NISTIR5742

Alkali-Silica Reaction and High

Performance Concrete

Chiara F. Ferraris

August 1995
Building and Fire Research Laboratory
National Institute of Standards and Technology
Gaithersburg, MD 20899

National Institute of Standards and Technology

Ronald H. Brown, Secretary
Technology Administration
Maq L. Good, Under Secretary for Technology
National Institute of Standards and Technology
Arati Prabhakar, Director
ABSTRACT

Damage due to alkali-silica reaction (ASR) in concrete is a phenomenon that was

first recognized in the U. S. since 1940 and has since been observed in many countries.
Despite numerous studies published, the mechanism is not yet clearly understood.
Nevertheless, the three major factors in concrete have been identified, i.e., the alkalies
contained in the pore solution, reactive amorphous or poorly crystallized silica present in
certain aggregates, and water.
In this study, we attempted to address the question: is high-performance concrete
(HPC) susceptible to ASR? Researchers have not reached an agreement on this matter
because factors other than the three major ones (pore solution alkalinity, aggregate
morphology and water presence) play a significant role in the occurrence of ASR, these
factors include aggregate gradation, w/c and compressive strength.
It was found that air content is the most important variable (other than the three
majors factors cited above) that increase expansion of concretes tiected by ASR. This
study indicates that even HPC should be susceptible to ASR if reactive aggregates are
used.

Keywords

Building Technology; Alkali-Silica reaction (ASR); High Performance Concrete (HPC);

Standard Testing.

...
111
 —

TABLE of CONTENTS

1. ~TRODUCTION ...................................................................................................... 1

2. BAcKGRom .........................................................................................................2
2.1. ASRMechanisms ..............................................................................................2
2.2. H.igh Performance concrete ~C) ................................................................... 11
3. AVAILABLE TESTS ............................................................................................... 13

4. E=EM~fi PROGW ................................................................................ 14

4.1. Design of High Performance Mortar (HPM) .................................................... 14
4.2. Test Coti~ration ........................................................................................... 15
5. RESULTS AND DISCUSSION ................................................................................ 16
5.1. Influence ofmixture design on expansion at 21 days ........................................ 16
5.2. Inikence of exposure conditions ...................................................................... 18
6. CONCLUSIONS ....................................................................................................... 18

7. AC~OWEDG~NTS ......................................................................................... 19

8. REFERENCES ......................................................................................................... 19

iv
 LIST of FIGURES

Figure 1: Schematic representation of the attack of alkaline solutions on silica ................ 3

Figure 2: Pore solution composition versus time expressed from cement pastes at ages up
to 1 day.[10] ................................................................................................... 3
Figure 3: Schema of the progression and consequence of the swelling of the ASR gel [3]. 4
Figure4: Cracks due to ASRin concrete [3] (Hoover Dam, USA) .................................4
Figure 5: Sketches displaying texture, morphology and compositions of the primary alkali
susceptible rocks and characteristic patterns when tiected by ASR [] .............5
Figure 6: Pessimum behavior; w/c= 0.4 and aggregate/cement= 2.75 . ............................ 6
Figure 7: Expansion versus aggregate ske[16] . ................................................................6
Figure 8: Expansion versus aggregate size; w/c=O.41 and aggregate/cement= 3 [3]... ......7
Figure 9: The influence of w/c ratio and aggregate size on mortar expansion fi-om
Baronio [16] suggests a relationship between w/c and expansion. .................... 8
Figure 10: The influence of water/cement ratio on the expansion of concrete due to ASR
shows greater expansion with increasing w/c. Aggregate/cement = 4.55 [] ...... 8
Figure 11: Influence of air content in hardened concrete on the relationship of expansion
and age: w/c= O.41; ale= 3; sodium oxide= 5 kg/m3[ 3] . ...............................9
Figure 12: Alkali concentration (expressed as equivalent NazO) drop in pore solution of
pastes as fi.mction of 1.5 year expansion of concretes A) rhyolitic tufi, B)
siliceous limestone (Spratt) [] ........................................................................ 10
F@ure 13: Schematic of theinteraction of ASRand HPC. ............................................. 12
Figure 14: Relation between selected parameters and expansion at 21 days. All expansions
presented are the average of at least 3 bars. The compressive strength is the
average of3 cubes . ........................................................................................ 17
Figure 15: Expansion of HPM bars under different exposure conditions .........................18

LIST OF TABLES

Table 1: Comparison of the alkali-silica reactivity tests .................................................. 13

