Analysis of Recycled Portland Cement Concrete Subbase Aggregate Using the Scanning Electron Microscopy and Field Investigation

watchpoorUrban and Civil

Nov 15, 2013 (3 years and 7 months ago)

209 views

Analysis of Recycled Portland Cement Concrete Subbase Aggregate Using the Scanning
Electron Microscopy and Field Investigation

By

David J. White, Ph.D.
(Tentative order)

Associate Professor, Iowa State University

4
22

Town Engineering Building

Ames, IA
50011
-
3232, Tel: (515) 294
-
1463, Fax: (515) 294
-
8216

Email:
djwhite@iastate.edu


Thang Huu Phan

Research Assistant

2711 South Loop Drive, Suite 4700

Ames, IA 5001
0
-
8664
, Tel: (515) 294
-
4510
, Fax: (515) 294
-
8216

Em
ail:
thphan@iastate.edu


Warren E. Straszheim
, Ph.D.

Associate
Scientist
, Iowa State University

46

Town Engineering Building

Ames, IA 50011
-
323
0
, Tel: (515) 294
-
8187
, Fax: (515) 294
-
4563

Email:
wesaia@iastate.edu


Peter Taylor, Ph.D.

Director II, Institute for Transportation, Iowa State University

2711 South Loop Drive, Suite 4700

Ames, IA 50010
-
8664, Tel: (515) 294
-
9333, Fax: (515) 294
-
0467

Email:
ptaylor@iastate.edu


Chuck T. Jahren, Ph.D.

Associate Professor, Iowa State University

428 Town Engineering Building

Ames, IA 50011
-
3232, Tel: (515) 294
-
3829, Fax: (515) 294
-
3845

Email:
cjahren@iastate.edu



Word Count

Text



=

4
146

Tables (2 x 250)

=

500

Figures (6 x 250)

=

1500

Total



=

6
146


Preparation for
presentation and publication by Journal of Materials in Civil Engineering

White
et al.

(tentative header)


2


Abstract

T
he m
icrostructure of
fresh

concrete
has
been investigated

using scanning electron microscopy (SEM)
analysis
,
but the
microstructure

of

old concrete
is rarely studied.
T
his paper
presents

the microstructure of
recycled Portland cement concrete (RPCC) aggregate materials
. This paper also compares
the stiffness of
RPCC aggregate subbases with crushed limestone aggregate subbases
.

Field investigations were
conducted
in Iowa
at

five

sites with
crushed limestone
subbases
and
four

sites with
RPCC aggregate
subbases.
Dynamic cone penetrometer (DCP) and
Clegg Impact Hammer tests were used to characterize
the

stiffness

of these subbases
.
These tests indicated

that
the
RPCC subbase
s

were generally stiffer than
the
crushed limestone aggregate subbase
s. The

average
CIV

and
DPI

values of the RPCC subbases were

180
%

to 200% higher than
the

crush
ed limestone aggregate subbases, which suggest
that
the
cementations of
the
un
-
hydrated cement grains in the RPCC aggregate materials increased the st
iffness of
those

subbase
s
.

Introduction

When roads must be replaced wit
h new construction, it is very expensive to dispose of old concrete
pavement material and to mine, process, and haul natural aggregate for new subbase layers. Since the
early 1980s, the desire for more economical and sustainable construction projects has l
ead to the
increased use of recycled Portland cement concrete (RPCC) as an alternative to natural aggregate (e.g.,
dolomite or limestone).
For example, u
sing RPCC aggregate
for subbase layers construction
has become a
common practice in Iowa

(White
et al.

2008)
.

RPCC
subbase
aggregate
that is
used
to
rehabilita
te

Iowa
roadway
s

i
s usually reclaimed locally from

the
existing

Portland cement concrete (PCC)
pavement

which
largely reduce
s

the cost
s

of hauling
the

waste
material
to

landfill and
purchasing

new
subbase
material
s
. However,
old PCC
pavement

normally

contain
s

cement particles that
were

not fully
hydrate
d

in

the
previous

paving process.
When old PCC
pavement is crushed and used as RPCC subbase aggregate, t
hese non
-
hydrated cement grains
are

expose
d

t
o pore water

from precipitation or groundwater that harden
s

the subbase (Bruinsma et al. 1997).

Further,
compaction during road construction and traffic loads breaks up the RPCC aggregate and provides more
opportunities for cementation.

F
ield and
laboratory stud
ies

investigating the mechanical and performance of RPCC subbase pavement
was conducted.
Stiffness and drainage characterization of the subbase layers were conducted using
dynamic cone penetrometer (DCP
)
and Clegg Impact Hammer

(Clegg 1986)
.

