You are on page 1of 10

Transportation Geotechnics 34 (2022) 100751

Contents lists available at ScienceDirect

Transportation Geotechnics
journal homepage: www.elsevier.com/locate/trgeo

Technical Note

Mechanical behavior and durability of a typical frictional cohesive soil from


Rio Grande do Sul/Brazil improved with Portland cement
Tennison Freire de Souza Junior a, *, Karla Salvagni Heineck, PhD b, Cezar Falavigna Silva c,
Francisco Dalla Rosa d
a
Civil Engineering, Post-Graduate Program in Civil Construction Engineering, Federal University of Rio Grande do Sul, Av. Osvaldo Aranha 99 – 3 floor, Porto Alegre, RS
90035-190, Brazil
b
Post-Graduate Program in Civil Engineering, Federal University of Rio Grande do Sul, Av. Osvaldo Aranha 99 –3 floor, Porto Alegre, RS 90035-190, Brazil
c
Consultant and Master’s in civil engineering, Post-Graduate Program in Civil Construction Engineering, Federal University of Paraná, Av. Coronel Francisco Heráclito
dos Santos n◦ 210, Curitiba, PR 81531-970, Brazil
d
Post-Graduate Program in Engineering, University of Passo Fundo, BR 285 – Km292,7, Passo Fundo, RS 90052-900, Brazil

A R T I C L E I N F O A B S T R A C T

Keywords: Lateritic soils are typical of tropical and subtropical climate regions, whose coverage is wide in Brazilian terri­
Lateritic soils tory. These soils are formed by the leaching of oxides and silicates and are found in unsaturated conditions with a
Shear strength high void ratio due to the presence of macropores and micropores. As a result, the soil has drained behavior and
Durability
high compressibility. Given these peculiarities, the use of cementing agents can be a suitable solution to promote
Shear modulus
modifications on the soil’s physicochemical properties to affect bearing capacity and stiffness. This study focuses
Portland cement
on the mechanical behavior of typical cohesive frictional soil in the city of Passo Fundo (Southern Brazil)
improved by Portland cement of high initial strength (Type III) based on the adopted amount of cement C (%)
and γd. Initially, unconfined compression strength tests (qu) were performed for cement contents of 3%, 6%, 9%,
and 12%, and γd equal to 16.4 kN/m3, 15.4 kN/m3, 14.4 kN/m3, 13.6 kN/m3, and 12.7 kN/m3, followed by
ultrasonic pulse velocity tests to determine the G0. Then, tensile tests were performed through Splitting Tensile
Test (qt) considering γd of 16.4 kN/m3, 15.4 kN/m3, and 14.4 kN/m3 for cement contents of 3%, 6%, 9%, and
12%. Finally, the improved soil was studied using durability tests under wetting/drying conditions with γd of
16.4 kN/m3, 15.8 kN/m3, 14.5 kN/m3 and cement content of 3%, 6%, 9%, and 12%. The results indicate that C
and γd are sufficiently significant factors in understanding soil’s behavior regarding wetting/drying of the
mixtures, in that order. Finally, normalized curves of qu, qt, G0 and LM and VR were developed as a function of the
adjusted η⁄Civ for practical purposes of dosages for ground improvement.

pavements, upstream slope protection for dams, protection of river


banks and associated grade-control structures, liners for channels, res­
Introduction ervoirs, lagoons, mass soil cement placements for dikes and foundation
stabilization [1]. Consoli et al. [8] proposed the shear strength
Soil improvement using binders such as Portland cement and lime normalization as a function of the η/Civ index (porosity/volumetric
are well-known alternatives in ground improvement to establish cement content). There are several studies already carried out with the
changes in the physical properties of the soil [34]. This practice has use of Portland cement improvement, such as in Botucatu sandstone [8],
become more common on soils with low load bearing capacity, high Osorio Sand [19], clays [7], different types of sandy soils [18], and the
compressibility or high permeability. lateritic soil of Maringa [15]. In general, these works study the me­
Catton [5], Dupas and Pecker [16] and Clough et al. [6] were pio­ chanical behavior of different types of cement improved soils, regarding
neers in the use of soil–cement. In geotechnical literature there are many the effect of the parameters ω, Civ and η, but there are no scientific re­
cases that support the application of this technique on foundations [e.g., ports about durability, mechanical strength and shear modulus of
[27,22,29,32,17]] as base material underlying bituminous and concrete

* Corresponding author.
E-mail addresses: tennisongeotec@gmail.com (T.F. de Souza Junior), karla.heineck@ufrgs.br (K. Salvagni Heineck), cezarfalavigna@gmail.com (C. Falavigna
Silva), dallarosa@upf.br (F. Dalla Rosa).

https://doi.org/10.1016/j.trgeo.2022.100751
Received 30 August 2021; Received in revised form 21 December 2021; Accepted 28 February 2022
Available online 16 March 2022
2214-3912/© 2022 Elsevier Ltd. All rights reserved.
T.F. de Souza Junior et al. Transportation Geotechnics 34 (2022) 100751

