MESRI, G. & OLSON, R. E. (1971). Gtotechnique 21, No. 4, 341-352.

CONSOLIDATION CHARACTERISTICS OF MONTMORILLONITE

GHOLAMREZA MESRI* and ROY E. OLSON?

SYNOPSIS
Artificially sedimented specimens of Wyoming On a utilise des Cchantillons artificiellement sedi-
bentonite (Volclay) were used to investigate the ment&s de bentonite du Wyoming (Volclay) pour
etudier l’effet des variables physico-chimiques sur
effect of physico-chemical variables on the consoli-
les proprietes de consolidation de la montmoril-
dation properties of montmorillonite. One-dimen- lonite. On a fait des essais de consolidation uni-
sional consolidation tests were performed on calcium dimensionnelle sur les formes calcium et sodium de
and sodium montmorillonite at various electrolyte la montmorillonite a diverses concentrations d’elec-
concentrations. The influence of pH and organic trolyte. On a Cgalement Ctudi& l’influence du pH
pore fluids on the consolidation properties of mont- et des liquides de pore organiques sur les proprietes
de consolidation de la montmorillonite.
morillonite was also investigated. Les courbes de consolidation de la forme calcium
The consolidation curves of calcium mont- de la montmorillonite sont essentiellement inde-
morillonite were essentially independent of both the pendantes de la concentration d’klectrolyte dans
electrolyte concentration in the free pore water and l’eau libre des pores et du pH de la suspension a
the pH of the suspension from which the specimens partir de laquelle on a fait sedimenter les cchantil-
were sedimented. The influence of pore water elec- Ions. L’influence de la concentration en t-lectrolyte
trolyte concentration on the swelling curves of cal- dans l’eau des pores sur les courbes de gonflement
cium montmorillonite was much smaller than would de la forme calcium de la montmorillonite est beau-
coup plus faible que ne permet de le prevoir la
be predicted from the Gouy-Chapman double layer thcorie de la double couche de Gouy-Chapman.
theory. The consolidation and swelling curves of Les courbes de consolidation et de gonflement de la
sodium montmorillonite were affected by the con- forme sodium de la montmorillonite sont influencees
centration of electrolyte in free pore water and there par la concentration en electrolyte dans l’eau libre
was reasonable agreement between the measured des pores et il y a un assez bon accord entre les
swelling curves and the double layer theory. Con- courbes de gonflement mesurees et la theorie de la
solidation tests in which organic pore fluids were used double couche. Des essais de consolidation effec-
demonstrated that swelling of montmorillonite is tues en utilisant des liquides de pore organiques ont
d6montrC que le gonflement de la montmorillonite
caused by either adsorption of the pore fluid or
est du soit a l’adsorption du liquide de pore, soit a la
formation of diffuse double layers. formation de doubles couches diffuses.
The coefficient of permeability of the sodium Le coefficient de permcabilite de la forme sodium
montmorillonite in water was as low as 2 x IO-l2 de la montmorillonite est aussi bas que 2 x IO-la
cm/s. The coefficients of permeability were higher cm/s. Les coefficients de permeabilite sont plus
for the montmorillonite in ethanol than in water, Cleves pour le systeme montmorillonite Cthanol que
pour le systeme montmorillonite eau, et on a obtenu
and considerably higher values were obtained when
des valeurs considkrablement plus &levees lorsque le
the pore fluid was non-polar, all compared at the liquide de pore Ctait non-polaire, toutes ces mesures
same void ratio. etant ramenkes a la m&me fraction de vide.

INTRODUCTION
A major part of the land area of Western Canada and Western United States is underlain
at shallow depths by highly plastic clay shales or by soils derived from the weathering of these
shales. These soils are generally characterized by relatively large volume changes when the
effective stresses are changed, and by relatively low effective angles of internal friction. The
source of these undesirable engineering properties is generally agreed to be the group of clay
minerals with expanding lattices, known as the montmorillonite group.
An experimental and theoretical investigation of the engineering properties of soils com-
posed of montmorillonite has been in progress at the University of Illinois at Urbana since
1963. It has involved both consolidation and shear testing of homo-ionic specimens of the
mineral montmorillonite (Mesri, 1969; Mesri and Olson, 1970). The experiments were per-
formed using sedimented specimens of either sodium or calcium montmorillonite, at a range of

* Department of Civil Engineering, University of Illinois, Urbana.

t Department of Civil Engineering, University of Texas at Austin,
341
342 G. MESRI AND R. E. OLSOK

Loading

/‘“p drainage

Fig. 1. One-dimen-
/ \ sional consolidation
Porous stone
Bottom drainage cell

values of pH and pore water electrolyte concentrations, and in various pore fluids. The pur-
pose of these experiments was to help clarify the influence of physico-chemical factors on the
engineering properties of highly plastic clays. The consolidation characteristics of the clay are
considered in this Paper.

