2013-018

2014
research-articleResearch ArticleXXX10.1144/qjegh2013-018J. Hussain et al.Effects of Moisture and Clay Minerals on Base Course

Downloaded from http://qjegh.lyellcollection.org/ at Purdue University Library on June 28, 2016

The effect of moisture and relative proportions of clay

minerals (smectite, chlorite and illite) on the performance
of unbound granular base course (UGM)
Jawad Hussain1, Douglas J. Wilson1* & Philippa M. Black2
1Department of Civil and Environmental Engineering, School of Engineering, University of Auckland,
Private Bag 92019, Auckland, New Zealand
2School of Environment, University of Auckland, Private Bag 92019, Auckland, New Zealand

*Corresponding author (e-mail: dj.wilson@auckland.ac.nz)

Abstract:  An increase in moisture accelerates the deterioration of a pavement and this is especially
so for unbound granular materials below the pavement surface. One of the possible reasons for this
accelerated deterioration can be the mineralogical makeup of the aggregates, which can include clays.
This research paper discusses the effect of the relative proportions of clay minerals on the performance
of different base course materials in repeated load triaxial (RLT) tests considering moisture variation.
The materials selected for this study were sourced from greywacke sedimentary rocks from both the
North and South Islands of New Zealand. The mineralogical makeup of the clay content present in the
base course materials was determined by X-ray powder diffraction (XRD). The performance of these
materials was further tested in RLT tests at different moisture and drainage conditions. The results of
the XRD and RLT tests showed that aggregates with high relative proportions of smectite clay minerals
can decrease the load-bearing capacity of the base course material when moisture is introduced into
the pavement materials.

The unbound base course material has an important load-bearing minerals present in the clay size fraction are converted into expand-
function in flexible pavements. In high traffic volume roadways, ing clays. The relative proportion of sand to clay in the finer aggre-
bound bituminous or concrete layers are laid over this unbound gate fractions and the presence of expanding clays are currently
base course to reduce the stresses at the top of the base course layer. tested in New Zealand (NZ) by performing the sand equivalent test
However, in chip sealed or very thin bituminous asphalt wearing (SE) and the clay index (CI) test respectively (Standards New
courses over unbound granular layers, the function of the wearing Zealand NZS 4407 1991). These current NZ test methods have
course is not as a structural layer but more for waterproofing and been shown to be inadequate in characterizing the effect of the clay
other properties such as skid resistance, drainage, maintenance and on the performance of the aggregate as they do not provide a
dust reduction. The unbound aggregates must therefore transfer the detailed insight into the mineralogy of the clays (Lowe 2010).
dynamic load stresses of traffic from the top layer to the subgrade Lowe recommended that the current tests (NZS 4407 tests 3.5 and
layer such that the stresses are reduced to a level at which the top of 3.6 for Clay Index and Sand Equivalent respectively), which were
the subgrade can withstand the distributed loads and thus prevent originally based upon US standards, need to be revised to take bet-
fatigue failures. These unbound materials when appropriately ter account of the clay mineralogy found in NZ and to take account
selected and designed demonstrate good resistance against perma- of both the quality and quantity of the clay fines in the aggregate.
nent deformation when they are drier than the optimum moisture Knowledge of minerals present in the clay size fractions can help to
content. However, the resistance to permanent deformation can identify the nature of the problems with the base course material.
decrease considerably with an increase in moisture above optimum Some clays, when immersed in or saturated with water, expand by
moisture contents. absorbing water in the interlayer position and this expansion can
There can be many reasons for the decrease in the strength of reduce the dynamic load-bearing capacity of the material when it is
these unbound base course materials with an increase in moisture, subjected to highly moist conditions for long periods.
such as the following. The strength of the aggregate and moisture susceptibility of the
• The expanding clays in the matrix of the compacted base clays contained in the aggregates that are used as base course mate-
course can absorb water and become plastic, which can cause rials are also affected by the amount of geological weathering of
excessive deformation. the source rock (Hodder & Hetherington 1991; Velde & Meunier
• The lubrication of the aggregates decreases the friction at the 2008) as well as the aggregates’ susceptibility to further weathering
aggregate particle to particle contact area. when in soaked condition while in service. The common clay min-
• Pumping of the finer fraction (fines) from the aggregate cre- erals found in the fine particles of NZ greywacke base course mate-
ates cavities within the base course material, decreasing the density rials are chlorite, smectite, illite and kaolin. This study was mostly
and increasing interstitial stresses. concerned with the swelling properties of these minerals, which
In this research paper only the expanding clays are addressed. result in a weakening of the base course material, thereby reducing
Expansion of clay is purely a material property, which can be the pavement life under loading. The smectite group minerals are
linked to a specific quarry, or can be linked to weathering of the highly swelling and can exist either as a separate mineral or as a
unbound material in which the existing clay and other unstable mixed interlayer mineral, usually with chlorite or illite (Murray

