SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 2 June 2005
Published in Soil Sci Soc Am J 69:1173-1184 (2005)
DOI: 10.2136/sssaj2004.0081
© 2005 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Ranville, J. F.
Right arrow Articles by Beckett, R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Ranville, J. F.
Right arrow Articles by Beckett, R.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Ranville, J. F.
Right arrow Articles by Beckett, R.
Related Collections
Right arrow Watershed-Scale Studies
Right arrow Colloids
Right arrow Colloid-Facilitated Transport
Right arrow Soil Methods/Instrumentation

Soil Mineralogy

Particle-Size and Element Distributions of Soil Colloids

Implications for Colloid Transport

James F. Ranvillea,c, David J. Chittleboroughb,* and Ronald Becketta

a CRC for Freshwater Ecology, Water Studies Centre, Department of Chemistry, Monash University, Clayton, Victoria, Australia
b School of Earth and Environmental Sciences, University of Adelaide, Glen Osmond, South Australia, Australia 5064
c Currently at Dep. of Chemistry and Geochemistry, Colorado School of Mines, Golden, CO 80401

* Corresponding author (David.Chittleborough{at}adelaide.edu.au)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Transport of colloids (diameter < 1 µm) through soil has implications for both horizon development and contaminant mobility. Colloid physical properties (size, shape), mineralogy, and surface chemistry are known to influence transport and deposition. Sedimentation field-flow fractionation-inductively coupled plasma–mass spectrometry (Sd FFF-ICP–MS) was used to examine size and element composition distributions of colloids collected from the profile of a texture-contrast soil located in South Australia. The morphology of the colloids was also examined by transmission electron microscopy (TEM). Colloids collected from the soil matrix, the source of mobile colloids, had significant differences in the size distributions among the three horizons sampled. These differences were consistent with long-term soil formation processes. Colloids mobilized by rainfall, which were collected from overland flow and infiltration through the soil profile, all showed very similar size distributions. This is consistent with the presence of preferential flow paths and suggests that colloids, and colloid-associated contaminants, can be transported rapidly through the vadose zone with minimal interaction with the soil matrix. The Sd FFF-ICP–MS analysis showed variation in element ratios (Fe/Al, Mg/Al), which were used to detect changes in surface coatings and mineralogy over the colloid size range. The study demonstrated the utility of Sd FFF-ICP–MS for examining the influence of colloid size on element composition and on elucidating colloid transport processes in soils.

Abbreviations: DOC, dissolved organic carbon • ICP–MS, inductively coupled plasma–mass spectrometry • PSD, particle-size distribution • Sd FFF, sedimentation field-flow fractionation • TEM, transmission electron microscopy • XRD, x-ray diffraction


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE CONDITIONS UNDER which colloids move through soils, regolith, fractured rock systems, and aquifers have been the subject of considerable interest in recent years because of the possibility that the transport of contaminants may be enhanced by their attachment to these colloids (McCarthy and Zachara, 1989; McDowell-Boyer et al., 1986). For a number of years inferences have been made that the unexpectedly rapid movement of contaminants was due to colloid-mediated transport. Recent research has confirmed that colloids can indeed act in this manner. In a study of Cu and Zn movement through undisturbed soil columns Karathanasis (1999) found that phyllosilicate colloids, especially those of high negative charge and organic C content, substantially enhanced the movement of Cu and Zn. Grolimund et al. (1996), also in a column study, demonstrated that in situ generated colloidal particles, principally vermiculite, kaolinite, illite, and muscovite, can transport large amounts of Pb. Despite these and other specific examples, the degree to which colloids can influence the long-term transport of contaminants is still debated (Honeyman and Ranville, 2002).

Apart from contaminants adsorbed to their surfaces, the colloids themselves may pose particular environmental problems. Dissolved organic carbon (DOC), originating principally in the O and A horizons, is a major problem in water treatment because of its reaction with additives such as chlorine and chloramines, used to make water for domestic purposes potable (Hine and Bursill, 1985). Phyllosilicates of submicron size can discolor water supplies and render such supplies less acceptable to the consumer. Soil-derived biocolloids such as bacteria and viruses may also pose a health risk if they become mobile: the literature on this is quite extensive (e.g., Kim and Corapcioglu 1996, Jin et al. 1997).