Table 2: Mortar sand gradation ...................................................................................... 15
Table 3: Suma~of samples tested .............................................................................. 16

v
1. Introduction
Damage due to alkali-silica reaction (ASR) in concrete is a phenomenon that was
first recognized in 1940 by Stanton [1,2] in North America and has since been observed in
many other countries. Many studies [3,4,5] were published since Stanton’s first paper, but
the mechanisms of ASR are not yet clearly understood [6]. Nevertheless, the major factors
have been identified, e.g., alkalies in the pore solution and reactive silica, present in
certain aggregates and the presence of water. Other factors can play a significant role,
such as environmental relative humidity (RH), porosity of the concrete and presence of
mineral admixtures.
High petiormance concrete (HPC) is usually defined as a concrete with high
strength and durability. As reported by Mindess [7] from SHRP-C-205 report [8], HPC
is defined by three factors:
1. Strength characteristics:
. 4-hour compressive strength 217.5 MPa (2500 psi) or very early strength (VES)
. 24-hour compressive strength235 MPa (5,000 psi) or high early strength (HES)
● 28-day compressive strength270 MPa (10,000 psi) or very high strength (VHS)
2. durability factor greater than 80% after 300 cycles of freezing and thawing
3. water-cementitious materials ratio <0.35.

Few studies have been reported on the relationship between the material design of
high perl?orrnance concrete (HPC) and ASR. It has been reported [9] that the low porosity
and the presence of mineral admixtures in most HPC, will protect HPC from deterioration
due to the ASR by the lack of water and by the beneficial influence of mineral admixtures.
Utiortunately, we lack long term field experience on the behavior of HPC because of the
young age of most HPC structures, the limited number of structures and the long time
taken for the effect to manifest itself Furthermore, we lack knowledge on the correct
method to test the durability of HPC containing aggregates possibly susceptible to ASR.
In this report, after a brief literature review, we will examine existing tests for
applicability to HPC. The results of some tests conducted at the National Institute of
Standards and Technology (NIST) on high petiorrnance mortars (HPM) will be discussed.
Recommendations for fi,n-ther studies and possible test methods are given.

1
2. Background

2.1. AS? Mechanisms

Most researchers agree that the main reaction of ASR is the reaction between
certain forms of silica present in the aggregates and the hydroxide ions ( OH) in the pore
water of a concrete [3,4]. Very early in the hydration of cement calcium ions are
incorporated in the hydration products but potassium and sodium stay in solution and
eventually they are partially incorporated into calcium silicate hydrate (C-S-H) and
monosulfate (AFJ. Hydroxide ions from the hydration of portland cement result in a pore
solution having a pH of at least 12.5 [10]. Soluble alkalies raise the pH to about 13 or
higher. Also, the amount of alkalis present in the pore water is related to the amount of
soluble alkalis present in the cement. The hydroxide ions will attack a silica surface. If the
silica is well crystallized the vulnerable sites are only at the exterior surface of the
aggregate (Figure 1a), but in the case of poorly crystallized silica, there are many
vulnerable sites in the aggregate structure, leading to disintegration of the silicate network
[11, 12] (Figure lb) . To keep a neutral charge balance, the cations Na+ and K+ difise
toward the hydroxide ions to react with them and the resulting product is a gel-like
material. According to Powers and Steinour [13] the migration of cations of Na+ and K+ is
slow, therefore the migration of Ca2+takes place. If the gel is high in calciu~ then the gel
is not expansive when exposed to water and, therefore, may not induce cracking in
concrete. This theory rests on the assumption that calcium could be available. Diamond
found that there is very little calcium in the pore solution (Figure 2). This is expected since
the high pH causes the volubility of Ca(OH)2 to be depressed. Nevertheless, calcium could
be dissolved from the solid phase of cement paste to produce a gel. Most researchers do
not mention the distinction between “safe” and “swelling” gel but there are
acknowledgments that there are more than one composition of gel produced by ASR [12].

2
 Jo ~lo

%
.
Si

Figure 1: Schematic representation of the attack of alkaline solutions on silica

a) Well crystallized silica b) Poorly crystallized silica. [ 11]

:’,

0.6 ~1 , 1 , I I 1 , 1 # 1 t I 1 I I I 1
I I I I [

~ 0.5 -
K+
A
: 9-”
s
z
~.4

~‘ 0.3
E’
s
!.- 0.2 -
z
8
“5 0.1

0
I 1 r I t , I
‘a 4 8 12 16 20 24

TIME, HOLRS

Figure 2: Pore solution composititm” VerSUStime expressed from cement pastes at ages up
to 1 day.[10]

3
 The formation of the gel per se is not deleterious. The deterioration of the concrete
structure is due to the water absorption by the gel and its expansion. R is reported [11,14]
that the RH must be higher than 80% for the gel to swell although it can be formed at
lower relative humidities. According to Hobbs [3] the progression of the swelling of ASR
gel follows the general patterns shown in Figure 3. As the tensile strength of the system is
exceeded, cracks will form and propagate. As there is not a preferential dkection for
cracks to propagate and also the sites of crack initiation are randomly distributed in the
specimen, map cracking will be characteristic of ASR deterioration (Figure 4). The sites of
the cracks are determined by the location of the reacting silica on the aggregates and the
availability of OH in the vicinity.

