Environ. Sci. Technol. 2004, 38, 5766-5772
Efficient Control System for Low-Concentration Inorganic Gases from a Process Vent Stream: Application of Surfactants in Spray and Packed Columns HUNGMIN CHEIN,* SHANKAR G. AGGARWAL, AND HSIN-HSIEN WU Environmental Nanotechnology Research Laboratory, Center for Environmental, Safety and Health Technology Development, Industrial Technology Research Institute, Chutung, Hsinchu, Taiwan 310
Control of low-concentration pollutants from a semiconductor process vent stream using a wet-scrubbing technique is a challenging task to meet Taiwan environmental emission standards. An efficient wet-scrubber is designed on a pilot scale and tested to control low concentration acid and base waste-gas emission. The scrubber system consisted of two columns, i.e., a fine spray column [cutoff diameter (based on volume), Dv(50) ) 15.63 µm; Sauter mean diameter (SMD) ) 7.62 µm], which is especially efficient for NH3 removal as the pH of the spraying liquid is ≈ 7 followed by a packed column with a scrubbing liquid pH ≈ 9.0 mainly for acids removal. It is observed that use of the surfactants in low concentration about 10-4 M and 10-7 M in the spray liquid and in the scrubbing liquid, respectively, remarkably enhances the removal efficiency of the system. A traditional packed column (without the spray column and the surfactant) showed that the removal efficiencies of NH3, HF, and HCl for the inlet concentration range 0.2 to 3 ppm were (n ) 5) 22.6 ( 3.4%, 43.4 ( 5.5%, and 40.4 ( 7.4%, respectively. The overall efficiencies of the proposed system (the spray column and the packed column) in the presence of the surfactant in the spray liquid and in the scrubbing liquid for these three species were found to increase significantly (n ) 5) from 60.3 ( 3.6 to 82.8 ( 6.8%, 59.1 ( 2.7 to 83.4 ( 4.2%, and 56.2 ( 7.3 to 81.0 ( 6.7%, respectively. In this work, development of charge on the gas-liquid interface due to the surfactants has been measured and discussed. It is concluded that the presence of charge on the gas-liquid interface is the responsible factor for enhancement of the removal efficiency (mass-transfer in liquid phase). The effects of the type of surfactants, their chain length, concentration in liquid, etc. on the removal efficiency are discussed. Since the pilot tests were performed under the operating conditions similar to most of the wet-scrubbers operated in semiconductors manufacturing facilities for inorganic pollutants, this study can be applied to modify the existing wet-scrubbers to enhance the removal efficiencies, especially for low-concentration pollutants. * Corresponding author phone: (886)3-5913853; fax: (886)35918753; e-mail:
[email protected]. 5766
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
Introduction Wet scrubbing is the conventional technology used to control waste-gas emission from various industrial processes. Many designs and configurations of wet-scrubbers have been proposed to control inorganic acids and bases, volatile organic compounds (VOCs), odorous components, particulates, etc. from the process vent stream (1). Commonly spray columns with or without packing material are used, especially for the removal of vapors and fine particulates from the waste stream. In the column, the scrubbing liquid is sprayed over the inlet stream so the selective transfer of material (absorption) from a gas to a contacting liquid takes place. Gas absorption involves the diffusion of material from a gas through a gas-liquid interface and ultimate dispersion in the liquid, and this refers the gas phase to liquid-phase masstransfer (2). To recognize the interfacial transport process between the gas- and the liquid-phase, a variety of theoretical models such as film theory (3), penetration theory (4), surface renewal theory (5), and eddy diffusivity theory (6) have been given. These theories have been employed successfully by some workers and discussed the effect of various factors on the mass-transfer in a column. Recently, Vaaraslahti et al. (7, 8) identified the natural charging of droplets during spraying, which may be another factor which influences the mass transfer. Laitinen et al. (9) studied the removal efficiency of wet scrubber after charging the fine particulate of