Movement of colloids is a three-stage process involving dispersion (or mobilization), transport and deposition. Much is known about dispersion processes in surface soils because aggregate stability has a large influence on: seedling emergence and plant viability, the resistance of soils to erosion, and water permeability (Quirk and Schofield, 1955). Less is known about colloid mobilization occurring at depth. The processes influencing colloid transport and deposition in soils and ground water are likewise poorly understood (Kretzschmar et al., 1999; Soil Survey Staff, 1999). One reason for this is that the significance of colloids has only recently been appreciated. In many aquifer and ground water studies, colloids have been considered artifacts resulting from sampling or other soil disturbance. Also, it was considered unfeasible to study these processes directly because colloid transport is slow compared with solute movement. For example, B soil horizons having very strong texture contrast, which arise from vertical clay transport (i.e., illuviation or lessivage), have taken 30000 to 100000 yr to form (Walker and Chittleborough, 1986).

Inferences about the processes of mobilization and deposition have been made in two ways. One is based on the nature and properties of clays in stream water but inferences of soil colloid transport processes based on this approach may be misleading. The other is based on the observation of clay coatings on voids, skeleton grains, and peds/aggregates. These coatings, known as argillans (Brewer, 1960, 1976), may form in ways other than deposition and care must be taken in interpreting their origins (Bullock et al., 1985). The assumption is that the nature and properties of the clay are an indication of the nature and properties of the clay that has moved. Recent experimental and observational research has thrown doubt on the validity of this assumption. Using capillary wick lysimeters emplaced at the A/B and B/C horizon boundaries of an Alfisol, Varcoe (1995) found that suspensions were enriched in illite compared with the soil matrix. In a study of petrosols, McSweeney and Fastovsky (1987) excised material along root traces, slickensides, and peds. The coatings, intercalations and infillings due to illuviation could not be definitively separated from those created by weathering and structural reorganization, caused by the forces of shrinking and swelling. Kaplan et al. (1997) also demonstrated mineralogical differences between mobile and immobile colloids.

In a study of changes in the nature of clays in B horizons with age, Walker and Hutka (1979) showed that there was a progressive increase in fine clay (<0.2 µm) in the B horizon whereas coarse clay (0.2–2 µm) dominated the clay fraction of A and C horizons. The argillans reflected the particle size of the B horizons in so far as they also were dominated by fine clay. Other pedologists have also observed that argillans and papules (argillaceous concentrations) have high proportions of fine and very fine clay (<0.08 µm) and the conventional wisdom is that illuviation is a process involving the transport of fine clay. A different view comes from the few studies designed to sample the suspensions actually moving. In a study of sediment flux through a hillslope plot, Pilgrim and Huff (1983) recorded that fine silt of size 4 to 8 µm was the predominant component moving, presumably along macropores. Cesium-137 analysis confirmed that the silts originated from the A1 horizon. Other workers have noted that particles of silt and coarse clay size were the dominantly mobile components (Kovenya et al., 1972; James and Chrysikopoulos, 2000). Mobile colloid size, composition, and abundance are highly dependant on hydrologic conditions (Kaplan et al., 1993; Ryan et al., 1998). This is especially true under unsaturated conditions where the presence of an air–water interface during the initial wetting of the soil can enhance colloid mobility (El-Farhan et al., 2000).

On theoretical and experimental grounds there is good reason to expect fractionation based on surface charge and size of the mobile colloids during their passage through pores. Kretschmar et al. (1995), in a laboratory study of colloid movement using columns of aquifer material, showed that the nature of the colloid exiting a profile is highly dependent on surface charge. Kaolinite coated with natural organic matter had a much higher mobility (i.e., was less likely to be adsorbed at the channel wall surface) than uncoated kaolinite and halloysite. Surface charge effects on colloid transport are also related to the water composition (Kaplan et al., 1996).

Most studies of colloid transport in ground water primarily focus on the surface chemical properties of the dispersed colloids and the aquifer matrix solids. Predictions of colloid transport, especially for simple laboratory systems (i.e., monodisperse colloids and monomineralogic columns), have been made by using DLVO (Derjaguin, Landau, Vervey, Overbeek) theory (Verwey and Overbeek, 1948), which describes the attractive and repulsive interactions that govern both the aggregation and adherence of colloids to surfaces (Ryan and Elimelich, 1996). These interactions involve the surface charge of the colloids, which depends on both the bulk composition of colloids and the presence of surface coatings, Fe oxide, and organic matter coatings being the most important in most soils. It has been observed that DLVO theory satisfactorily predicts colloid deposition rates when repulsive interactions are absent or when net attractive interactions occur due to the presence of oppositely charged surfaces (Elimelech, 1991). Measurements of colloid zeta potential ({zeta}) are useful in transport studies (Ryan et al., 1999) whereas these measurements for the porous media can be misleading (Elimelech et al., 2000). Elimelech et al. (2000) found that colloid deposition can be controlled by small patches of material that have charge favorable for colloid deposition, but do not contribute to the overall measured {zeta}.