Nomenclature η porosity
η/Civ porosity/volumetric cement content
cp specimen P load applied
CPT cone penetration test D diameter
C cement content (percentage of cement mass in relation to Le cp height
mass of dry soil) LM mass loss
CID consolidate isotropic drained triaxial test γ specific soil weight
Civ volumetric cement content (percentage of cement volume Vs S wave velocity
in total specimen volume) g gravitational acceleration
qu unconfined compressive strength Δt S wave time
qt tensile strength DF degrees of freedom
R2 coefficient of determination SS sum of squares
γg specific gravity MS mean square
γd dry unit weight Fc calculated F
ω moisture content Ft tabulated F
ωo optimum moisture content VR volume ratio

national lateritic clays. As an alternative for soil–cement, there are also In a natural state, the studied lateritic soil collected in the first meter
lime application studies in the geotechnical field [36,4,21]). Years later, presented as physical indexes: γg equal to 26.7 kN / m3, γnat equal to 16.1
Consoli et al. [9] proposed a general model that controls the strength - kN / m3, ω equal to 33%, LL equal to 42%, IP equal to 11% (Similar to
porosity/binder index behavior of sandy soil-pozzolan-lime blends. [32], void ratio equal to 1.21 and saturation degree of 73%[31]. The
Many soil improvement studies with lime have been carried out in γd value for modified Proctor energy compaction is equal to 16.43 kN/
different soils, such as clayey sand [9], dredged mud [23], soft clays [25] m3. Regarding the granulometric distribution, Consoli et al.[12]re­
and silty sand [30]. A recent paper by Consoli et al. [13] proposed the ported that the soil consists of 2.0% medium sand (0.425 mm ≤ diam­
stabilization of lateritic soil from northern Bolivia with Portland cement eter ≤ 2.0 mm), 20.0% fine sand (0.075 mm ≤ diameter ≤ 0.425 mm),
and sand. 22.0% Silt (0.002 mm ≤ diameter ≤ 0.075 mm) and 56% clay (diameter
This article aims to understand the mechanical performance of a ≤ 0.002 mm).
frictional cohesive soil typical of Passo Fundo/RS improved with Port­ In undisturbed samples collected from the surface of the experi­
land cement with high initial strength (Type III) based on cement con­ mental field, the average value of axial compression strength was equal
tent C (%) and dry unit weight (γd). to 54 kPa for deformations of 0.7%.[14]. Regarding the effective pa­
rameters, Rizzon[31]performed CID tests on this same soil with 20 kPa,
Experimental program 40 kPa, and 80 kPa confining stresses and obtained as stresses envelope
parameters φ’ equal to 26.6◦ and c’ equal to 14,9 kPa. Consoli et al. [12]
The experimental program was carried out in parts. First, the performed triaxial CID tests, with 20 kPa, 35 kPa, and 50 kPa confining
research of the geotechnical properties and parameters of the soil in stresses, and obtained 30.5◦ and 13 kPa for peak friction angle (φ’) and
other works [e.g., [31,35,12] and its complement with a modified cohesive intercept (c’), respectively. Regarding the plastic behavior and
Proctor compaction test. Subsequently, the C% and γd were determined, the stiffness of the material, for confining stresses of 20 kPa and 35 kPa,
and specimens were randomly chosen to assess the G0. Then, unconfined the soil behaved similarly to a ductile material, however, for 50 kPa
compression strength and splitting tensile tests were carried out with confining stress, the material showed a hardening behavior.
varying C% and γd. Finally, wetting and drying durability studies were
done on the soils improved with Portland cement. Water and cement. The experiment was performed with distilled water.
The binder used was Portland cement with high initial strength (Type
Materials III), which has high reactivity at low ages depending on its grinding
degree.
Lateritic soil
Lateritic soil in the experimental field is the result of in situ weath­ Methods
ering processes and has characteristics related to these processes, such as
low cementation and high porosity. The studied soil has a cohesive- The properties of cemented soil are influenced by several factors,
frictional behavior and hydraulic conductivity equivalent to granular such as: soil type, cement and water content, degree of compaction,
soils – k = 1.1 × 10-5 m/s [12]. Under natural conditions, these soils mixing uniformity, curing conditions and age of the compacted mixture
have a low bearing capacity, however, when compacted they can reach [1]. Therefore, next, all the materials and procedures adopted in the
high mechanical strengths, being suitable for landfills in road experiment are stated.
constructions.
Faro [17] performed in situ CPT tests on the respective soil. The Molding and curing of specimens (CP)
results showed that in superficial layers (depths less than 0.5 m), the tip The soil preparation was carried out following common procedures,
strength was between 4 MPa and 6 MPa, as well as values in a range of 1 such as: removing roots manually, drying by artificial heating, soil clod-
MPa to 2 MPa for the depths up to 15 m that followed the more resistant breaking, sieving (sieve # 10), and determining the moisture content of
layer. The referred soil is classified as A-5–8 (Silty Soil) by the American the material before reserving it in plastic bags. The amount of cement
Association of State Highway and Transportation Officials – AASHTO was calculated as a function of the soil dry mass, and the amount of
(LL = 41% and IP = 9.02%), and as CL (Clay of low plasticity) by the water was calculated as a function of the total mass of dry materials
Unified System of Soil Classification – SUCS[32]. As for the tropical (lateritic soil with cement). The curing time established was seven days,
classifications (MCT), it belongs to the group LG’ – clay soil with lateritic similar to that used by Consoli et al. [8], Fonseca et al. [19] and Consoli
behavior[35]. et al. [10].

