Removal of C₃F₈ Via the Combination of Non-Thermal Plasma, Adsorption and Catalysis

Abstract

The feasibility of C₃F₈ abatement via combining nonthermal plasma with adsorption and/or catalysis is investigated in this study. In terms of the simultaneous combination of plasma, adsorption and catalysis (CPAC), three different configurations including A/C layer (adsorbent layer prior to catalyst layer), C/A layer (catalyst layer prior to adsorbent layer) and A/C mixture (adsorbent and catalyst are mechanically mixed) are adopted. For all the experimental tests conducted in this study, the gas stream consists of 500 ppm C₃F₈, 2% O₂, and balanced N₂. The experimental results indicate that C₃F₈ removal efficiencies depend on what kind of packing material is adopted (adsorbent, catalyst or both) and how the material is packed within the plasma reactor. The removal efficiencies obtained with different reactors are in the order as: CPAC (A/C layer; AC mixture) > CPA (plasma with adsorbent alone) > CPC (plasma with catalyst alone) > CPAC (C/A layer). The indentified products after treatment include CO₂, CO, N₂O and CF₄. The formation of C₂F₆ is not observed in this study, which is encouraging since the global warming potential of C₂F₆ is actually higher than that of C₃F₈.

Introduction

With the characteristics of extremely long lifetimes in the atmosphere and relatively high GWPs (global warming potentials), PFCs are regulated as one of the GHGs (greenhouse gases) in Kyoto Protocol. PFCs are mainly emitted from semiconductor and TFT-LCD manufacturing industries. In terms of PFC abatement, the high temperature-based technologies including catalytic oxidation and thermal decomposition are still the mainstream. Although acceptable removal efficiency could be achieved, these equipments have to be maintained at a certain temperature even during the period without PFC emission. Different from the above-mentioned technologies, nonthermal plasma can be ignited and operated at room temperature, making it a potentially viable technique for PFC abatement. However, prior to further industrial application, the energy efficiency and the products’ selectivities still need to be improved [1, 2].

To resolve these challenges, a novel technique combining the nonthermal plasma, adsorption and catalysis (CPAC) simultaneously is developed in this study. It has been experimentally demonstrated in various studies that introducing catalyst into a nonthermal plasma reactor could effectively improve the abatement of gaseous pollutants [3–9]. Moreover, adsorbent is used to increase the concentration and retention time of pollutant molecules within discharge region, which is believed to be beneficial for energy efficiency and products’ selectivities. C₃F₈ is chosen as the target pollutant in this study because it has been gradually replacing CF₄ and C₂F₆ as the CVD chamber cleaning gas for semiconductor industry. The main purpose is to demonstrate the feasibility of PFC abatement via CPAC, which is constructed by introducing the same volumes of adsorbent and catalyst in the arrangements of two separate layers or a single layer (i.e. adsorbent and catalyst are mechanically mixed.). In addition to CPAC, the combinations of plasma with catalysis or adsorption alone, which are termed as CPC and CPA, are used as control as well.

Experimental

Physical Properties of the Catalyst and Adsorbents

The γ-Al₂O₃ and molecular sieves (MS-3A, MS-4A, MS-4A and MS-10A) pellets used as catalyst and adsorbent in this study are commercial products purchased from Alfa Aesae and Merck KGaA, respectively. The γ-Al₂O₃ and molecular sieve pellets are of spherical and cylindrical shape, respectively. To confirm the catalytic activity of γ-Al₂O₃, α-Al₂O₃ pellets are prepared in laboratory by calcinating γ-Al₂O₃ in air at 1,200°C for 3 h. As shown in Fig. 1, the results of XRD analysis demonstrated calcination of γ-Al₂O₃ in air could obtain the α-Al₂O₃ needed for the experiment.

