The Lower Cation Exchange Capacity Limit of Montmorillonite

Journal of Colloid and Interface Science 217, 77– 85 (1999) Article ID jcis.1999.6254, available online at http://www.idealibrary.com on

The Lower Cation Exchange Capacity Limit of Montmorillonite Lorenz Paul Meier 1 and Rolf Nu¨esch Department of Materials, Institute of Nonmetallic Materials, ETH, CH-8092 Zurich, Switzerland Received November 25, 1998; in revised form March 22, 1999; accepted March 30, 1999

known to have a charge density of 0.25 to 0.60 charges per half-unit cell. Considering a mean molar mass of 370 g per mol per half-unit cell, the charge per unit cell corresponds to a CEC of 67 to 162 mmol per 100 g smectite. Hectorite is an example of a naturally occurring low charged clay mineral, a Li-containing trioctahedral smectite, with a charge density of 0.28 charges per half-unit cell, and a corresponding CEC of 75 mmol per 100 g clay. Earlier studies proposed a critical interlayer charge limit, below which some collapsed interlayers can be observed (4 –7). According to the Hofmann–Klemen effect low-charged smectites are produced experimentally by neutralization of the negatively charged layer (8). In this procedure, lithium is incorporated into the empty octahedral sites of dioctahedral smectite. Smectites having reduced charges could plausibly develop regular interstratified structures (6). This can be observed by the development of superstructure signals in the X-ray diffraction pattern (9). The signals are indicative of the existence of collapsed clay interlayers with the superstructure sequence of expanded– collapsed– expanded interlayers. Regularly interstratified structures are described by the Reichweite (10, 11). The Reichweite expresses the crystallographic range of the regular arrangement of neighboring layers. The phenomenon has been interpreted in terms of the heterogeneous charge reduction of the smectites due to the effects of cation segregation (6). This has been suggested also as a result of interactions between the interlayer cations and the hydroxyl groups of the octahedral sheet (9). In nature, pyrophyllite and talc have interlayers that are both cation-free and nonexpanded. The reported distances are 9.1– 9.2 Å for pyrophyllite and 9.3–9.4 Å for talc, respectively. In nature, both of these minerals are known to form under hydrothermal or metamorphic conditions. Also, the rare mineral kerolite is known to form at low temperature (12). It has a fine-grained structure and is a nonexpanded, talc-like mineral, with a basal distance of 9.65 Å. A series of naturally occurring minerals having kerolite–Stevensite, from the Armagosa Desert, Nevada, were characterized as being mixed-layered materials (13). Simulated XRD patterns suggest the presence of a range of expanded interlayers from almost 0% to about 80%. The collapsed interlayers are considered as Kerolite with essentially no charges present. The expandable interlayers are interpreted as Stevensite with interlayer cations. The collapse

Wyoming montmorillonite (Volclay) with different charges were produced by Li-incorporation and the interlayer cations were replaced by tetramethyl ammonium. Their XRD pattern showed a regular sequence of expanded and collapsed interlayers. The regularly interstratified structure corresponds to a regularity of Reichweite R 5 1. The expanded part of the interlayers was calculated by comparing XRD pattern with simulations using NEWMOD software. The calculations of the cation exchange capacity CEC for the expanded interlayer part gives a constant value of 65 6 2 mmol/100 g fully swellable montmorillonite. This value is denoted as the lower CEC limit of montmorillonite. We propose a model which considers montmorillonite to be a stacked two-dimensional polyelectrolyte. The model propose that interlayers of the stack collapse spontaneously by cation shifting into the neighboring interlayers, if the charge density of a montmorillonite has a value below the lower CEC value. The shifted cations of the collapsed interlayer increase the charge density in the neighboring interlayers and prevent their collapse. A regularly interstratified structure arises with the sequence expanded/collapsed/expanded interlayer, which can be observed by XRD. The behavior of low charged montmorillonite is explained with the properties of a two-dimensional polyelectrolyte. Below the critical layer charge, the Van der Waals forces dominate over electrostatic repulsive forces and the interlayers collapse. © 1999 Academic Press Key Words: cation exchange capacity; charge reduced montmorillonite; critical charge density; regular mixed layer smectite/ kerolite; Reichweite; two-dimensional polyelectrolyte; fundamental particles.