.,
,.:::.
..\ “’ “;‘:”
..: -...
+,;

stage 1
“-””T
.. .,,.,,, ,, ....... .. . .
‘... .<
,.:‘:
Slage2 Stage 3 .,
Gelfilledm!CrOCrack
< Gel / C, Gel filledm!crocracksurrounded
%
&# byghlsaturated,mawe Stage 4

.,
...... . ,- ,. .
<-, -.’
.,.
‘,,
,-:..,.
.... .:.; ..
,.,
,. .-.,
.,, / .;; . “:,:.,’% F
.,
... ,.’ ?<>“
.\!.- .,
... .

Figure 3: Schema of the progression and corkequence of the swelling of the ASR gel [3].

Figure 4: Cracks due to ASR in concrete [3] (Hoover Dam, USA)

4
 Type of reactive aggregates include [3]:
. non-crystalline or impetiectly c~stalline silica, such as crystobalite, opaline, chert, flint
● strained quartz or microcrystalline quartz (often found in gneiss) shist, quartzite
. glassy to cryptocrystalline matrix of igneous rocks, such as pumice.
Figure 5 shows the various aggregates and their reaction patterns regarding ASR. A
factor that is reported of importance to the deleterious activity of the aggregates is the
gradation or size distribution.

OPAL IWTIXJTE DENSE FLINT Scnls” GRANITE ANO

GNE!SS WITH
OPAIJNE CHERT PORJWYRES CHERT WYIJJTE SEVERE TECTONIC
DISTORTION
VOLCANIC GLASS DENSE Ml- FOK4TED GNEISS
CRYSTALLINE
INDUSTRIAL GIASS SILJCECW ROCKS

ENTW PAJ7TICLE INTERNALCRACUNG CONFINED INTERNAL @ACX!NG FOLLOWS CRACKING

BUILDUP RESERVOIR WHERECIASTIC mROIJs, OPAJJNE WEAK INTERGR4NuMR ACCUMULATES
FOR SWELLING GEL GRAJNS PROVIDE REGIONS SEATS OF PATCHES WITH FROM REGIONS
BEFORE BURSTING. STIFFNESS ANO SWELLING GEL SWTZLING GEL - WITH THOROUGHLY
RkACTION RIM mNGTH OURING FORMATION, FOFWATION DISTORTED OUARTZ
OFTEN PRESENT GELFORMATION. WSING
SOMERIM RE4CTI(JN FFW3URING OF
MAY BE VISIBLE SURROUNDING ROCK
SUBSTANCE

Figure 5: Sketches displaying texture, morphology and compositions of the primary

alkali susceptible rocks and characteristic patterns when affected by ASR [14]
 The mix design should be considered, because the aggregate content and size, the
water/cement ratio and air content affect the reaction. For instance, Hobbs [3, 15] reports
that there is a pessimum amount of aggregate that will result in a higher expansion (Figure
6). Also, Hobbs [3] and 13aronio’s [16] results show that the greatest expansion occurs
with the finest gradation used (I?@ure 7 and Figure 8). Nevertheless, at very small sizes,
in the order of microns, a pozzolanic effect could be expected to reduce the amount of
expansion (as discussed later in this report).

OL!$JU—J
0 2 4 6 8 10 12

Opaline silica content%

Figure 6: Pessimum behavior; WIG= 0.4 and aggregatelcement = 2.75,

alkali content = 6 kg/m3 [3].

1.4

1.2

0.2 I

0.0~
O-o.lmm 0.1-1mm l-3mm

Sand sizes

Figure 7: Expansion versus aggregate size[16].

6
 25-
Concrete: moist cured. 20 ‘C:
opaline silica;
uncrushed coarse aggregate:
Na20 - 5.0kg/m3

2.0-

1.5-

8
c
.-0
fn
%
$
lil l.o–
A’ 0.3- 1.2

0 - 13

05-

m m m
A +
. 1 T
o n t a
o 50 100 150 200
Age: day

Figure 8: Expansion versus aggregate size; w/c=O.41 and aggregate/cement= 3 [3]

As mentioned earlier, other factors influencing the expansion due to the swelling of
the reaction gel include the water/cement ratio (w/c) and the air content in fi-esh concrete.
The influence of the water/cement ratio on expansion is not clear-cut; higher w/c ratio
leads to either higher or lower expansion depending on other factors such as age and
aggregate size. Figure 9 and Figure 10 show that the higher w/c ratio generally results in
a higher expansion, but in one case (gradation B: w/c =0.6 lower than w/c= 0.5), the
reverse was found.

