Effects of System Parameters on the Physical Characteristics of

Thornton Hall, University of Virginia,. Charlottesville, Virginia 22904-4742. Air sparging is a relatively new, cost-effective technology for the reme...
0 downloads 0 Views 72KB Size
Download PDF
Environ. Sci. Technol. 2001, 35, 204-208

Effects of System Parameters on the Physical Characteristics of Bubbles Produced through Air Sparging SUSAN E. BURNS* AND MING ZHANG Department of Civil Engineering, P.O. Box 400742, Thornton Hall, University of Virginia, Charlottesville, Virginia 22904-4742

Air sparging is a relatively new, cost-effective technology for the remediation of soil and groundwater contaminated with volatile organic compounds (VOCs). While the method has met with reasonable success at a large number of field sites, implementation of the technique is restricted to relatively coarse-grained soils with large values of air permeability, which significantly limits its applicability. An understanding of the fundamental parameters that control the formation and distribution of air in the sparging process is useful for optimizing the system implementation and extending its range of applicability. This work presents the results of an experimental investigation into the effect of process control parameters on the size and size distribution of air bubbles produced in aqueous and idealized saturated porous media systems. The experiments used digital image analysis to image and quantify the physical characteristics of the bubbles generated in a bench scale test cell. Results demonstrated that the average bubble size and range of size distribution increased as the injection pressure and size of the injection orifice were increased. Larger diameter bubbles with wider size distributions were produced in the presence of particles when compared to aqueous systems. As the particle size was decreased, the size of bubbles produced increased. Finally, the presence of trace quantities of the surfactant Triton X100 led to uniformly small diameter bubbles under all experimental conditions. The presence of the surfactant coating produced bubbles with physical characteristics that are more suited to in situ stripping of VOCs than the bubbles produced in the absence of a surfactant.

Introduction The formation and transport of gas bubbles through both aqueous and saturated soil systems is of critical importance in a variety of environmental systems, including air stripping and in-situ air sparging. Air sparging is an in-situ stripping process designed to remove volatile organic compounds (VOCs) from the saturated zone. Air is injected below the water table through an injection well with a slotted screen, and the injected air rises to the surface through buoyancy. Ideally, the injected air travels upward as discrete bubbles, acting as a collector for the dissolved VOCs. However, research has shown that channeling of the injected air occurs in soils * Corresponding author phone: (804)924-6370; fax (804)982-2951; e-mail: [email protected]. 204

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 1, 2001

that have a grain size smaller than approximately 4 mm (1, 2). When air channels form in the subsurface, contaminant removal becomes mass transfer limited as the VOCs must diffuse to the air channels for removal (3). This nonuniform distribution of stripping gases through the subsurface is a severe limitation in the implementation of air sparging (4). It is possible to affect the characteristics of the sparging process through alteration of the system and operating variables. In natural systems, the physical characteristics of the soil grain size distribution and pore throat geometry cannot be altered; however, the air injection characteristics and the properties of the interacting fluids, both wetting and nonwetting, can be manipulated. Changes to the injection velocity or the surface tension between the interacting air and water phases can significantly alter the formation and transport behavior of the air bubbles in the soil pore space. In the air sparging process, the rate of air injection is the most easily controlled system parameter. As the injection rate is decreased, the air/water interface will advance through capillary fingering, creating high levels of air saturation. In contrast, at high values of the air injection rate, as are used in air sparging, the air progresses through viscous fingering, which leads to preferential channel flow and relatively low values of air saturation (5). The interfacial tension between the interacting fluids is another parameter that can be controlled, although not quite as easily as the injection velocity. The addition of surfactants at the air injection nozzle during air sparging significantly alters the pattern of airflow in the sparging system by reducing the interfacial tension. Surfactants are surface-active solutes, characterized by a hydrophilic end and a hydrophobic end, that tend to concentrate at the interface of two phases. The most significant effect of the decrease in surface tension is a reduction in the threshold pressure required to overcome capillary effects at the pore throat. The high injection pressures used in air sparging to overcome the capillary barriers at the pore throats forces the advancement of the air into an unstable viscous fingering front; consequently, instability and nonuniform distribution of stripping gas result (6). These limitations can be overcome through a reduction in the surface tension by the addition of surfactants. The ability to form discrete air bubbles that do not readily coalesce in a porous media would significantly enhance the implementation of air sparging. In purely aqueous systems, it has been demonstrated that the addition of surfactants results in a significant decrease in the size of the bubbles produced during the mechanical injection of air (7), and similar trends have been observed at global scales in highly idealized saturated porous media (8). This is advantageous because the production of smaller diameter bubbles for a given gas flowrate results in a large surface area-to-volume ratio that is desirable in contaminant stripping/mass transfer processes. The presence of a surfactant coating will also increase drag on the bubbles (9), which will increase the bubble residence time in the system and will reduce the coalescence of bubbles (10-12), both of which are advantageous in stripping operations. While the presence of a surfactant coating can decrease the rate of mass transfer of contaminants into the vapor phase (7), only a trace concentration of surfactant is required to significantly reduce the interfacial tension. At these low concentrations, surfactants will accumulate at the trailing edge of a bubble translating vertically through an aqueous system, resulting in small diameter bubbles with top caps that are relatively surfactant-free and available for mass transfer (13, 14). Additional studies have shown that mass transfer is more 10.1021/es001157u CCC: $20.00

