Construction and Building Materials 314 (2022) 125720

Contents lists available at ScienceDirect

Construction and Building Materials

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

Autogenous and total shrinkage of limestone calcined clay cement

(LC3) concretes
Quang Dieu Nguyen a, *, Sumaiya Afroz b, Yingda Zhang b, Taehwan Kim b, Wengui Li a,
Arnaud Castel a
a
School of Civil and Environmental Engineering, University of Technology Sydney (UTS), Sydney, NSW 2007, Australia
b
Centre for Infrastructure Engineering and Safety, School of Civil and Environmental Engineering, University of New South Wales (UNSW), Sydney, NSW 2052,
Australia

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

Keywords: In this study, the developments of autogenous and total shrinkage of limestone calcined clay cement (LC3)
Calcined clay concretes were investigated. Three concrete grades including 25 MPa, 32 MPa and 45 MPa of both LC3 and
Limestone general purpose cement (GPC) concretes were considered. Compressive strength and tensile strength were
LC3
measured until curing of 28 days. In addition, pore size distribution of cementitious pastes was evaluated by
Shrinkage
using nitrogen adsorption. Several models were used to assess their applicability for LC3 concretes in predicting
Pore structure
Low-carbon concrete mechanical properties and shrinkage development. The LC3 concretes showed higher autogenous shrinkage at a
later age up to 100 days due to continuous refinement of the pore structure whilst the development of total
shrinkage was similar between LC3 and OPC concretes. All models underestimated LC3 concrete autogenous
shrinkage and the Bazant B4 model provided the best prediction of total shrinkage development.

1. Introduction environment affecting mostly concrete close to the surface [5].

Limestone calcined clay cement (LC3) has attracted global attention
The time dependent deformations due to volumetric change caused as a promising alternative for ordinary Portland cement (OPC) concrete
by shrinkage can significantly influence the durability and serviceability in the effort to reduce the carbon footprint and environmental impact of
of concrete structures. Under the restrained condition, volumetric the cement and concrete industry [6]. A combination of calcined clay
reduction due to shrinkage can induce tensile stresses and premature and limestone can replace OPC up to 50 wt% in the binder composition,
cracks if these tensile stresses exceed concrete tensile strength [1,2]. leading to a significant reduction of greenhouse gases emitted through
Early age cracking can impact the properties of concrete including the cement production without degradation of the mechanical proper­
microstructure development, mechanical properties development, and ties of concrete at an early age [7,8]. In addition, limestone and clays
structural integrity as well as long-term durability [3]. For instance, suitable for calcining, which are suitable for LC3, can be found in many
aggressive agents such as chloride, carbon dioxide and sulphate ions can places around the world, indicating the potential for widespread LC3
rapidly penetrate through cracks induced by restrained shrinkage, production [9]. The presence of calcined clay and limestone in binder
resulting in a significant reduction in the service life of concrete struc­ created AFm phases of monocarboaluminate and hemicarboaluminate
tures [4]. The shrinkage of cementitious materials after final setting can in microstructure [6,10]. The hydration products, mechanical and
be divided into autogenous shrinkage and drying shrinkage. The com­ durability properties of LC3-based materials were reported in previous
bination of these two shrinkage strains can be denoted as total studies [8,11–13]. Several deterioration mechanisms such as carbon­
shrinkage. Autogenous deformation in concrete after final setting time is ation, alkali-silica reaction, sulphate resistance and long-term rein­
described as a volumetric alteration without any moisture loss to the forcement corrosion have been studied to accelerate the adoption of LC3
external environment, which is dominantly governed by the so-called by the construction industry [13–16]. These results exhibited that LC3
chemical shrinkage and the self-desiccation process. On the other materials can perform comparably or even better than conventional
hand, drying shrinkage is caused by the water loss to the external concretes. However, as previously discussed, the performance of LC3

* Corresponding author.
E-mail address: [email protected] (Q.D. Nguyen).

https://doi.org/10.1016/j.conbuildmat.2021.125720
Received 7 July 2021; Received in revised form 9 November 2021; Accepted 14 November 2021
Available online 20 November 2021
0950-0618/© 2021 Elsevier Ltd. All rights reserved.
Q.D. Nguyen et al. Construction and Building Materials 314 (2022) 125720

