Rep

rodu

ced

from

Vad

ose

Zon

e Jo

urna

l. P

ublis

hed

by S

oil S

cien

ce S

ocie

ty o

f Am

eric

a. A

ll co

pyrig

hts

rese

rved

.

Transport and Retention Mechanisms of Colloids in Partially Saturated Porous Media

John T. Crist, Yuniati Zevi, John F. McCarthy, James A. Throop, and Tammo S. Steenhuis*

ABSTRACT tion and deposition has improved in the past 10 yr, scien-tific reviews emphasize the need for more research onThe transport, retention, and release of hydrophobic and hydro-the mechanisms controlling transport in the unsaturatedphilic polystyrene latex microsphere colloids were examined in 0.5-

cm-thick, 26-cm-long slab chambers filled with either regular (hydro- zone (Ouyang et al., 1996; Kretzschmar et al., 1999).philic) or weakly water-repellent sand. The water-repellent sand In a review of colloid transport in the vadose zone,consisted of a mixture of 0.4% strongly water-repellent grains with Lenhart and Saiers (2002) described the transport andunmodified regular sand for the remainder. The concentration of col- distribution of colloids in the vadose zone as advectionloids in the outflow water was measured at the same time as the pore- and dispersion, together with a sink–source term. Thescale distribution of colloids was recorded in still and video images. advection–dispersion is relatively well understood andAlthough the type of sand affected the flow pattern in the top of the

could be modified to include the effects of preferentialchamber, it did not affect the breakthrough for the same type offlow (Steenhuis et al., 2001). Difficulties arise from sizecolloids. More hydrophilic colloids were eluted in the drainage waterexclusion that limits the accessibility of colloids to por-than hydrophobic colloids. Images showed that there was a greatertions of the pore space. Determining the sink–sourceretention of the hydrophobic colloids due to strongly attractive hy-

drophobic interaction forces between colloids and subsequent filtering term is even more uncertain and currently an area of ac-of colloidal aggregates in the narrow passages between grains. Once tive research. As shown below, one of the major limita-filtered, these aggregates then served as preferred sites for attachment tions in understanding and modeling this term is theof other hydrophobic colloids. The hydrophilic colloids were retained difficulty of visualizing the processes in a medium whereprimarily in a thin film of water at the edge of the menisci, the air– the location and extent of the AW interface is a functionwater–solid (AWS) interface. Centrifugal motion within the pendular of many factors, such as matric potential and the previ-rings observed in the videos contributed to movement of the colloids

ous wetting history (Lenhart and Saiers, 2002). Visualiz-toward the AWS interface, where colloids were retained due to bothation is difficult; therefore, most studies have been lim-low laminar flow velocities near the grain surface and straining in theited to colloid breakthrough experiments, as well asthin water film at the edge of the meniscus. Except near the solidconceptual, analytical, and/or computer models. Colloidinterface, sorption at the air–water (AW) interface was not observed

and appeared unimportant to the retention of colloids. The findings breakthrough experiments in partly saturated mediaform an essential link between colloid retention and transport pro- (Schafer et al., 1998a; Jewett et al., 1999; Jin et al., 2000;cesses at the interfacial, pore, and Darcy scales. Chu et al., 2001; Lenhart and Saiers, 2002) show that more

hydrophobic colloids, compared with hydrophilic colloids,are retained in the porous media under otherwise similar

Mobile subsurface colloids have received con- conditions. Moisture content and interfacial energiessiderable attention in recent years because of their also play an important role. While under saturated con-

important role in the translocation of particle-reactive con- ditions all or most negatively charged hydrophilic col-taminants in soils (Wan and Tokunaga, 1997; Kretzschmar loids will be transported through clean sands, break-et al., 1999). Colloids are defined as suspended particu- through diminishes with decreasing moisture contents.late matter with diameters �10 �m (Stumm, 1977). Col- Reduced transport at lower moisture contents is oftenloidal sized materials may form stable complexes with vari- attributed to retention of colloids at the AW interface,ous pollutants previously considered to have very limited the area of which increases at lower moisture content.mobility in the subsurface (McCarthy and Zachara, 1989; This interpretation arises from the pore-scale visual-Ryan and Elimelech, 1996), including metals (Grolimund ization studies performed by Wan and Wilson (1994a,et al., 1996; Jordan et al., 1997; Karathanasis, 1999), pesti- 1994b). Those researchers, employing etched glass micro-cides (de Jonge et al., 1998; Sprague et al., 2000; Williams models, observed retention of microspheres and bacte-et al., 2000), and radionuclides (McCarthy et al., 1998; ria at the edges of air bubbles within the pores of theKersting et al., 1999; Flury et al., 2002). These complexes two-dimensional micromodel. Crist et al. (2004) ques-can significantly enhance the movement of contaminants tioned this interpretation on the basis of their pore-scalein both saturated and unsaturated porous media in a visualization in three-dimensional porous media. Theirprocess termed colloid-facilitated (or colloid-mediated) observations suggested that colloids were not retainedtransport. Though our understanding of colloid mobiliza- at the AW interfaces, but rather near the AWS interface

near the menisci of pendular rings.In unsaturated porous media, an additional mecha-J.T. Crist, Y. Zevi, J.A. Throop, and T.S. Steenhuis, Department of

nism of “film straining” in thin water films was postu-Biological and Environmental Engineering, Riley-Robb Hall, CornellUniversity, Ithaca, NY 14853; J.F. McCarthy, Department of Earth lated by Wan and Tokunaga (1997) and Veerapaneniand Planetary Sciences, University of Tennessee, Knoxville, TN 37996. et al. (2000). In the conceptual model of Wan and Toku-Received 29 Jan. 2004. Original Research Paper. *Corresponding naga (1997), when the water content is below criticalauthor (tss1@cornell.edu).

