Vous êtes sur la page 1sur 8

Influence of a Nonionic Surfactant on the Water Retention Properties

of Unsaturated Soils
Ahmet Karagunduz, Kurt D. Pennell,* and Michael H. Young

ABSTRACT polymers, and castor oil ethoxylates (Hochburg, 1996).

Surfactants are widely used in household products, industrial pro- These surfactants typically comprise 5 to 10% (50 to
cesses and as adjuvants to improve the delivery and effectiveness of 100 g L⫺1 ) of the formulated herbicide product and 0.1
agrochemicals. Due to their amphiphilic nature, surfactants tend to to 0.5% (1 to 5 g L⫺1 ) of the spray tank solution. As a
accumulate at gas-liquid and solid-liquid interfaces, and thus, have result, the amount of surfactant applied to agricultural
the potential to influence water flow and retention in unsaturated lands can be substantial, particularly if repeated herbi-
soils. The objective of this study was to investigate the effects of a cide applications are required throughout the growing
nonionic surfactant, Triton X-100, on the interfacial properties and season.
capillary pressure-water content relationships of F-70 Ottawa sand
Due to their amphiphilic properties, possessing both
and Appling soil. In the presence of surfactant, soil water contents
decreased incrementally as the surfactant concentration was increased
hydrophilic and lipophilic moieties, surfactants tend to
from 0 g L⫺1 up to the critical micelle concentration (CMC) of Triton accumulate or adsorb at gas-liquid and solid-liquid inter-
X-100 (0.15 g L⫺1 ). Over the same surfactant concentration range, faces. Below the surfactant concentration at which mi-
the surface tension of water decreased from 7.2 ⫻ 10⫺2 J m⫺2 to 3.2 ⫻ celles begin to form, the critical micelle concentration
10⫺2 J m⫺2 while solid-liquid contact angle decreased from 40ⴗ to 10ⴗ. (CMC), interfacial adsorption of surfactants typically
No further changes in interfacial properties or soil water characteris- results in a lowering of both the surface tension of water
tics were observed at surfactant concentrations above the CMC. The and the contact angle between the solid and aqueous
experimental results were used to develop and evaluate alternative phases (Rosen, 1989). Therefore, the water holding ca-
scaling approaches to describe concentration dependent changes in pacity of soils is likely to decrease in the presence of
soil water characteristics based on the van Genuchten model. A scaling
surfactants. This phenomenon could increase the depth
factor that incorporated both surface tension and content angle rela-
tionships provided accurate predictions of soil water retention curves of infiltration during spray events, and create a positive
over a range of surfactant concentrations. A simplified form of the pressure gradient between regions of surfactant-rich and
scaling factor also was developed, on the basis of a single fitting surfactant-free soil water (Tschapek et al., 1981, 1991;
parameter without the need for surface tension and contact angle Karkare et al., 1993; Karkare and Fort, 1993). Karkare
data. Although further validation of the simplified scaling factor will and Fort (1993) reported that soil water content was
be required, this approach offers an efficient means to describe the substantially altered in the presence of surfactants, with
effects of concentration dependent changes in interfacial properties higher water saturations observed in surfactant-free soil
on soil water characteristics. at the same pressure as soil containing surfactant. These
findings indicate that surfactant-induced changes of in-
terfacial properties may be an important factor influenc-
S urface-active agents (surfactants) are used exten-
sively in detergents, industrial processes, household
products, and pesticide formulations. As a consequence
ing water flow and retention in unsaturated soils.
Although the potential effects of surfactants on soil
water characteristics have been discussed in the litera-
of their widespread use and frequent resistance to bio- ture, only limited research has been conducted to di-
degradation, surfactants may persist in wastewater treat- rectly quantify the influence of surfactants on capillary
ment systems at relatively high concentrations (Wagener pressure-water saturation relationships. One of the most
and Schink, 1987; Holt and Bernstein, 1992; Zoller, commonly used soil water retention relationships was
1994; Moreno et al., 1994). Subsequent land application developed by van Genuchten (1980):
of wastewater treatment effluent and sludge as a source
of irrigation water and nutrients can result in the indirect ␪ ⫺ ␪r 1
⫽ [1]
release of relatively large quantities of surfactant to the ␪s ⫺ ␪r (1 ⫹ |␣h|n)(1⫺1/n)
environment, particularly in developing countries. In
addition, almost all herbicide formulations require some where ␣ and n are fitting parameters, h is the capillary
form of adjuvant to improve herbicide handling, delivery, pressure head (m), ␪ is volumetric water content (m3
and effectiveness. The most common types of surfactant m⫺3 ), and ␪r and ␪s are the residual and saturated volu-
adjuvants are tallow amine ethoxylates, nonylphenol metric water contents (m3 m⫺3 ), respectively. To account
and octylphenol ethoxylates, alcohol ethoxylates, block for differences in the surface tension of aqueous solu-
tions, the capillary pressure head term (h) in the van
Genucthen (1980) relationship was modified by Smith
Ahmet Karagunduz and Kurt D. Pennell, School of Civil and Environ- and Gillham (1994) using the following scaling factor:
mental Engineering, Georgia Institute of Technology, Atlanta, GA
30332; Michael H. Young, Division of Hydrologic Sciences, Desert ␴0
Research Institute, 755 E. Flamingo Road, Las Vegas, NV 89119. h0 ⫽ h1 [2]
Received 12 June 2000. *Corresponding author (kpennell@ce.gatech. ␴1
where ␴0 and ␴1 are the respective surface tension values
Published in Soil Sci. Soc. Am. J. 65:1392–1399 (2001). of the reference solution (e.g., water) and the solution of

