9.

ASSESSMENT OF QUICKLIME PILE BEHAVIOUR

A. HOLEYMAN, Assistant Manager, Research and Geotechnics, S.A. FRANKi N.V., Liege, Assistant Professor,
Free University of Bruxelles, Belgium
J.K. MITCHELL, Professor and Chairman, Department of Civil Engineering, University of California, Berkeley,
CA, U.S.A.

The authors present a mathematical model that describ es the physical phenomena involved in the stabilization of soil
masses with the help of pure quicklil!le piles. The finite-difference one-dimensional fonnula'tion makes provision for
the speed of the slaking of the quicklil!le pile, the expansion of the column , the development of lateral pressures in
the soil and the subsequent consolidation of the soil i n the radial direction. Physical parameters a r e deduced from
tests published by Kuroda et al (2.). The parametric analysis conducted substantiates the importance of the horizontal
permeability of the soil being: treated, the fineness of the lime u sed and its condition of placement.

INTRODUCTION Ca(OH)2 + 15. 5 kcal.

The g round temperature rises adjacent t o the p ile in
Lime treatment in one of the oldest and most widely used r esponse to the heat generation . The temperatur e g radient
types of admixture stabilization for soils. Its value for will depe nd on the speed of the reaction in relation to
improvement of road subgrades and bases has been apprecia- the specific heat and the therma l conductivities of the
ted for centuries. The value of quicklime for the rapid soil and the lime. This can result in increased water eva-
strengthening of soft, saturated clays has become apparent poration, provided there is some kind of ventilation.
more recently. Mixed-in-place lime piles or columns Since evaporation is possible only at the ground surface,
(1) have been developed in Sweden and Ja pan for support of and since the hydration of a lime pile typically takes 2
structures and open excavations. Yamanouchi et al. (2) to 10 days, as will be seen later, we .shall consider that
have described a sandwich method for quicklime the rise in temperature wil 1 have a limited impact on the
stabilization of soft clay used in embankment ove r al l water content of the soil to be stahilized.
construction . Expansion accompanying quic klime hydration can be esti-
A spec ial type of quicklime pile involves the placement mated. By cal culating the stoichiom.etric quantities, we
of quicklime directly in boreholes, rather than mixing the find :
lime with the surrounding clay. Loose lime can be compac-
ted or, alternatively, columns can be formed by stacking 56g CaO + 1 8g H20 - - - -·· 74g Ca{OH)z {l)
discs of compacted quicklime in the hole.
Al though the technique has been practised, there is Solid specific gravities are 3.37 for the quicklime and
little detailed information on the mechanisms of the 2.343 for the slaked lime, re·spectively. Thus, we obtain
stabilization and up to now no definite analyses of the the so lid volumes combination
significant phenomena. In this paper, the physics of the
stabilizat ion phenomena are desc ribed , a physic al model i s 17cc Cao + 18cc H20 - - - - 31.6cc Ca {OH)z (2)
formulat ed. and a mathematical so lution is pres ented. Some
computations are presented according to this method to which means that the solid phase volume increase is about
show the influence of significant parameters. 85 %.
The relative volume. of water necessary to hyd rate the
corresponding volume. of in-situ quicklime can be der ived
PHYSICS from equation (2), once the porosity of the lime placed in
the soil and the yield of the react ion are known. As suming
When quic klime (CaO) and moist soil are brought i nto a compac ted dry unit weight of 11280 N/m3 and a chemi cal
contact, several processes can take place that will affect efficiency of 93 % one gets a r elative volume of water
the properties of the soil : need ed e qual to 0.30.
1. Absorption of water from the soil t o hydrate the Expe rime n ts (6) s how that unde r atmosphe ric c onditions,
quicklime and fill its voids. an apparent v olume V of quicklime and corresponding volume
2. Heating caused by the hydration of the lime. (0.3V) of water combine into an apparent larg:er volume
3. Expansion of the lime upon hydration, which results (1.5 to 1.7 V) of slaked lime with a dry unit weight of
in compression of the soil in the horizontal direc- 10000 to 8500 N/m3 . This net vo lume increase is of crucial
t ion . importance, because the tende ncy for volume. expansion will
4. Hard en i ng of the annular zone around the pile into r esult in lat eral compr ession of the soil, wh.ich will, in
which there has been diffusion of lime. t urn, build up pore pressures. As these pressures
The process occur concurrently, and t h eir complete ana- dissipate, mo r e water will be expelled from the soil into
l ys is is complex. Therefore, some assumpt ions are neces- the draining lime, which will then expand further, un t il
sary. a ll the quicklime has hydrated. , We assume that the slaked
Considering that the migration o f lime through intact lime pile acts as a vertical drain for analy sis of the
clay is a very slow process ( of the order of 1 cm/year), consolidaton resulting from the lateral pressure.
we shall not consider process 4 as playing a significant Based on the previous discussion, we consider t h a t the
role in the stabilization, at lea s t during the first problem c an be treated as one of a cylindrical drain, the
year. tentative expansion of which is controlled by the amoun t
The hydration o f the qui c klime is expr essed by the of drained water.
c h emical react ion (])

