DIGITAL IMAGE ANALYSIS TO ASSESS MICROBUBBLE BEHAVIOR
`IN POROUS MEDIA
`
`By S. E. Burns,1 Associate Member, ASCE, and M. Zhang,2 Student Member, ASCE
`
`ABSTRACT: Air sparging is a commonly implemented technology for the remediation of volatile organic com-
`pounds from contaminated soil and ground water. In the sparging process, air is pressurized into the soil/ground-
`water matrix through injection wells. The air then travels to the ground surface through buoyancy, acting as a
`collector for volatile chemicals. To date, the design and implementation of air sparging has been largely empir-
`ical, based on the results of pilot studies. This paper uses digital image analysis to examine the transport and
`coalescence behavior of microbubbles in porous media, one of the most important control parameters for con-
`taminant removal in air sparging. This laboratory study compared the diameter of bubbles produced in aqueous
`systems with the diameters produced in uniform spherical particulate media (diameters of 14.5 mm and 27.0
`mm) and in elliptically shaped particulate media (equivalent spherical diameters of 14.5 mm). Results showed
`that the presence of a particulate media increased the average diameter and also increased the range of diameters
`of bubbles produced during sparging. As the diameter of the particulate media increased, the size of the bubbles
`decreased, indicating less coalescence in media with larger pore space. In addition, the effect of trace concen-
`trations of surface-active agents (surfactants) on the diameter and coalescence behavior of bubbles was examined.
`In both aqueous and aqueous/particulate matrices, the presence of surfactants significantly decreased the average
`diameter of the bubbles produced. Additionally, the degree of coalescence decreased in the surfactant systems,
`producing a very narrow range of bubble diameters in both aqueous and aqueous/particulate media.
`
`INTRODUCTION
`Air sparging is rapidly becoming one of the most commonly
`implemented remediation methods for the removal of volatile
`organic compounds (VOCs)
`from contaminated soil and
`ground water. The sparging process is essentially a stripping
`technique that mobilizes contaminants to the vapor phase
`through mass transfer into air bubbles. The process also stim-
`ulates biological degradation of contaminants through the in-
`troduction of oxygen into the ground water, a process known
`as biosparging. In the sparging process, air is pressurized into
`ground water through an injection well with porous openings.
`The gas travels through the soil/ground water in the form of
`either discrete bubbles or channels of air. Contaminants trans-
`fer into the vapor phase due to the concentration gradient and
`migrate to the surface through buoyancy. The VOC-laden va-
`por is then collected at the soil surface using soil vapor ex-
`traction (SVE), and treated for the removal of organics. While
`air sparging is being implemented at many field sites it is still
`a relatively new method, and implementation of the process
`to date has largely been empirical, based on the results of pilot
`studies (Marley et al. 1992). The initial implementation of air
`sparging systems has proven that the technique is a viable one;
`however, a more fundamental understanding of the parameters
`that control remedial performance is required in order to pro-
`vide more efficient and cost-effective designs. One of the most
`important aspects of the sparging process is that of the char-
`acteristics of the stripping gas introduced into the contami-
`nated media. This paper uses digital image analysis to quantify
`the effect of a porous medium on the growth and transport of
`mechanically produced bubbles; it also analyzes the effect of
`surfactants on the behavior of microbubbles.
`Fundamental to the success of air sparging is the delivery
`of gas bubbles for volatilization, with the objective being the
`
`1Asst. Prof., Dept. of Civ. Engrg., Thornton Hall, Univ. of Virginia,
`Charlottesville, VA 22903-2442. E-mail:sburns@virginia.edu
`2Grad. Res. Asst., Dept. of Civ. Engrg., Thornton Hall, Univ. of Vir-
`ginia, Charlottesville, VA.
`Note. Discussion open until June 1, 1999. To extend the closing date
`one month, a written request must be filed with the ASCE Manager of
`Journals. The manuscript for this paper was submitted for review and
`possible publication on October 13, 1998. This paper is part of the Jour-
`nal of Computing in Civil Engineering, Vol. 13, No. 1, January, 1999.
`qASCE, ISSN 0877-3801/99/0001-0043 – 0048/$8.00 1 $.50 per page.
`Paper No. 19425.