concretes can be strongly affected by premature cracks induced by method are shown in Fig. 1 respectively. Coarse aggregate was basalt
shrinkage. In addition, the development of shrinkage can be different with a maximum nominal size of 10 mm and water absorption of 1.08 %.
depending on the type of supplementary cementitious materials (SCMs) Sydney sand with water absorption of 3.5 % was used as fine aggregate.
used. For example, fly ash contributed to decreasing the autogenous Basalt and Sydney sand have specific gravities of 2.8 and 2.65 respec­
shrinkage whilst blast furnace slag and silica fume lead to an increase in tively. The grading curves of basalt aggregate and Sydney sand were
autogenous deformation [17–19]. Some studies stated that fly ash, blast reported in a previous study of the authors [32]. All aggregates were
furnace slag and silica fume produced lower drying shrinkage [20,21], oven-dried for at least 24 hrs at 105 ◦ C to achieve constant mass to
while Mokarem et al. [22] revealed that mixtures incorporating SCMs remove any moisture before concrete casting.
(fly ash, microsilica and blast furnace slag) presented higher drying Six concrete mixtures were fabricated to investigate the effect of
shrinkage strains than that of plain cement mixes. This can be attributed calcined clay and limestone (LC3) on the concrete properties including
to numerous factors of SCMs such as chemical, physical and mineral­ mechanical properties and shrinkage development. GPC and LC3
ogical properties. When using SCMs, pozzolanic reactions at later age denoted concretes containing only GP cement and the combination of
can lead to a significant refinement of the pore structure, affecting the GP cement, calcined clay and limestone as binder respectively. In LC3
shrinkage of cement paste [17,22,23]. The shrinkage development of concretes, 44 wt% of GP cement was substituted by the combination of
limestone-rich cement paste and concrete was significantly dependent calcined clay and limestone. In addition, in accordance with Australian
on the amount and chemical–mineralogical properties of the limestone Standard AS 3972, the composition of the GP cement used was 90 wt%
[24–26]. of OPC, around 5 wt% of limestone, 2 wt% of gypsum and 3 wt% of other
This study aims to assess the effect of calcined clay and limestone mineral additions. As a result, the actual proportion of OPC in the LC3
(LC3) on the development of autogenous, drying and total shrinkage binder was approximately 50 wt% of the total binder, which is consis­
strains of concrete. Dhandapani et al. [27] reported a higher autogenous tent with the replacement percentage of previous studies [8,27]. The
shrinkage and similar total shrinkage of LC3 concretes in comparison calcined clay and limestone content in LC3 binder were 30 wt% and
with conventional concretes but no explanation was proposed. Another 16.8 wt% respectively. It should be noted that limestone proportion was
study indicated that LC3 paste showed a similar autogenous shrinkage to the combination of Oyma limestone and limestone in GP cement.
reference cement paste after 28 days of hydration [10]. A replacement The mix design of the concrete used is presented in Table 2 with
rate at approximately 50 wt% binder by calcined clay and limestone was aggregates in surface saturated dry (SSD) condition. The study evaluated
used in this study. Three different concrete grades including 25 MPa, 32 three different concrete grades including 25 MPa, 32 MPa and 45 MPa
MPa and 45 MPa were fabricated by altering water/binder ratios for compressive strength after 28 days of standard water curing according to
both LC3 and reference plain cement concretes. The mechanical prop­ AS 1012.8.1 and ASTM C39 [33,34]. The numbers 25, 32 and 45 added
erties including compressive strength and tensile strength were evalu­ at the end of GPC and LC3 in Table 2 denoted the concrete grades of 25
ated for up to 28 days. Pore size distribution of paste was examined MPa, 32 MPa and 45 MPa respectively. All mixes were designed based
nitrogen adsorption to investigate the effect of calcined clay and lime­ on the same concrete volume and total binder content. Water/binder
stone on shrinkage strains. Finally, several standards and empirical ratio and coarse/fine aggregate volumetric ratio were also fixed within
models were reassessed for their applicability of both mechanical the same concrete grade. After oven-drying, concrete aggregates were
properties and shrinkage development in LC3 concretes. cooled down and SSD condition was obtained by adding calculated SSD
water prior to concrete casting. Superplasticiser (0.5 % of binder con­
2. Materials and mix design tent) was added during mixing to obtain acceptable workability. Pre­
vious studies on OPC concrete indicated that a dosage of superplasticiser
The binder used in this study includes general purpose (GP) cement, less than 0.6 % of the binder produced insignificant effect on shrinkage
calcined clay, and limestone. The characteristics of GP cement complies development [35–37]. After mixing, fresh concrete was poured into
with Australian Standard AS 3972 [28], which is presented in a previous
study [29]. The calcined clay was produced by rotary kiln calcination
and acquired as an industrial product. Hence, the kaolinite content in
the raw clay prior to calcination is unknown to the authors. However,
XRD-Rietveld analysis [30] showed that the calcined clay had 78.8 wt%
of amorphous content, which classifies it as high-grade calcined clay
[31]. The mineral composition of calcined clay was dickite, mullite,
kaolinite, quartz and anatase has been reported in a previous paper by
the authors [31]. The limestone was supplied by Omya Australia with
the brand name as Omyacarb containing more than 95 wt% CaCO3.
Table 1 summarized the chemical composition of the GP cement,
calcined clay and limestone determined by X-ray fluorescence (XRF).
Their particle size distribution measured using the laser diffraction

