Mechanisms of Soil-Lime Stabilization

An Interpretive Review
SIDNEY DIAMOND and EARL B. KINTER
Materials Division, Office of Research and Development,
U. S. Bureau of Public Roads

Literature on the mechanisms of the physicochemical and chem-

cal reactions responsible for soil-lime stabilization is critically
reviewed. The importance of cation exchange, flocculation, and
carbonation in the stabilization processes is discounted. The
rapid "ameliorating" effects of lime on the properties of plastic
soils are documented from the literature, and the distinction
between these short-time effects and the long-time cementitious
reactions responsible for the development of strength is noted.
The cation crowding hypothesis is rejected as an explanation
of the ameliorating effects. New results summarized show that
the cation crowding is more accurately described as physical
adsorption of calcium hydroxide on the clay surfaces. The
ameliorating effects may be due to very rapid cementation of
the primary clay particles at points of contact within the floes
formed on the addition of lime. This rapid cementation is
attributed to immediate formation of tetracalcium aluminate
hydrate by reaction of Al(OH)x groups at the edges of the clay
particles with adsorbed calcium hydroxide on the faces of ad-
jacent planar surfaces. This may be supplemented by some-
what slower formation of tobermorite gel at the same reaction
sites.
The nature and properties of the reaction products respon-
sible for slow development of strength in soil-lime systems are
discussed. In general, the products formed are usually tober-
morite gel and a calcium aluminate hydrate phase distinguished
by a 7. 6 A bas al spacing that is stable to moder ate drying;
other compounds formed under special circumstances include
tricalcium aluminate hexahydrate and calcium silicate hydrate
(I). Phases involving both silica and alumina in distinguishable
lattice positions are not usually formed.
The scant information in the literature on the chemical
mechanisms involved in the lime-clay mineral transforma-
tions is reviewed. Available literature would tend to suggest a
"through-solution" mechanism dependent on the rate of dis-
solution of the clay (or of silica and alumina from the clay),
since it is implied that the reaction depends on attainment of a
high pH and since addition of sodium hydroxide accelerates
strength development. On the other hand, evidence that the
reaction proceeds from an initial state of physical adsorption
of lime on the surface of the clay would suggest a surface chem-
ical reaction not dependent on prior dissolution of material from
the clay.

Paper sponsored by Committee on Lime and Lime-Fly Ash S~abilization and presented at the
44th Annual Meeting.
84

•SOIL STABILIZATION by the addition of quicklime or hydrated lime is an ancient

art, successfully adapted in modern times by perceptive engineers of the Texas High-
way Department and now widely practiced. It is not, however, the purpose of this
report to discuss either the history or the engineering aspects of soil-lime stabilization
as currently practiced, but to attempt to bring together and critically evaluate current
knowledge of the mechanism or mechanisms responsible for the stabilization process.
Once these are well understood, it should be possible to develop practical stabilization
prucedureB un a .tnore rational basis, so as to take full advantage of this useful and cheap
chemical agent.
The responses of soil to treatment with lime are complex and often dramatic. A
number of explanations have been proposed to account for these unusual responses,
including:
1. Cation exchange, Le., replacement of the exchangeable sodium, magnesium, or
other cations previously held by the soil clay by calcium cations derived from the lime;
2. Flocculation of the clay, and consequent increase in effective grain size;
3. Carbonation, i. e;, reaction of the lime with carbon dioxide from the atmosphere
to form calcium carbo~ate, which has been said to exert cementing action; and
4. So-called pozzol attic reactions with soil constituents to generate new minerals
of a cementitious natu"re .
Of these phenomena, the first three may occur but are inadequate explanations, as
we shall attempt to show. ·The latter explanation is at least partially correct , and in
this paper a more specific picture of the nature of these reactions is presented.
The factor of cation e:irohange has been mentioned by many authors, yet familiarity
with the cation exchange properties of soils should have eliminated this as a serious
explanation Ior the stabilizing effects oI lilne on soil. As soil scientists know, 111any
natural soils are largely calcium saturated. ·For example, recent work shows that
the montmorillonitic soils of the southwestern United States (Grumosols in soil science
classification systems); are normally two-thirds to three- quarters calcium saturated
(1). Despite this predominance of calcium in the exchange compl ex, these soils exhibit
all the classic deficiencies associated with montmorillonite soils when used as sub-
grades and, thus, require stabilizat,ion. Although it may be suggested that complete,
rather tha11 predomina...~t, calcium saturation is required for stabilization, it has been
demonstrated that even when lime is added in very large excess to dilute clay suspen-
sions, under conditions where cation movement is facilitated, complete exchange for
sodium does not take plac~ (~). Exchange in comparatively dry compacted soils is
undoubtedly less complete.
The concept that flocculation plays a major part in soil-lime stabilization is often
voiced, but careful examination of previously known facts shows that this also is an
inadequate explanation. Some soils, including most of the red and yellow soils of the
southeastern part of the United States, are flocculated in nature, as can be seen by
shaking the soil in water and examining the resulting floes. Despite this natural
flocculation these soils are not stable, and they do respond to lime treatment. F.urther-
more, it is well known that many chemical agents, including various salts, alcohols,
acids , and ketones, induce immediate flocculation when mixed with clays , yet are
valueless for stabilization. The fact that flocculation of clay occurs as a consequence
of the addition of lime is a well-known phenomenon, but the achievement of flocculation
is clearly not the mechanism by which lime stabilizes soils.
The hypothesis that soil-lime stabilization depends on the carbonation of the lime
to form calcium carbonate can be dismissed by reference to any of a great number of
studies in which reaction with atmospheric carbon dioxide was precluded by sealing
the samples and in which the characteristic modification of properties and development
of strength associated with lime stabilization were observed. As demonstrated by
Eades, Nichols, and Grim (3), carbonation does take place in the field, but the strength
gain said to accrue by virtue of cementation of soil grains by calcium carbonate has
yet to be conclusively demonstrated. It appears that the additional benefit by long-term
reaction of uncarbonated lime with the soil itself would far outweigh any such contribu-
tion, and carbonation is probably' a deleterious rather than a helpful phenomenon in
soil stabilization.
 85

Having rejected these inadequate hypotheses, we may now turn our attention to the
more nea,rly adequate, but still incomplete, explanations of the mechanisms of lime-
soil interaction that constitute the subject of this paper. Experience indicates that
there are at least two distinct stages of reaction involved: (a) the immediate or rapid
processes responsible for the "amelioration" of the water-sensitive properties of
untreated clay soil, and (b) the slower, long-term reactions resulting in formation of
the final cementitious products that are indicated by the gradual development of strength
in compacted soil-lime mixtures. These will be discussed separately.

RAPID AMELIORATION EFFECTS

Atterberg Limits
It has been shown by many workers (4, 5, 6) that addition of lime to plastic soils
results in reduction of the plasticity index (P. -1. ) . Usually an immedi at·e increase in
the plastic limit (P. L.) is observed on the addition of lime. The amount of this in-
crease varies directly with the amount of lime added, up to some limiting lime content;
further increments of lime usually bring little or no additional increase. The point of
inflection of the plot of lime added vs P. L. has been dignified by the appellation "lime
fixation point" (7). The implications involved in this term will be discussed later.
The effect 01-lthe liquid limit (L. L.) is less easy to summarize, since the published
data conflict. Some authors report decreases in L . L. on addition of lime ( 4, 5), but
others report that this par:l.meter may increase substantially ( 8, 9). Still other workers
(6, 10, 11) report that both increases and decreases occur, depending on the individual
soilbeing tested.
These discrepancies require some explanation. It should be pointed out that the
L. L. of a clay is far more sensitive to the kind of cation present than is the P. L.
Calcium-saturated clays have substantially lower L. L. than the same clays saturated
with sodium or certain other cations (12). Since some (although not complete) cation
exchange occurs on addition of lime, the effect of the lime may vary, depending on the
extent t o which preexisting cations other than calcium are exchanged. Studies on clays
of unknown original cation status are consequently difficult to interpret.
Clare and Cruchley (8) give data showing a drastic increase in L. L. with added
lime for a particular clay which was almost completely calcium saturated in its natural
state . They also state that these data are typical for the other clays tested as well;
the other clays varied in original calcium saturation from 35 percent upward. We, as
well as many others, have observed that mixtures of lime and calcium-saturated clay,
stored so as to prevent evaporation or carbonation, assume an increasingly dry appear-
ance; in extreme cases, a free-flowing slurry mixture may stiffen so that it will no
longer pour or take the shape of the container. This behavior would, of course, be
reflect ed in an increase in L. L. It seems that, in general, the L. L. will increase
on the addition of lime in the absence of a strong specific tendency in the opposite direc-
tion resulting from cation exchange effects.
When the L. L. increases on the addition of lime, the increase is not usually as
great as the accompanying increase in P. L. Thus, the separate effects on L. L. and
P. L. usually combine to yield a rather sharp decrease in P. I. Extra increments of
lime beyond some fairly small percentage do not usually result in significant further
decreases; in fact, sometimes the trend is reversed and incremental increases in
P. I. are found. It should be kept in mind that the P. I. is a composite parameter
that may not accurately reflect real changes in the system; for example, equal increases
in L. L . and P. L. following a given treatment would be reflected as "no change" in
the P. I.
Aging of the system generally results in further changes. Sometimes this effect is
small. Lund and Ramsey ( 6), for example, found that very little change occurred in
the Atterberg limit values Of soil-lime mixtures after the first hour. However, often
substantial time-dependent changes ensue. For example, Wolfe and Allen (13) report
substantial increases in P. I. for a number of lime-soil mixtures when cured for 2
days, as compared with samples tested immediately after the addition of lime. For
more extended curing periods (7 to 28 days) the effect was reversed, and significant
decreases in P. I. in most cases to a non plastic condition, were recorded.
86

