Construction and Building Materials 149 (2017) 406–415

Contents lists available at ScienceDirect

Construction and Building Materials

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

Drying shrinkage and cracking resistance of concrete made with ternary

cementitious components
Xiang Hu a,b, Zhenguo Shi a,c,⇑, Caijun Shi a,⇑, Zemei Wu a, Baihui Tong a, Zhihua Ou d, Geert de Schutter b
a
College of Civil Engineering, Hunan University, Changsha, Hunan 410082, China
b
Magnel Laboratory for Concrete Research, Department of Structural Engineering, Ghent University, Ghent B-9052, Belgium
c
Interdisciplinary Nanoscience Center (iNANO) and Department of Chemistry, University of Aarhus, Aarhus C 8000, Denmark
d
School of Civil Engineering, Hunan University of Technology, Zhuzhou 412007, China

h i g h l i g h t s

Compressive strength of concrete cured under drying shrinkage is investigated.

Slag and fly ash show different influences on drying shrinkage of concrete.

The cracking resistance of concrete is more controlled by total drying shrinkage.

Factorial design method is applicable for mixture proportion design.

a r t i c l e i n f o a b s t r a c t

Article history: The compressive strength, free and restrained drying shrinkage and cracking resistance of concrete under
Received 15 March 2017 drying condition were investigated for ternary cementitious systems containing Portland cement, slag
Received in revised form 9 May 2017 and fly ash. The restrained shrinkage test (ring test) was carried out following ASTM C1581. The results
Accepted 10 May 2017
showed that the increased replacement level of slag or fly ash from 0 to 50% led to a gradual decrease in
Available online 25 May 2017
28d compressive strength of concrete. The free drying shrinkage increased with the increase of slag con-
tent, but reduced with the fly ash content. The cracking resistance was well related to the free drying
Keywords:
shrinkage of concrete, other factors such as strength also showed certain effects on cracking resistance
Compressive strength
Ternary cementitious system
of concrete. The cracking resistance of concrete was enhanced with the addition of fly ash while weak-
Free drying shrinkage ened with the slag replacement level up to 50%.
Restrained shrinkage Ó 2017 Elsevier Ltd. All rights reserved.
Cracking resistance

1. Introduction physical properties, chemical and mineral component. Previous

Higher shrinkage induced cracking increases the risk of pene-
tration of aggressive substances into concrete. During the hydra-
tion or drying process of concrete, the decrease of absolute
volume for the materials leads to the generation of inner stress.
When this stress develops to an extent greater than the strength
of concrete, cracking initiates. The cracking of concrete will accel-
erate the penetration of water and other aggressive substances,
and thus exacerbate the deterioration of concrete [1,2].
Supplementary cementitious materials (SCMs), such as blast
furnace slag and fly ash are widely used to partially replace Port-
land cement. The influences of these SCMs on shrinkage and subse-
quent cracking of concrete are significantly related to their
⇑ Corresponding authors at: College of Civil Engineering, Hunan University,
Changsha, Hunan 410082, China.
E-mail addresses: zshi@chem.au.dk (Z. Shi), cshi@hnu.edu.cn (C. Shi).
studies [3] showed that the use of fly ash and slag with fineness
comparable to cement particles in concrete can efficiently restrain
the cracking and show superior durability performance than con-
crete without SCMs. However, some studies [4,5] also found that
fly ash only had small effects on reduction of drying shrinkage of
concrete at late stage. Obla et al. [6] showed that the drying shrink-
age were similar for concrete with and without fly ash, while the
addition of fly ash slightly increased the restrained cracking resis-
tance of concrete. For appropriately cured slag-blend concrete or
when ultrafine ground-granulated blast-furnace slag (GGBS) is
used, the amount of AFt (Ca6Al2(SO4)(OH)1226H2O) and calcium
silicate hydrate (C-S-H) and the density of hardened cement paste
can be greatly increased, which makes concrete more resistant to
shrinkage and cracking [7]. While a recent study [8] indicated that
the expansive strain due to the formation of AFt at early stages for
slag-blended concrete was cancelled by its higher creep. Low
Blaine slag was considered to be inappropriate for improvement