Quarterly Journal of Engineering Geology and Hydrogeology, Vol. 47, 2014, pp. 169­–176 © 2014 The Geological Society of London
http://dx.doi.org/10.1144/qjegh2013-018
Published Online First on April 22, 2014
 Downloaded from http://qjegh.lyellcollection.org/ at Purdue University Library on June 28, 2016

170 J. HUSSAIN ET AL.

 
&* $
 
0* %
 )* 
&
3HUFHQWDJH3DVVLQJ 

3HUFHQWDJH3DVVLQJ 
 
&
 
 
 
 
 
 
 
         
6LHYH6L]HPP 6LHYH6L]HPP

D E
Fig. 1. Gradation curves of base course materials: (a) CAPTIF; (b) road failure.
2006). There are many species of smectite; for example, sodium
montmorillonite, calcium montmorillonite, saponite (Mg, Fe-rich),
nontronite (Fe3+-rich) and beidellite (Al-rich). Sodium montmoril-
lonite has a 10–15 times swelling potential when placed in water.
Calcium montmorillonite has a swelling index of 2–3, which is
very much less than that of the sodium form (swelling index 24–
36); however, it can still be sufficient to cause an increase in vol-
ume of the base course material when subjected to moisture that
will accelerate pavement failure. In contrast, kaolin, illite and chlo-
rite have very low absorption and swelling properties (Murray
2006).
In this study we tested unbound base course materials in two
phases: first, to identify the mineral composition of the materials,
and, second, to compare the performances of the materials. X-ray
powder diffraction (XRD) data were collected from oriented sam-
ples of the clay fraction in their untreated and then glycolated state
to identify expanding minerals, such as smectite, in the clay size
fractions of the unbound granular base course (Wilson 1987). After
performing XRD tests on the materials (five greywacke aggregates:
three from the South Island that were used in coincident Accelerated
Pavement Testing (APT) tests and two North Island greywacke),
repeated load triaxial (RLT) tests (TNZ T/15 2010) were performed
to compare their relative engineering performance. The relative
proportions of the clay minerals smectite, chlorite, illite and kaolin
were estimated from the ratios of areas under the curves of their
basal XRD peaks following the methodology of Moore & Reynolds
(1997, pp 298-327).
The South Island aggregate materials used in this study are
taken from the test pavements of the Canterbury Accelerated
Pavement Testing Indoor Facility (CAPTIF) and from a known
road pavement failure in the North Island of New Zealand. In the
following sections, the research objectives and scope will be given,
along with the materials, methods, results and discussion, and
finally the conclusions.
Objectives and scope
The main objective of this paper was to investigate if there was a
difference in the mineralogy of the materials from the South
Island aggregates that were used in the CAPTIF experiments and
materials from a road failure that occurred in the North Island of
NZ. If a mineralogical difference could be shown then the effect
of different clay mineral proportions on the performance of

