- Email: [email protected]
Soil pore surface properties in managed grasslands
Soil & Tillage Research 55 (2000) 63±70
Soil pore surface properties in managed grasslands Meczislaw Hajnosa, Ludmila Korsunskaiab, Yakov Pachepskyc,* a Institute of Agrophysics, Lublin 20536, Poland Institute of Physical, Chemical, and Biological Problems in Soil Science, Pushchino 142292, Russia c USDA:ARS:BA:NRI:RSML, Hydrology Laboratory, Bldg. 007, Rm. 104, BARC-WEST, Beltsville, MD 20705, USA b
Received 3 August 1999; received in revised form 7 January 2000; accepted 28 January 2000
Abstract Properties of pore surfaces control adsorption and transport of water and chemicals in soils. Parameters are needed to recognize and monitor changes in pore surfaces caused by differences in soil management. Data on gas adsorption in soils can be compressed into parameters characterizing (a) area available to a particular adsorbate, and (b) surface roughness or irregularity. Our objectives were to see (a) whether models of adsorption on fractal surfaces are applicable to water vapor adsorption in soils in the capillary condensation range, and (b) whether differences in long-term management of grasslands are re¯ected by soil pore surface properties. Water vapor adsorption was measured in Gray Forest soil (Udic Argiboroll, Orthic Greyezem, clay loam) samples taken at four plots, where a long-term experiment on grassing arable land had been carried out for 12 years. The experiment had 22 design. Factors were `harvest±no-harvest' and `fertilizer±no-fertilizer'. The hay was cut after over-seeding in harvested treatments every year. Ammonium nitrate, superphosphate, and potassium chloride were applied annually after the snowmelt to get the total amount of nutrients of 60 kg haÿ1. The monolayer adsorption capacity was estimated from the Brunauer±Emmett±Teller model. A fractal Frenkel±Halsey±Hill model of adsorption on a fractal surface, and a thermodynamic adsorption model were applied in the range of relative pressures from 0.7 to 0.98 and provided good ®t of data. Values of the surface fractal dimension Ds were in the range from 2.75 to 2.85. Removal of carbohydrates resulted in increase of Ds. Differences in management practices did not affect values of Ds in the scale range studied, whereas the monolayer capacity was affected. Both fertilization and harvesting resulted in an increase of the monolayer capacity, with the largest increase observed in soil that was fertilized but not harvested. # 2000 Elsevier Science B.V. All rights reserved. Keywords: Water vapor; Adsorption; Fractal surface; Grassland; Gray Forest soil
1. Introduction Properties of pore surfaces control adsorption and transport of water and chemicals in soils. Parameters are needed to recognize and monitor changes in pore *
Corresponding author. Tel.: 1-301-504-7468; fax: 1-301-504-5823. E-mail address: [email protected] (Y. Pachepsky)
surfaces caused by differences in soil management. Once established, such diagnostic parameters can be used to quantify effects of management on soil functioning by means of land quality or soil quality indices (Singer and Ewing, 2000). Data on gas adsorption in soils can be compressed into parameters characterizing (a) area available to a particular adsorbate, and (b) surface roughness or irregularity. Speci®c surface area is known to char-
0167-1987/00/$ ± see front matter # 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 1 9 8 7 ( 0 0 ) 0 0 0 9 9 - 4
64
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
acterize the ability of soil to retain and transport nutrients and water. Speci®c surface area of soils correlates with the sorptive capacity and ion exchange capacity (Jurinak and Volman, 1957; Ben-Dor and Banin, 1995), retention and release of chemicals (Rhue et al., 1988; Valverde-Garcia et al., 1988; Ong and Lion, 1991), swelling (Sapozhnikov, 1985), water retention (Kapilevich et al., 1987), dry-aggregate stability (Skidmore and Layton, 1992). Tester (1990) observed an increase in speci®c surface area parallel to the increase in organic matter content after compost application. Pore surfaces in soils are irregular. Fractal models were found useful in the description of natural and man-made irregular surfaces. Fractal surfaces exhibit dependencies of their area S on the resolution, or scale, R at which the area is measured (Russ, 1994). These dependencies have a power law form S / R2ÿDs
(1)
where Ds is the surface fractal dimension. The surface fractal dimension characterizes the degree of roughness or disorder of the surface. Numerical values of surface fractal dimensions vary from 2.0 (perfectly smooth) to a maximum value of 3.0. The larger the value of Ds, the greater is the irregularity of the surface. Theoretical studies of adsorption on fractal surfaces predicted the effect of the surface fractal dimension on diffusion-limited adsorption kinetics (Seri-Levy and Avnir, 1993; Giona and Giustiniani, 1996). Douglas (1989) showed that adsorption of polymers on rough surfaces should be affected by the surface fractal dimension. The increase in fractal dimensions should cause tighter binding of polymer with a surface. Surface fractal dimensions were studied with relation to technological parameters in production of ionexchange ®bers (Kutarov and Kats, 1992), carbon blacks (Darmstadt et al., 1995), cements (Niklasson, 1993), food products (Aguerre et al., 1996), and wood (Hatzikiriakos and Avramidis, 1994). Fractal scaling was demonstrated for soil pore surfaces. Sokoøowska (1989) used a fractal model to analyze the soil particle surface heterogeneity with data on water vapor adsorption experiments. Toledo et al. (1990) showed that fractal dimensions of soil particle surfaces could be estimated from water retention data at low moisture contents. Bartoli et al. (1991)
introduced the use of mercury porosimetry to estimate surface fractal dimensions in soils. Soil constituents, such as humus materials and clay minerals, were found to be surface fractals (Rice et al., 1999; Sokoøowska and Sokoøowski, 1999). The management-related change in soil quality is a current topic of interest. Efforts were made to de®ne soil quality and suggest a minimum set of analyses to aid in evaluating soil quality as affected by long-term management practices (Doran and Parkin, 1994). Long-term and short-term management effects on soil pore surface fractal dimensions were observed using soil thin sections (Pachepsky et al., 1996b; GimeÂnez et al., 1997; Oleschko et al., 1998). Fractal dimension analysis provided additional information by describing the changes in pore shapes. Adsorption isotherms of gases present an attractive source of data on surface roughness. Several theories of adsorption on a fractal surface were developed during last 15 years. Two types of adsorption data analysis have been used most often. One of them uses several different adsorbates and employs the mean radius of each adsorbate's molecules as the resolution, or scale, R in Eq. (1). The logarithmic transform of this equation and the subsequent regression `radius vs. speci®c surface' in log±log scale provides the value of Ds as the slope subtracted from 2 (Farin and Avnir, 1989). With this approach, data on adsorption at low relative pressures are used at which the monolayer coverage is known to occur. Another type of the adsorption data analysis employs the data at high relative pressures, where capillary condensation occurs so that micropores become ®lled with adsorbate. Then the micropore radius becomes a measure of scale R in Eq. (1). Avnir and Jaroniec (1989) obtained a relationship between the adsorbed amount and vapor pressure by assuming the applicability of the Dubinin± Radushkievich equation to describe the adsorption in micropores, and assuming fractal pore volume distribution on the surface. The same equation was obtained by Neimark (1990) and Yin (1991) from the assumption that the capillary condensation was present. A different model was proposed by Neimark (1990). He did not assume validity of any adsorption isotherm equation, but used directly the scaling relationship (1) with the surface curvature radius as a scale measure. Although both models mentioned above were applied in experimental studies, no unambiguous way was
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
found to see which one is superior (Jaroniec et al., 1997). Our objectives were to see (a) whether models of adsorption on fractal surfaces are applicable to water vapor adsorption in soils in the capillary condensation range, and (b) whether differences in long-term management of grasslands are re¯ected by soil pore surface properties. 2. Materials and methods 2.1. Soil sampling and analyses Gray Forest soil (Udic Argiboroll) was sampled in April 1994 near Pushchino, Moscow Region, Russia. Samples were taken at a site, where an experiment on grassing arable land had been carried out since 1982 (Yermolayev and Shirshova, 1994). Grass mixture consisted of Festuca Pratensis, Huds., Phleum Pratense L., and Trifolium Pratense L. The experiment had 22 design. Factors were `harvest±no-harvest' and `fertilizer±no-fertilizer'. The hay was cut after over-seeding in harvested treatments every year. Ammonium nitrate, superphosphate, and potassium chloride were applied annually after the snowmelt to get the total amount of nutrients of 60 kg haÿ1. Area of each plot was 0.5 ha. The soil was in no-till, since 1982. The soil had the pH 6.3, a loamy texture with 45% particles less than 0.01 mm and 20% particles less than 0.001 mm, and contained illite, kaolinite, and smectites in its clay fraction. Sampling depth was 10± 30 cm. Five samples were taken from each plot, and the mixed sample was used in adsorption measurements for each of the four combinations of factors. Soil was air-dried and sieved, and aggregates less than 2 mm were used. Each sample was split, and carbohydrates were removed in one of the two subsamples. The oxidation of carbohydrates has been performed with 0.02 M NaIO4 solution (Cheshire et al., 1984). Water vapor adsorption isotherms were measured with a vacuum microbalance technique. The temperature was kept within the range of 2940.1 K. The adsorption isotherms were measured in duplicate at 14 relative pressure levels of 0.0217, 0.0902, 0.1246, 0.2142, 0.3008, 0.3404, 0.4112, 0.5324, 0.6075, 0.6533, 0.7320, 0.7727, 0.8680, 0.9826.