The influence of particle size on transport in saturated porous media arises primarily from three effects that were initially used to describe particle behavior in sand filters (Yao et al., 1971). The influence has been further elaborated for the transport of colloids in ground waters by numerous workers as reviewed by Ryan and Elimelech (1996). For larger particles, greater than a few microns, gravitational settling and straining by pores smaller than the dimension of the particle, combine to reduce the mobility of these particles. For very small colloids, less than a few tenths of a micrometer, the high diffusion coefficients of the colloids allow them to rapidly contact the surface of aquifer grains where they may be captured by attractive Van der Waals forces. This adhesion can be prevented by strong electrostatic repulsion. Thus, by inference, colloids of intermediate size are generally predicted to be most mobile. The exact size range of greatest mobility currently is not known due to a lack of data, especially for natural systems where polydispersity of colloid size and the chemical heterogeneity of both the colloids and the aquifer matrix confound simple models. Other effects such as particle shape have received even less attention (Mills et al., 1995).

An additional complication in the study of soil colloid transport lies in the presence of macropores in many soil types, especially texture contrast soils (Chittleborough, 1992). Macropores can act as relatively large conduits for water flow and thus enhance the transport of solutes and colloids. Models that may account for the transport of colloids through the pores between aquifer grains are likely to fail for macropore flow. Often these macropores are coated with clay materials, termed cutans that may have very different properties than the bulk soil. Because most water movement in these soils is through the macropores, solute and colloid transport is more likely to be influenced by interaction with the cutans than with the bulk soil matrix.

The objectives of the research were two-fold. The first was to characterize the particle size and elemental chemistry of mobile colloids in a system demonstrating significant macropore flow. In this way it was hoped that the source of these colloids, and the factors influencing their transport, could be deduced. A number of difficulties were associated with achieving this objective, especially the small amount of colloid available for analysis. Of the analytical techniques available, FFF (Giddings, 1993) offers the best prospect for rapid, high-resolution size separation. Due to its large size-based selectivity, FFF is generally considered to be a high-resolution technique. It has recently been applied to the characterization of environmental colloids in general and soil colloids in particular (Chittleborough et al., 1992; Beckett and Hart, 1993; Chen et al., 2001). Inductively coupled plasma–mass spectrometry offered the best method for elemental analysis. The direct coupling of these two techniques has been shown to be an effective means of determining elemental composition over the colloid size range (Murphy et al., 1993; Ranville et al., 1999). The second objective of this study was to further test the efficacy of FFF-ICP–MS, which is still a relatively new technique of environmental colloid analysis. Results of the FFF-ICP–MS analysis were compared with electron microscopy, which has been used extensively to observe the physical characteristics of soil colloids.

The strategy adopted was to collect colloidal samples from various positions in a soil horizon. The samples were from the soil matrix, the linings of soil macropores (cutans), and flow samples collected in the field during a single rainfall event. The size and element composition distributions were determined using FFF-ICP–MS and from their similarities and differences the relative significance and pathways of soil colloid transport were inferred.

The theory behind FFF and the specific sedimentation FFF subtechnique (Sd FFF-ICP–MS) that were developed for this project are described in detail elsewhere (Ranville et al., 1999). Briefly, FFF consists of a suite of high-resolution elution techniques, which, depending on the type of field applied and mode of operation, allow separation and sizing of macromolecules, submicron colloids, and particles of diameter 1 to 50 µm. This paper presents the results of FFF separations based on the sedimentation FFF subtechnique whereby a centrifugal field is produced by placing the channel in a centrifuge. The FFF appears similar to chromatography except that separation is based on physical forces arising from an applied field that distributes the injected particles of the sample into characteristic positions across a thin channel and with the accompanying migration down the length of the channel due to carrier flow.

The data obtained from a Sd FFF instrument are usually presented in the form of a plot of detector response, in this case a UV-visible spectrophotometer ({lambda} = 254 nm) and an ICP–MS, versus elution time or volume. This plot is termed a fractogram and is analogous to a chromatogram. For most colloid studies the UV detector records light attenuation, which is mainly because of particle scattering, not absorption. Detector response is used as an indirect measure of the mass of sample eluting at a given time during the FFF separation.

The equivalent spherical diameter at any given elution time or volume can be computed, assuming a constant particle density, using FFF theory. For the soil samples studied in this work, which are dominated by aluminosilicate minerals, a density of 2.5 g cm–3 was used. The full details of computing diameter from retention time are given elsewhere (Ranville et al., 1999), but briefly, if the field strength is constant the approximate form of the fundamental equation for this calculation is:

[1]
where k is Boltzmans constant, T is the absolute temperature, tr is the colloid elution time, {omega} is the centrifuge speed (rad s–1), r is the centrifuge radius, {Delta}{rho} is the density difference between the particle and the carrier solution, w is the channel thickness, and t° is the time required to pass one void volume through the channel. The more complicated case of field decay runs is discussed by Williams (2000). In this process the x-axis is converted from time or volume to diameter. The UV data on the y-axis is converted to relative mass. The resulting plot is a particle-size distribution. Integration of the area under this curve measures the relative mass of particles present in the injected sample. As this is usually not calibrated only the relative mass in arbitrary units is given.

