mechanism of adsorption and desorption of …clays.org/journal/archive/volume...

Clays and Clay Minerals, Vol. 43, No. 3, 324-336, 1995.

M E C H A N I S M OF A D S O R P T I O N A N D DESORPTION OF W A T E R V A P O R BY HOMOIONIC MONTMORILLONITES: 2. THE Li § Na § K § Rb § A N D

C s + - E X C H A N G E D FORMS

ISABELLE B~a~END,~ JEAN-MAuRICE CASES, I MICHI~LE FRAN~OIS, i JEAN-PIERRE URIOT, 2 LAURENT MICHOT, I ARMAND MASlON, | AND FABIEN THOMAS ~

Laboratoire Environment et Min6ralurgie et UA 235 du CNRS, BP 40, 54501 Vand0euvre Cedex, France

2 Centre de Recherches P6trographique et G6ochimique, Vand0euvre, BP.20, Vand0euvre, France

Abstract--Methods previously used to distinguish between water adsorbed on external surfaces and in the interlamellar space of Na-montmorillonite during adsorption and desorption of water vapor have been extended to a set of homoionic Li-, Na-, K-, Rb- and Cs-montmorillonite. The textural and structural features have been investigated at different stages of hydration and dehydration using controlled-rate thermal analysis, nitrogen adsorption volumetry, water adsorption gravimetry, immersion microcalorimetry and X-ray powder diffraction under controlled humidity conditions. During hydration, the size of the quasi-crystals decreases from 33 layers to 8 layers for Na-montmorillonite and from 25 layers to 10 layers for K-montmorillonite, but remains stable around 8-11 layers for Cs-montmorillonite. Each homoionic species leads to a one-layer hydrate, which starts forming at specific values of water vapor relative pressure. Li-, Na- and K-montmorillonite can form a two-layer hydrate. By comparing experimental X-ray diffraction patterns with theoretically simulated ones, the evolution of structural characteristics of montmorillonites during hydration or desorption can be described. Using structural and textural data, it is shown that during adsorption: (1) the rate of filling of interlamellar space of the one layer hydrate increases with the relative pressure but decreases with the size of the cations; and (2) the different hydrated states are never homogeneous. Key Words--Adsorption, Desorption, Exchangeable cations, Immersion microcalorimetry, Montmoril- lonite, Relative humidity, Surface area, XRD.

I N T R O D U C T I O N

The sorption of water by phyllosilicates is governed by the size and charge of saturating cation, by the value and the localization of layer charge of the adjacent silicate sheets and by textural characteristics like the extension of layers in the ab plane (Tessier 1984). The former property has been applied to limit swelling of clayey soils by exchanging natural saturating cations by potassium ions (Veniale et al 1986).

Monovalent and large cations stabilize the one-layer hydrate under a large range of relative humidity whereas small divalent cations stabilize the two-layer hydrate in the same domain (Tarasevitch and Ovcharenko 1975). Layer stacking type encountered in the case of well crystallized phyUosilicates such as vermiculite and saponite is determined by the nature of interlayer cations and the layer charge density (Suquet and Pezerat 1987). Montmorillonites exhibit a turbostratic stacking of quasi-crystals or tactoids (Aylmore and Quirk 1971). The basal spacings (d000 of montmoril lonites dried at 200~ increase with the size of the exchangeable cation from 9.5/~ for Li- to 11.2/k for Cs-montmorillonite (Martin-Vivaldi et al 1963). Larger cations could present symmetrical or non-symmetrical stacking types corresponding respectively to 10.7 and 11.5 ~ for Cs-

Copyright �9 1995, The Clay Minerals Society

montmoril lonite (Besson 1980). Homogeneous hydrates of montmoril lonites are characterized by a small range of dool- Numerous results assign 12.4 to 12.6 to the doo, of the one-layer hydrate of Na-montmori l - lonite and 15.4 to 15.6 /~ to the d0o~ the two-layer hydrate (Mooney et a11952, Glaeser and M6ring 1968, Keren and Shainberg 1974, Iwasaki and Watanabe 1988). Less numerous results for other cations are scat- tered around the same values but do not always deal with homogeneous states (Mamy 1968, Kamel 1981). Basal spacings of hydrated beidellites are some tenth of an ~ smaller than those of montmoril lonites saturated with the same cations (Glaeser and M6ring 1968, Kawano and Tomita 1991).

Furthermore, textural features such as specific surface area in the dry state are affected by cation exchange (Ben Ohoud and Van Damme 1990). In clay pastes, narrow and numerous pores have been described in the case of Na-montmoril lonite in dilute NaC1 solutions whereas Ca-montmoril lonite in CaCl2 aqueous solutions exhibit larger and less numerous pores (Tes- sier 1984). In montmoril lonite suspensions, mean Stokes' diameters of tactoids are dependent on saturating cations (Na < K < Mg, Ca) (Whalley and Mul- lins 1991). In the hypothesis that textural features are modified during the first stages of hydration, the nature

324

Vol. 43, No. 3, 1995 Adsorption and desorption of water vapor by homoionic montmorillonites 325

o f the exchangeable cation would have a strong influence.

In the first papers of this series (Cases et al 1992a, 1992b), it was shown that a mult iple analytical approach allowed us to reconsider earlier models for hydration-dehydration of Na-montmori l loni te in the range o f intracrystalline swelling (Mamy 1968, Sposito and Prost 1982, More and Hower 1986, Kraehenbuehl et al 1987). The main features were: (1) interstratified hydrates persist at all water vapor relative pressure values higher than 0.25. The one-layer hydrate is pre- dominan t for 0.25 < P/Po < 0.7 during adsorption, and 0.15 < P/Po < 0.6 during desorption, whereas the two-layer hydrate is predominant beyond these limits; (2) the specific surface area of quasi-cystals increases from 43 m2.g -t to 105 m2.g - t when water molecules adsorb mainly on external surfaces (P/Po < 0.25); (3) the driving force for intercrystalline swelling is a com- binat ion o f the heat o f immers ion of exchangeable cations (Na) and the surface pressure on external surfaces; and (4) different hydrat ion stages can be distinguished, corresponding to the solvation o f exchangeable cations and to the filling o f remaining interlamellar spaces. In the present paper the same approach was used to study the influence o f the exchangeable cations Li +, Na +, K +, Rb § Cs + on the textural and structural evolution ofmontmor i l lon i te upon adsorpt ion-desorpt ion of water vapor.

EXPERIMENTAL

Materials

The investigation was carried out with a Wyoming montmori l loni te (< 2 #m fractions) supplied by CECA S.A. (Paris-France), previously dispersed, centrifuged and N a exchanged. Homoionic samples o f Li, Na, K, Rb and Cs were prepared by dialysing 12 g.l i ter -1 montmori l loni te suspensions with 0.5 N chloride solutions (analytical grade salts) o f the suitable cation. The material was washed by dialysis until a chloride residual concentration o f 10- 3 N was obtained and then centrifuged. The suspended samples were then freeze- dried. After treatment, the purity o f the samples was 99%, the major impuri t ies being a luminum oxy-hy- droxides. An approximate structural formula has been calculated from the chemical and electron microprobe analyses o f the Na form according to a method proposed by J. Yvon et al (1990):

(Si3 s83Alo 117)IV(All.s30Mgo.257FeIIo.oo6Felllo.207Olo) vI- (OH)2Nao.3so

The above formula corresponds to a theoretical exchange capacity o f 101 meq/100 g whereas the cationic exchange capacity (CEC) determined by the ammo- nium displacement method (Calll~re et al 1982a) was found equal to 99.6 meq/100 g, with a percentage o f

homoionic i ty o f 97.6 (in particular, K § 0.4 meq/100 g and Ca 2+ 2 meq/100 g).