65
For comparison purposes, we also used published data on water vapor adsorption on clay minerals (Mooney et al., 1952; Orchiston, 1954; Jurinak, 1963). The BET equation (Adamson, 1967) N N0
c p=p0 1 ÿ p=p0 1 ÿ c ÿ 1p=p0
(2)
was ®tted to the experimental isotherms in the range of relative pressures p/p0 between 0.02 and 0.34. Here and below N is the adsorbate amount, p/p0 is the relative pressure de®ned as a ratio of the actual pressure p to the saturation vapor pressure p0. The monolayer capacity N0 and thermodynamic constant c were adjusted to achieve the best ®t. 3. Estimating fractal dimensions Avnir and Jaroniec (1989) developed the fractal analogue of the Frenkel±Halsey±Hill (FHH) equation in the following form: p (3) ln N C ÿ Ds ÿ 3 ln p0 Here C is the intercept. The same equation was derived by Yin (1991) and Neimark (1990) by assuming the sequential ®lling of pores from small size to large, that is capillary condensation-type local adsorption isotherm. A similar equation was derived by Pfeifer and Cole (1990) by generalization of the classical FHH adsorption theory for the capillary condensation regime. Neimark pointed out that Eq. (1) should be used with data obtained for p/p0>0.7. Results of Pfeifer and Cole (1990) also indicate that Eq. (1) is valid at high p/p0 values when the capillary condensation is the dominant mechanism. We estimated the surface fractal dimension Ds from the slope of the regression given by Eq. (1) at relative pressures p/p0 in the range from 0.7320 to 0.9826. Neimark (1990) and Neimark and Unger (1993) proposed a different model for calculating the fractal dimension Ds from the adsorption isotherm data. It was called the thermodynamic model. The model re¯ects the idea that during the adsorption on a fractal surface, the surface is smoothed by the adsorbate, and the surface of the adsorbent covered with an adsorbate is smaller than the surface of the uncovered adsorbent. The area of the `condensed adsorbate±
66
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
vapor' equilibrium interface is obtained from the Kiselev equation Z RT Nmax p ÿln dN (4) S g N p0 Here Nmax is the maximum adsorption, R the gas constant, T the Kelvin temperature, and g is the surface tension of the liquid adsorbate. This area is related by the fractal scaling law (1) to the mean scaling radius of this interface R which is calculated from Kelvin's equation R
2gVm RT ÿln p=p0
(5)
where Vm is the molar volume of the liquid adsorbate. The logarithmic transform of (1) with S from (4) and R from (5) results in the equation of the thermodynamic model: p (6) ln S a Ds ÿ 2 ÿln p0 Here a is the intercept. We estimated the surface fractal dimension Ds from the slope of the regression given by Eq. (6) at relative pressure p/p0 in the range from 0.7320 to 0.9826. The integral in the left-hand side of Eq. (4) was evaluated numerically with the trapezoid method. The maximum adsorbed amount was approximately equated to the adsorbed amount at p/p00.9826. Parameters of regression equations and their standard error estimates were found using the SigmaPlot
Fig. 1. Water vapor adsorption isotherms on Gray Forest soil (Udic Argiboroll) in grasslands; N is the adsorbed amount, p the actual vapor pressure, and p0 is the vapor pressure at saturation; (*, *) no harvesting, no fertilization, (&, & ) no harvesting, 60 kg NPK annually, (~, ~) harvested, no fertilization, (!, 5) harvested, 60 kg NPK annually; hollow symbols show data for the samples in which carbohydrates have been removed.
Version 2 software package. The statistical signi®cance of differences was tested with the t-test at the 0.05 signi®cance level. 4. Results and discussion Adsorption data are shown in Fig. 1. Shapes of the curves are similar. At the same relative pressure values, the adsorbed amount is the highest in soil from the non-harvested, fertilized plot, and is the lowest in soil from the non-harvested, non-fertilized plot. Intermediate values of adsorbed water amounts
Table 1 Parameters of soil pore surfaces from water vapor adsorption data Sample
Organic carbon Treatment content (%)
No harvest, no fertilizer 1.92 Ndb No harvest, 60 kg NPK 1.76 Nd Harvested, no fertilizer 1.68 Nd Harvested, 60 kg NPK 1.64 Nd a b
Original sample Carbohydrates removed Original sample Carbohydrates removed Original sample Carbohydrates removed Original sample Carbohydrates removed
Standard errors are shown after the `' sign. Not determined.
Surface fractal dimensions
Parameters of the BET equation
Fractal FHH model
Thermodynamic model
Monolayer BET capacity (g kgÿ1) c-value
2.770.01a 2.790.01 2.780.00 2.790.01 2.780.00 2.780.01 2.780.00 2.780.01
2.750.08 2.790.08 2.760.04 2.830.04 2.780.03 2.850.03 2.740.05 2.850.05
9.510.11 8.780.10 15.650.16 16.640.17 13.610.16 13.040.15 14.190.18 12.670.15
11118 14527 8611 15026 8512 7410 7410 719
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
Fig. 2. Fractal scaling of pore surface from data on water vapor adsorption on clay minerals; N is the adsorbed amount, p the actual vapor pressure, and p0 is the vapor pressure at saturation; (*) Namontmorillonite (Mooney et al., 1952), (*) Li-kaolinite (Jurinak, 1963), (~) illite (Orchiston, 1954).
are found in soils from harvested plots. These effects of treatments are re¯ected in values of the monolayer capacity N0 found from the BET equation (Table 1). The monolayer capacity did not correlate with the organic carbon content in this study. This could be related to the differences in composition of organic matter formed in soil under different management in this experiment (Yermolayev and Shirshova, 1988; Shirshova and Yermolayev, 1990).