`
`mass transfer of VOCs to the gaseous phase. Ideally, imple-
`mentation of the process is achieved through a uniform dis-
`tribution of gas bubbles throughout the subsurface; however,
`experimental evidence suggests that significant channeling of
`gas flow occurs when a gas is sparged into a porous media,
`especially when low permeability lenses are present (Baker et
`al. 1995; Leeson et al. 1995). Current methods of bubble gen-
`eration in sparging processes involve the mechanical pressur-
`ization of air, or another gas, into the contaminated ground
`water through a porous well screen. However, it is well doc-
`umented in the literature that the size of the bubbles produced
`during the mechanical sparging of air is large when compared
`with other methods of bubble generation (Burns et al. 1997).
`Large bubbles are a disadvantage in stripping operations be-
`cause the low surface area-to-volume ratios limit mass transfer
`into and out of the vapor phase. In air sparging applications,
`the smaller microbubbles are advantageous in the remedial
`process because small diameter bubbles have high surface
`area-to-volume ratios that facilitate mass transfer, are less
`buoyant, and have a longer residence in the system—having
`dimensions approximating those of the porous media, which
`facilitates transport through the pore network.
`Alternately, small amounts of a surface-active agent (sur-
`factant) can be used to produce smaller diameter gas bubbles
`in a porous medium. Surfactants are soaps, detergents, or long-
`chain alcohols that tend to accumulate at the interface between
`different phases. Because they can increase the apparent water
`solubility of nonaqueous phase liquids (NAPLs) (Kile and
`Chiou 1989; Haulbrook et al. 1993), surfactants are being
`studied for use in the remediation of contaminated ground wa-
`ter (Smith et al. 1997). This study examines the effect of trace
`concentrations of surfactant on the characteristics of bubbles
`produced in porous media. A more fundamental understanding
`of the removal mechanisms active in the sparging process re-
`quires knowledge of the behavior of gaseous bubbles in a po-
`rous medium. A variety of studies, both qualitative and quan-
`titative, have been performed to assess bubble flow behavior
`and bubble terminal velocity.
`Wehrle (1990) performed qualitative studies on the move-
`ment of air through gravels and coarse-grained sands. The re-
`sults showed that low permeability soils can develop pores that
`are filled with air essentially continuously, and that stratified
`soils can lead to significant lateral migration of airflow. Ji et
`al. (1993) also performed a qualitative study of airflow in a
`
`JOURNAL OF COMPUTING IN CIVIL ENGINEERING / JANUARY 1999 / 43
`
`J. Comput. Civ. Eng., 1999, 13(1): 43-48
`
`Downloaded from ascelibrary.org by Janelle Beitz on 12/10/20. Copyright ASCE. For personal use only; all rights reserved.
`
`Tennant Company Exhibit 1034
`
`

`

`simulated porous medium. Spherical silica beads, ranging in
`diameter from 4.0 mm to 0.2 mm, were used as the model
`porous medium in a thin, two-dimensional Plexiglas test cham-
`ber. Air was injected at the bottom of the test chamber and
`the flow pattern of the gas was monitored using a video re-
`corder. All experimental simulations were illuminated from the
`back of the chamber. Three simulations were performed with
`uniform packings of beads with diameters of 4.0 mm, 2.0 mm,
`and 0.75 mm; two simulations were performed with a mixture
`of bead sizes; and one simulation was performed with layered
`bead sizes in order to simulate low permeability zones. Ji et
`al. (1993) found that for uniform grain sizes with bubble di-
`ameters of approximately one to three bead diameters, air flow
`occurred as discrete bubbles; however, for smaller bead di-
`ameters, the flow transitioned from flow of bubbles to chan-
`nelized flow through the porous medium. Additionally, the
`mixtures of bead sizes and layers of small diameter beads led
`to nonuniform flow behavior and stagnant zones that were not
`exposed to airflow.
`Roosevelt and Corapcioglu (1998) used image analysis to
`measure the terminal velocity of air bubbles injected into a
`glass column randomly packed with 4.0 mm diameter glass
`beads. A single air bubble was injected into the saturated po-
`rous medium and the bubble movement up the length of the
`column was recorded using a camcorder. All experiments were
`performed using a backlit system, and the time of translation
`of the center of the bubble was monitored using a timer with
`an accuracy of 0.1 s. Image processing was used to back-
`ground subtract the glass beads from successive video images.