7
 ~grad. A; wtc= 0.4

~L
-
2.0- -
-grad. A; W/C= 0.5
7 -grad. A; W/C= 0.6
& 1.5--
-grad. B; WfC = 0.4
.-~
2 1.0- +grad. B; WC = 0.5
e
~
-A--grad. B; W/C= 0.6
w 0.5-

0.0
0 100 200 300

Time [days]

Figure 9: The influence of w/c ratio and aggregate size on mortar expansion from
Baronio [16] suggests a relationship between w/c and expansion.
Grad. A= O- O.lmm; Grad. B= O.l-lm.
The original graph included another gradation C with size 1-3 mm that
showed no influence of the w/c. -

;Iwzoiik
.50
=xpansion
Percent
.40

.30

.20

.10

.00
o 5 10 15 20

Time, Month

Figure 10: The influence of water/cement ratio on the expansion of concrete due to ASR
shows greater expansion with increasing w/c. Aggregate/cement = 4.55 [17]

8
 Concrete: moist cured, 20 “C;
opaline silica particle size 150–300~fm ●
20-

1.5 -

s
c
0
“g 1.0 -
a
~
Lu Symbol Air content: %

T
~
● 0.5 -
A 08
05- 1.3
// “0 : 28
0 36
/ —.. .—. -.—

00
K #--=” 5:
I
100
~-.
150
..—–.._ -1
200
Age: day

Figure 11: Influence of air content in hardened concrete on the relationship of expansion
and age: w/c = 0.41; a/c= 3; sodium oxide= 5 kg/m3 [ 3].

The air content influence is illustrated in Figure 11. It indicates that the higher air
content results in lower expansion. This could be explained by the gel expanding into the
available capillary pores or voids, without stressing the structure of the cement matrix.

The addition of pozzolanic fillers, such as fly ash [18], silica fime or other mineral
admixtures has been shown to reduce the expansion of concrete affected by ASR [18].
The mechanisms are not well understood, but it is agreed that the silica contained in the
mineral admixtures reacts faster [4, 19] with the 0?+ than the silica contained in the
aggregates. Therefore, the alkalis are rapidly consumed and the level of O?+_ reduced to a
level at which aggregates react very slowly or not at all The beneficial response of some
samples containing mineral admixtures is demonstrated in Figure 12. However, not all
mineral admixtures are beneficial, e.g., PFA-C or CSF-B in Figure 12. This inconsistency
is not clearly understood at the moment, but it has been suggested that the mineral
admixtures that do not counteract ASR are not removing alkalis from the solution. Also,
some mineral admixtures contain deleterious amounts of alkalies. ASTM has a limit on
alkali content of a mineral admixture (such as fly ash) if it is to be used with ASR-
susceptible aggregates.

9
 6
6:B 0.04%
m
5: +-+C

4:
CSF-B
I
control
x 3-
I PFA-A 10% : PFA-A
2:
‘CSF-A

PFA-B SI?RATT LIMESTONE

e_
0 I’”’’ l”’’’ t,, $, l,” o
-0.05 0% 0.35 ‘-0.05 0.05 0.15 0.25 0.35
EXP”&SION :tlL&GA A-23.2-14A) EXPANSION (%) (CAN/CSA A23.2-14A)

Figure 12: Alkali concentration (expressed as equivalent NazO) drop in pore solution of
pastes as fbnction of 1.5 year expansion of concretes A) rhyolitic tu~, B) siliceous
limestone (Spratt) [20]

From this brief overview of the literature on the mechanisms, the following
summarizes present knowledge on ASR:
e ASR is a reaction between the OH in the pore solution with amorphous or poorly
crystallized silica in the aggregates.
e The reaction product imbibes water and expands.
e The presence of water or RH higher than 80’% is necessa~ for the gel formed to
expand and induce concrete cracking.
e Some siliceous mineral admixtures deplete the alkalis horn the pore solution, lowering
the pH, therefore decreasing the likelihood of ASR.
e The aggregate type and size distribution play a significant role in the expansion
measured in concretes
o Other factors influencing the cracking due to ASR include air entrainment and possibly
Wlc.

10
 2.2. High Performance concrete (HPC)
High petiormance concrete (HPC) is usually define as a concrete with high strength
and durability. As reported by Mindess [21 ] ftom SHRP-C-205 report [22], HPC is
defined by three factors:
. Strength characteristics:
– 4-hour compressive strength 217.5 MYa (2500 psi) or very early strength
(vEs)
– 24-hour compressive strength z 35 MPa (5000 psi) or high early strength
(HEs)
– 28-day compressive strength270 MPa (10,000 psi) or very high strength
(VHs)
● durability factor greater than 80% after 300 cycles of freezing and thawing
● water-cementitious materials ratio <0.35.

