You are on page 1of 10

Journal of Testing and Evaluation

doi:10.1520/JTE20140334

Vol. 44

No. 3

May 2016

available online at www.astm.org

Kai-Wei Liu1 and Anol K. Mukhopadhyay2

Accelerated Concrete-Cylinder Test for


AlkaliSilica Reaction
Reference
Liu, K.-W. and Mukhopadhyay, A. K., Accelerated Concrete-Cylinder Test for AlkaliSilica Reaction,
Journal of Testing and Evaluation, Vol. 44, No. 3, 2016, pp. 12291238, doi:10.1520/JTE20140334. ISSN
0090-3973

ABSTRACT
Manuscript received August 29, 2014;
accepted for publication October 27,
2014; published online January 10, 2015.
1

The concrete-prism test (CPT, e.g., ASTM C1293-08b [ASTM C1293-08b: Standard Test
Method for Determination of Length Change of Concrete Due to Alkali-Silica Reaction,
Annual Book of ASTM Standards, ASTM International, West Conshohocken, PA, 2013]) has

Zachry Dept. of Civil Engineering, Texas


A&M Univ., 199 Spence St. 501H CE/TTI,
College Station, TX 77843, United States
of America (Corresponding author),
e-mail: kaiwei@e-mail.tamu.edu

been considered as the best index for alkalisilica reaction (ASR) eld performance, but

Ph.D., Texas A&M Transportation


Institute, Texas A&M Univ., 199 Spence
St. 501B CE/TTI, College Station, TX
77843, United States of America,
e-mail: anol@e-mail.tamu.edu

there is a signicant reduction in expansion associated with more alkali leaching in the ACPT

alkali leaching and test duration are still of concern. Reduction of test duration by increasing
testing temperature (e.g., 60 C) was proposed by several researchers to develop an
accelerated version of the concrete-prism test (ACPT). However, it has been found that
compared to the concrete-prism test (CPT). An accelerated concrete-cylinder test (ACCT)
has been proposed in this study with no involvement of errors because of operation and
temperature change along with arresting alkali leaching. The linear expansion of the ACCT
measured at a temperature of 60 C over a period of 1 month have been correlated with the
1-yr ASTM C1293 test to evaluate the ASR reactivity of aggregates. Cylinders made with
highly reactive borosilicate glass balls were tested rst to validate the proposed approach
before any concrete-cylinder testing. The results indicate that (1) an expansion limit of
0.04 % using 0.82 % Na2O equivalent (Na2Oe) cement without alkali boosting after a testing
period of 28 days is proposed for the ACCT to diagnose ASR aggregate reactivity, and (2)
the proposed method can serve as an alternative to validate the ASR-resistant mix design
(e.g., y-ash contents).
Keywords
alkalisilica reaction, accelerated concrete test, validation of ASR mix

Copyright V 2015 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
Copyright
by ASTM Int'l (all rights reserved); Sat Sep 3 21:34:16 EDT 2016
Downloaded/printed by
Pontificia Universidad Catolica de Chile (Pontificia Universidad Catolica de Chile) pursuant to License Agreement. No further reproductions authorized.
C

1229

1230

Journal of Testing and Evaluation

Introduction

to reduce testing period as well as alkali levels similar to job


mix), and (5) proposition of a concrete validation testing.

Since the alkalisilica reaction (ASR)-related problems were


identied in the early 1940s, extensive research has been carried
out on ASR over past decades. The development of test methods
to assess the ASR potential of concrete through an expansion
limit approach is one of the main focus areas of ASR research.
The main purpose of an ASR test method is to measure aggregate reactivity before their use in concrete structures. The
concrete-prism test (CPT) has been considered as the best index
for eld performance, but alkali leaching and test duration
impose major drawbacks [1,2]. Therefore, Ranc and Debray [3]
rst introduced the accelerated concrete-prism test (ACPT) in
the early 1990s. The concrete prisms were stored over water at
60 C, instead of 38 C. The results show a good correlation
between the 38 C and 60 C tests after the 56-day testing period.
Other researchers [4,5] also show a reasonably good correlation
between 1-yr concrete-prism expansion at 38 C and 2 to 4
months prism expansion at 60 C.
Although the test duration is shortened by simply increasing the test temperature, a signicant reduction in expansion
associated with high-alkali leaching was noticed in the ACPT
compared to the CPT [68]. When alkali leaches out of the
specimens, the sulfate ions replace the leached alkali hydroxides
and decrease the pH of pore solution. This eventually causes the
reduction of expansion.
An accelerated concrete-cylinder test (ACCT) has been
proposed in this study to overcome some of the above limitations (e.g., alkali leaching, test duration) and come up with an
alternative ASR concrete test method. The unique steps that are
taken for the proposed test to be considered as a rapid and reliable concrete ASR test method are (1) introduction of an automatic linear variable differential transducer (LVDT)-based
length change measurement system with no involvement of
errors as a result of operation and temperature change, (2)
measurements to avoid alkali leaching at relatively hightemperature (60 C), (3) testing cylinders made with borosilicate
glass balls and measuring its volume change (through the
change of solution level) and length change as a proof of concept, (4) tests at varying levels of alkali (alkali-boosted concrete