897
9.5

between two adjacent piles will not move ard will repre-
sent the raiius of influence of a pile. Let R2 be that
raiius am R1 the initial radius of the lime piles. The
z boundary conditions become :

(9)

in which U1 is the radial expansion of the cavity. The

solution to equation (8) can be shown to be :

- R2/r2
2
(IO)

The radial stress at the inner cavity can be obtained

from equation (6)
2 0. + 2G)
(11)
2 2
- R2/Rl

Figure l Cylindrical coordinates in which K1 is defined as K1 • (A + 2G )/ (1 - R~/Ri)

The volumetric deformation can also be obtained using
PHYSICAL MODEL equations (S) and (4)

Stiffness of the surrounding soil 2

E: a~+ (12)
V r
As the problem has cylindrical symmetry, we shall use
the cylindrical coordinates (fig.1) r, 8 and z. andeare r
the principal directions. One notices that for a given value of U1, the volume-
Equilibrium in the radial direction gives : tric strain is constant throughout the soil ring. Subs ti-
tuting.U1/R1 from equation (11), the volumetric strain
(3) of the soil is

in which o r am 0 8 are the principal radial am tan- or (Rl)

£v = _K_l__ _ (13)
gential stresses,
The correspoming strains E r and E8 are related to
the radial displacement u as follows: Comparing the volume of the cavity Ve to the volume of
au the soil ring V, one obtains :

{:: (4)
V/Vc ~ -(1 - R2/R1)
2 2
(14)

Considering the problem of an infinitely long lime pile, Since the volume decrease of the soil ring is equal to
we shall assume a plane strain comition : Ez ~ O, thus the volume increase of the cavity, the volumetric defonna-
the volumetric deformation Ev is equal to: tion of the cavity Eve is given by the simple relation
V
EV • Er + E:8 (S) £ £v·v or(Rl)/Kl (15)
vc
C
We shall further assume that the stresses am deforma- Thus K1 is the stiffness coefficient relating the
tions are related by the theory of elasticity, the elastic radial stress at the interface of the lime am the soil
behaviour of the material being fully described by : to the volumetric deformation of the cavity.

E : Young's modulus Under drained comi tions, K1 will correspom to the

V: Poisson's ratio final stage of the equilibrium between the lime expansion
(A + 2G) Er+ A E:8 (6) am the soil reaction. If we call D the horizontal cons-
[°r • trained modulus of the soil or the inverse of the coeffi-
08 • (A + 2G)E 8 + A £ cient of volume change as measured in a consolidation
r
with
test <.z)'
we fin! that
E. (I - V) I D (16)
(l +v) (l-2v) ·
G • E/2 (1 + \I)
(7) l-R 2 /R 2 l-R 2 /R 2
EV /(1 +\1)(1 - 2V) 2 l 2 I

With this hypothesis, the equilibrium comition (3) Since the strength of the soil is limited, we mi ght
becomes a differential equation of the radial displacement expect that a plastic zone develops from the inner cylin-
U(r) or u (i ). der towards the outer boundary as the radial stress
o ( R1) increases.
a2u
-- +
au
0 (8)
ar 2 r ar If we assume the following simple failure criterion :

The integration constants are determined by the boundary

(17)
conditions. Assuming a regular hexagonal,: pattern of the
locations of the piles, we can consider that the mid-point