interest (J m⫺2 ), and h0 and h1 are the capillary pressure were to (i) experimentally determine soil water reten-
heads reference solution and the solution of interest tion characteristics as a function of surfactant concentra-
(m), respectively. The scaling relationship given in Eq. tion; (ii) evaluate the utility of several possible scaling
[2] was incorporated into a numerical model to illustrate approaches for use with the van Genuchten (1980) equa-
the potential effects of a 7% n-butanol solution on un- tion; and (iii) develop a simplified scaling approach that
saturated water flow (Smith and Gillham, 1994). More can be employed in the absence of surface tension and
recently, Smith and Gillham (1999) conducted labora- contact angle data. For the experimental phase of this
tory column experiments and numerical modeling stud- study, soil water characteristics of a reference sand (F-
ies to investigate the effects of concentration dependent 70 Ottawa sand) and Appling soil were determined in
changes in surface tension on unsaturated water flow the presence of a representative nonionic surfactant
and solute transport. (Triton X-100). Surface tension and contact angle data
Capillary pressure-saturation relationships for two- were measured independently over a surfactant concen-
phase, organic liquid and water, systems are frequently tration range of 0 to 2.0 g L⫺1. Triton X-100 was selected
derived from soil water retention data (air and water) for study because it is widely used as a detergent and
using a modified form of the Leverett (1941) function as an adjuvant for pesticides and herbicides, exhibits
(e.g., Morrow, 1976; Lenhard and Parker, 1987; Kueper toxicity toward aquatic species and is persistent in the
and Frind, 1991). environment (Narkis and Ben-David, 1985; Ahel et al.,
1994; Renner, 1997).
␴org/H2O cos␻org/H2O
Pc(Se)org/H2O ⫽ Pc(Se)air/H2O [3]
where Pc is the capillary pressure (Pa) defined as Pn ⫺ Experimental Materials
Pw with the subscripts n and w referring to the non- A reference quartz sand, F-70 Ottawa sand, and Appling
wetting and wetting phases, respectively, Se is the effec- soil were used as the solid phases for the soil water retention
tive saturation defined as (Sw ⫺ Swr )/(1 ⫺ Swr ) with the experiments. F-70 Ottawa sand (40-270 mesh) was obtained
subscript wr referring to the wetting phase residual, ␴ from U.S. Silica (Ottawa, IL), and was used as received. Ap-
is the interfacial or surface tension (J m⫺2 ), ␻ is the pling soil was collected from the upper 30 cm of the soil profile,
contact angle, and the subscripts org, air, and H2O refer corresponding to the Ap1 and Ap2 horizons, at the University
to the organic liquid phase, gaseous phase, and aqueous of Georgia Agricultural Experiment Station, located near
phase, respectively. In practice, the contact angle is often Eastville, GA. The soil was classified as a loamy coarse sand
assumed to be zero and thus, cos␻ is equal to unity. of the Appling series (clayey, kaolinitic thermic Typic Haplu-
dult). Prior to use, the Appling soil was air-dried and ground
However, Demond and Roberts (1991) reported that to pass a 2 mm (9-mesh) sieve. The specific surface areas of
the value of cos␻org/H2O was significantly less than unity F-70 Ottawa sand and Appling soil were determined to be
for weakly water-wet systems. In addition, Morrow (1975) 1.5 ⫻ 102 m2 kg⫺1 and 3.5 ⫻ 103 m2 kg⫺1, respectively, on the
and Demond and Roberts (1991) modified Eq. [3] to basis of N2 adsorption at 77⬚K. Total organic carbon (TOC)
account for the effects of surface roughness and inter- analysis indicated that F-70 Ottawa sand contained no detect-
face curvature on the contact angle. The utility of the able organic carbon, while Appling soil contained 7.7 g OC
modified Leverett (1941) function to describe organic kg⫺1.
liquid and water pressure-saturation relationships for Triton X-100 (an octylphenol ethoxylate) was obtained
systems containing a strongly-sorbed cationic surfactant, from the Union Carbide Corp. (Houston, TX), and was used
as received without purification. The hydrophile-liphophile
cetyltrimethylammonium bromide (CTAB), was subse- balance (HLB) of Triton X-100 is 13.5, the average number of
quently demonstrated by Desai et al. (1992) and Demond ethylene oxide (EO) groups is 9.5, and the average molecular
et al. (1994). In an extension of this work, Lord et al. weight is 625 g mole⫺1 (Rosen, 1989). All aqueous surfactant
(1997a,b) investigated the effects of octanoic acid specia- solutions were prepared with deionized-distilled (DI) water
tion and concentration on interfacial properties, contact containing 0.5 g L⫺1 CaCl2 as a background electrolyte and
angle and capillary pressure-saturation relationships for 0.5 g L⫺1 NaN3 to prevent biological activity. The dynamic
two-phase systems (air-water and o-xylene-water). Al- viscosity of aqueous solutions, measured using an RS75 rhe-
though the utility of the modified Leverett (1941) scaling ometer (Haake, Paramus, NJ) at 20⬚C, ranged from of 1.00 ⫻
factor (Eq. [3]) was demonstrated in these studies, such 10⫺3 Pa s for distilled water to 1.05 ⫻ 10⫺3 Pa s for a 10 g L⫺1
solution of Triton X-100. Viscosities determined for the 10 g
an approach has not been evaluated using the van Gen- L⫺1 Triton X-100 solution were constant for shear rates rang-
uchten (1980) relationship (Eq. [1]) for systems con- ing from 200 to 1000 s⫺1, which is indicative of Newtonian be-
taining nonionic surfactants. havior.
Surface tension and contact angle are strongly depen-
dent on surfactant concentration, and have been shown Surface Tension and Contact Angle
to decline sharply until the CMC of the surfactant has
The surface tension of aqueous surfactant solutions, ranging
been reached (Rosen, 1989). For this reason, scaling
in concentration from 0 to 0.75 g L⫺1 Triton X-100, was deter-
factors used to describe soil water retention characteris- mined by the du Nouy ring method. The apparatus consisted
tics in the presence of surfactants must account for con- of a Cahn DCA 322 dynamic contact analysis system (Thermo-
centration dependent properties of the system, and Haake, Paramus, NJ) connected to a personal computer. A
should be applicable at surfactant concentrations above platinum-iridium du Nouy ring was immersed in the test so-
and below the CMC. Thus, the objectives of this study lution and then retracted through the gas-liquid interface. The