Materials and Methods


MATERIALS

Concrete cylinders made with pure phase material (i.e., borosilicate glass balls) and aggregates were tested in this study. Aggregates with different types of reactive silica characterized by
petrographic observation (i.e., ASTM C295 [9]) and varying
ranges of reactivity (e.g., ASTM C1260 [10], ASTM C1293 [11])
were selected. Table 1 summarizes the reactivity data that ASTM
C1260, ASTM C1293, ASTM C295, and compound activation
energy (CAE) [12] for these aggregates had determined. The
four levels of alkali (i.e., 1.8, 2.4/2.7, 4.0, and 5.3 kg/m3) were
selected. A low-alkali (CM1 0.57 % Na2O equivalent [Na2Oe])
and a high-alkali (CM2 0.82 % Na2Oe) Portland cement were
chosen to reach the desired alkali levels over the range from 1.8
kg/m3 (3.0 lb/yd3, 0.57 % Na2Oe without alkali boosting) to
5.3 kg/m3 (8.9 lb/yd3, 1.25 % Na2Oe with alkali boosting)
with varying levels of cement factor (CF).
TEST EQUIPMENT

A chemical approach that simulates the aggregate-pore solution


reaction that exists in concrete has been proposed earlier [1,12].
The authors have proved that solution volume change in a
closed system is in a form of chemical shrinkage as the ASR
between aggregates and alkali solution proceed. Use of the same
device (volumetric change measuring device, VCMD, in Fig. 1)
to measure solution volume change (Fig. 1(a)) and length change
(Fig. 1(b)) of cylinders made with borosilicate glass were conducted to have a comparative assessment between the previous
method [1,12] and the proposed cylinder method in this study.
A 7.6 cm  15.2 cm (3 in.  6 in.) concrete-cylinder specimen with cast-in place threaded rod (Fig. 1(b)) is placed inside
the container. The 7-day curing specimen is then immersed
with soak solution of specic alkalinity [equal to pore solution
alkalinity (PSA) of the specimen]. While the rod moves inside
the LVDT during ASR expansion of the specimen, the

TABLE 1 List of aggregate with relevant material data.


Aggregate

ASTM C1260 (14-Day Expansion, %)

ASTM C1293 (1-yr Expansion, %)

ASTM C295 (Reactive Constitute)

CAE [12] (KJ/mole)

Borosilicate glass
CA1


0.012


0.027

Amorphous silica
Few siliceous (e.g., chert) inclusions

5.53 (HR)
61.70 (NR)

CA2

0.417

0.078

Acid volcanic, chert

29.73 (HR)

FA1
FA2

0.317
0.381

0.058
0.391

Low-strained QTZ, chalcedony, chert


Acid volcanic, Chert

32.64 (R)
26.96 (HR)

FA3

0.079

0.035

Few siliceous (e.g., Chert) inclusions

60.36 (NR)

FA4

0.242

0.043

High-strained QTZ, Chert

36.39 (R)

Note: FA, Fine aggregate; CA, coarse aggregate; QTZ, quartz; NR, nonreactive; HR, highly reactive; R, reactive.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 3 21:34:16 EDT 2016
Downloaded/printed by
Pontificia Universidad Catolica de Chile (Pontificia Universidad Catolica de Chile) pursuant to License Agreement. No further reproductions authorized.

LIU AND MUKHOPADHYAY ON ACCELERATED CONCRETE

FIG. 1
VCMD test setup for (a) solution volume
change, and (b) length change.

method [14]. The extraction method consists of pressing a


cement paste cylinder with a loading of 181.4 kg (400 lb) to
extract the liquid contained in the specimen. The loading was
applied and released twice for all specimens to get a sufcient
quantity (i.e., 25 ml) [8] of the pore uid. The extracted pore
solution for Na and K ion concentration was analyzed by an
atomic absorption spectrometer. Table 3 presents the composition (Na and K) of pore solution extracted from the studied
cement pastes. A minimum of three cement paste specimens for
each mix in Table 2 was squeezed to extract pore solution followed by mixing the extracted solutions to get a representative
pore solution. The Na equivalent (Na
e ) [15] represents the total
alkali levels for each mix in this study. Note that the same alkali
level using two different types of cement with varying Na2Oe
percentages and different amounts of extra alkali addition does
not ensure the same PSA. For example, the PSA with 5.3 kg/m3
using CM1 is 0.88 normality (N) but the PSA with 5.3 kg/m3
using CM2 is 1.04 N. Therefore, cement composition (especially

displacements (inch) are recorded by the attached computer


automatically.
MIX DESIGN AND SPECIMEN PREPARATION

All of the ACCT mixes are ASTM C1293 type mix with or without alkali boosting. Table 2 presents a detailed description of the
mix designs. Concretes were mixed by hand following ASTM
C192 [13] procedures. Concrete cylinders using each mix (Table
2) were cast for ACCT.
PORE SOLUTION EXTRACTION

The cement paste cylinders (5.1 cm  10.2 cm) corresponding


to each mix in Table 2 were cast and covered with plastic foil,
and then stored under 100 % relative humidity (RH) at 25 C
6 2 C for 7 days. After the 7-day curing, the specimens were
de-molded and pore solutions were extracted from each paste
specimen. The pore solution extraction from cement paste
specimens was conducted by using a high-pressure squeezing

TABLE 2 Concrete mix design for conducting ACCT.