2
T.F. de Souza Junior et al. Transportation Geotechnics 34 (2022) 100751

For the unconfined compression strength test, the specimens were Pundit Test (Initial shear modulus)
molded in a cylindrical geometry and with dimensions of 5 × 10 cm Once removed from the immersion tanks, the specimens were dried
(diameter × height) based on Guedes et al.[20]. The materials were using absorbent papers applied to their surfaces in order to remove the
weighed and mixed until homogenization was achieved, then the excess of water. After that, three specimens were selected for Vs mea­
mixture was divided into three portions. The specimen was molded surements using the Pundit Lab® (Portable Non-Destructive Digital
through static compaction, and subsequently demolded, as well as the Indication Tester) equipment. The procedures for calibration, mea­
dimensions were registered. The specimen’s acceptability for the axial surement, and interpretation of the results generated applied to the
compression test and for tensile by diametrical compression met the specimens were based on studies by Bortolotto (2017). Fig. 1 illustrates
following aspects: the procedure for recording the propagation time of S waves (ts) in a soil
specimen improved with cement, which, through the S waves propa­
• Degree of compaction between 98% and 102%; gation speed was calculated due to the longitudinal distance (height) of
• The moisture content (ω): the specimen’s moisture content values the specimen.
obeyed a ± 2% variation with the optimum moisture content ob­ The shear modulus obtained from the Pundit is an idealized value
tained in the Modified Proctor curve of the soil (ω = 23.5%); and is referred as G0, [i.e. the modulus under very small strain][33].
• Specimen dimensions: diameter with 50 ± 0.5 mm and height with According to Schnaid and Odebrecht (2012), after obtaining the ultra­
100 ± 2 mm. sonic S waves propagation speeds, it is possible to calculate the shear
modulus through the elasticity theory. The formulation described in Eq.
After molded, specimens were introduced into plastic bags and the (1) is a classic equation of shear Modulus.
mass and dimensions were registered. They were subjected to 6 days of
γ Le
dry curing at 23◦ ±2◦ C. After this period, the specimens were immersed G0 = ρVs2 = ( )( )2 (1)
g Δt
in water for 24 h in order to reduce the suction effect, increasing the
degree of saturation (see Fig. 1a), and, finally, specimens were tested for
Durability test
unconfined compression strength. The moisture content value was ob­
The durability test for wetting/drying cycles was based on ASTM
tained by drying in an oven with a temperature of approximately 100◦ C
D559[2]procedures. Cylindrical specimens of 100 ± 0.4 mm in diameter
for a period of 24 h to 48 h after molding. For the durability tests the
and 127.3 ± 0.3 mm in height were molded and submitted to a curing
same procedures were adopted.
period of 7 days. The durability test was performed through 12 brushing
cycles with wetting/drying of the specimen. The water immersion time
Unconfined compression strength and splitting tensile tests
of the specimen was 5 h and the oven drying time was 43 h, totaling a
Unconfined compression strength tests help to evaluate the
total time/cycle of 48 h. The brushing cycles were performed 24 brushes
compressive strength of construction materials, such as wood and con­
(20 passes on the longitudinal side and 2 on each transverse side). The
crete. Both the unconfined compression strength test and the tensile test
weight of the specimen was measured in each cycle using a precision
by diametrical compression were adopted to verify the effectiveness of
balance and the volume was measured using a pachymeter at 3 points
soil improvement with the studied binders. In this way it is possible to
(transversal and longitudinal). Fig. 1b illustrates the specimens for the
analyze the effects of the influential parameters to establish the me­
durability tests produced before the brushing cycle and after wetting/
chanical behavior normalization as reported in other studies in the
drying procedures.
literature [e.g., [8,9,19.7]]. The tests followed the ASTM D 2166 [3]
standard, similar to the Brazilian standard NBR 5739 (ABNT, 1980),
Mechanical and durability tests
both simple, fast, reliable, and cheap. Diametrical compression tests
The program for unconfined compression strength and splitting
(known as the Brazilian test) were also performed to understand the
tensile testing was established to separately assess the influence of
tensile behavior of compacted mixtures.
cement content and porosity/cement content ratio (η/Civ) at the me­
An automatic load frame with a maximum capacity of 50 kN and test
chanical strength of the artificially cemented lateritic soil. The molding
rings with capacities of 10 and 50 kN, as well as resolutions of 0.005 and
points of the unconfined compression testing program are shown in
0.023 kN, respectively, were used for unconfined compression strength
Fig. 2, with the respective soil modified Proctor compaction curves with
tests. The displacement rate adopted was 1.14 mm/min. The curing and
90% and 100% saturation curves.
water immersion procedures followed the criteria in item 4.1. After both
It is known that the water needed for the soil–cement is in a range
tests, statistical studies were performed based on the ANOVA method in
between 5% and 13%, which would be enough to obtain the maximum
order to understand the significance of the investigated variables (γd, C
compaction and allow cement hydration at the same time, but to obtain
%) on the response variables (qu, qt and Go).
a satisfactory compaction, the optimum moisture content was adopted