unbound granular materials in RLT tests would be further inves-
tigated. A particular aim was to verify whether the presence of
smectite, which is more prone to expansion, if found in a high
percentages could cause a difference in the load-carrying capacity
of unbound base course.
Materials and methods
The aggregates used in the base course materials taken from the
CAPTIF pavement and the road failure were all sourced from
greywacke hard rock. The CAPTIF materials were sourced from
the South Island of NZ whereas the well-known road failure
material was sourced from North Island NZ greywacke source
rocks. The major geological difference between these materials
was found to be the proportion of the various minerals. In the
South Island materials, a significant proportion of chlorite and
illite minerals was found, whereas for the North Island materi-
als, higher proportions of smectite and kaolin minerals occurred.
Initially, the unbound base course materials used were obtained
from the test pavements constructed at CAPTIF, Christchurch. The
materials had three particle size distributions categorized as coarse
graded (CG), medium graded (MG) and fine graded (FG) and were
used as the base course in the CAPTIF test pavements. These three
materials were subjected to various laboratory tests to find out if
there were any significant mineralogical differences in the materi-
als, both in aggregates and in the clay contents of the mix. The test
results showed that the mineralogical compositions of the materials
used in the CAPTIF pavements were largely the same, hence the
difference in the performance of CAPTIF materials could not be
explained by their mineralogical composition. To investigate
whether the mineralogical composition of the materials can affect
the performance of the base course, the materials from a well-
known but politically sensitive road material failure in the North
Island (the location and material sources could therefore not be
identified) were collected and tested for their mineralogical compo-
sition. There was a significantly high proportion of smectite min-
eral in the materials obtained from this pavement failure in
comparison with the CAPTIF pavement materials. The particle size
gradation of the materials used in the CAPTIF and the North Island
pavement failure (taken after failure) is shown Figure 1a and b
respectively. The materials 2A, 4B, 8C (RLT test codes 3004–
3005) and 11C (RLT test codes 3010–3011) were sourced from
greywacke material quarries near the pavement failure location.
 Downloaded from http://qjegh.lyellcollection.org/ at Purdue University Library on June 28, 2016
Effects of Moisture and Clay Minerals on Base Course
Table 1. Test results for base course materials
Location and material Clay index test Sand equivalent test
(NZS 4407: 1991: 3.5) (NZS 4407: 1991: 3.6)
South Island from  
CAPTIF pavements
CG 1.4* 36†
MG 2.0* 24‡
FG 2.2* 17‡
North Island from a  
road failure
2A 3.6‡ 26‡
4B 2.8† 27‡
8C/3004–05 7.6‡ 26‡§
11C/3010–11 4.2‡ 27‡§
*Pass value.
†Near-failure value.
‡Failure value.
§Values for the material near pits, as exact values are not found in data.
These greywacke materials from the North Island of New Zealand
have larger amounts of smectite minerals in the material mixes
compared with South Island greywacke materials and have been
well described by Black (2009). The gradation curves show that the
materials used in CAPTIF test pavements have different gradation
curves. However, all the sample materials obtained from the road
failure are in the same range of particle distribution and met NZTA
M/4 specification particle size distribution (psd) envelopes. A com-
parison can be readily made between the mineralogical composi-
tions of all of the aggregate materials. However, as the psds for the
CAPTIF test pavement are rather varied only the closest psd sam-
ple from the CAPTIF test pavement will be compared with the
North Island road pavement failure samples (i.e. the CG sample).
RLT results that were taken as a post-construction material failure
test were obtained for the 11C road failure material. The 11C mate-
rial test results were labelled as 3010 when the RLT tests were
conducted in dry and drained conditions but as 3011 when the RLT
tests conditions were saturated and undrained.
The test methods usually specified for testing the quality of the
fine fraction of the base course materials are the clay index (CI) test
(3.5) and the sand equivalent (SE) test (3.6) (Standards NZ 1991). The
CI test is used to estimate the percentage of expansive clay minerals
in natural fines by using methylene blue titrations. The SE test is used
to find the relative proportions of silt and clay size particles in the fine
fractions of aggregates. The results of both of these tests are shown in
Table 1 for the base course materials tested and analysed in this study.
The specification for M/4 base course states that the material should
have a value of CI <3 and a value of SE >40. In the specifications
