Construction of pavements on expansive soil

Material Information

Construction of pavements on expansive soil
Huzjak, Robert John
Publication Date:
Physical Description:
122, [12] leaves : illustrations, charts, maps ; 28 cm


Subjects / Keywords:
Swelling soils -- United States ( lcsh )
Roads -- Design and construction -- United States ( lcsh )
Roads -- Maintenance and repair -- United States ( lcsh )
Roads -- Design and construction ( fast )
Roads -- Maintenance and repair ( fast )
Swelling soils ( fast )
United States ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 124-134).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Science, Department of Civil Engineering.
Statement of Responsibility:
by Robert John Huzjak.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
19783265 ( OCLC )
LD1190.E53 1988m .H88 ( lcc )

Full Text
A report submitted to the
Faculty of the Graduate School of the
University of Colorado at Denver in partial fulfillment
of the requirements for the degree of
Masters of Science
Department of Civil Engineering

This report for the Masters of Science degree by
Robert John Huzjak
has been approved for the
Department of
Civil Engineering

Appreciation and thanks are extended to Professor
Nien-Yin Chang, my major advisor, for his advice and
encouragement on my research, and Associate Professor
Tzong H. Wu and Professor John R. Mays for serving on the
committee, and to the Federal Highway Administration,
Region V, for their support in obtaining numerous publi-
cations. Special appreciation is extended to Laura
Bennett for typing the manuscript; and to my family for
their support and encouragement.

Huzjak, Robert John (M. S., Civil Engineering)
Construction of Pavements on Expansive Soil
Report directed by Professor Nien Y. Chang
Expansive soils are prevalent throughout a large
portion of the United States. Because of the great
extent of expansive soils in the United States, it is
impossible to alter all highway alignments to avoid
placing pavements on these materials.
An extensive literature search was performed to
determine the current design practices utilized for pave-
ments on expansive soils. This paper presents an over-
view of the expansive soil problems, describes typical
testing methods utilized for the design of pavements, and
provides a detailed description of the various construc-
tion methods and case histories which describe the suc-
cess of various treatment alternatives. This information
has been correlated to develop a recommended analysis.
Areas where additional research are needed are also
The form and content of this
recommend its publication.

I. INTRODUCTION.................................... 1
1.0 Effects of Expansive Soils on Pavement
Performance............................ 3
2.1 Physical Properties Affecting Swelling 5
2.2 Environmental Factors Affecting
Swelling............................... 12
2.3 Geology................................ . 15
2.4 Distribution of Expansive Soils.......... 19
III. SAMPLE IDENTIFICATION........................ 26
3.0 Sampling Techniques...................... 26
3.1 Laboratory Testing....................... 27
3.2 Indirect Techniques.................... 28
3.3 Combination Techniques................... 32
EFFECTS OF EXPANSIVE SOILS................... 38
4.0 Avoid the Expansive Material............. 39
4.1 Removal of Expansive Soil and
Replacement with Non^Expansive Soils 41
4.2 Applying a Surcharge Pressure............ 44
4.3 Pre-Wetting the Soil..................... 45
4.4 Chemical Stabilization................... 49
4.5 Lime Slurry Pressure Injection.......... 63
4.6 Compaction Control....................... 67

4.7 Membrane Control........................ 79
4.8 Explosive Treatment.................... 92
4.9 Post-Construction Techniques............ 94
4.10 State Highway Agent Practices.......... 94
4.10.1 Colorado........................... 95
4.10.2 Wyoming............................ 96
4.10.3 Montana........................... 97
4.10.4 Arizona............................ 98
4.10.5 Kansas............................. 98
4.10.6 North Dakota...................... 99
4.10.7 Texas........................... 100
4.10.8 Utah.............................. 100
4.10.9 Summary........................... 101
SOILS...................................... 103
5.1 Analysis Method..................... 103
5.2 Design and Construction
Recommendations.................... 112
5.2.1 Surface Drainage................. 112
5.2.2 Subsurface Drainage............. 113
5.2.3 Pavement Cross-Sections.......... 115
5.2.4 Pavement Types................... 117
VI. SUMMARY............................ 119
BIBLIOGRAPHY.......................................... 123

2.1 Molding water content versus dry density and
particle and particle orientation............... 7
2.2A Relationship for clay, silt, and sand.......... 11
2.2B Wetting and drying.............................. 11
2.3 Distribution of potentially expansive
materials in the United States: FHWA Regions
1, 3 & 5....................................... 20
2.4 Distribution of potentially expansive
materials in the United States: FHWA Region 4 21
2.5 Distribution of potentially expansive
materials in the United States: FHWA Region 6 22
2.6 Distribution of potentially expansive
materials in the United States: FHWA Regions
7 & 8......................................... 23
2.7 Distribution of potentially expansive
materials in the United States: FHWA Regions
9 & 10.......................................... 24
3.2 Correlation of percent swell, liquid limit,
and dry unit weight............................. 36
4.1 pH and water content versus percent lime for
samples from sampling site 12, Hayes, Kansas. 56
4.2 Percentage of expansion for various placement
conditions under a 1 psi surcharge............ 69
4.3 Total uplift pressure caused by wetting for
various placement conditions.................... 69
4.4 Pittsburg sandy clay, Effects of method of
compaction on swell pressure saturation........ 71
4.5 Vicksburg silty clay, Effects of method of
compaction on swell pressure saturation........ 71

4.6 Typical grading section......................... 73
4.7 Typical sprayed asphalt membrane applications
to minimize subgrade moisture variations from
surface infiltration............................ 81
4.8 Example of vertical membrane cutoff
construction.................................... 82
4.9 1-40 typical sections........................... 89
5.1 Chart for estimating the approximate
. potential vertical rise of natural soils..... 107
5.2 Four performance curves illustrating the
effect of foundation movements in the absence
of traffic..................................... 110
5.3 Diagram of surface drainage design and
construction recommendations for (a) trans-
verse and (b) and (c) longitudinal highway
sections....................................... 114

2.1 Typical values of free swell for common clay
minerals........................................ 18
3.1 Indirect techniques for identification/
classification of expansive soils............30-31
4.1 Methods for volume change control using
Summary of lime stabilization test procedures

Distortion, deterioration and cracking of highway
pavements caused by swelling or shrinking of expansive
clay soils are estimated in 1973 to cause damage in
excess of $1.1 billion annually in the United States
(Ref. 1). Expansive soils create problems throughout
many areas of the world and, in particular, in the
Western and Central United States. The State of Texas
estimates that the total cost of pavement maintenance due
to expansive clay soil varies from 6 to 10 million
dollars a year (Ref. 2). A 1972 survey of highway
departments in the 50 United States, District of Columbia
and Puerto Rico, indicated that 36 states have expansive
soils within their geographical area (Ref. 3). Because
of the great extent of expansive soils in the United
States, it is impossible to alter all highway routes to
avoid placing pavements on expansive materials.
Although the presence of expansive soils is vast
within the United States, there are no defined policies
regarding the design of pavements on expansive soils.
The behavior of expansive clayey soils under variable

conditions is not precisely known and, therefore, there
are numerous ways in which to mitigate the effects of the
expansive materials.
The major objectives of this study are to:
1. Identify the effect of expansive soils on
pavements, their primary physiographic areas within the
United States, and their properties.
2. Discuss in general the sampling and labora-
tory testing techniques used to identify expansive soils.
3. Describe in detail various successful methods
used to minimize the effects of expansive soils, and to
supplement the discussion with the use of case
4. Describe the current design practices used by
various state highway officials. Practices will be
discussed for new pavements, and for maintenance or
reconstruction of existing pavements on expansive clay.
5. Recommend areas which require future study.
The above topics have been studied by various
universities, independent researchers and highway agen-
cies. This paper will not address new methods to use for
expansive soils; however, will consist of a technical
literature review to determine what design and construc-
tion procedures are currently being used and what has
worked in the past. This report represents the results

of a review of available literature, combined with the
experience of various state highway agencies in dealing
with expansive soils.
1.0 Effect of Expansive Soils on Pavement Performance
As stated earlier, expansive soils have a signi-
ficant influence on the performance of pavements. The
primary effects of expansive clay are loss of riding com-
fort and road-holding ability, a reduction in pavement
service life, and the expenses of maintenance and rehabi-
litation. The principle forms of swell damage to a high-
way are:
1. A series of waves or unevenness which deve-
lops along a stretch of pavement. This usually occurs
without visible surface cracking or a reduction in
subgrade strength. These waves typically continue to
develop until the moisture equilibrium in the soil is
reached, and then they remain unchanged. This, results in
a severe loss of riding comfort.
2. Longitudinal cracks develop which run
parallel to the center line of the pavement. These
cracks develop because the pavement has effectively
sealed the soil, and the subgrade near the center line
becomes increasingly wet from moisture accumulation while
the soil at the shoulders fluctuates seasonally.

3. Transverse cracking occurs due to discon-
tinuities and irregularities in the subgrade soils.
These could occur due to changes in geologic structure,
vegetation, existing trees, or cut and fill transitions.
4. Localized failure of the pavements due to
decrease in strength and bearing capacity failure. This
typically occurs when a previously dry soil becomes near
saturated and loses the majority of its strength.
Pavement damage from the first and third forms of
failure identified is, by far, the most common (Ref. 2).
Bumps in the pavement can occur in groups, in cycles, or
at random. The bumps can vary in height to approximately
12 inches, which can seriously affect driving safety. If
the subgrade heaves uniformly, there would not be a
decrease in the quality of ride, nor would there be
structural deficiencies in the pavement. However, due to
the properties of soils and the climatic conditions pre-
viously discussed, differential movements are by far the
most common.