surface tension of the solution was calculated from the force measured with a water manometer, while pressures ranging
required to pull the ring through the interface, using a correc- from 10 to 100 kPa (1 to 10 m H2O) were recorded directly
tion factor which incorporates the dimensions of the ring and from a pressure meter installed in the manifold. At each pres-
solution density. Prior to use, the platinum-iridium ring was sure increment, the cell was removed from the apparatus and
placed in a flame to oxidize contaminants. the soil water content was determined gravimetrically by a
Contact angles between aqueous surfactant solutions and digital balance with a resolution of ⫾ 0.01 g (Model PB 3002,
solid surfaces were measured using a goniometer microscope Mettler Toledo, Columbus, OH). Equilibrium conditions were
equipped with a CCD camera and lens assembly (Ramè-Hart, assumed when the weight difference of the cells over two
Inc., Mountain Lakes, NJ). Glass (Superfrost pre-cleaned, consecutive days was less than 0.1 g. To determine the moisture
Fisher Scientific, Pittsburg, PA) and quartz microscope slides content of F-70 Ottawa sand and Appling soil at a pressure
(Technical Glass Products, Inc., Painesville TWP, OH) were of 1500 kPa (150 m H2O), pressure plate measurements were
utilized as the solid surfaces. Contact angle measurements conducted following the methods described by Klute (1986). All
were obtained using a sessile drop method and a captive air soil water retention experiments were performed in duplicate.
bubble method. For the sessile drop method, a quartz or glass Experimental moisture release curve data for each surfac-
microscope slide was positioned horizontally inside a rectangu- tant concentration and soil type were fit to Eq. [1] by a nonlin-
lar quartz chamber (0.075 m length ⫻ 0.075 m height ⫻ 0.025 ear, least squares regression procedure (SYSTAT, ver. 5.03).
cm depth) on two quartz stands (0.01 m in height). The cham- The saturated and residual volumetric water contents, ␪s and ␪r,
ber was filled with approximately 10 mL of water and sealed were determined from the endpoints of each moisture release
with parafilm to maintain high relative humidity within the curve, so that only the parameters ␣ and n were obtained from
chamber, and to minimize evaporative losses. A single drop the fitting procedure. A Wilcoxon Rank Sum test procedure
of aqueous solution was then formed on the microscope slide (SYSTAT, ver. 5.03) was used to evaluate statistical differ-
using a micro syringe (Gilmont, Great Neck, NY). A digitized ences between paired soil water retention curves as a function
image of the drop was captured using a personal computer of surfactant concentration. The ability of several different
connected to the Goniometer apparatus. The contact angle scaling approaches to describe experimental soil water reten-
was obtained from this image by adjusting the cross hair of tion curves as a function of surfactant concentration was evalu-
the telescope to attain tangency with the drop surface. Images ated by the Root Mean Square Error (RMSE):
were captured after an equilibration period of 5 to 10 min,
1 n

