A Study on the Development of Destruction or Removal Efficiency (DRE) Considering the Characteristics of Greenhouse Gas Abatement Technology Used in the Semiconductor and Display Industries in South Korea

Abstract

In this study, the Destruction or Removal Efficiency (DRE) of 10 types of F-gases used in the semiconductor and display industries in South Korea was measured. These industries use a large volume of F-gases with high Global Warming Potential (GWP), significantly contributing to national greenhouse gas emissions. Therefore, accurately calculating the greenhouse gas emissions from these industries and establishing appropriate mitigation plans is crucial. The current IPCC guidelines provide parameters for estimating greenhouse gas emissions for each gas, including DRE values. However, they present only a single coefficient for each gas, without considering the diverse abatement technologies that are commercially applied in practice. As a result, there is a potential for overestimating South Korea’s national greenhouse gas emissions, as these guidelines do not reflect the advanced abatement technologies used in each country’s semiconductor and display industries. To address this, the DREs of Combustion-type and Plasma-type abatement technologies, which are widely used in South Korea, were measured based on the Korean KS guidelines, developed from the U.S. EPA’s reduction efficiency measurement guidelines. The results showed that Plasma-type technologies, which are generally known to have better reduction efficiency, achieved higher DRE values compared with Combustion-type technologies. Furthermore, statistical analysis was conducted using SPSS 26 to assess whether it is significant to develop separate DRE values for different technologies. The analysis confirmed that developing distinct DREs for each technology is statistically significant. The findings of this study provide practical guidance for selecting optimal abatement technologies in South Korea’s semiconductor and display industries and serve as fundamental data for contributing to the achievement of sustainable carbon neutrality goals through more accurate greenhouse gas inventories in countries involved in semiconductor and display production.

1. Introduction

Under the Paris Agreement, countries agreed to keep long-term global average temperature increases well below 2 °C above pre-industrial levels, with efforts to limit the increase to 1.5 °C. As part of the Paris Agreement framework, international implementation reviews, known as Global Stocktakes, are conducted every five years starting in 2023 to assess progress toward the long-term goals. Under this framework, participating governments are required to set greenhouse gas (GHG) reduction targets and plans and to report their progress [1].

As part of its commitment to the Paris Agreement, South Korea declared “2050 Carbon Neutrality” in October 2020, and in December 2021, submitted its enhanced Nationally Determined Contribution (NDC), which aims to reduce greenhouse gas emissions by 40% compared with 2018 levels, to the UNFCCC. In March 2023, South Korea announced its “First Basic Plan for Carbon Neutrality and Green Growth”, which outlines specific sectoral and annual GHG reduction targets for achieving its 2030 goals and made some adjustments to these sectoral targets [2].

According to South Korea’s greenhouse gas inventory report, the industrial processes sector emitted 48.5 million tons of CO₂ equivalent in 2020, accounting for 7.4% of the nation’s total emissions. Among these, the semiconductor and display industries are major contributors, representing 31.2% of the emissions within the industrial processes sector [3].

In 2021, South Korea held a significant share in the global market for semiconductors (18.4%) and displays (33.3%), making it the second-largest contributor after China [4,5]. Additionally, as of 2022, these industries accounted for 22% of the country’s total exports [6], and they continue to grow steadily, driven by the expanding AI market and the recovery in demand for IT devices such as TVs and mobile phones [7]. The semiconductor and display industries, not only in Korea but globally, are expected to see continuous growth in energy and process gas demand driven by increasing production volumes and the shift toward higher value-added products [8,9,10]. Although the semiconductor and display industries rely heavily on electricity, with 60–70% of GHG emissions being indirect, there are limitations to reducing these indirect emissions due to the nature of the facilities [11,12]. Therefore, a key reduction strategy being studied involves controlling process emissions, which account for 20–30% of GHG emissions, through methods such as the installation of ultra-high-efficiency abatement technologies and the substitution of process gases [13,14,15,16].

In South Korea, the Tier 2a and 2b methods proposed by the 2006 IPCC Guidelines are used to estimate GHG emissions from the semiconductor and display industries, with emission factors largely based on the default values provided in the 2006 IPCC Guidelines. For the parameters required for GHG emission estimation, such as the fraction of gas i volume used in processes with emission-control technologies (a_i) and the fraction of gas i destroyed by the emission-control technology (d_i), the data from the site’s emission-control technologies are used when available; otherwise, IPCC default values are applied [17,18].