The study suggested that the
RPCC subbase although with generally low permeability, provides a relatively stiff and durable pavement
foundation for rigid pavement as demonstrated by adequate pavement performance (White et al
.

2008
)
.
White
et al.

(tentative header)


3


The study results have

lead to a question
whether

cementation of
the existing cement grains has
strengthened RPCC subbase stiffness.
RPCC aggregates were collected for scanning electron microscopy
(SEM) analysis in the laboratory.

In this study, SEM imaging and X
-
ray microanaly
sis techniques were used to analyze the phase
distribution and chemical composition of the
RPCC
specimens.

Four recycled Portland cement concrete
(RPCC) specimens collected from four different sites were examined. These RPCC aggregates were used
as subbase

aggregates in 1983, 1988, 1994, and 2003.
Field investigations were conducted on these four
RPCC aggregate subbase layers and five crushed limestone aggregate subbases.

Subbase stiffness was
characterized by field testing using dynamic cone penetrometer (
DCP), and Clegg Impact Hammer. Field
testing results were used to determine
whether

the cementation of residual cement grains in the RPCC
aggregates increased the stiffness of the RPCC subbase layers.

SEM Analysis

Background

Scanning electron microscopy (SEM) imaging and X
-
ray microanalysis techniques have been widely used
to study the microstructure of concrete in micrometer definition. High definition images with sub
-
micrometer allow analyzing the complex microstructure of c
oncrete, showing the bonding process of
cement particles. Quality of the cement paste can be analyzed by looking at the internal features of the
material using backscattered electron imaging. Despite several limitations of the method, such as
magnification
, two
-
dimensional surface analysis, and only a small portion of the surface exposed in a
given specimen is documented, the backscatter
-
mode SEM is the only method of observation
that
provides a clear assessment of the concrete microstructure (Diamond 2004)
.

In the SEM process, signals resulting from interaction between the focused electron beam, which is
scanned across the specimen, and the specimen are measured. Particle size, shape, surface roughness, and
fracture surfaces can be found by using images of
topography. Specimens can be polished into smooth
surfaces for phase distribution and chemical composition analyses (Stutzman 2001). The beam of
electrons strikes on the specimen producing signals detected as backscattered electrons (B
S
E), secondary
electr
ons (SE) and X
-
rays. SEs are low energy electrons resulting from the collision of a primary beam
electron with an electron of a specimen atom. These SEs come from the near surface layers, meanwhile
B
S
Es are high energy electrons that come from deep into sp
ecimen layers (Balendran et al. 1998,
Scrivener 2004).

Backscattered electrons are high
-
velocity beam electrons on a polished surface of a specimen. The B
S
E
image contrast is generated by the different phases’ compositions relating with their average atomic
number. Differences in backscatter reflect differences in chemical composition in the cement paste that
White
et al.

(tentative header)


4


can be observed by the different brightness in th
e image. Non
-
hydrated cement appears brightest followed
by calcium hydroxide, calcium silicate hydrate, and aggregate.
Voids appear dark in the image (Kjellsen
et al
.

1990, 1991
, Diamond 2004, Scrivener 2004).

X
-
radiation is a form of electromagnetic radi
ation and produced when a specimen is struck by high
-
energy electrons. The X
-
rays can be analyzed by energy dispersive or wavelength dispersive detectors to
give chemical composition of the scanned area. The X
-
ray signal is generally used to identify the p
resent
of elements and their concentrations and graphical distributions. The relative accuracy of quantitative
analysis ranges from 20% to 1% for the concentration degrees of the elements ranging from 1% to 50%,
respectively (Stutzman 200
4
, Scrivener 2004)
. For cementitious materials, X
-
rays are generated through
an interaction volume of about 1
-
2

m across, which is larger than the size of many hydrate phases,
allowing to analyze different phases (Scrivener 2004).

The microstructure of cementitious materi
als is formed by the hydration of
un
-
hydrated

cement grains
with the availability of water. In hydration process, cement grains react with water to produce hydrates.
These hydrates gradually increase, filling the spaces and forming a solid mass. Hydration
of Portland
cement is mainly dominated by the reactions of tri
-
calcium silicate (C
3
S) and di
-
calcium silicate (C
2
S),
producing calcium silicate hydrate (C
-
S
-
H) and calcium hydroxide (CH) (Mindess et al. 2002, Scriverner
2004). Calcium silicate hydrate is a

principle hydration product and is poorly crystalline material that
forms extremely small particles of less than 1

m in any dimension. In contrast, calcium hydroxide is a
crystalline material with fixed composition (Mindess et al. 2002).