DESCRIPTION OF CLAY AND SAMPLE PREPARATION

Clay
The clay used in this investigation was unfractionated Wyoming bentonite, known com-
mercially as Volclay, Micron Sized Grade BC, supplied by the American Colloid Company.
The soil is believed to contain 90-95% montmorillonite and amorphous material, with the
remainder mostly feldspar. A sedimentation analysis showed that all of the clay was finer
than 20 pm and 96% was finer than 0.5 pm.

Preparation of homo-ionic clay in water

Homo-ionic calcium and sodium montmorillonite were prepared by repeated washings with
1 N solutions of the chloride of the appropriate cation, followed by washings with dis-
tilled water, using a Sharples super-centrifuge when necessary. In all cases the purified clay
was more than 96% homo-ionic. The clay was prepared in batches of two thousand grammes
each. The chemical treatment required from two to eighteen months per batch.

Preparation of clay i% organic pore j&&i

Air dry montmorillonite was washed in ethyl alcohol, centrifuged to remove excess alcohol,
and dried at 110°C for twenty-four hours. The dry clay was cooled under vacuum in a
desiccator and mixed directly with the organic fluids.

DESCRIPTION OF APPARATUS AND TESTING PROCEDURES

Specimens containing water as a pore fluid were sedimented from suspensions with liquidity
indices of five. The suspensions were de-aired under vacuum for a week and then poured into
2.5 in. dia. x 12 in. high plastic consolidation tubes. The suspensions were consolidated to a
pressure of 90-125 lb/sq. ft, in increments, and then extruded directly into carefully aligned
Teflonl-lined consolidation rings of the same inside diameter, and trimmed to a height of one
inch.
1 Registeredtrade mark of the E. I. du Pont de Nemours Company.
 CONSOLIDATION CHARACTERISTICS OF MONTMORILLONITE 343
The rings were then mounted in the consolidation cell and placed in a loading frame for the
consolidation test. The applied pressures ranged from 35 to 64 000 lb/sq. ft. For virgin
consolidation, the pressure was doubled for successive loads; for swelling, the load was reduced
to one-fourth of its previous value. Loads were left in place until primary consolidation was
complete and the rate of secondary consolidation was very small.
Special care was taken to reduce experimental problems to an acceptable minimum. Both
the inside of the consolidation ring and the outside of the loading cap were lined with Teflon
(Fig. 1) to reduce side friction. The specimens were separated from the porous stones by
single thicknesses of Whatman No. 50 filter paper to prevent extrusion of the soil into the
stones and to eliminate the problem of partially draining boundaries (Newland and Allely,
1960). The rings were sealed in such a way as to eliminate evaporation from the top, and both
the upper and lower surfaces of the soil were connected to a reservoir containing the same fluid
as the pore fluid of the specimen.