INTRODUCTION

The swellability of clay minerals is primarily determined by their cation exchange capacity (CEC) and their charge density (1, 2). Clay minerals with or without a very high permanent layer charge are not swellable in water (e.g. illite and pyrophyllite). One of the main goals in this area of research is to find the lowest charged smectite which is still fully swellable. Another goal is the achievement of better understanding the behavior of smectite at this lowest CEC limit. The layer charge determines the cation exchange properties, such as selectivities for different cations (3). Smectites, are 1

To whom correspondence should be addressed. 77

0021-9797/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.

¨ ESCH MEIER AND NU

78

of the interlayer distance of Stevensite thus observed is owing to weaker interlayer bonding, in the presence of traces of water and lattice distortion. However, the behavior of low-charged smectite is not well known, yet. In the present investigation, attention will be focused on the behavior of low-charged smectites with less than 0.25 charges per half-unit cell. MATERIALS AND METHODS

Montmorillonite Volclay (American Colloid Company) having large surface area (800 m 2/g) and platelet shape was selected for the investigation of the behavior of low-charged clays. Effects of the edge charges and the disorder of the XRD specimens are minimal for this type of montmorillonite. The cationic exchange capacity was 93 mmol/100 g clay (12). Sodium- or lithium-saturated montmorillonites were prepared by ion exchange of the montmorillonite suspensions (20 g smectite per 1 liter aqueous solution) with a 1 M solution of sodium chloride or lithium chloride. The exchange procedure was repeated three times overnight. The excess salts were washed out with 95% ethanol until the conductivity in the supernatant solution was less than 5 mS/cm. After drying at 105°C, the Li- and Na-montmorillonites were mixed in various proportions to produce mixtures having Li-montmorillonite with weight fractions of lithium (F) of 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, and 1.0. One gram of each fraction has been dispersed in 200 ml of double distilled water, employing a probe ultrasonic treatment. They were stirred overnight and then dried out at 60°C. The dry samples were heated for 24 h at 220°C in order to incorporate Li 1 into the octahedral layers. The interlayer cations were exchanged against an ethanol solution of 0.5 M tetramethyl ammonium chloride solution in ethanol (6). Tetramethyl ammonium (TMA) is expected to produce a fixed expanded interlayer distance that is not sensitive to humidity. The properties of the TMA-exchanged montmorillonites are not that much different from those of hydrated inorganic cations. After heating the samples were ground and the fine powder was added to 25 ml of a 0.5 M solution of tetramethyl ammonium chloride in 95% ethanol and were stirred for 24 h. Ultrasonic treatment at this point would not be appropriate, because it reduces the thickness of the stacked particles and the low-charged montmorillonite particles flocculate. This results in a decrease in the orientation and quality of the XRD samples. The suspensions were centrifuged, decanted, and a fresh solution of tetramethyl ammonium chloride solution was added. After repeating the cation exchange procedure twice, the suspended smectites were washed with 95% ethanol until the conductivity of the supernatant solution was less than 5 mS/cm. The XRD samples were prepared by sedimentation in a centrifuge and dried slowly at room temperature at 50% r.h. (15).

XRD patterns were measured with a Phillips PW 1729 diffractometer, using an automatic theta-compensating slit, Cu K a radiation, a step size of 0.02° 2Q, and a counting time of 3 s per step. All XRD patterns were smoothed by a 15-point Savitsky–Golay filter and corrected for shifts in the measurement of 2 Q, by standardizing with respect to the 4.26 and 3.34 Å peaks, resulting from a small amount of quartz in the clay fractions. Simulations by NEWMOD. The following input values were used in a NEWMOD (16) simulation to determine the percentage of expanded interlayers in interstratified materials: 1. An interstratified structure was simulated for different mixtures of dioctahedral smectite (expanded fraction) and dioctahedral mica (collapsed fraction) in different proportions. 2. Potassium in the collapsed fraction was set to zero. It was assumed that there are no cations and no potassium in the interlayers. 3. A 9.7 Å basal distance was chosen for the nonexpanded interlayers (simulated mica fraction). 4. Tetramethyl ammonium, as an interlayer cation, produces a 13.8 Å basal spacing relevant to the expanded interlayers (simulated smectite fraction). 5. The position of the 002*-peak serves to provide a measure of the proportion of expanded to nonexpanded layers (peak nomenclature after (16)). 6. Iron content was set to zero. The cation exchange capacity (CEC of the charge reduced Volclay) was determined with ammonium acetate (14), which represents the exchangeable cations per 100 g smectite containing both expanded and collapsed interlayers. The CEC of the expanded interlayers (CEC expanded) represents the amount of exchangeable cations per 100 g expanded smectite and was calculated from the measured CEC and the percentage of expanded interlayers, according to CECexpanded 5