http://dx.doi.org/10.1016/j.conbuildmat.2017.05.113
0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
 X. Hu et al. / Construction and Building Materials 149 (2017) 406–415
of cracking resistance of concrete. Based on the investigation on
different factors contributing to early age shrinkage cracking in
concrete, Aly and Sanjayan [9] found that the influence of the ten-
sile elastic modulus is a major consideration for early age cracking
of slag-blended concretes.
Free drying shrinkage and restrained cracking resistance is very
important for mechanical properties and durability of concrete.
Different SCMs such as slag, fly ash and silica fume have been used
to decrease the drying shrinkage and enhance the cracking resis-
tance of concrete [10,11]. The drying shrinkage of cement-fly
ash-silica fume ternary mortars was measured in the study of
Wongkeo et al. [12], the drying shrinkage of cement mortar with
100% Portland cement was higher than all of other groups. Hale
et al. [13] investigated the drying shrinkage of concrete mixtures
with ternary system cement-slag-fly ash and found that the
shrinkage was decreased with slag addition while fly ash showed
no effects on the shrinkage of concrete.
In 1958, Simplex-lattice Design was introduced by Scheffe [14],
on which many design methods were developed, such as simplex
centroid design and the extremeness vertex design. Factorial
design method, a method to correlate the compositions of ternary
composite cements with different performance of samples with
only seven batches of experiments [15], has been applied in study-
ing hydration properties [16], strength [17], ASR expansions [18]
and chloride-ion permeability [19] for presenting the relationship
between material properties and components in a global way
and optimization of cementitious materials composition. In the
study of Shi et al. [18], the efficiency of the factorial design method
was verified by comparing the experimental and predicted results,
and the results correlated very well.
Slag and fly ash are two kinds of SCMs commonly applied in
civil engineering for partly replace Portland cement and improve
the mechanical and durability performance of concrete structure
[20,21]. The influences of slag or fly ash on shrinkage and cracking
resistance have been widely studied, the reduction of compressive
strength from fly ash [22] and increase of drying shrinkage from
slag [23] were considered to increase the risk of concrete cracking
when slag or fly ash was incorporated. The studies on how combi-
nation of slag and fly ash in cement concrete may affect the shrink-
age and cracking resistance is of great interest for application of
industrial waste and enhancement of concrete structure perfor-
mance. By applying factorial design method, the synergetic effects
of slag and fly ash on shrinkage and cracking resistance of concrete
can be studied integratedly and provide some information about
the determination and optimization of mixture proportion.
This paper studies the strength, drying shrinkage and cracking
resistance of concrete with cement-slag-fly ash ternary cementi-
tious materials. The single and synergetic influences of slag and
fly ash on shrinkage and cracking resistance of concrete are dis-
cussed. The effects of strength and drying shrinkage on cracking
resistance of concrete are also discussed. In the aid of factorial
designing method, optimal cementitious materials proportion for
higher compressive strength and cracking resistance can be
presented.
2. Experimental
2.1. Raw materials
2
A Portland cement (PO. 42.5) with the specific surface area of 336 m /kg was
used. Grades II fly ash and S95 slag according to GB/T18046-2000 were used as
SCMs. The chemical compositions of the Portland cement, fly ash and slag are
shown in Table 1. In accordance with GB/T1596-2005, the residue fly ash on sieve
with 45 lm square hole is 6.2%. The density and specific area of slag is 2900 kg/m3
and 446 m2/kg, respectively. The aggregates used in this study were river sand with
fineness modulus of 2.8 and gravel with continuous gradation between 5 and
16 mm. In order to improve the workability, polycarboxylate based superplasticizer
(water-reducing rate 30.6%, solid content 18.24%) was also applied.
407
2.2. Concrete mixture proportion and specimen preparation
Ternary binder materials comprising of Portland cement, fly ash and slag were
designed using a factorial design method as previously used in Ref. [18]. The mix
proportions of the binder materials are shown in Table 2 and projected in a ternary
diagram as shown in Fig. 1. Concrete mixtures with three different water-to-binder
(w/b) ratios, as shown in Table 3, were prepared.
In this study, three components x1, x2 and x3 are defined as the percentage of
Portland cement, fly ash and slag, and the sum of x1, x2 and x3 is 100%. Then the
value of responses Y, including any mechanical and durability characteristics of
ternary cementitious materials can be obtained as:
Y ¼ b1 x1 þ b2 x2 þ b3 x3 þ b12 x1 x2 þ b13 x1 x3 þ b23 x2 x3 þ b123 x1 x2 x3
bi bij and bjj (i, j = 1, 2, 3) are coefficients to be estimated. Based on the equation
shown above and experimental results of 7 mixtures, the values of 7 coefficients b
can be calculated by data fitting. Then the contour lines in ternary diagrams with dif-
ferent characteristic parameters, such as compressive strength, drying shrinkage and
cracking time were plotted. The software Surfer 8.0 was used for drawing.
Concrete samples were prepared according to the mixture shown in Tables 2
and 3. For compressive strength and drying shrinkage measurement, concrete
was cast in molds with the size of 75 mm  75 mm  285 mm. In addition, pre-
embedded probes were fixed at both ends of the specimen for drying shrinkage test.
After casting, the specimens were covered with plastic films and stored in a stan-
dard curing room at 20 °C and relative humidity higher than 98% for 24 h. After
demolding, samples were moved to a drying room at 23 ± 2 °C and relative humid-
ity of 50 ± 4%.
For restrained cracking measurements, fresh concrete was filled into the space
between two concentric steel rings. The concrete samples without demolding were
cured at temperature of 20 ± 2 °C and relative humidity above 95% for one day. After
that, the outer rings were removed and all samples with inner rings were moved to
a dry room with constant temperature (23 ± 2 °C) and relative humidity of
(50 ± 4%).
2.3. Testing methods
2.3.1. Compressive strength
The compressive strength was tested 28 days after mixing with one day curing
in standard room and 27d of drying under condition of 23 ± 2 °C and relative
humidity 50 ± 4%. The compressive strength was tested on two opposite lateral side
of samples and area of 75 mm  75 mm was pressed. Three samples were tested for
each group and the average value was reported.
2.3.2. Free drying shrinkage
In this study, drying shrinkage of concrete was measured according to method
described in ASTM C157-75. The length change of specimen was measured using
vertical comparator. Three specimens were prepared for each mixture. The initial
length of specimens (L0) was measured two hours after the samples were
demolded. The length (Lt) was measured at 1, 3, 7, 14, 21 and 28 days after the ini-
tial measurements. The drying shrinkage of the concrete at age of t days (est) can be
calculated as follows. The reported value for each mix is the average of three
measurements.
L0  Lt
est ¼  100% ð1Þ
L0
2.3.3. Restrained shrinkage
The restrained shrinkage of concrete was measured according to ring test
described in ASTM C1581. The mold for restrained cracking resistance measure-
ment are shown in Fig. 2 [24]. The outer ring of the steel mold has an inner diameter
of 406 mm, while the inner ring has an outside diameter of 330 mm. The inner steel
ring has a thickness of 13 mm with four strain gages mounted for strain measure-
ment. The cracking time to the nearest 0.25 day of each specimen was determined
after the drying was initiated. The measurements were continued until the sharp
change of the strain was observed.
3. Results and discussion
3.1. Compressive strength
The contour lines of compressive strength of samples for differ-
ent batches at 28 days in ternary diagrams are plotted in Fig. 3. The
compressive strength of concrete was gradually decreased with the
increase of total SCMs content, which was more remarkable for
samples with lower w/b ratio. The compressive strength of con-
crete in this study is more affected by fly ash than slag. It should
be noted that the contour plots of the part with cement content
408 X. Hu et al. / Construction and Building Materials 149 (2017) 406–415