materials are required to satisfy either the SE or the CI test. As the
CAPTIF materials (CG, MG, FG) are virgin materials, they are
selected on the basis of CI. However, in the material obtained from
pavement failures only the 4B material passes the M/4 requirements.
The test results show that there is no obvious correlation between CI
value and SE value. For example, the sand equivalent value shows
that CG has a higher clay-size content with respect to 11C although
the clay index of CG 1 is less than that for the 11C material.
The analytical method followed in this study was divided into
two stages. In the first stage, the mineral composition of the source
rocks and clays present in the fines of the base course material was
identified using petrographic thin-section studies and the standard
XRD test for identifying clay minerals (Wilson 1987). The clay size
fraction was separated using the standard density separation method
and deposited on a silica glass slide and allowed to sediment and
air dry to provide an oriented sample mount. The oriented sample
was then subjected to XRD analysis in three stages: in stage 1, the
171
sample was examined in the untreated state; in stage 2, the same
sample was exposed to ethylene glycol vapour at 40 °C (i.e. glyco-
lated) for 24 h; in stage 3, the glycolated sample was then reheated
to 550 °C for at least 1 h. The untreated XRD sample results show
the current clay mineral state present in the material prior to con-
struction or loading. The glycolated sample results show how these
minerals might behave when base course materials are subjected
to long contact with moisture.
In the second stage of the study RLT tests were carried out; the
experimental setup is shown in Figure 2. The samples were prepared
at optimum moisture content and were then tested at optimum and
higher moisture levels. The materials were tested in accordance with
the draft NZ standard T/15. The same specimens were subjected to
varying stress states in the staged RLT test and for 50000 loading
cycles applied to the sample for each of six stress states (each with
an increasing deviator load). The details of stress states are given in
Table 2. The materials were tested at optimum moisture drained
conditions and saturated undrained conditions. The materials tested
at optimum moisture content had the maximum density that can be
achieved in the field. The unbound materials are considered to have
their maximum load-bearing capacity at these conditions. The satu-
rated undrained conditions are the worst conditions, where material
loses friction between aggregates and has the least load-bearing
capacity. The aggregate clay minerals present in the finer portion of
the material mix come into contact with water in the RLT specimen,
and longer contact of clay minerals with water can show the effect
of moisture on finer particles present in the material.
Results and discussion
Mineralogy of the source rocks
The source for the South Island aggregates used in the CAPTIF
programme is quartzofeldspathic (i.e. arkosic) greywacke in
which micas dominate as both a detrital and a metamorphic
mineral. In contrast, the source quarries for the greywacke
aggregate from the North Island are volcaniclastic sandstones in
which chlorite is the dominant clay mineral. The difference in
the mineralogy of the clay fraction can be seen by comparison
of the relative intensities of the 14 and 7 Å peaks at c. 6 and c.
12.5° 2 theta (the first- and second-order basal spacing of chlo-
rite respectively) with the 10 Å peak at c. 9° 2 theta (first-order
basal spacing for illite–mica) in Figures 3a and 4a.
All the greywacke source rocks contain prehnite and pumpellyite,
and occasionally epidote minerals. The arkosic greywacke contains
abundant metamorphic albite + white mica and occasionally stilp-
nomelane and/or K-feldspar, whereas the volcaniclastic North Island
aggregate rocks contain abundant albite + chlorite as well as illite and
occasionally actinolite. In terms of metamorphism the North and
South Island greywackes have been metamorphosed under very low
metamorphic, sub-greenschist conditions. Estimates of the condi-
tions of metamorphism for rocks containing the observed assem-
blages would be 250–300 °C and 2–4 kbar (Frey et al. 1991).
X-ray diffraction study of clay size
fraction of samples
The results of XRD analysis for minerals present in the clay size
fraction of the base course materials are shown in Figures 3 and 4.
The first peak in the diffractogram at c. 6° 2 theta corresponds to
the basal spacing (i.e. the 001 plane) of both chlorite and air-dried
smectite minerals (at c. 14–15 Å). The second peak at c. 9° 2 theta
is the illite–mica basal spacing (c. 10 Å), whereas the third peak at
c. 12.5° 2 theta (c. 7 Å) coincides with the 002 lattice plane of
chlorite and the first basal spacing of the kaolin group minerals.
 Downloaded from http://qjegh.lyellcollection.org/ at Purdue University Library on June 28, 2016