CECmeasured . part expanded Interlayers

[1]

RESULTS

1. Influence of CEC on the 002* XRD Signal As the CEC is reduced more and more, it is clearly apparent that the d-value of the 002* mixed layer peak, corresponding to a regularity of the Reichweite R 5 1, shifts to lower values and exhibits decreased intensity (Fig. 1). Hence, the lower the CEC of the sample, the lower the amount of swellable interlayers, as well as their peak intensity. Decreasing peak intensity is another way to determine the collapsed interlayer (6), which can be calculated after theoretical derivation (17). In this work the peak position and the simulation using NEWMOD allow more precise determination of the collapsed interlayer part (18).

LOWER CATION EXCHANGE LIMIT OF MONTMORILLONITE

79

FIG. 1. XRD measurements show the influence of CEC on d-value and the intensity of the 002* peak (F 5 Li-montmorillonite proportion of the initial sample).

According to the experimentally determined CEC value, the d-value position of the 002* mixed layer peak (Reichweite R 5 1) shifts, and the intensity decreases as well. 2. Regularity Given by the Reichweite of the Interstratified Montmorillonite The XRD patterns and their NEWMOD simulations indicate that the interstratified structures are ordered sequences of col-

lapsed and expanded 2:1 interlayers with a regularity of the Reichweite 5 1 (15). In Fig. 2 the results of a measurement are compared to those obtained from simulations. The measured XRD pattern for the sample with a CEC of 42 mmol per 100 g shows a peak in the vicinity of a 2 Q value of 4°, which closely matches the position of the simulated 001* superstructure peak. Also, it is noteworthy that the randomly interstratified simulation (R 5 0) lacks the presence of any superstructure

FIG. 2. The measured XRD pattern of a charge-reduced smectite using the fraction with 0.8 Li and expanded smectite with a CEC of 42 mmol per 100 g montmorillonite shows a regularity of Reichweite R 5 1.

80

¨ ESCH MEIER AND NU

4. Lower Limit of CEC

FIG. 3. XRD pattern and their simulations of the different heat-treated and exchanged montmorillonite.

signals. In addition, similar differences in X-ray intensity are also apparent near the 008* peak. 3. The XRD Pattern of Different Charge-Reduced Montmorillonite In the XRD pattern (Fig. 3) the ordered mixed-layer superstructures are clearly observable for the treated samples F 5 0.7. The regularity of the XRD pattern, for samples having F 5 0.7, 0.8, and 0.9, corresponds to the regularity of the Reichweite R 5 1. Evidence for the value R 5 1 is not readily apparent at F 5 0.6. The treated sample with initial Lifraction of F 5 0.5 has no peak shift and, thus, has no collapsed interlayers. The sequence expanded– collapsed– expanded interlayers has the effect of limiting the part of expanded interlayers to be higher than 50%, as will be discussed later. However, there is still a 001* superstructure signal for F 5 1.0, as is demonstrated by a low angle peak in the XRD pattern. Such behavior could not be readily interpreted. In Table 1, the results for the NEWMOD simulations with maximum d-value XRD measurement for different Li-fractions. The percentage of smectite interlayers are determined by simulations of interstratification having different expanded to collapsed interlayer compositions, with a regularity corresponding to Reichweite R 5 1. The number (N) of smectite interlayers, in the MacEwan crystallite (“coherent scattering domain size”), is related to the peak width at the half maximum intensity of the 002*-peak (6).