which was sufficient for the shape of the drop to stabilize.
Prior to use, the glass and quartz microscope slides were RMSE ⫽ 兺 (Mi ⫺ Pi)2
n ⫺ 1i⫽1
cleaned with methanol and DI water, soaked in a 5 M HNO3
solution overnight, and rinsed thoroughly with DI water. For where, M is the measured value, P is the predicted value and
the captive air bubble method, the microscope slides were
n is the number of measured data points.
placed in the quartz chamber, submerged in test solution, and
allowed to equilibrate for 24 hr. An air bubble was then formed
on the underside of the microscope slide and the contact angle Analytical Methods
was determined following the procedure described above. For
each surfactant concentration and slide, contact angle measure- Aqueous-phase concentrations of Triton X-100 were mea-
ments were performed in duplicate or triplicate. sured with a Hewlett Packard (HP) Model 1100 high perfor-
mance liquid chromatograph (HPLC) equipped with an HP
Hypersil ODS column (12.5 cm length ⫻ 4 mm i.d., 5 ␮m
Soil Water Retention Curves particle size) and diode array detector. Triton X-100 is UV
Soil water pressure saturation relationships were deter- absorbent and was analyzed at a wavelength of 278 nm. The
mined for F-70 Ottawa sand and Appling soil at surfactant con- HPLC system was operated at a flow rate of 60 mL hr⫺1,
centrations ranging from 0 to 2.5 g L⫺1 and 0 to 0.75 g L⫺1, with a sample injection volume of 100 ␮L. The mobile phase
respectively, using a Tempe cell system (Soil Moisture Equip- initially contained 20% acetonitrile and 80% water, increased
ment Corp., Santa Barbara, CA). Each Tempe cell (0.06 m to 100% acetonitrile and 0% water over a period of 3 min,
height ⫻ 0.057 m i.d.) contained a porous ceramic plate with remained constant for the next 6 min, and decreased to the
a bubbling pressure exceeding 1 atm. The Tempe cells were initial values during the final 2 min of each analysis. All aque-
packed with air-dried soil under vibration in 0.01 m incre- ous surfactant samples were analyzed in duplicate.
ments. Prior to water saturation, the packed Tempe cells were
flushed with CO2 to allow for more rapid dissolution of the
entrapped gas phase. The cells were then saturated with aque- RESULTS AND DISCUSSION
ous solutions containing Triton X-100. The saturation process Surface Tension and Contact Angle
was performed using a low-speed piston pump (Model QG-
20, Fluid Metering Inc., Syosset, NY) equipped with a stainless The relationship between the surface tension of water
steel pump head module at a flow rate of 30 mL hr⫺1 (8.3 ⫻ and the concentration of Triton X-100 is shown in Fig.
10⫺9 m3 s⫺1 ) until the influent and effluent surfactant concen- 1. The measured surface tension decreased from 7.2 ⫻
trations were identical. After allowing the cells to equilibrate 10⫺2 to 3.2 ⫻ 10⫺2 J m⫺2 as the concentration of Triton
for one day, several additional pore volumes of surfactant X-100 was increased from 0 to approximately 0.15 g
solution were flushed through columns at a flow rate of 30 L⫺1. Above a surfactant concentration of 0.15 g L⫺1,
mL hr⫺1 (8.3 ⫻ 10⫺9 m3 s⫺1 ). Once complete water saturation
was achieved, pressure was applied to top of the cells in incre-
the surface tension remained essentially constant. This
ments over a pressure range of 0 to 100 kPa (0 to 10 m inflection point corresponds to the CMC of Triton
H2O). A low-pressure regulator, configured in series with a X-100, and indicates the concentration at which surfac-
nullmatic-type regulator (Soil Moisture Equipment Corp., tant monomers aggregate in solution to form micelles
Santa Barbara, CA), was used to apply pressure to the cells. (Rosen, 1989). Above the CMC, the concentration sur-
Pressures ranging from 0 to 10 kPa (0 to 1 m H2O) were factant monomers in solution remains constant, while

Table 1. Contact angle values for deionized water, the reference

electrolyte solution and Triton X-100 solutions.
Contact Angle (ⴗ) Contact Angle (ⴗ)
Triton X-100 Captive air droplet Sessile water droplet
(g L⫺1 ) Glass Quartz Glass Quartz
0† 37.9 (4.3)§ 36.0 (4.3) 42.4 (5.0) 46.8 (1.4)
0‡ 46.9 (4.2) 40.7 (2.3) 45.5 (5.5) 39.3 (1.6)
0.001 38.2 (3.9) 34.3 (2.8) 40.7 (8.4) 35.9 (3.7)
0.005 36.7 (2.1) 32.6 (2.8) 40.6 (3.8) 37.3 (4.4)
0.025 30.2 (3.9) 30.5 (6.5) 29.1 (5.6) 29.8 (8.8)
0.05 24.8 (2.9) 23.8 (5.4) 24.7 (4.0) 19.3 (4.4)
0.075 18.5 (1.7) 16.2 (3.6) 18.0 (1.3) 17.5 (6.2)
0.1 14.2 (2.3) 11.9 (2.7) 14.4 (5.7) 12.8 (1.5)
0.15 11.3 (1.4) 7.4 (1.5) 9.8 (0.5) 10.8 (1.0)
0.2 12.6 (1.1) 11.4 (1.4) 8.6 (2.1) 8.55 (2.6)
0.75 12.7 (1.4) 13.1 (3.6) 6.0 (1.9) 8.9 (2.5)
Fig. 1. Relationship between the surface tension of water and the † Deionized water.
concentration of Triton X-100. ‡ 0.5 g L⫺1 CaCl2 ⫹ 0.5 g L⫺1 NaN3 solution.
§ Numbers inside parentheses represent one standard deviation.