67
Removal of carbohydrates caused a small decrease of the monolayer capacity of water adsorption in soil from all but non-harvested, fertilized plots. Feller et al. (1992) observed that partial removal of carbon by treatments with H2O2 increased the surface area of clay-sized organo-mineral complexes while surface areas of silt and sand were less affected. Oades (1984) suggested that clays were embedded in a matrix of soil organic matter. Therefore, in our study, differences in composition of soil organic matter could be related to the type of particles to which it was bounded. Application of the fractal FHH model to data on soil minerals is shown in Fig. 2. The FHH model provides a reasonably good ®t to the adsorption data. Surface fractal dimensions increase in the sequence `illite±Likaolinite±Na-montmorillonite'. The values of Ds are 2.49, 2.62, and 2.74, respectively. Application of both the fractal FHH model and the thermodynamic model to soil data is shown in Fig. 3. The dependencies between transformed data are very close to linear ones as predicted by both models. Estimated values of surface fractal dimensions are given in Table 1. The standard errors of the fractal dimension estimates are larger for the thermodynamic model than for the fractal FHH model. Estimated values of Ds range from 2.73 to 2.78 for untreated samples and from 2.78 to 2.85 in samples, where carbohydrates were removed. Our data do not indicate that either of the two compared models is superior.
Fig. 3. Fractal scaling of the pore surface demonstrated with two models of vapor adsorption on fractal surfaces: (a) fractal Frenkel±Halsey± Hill model, (b) thermodynamic model; N is the adsorbed amount, Nmax the maximum adsorbed amount, p the actual vapor pressure, and p0 is the vapor pressure at saturation; (*) no harvesting, no fertilization, (&) no harvesting, 60 kg NPK annually, (~) harvested, no fertilization, (!) harvested, 60 kg NPK annually.
68
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
The values of fractal dimensions are close for both models applied to the same sample, and no statistically signi®cant differences are encountered. This might be expected since we used data in the range of relative pressures, where the capillary condensation was a dominating adsorption process. Jaroniec (1995) had shown that the Eq. (6) can be derived from the FHH equation under the assumption of the condensation approximation validity. Sahouli et al. (1996) observed the equivalence in results of application both the fractal FHH equation and the thermodynamic model (6) when the adsorption process was dominated by the capillary condensation in carbon black. Darmstadt et al. (1995) applied both the fractal FHH model (3) and the thermodynamic model (6) to data on N2 adsorption on carbon blacks. They found the fractal dimensions from both models to be very close in three of four samples. Their fourth sample had much lower fractal dimensions than the other three. It was suggested that this sample had mesopores where the capillary condensation occurred which led to overcompensating the reduced adsorption due to the fractal surface. Fractal dimension values did not differ between the experimental treatments. The soil surface irregularity remained the same. The intercepts in graphs of Fig. 2 show differences among the treatments. Soil in harvested plots displayed similarities in parameters of both fractal models. For non-harvested plots, the intercepts are different from each other and from the harvested plots. The differences in intercepts re¯ect the difference in ®lm surface area at the same capillary radius value. The larger the intercept, the larger is the ®lm surface area. Therefore, within the studied radii range from 3 to 60 nm, the number of pores of any given radius is the largest in the soil of the non-harvested fertilized plot. As the radius changes, this number scales in a similar manner in all plots. Several fractal models of pore surfaces were successfully used to distinguish effects of tillage and compaction on soil structure (Perfect and Kay, 1995). Scales in these studies corresponded to the aggregate sizes in dry sieving, i.e. 0.3±30 mm (GimeÂnez et al., 1997), or to the recognizable features in soil thick and thin sections, i.e. 0.008±3 mm (Pachepsky et al., 1996b; Oleszko et al., 1998). In the above studies, the authors measured soil structure directly
using variants of the box-counting method. Scales were much smaller in this study, and the resolution was de®ned indirectly by converting pressures to the radii. Different sensitivity of fractal parameters to long-term soil management and laboratory treatments was found in different ranges of scales (Pachepsky et al., 1996a; GimeÂnez et al., 1997). The absence of differences in fractal dimensions in the capillary condensation region does not imply that the grassland management does not affect surface roughness at all. The differences are absent in the speci®c range of pore radii from 3 to 60 nm, but it is possible that differences in surface roughness exist in other ranges of radii. Carbohydrate removal had an effect on the values of fractal dimensions estimated with the thermodynamic model but not with the fractal FHH (Table 1). No statistically signi®cant differences were found between the fractal dimension estimates in original samples and samples with removed carbohydrates. The increase in surface fractal dimension values was expected since the irregularity of the surface should increase after the removal of structure-forming organic compounds. Removal of organic matter caused an increase in surface fractal dimensions measured with the mercury porosimetry in three soils studied by Pachepsky et al. (1996a). Compared with pure minerals in Fig. 2, the soil in this study had larger values of surface fractal dimensions. Mixed mineral composition and the presence of organic matter are probable reasons for that. Application of other fractal models to quantify irregularity of soil composition and structure also resulted in obtaining high, close to three, values of mass and fragmentation fractal dimensions for ®ne textured soils (Perfect and Kay, 1995). A caution has to be exercised in interpreting values of fractal dimensions obtained with water adsorption because of speci®c interactions of water molecules with surfaces. Niklasson (1993) has found that data on nitrogen adsorption tend to give larger values of surface fractal dimensions than data on water adsorption. This author observed a close correspondence between fractal dimensions obtained from water adsorption and from small-angle neutron scattering. Pachepsky et al. (1996b) used mercury porosimetry to measure the surface fractal dimension under developing barley crop for the same soil as in this study but in agricultural use. In samples taken during spring, they found
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
values of Ds2.870.10 which did not differ signi®cantly from the values in Table 1. Fractal scaling presents additional parameters to characterize soil pore surfaces. Speci®c surface area was shown to be related to various physical and chemical properties of soil (Petersen et al., 1996). The region of relative pressures between 0.05 and 0.3 is used to ®nd the speci®c surface area (Adamson, 1967; Kutilek and Nielsen, 1994). Fractal scaling parameters as de®ned in this paper give information about different range of relative pressures, and therefore, different range of pore radii. There is a growing interest in competitive adsorption of volatile organic compounds and water in soils as related to soil bioremediation and fate of fumigants (e.g. Pennell et al., 1992; Poulsen et al., 1998). Fractal parameters of pore surfaces may ®nd applications in this ®eld. In conclusion, the models of adsorption on fractal surfaces appear to be applicable to water adsorption in soil in this study. Differences in long-term management of the grassland did not affect surface fractal dimensions, although the surface area was affected. References Adamson, A.W., 1967. Physical Chemical of Surfaces. Interscience, New York. Aguerre, R.J., Viollaz, R.E., Suarez, C., 1996. A fractal isotherm for multilayered adsorption in foods. J. Food Eng. 30, 227±238. Avnir, D., Jaroniec, M., 1989. An isotherm equation for adsorption on fractal surfaces of heterogeneous porous material. Langmuir 5, 1431±1433. Bartoli, F., Phylippy, R., Doirisse, M., Niquet, S., Dubuit, M., 1991. Structure and self-similarity in silty and sandy soils: the fractal approach. J. Soil Sci. 42, 167±185. Ben-Dor, E., Banin, A., 1995. Near-infrared analysis as a rapid method to simultaneously evaluate several soil properties. Soil Sci. Soc. Am. J. 59, 364±372. Cheshire, M.W., Sparling, J.P., Mundle, C.M., 1984. In¯uence of soil type, crop, and air drying on residual carbohydrate content. Plant and Soil 76, 339±347. Darmstadt, H., Roy, C., Kallaguine, S., Sahouli, B., Blasher, S., Pirard, R., Brouers, F., 1995. Fractal analysis of commercial and pyrolytic carbon blacks using nitrogen adsorption data. Rubber Chem. Technol. 68, 330±341. Doran, J.W., Parkin, T.B., 1994. De®ning and assessing soil quality. In: Doran, J.W., Coleman, D.C., Bezdicek, D.F., Stewart, B.A. (Eds.), De®ning Soil Quality for a Sustainable Environment. SSSA Special Publ. No. 35, Madison, WI, pp. 3±21. Douglas, J.F., 1989. How does surface roughness affect polymer± surface interactions? Macromolecules 22, 3707±3716.