Fig. 1

XRD spectra of γ-Al₂O₃ and α-Al₂O₃. The symbols “α” and “γ” represent crystalline of α-Al₂O₃ and γ-Al₂O₃, respectively

Nitrogen BET measurements were carried out for the materials applied. The results are listed in Table 1. For molecular sieves, the average pore diameter increased as the BET surface areas increased. For instance, the BET surface areas of MS-3A, MS-4A, MS-5A, and MS-10A are 22.50, 290.54, 370.86, and 456.56 m²/g, and the average pore diameters of MS-3A, MS-4A, MS-5A, and MS-10A are 3.52, 4.32, 5.20, and 10.12 Å, respectively. For aluminum oxides, α-Al₂O₃ is nonporous while γ-Al₂O₃ is porous. Therefore, the BET surface area of α-Al₂O₃ was relatively small (c.a. 9.79 m²/g) compared to that of γ-Al₂O₃, and the average pore diameters of α-Al₂O₃ was not analyzed.

Table 1 BET surface area and average pore size of materials

Experimental Tests of C₃F₈ Abatement

The laboratory-scale experimental system used to evaluate the effectiveness of C₃F₈ decomposition is schematically shown in Fig. 2. The experimental apparatus was comprised of a gas-generation system, high-voltage power supply system, a plasma reactor, and sampling detection system. The gas-generation system consisted of several gas cylinders, including C₃F₈/N₂, N₂, and O₂ and was regulated by a set of mass flow meters (MKS-1179A) to control the flow rates of feeding gases. The composition of the gas streams could be accurately controlled by mass flow meters.

Fig. 2

Schematic diagram of the experimental set-up of single-stage system

The high-voltage power supply system consists of a power meter (Chen-Hwa Co., Model 2100, Power-Win Technology) and a high-voltage transformer (Jui-Hsiang PTY Co. Ltd.). For all experiments carried out in this study, the applied frequency was fixed at 200 Hz and the applied voltage ranged from 12 to 18 kV (rms value). The discharge power was determined by the Lissajous diagram (V-Q plot) [10] which was measured by a digital oscilloscope (Tektronix 3014B) equipped with a high voltage probe (Tektronix P6015A) and a voltage probe (Tektronix 613B). A 10 μF capacitor was connected in series with the reactor for the measurement of the charge.

The laboratory-scale plasma reactor consisting of cylindrical tube, inner electrode, and outer electrode is schematically described in Fig. 3. The cylindrical tube was made of quartz with outer and inner diameters of 2.8 and 2.5 cm, respectively. The inner electrode was composed of a stainless steel rod (diameter, 0.3 cm) and was located along the tube’s centerline. The outer electrode, made of stainless steel mesh, was wrapped around the outside of the crystal quartz tube. The discharge length is 10 cm. For the simultaneous combination of plasma, adsorption and catalysis, three different configurations are tested in this study, including A/C (adsorbent layer prior to catalyst layer), C/A (catalyst layer prior to adsorbent layer) and A/C mixture (adsorbent and catalyst are mechanically mixed). The experimental tests were conducted at different temperatures varying from room to high temperature (ca. 25–250°C) and gas streams were kept at ambient pressure in the plasma reactor for all tests. For all the experimental tests for C₃F₈ abatement by plasma (including CPC, CPA and CPAC), the gas stream consists of 500 ppm C₃F₈, 2% O₂, and balanced N₂. O₂ content is fixed at 2% because our experimental results show that C₃F₈ removal efficiency increases with O₂ content up to 2% but further increase in O₂ content would lead to a lower removal efficiency instead.

