Cement and Concrete Composites 125 (2022) 104282

Contents lists available at ScienceDirect

Cement and Concrete Composites

journal homepage: www.elsevier.com/locate/cemconcomp

Using electrical resistivity to determine the efficiency of supplementary

cementitious materials to prevent alkali-silica reaction in concrete
Krishna Siva Teja Chopperla a, *, Jason H. Ideker b
a
Civil and Construction Engineering, Oregon State University, 101 Kearny Hall, Corvallis, OR, 97331, USA
b
Oregon State University, USA

A R T I C L E I N F O A B S T R A C T

Keywords: In this study, the miniature concrete prism test (MCPT) was used to evaluate expansion due to alkali-silica re­
Alkali-silica reaction action (ASR) in concrete mixtures with very highly reactive fine aggregates and supplementary cementitious
Supplementary cementitious materials materials (SCMs). The concrete pore solution alkalinity and transport properties of concrete measured through
Electrical resistivity
bulk electrical resistivity were determined to investigate their correlation to the MCPT expansions. The findings
Miniature concrete prism test
Pore solution alkalinity
showed that the concrete mixtures with SCM combinations that resulted in higher bulk electrical resistivity and
Formation factor lower pore solution alkalinity had less ASR expansion in the MCPT conditions. A potential new approach to
determine the efficiency of SCM combinations to prevent ASR was proposed by combining the hydroxide con­
centration of pore solution and electrical resistivity of concrete.

1. Introduction measure expansion due to ASR are the accelerated mortar bar test
(AMBT) and the concrete prism test (CPT). Recently, the miniature
Alkali-silica reaction (ASR) is a deleterious reaction that can cause concrete prism test (MCPT – AASHTO T 380 [10]) was shown to be more
expansion and damage in concrete with alkali-reactive aggregates. The reliable and faster than the CPT when benchmarked with outdoor
reaction occurs between the reactive silica of aggregate and alkali hy­ exposed blocks in terms of determining the SCMs’ efficiency to prevent
droxides in the concrete pore solution in the presence of moisture and ASR in concrete mixtures with highly and very-highly reactive aggre­
calcium hydroxide. The main source of alkalis in concrete is cement. The gates [11].
other possible sources are supplementary cementitious materials The increase in expansion due to ASR is often evaluated by studying
(SCMs), aggregates, and external sources such as deicing salts and the pore solution chemistry [2,12–14]. Even though pore solution
seawater [1]. alkalinity is considered as the major factor influencing ASR deteriora­
Limiting the alkali content and moisture in concrete with reactive tion, the extent of ASR in concrete could also depend on transport
aggregates can reduce the expansion due to ASR. One of the most properties as the ingress of water and alkalis could exacerbate the
common ways to reduce the alkali content in concrete other than using deterioration. However, the effect of transport properties of concrete on
low alkali cement is to replace a certain amount of cement with SCMs. ASR expansion has not been well studied. The electrical resistivity of
SCMs not only lower the alkali content of concrete by dilution but also concrete depends on porosity, pore connectivity, and pore solution
by binding alkalis due to additional C–S–H forming from the pozzolanic composition at a certain degree of saturation and temperature. Electrical
reaction [2–5]. Also, SCMs can refine the microstructure due to both resistivity, and further the formation factor, have been extensively used
filler and chemical effects [6]. The improvement in pore refinement due to evaluate the transport properties of concrete and as a measure of
to the use of SCMs can help reduce the ingress of moisture and alkalis, concrete durability [15–19]. Formation factor is a material property that
thus limiting the ASR [7–9]. The efficiency of SCMs to prevent ASR is represents the pore structure as it is inversely related to porosity and
generally studied in laboratory samples under accelerated conditions by connectivity of the material pore structure, and it is independent of
measuring expansion, confirming the formation of ASR products using cementitious material and pore solution chemistry. Apparent formation
scanning electron microscopy imaging, and/or quantifying the damage factor (formation factor determined on samples that are at matrix
using petrographic tools. Some of the well-known laboratory methods to saturation) and saturated formation factor are correlated as a function of

* Corresponding author.
E-mail addresses: [email protected] (K.S.T. Chopperla), [email protected] (J.H. Ideker).

https://doi.org/10.1016/j.cemconcomp.2021.104282
Received 8 May 2021; Received in revised form 27 September 2021; Accepted 28 September 2021
Available online 2 October 2021
0958-9465/© 2021 Elsevier Ltd. All rights reserved.
K.S.T. Chopperla and J.H. Ideker Cement and Concrete Composites 125 (2022) 104282