As stated in the introduction we lack knowledge of the behavior of HPC in regard

to ASR. There are no long-term petiorrnance data regarding ASR in HPC. Con.tlicting
viewpoints exist as to whether ASR gel will form in HPC and if it does form, whether the
higher tensile strength will be sufficient to prevent the concrete from cracking.
According to some researchers [9,23], the use of silica fbme will prevent the
reaction from occurring as noted in Section 2.1. Shah [9] states: “for the high-strength
concrete the effect of self-desiccation may reduce the moisture content to a level where no
alkali-aggregate reaction can take place”. Criaud et al.[23] state that the lower porosity
should prevent the diffision of 0?3 to the reaction sites. They imply that neither
hydroxide ions nor water can migrate to the reaction sites. However, it can be argued that,
at low w/c and low air content, the decreased porosities should result in a lack of space
(capillary pores and cracks) for the reaction gel to expand into. Also, it should be realized
that water is present at the beginning for the hydration of cement and some reaction could
occur before all the water is consumed.

The significant parameters are summarized in Figure 13. The shaded areas
correspond to parameters affected by the HPC design and processing.

11
 Alkali in Cement Reactive Silica
\ in Aggregates

.............
==:
............ Tensile Strength Low

H 1]

Figure 13: Schematic of the interaction of ASR and HPC.

According to our flow chart of the response of HPC to ASR (Figure 13), two areas
(marked 1 and 2 in Figure 13) need to be investigated to determine if HPC is susceptible
to ASR. The two issues are:
e “? 1“: Permeability:
If the permeability of HPC is such that water is not reaching the gel, then no
swelling and no cracks will occur.
e “?2”: Tensile strength:
If the tensile strength of HPC is higher than the stress generated by the
swelling of the gel, then no cracks will occur.

We have planned an experimental study to address both issues. In this report, a

response to only the first issue (“? 1“(Permeability)) is presented. Tests are underway for
responding to the second issue (“?2” (Tensile strength)) but. sufficient data are not
available yet.

12
3. Available tests
The first goal of this report is to determine which one of the available tests is
applicable to detect ASR in HYC. All tests could be classified in two categories:
. Tests on aggregates alone
. Tests on mortar or concrete samples containing aggregates
Only the second category of tests is potentially applicable to HPC because they address
the combination of cement and aggregates used. Therefore, in this report we will examine
closely the tests in the second category.

There are three tests that fall in this category: ASTM C1260 [24] , ASTM C227
[25] and a French rnicrobar method [26]. Other tests found in the literature are
modifications of the above tests. Table 1 gives the major characteristics and differences of
these three tests.

Table 1: Comparison of the alkali-silica reactivity tests

Sample design mortar bars mortar or concrete bars mortar bars

WJC= 0.50 WICnot specified Wlc=0.3
aggregates:fixedgradation aggregates: fixed gradation sand: fixed gradation
sand/cement= 2.25 sand/cement = 2.25 SICsuggested
Measurement In a IN NaOHsolutionat at 100?4oRH and 38°C at 150”C in 10!XOKOH
conditions 80”C solution
Criteria of Deleterious if Deleterious if Deleterious if
expansion expansion > 0.20/0at 14 ● expansion >0.170 at 12 expansion >0.1 10/0at
days months 3 days
● expansion > 0.05°/0at 6
months

A major difference between these tests is the duration of the measurements, i.e.,
from 3 days to 12 months. Of course, the rapid tests are more appealing but the following
variables could not be tested as specified, because they are specified by the test method:
. The alkali content of the cement: the alkalis supplied by the solution should be more
than enough to cause AS% therefore the alkali content of cement is irrelevant
● The mixture design of the sample.
● The aggregate grading.

In these tests, the samples cannot be classified as high performance materials

because the specified mixture designs are typical of conventional-based materials. This is
because the tests were designed to determine whether an aggregate is suitable for use in
concrete, i.e., whether it is subjected to ASR. They were not intended to test mixture
designs.

13
 Another test ofpossible relevance is ASTM C441 [27], usedto determine the
effectiveness of mineral admixtures in preventing ASR is. However, it is not suitable
because the sand gradation is fixed, the aggregate is a highly-reactive Pyrex glass, and the
w/c ratio is determined fi-om the flow measured by a flow table.