The potential of a soil to swell is influenced by
many factors. However, the effect of swell is primarily
controlled by moisture movement within the soils. The
soil properties such as permeability, plasticity,
moisture content, suction and density also have a great
influence. The geological aspects such as minerology,
origin, thickness of deposit, and geologic structure,
also influence the behavior significantly. In addition,
a similar clay with similar properties and origin placed
under different environmental condition will behave dif-
ferently. Environmental conditions such as climate,
humidity/ evaporation rate, rainfall, temperature, sur-
face vegetation and sources of moisture also have an
effect on the behavior of clays.
2.1 Physical Properties Affecting Swelling
Density: Research studies (Ref. 4, 5, 6, 7, 8 and 9)
have shown that the same clay compacted to a dense state
will, under similar conditions, heave more than soils
compacted to a looser state. This occurs because, when

the moisture content of the soil is increased, the indi-
vidual soil particles want to expand. If there is suf-
ficient void space, the soil particles can expand into
the void space and, therefore, will not cause an increase
in volume of the total soil mass. Generally, the lower
the density the lower the expansion potential. However,
very low density clays have low strength and may con-
solidate .
The density also affects the interparticle
arrangement. Details of the influence have been
described by Pacy C. and Chang (Ref. 9), and are shown in
Figure 2.1. In general, at a given compactive effort
and a low moisture content, a less orientated fabric is
obtained. As the moisture content and density increase,
the soil fabric is more oriented. This tends to create a
particle arrangement which concentrates the heave in a
single direction. (See Figure 2.1)

Figure 2.1 Molding water content versus dry density and particle and particle orientation
Ref. 9: "A Review of Engineering Experiences with Expansive Soils in Highway Subgrades"

Permeability (Ref. 9 and 10): Permeability is one of the
greatest factors affecting the rate of moisture movement
within a soil. The permeability, as identified in the
laboratory on either a natural or a remolded specimen, is
not applicable directly to field analysis. A typical
expansive clay is dessicated and highly fractured, and it
has been shown that the primary movement of moisture is
along the fissures and cracks. As can be expected, the
greater the permeability the greater the rate of water
movement and, therefore, the greater the rate of heave.
Typically, surface moisture will penetrate a more
permeable soil to a greater depth and, therefore, more
rapidly cause swelling to a greater depth. In a less
permeable soil, the moisture will migrate much slower
and, depending upon the evaporation rate and other fac-
tors, may never affect the soil to a great depth. The
permeability and the amount of fissuring and cracking
influence both the rate and depth of heave and, sub-,
sequently, the extent of vertical movement.
The depth of dessication and fracturing is the
depth to which evaporation influences are reflected in
the soil moisture content profile. Generally, the hotter
and drier the climate, the greater the depth of dessica-
tion .

cracks and fissures and the flow enters into the soil
mass, the permeability is a function of the initial
moisture content, dry density and soil fabric. "For com-
pacted soils, the permeability is greater at the lower
moisture contents and dry densities, and decreases to
some relatively constant value at about optimum moisture
content. Above optimum, the permeability is essentially
constant." {Ref. 9) The reason for this decrease near
optimum is that the void space is a minimum because of
close particle spacing.
Atterberg Limits (Ref. 11, 12, 13): Experience has shown
that volume change behavior correlates reasonably well
with liquid limit, plasticity index and shrinkage limit.
These indices tend to indicate the amount of clay par-
ticles, which then has an affect on the amount of swell.
The liquid limit and plastic index correlate reasonably
well with swell potential, primarily because of the
correlation between those indices and the type and amount
of clay minerals present. In general, the higher the
liquid limit, plasticity index and shrinkage limits, the
greater the potential volume change.
Moisture Content (Ref. 14): The initial moisture content
of a native deposit or a compacted fill will have an
effect on the swell potential. In general, the drier the

initial moisture content the greater the degree of
Soil Suction (Ref. 15, 16, 17, 18, 19): Soil suction can
be simplistically described as a soil property which
indicates the tendency of a soil to attract water. The
concept of soil suction is very complex, and only a very
brief discussion is presented here. More complete
descriptions of soil suction can be found in the referen-
ces .
Moisture migration will occur in soils due to
variations in soil suction. Moisture typically migrates
from areas of low suction to areas of high suction.
Moisture is redistributed until a new state of suction
equilibrium is obtained. This rate may be very fast or
very slow, depending upon the individual soil
It is possible that two different soils in con-
tact may be in suction equilibrium even though their
moisture contents may differ (see Figure 2.2A). It is
also possible for two portions of the same soil to be in
suction equilibrium but have different moisture contents.
This occurs when one soil portion is becoming wetter
while an adjacent portion is drying (see Figure 2.2B).

lems in
2.2A Relationship for clay, silt, and sand.
Figure 2.2B Wetting and drying
2.2A, 2.2B Soil suction versus moisture content
Ref. 2: "An Examination of Expansive Clay Prob-
Texas", 1971.

2.2 Environmental Factors Affecting Swelling
After the subgrade has been prepared and the
pavement structure placed, the primary influence on the
effect of heave is the environmental factors. The effect
of various environmental factors will be discussed.
Climate (Ref. 20, 21, 22, 23): Expansion of soil is
directly related to an increase in moisture content.
Climatic conditions, primarily rainfall and evaporation,
have a significant influence on the magnitude and time
rate of heaving.
Expansive soils are most prevalent in semi-arid
regions where evaporation exceeds rainfall. In these
climates, the soil is generally dry and increases in
moisture content are accompanied by an increase in
volume. The amount of moisture variation within the soil
is affected by the length of the evaporation and the
maximum rainfall duration. Where there is a definite wet
and hot dry season, there is a large seasonal variation
in the moisture content. In this type of climate, there
are definite seasonal variations in moisture content of
the soil. These are most pronounced near the shoulders
and outer edges of highway pavements.

Seasonal moisture variations have been reported
to depths of 10 to 12 feet; however, in semi-arid cli-
matic conditions, the depth is normally between 5 to 7
feet. Seasonal moisture variations are relatively
constant for given climatic conditions; however, there is
a general trend towards an accumulation of total moisture
content. This is particularly the case where the
majority of the rainfall occurs in the summertime when
there is also a high evaporation rate. This decreases
the amount of seasonal variations, but tends to be cumu-
lative, which generally tends to increase the expansion
with each season until an equilibrium condition
Vegetation (Ref. 24, 25, 26): The effect of vegetation,
both prior to and after construction, is a very impor-
tant factor. Vegetation such as trees, shrubs, and
grasses affect the pre-construction moisture content of
the soils. Tree roots typically tend to dry out the sub-
surface soils to depths of 6 to 14 feet because of the
moisture which they have pulled out, of the soil. When
the vegetation is removed and pavements are placed, the
moisture that was being used by the vegetation will tend
to accumulate beneath the pavement structure. This accu-
mulation of moisture tends to enhance volume change and,

in particularly differential heave due to the equaliza-
tion of moisture content based upon the density of the
root structures in various locations.
Burrange (Ref. 24) reported that the moisture
content of a forrested, silty clay soil at a depth of 18
inches was reduced from 20 percent in June to 10 percent
in September, even though 4 to 5 inches of rain fell over
that period. In contrast, the moisture content of an
open soil area nearby fluctuated very little, remaining
approximately at 20 percent at all times.
Trees placed near pavements also can cause loca-
lized settlement, as their root system develops and drys
out the soil. This is most predominant in soils where
the initial moisture content is moderate to high.
Effect of Pavements (Ref. 27, 28): When a surface is
sealed with a pavement or a building slab, the
equilibrium conditions of the subgrade soils' moisture
content are affected. According to DeBruijn (Ref. 28),
the subgrade beneath a sealed road surface will generally
have a higher moisture content and a lower suction poten-
tial than an area nearby. This is affected by the type
of cover on the shoulder areas and the extent of pavement
Poor surface drainage or cracking of a pavement
surface will allow easy access for rainfall to infiltrate

into the pavements. In addition, the construction of
ditches which are too close to the pavement or are not of
a sufficient depth or gradient, can also increase the
potential for moisture infiltration.
Other sources of moisture infiltration could be
from broken water lines, drains, culverts, or the ponding
of water along the roadway. If these conditions occur
during a seasonally dry period of the year, the effect
will be much greater than during a wet season. Poor
drainage or the infiltration of moisture through cracks
in the pavement can be very serious, and upset the
balance of an equilibrium soil mass.
2.3 Geology
Knowledge of the geologic structure and history
of the soil are necessary to provide an understanding of
the problems associated with expansive clay soils. Local
geology and local history of the geologic formation,
layering, and make-up have a great influence on the beha-
vior of that particular clay soil. If highway pavements
are constructed such that they cross a series of folds,
arches, or highly variable subsurface conditions, the
variation in geologic origin and history may cause dif-
ferential movement of the subgrade. The minerology of

the clay particles also has an influence on its expansive
behavior. In addition, whether the clay is of alluvial,
colluvial or residually weathered material influences its
behavior. In addition, if the expansive material is a
bedrock unit, its behavior will also be governed to some
degree by the geologic properties of the formation.
Expansive materials can generally be subdivided
into three categories on the basis of their physical
characteristics. These are expansive rocks, sediments
and residual soils. These divisions are primarily geolo-
gic in origin, and relate more to the geologic history
than to the specific minerology.
The term expansive rocks refers to relatively
hard clay-shales or claystones which have been buried,
consolidated and partially cemented. These rocks can
have different degrees of weathering; however, they are
classified as a rock and have a rock structure.
The sediments have generally been transported by
slow moving water, or as a result of gravity down a
hillside. The parent rock from which these materials
developed could be a great distance away, and their pro-
perties may not be reflected in the local rocks in the
immediate vicinity. The clays can vary from low to
highly plastic, and could be found in thin to thick depo-
sits. In addition, they typically have layering, and

have a variation in particle size depending upon the
velocity at which they were placed during that particular
time span.
Residual soils are materials which have been
weathered in place from the local bedrock of the site.
Their properties and texture can be predicted from the
parent rock; however, due to the change in structure from
the parent rock, their expansive properties could vary
For engineering purposes, clay generally refers
to inorganic, plastic material which is largely made up
of particles less than 0.005 mm in diameter. Most clay
contains two or more of the clay minerals, which are
illite, montmorillonite and kaolinite.
Most active clay minerals include montmorillonite
combined with some other type of clay mineral. Kaolinite
and illites are usually less active; however, they may
contribute to expansive properties if sufficient amounts
are present.
Minerology (Ref. 29, 30): Expansive rocks, sediments and
residual soils derive their expansive character from
their constituent clay mineral, past and present loading
history, and their natural and imposed environments. A
brief summary of the three basic clay minerology groups
are provided below. Other types of clay minerals exist;

however, they are not common and generally do not
substantially influence the swelling and will not be
discussed here.
Table 2.1 indicates the potential free swell of the
various clay materials.
Typical Values of Free Swell for Common Clay Minerals
Clay Mineral Free Swell %
Sodium Montmorillonite 1400-2000
Calcium Montmorillonite 45-145
Illite 60-120
Kaolinite 5-60
Table 2.1 Ref. 29: Clay Mineralogy, 1975
Montmorillonite is the most common of all clay
minerals, especially clays derived from weathering of
volcanic ash. Montmorillonite clay has a high potential
for expansion. Ordinarily, montmorillonite exists as
extremely small particles with dimensions on the order of
a few tens of angstrom units.
Illite is one of the more abundant clay minerals
occurring in modern marine clays, and in shales. The
clay layer is tightly bonded and, therefore, does not
readily admit sufficient amounts of water between the

clay layers. Therefore, illite is much more stable than
montmorillonite, and typically has low expansive proper-
Kaolinite is most common in residual clay soils.
Its bonding is fairly tight, and it is difficult to
separate into layers. As a result, the volume change
exhibited by this mineral is mainly due to water absorbed
on the periphery of the individual grains and, therefore,
shows very little swell.
2.4 Distribution of Expansive Soils
Soils of varying degrees of expansiveness are
present throughout the United States. The areal distri-
bution and degree of expansiveness of the materials
within the United States are shown on Figures 2.3 through
2.7 (Ref. 31).
The distribution of the expansive material shown
in Figures 2.3 through 2.7 was categorized on two bases.
(1) degree of expansiveness and (2) expected frequency of
occurrence of expansive materials.. The information
reviewed was correlated to provide four mapping cate-
gories that reflect the degree of expansiveness, and
expected frequency. The four categories are:

Figure 2.3 Distribution of potentially expansive mater-
ials in the United States: FHWA Regions 1, 3 & 5. Ref. 31
"A Review of Engineering Experiences with Expansive Soils
in Highway Subgrades", 1975.