the degree of saturation of the sample [15]. The higher the formation Table 1
factor of concrete is, the lower the ion diffusion and mass transport in The chemical oxide composition, D50, and D90 values of the cements.
concrete [19]. For instance, electrical resistivity and formation factor Cement OPC PLC
were used to determine chloride diffusion and water absorption [17].
SiO2 (%) 20.54 18.46
Therefore, electrical resistivity or formation factor as a measure of Al2O3 (%) 4.05 3.71
transport in concrete could be used as an additional parameter to assess Fe2O3 (%) 3.62 3.46
ASR expansion in concrete. CaO (%) 61.72 60.45
There are a few studies [20–22] where the influence of ASR on the MgO (%) 2.52 2.28
SO3 (%) 1.80 1.71
electrical properties of concrete was explored. Flores et al. [20] associ­ Na2O (%) 0.17 0.13
ated the lower (than expected) electrical resistivity with ASR expansion K2O (%) 0.69 0.63
in mortars with reactive aggregate. The lower electrical resistivity was Na2Oe (%) 0.62 0.54
attributed to cracks from ASR that increased the pore connectivity of the LOI (%) 1.96 6.75
CaCO3 (%) 1.79 13.11
mortar. Rivard and Saint-Pierre [21] measured p-wave velocities and
D50 (μm) 11.77 12.57
estimated Young’s modulus by resonant natural vibration frequencies D90 (μm) 29.94 37.06
and electrical resistivity to assess ASR progression. They reported that
electrical resistivity alone was not an effective tool to determine the ASR
progress, and they concluded that dynamic Young’s modulus and ul­
Table 2
trasonic pulse velocity yielded better results. However, it should be
The chemical oxide composition, D50, D90, and maximum reactivity of the
noted that the authors did not measure resistivity on the saturated SCMs used in the study (n.a. – not available).
samples, which could have influenced their resistivity measurements as
SCM Fly Fly Slag Silica Natural Limestone
degree of saturation affects the resistivity measurements. Also, the steel
Ash 1 Ash 2 (SL) Fume Pozzolan powder
pins at the ends of the specimens that were used for length measure­ (FA1) (FA2) (SF) (NP) (LS)
ments were used as electrodes to measure resistivity, which could have
SiO2 (%) 51.86 47.15 31.58 95.88 66.42 1.83
resulted in inconsistent measurements. Recently, Khajehnouri et al. [22] Al2O3 (%) 21.70 16.57 12.13 0.69 11.98 0.09
validated the use of the complex electrical measurements using spectral Fe2O3 (%) 5.04 5.88 0.55 0.12 0.86 0.05
induced polarization (SIP) to monitor ASR progression in concrete. The CaO (%) 8.61 12.54 41.34 0.70 4.06 55.62
authors found that the mixtures affected by ASR had a significant evo­ MgO (%) 2.58 4.80 6.97 0.26 0.18 0.47
SO3 (%) 0.78 0.60 3.75 0.15 0.19 0.05
lution of SIP responses such as phase lag, total chargeability, and mean
Na2O (%) 2.58 3.65 0.24 0.16 3.57 0.08
relaxation time at low frequency range. K2O (%) 1.45 1.72 0.28 0.49 4.35 0.05
While electrical resistivity alone may not show promise for moni­ TiO2 (%) 1.19 1.17 0.47 0.01 0.09 0.01
toring the progression of ASR, it may be effective for evaluating the P2O5 (%) 0.23 0.24 0.00 0.05 0.00 0.06
ZnO (%) 0.02 0.01 0.00 0.06 0.00 0.00
efficiency of SCMs for preventing ASR. To the author’s best knowledge,
Mn2O3 (%) 0.03 0.09 0.19 0.04 0.07 0.01
no studies were found in the literature that used electrical resistivity and Cl (%) 0.01 0.00 0.00 0.01 0.02 0.00
formation factor as indicators of transport properties of concrete to LOI (%) 1.42 n.a. n.a. 4.30 4.09 41.68
determine the efficiency of SCM combinations for ASR prevention. D50 (μm) 13.01 5.49 29.08 4.96 13.11 15.40
Therefore, this study focuses on i) determining the correlation between D90 (μm) 44.62 16.90 87.81 11.28 44.28 40.30
CH 83.7 72.8 38.6 187.4 87.0 n.a.
ASR expansion and transport properties of concrete and ii) investigating
consumed
if electrical resistivity through the formation factor of concrete can be (g/100 g
used along with hydroxide concentration of the pore solution to deter­ SCM)a
mine the efficiency of ASR preventive measures. a
Determined using thermogravimetric analysis according to the SCM pozzo­
lanicity test detailed in [23].
2. Materials and methods

2.1. Materials Table 3

Estimated major constituents of the reactive aggregates determined by petro­
In this study, two cements – ordinary portland cement and portland- graphic analysis.
limestone cement labeled as OPC and PLC were used. Both the cements Constituent Approximate % by weight
were made with the same (ASTM Type II/V) clinker with PLC having Fine aggregate – 1 (F1)
15% limestone (interground) content by mass. Several SCMs including Granitic Rock 44.7
two class F fly ashes with different calcium oxide contents (FA1 and Quartz 17.6
FA2), a slag (SL), a silica fume (SF), a natural pozzolan (NP), and a Feldspar 16.3
Diorite 8.8
limestone powder were used. The chemical oxide composition and
Volcanic Rock 6.4
particle size parameters (D50 and D90) determined using X-ray fluo­ Fine aggregate – 2 (F2)
rescence spectrometer (XRF) and laser diffraction particle size analyzer, Intermediate Volcanic Rock 59.2
respectively, are listed in Table 1 and Table 2. In addition, Table 2 lists Intermediate Volcanic Rock (Oxidized) 12.1
the pozzolanic reactivity of SCMs in terms of CH consumed in the Felsic Volcanic Rock 12.6
Feldspar 4.4
pozzolanic reactivity test as described in Ref. [23]. Quartz 2.3
Two very highly reactive fine aggregates (according to ASTM C1778 Granitic Rock 1.7
[24]), labeled as F1 and F2, were used in the study. The 14-day ASTM
C1260 expansions of the mixtures with F1 and F2 were 0.54% and
0.47%, respectively. A non-reactive high purity limestone rock was used comprised of granite, quartz, and feldspar. Whereas, F2 aggregate is
as a coarse aggregate for the concrete mixtures in this study. The major majorly comprised of volcanic rock fragments.
constituents of the reactive aggregates identified and estimated using
petrographic analysis (ASTM C295) are listed in Table 3. It was observed 2.1.1. Mixtures
that both the reactive aggregates have a different composition; there­ The SCM combinations used in the study were 1) 25% FA1 (25FA1),
fore, they are expected to react differently. F1 aggregate is majorly 2) 30% FA2 (30FA2), 3) 20% FA1 + 5% SF (20FA1-5SF), 4) 50% SL