waste-gas stream. However, Pilat et al. (10) reported theoretical calculations and experimental measurements of aerosol particles collected by electrostatic spray, but no study has ever been made on the effect of charge of spray droplets on the gas-droplet absorption or on the removal of gaseous pollutants by the spray column. Polat et al. (11) used the surfactants in spray water and reported that the surfactants can develop the electrostatic charge on spray droplets. Although they did not discuss the gas absorption phenomenon with a charged spray droplet, their findings encourage that the use of a surfactant in the scrubbing liquid could be a better way to enhance the gas absorption by considering its charging characteristics on a droplet surface. In the past decades, some reports have been published with the claim that the surfactant can be useful to enhance the efficiency of the waste-gas scrubbing process. For example, Mir and Zahka (12) reported the removal of solvent vapors by the scrubbing treatment using a surfactant in the scrubbing liquid. Rose and Stockman (13) described the sulfur dioxide scrubbing process using a surfactant to enhance the removal but stated that the exact mechanism by which the surfactants operate to enhance the removal in the scrubbing process is not understood. Similarly, Murray et al. (14) reported that the odorous organic compounds and ammonia could be removed efficiently by scrubbing with a dilute solution of sulfuric acid and surfactant without discussing the mechanism. Some reports are also available in favor of the enhancement in solubility for different compounds in the presence of surfactants, for example, Lo and Lee (15) reported enrichment in the solubility of hydrophobic organic compounds in the artificial fog droplets containing surfactant, and Liu and Chang (16) also reported improvement in the solubility of chlorophenols in the presence of surfactants. Tseng et al. (17) modified the desulfurization sorbents with surfactants showing different effects on the removal of PAHs, BTEX, and heavy metals and exploited the potential of calcium hydroxide sorbents for adsorbing air pollutants other than SO2. Huang and Lee (18-20) examined the enhancement 10.1021/es035241w CCC: $27.50
2004 American Chemical Society Published on Web 09/29/2004
FIGURE 1. Schematic diagram of the experimental setup. in solubility of vapor naphthalene, CO2, and SO2 in the presence of surfactants and suggested that these findings could be employed to develop a new control process for the toxic air pollutants by spray or packed column. In spite of these facts, some workers have also reported the discrimination of the effects of surfactants in gas absorption. Llorens et al. (21, 22) reported a decrease in the absorption of carbon dioxide induced by the presence of surfactants in water, due to the modification of a liquidphase mass-transfer coefficient, but they clearly mentioned that they did not consider other mechanisms, such as foam formation, wetting, etc., which can show positive effects on the mass-transfer. Similarly, Murray and Makota (23) reported that the surfactant significantly reduced mass-transfer rates, which mainly related to the surface tension changes induced by the surfactant in the system. A comprehensive explanation of these contradictory effects is still lacking, and more extensive experimental data are required to prove the exact mechanism. The goal of this study was to design a control system, especially for low-concentration inorganic gaseous pollutants to meet the Taiwan emission regulation (24). An experimental study on a pilot scale to enhance the removal efficiency of a spray and a packed column for low-concentration inorganic basic and acidic pollutants from waste-gas stream with surfactants was carried out. A study has been conducted to identify the effect of surfactants behavior on the removal efficiency of the control system. In this work, we evaluated the charge on the interface due to the presence of surfactants, which is suggested to be the responsible factor for the enhancement in the removal efficiency of the system. However, the enhancement depends on concentration of the surfactants, surface tension of the liquid, and other factors, which have been discussed here.