*S fcBSfSil
Figure 2.4 Distribution of potentially expansive
materials in the United States: FHWA Reqion 4. Ref. 31:
"A Review of Engineering Experiences with Expansive Soils
in Highway Subgrades", 1975.

O Z Of < Zb. u. 0 < o < ou < =*- -
Figure 2.5 Distribution of potentially expansive
materials in the United States: FHWA Region 6. Ref. 31:
"A Review of Engineering Experiences with Expansive Soils
in Highway Subgrades", 1975.

Figure 2.6 Distribution of potentially expansive
materials in the United States: FHWA Regions 7 & 8.
Ref. 31: "A Review of Engineering Experiences with
Expansive Soils in Highway Subgrades", 1975.


high: highly expansive ano/OR high
lap compiled by D. U. Patrick. H. K. Woods, art Frederick L. Sarth,
Engineering Geology and Rxk Mechanics Oivision, U. $. Amy
Engineer waterways Etpermenl Station, Vicksburg, Us.
Figure 2.7 Distribution of potentially expansive mater-
ials in the United States: FHWA Regions 9 & 10. Ref. 31:
"A Review of Engineering Experiences with Expansive Soils
in Highway Subgrades", 1975.

1. High highly expansive and/or high frequency
of occurrence.
2. Medium moderately expansion and/or moderate
frequency of occurrence.
3. Low Generally of low expansive character
and/or of low frequency of occurrence.
4. Non-Expansive These areas are primarily
underlain by materials which, due to their physical pro-
perties, doe not exhibit expansive properties.
As can be shown from these figures, expansive
soils extend throughout a majority of the continental
United States. However, the highest degree of frequency
of both low, moderate and high primarily exist in areas
west of the Mississippi River, in the Great Plains, and
the mountainous states.

3.0 Sampling Techniques
Prior to the construction of a roadway, soil
samples are obtained to determine the soil profile along
the alignment, and to obtain laboratory specimens for
testing. Prior to beginning a sampling program, the
engineer should have a thorough understanding of the
geology and localized conditions such that appropriate
sampling of the various materials present beneath the
highway can be obtained.
Typically, two types of sampling techniques are
utilized for roadway construction. The most common type
is disturbed sampling. Disturbed samples are obtained by
the use of bulk samples from auger cuttings, from backhoe
pits, or by the use of split-barrel samplers. The pur-
pose of the disturbed samples is to obtain samples to
determine the subsurface stratigraphy, and for classifi-
cation testing.
If tests such as consolidation, percent swell and
swell pressure are required, then relatively undisturbed
samples can be obtained. Undisturbed samples can be

obtained by the use of Shelby tubes, core barrels, or by
the use of driven ring samplers (i.e. California
Although these types of sampling procedures yield
relatively undisturbed samples, sample disturbance is a
factor in evaluating the laboratory test results. The
friction within the sampler at the soil interface can be
reduced by the use of silicone or teflon sprays. The
more problematic type of disturbance with regard to
measured volume change results from stress relief in the
sample when it is extruded from the sampling device and
stored prior to testing. The extrusion disturbance can
be reduced by testing the materials in rings cut directly
from the sampler, or by using a lined ring sampler in
which the rings are of sufficient diameter and thickness
to be directly used in the testing apparatus. Regardless
of which type of procedure is used, the exposure time
between extrusion and testing should be kept to an abso-
lute minimum such that volume and moisture changes are
3.1 Laboratory Testing
Laboratory tests are performed to determine the
properties of each soil layer. It is necessary to deter-

mine the expansive potential of these soils prior to
construction such that appropriate design and construc-
tion measures can be used to minimize their effect on the
Three general testing techniques are used for
determining the expansive potential of soils. The first
type is indirect techniques. Indirect techniques are
where one or more of the soil properties are measured and
compared with experience to determine the potential
volume changes. Typical indirect techniques are soil
composition, index properties, classification systems,
chemical and physical properties. Direct techniques
involve the measurement of volume change, typically in an
oedometer-type testing apparatus. Direct techniques
generally measure swell pressure or percent swell,
depending upon the information required. The third is
basically a combination technique in which data from both
indirect and direct techniques are correlated to develop
general categories with regard to probable severity of
3.2 Indirect Techniques
As described previously in this report, a large
number of soil properties and environmental conditions
influence volume change; therefore, a large variety of

indirect techniques used to determine potential volume,
change. Table 3.1 describes a majority of the published
techniques {Ref. 32).
The most common indirect technique used for the
identification and classification of expansive soils is
the index property. These properties can be routinely
determined, and experience has shown that the volume
change potential correlates reasonably well with liquid
limit, plasticity index and shrinkage limit. Although
there is a correlation between these factors, most state
agencies use a combination of the Atterberg Limits with
prior experience with similar geologic materials.
Furthermore, in one area a plasticity index less than 20
may be considered as only providing minimal problems; in
other areas the same plasticity index may cause moderate
to large problems.
Another simple test which could be used in
unknown areas to identify the potential for expansion is
the material's reaction with water. The extent to which
a material slakes, disintegrates when submersed in water
indicates its attraction for water, and also indicates a
relative degree of expansiveness. Materials which slake
immediately upon the introduction of water and which,
when stirred, become almost completely dispersed, usually
are highly expansive.

Indicator Group
Property and/or
Soil Composition Clay mineralogy by Measure of diffraction characteristics of clay minerals
X-ray diffraction when exposed to x-radiation. Procedure permits qualita-
tive and semiquantitative identification of clay mineral
components, based oh structural differences between the
clay minerals. Solvation techniques identify expansive
clay minerals.
Clay mineralogy by Identification is based upon exothermic and/or endothermic
differential ther- reactions which occur at particular temperatures. The
mal analysis (DTA) type of reaction and temperature are functions of minera-
logy. Heating rates, grain sizer and sample size
influence results. Multicomponent samples are difficult
to analyze.
Clay mineralogy by Measure of selective absorption of infrared radiation by
infrared radiation hydroxyls in clay minerals. Fair indicator, but not
Clay mineralogy by Qualitative indicator based on selective absorption of
dye absorption different types of dyes by different clay minerals.
Accuracy increases if more than one mineral is present.
Clay mineralogy by
Measure of the radiofrequency electric properties of clay-
water systems. Dispersion is the measure of the dielec-
tric constant at two frequencies. Good indicator of type
and amount of clay minerals. Some problems evolve when
mixtures of different expandable minerals are present in
the soil.
Physiochemical Cation exchange
Measure of the ion absorption properties of clay minerals.
CGC increases from a minimum for kaolinite to a maximum
for montmorillonite Good indicator of hydration proper-
ties of clay minerals.
Measure of the type of cations absorbed on the clay miner-
als. Does not directly relate to swell potential but
rather to the expected degree of swell from ion hydration
Table 3.1, Indirect Techniques for Identification/Classification of Expansive Soils,
Ref. 32: "A Review of Engineering Experiences with Expansive Soil in Highway Subarades
1975y~p. 54-.

Tndicator Group
Index Properties
Soil Classifica-
tion System
echnigues for Identificatlon/Classification of Expansive Soils (cont'd) (Ref. 32)
._________ _____.Description . . . _____
Property and/or
Colloidal content Measure of percent by dry weight basis of particles less
from hydrometer thari 1 micron in size. Indicator of amount of clay but no
analysis reference to type of mineral. Non conclusive.
Specific surface Measure of available clay mineral surface area for
area of clay hydration. Fair indicator of amount of clay mineral and
particles to some extent the type, since moritmorillonite minerals
are very fine and result in large specific surface areas
for given samples.
Soil fabric by ele- No direct measure of swell potential. Primarily used for
ctron microscopy studies of the influence of soil fabric on volume change.
Structure by X- Good for determining the extent of cracks and fractures of
radiography undisturbed materials which will influence moisture move-
ment. No direct measure of swell potential.
Atterberg limits Measures of the plasticity and shrinkage characteristics
of expansive soils. Liquid limit (LL) and plastic index
(PI) correlate reasonably well with swell potential pri-
marily because there are good correlations between them
and the type and amount of clay minerals present. For
shrinkage limit and shrinkage index (LL-SL) the property
of volume reduction is correlated with swell potential
because of similarities between the phenomena. Some of the
published classifications based on Atterberg limits axes
Degree of Expansion PI Shrinkage Index Shrinkage Index LL
Low 12 15 20 20-35
Medium 12-23 15-30 20-30 35-50
High 23-32 30-60 30-60 50-70
Very high 32 60 60 70-90
Extra high 90
Linear Shrinkage
Measure of shrinkage from a given moisture content.
Reasonably good indication of swell potential.
A-6 and A-7 and borderline soils to A-4, A-6, and A-7
generally have high swell potentials.
Pedological classification system in which the vertisol
order is by expansive soils.

To date, no uniform and reliable testing proce-
dure has been developed which accurately simulates the
field conditions. However, for most pavement projects,
typically tests which measure percent swell are used
since the confining overburden pressures are minimal. It
also should be realized that the laboratory test usually
over-predict the percent swell, since it is unlikely that
the soil mass in the field could be saturated at the same
rate as in the laboratory procedures.
Generally, the data from direct testing is used
to determine the relative volume changes of the soil to
determine the appropriate treatment techniques required.
However, the percent swell is seldom used directly in the
design of the pavement.
3.3 Combination Techniques
Combination techniques involve the correlation of
test results from both direct and indirect tests in an
attempt to provide a better estimate of ultimate volume
change. The following sections represent some of the
more widely used techniques for determining the probable
expansion percent, along with a very brief description of
the method.

3.3.1 Bureau of Reclamation Method (Ref. 34):
The Bureau of Reclamation Method correlates pro-
bable expansion with Colloid content, plasticity index
and shrinkage limit. The measured volume change was
taken from oedometer swell tests using a 144 pcf surcharge
pressure from air dried to saturation. The relationships
are shown below:
Colloid Content PI SL Expansion Percent Degree of Expansion
15 18 15 10 Low
13-23 15-28 10-16 10-20 Medium
20-31 25-41 7-12 20-30 High
28 35 11 30 Very High
Experience has indicated this method correlates
reasonably well with expected behavior, and provides a
good indicator of potential volume change. The major
problem is that the colloidal content indicates amount
but not type of clay constituents.