Fig. 3

Illustrations of the different CPACs tested in this study

Qualitative and quantitative analysis of C₃F₈ and associated end products including CO₂, CO and CF₄ were carried out with FT-IR (Nicolet 6700, Thermo). The C₃F₈ removal efficiency is calculated by:

$$ {\text{C}}_{ 3} {\text{F}}_{ 8} \,{\text{removal}}\,{\text{efficiency}} = \frac{{\left[ {{\text{C}}_{ 3} {\text{F}}_{ 8} } \right]_{\text{inlet}} - \left[ {{\text{C}}_{ 3} {\text{F}}_{ 8} } \right]_{\text{outlet}} }}{{\left[ {{\text{C}}_{ 3} {\text{F}}_{ 8} } \right]_{\text{inlet}} }} \times 100\% $$

(1)

Here, [C₃F₈]_inlet and [C₃F₈]_outlet are the concentration of C₃F₈ at the inlet and the outlet of reactor, respectively. The selectivities of CO₂, CO and CF₄ are determined by:

$$ {\text{CO}}_{ 2} \,{\text{selectivity}} = \frac{{\left[ {{\text{CO}}_{2} } \right]}}{{\left[ {C_{3} F_{8} } \right]_{\text{inlet}} - \left[ {C_{3} F_{8} } \right]_{\text{outlet}} }} \times \frac{1}{3} \times 100\% $$

(2)

$$ {\text{CO}}\,{\text{selectivity}} = \frac{{\left[ {\text{CO}} \right]}}{{\left[ {{\text{C}}_{ 3} {\text{F}}_{ 8} } \right]_{\text{inlet}} - \left[ {{\text{C}}_{ 3} {\text{F}}_{ 8} } \right]_{\text{outlet}} }} \times \frac{1}{3} \times 100\% $$

(3)

$$ {\text{CF}}_{4} \,{\text{selectivity}} = \;\frac{{\left[ {{\text{CF}}_{ 4} } \right]}}{{\left[ {{\text{C}}_{ 3} {\text{F}}_{ 8} } \right]_{\text{inlet}} - \left[ {{\text{C}}_{ 3} {\text{F}}_{ 8} } \right]_{\text{outlet}} }} \times \frac{1}{3} \times 100\% $$

(4)

where [CO₂], [CO] and [CF₄] are the concentrations of CO₂, CO, and CF₄ at outlet of reactor, respectively.

Results and Discussion

Thermal Catalysis Test

The light-off curves for C₃F₈ removal achieved with molecular sieve (MS-10A) and aluminum oxide (γ-Al₂O₃, and α-Al₂O₃) are presented in Fig. 4. The blank test was carried out without packing. Apparently, MS-10A and α-Al₂O₃ do not have significant catalytic activity for C₃F₈ abatement while γ-Al₂O₃ starts showing catalytic activity at 300°C.

Fig. 4

Effect of catalytic activity on C₃F₈ removal efficiency. Gas compositions: C₃F₈ = 500 ppm. O₂ = 2%, and N₂ balance. Gas flow rate = 125 sccm

Adsorption Test

For all adsorption capacity tests, the adsorbent was packed in a glass tube with an inner diameter of 2.5 cm and the packing height was kept at 2.0 cm. The feeding gases contained 500 ppm C₃F₈, 2% O₂, and balanced N₂. The gas flow rate, temperature, and pressure were controlled at 125 sccm, 25°C, and 1 atm, respectively. The adsorption capacity is determined by (5):

$$ {\text{Adsorption capacity }}\left( {{\text{mg C}}_{ 3} {\text{F}}_{8} /{\text{g Adsorbent}}} \right) = \frac{{P \cdot \left( {M.W.} \right) \cdot Q \cdot 10^{ - 3} }}{m \cdot R \cdot T}\int\limits_{0}^{t} {\left( {[{\text{C}}_{ 3} {\text{F}}_{ 8} ]_{\text{i}} - [ {\text{C}}_{ 3} {\text{F}}_{ 8} ]} \right)} dt $$

(5)

where P is the pressure (atm), M.W. is the molecular weight of C₃F₈ (g/mol), Q is the total gas flow rate (SLPM), m is the mass of adsorbent (g), R is the ideal gas constant (atm-L/mol-K), T is the gas temperature (K), t is the time required to reach adsorption saturation (min), and [C₃F₈] _i and [C₃F₈] are the inlet and outlet concentrations of C₃F₈ (ppm), respectively.