RPCC Aggregate

Samples

Forecasting sentence

Sample Collection

RPCC aggregate
specimen
s of the subbase layers
were collected from

four sites of the interstate highways
I
-
35 and I
-
80 in Iowa. S
pecimen

selection was made based on high variety of construction year,
geographic location, and site observation. At each test site, five core holes, including four 100
-
mm (4
-
in.)
and one 250
mm

(10
in.) cores, were drilled

through the PCC pavement layer to the s
ubbase surface
. Test
holes were cored at the middle of the concrete slabs in the travel lane and patched after testing to prevent
future damages by traffic. In
-
situ field tests were conducted on the subbase and subgrade layers through
core holes. Samples w
ere collected after a series of field testing was conducted.


White
et al.

(tentative header)


5


Engineering Properties

The materials were classified from either poor or well
-
graded sand to gravel (US
C
S) or A
-
1
-
a (AASHTO)
(White et al 2008).
Properties of the RPCC specimens are presented in

Table 1. Figure 1 shows the grain
size distributions of the specimens
. The surface of the crushed concrete particles was normally covered
with a significant amount of fines, which ranged in size from a few microns to several hundred microns.
These fines,
which were mostly the result of the crushing of the old concrete, were loosely connected to
the bulk aggregate (Katz 2004, Winkler and Mueller 1998).

S
pecimen

Preparation
for SEM Analysis

For each site, two

coarse RPCC aggregate samples were chosen for th
e SEM analyses
.
Specimen

surface
w
as

brushed, air
-
pressure cleaned

to clean fine aggregates attaching to the s
pecimen
.
One s
pecimen

was
used for the SEM surface analysis, while the other s
pecimen

was polished for the SEM section analysis.
In preparation of

the sample using for SEM section analysis, e
poxy was used to fill voids and keep the
microstructure stable. Epoxy also created contrast between pores and other chemical composition. The
s
pecimen

was

cut to expose a fresh surface for backscattered electron (B
S
E) imaging.
This surface

w
as

then polished using a series of successively finer grades of diamond paste to show the material’s
microstructure (Stutzman 2001).

In this study, electron
scanning
mi
croscopy was
analyzed

using a Hitachi S 2460 reduced
-
vacuum
scanning electron microscope. Back
-
scattered images were taken and mineral substances of C, O, Na,
Mg, Al, Si, S, Cl, K, Ca, and Fe were analyzed using energy dispersive analytical x
-
ray (EDAX) ar
ea
mapping. Qualitative mineral identification was performed at high magnification of EDAX point
analyses.

Image Acquisition Criteria

A set of SEM imaging criteria, including the selection of areas to be analyzed of a given specimen,
magnification and the
number of images need to be acquired, was set at the beginning of the SEM
analysis. The criteria were
applied

for SEM scanning of the specimens.
Care was taken to collect images
that most represented the specimen conditions. The images should show the micr
ostructure of the cement
paste, cracks, air voids, and
un
-
hydrated

cement grains.

For each specimen, two images at 30x magnification were taken. One image
focused

at the center area of
the specimen and the other was taken at the edge area of the specimen.

The
se

low magnified
images
provided a preliminary analysis of the specimen, allowing the researchers to
understand

how the
individual features fit together in the total structure
, which

wa
s difficult or impossible to observe with
other modes of electron m
icroscopy. Higher magnified images were then obtained
for

the microstructure

White
et al.

(tentative header)


6


analyses

at specific areas from these two images.
SEM images at 100x
,

300x
, and 1000x

magnifications
were
used

for the analysis.

SEM Results of the RPCC Aggregates

The formati
on of the hydrated Portland cement paste microstructure depends on different factors, such as,
the chemistry and fineness of the cement used, the water/cement ratio, the use of chemical admixtures,
variations in mixing procedures,
early curing temperatures, and variations in hydration conditions.
However, most cement pastes show common features (Diamond 2004).
The microstructure of cement
paste changes with time
under

the effects of temperature changes, loading conditions, and aging
.

Under
repeating traffic loads, more cracks appear and develop in the PCC pavement and/or subbase aggregates
.
Un
-
hydrated cement

grains in the cement paste
are able to expose to pore water through cracks and
gradually
hydrate.
The hydration of un
-
hydrated

cement

grains
may result in high pH values of drainage
water, precipitate that blocks drainage pipes,
decreasing in volume of voids in the concrete and/or subbase
layer that reduces freeze/thaw resistance.