`This method left an image of the moving bubble and allowed
`monitoring of the translation of the bubble center as a function
`of time. The rate of vertical rise of a bubble in a porous me-
`dium was found to be linearly dependent on time. The writers
`found that horizontal migration did occur but had a negligible
`effect on the vertical rise velocity. Additional results indicated
`that the rise time through a large-grained porous medium was
`approximately 20–25% slower than rise time through an un-
`bounded aqueous medium, and bubble rise velocity through a
`capillary tube decreased significantly as the bubble diameter
`approached that of the tube. This is significant because bubble
`flow through porous media has frequently been modeled as
`flow through a bundle of capillary tubes (Rimmer et al. 1996).
`Burns et al. (1997, 1998) used image analysis to measure
`the size of bubbles produced in aqueous media using mechan-
`ical and electrochemical generation techniques. The bubble
`images were videotaped using long-distance microscopy in
`combination with a video cassette recorder. Images were
`printed and imported into an image analysis system for deter-
`mination of their equivalent circular diameter. The studies
`quantified differences in the bubble size, size distribution, and
`power consumption for bubbles generated using the different
`methods.
`Other nonoptical imaging methods applied to monitor the
`transport of microbubbles through porous media include X-ray
`tomography (Chen et al. 1996) and electrical resistance to-
`mography (Schima et al. 1994; Lundegard and LaBreque
`1995; Bruell et al. 1997). Chen et al. (1996) used an X-ray
`computerized tomography (CT) medical scanner to image
`three-dimensional airflow patterns in uniform Ottawa sand soil
`deposits which allowed three-dimensional imaging of porosity
`and fluid saturations. The experimental results showed that in
`medium- to coarse-grained sands relatively few air channels
`are formed, and higher air injection rates enlarge and increase
`the number of channels, which increases air saturation. How-
`ever, larger and more uniform air saturations were measured
`in lower permeability soils, regardless of the air-injection sce-
`nario. Electrical resistivity tomography has also been used to
`monitor the distribution of airflow throughout the subsurface
`
`44 / JOURNAL OF COMPUTING IN CIVIL ENGINEERING / JANUARY 1999
`
`(Schima et al. 1994; Lundegard and LaBrecque 1995). Resis-
`tivity tomography applies electrical current across a set of
`electrodes and measures the voltage drop between the two
`points. An array of electrodes can be set up to measure the
`voltage drop in three dimensions, and the resistivity of the
`material through which the current flows can be calculated
`using inversion techniques. Because the resistivity of the me-
`dium is significantly affected by the concentration of air in the
`pore space, the method yields a global measurement of air
`saturation. The outer limit of influence of the sparged air is
`easily identifiable using resistivity techniques; consequently,
`the method has been applied for the delineation of radius of
`influence during sparging applications (Bruell et al. 1997).
`Several imaging studies have used etched glass to simulate
`the pore distributions present in porous media. Li and Yortsos
`(1995a, 1995b) used etched glass to visualize bubble growth
`during pressure depletion from a supersaturated solution. The
`glass model was constructed with pore sizes in the range of
`600–1200 mm and throat sizes of 100–600 mm (Satik and
`Yortsos 1996). The results showed that gas nucleated and clus-
`tered in a highly irregular manner in the large pore spaces,
`with a slow pressurization step followed by a rapid pore pen-
`etration step after the capillary barrier at the pore throat was
`overcome.
`Because the growth and transport of gas bubbles within po-
`rous media is a complex phenomenon, study of the funda-
`mental aspects of their behavior is critical to a more detailed
`understanding of the mechanisms that control the effectiveness
`of air sparging. Unlike previous studies, this paper presents
`quantitative results on bubble growth and coalescence behav-
`ior. The laboratory study addresses the effect of particle grain
`size on bubble growth and coalescence during transport
`through a porous medium; it also quantifies the effect of sur-
`factants on the characteristics and transport behavior of bub-
`bles produced in both aqueous solutions and saturated porous
`media.