The CEC of the reduced charge montmorillonite (RCM) is found to depend on the fraction of lithium in the sample and shows values from 92 mmol per 100 g (untreated sample) down to 34 mmol per 100 g for the 100% Li-treated montmorillonite (Table 2). Values of CEC of the different samples, calculated according to Eq. [1] for the expanded interlayers, remain essentially constant. This behavior presumably suggests that 65 mmol per 100 g is the lower CEC limit. Thus, below the lower CEC limit the swellable part of a charge reduced montmorillonite is associated with the CEC. In Fig. 4, the measured CEC values are plotted against the calculated CEC values in expanded interlayers. The CEC of the charge-reduced Volclay samples are expected to follow the line for fully swellable clays from untreated Volclay (filled square in Fig. 4) down to zero. However, the existence of collapsed interlayers has to be considered by the calculation of the CEC in the expanded interlayers. The CEC values in the expanded interlayers remains essentially unchanged, at a CEC value of 65 6 2 mmol per 100 g of clay. Above the critical CEC value, the charge-reduced Volclay is fully expanded, whereas below such value, the interlayers are partially collapsed. Nonetheless, the measured CEC of the partially collapsed montmorillonites varies from 37 to 66 mmol per 100 g. The calculated CEC values in the expanded interlayers of Volclay having expandable interlayers ranging from 59% to 86% remain essentially unchanged. The calculated CEC values in the expanded interlayers (according to Eq. [1]) of previously obtained (6) are less accurate, but they are around 65 mmol per 100 g clay (Fig. 5). The montmorillonite, used for charge reduction were obtained from Wyoming Upton, with tetramethyl ammonium and tetrapropyl ammonium as the interlayer cations. The expanded interlayer parts were calculated from the intensity of the 002* signal (17). Calculations based on Eq. [1] confirm that the lower CEC limit is about 65 mmol per 100 g montmorillonite. Similar results were obtained

TABLE 1 The Best Fit Results for the d-Values Using NEWMOD Simulations to Determine the Percentage of Expanded Interlayers for the Different Samples of F 5 0.5 to 1.0 Li-fraction F

d-value [Å]

Expanded interlayers %

Reichweite R

Layers N

0.4 0.5 0.6 0.7 0.8 0.8 0.9 1

13.81 13.85 13.64 13 12.63 12.63 12.49 11.89

100 100 86 72 62 50 59 43

1 1 1 1 1 0a 1 1

3–9 3–9 3–9 3–9 2–7 3–9 2–6 2–6

a

The other basal signals of this pattern do not match.

81

LOWER CATION EXCHANGE LIMIT OF MONTMORILLONITE

TABLE 2 Properties of the Different Samples

Li-fraction F

d-value [Å]

CEC measured [mmol per 100 g clay]

Percentage of expanded interlayers

CEC expanded layers [mmol per 100 g clay]

Volclay (untreated) 0.4 0.5 0.6 0.7 0.8 0.9 1.0 a

13.85 13.81 13.85 13.64 13.00 12.63 12.49 11.79

92 69 66 56 48 42 37 34

100 100 100 86 72 62 59 [43] a

92 69 66 65 67 65 63 [79] a

Note. The observed d-values, measured CEC, % expanded interlayers, and the calculated CEC in the expanded interlayers in TMA-treated RCM. a Expanded interlayers ,50%; see discussions.

for tetrapropyl ammonium, as another interlayer cation. The XRD pattern for tetrapropyl ammonium as the interlayer cation, exhibits less intense 001* superstructure peak. Also, the same reduction of this superstructure signal has been observed using tetrabutyl ammonium. Therefore, tetramethyl ammonium was used as the interlayer cation for the measurements and LCECL calculations. DISCUSSION