INDEX PROPERTIES

The Atterberg limits of the calcium and sodium montmorillonite in water were determined
using standard techniques (American Society for Testing and Materials, 1964) with the excep-
tion that the water content was defined as the ratio of the weight of the pore fluid (water plus
dissolved salts) to the weight of the solids. Atterberg limits of montmorillonite as functions
of the adsorbed cations, electrolyte concentration in the free pore water, and pH are given in
Table 1. In agreement with previous observations (Samuels, 1950; White, 1958; Warkentin,
1961), the valency of the adsorbed cation has a pronounced effect on the liquid limit; the
sodium montmorillonite has a liquid limit about five times that of the calcium montmorillonite.
For the sodium montmorillonite, the influence of electrolyte concentration on the liquid limit
is in qualitative agreement with double layer theory; when the concentration of electrolyte is
high the double layers are depressed and the liquid limit is low. The plastic limits were not
determined because the clay was too hard to rule out. One plastic limit test was performed on
a 0.05 N sodium montmorillonite, and a value of 47% was obtained, but the problems of
performing the test were such that this should be considered as only a reasonable approxi-
mation.
The liquid limits of the calcium montmorillonite at pH 5 are qualitatively in accordance
with double layer predictions but the data at pH 9 are not. In any case, the effects of electro-
lyte concentration on the limits of calcium montmorillonite are small.
The specific gravity values (Table 1) for the sodium and calcium montmorillonite are each
the mean of four tests; no single determination differed from the mean by more than 0.01.
Two measurements in carbon tetrachloride yielded a value of 2.78.
The specific surface was estimated using the glycerol retention method of Diamond and
Kinter (1958). A possible explanation for the lower surface area for the sodium montmorillo-
nite may be that some of the finest fraction of the soil was lost during centrifugation.

Table 1. Index properties

N 0~0001 0.001 0.01 0.1 1.0 G, S,

m21g
Cation pH LL PL LL I PL LL 1 PL LL ’ PL LL / PL

~_---___~__-___
5 - - / 220 33 200 33 1 193 31 189 32 2,80 680

9 - - 207 35 205 35 207 35 201 34

--- -I-! -- ---_

Sodium 7 1160 - 1140 - II960 - 880 - / - - 24s5 5oo

N = normality, LL=liquid limit, PL = plastic limit, G,= specificgravity of solids, S = specific surface.
344 G. MESRI AND R. E. OLSON

Fig. 2. One-dimen-
sional consolidation
curves for calcium
montxnorilloni~ at
Consolidation pressure:lb/s+ ft 0.001 N CaCl,

Fig. 3. One-dimen-
sional consolidation
curves for calcium
montznorillonite
Consolidation pressure: lb/t+ R (PH=5)
 CONSOLIDATION CHARACTERISTICS OF MONTMORILLONITE 345

Fig. 4. One-dimen-
sional consolidation I
curves for calcium IO IO 3 IO 4
montmorillonite Consolidation pressure: Ib/rq. ft
(PH =7)

,--

I-

I-

Fig. 5. One-dimen-
sional consolidation
,

curves for calcium IO‘ IO 2 IO’

montmorillonite
Consolidation pressure: Ib/sq. ft
(pH=V
 G. MESRI AND R. E. OLSOS

00

60

Fig. 6. Theoreti-

--!-io
cal and experi-
mental swelling
?-alu 2o curves for cal-
cium montmo-
\ riuonite (pH =7)
Cation exchangecapacity.
o=1OOm.eq./1OOg
I I - Specific surface,
IO2 I IO5 s=500 ma/g
IO 103
Temperature.
Swelling pressure: Ib/sq. ft T-298 A

35
Electrolyte cokntration

; ;:@NN
0 0.01 N
30 PO.1 N

15

2 20
z_
P-
3
15

10

5
Fig. 7. One-
dimensional con-
solidation curves
0 for sodium mont-
morillonite
Consolidation pressure: Ib/rq. It (PH =7)
 CONSOLIDATION CHARACTERISTICS OF MONTMORILLONITE 347
12 I . \I \\ I 1180

4 I 60
Fig. 8. Theoreti-
cal and experi-
mental swelling
curvee for sodium
montmorillonite
(pH=7)
Cation exchange capacity,
w=l00m.eq./1OOg
Specific surface.
S=500 ma/g IO 102 103 KY4 IOS
Temperature,
T=298 A Swelling
pressure:ib/sq. fc