Methods Experimental Setup. The pilot scale wet-scrubber system shown in Figure 1, consisted of countercurrent sprayand cross-flow packed-columns of identical dimensions (1.2 m × 0.3 m × 0.3 m). A waste-gas stream was generated by impinging the concentrated HF, HCl, and NH3 kept in
different impingers. The flow rate of the air passing through the impingers was fixed according to the desired concentration of a gas in the inlet stream. The waste-gas was entered into the spray column first. In the spray column, a fine watermist was generated with compressed air and RO water both passed through the flow-through autoproportional differential pressure regulator followed by the flow meter and then into the nozzle (type: ultrasonic fog nozzle, model: H2, Best A/V systems, Taiwan), where the atomization took place and provided the finest mist [Sauter mean diameter (SMD) ) 7.62 µm at air pressure (Pa) ) 490.3 kPa (5 kg/cm2) and water pressure (Pw) ) 392.3 kPa (4 kg/cm2) with water flow rate 0.15 L/min, pH ≈ 7]. The water-mist was sprayed over the inlet stream, and the stream was passed through the demister (>0.1 µm droplets can be removed with 50% efficiency) before entering the packed column. In the packed column, the waste-gas was scrubbed with flowing water on packing surfaces (Rasching ring, size ) 50 × 44 mm, surface area ) 155m2/m3, void fraction ) 0.96). The flow rate of the scrubbing liquid (pH ≈ 9) was adjusted using a flow meter and was kept at 18 L/min throughout the measurements. The waste-gas then was demisted and exhausted. The flow rate of the waste-gas stream was controlled by varying the speed of the exhaust fan. In this work, the flow rate of the waste-gas stream was kept at 10 m3/min with the residence time of 0.65 s for each of the columns. The liquid-to-gas ratio was calculated to be 0.015 and 1.8 L/m3 for the spray and the packed columns, respectively. The pH of the mist water and the scrubbing liquid was adjusted by dosing H2SO4/NaOH solutions. pH values and water and air pressures in the spray columns were recorded online. The system program is written using Visual SCADA Run time 98 (version 3.28) software to run the system through the computer (Windows98, PentiumII, 128 MB RAM). Water-Mist Size (Spray Column). The size distribution studies were carried out using insitec (0.5-200 µm measuring range) equipped with a laser beam (wavelength 670 nm; 2.6 mV, He-Ne source lamp) photodetector as described in the product manual (25). The droplet size generated through the nozzle was varied by changing Pa and Pw, and SMD was calculated using eq 1: VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5767
FIGURE 2. Experimental setup for the charge measurement.
TABLE 1. Characteristics of Droplet Size Distribution at Different Conditions
condition
air pressure (Pa), kPa
water pressure (Pw), kPa
volume median diameter [Dv(50)], µm
Sauter mean diameter (SMD), µm
1 2 3
294.2 392.3 490.3
196.1 294.2 392.3
12.34 15.87 15.63
7.08 7.48 7.62
N
∑n d
3
i i
SMD )
i)1
(1)
N
∑n d
2
i i
i)1
where i is the size range considered, ni is the number of droplets in the size range i, di is the median diameter of size range i, and N is the number of size ranges considered. The conditions applied in this work to study the effect of droplet size on removal efficiencies are shown in Table 1. Sampling and Analysis. To test the removal efficiency of the spray and the packed columns separately, the sampling was performed at three points simultaneously, i.e., at the inlet of the waste-gas stream of the spray column (A), at the outlet of the spray column (i.e. inlet of the packed column) (B), and at the outlet of the packed column (C) (Figure 1), using porous metal-disk denuders, which were constructed according to the literature (26), containing a filter pack of a 4.7 cm Teflon filter (Gelman Science, 2-µm) to collect fine particles, followed by two porous metal disks (diameter: 4.7 cm, thickness: 0.23 cm and pore size: 100-µm) coated with 5%, (w/v) sodium carbonate and 4% (w/v) citric acid to absorb acidic and basic gases, respectively. The flow rate was maintained at 2 L/min using a flow meter and a vacuum
pump with a sampling period of 30 min throughout the work. After sampling, the porous metal filter disks were extracted with 15 mL of deionized water, by an ultrasonic bath for 25 min, and the samples were kept in a refrigerator and subsequently analyzed by an ion-chromatograph (Model 4500i, Dionex Corp.). Efficiency Tests. The efficiency of the system was tested for NH3, HF, and HCl as they are the dominating gaseous pollutants in the exhaust of the semiconductor manufacturing process. The inlet concentrations of NH3, HF, and HCl were kept under 3 ppm throughout the tests, as the object of the work is to meet the emission regulations especially for lowconcentration gases. The sampling was performed in different test conditions, and removal efficiencies (RE) were calculated on the basis of inlet (Cin in ppm by volume) and outlet (Cout in ppm by volume) concentrations of a particular species using eq 2.