Montmorillonite is regarded to be a two-dimensional polyelectrolyte. The observed properties of the low-charged partially collapsed montmorillonite samples are in satisfactory agreement with this assumption. As would be expected, swelling properties of a two-dimensional polyelectrolyte (a sheet containing charges) are different from those of the one-dimensional case (a string containing charges). The water uptake and the swelling behavior of a one-dimensional polyelectrolyte depends largely on the charge. If the amount of charge is lower, both its swellability and expandability will be substantially lowered. The water uptake of a two-dimensional polyelectrolyte, at higher charge density, is quite independent of the charge density. It is, thus, employed as a method for characterization of the swellable surface of clay minerals (19). There is, however, a lower limit of charge density that is related to the cation exchange capacity, CEC. So, below a certain CEC value the swelling properties of a two-dimensional polyelectrolyte change and collapsed interlayers develop. This lower limit is explained in terms of the “lower cation exchange capacity limit model” (LCECL). The LCECL model can provide a reasonable explanation for the constant CEC values in the expanded interlayers, as well as the influence of the low charge density on the regularity observed in the XRD pattern. Description of the LCECL. A lower CEC limit exists, where the layer–layer attractive forces predominate over those of the ionic repulsive forces. Without losing their hydration shells the cations redistribute themselves from one interlayer into the neighboring ones and, so, cause the original interlayer

to collapse. The amount of cations in the neighboring interlayers increases as a result of such additionally migrating cations. Thus, the collapse of interlayers is prevented. As such, a well-defined interlayer sequence arises, with an expanded– collapsed– expanded interlayer superstructure, corresponding to a regularity of Reichweite R 5 1. This has been confirmed by XRD measurements. If the charge density, in the remaining expanded interlayers, is still too low, another interlayer within the stack will collapse in the same manner, resulting in the same structure sequence. The charge density in the expanded interlayers remains constant, as suggested by pertinent calculations of the CEC in the swellable interlayers. Reasons for the existence of a LCECL in two-dimensional polyelectrolytes. 1. Instead of swelling, an interlayer having no cations will collapse instead of swelling (e.g., pyrophyllite). Such collapse of an interlayer having no interlayer cations is expected to be favorable. Layer-to-layer attractive forces are

FIG. 4. Measured CEC versus the CEC in the expanded interlayers. The calculated CEC in the expanded interlayers remains constant at 65 6 2 mmol per 100 g Volclay.

¨ ESCH MEIER AND NU

82

FIG. 7. Two interlayer cation distributions are possible with the same negative octahedral layer charge. Despite anion–anion repulsion, collapsing will occur below the LCECL, as has been already discussed.

FIG. 5. Calculated CEC in the expanded interlayers with tetramethyl ammonium h and with tetrapropyl ammonium E as interlayer cation (6).

suggested to be responsible for collapsed interlayers. Van der Waals type forces operating between two flat surfaces are a function of 1/r 3 (20). 2. Within a coherent stack of layers, the following forces are effective: (a) attractive forces between negatively charged layers and the interlayer cations next to it; (b) attractive Van der Waals forces; (c) repulsive forces between the cations within the interlayer; (d) repulsive forces between the adjacent layers (Fig. 6). 3. A negatively charged octahedral layer is not centered on a certain side (Coulomb forces). Therefore, the hydrated cations can shift into the neighboring interlayers without a change in the attractive Coulomb forces within the same distance (Fig. 7). Below the LCECL, the resulting partially collapsed interlayer configuration is equal, or more stable, when compared to a fully expanded interlayer state. This final stage, as expected, results from a two-dimensional polyelectrolyte having characteristic shape properties and the balance of the Van der Waals and the Coulomb forces. 4. Finally, it is noteworthy here that a shift of cations into the neighboring interlayers leads to an increase in the cation– cation repulsion, which largely depends on the CEC. With a decrease in CEC, the repulsive forces become less relevant. The point at which the repulsive and the attractive forces are equal, corresponds to the lower limit of the CEC (critical

FIG. 6. The different forces within a stack of layers.