In this study, an ICP–MS was connected to the Sd FFF system after the UV-visible detector to examine the elemental composition across the particle-size range. When the Sd FFF is interfaced to an ICP–MS a raw ion current I(Eraw) is generated for each element. This is then corrected for noise and drift as described elsewhere (Ranville et al., 1999). The resulting normalized ion current I(Eo) is proportional to the mass concentration of the element present in the eluent dmic/dvi. By using a series of multi-element standards the ion currents are converted to µg L–1 concentrations. Just as the UV-based fractogram is converted to a size distribution, the element fractogram is converted to an element based particle-size distribution. In this case, the y-axis of the fractogram is converted from the µg L–1 concentrations to µg µm–1. Integration of the curve gives the total amount of the element present in the injected sample. Molar element ratios were determined to examine changes in mineralogy across the size range.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Sample Collection and Preparation
The study site was located in an agricultural region within the Onkaparinga catchment of the Mount Lofty Ranges just east of Adelaide in South Australia (Fig. 1) . The soil is the Prospect Hill sandy loam, an Alfisol with a strong textural contrast developed in Proterozoic micaceous siltstones and sandstones.



View larger version (31K):
[in this window]
[in a new window]
  		Fig. 1. Location map for the study site at Mt. Bold, South Australia.

 
A number of studies are underway at this site to investigate the relative importance of surface runoff (i.e., overland flow) versus infiltration (i.e., throughflow) as pathways for transporting nutrients (especially P), DOC, colloids and fine particulates. A series of hillside runoff plots, 20 m in length by 3 m in width, were established in 1989 to examine colloid, DOC, and P transport through the soil profile. At the downslope end of each plot a trench was dug, approximately 2 m deep, and a vertical face cut to allow installation of surface runoff and throughflow samplers at the boundaries between three soil horizons (Fig. 2) . Surface runoff and throughflow were collected at the A2/B1 (0.2 m), B2/B3 (0.55 m), and B3/C (0.85 m) horizon boundaries during a rainfall event. These positions will be referred to in this paper as the A2, B2, and B3 flow samples, respectively. In addition, soil samples were taken at the trench face from the A1, A2, and B3 horizons and will be referred to as A1, A2, and B3 horizon samples, respectively. Finally, two samples of material lining the macropores (cutans) within the B2 and B3 horizons were carefully collected by excision with a scalpel and the aid of a dissecting microscope: they will be referred to as B2 and B3 cutans. All samples were stored at 4°C until analysis.



View larger version (104K):
[in this window]
[in a new window]
  		Fig. 2. Cross-sectional view of the soil profile showing the method of collection of the overland flow and throughflow samples at the various soil horizon boundaries.

 
Sample Prefractionation
Soil particles exist over a size range far wider than can be analyzed in a single FFF analysis. For this reason sample prefractionation was applied to the soil sample by a combination of gravity settling and centrifugation using an Heraeus-Christ Macrofuge centrifuge with a swing-out head containing six 1-L capacity tubes. Size fractions of >2, 1 to 2, 0.08 to 1, and <0.08 µm were obtained. To obtain a sharply defined size cut, several replicate centrifugations were made. The 0.08- to 1-µm fraction is ideal for analysis by Sd FFF in the normal mode of elution. Thus detailed results are given for this size fraction only. The Sd FFF analyses were performed soon after sample collection. The other size fractions are amenable to other subtechniques and modes of FFF and await analysis by these techniques.

Sd FFF Instrumentation and Run Conditions
The instrument used was a Model S101 fractionator from FFFractionation LLC (now Postnova analytics, Salt Lake City, UT). The channel thickness was 0.0127 cm and the void volume was determined to be 2.5 mL. Carrier was delivered at a rate of 1.00 ± 0.05 mL min–1 and consisted of a 10–3 M sodium pyrophosphate solution, which also contained 0.02% (wt/v) sodium azide as a bacterial inhibitor. The carrier also contained 100 mg L–1 indium to provide an internal standard for the ICP–MS. A Spectra 100 UV-visible detector, operating at 254 nm, was used to monitor particles eluting from the FFF. Approximately 0.1 mg of sample is injected in 100 to 200 µL of sample.