Fig. 1. Laboratory procedures: a) mold and cure the specimens; b) wet and dry durability test; c) ultrasonic pulse velocity test (Pundit).

3
T.F. de Souza Junior et al. Transportation Geotechnics 34 (2022) 100751

Fig. 2. Lateritic soil compaction test with modified proctor energy.

[1]. However, the ω adopted in the mixture was equal to the ωo in the kN/m3, 15.8 kN/m3, 14.5 kN/m3 and cement contents C (%) of 3%, 6%,
modified proctor energy in order to reach the maximum γd of the 9%, and 12%.
mixture. Five γd values were established (16.4 kN/m3, 15.4 kN/m3, 14.4 Note: In addition to the twelve specimens established by the study
kN/m3, 13.6 kN/m3 and 12.7 kN/m3), which correspond respectively to matrix, other five specimens (reproduction of the specimens established
γd1, γd2, γd3, γd4, γd5 (see Fig. 2). Regarding the evaluation of the tensile in the test matrix) were produced because they did not meet the
behavior by Splitting Tensile, the γd values determined were 16.4 kN/ acceptability criteria, even though they were useful when interpreted in
m3, 15.4 kN/m3, 14.4 kN/m3. Specifically concerning Portland cement normalized curves.
of high initial strength (Type III), the levels adopted were 3%, 6%, 9%
and 12%, which is approximate to what is suggested by Mitchell [26] Results and discussions
and studied by Fonseca et al. [19] regarding granular soils.
For the purpose of statistical representation, five specimens were In what concerns unconfined compression strength, the tests carried
molded for each combination of C and γd range in axial compression out met the expectations based on previous studies with other granular
tests and for splitting tensile tests. For the shear modulus determination, and clay soils [e.g., [28,19,8,18]]. There was an increase in the ultimate
three specimens were randomly selected from the unconfined axial compression strength (qu) due to an increase in the dry soil’s
compression strength test set of samples. specific weight, and in the incorporated cement content adopted in the
Regarding the durability test under wetting/drying conditions, the study. In addition to these, soil particle size plays an important role in
same parameters were chosen, being adopted γd values equal to 16.4 shear strength behavior. Grain shapes and respective size distributions

Fig. 3. Unconfined compression strength variation in relation to the cement content introduced in the studied soil.