[
RE% ) 1 -
]
Cout × 100 Cin
(2)
For spray column efficiency calculation, Cin and Cout are the concentrations at (A) and (B), respectively; for packed column efficiency, Cin and Cout are the concentrations at (B) and (C), respectively; and for the overall efficiency calculation, Cin and Cout are the concentrations at (A) and (C), respectively (Figure 1). Surfactants. To study the behavior of a surfactant at the interface, anionic, cationic, and nonionic surfactants specified in Table 2 were tested. Charge Measurement. The charge (current) on the fine droplets in the spray column was measured as recommended in the literature (7) using an electrometer (Pro’sKit, 3PK8200C, Taiwan). The pilot system was connected to ground potential. Metallic net pieces (0.3 m × 0.3 m) were spread off on the surface of three sides in the spray column as shown in Figure 2 to collect the droplets. The potential difference
TABLE 2. Characteristics of Surfactants surfactant cetylpyridinium chloride (CPC) cetyltrimethylammonium bromide (CTAB) sodium dioctyl sulfosuccinate (SDS) sodium lauryl sulfate (SLS) poly(ethylene glycol) tert-octylphenyl ether (TX-100) polyoxyethylene(4) lauryl ether (Brij-30) a
formula
cationic, poorly soluble in water cationic, soluble in water anionic, poorly soluble in water anionic, soluble in water nonionic, soluble in water
10-4
9.0 × 9.2 × 10-4 2.5 × 10-3 8.2 × 10-3 3.3 × 10-4
358.01 364.45 444.55 288.38 646.85
C21H38ClN‚H2O CH3(CH2)15N(CH3)3Br C20H37NaO7S CH3(CH2)11OSO3Na C14H21(OCH2CH2)10OH
nonionic, soluble in water
6.4 × 10-5
362.54
C12H25(OCH2CH2)4OH
CMC ) critical micelle concentration.
5768
9
CMCa at 25 °C, M27
formula weight
characteristic
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
FIGURE 3. Comparison of the removal efficiencies (without surfactants).
FIGURE 4. Overall efficiencies with and without surfactants (background).
was measured at the water feeding point to the nozzle (a) and the point (b) which was the average of potentials measured at c, d, e, and f points on the net. Similarly, Figure 2 shows the arrangement to measure the charge developed on the packing surface in the packed column. The potentials were measured between the point just below the liquid outlet of the nozzle (a′) and the point (b′) which was the average of potentials measured at c′, d′, e′, and f ′ points on the nets arranged specifically within the layers of packing. The current and charge were calculated using eqs 3 and 5, respectively,
I)
V ∆q ) R ∆t
(3)
∆Vliq ∆t
(4)
Qliq ) q)
∆q I ) ∆Vliq Qliq
(5)
where I is the electric current (ampere), V is the potential (volt), R is the resistance (ohm), ∆q is the net charge generated on the droplets or on the surface of wetted packing during time ∆t, Qliq is the volumetric flow rate (liquid volume ∆Vliq passing through the nozzle during time ∆t), and q is the specific charge (coulomb/L). Surface Tension Measurement. A tensiometer (KRU ¨ SS, K8, Germany) was employed using the ring method for the determination of surface tension of surfactants solutions of different concentrations as recommended in the product manual (28). All the measurements were made at 22 °C (1 atm), which is the operating temperature of the pilot system.
