Geo-Frontiers 2011 © ASCE 2011 2444

A Case History of Expansive Clays in Southwest Puerto Rico

Yaurel Guadalupe-Torres1, MSCE, P.E. and Miguel Pando2, PhD, [Link], [Link].
1
PhD Student, University of Rhode Island; Former MS Student at the University of
Puerto Rico; PH (787) 467-3226; e-mail: yaurel_gt@[Link].
2
Associate Professor, Department of Civil and Environmental Engineering,
University of North Carolina, Charlotte, NC, 28233; PH (704) 687-6745; e-mail:
mpando@[Link].
Downloaded from [Link] by Univ of Okla Libraries on 01/23/15. Copyright ASCE. For personal use only; all rights reserved.

ABSTRACT

This investigation involves characterizing and evaluating the expansion

potential of several clayey soils from the area of Boquerón, Cabo Rojo, in southwest
Puerto Rico. The samples were collected from two test pits excavated as part of this
project and were also obtained from a geotechnical consultant that had performed a
study in the vicinity of the test pit locations. The experimental program involved in-
situ field tests and laboratory tests on selected representative samples. The laboratory
program involved geotechnical classification of the soils, gradation, Atterberg limits,
mineralogy characterization, and specialized tests for assessing swell potential of
clayey soils. The mineralogy characterization involved determination of specific
surface area, cation exchange capacity, and X-Ray diffraction. Results from these
tests revealed significant presence of montmorillonite. The mineralogy combined
with test results of swell pressure and free swell revealed that some of the soils tested
from the test site Cabo Rojo can experience free swell volume changes greater than
40%, and swell pressures greater than 32,000 psf. The results of this investigation
confirmed presence of highly expansive soils in this test area located in the southwest
region of Puerto Rico.

INTRODUCTION

In the United States, expansive soils have been reported as being the cause of
structural damages of up to $9 billion a year (e.g. Jones and Holtz, 1973, Jones and
Jones, 1987). Clayey soils expand when the moisture content increases and contract
when it decreases. The degree of volume change will depend greatly on the
mineralogy of the clay and is usually related to soil properties such as particle size,
surface area, structural charge, cation exchange capacity and plasticity. Expansive
soils typically involve clayey soils as they have the potential to significantly change
its volume with the gain or loss of moisture. The three most common groups of clay
minerals associated to expansive soils are kaolinite, micas, and smectite, being this
last one the most likely to present significant volume expansion potential.
In Puerto Rico (PR), the existing information evidence concerning damage
caused by expansive soils is scarce and often conflicting or even speculative. A
recent survey by Ochoa and Pando (2005) revealed a lack of consensus on this matter
within the PR geotechnical community. However, the municipality of Cabo Rojo was
identified as one of the regions where several geotechnical companies had
encountered expansive soils. This paper seeks to provide reliable information related

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 2445

to this type of soils in this region identified in the Ochoa and Pando (2005) survey as
potentially critical. It also involved the evaluation of expansion potential for
disturbed and undisturbed samples using different experimental methods.

STUDY AREA

General Description and Geology

The soil samples studied in this investigation were obtained from the area of
Downloaded from [Link] by Univ of Okla Libraries on 01/23/15. Copyright ASCE. For personal use only; all rights reserved.

Boquerón located in the municipality of Cabo Rojo, in SW PR (see Figure 1). The
study area was chosen near a residential project that was reported as possibly being
affected by the presence of expansive soils. The topography of the study area is
relatively gentle sloping ground with an elevation slightly higher in the northeastern
end of the project and descends towards the southwest. The elevation in the entire
study area ranges from approximately 5 to 20 meters above mean sea level.

1(a) N 1(b)

Fig. 1(b)
Main Island of Puerto Rico

(Adapted from: Departamento de Recursos Naturales y Ambientales, 2004) (Source: GoogleMaps 2008)

Figure 1. General location map of the study area

Figure 2 shows the relevant portion of the Puerto Real Geologic Quadrangle
prepared by Volckmann (1984) which describes the general surficial geology of the
area. This figure shows a box demarking the location of the study area. According to
this generalized geology map, the surficial geology of the study area falls within the
Guanajibo Formation (Tg) and Quartz Sand Deposits (Tqs). The Guanajibo
Formation (Tg) is comprised of loosely cemented, calcirudite and calcarenite, as well
as sands and gravels, with some silicified mudstone and layers of limy silt. The
quartz sand deposits (Tqs) consist of quartz sands with minor hematite or limonite
and clay (Volckmann, 1984). Based on this generalized geology map, the study area
would have little to no presence of clayey soils. However, the field investigation
described later in this paper, confirmed extensive presence of clayey soils within the
upper 30 feet of the study area.

