Environ. Sci. Technol. 2001, 35, 4947-4952
Effect of Cyclodextrins on Surface and Pore Properties of Soil Clay Minerals G R Z E G O R Z J O Z E F A C I U K , * ,† ATTILA MURANYI,‡ AND EVA FENYVESI§ Institute of Agrophysics of Polish Academy of Sciences, Doswiadczalna 4, 20-290 Lublin, Poland, Research Institute for Soil Science and Agricultural Chemistry of the Hungarian Academy of Sciences, Herman Otto´ 15, 1022 Budapest, Hungary, and Cyclolab R&D Laboratory Ltd., Illatos 7, 1097 Budapest, Hungary
Although cyclodextrins are increasingly used in soil decontamination, little is known about their effects on soil physicochemical properties. In this work, the surface and pore properties of randomly methylated β-cyclodextrin (RAMEB) and three typical clay minerals were characterized, and the effects of RAMEB concentrations on clay minerals were studied using water vapor adsorption-desorption and mercury intrusion porosimetry techniques. As compared to clay minerals, for pure RAMEB very large surface area and volume of nanometer-size pores (micropores) were determined. Energy of interaction with water vapor, volume of micrometer-size pores (mesopores), and fractal dimensions in both pore size ranges of RAMEB were lower than those of the minerals. When increasing amounts of RAMEB were added to the minerals, the surface area and micropore volume decreased and adsorption energy increased. The volume of mesopores decreased after RAMEB treatments for bentonite and kaolin and increased for illite. As deduced from the fractal dimensions increase, the pore structure of the minerals became more complex with RAMEB addition. The observed changes were in general contrary to these expected when RAMEB and minerals coexist as separate, nonreactive phases and suggested strong interaction of RAMEB with clay minerals.
Introduction Cyclodextrins are crystalline, water-soluble, nonreducing, cyclic, toroidially shaped oligosaccharides built from 6 to 12 glucopyranose units. The toroid resembles an open barrel where the internal cavity dimension is determined by the number of glucose units. The orientation of the glucose units is such that the outer edges of the barrel are ringed with primary and secondary hydroxyl groups enabling the hydrophilic character of the molecule. However, inside the cavity, high electron density of the glucosidic oxygen creates a hydrophobic trap into which a great variety of low-polarity guest molecules of sufficient size can be encapsulated (1). These inclusion-complexing properties of cyclodextrins make them useful in many research, industrial (2, 3), and environmental applications (4-8). * Corresponding author telephone: +48-81-7445061; fax: +4881-7445067; e-mail:
[email protected]. † Institute of Agrophysics of Polish Academy of Sciences. ‡ Research Institute for Soil Science and Agricultural Chemistry of the Hungarian Academy of Sciences. § Cyclolab R&D Laboratory Ltd. 10.1021/es010083z CCC: $20.00 Published on Web 11/17/2001
2001 American Chemical Society
Cyclodextrins and their derivatives remarkably increase the aqueous solubility and bioavailability of a great number of soil organic contaminants. Because of rapidly decreasing price, it is feasible to use these compounds to intensify ex situ and in situ soil-washing technologies and to improve bioremediation (9-11). Unmodified β- and γ-cyclodextrins enhance the biodegradation of PAHs and PCBs (12, 13). Hydroxypropyl β-cyclodextrin (HPBCD) is the best choice for in-situ flushing of nonaqueous-phase organic liquids and low-polarity organic compounds if minimal NAPL mobilization is desired (14). Randomly methylated 1-cyclodextrin (RAMEB) is the most promising derivative to enhance the biodegradation of mineral and motor oil contamination in the soil (15). Both HPBCD and RAMEB are effective in removal of explosives from contaminated soils of military areas (16). Most soil processes are governed by the physicochemical properties of soil solid phase, which depend in turn on soil mineral and organic composition. The complex structure and granular composition of the solid phase results in nonuniformity of the soil pore system. This is characterized by a pore size distribution function showing fractions of pores of different radii while the overall amount of the pores is characterized by the bulk and particle density or pore volume of the material (17). Complex geometry and chemical composition of the solid phase are the main reasons that surfaces of soil adsorbents are highly inhomogeneous (18). Such surfaces are characterized by the adsorption energy distribution function, showing fractions of surface sites of different adsorption energies, while the overall amount of the adsorbent surface is characterized by its surface area (19). The geometrical shape of natural media may appear similar at different scales of observation (magnification), which can be characterized by the fractal dimension (2022). As a rule, natural objects display the fractal behavior in a limited range of scales (23). Soil physicochemical properties described above are highly sensitive to various processes occurring in soils: organic matter leaching and oxidation, pH changes, silica accumulation, wetting-drying cycles, and many others (2330). A hypothesis of our research was that RAMEB modifies physicochemical properties of soils that can affect the efficiency of remediation practices. The behavior of soil minerals is a rational base for understanding soil processes. This work focuses on the effect of different RAMEB doses on selected surface and pore properties of typical clay minerals.