Currently, the IPCC provides single emission factors for each gas without distinguishing between different abatement technologies. However, as South Korea pursues strategies to enhance the efficiency of emission-control technologies in the semiconductor and display sectors to reduce GHG emissions, the IPCC factors do not adequately reflect the specific conditions of South Korea’s industrial processes [19,20,21].

Therefore, this study was conducted to evaluate the reduction rates of F-gases after passing through Combustion-type and Plasma-type abatement technologies, which are used to control direct emissions into the atmosphere following semiconductor and display manufacturing processes, and to determine whether there are significant differences between the abatement technologies, as foundational research for the development of Destruction or Removal Efficiency (DRE) considering the types of abatement technologies used in the semiconductor and display industries. By measuring and analyzing the abatement efficiency of the technologies currently in use at actual industrial sites, the study aims to determine whether there are significant differences among the types of abatement technologies. The results of this research will contribute to proposing optimal emission-control strategies tailored to South Korea’s industrial conditions, improving the accuracy of GHG emission estimations and enhancing the reliability of the national GHG inventory. This, in turn, is expected to aid in achieving South Korea’s carbon neutrality goals and serve as a foundational resource for the development of policies and regulatory standards that consider both environmental sustainability and economic efficiency across the industrial sector.

2. Literature Review

Although extensive research has been conducted globally in various areas of the semiconductor and display industries, studies on the parameters affecting greenhouse gas emissions control for air pollution management remain limited. Below is a summary of the relevant studies on DRE applicable to the semiconductor and display industries.

Yang et al. (2009) investigated a gas chromatography (GC) method to evaluate the DRE of local scrubbers for five perfluorinated compounds (PFCs) used in semiconductor manufacturing facilities: SF₆, NF₃, CF₄, C₂F₆, and C₃F₈. The study focused on abatement technologies of the combustion and electric-thermal types. The results indicated that for combustion systems, C₃F₈ achieved a removal efficiency of over 90%, while CF₄ showed an efficiency of 40–50%. In the case of electric-thermal systems, the removal efficiency for C₃F₈ was below 45%, and for CF₄, it was even lower [22].

Choi et al. (2012) observed that as incineration systems were being replaced with Plasma-type systems to more effectively decompose fluorinated compounds, the difference in DREs for non-decomposable greenhouse gases was significant. The study highlighted the need for research on reactive additives for various waste gases, optimization of torch design and operation, and the development of new thermal plasma torches suitable for decomposition processes, in order to achieve cleaner and more cost-effective thermal plasma decomposition of fluorinated compounds [23].

Ko et al. (2014) measured the DRE of a PLASMA-WET system designed to treat PFC gases generated in semiconductor manufacturing processes. The study found that as the concentration of CF₄ entering the chamber increased, the DRE improved due to the optimized operation of the scrubber and the efficient chemical reactions at high concentrations [24].

Kang et al. (2023) conducted a foundational study to develop DRE for greenhouse gases used in the semiconductor and display industries. Using field-measured DRE data, the study estimated the required sample size for different data processing methods for CF₄ and CHF₃ gases used in the semiconductor and display sectors. The research presented measurement-based DRE data by scrubber type and emphasized the need to consider scrubber types when aiming for representative DRE values for specific gases in these industries [25].

Lee et al. (2023) evaluated the decomposition characteristics of different types of plasma-wet scrubbers through experiments, assessing the DRE and by-product gas generation rates under varying parameters for both etching-type and WF-type plasma-wet scrubbers. The study found that WF-type scrubbers generated significantly fewer by-product gases compared with etching-type scrubbers. The research concluded that when selecting the optimal plasma-wet scrubber for PFCs, both the DRE and the by-product gas generation rate should be considered to minimize the final emission of PFCs [26].

3. Abatement Technologies and Measurement Methods

3.1. Types of Abatement Technologies in the Semiconductor and Display Industries

3.1.1. Point of Use (POU) Abatement

In the semiconductor and display manufacturing processes, abatement technologies designed to treat waste gases, including PFCs emitted during CVD, etching, and cleaning processes, function differently from typical environmental protection facilities, serving as local exhaust systems [27,28,29]. These facilities apply vacuum pump downstream gas scrubber systems to handle mixed gases containing various types of air pollutants. They are directly attached to the exhaust of the manufacturing line (FAB) and are commonly referred to as Point of Use (POU) scrubbers. The types of POU systems include Heat-type, Combustion-type, and Plasma-type, with their respective characteristics summarized in

Table 1. Characteristics of abatement technologies.