3.3.2 Chen Methods (Ref. 35):
Chen attempted to simplify the United States
Bureau of Reclamation method to eliminate the need for a
hydrometer analysis and to use the relative measure of
the soil density. A correlation was determined between
oedometer swell data obtained on undisturbed samples
under a 1,000 psf surcharge and percent passing a No. 200
sieve, liquid limit, and standard penetration resistance.
The resulting correlation is shown below:
Standard Probable
% Minus No. 200 LL % Penetration Blows Per Foot Expansion % Degree of Expansion
30 30 10 1 Low
30-60 30-40 10-20 1-5 Medium
60-95 40-60 20-30 3-10 High
95 60 30 10 Very High
This method has been used on similar geologic
soil conditions. However, attempts to correlate density
with standard penetration, although successful in cohe-
sionless material, typically are not meaningful in cohe-
sive materials.

3.3.3 Vijayvergiya and Sullivan Method (Ref. 36):
This method has used results of the oedometer
swell test under a 144 psf surcharge, and compared that
to the liquid limit and dry density. They developed a
family of curves which relates the above parameters with
quantitative volume change (see Figure 3.2). The corre-
lation with the laboratory data is good; however,
experience with the system is somewhat limited.
3.3.4 Nayak and Christiansen Method (Ref. 37):
This method develops a relationship for the pre-
dicted swell percent and the predicted swelling pressure
in terms of plasticity index, percent clay content and
initial moisture content. The two relationships to esti-
mate swell percentage and swell pressure are shown
Sp = (0.0229) (PI)1-45 + 6.38
SP = Predicted Swell Percentage
PI = Plasticity Index, Percent
C = Clay Content, Percent
Wi = Initial Moisture Content
/i 2
Pp = (0.035817) (PI)1^2 + 3.7912
Pp = Predicted Swelling Pressure, PSI

Figure 3.2 Correlation of percent swell, liquid
limit, and dry unit weight, Ref. 36: "Simple Tech-
nique for Identifying Heave Potential", 1973.

The correlation with measured oedometer data is
very good; however, field experience is somewhat

After an expansive soil has been identified and
characterized by using the testing procedures described,
a discussion must be made whether to modify the subgrade
soils to minimize expansion and pavement distress, or to
place the pavement directly on these soils and repair
pavement damage as remedial maintenance. All factors
must be considered when making this decision. The fac-
tors include, but are not limited to, the size of pro-
ject, amount of funding, use of the highway (i.e.
interstate, primary, secondary, county, etc.), and the
economics of various types of procedures along with their
various success rates. This section will attempt to
define and describe technical guidelines of various
treatment alternatives or combinations of treatments that
effectively minimize volume change and, therefore, damage
to pavements. The major options available to minimize
volume change and pavement damage can be grouped into the
following categories:

1. Avoid the expansive material and obtain a
different route for the pavement.
2. Remove the expansive soil and replace with
non-expansive soil.
3. Apply a surcharge pressure to the expansive
4. Pre-wetting the soil.
5. Chemically stabilize the soil.
6. Compaction control.
7. Membrane control.
8. Explosive treatment.
Every major literature review or conference on
swelling soils (Ref. 38 through 45) has reiterated these
various remedies. Each year, new strides and innovative
techniques have been developed for utilizing these
methods with literally hundreds of documents published.
It would be impossible to tabulate and review all of
these publications. However, the following paragraphs
summarize various case histories and general conclusions
concerning the above methods.
4.0 Avoid the Expansive Material
The optimum solution for expansive soils would be
to avoid the materials entirely. However, because the

selection of highway routes is primarily based on social,
economic and environmental conditions, the actual soils
which the roadway must traverse typically are of minor
concern in the planning process. With the vast amount of
information available regarding expansive soils, their
locations, distributions and severity, the devastating
impact of expansive soils on the pavements should be con-
sidered during initial conception of the highway. In
some instances, highways could be altered by less than
several miles and the expansive soils could be avoided.
This alternative for dealing with expansive soils should
not be overlooked because significant economic savings
could result.
The advantage of this method is that the expan-
sive soil can be avoided, and that special construction
procedures to deal with the expansive soil problem will
not be required. The major disadvantage is that the
location of the highways is generally determined based
upon transportation needs and right-of-ways and, there-
fore, it is normally not possible to relocate the road-
way .

4.1 Removal of Expansive Soil and Replacement with
Non-Expansive Soils
Removal of the expansive soil and the importing
and placement of non-expansive soils would seem to be an
obvious method to eliminate problems. This alternative
has the most success when the expansive soil layer is
relatively thin, and replacement materials are nearby.
If the import source is a far distance away, then the
cost of the haul becomes excessive and the economic
savings with this alternative no longer exists.
Typically, the expansive soils do not exist in
very thin lenses and, therefore, this procedure normally
removes a specified depth of the expansive soils and then
places a layer of non-expansive soils. The required
depth of excavation depends upon the expansiveness of the
soil and the weight of the proposed fill. The design
concept here is that the weight of the non-expansive fill
is sufficient to counteract the expansive pressures of
the native soils. For this type of analysis, a labora-
tory testing procedure which measures the swell pressure
under minimal volume changes is generally utilized.
The type of fill material placed over the expan-
sive soils is critical. It was suggested (Ref. 46) that
an impermeable soil be used since pervious materials may
collect water and actually provide a path of water entry

into the expansive materials.
The State of Colorado has utilized this alter-
native at the Strausberg East and West Clifton, Palisade
and Four Corners Project. The soil was a swelling
subgrade A-7-6 (20 +) (Ref. 47). This swelling clay
subgrade was replaced with sand or open graded sand and
gravel, in the top 2, 4 and 6 feet. It was found that
the clean, open graded materials provided a reservoir for
water beneath the pavement structure. Although the top
2, 4 or 6 feet did not swell, the subgrade material below
and on the sides of the clean material swelled even more
than would have been expected had the gravel not been
They concluded that the clay subgrade should not
be replaced with any material which is open graded and
will let moisture in, either by surface runoff or hydro-
genesis .
South Dakota (Ref. 48) indicated that limited
undercutting (6 to 18 inches) and replacement with addi-
tional sub-base did not solve their warping problems.
However, when lime was added to the.base course gravel,
results showed that the amount of warping was con-
siderably less than when the lime was not used (Ref.
49) .

Wyoming (Ref. 50 and 51) has had similar
experience with the use of gravel bases over expansive
subgrades to that of Colorado. An experimental project
performed on 1-25 near Kaycee resulted in a build-up of
moisture in the granular sections which was then followed
by heaving and surface cracking. Prior to the advent of
heavy loads and faster speeds of current traffic, Wyoming
constructed roads with a steep crown and dirty gravel
base layers. When they began to build straight align-
ments with shallow side slopes, several feet of exposed
gravel bases were used and were typically in the areas
where snow was piled. The result was swelling soils.
Wyoming's initial reaction was to thicken the
granular section. However, the thicker the gravel that
was placed, the more it heaved. They attempted to top
some of the gravel with some dirtier select material, and
had better results.
This alternative is extremely beneficial when the
expansive soil layer is very thin, and it can be com-
pletely removed. A partial removal of the expansive soil
has resulted in both success and failure, primarily
because it is difficult to determine how much of the
material must be removed. The major disadvantage of this
alternative is that, typically, the expansive soil layers
are very thick; and that adequate non-expansive, low per-

meable fill material usually is not readily available in
close proximity to the alignment.
4.2 Applying a Surcharge Pressure
An obvious alternative to reduce the volume
change of an expansive soil is to provide a surcharge
load of a sufficient amount of fill to counteract the
estimated volume change. This application is typically
only suitable for soils with low expansive properties
because of the large amount of fill required to coun-
teract the expansion potential. If a source of low or
non-swelling, low permeable material is available and the
final alignment is to be raised, this may be a valuable
alternative which should be considered. The weight of
the fill should be determined by the amount of vertical
pressure required to equal the measured swell. For this
type of analysis, a swell test which indicates a con-
fining pressure under zero or a very low volume change
should be used.
The R value method of pavement design developed
by California Division of Highways is partially based on
a requirement that the pavement weighs enough to prevent
expansion. An expansion pressure reading is given in
psi, and the weight of the pavement structure, including

wearing course, base course and any additional non-
expansive material being equal to or greater than the
tested expansion pressure.
The use of applying a surcharge load can be used
in conjunction with the replacement alternative discussed
in the previous section. Utilizing both of these alter-
natives together should provide a stable pavement struc-
One obvious advantage of this alternative is
that, if sufficient material can be applied to resist the
expansion potential of the soils, theoretically no expan-
sion should occur. This option is most advantageous when
the soils have a low estimated volume change and moderate
to large fill sections are required. One obvious disad-
vantage of this method is that, in cut sections, con-
siderable excavation and replacement may be required. In
addition, this method requires a precise estimate of
potential volume change, which is highly dependent upon
test methods and long-term surface and groundwater con-
4.3 Pre-Wetting the Soil
The object of this alternative is to allow the
swelling of the soils to occur and reach equilibrium

prior to the construction of the roadway. The most com-
monly used method for pre-wetting the soils is ponding,
and has been used in Texas since 1934 (Ref. 52). Wise
and Hudson further state that ponding has had rather
limited use due to its cost, length of time required, and
the wet, soft subgrades which develop. It takes between
1 to 3 months to produce satisfactory results and, there-
fore, may only be economically feasible when extremely
high volume changes are predicted.
It appears that (Ref. 53) ponding works best on
fractured or fissured soils which have subgrade moisture
contents which are higher and are more stable at greater
depths. Typically, this alternative works best when the
moisture content is stable and high at a depth below 8 to .
10 feet, and lower and variable at depths above 8 to 10
feet. It is also best to apply this alternative during a
dry season when the natural cracks and fissures are open.
To accelerate the moisture entry to clay soils,
the use of bore holes filled with sand has been used.
This reduces the vertical flow path of the water, and
accelerates vertical entry. These vwells are generally
spaced at less than 10 feet, and backfilled with sand.
In 1958 (Ref. 54), a section of 1-35 north of
Waco across the lower member of the Taylor Marl was
ponded for 22 to 41 days. After a period of 24 days,