4
T.F. de Souza Junior et al. Transportation Geotechnics 34 (2022) 100751

can affect the interparticle arrangements in the compaction, and in­ contacts between particles, just as a greater compaction reflects the
crease or reduce the specific area for chemical bond development [18]. decrease of porosity and the increase in interparticle effective stress,
Considering these aspects, the high fine content presented in the used and, therefore, an increase in the soil’s shear strength. Consoli et al. [8]
soil combined with the presence of natural oxides and silicates may have concluded that the qt/qu ratio for artificially cemented sand was be­
been beneficial for increasing the qu value for a certain ratio of η⁄ Civ. In tween 9% and 14%. However, when comparing the qt and qu values of
Fig. 3, it can be seen that the gain in the ultimate axial compression (qu) the present study through the adjusted equations, it is evident that the qt
followed an exponential growth trend for the studied cement content for values represented 16% of the qu value.
each γ d, although a linear growth in the qu (kPa) vs C (%) relation is
common.
Similarly, the gain in unconfined compression strength on splitting Effect of the porosity/volumetric cement content (η/Civ) ratio on the
tensile due to the incorporated cement content range in the studied lateritic soil shear modulus (Go)
γd values followed an exponential growth trend, as shown in Fig. 4.
As for the shear modulus (Go), as seen in Fig. 7, the normalized curve
Go vs η/Civ0.35 presents a potential function configuration, so that the
Effect of the porosity/volumetric cement content (η/Civ) ratio on lateritic stiffness of the cement improved soil decreases as the η/Civ0.35 index
soil shear strength behavior increases.
The drop in material stiffness occurs in an associated way, that is, the
The data were grouped and interpreted according to the η/Civ material cementation reduction affects the number of interparticle
parameter. A unique correlation seen between qu versus η/Civ predicts bonds (adhesion) allied to the increase of the soil porosity (random
the improved soil strength behavior, whose power applied to Civ for arrangement of grains and volume void), which in turn is linked to the
best-fit correlation (R2 = 0.96) was equal to 0.35 as shown in Fig. 5. This magnitude of γd. Although there are some points with higher shear
power value is consistent to the value obtained by Consoli et al. [7] in modulus in the η/Civ0.35 range equal to 23, the relationship can be
studies with silty clays and sandy clays. Through this relationship, the considered optimal (R2 = 0.79), also seen in Fig. 7 and Eq. (4). This
increase in the η values results in the reduction of the qu value is evident, behavior can be justified by internal suction generation on the improved
while this affects the mechanical strength increase of the improved soil, soil due to not having reached the complete saturation stage during the
as demonstrated by Consoli et al. [8,10]. immersion phase of the specimen, thus increasing the Go value quickly
Like the unconfined compressive strength (qu), a normalized curve based on the studies Georgetti et al. (2013).
was established between the tensile strength by splitting (qt) and the ( )− 4.487
η/Civ ratio, whose R2 value was equal to 0.90, as shown in Fig. 6. As G0 = 4.0 × 109
η
(4)
cement incorporation associated with the soil densification increased Civ0,35
(increase in γd), the η/Civ0.35 parameter increase as a consequence Fig. 6.
Another justification is due to the increased degree of saturation as a
In summary, the data correlations present high quality for uncon­
result of the reduction in the void index. The densification of the
fined compression tests [Eq. (2)] and tensile by diametrical compression
soil–cement mixture allowed the expulsion of air from the interparticles,
[Eq. (3)], based on the R2 coefficients, equal to 0.96 and 0.90,
raised the water/voids ratio and increased the proximity between
respectively.
cement and lateritic soil particles, creating more chemical bonds.
( )
η − 5.457 Apparently, for levels of η/Civ0.35 ≈ 23 these phenomena generate sig­
qu = 9.0 × 107 0.35 (2) nificant increase in the value of G0.
Civ
The correlation between Go/qu and η/Civ0.35 of the studied mixture is
(
η
)− 5.457 evinced in Fig. 8 and Eq. (5), and a good relationship between these
qt = 1.44 × 107 (3) parameters is observed when analyzing it (R2 = 0.70). This relationship
Civ0.35
is useful to assist in the soil improvement dosage starting from the
Both unconfined compression and tensile tests show that the intro­ cement amount balance and the appropriate minimum density,
duction of cementing agents provides the development of chemical providing a mixture that meets the strength and stiffness required by the

Fig. 4. Tensile strength by splitting tensile variation in relation to the cement content introduced in the studied soil.

5
T.F. de Souza Junior et al. Transportation Geotechnics 34 (2022) 100751

Fig. 5. Unconfined compression strength variation in relation to the porosity/volumetric cement content ratio in the studied soil.

Fig. 6. Tensile strength by splitting tensile variation in relation to the porosity/volumetric cement content ratio in the studied soil.

project[11]. There are other works in the literature addressing Go/qu assessing the durability of the studied soil, and a minimum cement
with binder elements in clayey sands improved with lime [7], as well as content of 6% was then established.
in Osório/RS sand and in Porto silty sand improved with Portland In terms of durability, the test results indicated that the higher the
cement[11]. value of γd and C of the improved soil, the lower was the mass loss and
( )2.04 the volumetric reduction of the cps. This indicates that the cement
Go η content and density significantly affects the durability of the improved
≅ 0.91 0,35 (5)
qu Civ material. Fig. 9 illustrates the improved soil mass loss behavior as a
function of the η/Civ parameter and, as predicted, it was possible to see
Effect of porosity/cement volumetric content (η/Civ) ratio on the that there was an exponential growth in the improved soil mass loss as
specimen’s mass and volume variation in durability test the η/Civ0.35 value increased. Conceptually, high η/Civ values represent
high levels of porosity and/or less incorporation of chemical binders in
When carrying out the durability tests with mixtures which the the mixture, which consequently make the geomaterials more suscep­
cement content was 3%, particle dispersion occurred in the second tible to deterioration by abrasion. Allied to this, since the adjusted curve
brushing cycle, regardless of the degree of compaction. Apparently, the was considered to be excellent (R2 = 0.86), it can be inferred that the
low artificial cementation combined to the geomaterial densification did parameter η/Civ has the potential to predict mass loss behavior of geo­
not reach a certain level of strength to support the capillary stresses that materials. In this way, designers and engineers, knowing the necessary
occurred during the immersion phase of the specimens. As the speci­ mass loss durability criteria, can specify the improvement dosage for in
mens were submerged, there was a decrease in effective stresses and this situ soils. Next, Eq. (6) informs the curve described in Fig. 9:
allowed the dispersion of solids due to low cementation increase[24].
The conclusion reach is that a 3% cement content is inadequate for

6
T.F. de Souza Junior et al. Transportation Geotechnics 34 (2022) 100751

Fig. 7. Shear modulus variation in relation to the porosity/volumetric cement content ratio in the studied soil.