Category	Combustion-Type	Plasma-Type
Characteristics	Direct combustion method using a high-temperature burner	Method using electric torch heat(plasma)
Advantages	The most commonly used method	Lower power consumption compared with Combustion-type
Disadvantages	High power consumption Generation of by-products Risk of explosion Combustion instability due to gas fuel	High power consumption Limited removal efficiency

[30,31,32].

Combustion-Type Abatement

The Combustion-type abatement technology was the first to be commercialized. It treats the gases emitted after the process by injecting them together with air into a high-temperature burner for direct combustion. In South Korea, Combustion-type systems were widely used; however, there has been a trend toward switching to Plasma-type systems as part of policies to improve the efficiency of process gas emission reduction technologies. Combustion-type abatement technology processes F-gases through combustible treatment by mixing LNG (Liquefied Natural Gas) with air (O₂) and utilizing an LNG flame. It is typically operated under LNG flow conditions of 6–20 LPM (liters per minute). The Mean Time Between Failures (MTBF) may vary depending on the manufacturer and process environment, but it generally ranges from 6 to 12 months. The typical flame temperature is between 1200 °C and 1500 °C, while the chamber temperature is approximately 900 °C to 1000 °C.

The Combustion-type abatement technology requires maintaining a high temperature, which leads to high power consumption and necessitates an LNG fuel supply line, posing risks such as gas fuel explosion and combustion instability. Additionally, the high-temperature combustion inevitably generates NOx, and the efficiency decreases at lower temperatures [33,34,35].

Plasma-Type Abatement

The Plasma-type abatement technology treats gases using the energy generated by arc plasma created through a high current. Working gases such as N₂, O₂, and air are used under conditions of 20–40 LPM, and the system typically operates within a power range of 6–23 kW. The Mean Time Between Failures varies depending on the manufacturer and process environment, typically ranging from 3 to 4 months. The usual flame maximum temperature is between 10,000 °C and 12,000 °C, while the chamber temperature ranges from 1000 °C to 1300 °C. Compared with other types of technologies, Plasma-type technology experiences significant wear on consumable parts such as cathodes and anodes, resulting in shorter replacement cycles.

This method allows for the efficient removal of fluorinated compounds at lower temperatures by plasma-induced degradation of the fluorinated gases, and it offers the advantage of a smaller size compared with combustion-based reduction systems [36,37]. Similar to other combustion technologies, it can connect to a wet zone at the downstream end to treat water-soluble gases and reaction by-products. While it operates at higher temperatures than direct combustion methods, offering high treatment efficiency, it also produces secondary pollutants such as NOx and has high power consumption [38,39]. Although the efficiency tends to vary depending on the gas load, its excellent PFC decomposition capability has led to increased adoption for greenhouse gas emission reduction.

3.2. Destruction or Removal Efficiency (DRE)

3.2.1. Selection of Target Facilities and Operating Conditions

In this study, to develop the DRE for F-gases used in the semiconductor and display industries, facilities responsible for over 85% of the greenhouse gas emissions in South Korea’s semiconductor and display sectors were selected as measurement sites. Over a period of six years, from 2018 to 2023, measurements were conducted across 1794 abatement units from 15 companies. The abatement technologies measured in this study are classified into Combustion-type and Plasma-type. Among these, the Plasma-type technology generates plasma through the supply of inert gases such as nitrogen and argon to the plasma torch.

3.2.2. Measurement Methods and Experimental Conditions

The DRE of the abatement equipment was measured in accordance with the Korean Industrial Standard KS I 0587 (Measurement Method for Volumetric Flow Rate of Non-CO₂ Greenhouse Gases (CF₄, NF₃, SF₆, N₂O) Used in Semiconductor and Display Processes) [40]. This standard was developed based on the U.S. EPA’s protocol EPA 430-R-10-003 (Protocol for Measuring Destruction or Removal Efficiency of Fluorinated Greenhouse Gas Abatement Equipment in Electronics Manufacturing, Version 1) [41]. The volumetric flow rate of process exhaust gases was measured using a quadrupole mass spectrometer (QMS; isepa-S, EL, Daejeon, Korea), while the concentration of F-gases was determined using a Fourier transform infrared spectrometer (FT-IR; DX4000, Gasmet, Vantaa, Finland) to calculate the DRE of the scrubbers.

Figure 1 shows the measurement setup for determining the concentration of greenhouse gases and the volumetric flow rate of process exhaust gases in semiconductor and display processes. A tracer gas injection line and a sampling inlet line were constructed, with a valve installed at the upstream end. The distance between the tracer gas injection line and the sampling inlet line should be as far as possible (at least 1 m) to ensure sufficient mixing of the tracer gas before it reaches the sampling inlet line. Krypton (Kr, Rigas 99.999%), a chemically stable inert gas, was used as the tracer gas injected into the pipeline to measure the flow rate of the process exhaust gases. The flow rate ratio between the inlet and outlet was not specified separately, as it varies depending on the type of abatement technology, the abatement manufacturer, and the process recipes of semiconductor and display manufacturers.