2
K.S.T. Chopperla and J.H. Ideker Cement and Concrete Composites 125 (2022) 104282

(50SL), 5) 25% FA1 + 25% SL (25FA1-25SL), and 6) 25% NP (25NP). 2.2.2. Electrical resistivity
The percent replacement of cements with SCMs was done on a mass Bulk electrical resistivity measurements were made at three different
basis. The interblended PLCs were denoted as OPC +10LS, implying test durations – 28, 56, and 84 days on the concrete cylinders. The
10% limestone powder by mass was added to the OPC at the time of concrete cylinders were cooled down to room temperature (23 ◦ C) from
mixing in the laboratory. Table 4 lists the twenty mixtures considered in 60 ◦ C before measuring their electrical resistivity. For cooling down the
this study. A control PLC mixture is not included in the test matrix as the concrete cylinder, the cylinder was taken out of the MCPT container and
preliminary investigation with the Pyrex mortar bar test showed that transferred to an empty plastic cylinder (75 mm diameter and 150 mm
PLC mixtures performed better than the OPC mixtures. length). The soak solution of twice the volume of the cylinder was
The concrete mixtures were designed and proportioned per AASHTO transferred from the MCPT container to the plastic cylinder to immerse
T 380. The water-to-cementitious material (w/cm) ratio of the mixtures the concrete cylinder in the same soak solution while it cooled down.
was 0.45. The total cementitious material content was 420 kg/m3. The The resistivity measurements were taken at 10 kHz frequency using a
alkali contents of all the mixtures were boosted to 1.25% of the cement commercially available resistivity meter. Before taking the resistivity
mass. The non-reactive coarse aggregate consisted of 9.5 mm sieve measurements, the excess moisture on the surface of the cylinders was
retained rock (57.5% by mass) and 4.8 mm sieve retained rock (42.5% removed using a damp cloth. To maintain a good contact between the
by mass). The dry mass of coarse aggregate per unit volume of concrete electrodes and the samples, a sponge soaked in lime saturated solution
was equal to 0.65 times its dry-rodded bulk density as per AASHTO T was used as a conductive medium in between the sample and the elec­
380. trodes. The error induced by the conductive medium on the resistivity
measurement was corrected as described in AASHTO TP 119. Required
2.2. Methods geometric corrections were made to calculate the bulk electrical re­
sistivity of the concrete cylinder as described in Ref. [25]. Two replicates
Three concrete prisms and two cylinders were cast per mixture with were tested for one mixture, and less than 5% coefficient of variation
the same mixture proportions. The prims were used for measuring (COV) was observed.
expansion due to ASR as described in section 2.2.1. The cylinders were
used to measure bulk electrical resistivity and express pore solutions as 2.2.3. Pore solution analysis
described in sections 2.2.2 and 2.2.3, respectively. The specimens of the The pore solutions of the concrete cylinders exposed to MCPT con­
same mixture were stored in the same container as shown in Fig. 1. ditions were expressed at two different test durations – 28 and 84 days. A
maximum stress of 870 MPa was applied on the crushed concrete cyl­
2.2.1. The miniature concrete prism test inder sample using a mechanical pore press to express the pore solution
The concrete prisms were cast according to AASHTO T 380 [10]. The [26,27]. The load rate was gradually increased and maintained between
dimensions of the prims were 50 mm × 50 mm x 285 mm. The prims 2.2 and 2.9 kN/s. All the expressed pore solutions were transferred to
were immersed in 1 N NaOH solution at 60 ◦ C after the initial curing and sealed syringes, de-aired to prevent carbonation, and they were stored at
conditioning procedure described in the standard. The volume ratio of 5 ◦ C. The solutions were later filtered through 0.45 μm cellulose filters to
the sodium hydroxide solution (also referred to as soak solution) to the remove any solids. The filtered solutions were analyzed to determine
specimens used was 2:1. The prisms were monitored for their length and their ion concentrations using XRF as described in Ref. [28]. The hy­
mass change for a period of 84 days, which is 28 days more than the droxide concentration was calculated from the sodium, potassium, cal­
recommended duration in the standard. The expansions of 3 bars per cium, sulfate, and chloride ion concentrations of the pore solutions.
mixture were recorded, and their average was reported. As per AASHTO
T 380 [10] expansion criteria, prevention measures in the concrete 2.2.4. Apparent formation factor
mixture are: The apparent formation factor of concrete cylinders was calculated
by dividing the bulk electrical resistivity of the specimen with the pore
• ineffective to prevent ASR if the 56-day average expansion is greater solution resistivity as shown in equation (1). In this study, the apparent
than 0.025% formation factor was determined for the concrete cylinders immersed in
• uncertain to prevent ASR if the 56-day average expansion is in be­ 1 N NaOH at 60 ◦ C for 28 and 84 days. As the measurements were taken
tween 0.020% and 0.025% at or after 28 days of sample immersion in the soak solution, it can be
• effective to prevent ASR if the 56-day average expansion is less than considered that the samples were at matrix saturation at the time of
0.020% measurements [15]. The pore solution resistivity was calculated from
the pore solution ion concentrations using the approach described in
Refs. [28,29].
Bulk resistivity of the specimen
Table 4 Formation factor = (1)
Experimental matrix. Pore solution resistivity

Cement SCM F1 fine aggregate F2 fine aggregate

3. Results
OPC Control (no SCM) x x
OPC 25FA1 x
3.1. The miniature concrete prism
–
20FA1-5SF x –
50SL x –
25FA1-25SL x – Fig. 2 shows the 28, 56, and 84-day MCPT expansions of the concrete
25NP – x mixtures with F1 reactive aggregate. The control mixture with no SCMs
30FA2 x x expanded the most compared to the mixtures with SCMs as expected.
PLC 25FA1 x x
20FA1-5SF x –
The mixtures with SCMs that failed the test by expanding more than
50SL x – 0.025% at 56 days were 30FA2 (OPC, PLC, OPC+10LS), 20FA1-5SF
25FA1-25SL x – (OPC), and 50SL (OPC). However, it should be noted that 56-day ex­
25NP – x pansions of OPC-20FA1-5SF and OPC-50SL mixtures were just above the
30FA2 x x
0.025% limit with average 56-day expansions of 0.028% and 0.027%
OPC+10LS 25FA1 x –
50SL x – respectively. The overall descending order of the mixtures in which the
30FA2 x – expansions observed were OPC >30FA2 > 20FA1-5SF > 50 SL > 25FA1

3
K.S.T. Chopperla and J.H. Ideker Cement and Concrete Composites 125 (2022) 104282

Fig. 1. Schematic of the concrete specimens in the MCPT container.

Fig. 2. 28, 56, and 84-day MCPT expansions of the mixtures with F1 aggregate.