Results and Discussion A spray column combined with a packed column was found to be more efficient to control the low-concentration basic and acidic gases other than any individual column, Figure 3. Waste-gas stream from various industrial processes contains low concentration ammonia either with high or low concentration acid constituents, treated with a packed column using an alkaline scrubbing liquid (pH>8). But it is reported that under atmospheric conditions almost all gasphase ammonia dissolves in slightly acidic/neutral cloudwater (29). So the system is designed to control, especially ammonia, by spraying the water-mist of pH nearly neutral over the waste-gas stream. Figure 3 discloses this fact that ammonia can be absorbed into the droplets more efficiently rather than acid constituents. The pH of the mist water in the spray column was kept at 7.2-7.5, whereas that of the scrubbing liquid was 8.5-9.1 in the packed column. The inlet (spray column) concentrations of the pollutants were kept within 0.2-3 ppm for each of the pollutants tested.
FIGURE 5. Effect of inlet concentration on the overall efficiency (the spray column liquid contained 1 × 10-4 M SLS, the packed column liquid contained 5 × 10-6 M CTAB). Enhancement in Removal Efficiency of the System Using Surfactants. Figure 3 shows that the removal efficiencies of a packed column only (without the spray column and the surfactant) for NH3, HF, and HCl were (n ) 5) 22.6 ( 3.4%, 43.4 ( 5.5%, and 40.4 ( 7.4%, respectively, whereas the efficiencies of the proposed system (the spray column with the packed column) without a surfactant for these three species become (n ) 5) 60.3 ( 3.6%, 59.1 ( 2.7%, and 56.2 ( 7.3%, respectively. The use of a surfactant in the mist water and in the scrubbing liquid is found to enhance the removal efficiencies of these species further and finally become (n ) 5) 82.8 ( 6.8%, 83.4 ( 4.2%, and 81.0 ( 6.7%, respectively, Figure 4. Thus on comparison with the packed column only efficiencies, the net enhancement in the efficiencies for these species are (n ) 5) 60.2 ( 8.4%, 40.0 ( 5.1%, and 40.6 ( 10.7%, respectively, especially in very low inlet concentration ranges, i.e., 0.2 to 3.0 ppm. However, the trend of efficiency versus inlet concentrations shows that efficiency decreases as the inlet concentration increases from 2.0 ppm, Figure 5. Effect of Surfactants on the Interface. It has been studied that in an aqueous medium, the hydrocarbon chain (the nonpolar part) of a surfactant molecule shuns interaction with water. This is due to the strong interactions between the water molecules arising from dispersion forces and hydrogen bonding, which act cooperatively to squeeze the nonpolar hydrocarbon chain out of the water (hydrophobic chain), whereas the polar or ionic portion of the molecule (headgroup) interacts strongly with water (hydrophilic head) by ion-dipole or dipole-dipole interactions. So, the surfactant molecule orients itself at the surface of the water such that the polar part or headgroup interacts with the water and the nonpolar part or hydrocarbon chain is held above the surface. This headgroup generates a positive charge (the cationic surfactant) or a negative charge (the anionic VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5769
FIGURE 6. Total charge on droplets [Dv(50) ) 15.63 µm, SMD ) 7.62 µm].