Therefore, the only test that could be easily modified to make it applicable to KPC is
ASTM C227. One modification is required, i.e., the fixed sand gradation needs
replacement by one that will increase flow (at low w/c) and minimize permeability. The
ASTM C 1260 test might be used to test some parameters such as porosity or mixture
design if mineral admixtures were not used, but the sand gradation would have to be
modified to reflect a HPM mix.

4. Experimental program

4.1. Design of High Performance Mortar (HPM)

We defined a high petiormance mortar, as a mortar with some of the characteristics

of HPC, i.e., low w/c, low porosity, high strength. In this report we did not consider the
use of mineral admixtures. Therefore, to design a high performance mortar, we imposed
the following conditions:
. low w/c, about 0.30
e low sand/cement ratio (s/c) of 1.4
0 low air content (< 2-3 %)

To design our HPM, we proportioned four graded sands from the same
manufacturer and with the same composition. The gradation was selected to obtain the
highest flow and the lowest air content. Carino et al. [28] had already found a s/c ratio
that gave a high compressive strength. We modified the mixture design slightly by using a
sand gradation that followed the Fuller curve [29]. This gradation should give the highest
compaction and the best flow. The w/c ratio was adjusted to obtain a flowable mixture as
measured by a flow table (ASTM C 230 [30]).
The sand was a fised silica (Tecosil*) and was available in a variety of particle sizes.
This selection gave us flexibility to determine the best gradation without the tedious work
of grinding the coarse aggregate to a specified size distribution as prescribed by ASTM
C 1260. Table 2 gives the sand gradation used.

* Tecosil is fhsed silica from C-E Minerals, 901 East Eighth avenue, King of Prussia, PA
19406. The name of the manufacturer is identified in this report to adequately describe the
experimental procedure. Such an identification does not imply recommendation or
endorsement by the National Institute of Standards and Technology, nor does it imply that
the material identified is necessarily the best available for the purpose.

14
 Table 2: Mortar sand gradation
Sand
Mortar Size range Mass
Designation Sieve size Dimension
ASTM El 1 [’m] [%]
HPM 50-100 300-150 25
20-50 850-300 25
10-20 2000-850 35
4-1o 4750-2000 15
ASTM C1260 50-100 300-150 5
20-50 850-300 30
10-20 2000-850 30
4-1o 4750-2000 35

The mortar w/c ratio varied between 0.29 to 0.47. S/C was either 1.411 [ 28] or
2.25 as described in ASTM C1260. High range water reducing agent (HRWR) was used
for all lYPM mixes.

4.2. Test Configuration

To measure expansion due to ASK bars (25 mm x 25 mm x 279 mm (1 in. xl in.

x 11 in.)) were prepa~ed using the above-mentioned mixture design. The cement used had
a high alkali content (about 1.2 0/0Na20 equivalent). As the high alkali content of the
cement should result in high alkali concentration in the pore solution, this cement should
promote ASR with reactive siliceous aggregates.
To test the influence of parameters such as w/c, s/c, air content, sand gradation and
duration of curing, the conditions of testing were as described in ASTM C1260, i.e., 80”C
in a solution of lN NaOH. These test conditions were intended to promote a reaction
faster than in normal field situation. Therefore, to test if HPC will be subject to AS~ we
exposed some bars to a less severe environment, i.e., 100 0/0RH and 20”C. This test has
the advantage of more closely simulating field conditions but it could lead to a relatively
slow reaction rate, increasing the time to obtain data. Most data were obtained using the
first set of conditions because we wanted a rapid test to determine the influence of some
of the mixture design parameters. Only one set was tested using the second conditions.

15
5. Results and discussion

5.1. Influence of mixture design on expansion at 24 days

Table 3 gives the mixture designs and other experimental variables that were
selected for this set of tests.

Table 3: Summary of samples tested

Mortar composition
Isample # (c Slc Sand gradation Curing Time air content
) (b) (c) [days] [%]
7Ei?z!- 0.295 1.411 HPM 1,14 5.9
~ASR2-25 0.30 1.411 HPM 1 13.7
ASR-142 0.32 1.411 HPM i 10.7
ASR-164 0.32 1.411 HPM 1,7,14,28 7.6
ASR2-20 0.40 1.411 HPM 1 12.4
‘ASR2-30 0.50 1.411 HPM 1 11.0
- 0.47 2.25 ASTM C1260
ASR2-3 0.47 2.25 ASTM C1260 1,14 I 8.9
ASR-166 I 0.47 2.25 ASTM C1260 I 1,13 I 7.3
Notes
a) wlc = water/cement ratio
b) s/c= sand/cement ratio
c) The sand used was Tecasii: see Table 2 for aradation

The relationships between the w/c, s/c, curing time, compressive strength, air
content and the expansion at 21 days are presented in Figure 14. From the data in Figure
14, it can be concluded that the main variable was air content, followed by curing time. No
larger expansion occurred with longer curing time. Expansion did not seem to be affected
by the following parameters: w/c, s/c and sand gradation. But it should be kept in mind
that air content is affected by those parameters (w/c, s/c and sand gradation). Therefore, it
is difficult to distinguish between the various variables tested. The compressive strength,
contrary to what was expected, does not appear to affect the expansion at 21 days
significantly. A trend could be seen toward higher expansion with longer curing. This
might be explained by the fact that expansion at 21 days means 21 days in I’JaOH, but the
specimens are older. They were exposed to water during the curing period. Therefore, the
reaction could have started immediately after casting and the exposure to NaOH and
higher temperature would only have accelerated it.