analysis indicated that water did not penetrate deeper
than 4 feet. However, ponding was continued for 41 days
and it did have an effect on soils at the 20-foot depth.
After 7 years of service, 2 of the 15 ponded test sec-
tions have become rough, while several unponded sections
in the area have heaved, and have been overlayed and
Ponding was used on a section of U. S. Highway 90
west of San Antonio, Texas, passing through the Taylor
Formation in 1970. This area was a 27-foot cut. An area
across the main driving lanes, median, shoulders, and 3
feet of the backslope of the hill was ponded for 45 days.
It was observed that little water reached below 3 feet
due to the lack of a fissured system in the clay, and the
resulting swelling was primarily confined to the upper 4
feet. After 45 days, the areas were lime stabilized to
aid in holding the moisture in the soils. Observations
performed through 1975 indicate that the effectiveness is
inconclusive at this time; however, the ponded sections
appear to have required less maintenance than adjacent
areas. ^
The Texas Highway Department performed a study to
record the depth of moisture variation due to ponding on
a 5-mile length of Highway 30 near New Boston, Texas.
This project has measured differential movements at

various depths, and variations in moisture content. The
initial moisture contents and densities to a depth of
approximately 20 feet were determined (Ref. 56). The
study was initiated November 12, 1968. Moisture content
measurements indicated that the time required to increase
the moisture contents to what appeared to be equilibrium
conditions varied from measuring point to measuring point
across the alignment. The study indicated, after a
7-month evaluation, the moisture content to a depth of
approximately 6 feet had an average moisture content of
30% (5% above original). The moisture content below the
6-foot depth increased approximately 2%, and was a fairly
uniform increase to a depth of approximately 20 feet.
The study also showed that the density increased approxi-
mately 4 pcf at a depth of 3 feet, and decreased to
approximately 2 pcf at a depth of 20 feet.
Vertical movement was also recorded at the moni-
toring points. For the 7-month study, the surface soils
moved upward approximately 1 inch more than the soil at
the 4-foot depth, and approximately 1.3 inches more, than
the soil at the 10-foot depth. This\ study further shows
that ponding is effective; however, the length of time
required to achieve uniform moisture and vertical rise
varies significantly and, generally, takes a substan-
tially long period of time to occur.

The advantage of this alternative is that it
allows the soils to swell and reach equilibrium prior to
the construction of the roadway. This eliminates dif-
ferential movement due to expansion, provided the
moisture content can remain in its wet condition. The
major disadvantage of this method is it only works well
on highly fractured and fissured soils. Also, it takes a
considerable amount of time and cost to adequately
increase the moisture content to a sufficient depth such
that expansive soil problems do not develop. In addi-
tion, this method usually results in a very soft and
unstable subgrade which may create construction dif-
ficulties .
4.4 Chemical Stabilization
Chemical stabilization is one common method used
for altering the properties of the clay soils to prevent
or minimize swelling. Various chemicals or additives
have been tried. Cementation by lime flyash and cement
has been used in addition to utilizing hydroxides,
chlorides, phosphoric acids, carbonate sulfates, silcona-
tes, asphalts, all have been attempted or used to reduce
the expansive properties of clays. Table 4.1, taken from
Mitchell (Ref. 57) summarizes the various chemical addi-
tives which have been used to control volume change.

Methods for Volume Change Control Using Additives (Ref. 57)
Method of Additive Effects on Soil . Method of Application Comments
Lime treatment
Reduce or eliminate swelling Remove# mix# replace
by ion exchange# floccula- or mix-in-place
tion# cementation, altera-
tion of clay minerals
Only suitable for shallow
depths. Mixing difficult
in highly plastic clays.
Delay between initial
addition of lime and
final mixing and place-
ment improving .ease of
handling and compaction.
2.6 percent lime usually
Treat depths to 36 in. Can
use conventional equip-
ment. Requires careful
quality control
Lime slurry injec- Controversial. Limited by
tion: lime piles slow lime diffusion rate.
May be effective in
fissured material.
Cement treatment Reduce or eliminate swell-
ing by cementation, ion
exchange, and alteration
of clay minerals
piles and walls
Not yet investigated. Might be
Might be suitable in
highly plastic soils for
treatment to large depth.
Could use dry lime, lime
mortar# or slurry
Remove, mix, replace Cement may be less effec-
Plant mix tive than lime in highly
Deep-plow (?) plastic clays. Mixing
difficult in highly
plastic clays
Deep-plow method not yet
investigated (?)
Reduction in swelling
noticeable for cement
contents > 4-6 %
Table 4.1, Methods for Volume Change Control Using Additives, Ref. 57: "Control of
Volume Changes in Expansive Earth Materials", 1973, pp. 200-216.

Table 4.1
Methods for Volume Change Control Using Additives (cont'd) (Ref. 57)
Method of Additive______Effects on Soil_________Method of Application_________ Comments
i Phosphoric acid
"Compaction aids"
Various effects have been
measured or hypothesized,
Reduced plasticity
Improved compaction
Reduced swell
Preservation of soil
Increased strength
Increased or decreased
Usually remove, mix,
and replace or mix-
in-place. In some
instances spraying
or injection is
used. ElectrO-
osmosis may be
useful in special
cases. Diffusion
may be effective
Problems of mixing or
injection may be signifi-
cant. No chemical addi-
tives for control of
volume change appear to
be available that are
effective, permanent and
economically competitive
with lime or cement when
large volumes of soil
must be treated
Calcium chloride may be
effective at least tem-
porarily in soils with
expanding lattice clays.
It may be useful in soils
with a high sulfate con-
tent. A number of
proprietary formulations
have been marketed. The
beneficial effects of
these materials have not
generally been documented

Lamb (Ref. 58 and 59) utilized the results of
several research projects to determine the most effective
stabilizing agent for clay soils. He concluded that the
ideal stabilizing agent should defuse and penetrate into
the expansive clay. Although this was the optimum pro-
perty, no chemical agent was found with this property.
However, hydrated lime was the most effective stabilizing
agent. Lime stabilizes the soil by replacing the active
sodium ions found in montmorillonite with less active
calcium ions. The results of tests reported by Lamb show
that a lime stabilizing agent will reduce volumetric
expansion from 5% to 9% to below 2%. In addition, swell
pressures will decrease from more than 1500 psf for an
untreated soil to about 500 psf for specimens compacted
at moisture contents above optimum.
In addition to reducing the expansion potential
of the clay soils, the lime-stabilized, clay soil is
stronger and, therefore, has a higher bearing capacity
than a natural clay. The stronger materials allow the
engineer to design a thinner and less expensive pavement
structure. In addition, the stabilized layer acts as a
membrane to minimize water entry from the top. However,
the membrane has less effect when cycles of wetting and
drying make it crack and allow moisture to enter from the
top again.

As stated previously, the lime does not have the
migrative property desired; therefore, it must be mecha-
nically mixed into the soils. This mixing is usually
done with a disc harrow or a small ripper. If the
material is not removed, the depth of penetration is
generally limited to 6 to 10 inches. However, for fill
construction, the lime can be applied and mixed in the
borrow area and, for subexcavation, the lime can be mixed
and applied in the backfill stockpiles.
Before determining if lime modification is
suitable for the project, it should be determined if (a)
the soil and lime will react, (b) how much lime (in per-
cent) is required to achieve the desired volume reduc-
tion, and (c) how much influence on soil properties will
be required to produce the volume change reduction (Ref.
To answer questions "a" and "b", the lime content
must be determined. This is referred to as the "lime
modification optimum" (LMO), which minimizes the volume
change. By definition, this refers to the lime percen-
tage that maximizes the reduction in plasticity (Ref.
61). The LMO can be accurately estimated using the
lime-pH test suggested by Eades and Grim (Ref. 62). The
procedure is summarized in Table 4.2.

1. Representative samples of air-dried, minus No. 40
soil to equal 20 gm of oven-dried soil are weighed to the
nearest 0.1 gm and poured into 150-ml (or larger) plastic
bottles with screw tops.
2. Since most soils will require between 2 and 5 per-
cent lime, it is advisable to set up five bottles with
lime percentages of 2, 3, 4, 5, 6. This will insure, in
most cases, that the percentage of lime required can be
determined in one hour. Weigh the lime to the nearest
0.01 gm and add it to the soil. Shake to mix soil and
dry lime.
3. Add 100 ml of C02~free distilled water to the
4. Shake the soil-lime and water until there is no
evidence of dry material on the bottom. Shake for a
minimum of 30 seconds.
5. Shake the bottles for 30 seconds every 10 minutes.
6. After one hour, transfer part of the slurry to a
plastic beaker and measure the pH. The pH meter must be
equipped with a Hyalk electrode and standardized with a
buffer solution having a pH of 12.00.
7. Record the pH for each of the lime-soil mixtures.
If the pH readings go to 12.40, the lowest percent lime
that gives a pH of 12.40 is the percent required to sta-
bilize the soil. If the pH did not go beyond 12.30 and 2
percent lime gives the same reading, the lowest percent
which gives a pH of 12.30 is that required to stabilize
the soil. If the highest pH if 12.30 and only 1 percent
lime gives a pH of 12.30, additional test bottles should
be started with larger percentages of lime.
Table 4.2, Summary of Lime Stabilization Test Procedure,
Ref. 62: "A Quick Test to Determine Lime Requirements for
Lime Stabilization", 1966.

Figure 4.1 shows the pH versus the lime content
and the influence of lime on the Atterberg Limits for
samples from Site 12, Hayes, Kansas (Ref. 60). From
Figure 4.1, the LMO is approximately 4%. The liquid
limit decreases and the plastic limit increases, which
results in a net decrease in the plasticity index. The
maximum reduction occurs at approximately 4%. The use of
Atterberg Limits combined with the pH test is recommended
since it provides verification of the lime percentage and
measures its effect on a given soil.
Several indicator tests and procedures utilize
the PI of the soil to determine the expansion potential.
This is also the case when lime stabilization is used.
If a 50% reduction in the plasticity index for soils with
a PI greater than 35 is not obtained at the LMO, then
using lime to minimize swell should not be utilized. If
the PI initially is less than 35, the reduction should be
to a PI of 15 or less.
An extensive lime stabilization project was
undertaken at the Dallas-Fort Worth Airport. This was
one of the world's largest lime stabilization projects,
which consumed approximately 300,000 tons (or approxi-
mately 30%) of the annual lime used in the United States
for stabilization (Ref. 63).

5 -
' 3 2
percent lim
4 e
0 L
2 4
Figure 4-1 pH and water content versus percent lime for
samples from sampling site 12, Hayes, Kansas
Ref. 60: "Technical Guidelines for Expansive
Soils in Highway Subgrades", 1979, pp. 92-95.