 2001 American Chemical Society Published on Web 11/15/2000

TABLE 1. Matrix of Bubble Size and Size Distribution Experiments expt no.

injector

media

surfactant

pressure (kPa)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

26S gauge needle 26S gauge needle 26S gauge needle 26S gauge needle glass frit glass frit glass frit 25G gauge needle 26S gauge needle 26S gauge needle 26S gauge needle glass frit glass frit 26S gauge needle 26S gauge needle 26S gauge needle

deionized water deionized water deionized water deionized water deionized water deionized water deionized water deionized water deionized water deionized water/14.5 mm particles deionized water/27.0 mm particles deionized water/14.5 mm particles deionized water/27.0 mm particles deionized water deionized water/14.5 mm particles deionized water/27.0 mm particles

no no no no no no no no no no no no no Triton X100 Triton X100 Triton X100

6.2 8.3 9.0 11.0 8.3 9.0 11.0 4.2 4.2 9.0 9.0 9.0 9.0 9.0 9.0 9.0

strongly controlled by the interfacial surface area than by the presence of a surfactant coating (11, 15), in part because diffusion can take place through the aqueous pores between the surfactant molecules (16).

Study Objectives This investigation quantified the effects of several system parameters on the physical characteristics of the air bubbles produced in a bench scale sparging system. Experiments were performed to quantify the effect of injection pressure, injection orifice, and particle grain size on the size of bubbles produced in a highly idealized porous medium. Additionally, the effect of trace concentrations of the surfactant Triton X100 on the size and size distribution of bubbles was also quantified.

Experimental Methods Materials. All water used in the experimental study was deionized and organic free (Barnstead, Nanopure). The nonionic surfactant, t-octylphenoxypolyethoxyethanol (Triton X100), was obtained from Sigma Aldrich Chemical Co. and was used as received. Two types of air injection orifices were tested in the experiments: a glass frit bubble diffuser (Fisher Scientific, Kimax coarse frit) and two syringe needles (Fisher Scientific, Gauges: 25G and 26S). The 25G needle had an inside diameter of 0.254 mm, and the 26S needle had an inside diameter of 0.114 mm. The bubble diffuser had multiple openings that produced a large number of bubbles simultaneously; in contrast, the syringe needle had a single orifice and produced discrete bubbles in series. All experiments were performed in a rectangular silica test chamber (45 mm by 295 mm by 260 mm) with flat cell walls in order to prevent optical distortion during the imaging experiments. Air was injected into the base of the test cell through Tygon tubing connected to the injection orifice. The injection pressure was controlled using a pressure regulator (Fairchild Industrial Products Company, Model 10). For the experiments performed in simulated soils, highly idealized porous media were used in the study: spherical silica beads (14.5 mm or 27.0 mm diameter). The large diameter beads were chosen for optical clarity in the experimentation. A summary of the conditions in the experiments is given in Table 1. Procedures. Digital image analysis was used to record and analyze the sizes and size distributions of the bubbles produced during the experiments. A complete description of the imaging procedure is given in ref 8, and a description of the sampling methodology is given in ref 17. A set of reference images of known areas was examined in order to