Abbreviations: AW, air–water; AWS, air–water–solid; BTC, break-Published in Vadose Zone Journal 4:184–195 (2005).© Soil Science Society of America through curve; DLVO, Derjaguin–Landau–Verwey–Overbeek forces;

PV, pore volume; SW, soil–water; 2D, distilled–deionized.677 S. Segoe Rd., Madison, WI 53711 USA

184

Rep

rodu

ced

from

Vad

ose

Zon

e Jo

urna

l. P

ublis

hed

by S

oil S

cien

ce S

ocie

ty o

f Am

eric

a. A

ll co

pyrig

hts

rese

rved

.

www.vadosezonejournal.org 185

make them hydrophobic and water-repellent, with a negativemoisture content, colloids are retained because thematric entry value. The hydrophobic water-repellent sandsthickness of the water film connecting one pendular ringused in these experiments consisted of a mixture of 0.4% ofto the next falls below the diameter of the colloid (Fig. 1the modified water-repellent grains with unmodified sand forin Wan and Tokunaga, 1997). However, calculations of the remainder.

film thicknesses for a range of matric potentials from The infiltration chamber was constructed from 0.5-cm-thick,�10 to 30 cm (Iwamatsu and Horii, 1996; Lenhart and clear acrylic sheets. Interior dimensions of the chamber wereSaiers, 2002) demonstrated that the thicknesses of the 26.0 cm high, 4.8 cm wide, and 0.5 cm deep. The front platefilms are tens of nanometers under equilibrium condi- interfered with the image analysis and was mounted with bolts

and wing nuts for disassembly after the sand was added. Thetions. Thus, such thin films can simply be interpretedinfiltration chamber was supported on a mounting assemblyas discontinuities in pendular rings between individualat a 45� incline from horizontal and perpendicular to the focusgrains (Lenhart and Saiers, 2002; Crist et al., 2004).of the camera. A 45� inclination was chosen to maximize gravi-The goal of our study is to extend understanding tational effects while preventing erosion of the packed sand

of colloid retention mechanisms at the AWS interface layers during infiltration and drainage. To allow visualizationsuggested by visual observations of Crist et al. (2004) of different locations, the front plate was removed and thefor microspheres and implied earlier by Thompson et al. viewing area was adjusted across the camera by sliding the(1998), Thompson and Yates (1999), and Chu et al. (2001, chamber along rails on the mounting assembly. Through a

sampling port at the bottom of the chamber, effluent samples2003) for virus transport. Specific objectives includefor the colloid BTCs were collected simultaneously with thecomparison of the behavior of negatively charged hy-visualizations.drophilic and hydrophobic colloids in water-repellent

The infiltration chamber was prepared by filling it with(hydrophobic) and unmodified hydrophilic coarse sands. sand; then one pore volume (PV, ≈26 mL) of water (0.1 mMResults will include colloid breakthrough curves (BTCs), CaCl2, pH 5.6) was delivered through the sampling port at anstill and video images, and calculations of the total po- inlet flow rate of 2 mL min�1. The pore volume was determinedtential energies of interaction of colloids with each other in initial experiments at a flow rate of 2 mL min�1 by measuringand with the AW and soil–water (SW) interfaces as a the volume required to obtain 50% of the initial Cl� concentra-

tion in the leachate. With the sand completely wetted, themeans of evaluating the mechanistic basis of the ob-chamber was placed on the inclined mounting assembly andserved behavior. The results thus form a link betweenleft to drain undisturbed for 30 min. The front plate wasinterfacial-, pore-, and Darcy-scale processes.removed, and using a peristaltic pump, a suspension of eitherhydrophilic or hydrophobic microspheres at a concentrationof approximately 3 � 105 particles mL�1 (in a solution ofMATERIALS AND METHODS0.1 mM CaCl2 and pH 5.6) was applied as a point source on

Apparatus and Experimental Design the top layer of sand. One pore volume of colloidal suspensionwas delivered at a flow rate of 2 mL min�1. Two pore volumesA similar experimental setup as Crist et al. (2004) was usedof colloid-free solution of the same ionic strength and pH wereand includes an infiltration chamber, light source, electro-applied at the same application rate immediately following theoptical equipment (lens, camera, and computer system), andinput of the colloidal suspension. Effluent from the samplingimaging software (Fig. 1). The electro-optical equipment in-port was collected every minute during the 3-PV injectioncluded a Zoom 6000 II lens assembly with 1X adaptor (Navitar,sequence. The samples were analyzed by measuring absor-Inc., Rochester, NY) and color charged-coupled device cam-bance at 380 nm using a spectrophotometer (Bausch and Lomb,era (Cohu, Inc., Poway, CA). An IBM-compatible computer,Inc., Rochester, NY). No correction for background levels wasmonitor, frame grabber card (Scion Corp., Frederick, MD),required because absorbance in effluent samples from theand Scion Image software were used for image processing andcontrol experiments was negligible. At the conclusion of thedisplay. Image resolution for the complete system was 212 000experiments with the unmodified hydrophilic sand, the verticalsquare pixels mm�2. In addition to capturing still digital imagesdistribution of retained colloids was determined by sectioningwith Scion Image, a videocassette recorder and monitor werethe sand in the chamber at 1-cm intervals. Each layer wasused to gather continuous real-time images for subsequentoven-dried at 105�C, and the retained colloids resuspended byreview and analysis. Several sets of colloid breakthrough ex-mixing with 7 mL of distilled–deionized (2D) water for 30 minperiments (described below) were performed with this visual-in a slow-speed agitator. The released colloids were quantifiedization system. The viewing area was illuminated from under-using the spectrophotometer. The sand was oven-dried atneath using a variable intensity, 150-W tungsten-halogen lamp105�C a second time and reexamined with the electro-opticalwith fiber optics cable (D.O. Industries, Inc.).equipment. The efficiency of colloid recovery in the effluentTwelve colloid breakthrough experiments were completedand mean arrival time of the colloids were evaluated using(six with unmodified hydrophilic sand and six with water-moments analysis of the observed BTCs for each replicate.repellent sand), producing six sets of replicate experiments

with hydrophilic colloids, hydrophobic colloids, and no col-loids. Nonfluorescent, blue-dyed polystyrene latex micro- Interfacial Potential Energiesspheres (Magsphere, Inc., Pasadena, CA) comparable in size

The interactions of colloids approaching each other or theto Cryptosporidium parvum oocysts were used in the experi-AW or SW interfaces were evaluated as the sum of Derjaguin–ments. The surfaces of the colloids were either negativelyLandau–Verwey–Overbeek (DLVO) forces, including van dercharged, hydrophilic 4.8-�m carboxylated microspheres or hy-Waals and double layer potential energies. Additionally, hy-drophobic 5.2-�m underivatized microspheres. Hydrophilicdrophobic forces were considered, although these interactionssand consisted of translucent quartz sand (Unimin Corp., Newwere important only for interactions of colloids with the AWCanaan, CT) with grain diameters equivalent to 0.85 tointerface and of hydrophobic colloids with each other. The1.70 mm, and was washed and rinsed 10 times in distilled watertotal potential energy, �tot, to these interactions was evaluatedto remove loose surface impurities. The procedure of Bauters

et al. (1998) was used to modify the surfaces of sand grains to as a function of the separation distance, x:

Rep

rodu

ced

from

Vad

ose

Zon

e Jo

urna

l. P

ublis

hed

by S

oil S

cien

ce S

ocie

ty o

f Am

eric

a. A

ll co

pyrig

hts

rese

rved

.