Materials and Methods Sample Preparation. Three typical clay minerals present in soils [bentonite (BEN), illite. (IT) and kaolin (KA)] and randomly methylated β-cyclodextrin (RAMEB, Wacker Chemie, Germany) containing on average 14 methyl groups per molecule were used in the experiments. Aqueous RAMEB solutions of different concentrations and the same volumes (equal to about 1 bedvolume of the minerals) were added to the minerals and slowly dried by about 1 week at a laboratory conditions. At the end of the drying process, the samples were carefully mixed with a glass rod to minimize nonuniform RAMEB distribution. Dry samples were gently ground in a mortar and used for the further experiments. The final content of RAMEB in the samples was 0.0 (control), 0.03, 0.10, 0.30, 1.0, 3.0, and 9.0 w/w % (dry mass). Surface Parameters. Water vapor adsorption isotherms were measured using vacuum chamber method at a temperature T ) 294 ( 0.1 K. To reach different relative water vapor pressures (p/p0), sulfuric acid of stepwise decreasing VOL. 35, NO. 24, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4947
concentrations (increase of p/p0) was used. The amount of adsorbed water, a (kg kg-1), at a given p/p0 was measured after 48 h of equilibration by weighing. The dry mass of the samples was estimated after completing the isotherms measurements, by 24 h oven-drying at 378 K. The Aranovich (31) isotherm was used for mathematical description of adsorption data. A linear form of this equation is
x/[a(1 - x)1/2] ) 1/(amC) + x/am
(1)
where x ) p/p0 at the temperature of the measurements; am (kg kg-1) is the amount of water molecules adsorbed in the first monolayer, i.e., the statistical monolayer capacity; and C ) exp((Ea - Ec)/RT is the constant related to the adsorption energy, Ea (J mol-1), and the condensation energy of water, Ec (J mol-1) and R is the universal gas constant. Unlike the standard Brunauer-Emmett-Teller (BET) isotherm, the Aranovich isotherm permits the presence of vacancies in the adsorbed layer, is thermodynamically correct, and fits the experimental polymolecular adsorption data within a broader range of relative pressures. After calculation of am values from the slopes (1/am) of eq 1, the surface areas of the studied samples, S, were found as
S ) LωamM-1
(2)
where L is Avogadro number, M (kg) is molecular mass of water, and ω is the area occupied by a single water molecule equal to 1.08 × 10-19 m2. The adsorption energy distribution functions showing fractions of surface sites of distinct adsorption energies, f(Ei), were calculated from adsorption data using a theory of adsorption on heterogeneous surfaces (32, 33) and applying the Aranovich isotherm to describe local adsorption effects and a condensation approximation (34):
f(Ei) ) [(1 - xi+1)1/2Θ(Ei+1) - (1 - xi)1/2Θ(Ei)]/(Ei+1 - Ei) (3) where Ei ) (Ea,i - Ec), Ea,i is the adsorption energy of the ith site, and Θ(p) ) a(p)/am is the total adsorption isotherm. The average water vapor adsorption energy, Eav, was calculated from f(Ei) values as n
Eav )
∑E f(E ) i
i
(4)
i)1
Details on the above experimental method and calculations can be found in refs 35 and 36. Micropore Parameters. Water vapor desorption isotherms were measured using the vacuum chamber method as described above, but sulfuric acid of stepwise increasing concentrations (decrease of p/p0) was used. From desorption values, characteristics of pores (17) ranging from ca. 1 to a few tens of nanometers were evaluated. These pores are later called micropores. The fractions of pores in the given range of radii, f(rav,i), were obtained from n
rav ) 1/2
∑f(r
av,i)(ri
+ ri+1)
(5)
i)1
and the average micropore radius, rav, was obtained from