Methods

To characterize the different states o f the l imited crystalline swelling during adsorpt ion and desorption of water vapor, different methods have been used, particularly to define the initial textural properties of the samples in the dry state in relation with pretreatment.

Controlled Transformation-Rate Thermal Analysis (CTRTA) (Rouquerol 1989, Gril let et al 1988). This process allows one to overcome partial overlaps o f successive dehydrat ion steps when using conventional thermal analysis. The system operates in the reverse way o f the conventional method, because the temper- attire is measured while a temperature dependent property o f the substance, here the water-content, is modified at a constant rate. This rate is controlled at a value low enough to ensure the satisfactory el iminat ion o f the temperature and pressure gradients within the sample. The flow o f gas evolved from the sample under a dynamic vacuum is used to control the heating of the furnace. For a constant composi t ion of the water vapor evolved, the temperature-vs- t ime data may be im- mediately converted into temperature-vs-mass-loss data, and therefore to a conventinoal thermogravi- metric curve. The experimental condit ions selected in the present study were a sample mass o f 300 rag, a residual pressure o f 1-2 Pa over the sample and a dehydrat ion rate o f 2.4 mg/h.

Nitrogen adsorption volumetry. Complete nitrogen gas adsorpt ion-desorpt ion cycles were determined using a conventional volumetry adsorption equipment (Delon 1970). Prior to each experiment, about 500 mg of sample was outgassed to a residual pressure of 1.3 10 -2 Pa at 100*C for 16 hours.

Water vapor adsorption measftrements. Water vapor adsorpt ion measurements were carried out in a con- t inuous gravimetric apparatus built around a Setaram MTB 10-8 symmetrical balance (Poirier et al 1987). Pr ior to each experiment, about 115 nag sample was outgassed under vacuum (1-2 Pa) to a temperature o f 100~ Water vapor was supplied from a source kept at 4 I~ through a Grandvil le-Phil l ips leak valve, at a slow flow rate to ensure quasi-equil ibrium condit ions at all times. The adsorpt ion isotherm, mass adsorbed at 30"C versus quasi-equil ibrium pressure, was directly recorded on an X-Y recorder. Pressures were measured using a Texas fused silica Bourdon tube automat ic gauge.

Immersion microcalorimetry at 30%7. Immers ion microcalorimetry at 30~ was used to determine the enthalpy o f immers ion in water o f the montmori l loni te sample (about 100 mg o f clay, outgassed under the

326 Brrend et al Clays and Clay Minerals

D

G

L

Figure 1. Schematic diagram of chamber and curved detector. a: Chamber, b: Curved detector, c: X-ray tube, d: Water vapor inlet, e: Aluminum window, f: micrometric slit, g: Ther- mocouple.

condit ions given above for the gravimetric adsorption o f water vapor, in a cell of about 6 cm 3) versus the pre-equil ibr ium relative pressure of water vapor following a method described by Grillet et al (1988). The enthalpy curve was then used to derive the external specific surface area o f quasi-crystals immersed in water (final state) by applying a modified Harkins and Jura procedure (Partyka et al 1979, Cases and Francois 1982, Fr ipia t et al 1982).

X-ray powder diffraction. Patterns of oriented samples were recorded with an apparatus specially built for this study. The oriented samples were prepared by pipetting aliquots of a suspension (0.35 g of montmori l loni te and 0.07 g of kaolinite (added to calibrate the curve detector) in 25 cm 3 o f demineral ised water) onto a mylar sheet covered with pharmaceutical gauze and evaporated to dryness in air. Dried clays films were separated from the mylar sheet, cut to appropriate size and set vertically on a sample holder enclosed in a chamber allowing the control o f relative humidi ty and temperature.

Pr ior to any experiment, samples were outgassed at 1-2 Pa and 100~ for 4-24 hours by connecting the chamber to vacuum pumps. Vacuum is checked on a Pirani gauge. The sample is heated to the desired temperature by radiat ion from two cupels set around the sample holder. During experiments, the temperature (30~ o f the sample is controlled by water circulation in the double-wall o f the chamber. The chamber is connected to a water vapor source. Acquisit ion of adsorpt ion data at a given water vapor pressure is ac- complished by recording the diffraction patterns at different t imes (1 hour, 2.30 hours and more than 4 hours (up to 72 hours depending o f the t ime required to reach the plateau corresponding to an apparent equilibrium).

Desorption data are obtained by recording the diffrac- t ion patterns for different t imes after the step of decrease of pressure induced by pumping 15 seconds and by the decrease o f the temperature of the water source.

X R D data were obtained with CoKa radiat ion using a CGR goniometer and an Inel CPS 120 curved detector (Figure 1). Data were collected, during 20 to 40 minutes, simultaneously over 60 degrees (20) and were processed using a mult ichannel Varrox analyser (2048 channels for 60 degrees). The diffractinel software im- planted in a Compaq Deskpro 286 calculator allowed transfer of data from the analyser, automatic peak search and automatic calibration using the doo~ and d0o2 peaks o f kaolinite as an internal standard.

X R D powder patterns were compared with simulated ones. The software used has been developed in the Laboratory of Crystallography of Orlrans, France, using the approach proposed by Drits and Sakharov (1976), Drits and Tchoubar (1990) and adapted by Besson and Kerm (Kerm 1988) for the study of the intensity diffracted along the 00 rod in reciprocal space from disordered phyllosilicates. The input data file in- cludes: (1) the abundances and posit ions of a toms in the different clay sheets (2:1 layer structure) and of water molecules in the interlamellar space (Data proposed by P~zerat and Mrring (1954) and by Drits (1975) and adapted from Ben Brahim (1985), Ben Brahim et al (1984, 1985, 1986) have been used); (2) the basal spacings corresponding to the different states of hydrat ion (zero, one, two or three layers of water); (3) the proport ions of the different hydrated states and the probabil i t ies of succession o f two kinds of layers; and (4) the distr ibution of the number of layers per quasi- crystals (tactoids). The detailed simulation procedure was described in previous works (Berend 1991, Cases et al 1992a).

RESULTS

Characterization o f the dry or initial state

Figure 2 shows the C T R T A curves for the different homoionic montmoriUonite samples. The successive dehydrat ion steps usually shown by conventional ther- mogravimetr ic analysis are more clearly evident (Cail- 16re et al 1982b). Region I, between room temperature and 100*C, corresponds to the desorption of physically adsorbed water. The next region (100~ 500~ corresponds mainly to the expulsion of water suppos- edly bound to exchangeable cations: the water loss decreases sharply with decreasing enthalpy of hydrat ion of exchangeable cations (Table 1). The final loss (around 5%) in region III (>500~ is due to the dehydroxy- lation of the structure.