Grain-Size Distribution
Changes in the effective grain size distribution (mechanical analysis) are observed
almost immediately following the addition of lime to a clay soil. Data secured by con-
ventional sieve and hydrometer analysis show that a major decrease in the content of
clay-size particles occurs within the first hour; subsequent results after periods
ranging up to 240 days show only small additional changes(_§). The new "grains"
produced as a result of the lime treatment are mostly sand sized and, although re-
latively weakly Uunded, are fo1· the n1ost pai-t capable of withstanding a 5-n1in dispe1:sion
period in a mechanical mixer (.§_, 13).

Moisture-Density Relations
It has been reported by many workers that the density to which a soil can be com-
pacted at a given moisture content is usually reduced significantly on addition of lime,
and delay in compaction results in further reduction. The moisture content needed to
achieve maximum density for a given compactive effort usually increases, sometimes
rather significantly. However, according to results summarized by Herrin and
Mitchell (14), lime in excess of a relatively limited amount, on the order of 5 percent
by weight of soil, generally produces little additional increase in the optimum moisture
requirement.

Volume Change Effects

Lund and Ramsey (6) report a drastic reduction in the volume change. This effect
was observed within the first hour after the addition of lime; with the soil tested, about
3 percent lime sufficed to give the maximum effect. A corresponding immediate in-
crease was noted on the shrinkage limit itself. Similar results were reported by Wolfe
and Allen (13).
Mitchelland Hooper (15) reported that lime (in this case, dolomitic lime) markedly
reduced the swelling of specimens tested under a modest surcharge pressure. The
effect was observed after an aging period of 24 hr. A reduction in expansion pressure
on addition of lime was observed by Wolfe and Allen for some soils, but not for others
(13).

Soil Suction
Clare and Cruchley ( 8) report significant increases in the amount of moisture
held against a given suction following the addition of lime.

Permeability
Although few data are available, the permeability of compacted soil-lime mixtures
has been reported to be much less than that of compacted soil alone (13 ).
These various short-term effects can be regarded as different aspects of the ame-
lioration effect of lime on the properties of clay soils. We will attempt to define
exactly what chemical or physico-chemical action is reflected in these sometimes
drastic changes.
Considerable evidence seems to point to the conclusion that these rapid responses
are not the result of the pozzolanic reactions to which permanent strength gains are
attributed. It has been well documented that the development of strength of specimens,
particularly after soaking, is a fair index of the amount of cementitious compound
formed (16). The experience of many workers, as summarized by Herrin and Mitchell
(14), andthe results documented by Anday (17) indicate that strength development in
soil-lime systems is a comparatively slow process characteristically requiring periods
of several weeks to many months at normal temperatures.
As pointed out by Mitchell and Hooper ( 15), if significant amounts of lime were
consumed by permanent cementing reactions at early ages, delay between mixing and
compaction should adversely affect the strength of the final product. They found no
such effect for delays up to a day, provided all samples were compacted to the same
density.
 87

Similarly, if significant permanent cementing were to occur at early ages, subsequent

remolding of the specimen should bring about a strong reduction in the ultimate strength
developed. Unpublished data by Shepard (18) indicate that remolding after as long as
a week of curing following initial compaction has no deleterious effect on the ultimate
strength of soil-lime mixtures. In contrast, Shepard found that soil-cement mixtures,
in which cementation occurs rapidly on hydration of the cement, did suffer severe
reductions in ultimate strength on remolding after a week of curing.
Thus, extensive development of pozzolanic reaction products seems to be ruled out
as a mechanism responsible for the rapid amelioration of the properties of plastic
soils by lime. What then is the nature of the mechanism responsible for these effects?
Extensive investigations undertaken by workers at Iowa State University (7, 19-23)
have led to the previously mentioned concept of the lime fixation point. This-was -
defined as that percentage of lime beyond which additional increments of lime produced
no appreciable increase in P. L. It was hypothesized that excess calcium cations, de-
rived from the lime, in some fashion "crowd onto" the clay particles and cause them
to become "electrically attracted," the process resulting in flocculation with weak
bonds between the floes. Additional lime, which produces calcium cations in excess
of those which could crowd onto the clay, produces no further change in the P. L.
Calcium held by the clay in amounts up to the lime fixation point was considered to
be immune to further reaction with the clay to form cementitious compounds. The
sedimentation velocity of floes formed by adding lime to a clay suspension reached a
maximum near the lime fixation point, indicating that floe size was at a maximum at
this point (21).
Ho and Handy (22) cited evidence that calcium was retained by bentonite from lime-
bentonite slurriesafter a very limited washing treatment. The amount of calcium
retained increased with increasing percentage of lime in the slurry, and for modest
percentages of lime was far more than the normal cation exchange capacity (C. E. C.)
of the clay. These writers hypothesized that the calcium in excess of the normal
C. E. C. was held at new exchange sites generated at the edges of clay particles by
increasing dissociation of acidic -Si(OH)x groups with increasing pH; hence, the term
pH-dependent exchange. However, it was noted (without explanation) that the amount
of calcium retained continued to increase steadily with increments of lime far beyond
the lime fixation point and no break in the curve was evident.
Ho and Handy also studied several additional features of the lime-bentonite system.
Small additions of lime (up to about 2. 2 percent) were found to increase the relative
viscosity of calcium bentonite slurries; however, aging of the samples continually
increased the lime content at which maximum relative viscosity was attained. These
authors also made differential thermal analyses (DTA) of dried lime-bentonite slurries,
and disclosed that samples to which only a small amount of lime had been added failed
to generate the normal endothermic response characteristic of the lime.
In another paper, Ho and Handy (23) reported the results of measurements of the
zeta potential (a quantity related to the net negative or positive charge carried by the
clay particles in dilute suspension) of lime-treated bentonite aged either for several
days ("fresh") or for a year ("aged"). Calcium-saturated bentonite had a zeta potential
of about -20 millivolts. Small additions of lime caused a slight increase in this value
and larger additions generated a modest decrease for both fresh and aged samples.
Sodium-saturated bentonite had a zeta potential of about -40 millivolts; additions of
lime to this clay up to about 6 percent caused strong proportional reductions to - 26
millivolts for the fresh samples and -22 millivolts for the aged samples, and higher
additions of lime caused no further change. Ho and Handy presented a detailed analysis
of their results which we consider highly speculative. It is sufficient to note at this
point that the large effect found for the sodium bentonite was apparently due to the
ordinary process of cation exchange of calcium for sodium, and in accord with the
e<Jrlier results of Prikryl and Esterka (2) this process was not quite complete.
We were far from satisfied that the interpretations of the early ameliorative effects
of lime on clay suggested by the Iowa State University workers could be correct in
detail. The apparently unlimited crowding of positively charged cations onto the limited
external surface of the montmorillonite would appear to be an unlikely phenomenon; if
88