Table 1
Chemical composition of cementitious materials (wt.%).

SiO2 Al2O3 Fe2O3 CaO MgO K2O Na2O SO3

Cement 21.91 5.30 3.67 65.10 1.51 0.62 0.19 –
Fly ash 54.29 22.55 5.53 1.34 2.56 0.71 0.19 0.44
Slag 33.00 13.91 0.82 39.11 10.04 1.91 – 0.16

Table 2 the evaporation of water at early stage of hydration may impede

Mix proportions of the ternary cementitious materials (wt.%).
No. Cement
1 100
2 50
3 50
4 75
5 75
6 50
7 66.6
the hydration of slag and fly ash contained concrete. Especially
Fly ash Slag for fly ash blended samples, the early hydration rate is much
0 0
slower and the loss of water at early time may affect the develop-
50 0 ment of hydration and compressive strength more significantly
0 50 [27]. Xu et al. [28] demonstrated that when part of cement was
0 25 replaced by fly ash, the cement percentage was reduced and can
25 0
increase the effective w/b ratio that dominates the hydration rate.
25 25
16.7 16.7 The increase of effective w/b ratio lowered Ca2+ concentration in
pore solution and weakened the interaction between particles.
Therefore, the compressive strength of concrete was decreased
more severely due to the addition of fly ash under this condition.
Comparing the results of batch B to C, the addition of superplas-
ticizer doesn’t show any influences on compressive strength of
concrete samples with w/b ratio 0.4 at 28 days of curing. In some
previous studies [29], similar results were also obtained for com-
pressive strength of concrete after the addition of superplasticizer.
However, superplasticizer addition was considered to slightly
increase the compressive strength because of the reduced air con-
tent in addition to the esteemed effect of proper placement and
compaction of concrete [30].
From the right side of the red lines (Portland cement > 50%) in
four contour plots, the compressive strength of concrete was
increased with the increase of Portland cement content, especially
within the scope of 50–75%. It can be known that under the drying
condition, it will be preferable for compressive strength to ensure
sufficient content of Portland cement higher than 75%.

3.2. Free drying shrinkage

Fig. 4 shows the free drying shrinkage of concrete with different

w/b ratios. In this paper, the free shrinkage of concrete after trans-
Fig. 1. Projection of mix proportions of ternary cementitious materials. ferring to drying condition (1 day after mixing) was measured. The
drying shrinkage of concrete developed fast at early ages and
almost reached to a constant value after about 14 days of curing.
Table 3 The contour plots of free drying shrinkage of concrete at 28 days
Mixture ratio of concrete.
of age are presented in Fig. 5. The drying shrinkage of batch A
Batch Binder Sand Gravel Superplasticizer w/b (w/b ratio 0.5) was the highest among the four groups and
A 1 1.63 2.77 – 0.5 decreased with the decrease of w/b ratio. The samples with higher
B 1 1.63 2.77 – 0.4 slag content gained higher shrinkage and faster shrinkage develop-
C 1 1.63 2.77 0.01 0.4 ment at the first 28 days. While the addition of fly ash reduced the
D 1 1.63 2.77 0.02 0.3
free drying shrinkage as the replacement level lower than 50%.
The reduction of w/b ratio led to the shortage of water within
concrete and restrained the development of autogenous shrinkage.
lower than 50% (left side of the red line in Fig. 3) are uncertain due Tongaroonsri and Tangtermsirikul [31] placed samples in a cham-
to the lack of experimental results. However, these parts of contour ber with controlled temperature (28 ± 1 °C) and relative humidity
lines were also plotted just to show the variation trend. In many (75 ± 5%). They obtained an opposite conclusion that the lower
previous studies [19,25], this method has been confirmed to be a w/b ratio was, the higher value of total shrinkage was within the
valuable tool to determine the optimal mixture proportions for first 28d curing ages. They accounted this to the significant
desired properties of concrete structure. increase of autogenous shrinkage. The lower w/b ratio stopped
In this study, the testing samples were cured under the condi- the migration of curing water into internal small pores, which
tion of 50% relative humidity, which was much lower than that resulted into internal drying (self-desiccation) and increase of
of the standard curing. Samples cured under this condition are autogenous shrinkage. However, the relative humidity of the cur-
much easier to lose the inner water than the standard cured con- ing condition was 50% in this study, much lower than that in the
crete and tend to insufficient hydration water. For fly ash or slag study of Tongaroonsri and Tangtermsirik [31]. For samples with
blended samples, the hydration reaction at early age was slower higher w/b ratio, the increased porosity and pore diameter acceler-
than that of non-additive Portland cement concrete [26]. Therefore, ated the moisture loss and increased capillary pressure within
 X. Hu et al. / Construction and Building Materials 149 (2017) 406–415 409