ONE-DIMENSIONAL CONSOLIDATION CURVES

Void ratio-effective stress relationships for 0.001 N calcium montmorillonite at pH values
of 5, 7 and 9 are shown in Fig. 2. The specimens at pH 5 and 9 were consolidated to
90 lb/sq. ft in the sedimentation tubes and allowed little time for secondary consolidation,
whereas the specimens at pH 7 were completely consolidated under a pressure of 125 lb/sq. ft.
Thus, the differences between the void ratios for pressures smaller than 150 lb/sq. ft resulted
mainly from minor variations in the procedures for sample preparation. Low values of pH are
normally expected to cause edge-to-face flocculation and higher void ratios. However, the
particles of montmorillonite are of the order of 10 A thick and have diameter-to-thickness
ratios of the order of 150-500 (Kahn, 1959; van Olphen, 1956) and thus may be considered more
like thin films than like stiff plates. Thus, even though the edges of the films may be attracted
to adjacent faces when the pH is 5, the dominant interaction is still face-to-face. The pH may
influence the viscous characteristics of dilute suspensions but it does not seem to influence the
consolidation curve for pressures greater than about 100 lb/sq. ft.
The influence of the electrolyte concentration on the consolidation curve of calcium mont-
morillonite is shown in Figs 3-5. The virgin curve is independent of electrolyte concentra-
tion within the range of experimental accuracy. The swelling curves are in the relative
positions predicted using double layer theory. However, the experimental swelling curves
and the theoretical curves, calculated using the classical double layer theory (Verwey and
Overbeek, 1948; van Olphen, 1963), do not compare favourably (Fig. 6). The most probable
explanation of the wide difference between the theoretical and the experimental curves is
irreversible domain formation (Aylmore and Quirk, 1959, 1962) where the particles do not
swell to spacing beyond about 19 A after being squeezed once to a closer spacing by the
application of high consolidation pressures (MacEwan, 1945, 1954).
Consolidation curves for sodium montmorillonite at a pH of 7 are shown in Fig. 7. The
positions of both the compression and the swelling curves are dependent on electrolyte con-
centration, suggesting a pronounced diffuse double layer effect. The experimental swelling
curves and the theoretical curves calculated using classical double layer theory are compared
in Fig. 8. The classical theory apparently predicts swelling of the right order of magnitude
 G. MESRI AND R. E. OLSON

Carbon tetrachlorid

Fig. 9. One-
dimensional con-
solidation curves
for montmorillo-
nite in organic
Consolidation pressure: Ib/sq. fc fluids

but at low pressures the specimens again seem to establish equilibrium at a void ratio smaller
than predicted by the theory, again suggesting some sort of potential minimum such as en-
countered for the calcium clay.
To check on the overall application of physico-chemical models to the consolidation
characteristics of montmorillonite, one-dimensional consolidation tests were performed on
specimens of montmorillonite in various organic pore fluids. The soil was consolidated from
slurries. The consolidation curves for specimens in ethanol, carbon tetrachloride and benzene
are shown in Fig. 9. The large hysteresis loop between the virgin and swelling curves indicates
that particle rearrangement is the major source of virgin compression. The very flat swelling
curves indicate that mechanical swelling is a minor component of swelling of montmorillonite
(Olson and Mesri, 1970). Swelling is apparently the result of adsorption of fluid and formation
of diffuse double layers, albeit not in conformance with classical theory.

COEFFICIENTS OF PERMEABILITY

The coefficients of permeability were calculated by fitting Terzaghi’s theory of consolidation

to the measured deformation versus log time curves (Casagrande and Fadum, 1940). These
calculated values are probably slightly lower than the actual values but the difference is
unlikely to exceed perhaps twenty per cent because the loading ratio was two during virgin
consolidation and the amount of secondary consolidation under each load was a very small
fraction of the total settlement.
The relationship between the coefficient of permeability and void ratio for the calcium
montmorillonite was independent of pH value (Fig. 10). The influence of the electrolyte con-
centration on the permeability of calcium montmorillonite is shown in Figs 11-13. The
coefficients of permeability are consistently lower when the electrolyte concentration in the
pore water is lower, an observation that suggests that water flow was retarded by the develop-
ment of diffuse double layers, even though the double layers were too weak to influence the
 CONSOLIDATION CHARACTERISTICS OF MONTMORILLONITE 349

Fig. 10. Permeabi-

lity-voidratiocurves
for calcium mont-
morillonite at 0.001
N CaCl, Permeability: cm/r

Fig. ll.gPermeabi-
lity-voidratiocurves
for calcium mont-
morillonite
(pH=5) Permeability: cm/r