FIGURE 1. Bubble cumulative size distributions as a function of injection pressure. Air was injected through a 26S gauge needle (single orifice). All experiments were performed in deionized water. quantify the error in the experimental measurements; all measured areas were within 2.6% of the reference values. Before each experiment, the test cell was washed and rinsed thoroughly to ensure that no trace surfactant was present during the testing. The chamber was filled with deionized water, and air was injected at the base of the cell through the injection orifice. Images of the bubbles produced were captured and analyzed using digital image analysis; the projected area and diameter of each bubble was determined. The bubble diameters were calculated by assuming that all bubbles were spherical. For the tests in porous media, the particles were packed into the test cell at a porosity of approximately 40% (ranged between 37% and 42%). Three inches of water was left above the packed test cell, and images were taken as the bubbles emerged from the soil column. All bubble images were taken under the same hydrostatic head in order to eliminate size differences due to pressure effects.

Results and Discussion A series of tests were performed in aqueous media to quantify the effect of injection pressure on the size and size distribution of air bubbles produced in deionized water without surfactant. The chamber was filled with deionized water, and air was injected at the base of the cell through the 26S gauge syringe needle at pressures of 6.2 kPa (0.9 psi), 8.3 kPa (1.2 psi), 9.0 kPa (1.3 psi), and 11.0 kPa (1.6 psi). The bubble size and size distributions produced in the single orifice injection regime are shown in Figure 1. The data are presented in terms of the dimensionless Bond number, Bo, which quantifies the ratio of the gravitational forces to the surface tension forces acting in the system VOL. 35, NO. 1, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

205

FIGURE 2. Comparison of bubble cumulative size distributions as a function of injection pressure for single and multiple orifice injectors. Air was injected through a glass frit bubble diffuser (multiple orifices) and a 26S gauge needle (single orifice). All experiments were performed in deionized water.

Bo )

(FL - FG)L2g σ

(1)

where FL ) density of the liquid phase (M/L3), FG ) density of the injected gas (M/L3), L ) characteristic length (bubble diameter) (L), g ) gravitational constant (L/T2), and σ ) surface tension (M/T2). Viscous effects on the formation of bubbles in water are negligible due to the low viscosity of the liquid phase (10). As anticipated, increases in the injection pressure and flow rate clearly led to the production of larger diameter bubbles with larger values of the Bond number. It is important to note that the syringe orifice produced a very uniform distribution of bubbles with a low standard deviation (e0.04 mm), as is characteristic of heterogeneous bubble formation by direct injection of a gas into a liquid through a single orifice. The formation of an air bubble in water is a surface controlled process that is highly reproducible if equipment variables such as orifice diameter and system variables such as surface tension and liquid density are held constant (10). Figure 2 compares the results obtained with the single orifice injector to an experiment performed under the same conditions but using the glass frit bubble diffuser as the injection orifice. The bubble diffuser is composed of sintered silica beads and has multiple openings for the injection of air. Results of experiments performed at injection pressures of 8.3 kPa (1.2 psi), 9.0 kPa (1.3 psi), and 11.0 kPa (1.6 psi) for both the diffuser (multiple openings) and the syringe (single opening) are shown in Figure 2. Several trends are noticeable from the data: average bubble diameter increases as the injection pressure increases for both injection orifices; the size distribution produced by the diffuser becomes significantly more broad as the injection pressure is increased; the size distribution for the diffuser shifts toward larger diameter bubbles as the injection pressure increases. Standard deviations for the diffuser experiments ranged from 0.20 mm to 0.74 mm, which is significantly larger than the single orifice experiments. The effect of the diameter of the injection orifice on the size and size distribution of the bubbles produced was quantified using two different syringe needles {25G (0.254 mm inner diameter) and 26S (0.114 mm inner diameter)}. The chamber was filled with deionized water and air was injected at a pressure of 4.2 kPa (0.6 psi). The results are shown in Figure 3. Even at this very low magnitude of injection pressure, the dominant effect of injection orifice on the characteristics of the bubbles produced is evident. When the diameter of the injection orifice was increased by a factor of 2.2, the average bubble diameter increased by over five times, 206

9

FIGURE 3. Effect of injection orifice diameter on the cumulative size distributions of air bubbles. Air was injection into deionized water through a 26S gauge needle (0.114 mm) and a 25G needle (0.254 mm) at a pressure of 4.2 kPa.