`
`MATERIALS AND METHODS
`
`The laboratory experiments in this study were performed in
`a rectangular test cell constructed out of Pyrex glass (dimen-
`sions of 45 mm 3 295 mm 3 260 mm). The test cell had flat
`glass walls in order to allow visualization of the generated
`microbubbles without distortion of the images. The top panel
`of the test cell was removable in order to allow access to the
`cell for cleaning and for placement of the porous media (Fig.
`1). A bubble diffuser (Kimax coarse frit, Fisher Scientific) was
`placed on the bottom of the test cell and connected to a pres-
`sure regulator (Model 10, Fairchild Industrial Products Com-
`pany) using Tygon tubing. In all experiments, air pressure was
`regulated using a pressure control panel (Trautwein Soil Test-
`ing Equipment) with a digital readout accurate to within
`60.25%.
`
`FIG. 1. Test Cell Configuration
`
`J. Comput. Civ. Eng., 1999, 13(1): 43-48
`
`Downloaded from ascelibrary.org by Janelle Beitz on 12/10/20. Copyright ASCE. For personal use only; all rights reserved.
`
`

`

`FIG. 2. Bubble Images: (a) Gray-Scale Image; (b) Binary Image
`
`images used for measurement are shown in Figure 2(b). All
`image processing was performed using Scion Image for Win-
`dows (Release Beta 2, Scion Corporation). In order to reduce
`bias in sampling, images were snapped at random intervals,
`and only clearly focused images were used for processing. The
`gray-scale images were converted to binary images and the
`cross-sectional area of the bubbles was measured. All areas
`were converted to an equivalent circular diameter to provide
`uniform comparison among experiments.
`At least one hundred bubbles were analyzed in each exper-
`iment in order to obtain representative sample populations.
`Sample size was determined according to (Hines and Mont-
`gomery 1990):
`
`Z sa/2
`E
`= confidence interval; s = standard
`where n = sample size;
`Za/2
`deviation; and E = error. Table 1 gives a detailed list of the
`experimental conditions along with the mean bubble diameter
`and standard deviation for the experiments.
`
`S
`
`n =
`
`D
`
`2
`
`(1)
`
`A series of experiments was performed in a deionized water
`medium, and another series of experiments was performed in
`a deionized water/silica bead media. In the aqueous experi-
`ments, the diffuser was placed in the test cell and submersed
`in fluid, while in the simulated soil experiments the diffuser
`was placed in the test cell and the silica beads carefully packed
`around it. In all packing procedures, the silica beads were wa-
`ter-wetted and placed in a small column of water in order to
`avoid trapped air in the void space. All specimens were packed
`to a porosity of approximately 40%.
`Three types of silica beads were used in the study: uniform
`spherical beads with a diameter of 14.5 mm; uniform spherical
`beads with a diameter of 27.0 mm; and uniform elliptical
`beads with an equivalent spherical diameter of 14.5 mm (di-
`ameter to width ratio of 2:1). After placement of the silica
`beads, an additional 64 mm of aqueous solution was left above
`the porous media in order to allow visualization of the air
`bubbles as they emerged from the simulated soil column.
`A series of experiments was performed in order to quantify
`the effect of surface-active agents on bubbles generated in
`deionized water and in deionized water and porous media. The
`surfactant used in the study was Triton X-100 (t-octylphen-
`oxypolyethoxyethanol, Sigma Chemical Company), a nonionic
`surfactant. In all the experiments, only trace amounts of sur-
`factant were added to the deionized water yielding a surfactant
`concentration of approximately 80 to 100 mg/L.
`Images of bubbles produced during the experiments were
`captured using a long-distance microscope (model QM1,
`Questar) configured for magnification of approximately 40
`times, in combination with a color CCD camera (model VCC-
`3972, Sanyo). The shutter speed on the camera was set to
`1,000 frames per second. Video output from the camera was
`connected directly to an image acquisition board (model PCI-
`1408, National Instruments) installed in a Pentium II com-
`puter. A fiber-optic light was used to illuminate the experi-
`ments from behind the test cell. In this configuration, the fluid
`transmitted light but the bubbles did not; this provided images
`of dark bubbles on a light background. In order to avoid dif-
`ferences in bubble size due to variations in hydrostatic pres-
`sure, images of the bubbles were taken at the same height in
`the test cell for all experiments. In the experiments with the
`porous media, this corresponded to the height at which the
`bubbles emerged from the beads. The translation of the mobile
`bubbles in and out of the focal plane was a consideration in
`obtaining high-quality images. In all experiments the focus of
`the long-distance microscope was fixed, and bubble images
`were acquired semicontinuously, using only clear images of
`bubbles within the focal plane for analysis. Calibration of the
`system was performed by capturing an image of a wire of
`known diameter.