Except for Li-montmori l loni te , the water contents after dehydrat ion at 100~ (Table 1) are slightly higher than results reported by Poinsignon and Cases (1978) and Cases et al (1992a). This difference suggests a con-


5

~o

20

+ M~t-Li

�9 M~t-Rb

~t4Zs

~ 25 I I I I : ~ I I

100 2oo 300 40o 5oo 6oo 7oo 800 9oo 1000 Temperature (~

Figure 2. Controlled-transformation-rate-thermal analysis curves of the different homoionic montmorillonites.

tr ibution of water adsorbed in the micropores to water desorbed above 100~

The nitrogen adsorption isotherms are all of type II with a very important type H3 hysteresis loop (Sing 1982), corresponding to capillary condensation in mesopores. The shape of the loop is usually associated with aggregates of plate-like particles giving rise to slit- shaped pores, and corresponds to a narrow size distr ibution of lamellar mesopores (Cases et al 1992a). A low pressure hysteresis also is observed extending to the lowest attainable pressures. It is weak for Li- (Fig- ure 3), Na- and Rb-montmoril lonites, and larger for K- (Figure 4) and Cs-montmorillonite. In addition, for the potassium exchanged form, the adsorption isotherms are not reproducible. Low pressure hysteresis except for Cs-montmorillonite, is assumed to be associated with the swelling and the change of the non- rigid porous structure formed by the organization of quasi-crystals upon N2 adsorption rather than to an irreversible uptake of molecules in micropores, present in the initial sample, having the same width as that of the adsorbate molecule. This assumption is supported by previous studies because the low pressure hysteresis loop was not observed on the Na-montmoril lonite sample obtained after air-drying (Cases et al 1992a). Particularly for K-montmoril lonite, physisorption is a destructive method and is not suitable for investigating mesoporosity. The total specific surface areas of samples calcula ted from the B r u n a u e r - E m m e t - T e l l e r

4O

3 0

2 o

> 2o

o

10

0 0.1 0 . 2 0 . 3 0 . 4 0 . 5 0 . 6 0 . 7 0 . 8 0 . 9 1.(

Relative Pressure

Figure 3. Adsorption-desorption isotherms of nitrogen at 77 K onto Wyoming Li-montmorillonite.

method (Table 2) increase with the radius size of the exchangeable cation. Assuming that powder contains discrete particles similar to square parallelepipeds of length ca. 3000 A and thickness ca. a multiple of the do01 given in Table 3, it is possible to calculate the mean size of the quasi-crystals (Cases 1985). The mean number of layers per quasi-crystals, decreases from 23 for Li-montmoril lonite to 11 for Rb-montmoril lonite. The model used to calculate the stacking of plates from adsorption data assumes the plates are perfectly stacked as in a deck of cards. This assumption is unlikely and leads to a coarse approximation because there is a good chance that the layers could be stacked in a stair case- like fashion or in a fractal like fashion as suggested by Ben Ohoud and Van Dame (1990). However the use of the procedure of de Boer et al (1966) reveals the presence of a small amount of micropores; nearly constant around 4 cm3.g - ~ (0.006 cm3.g - ~ converted to a liquid volume) except for Cs-montmorillonite, for which the basal spacing is large enough to allow some nitrogen molecules to be adsorbed in the interlamellar space. Surfaces available after the filling up of the micropores (real external surfaces areas of quasi-crystals)

Table 1. Main C.T.R.T.A. results.

E x c h a n g e a b l e c a t i o n s ( 1 ) L i + N a + K + R b + C s +

Water molecules per exchangeable cation (2) 1.54 1.36 1 .O0 1.18 1.02 Water bound to exchangeable cations (3) 3.0 2.6 1.9 2.1 1.7 Structural water (4) 5.3 5.2 5.1 4.9 4.7

1 = Nature of exchangeable cations. 2 = Number of water molecules per exchangeable cation at 100*C. 3 = Amount of water bound to exchangeable cations in % of the final mass. 4 = Structural water expressed in % of the final mass.

328 B6rend et al Clays and Clay Minerals

6 0

5 0

40

2 o 3O >

o 2 0

1 0

0 ~ r ~ I i r i I , I r I ~ I i I i I i

0 . 1 0 . 2 0 . 3 0 4 0 . 5 0 . 6 0 . 7 0 . 8 0 . 9 1

Relative Pressure

Figure 4. Adsorption-desorption isotherms of nitrogen at 77 K onto Wyoming K-montmorillonite and reproducibility.

can be de t e rmined fol lowing de Boer et a l (1966) procedure. The average size o f quasi-crystals calculated f rom the values o f external surface area corresponds to th icker stacks and decreases f rom 41 to L i -mon t - mor i l lon i t e to 11 layers for Cs-montmor i l lon i t e .

W a t e r vapor adsorp t ion-desorp t ion i so therms

Wate r v a p o r sorp t ion i so therms are i l lustrated in Figure 5. A m o u n t s o f water adsorbed suggest that a one- layer hydra te fo rms for all mon tmor i l lon i t e s corresponding to an a m o u n t o f adsorbed water 6 -7 m m o l - g-~. The two- layer hydrate state seems to be fo rmed only for Li- and N a - m o n t m o r i l l o n i t e (13-14 m m o l . g-~). The shape o f the i so therms is inf luenced by the nature o f exchangeable cation. The irregular shapes o f bo th the adsorp t ion and desorp t ion branches o f all i so therms suggest complex water adsorp t ion mecha-

Table 3. Determination of the evolution of the external surface area on homoionic montmorillonite for lower values of water vapor relative pressure taking into account the reference isotherms of Hagymassy et al 1969 (Berend 1991).

N a § K + R b + C s +

P/Po(1) 0.025 to 0.2 0.025 to 0.2 0.025 to 0.07 0.025 C (2) 23 12 5 5 Sm~g, (3) 42.5 to 83.4 44.6 to 77.5 50.5 to 74.3 79.0

1 = range of relative pressure considered. 2 = energetic constant of the B. E. T. theory. 3 = evolution of the external specific area of quasi-crystals with increasing relative pressure in the range considered.

n i sms (Cases et a l 1992a). The adsorp t ion p h e n o m e n a occur on surfaces whose specific surface areas are not cons tan t but d e p e n d on the relative pressure as is often the case for swelling clays. As expected f rom the high energy o f hydra t ion o f Li § the filling o f the interla- mel la r space o f L i -mon tmor i l l on i t e begins at very low rela t ive pressure. In these condi t ions, it is imposs ib le to de t e rmine the external specific surface area o f the quasi-crystal using the BET procedure. Fo r the o ther cations, the shape o f the i so therms and the X R D patterns (Figure 6) reveal that adsorp t ion occurs on external surfaces only at low rela t ive pressure. Adsorp- t ion in the inter lameUar space begins for par t icular values o f the re la t ive pressure decreasing with increasing the size o f the exchangeable cation. K a w a n o and T o m i t a (1991) repor ted that after dehydra t ion for one hour at 400"C, rehydra t ion proper t ies o f smect i tes are strongly dependen t on the size o f the inter layer cat ions (K + > N a + > Ca 2§ > Mg 2§ ) because large ionic radii impa i r the fixation o f cat ions in the hexagonal holes. The present results ob ta ined after dehydra t ion at 100~ can be expla ined by larger exchangeable cat ions cre- at ing larger basal spacings. This d imin i shes the interlayer forces o f cohes ion and therefore increases the accessibil i ty o f cat ions by water molecules .