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 1, 2001

FIGURE 4. Comparison of bubble cumulative size distributions resulting from air injection into saturated silica beads. Air was injected through a 26S gauge needle at a pressure of 9.0 kPa. Tests were performed in deionized water and 14.5 and 27.0 mm diameter beads saturated with deionized water. and the size distribution became larger (standard deviation ) 0.16 mm for the 25G needle versus 0.05 mm for the 26S needle). It was also noted during the experimentation that a lower magnitude of pressure was required to initiate bubble formation in the larger diameter orifice due to capillary effects. The effect of the presence of a porous media was quantified using the 26S (0.114 mm) needle as the injection orifice. The test chamber was packed with large silica beads of diameters either 14.5 mm or 27.0 mm. While the silica beads are a highly idealized porous media, the test results indicate useful trends. These tests were performed surfactant free at an injection pressure of 9.0 kPa (1.3 psi). Bubbles were imaged as they emerged into a small headspace of water at the top of the packed column. Visual observation ensured that the bubbles formed discretely at the injection orifice and then transported through the porous media. Figure 4 shows the results for the size distributions produced using single orifice injection in aqueous, 14.5 mm particle, and 27.0 mm particle systems. The average bubble size was larger for both of the columns packed with particulates than for the aqueous system, with an average bubble diameter of 1.54 mm produced in the 14.5 mm beads and 1.40 mm produced in the 27.0 mm beads, compared with an average diameter of 1.32 mm produced in the aqueous system. It is interesting to note that as the grain size is decreased, the average size of the bubbles produced increases, indicating increased coalescence within the porous media, even in particle sizes significantly larger than most natural soil grains.

FIGURE 5. Comparison of bubble cumulative size distributions as a function of particle size and injector. Air was injected through a 26S gauge needle (single orifice) and a glass frit bubble diffuser (multiple orifice) at 9.0 kPa. Tests were performed in deionized water and 14.5 and 27.0 mm diameter beads saturated with deionized water. The size distribution is also significantly affected by the presence of the porous media, with the distributions becoming more broad as the grain size is decreased. The effects are most significant for the largest bubbles produced in the system, with the smallest diameter bubbles showing relatively little change in distribution. Figure 5 compares the results of the single orifice injection regime with the multiorifice diffuser injection. These experiments were performed at an injection pressure of 9.0 kPa (1.3 psi), again using the 14.5 mm and 27.0 mm particles. These data demonstrate that the sizes and size distributions of bubbles produced by the diffuser are more significantly affected by the presence of the porous media. Average bubble diameters produced were 2.48 mm in the 14.5 mm particle test and 2.02 mm in the 27.0 mm particle test, both of which are larger than the average diameter of 1.05 mm produced in the aqueous test. Size distributions produced by the diffuser in the presence of the particles are more broad, with standard deviations of 1.19 mm for the 14.5 mm particles and 0.95 mm for the 27.0 mm particles. Figure 6a shows the results of a series of tests performed in the presence of the surfactant Triton X100, in terms of the Bond number. The addition of the surfactant to the system reduced the surface tension from 73.0 dynes/cm for deionized water to 33.5 dynes/cm for the surfactant/water solution. The experiments were performed using the 26S syringe needle as the injection orifice at a constant pressure of 9.0 kPa (1.3 psi). Tests were performed in the column with aqueous/ surfactant, aqueous/surfactant/14.5 mm particles, and aqueous/surfactant/27.0 mm particles. The presence of the surfactant in the system led to the production of smaller diameter bubbles that did not coalesce; consequently, a uniform size distribution was produced in both the aqueous and particulate experiments performed. In all experiments, the average bubble diameters were significantly reduced (average diameter < 1.0 mm). For the experiments in particulate media, the bubble size distribution was also reduced in the presence of surfactants, indicating reduced coalescence of the bubbles produced in the presence of trace surfactant. Figure 6b shows the data from the same experiment, but plotted in terms of surface area-to-volume ratios. The interfacial area available for mass transfer is of critical importance in any stripping process; consequently, large ratios of bubble surface area to volume of gas are desirable. Comparison between the bubbles generated in the presence of Triton X100 with those produced in deionized water