As mentioned by Diamond (2004), the use of bac
kscatter SEM for studying the microstructure of cement
paste and concrete, though pioneered by Scrivener and Pratt in early of 1980s (Scrivener and Pratt, 1984),
has been investigated by a few laboratories around the world.
In addition
, most of the investi
gations were
conducted on newly formed cement paste and concrete. In this study, the microstructure of old concrete
pastes from RPCC subbase aggregates is analyzed using SEM method. Four specimens 1, 2, 3, and 4
collected from RPCC subbase layers of inters
tates I
-
35 and I
-
80 were constructed in 1983, 1988, 1994,
and 2003, respectively.

Backscattered SEM images of the RPCC aggregate specimens
(
Figure
2
)

showed that residual cement
grains had existed in the recycled concrete used for the subbase layers. Crac
ks and voids were found
throughout of the specimens. These cement grains were able to expose to the surface or in cracks when
old concrete was crushed for RPCC materials. When the aggregates were used for subbase layers,
cementation would take place with t
hese cement grains when they contacted with meteoric or ground
water, increasing the stiffness of the subbase layers.

Residual Portland cement grains

U
n
-
hydrated

cement grains
we
re found in all or nearly all cement pastes. Most individual cement grains
contain
ed

fragments of several different kinds of crystals which
had been

adjacent to each other in the
clinker. These crystals include
d

impure C
3
S (alite), impure C
2
S (belite), impure C
3
A (aluminate), and
impure Ca
2
(Al, Fe) (ferrite solid solution, Fss) a
nd several minor components.
Clinkers
we
re normally
ground to sizes between 2

m to 80

m, with the mean diameter of about 10
-
12

m (Diamond 2004,
White
et al.

(tentative header)


7


Scrivener 2004).
Qualitative and quantitative features of Portland cement clinkers were well analyzed
using B
SE by Stutzman (2004).

In the hydration process, the cores of larger ground cement grains almost always remain un
-
hydrated for a
long time. These
un
-
hydrated

cement grains, which ha
ve

higher electron backscatter coefficients than the
hydrated products, ap
pear brightest in backscatter images. The size of
un
-
hydrated

cement grains
reportedly vari
es

from 5

m to 40

m (Diamond 2004, Scivener 2004).
Studies

conducted on concrete
specimens of several to 200
-
day old

suggested

that cement remnants were surrounded
by smooth
-
textured
uniformly gray hydration shells of varying thickness. These shells were separated from each other by a
groundmass of finer but less homogeneous hydra
ted components (Diamond 2004).

In this study, un
-
hydrated cement grains were observed in

three specimens, specimens 1, 2, and 3. Un
-
hydrated cement grains were not found in specimen number 4 (Figures 2 & 4). The irregularly textured
areas with bright color were calcium hydroxide that intermixed with C
-
S
-
H gel. The un
-
hydrated cement
grains ap
peared in bright white areas of different sizes, ranging from several

m to more than 100

m.
Unlike the hardened concrete specimens of within 200
-
day old where the hydration shells surrounding un
-
hydrated cement grains and other hydrated components were w
ell observed (Diamond 2004), hydrated
products of old cement pastes (specimens 1 to 4) seemed to intermix, resulting in a grey color. No
hydration shells were observed from BSE images of these specimens.

The SEM images of the specimen no.1
-
3 suggested th
at the specimen no. 2 had the highest amount of un
-
hydrated cement grains with the grain size was more than 100

m. Specimen no.1 with the un
-
hydrated
cement grain size of about 50

m had less amount of un
-
hydrated cement grain than that of the specimen
no.2. Specimen no.3 had the least un
-
hydrated cement grain compared to that of the specimen no. 1 and 2.
A little amount of un
-
hydrated cement grain was traced in this specimen.

C
-
S
-
H gel

C
-
S
-
H gel is produced by hydration of C
3
S and C
2
S in cement. The com
ponent contains calcium, silica,
and water. C
-
S
-
H is
the most important

hydration
component

that accounts for
half of the volume of a
hydrated paste. Two forms of C
-
S
-
H are indentified in the microstructure including
outer (
early
)

product
C
-
S
-
H and
“inner”

(
late
)

product C
-
S
-
H.

The outer product C
-
S
-
H is formed during early hydration

in the
originally water
-
filled space. This product, which
has a higher micro
-
porosity
,
is probably admixed with
monosulfoaluminate at the nanometer level

and contains a high le
vel of impurities.
Thus,

high unpurified
C
-
S
-
H is
also
called the “groundmass” or “undesignated product” (Mindess et al 2002).

Groundmass
areas were identified from all
four

specimens, although, with low proportions.

White
et al.

