^{-1}$), DMS remains more reactive than H₂S and can also undergo rapid oxidation by other species such as oxygen atoms ($k\left(O\right)\approx 3.7\times {10}^{-11}{ \mathrm{cm}}^3{\mathrm{molecule}}^{-1}{\mathrm{s}}^{-1}$). This high reactivity is attributed to their relatively weak C–S or S–H bonds (e.g., C–S ≈ 2.85 eV [38]), which allows for facile H-abstraction or adduct formation. These fast pathways (Equation (4)) render organic sulfides less sensitive to hydrodynamic bypass (Kinetic-Controlled Regime).

$${CH}_{3}SH/DMS/DMDS\stackrel{e^{-\mathrm{*}}}{\to }[\cdot {CH}_3,\cdot S,\cdot SH]\stackrel{\cdot OH,\cdot O}{\to }{SO}_2+{CO}_2+H_{2}O$$

(4)

To quantitatively connect intrinsic kinetics with reactor-scale performance, we define a residence-time-based Damköhler number, $Da$ = k[$\cdot \mathrm{O}\mathrm{H}$]_eff${\tau }_{\mathrm{res}}$. In the discharge zone, the residence time is ${\tau }_{\mathrm{res}}\approx 0.57\mathrm{s}$ at a flow rate of 2 m³h⁻¹. The critical radical level required to reach $Da\approx 1$, then ${\left[\cdot \mathrm{OH}\right]}^{\ast }=1/\left(k{\mathsf{\tau }}_{\mathrm{res}}\right)$. Calculated thresholds yield $\cdot {\mathrm{O}\mathrm{H}}^{\mathrm{*}}~8.8\times {10}^9{ \mathrm{cm}}^{-3}$ for DMDS, $5.3\times {10}^{10 }{\mathrm{cm}}^{-3}$ for CH₃SH, $2.7\times {10}^{11 }{\mathrm{cm}}^{-3}$ for DMS, $3.7\times {10}^{11} {\mathrm{cm}}^{-3}$ for H₂S, and $1.5\times {10}^{12}{ \mathrm{cm}}^{-3}$ for CS₂. Therefore, DMDS can reach $\mathrm{D}\mathrm{a}\gtrsim 1$ at much lower effective radical exposures than CS₂, providing a quantitative basis for the observed transition from kinetic-controlled to mass-transfer-limited (transport-/exposure-limited) behavior.

It is important to emphasize that [$\cdot \mathrm{O}\mathrm{H}$]_eff represents an exposure-averaged radical level, as OH radicals in atmospheric-pressure discharges are strongly localized in space and time. Absolute diagnostics, such as UV absorption and calibrated LIF/PLIF in air/H₂O DBDs and pulsed discharges, report OH densities spanning $~{10}^{13} {\mathrm{c}\mathrm{m}}^{-3}$ at the reactor scale, while substantially higher transient OH levels occur in microdischarge/filament high-field regions [39,40]. In particular, UV absorption measurements for kHz air DBDs show OH densities on the order of $~{10}^{19} {\mathrm{m}}^{-3}$ ($~{10}^{13}{\mathrm{cm}}^{-3}$) and indicate that OH is generated predominantly inside microdischarges [39]. Time-resolved PLIF studies further demonstrate pronounced filament-core-associated OH structures with local peaks significantly exceeding bulk averages [40,41]. Therefore, reaching the higher critical level required for CS₂ (${\left[\cdot \mathrm{OH}\right]}^{\ast }~1.5\times {10}^{12} {\mathrm{cm}}^{-3}$) is consistent with the necessity of experiencing sufficiently high local radical densities (on the order of the critical ${\left[\cdot \mathrm{OH}\right]}^{\ast }$) found in plasma-active micro-discharge/streamer regions. This physical constraint reinforces why CS₂ removal is highly sensitive to the flow-field optimization provided by the Deflector Rings.