172 J. HUSSAIN ET AL.

Fig. 2. Repeated load triaxial test setup.

One of the South Island aggregates used in the CAPTIF programme Table 2. Stress states of repeated load triaxial test
(FG) had weathered material added to it to provide a contrast to the
otherwise good quality aggregates used. The effect of weathering on Stress state σ1 (kPa) σ3 (kPa) P = (σ1 + 2σ3)/3 (kPa) Q = σ1 − σ3 (kPa)
the aggregate can be seen in the high background on the low-angle side 1 210 120 150 90
of the illite and chlorite peaks and the shoulder on the low-angle side 2 166.7 66.7 100 100
of the chlorite peak (i.e. hydration swelling of the basal spacing to 3 141.7 41.7 75 100
higher d-spacing) of the untreated sample, and the general lack of a 4 270 90 150 180
sharp expansive peak in the glycolated sample. These are common 5 470 140 250 330
features of weathered chlorites and illites, where interlayer minerals, 6 530 110 250 420
and also the expanding mineral known as vermiculite formed by leach-
ing of potassium from illite, are common (Moore & Reynolds 1997). P and Q are stress tensors.
North Island volcaniclastic greywackes that are altered or
weathered contain kaolin and smectite (Bartley et al. 2007). It is
very difficult to detect kaolin in samples containing chlorite, as the Percentage of relative area under
first-order kaolin basal peak overlaps with the second-order chlo- curve for XRD intensity graphs
rite basal peak. One of the characteristic features of iron-rich chlo-
rite is the very notable increase in the intensity of the 14 Å peak and As the smectite minerals are known to absorb water and revers-
corresponding decrease in intensity of the 7 Å peak on heat treat- ibly expand when they come into contact with moisture, in this
ment, and this feature appears in the diffractograms shown in section we focus on the relative percentage of smectite with
Figures 3c and 4c. There are small but significant differences in the respect to chlorite and illite. The area under the curve for each
position of the 002 and 004 peaks for iron-rich chlorite (at c. 12° mineral has been calculated and the relative percentage of the
and 25° two theta respectively) and the 001 and 002 peaks for kao- area under the curve of the respective minerals is shown in
lin, which allow discrimination of the two minerals (Moore & Figures 5 and 6.
Reynolds 1997, pp. 234–50) but also contribute to broadening of Figure 5 shows that for all CAPTIF test materials, the relative
the observed peaks (as seen particularly in Fig. 3a and b). Step percentage of the smectite mineral is around 20% for all XRD test
heating at 50 °C intervals from 400 to 550 °C also shows the pres- conditions. The dominant minerals (i.e. chlorite and illite miner-
ence of small amounts of kaolinite in the North Island samples. als) are found in approximately equal proportions. Figure 5 shows
Figures 3 and 4 show a clear difference in clay mineralogical that for all CAPTIF test materials the relative percentage of the
composition between the CAPTIF and the North Island failure sam- smectite mineral is around 20% of the total clay mineral content
ples. In the CAPTIF samples the 14 Å basal spacing develops a slight and that the dominant clay minerals (chlorite and illite) occur in
shoulder on the low-angle side of the peak on glycolation (Fig. 3b) approximately equal proportions when the relative intensities of
but is unaffected by heat treatment (Fig. 3c). The 10 Å illite peak the second (i.e. 002 or equivalent) basal spacings of chlorite and
shows little change in intensity although it sharpens with heat treat- illite are also taken into consideration. In the untreated (Fig. 5a)
ment. The exception is the sample FG, which on heat treatment and glycolated (Fig. 5b) diffractograms, the area under the curve
develops a clear shoulder on the low-angle side of the illite peak for the smectite basal peak is notably less than that for the chlorite
(between 7° and 9° two theta) indicating the presence of interlayered and illite minerals.
clay minerals. The North Island diffractograms show clearly that the Figure 6 shows the area under the curve for the basal spacings
untreated sample peaks at 14 Å (c. 6° two theta) shown in Figure 4a of the various minerals in the road failure materials, and it can be
are the combination of a smectite that expands on glycolation to c. seen that the percentage intensity of smectite is less than that of
16 Å (Fig. 4b) and collapses to 10 Å (c. 9° two theta) on heating (Fig. chlorite and higher than that for illite in samples 2A and 4B but is
4c) and a non-expanding 14 Å chlorite peak (Fig. 4a and b). at its highest level in samples 8C and 11C.
 Downloaded from http://qjegh.lyellcollection.org/ at Purdue University Library on June 28, 2016

