Session 1/16
Some Engineering Properties of Residual Clay Soils Occurring
in Southern Brazil
Quelques propriétés techniques des terrains argileux résiduels du Brésil méridional
by M il t o n V a r g a s , Civil and Electrical Engineer, Professor o f Soil M echanics and Foundation Engineering, E scola Politécnica,
University o f Sao Paulo, Chief, Soils D ivision o f the “ Instituto de Pesquisas T ecnológicas” , Sào Paulo, Brazil
Summary
The purpose of this paper is to make a contribution to the
knowledge of the main features and characteristics o f residual clay
soils occurring in southern Brazil, where the thickness o f the over­
burden of decomposed rock often reaches fifty metres more.
Soil profiles of decomposed gneiss, basalt and sandstone, occur­
ring in that region are discussed. It is assumed that—from the en­
gineering point of view—all residual soil profiles can be divided in
three principal layers : (a) a surface layer of mature residual soil
with a high void-ratio and a low degree of saturation, referred to,
in this paper, as “ porous layer” ; (b) a young residual soil whose
main characteristic is to show the original structure o f the parent
rock; (c) a desintegrated rock layer that can only be removed by
explosives.
Grain size distribution, Atterberg limits, porosity, consolidation
and shearing resistance of residual clays are discussed. The concept
o f “ virtual pre-consolidation pressure”, in such soils, is introduced
and justified.
Rock Weathering and Residual Soils
A s defined by H olm es: “ W eathering is the total effect o f all
the various sub-aerial processes that cooperate in bringing
about the decay and désintégration o f rocks, provided that no
large-scale transport o f the loosened products is involved.”
There are three phases in the process: (a) désintégration by
physical or m echanical changes; (b) chemical weathering res­
ponsible for the decay o f the desintegrated blocks ; (c) process
o f evolution which brings an intensely decom posed rock to the
state o f an hom ogeneous residual clay or sand often not
saturated.
Fig. 1 show s cross-sections through decom posed gneiss in
two sites where earth dams are being constructed. The first
dam-site (Fig. la ) is near the crest o f “ Serra do M ar” (State
o f Sào Paulo) a site still covered with dense forest in a heavy
Sommaire
L’objet de ce travail est de contribuer à la connaissance des as­
pects et caractéristiques les plus importants des terrains argileux
résiduels du Brésil méridional, où l’épaisseur des couches super­
ficielles décomposées atteint souvent plus de cinquante mètres.
Les profils de terrains de gneiss, basalte et grès décomposés, que
l’on trouve dans cette région, y sont décrits. Il est montré que tous
les terrains provenant de la décomposition de roches peuvent - du
point de vue technique - être divisés en trois couches principales:
a) la couche superficielle de terrain résiduel mûr, avec un indice de
vides élevé et un faible coefficient de saturation, laquelle dans le
présent travail, sera réputée «couche poreuse»; b) une couche de
terrain résiduel jeune, dont le principal caractère est qu’elle conserve
la structure originale de la roche-mère; c) une couche de roche
décomposée qui ne peut être extraite qu’au moyen d’explosifs.
La distribution granulométrique, les limites d’Atterberg, la poro­
sité, la consolidation et la résistance au cisaillement des argiles rési­
duelles y sont étudiées. L’auteur introduit et justifie le concept de
«pression virtuelle de pré-consolidation» comme une des caractéris­
tiques de ces terrains.
rainfall zone (about 5 m per year). The second (Fig. 1 b) is
a deforested zone (deforestation took place about one hundred
years ago) also near the crest o f a lower range o f m ountains
in the State o f R io de Janeiro. R ain precipitation is about 1 m
per year in this last site. The elevation above sea level o f the
first site is about double that o f the second.
The process o f decom position in the first site is at an earlier
stage than in the second; the decom posed overburden (in­
cluding desintegrated rock and residual soil) has a maximum
thickness o f about 30 m in the first and 75 m in the second site.
The average relation between the thickness o f residual soil and
desintegrated rock is 1/3 in the first and 4/1 in the second site;
finally the mantle o f residual clay (soil which has suffered an
evolution) is much thicker in the second site than in the first.
67
 From what is shown in Fig. 1 we can deduce that in the
decom position o f gneiss in Southern Brazil there is a tendency
towards increase in the thickness o f the residual clay top-soil
and o f the total decom posed overburden and towards decrease
in the thickness o f the disintegrated rock layer which overlies
sound rock. This confirms the follow ing interpretation o f the
rock weathering process: first there is a fracturing o f the rock
and then the chemical decom position o f the mineral consti­
tuents along fractures and clivage planes starts. The process
advances giving origin to the layer o f “ desintegrated rock ”
which in the beginning is formed by blocks and boulders and,
gineering point o f view— in three principal layers: (a) just
overlying sound rock, a desintegrated rock layer which can
only be removed by explosives; (b) a young residual soil layer
w hose principal character is to show the original structure o f
the parent rock (the hardness o f this material is extremely
variable: from very soft soils removed by spade to hard con­
cretionary mass o f gravel, sand and clay needing pick or ex­
plosives to be removed); (c) a superficial mantle o f mature
residual soil which includes the thin hum ic top-soil layer and
a redish or yellowish clayey or sandy soil with a high void-
ratio and a low degree o f saturation.