186 VADOSE ZONE J., VOL. 4, FEBRUARY 2005

Fig. 1. Experimental setup. Not shown are the color charged-coupled device camera and computer system.

�tot(x) � �vdW(x) � �edl(x) � �hyd(x) [1]� (2 2

0c) ln[1 � exp(�2x)]� [4]The van der Waals potential, �vdW, was estimated using Eq.[2]. Interactions of a colloid with another colloid or a grain

where ε is the dielectric constant of water (dimensionless), ε0surface were formulated as an unretarded sphere–sphere in-is the permittivity of free space, 0c is the surface potentialteraction, assuming pair-wise additivity of the interatomic po-of the colloid, and is the reciprocal double layer thicknesstentials (Eq. [2a]; Hamaker, 1937). Colloid interactions withcalculated from the valence and ionic strength of the electro-a macroscopically flat surface (the AW interface) were approx-lyte solution. The surface potentials, 0, were calculated basedimated using Eq. [2b] (Norde and Lyklema, 1989).on the �-potential. For small potentials, the potential decaysexponentially in the diffuse double layer, and the surface po-

�vdW � �A132

12 � yr 2 � ry � r

�y

r 2 � ry � r � y tential is related to the �-potential by

0 � ��1 �zac

�exp(�z) [5]� 2 ln� r 2 � ry � rr 2 � ry � r � y �� [2a]

where z is the distance between the surface of the charged�vdW � �

A132

6 �2ac(x � ac)x(x � 2ac)

� ln�x � 2ac

x � � [2b] particle and the slipping plane (van Oss et al., 1990). That dis-tance, z, is a theoretical construct, but is usually taken to be5�. However, the calculated interaction energy profile did notwhere y � ag/ac, r � x/ac, ac, and ag are the colloid and grain change substantially over a several-fold range of values of z.radii, respectively; x is the separation distance; and A132 is theThe �-potential of the colloids in the electrolyte solution usedcomplex Hamaker constant for solids 1 and 2 in medium 3.in the experiments was measured using a Zetasizer (MalvernInstruments, Southbough, MA) and found to be �18.6 andA132 � �√A11 � √A33��√A22 � √A33� [3]�23.4 mV for the hydrophilic and hydrophobic colloids, re-

where A11, A22, and A33 are the Hamaker constants for each spectively. The �-potential of the quartz sand was taken to becomponent. The value of A132 was calculated to be 4.8 � 10�21

�60 mV (Elimelech, 1985; Elimelech et al., 2000).J for the polystyrene–water–quartz system, 5.2 � 10�21 J for For the interaction of colloids with the sand grains or withthe polystyrene–water–polystyrene system, and �1.2 � 10�20

the AW interface, the double layer potential for a sphere andJ for the air–water–polystyrene system (Israelachvili, 1992). flat surface was approximated (Norde and Lyklema, 1989):

For the interaction of the colloids with each other, thedouble layer potential, �edl, was calculated for sphere–sphere

�edl � �εε0 ac( 20c

20s ) �20c0s

20c

20s

ln �1 � exp(�x)1 � exp(�x) �interaction for the constant potential case (Hogg et al., 1966):

�edl ��εε0 2ac

2ac � ag�22

0c ln �1 � exp(�x)1 � exp(�x) � � ln1 � exp(�2x)]� [6]

Rep

rodu

ced

from

Vad

ose

Zon

e Jo

urna

l. P

ublis

hed

by S

oil S

cien

ce S

ocie

ty o

f Am

eric

a. A

ll co

pyrig

hts

rese

rved

.

www.vadosezonejournal.org 187

Fig. 2. Interfacial potential energies for colloids as a function of distance to the grain surface, air–water (AW) interface, and another colloid.Left, hydrophilic colloids; right, hydrophobic colloids.

where 0s is the surface potential of the flat surface; the RESULTS AND DISCUSSION�-potential of the AW interface was taken to be �60 mV

Total Potential Energies of Interaction(Schafer et al., 1998b).In addition to the DLVO interactions, hydrophobic forces Figure 2 plots the total interaction energies for the

act between particles and AW interfaces. Asymmetrical hy- hydrophilic colloids (left panel) and hydrophobic col-drophobic interactions between two surfaces can be calculated loids (right panel) as a function of the separation dis-based on the respective water contact angles (Yoon et al., 1997;

tance. The interaction of the colloid and sand (Fig. 2;Schafer et al., 1998b). The hydrophobic interaction energydot-dash line) has a substantial repulsive energy barrierbetween small particles and a flat surface thus can be describedthat would act to prevent attachment of either the hydro-by (Schafer et al., 1998b)phobic or hydrophilic colloids. Interactions of the hydro-philic colloids with the AW interface are also strongly�hyd � K123 ac

x 2� �

K123 ac

x[7]

repulsive. However, attractive hydrophobic forces mod-erate the net repulsive energy barrier of the hydropho-

The force constant, K123, for asymmetric interactions be- bic colloids with the AW interface, and the interactiontween macroscopic bodies 1 and 2 in medium 3 can be de-energy becomes attractive at smaller separation distances.scribed as (Yoon et al., 1997; Schafer et al., 1998b)No repulsive energy barrier exists at any separation dis-tance for the interactions of the hydrophobic colloids

logK123 � a �cos 1 � cos 2

2 � � b [8] with each other, and there is a strong attractive potentialat closer approach distances. These results would predict

where 1 and 2 are the water contact angles for the AW that the hydrophilic colloids should be efficiently trans-interface (180�; Schafer et al., 1998b) and water–colloid inter- ported through the porous media, while the hydropho-face (10� and 100� for the hydrophilic and hydrophobic latex bic colloids should experience greater retention, primar-microsphere, respectively; Wan and Wilson, 1994a, 1994b; Butt ily due to aggregation of the colloids with each other.et al., 2002). The terms a and b are system-specific constants.