(tentative header)


8


The

term

“inner” product
refers to the dense C
-
S
-
H coating around
the hydrating
cement grains

that
develop within the boundaries of cement grains
. The coatings form diffusion barrier during later
hydration and
become thicker with time

by developing inwards and outwards
.
These hydration shells
have smooth texture and mostly uniformed gray color.
An alternative nomenclature
named as
“phenograins”
indicating un
-
hydrated
cement grains and surrounding hydration shells was
proposed
by
Diamond and Bonen (1995).

The term
is
used

in mineralogical nomenclature to
describe any significant
feature that is distinct from the groundmass regardless of their composition
.
However, this proposal has
not widely been accepted.
(Mindess et al 2002, Diamond 2004).


The BSE images of for specime
ns shown in Figure
2

were conducted on the center areas of the specimens
where hydrat
ed products
we
re denser and have less impurities. T
he
hydrated components in these areas
were mostly composed of
“inner


product C
-
S
-
H
, un
-
hydrated cement grains,
monosulf
oaluminate, and
ettringite.

From the BSE image of
the specimen 1

(Figure 2)
, the “inner” C
-
S
-
H
intermixed with other
hydration products in the “groundmass”, and the hydrated shells were not recognized.
Irregularly textured
areas surrounding cement grains s
uggest that they were originally formed from filled space. These areas
appear darker than the inner hydration shells
, cracks,

and

small pores.
The areas also host deposits of
calcium hydroxide (CH) and other minor hydration products, such as ettringite and monosulfate.
H
ydrated components with low level of impurities were found in BSE images of specimens 2, 3, and 4.
Single hydration

shells were
not recognized in these specimens, but instead, masses of “inner” C
-
S
-
H
surrounded the un
-
hydrated cement grains.

Calcium hydroxide

Calcium hydroxide

(CH)

is
an important product of cement hydration that accounts for about 20 to 25%
volume of cement paste
.
Calcium hydroxide (CH) accounts for between 20

25%
of the
volume of cement
paste.

CH

only
grows where
free space is available, but it
may stop growing or
it may
develop in another
direction if it encounters another calcium hydroxide
crystal
or a cement p
article. Many calcium hydroxide
crystals grow within the capillary pore space

(Mendess et al 2002,
Diamond
2004)
.
In SEM imaging
analysis, calcium hydroxide appears in gray color at a slightly brighter level than that of C
-
S
-
H gel. Since
CH normally stop g
rowing or growing around
hydrating

cement grains that it encounters
,
CH in cement
pastes usually appears as irregular masses of various sizes (Diamond 2004).

Calcium hydroxide of the old cement pastes was not easily observed from BSE images (Figure
2
). In this
figure, hydration products of specimen 1 had high level of impurities and no obvious CH crystals were
observed. The hydrated products of specimens 2 and 3 had low levels of impurities, appearing in a gray
White
et al.

(tentative header)


9


color at a similar level of brightness.
Calcium hydroxide appeared in slightly bright gray color and was
observed in irregular shapes (Figures
3

and
4
).

Aluminate containing phases

The hydration of C
3
A in the sulfate ions environment leads to the formation of hydration products
containing alumi
na.
At early ages, c
alcium sulfoaluminate hydrate (6
-
calcium aluminate trisulfate
-
32
-
hydrade), which is commonly called “ettringite”, is the product of tricalcium aluminate with sulfate ions
that are supplied by the dissolution of gypsum.
Ettringite in mas
ses in cement paste usually appear in
shrinkage pattern of curved cracks that resemble “tiger stripe” morphology.
At later ages, if the sulfate is
all consumed before the C
3
A has completely hydrated, then the ettringite reacts with the available C
3
A,
formi
ng monosulfate. Under some conditions, ettringite may re
-
crystallize in pores and voids into large
masses. Ettringite is found in old cement paste in needle
-
like shapes (
Diamond 2004,
Scrivener 2004).

Ettringite masses were

found in the RPCC aggregates sam
ples from all four sites

(Figures 2
-
5)
.
Individual
crystals of the ettringite masses from specimen no.1 seemed to intermix with C
-
S
-
H gel and were not
clearly identified. This finding is suitable with that of Scivener (2004) that ettringite persisted in ol
d
pastes.
In contrary
,
ettringite crystals of specimen nos. 2
-
4 were
very obvious on the SEM images
, though
they were old concrete materials
.
Ettringite crystals of old RPCC specimens (nos. 2 and 3) were mostly
damaged and became more porous, while
ettringite of the newer RPCC specimen (number 4) fully filled
the voids (Figures 3
-
5).


Pore space

and cracks

Visible pore spaces are normally observed in most cement pastes using SEM images. The actual content
depends mostly on
water/cement ratio and degr
ee of hydration (Diamond 2004). Visible voids observed
from the four specimens were either filled by epoxy
resin
or ettringite masses. The epoxy resin that filled
the pore space in prepared specimens causes dark areas on the SEM images.
Significant content
s of pore
space filled with ettringite masses were detected in specimen nos. 2, 3, and 4 in previous figures. These
pore space were interconnected with each other by cracks.