FIGURE 6. Effect of Triton X100 (surface tension ) 33.5 dynes/cm) on the cumulative size distributions of air bubbles resulting from the injection of air into deionized water and saturated porous media (14.5 mm and 27.0 mm diameters): (a) in terms of bubble size and (b) in terms of the surface area-to-volume ratio. Air was injected through a 26S gauge needle (single orifice) at a pressure of 9.0 kPa. illustrates the significant increase in available surface area, with the surfactant-coated bubbles having between two and three times greater surface area per volume of gas than the surfactant-free bubbles. The reduction in bubble size due to the decrease in interfacial tension also dominated the porous media effect, producing uniformly small bubble diameters even in the presence of particulates. The average bubble size and size distributions were essentially identical regardless of whether the experiment was performed in purely aqueous conditions or in the presence of the silica beads. Clearly, the presence of surfactants will alter the characteristics of the formation and transport of air within a particulate media. Surfactants in the system will also affect the capillary pressure curve, which will allow higher air saturation at the same pressure. As the air saturation increases, the mass removal will increase (18).

Acknowledgments The authors thank James E. Danberg for his assistance with the experimental setup. The authors also thank the anonymous reviewers for their insightful comments. Partial funding for this research was provided by the University of Virginia; this support is gratefully acknowledged.

Literature Cited (1) Ji, W.; Dahmani, M. A.; Ahlfeld, D. P.; Lin, J. D.; Hill, E. Ground Water Monitoring Remediation 1993, 13, 115-126. (2) Semer, R.; Adams, J. A.; Reddy, K. R. Geotechnical Geological Eng. 1998, 16, 59-75. VOL. 35, NO. 1, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

207

(3) Ahlfeld, D. P.; Dahmani, A.; Ji, W. Ground Water Monitoring Remediation 1994, 14, 132-139. (4) Kueper, B. H.; Frind, E. O. J. Contaminant Hydrol. 1988, 2, 95110. (5) Lenormand, R.; Touboul, E.; Zarcone, C. J. Fluid Mechanics 1988, 189, 165-187. (6) Miller, C. A. AIChE J. 1975, 21, 474-479. (7) Clift, R.; Grace, J. R.; Weber, M. E. Bubbles, Drops, and Particles; Academic Press: New York, 1978. (8) Burns, S. E.; Zhang, M. J. Computing Civil Eng. 1999, 13, 43-48. (9) Chhabra, R. P.; Kee, D. D. Hemisphere Publishing Corporation: New York, 1992; p 255. (10) Sadhal, S. S.; Ayyaswamy, P. S.; Chung, J. N. Transport Phenomena with Drops and Bubbles; Springer: New York, 1997. (11) Bischof, F.; Sommerfeld, M.; Durst, F. ASME Fluids Engineering Division: Washington, DC, 1993; pp 131-136. (12) Jeng, J. J.; Jer, R. M.; Yang, Y. M. Ind. Eng. Chem., Process Design Development 1986, 25, 974-978.

208

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 1, 2001

(13) Quintana, G. C. Intl. J. Heat Mass Transfer 1990, 33, 2631-2640. (14) Oguz, H. N.; Sadhal, S. S. J. Fluid Mechanics 1988, 194, 563579. (15) Molder, E.; Tenno, T.; Nigu, P. Crit. Rev. Anal. Chem. 1998, 28, 75-80. (16) Princen, H. M.; Overbeek, J. T. G.; Mason, S. G. J. Colloid Interface Sci. 1967, 24, 125. (17) Burns, S. E.; Yiacoumi, S. Z.; Tsouris, C. Separations Purification Technol. 1997, 11, 221-232. (18) Elder, C. R.; Benson, C. H.; Eykholt, G. R. J. Geotechnical Geoenvironmental Eng. 1999, 125, 947-958.

Received for review April 4, 2000. Revised manuscript received September 28, 2000. Accepted October 2, 2000. ES001157U