CEC). Below this value, the interlayers start to collapse and a regular mixed layered structure is generated, with the sequence expanded– collapsed– expanded interlayers (Fig. 8). Formation of a collapsed interlayer. In Fig. 9, the formation of a collapsed interlayer is displayed, a layered stack contains six layers (N 5 6) with one collapsing interlayer. In this example, there are 12 negative charges per layer, homogeneously dispersed. We arbitrarily preset 15 cations per interlayer as the minimum charge for a fully expanded montmorillonite. Of course, the negative charges per layer (122), in this example, is below the lower CEC limit. As such, the following behavior of the smectite is expected, according to the LCECL model: (1) First, a collapsed interlayer without any cations will be formed. The cations migrate from the collapsing interlayer. (2) Second, the cations move into interlayers, within the stack. (3) In addition, the cations are still next to the negatively charged layer. (4) Finally, a regular superstructure is observable by XRD. According to the lower CEC limit model, change in the charge in one interlayer influences the charge distribution in the whole MacEwan crystallite. Hence, we propose that the negative layer charge is neutralized over several layers as is illustrated in Fig. 9. The constant CEC value in the expanded layers plausibly suggests that the cations could be moved. It is expected that the two adjacent collapsed interlayers are energetically unfavorable. Moreover, a montmorillonite with a homogeneous negatively charged layer distribution could form heterogeneous cationic charge distribution within the stack. The numerical limit of the model. For the simulation of an XRD pattern with a regularity of Reichweite R 5 1 and less than 50% expanded interlayers, the stacks must have sequences where one collapsed interlayer is in contact with another one. The lower CEC limit model is based on the assumption that cations shift into the neighboring interlayer. Thus, the special

FIG. 8. Redistribution of cations due to collapsing generates a regularly interstratified structure.

LOWER CATION EXCHANGE LIMIT OF MONTMORILLONITE

83

FIG. 9. Generation of a collapsed layer and a new cationic distribution with homogeneously distributed anionic charge.

interlayer sequence expanded– collapsed– expanded has a minimal swellability. The theoretical minimum in swellability of stacks having various interlayer sequences is reported in Table 3. One clay sample contains stacks of a certain number of layers. The minimal swellability values are calculated for stacks of N 5 2–6 layers: The expanded layers in stacks of the particles having 2– 6 layers and a CEC lower than 55% of the critical value cannot be calculated according to Eq. [1]. Therefore, the behavior at values lower than half of the critical CEC is not known and so is the regularity of such montmorillonites. For the sample having F 5 1.0, a CEC of 34 mmol per 100 g has been determined, which is 52% of the critical value of 65 mmol per 100 g. Calculated values of CEC, in the expanded interlayers, were not within the limits set by the LCECL model. However, a superstructure was observed in XRD of Fig. 3. Thus, it is suggested that the observed regularity results from the ordering of some parts of the whole montmorillonite sample.

TABLE 3 The Minimum Swellability of Expanded Interlayers Realizable for Regularity R 5 1 for a Certain Number of Layers in a Stack Layers in a stack

Sequence with the maximum of coll. interlayers

Ratio of collapsed interlayers

Swellability with R 5 1 vs layers

2 3 4 5 6

s/c/s s/c/s/s s/c/s/c/s s/c/s/c/s/s s/c/s/c/s/c/s

1/2 1/3 2/4 2/5 3/6

50% 66% 50% 60% 50%

Note. Below this swellability value sequences of collapsed– collapsed interlayers arise which can not correspond to the LCECL model. Therefore the minimal swellability is for the stack size of N layers in a sample: N 5 2–6 layers: 55.3% expanded interlayers, for N 5 2–7: 55.6% and for N 5 3–9: 55.5% (mathematical average).

Further consequences of two-dimensional polyelectrolyte montmorillonite. 1. The ionic repulsive forces of a two-dimensional polyelectrolyte are a function 1/r 2 (r is the interlayer spacing). On the other hand, the Van der Waals forces between two flat surfaces are a function of 1/r 3 (20). Because of the order difference in the power of exponents between the attractive and repulsive forces, the distance of the collapsed layers (d-value) changes slightly for the layer having different charges. Therefore, the empty, but still slightly negatively charged layer of the sample with F 5 0.8 gives the best fitted result for a collapsed layer having a distance of 9.7 Å. The d-value for a completely uncharged layer is 9.3 Å (e.g., pyrophyllite and talc). This can be assumed to be a pertinent effect of the different charges of montmorillonite layers, in comparison with those d-values for the naturally occurring kerolite interlayers having d 5 9.65 Å (13). 2. The formation of either expanded or collapsed interlayers is supported by the elasticity of the layers which permit the squeezing out of the single remaining cations. A few adsorbed molecules, such as water, may not be able to inhibit such a collapsing. The forces of interaction between a molecule and a surface is a function of 1/r 4 , which means that Van der Waals forces still predominate over those of adsorbed water molecules. Siloxanes have hydrophobic surfaces that can result in the hydrated cations being repelled. However, the variation in the separation of the collapsed interlayers caused by the negatively charged layer repulsion was found to have little influence in the calculations of the swellable interlayer part in the NEWMOD simulations. Computer simulation. The collapse of an empty interlayer of low charged montmorillonite layers has been simulated using Cerius 2 3.5. The main goal of the simulation was to calculate the stability of two negatively charge layers having no interlayer cations in between. A periodic box containing four layers was generated. The layer charge of 0.083 per formula unit has been preset and one interlayer was without