Pore space was not well detected in the
specimen no.1, although an amount of ettri
ngite mass was observed in the specimen using the SEM
images. As a result, not many cracks were observed in the specimen no.1.

Comparison

of
Stiffness of
Limestone and
RPCC
Aggregate
Subbase
s


Field investigations
were conducted
to

limestone and
RPCC aggr
egate subbase sites, which were
geographically distributed throughout Iowa
.

The investigations were aimed to study the performance of
the pavement layers

and
compare
the stiffness of the RPCC subbase layers

to those of the limestone
aggregate subbases
.
At
each test site, five core holes, including four 100
-
mm (4
-
in.) and one 250
-
mm (10
-
White
et al.

(tentative header)


10


in.) cores, were drilled for each test site. Test holes were cored at the middle of the concrete slabs in the
travel lane and patched after testing to prevent future damages
by traffic. In
-
situ field tests were
conducted on the subbase layers through core holes.

Subbase stiffness was characterized by field testing
using dynamic cone penetrometer (DCP), and Clegg Impact Hammer (ASTM D6951, D5874, Al
-
Amoudi
et al 2002, and Clyne

2001).
Field testing

values were used to
determine whether the cementation of the
residual cement grains in the RPCC aggregates
increased

the stiffness of the

RPCC

subbase layers.

Field testing results of four RPCC aggregate subbase layers, of which RPCC

aggregates were used
for

SEM analys
i
s
,
were used to compare with those of the
crushed
limestone aggregate subbases. Subbase
stiffness was characterized by
weighted
average
value of
dynamic penetration index (
DPI
avg
) and Clegg
impact value (
CIV
). The
DPI
av
g

of a subbase layer was obtained by averaging
DPI

values with depths for
the whole thickness of the layer.
Since
DPI
avg

values
we
re measured in mm per blow, lower
DPI
avg

value
mean
t

higher stiffness of the subbase layer.
CIV

values were also used to give indications of the stiffness
that higher
CIV

values reflected higher stiffness of the subbase layers
.
In case a subbase layer was very
stiff, results of the Clegg Impact hammer tests were outside of measurement range.

Althou
gh the Clegg impact hammer tests conducted on the subbase surface
s while the DCP presented the
subbase strength with depth, their results consistently reflected the subbase stiffness

(Figure 6)
. Stiffer
subbases resulted in higher
CIV

and lower
DPI
avg

valu
es
, and contrarily
.

According to the test results,
RPCC subbase layer from site no. 7 is the stiffest subbase layer compared with those of the other sites.
This layer

provided the lowest value of
DPI
avg
, while its
CIV

value was out of measurement range due

to
hard material surface of the layer.

V
ariability
in

the stiffness values of the RPCC subbases

seemed to be higher than
variability in
the
stiffness
values

of the crushed limestone subbases.

The variability of
DPI

values of the crushed limestone
subbases (44%) was lower than that of the RPCC subbases (56%)
. Although the variability of the
CIV

values of the crushed limestone subbases (24%)
was

equal to that of the RPCC subbases, it did not reflect
the fact that one
RPCC subbase layer (site no.7) was very stiff that the
CIV

value was out
side of the
measurement range.

Overall comparison of the stiffness of the crushed limestone and RPCC subbase layers suggested that
RPCC subbases were generally two times stiffer than
the crushed limestone subbases. The average
DPI

value of the RPCC subbases was about 50% (3.8), compared to that of the crushed limestone subbases
(7.5). The stiffness of the subbase layers presented in
CIV

values also presented in a similar manner when
th
e average
CIV

value of the RPCC subbases (58) was about 190% of that of the crushed limestone
subbases (31). In addition, by comparison of
DPI
avg

and
CIV

values (Figure 6), it was understood that the
White
et al.

(tentative header)


11


stiffness of individual RPCC subbase layer was higher th
an that of a crushed limestone subbase. This
could be explained that cementation of un
-
hydrated cement grains in the RPCC aggregates had partly
provided strength to the RPCC subbase layers.