it were to occur, one would expect that a large decrease and eventual reversal of the
negative charge of the clay would take place, due to the limited number of -Si(OH)x
groups on the edges of the clay particles that could dissociate to generate negative
charges to balance the cation uptake.
We carried out a series of investigations in an attempt to shed more light on these
matters. Details of these investigations will be submitted for publication elsewhere,
but some of their results can be summarized here.
Measurements were carried out on Wyoming bentonite which had been saturated
with calcium cations by standard laboratory procedures and checked for completeness
of saturation. Small quantities of the calcium-saturated clay were shaken with a large
amount of saturated calcium hydroxide solution for various periods at room temperature,
and then separated by centrifugation. The supernatant solutions were recovered and
promptly analyzed for calcium content (using versenate titration) and hyd1·oxyl content
(by titration against standard acid). The following results were obtained:
1. A very rapid reduction of the concentration of both calcium and hydroxyl ions in
the solution was observed.
2. The proportion of the ions remaining in solution was stoichiometric, indicating
that equivalent proportions of hydroxyl and calcium ions (two hydroxyl for each calcium)
had been taken up by the clay.
3. About 3 percent of lime (by weight of the clay) was found to be adsorbed within
the first 5 min, the minimum time in which the manipulations could be performed.
This amount corresponds roughly to the lime fixation point of the Iowa State workers.
4. Further sorption continued at a declining rate as long as the experiments were
continued ( 3 wk). At that time the total amount of lime removed from the solution
amounted to about 20 percent of the weight of the clay.
5. \Vashing the centrifugally separated clay with an an1ount of distilled water equal
to the volume of the saturated lime solution originally used resulted in essentially
complete removal of the adsorbed lime, if carried out shortly after the onset of the
experiment. However, the ratio of hydroxyl to calcium ions removed was slightly
less than stoichiometric.
6. After a sample had been in contact with the lime solution for several days and
had adsorbed much additional lime, a single wash as outlined removed only part of
the lime; subsequent additional washings removed smaller and smaller a.111ounts. The
ratio of hydroxyl to calcium ions removed in successive washings declined from values
fairly close to stoichiometric (about 1. 7) to about 1.
7. In DT A of dried clay which had sorbed about 4 percent of lime from solution
(in about an hour of contact), no lime was detected. A mechanical mixture of dry clay
with the same amow1t of dry lime gives a readily detectable endothermic response at
500°, characteristic of crystalline lime.
8. No X-ray diffraction peaks for crystalline calcium hydroxide were observed for
the clay recovered from lime solution a.S mentioned previously; the corresponding dry
mixture produced easily detectable peaks under the same diffraction conditions.
9. Measurements of surface charge of the calcium bentonite were carried out on
untreated clay and on clay recovered after immersion in saturated lime solution for
an hour. The method used was that which Pike and Hubbard used for a similar deter-
mination of the surface charge of hydrating cement particles (24). No significant
change in negative charge was found. -
10. In other experiments, lime was added to suspensions of calcium-saturated
bentonite of restricted water content. At lime contents of 4 percent or less by weight
of clay, the electrical conductivity of the suspension was found to decrease to very low
values in less than 24 hr. This was regarded as an indication that the lime was being
rapidly adsorbed, with the consequent decrease in concentration of the lime electrolyte
in the pore solution. Removal of some of the pore solution by high-speed centrifugation
and subsequent chemical analysis confirmed this interpretation.
These data indicated that the response to lime in clay-water systems of restricted
water content was similar to that occurring in dilute suspensions.
 89

No evidence of heat production or removal was found to accompany the addition of

lime to calcium bentonite suspensions of restricted water content, in tests carried
out over several hours with a moderately sensitive calorimeter. This is a further
indication that the lime sorption process is one of physical adsorption rather than
chemical reaction.
These results are interpreted to yield a picture of the initial reactions somewhat
different from that previously developed elsewhere.
1. Calcium hydroxide is physically adsorbed from solution at a very rapid rate by
calcium-saturated clay, and presumably also by clay saturated with other cations.
This adsorption is largely reversible at very early stages, but is soon followed by
reaction to produce calcium silicate hydrates. The conclusion that lime is physically
adsorbed on clay surfaces was also reached by Prikryl and Esterka ( 2).
2. The adsorption removes calcium ions and hydroxyl ions from solution concur-
rently, and does not reflect a crowding of cations only onto new exchange sites generated
at high pH levels, as previously postulated.
3. Since both ions are sorbed in equivalent amounts (within the limits of the accuracy
of the analysis), no significant change occurs in the net negative charge of the clay
particles. This does not preclude small changes in the zeta potential after some period
of chemical reaction.
4. The amount of calcium and hydroxyl ions sorbed immediately (about 3 percent
by weight of clay) would correspond roughly to sorption of a little more than a mono-
molecular layer of calcium hydroxide on the external surfaces of the clay. This sorption
corresponds roughly to the lime fixation point and suggested that the latter represents
approximately a monolayer of lime on the external surfaces of the particular clay
concerned.
5. The very large amount of slow sorption beyond this value is thought to reflect
several additional processes, but principally the slow reaction of adsorbed lime with
the clay surfaces to produce calcium silicate and calcium aluminate hydrates. We
think that, at least in wet systems, these products spall from the clay surface, thus
liberating fresh clay surface for further adsorption and reaction. Physical adsorption
of lime onto the newly formed reaction products is also likely. Finally, a slow, re-
stricted entry of lime into the interlayer spaces of the clay may take place.
The observation that adsorption of lime occurs in clay-lime systems does not in
itself provide any particular indication of how the clay properties are so drastically
changed within a short time after the addition of lime. The mechanics of this process
requires further elucidation.
Strong flocculation is commonly observed when lime is added to clay. In general,
according to current colloid-chemical concepts (25), clays flocculate on addition of
electrolytes because of the modifying effect of theelectrolyte on the extension of the
electrical double layer from the surfaces of the clay particles. The electrolyte
represses the double layer, thus reducing the electrostatic repulsive forces between
clay particles. This results in a net attraction, especially between negatively charged
faces and positively chargededges of adjacent particles, anda "card-house" or "double-T"
structure develops. The flocculated clay is still sensitive to water and, generally
speaking, removal of the electrolyte from the pore solution results in a process whereby
the double layer again spreads out, the repulsive forces between particles increase, the
floes weaken and are reduced in size, and eventually the system deflocculates.
In our experiments it was noted that removal of lime from the pore solution results
in only low concentrations of electrolyte being present after a few hours; yet the initially
formed floes persist and become, if anything, more pronounced and stable with time.
Obviously, the properties of the clay are altered more profoundly by lime than by
1- nonreactive electrolytes such as salts.
One of the authors has demonstrated (26) that lime reacts almost instantaneously
with hydrous alumina of high surface areato generate a well-crystallized compound,
tetracalcium aluminate hydrate, C4AH1a (in the shorthand notation commonly used for
these compounds, C = CaO, A= AlaOa, H = H20, S = Si02, etc.). This compound is
also probably formed on the hydration of CaA in portland cement (27). We postulate
90

that in lime-clay systems a similar immediate reaction takes place between the alumina-
bearing edges of the clay particles and the lime adsorbed on the clay surfaces. In
particular, we postulate that this reaction occurs at the points of contact between the
edges of one particle and the faces of adjacent particles in the card-house structure
of the floes. Calcium silicate hydrate (tobermorite gel) probably forms at these points
of contact also, but more slowly, perhaps over a period of s6me hours. Formation
of very small quantities of these cementing products at the points of contact is thought
to be sufficient to stabilize the floes aJ1d kr1it the particles together so that plasticity,
shrinkage and swelling, and other normal clay-water interactions are distinctly in-
hibited. Thus, the clay becomes ameliorated but not really stabilized, since the floes
have not been bound together well enough for the mass to retain any significant strength
after soaking. However, the in di vi dual particles are cemented together well enough
within the floes to resist dispersion, and the floes may act as single grains in mechan-
ical analysis ( 6).
a
To develop material capable of holding together and resisting applied loads after
soaking, compaction is required to obtain a minimum void volume, and time must be
allowed to permit the slow continuing chemical reaction to develop additional cementing
products sufficient to fill the voids at least partially.
As shown by Jambor (16), the strength developed in lime-pozzolan systems depends
to some extent on the kindof cementing agent formed, but also in great degree on the
proportion of void space occupied by the cementing agent-the "gel/space ratio"
familiar in portland cement technology.
The formation of calcium aluminate and calcium silicate hydrates at particle contacts
very early in the reaction process does not lend itself to easy demonstration, due to
the very small amounts of these products involved and the nature of the system. The
concept is offe1·ed as a working iiyi:;olhesis cun::;i::;le11t wilh the known properties oi the
system. Further work is under way to provide additional information in an effort to
prove or disprove this hypothesis.