> 25FA1-25SL. Fig. 3. 28, 56, and 84-day MCPT expansions of the mixtures with F2 aggregate.
Among the mixtures with SCMs, the mixtures with 30FA2 expanded
the most and failed the test and the mixtures with 25FA1 passed the test
even though both the fly ashes are of Class-F classification. The reason
for high expansion of mixtures with 30FA2 is likely due to higher alkali
content (4.0% Na2Oe) and relatively lower alumina content (16.6%) of
FA2. In addition, FA2 has lower pozzolanic reactivity compared to FA1,
thus implying lower alkali binding capacity of FA2. Overall, the mix­
tures with Portland-limestone cement expanded slightly less than the
OPC mixture, which might be attributed to relatively lower pore solu­
tion alkalinity.
Fig. 3 shows the 28, 56, and 84-day MCPT expansions of the concrete
mixtures with F2 reactive aggregate. The mixtures with SCMs expanded
much lower than the control mixture as expected. The mixtures with
30FA2 failed the test by expanding more than 0.025% at 56 days. The
descending order of the mixtures in which the expansions observed were
OPC >30FA2 > 25NP > 25FA1. Similar to the mixtures with F1
aggregate, the mixtures with 30FA2 expanded the most among the
mixtures with SCMs and failed the test. It was observed that FA1 per­
formed slightly better than NP even though NP has slightly better poz­
zolanicity than FA1 according to the pozzolanic reactivity test. The Fig. 4. 28, 56, and 84-day bulk electrical resistivities of the mixtures with
reason for the slightly better performance of FA1 than NP is likely due to F1 aggregate.
FA1 mixture’s lower transport properties, thus limiting alkali ingress
from the soak solution, and this will be explained further in sections 3.3 concrete mixtures with F1 reactive aggregate. The control mixture (with
and 3.4. no SCMs) had the lowest bulk electrical resistivity compared to the SCM
mixtures as SCMs, in general, refine the pore structure and increase the
resistivity of the pore solution by binding alkalis. The overall descending
3.2. Electrical resistivity order of the mixtures in which the bulk electrical resistivities observed
were 25FA1-25SL > 20FA1-5SF > 25FA1 > 50SL > 30FA2 > Control.
Fig. 4 shows the 28, 56, and 84-day bulk electrical resistivity of the

4
K.S.T. Chopperla and J.H. Ideker Cement and Concrete Composites 125 (2022) 104282

Over the test duration, while the increase in SCMs’ reaction contribute
to increase in the resistivity of concrete, cracking due to ASR could
reduce the resistivity of concrete due to higher diffusion of the soak
solution into the specimens.
Among the SCM mixtures, the mixtures with 25FA1-25SL and 30FA2
had the highest and lowest bulk electrical resistivity, respectively. This
can be explained by studying the amount of SCM in the mixture, SCM
pozzolanic reactivity, and pore solution resistivity, as they can affect the
electrical resistivity of concrete. The pozzolanicities of the SCM blends
(i. FA1 and SF, ii. FA1 and SL) were determined the same way as the
pozzolanicities of individual SCMs using the pozzolanic reactivity test
(as listed in Table 2). The amount of CH consumed by FA1-SF blend
(80% FA1 and 20% SF to obtain 20FA1-5SF) and FA1-SL blend (50%
FA1 and 50% SL to obtain 25FA1-25SL) in the pozzolanic reactivity test
were 88.0 g and 50.2 g per 100 g SCM blend. The amount of SCM(s) in
the mixture and the pozzolanic reactivities of SCM/SCM blends were
correlated well (R2 = 0.92–0.96) to bulk electrical resistivities of the
OPC concrete specimens as shown in Fig. 5. The mixture with 30FA2 was
excluded from the correlation as the lower pore solution resistivity (due
to higher available alkali from FA2; explained further in section 3.3) of
Fig. 6. 28, 56, and 84-day bulk electrical resistivities of the mixtures with
the mixture reduced the bulk electrical resistivity of the mixture. The
F2 aggregate.
general observation was the higher the pozzolanic reactivity of SCM/
SCM blend and higher SCM fraction in the mixture, the higher the bulk
3.3. Pore solution analysis
electrical resistivity of the concrete mixture – likely due to better pore
refinement and alkali binding contributing to higher electrical resistivity
Table 5 shows the hydroxide ion concentration of the pore solutions
when the comparison is made among the mixtures with SCMs at a test
expressed from the concrete cylinders at 28 and 84-day test durations
duration. Similar observations were made by Ramanathan et al. [30],
and the pore solution resistivities. It should be noted that the mixtures
where the calcium hydroxide consumed by SCM in the pozzolanic
with 50SL at 28 days were not expressed for pore solution due to tech­
reactivity test was correlated well (R2 = 0.62) with the bulk resistivity of
nical issues in the laboratory. It was observed that the control mixtures
cement pastes with SCMs.
had the highest hydroxide ion concentration compared to the mixtures
The bulk electrical resistivity of the mixtures with limestone was
with SCMs. The observed descending order of 28-day hydroxide con­
found to be similar or higher than the bulk electrical resistivity of the
centrations of the mixtures with F1 aggregate was OPC >30FA2 >
OPC mixtures, likely due to lower transport in the PLC mixtures (this
25FA1 > 20FA1-5SF > 25FA1-25SL. When 84-day hydroxide ion con­
observation was confirmed and reported in section 3.4). Fig. 6 shows 28,
centrations of with F1 aggregate were compared, the observed order was
56, and 84-day bulk electrical resistivity of the concrete mixtures with
OPC >30FA2 > 20FA1-5SF > 50SL > 25FA1 > 25FA1-25SL. It was
F2 reactive aggregate. Similar to the mixtures with F1 aggregate, the
observed the mixtures with 30FA2 had the highest alkalinity among the
control mixture had lower bulk electrical resistivity than the mixtures
SCM mixtures due to the high alkali content (4.8% Na2Oe) of FA2. The
with SCMs. Also, the mixtures with 30FA2 were observed to have the
mixtures with 25FA1-25SL had the lowest hydroxide ion concentration
lowest resistivity among the mixtures with SCMs.
likely due to their higher cement replacement and relatively high
pozzolanic reactivity. It is interesting to note that the 25FA1 mixture had
higher alkalinity than 20FA1-5SF mixture at 28 days but lower alkalinity
at 84 days. Similar mixtures with silica fume were previously shown to
release the initially bound alkali at later ages [31], which may explain
this observation.
Three mixtures of pastes were prepared with w/cm of 0.47 to
confirm the available alkali from FA2. The mixtures include OPC-control
(100% OPC), OPC-25FA1, and OPC-25FA2. The paste mixtures were
cured in sealed condition at 23 ◦ C for 28 days. Later, their pore solutions
were expressed and analyzed for their sodium and potassium ion con­
centrations. The pore solution analysis of the pastes showed that the
concentration of the sum of Na+ and K+ in the pore solutions of OPC-
control mixture was 0.529 mol/l, in OPC-25FA1 mixture was 0.350
mol/l, and in OPC-25FA2 mixture was 0.490 mol/l. The alkali content of
the OPC-25FA2 mixture (0.490 mol/l) was higher than the 25% dilution
of alkali content of the OPC-control mixture (0.75*0.529 = 0.396 mol/
l); this confirmed that FA2 has higher available alkali than it binds.
When the pore solution alkalinity of the mixtures with F2 aggregates
were compared, the observed descending order of alkalinity was OPC
>30FA2 > 25NP > 25FA1. Even though NP is slightly more pozzolanic
reactive than FA1, the pore solution alkalinity of 25FA1 was slightly
lower than 25NP. This could be due to better refinement of the pore
structure of FA1 mixtures leading to lower soak solution ingress (this
explanation was confirmed with the results shown in section 3.4). The
Fig. 5. Correlation between SCM pozzolanicity (CH consumed in the pozzo­
pore solution resistivities of all the mixtures were inversely correlated to
lanic reactivity test) and amount of SCM and 28-day bulk electrical resistivity of
the hydroxide ion concentrations as the pore solutions, in general, are
the OPC concrete specimens.