`Sample gray-scale images of the bubbles produced are
`shown in Fig. 2(a), and the accompanying processed binary
`
`RESULTS
`
`The first series of experiments measured the equivalent cir-
`cular diameter of bubbles produced by mechanically pressur-
`izing air through a bubble diffuser into deionized water. The
`experiments were performed in order to quantify the effect of
`injection pressure on the size of individual bubbles formed in
`the system. During this series of experiments, care was taken
`to ensure that no surfactant was present in the test cell. The
`cumulative number distribution as a function of equivalent cir-
`cular diameter for the bubbles produced using injection pres-
`sures of 8.3 kPa (1.2 psi), 9.0 kPa (1.3 psi), 9.7 kPa (1.4 psi),
`and 11.0 kPa (1.6 psi) is shown in Fig. 3. The two distributions
`for the lowest pressures (8.3 and 9.0 kPa) are essentially iden-
`tical with similar means (0.91 mm versus 1.05 mm), similar
`standard deviations (0.22 mm and 0.20 mm), and similar
`ranges of size (0.3 mm–2.0 mm). However, as the injection
`pressure is increased, three trends can be noted from the data:
`the average bubble diameter increases; the bubble size distri-
`bution becomes more broad with larger bubbles being pro-
`
`TABLE 1. Summary of Experiments
`
`Experiment
`number
`(1)
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`
`Pressure
`(kPa) (psi)
`(2)
`8.3 (1.2)
`9.0 (1.3)
`9.7 (1.4)
`11.0 (1.6)
`9.0 (1.3)
`9.0 (1.3)
`9.0 (1.3)
`9.7 (1.4)
`9.0 (1.3)
`9.7 (1.4)
`9.0 (1.3)
`9.7 (1.4)
`
`Particle size
`(mm)
`(3)
`not applicable
`not applicable
`not applicable
`not applicable
`14.5
`27.0
`Elliptical
`not applicable
`Elliptical
`14.5
`14.5
`14.5
`
`Surfactant concentration
`(mg/L)
`(4)
`not applicable
`not applicable
`not applicable
`not applicable
`not applicable
`not applicable
`not applicable
`100
`80
`100
`80
`not applicable
`
`Mean bubble diameter
`(mm)
`(5)
`0.91
`1.05
`1.37
`2.98
`2.48
`2.02
`2.07
`0.39
`0.55
`0.88
`0.71
`1.63
`
`Standard deviation
`(mm)
`(6)
`0.22
`0.20
`0.61
`0.74
`1.19
`0.95
`0.66
`0.08
`0.17
`0.23
`0.21
`0.80
`
`JOURNAL OF COMPUTING IN CIVIL ENGINEERING / JANUARY 1999 / 45
`
` J. Comput. Civ. Eng., 1999, 13(1): 43-48
`
`Downloaded from ascelibrary.org by Janelle Beitz on 12/10/20. Copyright ASCE. For personal use only; all rights reserved.
`
`

`

`FIG. 3. Cumulative Size Distribution of Bubbles as Function of
`Injection Pressure
`
`FIG. 4. Cumulative Size Distribution of Bubbles as Function of
`Particle Size
`
`duced; and the entire distribution shifts toward larger diameter
`bubbles at the highest pressure level.
`When gas is injected through an orifice into a fluid, bubble
`formation occurs in three distinct phases—nucleation, growth,
`and detachment (Lubetkin 1994). The formation of bubbles at
`an orifice under constant pressure conditions can be described
`by the following (Clift et al. 1978):
`
`D
`
`1/2
`
`(2)
`
`chS
`
`Q = K p 2 rgH 1 rga 2
`
`2s
`a
`where Q = flowrate; K = orifice constant; pch = chamber pres-
`sure; r = fluid density; g = gravitational constant; H = sub-
`mergence; a = instantaneous bubble radius; and s = surface
`tension. Raising the injection pressure affects the growth
`phase, and it increases the rate at which gas is transferred into
`the bubble, which increases the bubble size before it detaches
`(Sadhal et al. 1997).