Alkali (kg/m3)

CA

FA

w/c

CAF

1a

2.4

CA1

FA4

0.45

0.76

1b
1c

4.0
5.3

Mix

Cement Type

CF (kg/m3)

Additional Alkali (NaOH, kg/m3)

Na2Oe (%)

CM1

420

0.57

CM1
CM1

420
420

1.6
2.9

0.95
1.25

2a

2.4

CM1

420

0.57

2b

4.0

CM1

420

1.6

0.95

2c

5.3

CM1

420

2.9

1.25

3a
3b

1.8
2.7

CM1
CM2

312
325




0.57
0.82

3c

5.3

CM2

420

1.8

1.25

4a

1.8

CM1

312

0.57

4b
4c

2.7
5.3

CM2
CM2

325
420


1.8

0.82
1.25

CA2

FA3

CA1

FA1

CA1

FA2

2.7

CA2

FA2

CM2

325

0.82

2.7

CA1

FA3

CM2

325

0.82

2.7

CA2

FA3

CM2

325

0.82

Note: w/c, Water-to-cement ratio; CAF, coarse aggregate factor.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 3 21:34:16 EDT 2016
Downloaded/printed by
Pontificia Universidad Catolica de Chile (Pontificia Universidad Catolica de Chile) pursuant to License Agreement. No further reproductions authorized.

1231

1232

Journal of Testing and Evaluation

TABLE 3 Concentration of the extracted pore solution from cement paste.


Mix (Table 2)

Na (ppm)

K (ppm)

Na (N)

K (N)

Na
e (N)

0.57

1a, 2a

1800

19000

0.08

0.49

0.37

0.95

1b, 2b

4898

16100

0.21

0.41

0.46

420
312

1.25
0.57

1c, 2c
3a, 4a

14132
1539

17300
21031

0.61
0.07

0.44
0.54

0.88
0.38

2.7 (CM2)

325

0.82

3b, 4b, 5, 6, 7

4153

31562

0.18

0.81

0.66

5.3 (CM2)

420

1.25

3c, 4c

12755

31865

0.55

0.81

1.04

Alkali (kg/m3)

CF (kg/m3)

Na2Oe (%)

2.4 (CM1)

420

4.0 (CM1)

420

5.3 (CM1)
1.8 (CM1)

the type of alkali-bearing phases in cement) plays an important


role in controlling the PSA and the same alkali level using different cements do not.
Pore solution chemistry data in the published literature
[1619] are very similar to the alkali concentrations in Table 3
that the pore-solution extraction method has determined in this
study. Therefore, these pore solution concentrations were used
to generate the soak solutions for the concrete-cylinder tests
that correspond to each mix in Table 2. The quantities of NaOH
and KOH pellets needed to generate an articial solution of the
same composition for each mix in Table 3 were rst calculated
and then dissolved in deionized water to prepare the soak solutions. Calcium hydroxide crystals were then added (1 g per liter
solution) to the above respective solutions slightly above saturation to prepare an alkaline solution saturated with calcium hydroxide. Adding calcium hydroxide crystals slightly above the
saturation point ensures the presence of undissolved calcium
hydroxide crystals, which represents a situation similar to concrete pore solution.

TEST PROCEDURE

The ACCT procedure is briey described below:

A 27.9-cm (11-inch) stainless steel threaded rod was embedded (2.54 cm depth) on top of each concrete cylinder
(7.6 cm  15.2 cm) during specimen casting. After casting, the molds were covered with plastic foil and kept
inside a 100 % RH chamber at 25 C 6 2 C for 7 days.
After 7 days, the concrete cylinders were de-molded and
placed inside the VCMDs, which were lled up by soak
solution with chemistry that is equal to the pore solution
chemistry of each mix (each mix has a specic level of
alkalis according to Table 2). The purpose of creating
soak solution chemistry equal to pore solution chemistry
is to prevent alkali leaching from the specimen.
Each VCMD was tightly closed and placed inside an oven
at 60 C.
Expansion measurements were recorded every 15 min
automatically through a data-acquisition computer system over time.

Initially, the concrete specimen expands because of temperature increase from the starting temperature to the target temperature (60 C). The subsequent LVDT readings after

temperature stabilization (the reference/initial LVDT reading)


represent displacement because of ASR. The displacement
because of ASR over time divided by the original length at the
reference point multiplied by 100 represents the % expansion of
the concrete cylinder because of ASR over time.