16
Figure 14: Relation between selected parameters and expansion at21 days. All expansions
presented are the average of at least 3 bars. The compressive strength is the average of 3
cubes.

17
 5.2. Influence of exposure conditions
As stated earlier, two types of exposure were used, ASTM C 1260 and room
temperature over water. Figure 15 shows expansion data for the first 40 days of
exposure. We obsewed that the expansion for the sample exposed to 100°/0 RH is about
half that for the samples in NaOH, as was expected. What is more surprising is that the
lowest w/c ratio gave the highest expansions. This could imply that the HFM is susceptible
to AS% even under the milder exposure conditions. It should be remembered, of course,
that reactive aggregates and high alkali cement were used. From these data, it might be
expected that HPC would be susceptible to ASR if alkali-susceptible aggregates were
used.

Expansion at 20 C over water ~w IC-4.30 Expansion at 80 C in NaOH ~ w IC=O.3$1

Influence of wic + w IC=O.40 Influence of W/C + w IC=O.40

+ w IC=O.50 -A-- w IG=O.50

1.0 , a

0.0 ~
o 10 20 30 40 50 o 10 20 30 40 50
Tim e [days] Time [days]

Figure 15: Expansion of HPM bars under different exposure conditions

6. Conclusions

The following conclusions can be drawn fi-om this report:

e The existing rapid test ASTM C 1260 cannot be applied for HIT without modification,
because the mixture design is pre-deterrnined. T~s-test was designed to detect
deleterious aggregates and the expansion with different mixture design, i.e., w/c, s/c,
sand gradation.
e HPM is likely to be susceptible to ASR expansion if deleterious aggregates are used.
e There seems to be a dependence on air content such that maximum expansion was
observed at 6-7°/0 air.
e There is a trend suggesting that a longer curing increases expansion.
e The other variables tested, -- WIC,s/c, sand gradation and compressive strength -- did
not seem to control the expansion due to ASR significantly. But, it should be
remembered that air content is modified by changes in w/c, s/c and sand gradation.

Further studies should include a wider range of variables to significantly modi$ the
permeability of the mortar, and to include mixture designs that are currently used for
concretes. The design of a test specifically adapted to HYC should be the first priority.
Other characteristics of HPC, such as tensile strength, should be investigated.

18
7. Acknowledgments

This work was sponsored by the “High Petiorrnance Construction Materials”

program at NIST. I would like to thank Dr. James Clifton for the invaluable help obtained
through discussions and comments on this work. John Winpigler is also acknowledged for
peflorrning the measurements. Support for this work was also obtained from the Nuclear
Regulatory Commission (NRC).

8. References

1. Stanton T. E., “Expansion of Concrete Through Reaction Between Cement and

Aggregate” Proc. of the American Sot. of Civil Eng. Vol. 66, #10, 1940, p. 1781-1811.
2. FrohnsdorffG., Cliflon J., Brown P., ‘TI.istov and Status of Standards Relating to
Alkalies in Hydraulic Cements” Special Technical Publication # 663, ASTM 1979, p 16-
34
3. Hobbs D. W., Alkali-Silica Reaction in Concrete, Thomas Telford, London 1988
4. Diamond S., Penko M., “Alkali Silica Reaction Processes: The Conversion of Cement
Alkalis to Alkali Hydroxide” ,G. M. Idorn Inter. Symposium, Durability of Concrete ACI
SP-131, 1992
5. Helmuth R., “Alkali-Silica Reactivity: An Overview of Research” SHRP Report C-342,
1993.
6. Capra Bruno, Bournazel Jean-Pierre, “Perspective nouvelles pour la prise en compte
des alcali-reactions clans Ie calcul des structures” Materials and Structures, vol. 28, p. 71-
73, 1995
7. Mindess S., “Materials selection, proportioning and quality control”, High Performance
Concrete: Properties and Applications, Ed. by Shah S.P. , Ahmad S.H. McGraw-Hill,
1994, p. 1-26
8. SHRP-C/FR-91 -103, Hkh Ped?ormance Concretes, a state of the art reuort, Strategic
Highway Research Program, National Research Council, Washington DC., 1991
9. Shah S., Ahmad S., Hkh Performance Concrete: Properties and Applications,
McGraw-Hill, 1994
10. Diamond S., “Alkali Reactions in Concrete Pore Solutions Effects”, Proc. 6th Int.
Conf, Alkalis in Concrete, Idom G.M. and Rostam S. eds. 1983, p. 155-166
11. Figg J., “An Attempt to Provide an Explanation for Engineers of the Expansive
Reaction between Alkalis and Siliceous Aggregates in Concrete”, 6th Int. Conf Alkalis in
Concrete Copenhagen 1983.
12. Helmuth R., Stark D. “Alkali-Silica Reactivity Mechanisms”, in Materials Science of
Concrete III, J. Skalny cd., ACS 1992
13. Stark D., “Alkali-Silica Reactions in Concrete”, in Simificance of Tests and
Properties of Concrete and Concrete Makirw Materials, Klieger P. and Lamond J. eds.,
ASTM STP 169C, 1994
14. Idom G. M., Johansen V., Thaulow N., “Assessment of Causes of Cracking in
Concrete”, Materials Science in Concrete III, Amer. Ceramic Sot., New York, 1992