f(rav,i) ) (v(ri+1) - v(ri))/vt
(6)
For the calculations, we assumed that the micropore radius, r, in the capillary condensation process is related to the vapor pressure, p, by the Kelvin equation for the hemispherical meniscus with zero water/solid contact angle and that the 4948
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 24, 2001
condensation in micropores occurs above p/p0 ) 0.35 (below this value, surface adsorption processes dominate). Thus, the total micropore volume, vt, was taken as the volume of adsorbed water at the maximum p/p0 value applied minus this at p/p0 ) 0.35. Details on the above calculations can be found in ref 26. The micropore fractal dimensions, D, were calculated from the slopes of the linear parts of the ln-ln plots of adsorption a vs adsorption potential A ) RT ln(p/p0), using (21, 37)
ln(a) ) constant + (D - 3) ln(A)
(7)
This equation is derived for multilayer adsorption region, i.e., for the values of ln(A) lower than zero. Details on this equation are presented in ref 38. Mesopore Parameters. Mercury intrusion porosimetry (MIP) measurements were performed for aggregates of the minerals. The aggregates of ca. 10 mm in diameter were handmade from mineral-water-saturated pastes, air-dried and subjected to four wetting-drying cycles to stabilize the structure. One cycle consisted of 2 days wetting in saturated water vapor atmosphere at a lab temperature following by 30 °C overnight drying. The aggregates were crushed onto a few millimeter (diameter) pieces. Prior to the measurements, these pieces were dried overnight at 378 K and outgassed in a vacuum at the lab temperature, which is the standard requirement for the MIP technique. The MIP curves relating the intruded volume of mercury, V, to the mesopore radius, R, were used to evaluate mesopore size distribution functions, average radii (39), and fractal dimensions of mesopore surface, Ds. Two former characteristics were calculated similarly as for micropores and the latter one from the linear parts of log-log plots of the first derivative of pore volume on radius vs radius using the following dependence (23):
Ds ) 2 - d log[dV(R)/dR)/d log R
(8)
More details on this equation can be found in refs 23 and 38. We decided to look for the fractal scaling only in the dominant mesopore range, i.e., in this range of pore sizes for which the increase of the pore volume vs radii is highest. In this range, the linear parts of log(dV/dR) vs log(R) plots of the minerals overlap with this of RAMEB. All measurements mentioned above were performed in three replicates. A number of the calculations requested defining linearity ranges. To do this, we used the procedure introduced by Yokoya et al. (40) (cf. also 41). According to this procedure, the measure of the linearity L for the set of the points in a plane is
L ) (4σxy2 + (σyy - σxx)2)1/2(σyy + σσyy)-1 where σxx, σyy, and σxy are the variances of x-coordinates, y-coordinates, and the covariance between x and y coordinate sets, respectively. The L value falls between 0 (for uncorrelated and random points) and 1 (for points on a straight line). To separate linearity ranges, the value of L is computed for the first three points, then for the first four, five, and so on while the value of L increases. The end of the linearity range is in the points after which the value of L begins to decrease. To detect eventually the next linearity range, the procedure is repeated considering the latter points to be the first points of the next linearity interval. Once the linearity range was detected, we checked if the linear regression coefficient decreases after excluding the first and/or the last point. If so, these points were rejected.