Adsorp t ion i so therms on the external surfaces o f mon tmor i l l on i t e can be c o m p a r e d to the reference ad-

Table 2. Specific surface areas calculated from the B.E.T. method and volumes of micropores evaluated following the procedure of de Boer et al (1966).

L i § N a § K + R b + C s +

Vm (cm3-g - ' ) (1) 8.6 9.65 10.65 15.4 29.8 S,o~, (m2-g -~) (2) 37.7 42.2 46.5 67.4 130.3 V (cm3-g -~) (3) 4.0 3.4 3.2 5.6 19.1 Sox, (m2"g - ' ) (4) 23.2 26.3 34.7 46.7 64.2 M(5) 41 33 25 17 11

1 = monolayer capacity calculated from the B. E. T. equation per unit mass of adsorbent. 2 = corresponding total specific surface area. 3 = total microporous volume per unit mass of adsorbent. 4 = external specific surface area calculated from de Boer et al, 1966 procedure, and 5 = average number of layers per particles or quasi-crystals derived from external surface area (average thickness of the elementary stacks).

_r Li ~i:i N, �9

_

~elmtlve P r ~ . ~ aelall~e r~s~u~

~ l l s ~ K ~ s 6 i R b ~ i ~ C s

:i ~ - ~ 2 . ~ o ~ f f o . . . . 0 o: o4 g6 ~s 1 0 o~ 0a ~s os 1 o 0~ g4 o6 0~

Relative P ~ u ~ Relative Pressu~ Re]ative p ~ u r e

Figure 5. Water adsorption and desorption isotherms onto Li-, Na-, K-, Rb-, and Cs-montmorillonites.


n . K ~ ~6 R b ~6 Cs

Relative P~SU~ Relative le~enar t Rellti~ P r ~ u ~

Figure 6. Evolution of the values of d(001) for montmorillonites saturated by the indicated cation during hydration and dehydration.

sorption isotherms of water vapor of non-porous samples with the same BET energetic constant, C (Hagy- massy et a! 1969, Berend 1991). For a given value of P/Po, i f Q a is the amount of water adsorbed per gram of montmoril lonite, 0 the surface coverage of the reference isotherm taken into account, N A the Avogadro number and r the cross-sectional area of the water molecule strongly adsorbed on the external surface, the specific surface area of montmoril lonite S is given by the relationship:

S = (Qa/0)*NA*r (1)

The specific surface areas of Na-, K- and Rb-montmorillonite are found to increase with P/Po for every value of the constant C (Cases et al 1992a) (Table 3). Subsequently, the value of the constant C is chosen in order to obtain the same surface area by applying the BET procedure to the beginning of water adsorption isotherms and to nitrogen adsorption isotherms. The classical value of 14.8 ~2 proposed by Harkins and Jura (1944) was used for a.

X R D patterns

Typical XRD patterns of oriented samples are given in Figures 7-9 for samples in the initial state (zero- layer hydrate state), and for samples in quasi equilib- r ium with relative pressures between 0.4-0.6 and 0.92- 0.98 respectively. The evolutions of basal spacings, d(001), during hydration and dehydration are presented in Figure 6. Usually for low relative pressures, the kinetics of swelling of most montmorillonites are rapid. The equilibrium time can reach 24 hours and sometimes 72 hours for highest relative pressures. Dur- ing adsorption, interstratified structures are present for the whole range of relative pressures because the curves present no steps (Figure 6). Only the most hydrated states seem to be homogeneous. Another evidence for the presence of interstratified structures is given by the absence of plateaus and steps on the water vapor ad-

I N I T I A L S T A T E ( d r y s t a t e )

K

I I I I i I

Figure 7. Observed X Ray diffraction tracings of samples in the initial state (dry state). K represents kaolin peaks added in the XRD experiments for internal calibration of the curved detector and M, the montmorillonite peaks.

sorption isotherms (Figure 5). During desorption quasi-homogeneous states expected at about 10 A for the dry state (OS), ~ 12.4 A for the one-layer hydrate state (1S) and ~ 15.5 A for the two-layer hydrate state (2S) alternate with interstratified structures. To understand the mechanism of water uptake, it is now important to consider the possible change of the external surface of quasi-crystals during hydration.

Characterization o f the f inal size o f quasi-crystals

The final size of the quasi-crystals can be determined by measuring the heat of immersion of homoionic montmoril lonites after preadsorption of increasing numbers of water layers (Cases and Francois 1982, Fripiat et a! 1982, Cases et a! 1992a). The immersion enthalpies of montmoril lonites obtained from the dry state equal 70 J -g- ' for Li- and Na-montmori l lonite and around 30-40 J.g-~ for the other montmori l lonites. Enthalpies of immersion decrease with precoverage up to relative pressures varying between 0.5 and 0.7 and then remain approximately constant (Figure 10). The decrease of the enthalpy of immersion with precoverage is correlated with the amount of water adsorbed which progressively saturates the most en-

330 B6rend et a l Clays a n d Clay M i n e r a l s

0.4 < P / P o < 0 .6

K

Li ~ ~ l 1.39 K M

o

Rb I 2,01 K

0 .92 ~ P / P o

El

K X 1~.2S IS,34

M

16.43

M

;o ;1, ,', ,'o; '~ ~ ', ', " ~'~ ; ~

Figure 8. Observed X Ray diffraction tracings of samples under a relative pressure between 0.4 and 0.6 during adsorption and after reaching the quasi equilibrium time: Li-montmorillonite, P/Po = 0.4; t = 90 h; Na-montmorillonite, P/Po = 0.5; t = 18.45 h; K-montmorillonite, P/Po = 0.5; t = 41 h; Rb-montmoriUonite, P/Po = 0.6; t = 67 h; Cs-montmorillonite, P/Po = 0.5; t = 19 h. K represent kaolin peaks added in the XRD experiments for internal calibration of the curved detector and M, the montmorillonite peaks.

ergetic sites on the external surfaces and in the inter- lameUar space. Cons tan t va lues are measured dur ing the hydra t ion o f the quas i -cons tan t interstrat if ied 0S/ 1 S s tructure o f L i -mon tmor i l l on i t e and dur ing adsorp- t ion o f water molecules on the external surface o f Na- and K-mon tmor i l l on i t e . In this lat ter case, this phe- n o m e n o n is ma in ly due to the large difference be tween the surface area o f the sample before i m m e r s i o n (Table 3) and the cor responding va lue after i m m e r s i o n in water (about 800 m2.g- I ) .