5
K.S.T. Chopperla and J.H. Ideker Cement and Concrete Composites 125 (2022) 104282

Table 5
The hydroxide ion concentration of the pore solutions and pore solution resistivities determined at 28-day and 84-day test durations (n.a. – not available).
Reactive Mixture 28-day [OH− ] concentration, 84-day [OH− ] concentration, 28-day pore solution resistivity, 84-day pore solution resistivity,
aggregate mol/l mol/l kohm.cm kohm.cm

F1 OPC_Control 0.688 0.851 0.0076 0.0063

OPC_25FA1 0.491 0.571 0.0101 0.0089
PLC_25FA1 0.497 0.513 0.0100 0.0097
OPC+10LS_25FA1 0.501 0.607 0.0100 0.0084
OPC_30FA2 0.657 0.729 0.0079 0.0072
PLC_30FA2 0.569 0.663 0.0090 0.0078
OPC+10LS_30FA2 0.549 0.598 0.0092 0.0086
OPC_50SL n.a. 0.599 n.a. 0.0086
PLC_50SL n.a. 0.525 n.a. 0.0096
OPC+10LS_50SL n.a. 0.490 n.a. 0.0102
OPC_20FA1-5SF 0.443 0.706 0.0111 0.0076
PLC_20FA1-5SF 0.427 0.677 0.0114 0.0079
OPC_25FA1-25SL 0.400 0.565 0.0121 0.0093
PLC_25FA1-25SL 0.358 0.550 0.0134 0.0095
F2 OPC_Control 0.609 0.708 0.0085 0.0074
OPC_30FA2 0.480 0.536 0.0103 0.0094
PLC_30FA2 0.503 0.595 0.0099 0.0086
OPC_25NP 0.491 0.534 0.0101 0.0093
PLC_25NP 0.445 0.513 0.0109 0.0096
PLC_25FA1 0.430 0.504 0.0114 0.0098

dominated by alkali ions; the higher the alkali ions, the lower the re­
sistivity [32].
When the pore solution alkalinities of the OPC and PLC mixtures
were compared, it was observed that the PLC mixtures had slightly lower
pore solution alkalinities compared to the OPC mixtures even though the
mixtures were boosted to the same alkali content of the cement mass.
The reason for this is due to better pore refinement in PLC mixtures than
the OPC mixtures (as explained further in section 3.4 and discussed in
detail in Ref. [33]) that could limit the ingress of the soak solution.

3.4. Apparent formation factor

Fig. 7 and Fig. 8 show the calculated apparent formation factor (Fapp)
of the concrete specimens at 28 and 84 days of test duration. The control
mixture had the lowest Fapp compared to the SCM mixtures as SCMs, in
general, improve the pore structure by both filler and chemical effects as
discussed previously. The descending order of Fapp of the mixtures with
F1 aggregate was 25FA1-25SL > 20FA1-5SF > 25FA1 > 30FA2 ~ 50SL

Fig. 8. 28 and 84-day apparent formation factor of the mixtures with

F2 aggregate.

> Control. Similarly, the descending order of Fapp determined for the
mixtures with F2 aggregate was 25FA1 > 25 NP > 30FA2 > Control.
The SCM properties that could affect the pore structure are its
reactivity and particle size distribution. The higher the reactivity of
SCMs up to a certain replacement level of cement, the higher the
amounts of hydration products form that could fill in the pores, thus
reducing porosity and pore connectivity. The order in which Fapp was
observed was similar to the observed bulk electrical resistivity with the
exception of 30FA2 and 50SL mixtures. The 30FA2 mixture had a
marginally higher Fapp than 50SL but the bulk electrical resistivity of
50SL was slightly higher than 30FA2; the reason for this could be due to
the lower pore solution resistivity of the 30FA2 mixture.
When Fapp of the OPC and PLC mixtures are compared, it was
observed that Fapp of the PLC mixtures was similar or higher than the
OPC mixtures with the exception of the mixtures with 30FA2. The
reason for overall higher Fapp for PLC mixtures could be better particle
packing, slightly improved clinker hydration due to the nucleation effect
of limestone particles, and the formation of secondary hydration prod­
ucts from the partial reaction of limestone with aluminates [34–37].
Fig. 7. 28 and 84-day apparent formation factor of the mixtures with
F1 aggregate.