Effects of Moisture and Clay Minerals on Base Course 173

 0*87  0**/

&OD\)UDFWLRQ &OD\)UDFWLRQ
 &*87  &**/
 2ULHQWHG8QWUHDWHG  2ULHQWHG*O\FRODWHG
)*87 )**/

$
 

$
$

$
 

&RXQWV
&RXQWV

$
 

$
 
 
 
 
 
               
$QJOH'HJUHHV7KHWD $QJOH'HJUHHV7KHWD
D E

 0*+7
&OD\)UDFWLRQ
 &*+7
 2ULHQWHG+HDWHGž&
)*+7


$

&RXQWV


$






       
$QJOH'HJUHHV7KHWD
F
Fig. 3. CAPTIF materials, XRD test, oriented clay fraction: (a) untreated; (b) glycolated; (c) heated at 550 °C.

  $JO

&OD\)UDFWLRQ $XW &OD\)UDFWLRQ
$

 %XW  %JO

2ULHQWHG8QWUHDWHG 2ULHQWHG*O\FRODWHG

$
 &XW  &JO
$

$

&JO
$

 &XW 

&RXQWV
&RXQWV

 
$

$

 
 
 
 
               
$QJOH'HJUHHV 7KHWD $QJOH'HJUHHV 7KHWD
D E

 $K
&OD\)UDFWLRQ
 %K
2ULHQWHG+HDWHGž&
 &K
$

 &K
&RXQWV

$






       
$QJOH'HJUHHV7KHWD
F
Fig. 4. Pavement failure materials, XRD test, oriented clay fraction: (a) untreated; (b) glycolated; (c) heated at 550 °C.
 Downloaded from http://qjegh.lyellcollection.org/ at Purdue University Library on June 28, 2016

174 J. HUSSAIN ET AL.



&$37,) $UHD 8QGHU&XUYH87 &$37,) $UHD 8QGHU&XUYH*/
 
 
 
 
 
 
 
 
 
 
&*87 0*87 )*87 &**/ 0**/ )**/
6PHFWLWH &KORULWH ,OOLWH 6PHFWLWH &KORULWH ,OOLWH
D E
Fig. 5. CAPTIF materials, XRD test, percentage of area under curve: (a) untreated; (b) glycolated.