the number of micelles increases with increasing surfac- observed a sharp reduction in contact angle with increas-
tant concentration. Literature values for the CMC of ing concentration of nonylphenol ethoxylate on methyl-
Triton X-100 range from 0.11 to 0.16 g L⫺1 (Ross and ated (hydrophobic) quartz plates. For untreated (hydro-
Olivier, 1959; Kile and Chiou, 1989; Edwards et al., philic) quartz plates, the measured contact angle increased
1991). gradually from 12⬚ to 25⬚ as the surfactant concentration
Contact angle measurements as a function of Triton was increased to the CMC (Warr et al., 1984). The
X-100 concentration are presented in Table 1 for both observed changes in contact angle with surfactant con-
glass and quartz slides. In the absence of surfactant centration were attributed to the formation of an ad-
and electrolyte, measured contact angles ranged from sorbed surfactant monolayer on the quartz surface; the
approximately 36⬚ to 47⬚, depending on the experimen- hydrophobic surfactant moiety oriented outward for
tal method and the solid surface. In general, the addition untreated (hydrophilic) quartz and the hydrophilic sur-
of background electrolyte (0.5 g L⫺1 NaN3 ⫹ 0.5 g L⫺1 factant moiety oriented outward for methylated (hy-
CaCl2 ) caused a slight increase in the measured contact drophobic) quartz. Similar conceptual models were pre-
angle. These findings are consistent with data reported sented by Clunie and Ingram (1983) to describe the
by Wu et al. (1994), who determined the contact angle adsorption of nonionic surfactants on polar and nonpo-
between DI water and glass powder, on the basis of lar surfaces.
thin layer wicking, to be 49⬚ and 66⬚ in the absence and Several other researchers, however, have observed
presence of electrolyte (0.28 g L⫺1 CaCl2 ), respectively. an increase and subsequent decrease in the contact angle
Ethington (1990) also reported contact angles of 29⬚ as the concentration of Triton X-100 increased from 0 g
and 42⬚ for water drops on two different quartz slides. A L⫺1 to the CMC. For example, Gonzalez and Travalloni-
number of researchers, however, obtained much smaller Louvisse (1989) found that the contact angle between
contact angle values, ranging from 2⬚ to 7⬚, for water quartz and Triton X-100 solutions increased from 20⬚
drops and air bubbles on quartz slides (e.g., Desai et at 0.0001 g L⫺1, remained constant at 38⬚ from 0.002 to
al., 1992; Lord et al., 1997b). The larger contact angle 0.02 g L⫺1, and then decreased to 10⬚ at a concentration
values reported herein are most likely the results of of approximately 0.1 g L⫺1. Li and Gu (1985) observed
differences in the experimental method, cleaning proce- a similar relationship between concentration of Triton
dure and properties of the solid surface, and indicate X-100 and contact angle, which increased to a maximum
that the surfaces of the quartz and glass slides were value of 34⬚ at 0.05 g L⫺1 and then decreased to 5⬚ at
slightly hydrophobic or weakly water-wetting. This con- 1.0 g L⫺1. At low surfactant concentrations, adsorption
dition may, in fact, be more representative of natural of Triton X-100 was attributed to interactions between
soils, which are likely to contain organic matter and hydrophilic surfactant moiety and silanol or hydoxyl
would not be subject to rigorous cleaning. For example, groups on the quartz surface. In this scenario, the hy-
Bachmann et al. (2000) obtained advancing contact drophobic tail of the adsorbed surfactant monomer
angles values ranging from 22⬚ to 94⬚ for a sandy soil, would be oriented outward, thereby increasing the con-
using a modified sessile drop approach in which a glass tact angle with water. When the concentration of Triton
slide was covered with a single layer of soil particles. X-100 was further increased, the authors proposed that
In the presence of Triton X-100 the contact angle interactions between the hydrophobic moieties of sur-
decreased from approximately 40⬚ to 10⬚ as the surfac- factant monomers led to the formation of an adsorbed
tant concentration increased from 0.001 to 0.15 g L⫺1, surfactant bilayer. Under these conditions, the hydro-
regardless of the experimental method. At concentra- philic moiety of the surfactant would be oriented out-
tions greater than 0.15 g L⫺1, which corresponds to the ward, thereby reducing the contact angle with water.
measured CMC of Triton X-100, further reductions in These experimental data and proposed explanations are
contact angle were not observed. These results are con- consistent with contact angle data obtained in the pres-
sistent with data reported by Warr et al. (1984), who ence of cationic surfactants (e.g., Desai et al., 1992).

Soil Water Retention Curves observed, with the exception of F-70 Ottawa sand at a
surfactant concentration of 1.5 g L⫺1.
Soil water retention curves for F-70 Ottawa sand and
Appling soil measured at Triton X-100 concentrations These results demonstrate that the presence of Triton
ranging from 0 to 2.5 g L⫺1 and 0 to 0.75 g L⫺1, respec- X-100 significantly altered soil water retention proper-
tively, are shown in Fig. 2. Duplicate capillary pressure- ties of both F-70 Ottawa sand and Appling soil. The
soil water saturation experiments (Tempe cells) were observed changes in soil water characteristics corre-
conducted for each soil and surfactant concentration, spond to independently measured reductions in surface
although average values are plotted in Fig. 2 due to the tension and contact angle as a function of increasing
large number of data points. In the presence of Triton Triton X-100 concentration below the CMC. To account
X-100, soil water retention curves for both F-70 Ottawa concentration dependent changes in surface tension and
sand and Appling soil were significantly different (P ⬍ contact angle on soil water characteristics, a scaling fac-
0.05) from those obtained for the reference electrolyte tor similar in form to Eq. [3] was employed:
solution, with the exception of F-70 Ottawa sand at a
surfactant concentration of 0.05 g L⫺1. For example, at h0 ⫽ 冢␴␴ cos␻