PRODUCTS OF LONG-TERM REACTION BETWEEN LIME AND CLAY

The reactions occuring over a long period of time and, in the last analysis, re-
sponsible for the stabilization of the soil are of at least as much concern as the amel-
/
iorating responses just described. Most soils consist largely of uncombined silica
and of silicates of various kinds; alumino-silicates usually predominate in the clay
fraction. Consequently, it is not surprising that reaction with lime produces compounds
falling largely into two classes: hydrated calcium silicates and hydrated calcium
aluminates.
Except when formed wider hydrothermal conditions, the calcium silicate hydrates
'lre invariably poorly crystallized and difficult to detect. Three such phases are known
in the literature: (a) tobermorite gel, also called C-S-H (gel); (b) C-S-H (I); and
( c) C-:-S-H (II). The gel is a high-calcium phase normally generated in, and responsible
for, the strength of portland cement concrete. C-S-H (I) has a distinctly lower calcium
content, a somewhat different morphology, and is distinguishable by a very strong
exothermic peak at about 850 C on DTA. Although C-S-H (I) may be synthesized in a
fairly well-crystallized form in the laboratory, the gel phase is invariably more poorly
crystallized and has a maximum of only three peaks on X-ray diffraction. C-S-H (II)
is a high-calcium phase whose exact properties are open to some doubt. All of these
compounds are presumed to have a layer structure similar but not necessarily identical
to that of the well-crystallized mineral tobermorite . The latter may be synthesized
readily by hydrothermal means, but its occurrence in soil-lime reaction products
produced under normal atmospheric conditions is extremely unlikely.
There are several types of calcium aluminate hydrates. The one commonly formed
at normal temperatures is a member of the tetracalcium aluminate hydrate group.
This group comprises a number of crystalline modifications and partially dehydrated
states, as well as phases of the same basic structure but incorporating carbonate
groups in place of some of the hydroxyls. The form generally produced by clay-lime
reactions is most similar to a phase with the composition C4 AH12 · C02 and has a
 91

basal spacing of 7. 6 A whicjl remains unchanged on drying. In contrast, pure C4 AH1J

has a basal spacing of 7. 9 A when moist, and the spacing diminishes stepwise to about
7. 4 A or less on drying. It is possible that the soil-lime product may incorporate
some silicon in isomorphous substitution for aluminum.
Reactions of lime with clay minerals at slightly elevated temperatures generally
give rise to a quite different calcium aluminate hydrate, CJAHe. This is a cubic phase
with a crystal structure different from that of the tetracalcium aluminate hydrates; it
forms preferentially at temperatures above about 30 C and once formed is stable at
room temperature. A good recent monograph covering the status of knowledge of both
calcium silicate and calcium aluminate hydrates is that edited by Taylor (28).
Goldberg and Klein (29) carried out the first published X-ray study of lime-clay
reaction products, but 8Ucceeded in detecting only calcium carbonate, which was
probably produced during air drying of the sample prior to X-raying. Eades and Grim
(30) reacted lime with pure clay minerals at 60 C and found that new minerals formed.
Kaolinite so treated yielded peaks for a poorly crystallized calcium silicate hydrate of
unspecified type, and a peak at 5.1 A (and others) which we interpret as being due to
C 3 AH 6 • Eades and Grim could not detect any new crystalline product of lime-mont-
morillonite reaction, although it was obvious that reaction had occurred. Illite reacted
somewhat, but no positive statement was made as to the nature of the reaction product '
or products.
Hilt and Davidson (31) examined the product of a long-time reaction of lime and
montmorillonite at room temperature, and finally identified it as a C4AH1J type ma-
teriql. Its properties were consistent with the previously mentioned product having a
7. 6 A spacing. The published X-ray diffractometer traces indicated weak peaks attrib-
utable to poorly crystallized calcium silicate hydrates as well. Glenn and Handy (32)
studied the products of reaction between several clay minerals and various forms of
lime at room temperature, but their results were not completely interpreted. They
found in general that poorly crystallized calcium silicate hydrates and the 7. 6 A cal-
cium aluminate hydrate were found in the kaolinite systems; in addition, an unknown
12. 6 A compound was formed when dolomitic lime was used. Montmorillonite yielded
both the 7. 6 A calcium aluminate hydrate and materials whose diffraction peaks were
ascribed to a more nearly pure C4AH1J phase.
Glenn and Handy observed no reaction of lime with quartz, and little reaction with
muscovite or vermiculite. In contrast, in studies of the products formed in field soil
stabilization projects, Eades, Nichols, and Grim (3) found that considerable reaction
had occurred in quartz- and mica-bearing soils. When examined under the petrographic
microscope, the quartz and mica particles were found to have developed fuzzy outlines,
and visible cementing gel was present not only as a coating on the grains but also in
cracks within the grains. Examination by X-ray methods disclosed the presence of
calcium silicate hydrates as well as of calcium carbonate. The authors did not refer
to the possibility that calcium aluminate hydrates might also have formed.
Diamond, White, and Dolch (33) reported that kaolinite and montmorillonite reacted
with lime to produce different products under different reaction conditions. At 60 C,
both clays produced calcium silicate hydrate classified as C-S-H (I) and the kaolinite
produced the cubic CJ AHe; no crystalline calcium aluminate compound was formed
from montmorillonite. At lower temperatures the products from both clays were
considered to be tobermorite gel and the 7. 6 A calcium aluminate hydrate. Small peaks
at about 9 A were found in the montmorillonite products under both conditions; but they
disappeared on washing, and no explanation for their presence could be offered. These
authors found that both the tobermorite gel and the 7. 6 A calcium aluminate hydrate
product were formed by reaction of lime with mica, illite, and even pyrophyllite.
Reaction with quartz generated tobermorite gel, even at 60 C.
Glenn and Handy (34) recently reported further results on long-term studies of
reaction products of lime and montmorillonite at room temperature. In general, the
formation of tobermorite gel, C-S-H (I), C4AHlJ, and possibly C-S-H (II) were re-
ported. Interestingly enough, both fresh mixtures of clay and lime and mixtures which
had been allowed to react at room temperature for several years yielded only well-
crystallized tobermorite on hydrothermal treatment. The aluminum present in the
92

product was thought to be incorporated as an isomorphous replacement for silicon

within the tobermorite lattice.
It is noteworthy t hat in all t he s tudie s dis cussed so far, no quaternary compounds
(that is, compounds in t he lime - silica- alumin a- water system) were reported. App ar-
ently, where separate calcium aluminate phases were not developed, the aluminum
present was incorporated isomorphously in the calcium silicate phase. In contrast to
these results, Benton (35) reported that reactions of lime with calcined kaolinite
yieided the quaternary compound gehlenite hydrate, C2 ASx, also called Stratling' s
compound. This compound was formed in addition to the more usual products, a
poorly crystallized calcium silicate hydrate and the 7. 6 A calcium ¥uminate hydrate.
In a Japanese study (36) the same gehlenite hydrate and the 7. 6A calcium aluminate
hydrate were reported toresult from reactions at normal temperatures between lime
and Kanto loam soils, which consist mostly of the amorphous clay mineral allophane.
At higher temperatures the same soils produced C3AHa and a hydrogarnet quaternary
phase instead of these compounds. When gypsum was also included in the treatment
of these soils with lime, excellent stabilization results were obtained. In this case,
ettringite or so-called cement bacillus, a sulfate-bearing phase formed in hydrating
cements, was the major cementing agent formed.