(A) Mold (B) Specimen

Fig. 2. Pictures of the mold and specimen for cracking time testing.

(A) Batch A (w/b = 0.5) (B) Batch B (w/b = 0.4)

(C) Batch C (w/b = 0.4) with superplasticizer (D) Batch D (w/b = 0.3) with superplasticizer
Fig. 3. Compressive strength contours of different batches at 28 days (MPa).

samples under this extreme drying condition. These may decrease
the water content within samples and therefore increase the
shrinkage, especially drying shrinkage.
Some previous studies [32,33] have shown that the chemical
shrinkage caused by the reaction of slag was higher than that
caused by cement hydration. Moreover, the high reactivity of blast
furnace slag promoted the formation of hydration products and
autogenous shrinkage development. The results of this study
showed that the addition of slag didn’t present any obvious nega-
tive effects on compressive strength of concrete curing under dry-
ing condition, while the shrinkage problem was more severe than
concrete without slag. The hydration rate and degree of fly ash was
much lower than that of cement and slag and the hydration prod-
ucts of cement are needed to activate the pozzolanic reaction of fly
410 X. Hu et al. / Construction and Building Materials 149 (2017) 406–415

450 450

400 1 400
1
2 2
350 350
Free shrinkage strain (10-6)

Free shrinkage strain (10-6)

3
300 4 300 4
5 5
250 6 250 6
7 7
200 200

150 150

100 100
50 50
0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Age (d) Age (d)
(A) Batch A (w/b = 0.5) (B) Batch B (w/b = 0.4)
450 450
400 1 400 1
2 2
350 3 350
Free shrinkage strain (10-6)

3
Free shrinkage strain (10-6)

4 4
300 300
5 5
250 6 6
250
7 7
200 200

150 150

100 100

50 50

0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Age (d) Age (d)
(C) Batch C (w/b = 0.4) with superplasticizer (D) Batch D (w/b = 0.3) with superplasticizer
Fig. 4. Development of drying shrinkage with testing time.