Results and Discussion General Results. Isotherms and Adsorption Energy Distribution. The experimental desorption isotherms for the RAMEB
FIGURE 1. Water vapor desorption isotherms for RAMEB and RAMEB-treated minerals. Abbreviations: BEN, bentonite; IT, illite; KA, kaolin. The number following the mineral abbreviation is the dose of RAMEB (%). and RAMEB-enriched minerals are presented in Figure 1. The isotherms were measured for the RAMEB in the forms of powder and crystals; however, these were the same within the range of experimental error. A very high amount of water is sorbed by RAMEB, so we expected an increase in water sorption on minerals after RAMEB addition. However, the isotherms for RAMEB-enriched minerals show lower adsorption as compared to the pure minerals for all but KA + 9% RAMEB samples. The water vapor sorption in RAMEB is a low-energy process, which was seen from the water vapor sorption energy distribution function. The highest amount of low-energy sites was present. Fractions of high- and medium-energy sites were small. The shape of this function reflects that water molecules interact with cyclodextrins practically only via hydrogen bonds, with energies close to water-water interactions (condensation energy). β-Cyclodextrin crystallizes with 6.5 water molecules in the internal cavity, involved in a well-defined network of hydrogen bonds (42). The sorption energy of the internal waters should be low also. However, for the minerals enriched with CD, we observe an increase of higher energy centers and a decrease of the lower energy ones with RAMEB concentration increase. Size Distribution and Fractal Scaling of Micropores. The micropore size distribution function for RAMEB has a shape of a single, rather symmetrical peak. The median micropore radius of RAMEB is ca. 10 nm. In general, for the studied minerals, fractions of smaller micropores decrease and of larger micropores increase with the increase of the RAMEB load. At intermediate RAMEB doses, an increase in smaller micropore fractions was observed for BEN and IT. In general, the volume of the micropores exhibits fractal scaling for all of the samples as evidenced by the linearity of ln-ln plots of adsorption vs adsorption potential. However, the range of fractal scaling of the studied materials extends below the traditional multilayer regime and covers pore radii down to ca. 1 nm. Such small pores can be referred to also as surface roughness. The ranges of the fractal scaling (linearity intervals) are different for different minerals, and these do not change significantly after RAMEB addition. The slope of the linear part of these plots is highest for RAMEB, indicating that the water sorption process by RAMEB exhibits
FIGURE 2. Mercury intrusion porosimetry curves for RAMEB and RAMEB-treated minerals. Abbreviations: BEN, bentonite; IT, illite; KA, kaolin. The number following the mineral abbreviation is the dose of RAMEB (%). apparently fractal scaling with a very low fractal dimension. The slopes and fractal scaling ranges for the minerals do not change markedly with RAMEB addition. MIP Curves and Fractal Scaling of Mesopores. Figure 2 shows mercury intrusion porosimetry curves. The curve for RAMEB was measured for its crystalline form. As seen from the MIP curve, the largest increase in mesopore volume of the RAMEB occurs at the “pores” around 0.01 µm, i.e., these pores dominate. There seems to be a reason neither for the presence of such pores within the crystals of the RAMEB used for the MIP measurements nor for their occurrence as the pores of the sample bed (between crystal grains) because the RAMEB crystals used were around 0.1-2 mm in diameter. Most probably, RAMEB crystals shrink under high mercury pressures. The MIP curves for RAMEB-enriched KA and BEN go below these for the initial ones. This probably happens due to shrinking of aggregates or pore filling with RAMEB. However, the opposite trend of the MIP curves for illite indicates aggregate structure loosening due to RAMEB treatment. For BEN and KA, one can observe a shift in dominant mesopores toward smaller sizes, whereas for the IT a tendency of finer mesopores decrease and an increase of larger pores is clearly seen. Fractal scaling seems to be applicable to the mesopores of the studied materials, which was seen from the linearity of dV/dR vs R log-log plots. In general, the linearity ranges are found for all the samples, as this was observed also for the micropores. For the RAMEB, two linearity ranges of similar slopes can be distinguished. These slopes are lower than for the minerals, indicating that the RAMEB fractal dimension is smallest. The linearity ranges for the initial minerals occur for pore radii from 0.35 to 0.01 µm for BEN, from 0.45 to 0.02 µm for IT, and from 0.71 to 0.14 µm for KA, and in general these become narrower with RAMEB addition increase (which is particularly seen for IT) and shift toward smaller pore radii (which is particularly seen for BEN). Surface and Pore Parameters. Surface and pore parameters of the initial materials, calculated from adsorptiondesorption isotherms and MIP curves, are presented in Table 1, while changes of these parameters for minerals after various RAMEB additions are summarized in Figures 3-6. In these figures, the ratio of a given parameter for the RAMEB-treated VOL. 35, NO. 24, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4949
TABLE 1. Values of Surface and Pore Parameters of the RAMEB and Initial Samples of Minerals Studied (95% Coincidence Intervalsa pore properties surface properties
S RAMEB BEN IT KA
(m2
g-1)
428 ( 15.8 305 ( 8.5 139 ( 3.9 55 ( 1.5
Eav/RT 1.5 ( 0.05 2.6 ( 0.07 2.9 ( 0.08 4.8 ( 0.13
micropores
v
(mm3
g-1)
630 ( 23 141 ( 3.9 130 ( 3.6 69 ( 1.9
mesopores
rav (nm)
D
19.3 ( 0.71 15.8 ( 0.44 19.0 ( 0.53 27.0 ( 0.75
1.93 ( 0.011 2.35 ( 0.065 2.39 ( 0.066 2.54 ( 0.070
V
(mm3
g-1)
64 ( 1.8 267 ( 4.9 128 ( 2.4 101 ( 1.9
Rav (µm)
Ds
0.74 ( 0.020 0.37 ( 0.007 0.15 ( 0.003 0.16 ( 0.003
3.57 ( 0.099 3.76 ( 0.069 4.33 ( 0.080 4.76 ( 0.069
a Abbreviations: S, surface area; E /RT, average water vapor adsorption energy (in RT units; R is universal gas constant, T is temperature of av the measurements); v, volume of micropores; rav, average micropore radius; D, fractal dimension of micropores; V, volume of mesopores; Rav, average mesopore radius; Ds, scaling factor of mesopore surface.