When analyzing a wide size range, a high field is applied initially to separate the smallest particles from the void peak, which contains species too small to be analyzed under the FFF conditions chosen. To elute larger colloids, this field is reduced over time. In these experiments the speed of the centrifuge was decreased from 107 x g (800 rpm) initially to approximately 1 x g (50 rpm) following a power decay program (Ranville et al., 1999). Under these conditions particles less than about 0.08 µm are not resolved but particles >0.08 µm are well-retained and particles up to about 1 µm are eluted in under 60 min.

ICP–MS Instrumentation and Operating Conditions
A VG PlasmaQuad ICP–MS was directly coupled to the Sd FFF using a simple Teflon T-fitting. The ICP–MS peristaltic pump was used to deliver the Sd FFF eluent to the pneumatic nebuliser. The operating conditions used for the ICP–MS are given in Ranville et al. (1999) and were the same as those used for the routine operation of the ICP–MS in our laboratory. The ICP–MS was calibrated using SPEX multi-element standards prepared in the Sd FFF carrier solution.

X-Ray Diffraction and Electron Microscopy
The samples were Mg saturated twice with MgCl2, washed five times with deionized water then five drops of glycerol added. The clay and membrane was fixed to 32-mm aluminum disks using double-sided tape. X-ray diffraction patterns were collected on a Philips PW1800 microprocessor-controlled diffractometer using CoK{alpha} radiation, variable divergence slit, and graphite monochromator. The diffraction patterns were recorded from 3 to 33° in steps of 0.05° 2{theta} with a 2.0-s counting time per step, and logged to data files on PC for analysis.

Small samples of the suspensions following fractionation were pipetted onto carbon-coated copper grids and observed in an Hitachi 200 KV analytical electron microscope fitted with an energy dispersive x-ray analyzer.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Particle Size and Element Distributions
Apart from some initial runs with 1- to 0.2-µm fractions most work in this paper refers to the fraction 0.08 to 1 µm. In Fig. 3 , the efficacy of Sd FFF to separate layer lattice clay minerals is illustrated with some TEM micrographs of fractionated 1- to 0.2-µm colloids from a B2/B3 boundary flow sample. Subsamples were taken during the fractionation at times equivalent to elution of diameters 0.15, 0.3, and 0.8 µm, as calculated with the assumption that the particles are spherical. The particles generally increase in size across the fractogram. Smaller particles in each fraction may be the result of aggregate dispersion during post-fractionation. Particles with larger cross-section than the nominal equivalent spherical diameter are expected for platy morphologies.



View larger version (104K):
[in this window]
[in a new window]
  		Fig. 3. Transmission electron micrographs of fractionated 0.2- to 1.0-µm colloids. The colloids were isolated from flow along the B2/B3 horizon boundary of the Prospect Hill sandy loam at elution volumes corresponding to mean equivalent spherical diameters of (a) 0.15, (b) 0.3, and (c) 0.8 µm.

 
Figure 4 shows the particle mass, Al-, Fe-, and Mg-based particle-size distributions (PSDs) for the 0.08- to 1-µm colloids from the three soil horizons. In general the measured PSDs are consistent with the 0.08- to 1.0-µm prefractionation although the distributions did not extend past about 0.7 µm. For the A1 sample (and also the B2 and B3 cutans shown in Fig. 5) some material was present that was <0.08 µm. For these samples the data smaller than 0.08 µm were removed from the fractogram along with the void peak, thereby affecting a sharp rise in the PSD at this point. A possible explanation for the presence of material <0.08 µm is particle disaggregation after the prefractionation performed before Sd FFF-ICP–MS analysis.



View larger version (42K):
[in this window]
[in a new window]
  		Fig. 4. UV detector (relative mass) and Al-, Fe- and Mg-based particle-size distributions (PSDs) for the 0.08- to 1.0-µm soil colloids from A1, A2, and B3 horizons.

 


View larger version (31K):
[in this window]
[in a new window]
  		Fig. 5. UV detector (relative mass) and Al, Fe and Mg concentration-based PSDs for the 0.08- to 1.0-µm colloids from the B2 and B3 cutans.

 
In each sample the PSD based on the UV detector signal (i.e., relative mass) and amount of Al are almost identical. This observation would be obtained if there is little distortion in the mass-based size distributions as a result of size-dependant scattering effects and also that the Al contents of the particles are relatively constant across the range 0.1 to 0.7 µm. Some support for these hypotheses has been obtained using evaporative light scattering detection for river water colloids (Beckett et al., 1988) but they would not be applicable to all samples. Reschiglian and coworkers (Zattoni et al., 2000) have recently suggested an approach for calibrating the mass response using a photodiode array detector and this may improve the accuracy of the PSD determination.