`The results of the experiments performed in spherical glass
`beads are shown in Fig. 4, with the results of the bubble sizes
`produced in an aqueous matrix shown for comparison. All ex-
`periments were performed using an injection pressure of 9.0
`kPa (1.3 psi). Experiments were performed using spherical sil-
`ica beads with diameters of 6.4 mm, 14.5 mm, and 27.0 mm.
`Only the results from the two larger diameter bead experi-
`ments are shown, because the bubbles produced in the 6.4 mm
`were too large to image using the test system configuration,
`and it wasn’t possible to obtain a representative sample.
`The test results in the spherical media show a significant
`degree of coalescence, even in beads with diameters signifi-
`cantly larger than most soil grains. Average bubble diameters
`produced were 2.48 mm (14.5 mm beads) and 2.02 mm (27.0
`mm beads) which is significantly larger than the bubbles pro-
`duced in aqueous media (average diameter of 1.05 mm). Pre-
`vious experimental results observed discrete bubble flow in
`particle sizes on the order of 4.0 mm (Wehrle 1990; Ji et al.
`1993); however, the results of this study show that coalescence
`can occur, even when the particle sizes are many times larger
`than the bubble diameter. It is significant to note that the bub-
`ble diameters emerging from the porous media decrease as the
`particle size increases. This is clear from the data for the 27.0
`mm and 14.5 mm beads and is also supported by the fact that
`the bubbles emerging from the 6.4 mm beads were too large
`to capture in the imaging frame.
`An experiment was performed to compare the effect of par-
`ticle shape on the size of bubbles generated in the system (Fig.
`5). The experiments were performed with spherically and el-
`liptically (diameter-to-width ratio of 2:1) shaped particles at
`an injection pressure of 9.0 kPa. The average bubble diameter
`
`46 / JOURNAL OF COMPUTING IN CIVIL ENGINEERING / JANUARY 1999
`
`was significantly smaller in the elliptical particles than in the
`spherical particles (2.07 mm versus 2.48 mm), indicating a
`lower degree of bubble interaction and less coalescence. The
`difference in the distribution is most pronounced at the larger
`bubble sizes, with the largest bubbles having a size differential
`of approximately 1.5 mm.
`Experiments were performed in order to evaluate the effect
`of low concentrations of surfactant in an aqueous system. Fig.
`6 shows the results of two experiments performed at an injec-
`tion pressure of 9.7 kPa, one without surfactant and one with
`surfactant present in the system. The presence of the surfactant
`makes a dramatic difference in bubble size with average di-
`ameters of 0.39 mm (standard deviation of 0.08 mm) with
`surfactant, and 1.37 mm (standard deviation of 0.61 mm) with-
`out surfactant. Additionally, the size distributions of the two
`experiments were significantly different, as reflected in the
`standard deviation.
`Surfactants in the system cause a reduction in the interfacial
`tension between the fluid and the gas phases, which in turn
`causes a reduction in the diameter of bubbles produced under
`constant pressure injection scenarios (2). The accumulation of
`surfactants at the bubble/fluid interface also sets up a tangen-
`tial force that increases the drag on the bubble (Sadhal et al.
`1997), which in turn increases the bubble residence time in
`the system. For bubbles translating through an aqueous sys-
`
`FIG. 5. Cumulative Size Distribution of Bubbles as Function of
`Particle Shape
`
` J. Comput. Civ. Eng., 1999, 13(1): 43-48
`
`Downloaded from ascelibrary.org by Janelle Beitz on 12/10/20. Copyright ASCE. For personal use only; all rights reserved.
`
`

`

`distributions show that the surfactant limits the degree of bub-
`ble coalescence in the porous media. It is interesting to note
`that the presence of surfactant in the system appears to dom-
`inate many of the other system variables in terms of bubble
`coalescence behavior. Neither the presence of the porous me-
`dium nor the increase in injection pressure had a significant
`impact on the size of the bubbles generated when there was
`surfactant present in the system.