In contrast, inorganic sulfides show significantly slower kinetics. The rate constant for H₂S with $\cdot \mathrm{O}\mathrm{H}$ is roughly $4.7\times {10}^{-12} {\mathrm{cm}}^3{\mathrm{molecule}}^{-1}{\mathrm{s}}^{-1}$ (Equation (5)). CS₂ is the most recalcitrant, with a rate constant of only $\sim 1.2\times {10}^{-12} {\mathrm{cm}}^3{\mathrm{molecule}}^{-1}{\mathrm{s}}^{-1}$. This kinetic inertness stems from the combined effect of the strong C=S double bonds (BDE ≈ 4.5 eV [42]) and a slow initial $\cdot \mathrm{O}\mathrm{H}$-addition mechanism, presenting a high kinetic barrier to radical oxidation. Consequently, CS₂ degradation may additionally involve plasma-specific pathways (e.g., electron-impact dissociation) in high-energy regions (Equation (6)). In this work, the term “Mass Transfer-Limited Regime” refers to limitations by transport and effective exposure—specifically, the probability of molecules entering plasma-active streamer cores to experience sufficiently high local radical/electron densities (on the order of the critical [⋅OH])—rather than molecular diffusion alone. This explains why CS₂ removal is highly sensitive to flow-field optimization by the Deflector Rings.

$$H_{2}S\stackrel{e^{-\mathrm{*}}}{\to }[\cdot H,\cdot S,\cdot HS]\stackrel{\cdot OH,\cdot O}{\to }{SO}_2/{SO}_3/S+H_{2}O$$

(5)

$${CS}_2\stackrel{e^{-\mathrm{*}}}{\to }[\cdot S,\cdot CS,\cdot C]\stackrel{\cdot OH,\cdot O}{\to }{SO}_2+{CO}_2+(OCS,{NO}_x\dots )$$

(6)

3.7.2. Chemical Synergy and Potential Gas-to-Particle Conversion

Beyond the independent degradation pathways, a distinct synergistic effect exists between NH₃ and sulfur-containing compounds, which significantly enhances the overall removal efficiency [12].

Radical Propagation: The co-existence of NH₃ and H₂S may establish a self-sustaining radical cycle. The reactive hydrogen atoms ($\cdot \mathrm{H}$) released from H₂S dissociation (Equation (5)) may act as cross-initiators. Concurrently, bulk radicals ($\cdot \mathrm{O}\mathrm{H}$, $\cdot \mathrm{O}$) accelerate the decomposition of NH₃ into amino radicals (${\cdot \mathrm{N}\mathrm{H}}_2$) via hydrogen abstraction, as shown in Equation (7). Note that the reaction with $\cdot \mathrm{O}$ generates secondary $\cdot \mathrm{O}\mathrm{H}$ radicals, further propagating the chain reaction. The generated NH₂ subsequently contributes to De-NO_x processes (Equation (8)).

$${NH}_3+\cdot OH/\cdot O\to {\cdot NH}_2+H_{2}O/\cdot OH$$

(7)

$$\cdot {NH}_2+NO\to N_2+H_{2}O$$

(8)

Gas-to-Particle Conversion: A plausible terminal pathway is the in situ conversion of gaseous pollutants into solid deposits. As depicted in Equations (9) and (10), the acidic oxidation products (HNO₃ and H₂SO₄) are proposed to react with excess NH₃ to form ammonium nitrate and ammonium sulfate.

$${NO}_2+\cdot OH\to {HNO}_3\stackrel{{NH}_3}{\to }{NH}_4{NO}_3\left(s\right)$$

(9)

$${SO}_2+\cdot O\to {SO}_3\stackrel{H_{2}O}{\to }{H_{2}SO}_4\stackrel{{NH}_3}{\to }{\left({NH}_4\right)}_2{SO}_4\left(s\right)$$

(10)