Results and Discussion


MORTAR CYLINDER TEST WITH BOROSILICATE GLASS
BALLS

Mortar cylinders with highly reactive borosilicate glass balls


(SiO2: 81 %, Na2O: 4 %, Al2O3: 2 %, B2O3: 13 %) were tested
rst to verify its applicability to measure ASR expansion before
any concrete testing. The glass-mortar cylinders were cast, and
measurements of net solution volume change (Fig. 1(a)) and the
length change (Fig. 1(b)) were conducted to have a comparative
assessment between the previous method and the proposed cylinder method in this study. The mortar mix used alkali level 2.7
kg/m3 (4.5 lb/yd3 equivalent to 0.82 % Na2Oe) with 40 % glass.
The curing conditions (i.e., 7-day fog-room curing before demolding) and testing conditions (i.e., 60 C, immersing in PSA)
were the same for both cylinder tests. For the cylinder of solution volume change measurement, the device is rst lled up
with PSA overnight at room temperature to allow maximum
saturation of voids in the PSA. The next day, the device is
placed on a vibrating table and conducts vacuuming under
vibration for 2 h to mainly remove entrapped air bubbles in the
solution and also help to saturate the unlled voids in the cylinder after overnight saturation. The device is then placed inside
an oven and heated to 60 C to eliminate the effect of crack porosity [1] as a result of temperature change. When the solution
reaches 60 C, the device is removed from the oven and applied
a second stage vacuuming under vibration of 45 min to facilitate
further removal of air bubbles (may be generated during preheating at 60 C) from solution. The device is placed back inside
the oven whose temperature is then raised to 60 C. The oat
movements caused by solution volume changes as initial thermal expansion and the chemical reaction between cylinder and
alkaline solution progress are then recorded through a dataacquisition system. Fig. 2 shows (1) the solution volume change
and length change of glass-mortar cylinders at 60 C over time
(Fig. 2(a)), (2) the macro-crack pattern of the cylinder (Fig.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 3 21:34:16 EDT 2016
Downloaded/printed by
Pontificia Universidad Catolica de Chile (Pontificia Universidad Catolica de Chile) pursuant to License Agreement. No further reproductions authorized.

LIU AND MUKHOPADHYAY ON ACCELERATED CONCRETE

FIG. 2 (a) Solution volume change and length change, (b) macro-crack
pattern, and (c) thin section of borosilicate glass-mortar cylinders.

proposed cylinder test is capable of measuring ASR expansion


in a short period of time.
ACCELERATED CONCRETE-CYLINDER TEST (ACCT)

2(b)),

and (3) nature of micro-crack under a microscope (Fig.

2(c)).

The measurement of the length change of the glass-mortar


cylinder shows expansion, but the measurement of the solution
volume change shows reduction (Fig. 2(a)), which might be
because of the consumption of reactants (e.g., water, ionic species) and the formation of micro-/macro-cracks [1]. These
cracks create fresh surfaces and act as open passages for alkali
solution to ow into the matrix of the cylinder leading to more
solution level reductions. Therefore, the reduction of solution
volume in Fig. 2(a) might be caused by new open passages
caused by new crack formations, and measurement of the
change of solution level cannot guarantee measurement of the
expansion of the specimen as a result of ASR.
The following conrm the presence of a high degree of
ASR: (1) presence of macro-cracks in the cylinder (Fig. 2(b)),
(2) presence of micro-cracks in a reacted glass ball (Fig. 2(c)),
(3) micro-cracks passing through both the reacted glass ball and
cement paste (Fig. 2(c)), and (4) presence of ASR gel at the periphery of the reacted glass ball (Fig. 2(c)). The ASR features
(both macro- and micro-scale) supports the high linear expansion measurement (Fig. 2(a)) in the cylinder test. Therefore, the