MECHANISM OF DEVELOPMENT OF CEMENTATION PRODUCTS

The early physico-chemical reactions of lime that produce the amelioration effects
on soil clays and the nature of the cementing compounds that produce the final cemented
product have been discussed, but little has been said about the mechanics of the chemical
and structural transformations that generate the final compounds. Very little is known
about the details of the reaction processes involved, a..11d most i..11formation is speculative
in nature. This lack of knowledge is one of the major gaps in our understanding of the
soil-lime system.
Eades and Grim (30) have suggested that with kaolinite, "the reaction seems to
take place by lime eating into the kaolinite particles around the edges with a new phase
forming around a core of kaolinite. " Electron micrographs by Diamond, White and
Dolch (33) tend to confirm this idea, in part, since the edges of residual kaolinite
particles were ragged and irregular, as if they had been att:icked chemk;i lly The
probability that partial exfoliation had taken place was also raised by the latter authors,
who called attention to the fact that since two distinct crystalline hydration products
were produced, a simple topotactic solid-state mechanism was not a reasonable ex-
planation.
Recently Sloan (37) reported an electron-microscope study of the effects of treating
kaolinite with sodium hydroxide and with lime in dilute suspensions. He confirmed
that under such conditions the primary attack on the kaolinite particles took place at
the particle edges. With lime he observed what appeared to be nucleation of a reaction
product at or near the edges of the kaolinite particles; however, this product did not
appear to be one previously reported to occur,in lime-clay reactions, and the exact
significance of this observation is uncertain.
With illite and montmorillonite, Eades and Grim concluded: "Following the satura-
tion of the interlayer positions with calcium ions the whole clay mineral structure
deteriorates without the formation of substantial new crystalline phases." In contrast,
Diamond, White and Dolch reported that new crystalline phases (tobermorite gel and
the 7. 6 A calcium aluminate hydrate) were formed from these two minerals and, even
in advanced stages of the reaction when most of the clay had been decomposed, the
clay still remaining retained its crystallinity almost intact. For example, a 2 M illite
retained all the X-ray peaks characteristic of this polymorph, and montmorillonite
retained its characteristic (060) spacing.
Eades, Nichols, and Grim (3) suggested that a pH high enough to dissolve silica is
an essential feature of the process of formation of calcium silicate hydrates by lime-
clay reactions; this idea was seconded by Diamond, White and Dolch (33) who envisioned
solution from the edges of the clay crystals as a possible mode of reaction. The
addition of sodium hydroxide to a lime-soil mixture has been found by several workers
 93

to accelerate strength development (13, 38). Moh (39) bases his explanation of these
benefits in both soil-lime and soil-cement systems partly on the hypothesis of an
increased rate of solubilization of potentially reactive silica.
These arguments imply that the reaction path proceeds through solution, with silica
liberated from the clay reacting with dissolved lime. fu contrast our results seem to
suggest that reaction proceeds between adsorbed lime and the surface layers of the
clay with which it is in contact. Such a mechanism has been proposed to account for
the rapid generation of tobermorite gel by lime sorbed on silica gel surfaces at mod-
erately elevated temperatures ( 40).
It may be that, as seems to be the case in portland cement hydration, both "adsorbed-
state" and "through-solution" reactions take place and both contribute to the development
of the final product. However, any ultimate resolution of these rather difficult problems
must await further intensive experimental work.

SUMMARY
Among physico-chemical mechanisms commonly suggested as explanations of the
stabilizing effects observed in soil-lime stabilization, cation exchange replacement of
existing cations by calcium, flocculation, and carbonation have been prominent. Since
many soils in need of stabilization are naturally calcium saturated, flocculated, or
both; the first two mechanisms can no longer be seriously considered. Since soil-lime
systems sealed from contact with carbon dioxide develop the normal indications of
stabilized soils, carbonation must also be rejected.
The effects of lime on soils are such that two ' stages of reaction can be detected:
(a) an early stage in which the properties of the plastic soil are greatly ameliorated but
little permanent strength is developed, and (b) a subsequent stage marked by the slow
development of strength and the accumulation of soil-lime reaction products. Among
the effects observed in the first stage are large increases in the P. L., generally
leading to a reduction in the P. I. ; a sharp reduction in the apparent content of clay
size particles as they are bound into floes stable against the dispersion incident to
/ mechanical analysis; increase in the moisture and the compactive effort required to
achieve a given density; and reduction in such parameters as swell pressure, volume
change on drying, and permeability. These changes are commonly produced in periods
ranging from minutes to a few hours after the addition of lime.
Recent work by various workers has suggested the existence of a lime fixation point,
that is, an amount of lime that must be added to a given soil to maximize these effects.
This lime was considered to be unavailable for further reaction. The mechanism of
the effect was discussed in terms of an apparent crowding of excess calcium cations
onto the clay following the addition of lime. This in turn was tentatively attributed to
pH-dependent exchange sites originating at the edges of the Glay particles.
The results of our recent experimental work contradict these hypotheses. It is
shown that the cation crowding effect is, in reality, one of physical adsorption of
calcium hydroxide onto the clay surfaces. We postulate that the ameliorative effects
are due to almost immediate but limited chemical reaction at the points of contact
between the edges and faces of primary clay particles within the floes formed by the
normal electrolyte effect of added lime. This reaction is visualized as the formation
of small amounts of tetracalcium aluminate hydrate by reaction of the exposed Al(OH)x
groups at the edges of the clay surfaces with lime sorbed on the faces of adjacent sur-
faces. This immediate reaction is supplemented by somewhat slower reaction of the
silica with lime to generate tobermorite gel.
The nature of the compounds held responsible for the slow development of strength
in soil-lime systems has been discussed in some detail. The exact products formed
vary somewhat with the kind of clay and the reaction conditions, especially temper'!-1!,!DL
There are commonly at least two phases produced, a calcium silicate hydrate and a
calcium aluminate hydrate. The former is usually tobermorite gel; the latter is a
well-crystallized hexagonal compound, which is probably an il:ppure (substituted)
tetracalcium aluminate hydrate and is characterized by a 7. 6 A basal spacing independent
of drying conditions. At temperatures only slightly above normal room temperature
94

a different calcium aluminate hydrate phase, the cubic tricalcium aluminate hexahy-
drate, is produced.
It was found that quartz, mica, and other phases considered less reactive than
clays may also react under appropriate conditions and give rise to similar cementitious
products. Quaternary phases in which silicon and aluminum atoms occur in distin-
guishable lattice positions are not commonly formed, except when lime reacts with
previously calcined clay or with the amorphous clay mineral allophane.
The present. lack of detailed knowledge of the mechanisms of the chemical n'!actions
that produce these final cementitious products is discussed. Evidence exists that the
reactions are favored by conditions of high pH which would make silica more soluble;
addition of sodium hydroxide to lime-clay systems produces significant strength gains
at early ages. This implies a mechanism involving reaction of dissolved silica and
alumina with calcium ions. On the other hand, evidence that the reaction is preceded
by sorption of the calcium hydroxide from solution strongly implies a direct surface
reaction with the clay. It may be that both mechanisms are operative.

ACKNOWLEDGMENTS
The authors acknowledge with thanks the contributions made by W. C. Ormsby,
National Bureau of Standards, and Donald G. Fahs, U. S. Bureau of Public Roads, in
discussions of the literature of soil - lime stabilization.

REFERENCES
1. Johnson, W. M., Cady, J. G., and James, M. S. Characteristics of Some Brown
Grumosols of Arizona. Proc. Soil Science Soc. of Amer., Vol. 26, pp. 389-
393; 1962.
2. Prikryl, J., and Esterka, F. Equilibrium Studies of Sorbed Clay Mineral Cations
and Electrolytes in Alkaline Solutions. (In Czech). Silikaty, Vol. 4, pp. 135-
145, 1960.
3. Eades, J. L., Nichols, F. P., Jr., and Grim, R. E. Formation of New Minerals
with Lime Stabilization as Proven by Field Experiments in Virginia. Highway
Research Board Bull. 335, pp. 31-39, 1962.
4. Wang, J. W. H., Mateos, M., and Davidson, D. T. Comparative Effects of
Hydraulic, Calcitic, and Dolomitic Limes and Cement in Soil Stabilization.
Highway Research Record No. 29, pp. 42-54, 1963.
5. Jan, M. A., and Walker, R. D. Effect of Lime, Moisture, and Compaction on a
Clay Soil. Highway Research Record No. 29, pp. 1-12, 1963.
6. Lund, 0. L., and Ramsey, W. J. Experimental Lime Stabilization in Nebraska.
Highway Research Board Bull. 231, pp. 24-59, 1959.
7. Hilt, G. H., and Davidson, D. T. Lime Fixation in Clayey Soils. Highway Re-
search Bull. 262, pp. 20-32, 1960.
8. Clare, K. E., and Cruchley, A. E. Laboratory Experiments in the Stabilization
of Clays with Hydrated Lime. Geotechnique, Vol. 7, pp. 97-111, 1957.
9. Zolkov , E. Influence of Chlorides and Hydroxides of Calcium and Sodium on Con-
sistency Limits of a Fat Clay. Highway Research Board Bull. 309, pp. 109-
115, 1962.
10. Woods, K. B., and Yoder, E. J. Stabilization with Soil, Lime, or Calcium Chlo-
ride as an Admixture. Proc. Conf. on Soil Stabilization, M. I. T., pp. 3-19,
1952.
11. Taylor, W. H., Jr., and Arman, A. Lime Stabilization Using Preconditioned
Soils. Highway Research Board Rull. 262, pp. 1-19, 1960.
12. Grim, R. Applied Clay Mineralogy. p. 207. New York, McGraw-Hill, 1962.
13. Wolfe, R. E., and Allen, J. R. Laboratory Evaluation of Lime Treatment of
Typical Minnesota Soils. Minnesota Dept. of Highways, Final Rept. of In-
vest. No. 601, 1964.
14. Herrin, M., and Mitchell, H. Lime-Soil Mixtures. Highway Research Board
Bull. 304, pp. 99 - 138, 1961.
 95