The subsurface conditions consisted of 8 to 16
feet of highly plastic, expansive clay over blue shale of
the Eagle Fort Formation. The plasticity index varied
from 40 to 70, and had an estimated potential vertical
rise equivalent to 10% of the thickness of the layer (0.8
to 1.6 feet). It was determined that approximately 75
days of waiting time after the completion of the lime
stabilization was required for the moisture contents to
reach equilibrium. The final pavement section consisted
of 9 inches of stabilized base and 15 to 17 inches of
concrete pavement. After the lime stabilization was per-
formed, an asphalt emulsion was used to seal the surface
and reduce moisture loss due to evaporation.
The thickness of the stabilized layer throughout
the airport varied in depth depending upon the applica-
tion. The aprons are underlain by an 18-inch stabilized
layer, while the taxiways and runways have a 9-inch sta-
bilized layer. The highways have 24 inches of modified
clay subgrade composed of 6 inches of lime-stabilized
material and 18 inches of lime-modifi^ed material. Six to
7% lime was used for stabilization, and 3 to 4% for modi-
fications .
The mixing was not considered complete until 100%
of the material passed through a l-l/2-inch sieve and 60%

passed through a No. 40 screen. This amount of pulveri-
zation was required such that large unstabilized clay
chunks were not present in the stabilized layer.
Stabilization was performed using both dry lime
and lime slurry. Dry bulk spreading was the fastest
method of placing the lime. The dry lime was placed by
utilizing large truck transport.
To convert the dry lime to a slurry, the lime was
blown to the transport through a pipe into a mixing
device where it was combined with water to produce a mix-
ture containing 1 part lime to 2 parts water by weight.
After mixing, the slurry was applied to the subgrade at a
40 to 60 psi pressure using a typical water truck. The
application rate was sufficient to produce a dry lime
content in the stabilized layer of 6%. During placement,
additional water was added because the lime required
excess water in order to react properly.
To prepare the subgrade for the lime, the soil
was initially broken into maximum sizes of 4 to 6 inches.
The thicker 18-inch stabilized layer was placed in two
consecutive lifts utilizing disc harrows to mix the
Although this process involved a considerable
amount of work, Kelly (Ref. 63) felt the money was well
spent, considering the impact an uneven pavement surface

could have. Results indicate adequate performance.
The South Dakota Department of Transportation
constructed an experimental road section to evaluate the
performance of various stabilized agents (Ref. 64). The
test section was composed of 2 inches of asphalt, 5
inches of base course, 3 to 6 inches of sub-base, and
select import varying from 6 to 18 inches, over stabi-
lized sections of the Pierre Shale. Several stabilizing
agents were used:
1. Lime 6%
2. Lime-Asphalt, 6% Lime Plus 4% RC1
3. 5% Phosphoric Acid Plus 2% ferric sulfate
4. PDC Formula, 4:2:1 Lime, Portland Cement,
Soy Flour for a total of 5% additive.
The results of the experimental testing show that
all stabilizing agents altered the physical properties to
some degree. However, lime had a more permanent effect
in altering the properties. Lime, Lime plus RC1 and the
PDC Formula, in that order, caused significant increases
in the CBR value, while after 4 years the phosphoric acid
section was only slightly higher than the untreated soil.
Serviceability ratings for the stabilized sections except
the phosphoric acid were better than those of the stan-
dard design. Lime treated sections had the best rating,
while PDC and lime plus RC1 closely followed. The study

concluded that the phosphoric acid was not an effective
stabilizing agent for the Pierre Shale, and that the
effect of the PDC Formula was primarily caused by the
lime-cement combination rather than the soy flour addi-
tive. It was concluded that 6% lime altered the expan-
sive properties the greatest, and also was the most
economical alternative.
South Dakota has completed a study of 13 projects
where lime stabilization was used between 1965 and 1975
(Ref. 65). These projects were built prior to 1968 when
the South Dakota Department of Highways was just
beginning to use lime as a stabilization area. Their
design concept at that time was that a small amount of
lime could modify the physical characteristics signifi-
cantly to reduce the expansion potential.
They concluded that a small addition of lime only
adds, to the fines, especially if there is not a suf-
ficient amount of clay minerals to provide the pozzolanic
reaction. Also, cohesionless soils treated with minimal
amounts of lime were prone to water accumulation and
frost heave. a
The results of their analysis completed in 1975
indicates that the plasticity index of the treated soils
for all projects varied from 4 to 11 points under the
untreated soils in 1968, and that, in 1975, the PI is

still 2 to 13 points below that in 1975. The higher per-
centage points are the projects where 5 to 6% lime was
used, and the lesser reduction in PI is typically where
3% lime was used. They also indicated that the lime
treated soil on all 13 projects, regardless of the amount
of lime used, has maintained a fairly uniform moisture
content throughout the 11-year period. They ultimately
concluded that, if a sufficient amount of lime is mixed
with a reactive clay subgrade, a permanent increase in
shearing strength is obtained. In addition, a reduction
of the expansive property which results in improved ride,
serviceability is also obtained. The use of lime stabi-
lization reduced warping and surface irregularities due
to expansive soils, and also increases the CBR value
which provides a greater support for the pavement and
ultimately reduces the base, sub-base and wearing surface
thicknesses of the pavement structure.
The methods previously described are typically
used when the lime is mixed with a set of ripping teeth
or a disc harrow into the soil. Another technique for
utilizing lime stabilization is the, drill hole method.
With this method, holes are drilled into the subgrade and
backfilled with a lime slurry or lime slurry-sand mix-
ture. This process can be quite slow unless a sufficient
amount of cracks and fissures extends throughout the

depth which is being treated. This drill hole technique
has been used for both new construction and for remedial
measures by several highway agencies.
The South Dakota Department of Transportation
(Ref. 66) in a remedial measure used a lime slurry com-
posed of 1 part lime, 1 part water and 1 part sand. This
was placed in 4-foot deep holes on 5-foot centers into
Pierre Shale subgrade. The test results showed some
reduction in the frequency and sharpness of surface
bumps. A definite improvement in serviceability in time
was noted in the stabilized areas compared to adjacent,
non-stabilized areas. These studies further showed that
lime migration from the hole basically was confined to
the periphery of the hole. The success with this tech-
nique arises from (a) an increase in moisture content of
the surrounding subgrade due to migration of water, aided
by lime from the hole; and (b) stress relief of the
lateral expansive pressure due to the placement of the
holes, thereby reducing upward swell pressures.
The chemical which has had the most success in
stabilizing expansive clay soils has been lime. The
advantages of a lime stabilized subgrade is that the
addition of lime substantially reduces the expansion
potential of the clay soils. In addition, the lime sta-
bilized soil is stronger, and has a higher bearing capa-

city which will allow a thinner, less expensive pavement
structure. The lime can be mixed in the borrow area,
with a disc harrow in cut areas (for thin stabilized sec-
tions), or in backfill stockpiles. Also, the amount of
lime required to stabilize the soil can be readily deter-
mined. The major disadvantage is that typically the clay
soils are very dry, and the clay chunks must be broken
down into relatively small particles such that the lime
can be uniformly blended into the soil.
4.5 Lime Slurry Pressure Injection
To obtain greater distribution of lime in the
soils, a technique of lime slurry pressure injection
(LSPI) was developed. This technique consists of pumping
lime under pressures up to 200 psi through hollow injec-
tion rods into the subgrade. The actual soil pressure
used depends upon the soil conditions. The injection
holes are typically placed at 3 to 5-foot on centers
(Ref. 67 and 68). The injection rods penetration the
soil at approximately 1 foot intervals, and slurry (2.5
to 3 pounds of lime per gallon of water) is injected
until refusal. Refusal is defined as:
1. The soil will not take additional slurry.
2. Slurry is running freely out of the other

injection holes or around the pipe.
3. The slurry has fractured the surface at some
point and is flowing outward.
The movement of the lime slurry through fine
grained soils occurs through the cracks, fissures and
other discontinuities. Therefore, the actual penetration
into the soil mass is limited. Most of the reaction
occurs at the surface of the crack or fissures. To uti-
lize this treatment, as with other lime treatment
programs, the clay soil must be reactive in lime, and
also must have sufficient fractures and fissures to
accept the injected slurry and allow movement through the
soil mass.
The two most commonly used injection plans are
single injection (4 feet deep) and double injection (3
feet deep). The single injection method is the minimum
recommended treatment, and is approximately 20% less in
cost than 6-inch deep stabilization with 6% lime on a
recompaction basis (Ref. 69). For this method, the
injection holes are spaced on 5-foot centers, and a
volume of 2.5 pounds of lime per sqiiare foot of soil is
used. Approximately one-fourth of the total quantity is
left on the surface from refusal in the injection proce-
dure. This lime is then mixed into the top 6 inches, and
recompacted following completion of the injection.

Although the amount remaining on the surface is not of
sufficient quantity to obtain full stabilization of the
layer, it does significantly modify the plasticity, and
improves the performance of the expansive layer. Also,
the inclusion of the lime slurry in the top 6 inches
reduces the permeability of the initial layer.
For more highly expansive clay soils, double
injection system is used. The initial injection grid is
approximately 5-foot on center, and a second grid is
diagonally offset after 48 hours from the first grid.
Approximately 3.5 pounds of lime per square yard is used
with this method. The advantages of the double injection
1. More lime is installed
2. Secondary fissures are filled with lime.
3. Lime is more evenly distributed into the
4. The soils are additionally pre-swelled due to
the second injection.
5. More lime is available on the surface for
mixing into the top 6 inches.
Estimates show that the cost for double injection
3 feet deep is approximately the same as 8 inches of a
conventional 6% lime and recompaction (Ref. 69). The
overall quantity of the stabilized layer depends upon the

quantity of lime installed, injection spacing, depth, and
the number of injection passes. Because of the effects
of lime concentrating around the seams, the quality of
the LSPI treatment cannot be evaluated by conventional
tests such as Atterberg Limits, pH, swell, or strength
tests on recovered samples.
The literature indicates conflicting reports con-
cerning the effectiveness of LSPI, and it seems logical
to conclude that LSPI is an effective swell control pro-
cedure under certain circumstances. However, it is not a
cure-all for all conditions. The conditions most
favorable to a successful LSPI treatment appear to be the
presence of an extensive fissure and a crack network in
which the slurry can be successfully injected. The
treatment mechanisms explaining the LSPI effectiveness
include prewetting, the development of lime moisture
barriers which effectively reduce moisture migration, and
a reduction in swell due to the soil-lime reaction at the
cracks and fissures.
The Woodbine Corporation, since 1968, has stabi-
lized a significant amount of highways, streets and
parking lots using the LPSI system. These projects range
from interstate highways to parking lots for small
shopping centers. The depth of treatment varied from 3
to approximately 12 feet, but generally ranged from 3 to