The selected mixes in Table 2 were tested using ACCT at 60 C


and varying levels of alkalinity. For each test corresponding to
each mix (Table 2), the soak solution chemistry was equal to
pore solution chemistry (Table 3).
Mix 4 was used to cast two cylinders for each alkali level
(i.e., 1.8, 2.7, and 5.3 kg/m3) to verify the variability (within the
lab) of ACCT test results. The expansion corresponding to two
replicas were used to check the variations, and the expansion
results are presented in Fig. 3. The majority of expansion-based
coefcient of variation is within 10 % after the 28-day expansion for the tested mixes at all alkali levels, which indicates that
the variations of measured expansion of ACCT are low. Fig. 4
shows the expansion curves of all mixes at each alkali level. For
each mix, the higher the alkali level, the higher the level of
expansion.
Proposed expansion limits of ACPT at 60 C were found
from a range of 0.02 % to 0.08 % at 8 weeks [3,4,20] and 0.03 %
to 0.04 % at 13 weeks [5,20]. In this study, same exposure
block/CPT expansion limit of 0.04 % [2] is chosen and shown
in Fig. 4. Table 4 summarizes the aggregate reactivity based on
the ACCT expansions in Fig. 4 and compares with ASTM
C1293 1-yr expansion data. For the mixes 1 to 4 with highalkali levels (2.7, 4.0, 5.3 kg/m3 with Na2Oe > 0.6 %), the expansion limit of 0.04 % can be reached within 49 days (Figs. 4(b) to
3
4(d)). The alkali level 5.3 kg/m (1.25 % Na2Oe) in Fig. 4(d) is
the same as a conventional concrete prism (i.e., ASTM C1293).
If the same expansion limit of 0.04 % and the same alkali level
1.25 % Na2Oe in ASTM C1293 are applied in this study, the
ACCTs have identied these aggregates as reactive aggregates
within 4 weeks (mixes 1c, 2c, 3c, and 4c in Table 4), which
matches with the ASR reactivity diagnostic of ASTM C1293
1-yr values. This suggests that ACCT in VCMD can be used as
an alternative approach of concrete test for ASR.
The outdoor exposure blocks are generally cast with a highalkali cement (0.9 6 0.1 Na2Oe) with a CF of 420 kg/m3 [21].
Fournier et al. [21] found that the exposure blocks using highalkali cement without alkali boosting (nominally 0.95 % Na2Oe)
correlate best with boosted (1.25 % Na2Oe) concrete prisms.
Based on the results in Table 4, it seems that ACCT with an
alkali level of 2.7 kg/m3 (4.5 lb/yd3 , 0.82 % Na2Oe without
alkali boosting) and with a CF of 325 kg/m3 might be a good
choice to pass/fail a concrete mix with the same exposure block/
CPT expansion limits (i.e., 0.04 %). A concrete mix with a conventional CF (e.g., 325390 kg/m3) will be sufcient to achieve
2.7 kg/m3 alkali levels if the Na2Oe of the cement is relatively
high (e.g., 0.6 % < Na2Oe  0.82 %). However, if the Na2Oe of
the cement is low (e.g., <0.6 %), a high CF may be needed to
achieve 2.7 kg/m3 alkali levels.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 3 21:34:16 EDT 2016
Downloaded/printed by
Pontificia Universidad Catolica de Chile (Pontificia Universidad Catolica de Chile) pursuant to License Agreement. No further reproductions authorized.

1233

1234

Journal of Testing and Evaluation

FIG. 3
Expansion curve of mix 4 with different alkali
levels.

Based on the above discussions [i.e., the variations of


measured ACCT expansion are low after the 28-day expansion
(Fig. 3(a)), the alkali level 5.3 kg/m3 (1.25 % Na2Oe) is same as
conventional CPT, the alkali level 2.7 kg/m3 (0.82 % Na2Oe
without alkali boosting) might be a good choice to pass/fail a
concrete mix in this study], Fig. 5 only shows plots of 1-yr
expansions in ASTM C1293 versus the 28-, 35-, 42-, and 49-day

ACCT expansions at alkali levels of 2.7 (0.82 % Na2Oe) and 5.3


kg/m3 (1.25 % Na2Oe). The linear regression line is shown how
expansions at 1 year of ASTM C1293 correlate well with ACCT
expansion at each testing time. A better correlation factor (R2)
is found at 28 days than other days for both alkali levels. At 28
days expansion, the correlation is also better at alkali level 2.7
kg/m3, compared to 5.3 kg/m3 with respect R2 to 0.99 and

FIG. 4
Expansion curve of ACCT at each alkali level:
(a) 1.8 kg/m3, (b) 2.4/2.7 kg/m3, (c) 4.0 kg/
m3, and (d) 5.3 kg/m3.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 3 21:34:16 EDT 2016
Downloaded/printed by
Pontificia Universidad Catolica de Chile (Pontificia Universidad Catolica de Chile) pursuant to License Agreement. No further reproductions authorized.

LIU AND MUKHOPADHYAY ON ACCELERATED CONCRETE

TABLE 4 Aggregate reactivity based on the ACCT expansion.


Alkali (kg/m3)

Na2Oe (%)

PSA (N)

Time to Reach 0.04 % Expansion (day)

ASTM C1293 1-yr Expansion (%)

CAE Classification [12] (Table 1)

1a

2.4

0.57

0.37

NIR until 49

0.043

Reactive

1b

4.0

0.95

0.46

32

1c

5.3

1.25

0.88

18

2a
2b

2.4
4.0

0.57
0.95

0.37
0.46

46
36

0.078

Highly reactive

2c

5.3

1.25

0.88

20

3a

1.8

0.57

0.38

NIR until 49

0.058

Reactive

3b
3c

2.7
5.3

0.82
1.25

0.66
1.04

27
12

4a

1.8

0.57

0.38

19

0.391

Highly reactive

4b

2.7

0.82

0.66

10

4c

5.3

1.25

1.04

2.7

0.82

0.66

Highly reactive

2.7

0.82

0.66

NIR until 49

0.027/0.035

Nonreactive

2.7

0.82

0.66

20

0.078

Highly reactive

Mix

NIR, Not identied as reactive.