The PSDs for the three soil horizon samples are quite different. The PSD for the A1 horizon was somewhat bimodal with a sharp peak below 0.15 µm and an extended shoulder between about 0.2 to 0.5 µm. In contrast, most of the material in the lower size peak is absent from the A2 sample. The B3 horizon sample also has a peak with a maximum at 0.12 µm and most of the larger material seen in the A1 and A2 horizons is absent.

The Fe and Mg PSDs generally do not follow the mass or Al distribution closely indicating a change in the chemical composition across the PSD of the samples. For the A1 and A2 samples Fe appears to be more concentrated in smaller sizes whereas Mg is generally enriched in the larger particles.

As can be seen in the plots in Fig. 5, the mass, Al, Fe, and Mg PSDs for the cutans of size 0.08 to 1.0 µm collected from pores within the B2 and B3 soil horizons are almost identical. Furthermore the PSDs of the two cutans are very similar to both the mass and Al PSDs of the B3 soil sample, the majority of which are dominated by the smaller particles. Unlike the B3 horizon, significant amounts of material <0.08 µm were observed in the PSDs of the cutans. This material was removed from the fractograms with the void peaks before conversion to PSD and is thus not shown in Fig. 5.

The Fe/Al and Mg/Al atomic ratios for the A1 and B3 horizons and B3 cutans are compared in Fig. 6 . The Fe/Al ratio decreases with increasing size for all of the soil horizon samples, including A2, which is not shown here. The A1 and B3 distributions are distinguished by a rapid decrease in Fe/Al ratio of nearly three-fold between 0.1 and 0.2 µm: the ratio is almost constant from 0.2 to 0.4 µm. The rise in Fe/Al with increase in particle-size above 0.4 µm in the B1 horizon is probably due to an increase in the proportion of illite relative to kaolinite. The Fe content of kaolinite in the Prospect Hill sandy loam is quite low (<1%) whereas in illite it is approximately 7%. The Fe/Al ratio for the B3 cutans also decreases but there is a more even decline over the range 0.1 to 0.4 µm.



View larger version (23K):
[in this window]
[in a new window]
  		Fig. 6. Element atomic ratios versus PSDs for colloids from the A1 and B3 horizons and for the B3 cutans (a) Fe/Al and (b) Mg/Al

 
The Mg/Al ratio increases for all samples above 0.25 µm. In the case of the cutan there is a small decrease in Mg/Al from 0.1 to 0.2 µm. The A1 horizon is a little more enriched in Mg than the B3 horizon, an observation consistent with the x-ray diffraction (XRD) plots in Fig. 7 , which show that the A1 horizon has a greater proportion of vermiculite and illite than the B3 horizon. Note that in the fine clay fraction (<0.2 µm) of the B3 horizon there is no vermiculite and only a very small amount in the A1. The difference in the proportion of illite to kaolinite between horizons is even greater in the fine clay fraction.



View larger version (18K):
[in this window]
[in a new window]
  		Fig. 7. X-ray diffraction plots of (a) coarse (2-0.2 µm) and (b) fine (< 0.2 µm.) clays from the A1 and B3 horizons of the Prospect Hill sandy loam at Mount Bold.

 
The increase in Mg/Al ratio with particle size for the soils and cutans shown in Fig. 6 reflects the increase in vermiculite as a function of particle size (see XRD data in Fig. 7). The TEMs, some of which are shown in Fig. 8 , indicate that about 10% of the kaolinite in the fine clay fraction is tubular but it is absent in the coarse clay (Fig. 8).



View larger version (103K):
[in this window]
[in a new window]
  		Fig. 8. Transmission electron micrographs of clays of size 0.2–1 µm fractionated by sedimentation field-flow fractionation (Sd FFF). Clay colloids were isolated from cutans in the B3 horizon of the Prospect Hill sandy loam (a) 0.3 µm; (b) 0.5 µm.

 
The relatively high concentration of Fe in the finest fraction may indicate that a significant proportion of the Fe is present as Fe oxyhydroxide surface coatings. Similar trends have been reported in studies of suspended particles in rivers and reservoirs (van Berkel and Beckett, 1996). The fact that the Fe content does not continue to decrease as the size increases above 0.3 µm could be explained if the surface coating effect was balanced by a change in mineralogy such that the bulk composition was richer in Fe. This proposition is supported by the XRD data that show an increase in illite/kaolinite (Fig. 7). In contrast to the A1 and B3 soil samples the Fe/Al ratio for the B3 cutans shows only a slight decrease between 0.1 and 0.3 µm followed by a pronounced reduction between 0.3 and 0.5 µm. The latter decrease and corresponding increase in Mg/Al ratio shown in Fig. 6, indicate a change in clay mineralogy above 0.3 µm. Apart from the very smallest and largest particles in the size range, the Fe/Al ratio in the cutans is significantly greater than the soil horizon samples. Micromorphological examination confirmed that a more precise designation for these cutans was ferriargillans (Brewer 1976).