`Experimental studies detailing the effect of surfactant con-
`centration on the contaminant removal efficiency of air sparg-
`ing are ongoing (Burns and Zhang, in preparation).
`
`CONCLUSIONS
`
`The objective of this experimental study was to examine the
`characteristics of bubbles generated for application in the air
`sparging process. It produced quantitative measurements of
`bubble sizes in both aqueous and aqueous/particulate media
`and it quantified the effect of surfactants on the characteristics
`of the bubbles produced. Based on the results of this investi-
`gation, the following observations can be made:
`
`1. Increases in the injection pressure of air lead to signifi-
`cant increases in the average bubble diameter produced.
`2. Increases in the injection pressure of air lead to a wider
`distribution of bubble sizes produced.
`3. At the highest injection pressure (11.0 kPa), the size dis-
`tribution of bubble diameters increased, with the smallest
`bubble produced being approximately three times larger
`than the smallest bubble produced at 8.3 kPa.
`4. Bubble coalescence occurs even in very large diameter
`(27.0 mm) particulate materials.
`5. Bubble diameters produced in porous media decrease as
`the diameter of the particulate material increases.
`6. Less bubble coalescence occurred in elliptically shaped
`particulate materials than in spherically shaped ones.
`7. Surfactants in the aqueous systems decrease the average
`bubble diameter produced and also produce more uni-
`form bubble sizes.
`8. Surfactants in aqueous/particulate media systems de-
`crease the average bubble diameter and produce a nar-
`rower range of bubble sizes than do surfactant free sys-
`tems.
`
`ACKNOWLEDGMENTS
`
`Thanks to Carl T. Herakovich for the use of his long-distance micro-
`scope and to James E. Danberg for his assistance with the experimental
`setup. Partial financial support for this research was provided by the Uni-
`versity of Virginia. This support is gratefully acknowledged.
`
`APPENDIX. REFERENCES
`
`Baker, R. S., Hayes, M. E., and Frisbie, S. H. (1995). Evidence of pref-
`erential vapor flow during in situ air sparging. In situ aeration: Air
`sparging, bioventing, and related remediation processes, Battelle Press,
`Columbus, Ohio, 63 – 73.
`Bruell, C. J., Marley, M. C., and Hopkins, H. H. (1997). ‘‘American
`Petroleum Institute in situ air sparging database.’’ J. Soil Contamina-
`tion, Vol. 6, No. 2, 169 – 185.
`Burns, S. E., Yiacoumi, S., and Tsouris, C. (1997). ‘‘Microbubble gen-
`eration for environmental and industrial separations.’’ Separations and
`Purification Technol., Vol. 11, 221 – 232.
`Burns, S. E., Yiacoumi, S., Frost, J. D., and Tsouris C. (1998). ‘‘Appli-
`cation of digital image analysis for size distribution measurements of
`microbubbles.’’ Proc. Engrg. Found. Conf. on Imaging Technologies:
`Techniques and Applications in Civ. Engrg., Davos, ASCE, Reston, Va.,
`100 – 107.
`Chen, M.-R., Hinkley, R. E., and Killough, J. E. (1996). ‘‘Computed
`tomography imaging of air sparging in porous media.’’ Water Resour.
`Res., Vol. 32, No. 10, 3013 – 3024.
`Clift, R., Grace, J. R., and Weber, M. E. (1978). Bubbles, drops, and
`particles, Academic Press, New York.
`
`JOURNAL OF COMPUTING IN CIVIL ENGINEERING / JANUARY 1999 / 47
`
`FIG. 6. Cumulative Size Distribution of Bubbles in Aqueous
`and Surfactant Systems
`
`tem, the accumulation of surfactant is nonuniform, with higher
`concentrations at the rear of the bubble (Sadhal et al. 1997).