However, for both the hydrophobic and hydrophilic col-For a system of silica surfaces with different contact angles,loids, a substantial repulsive energy barrier at largera was �7 and b was �18 (Schafer et al., 1998b). We adjustedseparation distances hinder colloid attachment at thea and b until the �hyd approached zero at large separation dis-

tances, which yielded a � �6 and b � �22. AW interface.

Rep

rodu

ced

from

Vad

ose

Zon

e Jo

urna

l. P

ublis

hed

by S

oil S

cien

ce S

ocie

ty o

f Am

eric

a. A

ll co

pyrig

hts

rese

rved

.

188 VADOSE ZONE J., VOL. 4, FEBRUARY 2005

Fig. 3. Image of the sand layers after application of hydrophobic colloids and water, showing the distribution of the blue-dyed colloids withinthe preferential (fingered) flow path.

Breakthrough Experiments and ences in the flow pattern may have had an effect on theextent of colloid retention.Colloid Retention Measurements

At the Darcy scale, the colloid BTCs were signifi-Moisture contents and bulk densities were deter- cantly different for the hydrophilic and hydrophobicmined after 30 min of drainage. Water contents ranged colloids while the type of sand had a much smaller effectfrom 0 at the top layer to saturation at the lowest depth (Fig. 4). Breakthrough of hydrophilic colloids was firstdirectly after the infiltration period. Bulk densities detected at 0.3 PV (Fig. 4a), and effluent concentrationsranged between 1.66 and 1.74 g cm�3. Three zones can increased steadily to a peak value at 1.2 PV. Hydropho-be delineated: 0 to about 0.08 cm3 cm�3 for the 0- tobic colloids first appeared at 0.15 PV, reached a plateau of14-cm depth, approximately 0.08 to 0.29 cm3 cm�3 forapproximately of 0.03 C/C0, increased again at 0.4 PV,the 14- to 19-cm depth, and approximately 0.29 to theand then reached the peak concentration at 1.2 PV,saturated moisture content which ranged between 0.35similar to the hydrophilic colloids (Fig. 4b). For both typesand 0.37 cm3 cm�3 for the 19- to 25-cm depth. Standardof sand, the retention of the hydrophobic colloids wasdeviations in moisture contents were greatest in thealmost twice that of the hydrophilic colloids.intermediate and lower zones. Moisture contents during

The type of sand had no effect on either the shapethe colloid experiments were greater, increasing shortlyof the BTC or the extent of the hydrophilic colloidafter water was added initially and then remaining es-retention (p � 0.27). Approximately one-half (49.6 �sentially constant since the visualization showed that1.8%) of the hydrophilic colloid mass was recovered inmenisci were stationary. After the water flux was stopped,the effluent for the replicate columns of the two typesdrainage stopped within a few minutes. The exact mois-of sand (Fig. 4). There was a small but significant differ-ture contents during the experiments could not be mea-ence in hydrophobic colloid retention between the twosured because the soils drained much faster than thesands, with less retention (p � 0.05) of the hydrophobicsamples could be taken.colloids on the water-repellent sand (30.1 � 1.3% ofAs expected from the results of Bauters et al. (2000),the colloid mass was recovered in the effluent, or ≈70%the presence of a few water-repellent grains affected theretained in the columns), compared with the hydrophilicwater flow pattern. Infiltration in the unmodified hydro-sand (25.6 � 0.5% recovery in the effluent) (Fig. 4). Itphilic sand produced one fingered flow path for colloidis unlikely that the differences in the flow pattern be-and water movement, measuring approximately 2 cmtween the two types of sand played a role in the extentwide in the upper packed sand layers and increasing toof the retention of hydrophobic colloids because thethe width of the chamber below the 11- to 13-cm depthhydrophilic colloids should have been affected in a simi-(Fig. 3). In contrast, for infiltration in the weakly water-lar way. Thus, the difference may be related to greaterrepellent sand, flow across the whole width of the cham-attraction between the hydrophobic colloids and theber was established within 2 to 4 cm below the point ofstrongly water-repellent grains despite that these grainsapplication. The type of colloid did not affect the infil-

tration pattern, although, as will be discussed, differ- constituted only 0.4% of the porous media. We did not

Rep

rodu

ced

from

Vad

ose

Zon

e Jo

urna

l. P

ublis

hed

by S

oil S

cien

ce S

ocie

ty o

f Am

eric

a. A

ll co

pyrig

hts

rese

rved

.

www.vadosezonejournal.org 189

Fig. 4. Colloid breakthrough curves, the solid line represents the mean C/C0 values for regular (hydrophilic) sand, and the dashed line representsthe slightly water-repellent sand. (a) Hydrophilic colloids, (b) hydrophobic colloids.

attempt to estimate the potential energy of interaction (open and solid squares in Fig. 5). Although the trendwas correct, the absolute concentrations are underesti-of the colloids with the water-repellent grains.mated for the hydrophilic colloids because when theThe depth distribution of the colloids retained in theamount of colloids in the effluent water and that re-porous media but capable of being detached by extrac-tained in the soil were summed, the mass balance couldtion with 2D water was determined for the unmodifiedonly account for 73 and 75% of the total amount ofhydrophilic sand (Fig. 5). No measurements were madecolloids for the replicate columns. The mass balance forfor the water-repellent sand. For the unmodified hydro-the hydrophobic colloids was 95 and 116%. That is, afterphilic sand the relative greatest amount of colloids re-drying and resuspension in 2D water, all the hydropho-tained was around the 14-cm depth, where the capillarybic colloids could be released, but only one-half of thefringe begins. Besides this similarity, the trends of reten-hydrophilic colloids retained on the unmodified sandtion with depth varied with colloid type. The retentiongrains could be removed by the same procedure.of hydrophobic colloids did not show a clear trend with

depth in the first 10 cm, where the moisture contents Visualization of Colloid Retention in Videoare the smallest, and then decreased farther down (openand Still Imagesand solid circles in Fig. 5). The concentration hydrophilic

colloids retained on the sand increased first and then Visualization of the colloids provided unique and val-uable information that is critical in understanding thebecame approximately constant below the 15-cm depth

Rep

rodu

ced

from

Vad

ose

Zon

e Jo

urna

l. P

ublis

hed

by S

oil S

cien

ce S

ocie

ty o

f Am

eric

a. A

ll co

pyrig

hts

rese

rved

.