The configuration for measuring the destruction efficiency of abatement equipment is shown in Figure 2. Sampling ports are installed in the pipelines connected before and after the abatement equipment to create tracer gas injection lines and sample gas inlet lines, with valves installed at the upstream ends. Similar to the flow rate measurement setup, the distance between the tracer gas injection line and the sample gas inlet line should be as far apart as possible to allow sufficient mixing of the tracer gas before it is measured at the sample gas inlet. The results from the QMS and FT-IR were monitored in real time, and the volumetric flow rate and concentration of F-gases were continuously measured over a one-hour period.

The measurement conditions varied depending on the type of abatement technology, as shown in

Table 2. Sampling conditions of target facilities.

Category	Combustion-Type	Plasma-Type
LNG Flow (L/min)	14.8~25.0	N/A
O2 Flow (L/min)	10.0~63.0	N/A
Input Plasma Power(kW)	N/A *	6.4~13.0
Temperature (°C)	N/A	N/A
Total Gas Flow Rate (SLM)	300~500 SLM	300~600 SLM

* N/A means Not Applicable.

. The Combustion-type operated with a total gas flow rate of 300–500 SLM and an LNG input flow rate of 14.8–25.0 L/min. The Plasma-type operated with a total gas flow rate of 300–600 SLM and plasma power of 6.4–13.0 kW. The Catalyst-type operated with a total gas flow rate of 25–129 CMM at a temperature condition of 700 °C.

3.2.3. Calculation of DREs

The volumetric flow rate of F-gases is determined by injecting an inert gas (such as helium or krypton) as a tracer gas into the pipeline through which the process exhaust gas flows using an MFC. The concentration of the tracer gas is then measured using a QMS. The volumetric flow rate of the process exhaust gas is calculated using Equation (1) as follows:

$$F={\displaystyle \frac{S_f}{C_{tracer}\times {10}^{-6}}}$$

(1)

In the formula, F represents the volumetric flow rate of the process exhaust gas for a single concentration (L/min), S_f represents the volumetric flow rate of the tracer gas injected through the gas inlet (L/min), and C_tracer represents the measured concentration of the tracer gas (μmol/mol).

The average volumetric flow rate of the tracer gas is calculated using Equation (2), and the relative standard deviation of the volumetric flow rate of the process exhaust gas is calculated using Equation (3).

$$F_m=\sum _1^n{\displaystyle \frac{F_i}{n}}$$

(2)

$${\sigma }_{F_m}=\sqrt{{\displaystyle \frac{1}{n}}\sum _1^n{(F_i-F_m)}^2}$$

(3)

In the formulas, F_m represents the average volumetric flow rate of the process exhaust gas from the n measured data points (L/min), F_i represents the volumetric flow rate of the process exhaust gas for the i-th measurement (L/min), n represents the number of measurements, and σ_Fm represents the relative standard deviation.

The volumetric flow rate of the target substance is calculated using the concentration measured by FT-IR, as shown in Equation (4):

$$V_i=\sum _{i=1}^n{F}_{m,i}{C}_{i,j}=F_{m,i}\sum _{i=1}^n{C}_{i,j}$$

(4)

In the formula, V_i represents the volumetric flow rate of the target substance (L/min), F_m represents average volumetric flow rate of the process exhaust gas based on single concentration data points (L/min), C_i represents the concentration of the target substance (μmol/mol), and C_i,j represents the concentration of F-gas in the supply line or exhaust outlet of the facility (μmol/mol).

The DRE is calculated using Equation (5) by measuring the volumetric flow rates of the F-gas at both the inlet and outlet of the abatement equipment during a specific period of normal operation.

$$DRE=(1-{\displaystyle \frac{V_{in}}{V_{out}}})\times 100$$

(5)

In the formula, V_in represents the volumetric flow rate of F-gas entering the abatement equipment during normal operation (L/min) and V_out represents the volumetric flow rate of F-gas exiting the abatement equipment during normal operation (L/min).

3.3. Statistical Analysis of Differences in DREs by Abatement Technology

In this study, a statistical analysis was conducted to determine whether there are significant differences in DREs across different abatement technologies. The average DRE of each abatement technology was compared. Statistical analysis was performed using the SPSS 26 software program.