FIGURE 3. Changes of surface areas (S) and average water vapor adsorption energies (Eav) of the studied minerals vs cyclodextrins content. On y-axis, the ratio of the given parameter of the mineral + RAMEB to this of the control mineral is shown. On x-axis log(CD) ) -2 denotes control samples (without RAMEB addition). Abbreviations: BEN, bentonite; IT, illite; KA, kaolin. mineral to its value for the control mineral is presented on the y-axis. Surface Area and Average Adsorption Energy. At first, one has to mention that the surface area of RAMEB could not be found by plotting adsorption data in the linear coordinates of the Aranovich isotherm because no satisfactory linear fit was obtained. Most probably, the process of water vapor absorption rather than adsorption occurs for RAMEB. Nevertheless, to have a reference value for comparison with the minerals, we calculated the “surface area” for RAMEB using the formula
S ) LωM-1max{a(l - x)1/2}
(9)
The value of RAMEB surface area of ca. 400 m2 g-1 was calculated. This value has no physical meaning. This is presented only to show to what extent the surface area of the minerals should increase with the CD load. It is worth noting that the surface areas of the studied minerals calculated using eqs 1 and 9 were very close to each other. The decrease in the surface area of the minerals after RAMEB addition (Figure 3) seems somewhat surprising taking into account the very high water sorption of the RAMEB. The decrease of the surface area is most pronounced for BEN and IT, which have high surface areas. For KA + 9% RAMEB, the surface area increases. 4950
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 24, 2001
FIGURE 4. Changes of the micropore (v) and mesopore (V) volumes of the studied minerals vs cyclodextrins content. On y-axis, the ratio of the given parameter of the mineral + RAMEB to this of the control mineral is shown. On x-axis log(RAMEB concentration) ) -2 denotes control samples (without RAMEB addition). Abbreviations for points the same as in Figure 3: triangles, bentonite; diamonds, illite; circles, kaolin. The decrease in surface area indicates that strong interactions of the RAMEB and the surface of the studied minerals occur. These may involve association of mineral plates as well as of amorphous mineral components, by RAMEB. The possibility of RAMEB entering into interlamellar spaces of 14 Å minerals (the thickness of a cyclodextrin ring is around 8 Å) cannot be excluded. Strong binding of the cyclodextrins to mineral phases has been demonstrated. ′This process is the basis for preparing specific sorbents by intercalation of natural or artificial minerals with cyclodextrins (43, 44). The increase in surface area for KA at high RAMEB concentration may indicate that, at a certain RAMEB level, the “cyclodextrin-binding capacity” of the KA becomes overloaded, and the individual properties of RAMEB start to be expressed. Strong interactions of the RAMEB and minerals apparently affect the average adsorption energy patterns (Figure 3) as well. Instead of an adsorption energy decrease, as was expected because of the very low water vapor sorption energy of the RAMEB, the adsorption energy of the minerals increases after RAMEB addition, consistent with the changes in adsorption energy distribution functions described above. The adsorption energy increase indicates that the overall water binding forces become higher after cyclodextrins treatments. This increase in interaction energy of polar water
FIGURE 5. Changes of average micropore (r) and mesopore (R) radii of the studied minerals vs cyclodextrins content. On y-axis, the ratio of the given parameter of the mineral + RAMEB to this of the control mineral is shown. On x-axis log(RAMEB concentration) ) -2 denotes control samples (without RAMEB addition). Abbreviations for points the same as in Figure 3: triangles, bentonite; diamonds, illite; circles, kaolin.