190 VADOSE ZONE J., VOL. 4, FEBRUARY 2005

Fig. 5. Distribution of colloids with depth in regular hydrophilic sand. The squares represent the hydrophilic colloids and the circles are thehydrophobic colloids. The open and solid represent the two replicates. The solid line is the average of the two replicates. The triangles rep-resent the water content.

different mechanisms of colloid retention and in com- Retention of Hydrophilic Colloidsprehending the observed breakthrough pattern of col- The processes affecting the retention of hydrophilicloids. Proper interpretation is aided by both video re- colloids are well illustrated in the photograph, taken atcording that portrays the dynamic behavior of colloid the 10-cm depth in the regular hydrophilic sand 30 min af-attachment and still images during and after colloid ad- ter the hydrophilic colloid suspension was added (Fig. 7).dition that show the locations where colloids were im- In the figure, grains are clearly visible. Most of the poremobilized. To prevent interference by condensation on spaces contain water, except in the right-hand cornerthe front plate (of the chamber) through which the im- where one of the pore spaces between four grains is filledages were taken, the front plate was removed after thechamber was drained for 30 min and before the colloidswere added. The disturbance on the sand was minimalsince it was unsaturated and the pendular rings heldthe sand in place. Removing the plate resulted in anincreased number of AWS interfaces, especially in thelower depths of the chamber, and might have providedadditional surfaces for retention of the colloids.

Selected video images of the experiment where hy-drophilic colloids were added to the unmodified hydro-philic sands are provided in the supplemental material.These three files are large and can best be downloadedwith a fast internet connection. Two videos were takenin the viewing area shown in Fig. 6, which was located6 cm below the top of the column. In these videos (Vid-eos 1 and 2), we observe four grains (Fig. 6) with fourpendular rings partly visible. The large space betweenthe grains is devoid of water and filled with air. Colloidsmoved through three of the pendular rings (labeled A,B, and C), with water flowing down from the top of thefigure. These pendular rings are apparently connected ata level below the visual capabilities of the experimental

Fig. 6. Four isolated menisci (or a pendular ring of water) betweensetup. Several processes affecting colloid transport are four sand grains associated with static air–water and air–water–solidevident in the still images and videos. These retention interfaces. These are the same sand grains that are shown in Videos

1 and 2.and mobilization processes are discussed below.

Rep

rodu

ced

from

Vad

ose

Zon

e Jo

urna

l. P

ublis

hed

by S

oil S

cien

ce S

ocie

ty o

f Am

eric

a. A

ll co

pyrig

hts

rese

rved

.

www.vadosezonejournal.org 191

Fig. 7. Various retention mechanisms of hydrophilic colloids: gravitational settling, filtration, colloids collection in bridges, and colloid retentionat the air–water–solid interface. The picture was taken at the 10-cm depth at 30 min after the colloid suspension was added.

with air. Partial pendular rings can be observed border- sive barrier during the collision with the stationarycolloid resulting in coagulation. This could be aided bying this pore space. In the remainder of the photograph

(Fig. 7), the grains stick out above the water as pebbles. certain functional groups of the carboxylated micro-spheres that attach to each other. The “v-shaped strings”The most distinct feature and the main mechanism of re-

tention are the blue bands of colloids at the fringes of of colloids visible in a close-up image of the bridge inFig. 9 would support this assumption. Finally, of minorthe menisci, associated with the AWS interface (Fig. 7).

Although the details of the menisci are difficult to dis- importance are some colloids at the SW interface, whichwe have attributed to gravitational settling.cern, it was clear from direct observations that the dark

band of blue hydrophilic colloids was located near the Since the negatively charged, hydrophilic colloids arerepelled by both the AW and SW interfaces (Fig. 2),grain surfaces at the edge of the meniscus where the

thickness of the water film is smallest. Preferential at- the colloids are also repelled by the AWS interface. De-spite this hydrophilic colloids are retained at the AWStachment at the AWS interface is also apparent in Fig. 8a

and in Fig. 5 at locations labeled B and C. Crist et al. interface. Thus an additional mechanism is needed tocounteract the repellent force of the AWS interface.(2004) also found the attachment at or near the AWS

interface to be the primary retention mechanism for This mechanism can best be deduced from the move-ment of colloids in the pendular ring labeled B in1-�m colloids. In addition, although of less importance,

Fig. 7 shows the retention of colloids as a result of fil- Video 1. In this video, single colloids can be seen movingthrough the pendular rings as small black dots and atration in the small pore spaces where the grains come

together. This is evidenced by the dark blue band at the few as larger colloidal aggregates. Although most of thesecolloids move through the middle of the pore, a few oflocation where the grains touch (Fig. 7). Another form

of retention that has not often been mentioned in the these colloids veer off the path, bringing them close tothe edge of the meniscus. This veering off perhaps isliterature is coagulation of the hydrophilic colloids, the

forming of a “bridge” between grains. In Fig. 7, the pen- caused by the force imparted by the centrifugal motionof the circular flow through the pendular ring. The largerdular rings bordering the air space at the lower right side

contains a dark blue concentrated “patch” of colloids colloidal aggregates suddenly become trapped at theAWS interface, and single colloids are caught by the col-extending across the meniscus between two grains. We

do not think that this represents attachment at the AW loids already present at the interface. There is no regu-larity in the interval between “catches,” and we are notinterface between the grains, but rather an accumulation

of coagulated colloids that bridge from one AWS inter- sure why the colloids are caught; it is likely related to thesmall thickness of the meniscus at the AWS interface.face to another. Hydrophilic colloid aggregation is coun-

ter to the DLVO theory presented in Fig. 2, since it shows In addition, Video 1 shows that colloids are unlikely tobe retained at the AW interface as proposed originallya strong repulsive force at distances �1 nm between

colloids. However, at distances �1 nm there is an attrac- by Wan and Wilson (1994a). The water in the AW inter-face, except close to the edges near the solid interface,tive force. The velocity of the moving colloid could pro-

vide sufficient momentum to break through the repul- is in motion, and any colloid at the AW interface will

Rep

rodu

ced

from

Vad

ose

Zon

e Jo

urna

l. P

ublis

hed

by S

oil S

cien

ce S

ocie

ty o

f Am

eric

a. A

ll co

pyrig

hts

rese

rved

.