3.4. Uncertainty Analysis of F-Gases’ DREs

The uncertainty in estimates derived from calculated values or from data relies on various factors such as measurement variability, differences in standard conditions, environmental factors during measurement, and other sources of uncertainty like the standard deviation associated with the equipment used. The variability of such uncertainty can affect the final results. Bias may occur due to systematic errors, while random errors cause the results to deviate symmetrically around the mean [42]. Such random errors are generally expressed as uncertainty, which quantifies the degree of reliability or confidence in the measurement results [43,44].

The 2019 IPCC Guidelines suggest various methods for selecting uncertainty. In this study, we adopted the Tier 1 approach for selecting uncertainty to calculate the DRE uncertainty [45]. When calculating uncertainty, if the sample size is sufficiently large and the number of measured data points is sufficiently large, the Central Limit Theorem (CLT) can be applied. According to the CLT, the sample mean $\overline{X}$ will follow a normal distribution with a mean $\mu$ and a standard deviation of σ/$\sqrt{n}$. This allows for the estimation of the uncertainty in the mission factor. Thus, using the z-value of 1.96, corresponding to a 95% confidence level, the relative uncertainty of the average can be calculated using Equation (6):

$$U=\pm ({\displaystyle \frac{1.96\times \sigma }{X}})\times 100$$

(6)

In the formula, U represents the uncertainty; 1.96 represents a z score with 95% confidence interval, assuming normal distribution; σ represents standard deviation; and X represents the mean of sample measurements.

4. Result and Discussion

4.1. DRE Estimation Results by Gas Type

In this study, experiments were conducted to develop the DRE for F-gases used in the semiconductor and display industries. As shown in

Table 3. Summary of collected data by gas type.

No.	Target Gas	Sample	Sampling Period
1	C2F6	233	2018~2023
2	C3F8	55
3	c-C4F8	73
4	CF4	720
5	CHF3	144
6	SF6	232
7	C2HF5	83
8	CH2F2	63
9	N2O	153
10	C4F6	38
Total samples		1794

, facilities using each type of gas were selected as the target sites. The selected facilities account for over 85% of the greenhouse gas emissions from the semiconductor and display industries in South Korea. The experiments were conducted at facilities from 10 companies that emit fluorinated gases (C₂F₆, C₃F₈, C-C₄F₈, CF₄, CHF₃, SF₆, C₂HF₅, CH₂F₂, N₂O, C₄F₆) from 2018 to 2023, covering 1794 abatements. All gases measured in this study are process gases used in semiconductor and display manufacturing processes, and they are collectively referred to as F-gases, including HFCs, PFCs, SF₆, and N₂O. The selection process considered the production processes and the characteristics of semiconductor and display products, which involve complex steps and confidential information. As many of the products are already commercialized, the number of gases included in the study was limited due to the availability of consistent sampling conditions.

The DREs for 10 F-gases used in the semiconductor and display industries were calculated and compared with the values provided by the IPCC, as shown in

Table 4. Comparison of DREs for F-gases by gas type.

Target Gas	2006 IPCC GLs	2019 IPCC GLs	This Study
			DRE (Mean)	SD
C2F6	0.9	0.98	0.97	0.04
C3F8	0.9	0.99	0.98	0.03
c-C4F8	0.9	0.98	0.98	0.03
CF4	0.9	0.89	0.92	0.11
CHF3	0.9	0.98	0.98	0.05
SF6	0.9	0.96	0.98	0.07
C2HF5	N/A	0.98	0.997	0.01
CH2F2	N/A	0.99	0.97	0.02
N2O	N/A	0.60	0.88	0.15
C4F6	N/A	0.98	0.98	0.02

[46]. The 2006 IPCC Guidelines do not provide separate default DREs for the four gases: C₂HF₅, CH₂F₂, N₂O, and C₄F₆. Among the F-gases analyzed, C₂HF₅ showed the highest DRE at 0.997 ± 0.012, followed by C₄F₆ at 0.98 ± 0.02, and both C₃F₈ and C₄F₈ at 0.98 ± 0.03. N₂O exhibited the lowest DRE at 0.88 ± 0.15, followed by CF₄ at 0.92 ± 0.11.