ash. The replacement of fly ash to cement inhibited the hydration shrinkage, while that with 50% cement and 50% fly ash showed
reaction at the early age and decreased the autogenous shrinkage the lowest. However, the maximum shrinkage value in Fig. 5 (C)
of concrete caused by hydration reaction. Moreover, the refine- placed in the mixture point of 75% cement and 25% slag. When less
ment effects and pozzolanic reaction of fly ash at later curing time than 50% replacement, fly ash decreased the measured shrinkage
improved pore structure and resistance to shrinkage of concrete value. After that, the drying shrinkage slightly increased as the
[34]. The higher elasticity modulus of cement concrete incorpo- content of fly ash continuously increased. The results are in agree-
rated with fly ash also restrained the development of shrinkage ment with previous study [37]. Another important information
and significantly decreased the drying shrinkage [35]. which can be obtained from the contour plots is that the content
Comparing the results of batch B and C, the polycarboxylate of slag in cement-slag-fly ash system governs the total drying
superplasticizer obviously reduced the free drying shrinkage of shrinkage of samples. Fixing the slag content in contour plots,
concrete samples. The shrinkages of batch B samples were 20– the variation of cement and fly ash content showed almost no
30% higher than that of batch C, especially for the samples blended influences on the value of total drying shrinkage.
with slag. The reduction of drying shrinkage from the addition of
superplasticizer may be attributed to the fact that superplasticizer 3.3. Cracking resistance
is efficient in dispersing the cement pastes particles in water and
increasing hydration rate and leading to higher strength built up The strains vs. testing time curves obtained from the ring test
and resistance to drying shrinkage [36]. It can be seen from the experiment are shown in Fig. 6. The data acquisition was started
results that the superplasticizer used in this study was more effec- from shifting the samples into drying condition and taking off
tive to slag blended samples due to the higher hydration rate of the outer ring of the mold, in this study 1d after mixing. The slight
slag than fly ash at early curing age. expansion on the initial testing period of time may be the result of
From Fig. 5 (A), (B) and (D), it can be known that the mixture the early hydration heat and adaption of samples to environment
with 50% cement and 50% slag obtained the highest drying and testing equipment. After that, the strain developed and gradu-
 X. Hu et al. / Construction and Building Materials 149 (2017) 406–415
(A) Batch A (w/b = 0.5)
(C) Batch C (w/b = 0.4) with superplasticizer (D) Batch D (w/b = 0.3) with superplasticizer
Fig. 5. Drying shrinkage contours of different batches at 28 days (106).
ally increased with testing time until a sudden change, which cor-
responds to the cracking of concrete. After that, the strains of sam-
ples kept unchanged due to the formation of inner cracking. A
comparison of free drying shrinkage and the restrained ring strain
showed that the restrained ring strain was less than the free drying
shrinkage due to the restriction of steel ring to the development of
concrete shrinkage. The steel ring strain can be considered as the
portion of the free drying shrinkage occurring under the condition
of restriction. Lomboy et al. [38] comparatively analyzed the
shrinkage performance evaluated by prism length test, free and
restrained ring tests. They found that the development trends of
shrinkage strain were similar for these three methods. The shrink-
age strain rate from the results of free ring test was the highest, fol-
lowed by the prism length test. The restrained ring test gave the
lowest value of shrinkage strain rate.
The cracking strain at the sudden changing point was consid-
ered closely related to the cracking resistance of the concrete. It
is influenced by compressive strength, elasticity modulus and
creep properties of concretes. Comparing the cracking strain of
concrete samples shown in Table 4, the decrease of w/b ratio
(batch A to B) tends to decrease the cracking strain of concrete
samples. The decrease in w/b ratio is mostly accompanied by the
increase of compressive strength and improvement of pore struc-
ture, which may increase the ultimate cracking stress of concrete.
However, the weakened of stress relaxation effects from lower w/b
ratio resulted into the decrease of cracking strain [31]. The use of
superplasticizer in the samples of batch C and D significantly
increased the cracking strain. It has been confirmed that the
411
(B) Batch B (w/b = 0.4)
employment of superplasticizer improved the workability of con-
crete and increased the modulus of elasticity value up to 30%
[39]. The enhancement of stiffness decreased the rate of strain
development, recorded as cracking strain rate, and then increased
the value of cracking strain [31]. The addition of slag and fly ash
seems to affect the cracking strain in different ways. Slag decreased
the cracking strain while it was increased with fly ash replacement.
The flexibility of concrete was improved by fly ash while the addi-
tion of slag made concrete more fragile.
Based on the value of cracking strain enet (lm) measured by the
ring test, a general relation was applied to link the shrinkage strain
value and curing age t (day), which can be also applied for the dry-
ing shrinkage measured by length measurement [24].
pffiffi
enet ¼ a t þ k ð2Þ
According to the formula above, a linear relation can be
obtained between enet and square root of curing age, and the slope
of this linear relation a (lm/day1/2) was defined as shrinkage strain
rate. In ASTM C 1581, this method has been standardized to esti-
mate the shrinkage performance and cracking resistance of ring
concrete samples. The development rate of stress calculated by this
method is also introduced in the standards.
The value of shrinkage strain rate in formula (2) was calculated
according to the results of cracking resistance measurement (ring
test). From the results shown in Table 5, the shrinkage strain rate
of different samples showed a totally different trend compare with
that of cracking strain shown in Table 4. The decrease of w/b ratio
412 X. Hu et al. / Construction and Building Materials 149 (2017) 406–415

40 40
20 20

Steel ring strain (10-6)

0 0 1
Steel ring strain (10-6)

4
-20 3 -20
-40 7 -40 3
-60 4 -60 6
-80 -80 5
5
-100 -100 2
-120 6 2 7
1 -120
-140 -140
-160 -160
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Age (d) Age (d)
(A) Batch A (w/b = 0.5) (B) Batch B (w/b = 0.4)

40 40
20 20
6
0 0
Steel ring strain (10-6)
Steel ring strain ( 10-6)

4 7
-20 -20
3 4
-40 -40 6
-60 7 5 -60
3
-80 -80 1
-100 -100
2
-120 -120 2
1
-140 -140 5
-160 -160
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Age (d) Age (d)
(C) Batch C (w/b = 0.4) with superplasticizer (D) Batch D (w/b = 0.4) with superplasticizer

Fig. 6. Steel ring strain of different batches with curing age under dry condition.

Table 4
Cracking strain of different concrete mixtures (106).

Batch 1 2 3 4 5 6 7
A 97.0 119.0 73.0 98.5 99.0 107.0 86.0
B 82.0 107.0 76.0 88.0 117.5 97.0 106.0
C 121.0 –* 90.0 80.0 113.0 102.0 112.0
D 157.0 107.0 124.0 130.0 138.0 124.0 129.0
*
The cracking time of this sample exceeded the testing time.

Table 5
Regressed a obtained by the steel ring method (106/day1/2).