 
$UHD 8QGHU&XUYH87 $UHD 8QGHU&XUYH*/
 
 
 
 
 
 
 
 
 
 
$ % & & $ % & &
6PHFWLWH &KORULWH ,OOLWH 6PHFWLWH &KORULWH ,OOLWH
D E
Fig. 6. Pavement failure materials, XRD test, percentage of area under curve: (a) untreated; (b) glycolated.



 &*68'
&*20
3HUPDQHQW6WUDLQ 
















      
/RDGLQJ&\FOHV
Fig. 7. Permanent deformation of CG M/4 AP40 base course material used in CAPTIF.

Permanent deformation results from RLT test results for the sample material 11C (3010–3011) from
RLT tests the North Island road failure specimens (Fig. 8). The material
shown as 3004–3005 is a base course from a similar failure site
The RLT test results for the CG materials from the CAPTIF with no weathered aggregates and the least fines, which results
tests are shown (Fig. 7) for the purpose of comparison with the in the least deformation when compared with sample 11C
 Downloaded from http://qjegh.lyellcollection.org/ at Purdue University Library on June 28, 2016

Effects of Moisture and Clay Minerals on Base Course 175

E
Fig. 8. Permanent deformation of materials taken from a pavement failure: (a) saturated undrained RLT; (b) dry drained RLT.
 Downloaded from http://qjegh.lyellcollection.org/ at Purdue University Library on June 28, 2016
176 J. HUSSAIN ET AL.
(3010–3011), which has higher proportions of smectite minerals.
The other materials, 3001–3002 and 3007–3008, are the sub-
base materials, which demonstrate the difference in performance
of materials at optimum and saturated moisture conditions; there
is less resistance to deformation at high moisture conditions.
The mineralogy of the CG material shows that it has lower pro-
portions of smectite mineral in comparison with the chlorite,
illite and kaolin minerals. The RLT test results (Fig. 7) show
that the material performed well at optimum moisture content in
drained conditions, resulting in a 0.3% permanent deformation.
The highest intensity ratio of the smectite was found when com-
paring the remaining minerals that included smectite, which is
known as a problematic swelling mineral. The 11C material
showed a permanent deformation of 0.4% and is shown as 3010
in Figure 8b. The performance of the CG material is better than
that of the 11C material in terms of resisting cyclic loads.
In saturated drained conditions, the CG material resisted the
dynamic loading until the sixth stress state and resisted failure until
more than 260000 loading cycles (Fig. 7). In comparison, the 11C
material, which is shown as 3011 in Figure 8a, failed after only
100000 loading cycles. In the saturated drained condition, the CG
material has clearly performed better than the 11C material. The
apparent reason for the good performance of the CG material in
comparison with the 11C material is the lower percentage of smec-
tite to the other considered minerals. Other possible reasons for the
difference in performance are the gradation and maximum dry den-
sity, which were 2.34 t m−3 and 2.28 t m−3 for the CG and 11C mate-
rials respectively.
Summary and conclusions
This study was conducted to investigate the type and quantity of
minerals present in the unbound base course materials and the
effect of the relative proportion of these minerals on the RLT
performance of various materials. XRD tests were also con-
ducted on the greywacke base course aggregates from the
CAPTIF test pavement in the South Island and a well-known
road pavement failure in the North Island. The relative propor-
tions of the minerals were calculated by using the area under the
curve technique from the results for untreated oriented clay sam-
ples, glycolated samples and samples heated at 550 °C. A further
two greywacke materials were subjected to RLT tests and a
comparative analysis was undertaken for materials from the
North and South Islands. The following conclusions can be
drawn from the test results and analysis.
• The XRD results for materials from the CAPTIF test pave-
ment and the road pavement failure showed the presence of similar
minerals in the greywacke rocks of the North and South Islands
(e.g. chlorite, illite, smectite) but with varying proportions.