898
 9.5

corresponding to a purely cohesive· aoil, in which Qu is

the unconfined compression strength, the fundamental
equation (3) becomes
R
.. .
aa /•r + qjr - o (18) ·-:·
which leads to
= oRu + <I,,-2n ~
R
{19)
I SOLUTION 3
cr(r)
r I
in which oRu is the radial stress for the radial distan.ce
Ru, which defines the extent of the plastic zone. t~~ 2
Expressing that, at the inner boundary of the elastic
zone which is a cylinder of radius R,_,, the failure cotrli-
I
tion (17) is reached, an:! calling Ur the correspon:ling
radial displacement, one gets the relationship between
I
qu and Ru : I
u
r 2 i2 i
Ru iu
(20)
Ja 02 0,6 a.a 1.0
calling 1\, /R2= a, an:! substituing for U /R , from equa-
tion (ll), which becomes CTR = 2K 1 . U /R , roneu obtains :
u r u
Figure 2 : Comparison between elastic an:! elasto-pla9tic
1 - V 2 solutions.
Cl (21) lla
~ r
volumetric stiffness coefficient = llV/V (see eq.15)
The complete radial stress distribution can be worked
out in the following way : by assuming a value of a , one compressibility of the quicklime_ llVCaO/VCaO
obtains the value of CTRu through the relationship (21). - Zlar .
The solution is obtained in the outer elastic region with ratio of the compressibility of the slaked lime to
the equation : the quicklime.

a + G(l + R~//) We fire the equilibrium incranental pressure

r
(22)
OR + G{l + R2 /R2)
u " 2 u
an! in the inner plastic region using equation (19), which
l\a
r
(24 )

gives the plastic radial stress as a function of the ra-

dius of the cavity. in which µP is the ratio of the actual expansion to the
Fig.2 gives two examples of the radial stress clistribu- unrestrained expansion at the same stage of the hydratior
tion an:! comparisons with the elastic solution, thus not reaction.
taking any strength criterium into consideration. I t can
be seen from these examples that lohen the value of a va- If we assume that the increase in radial pressure at any
ries from o.4 to O .8, the radial stresses at the cavity, time can be expressed by
for this case where R 1 /R 2 = o. 2, differ by the same
amount (lip£ = fl 2 ) whether shear failure is taken into lla
consideration or :fot. r . dµ (25)
llor,t{µ)
If we further assume that the soil undergoes plastic de- = t 4lJ
formation at zero volume change, it means that the radial
displacement U1 will be equal to that corresporoing to
the elastic problem. It can therefore be concluded that R2
within the range of the plastic zone extension mentioned,
the elastic stiffness coefficient K1 is not significant-
R1 ...,
ly affected by the inclusion of a failure criterion in the
analysis. INITlAL
CaO SOIL
STAGE
Expanding cylinder

Let us schematize the situation as follows : ( fig .3). I

Initially, the soil an:! the quicklime are at a zero
stress con:lition. As a part of the available lime reacts,
it expands. If we call C1, the volumetric expansion
TENTATIVE CaO SOIL
under atmospheric con:li tions an:! µ the proportion of the Ca~~
quicklime which has reacted to the initial volume, we can
say that :
I I
(23) ~ J,o...l

if llVf is the free volume expansion corresporoing to the RESULTANT j

variation of µ.
i
In fact, due to the restraint prov ided by the soil, the
effective volume variation will be smaller. If we call Figur e 3 Expansion o f the Cylinder

899
9.5

we know at that time the total stress increment at the

interface between the lime pile and the surrounding soil
and also the resulting expansion, JJpCI •
Equations (24) arrl (25) have been applied to the results O.G
of experiment conducted by Kuroda et al (~) on pure lime
which was allowed a set volumetric expansion : JJpCJ. When 04
the lime is not in contact with the fixed boundary K1 = 0 > OJ
and 6 P = 0 and when it is, l/K1 = O.
w
1MPa1
The ?esulting pressure __ Pe is derived from equations
z 0.2
c1 =