Fig. 8. Variation of the shear modulus and unconfined compression strength ratio in relation to the porosity/volumetric cement content in the studied soil.

( )1.469 ( )
η
LM (%) = 1.7 × 10− 3
(6) η
Civ0.35 VR (%) = 1.315e − 0.0012
(7)
Civ0,35
Fig. 10 illustrates the VR of the improved soil as a function of the
η/Civ parameter and, as predicted, there was a decrease in the volume of
the improved soil with the increase of η/Civ0.35 value. The value of VR Statistical analyses
represents ratio of the final volume to the initial volume of the cp. As
previously mentioned, when η/Civ presents higher values, there is a Based on the investigated variables, statistical analyzes were per­
tendency to abrasion degradation susceptibility and, in this case, this is formed through two-way ANOVA method. All the factor variations were
reversed into a volumetric reduction of the specimen due to the abrasion significant on the three statistical evaluations, since all the cases pre­
cycles. Equation (7) describes the VR behavior as a function of the sented Fc (calculated factor) values higher than the Ft value (tabulated
parameter η/Civ with an excellent relationship between its data (R2 = factor). As well, the p-value found was less than 5% in all sources of
0.84). variation, which leads to the rejection of the null hypothesis of equal
means. In the unconfined compressive strength (see Table 1) and tensile
strength by diametrical compression (see Table 2) tests, there was a

7
T.F. de Souza Junior et al. Transportation Geotechnics 34 (2022) 100751

Fig. 9. Relationship between LM (%) and porosity/volumetric cement content parameter in the studied soil.

Fig. 10. Relationship between volumetric ratio VR and porosity/volumetric cement content parameter in the studied soil.

Table 1 Table 2
Statistical evaluation of the investigated variables of unconfined compression Statistical evaluation of the investigated variables of tensile strength (qt) by
strength (qu) tests using the ANOVA method. splitting tensile tests using the ANOVA method.
Variable DF SS (Aj.) MS (Aj.) Fc Ft p- Variable DF SS MS Fc Ft p-
value value value (Aj.) (Aj.) value value value

C(%) 3 48.309 16.103 316.25 2.719 0.000 C(%) 2 1.174 0.391 125.8 3.191 0.000
γd (kN/m3) 4 35.478 8.8696 174.19 2.486 0.000 γd (kN/m3) 3 0.545 0.273 87.62 2.798 0.000
Interaction –γd (kN/ 12 13.157 1.0964 21.53 1.875 0.000 Interaction. γd (kN/ 6 0.185 0.031 9.89 2.295 0.000
m3) vs C(%) m3) vs C(%)
Error 80 4.073 0.0509 Error 48 0.149 0.003
Total 99 101.017 Total 59 2.054

major significant effect of the parameter C% compared to the significant shown in Table 3.
effect γd and the interaction (γd vs C%), as well as in the studies by Foppa
(2005) and Consoli et al. [8]. However, regarding the assessment of the Final considerations
shear module, the main significant effect was the γd parameter
compared to the significant effect C% and the interaction (γd × C%), as Based on the results obtained and the discussions presented, it was