15. Mitchell, J. K., and Hooper, D. R. Influence of Time Between Mixing and Com-
paction on Properties of a Lime-Stabilized Expansive Clay. Highway Research
Board Bull. 304, pp. 14-31, 1961.
16. Jambor, J. Relation Between Phase Composition, Over-all Porosity, and Strength
of Hardened Lime-Pozzolana Pastes. Mag. Coner. Res., Vol. 15, pp. 131-
142, 1963.
17. An day, M. C. Curing Lime-Stabilized Soils. Highway Research Record No. 29,
pp. 13-26, 1963.
18. Shepard, F. D. Effects of Recompaction on the Strength of a Lime and Cement
Stabilized Soil. Unpublished report of study carried out in laboratories of
Virginia Council of Highway Investigation and Research, Charlottesville, 1963.
19. Laguros, J. G., Handy, R. L . , and Reign, L. L. Effect of Exchangeable Calcium
on Montmorillonite Low-Temperature Endotherm and Basal Spacing. Highway
Research Board Bull. 349, pp. 51-58, 1962.
20. Pietsch, P. E. , and Davidson, D. T. Effects of Lime on Plasticity and Com-
pressive Strength of Representative Iowa Soils . Highway Research Board Bull.
335, pp. 11-30, 1962.
21. Pinto, C. de S., Davidson, D. T., and Laguros, J. G. Effect of Lime on Cement
Stabilization of Montmorillonitic Soils. Highway Reaearch Board Bull. 3 53,
pp. 64-83, 1962.
22. Ho, C., and Handy, R. L. Characteristics of Lime Retention by Montmorillonitic
Clays . Highway Research Record No . 29, pp. 55-69, 1963.
23. Ho, C., and Handy, R. L. Electrokinetic Properties of Lime-Treated Bentonites.
Paper presented at 12th Nat. Conf. on Clays and Clay Minerals, Atlanta, Ga.,
Oct. 1963.
24. Pike, R. G., and Hubbard, D. Source of the Non-Migratable Ionic Charges De-
veloped by Portland and High-Alumina Cements During Hydration. Highway
Research Board Proc., Vol. 37, pp. 256-270, 1958.
25. Van Olphen, H. An Introduction to Clay Colloid Chemistry. New York, Inter-
science Publishers, 1963. 301 pp.
26. Diamond, S. Rapid Reaction of Lime with Hydrous Alumina. Nature, Vol. 204,
p. 183, 1964.
27. Buttler, F. G., Dent Glasser, L. S., and Taylor, H. F. W. Studies on 4Ca0 ·
AhOJ · 13 H20 and the Related Natural Mineral Hydrocalumite. Journal of
Amer. Ceram. Soc., Vol. 42, pp. 121-126, 1959.
28. T aylor, H. F. W. , ed. The Chemistry of Cements. New York, Academic Press,
1964.
29. Goldberg, I., and Klein, A. Some Effects of Treating Expansive Clays with
Calcium Hydroxide. ASTM Spec. Publ. 142, pp. 53-71, 1952.
30. Eades, J . L. , and Grim , R. Reactions of Hydrated Lime with Pure Clay Minerals
in Soil Stabilization. Highway Research Board Bull. 262, pp. 51-63, 1960.
31. Hilt, G. H. , and Davidson, D. T. Isolation and Investigation of a Lime-Montmo-
rillonite Crystalline Reaction Product. Highway Research Board Bull. 304,
pp. 51-64, 1961.
32. Glenn, G. R., and Handy, R . L. Lime-Clay Mineral Reaction Products. Highway
Research Record No. 29, pp. 70-82, 1963.
33. Diamond, S., White, J. L., and Dolch, W. L. Transformations of Clay Minerals
by Calcium Hydroxide Attack. In Clays and Clay Minerals, Proc. 12th Conf.
New York, Pergamon Press, inpress.
34. Glenn, G. R., and Handy, R. L. X-Ray Studies of Lime-Bentonite Reaction
Products. Paper presented at 66th Ann. Meeting, Amer. Ceram. Soc., Chicago,
Ill. ' 1964.
35. Benton, E. J. Cement-Pozzolan Reactions. Highway Research Board Bull. 239,
pp. 56-65, 1960.
36. Ariizumi, A., Maki, T., and Oba, M. On Stabilization of Kanta Loam by Means
of Lime Gypsum Admixtures. Ann. Rept. of Roads, Japan Roads Assoc.,
Tokyo, pp. 22-31, 1962.

I
96

37. Sloan, R. L. Early Reaction Determination in Two Hydroxide-Kaolinite Systems

by Electron Microscopy and Diffraction. Paper presented at 1st Ann. Meeting,
Clay Minerals Soc., Madison, Wis., 1964.
38. Davidson, D. T., Mateos, M., and Barnes, H. F. Improvement of Lime Stabiliza-
tion of Montmorillonitic Clay Soils with Chemical Additives. Highway Research
Board Bull. 262, pp. 33-50, 1960.
39. Moh, Za-Chieh. Soil Stabilization with Cement and Sodium Additives. Jour. Soil
Mech, and F01md, Div , ; A.BCE, Vol. 88, Pt. SM-6, pp. 81-105, 1962.
40. Greenberg, S. A. The Chemical Reactions of Calcium Hydroxide, Silica, and
Water Mixtures at 82°C. Jour. Phys. Chem., Vol. 61, pp. 373-374, 1957.

Discussion
R. L. HANDY, T. DEMIREL, CLARA HO, R. M. NADY, C. G. RUFF, ET AL.,
Iowa State University, Ames. -We sincerely congratulate the authors for this timely
review and skillful appraisal of the mechanisms of soil-lime stabilization. In this
discussion we would like to call attention to two statements in the paper.

The point of inflection of the plot of lime added vs P. L.

has been dignified by the appellation "lime fixation point."

This is quite correct. However, for reasons of purification we later substituted "lime
retention point" for "lime fixation point" (22). The term "fixation" was a rather nn-
fortunate apRellation because analogy with K+ or NH 4 + fixation implies an irreversible
cation adsorption phenomenon. As pointed out by the authors, this is not so; the
CA(OHh is readily removable at early ages even though the increase in plastic limit
(P. L.) appears to be immediate.

Ho and Handy presented a detailed analysis of their results

which we consider highly speculative.

We agree completely. In addition we would paraphrase this with only the names changed,
perhaps with "Diamond and Kinter" for "Handy and Ho." Not that we regard speculation
as sinful; on the contrary, speculation is high art, vital for exiting resonant vibrations
in others and for decreasing free energy of the authors. With this in mind, let us dis-
::uss in detail the speculations of Diamond and Kinter, and also include a few from Iowa
State University.
The authors' speculation of physical adsorption of lime is not in disagreement with
other hypotheses, since all evidence points to physical rather than chemisorption. How-
ever, this adsorption should be explained. If it is an ionic adsorption, it should be con-
sistent with double layer theory, since one cannot adsorb 60 meq or more of negative
OH ions on 100 g of clay surface without some kind of incentive. Also, the love for
OH- appears rather special, since other anions (sulfate, chloride, etc.) are rejected,
even when in company with calcium. We, therefore, have suggested an ionic adsorption
whereby OH- and ca++ ions repress the double layer (ion "crowding"), a widely accepted
trend when an electrolyte is added to a clay suspension.
Then, since the lime adsorption reaction is uniquely dependent on pH of the suspen-
sion, it would appear that OH- ions react at broken bonds or react with H+ ions of the
clay, increasing the negative charge so it can adsorb more calcium in stoichiometric
proportion. Actually, the occurrence of a pH-dependent charge was not invented by us
especially for this purpose; it is recognized and well established in the literature, col-
loid chemists preferring to call it "an electrical double layer created by adsorption of
potential-determining ions" (25). The amount of charge available is believed to depend
on the pH, broken bonds, andthe dissociable hydrogen ions available. We may even
speculate that with a long enough time and a high enough pH, montmorillonite could even
 97