f
(24) and (25) 0.5 _ 0.2
4:
0:: 1
t- Ql C3 = 2 MPa
• dµ (/)
C5 = 0.2
p· z
0 .... KURODA AT AL
)JP (/) Q
z • KURODA'S RFEREN
(26) 4:
a..
X
UJ
The value of C5 should be a constant for any type of
0.04 006 0.10 0.20
lime whereas C1 and C3 depend on the grinding and on
the compaction. C5 was chosen on .the basis of relative EXPANSIVE PRESSURE Pe
pressures developed at various values of JJp.CJ • Figure 4 : Relation between the expansive pressure
For the experiment described by Kuroda et al (6), the and the expansion strain.
pressure generated after 2 hours corresponded -to
c1 = 0.7 and c 3 = I MPa-1: whereas another type of velocity Vh of water upon a radial gra'.l.ient i 9h/ 3r ,
lime, reported in the same paper, yielded the values of being the piezometric level of the water.
c 1= 0.5 and Cr 2 MPa-l(see fig.4).It can be seen that the
agreement is _good between the values predicted according
to the theory developed herein and the measured values.
As 6orJ JJ ) is a radial stress increase due to the expan- Kh hydraulic corrluctivity in the horizontal
sion of 'the lime pile, we know the effective stress state
towards which the soil is moving, when it is fully drai- If the volume of water expelled during a time incr ane nt
ned, if the value of Ki is that corresporrling to drai- corresponds to the volume decrease of the soil elemen t it
ned conditions. is flowing in arrl out of, one obtains the equation of ra-
The pore pressure generation l:opw corresponding to the dial consolidation
radial stress increment must be known in order to analyze
the tendency of the water to flow into the lime. A general
expression for the pore pressure for our system is ~2~ + -r ah (32 )
or ar
(27)
Since Er + Ee is constant throughout the soil ring , t he
tn which A is the ratio of the pore pressure generation to equation becanes similar to that corresponding to the ver-
the deviatoric stress increment. tical consolidation resulting from ra:Hal drainage
For the consolidation test (E e = O), A = 1 whereas for
the triaxial test (Ee i- O ) , A = 1/3 according to elastic a2h l .dh 3h
theory. We assume that an expanding lime pile is an -:-7 + - - (3 3)
intermediate condition and A = 1/2 : 3r r cl r ch cl t

(28) but in which

This hypothesis leads to the conclusion that the pore

pressure generation is proportional to the volume incre- Ch =
¾·°' + G) Kh
(34)
mental deformation, and is constant throughout the soil y
mh [ 2(1 - v )] y
ring : using elastic theory, it can be shown that where Y is the unit weight of water (9810 N/m 3 ), and
mh is the coefficient of volume change as obta ined
(29) with a consolidation test on a sample loaded
the horizontal direction.

Asswning that v = 0,5 in undrained con:l.itions, this

leads to a simple form of t~e _. _equilibrium _corrlition Rate of slaking

(30) From the radial pore pressure distribution resul ti ng

from the consolidation calculation, one gets the flow rat e
of water Q H entering per unit length of lime pile.
O
2
Radial consolidation dh
QH O = 21r.Rl .Kh. ( ;:J r \ (35 )
The pore pressure which will be generate:1. through uni- 2
form increments corresponding to the variation of the
boundary conditions (eq. 30) will dissipate as water is
expelled towards the pile. To examine this problem, we If we recall that C2 is the relative volume of wa ter
shall assume again that the soil is elastic, an:l. that the needed to react with quicklime, the volume of lime
vertical strain E2 = 0, reacting during a time increnent dt is :
Darcy's law is applicable_ . to describe the horizontal

900
 9.5

the initial values of µt and µ P, t are O.

µt+6T µt + 6 µ
t
t.a
r,t
l: u
t = 0 g,t
Therefor", equa.tion (36) gives the upper li.mit value of
the reaction parameter, µp,t • t'IC\,/CKI .Cl)