6
K.S.T. Chopperla and J.H. Ideker Cement and Concrete Composites 125 (2022) 104282

4. Discussion

Correlations between the MCPT expansion and pore solution alka­

linity and bulk electrical resistivity were explored. A strong correlation
(R2 ≥ 0.85) between the MCPT expansion and bulk electrical resistivity
was observed as shown in Fig. 9. It was observed that the higher the bulk
electrical resistivity, the lower the expansion observed in the MCPT with
respect to the different mixtures.
As all the bulk electrical resistivity measurements were made at room
temperature and at matrix saturation of specimens, the parameters that
affect the bulk electrical resistivity most are pore solution composition,
porosity, and connectivity of the pore structure. Because the pore so­
lutions, in general and especially in the case of MCPT as the soak solu­
tion is NaOH, are dominated by Na+, K+, and OH− , a higher bulk
resistivity associated with fewer alkali ions could indicate lower ASR
expansion. If the bulk resistivity increases due to pore refinement from
SCM usage, which implies lower ion diffusion and mass transport, the
reduction of moisture or/and alkali ingress could result in lower ASR
expansion. For these reasons, bulk electrical resistivity could be a po­
tential parameter to evaluate the efficiency of SCMs to prevent ASR.
However, it should be noted that the previous statement might not hold Fig. 10. Correlation between the 28 and 84-day apparent formation factor and
true if the admixtures are used in concrete that could alter the pore bulk electrical resistivity of all the concrete mixtures.
solution resistivity and electrical resistivity.
Fig. 10 shows the correlation between Fapp and bulk electrical re­
sistivity of all the concrete mixtures. It was observed that Fapp and bulk
electrical resistivity correlated very well (R2 = 0.98). However, if the
concrete mixture has admixtures or alternative SCMs that could alter the
pore solution conductivity, the electrical resistivity may not represent
the transport properties of concrete well [19,38]. Formation factor is a
better indicator of transport properties than the electrical resistivity as
formation factor represents the pore structure and accounts for the in­
fluence of pore solution chemistry [19].
The correlation between the MCPT expansion and [OH− ] for the
mixtures with F1 aggregate has R2 equal to 0.89 and 0.71 at 28-day and
84-day, respectively, as shown in Fig. 11. Also, the correlation between
the MCPT expansion and [OH− ] for the mixtures with F2 aggregate has
R2 equal to 0.99 and 0.96 at 28-day and 84-day, respectively, as shown
in Fig. 12. It was observed that higher pore solution alkalinity resulted in
higher MCPT expansion. It should be noted that even though the cor­
relation between the expansion and pore solution alkalinity are similar
for the mixtures with F1 and F2 aggregates, they are not the same. This
could be due to different mineralogical composition of the reactive

Fig. 11. Correlation between the 28 and 84-day MCPT expansion and the hy­
droxide concentration of the pore solutions of the concrete mixtures with F1
reactive aggregate.

aggregates as shown in section 2.1.

As the mixtures with higher pore solution alkalinity had higher
MCPT expansion overall, and the mixtures with higher bulk electrical
resistivity and apparent formation factor had lower MCPT expansion, a
new parameter: [OH− ]/Fapp was correlated to the 56-day MCPT
expansion of all the mixtures as shown in Fig. 13.
Based on the failure criteria of the MCPT method, a threshold 28-day
[OH− ]/Fapp can be proposed for different reactive aggregates. For
instance, for the reactive aggregates used in this study, threshold 28-day
[OH− ]/Fapp of 0.0003 mol/l (or 0.300 mmol/l) was proposed as shown
in Fig. 13. Above the threshold value, higher the [OH− ]/Fapp higher the
expansion was observed. To verify the proposed threshold 28-day
[OH− ]/Fapp, two concrete mixtures, that were expected to exceed the
MCPT expansion criteria, OPC-15FA1 and OPC-30SL with F1 aggregate
were tested. The MCPT expansion, hydroxide concentration in the pore
solution, and the apparent formation factor are listed in Table 6.
Fig. 9. Correlation between the 28, 56, and 84-day MCPT expansion and bulk
It was determined that 28-day [OH− ]/Fapp of OPC-15FA1 and OPC-
electrical resistivity of all the concrete mixtures.

7
K.S.T. Chopperla and J.H. Ideker Cement and Concrete Composites 125 (2022) 104282

propose a more robust threshold [OH− ]/Fapp values for different reac­
tive aggregates.
As monitoring field exposed blocks are considered to be one of the
most reliable methods to assess ASR, a similar approach could be used to
correlate field exposed block expansions to [OH− ] and Fapp of the con­
crete specimens (with the same mixture proportions as in the blocks)
tested in the laboratory and then to establish threshold [OH− ]/Fapp for
different aggregate reactivities. The proposed approach shows the po­
tential to assess preventive measures for ASR. Also, the plot between
ASR expansion and [OH− ]/Fapp shows that performance-based limits
such as threshold hydroxide concentration and minimum Fapp can be
proposed for concrete mixtures with different reactive aggregates and
various risk levels in guidelines for ASR prevention. In addition, with
recent advances in the application of thermodynamic modeling to pre­
dict concrete properties such as formation factor and pore solution
composition [39–41], thermodynamic modeling combined with the
proposed approach in this study could allow us to determine efficient
SCM combinations for ASR with minimal laboratory testing.
In this study, two critical factors that can affect ASR expansion – pore
solution alkalinity and transport properties were correlated to MCPT
expansions. An important factor that was not explored in this study was
Fig. 12. Correlation between the 28 and 84-day MCPT expansion and the hy­ the effect of aluminum concentrations in the pore solution on reducing
droxide concentration of the pore solutions of the concrete mixtures with F2 the ASR expansion. Including the aluminum concentration in the pro­
reactive aggregate. posed parameter, [OH− ]/Fapp, may improve the robustness of the pro­
posed approach. Additional future work should include using the
proposed approach for concrete mixtures with low and moderately
reactive aggregates and non-conventional SCMs.