Table 1
Chemical composition of GP cement, calcined clay and limestone.
Chemical composition (wt%) GP cement Calcined clay Limestone

SiO2 19.74 48.15 1.10

Al2O3 4.70 41.63 0.24
Fe2O3 2.98 2.27 0.17
CaO 64.62 0.12 54.84
MgO 1.48 0.09 1.53
Na2O 0.21 0.32 0.04
K2O 0.64 0.05 0.01
TiO2 0.31 3.39 –
SO3 2.24 0.09 0.03
Fig. 1. Particle size distribution curves of GP cement, calcined clay
Loss on ignition (LOI) 3.18 3.21 43.11
and limestone.

2
Q.D. Nguyen et al. Construction and Building Materials 314 (2022) 125720

Table 2 The reference lengths of the specimens were measured within a few
Mix composition details of concretes. minutes after demoulding for total shrinkage, after wrapping specimens
Materials (kg/ GPC- LC3- GPC- LC3- GPC- LC3- for autogenous shrinkage, and after removing specimens from a water
m3) 25 25 32 32 45 45 bath for drying shrinkage. It should be noted that total and autogenous
Coarse aggregate 1059 1046 1025 1010 1221 1202 shrinkage were measured after demoulding whilst drying shrinkage was
Fine aggregate 866 856 839 827 621 611 determined after 7 days of water immersion. All shrinkage specimens
Total binder 310 310 360 360 388 388 were then stored in an environmental chamber at a fixed temperature of
GP cement 310 173.6 360 201.6 388 217.3 23 ± 2 ◦ C and RH of 50 ± 3 %. All shrinkage strains were then measured
Calcined clay 0 93.0 0 108.0 0 116.4
Limestone 0 43.4 0 50.4 0 54.3
up to 100 days. Various models and standards [41–44] consider that
Water 173.6 173.6 176.4 176.4 174.5 174.5 total shrinkage is the sum of autogenous shrinkage and drying shrinkage
Water/binder 0.56 0.56 0.49 0.49 0.45 0.45 as a simplified approach. Nonetheless, the concepts of total, autogenous
ratio and drying shrinkage of concrete in this study were discussed in more
details in the results and discussion section.
different moulds and covered using appropriate lids to prevent surface
3.3. Pore size distribution of cementitious pastes
moisture loss. The concrete casting procedure was conducted in a
controlled room with the temperature at 23 ± 2 ◦ C and relative humidity
Plain cement and LC3 pastes were fabricated to investigate the
of 50 ± 3 %. Hardened concrete was demoulded after 1 day and fol­
microstructure development using the paste mix compositions shown in
lowed different curing and pre-condition regimes complying with the
Table 3. Nitrogen adsorption technique was conducted to evaluate pore
different standard protocols detailed in next section.
size distribution (PSD). The cementitious pastes were fabricated and
cured in sealed condition until being tested at 28 days and 90 days.
3. Experimental programs
Before conducting the experiment, the paste samples were immersed in
isopropanol for 1 h to stop the hydration, followed by 30 min of drying
3.1. Mechanical properties
at 40 ◦ C [45]. The specimens were then ground by a ball mill and passed
850 µm sieve size. The degassing stage was conducted at a temperature
Standard concrete cylinders with 200 mm in height and 100 mm in
of 40 ◦ C for 16 hrs [46]. A Novatouch LX4 from Quantachrome Instru­
diameter were used to assess the mechanical properties including
ment was used and PSDs of samples were calculated by BJH (Barrett-
compressive strength and indirect tensile strength. Specimens were
Joyner-Halenda) method using the Kelvin model of pore filling [47]. The
continuously cured in lime water up to 28 days at 23 ± 2 ◦ C according to
desorption branches of the isotherms were selected to analyze the PSDs
Australian Standard AS 1012.8.1 [33]. Compressive strength and indi­
based on the previous studies [48,49]. The mesopore range with 2–50
rect tensile strength of LC3 concrete specimens were measured at 1, 3, 7,
nm of pore width was the most accurate pore size range as the BJH
14, 21 and 28 days after casting whilst GPC concretes compressive
method was employ to determine pore size distribution in this range
strength was measured only at 28 days as reference values. Compressive
[50].
strength and indirect tensile strength test protocols were following
ASTM C39 and ASTM C496 respectively [34,38].
4. Results and discussion
3.2. Shrinkage measurement
4.1. Mechanical properties
Concrete prisms with a dimension of 280 mm ×75 mm ×75 mm were
The compressive strength development of LC3 concretes for three
fabricated to monitor the shrinkage development up to 100 days. The
different grades is presented in Fig. 2. GPC concretes (GPC-25, GPC-32
shrinkage strain was measured by using a digital gauge meter with an
and GPC-45) compressive strength was measured only at 28 days.
accuracy of 1 µm/m. Australian Standard AS 1012.13 [39] specifies that
Fig. 2 shows that the compressive strength of LC3 concrete was com­
concrete drying shrinkage can be determined by measuring the concrete
parable to or even higher (in the case of 32 MPa concrete grade) than
prisms shrinkage following 7 days of lime water curing. However, the
that of GPC concrete after 28 days of water curing. The results were
development of autogenous shrinkage of LC3 concrete after 7 days can
consistent with the previous study using a 50 wt% binder replaced by a
be significant due to the occurrence of pozzolanic reactions [23].
combination of calcined clay and limestone [7,8,31]. The compressive
Therefore, an alternative approach was conducted to evaluate concrete
strength of LC3 concrete considerably increased from day 1 to day 7. The
shrinkage. After 1 day, concrete prisms were divided into three different
values at 7 days achieved more than 70 % of the compressive strength at
categories as follows:
28 days. Fig. 3 shows the increase of tensile strength of LC3 concretes up
to 28 days of water curing. The tensile strength (fct) of concrete was
• Category 1 denoted as “autogenous shrinkage” included prisms of
determined using: fct = 0.9 × fct,sp where fct,sp is the measured splitting
which all surfaces were covered by self-adhesive water-proof
tensile strength of concrete. A similar development trend was observed
aluminum foil after demoulding to prevent moisture loss during the
between compressive strength and tensile strength of LC3 concretes with
experimental period. “Simplified procedure for the determination of
70 % of the 28 days tensile strength obtained at 7 days.
autogenous shrinkage” in European Standard EN 12390-16 [40] was
utilized as a standardised protocol. The weight of autogenous
shrinkage specimens has been monitored. The variation of weight
was negligible during the whole test duration, indicating the tight­
ness of the self-adhesive aluminum foil.
• Category 2 denoted as “total shrinkage” included unsealed speci­ Table 3
mens exposed to a controlled environment after demoulding. Mix composition details of cementitious pastes.
• Category 3 denoted as “drying shrinkage” included specimens that
GPC LC3
were stored in a lime water bath for additional 7 days after
demoulding before the first shrinkage measurement as described in GP cement (wt%) 100 56
Calcined clay (wt%) 0 30
AS 1012.13 [39].
Limestone (wt%) 0 14
Water/binder ratio 0.4 0.4