0.87. Therefore, ACCT with cement without alkali boosting


(2.7 kg/m3, 0.82 % Na2Oe) is a good candidate to pass/fail a
concrete mix in this study.
The measured ACCT expansion with alkali levels of 2.7 and
5.3 kg/m3 at 28 days along with ASTM C1293, 1-yr expansions
are graphically presented in Fig. 6. The line of equity [8] is
shown where expansions at alkali level 2.7 kg/m3 match better
than alkali level 5.3 kg/m3 to 1-yr ASTM C1293 expansions.
Based on the above discussion, it is recommend to use cement
without alkali boosting (2.7 kg/m3, 0.82 % Na2Oe) to evaluate
the ASR reactivity in ACCT.
Fig. 7 shows the expansion curve of mixes 3 to 7 using alkali
level 2.7 kg/m3 (0.82 % Na2Oe). The 28-day ACCT expansions
(%) are 0.047, 0.269, 0.006, and 0.064 for mixes 3, 4, 6, and 7,
respectively, which match well with the reactivity diagnostic of
1-yr ASTM C1293 (%) (i.e., 0.058, 0.391, 0.027/0.035, 0.078)
and CAE classication in Table 1 (i.e., the lower the ASR CAE,
the higher the reactivity). This indicates that a straight cement
concrete mix with alkali level 2.7 kg/m3 (0.82 % Na2Oe) is sufcient to delineate ASR reactivity at 28 days. Therefore, ACCT
with relatively low alkali levels (as opposed to high-alkali levels,
i.e., 4.05.3 kg/m3 in the current CPT test/exposure block) can
be effective to identify the ASR reactivity in a relatively short period of time. It may be close to testing a job mix if the job mix is
a straight cement mix.

FIG. 5 Comparison of expansion at 1 yr in ASTM C1293 test versus expansion at 28,


35, 42, and 49 days in the ACCT test with alkali levels of 2.7 and 5.3 kg/m3.

VERIFICATION OF ALKALI LEACH PROOF

The changes of soak solution chemistry were monitored to verify the leach-proof situation as well as the possibility of ion
migration from soak solution to the specimen. An increase of
OH, Na, and K ion concentrations in soak solution represents leaching of these ions from the specimen. On the other
hand, a reduction in concentration of these ions in soak

Copyright by ASTM Int'l (all rights reserved); Sat Sep 3 21:34:16 EDT 2016
Downloaded/printed by
Pontificia Universidad Catolica de Chile (Pontificia Universidad Catolica de Chile) pursuant to License Agreement. No further reproductions authorized.

1235

1236

Journal of Testing and Evaluation

FIG. 6 Comparison of expansion at 1 yr in ASTM C1293 versus expansion at


28 days in the ACCT with alkali levels of 2.7 and 5.3 kg/m3.

solution indicates ion migration from the soak solution to the


specimen while ASR is in progress in the specimen. The results
are shown in Fig. 8 for the change of OH, Na, and K concentrations of soak solution of mixes 3 and 4 with alkaline levels
1.8, 2.7, and 5.3 kg/m3 caused by ASR after a testing period of
49 days.
Initially, the ionic concentrations in pore solution and soak
solution are equal, which does not allow ion migration between
pore solution and soak solution. As ASR progresses, Na, K,
and OH concentrations are reduced in the pore solution of the
specimen, which triggers ion migration from the soak solution
to the pore solution. The decrease in Na, K, and OH concentrations in soak solution after the test (Fig. 8) suggests ion
migration from soak solution to the specimen. However, the
degree of reduction of the ions in soak solution is not that high.

It seems that the effect of leach-proof situations on the measured expansion is more pronounced than the effect caused by
ion migration from the soak solution into the specimens. However, a continuous increase of expansion instead of gradually
approaching an asymptotic shape may be supported by ingress
of ions from soak solution. Note that this effect is very pronounced in the case of ASTM C1260, where the soak solution is
1 N NaOH, which is way higher than the mortar bar pore solution and might lead to a false positive. The use of a soak solution that is equal to pore solution may cause a little accelerating
effect on the test result itself. However, it can be used as a
method of simulating a eld situation where an external source
of alkalis (e.g., sea water ingress, deicing chemicals) and alkali
redistribution (i.e., concentration of alkalis in certain zones
within the main concrete body because of intense ambient
temperature and RH variations) are signicant factors.

FIG. 8 The change of Na, K, and OH of soak solution of (a) reactive mix
3, and (b) highly reactive mix 4 with alkali levels 1.8, 2.7, and 5.3 kg/
m3 after testing period of 49 days.

FIG. 7 Expansion curves of ACCT (mixes 3 to 7 with alkali level 2.7 kg/m3)
over time.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 3 21:34:16 EDT 2016
Downloaded/printed by
Pontificia Universidad Catolica de Chile (Pontificia Universidad Catolica de Chile) pursuant to License Agreement. No further reproductions authorized.