2400
+ 50~60°C Curing
0 40°C •
2000 8 5 ~ 23°C •

1600
Unconfined
Compressive
Strength, 1200
psi

800

400

0
0 10 . 20 30
Crystalline Ca (OH~ depletion, gm/IOOgm dry mix
Figure 1. Relation of unconfined strength to Ca(OH) 2 reacted in lime-Ca-bentonite-water
cylinders cured at various times and temperatures.

dehydrolyze from the octahedral layer to contribute as much or more charge than the
permanent charge derived from isomorphous substitution. Unfortunately, this is pure
speculation, and soil chemists usually work in the pH range from 3 to 8. Recent work
on 60 Wisconsin soil samples averaging 13. 3 percent clay gave C. E. C. =30 + 4. 4 X,
where Xis the pH; the correlation coefficient was 0. 979 (3). That is, increasing the
pH 5 points (corresponding to an increase in OH ion concentration of 10 5 times) in-
creased the cation exchange capacity 22 meq/100 g. Although applying this relation-
ship to a lime-bentonite mixture involves extrapolation to higher clay content and higher
pH range, the prediction is at least in the correct order of magnitude.
Apparently Diamond and Kinter favor molecular rather than ionic adsorption. They
also suggest that adsorption results in pozzolanic reaction where the lime coats the clay
crystal edges, and cementation of these edges to faces to give a stable floe "card-house"
structure, a very attractive and valuable speculation, certain to excite some resonant
vibrations. However, we feel that their adsorption mechanism needs additional clari-
fication. From the surface area calculation the authors apparently assume adsorption
of a monolayer OH- and ca++ ion patchwork in the ionic ratio 2 to 1. Perhaps they will
entertain some questions: Why aren't other salts adsorbed (except at high pH)? Why
did repeated leaching eventually remove more Ca++ than 2(0H)-? (This is predicted by
double layer theory in which OH' s react with clay.) What is the shape of the Ca(OH)z
molecule, and how does the molecular adsorption pattern fit the clay mineral structure,
both on the silica tetrahedral sites and over the holes? Also, if a Ca(OH)z is adsorbed
as a monolayer, we would hope that X-ray diffraction will eventually indicate some be-
tween the layers, since the same forces of adsorption presumably exist there as on the
outside.
The final point we wish to make is the "lime retention" point. Regardless of the de-
tails of the adsorption mechanism, the lime retention or adsorption point is a valuable
98

3.4
3.2
3.0
0~
.. 2.8

-
~
If)

1.\1
2.6
2.4
14
13 c
<t
2.2 12
,,,
0

0
~ c
2.0 II
,,,8.
0
1.\1
1.8 ,,,
:J

iii
1.6 9 E
Q)
:J
..c 1.4 8 ·c
:J ..c
i
1.2 7 -:J
(.)
1.0 S;~*J-*/* CT
Q)

:::c 0.8 5
* * •
~

-
0
Q)
::::;
0.6
0.4
0.2
4
3
2
0

:::c
a.

I
00 2 3 4 5 6 7 8 9 10 II 12

Ca(OH~ added, 0/o

Figure 2. Relations of dilute HCl-soluble Siq, , Al;a 0 3 and equilibrium pH of Ca( OHh -Na-
bentoni te suspensions cured at room temperature for more than 2 yr.

concept, particularly useful for engineering purposes. To be sure, the lime retention
point slowly changes on curing, probably as a result of localized pozzolanic reactions
as suggested by Diamond and Kinter. Of more practical importance is that pozzolanic
reactions sufficient for cementation do not occur unless lime is added in excess of the
lime retention point, readily estimated for the P. L. test. This is further illustrated
by new data presented in Figures 1 and 2.
Figure 1 shows the compressive strength of 1/2-in. diameter by 1-in. lime-ca++_
bentonite-water cylinders vs loss of crystalline lime from the system as measured by
X-ray diffraction. Curing was at various times and temperatures. The initial mixes
contained 41. 6 percent lime, corresponding to a C/S ratio of 1. 0 with all lime and
clay reacted. Extrapolation of the strength curves to zero strength suggests that about
6 to 7 percent lime was unavailable for much cementation reaction, the strength being
about the same as for the pure clay. This lime percentage is somewhat higher than the
lime retention point from plasticity tests.
Figure 2 shows silica and alumina soluble by dilute acid treatment of Na-bentonite-
lime suspensions which had been sealed and cured at room temperature for over 2 yr.
The soluble silica and alumina are assumed to be mainly from pozzolanic reaction
products, although as can be seen, about 0. 4 percent of each was leached from untreated
clay. The amount of reacted silica is approximately doubled by addition of 2 to 7 per-
cent lime, perhaps in part due to glassy or silicious impurities in the bentonite. The
 99

amount of reacted alumina is increased only slightly by addition of 2 to 8 percent lime.

However, in the neighborhood of 8 to 9 percent lime, both silica and alumina take a
sharp jump, indicative of pozzolanic reaction. It is interesting that in this region pH
has stabilized at 10. 5, which other evidence suggests is the threshold for pozzolanic
reaction (42).

References
41. Helling, Charles S., Chesters, G., and Corey, R. B. Contribution of Organic
Matter and Clay to Soil Cation-Exchange Capacity as Affected by pH of the
Saturating Solution. Proc. Soil Sci. Soc. Amer., Vol. 28, No. 1, pp. 517-520,
1964.
42. Davidson, L. K., Demirel, T., and Handy, R. L. Soil Pulverization and Lime
Migration in Soil-Lime Stabilization. Highway Research Record No. 92, 1965.

SIDNEY DIAMOND and EARL B. KINTER, Closure-We thank Drs. Handy, Demirel,
Ho, Nady, Ruff, et al. of the Iowa State University staff for their "resonant vibrations."
Differences in interpretation of data always occur in fields of active scientific inves-
tigation, and we are not surprised that our interpretations are not universally accepted.
With regard to specific points raised by the discussors, we would offer the follow-
ing statements:
1. We concur that the "love (of clay surfaces) for OH- ions appears rather special,"
not only because other anions are rejected, but also because this love is far less torrid
when the OH- ions are accompanied by cations other than calcium. For example, we
have immersed calcium-saturated Wyoming bentonite in solutions of sodium hydroxide
of the same concentration as saturated lime solutions (0. 04 N). Such sodium hydroxide
solutions ha ve a pH distinctly higher than those of saturated llm ewater, approximately
13. 6 as compared to 12. 4. In a typical experiment in which Y2 g clay was shaken in
100 ml of solution, we found that 48 meq of sodium ions per 100 g of clay were re-
moved from solution by the clay, and 23 meq of calcium ions per 100 g of clay were
returned to the solution. Thus, after correcting for cation exchange, an excess of
about 25 meq sodium ions per 100 g of clay was adsorbed by the clay, along with ap-
proximately the equivalent amount of hydroxyl ions. Thus, in this higher pH system,
the clay picked up only about a quarter as many hydroxyls as it does under comparable
conditions in saturated lime solutions of lower pH. Obviously, adsorption of OH- ions
is not uniquely dependent on the pH of the system, but increases markedly in the spe-
cific joint presence of calcium and hydroxyl ions. This is not consistent with the ar-
gument advanced by the discussors that the process is one of formation of an augu-
mented double layer created by adsorption of potential-determining (hydroxyl) ions,
coupled with indifferent adsorption of accompanying cations.
Prikryl and Esterka (2) performed the reverse experiment, that is, immersion of
sodium-saturated bentonlte (and other clays) in calcium hydroxide solutions. They
found that the ratio of equivalents of Ca adsorbed from solution to equivalents of Na
released to solution (which would be 1 if only ordinary cation exchange were taking
place) increased uniformly with increasing initial concentration of calcium hydroxide.
Again, these results are not explainable on the hypothesis suggested by the discussors,
but were taken by Prikryl and Esterka (and by the present writers) to be indicative of
physical adsorption of calcium hydroxide.
2. The Wisconsin soil data cited by the discussors (41) yield a statistical pre-
diction of a 22 meq/ 100 g pH-dependent gain in cation exchange capacity, from pH 7
to pH 12. This is not really of the same magnitude as the amount of lime removed
from solution by montmorillonite. Our data s how adso rption of the order of 100 meq/
100 g clay (3 % percent lime by weight of clay) in 10 min a nd more than 120 meq/100
g of clay (4 % pe r cent lime) in less than an hour. However, the apparent adsorption
of sodium hydroxide, cited under point 1 of this closure, is of the same magnitude,
and it may well be that adsorption of sodium hydroxide by bentonite is explainable on
100