69
Farin, D., Avnir, D., 1989. The fractal nature of molecule±surface interactions and reactions. In: Avnir, D. (Ed.), The Fractal Approach to Heterogeneous Chemistry. Wiley, New York, pp. 271±293. Feller, C., Shouller, E., Thomas, F., Rouiller, J., Herbillon, A.J., 1992. N2-BET speci®c surface areas of some low activity clay soils and their relationships with secondary constituents and organic matter contents. Soil Sci. 153, 293±299. GimeÂnez, D., Allmaras, R.R., Nater, E.A., Huggins, D.R., 1997. Fractal dimensions for volume and surface of interaggregate pores: scale effects. Geoderma 77, 19±38. Giona, M., Giustiniani, M., 1996. Adsorption kinetics on fractal surfaces. J. Phys. Chem. 100, 16690±16699. Hatzikiriakos, S.G., Avramidis, S., 1994. Fractal dimension of wood surfaces from sorption isotherms. Wood Sci. Technol. 28, 275±284. Jaroniec, M., 1995. Evaluation of the fractal dimension from a single adsorption isotherm. Langmuir 11, 2316±2317. Jaroniec, M., Kruk, M., Olivier, J., 1997. Fractal analysis of composite adsorption isotherms obtained by using density functional theory data for argon in slitlike pores. Langmuir 13, 1031±1035. Jurinak, J.J., 1963. Multilayer adsorption of water by kaolinite. Soil Sci. Soc. Am. Proc. 27, 269±272. Jurinak, J.J., Volman, D.H., 1957. Application of Brunauer, Emmett and Teller equation to ethylene dibromide adsorption by soils. Soil Sci. 83, 487±496. Kapilevich, Zh.A., Pisentsky, G.A., Vysochenko, A.V., 1987. Speci®c surface as a basic parameter for evaluating hydromelioration characteristics of mineral soils. Sov. Soil Sci. 19, 68±76. Kutarov, V.V., Kats, B.M., 1992. Determination of the fractal dimension of ion-exchange ®bers from adsorption data. Russ. J. Phys. Chem. 67, 1854±1856. Kutilek, M., Nielsen, D., 1994. Soil Hydrology. Catena Verlag, Cremlingen-Destedt, Germany. Mooney, R.W., Keenan, A.G., Wood, L.A., 1952. Adsorption of water vapor by montmorillonite. Part I. Heat of desorption and application of BET theory. J. Am. Chem. Soc. 74, 1367± 1371. Neimark, A.V., 1990. Determination of surface fractal dimension from the results of an adsorption experiment. Russ. J. Phys. Chem. 64, 1398±1403. Neimark, A.V., Unger, K.K., 1993. Method of discrimination of surface fractality. J. Colloid Interf. Sci. 158, 412±419. Niklasson, G.A., 1993. Adsorption on fractal structures: applications to cement materials. Cem. Conc. Res. 23, 1153±1158. Oades, J.M., 1984. Soil organic matter and structural stability: mechanisms and implications for management. Plant and Soil 76, 319±337. Oleschko, K., Brambila, F., Acess, F., Mora, L.P., 1998. From fractal analysis along a line to fractals on the plane. Soil and Tillage Research 45, 389±406. Ong, S.K., Lion, L.W., 1991. Trichloroethylene vapor sorption onto soil minerals. Soil Sci. Soc. Am. J. 55, 1559±1568. Orchiston, H.D., 1954. Adsorption of water vapor: clays at 258C. Soil Sci. 78, 463±480.
70
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
Pachepsky, Ya.A., Korsunskaia, L.P., Hajnos, M., 1996a. Fractal parameters of soil pore surface area under a developing crop. Fractals 4, 97±110. Pachepsky, Ya.A., Yakovchenko, V., Rabenhorst, M.C., Pooley, C., Sikora, L.J., 1996b. Fractal parameters of pore surfaces as derived from micromorphological data: effect of long-term management practices. Geoderma 74, 305±325. Pennell, K.D., Rhue, R.D., Rao, P.S.C., 1992. Vapor-phase sorption of p-xylene and water on soils and clay minerals. Environ. Sci. Technol. 26, 756±765. Perfect, E., Kay, B.D., 1995. Applications of fractals in soil and tillage research: a review. Soil Till. Res. 36, 1±20. Petersen, L.W., Moldrup, P., Jacobsen, O.H., Rolston, D.E., 1996. Relation between soil speci®c area and soil physical and chemical properties. Soil Sci. 161, 9±21. Pfeifer, P., Cole, M.W., 1990. Fractals in surface science: scattering and thermodynamics of adsorbed ®lms. II. New J. Chem. 14, 221±232. Poulsen, T.G., Moldrup, P., Hansen, J.A., 1998. VOC vapor sorption in soil: soil type dependent model and implications for vapor extraction. J. Environ. Eng. 124, 145±157. Rhue, R.D., Rao, P.S.C., Smith, R.E., 1988. Vapor phase adsorption of alkylbenzenes and water on soils and clays. Chemosphere 17, 727±741. Rice, J.A., Tombacz, E., Malekani, K., 1999. Applications of light and X-ray scattering to characterize the fractal properties of soil organic matter. Geoderma 88, 251±267. Russ, J.C., 1994. Fractal Surfaces. Plenum Press, New York. Sahouli, B., Blacher, S., Brouers, F., 1996. Fractal surface analysis by using nitrogen adsorption data: the case of the capillary condensation regime. Langmuir 12, 2872±2874. Sapozhnikov, P.M., 1985. Relationship between the swelling of some soils and speci®c surface categories and the energetics of soil moisture. Sov. Soil Sci. 17, 87±93. Seri-Levy, A., Avnir, D., 1993. Kinetics of diffusion-limited
adsorption on fractal surfaces. J. Phys. Chem. 97, 10380± 10385. Shirshova, L.T., Yermolayev, A.M., 1990. Pecularities of humus accumulation in arable Gray Forest Soil during grassing. J. Gen. Biol. 51, 642±650. Singer, M.J., Ewing, S., 2000. Soil quality. In: M.E. Sumner (Ed.), Handbook of Soil Science. CRC Press, Boca Raton, FL, pp. G271±G297. Skidmore, E.L., Layton, J.B., 1992. Dry-soil aggregate stability as in¯uenced by selected soil properties. Soil Sci. Soc. Am. J. 56, 557±561. Sokoøowska, Z., 1989. On the role of energetic and geometric heterogeneity in sorption of water vapor by soils: application of a fractal approach. Geoderma 45, 251±265. Sokoøowska, Z., Sokoøowski, S., 1999. In¯uence of humic acid on surface fractal dimension of kaolin: analysis of mercury porosimetry and water vapor adsorption data. Geoderma 88, 233±245. Tester, C.F., 1990. Organic amendment effects on physical and chemical properties of a sandy soil. Soil Sci. Soc. Am. J. 54, 827±831. Toledo, P.G., Nony, R.A., Davis, H.T., Scriven, L.E., 1990. Hydraulic conductivity of porous media at low water content. Soil Sci. Soc. Am. J. 54, 673±679. Valverde-Garcia, A., Gonzales-Pradas, E., Villafranca-Sanchez, M., del Rey-Beno, F., Garcia-Rodrigues, A., 1988. Adsorption of thiram and dimethoate on Almeria soils. Soil Sci. Soc. Am. J. 52, 1571±1574. Yermolayev, A.M., Shirshova, L.T., 1988. Dynamics of plant organic matter and certain humus fractions in Gray Forest soil under arti®cial grassland. Sov. J. Ecol. 19, 10±16. Yermolayev, A.M., Shirshova, L.T., 1994. Productivity and functioning of a perennial sown meadow of various management type. Pochvovedenie 12, 97±105 (in Russian). Yin, Y., 1991. Adsorption isotherm of fractally porous material. Langmuir 7, 216±217.