' -!:" -J:,,~.,j,c,o "" 4. Consolidation with pore pressure generation takes place
untilµ= 1. When this value is obtained, all the lime
has reacted and , Ug, t = 0
6µ =0
the lower limit being set by the speed at which the sla-
king can proceed, From available infonnation, it appears The final resultant expansion is given by µ p•Cl and
that the build up of the internal pressure is not instan- the radial consolidation continues thanks to vertical
drainage through the lime column.
taneous. If we call Vlim the constant limiting speed of
the reacting front separating slaked from quicklime in a
mass of lime in excess water, we can deduce from published
experiments (.§.,l) a consistent value of about O, 02 CALCULATIONS
mm/sec.
A set of examples has been calculated in order to inves-
The speed of the reacted front being limited, the allow- tigate the realism of the model chosen and to evaluate the
able water at the pile interface will be expressed by : influence of the various parameters involved in the analy-
sis,
vmax = vT--=---ii · Vlim (38) This parametric stmy included four types of soil canbi-
ning the following properties
since the radius of the cylindrical front decreases as the
reaction is taking place. E = ,4 MP a ard E = 4 MPa
This means that the hydraulic boundary condition at the Kh = 10-7 m/s and Kh = 10-9 m/s
drain will be h = 0 provided the gradient does not exceed
a maximum value. If this value is exceeded with the nonnal Poisson's ratio was kept constant : v = 0.4.
drainage, then the condition on the maximum velocity is
applied. The corresponding values of the coefficient of radial
consolidation were :

7.3 10-6 , 7 .3 7.3 10-8 and

ALGORITHM
-7 2 -I
7.3 10 ms •
The algorithm to calculate the phenomena as a function
The two lime types taken into consideration had the fol-
of time is based on the following steps :
lowing characteristics
Lime 1 Lime 2
1. An initial pore pressure distribution is chosen
C1 = 0.7 C1 0.5
Pw = 0 for O< r.;; R1 for t ;;.O
C2 0.3 C2 0.3
Pw = Dp 0 for R1~,;; R2 for t = O C3 l MPa-l C3 2 MPa-l
2, Consolidation according to equation (33) is calcula- C5 = 0.2 C5 0.2
ted during a time interval f-t, using the method of
The limiting speeds of the reaction front were taken as
finite differences where 6R is the difference
2xl0-S and 2xl0- 4 m/s.
between radial nodal points
The initial pore pressures Dp 0 considered in the soil
U.
lrl,t
-2U. +U. 1 u. -u. were .1 and • lxlo- 3 MPa.
U = U. +C ( i, t Hf: +- l."'l: 1.,~ + ug,t Four sets of pile radius and radius of influence were
i,t+lt 1,t h 6 Rz
Ri M investigated R1,R2 = 0.2, 0.6/ 0.2, 0.8/ 0.2,1,60/
0.1,0.8 m.
Fig. 5 shows the results of a sample calculation,
in which perfonned with a radius incrE!llent of 5 cm and a time step
of 51.5 s, corresponding to the following c ase :
Ui t is the pore pressure at time t at radius
R1'+ (i - l).6R
R1 = 0.2 Vlim = 2 10-S m/s E = • 4 MP a
R2 = o.s
U~,t is the pore pressure generatei during the
Dpo = .1 10-3 MPa Lime type I K = I0- 7m/s
time lt. h
The horizontal lines correspond to 6or, tor the und raine:i
3. Ug,t is computei in the following way pore pressure, whereas the curved lines correspond to the
u = cl .t.lJ pore pressure as a function of the radial distance,
It can ~e seen on fig.5 that the lateral contact pressure
g ,t+6t C} I + lJp,t ' CJ + (I - 1\)c5] + l/K 1 g eneration amou nts to . 5 MP a and t h at two days are n!!ces-
sary for the lime to react completely.
in which 6µ = 6Q/C2,Vc
The general treo1s revealed by the results of this parame-
where 6Q is the smaller of the two expressions
tric s tm y are as follows :
u - u
211Rl .!rt. ( 2, t 1, t) i'I - the lateral contact pressure increase corresponding to
Y 6R t
the full reaction of a type of lime deperds mainly on
the modulus of elasticity of the soil, but only margi-
nally on the value of Rz/R1, provided it is larger
than 3.