FIGURE 6. Changes of micropore fractal dimension (D) and mesopore scaling factor (Ds) of the studied minerals vs cyclodextrins content. On y-axis, the ratio of the given parameter of the mineral + RAMEB to this of the control mineral is shown. On x-axis log(RAMEB concentration) ) -2 denotes control samples (without RAMEB addition). Abbreviations for points the same as in Figure 3: triangles, bentonite; diamonds, illite; circles, kaolin. molecules with RAMEB-treated minerals can enhance desorption of nonpolar compounds and their inclusion into the cyclodextrin. The decrease of the average adsorption energy for KA is noted at the extreme RAMEB content, which can be related to the overload of RAMEB. Pore Volumes and Radii. Changes in micro- and mesopore volumes of the studied minerals vs RAMEB addition are illustrated in Figure 4. Although the RAMEB micropore volume is much higher than of the minerals (over 600 mm3 g-1), for all minerals this tends to decrease slightly with RAMEB load increase. The exception is, as usual, KA with 9% RAMEB, for which the water sorption by the RAMEB excess
can occur. The mesopore volumes of KA and BEN decrease after RAMEB addition, indicating that the aggregate structure of these minerals becomes generally more compacted. However, for the IT an increase in mesopore volumes is observed, indicating aggregate structure loosening, which may be connected with a rearrangement of the mineral particles within aggregate network. Changes in average micro- and mesopore radii of the studied minerals due to RAMEB addition are illustrated in Figure 5. The micropore radii for BEN and IT increase at low RAMEB doses. The radii first decrease and then start to rise sharply at higher RAMEB doses. For KA, the micropore radii increase with RAMEB dose increase. The increase of micropore radii and the decrease of volume may be connected with filling of the smaller pores by RAMEB as well as by gluing the smaller pore walls together. In this case, larger pores remain intact. The average mesopore radius for KA rises at the lowest RAMEB dose and then tends to decrease with RAMEB concentration increase. A similar trend is observed for the IT; however, the average mesopore radius starts to increase again at 3% RAMEB load. For BEN, the average mesopore radius initially decreases and next increases at 1% RAMEB dose. The initial increase in the mesopore radius of KA and IT may be due to the filling of the smallest mesopores with RAMEB. A further decrease in the mesopore radius may occur due to covering of mesopore walls by a layer of RAMEB. Because IT have smaller-size particles than KA, its aggregates have smaller mesopores, faster becoming narrower due to the pore walls coverage. The finest from these pores may be again filled at larger RAMEB concentrations, leading to the increase in average mesopore radius. For BEN, having the smallest particles, the mesopore walls covering may dominate just at the beginning of RAMEB addition, and further pore filling starts earlier than for the IT. Fractal Scaling of Pores. A rearrangement of the micropore structure (changes in surface roughness) of the minerals due to RAMEB addition is possible. This can be reflected in a slight increase in micropore fractal dimension at high RAMEB contents, as shown in Figure 6. This increase indicates a rise in micropore complexity. Note that the micropore fractal dimension of the RAMEB is much lower than for the minerals studied (Table 1). Aggregates of the control minerals exhibit more complex mesopore structure than the RAMEB (see Table 1). However, in all cases the fractal dimension values are greater than 3, i.e., these are out of the theoretical Ds ∈ range. The mesopores in the mineral aggregates can have ink bottle shapes or be formed from larger voids connected by narrower necks. Application of the cylindrical pore model for such pores can lead to Ds values higher than 3. The pore volume V(R), representing now the volume of the void + the volume of the neck, can rise faster with the pore radius R, i.e., the radius of the neck, and thus dV(R)/dR in eq 8 can be higher than for the cylindrical pore body of Ds ) 3(2-dimensional simulations of such phenomena are presented in ref 45). Therefore, the calculated Ds value should be referred to as a scaling factor rather than the fractal dimension. The scaling factor Ds of BEN and IT decreases at low RAMEB additions and rises at higher RAMEB doses (Figure 6). For KA, the Ds decreases again at higher RAMEB doses. For all of the minerals, the decrease of the scaling factor may be related to filling of the smallest pores and pore walls coverage by RAMEB, both leading to the smoothing of pore surfaces. As a possible explanation of the Ds increase, one may consider the accumulation of cyclodextrins at the necks of the mesopores. This leads to a decrease of the radii of the necks while the volumes of the voids remain large. This is apparently consistent with the observed decrease of the scaling range and its shift toward smaller mesopore radii. At VOL. 35, NO. 24, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4951
higher RAMEB doses, the KA mesopores can become closed or filled with CD.