3
Q.D. Nguyen et al. Construction and Building Materials 314 (2022) 125720

Fig. 2. Compressive strength development up to 28 days of LC3 concretes and GPC concrete compressive strength at 28 days as reference.

4.2. Autogenous shrinkage

The average autogenous shrinkage strains up to 100 days after

demoulding of the 6 mixes of LC3 and GPC concretes is shown in Fig. 4.
The development of autogenous shrinkage from day 1 to day 7 is
enlarged and presented in the inset of Fig. 4. The negative and positive
values in Fig. 4 denote shrinking and expansion (or swelling) of con­
cretes respectively. Except for GPC-32 concrete, the expansion of sealed
concrete prisms was observed in both LC3 and GPC concretes from day 1
to day 5 after demoulding. The expansion behaviour was more
remarkable in LC3 concretes than in GPC concretes with the highest
value of 38 µm/m in LC3-45 concrete after one day. Early age expansion
of cementitious materials with medium to high water/binder ratios were
reported in previous research [3,51] and can be attributed to several
phenomena. In this study, the swelling peaks of GPC concretes can be
explained by the formation or development of different hydration
products so-called structural swelling, such as portlandite and ettringite
[51,52]. However, the autogenous swelling mechanism of LC3 concretes
at early age has been unresolved. The initial autogenous expansion up to
4 days was reported in a previous study [10]. Therefore, further study is
recommended to investigate the influence of calcined clay and limestone
on the expansion of LC3 concretes at early age.
In long-term autogenous shrinkage up to 100 days, LC3 concretes
showed noticeably higher shrinkage strains in comparison to GPC con­
cretes. Specifically, at 100 days, GPC-25, GPC-32 and GPC-45 concretes
Fig. 3. Tensile strength development up to 28 days of LC3 concretes. presented autogenous shrinkage values of − 67, − 123 and − 178 µm/m
respectively whilst LC3 concretes exhibited higher shrinkage strains
ranging from − 235 to − 278 µm/m. Decreasing the water/binder ratios
considerably increased the autogenous shrinkage of GPC concretes with
more than 50 µm/m in difference within the 3 concrete mixes at 100