This hypothesis is supported by the qualitative observation of yellowish-white solid residues accumulating on the inner surface of the dielectric barrier tube. The white precipitates are consistent with ammonium salt formation, whereas the yellowish traces may suggest the co-deposition of sulfur-containing species (e.g., elemental sulfur) derived from H₂S-related pathways (Equation (5)). These observations suggest that a fraction of the pollutants may be removed via gas-to-particle conversion (phase transition) rather than solely by gaseous oxidation. However, the deposits were not collected or characterized (mass loading, morphology, or composition) in this study, and the long-term operational impacts (e.g., fouling/pressure-drop increase) were not evaluated; therefore, these issues warrant dedicated future investigations, including deposit characterization (e.g., SEM–EDS/XRD/ion chromatography) and extended durability tests with pressure-drop and discharge-stability monitoring.

3.7.3. Spatial Mismatch and Flow Field Reconstruction

In the baseline MCDBD configuration without the DR, the CFD velocity profiles presented in Figure 8a reveal a critical “Bypass Effect.” A significant fraction of the gas stream flows through the interstitial spaces, effectively bypassing the active high-energy zones at the needle tips. This confirms the spatial mismatch where the bulk gas velocity is decoupled from the discharge energy density.

The introduction of the DR functions as a passive process intensification strategy. As shown in the flow topology in Figure 8b, the DR mechanically blocks the low-field zones and forces the gas stream to converge toward the high-energy electron avalanche zone (indicated by the high-velocity regions). Furthermore, the recirculation vortices observed downstream of the DR structure (refer to Figure 8b) extend the residence time of pollutants. This hydrodynamic reconstruction ensures that intermediates (like carbonyl sulfide (OCS)) and precursors for salt formation remain within the radical-rich environment long enough to undergo complete conversion.

3.7.4. Unified Synergistic Mechanism

Based on the distinct responses to flow field modulation (refer to Figure 8), the degradation process can be categorized into two distinct regimes, as visualized in the upper panel of Figure 11:

• Kinetic-Controlled Regime (Organic Sulfides);

For organic sulfides possessing low BDE, the overall reaction rate is primarily governed by intrinsic chemical kinetics rather than physical transport. As shown in the left panel of Figure 11, these highly reactive molecules can be effectively degraded by radicals dispersed in the bulk flow. This explains the negligible enhancement provided by the DR as observed in Figure 9b.

2.

Mass Transfer-Reaction Co-Controlled Regime (Inorganic Sulfides).

For refractory molecules like CS₂ (High BDE), degradation is strictly confined to the high-energy streamer channels. The DR establishes a co-controlled regime (right panel of Figure 11): it enhances mass transfer by directing pollutants into the discharge zone and promotes reaction completeness by extending residence time. This mechanism perfectly elucidates the experimental results in Figure 9a, where inorganic sulfide removal rose sharply with flow optimization.

4. Conclusions

This study developed a novel Dielectric Barrier Discharge reactor (DR-MCDBD) to resolve the conflict between gas throughput and energy efficiency. The reactor features a Multi-Needle to Cylinder configuration integrated with Flow Deflector Rings. Optimization tests identified the 8-needle topology as the ideal geometry, as it effectively balances discharge density with electric field shielding. Furthermore, the Flow Deflector Rings successfully reconstructed the hydrodynamic field. By mitigating the “bypass effect” and inducing local turbulence, this modification extended gas residence time and enhanced collision probabilities.

Crucially, the study elucidated distinct degradation behaviors. Organic sulfide removal was primarily kinetically controlled. In contrast, refractory inorganic sulfides (e.g., CS₂) followed a mass transfer-reaction co-controlled mechanism. This distinction highlights the necessity of hydrodynamic modulation to overcome mass-transfer limitations. Under optimal conditions, the system removed over 80% of H₂S and nearly 100% of organic sulfides. Notably, these results were achieved at a remarkably low SIE of 6.26 J/L, with O₃/NO_x/SO₂ observed at the ppm level. Overall, the synergistic regulation of discharge and flow fields establishes the DR-MCDBD reactor as a scalable and readily integrable concept for industrial odor control, when coupled with appropriate downstream abatement for O₃/NO_x/SO₂.