Batch 1 2 3 4 5 6 7
A 20.1 16.6 21.2 21.5 20.1 13.9 21.0
B 28.7 28.1 22.0 28.5 27.1 24.4 32.2
C 19.2 14.1 18.3 15.6 19.6 11.7 19.8
D 22.8 19.5 20.8 19.4 17.6 17.4 19.1

from 0.5 to 0.4 significantly increased the strain rate of samples, therefore, for samples with lower strain rate, cracks have more
while the incorporation of superplasticizer slightly decreased the time to seek the least resistance path and showed higher cracking
value of shrinkage strain rate. Tongaroonsri and Tangtermsirikul strain. Mostly, the cracks of low strain rate samples were formed
[31] studied the influences of strain rate on cracking strain and within paste and interface between paste and coarse aggregate
found that the higher shrinkage strain rate resulted into lower and rougher fracture surfaces were obtained.
cracking strain due to the different fracture behavior of concrete. Generally, the cracking time of samples were adopted to esti-
The cracking velocity of concrete was increased with strain rate, mate the cracking resistance of concrete. Fig. 7 shows the contour
 X. Hu et al. / Construction and Building Materials 149 (2017) 406–415
(A) Batch A (w/b = 0.5)
(C) Batch C (w/b = 0.4) with superplasticizer
Fig. 7. Cracking time contours of different batches.
plots of cracking time of concrete with different mix proportion.
The cracking time of concrete was almost linearly decreased with
the increase of slag content, while the addition of fly ash gradually
enhanced the cracking resistance of concrete when total replace-
ment level lower than 50%. The incorporation of superplasticizer
and increase of w/b ratio both increased the cracking resistance
of samples.
It can be summarized from many previous studies [40,41] that
many factors including free drying shrinkage, shrinkage develop-
ing rate, creep properties, elasticity modulus, testing method for
cracking resistance (free or restrained) and tensile strength all
had influences on cracking resistance of concrete. Among these,
the shrinkage properties were considered to be the most important
factor controlling the cracking of concrete [35]. According to the
results of drying shrinkage and cracking resistance, the samples
with 50% slag replacement showed higher shrinkage value and
shorter cracking time than any other samples. The addition of
50% fly ash resulted in the lowest shrinkage value and longest time
before cracking. For slag blended samples, the formation of AFt at
early ages produced expansive stress and may lower the cracking
risk [7]. However, the higher drying shrinkage strain and smaller
creep coefficient [8] resulted into lower cracking resistance of slag
blended concrete. Meanwhile, the lower tensile strength, compara-
ble elasticity modulus and lower ultimate stress-strength ratio
comparing to non-slag concrete also increase the risk of cracking
[42]. Massive studies have shown that the decrease of hydration
413
(B) Batch B (w/b = 0.4)
(D) Batch D (w/b = 0.3) with superplasticizer
heat and increase of flexibility and creep of fly ash blended con-
crete made the fly ash very attractive when considering both
shrinkage reduction and cracking resistance enhancement [41,43].
Superplasticizer, as a surfactant, is effective for the decrease of
internal surface tension of capillary pore and drying shrinkage of
the concrete. The addition of superplasticizer improved the work-
ability of cement-based materials and may improve the pore struc-
ture of concrete, which potentially lowers the risk of cracking [22].
When comparing the 4 cracking contour plots in Fig. 7, it can be
seen that the distances between every two contour lines in Fig. 7
(A) and (B) were identical with each other. The influences of slag
and fly ash on cracking resistance of concrete have not much to
do with the w/b ratio of materials. While for concrete with super-
plasticizer addition (batch C and D), the cracking performance of
samples was more sensitive. That is to say, slag and fly ash replace-
ment shows more influences on the cracking resistance of samples
when superplasticizer is applied.
Theoretically, the higher the drying shrinkage value is, the ear-
lier the concrete cracking initiates. When other factors are ignored,
the drying shrinkage and cracking resistance of concrete may be
reversely correlated. Fig. 8 shows the relationship between drying
shrinkage and cracking time of concrete samples in this study. The
relationship between these two parameters was not that obvious
due to the effects of other factors. However, we can still see from
the figure that the increase of free shrinkage decreased the crack-
ing resistance of concrete. Under some condition, the improvement
414 X. Hu et al. / Construction and Building Materials 149 (2017) 406–415
Fig. 8. Free drying shrinkage as a function of cracking time.
of creeping performance may reduce the cracking risk of high
shrinkage concrete samples, while the decrease of tensile strength
leads to lower cracking resistance [44]. In the study of Akkaya et al.
[45], the cracking resistance of cement-fly ash-silica fume was
more affected by compressive strength while the total shrinkage
of concrete almost not varied with different proportions of ternary
binder.
According to experiment results and analysis in this study, from