8
T.F. de Souza Junior et al. Transportation Geotechnics 34 (2022) 100751

Table 3 References
Statistical evaluation of the investigated variables of shear module (G0) tests
using the ANOVA method. [1] ACI 230 (2009). Report on soil cement. 5, American Concrete Institute, Farmington
Hills, MI, USA.
Variable DF SS (Aj.) MS (Aj.) Fc value Ft p- [2] ASTM D559 (2003). Standard test methods for wetting and drying compacted soil-
value value cement mixtures. American Society for Testing and Materials, Pennsylvania, USA.
[3] ASTM D2166 (2006). Standard test method for unconfined compressive strength of
C(%) 4 1.35E + 08 33,712,130 1,563.19 2.606 0.000
cohesive soil. American Society for Testing and Materials, Pennsylvania, USA.
γd (kN/m3) 3 62,324,209 20,774,736 963.3 2.839 0.000
[4] Attoh-Okine NO. Lime treatment of laterite soils and gravels. Constr Build Mater
Interaction – 12 1.09E + 08 9,110,475 422.44 2.003 0.000 1995;9(5):283–7. https://doi.org/10.1016/0950-0618(95)00030-J).
γd (kN/m3) [5] Catton, M. D. (1962). Soil-cement technology—A résumé. Research and Development
vs C(%) Laboratories of the Portland Cement Association, Portland, OR.
Error 40 862,651 21,566 [6] Clough GW, Sitar N, Bachus RC, Rad NS. Cemented Sands under Static Loading.
Total 59 3.07E + 08 J. of the Geotech. Engineering Division, ASCE 1981;107(6):799–817. https://doi.
org/10.1061/AJGEB6.0001152.
[7] Consoli, N. C.; Dalla Rosa, A.; Corte, M. B.; Lopes, L. S. Jr; & Consoli, B. S. (2011).
possible to state the following considerations: Porosity/cement ratio controlling strength of artificially cemented clays. J. Mater.
Civ. Engng., 28(8), 1249–1254. 10.1061/(ASCE)MT.1943-5533.0000283).
[8] Consoli NC, Foppa D, Festugato L, Heineck KS. Key parameters for strength control
a) An excellent exponential relationship (R2 = 0.96) was obtained be­ of artificially cemented soils. J Geotech Geoenviron Engng ASCE 2007;133(2):
tween the unconfined compressive strength (qu) and the η/Civ0.35 197–205. https://doi.org/10.1061/(ASCE)1090-0241(2007)133:2(197)).
[9] Consoli NC, Lopes Jr LS, Heineck KS. Key parameters for strength control of lime
parameter, indicating that the power value equal to 0.35 represents a stabilized soils. J Materials Civil Engineering ASCE 2009;21(5):210–6. https://doi.
satisfactory adjustment in the prediction in terms of compression org/10.1061/(ASCE)0899-1561(2009)21:5(210).
mechanical behavior, as in Consoli et al. [7], [10] Consoli NC, Lopes Jr LS, Prietto PDM, Festugato L, Cruz RC. Variables controlling
stiffness and strength of lime-stabilized soils. J Geotech Geoenviron Eng 2011;137
b) An excellent exponential relationship (R2 = 0.90) was obtained be­ (6):628–32. https://doi.org/10.1061/(ASCE)GT.1943-5606.0000470.
tween the tensile strength by qt and the η/Civ0.35 parameter, indi­ [11] Consoli NC, Fonseca AV, Silva SR, Cruz RC, Fonini A. Parameters controlling
cating that the power value equal to 0.35 represents a satisfactory stiffness and strength of artificially cemented soils. Géotechnique 2012;62(2):
177–83. https://doi.org/10.1680/geot.8.P.084.
adjustment in the prediction in terms of tensile mechanical behavior,
[12] Consoli NC, Giese DN, Scheuermann Filho HC, Festugato L, Rocha MM,
as in Consoli et al. [7]. In addition, the value of qt represented 16% of Heineck KS, et al. On porous bonded residual soil in natural and dynamically
the qu of the soil; compacted states through plate load tests. J Geotech Geoenviron Engng ASCE 2020;
146(8):06020011. https://doi.org/10.1061/(ASCE)GT.1943-5606.0002321.
c) As well as on the results of unconfined compression strength and
[13] Consoli NC, Párraga Morales D, Saldanha RB. A new approach for stabilization of
tensile strength by diametrical compression, the Go showed good lateritic soil with Portland cement and sand: strength and durability. Acta Geotech
quality in the exponential relation (R2 = 0.79). Also, it was possible 2021;16(5):1473–86.
to establish a normalized curve between G0⁄qu vs η/Civ0.35 with good [14] Dalla Rosa F, Thomé A, Ruver CA. Determinação de alguns parâmetros geotécnicos do
campo experimental de Geotecnia da Universidade Passo Fundo [in Portuguese].
quality (R2 = 0.70), being very useful for soil improvement dosages Simpósio Brasileiro de Jovens Geotécnicos 2004;1:2004.
that consider aspects such as stiffness combined with unconfined [15] De Souza JM, Rudnick RC, Lukiantchucki JA. Evaluation of the incorporation of
compression strength; construction waste (CW) for the stabilization of soil-cement mixtures. Ambiente
Construído 2020;28(4):261–80. https://doi.org/10.1590/s1678-
d) After carrying out the wetting/drying durability tests, there was a 86212020000400471).
potential growth ratio between mass loss (LM) and η/Civ0.35. The [16] Dupas J, Pecker A. Static and dynamic properties of sand cement. J. of the Geotech.
ratio obtained as a result of the performed tests presented excellent Engineering Division, ASCE 1979;105(3):419–36. https://doi.org/10.1061/
AJGEB6.0000778.
quality (R2 = 0.86) and reflects the idea that the material with low [17] Faro VP. Laterally loaded deep foundations associated with treated soils:
cementation and/or greater mixture porosity becomes more sus­ conception, load tests and design guidelines [In Portuguese.]. In: DSc. Thesis, Post
ceptible to easily wear, and loses mass compared to denser and Graduate Program in Civil Engineering, Universidade Federal do Rio Grande do
Sul; 2014.
cemented mixtures;
[18] Floss, M. F. (2012). Parameters for controlling the strength and stiffness of artificially
e) After carrying out the wetting/drying durability tests, an exponential cemented granular soils [In Portuguese]. D.Sc. Thesis, Graduate Program in Civil
decreasing relation between VR and η/Civ0.35 occurs. The relation­ Engineering, Universidade Federal do Rio Grande do Sul.
[19] Fonseca AV da, Cruz RC, Consoli NC. Strength properties of sandy soil-cement
ship obtained from the tests has an excellent R2 relation (R2 = 0.84)
admixtures. Geotechnical and Geological Engineering 2009;27:681–6. https://doi.
and reflects the idea that material with low cementation and/or org/10.1007/s10706-009-9267-y.
greater mixture porosity becomes more susceptible to volume [20] Guedes SB, Coutinho RQ, Fonseca AV. Details about unconfined compression tests
reduction compared to denser and more cemented mixtures; and diametral compression tests for tensile strength determination. Geotecnia
2016;136:127–42. https://doi.org/10.24849/j.geot.2016.136.08.
f) When applying the ANOVA method, all variables studied were [21] Herrier G, Chevalier C, Froumentin M, Cuisinier O, Bonelli S, Fry JJ. Lime treated
considered significant. The p-value was less than 5% in all sources of soil as an erosion-resistant material for hydraulic earthen structures. In: 6th
variations, which leads to the equal means null hypothesis rejection, International Conference on Scour and Erosion; 2012. p. 582–95.
[22] Ingles, O. G.; & Metcalf, J. B. (1973). Soil stabilization: Principles and practice.
being the C (%) parameter considered the main one due to its level of Butterworth-Heinemann Ltd, 384p.
significance in the study. [23] Jin, Q.; Li, B. (2019) Effects of lime treatment on the geotechnical properties of
dredged mud. Marine Georesources & Geotechnology, v. 37, p. 1083 - 1094.
https://doi.org/10.1080/1064119X.2018.1527421.
Declaration of Competing Interest [24] Lambe, T. W.; & Whitman, R. V. (1969). Soil mechanics. John Wiley and Sons: New
York, 553p.
The authors declare that they have no known competing financial [25] Liu MD, Indraratna B, Horpibulsuk S, Suebsuk J. Variations in strength of lime-
treated soft clays. Proc Institution of Civil Engineers - Ground Improvement 2012;
interests or personal relationships that could have appeared to influence
165(4):217–23.
the work reported in this paper. [26] Mitchell, J. K. (1981). Soil improvement: State-of-the-art report. Proc. 10th
International Conference on Soil Mechanics, Vol. 4, Stockholm, pp. 509-565.
[27] Nussbaum, P. J.; & Colley, B. E. (1971). Dam construction and facing with soil-
Acknowledgments
cement. Research and Development Laboratories of the Portland Cement
Association: Bulletin.
The authors wish to express their gratitude to CAPES, FAPERGS and [28] Prietto, P. D. (2004). Strength and dilatancy of cemented soils: A theoretical-
CNPq—Brazilian Council of Scientific and Technological Research experimental approach [In Portuguese.] D.Sc. Thesis, Post-Graduate Program in
Civil Engineering, Federal University of Rio Grande do Sul, Brazil.
(Project Produtividade em Pesquisa) for the financial support to the [29] Rattley MJ, Lehane BM, Consoli N, Richards DJ. Uplift of shallow foundations with
research group. cement-stabilised backfill. Proc Institution Civil Engineers - Ground Improvement
2008;161(2):103–10.