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 2446
Downloaded from [Link] by Univ of Okla Libraries on 01/23/15. Copyright ASCE. For personal use only; all rights reserved.

Figure 2. Generalized geology map of the study area (adapted from Volckmann, 1984).

Site Characterization and Sampling

The study area was selected in the close vicinity of a borehole carried out as
part of a forensic geotechnical investigation for a housing development where cracks
and defects were reported and were presumably attributed to the presence of
expansive soils. Some of the undisturbed samples from this geotechnical
investigation at the housing development were made available for this study.
Additionally as part of this research project, two trenches were excavated near the
housing development presumably affected by the presence of expansive soils. These
two trenches were used by Guadalupe (2008) to obtain disturbed and undisturbed soil
samples and perform additional in-situ tests such as dynamic cone, pocket
penetrometer, and natural moisture content. The generalized geotechnical soil profile
of these two trenches consisted of a 2 ft thick upper layer of brown sand with silt and
low plasticity clay, followed by approximately 4 ft of white, high plasticity clay with
few fragments of oxidized red soil particles. The white clay layer was underlain by
about 2 ft of a yellowish orange, high plasticity clay. The last 4 ft of the trenches
consisted of yellow low plasticity silt.
Figure 3 shows a summary plot of the three in-situ tests performed in one of
the trenches. The dynamic cone penetration and pocket penetrometer results show
very similar trend as both indicate higher resistance in the upper 4 to 5 feet, and a
decrease in resistance below a depth of about 5 feet. The penetration resistance
increased again at a depth of about 10 ft.

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 2447

Dynamic Cone Pocket Penetrometer Natural Moisture

2
Blows per m*10-1 Strength, ton/ft Content, %
0 4 8 12 16 4.0 4.2 4.4 4.6 4.8 5.0 0 10 20 30 40 50
0

3
Downloaded from [Link] by Univ of Okla Libraries on 01/23/15. Copyright ASCE. For personal use only; all rights reserved.

5
Depth, ft

10

11

12

Figure 3. Summary of in-situ tests performed at one test trench

Table 1 shows a summary of the samples selected for a detailed study. Most
of the samples showed high percentages of clay and high Atterberg limits. A total of
seven samples were chosen for a detailed analysis based on their high plasticity. The
five samples labeled A-# were obtained from the private geotechnical study, and the
two samples labeled B-# were selected from the two exploratory trenches carried out
by Guadalupe (2008). The seven samples selected all have over 75% of fine particles
(fraction < #200), of which six have more than 50% clay fraction (< 2µm). Based on
Atterberg limit values, the samples likely to exhibit the highest expansive potential,
are A-3 (LL = 138 % and PI = 81 %), B-2 (LL = 155 % and PI = 112 %), and B-1
(LL = 169 % and PI = 140 %).

Table 1. Summary of soil characterization for selected samples

Atterberg Grain Size Composition Specific
Depth Wn Activity
Sample (%) (%) Gravity USCS
(%) Ac
(ft) LL PI Sand Silt Clay (Gs)
A-1 6–8 21.4 101 69 10.34 23.23 66.43 1.04 2.74 CH
A-2 10 – 12 26.0 92 67 20.51 28.17 51.32 1.31 2.88 CH
A-3 16 – 18 47.9 138 81 2.42 16.93 80.65 1.00 2.77 MH
A-4 18 – 20 46.5 86 52 2.09 31.98 65.93 0.79 2.83 CH
A-5 27 – 29 12.7 82 58 3.27 50.52 46.21 1.26 2.78 CH
B-1 2–6 27.0 169 140 13.1 15.48 71.42 1.96 2.65 CH
B-2 6–8 40.4 155 112 1.61 12.30 86.09 1.30 2.73 CH

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 2448

Figure 4 shows the plasticity chart for the seven selected samples. All
samples are in the high plasticity region of this chart. Sample A-3 falls in the
plasticity chart in the region corresponding to high plasticity silt, however the grain
size analysis of this sample indicate this sample has 81% clays and 17% silts.