901
9.5

IMPa A<i"r,t presented,

-+++-:-t-+- The main parameters to describe the expansion beha..;
t= 2.10 4 s viour of a lime pile are the coefficient of free vo ! ume.-·
,-5
-;:-,~~- tric expansion and the canpressibility of the sl.aked
lime. This model has been successfully applied in ~ par'-',,,
ticular case to reproduce the results of LabPratory :
t e sts done in Japan,
Another parameter of importance introduced in this
.4 analysis is the speed of the reaction front. It increa-'- ·
ses with increasing fineness of the quicklime.
A parametric study has been comucted using differertt
t = s.10 3s pile diameters am spacings, different types of so ils
,3 am different types of lime. The main conclusions to be
drawn from the results are that

- the permeability of the soil to be treated has t o ',b e,

sufficient, (Kh > 10-7m/s) otherwise, quite h;!gli
,2 pore pressures are developed, especially if the grimi,iig .,,
of the lime is fine.
- the final expansion of the lime does not depenl very,
much on the pile spacing, which means that the relative
volume reduction ( or soil improvement), is uniform all
.1 inversely proportional to the influence area of a pil.e, ,
t:103S-
,- - In g;ost of the cases ex,u11ined, the lime has canple te.l_y,
reacted after two days. · ·· ' ·
- the conlitions of canpaction of the lime in-situ are pf
r utmost importance.
0 .2 .4 .6 .8 (ml REFERENCES

Figure 5 Pore pressure distribution B RCMS B ,B., Lime Columns in Theory am Practice, Proc.
Int. Conf. of Soil Mechanics, Reunion Cormemorativa, .
- this lateral pressure build-up i nc r eases with the fre e Mexican Soil Mechanics Society, Mexico City, August /
volumetric expansion coefficient C1 but decreases with 1982, pp. 149-165. ·,
increa sing compressibility C3.C5 of the slaked lime, (1) YAMANOUCHI T., MIURA N., MATSlJBAYASHI N., ab4
This means that the density of the quick lime in situ is FUKUDA N. , Soil Improvement with Quicklime am Filt.er
very important as a higher compaction will increase C1 Fabric, J. Geot. Eng. Division, ASCE, Vol.108, No GT7,
and decrease C3, pp 953-965, 1982.
AZIE v.J. Fundamental mechanics of calcination am
- the time required for the complete slaking of the lime hydration of lime and methods of control. Symposium O'll .
is not dependent on the initial pore pressure, provided lime AS'!M, MAR 1939.
Kh = 10-7 m/s, For Kh = 10-9 m/s, the cor- VESIC A. (1972), Expansion of cavities in infinit:e '
responding time is larger. It is greatly reduced when mass of soil. A.S.C,E, Journal of S.M.F,E., Vol.9S:, ,
the limit speed of the reaction front is increased, es- N°SM3, MAR 1972.
pecially when the soil possesses a large permeability. (1) LAMBE T. W. & l\l!I'lMAN R, V. Soil Mechanics, J. Utley
This means that if the lime is more finely ground, the chap 12, (1969). ·•·· ,'.
time required for the total reaction will decrease, (.§.) KURODA E. , OHASHI Y. & TSUYOSHI M. (1980), Expansive '
unless the soil is very impervious, pressure of quick lime pile. Research Institut e of
Technology, Tokyu Construction Co, Ltd, Japan (1980).
- when the initial pore pressure equals 0.1 MPs, the time (].) CHIU K.H. & CHIN K.Y. (1963), The study of imprcwi,gg
required for the complete slaking, to occur does not de- Bearing capacity of Tapei Silt by using quicl<Hme
pend on Ch, unless it is smaller than 10-6 m2s- 1 • For an piles. 2nd Asian 'Regional Conference in S. M.F.!'l.
initial pore pressure equal to 0, 1 kPa, this time in- Tokyo, Vol. 1, pp 387-393, (1963).
creases when the value of the coefficient of consolida-
tion decreases. This means that an initial pore pressure
as induced by a vertical load, for example, wil 1 speed up
the process.
- when Kh 10-7 m/s, the consolidation proceeds
faster than the build-up of the pore pressure. The maxi-
mlJIL pore pressure at the radius of influence is equal to
O. 3 to O .4 the undrained pore pressure increase. When
Kh = 10-9 m/s, the consolidation is slower than
the build-up of the pore pressure, which means that for
most cases, the maximum pore pressure obtained at the
radius of influence will be equa l to the undrained pore
pressure increase. Therefore, to avoid an excessive
build-u p of pore pressure, it is necessary for the soil
to possess a sufficient horizontal permeability.

CONCLUSIONS

A physical model represe nting the expans ion of a s la~

king li1ue pile in a n ela stic soil ring has been

902