S o u n d r o c k (g n e / s s )

à ) — ^/oung d e co m p o se d so //
LEGEND:
• T o p -s o i/ w i/ / h um us Ctm /hickness)
[] M àiu re residue/soi/ H . ■ P orous ye//ow o r re d /s /r c/ay.
R esidua/ s o i/ fro/n
idua/cc/ay
- Resic/ua/ /a y oorr c/ayey
c/ayey sand,
sand, £
s /o w in g > d e co m p o sifio n
----- ... . . , ./ r s/ruc/ure
s r r u c / u r e of
o f paren/
p a r e n / rock.
ro c k . o f ro c k
I lOl/ng restaur/SO// \2-C/ay or c/ayeysand d w
w i/h
i/h pebb/es
pebb/es a a,n d
I
6/ocks
b/o cfts of
n/
o f rock
ro c m sshow
/•/V- /c*
/ ~ ing s/ruc/ure
'------—
o f p a re n / ro c k .

/fo rizo ntef s c # /? :

ro c k (g n e is s )
O 25m
b) Ma/ure decom posed so i/

Fig. 1 C ro ss-S ectio n s through D e co m p o se d G n eiss

S ection s transversales de terrains de d éc o m p o sitio n de g neiss

later on, by boulders bound together in a matrix o f residual
soil. The process continues until the rock has been transformed
into a peculiar type o f soil. In the case o f igneous rocks the
feldspars and som e micas, under the action o f slightly acid
water, gives origin to the clay fraction o f the residual soil and
the quartz crystals form the sandy fraction o f the soil. In the
early phases o f weathering there are many intact feldspars and
mica crystals in the sandy fraction but, as those crystals are
being decom posed, an increase in the clayey character o f the
residual soil is observed. The structure o f the parent rock is
preserved until a very advanced stage o f decom position is
reached.
After throroughly decom posed, or sim ultaneously with the
final stages o f decom position o f the rock, a new process o f
transformation o f the residual soil begins, which is referred to
here as “ evolution o f the residual s o il” and which gives origin
to the yellowish or reddish superficial soil very com m on in
Southern Brazil.
From what was said, soils originated from the decom posi­
tion o f rock in Southern Brazil, can be divided— from the en-
P o ro u s R e sid u a l C la y o f S o u th ern B razil
In the highlands o f Southern Brazil that superficial layer
reaches tens o f metres o f thickness, giving rise to many troubles
connected with building foundations, earth dams and cuts,
due to its high com pressibility and permeability.
Lixiviation and oxidation occurs in that layer with a lateritic
process o f evolution o f the soil which gives to it a peculiar
redish color. Lixiviation causes soil porosity to increase co n ­
siderably; in m any cases the volum e o f the voids reaches about
60% o f the total volum e. A s its water contents are low , there
is not enough water to saturate the voids. This is the cause
o f its high compressibility.
Underlying this layer there is often a hard clay layer where
the lixiviated elem ents o f the superior horizons precipitate;
this causes an increase in the soil density resulting in an un­
usual com pactness or hardness. There is also an intense oxida­
tion resulting in a peculiar form ation o f lim onite concretions.
Evolution is the principal cause o f increase in thickness o f
the “ porous clay layer” . In young soils, it is alm ost absent