Soil pore surface properties in managed grasslands Meczislaw Hajnosa, Ludmila Korsunskaiab, Yakov Pachepskyc,* a Institute of Agrophysics, Lublin 20536, Poland Institute of Physical, Chemical, and Biological Problems in Soil Science, Pushchino 142292, Russia c USDA:ARS:BA:NRI:RSML, Hydrology Laboratory, Bldg. 007, Rm. 104, BARC-WEST, Beltsville, MD 20705, USA b
Received 3 August 1999; received in revised form 7 January 2000; accepted 28 January 2000
Abstract Properties of pore surfaces control adsorption and transport of water and chemicals in soils. Parameters are needed to recognize and monitor changes in pore surfaces caused by differences in soil management. Data on gas adsorption in soils can be compressed into parameters characterizing (a) area available to a particular adsorbate, and (b) surface roughness or irregularity. Our objectives were to see (a) whether models of adsorption on fractal surfaces are applicable to water vapor adsorption in soils in the capillary condensation range, and (b) whether differences in long-term management of grasslands are re¯ected by soil pore surface properties. Water vapor adsorption was measured in Gray Forest soil (Udic Argiboroll, Orthic Greyezem, clay loam) samples taken at four plots, where a long-term experiment on grassing arable land had been carried out for 12 years. The experiment had 22 design. Factors were `harvest±no-harvest' and `fertilizer±no-fertilizer'. The hay was cut after over-seeding in harvested treatments every year. Ammonium nitrate, superphosphate, and potassium chloride were applied annually after the snowmelt to get the total amount of nutrients of 60 kg haÿ1. The monolayer adsorption capacity was estimated from the Brunauer±Emmett±Teller model. A fractal Frenkel±Halsey±Hill model of adsorption on a fractal surface, and a thermodynamic adsorption model were applied in the range of relative pressures from 0.7 to 0.98 and provided good ®t of data. Values of the surface fractal dimension Ds were in the range from 2.75 to 2.85. Removal of carbohydrates resulted in increase of Ds. Differences in management practices did not affect values of Ds in the scale range studied, whereas the monolayer capacity was affected. Both fertilization and harvesting resulted in an increase of the monolayer capacity, with the largest increase observed in soil that was fertilized but not harvested. # 2000 Elsevier Science B.V. All rights reserved. Keywords: Water vapor; Adsorption; Fractal surface; Grassland; Gray Forest soil
1. Introduction Properties of pore surfaces control adsorption and transport of water and chemicals in soils. Parameters are needed to recognize and monitor changes in pore *
Corresponding author. Tel.: 1-301-504-7468; fax: 1-301-504-5823. E-mail address: [email protected] (Y. Pachepsky)
surfaces caused by differences in soil management. Once established, such diagnostic parameters can be used to quantify effects of management on soil functioning by means of land quality or soil quality indices (Singer and Ewing, 2000). Data on gas adsorption in soils can be compressed into parameters characterizing (a) area available to a particular adsorbate, and (b) surface roughness or irregularity. Speci®c surface area is known to char-
0167-1987/00/$ ± see front matter # 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 1 9 8 7 ( 0 0 ) 0 0 0 9 9 - 4
64
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
acterize the ability of soil to retain and transport nutrients and water. Speci®c surface area of soils correlates with the sorptive capacity and ion exchange capacity (Jurinak and Volman, 1957; Ben-Dor and Banin, 1995), retention and release of chemicals (Rhue et al., 1988; Valverde-Garcia et al., 1988; Ong and Lion, 1991), swelling (Sapozhnikov, 1985), water retention (Kapilevich et al., 1987), dry-aggregate stability (Skidmore and Layton, 1992). Tester (1990) observed an increase in speci®c surface area parallel to the increase in organic matter content after compost application. Pore surfaces in soils are irregular. Fractal models were found useful in the description of natural and man-made irregular surfaces. Fractal surfaces exhibit dependencies of their area S on the resolution, or scale, R at which the area is measured (Russ, 1994). These dependencies have a power law form S / R2ÿDs
(1)
where Ds is the surface fractal dimension. The surface fractal dimension characterizes the degree of roughness or disorder of the surface. Numerical values of surface fractal dimensions vary from 2.0 (perfectly smooth) to a maximum value of 3.0. The larger the value of Ds, the greater is the irregularity of the surface. Theoretical studies of adsorption on fractal surfaces predicted the effect of the surface fractal dimension on diffusion-limited adsorption kinetics (Seri-Levy and Avnir, 1993; Giona and Giustiniani, 1996). Douglas (1989) showed that adsorption of polymers on rough surfaces should be affected by the surface fractal dimension. The increase in fractal dimensions should cause tighter binding of polymer with a surface. Surface fractal dimensions were studied with relation to technological parameters in production of ionexchange ®bers (Kutarov and Kats, 1992), carbon blacks (Darmstadt et al., 1995), cements (Niklasson, 1993), food products (Aguerre et al., 1996), and wood (Hatzikiriakos and Avramidis, 1994). Fractal scaling was demonstrated for soil pore surfaces. Sokoøowska (1989) used a fractal model to analyze the soil particle surface heterogeneity with data on water vapor adsorption experiments. Toledo et al. (1990) showed that fractal dimensions of soil particle surfaces could be estimated from water retention data at low moisture contents. Bartoli et al. (1991)
introduced the use of mercury porosimetry to estimate surface fractal dimensions in soils. Soil constituents, such as humus materials and clay minerals, were found to be surface fractals (Rice et al., 1999; Sokoøowska and Sokoøowski, 1999). The management-related change in soil quality is a current topic of interest. Efforts were made to de®ne soil quality and suggest a minimum set of analyses to aid in evaluating soil quality as affected by long-term management practices (Doran and Parkin, 1994). Long-term and short-term management effects on soil pore surface fractal dimensions were observed using soil thin sections (Pachepsky et al., 1996b; GimeÂnez et al., 1997; Oleschko et al., 1998). Fractal dimension analysis provided additional information by describing the changes in pore shapes. Adsorption isotherms of gases present an attractive source of data on surface roughness. Several theories of adsorption on a fractal surface were developed during last 15 years. Two types of adsorption data analysis have been used most often. One of them uses several different adsorbates and employs the mean radius of each adsorbate's molecules as the resolution, or scale, R in Eq. (1). The logarithmic transform of this equation and the subsequent regression `radius vs. speci®c surface' in log±log scale provides the value of Ds as the slope subtracted from 2 (Farin and Avnir, 1989). With this approach, data on adsorption at low relative pressures are used at which the monolayer coverage is known to occur. Another type of the adsorption data analysis employs the data at high relative pressures, where capillary condensation occurs so that micropores become ®lled with adsorbate. Then the micropore radius becomes a measure of scale R in Eq. (1). Avnir and Jaroniec (1989) obtained a relationship between the adsorbed amount and vapor pressure by assuming the applicability of the Dubinin± Radushkievich equation to describe the adsorption in micropores, and assuming fractal pore volume distribution on the surface. The same equation was obtained by Neimark (1990) and Yin (1991) from the assumption that the capillary condensation was present. A different model was proposed by Neimark (1990). He did not assume validity of any adsorption isotherm equation, but used directly the scaling relationship (1) with the surface curvature radius as a scale measure. Although both models mentioned above were applied in experimental studies, no unambiguous way was
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
found to see which one is superior (Jaroniec et al., 1997). Our objectives were to see (a) whether models of adsorption on fractal surfaces are applicable to water vapor adsorption in soils in the capillary condensation range, and (b) whether differences in long-term management of grasslands are re¯ected by soil pore surface properties. 2. Materials and methods 2.1. Soil sampling and analyses Gray Forest soil (Udic Argiboroll) was sampled in April 1994 near Pushchino, Moscow Region, Russia. Samples were taken at a site, where an experiment on grassing arable land had been carried out since 1982 (Yermolayev and Shirshova, 1994). Grass mixture consisted of Festuca Pratensis, Huds., Phleum Pratense L., and Trifolium Pratense L. The experiment had 22 design. Factors were `harvest±no-harvest' and `fertilizer±no-fertilizer'. The hay was cut after over-seeding in harvested treatments every year. Ammonium nitrate, superphosphate, and potassium chloride were applied annually after the snowmelt to get the total amount of nutrients of 60 kg haÿ1. Area of each plot was 0.5 ha. The soil was in no-till, since 1982. The soil had the pH 6.3, a loamy texture with 45% particles less than 0.01 mm and 20% particles less than 0.001 mm, and contained illite, kaolinite, and smectites in its clay fraction. Sampling depth was 10± 30 cm. Five samples were taken from each plot, and the mixed sample was used in adsorption measurements for each of the four combinations of factors. Soil was air-dried and sieved, and aggregates less than 2 mm were used. Each sample was split, and carbohydrates were removed in one of the two subsamples. The oxidation of carbohydrates has been performed with 0.02 M NaIO4 solution (Cheshire et al., 1984). Water vapor adsorption isotherms were measured with a vacuum microbalance technique. The temperature was kept within the range of 2940.1 K. The adsorption isotherms were measured in duplicate at 14 relative pressure levels of 0.0217, 0.0902, 0.1246, 0.2142, 0.3008, 0.3404, 0.4112, 0.5324, 0.6075, 0.6533, 0.7320, 0.7727, 0.8680, 0.9826.