The 2006 IPCC Guidelines (IPCC GLs) provide DREs for six gases (C₂F₆, C₃F₈, c-C₄F₈, CF₄, CHF₃, SF₆), with an average of 0.92. When comparing the DREs obtained in this study, most gases demonstrated values higher than those suggested in the IPCC GLs. The 2019 IPCC GLs extended the list to 15 gases, including CF₄, CHF₃, SF₆, C₂HF₅, CH₂F₂, CH₃F, N₂O, C₄F₆, C₂F₄, and other gases such as C₅F₈ and NF₃. The results of this study show that the DREs for CF₄, SF₆, and C₂HF₅ exceeded the values suggested in the GL, while c-C₄F₈, CHF₃, and C₄F₆ showed similar results. On the other hand, gases such as C₂F₆, C₃F8, and CH₂F₂ showed lower DREs than those reported in the 2019 IPCC GLs. Additionally, N₂O exhibited a significant deviation, showing the lowest DRE, with a difference of approximately 47% compared with the values provided by the GL.

This discrepancy is likely due to the lower reactivity of N₂O within abatement, making it difficult to achieve efficient removal [47,48]. Furthermore, the values suggested by the IPCC reflect globally averaged process performance under typical conditions, which may lead to differences from the results of this study.

4.2. DRE Estimation Results by Abatement Technology

To determine whether there are significant differences in DRE based on the type of abatement technology, the collected data were categorized as shown in

Table 5. Summary of collected data by abatement technology.

No.	Target Gas	Abatement Technology		Total Samples
		Combustion-Type	Plasma-Type
1	C2F6	76	157	233
2	C3F8	N/A	55	55
3	c-C4F8	N/A	73	73
4	CF4	301	419	720
5	CHF3	63	81	144
6	SF6	146	86	232
7	C2HF5	83	N/A	83
8	CH2F2	N/A	63	63
9	N2O	66	87	153
10	C4F6	N/A	38	38
Total samples		735	1059	1764

. The results indicated that abatement technologies could be classified into Combustion-type and Plasma-type systems. Among the 1794 abatements analyzed, 735 were classified as Combustion-type and 1059 as Plasma-type. The Combustion-type systems primarily targeted six gases (C₂F₆, CF₄, CHF₃, SF₆, C₂HF₅, N₂O), while the Plasma-type systems covered nine gases (C₂F₆, C₃F₈, c-C₄F₈, CF₄, CHF₃, SF₆, CH₂F₂, N₂O, C₄F₆). Due to differences in abatement technology usage, practical constraints during actual measurements, and site-specific confidentiality concerns, the number of samples available for each gas was limited.

The DREs for 10 F-gases used in the semiconductor and display industries were estimated based on abatement technologies, and the results were compared with the values provided by the IPCC, as shown in

Table 6. Comparison of DRE for F-gases by gas type.

Target Gas	2006 IPCC GLs	2019 IPCC GLs	This Study
			Combustion-Type		Plasma-Type
			DRE (Mean)	SD	DRE (Mean)	SD
C2F6	0.9	0.98	0.95	0.05	0.98	0.03
C3F8	0.9	0.99	N/A	N/A	0.98	0.03
c-C4F8	0.9	0.98	N/A	N/A	0.98	0.03
CF4	0.9	0.89	0.86	0.14	0.96	0.05
CHF3	0.9	0.98	0.97	0.06	0.99	0.03
SF6	0.9	0.96	0.98	0.08	0.97	0.05
C2HF5	N/A	0.98	0.997	0.01	N/A	N/A
CH2F2	N/A	0.99	N/A	N/A	0.97	0.02
N2O	N/A	0.60	0.81	0.13	0.94	0.14
C4F6	N/A	0.98	N/A	N/A	0.98	0.02

. Among the Combustion-type technologies, the highest DRE was observed for C₂HF₅ at 0.997 ± 0.012, followed by SF₆ at 0.98 ± 0.03 and CHF₃ at 0.97 ± 0.02. On the other hand, N₂O had the lowest DRE at 0.88 ± 0.13, and CF₄ showed a relatively low DRE at 0.86 ± 0.14.

For Plasma-type technologies, CHF₃ exhibited the highest DRE at 0.99 ± 0.03, followed by C₄F₆ at 0.98 ± 0.02. C₂F₆, C₃F₈, and c-C₄F₈ also showed relatively high DREs at 0.98 ± 0.03. Conversely, N₂O had a lower DRE at 0.94 ± 0.14, while CF₄ had a DRE of 0.98 ± 0.05, indicating a comparatively lower removal efficiency.

When comparing DREs by abatement technology, both Combustion-type and Plasma-type technologies showed consistently high DREs for five gases (C₂F₆, CF₄, CHF₃, SF₆, and N₂O), with Plasma-type technologies generally outperforming Combustion-type technologies, except for SF₆. The DRE for SF₆ was slightly lower in Plasma-type technologies compared with Combustion-type technologies, but this difference was within an acceptable range. For gases such as C₂F₆, CF₄, CHF₃, and N₂O, the DRE difference by abatement technology was minimal, with differences of approximately 0.02 to 0.10. However, for N₂O, a difference of approximately 10% was observed between the two technologies.