The adsorption capacities of molecular sieves (MS-3A, MS-4A, MS-5A, and MS-10A) and aluminum oxides (γ-Al₂O₃ and α-Al₂O₃) for C₃F₈ are presented in Fig. 5. C₃F₈ adsorption capacity is in the order of MS-10A > MS-5A > MS-4A > MS-3A > γ-Al₂O₃ > α-Al₂O₃.

Fig. 5

Comparison of the adsorption capacities of various materials (MS-3A, MS-4A, MS-5A, MS-10A, γ-Al₂O₃, and α-Al₂O₃) for C₃F₈

In general, the BET surface area and average pore diameter are the important parameters in determining adsorption capacity, suggesting that it can be used as an indicator for selecting the appropriate adsorbent. Figure 6a and b shows the relationships between the BET surface area and average pore diameter and the adsorption capacity obtained with all the materials tested in this study. If one just focuses on the results obtained with molecular sieves, it seems that the adsorption capacity is strongly related to the average pore size. For instance, the average pore diameter of MS-3A, MS-4A, MS-5A and MS-10A are 3.52, 4.32, 5.2, 10.12 Å and the corresponding adsorption capacities are 0.92, 0.93, 0.98, 13.90 mg/g, respectively. Since the size of C₃F₈ molecule is 5.450 Å [11], the adsorption capacity should be low when the average pore size is smaller than 5.450 Å, which is consistent with the experimental results. However, it needs to be emphasized that the average pore size seems not the only factor dominating the adsorption capacity. For instance, the average pore size of γ-Al₂O₃ is actually greater than the molecule size of C₃F₈, however, the adsorption capacity for C₃F₈ is still insignificant. Further study on this topic is needed to establish a suitable indicator to provide useful information for the selection of an appropriate adsorbent.

Fig. 6

Relationship between BET surface area (a) and average pore diameter (b) and C₃F₈ adsorption capacity for various materials tested in this study. Experimental conditions are the same as that described in Fig. 5

Comparison of C₃F₈ Abatement Via CPC, CPA and CPACs

γ-Al₂O₃ and MS-10A are selected for the experimental tests for CPC and CPA because of their good adsorption capacity and catalytic activity for C₃F₈, respectively. The dependence of C₃F₈ removal efficiency on the operating temperature of the system combining plasma with catalyst or adsorbent is presented in Fig. 7.

Fig. 7

Effect of operating temperature combined plasma with catalyst or adsorbent on C₃F₈ removal efficiency. Gas compositions: C₃F₈ = 500 ppm. O₂ = 2%, and N₂ balance. Gas flow rate = 50 sccm, discharge power = 20 W

As shown in Fig. 7, the performance of CPA is better than that of CPC when the operating temperature is between 50 and 150°C. The C₃F₈ removal efficiencies of α-Al₂O₃ and CPC were almost the same, suggesting that catalysis is ineffective for C₃F₈ abatement at this temperature window. The reason is mainly attributed to the low activity of catalysts when the temperature is controlled at the range of 50–150°C. On the contrary, CPC could achieve better performance instead as the temperature is raised to 250°C. Higher C₃F₈ removal efficiency achieved with CPC results from the higher activity of catalysts. The C₃F₈ removal efficiency of CPA decreases with increasing operating temperature because higher temperature is unfavorable for physical adsorption.

In order to combine the advantages of the two types of packing materials tested in this study, the co-operation of adsorption and catalysis was studied. MS-10A was mainly applied to enhance the adsorption, whereas γ-Al₂O₃ was used as the catalytically active component. The combination was realized by introducing the same volumes of adsorbent and catalyst in the arrangement of separate layers (A/C layer or C/A layer) or a layer of mechanical mixture (A/C mixture). Here, A/C layer represents the adsorbent is placed upstream of the catalyst while C/A layer stands for the location of catalyst is ahead of adsorbent.