65
For comparison purposes, we also used published data on water vapor adsorption on clay minerals (Mooney et al., 1952; Orchiston, 1954; Jurinak, 1963). The BET equation (Adamson, 1967) N N0
c p=p0 1 ÿ p=p0 1 ÿ c ÿ 1p=p0
(2)
was ®tted to the experimental isotherms in the range of relative pressures p/p0 between 0.02 and 0.34. Here and below N is the adsorbate amount, p/p0 is the relative pressure de®ned as a ratio of the actual pressure p to the saturation vapor pressure p0. The monolayer capacity N0 and thermodynamic constant c were adjusted to achieve the best ®t. 3. Estimating fractal dimensions Avnir and Jaroniec (1989) developed the fractal analogue of the Frenkel±Halsey±Hill (FHH) equation in the following form: p (3) ln N C ÿ Ds ÿ 3 ln p0 Here C is the intercept. The same equation was derived by Yin (1991) and Neimark (1990) by assuming the sequential ®lling of pores from small size to large, that is capillary condensation-type local adsorption isotherm. A similar equation was derived by Pfeifer and Cole (1990) by generalization of the classical FHH adsorption theory for the capillary condensation regime. Neimark pointed out that Eq. (1) should be used with data obtained for p/p0>0.7. Results of Pfeifer and Cole (1990) also indicate that Eq. (1) is valid at high p/p0 values when the capillary condensation is the dominant mechanism. We estimated the surface fractal dimension Ds from the slope of the regression given by Eq. (1) at relative pressures p/p0 in the range from 0.7320 to 0.9826. Neimark (1990) and Neimark and Unger (1993) proposed a different model for calculating the fractal dimension Ds from the adsorption isotherm data. It was called the thermodynamic model. The model re¯ects the idea that during the adsorption on a fractal surface, the surface is smoothed by the adsorbate, and the surface of the adsorbent covered with an adsorbate is smaller than the surface of the uncovered adsorbent. The area of the `condensed adsorbate±
66
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
vapor' equilibrium interface is obtained from the Kiselev equation Z RT Nmax p ÿln dN (4) S g N p0 Here Nmax is the maximum adsorption, R the gas constant, T the Kelvin temperature, and g is the surface tension of the liquid adsorbate. This area is related by the fractal scaling law (1) to the mean scaling radius of this interface R which is calculated from Kelvin's equation R
2gVm RT ÿln p=p0
(5)
where Vm is the molar volume of the liquid adsorbate. The logarithmic transform of (1) with S from (4) and R from (5) results in the equation of the thermodynamic model: p (6) ln S a Ds ÿ 2 ÿln p0 Here a is the intercept. We estimated the surface fractal dimension Ds from the slope of the regression given by Eq. (6) at relative pressure p/p0 in the range from 0.7320 to 0.9826. The integral in the left-hand side of Eq. (4) was evaluated numerically with the trapezoid method. The maximum adsorbed amount was approximately equated to the adsorbed amount at p/p00.9826. Parameters of regression equations and their standard error estimates were found using the SigmaPlot
Fig. 1. Water vapor adsorption isotherms on Gray Forest soil (Udic Argiboroll) in grasslands; N is the adsorbed amount, p the actual vapor pressure, and p0 is the vapor pressure at saturation; (*, *) no harvesting, no fertilization, (&, & ) no harvesting, 60 kg NPK annually, (~, ~) harvested, no fertilization, (!, 5) harvested, 60 kg NPK annually; hollow symbols show data for the samples in which carbohydrates have been removed.
Version 2 software package. The statistical signi®cance of differences was tested with the t-test at the 0.05 signi®cance level. 4. Results and discussion Adsorption data are shown in Fig. 1. Shapes of the curves are similar. At the same relative pressure values, the adsorbed amount is the highest in soil from the non-harvested, fertilized plot, and is the lowest in soil from the non-harvested, non-fertilized plot. Intermediate values of adsorbed water amounts
Table 1 Parameters of soil pore surfaces from water vapor adsorption data Sample
Organic carbon Treatment content (%)
No harvest, no fertilizer 1.92 Ndb No harvest, 60 kg NPK 1.76 Nd Harvested, no fertilizer 1.68 Nd Harvested, 60 kg NPK 1.64 Nd a b
Original sample Carbohydrates removed Original sample Carbohydrates removed Original sample Carbohydrates removed Original sample Carbohydrates removed
Standard errors are shown after the `' sign. Not determined.
Surface fractal dimensions
Parameters of the BET equation
Fractal FHH model
Thermodynamic model
Monolayer BET capacity (g kgÿ1) c-value
2.770.01a 2.790.01 2.780.00 2.790.01 2.780.00 2.780.01 2.780.00 2.780.01
2.750.08 2.790.08 2.760.04 2.830.04 2.780.03 2.850.03 2.740.05 2.850.05
9.510.11 8.780.10 15.650.16 16.640.17 13.610.16 13.040.15 14.190.18 12.670.15
11118 14527 8611 15026 8512 7410 7410 719
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
Fig. 2. Fractal scaling of pore surface from data on water vapor adsorption on clay minerals; N is the adsorbed amount, p the actual vapor pressure, and p0 is the vapor pressure at saturation; (*) Namontmorillonite (Mooney et al., 1952), (*) Li-kaolinite (Jurinak, 1963), (~) illite (Orchiston, 1954).
are found in soils from harvested plots. These effects of treatments are re¯ected in values of the monolayer capacity N0 found from the BET equation (Table 1). The monolayer capacity did not correlate with the organic carbon content in this study. This could be related to the differences in composition of organic matter formed in soil under different management in this experiment (Yermolayev and Shirshova, 1988; Shirshova and Yermolayev, 1990).