`Figs. 7 and 8 show the results of experiments performed in
`aqueous/particulate media in the presence of surfactant. The
`data clearly show that the reduction in average bubble diam-
`eter due to the presence of surfactant occurs in particulate sys-
`tems as well as in aqueous stems. In addition, the narrow size
`
`FIG. 7. Effect of Surfactant on Bubble Coalescence in Porous
`Media
`
`FIG. 8. Cumulative Size Distribution for Surfactant/Porous
`Media as Function of Pressure
`
` J. Comput. Civ. Eng., 1999, 13(1): 43-48
`
`Downloaded from ascelibrary.org by Janelle Beitz on 12/10/20. Copyright ASCE. For personal use only; all rights reserved.
`
`

`

`Haulbrook, W. R., Feerer, J. L., Hatton, T. A., and Tester, J. W. (1993).
`‘‘Enhanced solubilization of aromatic solutes in aqueous solutions of
`N-vinylpyrrolidone/styrene.’’ Envir. Sci. and Technol., Vol. 27, No. 13,
`2783 – 2788.
`Hines, W. W., and Montgomery, D. C. (1990). Probability and statistics
`in engineering management and science, 3rd Ed., Wiley, New York.
`Ji, E., Dahmani, A., Ahlfeld, D. P., Lin, J. D., and Hill, E. (1993). ‘‘Lab-
`oratory study of air sparging: Air flow visualization.’’ Groundwater
`Monitoring and Remediation, Fall, 115 – 126.
`Kile, D. E., and Chiou, C. T. (1989). ‘‘Water solubility enhancement of
`DDT and trichlorobenzene by some surfactants below and above the
`critical Micelle concentration.’’ Envir. Sci. and Technol., Vol. 23, No.
`7, 832 – 838.
`Leeson, A., Hinchee, R. E., Headington, G. L., and Vogel, C. M. (1995).
`‘‘Air channel distribution during air sparging: A field experiment.’’ In
`situ aeration: Air sparging, bioventing, and related remediation pro-
`cesses, Battelle Press, Columbus, Ohio, 215 – 235.
`Li, X., and Yortsos, Y. C. (1995a). ‘‘Visualization and simulation of bub-
`ble growth in pore networks.’’ AIChE J., Vol. 41, No. 2, 214 – 222.
`Li, X., and Yortsos, Y. C. (1995b). ‘‘Theory of multiple bubble growth
`in porous media by solute diffusion.’’ Chemical Engrg. Sci., Vol. 50,
`No. 8, 1247 – 1271.
`Lubetkin, S. D. (1994). Bubble nucleation and growth. Controlled par-
`ticle, droplet, and bubble formation, D.J. Wedlock, ed., Butterworth-
`Heinemann, Stoneham, Mass., 159 – 190.
`Lundegard, P. D., and LaBrecque, D. (1995). Air sparging in a sandy
`aquifer (Florence, Oregon, USA): Actual and apparent radius of influ-
`ence. J. Contaminant Hydrology, Vol. 19, No. 1, 1 – 27.
`
`Marley, M. C., Hazebrouck, D. J., and Walsh, M. T. (1992). ‘‘The appli-
`cation of in situ air sparging as an innovative soils and ground water
`remediation technology.’’ Ground Water Monitoring Rev., Spring,
`107 – 115.
`Rimmer, A., Parlange, J.-Y., Steenhus, T. S., Darnault, C., and Condit, W.
`(1996). ‘‘Wetting and nonwetting fluid displacements in porous me-
`dia.’’ Transport in Porous Media, Vol. 25, No. 2, 205 – 215.
`Roosevelt, S. E., and Corapcioglu, M. Y. (1998). Air bubble migration in
`a granular porous medium: Experimental studies, Water Resour. Res.,
`Vol. 34, No. 5, 1131 – 1142.
`Sadhal, S. S., Ayyaswamy, P. S., and Chung, J. N. (1997). Transport
`phenomena with drops and bubbles, Springer-Verlag, New York.
`Satik, C. and Yortsos, Y. C. (1996). ‘‘A pore-network study of bubble
`growth in porous media driven by heat transfer.’’ J. Heat Transfer, Vol.
`118, No. 2, 455 – 462.
`Schima, S., LaBrecque, D. J., and Lundegard P. (1994). ‘‘Using resistivity
`tomography to monitor air sparging.’’ Proc., SAGEEP, Environmental
`and Engineering Geophysics Society, Wheat Ridge, Colo., 757 – 774.
`Smith, J.