The mass, Al, Fe, and Mg PSDs for the overland, A2/B1 boundary and B3/C boundary flow samples are shown in Fig. 9 . As for the soil horizons and cutans the Al and UV PSDs of colloids in overland and throughflow are almost coincident (Fig. 9). The PSDs for all of the flow samples are very similar and closely resemble the plots of the B3 horizon and cutan rather than the A horizon. This is good evidence that a significant proportion of the clay in the B horizon originated from the A horizon. The mobile clay is different from that which remains in the A in that it has a higher proportion of illite.



View larger version (52K):
[in this window]
[in a new window]
  		Fig. 9. UV detector (relative mass) and Al-, Fe- and Mg-based PSDs for colloids from (a) the overland flow and throughflow samples from the A2/B1, B2/B3 and B3/C soil horizon boundaries.

 
The Fe/Al and Mg/Al atomic ratios are presented in Fig. 10 . Although the Fe/Al ratios of the mobile colloids are generally higher than those of the soils, and hence closer to the values for the cutans, the distributions of the Fe/Al ratio are similar to that found for the soils and are quite unlike those of the cutans (Fe/Al data in Fig. 6). Iron is enriched in the smaller particle sizes from which we infer that Fe oxyhydroxide coatings may be significant in the size range lower than about 0.2 µm. The flow samples then display a slight increase in Fe/Al ratio above a size of 0.3 to 0.4 µm depending on the exact sample which is consistent with there being a change in mineralogy for the larger particles. Note that for clarity only Fe/Al plots for the overland flow and the B2/B3 samples are given in Fig. 10 but the other throughflow samples show similar trends.



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 10. Element atomic ratios versus PSDs for colloids collected from the flow samples (a) Fe/Al ratios from the overland flow and B2/B3 through flow, and (b) Mg/Al ratios from the overland flow and A2/B1 and B3/C infiltration flows.

 
Figure 10 gives the Mg/Al molar ratio distributions for the overland flow and A2/B1 and B3/C throughflow. The particles in the overland and A2/B1 flows show varying degrees of increase in Mg composition across the size range except perhaps for the very smallest particles (<0.15 µm). The Mg/Al ratios for the overland flow samples are always higher than the throughflow suspensions suggesting larger proportions of vermiculite, a proposition also supported by the XRD data for these samples. For simplicity the Mg/Al distribution of the B2/B3 interface flow sample is not shown on Fig. 10 but it is very similar to the A2/B1 throughflow sample.

Implications for Colloid Transport Processes in Soils
It has long been observed that for many soil types, soil formation processes result in a coarsening of the near-surface horizons (e.g., A and E) and an accumulation of fine particles in deeper horizons (Birkeland, 1974). This has generally been observed by a change in the sand, silt, and clay ratio. It is interesting to note that these changes were also observed within the colloid size range as shown in Fig. 4. The fine particles present in the A1 horizon may be a result of high biological activity that is likely to be occurring in the upper part of the profile and which results in the formation of colloids by weathering. These finest colloids are very enriched in Fe and may represent an organic colloid–iron oxyhydroxide association that is a result of biological processes. The coarser PSD of colloids in the A2 horizon may reflect a loss of the fine particles by translocation with a subsequent enrichment of the larger colloids in this horizon. Comparison of the A1 and A2 PSDs show a loss of particles in the size range that corresponds to that seen for the mobile colloids (Fig. 8). These fine colloids may be deposited in the B3 horizon, which is consistent with the observed PSD for this horizon. An alternative hypothesis is that the fine colloids are dissolved in the A2 horizon and then reprecipitated within the B3.

The general observation that for the mobile and soil matrix colloids iron content increases with decreasing size shows that Fe/Al atomic ratio increases from about 0.08 to 0.25 (Fig. 6), could indicate that a significant amount of Fe is present as surface coatings. This could have several important implications. Iron oxyhydroxides have been shown to be very important in the adsorption of pollutants in soils, especially P and trace metals. Work with these same soils has demonstrated a close correspondence between Fe content and P binding (van Berkel and Beckett, 1996). Iron oxide coatings also play an important role in colloid transport (Ryan and Elimelich, 1996). The mobility of these colloids, and associated pollutants may be strongly influenced by their Fe content. The presence of Fe coatings, which can potentially have a net positive charge at most soil pHs, can result in attractive interactions of the colloids with negatively charged clay minerals. This may explain the observation that the cutans, which may represent mobile colloids that have been captured on the macropore walls, have a higher, more uniform Fe content across the size range. The Fe/Al ratio of the cutans and the mobile soil colloids are very similar (both about 0.2 atomic ratio) suggesting a possible colloid capture mechanism. An alternative explanation could be that elevated rates of weathering occur along the macropores due to the greater amount of water contact, which results in the formation of more iron oxides in the cutans. Combining the higher Fe content of the cutans with the fact that most water flow occurs in the macropores, suggests that the cutans may play a more important role in the sorption of pollutants than the soil matrix colloids. When attempting to predict the interaction of soil and pollutants such as P or trace metals, additional knowledge about the soil composition (i.e., the chemistry of the cutans as well as the soil matrix) may be needed.