Concluding Remarks

This paper aimed to give the reader some insigh
ts of the microstructures of old cement pastes as seen in
backscatter
-
mode SEM. The results of the SEM analysis were related to the stiffness of the subbase
layers, whose specimens were obtained for the SEM analysis
, to determine
whether

the existing un
-
hy
drated cement grains in the old concrete improve the stiffness of the subbase layers.
SEM analysis
showed that
most
RPCC aggregates contain
ed

some certain

amount
s

of un
-
hydrated cement grains
.
Analysis results showed that RPCC subbase layers were generally

stiffer than crushed limestone
aggregate subbase layers. Stiffness values in terms of the average values of
CIV

and
DPI

of the RPCC
subbases were about 180 to 200% higher than those of crushed limestone aggregate subbases. This
suggested that cementations of un
-
hydrated cement grains in the RPCC aggregate materials increased the
stiffness of RPCC subbase layers.



White
et al.

(tentative header)


12


Referenc
es

1.

White, D. J., Ceylan, H., Jahren, C. T., Phan, T. H., Kim, S., Gopalakrisnan, K., and M. T.
Suleiman.
Performance Evaluation of Concrete Pavement Granular Subbase


Pavement Surface
Condition Evaluation
. Final Report, Iowa DOT Project TR
-
554, CTRE Proje
ct 06
-
250, 2008.

2.

ASTM D6951
-
03, “Standard Test Method for Use of the Dynamic Cone Penetrometer in Shallow
Pavement Applications”.

3.

ASTM D5874
-
02, "Standard Test Method for Determination of the Impact Value (IV) of a Soil”.

4.

Al
-
Amoudi, S. B. O.; Asi, M. I.; W
ahhab, A
-
A. I. H.; and A. Z. Khan.
Clegg Hammer

California
-
Bearing Ratio Correlations.
Journal of Materials in Civil Engineering
, Nov.


Dec.,
2002, pp. 512

523.


5.

Clyne, T. R., Voller, V. R., and B. Birgisson.
Evaluation of a Field Permeameter

to Measure
Saturated Hydraulic Conductivity of Base/Subgrade Materials
. Final Report No. MN/RC


2001
-
19, Minnesota Department of Transportation, 2001.

6.

Clegg, B. (1986). Correlation with California Bearing Ratio.
News Letter 2
,
http://www.clegg.com.au/inf
ormation_list12.asp, Date Accessed: 04/30/2008.

7.

Iowa Department of Transportation.
General Supplementation of Specifications for Highway and
Bridge Construction
, GS
-
01016. Ames, Iowa, 2009.

8.

Balendran, R. V., Pang, W. H., and
X. H.
Wen (1998). Use of scanni
ng electron microscopy in
concrete studies. Structural Survey, Vol. 16 No. 3, pp 146


153.

9.

Scrivener, K. L. (2004). Backscattered electron imaging of cementitious microstructures:
understanding and quatification. Cement & Concrete Composites, Vol. 26, pp.

935


945.

10.

Brui
n
sma, J.E., Peterson, K.R., and
M.B.
Snyder

(
1997
). Chemincal Approach to Formation of
Calcite Precipitate from Recycled Concrete Aggregate Base Layers. Transportation Research
Record 1577.

11.

Mindess, S., Young
, J.F., and D. Darwin (
2002
)
.
Concrete (2nd

Edition). Prentice Hall, Upper
Saddle River, NJ 07458.

12.

Kjellsen, K.O., Detwiler, R.J. and Gjorv, O.E. (1990),

“Backscattered electron imaging of cement
pastes

hydrated at different temperatures”, Cement and

Concrete Research, Vol. 20 No. 2, U
SA,
pp. 308
-
11.

13.

Kjellsen, K.O., Detwiler, R.J. and Gjorv, O.E. (1991), “Development

of microstructures in plain
cement pastes

hydrated at different temperatures”, Cement and

Concrete Research, Vol. 21 No. 1,
USA, pp. 179
-
89.

14.

Stutzman P.

Scanning electron m
icroscopy imaging of hydraulic

cement.

Cem Concr Compos,
this issue [doi:10.1016/j.cemconcomp.

2004.02.043].

15.

Diamond S, Bonen D.

A re
-
evaluation of hardened cement

paste microstructure based on
backscatter SEM investigations.

In: Diamond S, Mindess S, Glas
ser FP, Roberts LW, Skalny

JP,
White
et al.

(tentative header)


13


Wakeley LD, editors.

Microstru cture of Cement
-
Based Systems/Bonding and Interfaces in
Cementitious Materials,

Materials Research Society Symposium Proceedings,

vol 370.

Pittsburgh
: Materials Research Society; 1995.

p.13

22.

16.

Diamond, S. (2004). The Microstructure of Cement Paste and Concrete


A Visual Primer.
Cement & Concrete Composites, Vol. 26, pp. 919


933.

17.

Stutzman, P. E. (2001). Scanning Electron Microscopy in Concrete Petrography. National
Institute of Standards and

Technology, 100 Bureau Drive, Stop 8621,
Gaithersburg, Maryland
20899
-
8621, pp 59


72.