1 [5]
a negative pressure head of 0.23 m H2O (2.3 kPa) the
volumetric water content of F-70 Ottawa sand decreased where the subscripts 0 and 1 refer to the reference solu-
by 45.5% at a surfactant concentration of 0.750 g L⫺1, tion, in this case DI water ⫹ 0.5 g L⫺1 NaN3 ⫹ 0.5 g L⫺1
compared with water content values obtained in the CaCl2, and the aqueous surfactant solution, respectively.
absence of surfactant (Fig. 2a). Similar behavior was The pressure head scaling factor given in Eq. [5] was
observed for Appling soil, although the reductions in substituted for the pressure head (h) in the van Genuch-
volumetric water content were not as dramatic as those ten (1980) equation (Eq. [1]), to obtain the following
observed for F-70 Ottawa sand (Fig. 2b). In all but capillary pressure-volumetric water content expression.
two cases, soil water retention curves determined in the ␪ ⫺ ␪r 1
presence of Triton X-100 were significantly different ⫽ [6]
冢 兩 兩冣
n (1⫺1/n)
␪s ⫺ ␪ r ␴ cos␻0
(P ⬍ 0.1) from one another at surfactant concentrations 1⫹ ␣ 0 h1
below the CMC (0.15 g L⫺1 ). For surfactant concentra- ␴1cos␻1
tions above the CMC, however, no significant differ-
The relationship between the surface tension-contact
ences (P ⬎ 0.05) in soil water retention curves were
angle scaling factor (Eq. [5]) and the concentration of
Triton X-100 below the CMC is shown graphically in
Fig. 3. Contact angle data obtained for captive air bub-
bles on glass microscope slides were used to calculate
the scaling factors shown in Fig. 3. Although a slightly
non-linear relationship was observed, as a first approxi-
mation, the scaling factor was assumed to be linear over
a surfactant concentration range of 0 g L⫺1 to the CMC
(0.15 g L⫺1 ). By definition, the pressure head scaling
factor is equal to unity in the absence of surfactant, and
thus, the scaled van Genuchten (1980) relationship (Eq.
[6]) can be expressed as:
␪ ⫺ ␪r 1
⫽ [7]
␪s ⫺ ␪r (1 ⫹ |␣(1 ⫹ ␤C)h1|n)(1⫺1/n)
where C is the aqueous concentration of Triton X-100
below the CMC, and is equal to the CMC at concentra-

Fig. 2. Soil water retention curves for (A) F-70 Ottawa sand and (B) Fig. 3. Relationship between the surface tension-contact angle scaling
Appling soil as a function of Triton X-100 concentration. factor and the concentration of Triton X-100.

Table 2. van Genuchten parameters for soil water retention curves at each concentration of Triton X-100.
Triton X-100 F-70 Ottawa Sand Appling Soil
(g L⫺1 ) ␪ s† ␪ r† n‡ ␣‡ ␪s † ␪r† n‡ ␣‡
0 0.3520 0.00008 3.6762 4.1966 0.4108 0.00726 1.5470 4.3668
0.0025 – – – – 0.4185 – 1.4313 7.4135
0.05 0.3570 0.00008 4.7688 4.5561 0.4291 0.00663 1.4892 6.4652
0.075 0.3568 0.00005 5.3164 5.5323 0.4207 0.00730 1.4696 7.7002
0.1 0.3563 0.00005 5.8925 5.5301 0.4293 0.00694 1.4850 6.7037
0.2 0.3565 0.00012 3.6463 6.1259 0.4193 0.00524 1.4620 8.8079
0.75 0.3530 0.00010 3.5607 6.3955 0.4130 0.00534 1.4136 10.5305
1.5 0.3495 0.00007 3.6070 6.0542 – – – –
2.5 0.3482 0.00007 3.6317 5.5155 – – – –
† Measured value.
‡ Fitted parameter.

tions equal to and above the CMC, and ␤ is the slope data. Predictions of soil water characteristics using the
of the linear regression shown in Fig. 3. As can be seen final approach (Case V), in which a single scaling factor
from Eq. [7], the term ␣(1 ⫹ ␤C) is equal to ␣ at a was employed, were similar to those obtained for Case
surfactant concentration 0 g L⫺1, and is equal to ␣CMC IV, even though Case V required one less fitting pa-
at the CMC. Thus, if water retention characteristics of rameter.
a soil are known at the CMC of a surfactant, ␤ can be Results obtained for Appling soil (Fig. 5a and 5b)
estimated by fitting the experimental data to Eq. [7]. A were similar to those observed for F-70 Ottawa sand,
␤ value of 4.47 was obtained from the linear regression where predictions based on Case II and Case III led to
shown in Fig. 3. overestimation and underestimation of measured soil
water contents, respectively. The degree of underesti-
Comparison of Scaling Approaches mation for Appling soil (Case III), however, was much
Measured soil water retention curves were described less than that observed for F-70 Ottawa sand. Case IV
using both un-scaled and scaled van Genuchten (1980) and Case V predictions for Appling soil were virtually
relationships. For the reference case (Case I), each soil identical, and could not be distinguished (Fig. 5b). These
water retention curve was fit to the van Genuchten findings demonstrate that the single fitting parameter
(1980) equation to obtain unique values of ␣ and n (data (␤), Eq. [7], provided excellent fits to the experimental
shown in Table 2). In the second approach (Case II), data without the need for concentration dependent sur-
the effects of surfactant concentration were ignored and face tension and contact angle data.
soil water retention curves were based on the values of Results of the RMSE analysis, based on measured
␣ and n obtained for the reference electrolyte solution. and predicted pressure-saturation data for F-70 Ottawa
This approach neglected any dependence of surface ten-
sion and contact angle on surfactant concentration, and
represents a worst-case scenario in which a single water
retention relationship was applied, regardless of surfac-
tant concentration. In the third approach (Case III),
the pressure head was scaled using only the measured
surface tension ratio (␴0/␴1), in a manner similar to that
employed by Smith and Gillham (1994, 1999) for a 7%
butanol solution. In the fourth approach (Case IV), both
the surface tension and contact angle relationships were
incorporated in the scaling factor as shown in Eq. [5].
The final approach (Case V) was based on the use of
a single scaling factor, ␤, determined from soil water
retention curves obtained in the absence of surfactant
and at the CMC of the Triton X-100 (Eq. [7]).
Comparisons between measured data and predicted
soil water retention curves for F-70 Ottawa sand at
surfactant concentrations below (0.075 g L⫺1 ) and above
(0.75 g L⫺1 ) the CMC of Triton X-100 are shown in Fig.
4a and 4b, respectively. When a scaling factor was not
included (Case II), predictions of soil water content
were much larger than the measured values. In contrast,
the use of a scaling factor that included only the surface
tension ratio (Case III) resulted in underestimation of
soil water content values. Incorporating the combined Fig. 4. Predicted and measured soil water retention curves for F-70
surface tension-contact angle scaling factor (Case IV) Ottawa sand at Triton X-100 concentrations of (A) 0.075 g L⫺1
resulted in greatly improved fits to the experimental and (B) 0.75 g L⫺1.