84

¨ ESCH MEIER AND NU

FIG. 10. The simulated periodic box represents the collapsing of the empty interlayer. The cationic charge is shifted into the neighbored interlayers.

interlayer cations. The charge was neutralized from the two adjacent interlayers by the additional charge. The starting condition was a state with a collapsed interlayer distance slightly bigger (0.35 Å), then the final interlayer distance. The energy minimization showed an energy minimum for the state described in Fig. 10. The attractive forces within the collapsed interlayer dominate over the repulsive forces despite the repulsive Coulomb forces. The equilibrated system shows the collapsed state. Discussion of other mechanisms causing collapsed interlayers. 1. The presented model presupposes that the charge reduction procedure works uniformly for the layers. The formation of a regular mixed layer was explained in terms of the heterogeneous charge reduction of the montmorillonite due to the segregation of the lithium and sodium ions before the heating procedure (6). We cannot concur with these suggestions for the following reasons: The selectivity difference between Li and Na ions in montmorillonite is too small for ion segregation (21, 22). It is not expected to have interlayers containing just a single species, i.e., having either Li or Na ions. Furthermore, the model is able to explain the occurrence of randomly interstratified structures, but not those with the regularity described by a Reichweite R 5 0. In addition the sample with F 5 0.9 has 41% expanded interlayers. If charge segregation would occur, 90% of the interlayers would contain just lithium, before the charge is reduced. The reduction in charge is expected to work on both adjacent layers. Therefore, the negative charge, almost present in every layer of the whole stack, will be lowered in a homo-

geneous fashion. The calculated CEC is in satisfactory agreement with the effects expected by the LCECL model. Consequently, we conclude that the regularity cannot be interpreted in terms of a heterogeneous cations distribution. 2. Regular superstructure having a strictly alternating sequence of expanded– collapsed interlayers was suggested to be caused due to edge deformation of the layers (23). The swelling of one interlayer results in a deformation of the layer edges and obstructs the accessibility to the adjacent interlayers (Fig. 11). Nonetheless, this mechanism can be excluded for montmorillonite, because it would favor the collapse of exactly every other interlayer. Such a rectorite type alternating superstructure, should be visible. This is in clear contradiction to the XRD patterns of the different samples. Furthermore, strictly alternating structures are known for well crystallized materials like mica and vermiculite. This mechanism requires a stack of layers of exactly the same size and shape, which is not the case for montmorillonite. Hence, an increase of collapsed interlayers cannot be explained in terms of the suggested edge deformation. 3. Slightly different positioned hydroxyls in the octahedral layer structure may cause regularly collapsed interlayers (9). We suggest that this effect is small in comparison to ionic forces. The LCECL model is plausible by the following reasons: Swellability of smectites is connected to interlayer cations and layer charge density. The smectite lattice structure and the negative layer charge mean that the Coulomb attraction force is not focused on one interlayer. Therefore, interlayer cations have two energetically favorable positions within the neighbored interlayers next to the negative charge. Partial collapse of interlayers is always possible if the layer charge density is too low. Computer simulations of the energetic stability of the stack of smectitic layers support the existence of asymmetric charge distribution and the following regular collapse of interlayers.

FIG. 11. The edges of strictly alternating structures of expanded– collapsed interlayers are suggested to be deformed (23).