192 VADOSE ZONE J., VOL. 4, FEBRUARY 2005

Fig. 8. (a) Hydrophilic colloids, mainly deposited at the air–water–solid interface. (b) Hydrophobic colloids, mainly deposited within the pendularring at the solid–water interface with a few at the air–water–solid interface.

move with the water in the main flow direction. The ment, aggregated particles are seen accumulated in thelowest part of the pendular ring labeled A. This is likelywater in the pendular ring labeled D is stationary and,

at the same time, is unconnected during infiltration and due to gravitational settling of the aggregates near themeniscus even though it is apparent that the water isdrainage. Therefore, colloids were not present at any

time, and the pendular ring cannot retain any colloids. flowing upward at this time. It is interesting to see thatsometimes the water movement stops and even movesFinally, in Video 2 another mechanism of colloid re-

tention can be observed that was not visible in the still backward. This has a large effect on the stability ofthe accumulated colloids. It is not clear if a blockageimages. In the images of Video 2, which is at the location

depicted in Fig. 6 and appears later in the flow experi- upstream or the pump caused the change in flow condi-

Fig. 9. Coagulated hydrophilic colloids forming “bridges” between sand grains at the 13-cm depth 98 min after adding colloid suspension.

Rep

rodu

ced

from

Vad

ose

Zon

e Jo

urna

l. P

ublis

hed

by S

oil S

cien

ce S

ocie

ty o

f Am

eric

a. A

ll co

pyrig

hts

rese

rved

.

www.vadosezonejournal.org 193

tions. In this video, filtering of colloids can also be seen at the AWS interface, but as soon as the water sur-face expands due to an increase in flow rate, the depos-in pendular rings labeled B and C in Fig. 6. It is a dy-

namic process where colloids sometimes move slowly ited colloids are swept away. Moreover, the video im-ages show an air bubble trapped in the pore space. Whileand sometimes quickly, almost similar to rocks as bed

load in a fast flowing river. a few colloids are trapped within the ring, the majorityof the colloids pass by the bubble.

Retention of Hydrophobic ColloidsMechanisms of Colloid Retention and TransportFor the hydrophobic colloids (Fig. 8b), retention at the

AWS interface was present but minor compared with the On the basis of the results of the Darcy-scale BTCshydrophilic colloids. The major deposition mechanisms and the video and still images at the pore scale, weare due to the strong attachment force that exists be- can identify the mechanisms of colloid transport in ourtween hydrophobic colloids, resulting in the formation partially saturated porous media.of larger colloid aggregates that can be more easily fil- The observed retention of a greater amount of hy-tered by the relatively narrow pore spacing close to drophobic than hydrophilic colloids is consistent withwhere the grains are touching. Also, the colloidal aggre- the calculated balance of attractive and repulsive elec-gates can attach to those colloids already present at the trostatic forces. Repulsive energy barriers limit attach-grain surface. The dynamic nature of this process was ment between the colloids and the hydrophilic surfacesnot captured on video, but it is well illustrated in still of the unmodified sand grains, as well as with the AWimages (Fig. 8b). Even though the injection solution interface (Fig. 2). The absence of an energy barrier be-contained colloids of a uniform size, it is apparent from tween hydrophobic colloids and attractive interactionsvisual observations during the experimental run that the at shorter separation distances favor aggregation of themobile colloids occur in a range of sizes, suggesting that hydrophobic colloids. Rapid aggregation would be ex-the colloids have aggregated during transport. pected in the absence of energy barriers, resulting in

The images of colloid retention are in agreement with formation of the extended structures seen in Fig. 8b.the BTCs. The presence of fewer hydrophobic than Kim and Berg (2000) also observed that as the aggregatehydrophilic colloids in the drainage water is consistent grew, new particles were more likely to contact andwith rapid aggregation of the hydrophobic colloids and immediately attach to the periphery of existing aggre-retention of the aggregates by straining at pore throats gates. This is presumably the most significant mecha-as postulated by Bradford et al. (2002, 2003). Although nism of hydrophobic colloid retention.some aggregate formation was observed for the hydro- The retention of hydrophilic colloids in the very thinphilic colloids, the images suggest that aggregation was film of water at the edge of the menisci is probably thefar less extensive than for the hydrophobic colloids, which result of hydrodynamic processes. Saiers and Lenhartis consistent with the existence of a relatively small re- (2003b) also reasoned that silica colloids were trappedpulsive energy barrier for hydrophilic colloid–colloid in the narrow wedges near the three-phase contact ofinteractions. pore-corner menisci and at the termini of discontinuous

corner water, whereas colloid retention did not occur atAbsence of Film Straining water films. The centrifugal motion within the pendular

rings observed in the videos would tend to force colloidsThe resolution of the visualization equipment made ittoward the AW interface, and then any deviation fromdifficult to prove the absence or presence of film straining,the primary direction of flow could move the particlebecause water films thinner than approximately 1 �mtoward the AWS interface, as shown in Video 1. Undercould not be observed. Despite that, we did not find anylaminar flow, the particle velocity approaches zero nearmicrospheres on the grains itself away from the AWSthe grain surface and increases at distances away frominterface. Thus, the relatively large 5-�m microspheresthe surface. Thus, colloids propelled toward the edgewere not strained by the films where water moves fromof the meniscus would encounter the very slow movingone pendular ring to the next via film flow. Further, wewater near the AWS interface and become immobilized.would expect that, as water covering the grain thins toAlternately, or in addition, the retention in the thinnestfilms during drainage and films approach the diameterportion of the SW interface at the edge of the meniscusof the colloids, capillary pressure would push the col-may represent a form of film straining, with colloids be-loids toward the bulk solution, as shown in Sur and Pakcoming immobilized in films as the film thickness ap-(2001) for suspended films.proaches that of the colloid diameter. Assuming a con-tact angle for water and quartz sand of 30� (Freitas andMobilization of Colloids Deposited at Sharma, 1999), the meniscus thickness will equal that

the Air–Water–Solid Interface of the colloid diameter (5 �m) at a distance of only 5 to7 �m from the edge of the AWS interface. However,During a trial run, the water flow was increased, andwhen the immobilized colloids are aggregated, we findthe accompanying video (Video 3) demonstrates colloidthat for an aggregate 10-fold larger than the primarybehavior consistent with the proposed mechanism ofparticle, the theoretical distance from the AWS inter-Saiers and Lenhart (2003a) that immobilization of theface (≈50 �m) is in better agreement with the observedhydrophilic colloids at the AWS interface was easily re-

versible. In this video, coagulated colloids are deposited distances (e.g., Fig. 8A).