Fig. 6. Root Mean Square Error (RMSE) values based on predicted

and measured soil water retention data for (A) F-70 Ottawa sand
Fig. 5. Predicted and measured soil water retention curves for Ap- and (B) Appling soil.
pling soil at Triton X-100 concentrations of (A) 0.075 g L⫺1 and
(B) 0.75 g L⫺1.
ever, the use of a scaling factor that incorporated surface
sand and Appling soil, are presented in Fig. 6a and 6b, tension and contact angle dependence on surfactant
respectively, for all surfactant concentrations consid- concentration (Case IV) accurately described soil water
ered. As expected, the reference approach (Case I), in retention curves for F-70 Ottawa sand and Appling soil
which van Genuchten equation was directly fit to each over a wide range of surfactant concentrations. In the
experimental data set, yielded the lowest RMSE values. absence of surface tension and contact angle data, a
In general, Case II (no scaling) and Case III (scaling simplified scaling factor was developed on the basis of
with surface tension only) yielded the highest RMSE a single fitting parameter (␤), derived from pressure-
values. In most cases, use of the combined surface ten- water saturation data measured in the absence of surfac-
sion-contact angle scaling factor (Case IV) gave lower tant and at the CMC of the surfactant. The simplified
RMSE values than were observed for Case II and III. scaling approach provided an efficient means to predict
Furthermore, the RMSE values obtained for Case IV soil water characteristics in the presence of Triton
and Case V (single fitting parameter, ␤) were similar, X-100, and may be applicable to other solute-soil sys-
indicating that the latter approach provided accurate tems subject to concentration dependent changes in sur-
predictions of soil water characteristics for these surfac- face tension and contact angle.
tant-soil systems.
CONCLUSIONS The authors thank Ms. Lee Ogden, Daniel B. Warnell
School of Forest Resources at the University of Georgia, for
The presence of a nonionic surfactant (Triton X-100), conducting the 1500 kPa pressure plate analysis, and the Union
even at relatively low concentrations, strongly influ- Carbide Corporation for supplying Triton X-100. Funding for
enced interfacial properties and soil water characteris- this research was provided in part by the Herty Foundation,
tics of F-70 Ottawa sand and Appling Soil. When the Research and Development Center, Traditional Industries Pro-
van Genuchten (1980) relationship obtained in the ab- gram in Pulp and Paper. The content of this publication does
sence of surfactant was employed without a concentra- not necessarily represent the views of the Foundation and no
tion-dependent scaling factor (Case II), predicted soil endorsement should be inferred.
water contents were substantially larger than values
measured in the presence of surfactant. A scaling ap- REFERENCES
proach based only on changes in surface tension as a Ahel, M., W. Giger, and C. Schaffner. 1994. Behavior of alkylphenol
function of surfactant concentration (Case III) did not polyethoxylate surfactants in the aquatic environment-II. Oc-
curance and transformation in rivers. Water Res. 28:1143–1152.
substantially improve the predictive capability of the Bachmann, J., R. Horton, R.R. van der Ploeg, and S. Woche. 2000.
van Genuchten (1980) relationship, and resulted in un- Modified sessile drop method for assessing initial soil-water contact
derestimation of measured soil water contents. How- angle in sandy soil. Soil Sci. Soc. Am. J. 64:564–567.