420 420
'
-
,:;.
u-,
360 HM-67 SO IL
Cl:
£ 28-DAYCURE
1-11c- 44SO I L
28-0AY CUR
Ht<-30 SOIL I
2 8- DAY CU RE! 360
Cf)

a;_
I
f- -( f-
l? 1"> CJ
z z
::> I)
~ 300 0 t 300 ~
f- f-
(f) 0 (f)
'.) ( (
0
'>240
(]j
Cf) ()
,, 240 >
Vi
Cf)
w w
a:: a::
~ 180 ) 180 ~
0 0 0
Cl ( 0 Cl
0
0 0 0
w 120 120 w
z z
LL ii:
z z
0
Cl
z
60 60 8z
:::> LIME RETENTION POINT, LIME RETENJJON POINT LIME RETENTION POINT :::>
I T
PE ,CENT
1
I 2 PEROE T
I I I I
'
I
T PERICENTI I
0
2 4 6 8 10 12 0 2 4 6 8 10 12 0 2 4 6 8 10 12
PERCENT LIME PERCENT LIME PERCENT LIME

Figure 3. Unconfined compressive strength as function of lime content for a montmoril-

lonite, an illite-chlori te, and a kaolinite soil ('l, Figs. 3, 4, 5).

the basis of pH-dependent exchange. We

-do not imply that this is so, only that the
possiblity exists.
3. In response to the question of why
500 other salts (we take this to mean "other
electrolytes") are not adsorbed, we are
-'
ui
Ie
HEAVY LAY:
4-WEEK CURE
c I forced to plead ignorance. Unfortunately,
Q_
O 26-WEEK CURE ( we are not alone, since very little pub-
£ 400 lished research has been done on adsorp-
f-
l?
tion from solution, and most of this con-
z cerns adsorption of organic compounds.
w 0
a:: A useful discussion of the available lit-
f-
en erature was given by Brunauer and Cope-
w
> 300 land (43).
iii
en
w
a::
• 4. With regard to the question of the
balance of 2 OH- and ca++ ions released
into solution on repeated washing of lime-
Q_
::;;;
0
• treated clay, we consider that two factors
0 200 - are responsible. First, during the time
0 of processing, extending over periods of
w
z
lJ._
• one to several hours, some reaction be-
tween the adsorbed lime and the clay has
z undoubtedly occurred to produce small
0
0 100 amounts of tobermorite gel and probably
z 0
::::i C4AH13-type compounds. These sub-
• stances release hydroxyl and calcium ions
LIME RETENTION POINT, 3 PERCENT in ratios less than the 2:1 ratio char-
(EST. FROM AUTHOR'S DATA)
I I I I acteristic of lime. The second cause of
0
0 2 4 6 8 10 the discrepancy is the increasing degree
PERCENT LIME of carbonation one must expect on re-
pea t ed handling of lime- bearing materials
Figure 4. Unconfined compressive strength in the atmosphere.
as function of lime content for heavy clay 5. The shape of the calcium hydroxide
(London clay) (46, Fig. l). molecule (or ion-triplet) is not, of course,
 101
260 260
Ul ~
n:
£
I15· 0AY CURE I
!(AOL.IN Ii l l(AQUN E
15• DAY CURE
I 0:

:£
t- ~ t-
f.!) 240 240 '-'
z I> z
w w
a::
t-
<f> ~
~ 220 220 ~
iii ( iii
ff)

..15 ..
<f>
w w
a:: a::
0
200 200 ~
u 0

0
0
w
z 0
0 "z
w
,;: 1eo ,_ 0 180,;:
z z
0 0
u 0
z 0 z
::J 0 ::J
160 160
0 3 6 0 4 6
PERCENT LIME PERCENT LIME

Figure 5. Unconfined compressive str ength as function of lime content for two kaolins;
lime re tention point not available ('.t.7, Fig. 7).

amenable to direct determination. Some clue may be.obtained from the crystal struc-
ture of portlandite, which is hexagonal with a= 3. 59 A and c = 4 . 91 A. The structure
is built up of layers of octahedra, each of which has a calcium ion at its center and a
hydroxyl group at each of its six apices; each hydroxyl group is shared by three octa-
hedra (28).
If this arrangement were to be preserved in calcium hydroxide ads9rbed on a clay
su rlace, one would expect th e thickne ss or each layer t o b e about 4. 9 A and the .coverage
area per molecule to be that of a 6 Cf rho mbu s with 3. 59 A. sides, that i s , 11. 2 A2 • Ip a
s tudy of the chemisorption of lime on silica gel, Gree!lberg (44 ) us ed a n area of 25 A2
per Ca(OHh molecule and found that this gave reasonable r esults. The pr.esent writers
have adsorption isothe rm data which indicate a coverage area of about 25 A2 on mont-
morillonite and a somewhat higher value on kaolinite. Thus, the indications are that
adsorbed calcium hydroxide molecules are not as closely packed on the clay surface
as they are in crystalline portlandite.
The potential fit of the portlandite structure, that is, Ca06 polyhedra, to the hexa-
gonal arrangement of silica tetrahedra that constitutes an idealized representation of
most clay mineral surfaces was diagrammed by Taylor and Howison (45), and was
shown to be poor, the calcium polyhedra being too large for the hexagonal silica net-
work.
6. Finally, we must address ourselves to the question of the lime retention point.
The hypothesis postulates that no significant reaction occurs, and hence, no strength
gains accrue, for treatment with lime in amounts less than the lime retention point.
The discussors kindly inclosed two new figures to reinforce this argument. Un-
fortunately, the present writers have considerable difficulty in appreciating the signi-
ficance of these figures.
Figure 1 relates strength development to depletion of crystalline lime in systems,
all of which contained 41 percent lime by weight of clay. Only very low densities are
obtainable in systems with such unusually high lime contents. We suggest that because
of these low densities, strength gains obtainable in such systems bear little relation-
ship to those that can be obtained in well-compacted systems of lime contents near the
lime retention point, that is, 2 to 4 percent. The lines plotted on this figure intersect
the zero strength axis at more than 6 percent lime. Surely the discussors do not mean
to imply that a calcium bentonite treated with 6 percent lime and properly compacted
and cured will not develop any significant strength.
The relevance of the discussors' Figure 2 to the question at issue is unfortunately
not apparent to the present writers.
102

Data are available from the literature, however, which bear directly on the question
of strength gains, and hence reactivity, in soils treated with amounts of lime below the
lime retention point. If the concepts of the discussors are correct, a plot of strength
vs lime content for properly compacted and cured specimens should show no strength
gain below the lime retention point, a distinct infl ection at this point, and strength in-
creases from then on as some function of increased lime content. In contradistinction,
we suggest that with many soils, small increments of lime less than the lime retention
point du ln fact add to the strength of such samples as a result of chemical reaction. On
this basis, a plot of strength vs lime content should show incr eases in strength with
very small increments of lime starting at or near zero.
Figures 3, 4, and 5 are offered from data in the literature (7, 46, 47). Only the
data points are submitted; no tr end lines are drawn. Readers Of this discussion are
invited to draw their own conclusions.

References
43. Brunauer, S., and Copeland, L. E. Surface Tension, Adsorption. In Handbook
of Physics, Chapter 7. New York, McGraw-Hill, 1958.
44. Greenberg, S. The Chemisorption of Calcium Hydroxide by Silica. Jour.
Phys. Chem., Vol. 60, pp. 325-330, 1956.
45. Taylor, H.F. W., and Howison, J. W. Relationships Between Calcium Silicates
and Clay Minerals. Clay Minerals Bull., Vol. 3, pp. 98-111, 1956.
46. Dumbleton, M. J. Lime-Stabilized Soil for Road Construction in Great Britain.
Roads and Road Construction, Vol. 40, pp. 321-325, 1962.
47. Ormsby, W. C. Kaolin-Lime-Water Systems: Strength and Rheological Prop-
erties. NBS Tech. Rept .. 1964.