Fig. 12. Permeabi-

lity-void ratio curves
for calcium mont-
morillonite
(PI-I=71 Permeability: cmIr
350 G. MESRI AND R. E. OLSON

virgin consolidation curves significantly. The relationship between void ratio and per-
meability for sodium montmorillonite at pH 7 is shown in Fig. 14. At the higher void ratios
the diffuse double layers seem to exert a small influence on water flow, as for calcium mont-
morillonite, but the effect is not large.
The coefficients of permeability for montmorillonite in ethanol, carbon tetrachloride and
benzene are compared with the values previously considered for sodium and calcium mont-
morillonite in water, in Fig. 15, where a log-log scale is used. The most striking feature of
this comparison is the large range in the coefficient of permeability at a given void ratio, of the
order of lo6 times for a void ratio of two. The explanation for this wide range is as follows: the
montmorillonite in non-polar pore fluids, such as carbon tetrachloride and benzene, is floccu-
lated into very tight, more-or-less spherical domains (Mesri and Olson, 1971), leaving large
pores through which the fluid can flow. These pores are not constricted either by the pre-
sence of adsorbed layers of fluid molecules or by the presence of diffuse double layers, and are
thus completely open for fluid flow. The low coefficient of viscosity of the fluids promotes
even more rapid flow. In the case of ethanol, the fluid is adsorbed onto the clay surfaces, thus
separating the individual sheets of clay and preventing the formation of the tight domains
encountered with the non-polar pore fluids. Fewer channels are likely to exist when the pore
fluid is alcohol, and the channels are constricted by the presence of adsorbed layers, but no
diffuse double layers form. When water is the pore fluid, further dispersion occurs, with
fewer channels available for flow, and with the available channels under the influence of
diffuse double layers containing hydrated ions.
The reduction in the coefficient of permeability as the void ratio is reduced results from the
reduced amount of void space for flow, the greater influence of the diffuse double layers and
adsorption of fluid on the available void space, and probably a re-orientation of the particles
perpendicular to the direction of fluid flow (Quigley and Thompson, 1966) thus increasing the
tortuosity factor. The convergence of the curves for sodium and calcium montmorillonite in
water, at low void ratios, indicates that the particles are being compressed into such close
proximity that all of the void space is clogged by the presence of large concentrations of
hydrated cations.

SUMMARY AND CONCLUSIONS

One-dimensional consolidation tests were performed using saturated specimens of sodium

and calcium montmorillonite. All specimens were consolidated from slurries. Tests with
calcium montmorillonite at a range of pH from 5 to 9 showed that tendencies towards edge-to-
face flocculation at low pH, if they occurred at all, did not influence the consolidation charac-
teristics. The filmy nature of the particles would lead to a dominant face-to-face interaction
regardless of any tendency towards attraction between edges and faces.
Variations in pore water electrolyte concentration had little effect on the void ratio-
effective stress relationships for the calcium montmorillonite in water, apparently because of
double layer effects smaller than predicted by classical theory and because of the formation of
permanent domains. The consolidation curves of sodium montmorillonite in water were in
qualitative conformance with double layer theory, but even in this case there was evidence of
domain formation during swelling at low values of effective stress.
Tests with montmorillonite in ethanol yielded much lower void ratios than for the mont-
morillonite in water, apparently as a result of destruction of the diffuse double layers. Even
lower void ratios were obtained for montmorillonite in non-polar fluids. Further, the amount
of swelling of montmorillonite in the non-polar fluids was only about one per cent of the
swelling of the sodium montmorillonite in water.
The coefficients of permeability of the montmorillonite in water were extremely low, with a
minimum value of about 2 x lo-l2 cm/s. These values were only slightly influenced by the
 CONSOLIDATION CHARACTERISTICS OF MONTMORILLONITE 351

Fig. 13. Permeabi-

lity-voidratio curves
for calcium mont- I
10-I’ IO_‘0 10-g 10-a 10-7 10-6
morillonite
(PH =3) Permeability: cm/n

35

30

25

.g 20

e
’z
15

IO

Fig. 14. Permeabi-

lity-voidratio curves
0
for sodium mont- 10-12 -II
10 IO_‘0 lo-’ 10-a I(
morillonite
(PH =7) Permeability: cm/s

A Water(calcium montmorillonlte)

Fig. 15. Coeffhients

of permeability of
montmorillonite in
various fluids Permeability: cm/s
352 G. MESRI AND R. E. OLSON

pore water electrolyte concentration but were greatly influenced by the valency of the ions.
The sodium montmorillonite was much less permeable than the calcium montmorillonite.
Higher coefficients of permeability were calculated for the montmorillonite in ethanol than
in water, and even higher values were obtained when the pore fluid was non-polar, all compared
at the same void ratio. The wide range in coefficients of permeability, from about 2 x lo-l2
to 2 x lOA cm/s can be explained, qualitatively, in terms of diffuse double layers, fluid ad-
sorption, and dispersion or aggregation of clay particles.