LOWER CATION EXCHANGE LIMIT OF MONTMORILLONITE

CONCLUSIONS

85

REFERENCES

The lower CEC limit model provides a satisfactory explanation of the regularity effects accompanying the collapse of interlayers, as well as the formation of regular interstratification, for charge-reduced montmorillonite from Wyoming (Volclay). The behavior of a low-charged montmorillonite can be explained in terms of the properties of a stacked two-dimensional polyelectrolyte. The results, thus obtained, are indicative of a lower limit of CEC and a critical value of the charge density of a completely expanded montmorillonite in water of around 65 mmol per 100 g clay and 0.24 charges per half unit cell. A montmorillonite providing a charge below this CEC value develops a new interlayer cation distribution by shifting the cations into the adjacent interlayers and a sequence of expanded– collapsed–expanded interlayers with the regularity (Reichweite) R 5 1 is formed. Thus, a montmorillonite with homogeneously distributed negatively charged layers may spontaneously form heterogeneously distributed cationic countercharges. Regular interstratification with Reichweite R 5 1 occurs within the range of the critical CEC value to half of that value. Computational simulations of low charged clay layers without interlayer cations describe the collapsing. The charge distribution mechanisms below half the critical CEC value are not determined yet.

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

ACKNOWLEDGMENTS

20.

We thank the “Komission fu¨r Technologie und Innovation” for their financial support, I. T. Madsen, U. W. Fute, and G. Lagaly for fruitful discussions and K. Emmerich for her accurate measurements.

21. 22. 23.

12. 13. 14. 15.

16.

17. 18.

19.

Lagaly, G., and Weiss, A., in “Proc. Int. Clay Conf. XX, 1975,” pp. 157. Sato, T., Watanabe, T., and Otsuka, R., Clays Clay Minerals 40, 103 (1992). Maes, A., and Craemers, A., J. Chem. Soc. Farad. I 74, 1234. Calvet, R., and Prost, R., 175–186 (1971). Brindley, G. W., and Ertem, G., Clays Clay Minerals 19, 399 (1971). Clementz, D. M., and Mortland, M. M., Clays Clay Minerals 22, 223 (1974). McBride, M. B., and Mortland, M. M., Soil Sci. Soc. Amer. Proc. 38, 408 (1974). Hofmann, V., and Klemen, R., Z. Anorg. Allg. Chem. 262, 95 (1950). Moore, D. M., and Hower, J., Clays Clay Minerals 34, 379 (1986). Jagodzinski, H., Acta Crystallogr. 2, 201 (1949). Moore, D. M., and Reynolds, R. C., “X-Ray-Diffraction and the Identification and Analysis of Clay Minerals,” p. 174. Oxford Univ. Press, Oxford, 1997. Brindley, G. W., Bish, D. L., and Wan, H.-M., Min. Mag. 41, 443 (1977). Eberl, D., Jones, B., and Khoury, H. N., Clays Clay Minerals 30, 321 (1982). Wilson, M. J. (Ed.), “A Handbook of Determinative Methods in Clay Mineralogy,” Chap. 7.3. Blackie, Glasgow/London, 1987. Moore, D. M., and Reynolds, R. C., “X-Ray-Diffraction and the Identification and Analysis of Clay Minerals,” p. 307. Oxford Univ. Press, Oxford, 1997. Moore, D. M., and Reynolds, R. C., “X-Ray-Diffraction and the Identification and Analysis of Clay Minerals,” p. 174. Oxford Univ. Press, Oxford, 1997. Hendricks, S. B., and Teller, E., J. Chem. Phys. 10, 147 (1942). Moore, D. M., and Reynolds, R. C., “X-Ray-Diffraction and the Identification and Analysis of Clay Minerals,” p. 178. Oxford Univ. Press, Oxford, 1997. Kraehenbuehl, F., Stoeckli, H. F., Brunner, F., and Mu¨ller-Vonmoos, M., Clay Min. 22, 1 (1987). Israelachvili, J., “Intermolecular & Surface Forces,” 2nd ed., p. 176. Academic Press, London, 1992. Gast, R. G., and Klobe, W. D., Clays Clay Minerals 19, 311 (1971). Gast, R. G., Soil Science Soc. Am. Proc. 36, 14 (1972). Sawhney, B. I., Clays Clay Minerals 20, 93 (1972).