Rep

rodu

ced

from

Vad

ose

Zon

e Jo

urna

l. P

ublis

hed

by S

oil S

cien

ce S

ocie

ty o

f Am

eric

a. A

ll co

pyrig

hts

rese

rved

.

194 VADOSE ZONE J., VOL. 4, FEBRUARY 2005

deposition of Brownian particles in porous media. Ph.D. diss. TheLenhart and Saiers (2002) found that colloid transportJohns Hopkins University, Baltimore, MD.in unsaturated porous media depended principally

Elimelech, M., M. Nagai, C.H. Ko, and J.N. Ryan. 2000. Relativeon the degree of pendular ring discontinuity, pore water insignificance of mineral grain zeta potential to colloid transportvelocity, and the retention capacity of the AW interface. in geochemically heterogeneous porous media. Environ. Sci. Tech-

nol. 34:2143–2148.Since we worked with relatively large colloid sizesFlury, M., J.B. Mathison, and J.B. Harsh. 2002. In situ mobilization(≈5 �m), our interpretation of the transport processes

of colloids and transport of cesium in Hanford sediments. Environ.in the vadose zone is not inconsistent with this descrip- Sci. Technol. 36:5335–5341.tion, but requires the following modifications. Retention Freitas, A.M., and M.M. Sharma. 1999. Effect of surface hydropho-

bicity on the hydrodynamic detachment of particles from surfaces.occurs at the AWS interface, and depends on the reten-Langmuir 15:2466–2476.tion capacity of this triple-point interface. Second, parti-

Grolimund, D., M. Borkovec, K. Barmettler, and H. Sticher. 1996.cle motion within the curved pendular rings may beColloid-facilitated transport of strongly sorbing contaminants in

important in attachment at the AWS interface, as shown natural porous media: A laboratory column study. Environ. Sci.in the video images. Furthermore, the mechanisms for Technol. 30:3118–3123.

Hamaker, H.C. 1937. The London-van der Waals attraction betweencolloid retention at the AWS interface are not clear,spherical particles. Physica 4:1058–1072.but may be related to factors such as pore water motion

Hogg, R., D.S. Cahn, T.W. Healy, and D.W. Fuerstenau. 1966. Diffu-with the pendular rings, low laminar flow velocities near sional mixing in an ideal system. Chem. Eng. Sci. 21:1025–1038.the grain surface, and/or retention of colloids or col- Israelachvili, J. 1992. Interfacial forces. J. Vac. Sci. Technol., A 10:loidal aggregates in the thin water films near the AWS 2961–2971.

Iwamatsu, M., and K. Horii. 1996. Capillary condensation and adhe-interface. The latter is affected by changes in flow regimesion of two wetter surfaces. J. Colloid Interface Sci. 182:400–406.(Video 3). More studies are needed to examine the im-

Jewett, D.G., B.E. Logan, R.G. Arnold, and R.C. Bales. 1999. Trans-portance of these processes. port of Pseudomonas fluorescens strain P17 through quartz sandcolumns as a function of water content. J. Contam. Hydrol. 36:73–89.

Jin, Y., Y.J. Chu, and Y.S. Li. 2000. Virus removal and transport inAPPENDIXsaturated and unsaturated sand columns. J. Contam. Hydrol. 43:

The following videos are available as supplemental mate- 111–128.rial. Use of high speed internet connection is recommended Jordan, R.N., D.R. Yonge, and W.E. Hathhorn. 1997. Enhanced mo-since each video is around 30 MB. bility of Pb in the presence of dissolved natural organic matter.

J. Contam. Hydrol. 29:59–80.Video 1. Retention of hydrophilic colloids at AWS interface.Karathanasis, A.D. 1999. Subsurface migration of copper and zincVideo 2. Mechanisms of colloid retention in unsaturated po-

mediated by soil colloids. Soil Sci. Soc. Am. J. 63:830–838.rous media.Kersting, A.B., D.W. Efurd, D.L. Finnegan, D.J. Rokop, D.K. Smith,Video 3. Instability of colloid retention at AWS interface

and J.L. Thompson. 1999. Migration of plutonium in ground waterwith changing flow rate.at the Nevada Test Site. Nature (London) 397:56–59.

Kim, A.Y., and J.C. Berg. 2000. Fractal aggregation: Scaling of fractalREFERENCES dimension with stability ratio. Langmuir 16:2101–2104.

Kretzschmar, R., M. Borkovec, D. Grolimund, and M. Elimelech.Bauters, T.W.J., D.A. DiCarlo, T.S. Steenhuis, and J.-Y. Parlange.1999. Mobile subsurface colloids and their role in contaminant1998. Preferential flow in water-repellent sands. Soil Sci. Soc.transport. Adv. Agron. 66:121–193.Am. J. 62:1185–1190.

Lenhart, J.J., and J.E. Saiers. 2002. Transport of silica colloids throughBauters, T.W.J., T.S. Steenhuis, D.A. DiCarlo, J.L. Nieber, L.W. Dek-unsaturated porous media: Experimental results and model com-ker, C.J. Ritsema, J.-Y. Parlange, and R. Haverkamp. 2000. Physicsparisons. Environ. Sci. Technol. 36:769–777.of water repellent soils. J. Hydrol. (Amsterdam) 231–232:233–243.

McCarthy, J.F., K.R. Czerwinski, W.E. Sanford, P.M. Jardine, andBradford, S.A., J. Simunek, M. Bettahar, M.Th. van Genuchten, andJ.D. Marsh. 1998. Mobilization of transuranic radionuclides fromS.R. Yates. 2003. Modeling colloid attachment, straining and exclu-disposal trenches by natural organic matter. J. Contam. Hydrol.sion in saturated porous media. Environ. Sci. Technol. 37:2242–2250.30:49–77.Bradford, S.A., S.R. Yates, M. Bettahar, and J. Simunek. 2002. Physi-

McCarthy, J.F., and J.M. Zachara. 1989. Subsurface transport of con-cal factors affecting the transport and fate of colloids in saturatedtaminants—Mobile colloids in the subsurface environment may alterporous media. Water Resour. Res. 38(12):1327. doi:10.1029/2002the transport of contaminants. Environ. Sci. Technol. 23:496–502.WR00134.