150
A-1
A-2
Downloaded from [Link] by Univ of Okla Libraries on 01/23/15. Copyright ASCE. For personal use only; all rights reserved.

A-3
125
A-4
A-5
B-1 High Plasticity
100 B-2
Plasticity Index, %

n e
Li
U
75
e
Lin
A
50

25

0
0 25 50 75 100 125 150 175

Liquid Limit, %
Figure 4. Plasticity of selected samples

CLAY MINERALOGY OF SELECTED SAMPLES

The clay mineralogy of the selected samples was evaluated by means of

specific surface area, cation exchange capacity, and X-Ray diffraction.

Specific Surface Area

Specific surface area (SSA) was measured using the Ethylene Glycol
Monoethyl Ether (EGME) method (Carter et al., 1965). The SSA results are
summarized in Figure 5. The three samples with highest surface area were A-3, B-1,
and B-2 with SSA values of 446.4, 456.2 and 464.8 m2/g, respectively. These three
samples were also the samples with the highest values of liquid limits and plasticity
index. For comparison purposes, typical ranges of specific surface area values of
highly expansive pure clay minerals vermiculite and smectite are 300 to 500 m2/g,
and 600 to 800 m2/g, respectively. The highest specific surface area values obtained
from the selected samples fall in the range of pure clay mineral vermiculite.
However, as discussed later, the selected samples were not pure, but rather
heterogeneous and therefore are composed of different minerals. The reported values

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 2449

of specific surface area represent an average surface area of the different minerals
present in the test sample. For this reason, the presence of clay mineral smectite is
possible to some degree.

800
Range for
Pure
700 Smectites
Downloaded from [Link] by Univ of Okla Libraries on 01/23/15. Copyright ASCE. For personal use only; all rights reserved.

Specific Surface Area, m /g

600
2

500
Range for
Pure
400 Vermiculites

300

200

100

0
A-1 A-2 A-3 A-4 A-5 B-1 B-2

Sample ID
Figure 5. Values of specific surface area of selected samples

Cation Exchange Capacity (CEC)

The cation exchange capacity (CEC) is defined as the ability of a mineral to

exchange one cation by another and it is measured in cmolc/kg. The exchangeable
cations of the samples were extracted using ammonium chloride (NH4Cl) and then the
CEC from the extraction was measure by atomic absorption. Previous research by
Holtz (1969), McKeen and Hamberg (1981) and Hamberg (1985) have proposed
correlations between the parameter CEAc (Cation Exchange Activity) and the activity
of the soil. The CEAc parameter is defined as the total clay CEC normalized by the
percentage of clay in the soil. The activity of the clay (Ac) is defined as the ratio of
the plasticity index and percentage of clay fraction in the soil. The values of CEAc
can be used to identify the predominant clay mineralogy of the soil using a chart like
the one proposed by Holtz (1969). Figure 6 shows such a chart, and for this study the
CEAc and corresponding activity (Ac) values for the seven samples selected for
detailed study. This figure shows that six of the samples fall in the region identified
by these references as corresponding to montmorillonite which belongs to the
smectite group.

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 2450
Downloaded from [Link] by Univ of Okla Libraries on 01/23/15. Copyright ASCE. For personal use only; all rights reserved.

Figure 6. Cation exchange activity versus activity plot (adapted from Holtz, 1969)

X-Ray Diffraction

Four samples were selected for X-Ray diffraction (XRD). X-Ray diffraction
characterization was performed by Evans Analytical Group Laboratory in Texas,
USA. Table 2 summarizes the results of quantitative analysis of the XRD.