LIU AND MUKHOPADHYAY ON ACCELERATED CONCRETE

FIG. 9 Expansion of mix 4b (2.7 kg/m3) with and without y-ash


replacement.

the highly reactive ASR aggregate in mix 4b; instead, 35 % yash replacement will ensure a safe mix (i.e., ACCT expansion
<0.04 %). This is also an indication that the proposed concretecylinder test might be effectively used to determine y-ash content, cement content, and content of ternary blends to develop
safe ASR-resistant mixes.
PROPOSITION OF AN ASR-RESISTANT MIX DESIGN
VALIDATION PROCEDURE

VALIDATION OF FLY-ASH CONTENT

Mix 4b (2.7 kg/m3, 0.82 % Na2Oe) with and without y-ash F


replacement was conducted to validate the adjusted mix in
ACCT. Fly-ash F replaced 25 % and 35 % of cement in the mix,
and pore solution was extracted to generate the soak solution.
The pore solution (Na
e ) is reduced from 0.66 N to 0.43 N with
25 % y-ash replacement and from 0.66 N to 0.33 N with 35 %
y-ash replacement. It can be expected that the expansion can
be reduced because of a reduction of alkali concentration in
pore solution. Fig. 9 shows the expansion curves of mix 4b with
and without y-ash replacement. The expansion at 28 days is
reduced from 0.269 % to 0.067 % with 25 % y-ash replacement
and reduced to 0.004 % with 35 % y-ash replacement. This
indicates that 25 % y-ash replacement does not provide sufcient protection measure (i.e., ACCT expansion >0.04 %) for

For the aggregates where the reactivity prediction based on the


current test methods (e.g., ASTM C1260, ASTM C1293), mix
design validation (Fig. 10) through direct VCMD concrete testing is recommended. The mix design based on the current test
methods and mitigation practices can be tested with or without
added alkalis and verify whether the expansion stays below the
assigned limit (i.e., 0.04 %) at the specied testing period (i.e.,
28 days). If the expansion is above the assigned limit, the mix
needs further adjustment before placement.

Conclusions
Based on the results, the following conclusions can be drawn:

Because the data collection in the proposed test is automatic through LVDT (no error because of operation)
under constant temperature (no error because of temperature difference), the reliability of the proposed cylinder
test is expected to be high. Creating a leach-proof situation in the proposed cylinder test is another advantage
that enhances the reliability of the proposed cylinder test.
The proposed ACCT method was developed to determine
the length change of concrete cylinder (7.6 cm by 15.2
cm) caused by ASR at a temperature of 60 C. ACCT with
relatively low alkali levels (2.7 kg/m3, 0.82 % Na2Oe as

FIG. 10
ASR-resistant mix design validation through
ACCT.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 3 21:34:16 EDT 2016
Downloaded/printed by
Pontificia Universidad Catolica de Chile (Pontificia Universidad Catolica de Chile) pursuant to License Agreement. No further reproductions authorized.

1237

1238

Journal of Testing and Evaluation

opposed to high-alkali levels, i.e., 4.0/5.3 kg/m3, 0.95 %/


1.25 % Na2Oe in the current exposure block/CPT test) at
60 C can effectively be used to pass/fail a concrete mix in
a relatively short time (i.e., 28 days) with an expansion
limit of 0.04 %.
The proposed ACCT has the ability to emerge as a potential method to test job mix (e.g., a mix with typical ASR
mitigation measures) in the laboratory and serve as an alternative method to validate an ASR-resistant mix. It also
provides an effective measure to determine y-ash content, cement content, and content of ternary blends to develop safe ASR-resistant mixes.
For the aggregates where the reactivity prediction based
on the current test methods is satisfactory, mix design
verication/validation through the proposed concrete
testing is highly recommended.
Although there are no data on eld performance (i.e., exposure block) of the same mixes at this time, an attempt
will be made to verify the applicability of the proposed
expansion limit (i.e., 0.04 % at 28 days) by testing more
numbers of aggregates in ACCT and exposure block.
This can be accomplished by generating more concrete
data using different mixes, followed by a comparative
assessment between the proposed cylinder test and eld
performance.

[7]

[8]

[9]

[10]

[11]

[12]

[13]

ACKNOWLEDGMENTS

The writers acknowledge and thank the sponsors of this


research: the Texas Department of Transportation, Texas A&M
Transportation Institute, and Texas A&M University.

[14]

[15]

References
[1] Liu, K. W. and Mukhopadhyay, A., Alkali-Silica Reaction
in a Form of Chemical Shrinkage, Civil Eng. Arch., Vol. 2,
2014, pp. 235244.
[2] Ideker, J. H., Drimalas, T., Bentivegna, A. F., Folliard, K. J.,
Fournier, B., Thomas, M. D. A., Hooton, R. D., and Rogers, A.
C., The Importance of Outdoor Exposure Site Testing, Fourteenth International Conference on Alkali-Aggregate Reaction in
Concrete, Paper 051412-IDEK, Austin, TX, May 2025, 2012.
[3] Ranc, R. and Debray, L., Reference Test Methods and a
Performance Criterion for Concrete Structures, Ninth
International Conference on Alkali-Aggregate Reaction in
Concrete, London, U.K., July 2731, 1992, pp. 824831.
[4] De Grosbois, M., Evaluation of the Potential Alkali-Reactivity
of Concrete Aggregates: Performance of Testing Methods and
a Producers Point of View, Eleventh International Conference
of Alkali-Aggregate Reactivity in Concrete, Quebec City, QC,
Canada, June 1116, 2000, pp. 267276.
[5] Touma, W., Fowler, D., Carrasquillo, R., Folliard, K. J., and
Nelson, M. R., Characterizing the Alkali-Silica Reactivity
of Aggregates Using ASTM C1293, ASTM C1260, and
Their Modications, J. Trans. Res. Board, No. 1757, 2001,
pp. 157165.
[6] Folliard, K. J., Ideker, J. H., Thomas, M. D. A., and Fournier,
B., Assessing Aggregate Reactivity Using the Accelerated
Concrete Prism Test, Seventh CANMET/ACI International