Us ing the Hark ins and Jura formula, the plateau obse rved at high re la t ive pressure can be used to der ive the external specific surface o f quasi-crystals immersed in water, S . j (Table 5). The value obta ined for Na- m o n t m o r i l l o n i t e is lower than that presented recently (105 m2.g -1, Cases e t a l 1992a). The defini t ion o f the pos i t ion o f the pla teau in the high precoverage d o m a i n lacks precis ion because preadsorp t ion and i m m e r s i o n p h e n o m e n a are not total ly reproducib le as they are

Figure 9. Observed X Ray diffraction tracings of samples under a relative pressure around 0.95 during adsorption and after reaching the quasi equilibrium time: Li-montmorillon- ire, P/Po = 0.92; t = 4 h; Na-montmorillonite, P/Po = 0.98; t = 18.30 h; K-montmorillonite, P/Po = 0.97; t = 72 h; Rb- montmoriUonite, P/Po = 0.97; t = 24 h; Cs-montmoriUonite, P/Po = 0.97; t = 24 h. K represent kaolin peaks added in the XRD experiments for internal calibration of the curved detector and M, the montmorillonite peaks.

affected by small changes in the exper imenta l condi- t ions. Indeed, in this range o f precoverage a compro - mise mus t be chosen as high clay weights are needed for an acceptable the rmic response o f the apparatus whereas low clay quant i t ies ensure a perfect swelling o f the clay in the cell. A longer precoverage t ime was considered for L i -montmor iUoni te . The m e a n value o f the enthalpy o f i m m e r s i o n was used for Rb- and Cs- montmor i l lon i te . The values o f the external surface areas de r ived in this m a n n e r are close to those obta ined by treat ing the water adsorp t ion i so therms jus t before water molecules start enter ing the in ter lamel lar space (Table 3). These values which are related to the m e a n size o f quasi-crystals dur ing adsorp t ion will therefore be used in subsequent model l ings. They are calculated wi thou t any assumpt ion concerning the value o f the cross-sect ional area o f the water molecule adsorbed on the solid. Fo r Na- , K-, Rb- and Cs-montmor i l lon i tes , the external surface areas o f quasi-crystals are constant dur ing the filling o f the in ter lamel lar space.

Vol. 43, No. 3, 1995 Adsorption and desorption of water vapor by homoionic montmorillonites

Table 4. Parameters used for the simulation of XRD patterns.

331

M in initial state (I) doo, (A) (2)

apparent doo~ Scherrer formula zero-layer hydrate one-layer hydrate two-layer hydrate

Li § 10 8 9.6-10.0 12.0-12.2 15.6 Na § 12 9 9.55 12.5 15.55 K § 9 7 9.95 12.5 16.0 Rb § 18 10 10.25-10.6 12.5 Cs § 20 11 10.7-11.5 12.4

1 = the range of thicknesses of unit layers making the diffracting domain in the interference function. 2 = basal spacings corresponding to the 001 reflections.

DISCUSSION

Simulation of the X-ray powder patterns in the small-angle X-ray scattering domain (the 001 reflection) at various hydration states

The comparison of simulated XRD patterns with the experimental ones is not an easy task because of the disordered structure of the samples and of the experimental procedure used i.e., oriented preparations analyzed in reflection. This results in four main problems: 1) measured and calculated intensities do not agree for small angles which make the determination of the mean number of layers in a stack (quasi-crystal), M, inaccurate; 2) relative intensities of first order reflection and harmonics are often not consistent with the models predicting less intense harmonics; 3) back- ground noise is sometimes important for higher relative pressures; and 4) large differences in the clay structure in the range of the two-layer hydrate often cor- respond to small changes in the XRD pattern due to the poor accuracy obtained in the small angle region. Consequently the positions of the reflections were taken into account rather than the widths or relative intensities of the reflections. The precision obtained was better than 10% except in the range of the two-layer hydrate, where it is difficult to choose between a quasi- homogeneous state and an interstratified one (for ex- ample 100% 2S or 10% 3S; 70% 2S; 20% 1S).

The parameters used for the simulation are given in Table 4. For Li-montmoril lonite two states in the zero- layer state have been taken into account (Li-H20, 10 ik and Li, 9.4/~ (Suquet et al 1982). Small changes in the positions of interlayer cations or water molecules do not affect the XRD pattern except for the Cs + cation.

Table 5. Specific surface areas of the quasi-crystals after immersion in water.

Li + Na § K + Rb + Cs +

SHj (m2"g -') (I) 75.3 84.0 41.8 55.4 85.6 M (2) 9 8 18 13 8

1 = surface calculated using the Harkins and Jura formula. 2 = corresponding mean number of layers in the quasi-crystals assuming a model of parallelepiped.

Basal spacings of the models have been set taking into account the harmonics of the more homogeneous states and results from the literature. The mean numbers (M) of layers per stacks were determined in order to obtain a good superimposition of the positions of the experimental and calculated reflections. These values are higher than values obtained using the width of the doo~ reflection particularly for Rb- and Cs-montmoril lonites which present symmetrical (hexagonal cavities face to face) and non-symmetrical (hexagonal cavities shift- ed) stacking types, noted 0S Sym and 0S Non Sym respectively in Figures 11 and 12. During adsorption, the amounts of symmetrical and non-symmetrical dry state were taken equal for Cs-motmorillonite. But in order to explain the broadening of peaks d(002) and d(004) and the high value of the position of the peak d(002), the different states (0S Sym, 0S Non Sym, 1S) were assumed to be separated (unmixed) for values of relative pressure from 0.15 to 0.50 (Figure 11).

Proportions of the different hydrates versus P/Po during adsorption and desorption are illustrated on Figures 11 and 12, respectively. A distinction has been made between symmetr ica l and non- symmet r i ca l stacking types for Rb- and Cs-montmoril lonite in the dry state. This determination, only a rough estimate for the hydrated state, will not be taken into account

K t':' :t R, :! T r,

Figure 10. Enthalpy of immersion (dots and line) and water adsorption isotherms (line) vs. precoverage relative pressure for Li-, Na-, K-, Rb-, and Cs-montmorillonites.


~ L i ~ too N a 7 !oo L I L i

~ .+~ o o2 o+ o6 o,B i o o.2 o l o~ oa i o o2 o+ o~ os i

Re l l t i v , P r r Relatlv~ pressure Relative Pre~u~

+_+ ++++ R,, +,++ c, -+,++'+ ,,

,+p .+ , +

-++,o ,

. . . . . . . . . . . ; . . . . . . . . . . . . i . . . . . . . . . . . ; o ' Relative P n m a ~ Re la t l~ Pre~ure l e h f i v e P ~ S u ~ 00 02 0 4 06 O8 ]

a d a a ~ . P m s . ~

Figure 11. Evolution of the different hydrate states (zero-, one-, and two-layer hydrates) with water relative pressure during adsorption for Li-, Na-, K-, Rb-, and Cs-montmorillonites.