4 feet. They indicated that jobs completed in 1974 and
1975 have had recent observations in 1986. Their recent
observation indicated that these pavements are in good
condition, and that the treatment has reduced expansive
soil related distress (Ref. 70).
The benefit of lime slurry pressure injection is
that you can obtain a greater distribution of lime in the
soils without removing the material, and utilizing a
disc. This method is most advantageous in cut sections
where conventional methods would require complete removal
and replacement of the materials. The major disadvantage
is that the slurry primarily moves through cracks and
fissures, and if the soil is not highly cracked and
fissured most of the reaction occurs at the interface of
the cracks. If this method is used, the injection type
and sequence for each installation must be carefully
developed, and the effective depth of penetration around
the bore holes should also be monitored to determine that
adequate penetration into the soil mass has occurred.
4.6 Compaction Control \
The placement, moisture content, density and the
method of compaction used have significant effects on the
amount and rate of swelling of a compacted expansive soil

(Ref. 71 through 76, 17). Research performed by Holtz
and Gibbs (Ref. 73) describes some of the influences.
Their tests indicate that an increase in water content at
compaction for a given density decreased the swell and
the swell pressure. Also, that an increase in compaction
density at any given moisture content usually increases
but may decrease the swell, especially at a high moisture
content, depending upon the range of densities involved.
However, generally, an increase in density caused an
increase in swell. From their results, it could be
concluded that low densities and high water contents
result in lower degrees of expansion. Figure 4.2 plots
the percent expansion for various moisture-density com-
paction conditions under a 1 psi surcharge load. Figure
4.3 plots the total uplift pressure caused by wetting for
various placement conditions.
Seed and % Chan (Ref. 44) observed, in their
experiments, that soils compacted dry of optimum have
higher swelling characteristics and increase to a higher
moisture content than do soils at the same density com-
pacted wet of optimum. \

1 iov.^e.6 : 0. 1.7 6.5 & ft [L-l VE
0% "7I 6% 6.9 7/ 5. ^7"-- . 2 t7~7 D% \
44/ 0 _35_ J.8_ ^7
SA mple no. rM-X247 i-
10 15 20 25 30
W 100
r-^r _
19.2 100% SA1 ( SPECIFIC U AT ION C GRAVITY URVE = 2.749 )
10 10 | - ~ 4 1 17.9 A 10.3 95\.
6 4 *T.Z 4.7 '4.8^_
2 >1 L 2.1 *Ti SAMPLE n t D.7M-X24I i -
10 15 20 25 30 35
Figure 4.2 Percentage of expansion for
various placement conditions under a 1 psi
surcharge, Ref. 73: "Engineering Properties
of Expansive Clays", 1954
Figure 4.3 Total uplift pressure
caused by wetting for various place-
ment conditions, Ref. 73: "Engineer-
ing Properties of Expansive Clays"

Seed's and Chan's tests were performed using a
dynamic compaction process. The increase or decrease in
swell upon an increase in density for any given water
content occurs because changes in densities at high
degrees of saturation are accompanied by changes in soil
structure. The swelling characteristics reflect the
effect of both a change in soil structure and moisture
The method of compaction also influences the
swelling characteristics of a compacted soil. An expan-
sive soil with a dispersed (defloculated) structure
swells less than the one with a floculated structure
under the same water content and density. Kneading com-
paction leads to a defloculated structure and, hence,
less swell. Figures 4.4 and 4.5 plot the swell pressure
versus dry density for two types of soils with both a
static and kneading compaction process.
The results of the compaction studies appear to
indicate that swell and/or swell pressure can be mini-
mized if the soil is compacted to a moderate density at a
water content greater than optimum using a compaction
process which produces a defloculated structure. Typical
compaction equipment which produces a kneading action is
a sheepsfoot roller. It has also been determined (Ref.
77) that compaction in this manner with a sheepsfoot

Figures 4.4 and 4.5 Effects of method of compaction
on swell pressure saturation, Ref. 74: "Study of
Swell and Swell Pressure Characteristics of Compac-
ted Clays",1962, pp. 12-39.

roller slightly above optimum moisture will result in a
soil permeability which is at or very close to the mini-
mum for that soil. Therefore, any subsequent swelling
due to the movement of water through this soil will be
very slow. One disadvantage to volume control by com-
pacting the soil to a high moisture-low density condition
is that it produces a fill which is considerably weaker
and has lower strength properties than for compaction dry
of optimum;
North Dakota (Ref. 78) prior to 1967 had a com-
paction specification equal to 90% of AASHTO T-180 maxi-
mum dry density, and a minimum moisture content of 75% of
optimum moisture. This criteria resulted in compacting
the soil to a relatively high density at a fairly low
moisture content. They concluded that this condition was
conducive to expansion. Since 1967, the compaction spe-
cifications have been changed to 85% of AASHTO T-180, and
the minimum moisture content is optimum moisture. North
Dakota believes that using this standard, the soils are
compacted at a higher moisture content and, thereby, the
expansion potential should be reduced. Experience has
shown, particularly in one project, that a section of the
highway constructed under the old standards developed
some very rough joints. An adjacent section of the road

constructed in accordance with the new compaction stan-
dards, had virtually no rough joints and the ride on this
pav-ement is noticeably different. North Dakota believes
that, by the use of compaction control ana continuously
reinforced pavement, they have virtually eliminated pave-
ment roughness due to expansive soils.
The South Dakota Department of Transportation
(Ref. 79) has experienced difficulties in the long term
performance of roadways over the expansive Pierre Shale.
The specifications require the undercutting of the top 3
feet of earth subgrade for the full roadbed width,
shoulder slope to shoulder slope. The backfill material
for this 3 feet is composed of select, weathered soil
suited for that purpose. The lower (3 to 6-foot deep)
zone is confined to the area between the subgrade
shoulder lines (see Figure 4.6 below).
Figure 4.6 Typical grading sectionRef. 79: "Stabiliza-
tion of Expansive Shale Clay by Moisture-Density Control",

The material between the 3 to 6-foot depth is constructed
of normal soil using higher moistures and lower minimum
densities than for the underlying embankment (depths
greater than 6 feet). The specifications further
required that the target low density of 92% of AASHTO
T-99, and the water content at compaction should be 3%
above optimum mositure. Because of the heavy construc-
tion equipment used to compact the soil, the target den-
sity was not obtainable, and the specifications were
revised to a minimum density in the upper 6 feet of
subgrade to 92% of AASHTO T-99, with a target density of
95%. The water content was set not lower than optimum
with a target content of 3% above optimum.
The construction tests indicated that the average
moisture and densities were reasonably close to the goals
set; however, in-depth investigations of the subject on
one project showed that the results were not quite as
close as the construction tests indicated. A review of
89 tests taken within the upper 6 feet indicated an
average density of 99.5% of T-99, and a moisture content
of approximately 2% above optimum moisture. The increase
in density was due to the construction equipment, and a
portion of the loss of moisture content occurred during
the compaction process.

Roughness index measurements have been performed
since the initial construction (1968 through 1975). The
roughness index measurements indicate that the special
moisture and density controls used appear to have
retarded the adverse effects of the expansive soils.
South Dakota is very pleased with the results, since the
roughness ratings measured after 8 years previously could
not be maintained for 3 to 4 years with their previous
type of construction. Although these moisture-density
controls probably will not completely reduce the effect
of the expansive soils, the study after 8 years indicates
that the ridability will be much greater over a longer
period of time, and will result in significantly lower
maintenance costs.
The Wyoming State Highway Department has experi-
mented with moisture density control (Ref. 80). In areas
where interbedded layers intersect subgrade, Wyoming
feels that moisture density control is an excellent
alternative. The use of moisture density control in this
situation produces a more uniform subgrade, and elimina-
tes the sharp, choppy heaves often produced by inter-
bedded layers encountered at the surface. However, they
also believe that the use of moisture-density control in
the harder shales places moisture in an area where it
ordinarily would not reach, and that a better approach

would be to prevent moisture intrusion.
The requirement of a moisture percentage of opti-
mum moisture has theoretical justification; however,
there are some limits to its practical application with
regard to the materials encountered in Wyoming. The
laboratory material used for moisture-density curves is
thoroughly broken up, and is limited to minus #4 sieve
material. The material in the field is never broken up
to this degree so, in effect, a more granular fill at a
moisture content far above optimum for the actual field
condition is placed. The result is shale fragments or
clay "clods" dry on the inside, with free water in the
voids between them. This results in an unstable material
which can cause internal breakdowns or expansion after
surfacing is placed, even if additional moisture content
is kept out.
The Colorado Department of Highways has performed
studies regarding construction in arid regions with and
without moisture-density control (Ref. 81). Their work
has shown that satisfactory fills can be constructed with
swelling soils if good moisture-density control is used.
However, moisture-density control for the entire depth of
deep fills may not be economically justified if water is
at a premium. "A comparision of two similar roadways,
one constructed with moisture-density control and the

other without, shows that 33% of the distress observed in
the latter occurred in fills where no moisture-density
control was used, while no distress was observed in the
fills where moisture-density control was used."
(Ref. 81)
The research of the Colorado Department of
Highways has resulted in a recommended implementation
package for swelling soils in Colorado (Ref. 82).
Pavement distortions from swell have been found to be
most prevalent on soils which classify as A-6 and A-7
groups, and on borderline soils between A-4 and the A-6
and A-7 groups. In addition, certain A-2-6 and A-2-7
soils which are borderline to the A-6 and A-7 have pro-
duced swell. The suggested depth of moisture-density
control below grade for cuts and the top of fills for
interstate and primary highways are described below:
Interstate and Primary Highways:
Depth of
Treatment (feet)
Over 50

Secondary and State Highways:
Depth of
Treatment (feet)
Over 50
The obvious benefit of moisture-density control
is it can be conveniently utilized in fill sections.
Moisture-density control must be maintained in all fills
to some degree; therefore, both the contractors, engi-
neers and field technicians are very familiar with the
procedures. Numerous highway departments have utilized
moisture-density control with generally good success.
The disadvantage primarily occurs in areas where water
may be at a premium. In addition, more extensive field
testing is required to control the moisture and density
within the specifications. Also, there may be some long-
term problems if only moisture-density control is used,
since the moisture content of the in-place soils may
change with time. For moisture-density control to per-
form adequately, the clay soils must he adequately broken
down such that there is a uniform moisture content
throughout the fill, and that dry clay clods are not pre-
sent .