ACKNOWLEDGEMENTS
The experiments reported in this Paper were performed in the soil mechanics laboratories
of the University of Illinois at Urbana, during the period from 1963 to 1969. They represent
part of the research performed under Grants GP325 and GK531 from the National Science
Foundation.

REFERENCES
AMERICAN SOCIETY FOR TESTING AND MATERIALS (1964). Procedures for testing soils. Philadelphia,
Pa: American Society for Testing and Materials.
AYLMORE, L. A. G. & QUIRK, J. P. (1959). Swelling of clay-water system. Nature, Lond. 183, 1752-1753.
AYLMORE, L. A. G. & QUIRK, J. P. (1962). The structural status of clay systems. CZays CZay Miner. Proc.
9th Nat. Conf., pp. 104-130. New York: Pergamon Press.
CASAGRANDE,A. & FADUM, R. E. (1940). Notes on soil testing for engineering pm-poses. Soil Mechanics
Series No. 8, Harvard Graduate School of Engineering.
DIAIMOND,S. & KINTER, E. B. (1958). Surface areas of clay minerals as derived from measurements of
glycerol retention. Clays Clay Miner. Proc. 5th Nat. Conf., Nat. Acad. Sci., Nat. Res. Council Publ.
566, pp. 334-47.
KAHN, A. (1959). Studies on the size and shape of clay particles in aqueous suspension. Clays Clay Miner.
Proc. 6th Nat. Conf., pp. 220-36. New York: Pergamon Press.
MACEWAN, D. M. C. (1948). Adsorption by montmorillonite and its relation to surface adsorption. Nature,
Lond. 162, 935-936.
MACE~AN, D. M. C. (1954). Short range electrical force between charged colloid particles. Nature, Lond.
174, 39-40.
MESRI, G. (1969). Engineering firoperties of montmoriZZonite. Ph.D. thesis, University of Illinois.
MESRI, G. & OLSON, R. E. (1970). Shear strength of montmorillonite. Gdotechnique 20, No. 3, 261-270.
MESRI, G. & OLSON, R. E. (1971). Mechanisms controlling the permeability of clays. CZuys Clay Miner.
19, No. 3, 151-158. London: Pergamon Press.
NEWLAND, P. L. & ALLELY, B. H. (1960). A study of the consolidation characteristics of a clay. G&o-
technique 10, No. 2, 62-74.
OLSON, R. E. & MESRI, G. (1970). Mechanisms controlling the compressibility of clays. JnZ Soil Mech.
Fdns Div. Am. Sot. Civ. Engrs 96, No. SM6, 1863-1878.
QUIGLEY, R. M. & THOMPSON,C. D. (1966). The fabric of anisotropically consolidated sensitive marine clay.
Can. Geotech. JnZ 3, 61-73.
SAMUELS, S. G. (1950). The effect of base exchange on the engineering properties of clays. Note C-176,28 pp.
Garston, England: Building Research Station, Dept. of Sci. and Ind. Research.
VAN OLPHEN, H. (1956). Forces between suspended bentonite particles. Clays Clay Miner. Proc. 4th Nat.
Conf., Nat. Acad. of Sci., Nat. Res. Council, Publ. 456, pp. 204-224.
VAN OLPHEN, H. (1963). An introduction to colloid chemistry. h’ew York: Inter-science.
VERWEY, E. J. W. & OVERBEEK, J. Th. G. (1948). Theory of the stability of Zyophobic coZZoids. New York:
Elsevier Publishing Company Incorporated.
WARKESTIN, B. P. (1961). Interpretation of the upper plastic limit of clays. Nature. Lond. 190,287-288.
WHITE, W. A. (1958). Water sorption properties of homoionic clay minerals. Illinois State Geological
Survey, Report of Investigation No. 208, 46 pp.