Table 2. X-Ray diffraction test results

% Total
% Approx. Weight of
Sample % Crystallinity Crystalline Phase Approx.
Crystalline Fraction
Weight
A-2 53 Montmorillonite 29.3 15.53
Kaolinite 27.9 14.8
Quartz 41.5 22.0
Goethite 1.2 0.64
A-3 30 Montmorillonite 67.8 20.34
Kaolinite 8.7 2.61
Quartz 20.0 6.00
Goethite 3.5 1.05
B-1 40 Montmorillonite 63.6 25.44
Kaolinite 16.9 6.76
Quartz 19.0 7.6
Goethite 0.4 0.16
B-2 25 Montmorillonite 62.2 15.55
Kaolinite 10.0 2.5
Quartz 21.5 5.38
Goethite 6.3 1.57

The crystalline fraction of samples A-3, B-1 and, B-2 are mainly
montmorillonite with percentages by weight of 67.8, 63.6, 62.2%, respectively.
These three samples also have some presence of kaolinite ranging between 8.7 and
16.9%. Sample A-2, on the other hand, has similar percentages of montmorillonite

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 2451

(29.3%) and kaolinite (27.9%). From the quantitative XRD results it is concluded
that the main mineral present in the crystalline phase of three of the four samples
selected for XRD testing was montmorillonite.

MEASUREMENTS OF EXPANSIVE POTENTIAL

The expansive potential of selected samples was also evaluated using the
modified free swell index (MFSI) and the ASTM test of swell potential. These tests
and their results are discussed below.
Downloaded from [Link] by Univ of Okla Libraries on 01/23/15. Copyright ASCE. For personal use only; all rights reserved.

Modified Free Swell Index (MFSI)

This method consists in pouring 10 grams of oven-dry sample into a calibrated

100-mL test tube. Then the test tube is filled with water. Then the sample and water
mixture is allowed to swell free of external loading or stresses. Once the swell is
fully developed, measurement of the final volume is taken. The initial volume of the
soil sample is obtained using the specific gravity. The change in volume divided by
the initial value is the Modified Free Swell Index. The MFSI results are summarized
in Table 3. The samples with the highest values of MFSI are A-3, B-1, and B-2.
These samples all showed values of MFSI above 7. Table 3 also shows swell
potential adjectives as proposed by Sivapullaiah et al. (1987). These authors
correlated the values of MFSI with expansive potential of soil. According to this
correlation all samples classified as having a moderate expansive potential. These
authors also proposed correlating the liquid limit by volume (gravimetric Liquid
Limit times Gs) with the potential volume of soil. According to this second proposed
correlation, samples B-1 and B-2 would have a high expansive potential.

Table 3. Modified Free Swell Index and Liquid Limit by volume

Initial Final Liquid
Swell Potential1
Sample Volume Volume MFSI Limitvolume
cm3 cm3 (%) MFSI LLvolume
A-1 4.00 21 4.25 276 Moderate Moderate
A-2 3.50 26.25 6.50 263 Moderate Moderate
A-3 4.00 32 7.00 382 Moderate Moderate
A-4 4.00 28 6.00 243 Moderate Moderate
B-1 4.00 40 9.00 447 Moderate High
B-2 4.00 33 7.25 426 Moderate High
1
Swell potential adjectives correlated to MFSI and LLvolume as proposed by Sivapullaiah et al. (1987).

Swell Potential Tests

This type of test is performed using the oedometer equipment with

undisturbed shelby tube samples. According to ASTM Standard D-4546 (ASTM,
2003) this test can be done following three possible methods. Method A consists of
saturating the sample and taking deformation reading versus time until the full 1-D
swelling is achieved (usually between 3 to 16 days). Once the 1-D swell is fully
developed, increasing load is then applied gradually until the sample returns to its

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 2452

original height. An initial surcharge of approximately 32 psf was applied for this
method. This test allows us to determine the percentage of lift or heave percent, free
swell, and swell pressure which is the pressure required to return the sample to its
initial height. Table 4 summarizes the results of the Swell Potential test by Method A
(ASTM D-4546). It can be observed that sample B-1 resulted with the highest swell
pressure (32,483 lb/ft2), followed by samples B-2 and A-2, with swell pressure values
close to 13,000 lb/ft2.
Downloaded from [Link] by Univ of Okla Libraries on 01/23/15. Copyright ASCE. For personal use only; all rights reserved.

Table 4. Swell Potential Results using ASTM Standard D-4546 (Method A)

Initial Initial
Depth Swell Pressure Free Swell Heave
Sample Unit Weight Moisture
ft psf Method A Method A
pcf %
A-1 126.7 23.23 7 9,462 5.03 4.38
A-2 122.3 30.20 11 12,491 14.25 11.06
A-3 103.3 43.99 17 4,253 8.17 3.43
A-4 103.2 51.33 19 4,612 7.74 3.40
B-1 116.0 34.58 4 32,483 42.16 39.43
B-2 109.8 41.45 7 13,098 18.23 14.26

These three samples also showed the highest values of free swell and heave
percent. Free swell values of 42.16%, 18.23%, and 14.25%, were measured for
samples B-1, B-2, and A-2, respectively. The results of heave percentage are similar
in magnitude to these from free swell. Figure 6 shows the recorded expansion
behavior for the different samples using test Method A.