The 2006 IPCC GLs provide DRE values for six gases (C₂F₆, C₃F₈, c-C₄F₈, CF₄, CHF₃, SF₆) without distinguishing between abatement technologies. These values represent average removal efficiencies across Combustion-type, Plasma-type, and Catalyst-type technologies. Compared with these reference values, the DREs observed in this study were generally higher, except for CF₄, which showed a slightly lower DRE for the Combustion-type technologies.

The 2019 IPCC GLs extended the list to include 15 gases, including C₂F₆, C₃F₈, c-C₄F₈, CF₄, CHF₃, SF₆, CH₂F₂, N₂O, C₄F₆, C₂F₄, C₄F₈O, C₂HF₅, C₅F₈, and NF₃. The DREs provided in the 2019 IPCC GLs are mostly lower than those found in this study, indicating that the abatement technologies used here offer better removal efficiency for most gases. However, when comparing the 2006 IPCC GLs and the study results, the DREs differed by 3% to 57%, with Plasma-type technologies showing consistently higher efficiency than Combustion-type technologies.

Combustion-type technology utilizes combustion reactions to decompose gases at high temperatures, but for gases with low reactivity, combustion alone may not be sufficient for complete decomposition. In contrast, Plasma-type technology promotes more efficient gas decomposition compared with the Combustion-type through electrical reactions. CF₄ and N₂O are chemically stable and have low reactivity, which likely resulted in lower DREs for the Combustion-type technology.

4.3. Statistical Analysis Results on DRE Differences by Abatement Technology

In this study, an independent t-test was conducted to determine whether there are significant differences in DRE by abatement technology. The analysis focused on five gases (C₂F₆, CF₄, CHF₃, N₂O, SF₆) and was performed using SPSS 26 software. The results of the statistical analysis are shown in

Table 7. Results of independent samples t-test for DRE differences between Combustion-type and Plasma-type technologies.

Target Gas	Null Hypothesis	Levene’s Test (p-Value)	Independent Samples t-Test (p-Value)	Decision
C2F6	The mean DRE for Plasma-type and Combustion-type are the same.	<0.001	<0.001	Reject the null hypothesis
CF4	The mean DRE for Plasma-type and Combustion-type are the same.	<0.001	<0.001	Reject the null hypothesis
CHF3	The mean DRE for Plasma-type and Combustion-type are the same.	<0.001	0.062	Retain the null hypothesis
N2O	The mean DRE for Plasma-type and Combustion-type are the same.	0.098	<0.001	Reject the null hypothesis
SF6	The mean DRE for Plasma-type and Combustion-type are the same.	0.746	0.651	Retain the null hypothesis

.

First, a Levene’s test was conducted to determine whether the variances in DRE between the Combustion-type and Plasma-type technologies are equal. The null hypothesis for the Levene’s test is that “the variances in DRE by abatement technology are equal [49,50]”. The test results showed that the p-values for C₂F₆, CF₄, and N₂O were less than 0.001, leading to the rejection of the null hypothesis and confirming that the variances are not equal. Based on these results, the independent t-test was conducted under the assumption of unequal variances, showing that significant differences exist for C₂F₆, CF₄, and N₂O, with p-values less than 0.001, indicating that DRE varies depending on the abatement technology used. On the other hand, for CHF₃ and SF₆, the p-values were greater than 0.001, suggesting that the null hypothesis of equal variances could not be rejected, implying no significant differences in DRE by abatement technology.

4.4. Uncertainty Analysis Results

The uncertainty of F-gas DREs was estimated using the sampling uncertainty method suggested in the 2019 IPCC Refinement, considering a 95% confidence interval. The results are presented in

Table 8. Uncertainty in DRE for F-gases by abatement technology.

Target Gas	This Study
	Combustion-Type	Plasma-Type
C2F6	±1.29%	±0.43%
C3F8	N/A	±0.70%
c-C4F8	N/A	±0.70%
CF4	±1.79%	±0.52%
CHF3	±1.52%	±0.61%
SF6	±1.33%	±1.03%
C2HF5	±0.28%	N/A
CH2F2	N/A	±0.59%
N2O	±3.86%	±3.09%
C4F6	N/A	±0.61%

. Currently, the IPCC GLs do not provide separate uncertainty values for DRE, making direct comparison difficult.