Clunie, J.S., and B.T. Ingram. 1983. Adsorption of nonionic surfac- Influence of organic acid solution chemistry on subsurface transport
tants. p. 105–152. In G.D. Parkitt and C.H. Rochester (ed.) Adsorp- properties. 1. Surface and interfacial tension. Environ. Sci. Tech-
tion from Solution at the Solid-Liquid Interface. Academic Press, nol. 31:2045–2051.
New York. Lord, D.A., A.D. Demond, A. Salehzadeh, and K.F. Hayes. 1997b.
Demond, A.H., and P.V. Roberts. 1991. Effect of interfacial forces Influence of organic acid solution chemistry on subsurface transport
on two-phase capillary pressure-saturation relationships. Water properties. 2. Capillary pressure-saturation. Environ. Sci. Tech-
Resour. Res. 27:423–437. nol. 31:2052–2058.
Demond, A.H., F.N. Desai, and K.F. Hayes. 1994. Effect of cationic Morrow, N.R. 1975. The effects of surface roughness on contact angle
surfactants on organic liquid-water capillary pressure-saturation with special reference to petroleum recovery. J. Can. Petrol. Tech-
relationships. Water Resour. Res. 30:333–342. nol. 14:42–53.
Desai, F.N., A.H. Demond, and K.F. Hayes. 1992. Influence of surfac- Morrow, N.R. 1976. Capillary pressure correlations for uniformly
tant sorption on capillary pressure-saturation relationships. p. 133– wetted porous media. J. Can. Petrol. Technol. 15:49–69.
148. In D. Sabatini and R. Knox (ed.) Transport and Remediation Moreno, A., J. Ferrer, F.R. Bevia, D. Prats, B. Vazques, D. Zarzo.
of Subsurface Contaminants: Colloidal, Interfacial and Surfactant 1994. LAS monitoring in a lagoon treatment plant. Water Re-
Phenomena. American Chemical Society, Washington, D.C. sour. 28:2183–2189.
Edwards, D.A., R.G. Luthy, and A. Lui. 1991. Solubilization of polycy- Narkis, N., and B. Ben-David. 1985. Adsorption of non-ionic surfac-
clic aromatic hydrocarbons in micellar nonionic surfactant solu- tants on activated carbon and mineral clay. Water Resour. 19:815–
tions. Environ. Sci. Technol. 25:127–133. 824.
Ethington, E.F. 1990. Interfacial contact angle measurements of water, Renner, R. 1997. European bans on surfactant trigger transatlantic
mercury and 20 organic liquids on quartz, calcite, biotite, and Ca- debate. Environ. Sci. Technol. 31:316A–320A.
montmorillonite substrates. USGS-OFR 90-409. Rosen, M.J. 1989. Surfactants and Interfacial Phenomena. John Wi-
Gonzalez, G., and A.M. Travalloni-Louvisse. 1989. The effect of Tri- ley & Sons, New York.
ton X-100 and ethanol on the wettability of quartz. Langmuir Ross, S., and J.P. Olivier. 1959. A new method for the determination
5:26–29. of critical micelle concentration of un-ionized associated colloids in
Hochberg, E.G. 1996. The market for agricultural pesticide inert ingre- aqueous or in non-aqueous solution. J. Phys. Chem. 63:1671–1674.
dients and adjuvants. p. 203–208. In C.L. Foy and D.W. Pritchard Smith, J.E., and R.W. Gillham. 1994. The effect of concentration-
(ed.) Pesticide Formulation and Adjuvant Technology, CRC Press, dependent surface tension on the flow of water and transport of
Boca Raton, FL. dissolved organic compounds: A pressure head-based formulation
Holt, M.S., and S.L. Bernstein. 1992. Linear Alkylbenzenes in sewage and numerical model. Water Resour. Res. 30:343–352.
sludges and sludge amended soils. Water Res. 26:613–624. Smith, J.E., and R.W. Gillham. 1999. Effect of solute concentration-
Karkare, M.V., H.T. La, T. Fort. 1993. Criteria for effectiveness of dependent surface tension on unsaturated flow: Laboratory and
surfactants as water moving agents in unsaturated wet sand. Lang- sand column experiments. Water Resour. Res. 35:973–982.
muir 9:1684–1690. Tschapek, M., C. Wasowski, and R.M. Torres Sanchez. 1981. Tangen-
Karkare, M.V., and T. Fort. 1993. Water movement in unsaturated tial force and water migration in unsaturated disperse material.
porous media due to pore size and surface tension induced capillary Colloids and Surfaces 3:295–298.
pressure gradients. Langmuir 9:2398–2403. Tschapek, M., C. Wasowski, and S. Falasca. 1991. Boundary water
Keuper, B.H., and E.O. Frind. 1991. Two-phase flow in heterogeneous (solution) as a medium for spreading of surfactants. Colloids and
porous media. 2. Model application. Water Resour. Res. 27:1059– Surfaces 55:1–8.
1070. van Genuchten, M.Th. 1980. A closed-form equation for predicting
Kile, D.E., and C.T. Chiou. 1989. Water solubility enhancements of the hydraulic conductivity of unsaturated soils. Soil Sci. Soc. Am.
DDT and trichlorobenzene by some surfactants below and above J. 44:892–898.
the critical micelle concentration. Environ. Sci. Technol. 23:832– Wagener, S., and B. Schink. 1987. Anaerobic degradation of nonionic
838. and ionic surfactants in enrichment cultures and fixed bed reactors.
Klute, A. 1986. Water Retention: Laboratory Methods. p. 635–660. Water Resour. 21:615–622.
In A. Klute (ed.) Methods of soil analysis. Part 1. Physical and Warr, C.G., P. Scales, F. Grieser, J.R. Aston, D.R. Furlong, and T.W.
mineralogical methods. 2nd ed. Agron. Monogr. No. 9. ASA and Healy. 1984. Polydisperse non-ionic surfactants: Their solution
SSSA, Madison, WI. chemistry and effect on the wettability of solid surfaces. p. 1329–
Lenhard, R.J., and J.C. Parker. 1987. Measurement and prediction 1338. In K.L. Mittal and B. Lindman (ed.) Surfactants in Solution,
of saturation-pressure relationships in three-phase porous media Volume 2. Plenum Press, New York.
systems. J. Contam. Hydrol. 1:407–424. Wu, W., R.F. Giese, Jr., and C.F. van Oss. 1994. Linkage between
Leverett, M.C. 1941. Capillary behavior in porous solids. Trans. Am. x–potential and electron donicity of charged polar surfaces. 1. Im-
Inst. Min. Metall. Pet. Eng. 142:152–169. plications for mechanisms of flocculation of particle suspensions
Li, W., and T. Gu. 1985. Equilibrium contact angles as a function of with plurivalent counterions. Colloids Surfaces 89:241–252.
the concentration of nonionic surfactants on quartz plate. Colloid Zoller, U. 1994. Non-ionic surfactants in reused water: Are activated
Polymer Sci. 263:1041–1043. sludge/soil aquifer treatments sufficient? Water Resour. 28:1625–
Lord, D.A., K.F. Hayes, A.D. Demond, and A. Salehzadeh. 1997a. 1629.