18.

Packard, R. G. Thickness Design for Concrete Highway and Street Pavements. Paving
Transportation Department, Portland Cement Association, 1984.

19.

Katz, A. (2004) Treatments for the improvement of recycled aggregate. J Mater Sci Civ Eng
16(6):597

603. doi:10.1061/(ASCE)0899
-
1561(2004)16:6(597)





























White
et al.

(tentative header)


14


Appendices

Table 1.

Properties of Recycled Aggregates

Site

Specimen
No.

Construction
year
(a)

Classifications

OD specific
gravity
(b)

SSD specific
gravity
(c)

Absorption

(d)

(%)

No.

USCS

AASHTO

6

1

1983

SW
-
SM

A
-
1
-
a

2.18

2.36

8.2

7

2

1988

GP
-
GM

A
-
1
-
a

2.14

2.33

8.9

8

3

1994

GW
-
GM

A
-
1
-
a

2.20

2.36

7.2

9

4

2003

GP

A
-
1
-
a

2.21

2.35

6.4

(a)
: Construction years that RPCC aggregates were used for the subbase layers of new constructed pavements

(b)
: Oven
-
dry specific gravity

(c)
: Saturated
-
surface
-
dry specific gravity

(d)
: Dry to saturated
-
surface
-
dry

Table 2.
Summary of
CIV

Values and the Weighted Average Values of
DPI

Testing
site No.


Subbase
material

Interstate/
Highway

Direction

Mile
marker

Year of
construction

CIV


DPI
avg
(a)

(mm/blow)

1

C.L.
(b)

US 20

East

122.50

1990

30

9.0

2

C.L.

US 20

West

116.80

2005

22

12.0

3

C.L.

US 30

East

194.35

2005

38

4.7

4

C.L.

I 235

South

10.90

1968

27

7.6

5

C.L.

IA 92

East

132.16

1993

39

4.0

6

RPCC

I 35

North

131.40

1983

52

4.0

7

RPCC

I 80

East

65.10

1988

-

(c)

1.5

8

RPCC

I 80

East

165.20

1994

74

3.0

9

RPCC

I 35

North

140.75

2003

48

6.5

Notes
:

(a)

the weighted average value over the subbase thickness, which ranges from 20 to 28 cm (8 to 11 inches);

(b)
crushed limestone aggregate;

(c)

out of measurement range due to hard material surface.



White
et al.

(tentative header)


15



Figure 1.

Sieve analysis of four RPCC subbase aggregate materials


Figure 2. SEM images of RPCC aggregate specimen nos. 1, 2, 3, and 4 (x300 magnification)

Particle Size (mm)
0.01
0.1
1
10
100
Percent Finer (%)
0
20
40
60
80
100
Site 6
Site 7
Site 8
Site 9
Iowa DOT
Middle Gradation
1
2
3
4
White
et al.

(tentative header)


16



Figure 3. SEM image of specimen 2 (recycled in 1988) (x100 magnification)


Figure 4. Irregularly shaped c
alcium hydroxide in a cement paste
-

specimen number 4 (recycled in
2003)

cement
grain
ettringite
sand
(aggregate)
intermixed
C
-
S
-
H, CH
and
other products
void
space
CH
CH
White
et al.

(tentative header)


17



Figure 5. Ettringite masses deposited in air voids in a field of concrete. The EDX spectrum of
ettringite (specimen no. 3)



Figure 6. Comparison of
DPI
avg

and
CIV

values of crushed limestone and RPCC aggregate subbases

2
4
6
E n e r g y ( k e V )
0
5 0
1 0 0
1 5 0
c p s
C
O
F e
N a
M g
M g
A l
S i
S
C l
K
C a
C a
F e
Average Dynamic Penetration Index Values
Site Number
1
2
3
4
6
7
8
9
0
5
DPI
avg
0
2
4
6
8
10
12
14
16
Limestone subbases
RPCC subbases
Statistic Summary for:
Limestone subbases
Average DPI= 7.5
Standard deviation = 3.3
C.O.V = 44%
RPCC subbases
Average DPI = 3.8
Standard deviation = 2.1
C.O.V = 56%
LWD Modulus of Elasticity of Subbase Layers
Site Number
1
2
3
4
6
7
8
9
0
5
Clegg Impact Value,
CIV

0
20
40
60
80
100
Limestone subbases
RPCC subbases
Statistic summary:
Limestone subbases
Average CIV = 31
Standard deviation = 7
C.O.V = 24%
RPCC subbases
Average CIV = 58
Standard deviation = 14
C.O.V = 24%