6 8
and, consequently, no hard layer occurs. M ature soils, in turn,
processes a very thick “ porous clay layer” .
Fig. 2 show s as an exam ple, a soil profile in a zone o f de­
com posed basalt, which is very com m on in the interior of
Southern Brazil. There is, under a 2.5 m thick fill, a “ porous
cla y ” layer o f 5 m thickness; underlying that layer there is
a stiff clay layer and finally a desintegrated basalt layer over-
lying sound rock. The figure show s also results o f grain size
analyses o f the several layers. By these analyses we can see
that the clay content decreases with depth, in accordance with
Penetration res/s/ance
M ows/' / f o o f p e n e f-ra fio n
IP 20 30
Om _ Void ra tio and
G ra /n S iz e moisture co nte nt'
= F ill A n a lyse s: £ =2OS ; 0J - 32%,
grains were observed to be fragments o f intact or incipiently
weathered rock; the sand fraction was constituted, in all cases,
by intact crystals o f the mineral constituents o f the parent rock
as indicated in Fig. 3.
Fig. 4 show s a plasticity-chart where 589 liquid-limit and
plasticity index pairs o f values obtained from Atterberg tests
on residual clays from Southern Brazil were plotted. Each
point was classified according to the nature o f the parent rock
in zones limited by two inclined parallel lines and tw o vertical
lines. Each o f these zones covers the total range o f occurrence
for soils originated by decom position o f four types o f rock.
In each o f these zones the regions where the m ost frequent

28% fin e sand go/ id A ir Wafer

LL 8 2 : P I 29
Porous
5m ' 1 — £=¿76 ; Cü-2&%
m .
3/%> fin e S3nd S o /id A ir V/ater
U S3; P I !0

32% c /a y Y /////A ^ l
32% s i / t S c / / d /,/• Wafer
36% s a n d
L L S 3 ;P I 9

h-0 50 , 60
L iq u id L im ii C '/ ° ) —

F ig. 2 S o il Profiles in D esin teg ra ted B asalt Fig. 4 R esid u al P orous C lays— P lasticity C hart
Profils du so l en terrain de basalte d éco m p o sé A rgiles résiduelles p oreuses - D iagram m e de p lasticité

w hat w as previously said. It show s also graphs indicating the
relative volum es occuppied by solid grains, water and air, in
the several layers. The voids are larger at the surface o f the
soil and decrease with depth, but the saturation increases with
depth.
Grain-Size Distribution, Atterberg Limits, and Porosity
o f Residual Clays
Fig. 3 show s grain size distribution curves o f soils for the
decom position o f ten different parent rocks. The coarser
values fall are indicated by dashed lines. It is interesting to
notice that alm ost all soils plot below the /4-line and show a low
plasticity. This fact can be understood by remembering that, in
all the studied soils, the clay fraction was o f the kaolin group.
Fig. 5 presents the variation with depth o f the Atterberg
lim its, moisture content, grain-size distribution and porosity
o f a residual clay from the decom position o f a clayey-sand-
stone, from Campinas (State o f S. Paulo). The superficial
layer o f porous sandy-clay, with a constant and low water
content, over a layer o f stiff clay is clearly shown. T he grain-
size distribution is alm ost uniform but the porosity varies with