Norde, W., and J. Lyklema. 1989. Protein adsorption and bacterialButt, H.J., S. Ecke, L.O. Heim, K. Heiskanen, M. Preuss, R. Raiteri,adhesion to solid-surfaces—A colloid-chemical approach. ColloidsN. Schreithofer, O.I. Vinogradova, and G.E. Yakubov. 2002. Mi-Surf. 38:1–13.crosphere tensiometry to determine the contact angle of individual

Ouyang, Y., D. Shinde, R.S. Mansell, and W. Harris. 1996. Colloid-particles. p. 1–14. In K.L. Mittal (ed.) Contact angle, wettabilityenhanced transport of chemicals in subsurface environments: Aand adhesion, V.2. Proceedings of the Second International Sympo-review. Crit. Rev. Environ. Sci. Technol. 26:189–204.sium on Contact Angle, Wettability and Adhesion, Newark, NJ.

Ryan, J.N., and M. Elimelech. 1996. Colloid mobilization and transport21–23 June 2000. VSP Publ., Zeist, The Netherlands.in groundwater. Colloids Surf., A 107:1–56.Chu, Y., Y. Jin, M. Flury, and M.V. Yates. 2001. Mechanisms of

Saiers, J.E., and J.J. Lenhart. 2003a. Colloid mobilization and trans-virus removal during transport in unsaturated porous media. Waterport within unsaturated porous media under transient-flow condi-Resour. Res. 37(2):253–264. doi:10.1029/2000WR900308.tions. Water Resour. Res. 39(1):1019. doi:10.1029/2002WR001370.Chu Y., Y. Jin, T. Baumann, and M. V. Yates. 2003. Effect of soil

Saiers, J.E., and J.J. Lenhart. 2003b. Ionic-strength effects on colloidproperties on saturated and unsaturated virus transport throughtransport and interfacial reactions in partially saturated porouscolumns. J. Environ. Qual. 32:2017–2025.media. Water Resour. Res. 39(9):1256. doi:10.1029/2002WR001887.Crist, J.T., J.F. McCarthy, Y. Zevi, P. Baveye, J.A. Throop, and T.S.

Schafer, A., P. Ustohal, H. Harms, F. Stauffer, T. Dracos, and A.J.B.Steenhuis. 2004. Pore-scale visualization of colloid transport andZehnder. 1998a. Transport of bacteria in unsaturated porous media.retention in partly saturated porous media. Available at www.J. Contam. Hydrol. 33:149–169.vadosezonejournal.org. Vadose Zone J. 3:444–450.

Schafer, A., H. Harms, and A.J.B. Zehnder. 1998b. Bacterial accumu-de Jonge, H., O.H. Jacobsen, L.W. de Jonge, and P. Moldrup. 1998.lation at the air-water interface. Environ. Sci. Technol. 32(23):Particle-facilitated transport of prochloraz in undisturbed sandy3704–3712.loam soil columns. J. Environ. Qual. 27:1495–1503.

Elimelech, M. 1985. The effect of particle size on the kinetics of Sprague, L.A., J.S. Herman, G.M. Hornberger, and A.L. Mills. 2000.

Rep

rodu

ced

from

Vad

ose

Zon

e Jo

urna

l. P

ublis

hed

by S

oil S

cien

ce S

ocie

ty o

f Am

eric

a. A

ll co

pyrig

hts

rese

rved

.

www.vadosezonejournal.org 195

Atrazine adsorption and colloid-facilitated transport through the van Oss, C.J., R.F. Giese, and P.M. Costanzo. 1990. DLVO and non-DLV interactions in hectorite. Clays Clay Miner. 38(2):151–159.unsaturated zone. J. Environ. Qual. 29:1632–1641.

Veerapaneni, S., J. Wan, and T.S. Tokunaga. 2000. Motion of particlesSteenhuis, T.S., Y.-J. Kim, J.-Y. Parlange, M.S. Akhtar, B.K. Richards,in film flow. Environ. Sci. Technol. 34:2465–2471.K.-J.S. Kung, T.J. Gish, L.W. Dekker, C.J. Ritsema, and S.A. Abur-

Wan, J., and T.K. Tokunaga. 1997. Film straining of colloids in unsatu-ime. 2001. An equation for describing solute transport in field soilsrated porous media: Conceptual model and experimental testing.with preferential flow paths. p. 137–140. In D.D. Bosch and K.W.Environ. Sci. Technol. 31:2413–2420.King (ed.) Preferential flow, water movement and chemical trans-

Wan, J., and J.L. Wilson. 1994a. Visualization of the role of the gas-port in the environment. ASAE, St. Joseph, MI.water interface on the fate and transport of colloids in porousStumm, W. 1977. Chemical interaction in particle separation. Environ.media. Water Resour. Res. 30:11–23.Sci. Technol. 11:1066–1070.

Wan, J., and J.L. Wilson. 1994b. Colloid transport in unsaturatedSur, J., and H.K. Pak. 2001. Capillary force on colloidal particles in a porous media. Water Resour. Res. 30:857–864.freely suspended liquid thin film. Phys. Rev. Lett. 86(19):4326–4329. Williams, C.F., M. Agassi, J. Letey, W.J. Farmer, S.D. Nelson, and M.

Thompson, S.S., M. Flury, M.V. Yates, and W.A. Jury. 1998. Role of Ben-Hur. 2000. Facilitated transport of napropamide by dissolvedthe air–water–solid interface in bacteriophage sorption experi- organic matter through soil columns. Soil Sci. Soc. Am. J. 64:ments. Appl. Environ. Microbiol. 64(1):304–309. 590–594.

Thompson, S.S., and M.V. Yates. 1999. Bacteriophage inactivation Yoon, R.-H., D.H. Flinn, and Y.I. Rabinovich. 1997. Hydrophobicat the air–water–solid interface in dynamic batch systems. Appl. interactions between dissimilar surfaces. J. Colloid Interface Sci.

185:363–370.Environ. Microbiol. 65(3):1186–1190.