I N a

o 02 o+ 0+ 08 i

B e , , , 3 0 ,

o 02 0+ 06 0s 1 o o2 04 06 o8 i

ReIIUvr l ' r ~ u m melJlllve P r ~ +

Figure 12. Evolution of the different hydrate states (zero-, one-, and two-layer hydrates) with water relative pressure during desorption for Li-, Na-, K-, Rb-, and Cs-montmorillonites.

later on. In the initial state, 100% of 0S is assumed. During adsorption, 10% 1S is observed for P/Po equals to 0.02 for Li-, 0.33 for Na-, 0.30 for K-, 0.20 for Rb-, and less than 0.08 for Cs-montmorillonite. About 50% 1S is observed for all homoionic montmorillonites at P/Po = 0.5. For higher values of P/Po: 1) a two-layer state hydrate is formed for Li-, Na- and to a lesser extent for K-montmoril lonite; (2) the one-layer state develops for Cs-montmorillonite; and (3) seems to be unchanged for Rb-montmoril lonite. Quasi-homogeneous 2S of Li- and Na-montmori l lonite are observed for P/Po higher than 0.9 taking into account the previous reservations. Desorption of two-layer states for Li-, Na-, and K-montmori l loni te leads to a large proportion of one-layer state and a small proportion of zero-layer state (Figure 12). Complete closure of the layers takes place at P/Po values which decrease with decreasing cation size. For Rb- and Cs-montmoril lonite a small decrease of 1S is observed for P/Po higher than 0.2 and a complete interlamellar space closure for lower values of P/Po-

Textural evolution during water adsorption

From the previous results and according to Cases et al (1992a), it can be assumed that during the first stage of water adsorption the elementary tactoids of Li- and Na-montmoril lonite , of about 40-30 layers thick in the dry state, are split into smaller ones corresponding to about 8-9 layers thick. The specific surface area of these tactoids remains constant even at the highest relative pressures investigated.

The specific surface area of Cs-montmorillonite remains constant in the dry state and during hydration at about 65-80 m2-g - ' , corresponding again to about 8 layers per tactoid. This value agrees with the result obtained by applying the Scherrer formula to the dool line broadening of Cs-montmorillonite in the dry state. It is the thickest stack found for all the cations investigated.

K- and Rb-montmori l lonite show an intermediate evolution. A trend to splitting is observed (Table 3) in the first stage of water hydration but the opposite trend appears for higher relative pressure resulting in a more compact texture of the hydrated material.

Literature results (Cebula et al 1980, Stul and Van Leemput 1982, Schramm and Kwak 1982) show that the surface area of dry montmoril lonite increases with the size of the cation and that in dilute suspensions, the mean number of layers per tactoid increases with the cation atomic weight from 1 for Li § to 4 (or 5 depending of the equation used) in the case of Cs § Such results involve a splitting of the tactoids saturated with the smaller cations, in agreement with the present work.

A large discrepancy is noticeable between the number of layers per tactoid obtained with the Harkins and Jura formula (Table 5) and results reported by Schramm and Kwak (1982). The difference can be assigned to the fact that the bulk concentration of the final state obtained with immersion microcalorimetry is 8 times higher than the dilute bentonite suspensions (2 g per liter) investigated by these authors.

Connection between water adsorbed on external surface and water present in interlamellar space

For a given relative pressure, the total amount of water present in montmoril lonite is the sum of the initial water content (Q~at, Table 1, line 2) and of the experimental adsorbed amount (Qa). The surface coverage 0ext of the external surface area of quasi-crystals is given by the reference isotherm of a non-porous adsorbent having the same BET energetic constant (Ta- ble 3, Hagymassy et al 1969). The surface areas (S) are known from nitrogen adsorption, water vapor adsorption and microcalorimetry measurements. The external water amount (Qext) can be derived using the reference isotherm and equation (1).


a . . : - "

I ~ ~.=,~ j i ~ ..---:::: ::i/'~

'2i 1,0 i,0t

Rel~ive Pftssu~ IIElall~ pl,eDlute Rellti~ P ~

Figure 13. Calculated water adsorption-desorption isotherms on the external surface and in the interlamellar space of Li-, Na-, K-, Rb-, and Cs-montmorillonites.

The remaining water is internal. It is mainly adsorbed in the interlamellar space (about 7-8 mmol .g for a monolayer) and to a l imited extent in the microporosity. The l imit between micropores and interla- mel lar space is somewhat imprecise for Cs-montmorillonite which has a high do0t value in the dry state. In this case, the interlamellar space could become part o f intra-aggregate porosity.

Plotting o f the internal and external adsorption-desorption isotherms (Figure 13) shows that the internal water content is higher than the external one in the whole range of relative pressures investigated. External water cannot be neglected. It increases sharply at low and at high relative pressures.

Filling of the interlamellar space

As the amount of the external water and the con- centrations o f the different hydrate states are known over the whole range of relative pressure, the uptake of water in the interlamellar space can be calculated assuming that:

(1) The internal lamellar specific area Sin, varies according to the relation

s~,t = (801.3 - ( so=, - Se=, ~,)) (2)

where Se=t represents the total external specific surface area and S e x t l a t , the lateral external specific surface area. I f the tactoids are assumed parallepipeds, Sex, la, is equal to 5.1 m2.g -~, 6.7 m2.g ~ and 8.3 m2.g -I in the dry state, monolayer hydrate, two-layer hydrate respec- t ively (Cases 1985).

(2) Models o f the one-layer hydrate and two-layer hydrate proposed by Ben Brahim et al. (1985) for Na- beidelli te are valid for homoionic montmoril lonites. Cross sectional areas o f 7.8/~2 (a~) for the water molecule in the one-layer hydrate and 8.7/~2 (tr2 and (r3) in the two- and three-layer state are assumed.

The a m o u n t of water, Qm~, in mmol g - ' , adsorbed

120

100

8O

4O

20

0

0 0

r �9

�9 ql~ �9 A ~ D

[]

A r a

0 50 100 1 5 0 2 0 0

% of interlamellar swelling

Figure 14. Evolution of the filling up of interlayer water versus mean interlamellar swelling during water adsorption of Li-, Na-, K-, Rb-, and Cs-montmorillonites.

in the interlamellar space as a monolayer (i = 1), bilayer (i = 2) and third layer (i = 3) is given by the relation

Qmi = i'Smt/2"cq'NA (3)

For given abundances of relative proport ions Wi o f each type o f layer, the theoretical adsorbed amount in the interlamellar space Qin, th is given by the relation

Q i n t th = W 0 * Q i n i t q'- W,*Qm~ + W 2 * Q m 2 -t- W a * Q m 3

(4)

140

120 &

100

~ so

~ 60

Q 40

2o

o

Figure 15.

A O A

A~ D Q

I i I i I | I i

50 100 150 200 250

% Of intedamellar swelling

Evolution of the filling up of interlayer water versus mean interlamellar swelling during water desorption of Li-, Na-, K-, Rb-, and Cs-montmorillonites.


The filling up of the interlamellar space is obtained from the ratio Qi,t/Qint ~h where Qint is the experimental amount of water in the internal space previously calculated. The mean interlamellar swelling is defined as the sum of W~ balanced by i (Figures 14 and 15). The lack of accuracy at small diffraction angles makes the filling-up estimate less precise at high relative pressures. Filling-up of the interlamellar space of montmorillonites lies mainly in the range between 50% and 100%. Higher values can be obtained. In the case of Li-montmorillonite, the evolution with time of XRD patterns suggests slower swelling kinetics than for isotherms. In the case of K-montmoril lonite, the external amount of adsorbed water seems to be underestimated taking into account the Harkins and Jura surface at low relative pressures. An increase of the external surface area occurs probably as it has been observed during adsorption for the same relative pressure.