19
 15. Leslie Struble, The Influence of Cement Pore Solution on Alkali-silica Reaction
NISTIR 87-3632, 1987
16. Baronio, Barra, Montanaro, Delmastro, Bachiornni, “Couplage d’action de certains
parametres physiques sur le developpement de la reaction akxdis-granulate”, in From
Material Science to Construction Materials EnEineenng, 1st Int. Conf Versailles Vol. 3,
1987
17. Stark D., “Effects of Water-Cement Ratio on Expansion Due to ASR”, Concrete
Technology Today 1995, p.3-5
18. Schumann David C., Carrasquillo Ramon L. and Farbiarz Jose~ State-of-the-art
re~ort on the Mechanism of Alkali-Aggre~ate Reaction in Concrete containing Fly Ash,
Research report 450-2, Center for Transportation Research 1988.
19. Carrasquillo R.L. Farbiaz J., “Alkali-Aggregate Reaction in Concrete Containing Fly
Ash: Final Report” Center of Transportation Research, Research Report #450-3F, 1988
20. Duchesne J. Bembe M. A., “The Effectiveness of Supplementary Cementing Materials
in Suppressing Expansion due to ASR: Another Look at the Reaction Mechanisms Part 2:
Pore Solution Chemistry” Cem. and Concr. res. Vol. 24 No 2, 1994, p 221-230
21. Mindess S., “Materials selection, proportioning and quality control”, ~
Performance Concrete: Properties and Applications, Ed. by Shah S.P. , Ahmad S.H.
McGraw-Hill, 1994, p. 1-26
22. SHRP-C/FR-91 -103, High Pefiormance Concretes, a state of the art re~ort, Strategic
Highway Research Program, National Research Council, Washington DC., 1991
23. Criaud A., Cadoret G., “HPCs and Alkali Silica Reactions. The double role of
pozzolanic materials” in High Perllormance Concrete: fi-om Materials to Structure ,Y.
Malier cd., 1992
24. Standard Test Method for Potential Alkali Reactivity of Aggregates (Mortar-Bar
Method), ASTM designation C 1260-94, 1994 Annual Book of ASTM Standards Vol.
04.02
25. Standard Test Method for Potential Alkali Reactivity of Cement-Aggregate
Combinations (Mortar-Bar Method), ASTM designation C227-90, 1994 Annual Book of
ASTM Standards Vol. 04.02
26. Criaud A., Vernet C. Defosse C. “The Microbar Method, an Accelerated Expansion
Test for Evaluating Aggregates, Assessment of Canadian Aggregates”, Workshop
Canadian developments in Testing Alkali-A~gregate reactivity Toronto 1990.
27. Standard Test Method for Effectiveness of Mineral Admixtures of Ground Blast-
Furnace Slag in Preventing Excessive Expansion of Concrete Due to the Alkali-Silica
Reaction, ASTM designation C 441-89,1994 Annual Book of ASTM Standards Vol.
04.02
28. Carino N. J., Knab L., Clifton J., A~plicability of the Maturity Method to HiRh-
Ped?ormance Concrete, NIST-IR #4819, 1992
29. Fuller W.B. “The laws of Proportioning Concrete”, Amer. Sot. of Civil Eng.
Transactions Vol. 59, 1907, p. 67-143
30. Standard Specification for Flow Table for Use in Tests of Hydraulic Cement, ASTM
designation C 230-90,1994 Annual Book of ASTM Standards Vol. 04.01

20