the point of cracking resistance enhancement, the content of slag
should be reduced while relatively higher replacement level of
fly ash is recommended. While, the reduction of compressive
strength due to the addition of fly ash is greater than slag. The fac-
torial design method provide a possibility to comprehensively con-
sider different performance and requirements on concrete
structure when designing the mixture proportion. However, series
of properties and performance of concrete have to take into
account when determining the mix proportion of concrete. Facto-
rial design method, no matter for ternary cementitious materials
system or more than 3 factors studied, can be a good approach
to decide an appropriate mix proportion to meet different request
practically.
4. Conclusions
Based on the results from this study, the following conclusions
can be drawn:
(1) The compressive strength of concrete cured under drying
condition was decreased as the partly replacement of
cement by slag and/or fly ash. The influences of SCMs on
compressive strength of concrete were more significant
when the replacement level was higher than around 25%.
(2) Slag increased the drying shrinkage of concrete while the
drying shrinkage was decreased by the addition of fly ash.
The addition of superplasticizer significantly decreased the
drying shrinkage of concrete. The total drying shrinkage of
concrete was mainly controlled by the content of slag in
mixture proportion.
(3) The increase of w/b ratio and addition of superplasticizer
increased the cracking strain of concrete, while decreased
the strain rate obtained from ring test. The cracking strain
of concrete was reversely related to the strain rate.
(4) The cracking resistance of concrete is influenced by many
factors, while the free shrinkage is the most important one.
The addition of slag and fly ash showed different influences
on cracking time of concrete in dry environment, which
mostly comes from the different free drying shrinkage, creep
and mechanical properties.
(5) Based on the results of factorial design method, in ternary
system of Portland cement-slag-fly ash, the addition of fly
ash decreased the compressive strength of concrete, while
it’s preferable to reduce the replacement level of slag when
considering the drying shrinkage reduction. .
Acknowledgements
The financial supports from the National Natural Science Foun-
dation of China under contract Nos. 51638008 and 51461135001
and Chinese Scholarship Council (CSC) are gratefully
acknowledged.
References
[1] G. De Schutter, Durability of marine concrete structures damaged by early age
thermal cracking//International RILEM Workshop on Life Prediction and Aging
Management of Concrete Structures, Cannes, Oct. 2000, RILEM Proceedings
PRO16 (Ed. D. Naus). 2000: 185–194.
[2] K.J. Wang, D.C. Jansen, S.P. Shah, et al., Permeability study of cracked concrete,
Cem. Concr. Res. 27 (3) (1997) 381–393.
[3] X.M. Zhou, J.R. Slater, S.E. Wavell, et al., Effects of PFA and GGBS on early-ages
engineering properties of Portland cement systems, J. Adv. Concr. Technol. 10
(2) (2012) 74–85.
[4] T. Seo, M. Lee, C. Choi, et al., Properties of drying shrinkage cracking of concrete
containing fly ash as partial replacement of fine aggregate, Mag. Concr. Res. 62
(6) (2010) 427–433.
[5] D.T. Nguyen, R. Sahamitmongkol, L.N. Trong, et al., Prediction of shrinkage
cracking age of concrete with and without expansive additive, Songklanakarin
J. Sci. Technol. 32 (05) (2010).
[6] K.H. Obla, R.L. Hill, M.D.A. Thomas, et al., Properties of concrete containing
ultra-fine fly ash, ACI Mater. J. 100 (5) (2003) 426–433.
[7] J.Y. Li, Y. Yao, A study on creep and drying shrinkage of high performance
concrete, Cem. Concr. Res. 31 (8) (2001) 1203–1206.
[8] T. Kanda, H. Momose, K. Imamoto, Shrinkage cracking resistance of blast
furnace slag blended cement concrete-Influencing factors and enhancing
measures, J. Adv. Concr. Technol. 13 (1) (2015) 1–14.
[9] T. Aly, J.G. Sanjayan, Factors contributing to early age shrinkage cracking of
slag concretes subjected to 7-days moist curing, Mater. Struct. 41 (4) (2008)
633–642.
[10] X.H. Wang, K.J. Wang, F. Bektas, et al., Drying shrinkage of ternary blend
concrete in transportation structures, J. Sustain. Cement-Based Mater. 1 (1–2)
(2012) 56–66.
[11] J.G. Han, H. Fang, K.J. Wang, Design and control shrinkage behavior of high-
strength self-consolidating concrete using shrinkage-reducing admixture and
super-absorbent polymer, J. Sustain. Cement-Based Mater. 3 (3–4) (2014)
182–190.
[12] W. Wongkeo, P. Thongsanitgarn, A. Chaipanich, Compressive strength and
drying shrinkage of fly ash-bottom ash-silica fume multi-blended cement
mortars, Mater. Design (1980–2015) 36 (2012) 655–662.
[13] W.M. Hale, S.F. Freyne, T.D. Bush, et al., Properties of concrete mixtures
containing slag cement and fly ash for use in transportation structures, Constr.
Build. Mater. 22 (9) (2008) 1990–2000.
[14] H. Scheffé, Experiments with mixtures, J. Roy. Stat. Soc.: Ser. B (Methodol.)
(1958) 344–360.
[15] H.M. Wadsworth, Handbook of Statistical Methods for Engineers and Scientist,
McGraw-Hill, New York etc, 1990.
[16] C.J. Shi, D.H. Wang, L.M. Wu, et al., The hydration and microstructure of ultra
high-strength concrete with cement–silica fume–slag binder, Cement Concr.