FIGURE 7. Overall charge on the droplets in different conditions, Table 1 (SLS ) 1 × 10-4 M, CTAB ) 0.5 × 10-4 M, TX-100 ) 1 × 10-4 M). surfactant) or polarizes (the nonionic surfactant) the surface of the water. Surfactant Behavior on a Gas-Droplet Interface in a Spray Column. An experiment to measure the charge developed on droplets was conducted to explain the enhancement in the efficiency in the presence of a surfactant. It was observed that the overall charge generated on the droplets is the function of the concentration of the surfactant (Figure 6) as well as the droplets size range used in the experiment (Figure 7). It is clear that the presence of ionic heads at the interface results in a net contribution to the overall charge of the droplets. The variation in overall charge with the concentration of ionic surfactants shows that the overall charge was maximum at the concentration of almost 10-100 times diluted than their critical micelle concentration (CMC) values, which is the concentration of a surfactant at which a micelle (a colloidal aggregate of a surfactant molecule) begins to form. This can be elucidated similar to Polat et al. explanation (11). In the column, as the droplet comes forward, hydrocarbon chains experience a drag force due to the incoming waste-gas stream (the nozzle was placed at the countercurrent direction with the waste-gas stream) and thus oriented on one side of the droplet opposite to the droplet’s path (Marangoni effect). Close packing of ionic surfactant molecules on one side of the droplets makes the droplets similar to a tiny dipole. Thus some electrostatic forces develop on the gas-droplet interface resulting in the enhancement of the polar gas absorption to liquid droplets (the dipole moment for NH3, HF, and HCl is 1.46, 1.98, and 1.03 D, respectively). However, an increment of the absorption, i.e., removal efficiency, depends on the type of the surfactant and the surfactant concentration. Figure 6 shows that with increasing the ionic surfactant concentration, the charge on the droplet increases, which may be due to the accumulation of surfactant molecules as one side of the droplets becomes more pronounced, imparting a stronger dipole interaction. Further increase in the concentration of the surfactant (approaching toward CMC), results in the 5770
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
molecules gradually occupying all the available sites on the droplet surface, and at last total coverage is reached so the dipole structure of the droplet disappears. At this stage, surfactant molecules are uniformly distributed on the droplet surface. This leads to a decrease in the apparent charge of the droplet due to neutralization of the ionic character of the surfactant molecules (Figure 6). The charge on the droplet shows a decline after a certain concentration value where the dipole effect is most emphasized and reaches a minimum around the total surface coverage concentration. The droplet size-range with Dv(50) ) 15.63 µm (condition 3 in Table 1) showed the highest overall charge generated in the presence of the ionic surfactant as compared to size-ranges Dv(50) ) 12.34 µm and 15.87 µm, Figure 7. This may be due to the droplets with Dv(50) ) 15.63 µm which hold appropriate molecules of the surfactant to be oriented to themselves for generating maximum charge density. So, condition 3, i.e., Pa ) 490.3 kPa (5 kg/cm2) and Pw ) 392.3 kPa (4 kg/cm2), was used in this study. Results of the experiment show (Figure 6) that anionic surfactants imparted the highest magnitude of overall charge in comparison with cationic surfactants which can be explained by the inductive effect of the chain. The electron releasing tendency of the hydrocarbon chain accumulates a negative charge on the headgroup of the anionic surfactant resulting in the increased strength of the dipole, whereas in the case of the cationic surfactant, this effect acts reversibly and minimizes the magnitude of the positive charge. Evidently the magnitude of the negative charge also depends on the structure of the hydrophobic tail, i.e., the straight or branched chain. On comparing the charge-generating behavior of anionic surfactants (Figure 6), i.e., sodium lauryl sulfate (SLS) and sodium dioctyl sulfosuccinate (SDS), it is predicted that due to straight-chain structure, SLS molecules can be more closely packed on one side of the droplet rather than branch-chain structured SDS. Since the ionic head of both the surfactants is similar in character, the charge density per area of droplet should be higher for densely packed SLS molecules, and resulting droplets with SLS show a higher overall charge than SDS. On