General Comments The following behaviors of the minerals were observed after RAMEB addition: decrease in surface areas, increase in adsorption energy, decrease in volumes of micro and mesopores, increase in micro and mesopore radii, and increase of pore complexity. The opposite behaviors would generally be expected if cyclodextrins did not react with the solid phase of the minerals. Only the behavior of KA with the highest RAMEB dose (9 % w/w) suggested a presence of the nonreacted cyclodextrin excess. The intensity of the changes in surface and pore properties of the studied minerals generally increase with cyclodextrins dose increase. Our findings suggest that the surface properties and pore structure of minerals change dramatically upon RAMEB addition. These observations may be important for understanding the effects of RAMEB on mineral soils and can give an additional explanation to the bioremediation-accelerating effect of RAMEB. This additive has been so far considered to be primarily a carrier (solubilizer) of the poorly soluble organic contaminants. According to our findings, this cyclic oligosaccharide plays an additional role by affecting the surface and pore properties of the solid phase. In the experiments performed, samples amended with RAMEB were used. It was not determined if certain surface and pore properties return to the original state after the removal of RAMEB, either by leaching or by biodegradation. It is also interesting if the phenomena observed for RAMEB are generally true for other cyclodextrins or carbohydrates. The primary interest is to study the behavior of soils under cyclodextrin addition. We completed such experiments using RAMEB as the additive.
Acknowledgments Part of this work was funded by the Polish Research Committee KBN under Project 6P04G05613, by the NATO Science for Peace Program SFP-973720, and by the Ministry of Education of Hungary (OM-NATO 00009/20).
Supporting Information Available Examples of water vapor adsorption energy distribution functions, micropore size distribution functions, ln-ln plots of adsorption vs adsorption potential (fractal scaling of micropores), and log-log plots of the first derivative of mesopore volume on radius vs radius (fractal scaling of mesopore surface) for RAMEB and RAMEB-treated minerals. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Szejtli, J. Chem. Rev. 1998, 98, 1743-1753. (2) Bar, R. Supramol. Chem. 1996, 3, 603-615. (3) Cyclodextrins and Their Industrial Uses; Duchene, D., Ed.; Editions de Sante: Paris, 1987; pp 1-665. (4) Dailey, O. D., Jr.; Dowler, C. C.; Glaze, N. C. Pesticide Formulations in Applied Systems; ASTM Special Technical Publications 10; ASTM: West Conshohocken, PA, 1990; pp 26-37. (5) Szejtli, J. Minutes, 6th International Symposium on Cyclodextrins; Hedges, A. R., Ed.; Sante: Paris, 1992; pp 380-389. (6) Szente, L.; Fenyvesi, E.; Szejtli, J. Environ. Sci. Technol. 1999, 33, 4495-4498. (7) Wang, X.; Brusseau, M. L. Environ. Sci. Technol. 1993, 27, 28212830.
4952
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 24, 2001