Compos. 61 (2015) 44–52.
[17] W. Sun, H. Yan, Studies on the quantitative relationships between
compositions of blended cementitious materials and compressive strength of
concrete, J. Southeast Univ. (Natural Science Edition) 33 (4) (2003) 450–453.
[18] Z.G. Shi, C.J. Shi, R. Zhao, et al., Factorial design method for designing ternary
composite cements to mitigate ASR expansion, J. Mater. Civ. Eng. 28 (9) (2016)
04016064.
[19] S. Math, Optimizing SCM proportions to meet multiple performance characters
of ternary concrete mixtures using simplex-centroid design and analysis
techniques, Clemson University, 2013.
[20] P.T. Durdziński, M.B. Haha, S.A. Bernal, et al., Outcomes of the RILEM round
robin on degree of reaction of slag and fly ash in blended cements, Mater.
Struct. 50 (2) (2017) 135.
[21] D.A. Salas, A.D. Ramirez, C.R. Rodríguez, et al., Environmental impacts, life
cycle assessment and potential improvement measures for cement
production: a literature review, J. Clean. Prod. 113 (2016) 114–122.
 X. Hu et al. / Construction and Building Materials 149 (2017) 406–415
[22] T. Hemalatha, A. Ramaswamy, A review on fly ash characteristics-towards
promoting high volume utilization in developing sustainable concrete, J. Clean.
Prod. (2017).
[23] E. Özbay, M. Erdemir, H.I. _ Durmusß, Utilization and efficiency of ground
granulated blast furnace slag on concrete properties–a review, Constr. Build.
Mater. 105 (2016) 423–434.
[24] H.T. See, E.K. Attiogbe, M.A. Matthew, Potential for restrained shrinkage
cracking of concrete and mortar, Cem. Concr. Aggregates 26 (2) (2004) 1–8.
[25] F. Bektas, X. Wang, Statistical mixture design of ternary blends for controlling
ASR, Mag. Concr. Res. 67 (2) (2015) 63–70.
[26] Y. Luan, T. Ishida, T. Nawa, et al., Enhanced model and simulation of hydration
process of blast furnace slag in blended cement, J. Adv. Concr. Technol. 10 (1)
(2012) 1–13.
[27] B.W. Langan, K. Weng, M.A. Ward, Effect of silica fume and fly ash on heat of
hydration of Portland cement, Cem. Concr. Res. 32 (7) (2002) 1045–1051.
[28] A. Xu, S.L. Sarkar, L.O. Nilsson, Effect of fly ash on the microstructure of cement
mortar, Mater. Struct. 26 (7) (1993) 414–424.
[29] O. Boukendakdji, E.H. Kadri, S. Kenai, Effects of granulated blast furnace slag
and superplasticizer type on the fresh properties and compressive strength of
self-compacting concrete, Cement Concr. Compos. 34 (4) (2012) 583–590.
ß . Erdoǧdu, Effect of retempering with superplasticizer admixtures on slump
[30] S
loss and compressive strength of concrete subjected to prolonged mixing,
Cem. Concr. Res. 35 (5) (2005) 907–912.
[31] S. Tongaroonsri, S. Tangtermsirikul, Effect of mineral admixtures and curing
periods on shrinkage and cracking age under restrained condition, Constr.
Build. Mater. 23 (2) (2009) 1050–1056.
[32] A.A. MeloNeto, Estudo da retração em argamassa com cimento de escória
ativada (Shrinkage of alkali-activated slag), Universidade de São Paulo, 2002
[in Portuguese].
[33] E. Douglas, A. Bilodeau, V.M. Malhotra, Properties and durability of alkali-
activated slag concrete, Mater. J. 89 (5) (1992) 509–516.
[34] P. Nath, P. Sarker, Effect of fly ash on the durability properties of high strength
concrete, Proced. Eng. 14 (2011) 1149–1156.
415
[35] K.V. Subramaniam, R. Gromotka, S.P. Shah, et al., Influence of ultrafine fly ash
on the early age response and the shrinkage cracking potential of concrete, J.
Mater. Civ. Eng. 17 (1) (2005) 45–53.
[36] S.H. Alsayed, Influence of superplasticizer, plasticizer, and silica fume on the
drying shrinkage of high-strength concrete subjected to hot-dry field
conditions, Cem. Concr. Res. 28 (10) (1998) 1405–1415.
[37] P. Nath, P.K. Sarker, Effect of mixture proportions on the drying shrinkage and
permeation properties of high strength concrete containing class F fly ash,
KSCE J. Civil Eng. 17 (6) (2013) 1437–1445.
[38] G. Lomboy, K.J. Wang, C.S. Ouyang, Shrinkage and fracture properties of
semiflowable self-consolidating concrete, J. Mater. Civ. Eng. 23 (11) (2010)
1514–1524.
[39] P. Pereira, L. Evangelista, J. De Brito, The effect of superplasticizers on the
mechanical performance of concrete made with fine recycled concrete
aggregates, Cement Concr. Compos. 34 (9) (2012) 1044–1052.
[40] K. Wiegrink, S. Marikunte, S.P. Shah, Shrinkage cracking of high-strength
concrete, Mater. J. 93 (5) (1996) 409–415.
[41] T. Zhang, W.Z. Qin, Tensile creep due to restraining stresses in high-strength
concrete at early ages, Cem. Concr. Res. 36 (2006) 584–591.
[42] T. Kanda, H. Momose, K. Imamoto, et al., Quantitative evaluation of crack
resistance mechanism of blast furnace slag blended cement concrete via
restrained shrinkage stress analysis, J. Adv. Concr. Technol. 13 (9) (2015) 421–
437.
[43] P. Chindaprasirt, U. Rattanasak, Shrinkage behavior of structural foam
lightweight concrete containing glycol compounds and fly ash, Mater. Des.
32 (2) (2011) 723–727.
[44] K.J. Wang, S.M. Schlorholtz, S. Sritharan, et al., Investigation into Shrinkage of
High-Performance Concrete Used for Iowa Bridge Decks and Overlays, 2013.
[45] Y. Akkaya, C. Ouyang, S.P. Shah, Effect of supplementary cementitious
materials on shrinkage and crack development in concrete, Cement Concr.
Compos. 29 (2) (2007) 117–123.