4.7 Membrane Control
It has been determined that changes in moisture
content is the primary factor which determines the amount
of volume changes of swelling soils. Therefore, it
appears obvious that, if expansive soils can be isolated
from moisture changes, volume changes can be reduced or
minimized significantly. Waterproofing membranes have
been used, and are a promising method for limiting access
of water infiltration and minimizing moisture changes.
They are primarily used when the effect of the moisture
content is related to surface intrusion, and not regional
and seasonal variations in groundwater flow.
The preconstruction application of waterproofing
membranes can be achieved in several ways (Ref. 84).
1. Continuous sprayed asphalt membrane over the
entire subgrade and ditches.
2. Full-depth asphalt pavement with a sprayed
asphalt or synthetic membrane beneath the roadway and
3. Full-depth asphalt pavement with paved
ditches and cut sections.
4. Vertical synthetic fabric membrane cutoffs.
Waterproofing membranes, particularly con-
tinuously sprayed asphalt membranes, have been used sue-

cessfully in many states. Asphalt membranes perform best
when applied over the entire subgrade section, down the
slopes and up the backslope (typically 1.5 feet above the
ditch invert or 6 to 12 inches above the finished pave-
ment elevation). Membranes typically perform best where
the soil profile is relatively dry, the moisture profile
is relatively uniform with depth, the groundwater table
is at a sufficient depth and has no influence on the
near-surface behavior, and the climate is a dry, semi-
arid to arid climate.
It was recommended (Ref. 83) that for interstate
highways the asphalt membrane should be applied from
backslope to backslope for all cut sections. For fill
sections, the membrane should extend a sufficient
distance down the slope to assure that moisture cannot
reach the compacted fill (typically 1.5 to 2 feet below
finished pavement grade). For divided highways where the
median is less than 2 highway lanes in width, the treat-
ment should be continuous through both lanes and the
median. Typical examples of continuous sprayed asphalt
membrane and full-depth asphalt pavement with sprayed
asphalt membrane beneath the ditches is shown in
Figure 4.7.

' -----1 ...........r 11, i' . '

*1 v ii1; i' 1 Figure 4.7 Typical sprayed asphalt membrane applications to minimize subgrade moisture
variations from surface infiltration, Ref. 83: "Technical Guidelines for Expansive Soils
in Highway Subgrades", FHWA-RD-79-51, 1979, pp. 98-103.

Figure 4.8 Example of vertical membrane cutoff construction, Ref. 83: "Technical
Guidelines for Expansive Soils in Highway Subgrades", FHWA-RD-79-51, 1979, pp. 98-103

Vertical membrane cutoffs of synthetic materials
placed vertically along the edge of the pavement have not
been used extensively, primarily because of construction
problems and the availability of high-strength fabrics.
Typically, they are installed by trenching and installing
the membrane; then the trench is backfilled as discussed.
An example of a vertical membrane cutoff applied to a
two-lane road is shown in Figure 4.8. The depth of the
vertical membrane should extend to the depth of the
active zone. Typically, this is not done due to
construction difficulties and economics. However, it has
been found that membrane depths of less than 2 to 3 feet
will generally not provide enough protection to warrant
their application. The backfill used in the trench is
also critical. The backfill for at least 24 inches below
the surface should be an impermeable soil. If the trench
is deep (greater than 5 feet), granular backfill may be
used in the lower several feet. The upper material, as
previously stated, should be impermeable and should be
sprayed with asphalt.
The Colorado Department of Highways (Ref. 84
through 88) has had significant success with asphaltic
membranes in controlling moisture changes and heaving.
The Clifton project (Ref. 88), which was performed on
cuts through the Mancos Shale, used several types of

treatments to reduce expansion. A description of the
treatments is given.
Section Number Treatment
1 and 7 Subexcavated 2 feet, backfilled with coarse aggregate (3/4" to 1/4").
2 and 8 Subexcavated 2 feet, backfilled with Class 2 sub-base (95-100% minus 3-inch and 3-15% minus No. 200).
3 and 5 Subexcavated 2 feet, backfilled with fine sand.
4 and 6 Subexcavated 2 feet, backfilled with structural backfill (granular).
9 and 11 Scarified 1 foot deep, wetted and com- pacted .
10 4% hydrated lime mixed in the top 1 foot of subgrade.
12 Lime shafts 6" diameter, filled with hydrated lime paste.
13 1% hydrated lime mixed in the top 3 feet of subgrade, sprinkled for three weeks, and compacted.
14 Scarified top 3 feet, sprinkled for three weeks, and compacted.
16 Subexcavated 2 feet, top of subgrade covered with asphalt membrane and back- filled with fine sand.
17 Subexcavated 2 feet,, backfilied with A-4 material (silty soil).
18 Subexcavated 2 feet, applied asphalt membrane, and backfilled with silty soil.

CBR tests indicated this shale had a potential
swell of 0.4 to 3.1%, with an average of 1.6%. This pro-
ject on 1-70 concluded that two sections using 3/8-inch
thick, catalytically blown asphalt membranes out-
performed the other treatment measures used in the study.
The tests further show that the granular base course over
the membranes were wet and that, after 5 years the
membranes were in excellent condition, pliable, uniform
in character, and showed no signs of movement or
cracking. Test sections on State Highway 96 near Ordway,
Colorado, were constructed to evaluate several full-depth
asphalt-stabilized base mixtures in various thicknesses,
as related to each other and to standard untreated aggre-
gate designs. Although the thinner asphalt-stabilized
base sections were designed for a 3-year life span, ser-
viceability indexes after more than 7 years of service
indicated that none of the sections are close to failure.
Moisture studies indicate that the subgrades under the
asphalt bases are continuing to remain drier than the
untreated bases. In addition, the asphalt-stabilized
bases have resulted in the underlying subgrades having
field CBR values near 14, and moisture contents of
approximately 14% after 7 years of service. The sub-
grades under the untreated bases have field CBR values of
9 at 18% moisture.

Test sections on U. S. Highway 40 west of Elk
Springs indicate that sections constructed with a
3-l/2-inch or thicker asphalt stabilized base and
membrane lined ditches were performing well. However,
sections constructed with a 4-inch wearing surface and no
base, and sections with 12-inch thick clay bases enve-
loped with catalytically blown asphalt membranes, have
experienced both structural and heaving failures.
An experimental section of U.S. Highway 12 over
the Pierre Shale was evaluated by the South Dakota
Department of Transportation (Ref. 89). The upper 6
inches of subgrade was treated with a mixture of lime and
RC-1 asphalt to form a waterproof cover. Also, a
polyurethane plastic blanket was placed vertically to a
depth of 4 feet at a distance of 20 feet either side of
the center line, just inside the shoulder. The results
indicated that there were no significant differences in
moisture contents of the sections with moisture barriers
than those without. It appears that the polyurethane
plastic cutoff was not placed deep enough, and that the
fractured shale permitted moisture to move beneath the
wall. The moisture seemed to be higher, and fluctuated
more in the area close to the barrier itself, possibly
indicating that a thermal change may be causing conden-
sation near the plastic cutoff.

The Wyoming Highway Department has used plant mix
bituminous bases and membranes, in effect to reduce the
moisture content increase of expansive soils (Ref. 90).
Wyoming believes this is an economical means of dealing
with expansive soils, and this has been used as a normal
design consideration for several years. The results have
been very good, and appear to be far better than anything
else that they have used. This treatment should not only
be applied beneath the surface; however, should extend to
ditches and other areas where seepage may enter beneath
the pavement structure.
The Arizona Department of Transportation also has
used membrane control. In the summer of 1973, an experi-
mental project was constructed on U. S. 180 near the
southern boundary of Petrified Forrest National Park. In
one cut section, a catalytically blown asphalt membrane
was placed on the subgrade prior to placement of the
select material and asphaltic concrete (Ref. 91). In the
other cut sections, full-depth asphaltic concrete was
placed directly on the subgrade. In all cut sections,
the shoulders were paved with asphaltic concrete to the
ditch bottom and up the back slope 1 foot. These treat-
ments worked very well in preventing moisture intrusion
into the subgrade.

In addition to utilizing membranes for new
construction, membranes have also been used to rehabili-
tate pavements over expansive soils. The Arizona
Department of Transportation has performed several
experimental studies using membranes for overlay pro-
jects. Results of their studies have led Arizona to
implement the membrane process of control on an opera-
tional basis for all expansive soil subgrades (Ref. 91).
Initially, a leveling course is applied to the
old pavement, and then the shoulder slopes are bladed and
compacted to a uniform slope. Before the asphalt rubber
membrane is applied, the soil slopes are primed with a
light coating of emulsion {0.08 gallons per square yard).
The asphaltic rubber mixture is then applied to the
slopes. The asphalt mixture consisted of 25% ground
vulcanized rubber, which reacted with asphalt at tem-
peratures of approximately 350 F. The asphaltic rubber
is applied at a rate of 0.75 gallons per square yard on
the earthen shoulder slopes, and 0.6 gallons per square
yard on the leveled asphaltic concrete surface. The rub-
berized asphalt applied to the asphaltic concrete is
covered with chips to provide a wearing surface, and
facilitate load transfer. The membrane on the shoulders
is protected with 6 inches of soil (see Figure 4.9).

1.5" AC
1.5" AC
2.75" AC AZ-MO
Figure 4.9 1-40 Typical Sections,!Ref. 91: "Membrane
Technique for Control of Expansive Clays", 1979.

Prior to rehabilitating the roadways, they were
distorted, cracked, and required excessive maintenance.
The condition the roadways were in took approximately 10
to 12 years to reach an unacceptable level of roughness
and maintenance. Field observations plus measurements
indicate that the membrane treatment over the badly
distorted highways has improved the overlay performance.
The improvement is attributed to the membrane's ability
to seal out moisture. Although moisture redistribution
is taking place, the significant benefit of using membra-
nes for existing pavement is to decrease moisture fluc-
tuation from seasonally dry to wet conditions. In
addition, it also reduces reflective cracking to some
degree. Based on their extrapolated calculations, it
appears that the roughness is increasing at a rate that
it will take approximately 33 years to reach the objec-
tionable degree of roughness which the original pavement
experienced in 10 to 12 years. Arizona's experience with
the use of membranes on existing pavement has lead to the
following conclusions:
1. Such treatment will reduce reflective
cracking by adding some tensile reinforcement at the con-
nection and reducing water entry into the cracked pave-
ment. The proper membrane will redistribute and relieve
stresses caused by the existing pavement.

2. Shoulder and ditch membrane treatment
{asphalt, rubber, paved shoulders and ditches, or any
other suitable waterproofing membrane layer) can and will
improve the long-term ride performance of a highway over
expansive soils.
3. If ride is improved and cracking severity is
reduced, maintenance costs are also lower.
The obvious advantage to membrane control is that
it reduces changes in moisture content, which is the pri-
mary cause of expansive soil problems. Another benefit
is that membranes can be used for both new construction
and for rehabilitation of existing pavements which do not
require complete removal and reconstruction. The use of
membranes can also control both surface and subsurface
water flow beneath pavements. Generally, membranes
beneath the pavement work best in dry climates where the
primary source of water is from precipitation and surface
runoff, and not regional or seasonal variations in
groundwater flow. One disadvantage is that when vertical
moisture barriers are used, they must be placed deep
enough to vent moisture flow beneath the pavement struc-
ture. The cost of the membrane may be more expensive
than other forms of stabilization, depending upon the
thickness and type of membrane required. In addition,
the membrane must be protected such that it maintains its