1
1 ofrock -
70

î / Rock
I
v so C layey san dston cfCam pinas)
B a sa // ( Londrina)
^S 40
/ G n eiss f R ib eira o do Campo)
Schisi( Cubatëo)

£sl
Granité ( M&ndaqji)
G ranife( Itapevi)
Gneiss (Vi/a Anas^acioJ
GneissCCorrego Zïgario)
L im e s Z o n e f P e r d s ) F ig. 3 G rain S ize D istribution or
S c / rfs - / Ç ^ e r d s jj R esidual C lays
D istrib u tion granulom é-
OJ / 10
trique d ’argiles résiduelles
o f G ra in , D f m m )

69
 depth. It has an average value o f 55% in the upper layer and
only 40% in the lower one. A lso, as the water content is
alm ost constant, the air content is high in the upper layer and
alm ost absent in the lower layer.
High porosity and non-saturation o f residual clay only occur
Virtual Pre-Consolidation Load in Residual Clays
The concept o f pre-consolidation load, introduced by
A. Casagrande, is plainly defined and explainable in the parti­
cular case o f sedimentary clays. It has no meaning on residual

Afterberg Limits
and Moisture Confe/rffy.) Grain Size Distribution f a ) Porosity C/ i )
25" SO o so /oo O 50 lo o

Sandy
porous c/ay
yellow,
re d c
and
brown "S'
1
I
S tiff *
sandy clay,
yellow
and red

'5 -, Compact
coarse Sard

Fig. 5 P o ro u s R esid u a l C lay (C am p inas) from D e co m p o sitio n o f a C layey S an d ston e. V ariation o f C on sistan ce, G rain-Size D istrib u tion and
P o ro sity w ith D ep th
A rg ile résidu elle p o reu se (C am p inas) p rovenant de la d écom p osition d ’un grès argileux. V ariation de la con sistan ce, distribu tion gra-
nu lom étriq u e et p o r o sité selon la profon deur

in superficial layers. Fig. 6 show s the sam e variation o f Atter-
berg limits, grain-size distribution and porosity with depth for
a layer o f residual clay from the decom position o f gneiss, never
exposed to the atmosphere. In that case the evolution which
brings the residual soil to the porous state could not progress
and the porosity is alm ost constant with depth.
Similar distribution o f Atterberg lim its and grain-size distri­
bution was observed in several borings where such observa­
tions were made. The occurrence o f the porous state was ob­
served in superficial layers o f granite, gneiss, schist, sandstones,
clays. H owever, consolidation tests show that there is, for the
latter clays, som ething similar to the pre-consolidation pres­
sures. T hat is, up to a certain load the decrease in void-ratio
by consolidation is very sm all; after the applied load exceeds
that limit, the relation between void-ratio and applied pressures
follow s the consolidation law. This critical load which lim its
both behaviours o f clays can be determined by Casagrande's
Sedimentary
c/ays
. . W .L.

claystones and basalt. A lso the superficial layers o f tertiary £ 6,15

Sa n d and A tte rb e rg L im its G ra in S / z e
clays suffer the sam e phenom enon and the result is a porous g ra v e t a n d M oisture C onte# /fâ) D is trib u tio n (° / i) P o ro sity ('/ • )

clay w hose properties are very similar to those originated from . 25* 50 „r .0 SO 100 .o so too
igneous and sedimentary rocks. rCJay tr<action A,
1
>
Consolidation and Shearing Resistance o f Residual Clay ) i Sotd
matter
Experience has shown that the decrease o f void-ratio with
} l
Sand
Water '