Upon adsorption a large filling-up of the interlamellar space (> 40%), decreasing with increasing cation size, is observed at the very beginning of innercrystalline swelling. The filling-up increases with increasing swelling particularly for the larger cations. In the case of Li- and Na-montmoril lonite, a decrease is observed during the formation of the two-layer hydrate state.

For Na- and K-montmoril lonite, the completion of the water monolayer on the external surfaces (Figure 13) occurs at the same relative pressure as the beginning of the formation of the one-layer hydrate. The same correspondence can be noted for the completion of the hilayer on the external surface of Li- and Na-montmorillonite and the beginning of the formation of the two-layer hydrate state. The surface pressure seems to be a driving force for swelling in addition to the heat of hydration. This was previously demonstrated in the case of Na-montmori l loni te by Cases et al (1992a).

Upon desorption a sharper decrease of the filling-up is observed for the closure of the layers except in the case of K-montmoril lonite, which seems to keep water enclosed in the inter-aggregate porosity or interlamellar space. The final remaining filling-up decreases with increasing cation size, except for K +. In the case of Li- and Na-montmori l lonite a decrease of the filling-up is also observed during the transition from the two-layer hydrate to the one-layer hydrate.

CONCLUSIONS

The study of five homoionic montmoril lonites saturated by alkaline cations allows us to support the mechanism of adsorption and desorption of water vapor obtained on a single Na-montmoril lonite by Cases et a l (1992a):

1) All the montmoril lonites form interstratified hydrated states during adsorption. The opening of the layers occurs as defined relative pressure values which depend on the hydration energy of the cation and on

dora in the dry state. More homogeneous states are encountered during desorption.

2) Surface areas of Li- and Na-montmoril lonite increase during water adsorption, as expected from the splitting of quasi-crystals. However a tendency to ag- gregation can be noted during the formation of the one- layer hydrate of K- and Rb-montmoril lonite. Cs-montmorillonite remains unchanged during water vapor adsorption.

(3) The bidimensional surface pressure acts as a driving force for the innercrystalline swelling only for small basal distances (dool < 10 ,~) and for average energies of hydration, as is typically the case for Na- montmorillonite.

(4) For all montmorillonites, the filling-up of the interlamellar space varies during the completion of a hydrated state. This corresponds to the solvation of exchangeable cations and to the filling-up of the remaining interlamellar space. Solvation of the larger cations involves a lower number of water molecules per cation.

ACKNOWLEDGMENT

We thank the French Ministery of Research and Technology (Phygys program), the PIRSEM-ARTEP program and the CNRS (Departement Sciences de l 'U- nivers) for support of this work, Yves Grillet (Centre de Thermodynamique et de Microcalorim6trie du C. N. R. S.-Marseille), F. Lhote (Centre de Recherche P&rographique et G6ochimique-Vandteuvre), for their help and advice during the course of this study and T. Pinnavaia for his valuable discussions and correction of the English.

REFERENCES

Aylmore, L. A. G., and J. P. Quirk. 1971. Domains and quasi-cristalline regions in clay systems. Soil Sci. Soc. Amer. J. 35: 652-654.

Ben Brahim, J. 1985. Contribution ~ l'6tude des syst~mes eau-argile par diffraction des rayons X. structure des couches ins6r6es et mode d'empilement des feuillets dans les hydrates homog6nes h une ou deux couches d'eau de la beidilite-Na: Th~se de doctorat d'6tat es sciences physiques, Orl6ans, 183 pp.

Ben Brahim, J., G. Besson, and C. Tchoubar. 1984. Etude des profils des bandes de diffraction X d'une beidillite-Na hydrat6e ~t deux couches d'eau. D6termination du mode d'empilement des feuillets et des sites occup6s par l'eau. J. Appl. Cryst. 17: 179-188.

Ben Brahim, J., G. Besson, and C. Tchoubar. 1985. Layer Succession and Water Molecules Arrangement in a Ho- mogeneous Two-Layer Na-Smectite. 5th Meeting of the European Clay Groups, Prague. J. Konta, ed. Praha: Uni- verzita Karlova, 65-75.

Ben Brahim, J., N. Armagan, G. Besson, and C. Tchoubar. 1986. M6thode diffractom6trique de caract6risation des 6tats d'hydratation des smectites et stabilit6 relative des couches d'eau ins6r6es. Clay Miner. 21:111-124.

Ben Ohoud, M., and H. Van Damme. 1990. La texture fractale des argiles gonflantes. C. R. Acad. Sci. Paris 311: S6rie II, 665-670.

Vol. 43, No. 3, 1995 Adsorption and desorption of water vapor by homoionic montmoriUonites 335

Berend, I. 1991. Les m6canismes d'hydratation de montmorillonites homoiniques pour des pressions relatives in- f6rieures a 0,95. Th~se de doctorat. INPL-Nancy, France, 330 pp.

Besson, G. 1980. Structure des smectites diocta6driques. Param~tres conditionnant les fautes d 'empilement des feuiUets. Th~se de doctorat d'Etat. Universit6 d'Orl6ans, France, 146 pp.

Caill~re, S., S. H6nin, and M. Rautureau. 1982a. Mindral- ogle des argiles. I. Structure et propri(t(s physico-chimiques, 2 ~ 6dition, Paris: Masson, 184 pp.

Caill~re, S., S. H6nin, and M. Rautureau. 1982b. Mindral- ogie des argiles. 2. Classification et nomenclature, 2 e 6dition, Paris: Masson, 189 pp.

Cases, J. M., and M. Franqois. 1982. Etude des propri6t6s de l 'eau au voisinage des interfaces. Agronomie 2:931-938.

Cases, J. M. 1985. Personal communication. Cases, J. M., I. Berend, G. Besson, M. Franqois, J. P. Uriot,

F. Thomas, and J. E. Poirier. 1992a. Mechanism of adsorption-desorption of water vapor by homoionic montmorillonite. 1. The sodium exchanged form. Langmiur 8: 2730-2739.

Cases, J. M., C. H. Pons, I. Berend, M. Francois, Jin Hong Min, D. Tchoubar, G. Besson, F. Thomas, and J. Y. Bot- tero. 1992b. Fluid-swelling clays interaction, Revue de l'Institut Franqais de P(trole 47:217-238.

Cebula, D. J., R. K. Thomas, J. W. White. 1980. Small angle neutron scattering from dilute aqueous dispersion of clays. J. C. S. Faraday 176: 314-321.

de Boer, J. H., B. C. Lippens, B. G. Linsen, J. C. P. Broekhollf, A. Van der Heuvel, and Th. J. Osinga. 1966. The t-curve of multimolecular N2 adsorption. J. Coll. Interface Sci. 21: 405-414.

Delon, J .F . 1970. Contribution ~ l'6tude de la surface sp6- cifique et de la microporosit6 des min6raux et des roches. Th~se de doctorat 6s sciences physiques. Nancy, France: INPL, 187 pp.

Drits, V.A. 1975. The structural and crystallochemical features of layer-silicates. In Crystallochemistry of Minerals and Geological Problems. A. G. Kossovskaya, ed. Novo- sibirsk: Nauka, 35-51 (in Russian).