9
T.F. de Souza Junior et al. Transportation Geotechnics 34 (2022) 100751

[30] Rios S, Fonseca AV, Bangaru SS. Silty sand stabilized with different binders. residual soils, Behavioral Characteristics of Residual Soils, GSP 92, ASCE, Reston,
Procedia Eng Science Direct 2016;143:187–95. https://doi.org/10.1016/j. VA, pp. 12-25.
proeng.2016.06.024). [34] Sherwood, P. (1993). Soil stabilization with cement and lime. State of the art review.
[31] Rizzon, M. M. (2016). Effects of dynamic compaction in residual soil [In Portuguese.] Transport Research Laboratory.
M.Sc. dissertation, Post-Graduate Program in Civil Engineering, Federal University [35] Souza Junior, T. F.; Jesus, F. F. S.; Ruver, C. A.; Heineck, K. S.; & Luz, C. C. (2019).
of Rio Grande do Sul, Brazil. Aplicação do método de classificação MCT ao solo coesivo friccional do município
[32] Ruver, C. A. (2011). Pullout study of isolated foundations in soil treated with cement de Passo Fundo - RS [in Portuguese]. Seminário de Engenharia Geotécnica do Rio
[In Portuguese.] D.Sc. Thesis, Post-Graduate Program in Civil Engineering, Federal Grande do Sul,10., Santa Maria. Anais... Santa Maria: UFSM, 2019.
University of Rio Grande do Sul, Brazil. [36] Transportation Research Board (1987). Lime stabilization—Reactions, properties,
[33] Schneider, J. A.; Hoyos, L., Jr.; Mayne, P. W.; Macari, E. J.; & Rix, G. J. (1999). design, and construction. State of the Art Rep. No. 5, Transportation Research Board,
Field and laboratory measurements of dynamic shear modulus of Piedmont Washington, D.C.

10

You might also like