3200
A-1
2800 A-2
A-3
A-4
2400 B-1
Vertical Deformation, *10 in

B-2
-4

2000

1600

1200

800

400

0
0.1 1 10 100 1000 10000 100000

Time, min
Figure 7. 1-D swell behavior with time of selected samples under 32 psf

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 2453

SUMMARY DISCUSSION AND CONCLUSIONS

The detailed field and lab investigation carried out on clay soil samples from a
site in southwest Puerto Rico revealed that the studied soils have high values of
plasticity, clay content, and clay activity. Additionally, some of the tested samples
showed moderate to high values of specific surface area and their CEAc suggests high
expansive behavior. XRD tests on selected samples revealed high content of
montmorillonite in some of these samples. Swell potential tests, following ASTM D-
4546, yielded swell pressure values up to 32,483 psf, and 1-D free swell values up to
Downloaded from [Link] by Univ of Okla Libraries on 01/23/15. Copyright ASCE. For personal use only; all rights reserved.

42%. All these results suggest these soils have a high expansive potential. However,
correlations of swell potential with the MFSI index suggest these soils would only
have a moderate expansive potential. The correlation by Sivapullaiah et al. (1987),
based on using the volumetric liquid limit (LLvolume), suggests the expansive potential
of these soils would be between moderate to high. However, MFSI has the drawback
that it is based on testing oven dried soil sample that have lost their structure. If
possible, the assessment of swell potential of clay soils should include tests on
undisturbed samples. Finally, the results from this study confirm results in the survey
by Ochoa and Pando (2005) which indicated some participants reported presence of
expansive soils in regions of SW Puerto Rico.

REFERENCES

Carter, D. L., Heilman, M.D. and González, C.L. (1965). Ethylene glycol
monoethylether for determining surface area of silicate minerals. Soil Sci. 100 (5),
pp. 356-360.
Departamento de Recursos Naturales y Ambientales de Puerto Rico. (2004).
“Inventario de Recursos de agua de Puerto Rico.” Draft Document.
Guadalupe, Y. (2008). “Caracterización y evaluación del potencial expansivo de
suelos arcillosos en Cabo Rojo.” In Spanish, MS Thesis, UPRM, 190 p.
Hamberg, D. J. (1985). “A simplified method for predicting heave in expansive
soils.” MS Thesis, Colorado State University, Fort Collins, CO.
Holtz, W.G. (1969). “Volume change in expansive soils and control by lime
treatment”, 2nd Intl. Res. & Eng Conf. on Expansive Clay Soils, Texas A&M, TX.
Jones, D. E., and Holtz, W. G. (1973). “Expansive Soils – The Hidden Disaster,”
Civil Engineering – ASCE, 43 (8), New York, NY, August, pp 49-51.
Jones, D. E. and Jones, K. A. (1987). “Treating Expansive Soils”, Civil Engineering,
ASCE, 57 (8), August 1987, pp. 62-65
Lambe, William T. (1960). The Character and Identification of Expansive Soils.
Technical Studies Program, Federal Housing Administration, 116 p.
McKeen, R. G. and Hamberg, D. J. (1981). “Characterization of expansive soils.”
TRB, Trans. Res. Rec., 790, pp. 73-78.
Ochoa, M. and Pando, M. A. (2005). “Survey to Geotechnical Consultants and
Experts on Expansive Soils in Puerto Rico.” REU Report, UPRM, 45 p.
Sivapullaiah, P. V., Sitharam, G. T. and Subba Raol, K. S. (1987). “Modified Free
Swell Index for Clays.” Geotech. Testing J., ASTM, 10 (2), pp 80-85.
Volckmann, R. P. (1984). “Geologic Map of the Puerto Real Quadrangle, Southwest
Puerto Rico.” USGS Miscellaneous Investigations Series Map I-1559.

Geo-Frontiers 2011