67
Removal of carbohydrates caused a small decrease of the monolayer capacity of water adsorption in soil from all but non-harvested, fertilized plots. Feller et al. (1992) observed that partial removal of carbon by treatments with H2O2 increased the surface area of clay-sized organo-mineral complexes while surface areas of silt and sand were less affected. Oades (1984) suggested that clays were embedded in a matrix of soil organic matter. Therefore, in our study, differences in composition of soil organic matter could be related to the type of particles to which it was bounded. Application of the fractal FHH model to data on soil minerals is shown in Fig. 2. The FHH model provides a reasonably good ®t to the adsorption data. Surface fractal dimensions increase in the sequence `illite±Likaolinite±Na-montmorillonite'. The values of Ds are 2.49, 2.62, and 2.74, respectively. Application of both the fractal FHH model and the thermodynamic model to soil data is shown in Fig. 3. The dependencies between transformed data are very close to linear ones as predicted by both models. Estimated values of surface fractal dimensions are given in Table 1. The standard errors of the fractal dimension estimates are larger for the thermodynamic model than for the fractal FHH model. Estimated values of Ds range from 2.73 to 2.78 for untreated samples and from 2.78 to 2.85 in samples, where carbohydrates were removed. Our data do not indicate that either of the two compared models is superior.
Fig. 3. Fractal scaling of the pore surface demonstrated with two models of vapor adsorption on fractal surfaces: (a) fractal Frenkel±Halsey± Hill model, (b) thermodynamic model; N is the adsorbed amount, Nmax the maximum adsorbed amount, p the actual vapor pressure, and p0 is the vapor pressure at saturation; (*) no harvesting, no fertilization, (&) no harvesting, 60 kg NPK annually, (~) harvested, no fertilization, (!) harvested, 60 kg NPK annually.
68
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
The values of fractal dimensions are close for both models applied to the same sample, and no statistically signi®cant differences are encountered. This might be expected since we used data in the range of relative pressures, where the capillary condensation was a dominating adsorption process. Jaroniec (1995) had shown that the Eq. (6) can be derived from the FHH equation under the assumption of the condensation approximation validity. Sahouli et al. (1996) observed the equivalence in results of application both the fractal FHH equation and the thermodynamic model (6) when the adsorption process was dominated by the capillary condensation in carbon black. Darmstadt et al. (1995) applied both the fractal FHH model (3) and the thermodynamic model (6) to data on N2 adsorption on carbon blacks. They found the fractal dimensions from both models to be very close in three of four samples. Their fourth sample had much lower fractal dimensions than the other three. It was suggested that this sample had mesopores where the capillary condensation occurred which led to overcompensating the reduced adsorption due to the fractal surface. Fractal dimension values did not differ between the experimental treatments. The soil surface irregularity remained the same. The intercepts in graphs of Fig. 2 show differences among the treatments. Soil in harvested plots displayed similarities in parameters of both fractal models. For non-harvested plots, the intercepts are different from each other and from the harvested plots. The differences in intercepts re¯ect the difference in ®lm surface area at the same capillary radius value. The larger the intercept, the larger is the ®lm surface area. Therefore, within the studied radii range from 3 to 60 nm, the number of pores of any given radius is the largest in the soil of the non-harvested fertilized plot. As the radius changes, this number scales in a similar manner in all plots. Several fractal models of pore surfaces were successfully used to distinguish effects of tillage and compaction on soil structure (Perfect and Kay, 1995). Scales in these studies corresponded to the aggregate sizes in dry sieving, i.e. 0.3±30 mm (GimeÂnez et al., 1997), or to the recognizable features in soil thick and thin sections, i.e. 0.008±3 mm (Pachepsky et al., 1996b; Oleszko et al., 1998). In the above studies, the authors measured soil structure directly
using variants of the box-counting method. Scales were much smaller in this study, and the resolution was de®ned indirectly by converting pressures to the radii. Different sensitivity of fractal parameters to long-term soil management and laboratory treatments was found in different ranges of scales (Pachepsky et al., 1996a; GimeÂnez et al., 1997). The absence of differences in fractal dimensions in the capillary condensation region does not imply that the grassland management does not affect surface roughness at all. The differences are absent in the speci®c range of pore radii from 3 to 60 nm, but it is possible that differences in surface roughness exist in other ranges of radii. Carbohydrate removal had an effect on the values of fractal dimensions estimated with the thermodynamic model but not with the fractal FHH (Table 1). No statistically signi®cant differences were found between the fractal dimension estimates in original samples and samples with removed carbohydrates. The increase in surface fractal dimension values was expected since the irregularity of the surface should increase after the removal of structure-forming organic compounds. Removal of organic matter caused an increase in surface fractal dimensions measured with the mercury porosimetry in three soils studied by Pachepsky et al. (1996a). Compared with pure minerals in Fig. 2, the soil in this study had larger values of surface fractal dimensions. Mixed mineral composition and the presence of organic matter are probable reasons for that. Application of other fractal models to quantify irregularity of soil composition and structure also resulted in obtaining high, close to three, values of mass and fragmentation fractal dimensions for ®ne textured soils (Perfect and Kay, 1995). A caution has to be exercised in interpreting values of fractal dimensions obtained with water adsorption because of speci®c interactions of water molecules with surfaces. Niklasson (1993) has found that data on nitrogen adsorption tend to give larger values of surface fractal dimensions than data on water adsorption. This author observed a close correspondence between fractal dimensions obtained from water adsorption and from small-angle neutron scattering. Pachepsky et al. (1996b) used mercury porosimetry to measure the surface fractal dimension under developing barley crop for the same soil as in this study but in agricultural use. In samples taken during spring, they found
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
values of Ds2.870.10 which did not differ signi®cantly from the values in Table 1. Fractal scaling presents additional parameters to characterize soil pore surfaces. Speci®c surface area was shown to be related to various physical and chemical properties of soil (Petersen et al., 1996). The region of relative pressures between 0.05 and 0.3 is used to ®nd the speci®c surface area (Adamson, 1967; Kutilek and Nielsen, 1994). Fractal scaling parameters as de®ned in this paper give information about different range of relative pressures, and therefore, different range of pore radii. There is a growing interest in competitive adsorption of volatile organic compounds and water in soils as related to soil bioremediation and fate of fumigants (e.g. Pennell et al., 1992; Poulsen et al., 1998). Fractal parameters of pore surfaces may ®nd applications in this ®eld. In conclusion, the models of adsorption on fractal surfaces appear to be applicable to water adsorption in soil in this study. Differences in long-term management of the grassland did not affect surface fractal dimensions, although the surface area was affected. References Adamson, A.W., 1967. Physical Chemical of Surfaces. Interscience, New York. Aguerre, R.J., Viollaz, R.E., Suarez, C., 1996. A fractal isotherm for multilayered adsorption in foods. J. Food Eng. 30, 227±238. Avnir, D., Jaroniec, M., 1989. An isotherm equation for adsorption on fractal surfaces of heterogeneous porous material. Langmuir 5, 1431±1433. Bartoli, F., Phylippy, R., Doirisse, M., Niquet, S., Dubuit, M., 1991. Structure and self-similarity in silty and sandy soils: the fractal approach. J. Soil Sci. 42, 167±185. Ben-Dor, E., Banin, A., 1995. Near-infrared analysis as a rapid method to simultaneously evaluate several soil properties. Soil Sci. Soc. Am. J. 59, 364±372. Cheshire, M.W., Sparling, J.P., Mundle, C.M., 1984. In¯uence of soil type, crop, and air drying on residual carbohydrate content. Plant and Soil 76, 339±347. Darmstadt, H., Roy, C., Kallaguine, S., Sahouli, B., Blasher, S., Pirard, R., Brouers, F., 1995. Fractal analysis of commercial and pyrolytic carbon blacks using nitrogen adsorption data. Rubber Chem. Technol. 68, 330±341. Doran, J.W., Parkin, T.B., 1994. De®ning and assessing soil quality. In: Doran, J.W., Coleman, D.C., Bezdicek, D.F., Stewart, B.A. (Eds.), De®ning Soil Quality for a Sustainable Environment. SSSA Special Publ. No. 35, Madison, WI, pp. 3±21. Douglas, J.F., 1989. How does surface roughness affect polymer± surface interactions? Macromolecules 22, 3707±3716.