Drits, V. A., and B. A. Sakharov. 1976. X-ray Structure Analysis of lnterstratified Minerals. Moscow: Nauka, 255 pp. (in Russian).

Drits, V. A., and C. Tchoubar. 1990. X-ray Diffraction by Disordered Lamellar Structures. Berlin: Springer-Verlag, 37 l PP.

Fripiat, J. J., J. M. Cases, M. Franqois, and M. Letellier. 1982. Thermodynamic and microdynamic behavior of water in clay suspensions and gels..L Coll. Interface Sci. 89: 378- 400.

Glaeser, R., and J. M6ring. 1968. Domaines d'hydratation homog~ne des smectites. C. R. Acad. SC. Paris sPrie D 267: 463-466.

Grillet, Y., J. M. Cases, M. Franqois, J. Rouquerol, and J. E. Poirier. 1988. Modification of the porous structure and surface area of sepiolite under vacuum thermal treatment. Clays & Clay Miner. 36: 233-242.

Hagymassy, J., S. Brunauer, and R. S. Mikhail. 1969. Pore structure analysis by water vapor adsorption. 1. T-curves for water vapor. J. Coll. Interface Sci. 29: 485-491.

Harkins, W. D., and G. Jura. 1944. An absolute method for the determination of the area of a finely divided crystalline solid. J. Am. Chem. Soc. 66: 1366-1375.

Iwasaki, T., and T. Watanabe. 1988. Distribution of Ca and Na ions in dioctahedral smectites and interstratified dioctahedral mica-smectites. Clays & Clay Miner. 36: 73-82.

Kamel, M . W . 1981. Etude de l ' imbibition, du gonflement

et du dess6chement de quelques argiles. Th6se doctorat. Universit6 Toulouse, 187 pp.

Kawano, M., and K. Tomita. 1991. X-ray powder reflection studies on the rehydration properties of beideilite. Clays & Clay Miner. 39: 77-83.

Keren, R., and I. Shainberg. 1974. Water vapor isotherms and heat of immersion ofNa/Ca- montmorillonite systems. I: Homoionic clay. Clays & Clay Miner. 23: 193-200.

Kerm, A. G. 1988. Etude et caract6risation des premiers stades d 'hydratation d'une nontronite. Th6se d'Universit6, Ofl6ans, France, 56 pp.

Kraehenbuehl, F., H. F. Stoeckli, F. Brunner, G. Kahr, and M. Mueller-vonMoos. 1987. Studyofthewater-bentonite system by vapour adsorption, immersion calorimetry and X-ray techniques: I. Micropore volumes and internal surface areas, following Dubinin 's theory. ClayMiner. 22: 1-9.

Mamy, J. 1968. Recherches sur l 'hydratation de la montmorillonite: Propri6t6s di61ectriques et structure du film d'eau. Ann. Agron. 19: 175-292.

Martin-Vivaldi, J. L. M., D. M. C. Macewan, and M. Rod- riguez Gallego. 1963. Effects of thermal treatment on the c-axial dimension of montmorillonite as a function of the exchange cation. In Proc. Int. Clay Conf., Stockholm. Vol. 1. New York: Pergamon Press, 45-51.

Mooney, R. W., A. G. Keenan, and L. A. Wood. 1952. Adsorption of water vapor by montmorillonite. II--Effect of exchangeable ions and lattice swelling as measured by X-ray diffraction. J. Am. Chem. Soc. 74: 1371-1374.

More, D. H., and J. Hower. 1986. Ordered interstratification of dehydrated and hydrated Na-smectite. Clays & Clay Miner. 34: 379-384.

Partyka, S., F. Rouquerol, and J. Rouquerol. 1979. Calor- imetric determination of surface areas. Possibilities of a modified Harkins and Jura procedure. J. Coll. Interface Sci. 68: 21-31.

P6zerat, H., and J. Meting. 1954. Recherches sur la position des cations 6changeables et de l 'eau dans les montmorillonites. CRAS Paris 265: D, 529-532.

Poinsignon, C., and J. M. Cases. 1978. Etude de l'eau d'hydratation des cations compensateurs de smectites homoio- niques. Bull. Mineral. 101: 469-477.

Poirier, J. E., M. Francois, J. M. Cases, and J. Rouquerol. 1987. Study of water adsorption on Na-montmorillonite. New data owing to the use of a continuous procedure. In Proceedings of the Second Engineering Foundation Con- ference on Fundamental of Adsorption. A. I. Liapis, ed. New York: AIChE. Pub,, 473-482.

Rouquerol, J. 1989. Controlled transformation rate thermal analysis: The hidden face of the thermal analysis. Ther- modynamica Acta 144: 209-224.

Suquet, H., R. Prost, and H. Pezerat. 1982. Etude par spec- trom&rie infrarouge et diffraction des rayons X des inter- actions cations-feuillets dans les phases ~ 14.6, 12.2 et 10. l

d 'une saponite-Li de synth~se. Clay Miner. 17: 225-235. Schramm, L. I., and J. C. T. Kwak. 1982. Influence of ex-

changeable cation composition on the size and shape of montmorillonite particles in dilute suspension. Clays & Clay Miner. 30: 40-48.

Sing, K. S .W. 1982. Reporting physical data for gas/solid systems with special reference to the determination of surface area and microporosity. Pure and Applied Chem. 54: 2201-2218.

Sposito, G . , andR . Prost. 1982. Structure ofwater adsorbed on smectites. Chemical Reviews 82: 554-573.

Stul, M. S., and L Van Leemput. 1982. The texture of montmorillonites as influenced by the exchangeable inor- ganic cation and the drying method. I. External surface area related to the stacking units of the aggregates. Surface Tech- nology 16: 89-100.


Suquet, H., and H. Pezerat. 1987. Parameters influencing layer stacking types in saponite and vermiculite: A review. Clays & Clay Miner. 35: 353-362.

Tarasevitch, J. I., and F. D. Ovcharenko. 1975. Adsorbtisiya na glinitykh mineralakh Naukova Dumka, Kiev, 352 pp. (in Russian).

Tessier, D. 1984. Etude exp6rimentale de l 'organisation des mat6riaux argileux. Th6se doct. Universit6 Paris VI, 361 pp.

Veniale, F. ,M. Setti, andM. Tortelli. 1986. Stabilizzazione- consolidamento di terreni argillosi mediante diffusione di

sali. In XVI Convegno Nazionale di Geotecnica, Bologna, 1986. Associazione geotecnica italiana. 387-396.

Whalley, W. R., and C. E. Mullins. 1991. Effect of saturating cation on tactoid size distribution in bentonite suspensions. Clay Miner. 26:11-17.

Yvon, J., J. Baudracco, J. M. Cases, J. Weiss. 1990. El& ments de minrmlogie quantitative en microanalyse des argiles. In Materiaux argileux, structure, proprt~t~s et appli- cations. Decarreau, ed. Paris: SFMCGFA, 473--488.

(Received 28 March 1994; accepted 11 July 1994; Ms. 2485)

mechanism of adsorption and desorption of …clays.org/journal/archive/volume...

Documents