f/action
applied normal pressure in residual clays follow s the sam e law D > 50fj
w hich governs the consolidation phenom enon for sedimentary -W Xfradioi
Y& sû
clays. T hat is, there is a minor decrease o f void-ratio when
the applied loads are smaller than a certain value and when
V5
they exceed this value the decrease o f void-ratio is proportional ->
to the logaritm o f the relation between the final and the initial 1
pressure. 1
Experience has shown also that the shearing resistance o f
a sam ple o f residual clay is a function only o f the void-ratio 1
in the m om ent o f rupture, whatever the type o f test, provided
that the samples submitted to the test had all the sam e degree Fig. 6 R esid u al C lay (B elo H o rizo n te) from D e co m p o sitio n o f G n eiss.
o f saturation at the m om ent o f rupture. V ariation o f C on sistan ce, G rain -S ize D istrib u tion and P orosity
w ith D epth
Shearing resistance o f residual clays w as treated with som e A rgile résidu elle proven an t de la d écom p osition du gneiss.
detail by the author in another paper submitted to the 3rd V ariation de la con sistan ce, distribu tion granulom étrique et
C onference on Soil M echanics and Foundation Engineering. p o ro sité selon la p rofon d eu r

70
empirical m ethod in the sam e way as for the sedimentary clays.
U nder such consideration we decided to call this limit load,
“ virtual pre-consolidation pressure” . The explanation o f the
existence o f such value can be m ade as fo llo w s: Let us suppose
a certain superficial portion o f a rock decom poses and trans­
form s into a uniform and hom ogeneous soil. Such a soil will
have a certain void-ratio, w hich depends entirely on the
nature and structure o f the parent rock. If we remould this soil,
increasing its water-content until slightly above the liquid limit
we, certainly, will obtain a sam ple with a void-ratio eLL su­
perior to e0. If we run a consolidation test on this sam ple we
shall obtain a consolidation curve similar to A ' B ' (see Fig. 7).
If w e com pact the soil in its optim um m oisture and maximum
density and then carry out a consolidation test we will obtain
a curve like A" B ". Finally if we remould the clay at its natural
water content, w e shall have a consolidation curve, intermediate
between A' B ' and A" B" (see curve A " 'B " ' in Fig. 7).
If, however, w e run a consolidation test on an undisturbed
sam ple o f the residual clay soil w e w ill see that the resulting
—( Voie/- ratio in th e fo rm a tio n
o f c/a y )
R e m o ld e d a t th e liq u id
tim /t-
Undisturbed
fP e m o td e d a t
na tu re! water
c o n te n f.
C o m p a cted <st
optim u m m o is tu re .
Fig. 7 C o n so lid a tio n Curves fo r R esid ual C lays
C ourbes de c o n so lid a tio n pour argiles résidu elles
curve A B cannot coincide with any o f the three above m en­
tioned curves (except in extreme cases with A'" B ’"). The
only possibility is a curve like A B, which w e shall call “ basic
consolidation curves” ; in connection with it we can observe
a “ virtual pre-consolidation pressure” defined as above.
For a shallow sam ple o f a residual clay, when the earth
pressure above the sam ple is less than the “ virtual pre-
V irtu a l P re - C onsolidation P re ssu re s ( Ag/cm 2) -~-
0 ____________ ___________ i ___________6
¡5
€
S'
“5>
I
r /o
F ig. 8 V irtual P re-C on solid ation Pressures A gain st D ep th o f Sam ples
Pressions virtuelles de p r é co n so lid a tio n dans le sens contraire
à la p rofon d eu r des éch an tillon s
consolidation lo a d ” , the consolidation test will show the real
“ virtual pre-consolidation pressures” o f the residual soil. But
for deep samples, when the earth pressure is bigger than the
“ virtual pre-consolidation pressure” , the pre-consolidation
pressures observed in the test will constitute a real pre-con­
solidation load equal to the earth pressure above the sample.
In the majority o f cases residual soils cannot be represented
by only one “ basic consolidation cu rve” . There will be at
least a basic curve for each layer o f parent rock o f different
kind. So there will be no correlation between the depth o f
sam ple and the “ virtual pre-consolidation pressure” above
a certain elevation. U nder such an elevation there w ill be
observed a pre-consolidation pressure similar to the classical
one, regardless o f the soil nature and properties (see Fig. 8 ).
71