• The smectite mineral was seen to expand when in contact with
moisture by absorbing water and to contract when heated. Smectite
was found in relatively high proportions in the North Island aggre-
gates taken from a road failure and in relatively low proportions in
the South Island aggregates taken from CAPTIF test pavements.
• The RLT test results demonstrate that the road failure materi-
als from the North Island greywacke showed relatively higher
deformations than the South Island materials when tested in satu-
rated undrained conditions, owing to the presence of smectite in
Received 1 March 2013; accepted 21 February 2014.
higher proportions in North Island aggregates than in South Island
aggregates.
• High proportions of smectite can cause a significant reduction
in the performance of the same material in high moisture environ-
ments, as the North Island material failed after only 100000 loading
cycles in saturated undrained conditions whereas it survived more
than 300000 loading cycles at optimum moisture content.
The conclusions of this study are based on a programme of lim-
ited experimental work owing to limited resources; however, the
results clearly show that the presence of smectite minerals in higher
relative proportions can lead to the earlier failure of a base course
material in high moisture environments. There is a need for further
research on the comparison of the performance of the aggregates
with different relative mineral proportions, as extensive experi-
mentation is required to determine the point at which a change in
performance behaviour occurs based on the relative proportion of
smectite mineral. The XRD results clearly show the relative pro-
portions of smectite mineral to the other minerals present in a base
course and can help the construction industry in selecting appropri-
ate base course material.
Acknowledgements. The authors acknowledge the input of H.
Arampoorthy and J. Patrick from OPUS laboratories; G. Arnold;
J. Wilmshurst from the Geology Department, The University of
Auckland; and A. Fussell from Canterbury Accelerated Pavement
Testing Indoor Facility (CAPTIF). J.H. is also grateful to Higher
Education Commission (HEC) of Pakistan for funding his PhD
at The University of Auckland, as this paper is a part of his PhD
studies.
References
Bartley, F.G., Harvey, C.C., et al. 2007. Clay Mineralogy of Modified
Marginal Aggregates. Land Transport NZ Research Report No 381.
Black, P. 2009. Geologic Inventory of North Island and Aggregate
Resources: Influences on Engineering Materials Properties. Mineral
Wealth of North Island, Auckland.
Frey, M., De Capitani, C. & Liou, J.G. 1991. A new petrogenetic grid for
low-grade metabasites. Journal of Metamorphic Geology, 9, 497–509.
Hodder, A.P.W. & Hetherington, J.R. 1991. A quantitative study of the
weathering of greywacke. Engineering Geology, 31, 353–368.
Lowe, J.S., et al. 2010. The Proficiency of Sand Equivalent and Methylene
Blue (Clay Index) Test Methods for Determining the Deleterious
Mineral Content of Weakly Metamorphosed Sedimentary Rock. Road &
Transport Research: A Journal of Australian and New Zealand Research
and Practice, 19, 23–36.
Moore, D.M. & Reynolds, R.C. 1997. X-ray Diffraction and the
Identification and Analysis of Clay Minerals, 2nd Edn. Oxford University
Press, Oxford.
Murray, H.H. 2006. Structure and composition of the clay minerals and their
physical and chemical properties. In: Haydn, H.M. (ed.) Developments in
Clay Science, Vol. 2. Elsevier, Amsterdam, 7–31.
Standards New Zealand NZS 4407. 1991. Methods of sampling and testing
road aggregates. Standards Association of New Zealand, Wellington.
Transit New Zealand. 2010. Specification for Repeated Load Triaxial (RLT)
Testing of Unbound and Modified Road Base Aggregates (No. TNZ T/15
(Draft)). Transit New Zealand, Christchurch.
Velde, B. & Meunier, A. 2008. The Origin of Clay Minerals in Soils and
Weathered Rocks. Springer, Berlin.
Wilson, M.J. 1987. X-ray powder diffraction methods. In: Wilson, M.J. (ed.)
A Handbook of Determinative Methods in Clay Mineralogy. Chapman &
Hall, New York.