comparison between cationic surfactants (Figure 6), it was observed that cetyltrimethylammonium bromide (CTAB) [headgroup: trimethylamine, (CH3)3N+] imparted a higher overall charge than cetylpyridinium chloride (CPC) (headgroup: pyridine, C6H5N+). Although their chain characteristic is the same, but they differ by headgroup character, hence the charge density per area of droplet is higher for CTAB molecules. Nonionic surfactants [poly(ethylene glycol) tert-octylphenyl ether (TX-100), polyoxyethylene (4) lauryl ether (Brij-30)] failed to generate an appreciable charge and exhibited a slightly increased trend within the concentration ranges tested, regardless of the CMC concentration (Figure 6). It can be explained by the interaction behavior of a highly polar ethylene oxide group (CH2-CH2-O) with water. In water the ethylene oxide group of TX-100 and Brij-30 appeared as dipole with CH2 as a positive and O as a negative pole and dissolves in water via H-bonding forms with O (negative pole). So the overall charge on the group remains positive and increases steadily as the concentration increases. Anionic surfactant SLS was used in a spray column targeted mainly to remove ammonia from the waste-gas. A known amount of SLS was mixed with a spraying liquid (nearly neutral pH), whereas a scrubbing liquid (alkaline pH) was used without a surfactant. Figure 8 shows that the overall removal efficiency for NH3 increases as the SLS concentration in the spraying liquid increases up to a certain concentration which imparted the maximum charge on the droplets (Figure 6), unlike HF and HCl removal. This may be explained as SLS molecules [exist as monomers (30)] generate a negative charge on the droplets, and according to the two-film theory,
FIGURE 8. Effect of SLS concentration on the overall removal efficiency, the inlet concentration of the gas pollutants was within 0.2-3 ppm (SLS mixed with the spray column liquid, whereas the packed column liquid did not contain surfactant).
FIGURE 10. Surface tension (σ) of surfactant solutions at 22 °C (σ of the scrubbing liquid without the surfactant was 71.8 ( 0.33 mN/m).
FIGURE 9. Correlation between the overall removal efficiencies of NH3, HF, and HCl and the charge generated on the droplets due to different concentrations of SLS (e1 × 10-4 M) in the spray column (SLS mixed with the spray column liquid, whereas the packed column liquid did not contain surfactant).
FIGURE 11. Overall charge measured on the interface in the packed column (the charge measured with the scrubbing liquid was 31 ( 2.5 nC/L).
ammonia is present as NH3‚H2O at the interface of gas-film and liquid-film and then as NH4+ at liquid-film [ammonia absorption (29, 2)], which travels faster due to the ion-ion interaction and gets neutralized with a charged head (OSO3-) and further goes to the bulk liquid [some other SLS molecules can be free monomers suspended inside the droplet (15)]. So the absorption of NH3 is supposed to be increased as compared to HF and HCl gas in the same duration. Figure 9 shows the correlation between the overall removal efficiencies of NH3, HF, and HCl and the charge generated on the droplets when SLS (e1 × 10-4 M) was mixed with a spray column liquid. A positive correlation was observed in the case of NH3 (R2 ) 0.961), whereas HF and HCl removal efficiency was found to be decreased with R2 ) 0.785 and R2 ) 0.702, respectively. The result of the optimization of concentration for the efficiency shows that 0.0001 M SLS was appropriate to get the maximum overall removal efficiency for NH3, Figure 8. This is the concentration at which SLS imparted the highest overall charge on the droplets. Surfactant Behavior on a Gas-Liquid Interface in a Packed Column. It is reported that the surfactants exert two major effects, which modify the gas absorption process: the hydrodynamic effect and the interfacial effect. The hydrodynamic effect modifies interfacial tension and thus interfacial pressure, which is responsible for mass-transfer. The interfacial resistance effect also called the barrier effect is due to the fact that the surfactant concentrates at the interface form a monolayer, which offers a resistance to the gas molecules. Surfactants reduce the surface tension of the liquid, which is the cause of the reduction of liquid holdup in packings, and thus reported to adversely effect the gas-liquid mass transfer (31, 32). The quantitative relationship of surfactant effects on gas-liquid mass transfer in the column has been