(8) Wang, M. J.; Marlowe, E.; Miller-Maier, R.; Brusseau, M. L. Environ. Sci. Technol. 1998, 32, 1907-1912. (9) Szejtli, J.; Fenyvesi, E. Extraction of Organic Pollutants from Contaminated Soils Using Aqueous Cyclodextrin Solutions. European Patent EP 613735, 1994. (10) McCray, J. E.; Boving, T. B.; Brusseau, M. L. Ground Water Monit. Rem. 2000, 20, 94-103. (11) Gruiz, K.; Fenyvesi, E.; Kriston, E.; Molnar E. J. Inclusion Phenom. Mol. Recogn. Chem. 1996, 25, 233-236. (12) Bardi, L.; Mattei, A.; Steffan, S.; Marzona, M. Enzyme Microb. Technol. 2000, 27, 709-713. (13) Fava, F.; Di Gioia, D.; Marchetti, L. Biotechnol. Bioeng. 1998, 58, 345-355. (14) McCray, J. E.; Brusseau M. L. Environ. Sci. Technol. 1998, 32, 1285-1293. (15) Gruiz, K.; Kriston, E. J. Soil Contam. 1995, 4, 163-173. (16) Sheremata, T. W.; Hawari, J. Environ. Sci. Technol. 2000, 34, 3462-3468. (17) Roquerol, R.; Avnir, D.; Fairbridge, C. W.; Everett, D. H.; Haynes, J. H.; Pernicone, N.; Ramsay, J. D. F.; Sing, K. S. W.; Unger, K. K. Pure Appl. Chem. 1994, 66, 1739-1758. (18) Sokolowska, Z. Probl. Agrophys. 1989, 58, 1-145 (in Polish). (19) Oscik, J. Adsorption; PWN: Warsaw, 1979 (in Polish). (20) Avnir, D.; Farin, D.; Pfeifer, P. J. Colloid Interface Sci. 1985, 103, 112-123. (21) Neimark, A. V. Russ. J. Phys. Chem. 1990, 64, 2593-2605. (22) Pfeifer, P.; Obert, M. Fractals. Basic Concepts and Terminology. In The Fractal Approach to Heterogeneous Chemistry; Avnir, D., Ed.; John Wiley and Sons: New York, 1989. (23) Pachepsky, Ya. A.; Polubesova, T. A.; Hajnos, M.; Sokolowska, Z.; Jozefaciuk, G. Soil Sci. Soc. Am. J. 1995, 59, 68-75. (24) Jozefaciuk, G.; Sokolowska, Z.; Sokolowski, S.; Alekseev, A.; Alekseeva, T. Clay Miner. 1993, 28, 145-148. (25) Jozefaciuk, G.; Sokolowska, Z.; Hajnos, M.; Hoffmann, C.; Renger, M. Geoderma 1996, 74, 125-137. (26) Jozefaciuk, G.; Muranyi, A.; Szatanik-Kloc, A.; Csillag J.; Wlodarczyk, T. Pol. J. Soil Sci. 1999, XXXII, 23-32. (27) Lipiec, J.; Hatano, R.; Slowinska-Jurkiewicz, A. Soil Tillage Res. 1998, 47, 61-66. (28) Pachepsky, Ya. A.; Polubesova, T. A.; Hajnos, M.; Jozefaciuk, G.; Sokolowska, Z. Soil Sci. Soc. Am. J. 1995, 59, 410-417. (29) Perfect, E.; Blevins, R. L. Soil Sci. Soc. Am. J. 1997, 61, 896-900. (30) Wilczynski, W.; Renger, M.; Jozefaciuk, G.; Hajnos, M.; Sokolowska, Z. Z. Pfl. Bodenkunde 1993, 156, 235-238. (31) Aranovich, G. L. Langmuir 1992, 8, 736-739. (32) Jaroniec, M.; Rudzinski, W.; Sokolowski, S.; Smarzewski, R. J. Colloid Polym, Sci. 1975, 253, 164-166. (33) Jaroniec, M.; Brauer, P. Surf. Sci. Rep. 1986, 6, 65-117. (34) Harris, L. B. Surf. Sci. 1968, 10, 129-145. (35) Jozefaciuk, G.; Shin, J. S. Korean J. Soil Sci. Fertilizer 1996, 29, 86-91. (36) Jozefaciuk, G.; Shin, J. S. Korean J. Soil Sci. Fertilizer 1996, 29, 218-225. (37) Jaroniec, M.; Kruk, M.; Olivier, J. Langmuir 1997, 13, 10311035. (38) Sokolowska, Z.; Sokolowski, S. Geoderma 1999, 88, 233-249. (39) Sridharan, A.; Venkatappa Rao, G. Soil Sci. Soc. Am. Proc. 1972, 36, 980-981. (40) Yokoya, N.; Yamamoto, K.; Funakuro N. Comput. Vision, Graphics Image Process. 1989, 46, 284-302. (41) Pachepsky, Ya. A.; Ritchie, J. C.; Gimenez, D. Remote Sensing Environ. 1997, 61, 150-161. (42) Lindner, K.; Saenger, W. Carbohydr. Res. 1982, 99, 103-115. (43) Kijima, T.; Tanaka, J.; Matsui Y. Nature 1984, 310, 533-534. (44) Zhao, H.; Vance, G. F. Clays Clay Miner. 1998, 6, 712-718. (45) Kozak, E. Methodical aspects of Pore Size Distribution and Fractal Dimension Estimation of Soil Materials. Ph.D. Thesis, Institute of Agrophysics, Lublin, Poland, 1994; pp 1-130 (in Polish).
Received for review March 21, 2001. Revised manuscript received September 13, 2001. Accepted September 25, 2001. ES010083Z