5. Conclusions

The miniature concrete prism test was used to evaluate the efficiency
of several SCM combinations to prevent ASR in concrete with two
different very-highly reactive aggregates. Bulk electrical resistivity, pore
solution alkalinity, and apparent formation factor of the concrete cyl­
inders exposed to the MCPT conditions were determined. The correla­
tions between the MCPT expansion and pore solution hydroxide
concentration and bulk electrical resistivity were evaluated. The major
conclusions of this study are:

• The concrete mixtures with SCM combinations that resulted in

higher bulk electrical resistivity and lower pore solution alkalinity
had less expansion due to ASR.
• The correlations between the 56-day MCPT expansion, hydroxide
concentration in the pore solution and apparent formation factor
determined at 28 days suggest a threshold [OH− ]/Fapp that is
depending on aggregate minerology. This threshold incorporates
both pore solution alkalinity and transport properties of concrete,
Fig. 13. Correlation between 56-day MCPT expansion and 28-day [OH− ]/Fapp
of all the concrete mixtures ([OH− ]: hydroxide concentration of the and after additional research, could be specified to identify SCM
pore solution). combinations to prevent ASR in concrete with reactive aggregates.
Proposing a threshold 28-day [OH− ]/Fapp based on reactive aggre­
gate mineralogy could help reduce the testing duration (e.g. from 56
30SL were 0.425 and 0.612 mmol/l which were greater than the pro­
to 28 days) and the amount of performance testing needed to eval­
posed threshold value for the reactive aggregates used in the study.
uate the SCM combinations’ efficiency to prevent ASR.
Therefore, 15% FA1 and 30% SL were ineffective based on the proposed
approach that agrees with the MCPT expansion criteria; thus, confirming
the predicted outcome using the threshold [OH− ]/Fapp for the reactive
aggregates in this study. However, more work is needed to confirm and

Table 6
Confirmation of the proposed threshold 28-day [OH− ]/Fapp for the reactive aggregates used in the study.
Mixture Average 56-day MCPT SCM effectiveness according to 28-day [OH− ] 28-day 28-day [OH− ]/Fapp SCM effectiveness criteria:
expansion (%) AASHTO T 380 (mmol/l) Fapp (mmol/l) 28-day [OH− ]/Fapp > 0.300 mmol/l
= ineffective