Dependence of C₃F₈ removal efficiency on the operating temperature achieved with CPACs with different configurations is presented in Fig. 8. When the operating temperature is below 150°C, the performance of C₃F₈ abatement is in the order of A/C layer ≈ A/C mixture > C/A layer at a fixed temperature. The purpose of packing adsorbent within the plasma reactor is to increase C₃F₈ concentration and retention time in the discharge zone. In the configuration of C/A layer, C₃F₈ is immediately adsorbed once the gas stream enters the plasma reactor. Therefore, compared with CPAC (C/A layer), it can be expected that the probability of the collision between C₃F₈ and the active species generated in plasma would be higher in CPAC (A/C layer). As for CPAC (A/C mixture), it can be considered that half of the adsorbent is placed in the first of the plasma reactor with the assumption that adsorbent and catalyst are homogeneously mixed. As a result, it is not surprising that CPAC (A/C mixture) achieved better performance than CPAC (C/A layer). Nevertheless, it needs to be noticed the differences between the C₃F₈ removal efficiencies achieved with different CPACs are less significant as the temperature is increased.

Fig. 8

Effect of the packing way (single-component layers and mechanical mixture) on C₃F₈ removal efficiency. Gas compositions: C₃F₈ = 500 ppm. O₂ = 2%, and N₂ balance. Gas flow rate = 50 sccm, discharge power = 20 W

On the other hand, by comparing the results presented in Figs. 7 and 8, it can be concluded that C₃F₈ removal efficiencies obtained with different plasma reactors are in the order as: CPAC (A/C layer; AC mixture) > CPA (plasma with adsorbent alone) > CPC (plasma with catalyst alone) > CPAC (C/A layer), confirming the benefits of the simultaneous combination of adsorption, catalysis and plasma for C₃F₈ removal.

Products Selectivity and Carbon Balance

The products’ selectivities obtained with CPACs (A/C layer and AC mixture) are shown in Fig. 9. The major end products identified by FT-IR include CO₂, CO, N₂O and CF₄. The concentrations of N₂O ranged from 150 to 300 ppm, depending on the reactor configuration and operating temperature. Moreover, it is worth noticing that both NO and NO₂ are not detected in this study; however, the reasons for this interesting phenomenon is still unclear and further study is needed. The carbon balance ranges from 95 to 100%, meaning that most of the carbon from C₃F₈ removal has been transformed to CO₂, CO, and CF₄. Furthermore, it needs to be emphasized that the formation of C₂F₆ is not observed in this study, indicating that the novel technique proposed is promising for C₃F₈ abatement in that the GWP of C₂F₆ is higher than that of C₃F₈.

Fig. 9

Comparison of the selectivities of CO₂, CO, and CF₄ achieved with different CPACs. Gas compositions: C₃F₈ = 500 ppm. O₂ = 2%, and N₂ balance. Gas flow rate = 50 sccm, discharge power = 20 W

Conclusions

Several novel plasma reactors, including CPC, CPA and CPACs (A/C layer, C/A layer and A/C mixture) are experimentally tested for C₃F₈ abatement in this study. The dependences of C₃F₈ removal efficiencies obtained with CPC and CPA on the temperature are completely different. For CPC, increasing temperature is beneficial for C₃F₈ decomposition due to the increased catalytic activity. On the contrary, increasing temperature is unfavorable for C₃F₈ abatement in CPA because higher temperature would result in a lower adsorption capacity. Moreover, the experimental results show that CPAC could achieve better performance than CPC and CPA. However, one must be careful in choosing the right configuration. For example, CPAC (C/A layer) would lead to poorer performance compared with CPC and CPA. On the other hand, it needs to be mentioned that the global warming potential of C₂F₆ is actually higher than that of C₃F₈. Therefore, the plasma treatment would be meaningless if most of the C₃F₈ molecules decomposed are converted into C₂F₆. Fortunately, the generation of C₂F₆ is not observed in this study, suggesting the feasibility of C₃F₈ abatement via CPAC.