[16]

[17]

[18]

[19]

[20]

[21]

Conference on Recent Advances in Concrete Technology, Las


Vegas, NV, May 2629, 2004, pp. 269283.
Ideker, J. H., Folliard, K. J., Fournier, B., and Thomas, M. D.
A., The Role of Non-Reactive Aggregates in the Accelerated
(60 C) Concrete Prism Test, Eighth CANMET/ACI International Conference on Recent Advances in Concrete Technology,
Montreal, Canada, May 31June 3, 2006, pp. 4570.
Ideker, J. H., East, B. L., Folliard, K. J., Thomas, M. D. A.,
and Fournier, B., The Current State of the Accelerated
Concrete Prism Test, J. Cement Concrete Res., Vol. 40,
No. 4, 2010, pp. 550555.
ASTM C295-12, Standard Guide for Petrographic Examination of Aggregates for Concrete, Annual Book of ASTM Standards, ASTM International, West Conshohocken, PA, 2013,
www.astm.org
ASTM C1260-07, Standard Test Method for Potential Alkali Reactivity of Aggregates (Mortar-Bar Method), Annual Book of ASTM Standards, ASTM International, West
Conshohocken, PA, 2013, www.astm.org
ASTM C1293-08b, Standard Test Method for Determination of Length Change of Concrete Due to Alkali-Silica
Reaction, Annual Book of ASTM Standards, ASTM International, West Conshohocken, PA, 2013, www.astm.org
Liu, K. W. and Mukhopadhyay, A., A Kinetic-Based ASR
Aggregate Classication, J. Constr. Build. Mater., Vol. 68,
2014, pp. 525534.
ASTM C192-13, Standard Practice for Making and Curing
Concrete Test Specimens in the Laboratory, Annual Book
of ASTM Standards, ASTM International, West Conshohocken, PA, 2013, www.astm.org
Barneyback, R. S. and Sidney, D., Expression and Analysis
of Pore Fluids from Hardened Cement Pastes and Mortars,
J. Cement Concrete Res., Vol. 11, No. 2, 1981, pp. 279285.
French, W. J., Reactions between Aggregates and Cement
PasteAn Interpretation of the Pessimum, Quart. J. Eng.
Geol. Hydrogeol., Vol. 13, No. 4, 1980, pp. 231247.
Lorenzo, P., Goni, S., Hernandez, S., and Guerrero, A., Effect
of Fly Ashes with High Total Alkali Content in the Alkalinity
of the Pore Solution of Hydrated Portland Cement Paste, J.
Am. Ceram. Soc., Vol. 79, No. 2, 1996, pp. 470474.
Brouwers, H. J. H. and van Eijk, R. J., Alkali Concentrations of Pore Solution in Hydrating OPC, J. Cement
Concrete Res., Vol. 33, No. 2, 2003, pp. 191196.
Berube, M. A., Tremblay, C., Fournier, B., Thomas, M. D.,
and Stokes, D. B., Inuence of Lithium-Based Products
Proposed for Counteracting ASR on the Chemistry of Pore
Solution and Cement Hydrates, J. Cement Concrete Res.,
Vol. 34, 2004, pp. 16451660.
Lothenbach, B. and Winnefeld, F., Thermodynamic Modelling of the Hydration of Portland Cement, J. Cement
Concrete Res., Vol. 36, 2006, pp. 209226.
Bolotte, B., Development of an Accelerated Performance
Test on Concrete for Evaluating Its Resistance to AAR,
Ninth International Conference on AAR in Concrete, London, U.K., July 2731, 1992, pp. 110116.
Fournier, B., Nkinamubanzi, P. C., and Chevrier, R.,
Comparative Field and Laboratory Investigations on the Use
of Supplementary Cementing Materials to Control AlkaliSilica Reaction in Concrete, Twelfth International Conference
on Alkali-Aggregate Reaction in Concrete, Beijing, China, Vol.
1, World Publishing Corporation, Beijing, 2004, pp. 528537.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 3 21:34:16 EDT 2016
Downloaded/printed by
Pontificia Universidad Catolica de Chile (Pontificia Universidad Catolica de Chile) pursuant to License Agreement. No further reproductions authorized.