OPC- 0.260 Ineffective 591 1391 0.425 Ineffective

15FA1
OPC-30SL 0.285 Ineffective 597 976 0.612 Ineffective

8
K.S.T. Chopperla and J.H. Ideker Cement and Concrete Composites 125 (2022) 104282

Declaration of competing interest [18] W.J. Weiss, T.J. Barrett, C. Qiao, H. Todak, Toward a specification for transport
properties of concrete based on the formation factor of a sealed specimen,
Advances in Civil Engineering Materials 5 (2016) 179–194.
The authors declare that they have no known competing financial [19] J.W. Weiss, C. Qiao, B. Isgor, J. Olek, Implementing rapid durability measure for
interests or personal relationships that could have appeared to influence concrete using resistivity and formation factor, Joint Transportation Research
the work reported in this paper. Program, Purdue University, West Lafayette, 2020.
[20] J. Flores, M. Kamali, A. Ghahremaninezhad, Electrical Resistivity Measurement to
Study Alkali-Silica-Reaction Cracking in Mortar, Forensic Engineering, 2015,
Acknowledgements pp. 230–241.
[21] P. Rivard, F. Saint-Pierre, Assessing alkali-silica reaction damage to concrete with
non-destructive methods: from the lab to the field, Construct. Build. Mater. 23
The research was an extension of a project funded by the California (2009) 902–909.
Department of Transportation (CalTrans), USA (Task 3136, Agreement [22] Y. Khajehnouri, P. Rivard, M. Chouteau, C.L. Bérubé, Validation of complex
65A0677). The funding source was not involved in the data collection, electrical properties of concrete affected by accelerated alkali-silica reaction,
Cement Concr. Compos. 113 (2020) 103660.
analysis and interpretation of data, and preparation of the article. The [23] P. Suraneni, J. Weiss, Examining the pozzolanicity of supplementary cementitious
authors thank Wood Environment and Infrastructure Solutions, Inc. in materials using isothermal calorimetry and thermogravimetric analysis, Cement
San Diego, USA for the petrographic examination of the reactive ag­ Concr. Compos. 83 (2017) 273–278.
[24] ASTM C1778-20 Standard Guide for, Reducing the Risk of Deleterious Alkali-
gregates. The authors thank Jeremy Smith for the laboratory assistance Aggregate Reaction in Concrete, ASTM International, West Conshohocken, PA,
and Antara Choudhary for characterizing the cementitious materials. 2020.
[25] R. Spragg, C. Villani, K. Snyder, D. Bentz, J. Bullard, W. Weiss, Factors that
influence electrical resistivity measurements in cementitious systems,
References
transportation research record, Journal of the Transportation Research Board 2342
(2013) 90–98.
[1] F. Rajabipour, E. Giannini, C. Dunant, J.H. Ideker, M.D.A. Thomas, Alkali–silica [26] R. Barneyback, S. Diamond, Expression and analysis of pore fluids from hardened
reaction: current understanding of the reaction mechanisms and the knowledge cement pastes and mortars, Cement Concr. Res. 11 (1981) 279–285.
gaps, Cement Concr. Res. 76 (2015) 130–146. [27] M.P. Adams, J.H. Ideker, Pore solution chemistry of calcium aluminate cement
[2] M. Thomas, The effect of supplementary cementing materials on alkali-silica concrete systems undergoing accelerated conversion, Factors Influencing
reaction: a review, Cement Concr. Res. 41 (2011) 1224–1231. Conversion and Volume Stability in Calcium Aluminate Cement Systems, 98.
[3] F.P.G. Sung-Yoon Hong, Alkali binidng in cement pastes Part I. The C-S-H phase, [28] M.T. Chang, P. Suraneni, O.B. Isgor, D. Trejo, W.J. Weiss, Using X-ray fluorescence
Cement Concr. Res. 29 (1999) 1893–1903. to assess the chemical composition and resistivity of simulated cementitious pore
[4] M.S.Y. Bhatty, N.R. Greening, Interaction of alkalis with hydrating and hydrated solutions, International Journal of Advances in Engineering Sciences and Applied
calcium silicates. The Fourth International Conference on the Effects of Alkalis in Mathematics 9 (2017) 136–143.
Cement and ConcretePurdue, 1978, pp. 87–112. [29] K.A. Snyder, X. Feng, B.D. Keen, T.O. Mason, Estimating the electrical conductivity
[5] F. Rajabipour, E. Giannini, C. Dunant, J.H. Ideker, M.D.A. Thomas, Alkali-silica of cement paste pore solutions from OH− , K+ and Na+ concentrations, Cement
reaction: current understanding of the reaction mechanisms and the knowledge Concr. Res. 33 (2003) 793–798.
gaps, Cement Concr. Res. 76 (2015) 130–146. [30] S. Ramanathan, M. Kasaniya, M. Tuen, M.D.A. Thomas, P. Suraneni, Linking
[6] B. Lothenbach, K. Scrivener, R.D. Hooton, Supplementary cementitious materials, reactivity test outputs to properties of cementitious pastes made with
Cement Concr. Res. 41 (2011) 1244–1256. supplementary cementitious materials, Cement Concr. Compos. 114 (2020)
[7] S.M.H. Shafaatian, A. Akhavan, H. Maraghechi, F. Rajabipour, How does fly ash 103742.
mitigate alkali–silica reaction (ASR) in accelerated mortar bar test (ASTM C1567)? [31] M.H. Shehata, M.D.A. Thomas, Use of ternary blends containing silica fume and fly
Cement Concr. Compos. 37 (2013) 143–153. ash to suppress expansion due to alkali–silica reaction in concrete, Cement Concr.
[8] R.D. Moser, A.R. Jayapalan, V.Y. Garas, K.E. Kurtis, Assessment of binary and Res. 32 (2002) 341–349.
ternary blends of metakaolin and Class C fly ash for alkali-silica reaction mitigation [32] Y. Bu, J. Weiss, The influence of alkali content on the electrical resistivity and
in concrete, Cement Concr. Res. 40 (2010) 1664–1672. transport properties of cementitious materials, Cement Concr. Compos. 51 (2014)
[9] J. Lindgård, E.J. Sellevold, M.D.A. Thomas, B. Pedersen, H. Justnes, T.F. Rønning, 49–58.
Alkali–silica reaction (ASR)—performance testing: influence of specimen pre- [33] K.S.T. Chopperla, J. Smith, J. Ideker, The efficacy of portland-limestone cements
treatment, exposure conditions and prism size on concrete porosity, moisture state with supplementary cementitious materials to prevent alkali-silica reaction
and transport properties, Cement Concr. Res. 53 (2013) 145–167. (submitted), CEMENT (2021).
[10] A. T380, Standard Method of Test for Potential Alkali Reactivity of Aggregates and [34] S. Adu-Amankwah, M. Zajac, C. Stabler, B. Lothenbach, L. Black, Influence of
Effectiveness of ASR Mitigation Measures (Miniature Concrete Prism Test, MCPT), limestone on the hydration of ternary slag cements, Cement Concr. Res. 100 (2017)
American Association of State Highway and Transportation Officials (AASHTO), 96–109.
2019, p. 10. [35] E. Berodier, K. Scrivener, Understanding the filler effect on the nucleation and
[11] J. Tanesi, T. Drimalas, K.S.T. Chopperla, M. Beyene, J.H. Ideker, H. Kim, growth of C-S-H, J. Am. Ceram. Soc. 97 (2014) 3764–3773.
L. Montanari, A. Ardani, Divergence between performance in the field and [36] B. Lothenbach, G. Le Saout, E. Gallucci, K. Scrivener, Influence of limestone on the
laboratory test results for alkali-silica reaction, Transport. Res. Rec. 2674 (2020) hydration of Portland cements, Cement Concr. Res. 38 (2008) 848–860.
120–134. [37] T. Matschei, B. Lothenbach, F.P. Glasser, The role of calcium carbonate in cement
[12] T. Kim, J. Olek, H. Jeong, Alkali–silica reaction: kinetics of chemistry of pore hydration, Cement Concr. Res. 37 (2007) 551–558.
solution and calcium hydroxide content in cementitious system, Cement Concr. [38] C.M. Tibbetts, J.M. Paris, C.C. Ferraro, K.A. Riding, T.G. Townsend, Relating water
Res. 71 (2015) 36–45. permeability to electrical resistivity and chloride penetrability of concrete
[13] M.H. Shehata, M.D.A. Thomas, The role of alkali content of Portland cement on the containing different supplementary cementitious materials, Cement Concr.
expansion of concrete prisms containing reactive aggregates and supplementary Compos. 107 (2020) 103491.
cementing materials, Cement Concr. Res. 40 (2010) 569–574. [39] D. Glosser, V.J. Azad, P. Suraneni, O.B. Isgor, W.J. Weiss, Extension of powers-
[14] Y. Kawabata, K. Yamada, Evaluation of alkalinity of pore solution based on the brownyard model to pastes containing supplementary cementitious materials, ACI
phase composition of cement hydrates with supplementary cementitious materials Mater. J. 116 (2019).
and its relation to suppressing ASR expansion, J. Adv. Concr. Technol. 13 (2015) [40] K. Bharadwaj, D. Glosser, M.K. Moradllo, O.B. Isgor, W.J. Weiss, Toward the
538–553. prediction of pore volumes and freeze-thaw performance of concrete using
[15] M.K.H.T.L. Chunyu Qiao, W.J. Weiss, Electrical resistivity and formation factor of thermodynamic modelling, Cement Concr. Res. 124 (2019) 105820.
air-entrained concrete, ACI Mater. J. 116 (2019). [41] D. Glosser, P. Suraneni, O.B. Isgor, W.J. Weiss, Estimating reaction kinetics of
[16] M.K. Moradllo, C. Qiao, O.B. Isgor, S. Reese, W.J. Weiss, Relating the formation cementitious pastes containing fly ash, Cement Concr. Compos. 112 (2020)
factor of concrete to water absorption, ACI Mater. J. 115 (2018) 887–898. 103655.
[17] C. Qiao, A. Coyle, B. Isgor, J. Weiss, Prediction of chloride ingress in saturated
concrete using formation factor and chloride binding isotherm, submitted for
publication. Advances in Civil Engineering Materials, 2018.