69
Farin, D., Avnir, D., 1989. The fractal nature of molecule±surface interactions and reactions. In: Avnir, D. (Ed.), The Fractal Approach to Heterogeneous Chemistry. Wiley, New York, pp. 271±293. Feller, C., Shouller, E., Thomas, F., Rouiller, J., Herbillon, A.J., 1992. N2-BET speci®c surface areas of some low activity clay soils and their relationships with secondary constituents and organic matter contents. Soil Sci. 153, 293±299. GimeÂnez, D., Allmaras, R.R., Nater, E.A., Huggins, D.R., 1997. Fractal dimensions for volume and surface of interaggregate pores: scale effects. Geoderma 77, 19±38. Giona, M., Giustiniani, M., 1996. Adsorption kinetics on fractal surfaces. J. Phys. Chem. 100, 16690±16699. Hatzikiriakos, S.G., Avramidis, S., 1994. Fractal dimension of wood surfaces from sorption isotherms. Wood Sci. Technol. 28, 275±284. Jaroniec, M., 1995. Evaluation of the fractal dimension from a single adsorption isotherm. Langmuir 11, 2316±2317. Jaroniec, M., Kruk, M., Olivier, J., 1997. Fractal analysis of composite adsorption isotherms obtained by using density functional theory data for argon in slitlike pores. Langmuir 13, 1031±1035. Jurinak, J.J., 1963. Multilayer adsorption of water by kaolinite. Soil Sci. Soc. Am. Proc. 27, 269±272. Jurinak, J.J., Volman, D.H., 1957. Application of Brunauer, Emmett and Teller equation to ethylene dibromide adsorption by soils. Soil Sci. 83, 487±496. Kapilevich, Zh.A., Pisentsky, G.A., Vysochenko, A.V., 1987. Speci®c surface as a basic parameter for evaluating hydromelioration characteristics of mineral soils. Sov. Soil Sci. 19, 68±76. Kutarov, V.V., Kats, B.M., 1992. Determination of the fractal dimension of ion-exchange ®bers from adsorption data. Russ. J. Phys. Chem. 67, 1854±1856. Kutilek, M., Nielsen, D., 1994. Soil Hydrology. Catena Verlag, Cremlingen-Destedt, Germany. Mooney, R.W., Keenan, A.G., Wood, L.A., 1952. Adsorption of water vapor by montmorillonite. Part I. Heat of desorption and application of BET theory. J. Am. Chem. Soc. 74, 1367± 1371. Neimark, A.V., 1990. Determination of surface fractal dimension from the results of an adsorption experiment. Russ. J. Phys. Chem. 64, 1398±1403. Neimark, A.V., Unger, K.K., 1993. Method of discrimination of surface fractality. J. Colloid Interf. Sci. 158, 412±419. Niklasson, G.A., 1993. Adsorption on fractal structures: applications to cement materials. Cem. Conc. Res. 23, 1153±1158. Oades, J.M., 1984. Soil organic matter and structural stability: mechanisms and implications for management. Plant and Soil 76, 319±337. Oleschko, K., Brambila, F., Acess, F., Mora, L.P., 1998. From fractal analysis along a line to fractals on the plane. Soil and Tillage Research 45, 389±406. Ong, S.K., Lion, L.W., 1991. Trichloroethylene vapor sorption onto soil minerals. Soil Sci. Soc. Am. J. 55, 1559±1568. Orchiston, H.D., 1954. Adsorption of water vapor: clays at 258C. Soil Sci. 78, 463±480.
70
M. Hajnos et al. / Soil & Tillage Research 55 (2000) 63±70
Pachepsky, Ya.A., Korsunskaia, L.P., Hajnos, M., 1996a. Fractal parameters of soil pore surface area under a developing crop. Fractals 4, 97±110. Pachepsky, Ya.A., Yakovchenko, V., Rabenhorst, M.C., Pooley, C., Sikora, L.J., 1996b. Fractal parameters of pore surfaces as derived from micromorphological data: effect of long-term management practices. Geoderma 74, 305±325. Pennell, K.D., Rhue, R.D., Rao, P.S.C., 1992. Vapor-phase sorption of p-xylene and water on soils and clay minerals. Environ. Sci. Technol. 26, 756±765. Perfect, E., Kay, B.D., 1995. Applications of fractals in soil and tillage research: a review. Soil Till. Res. 36, 1±20. Petersen, L.W., Moldrup, P., Jacobsen, O.H., Rolston, D.E., 1996. Relation between soil speci®c area and soil physical and chemical properties. Soil Sci. 161, 9±21. Pfeifer, P., Cole, M.W., 1990. Fractals in surface science: scattering and thermodynamics of adsorbed ®lms. II. New J. Chem. 14, 221±232. Poulsen, T.G., Moldrup, P., Hansen, J.A., 1998. VOC vapor sorption in soil: soil type dependent model and implications for vapor extraction. J. Environ. Eng. 124, 145±157. Rhue, R.D., Rao, P.S.C., Smith, R.E., 1988. Vapor phase adsorption of alkylbenzenes and water on soils and clays. Chemosphere 17, 727±741. Rice, J.A., Tombacz, E., Malekani, K., 1999. Applications of light and X-ray scattering to characterize the fractal properties of soil organic matter. Geoderma 88, 251±267. Russ, J.C., 1994. Fractal Surfaces. Plenum Press, New York. Sahouli, B., Blacher, S., Brouers, F., 1996. Fractal surface analysis by using nitrogen adsorption data: the case of the capillary condensation regime. Langmuir 12, 2872±2874. Sapozhnikov, P.M., 1985. Relationship between the swelling of some soils and speci®c surface categories and the energetics of soil moisture. Sov. Soil Sci. 17, 87±93. Seri-Levy, A., Avnir, D., 1993. Kinetics of diffusion-limited
adsorption on fractal surfaces. J. Phys. Chem. 97, 10380± 10385. Shirshova, L.T., Yermolayev, A.M., 1990. Pecularities of humus accumulation in arable Gray Forest Soil during grassing. J. Gen. Biol. 51, 642±650. Singer, M.J., Ewing, S., 2000. Soil quality. In: M.E. Sumner (Ed.), Handbook of Soil Science. CRC Press, Boca Raton, FL, pp. G271±G297. Skidmore, E.L., Layton, J.B., 1992. Dry-soil aggregate stability as in¯uenced by selected soil properties. Soil Sci. Soc. Am. J. 56, 557±561. Sokoøowska, Z., 1989. On the role of energetic and geometric heterogeneity in sorption of water vapor by soils: application of a fractal approach. Geoderma 45, 251±265. Sokoøowska, Z., Sokoøowski, S., 1999. In¯uence of humic acid on surface fractal dimension of kaolin: analysis of mercury porosimetry and water vapor adsorption data. Geoderma 88, 233±245. Tester, C.F., 1990. Organic amendment effects on physical and chemical properties of a sandy soil. Soil Sci. Soc. Am. J. 54, 827±831. Toledo, P.G., Nony, R.A., Davis, H.T., Scriven, L.E., 1990. Hydraulic conductivity of porous media at low water content. Soil Sci. Soc. Am. J. 54, 673±679. Valverde-Garcia, A., Gonzales-Pradas, E., Villafranca-Sanchez, M., del Rey-Beno, F., Garcia-Rodrigues, A., 1988. Adsorption of thiram and dimethoate on Almeria soils. Soil Sci. Soc. Am. J. 52, 1571±1574. Yermolayev, A.M., Shirshova, L.T., 1988. Dynamics of plant organic matter and certain humus fractions in Gray Forest soil under arti®cial grassland. Sov. J. Ecol. 19, 10±16. Yermolayev, A.M., Shirshova, L.T., 1994. Productivity and functioning of a perennial sown meadow of various management type. Pochvovedenie 12, 97±105 (in Russian). Yin, Y., 1991. Adsorption isotherm of fractally porous material. Langmuir 7, 216±217.