The uniformity of the PSDs of the mobile colloids with depth is in contrast to the significant differences in the PSDs of the soil matrix colloids. Assuming that the colloids generated during the rainfall arose from the A1 horizon, an intermediate-sized colloid appears to be most favored for transport, although the size distribution is skewed to the smaller end of the distribution (0.2 µm). In general this observation of a narrow size distribution for the mobile colloid, as compared with the source of the colloids, is consistent with the predictions from filtration theory (Ryan and Elimelich, 1996). The uniformity in PSD and the lack of correlation between the PSD of the mobile soil colloid and the soil matrix colloid within the same horizon suggests that the colloids transported in the macropores do not interact substantially with the soil matrix. One implication is that macropore flow then can act as a conduit for mobile colloids, and any associated pollutants, that are released from the upper soil horizons even in horizons high in clay. This may allow colloid-associated pollutants to migrate into local and perhaps regional ground waters even through soils with horizons high in clay such as Haploxeralfs.

However, not only are PSDs for throughflow samples collected from different horizons similar, but the overland flow sample is also similar. Because the overland flow sample is likely to experience a different hydrologic environment than the throughflow samples, the observed uniformity of the PSDs may explain more about the relative dispersability of certain sized colloids than the effect of transport on colloid size. The results suggest that in terms of size, the intermediate size particles (approximately 0.2 µm) appear to be most dispersible.

An effect of mineralogy on dispersability and/or transport might be seen in the Mg/Al ratios of the various colloids. The observed higher Mg/Al ratio for the A1 soil versus the B3 soil was reflected in the greater amount of vermiculite and illite in the upper soil, which was also observed in the XRD data (Fig. 8). Differences in the Mg/Al ratio are also seen to occur across the size range for the soil matrix colloids, especially for the A1 horizon. The increase in Mg/Al ratio with increasing size suggests a greater content of illite and vermiculite in the larger colloids with kaolinite being the dominant clay mineral in the smaller colloids. A similar trend with increasing size is seen for the mobile colloids. In general the Mg/Al ratio is higher in the mobile colloids, especially in the 0.2- to 0.3-µm range, which corresponds to the peak maximum of the mobile colloids. For example, the Mg/Al ratio for the B3 horizon is about 0.02 versus 0.03 for the B3/C mobile colloids at a diameter of 0.25 µm. Therefore it appears that a greater amount of vermiculite and/or illite is being transported in the mobile soil colloids. The difference is even greater between the A1 horizon and the overland flow samples: 0.025 versus 0.05 at 0.2 µm, respectively. Therefore differences in chemistry of various clay minerals, especially the higher CEC of the 2:1 clays, may have a marked influence on the relative mobility of pollutants associated with clay minerals.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Development of Sd FFF-ICP–MS, and its application to soils provides a powerful tool to investigate soil processes involving colloids. Sd FFF-ICP–MS produces elemental size distributions having a great level of detail in the submicrometer range. This detail could only be produced with classical procedures if laborious repetitive centrifugation steps were performed. The small quantity of material needed for Sd FFF-ICP–MS also makes it advantageous over classical approaches. Only by having this level of resolution, could the difference in the depth-dependant trends between the matrix, cutans, and mobile colloids be discerned. Furthermore the physical distribution of metals, for example their presence as surface coatings as opposed to their location within the mineral structures, can be inferred from the size-dependence of elemental ratios (i.e., Fe/Al). Further work with contaminated soils should provide new insights into pollutant fate, transport and bioavailability in soils.


   ACKNOWLEDGMENTS
 
Funding was provided by the Land and Water Resources Research and Development Corporation and the Australian Research Council.

Received for publication March 1, 2004.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Ranville, J. F.
Right arrow Articles by Beckett, R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Ranville, J. F.
Right arrow Articles by Beckett, R.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Ranville, J. F.
Right arrow Articles by Beckett, R.
Related Collections
Right arrow Watershed-Scale Studies
Right arrow Colloids
Right arrow Colloid-Facilitated Transport
Right arrow Soil Methods/Instrumentation

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Vadose Zone Journal Journal of Plant Registrations
Journal of Natural Resources
and Life Sciences Education		 Journal of
Environmental Quality