When comparing the uncertainty by abatement technology, the range for Combustion-type technology was found to be from ±0.28% to ±3.86%, while for Plasma-type technology, the range was from ±0.43% to ±3.09%, with Plasma-type technology generally showing smaller uncertainties compared with Combustion-type technology. The smallest uncertainty in Combustion-type technology was observed for C₂HF₅ at ±0.28%, while for Plasma-type technology, the smallest uncertainty was for C₂F₆ at ±0.43%. On the other hand, N₂O exhibited the highest uncertainties in both Combustion-type and Plasma-type technologies, with ranges of ±3.86% and ±3.09%, respectively. Due to the chemical stability and low reactivity of N₂O compared with other gases, it is predicted that slight changes in process conditions could lead to significant variability in uncertainty.

Uncertainty tends to decrease as the number of samples increases, and in cases where more samples were collected, the uncertainty was lower [51,52]. Plasma-type technology generally involved collecting more samples compared with Combustion-type technology, leading to lower uncertainties for the Plasma-type technology. For instance, 157 samples of C₂F₆ were collected for Plasma-type technology compared with 76 samples for Combustion-type technology, resulting in lower uncertainty for Plasma-type (±0.43%) compared with Combustion-type (±1.29%). Similarly, 419 samples of CF₄ were collected for Plasma-type technology compared with 301 samples for Combustion-type technology, resulting in lower uncertainty for Plasma-type (±0.52%) compared with Combustion-type (±1.17%). However, for CHF₃, even though Plasma-type technology collected 87 samples compared with 86 samples for Combustion-type technology, the uncertainty remained similar, with Plasma-type at ±3.09% and Combustion-type at ±3.86%.

5. Conclusions

This study conducted a comparative analysis of the Destruction or Removal Efficiency (DRE) and the uncertainty of abatement technologies used in the semiconductor and display industries in South Korea. The experimental results comparing Plasma-type and Combustion-type abatement technologies showed that, overall, the Plasma-type systems exhibited higher DREs than the Combustion-type systems. Notably, for chemically stable and low-reactivity gases like CF₄ and N₂O, the Plasma-type systems demonstrated significantly higher DREs compared with the Combustion-type systems. The DRE results indicated that the average DRE of Plasma-type systems was generally above 0.95, indicating stable abatement performance. In contrast, the Combustion-type systems recorded average DREs ranging from 0.85 to 0.95, showing relatively lower performance. Particularly for N₂O, the Combustion-type system exhibited a DRE of approximately 0.80, while the Plasma-type system achieved a DRE above 0.93, highlighting a clear performance difference between the two systems. Based on statistical analysis, the null hypothesis was rejected for more than half of the gases tested, confirming that there are performance differences in greenhouse gas abatement depending on the type of abatement technology used.

In the uncertainty analysis, the Combustion-type systems showed a range of ±0.28% to ±3.86%, while the Plasma-type systems showed a range of ±0.43% to ±3.09%, with the Plasma-type systems generally having lower uncertainty than the Combustion-type systems. This result suggests that the lower uncertainty observed in Plasma-type systems may be attributed to the larger number of samples collected. Alternatively, the variability in the characteristics of gases and the abatement technologies themselves could also influence the differences in uncertainty, indicating the need for further research to draw more detailed conclusions.

Additionally, in other industries in Korea, such as industrial facilities or power plants, TMSs (Tele-Monitoring Systems) are installed on individual stacks to automatically monitor flow rates and concentrations, assess facility abnormalities, and control emission concentration standards according to relevant regulations in real time. However, in the semiconductor and display industries, the complexity of the manufacturing facilities and concerns over recipe leakage prevent continuous automated measurement and the disclosure of such information. In this study, field measurements were conducted at industrial sites and for each abatement type using QMS and FT-IR equipment, reflecting Korea’s technological level for the development of DRE standards. However, these instruments are expensive and challenging to manage. It is anticipated that research in this area could be further invigorated if future advancements in related technologies enable simplified measurement processes.

In conclusion, this study provides a foundational analysis for developing DRE standards tailored to the abatement technologies used in the semiconductor and display industries. By measuring the reduction efficiency of abatement equipment currently in operation, this research verifies the differences between the technologies and offers valuable data for selecting and implementing optimal abatement solutions to support sustainable development in the industry. Furthermore, by segmenting DRE by technology type, the study contributes to improving the accuracy of greenhouse gas emissions calculations, enhancing the reliability of emissions inventories, and enabling more precise emissions estimates at both the national and industrial levels. It is anticipated that further research in this area could play a significant role in helping South Korea achieve its carbon neutrality goals.