established and reported by some workers (21, 31-34). However, in this study it was observed that use of the surfactant in a very dilute concentration in the scrubbing liquid can increase the gas absorption significantly. This can be explained on the basis of results of surface tension experiments, which show that very dilute solutions of surfactants (1 × 10-7 to 5 × 10-6 M) do not appreciably change the surface tension of water (Figure 10), whereas charge measurements results show that within these concentration ranges they develop an electrostatic charge on the interfacial surface in a packed column (Figure 11). This discloses that the surfactant present in very low concentrations in a scrubbing liquid should not change the hydrodynamics of the liquid, i.e., the surface renewal rate, the ripples appearing rate on the liquid film, etc. Moreover, the present experimental system was operated with a waste-gas flow rate ) 10 m3/min and a circulating water flow rate per cross section area ) 50 L/min/m2, i.e., in the turbulence region, which is the operating condition of the control systems mostly in practices. So the interfacial turbulence was very high [the Reynolds number for waste gas in a duct (a packed column without packing) was calculated to be 46920], the interfacial renewal was very fast, and the time the liquid constituents remain in the interface was very short (22). In this case the surfactant did not have enough time to be oriented sufficiently and made a monolayer on the surface. Thus the interfacial or barrier effect is no longer applicable with a low concentration of surfactant and high flow conditions. Consequently in such conditions the enhancement in the removal efficiency is due to the interfacial electrostatic charge exerted by the surfactant. It was observed that CTAB generated a higher overall charge as compared with other cationic surfactant (CPC) tested in this study and hence was used in a packed column targeted to control mainly acidic gases. Optimization of the CTAB concentration results show that maximum VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5771
of surfactants in low concentration, and reproducibility of the results imply that the water-mist system accompanied with a surfactant dosing system can be applied for modification of existing wet-scrubbers of semiconductor manufacturing facilities operating with certain conditions to reduce the emission amount of inorganic pollutants.
Acknowledgments Authors are thankful to the Ministry of Economic Affairs (MOEA), Taiwan, R.O.C., for providing partial funds for this study (Contract 92-EC-17-A-10-R7-0523). FIGURE 12. Effect of CTAB concentration on the overall removal efficiency, the inlet concentration of the gas pollutants was within 0.2-3 ppm (CTAB mixed with the packed column liquid, whereas the spray column liquid did not contain surfactant).
FIGURE 13. Correlation between overall removal efficiencies of HF, HCl, and NH3 and the charge generated on the surface of the packings due to different concentrations of CTAB (1 × 10-7 to 5 × 10-6 M) in the packed column (CTAB mixed with the packed column liquid, whereas the spray column liquid did not contain the surfactant). removal efficiency was found at a concentration ranging from 1 × 10-7 to 5 × 10-6 M (Figure 12), where the surface tension values are almost similar to water and the overall charge measured was maximum. A further increase in concentration shows a sharp decrease in efficiency due to the domination of the hydrodynamic effect (the surface tension found to be decreased). Figure 13 shows the correlation between the overall removal efficiencies of HF, HCl, and NH3 and the charge generated on the surface of the packings when CTAB was present in 5 × 10-6 M in a packed column liquid. A positive trend was observed in the case of HF and HCl with R2 ) 0.954 and 0.989, respectively, while a negative correlation was observed for NH3 (R2 ) 0.988). This can also be explained as the NH3 absorption discussed in the previous section referring to a two-film theory. At liquid film, the F-/Cl- ion travels faster and gets neutralized with a charged surfactant [(CH3)3N+] and finally goes to the bulk liquid. Thus the absorption of HF and HCl is supposed to be increased. Figure 5 shows the trend of the efficiency change with the inlet concentration of the pollutants, and this may be explained as the charged surfactant on the interface gets neutralized with a certain concentration of the ions (NH4+/ F-/Cl-), the charge effect is no longer effective for the excess ions if any at liquid film, thus the removal efficiency gradually decreases as the concentration of the pollutant increases after ≈2 ppm. Reproducibility of the System Results. The system was run for 2 h continuously (without blow-down the scrubbing water) with reproducible efficiencies (n ) 3, RSD