4
Q.D. Nguyen et al. Construction and Building Materials 314 (2022) 125720

Fig. 4. Autogenous shrinkage development of concretes up to 100 days. Inset: Enlargement of autogenous shrinkage development from demoulding to 7 days.

days. Decreasing the water/binder ratios seemed to have a lower in­ Autogenous shrinkages for all GPC concrete mixes were approximately
fluence on LC3 concrete autogenous shrinkage. To be specific, LC3-25 stabilized after 50 days whilst the stabilization was observed at around
and LC3-32 concretes showed relatively similar autogenous shrinkage 90 days in LC3 concretes. The increase in autogenous shrinkage
developments from 6 days to 50 days and LC3-32 concrete exhibited 10 observed on LC3 concretes is consistent with the previous study where
µm/m shrinkage strain higher than LC3-25 at 100 days. LC3-45 dis­ the autogenous shrinkage value of LC3 concrete was measured for up to
played the highest shrinkage value of all concrete mixes after 10 days one year [27]. The higher autogenous shrinkage can be attributed to a
and obtain around − 275 µm/m at 100 days. Moreover, the autogenous long-term hydration process in LC3 system. Decreasing portlandite
shrinkage developed at a decreasing rate over the testing period. content and increasing AFm phases up to 90 days were reported in

Fig. 5. Pore size distribution of pastes determined by nitrogen adsorption and BJH method.

5
Q.D. Nguyen et al. Construction and Building Materials 314 (2022) 125720

previous studies [7,53]. This observation indicates continuous reactions 4.3. Total and drying shrinkage
between cement, calcined clay and limestone occurring in the micro­
structure, resulting in a long-term autogenous shrinkage. Moreover, the Fig. 6 presents the development of the total shrinkage of all concretes
different autogenous deformation between GPC and LC3 concrete can be after demoulding. No swelling peak was observed in all concrete mixes,
explained by the pore structure refinement with the addition of calcined indicating that the shrinkage due to drying could balance (or mask) the
clay and limestone. The PSD of GPC and LC3 pastes at 28 days and 90 expansion observed in the autogenous shrinkage at an early age. For 25
days after casting is exhibited in Fig. 5. The pore size ranges can be MPa and 32 MPa, LC3 concretes exhibited higher shrinkage than GPC
classified as: from 0.5 nm to 2.5 nm (micropores), 2.5 nm to 10 nm concretes up to 20 days. Then GPC concretes shrinkage became superior
(small capillary pores or gel pores), 10 nm to 50 nm (medium capillary to that of LC3 concretes from 20 days to 80 days. LC3 and GPC concretes
pores) and 50 nm to 10 µm (large capillary pores) [54] and these ranges with 25 MPa and 32 MPa showed relatively equivalent shrinkage strains
are also integrated into Fig. 5. Micropores, gel pores and medium at 100 days. LC3-45 concrete displayed a higher total shrinkage than
capillary pores can influence the autogenous shrinkage of cementitious that of GPC-45 concrete, but GPC-45 concrete started to slightly increase
materials. Indeed, pore size in these ranges (less than 50 nm) can after 80 days. Consequently, GPC-45 and LC3-45 concretes achieved
generate capillary pressure and considerably increase the self- similar total shrinkage about − 600 µm/m at 100 days. The increase in
desiccation process in the microstructure [55]. Due to the limitation of binder content (Table 2) could explain the increase in total shrinkage
the BJH method in nitrogen adsorption (pore size ranging from 2 nm to strains between 25 MPa and 45 MPa concrete. The total shrinkage of
50 nm) as mentioned in section 3.3, the gel pores and medium capillary concrete specimens at a constant temperature and relative humidity is
pores are primarily examined in this study. From Fig. 5, LC3 specimens often considered as the combination of autogenous shrinkage and drying
exhibited a considerably higher amount of gel pores and medium shrinkage, in various studies and by standards [40–44]. However, it
capillary pores than GPC pastes at both 28 days and 90 days of water should be noted that the “drying shrinkage” strains of concretes are not
curing. The self-desiccation process which is considered as the main the direct subtraction of autogenous shrinkage (Category 1) from total
cause of autogenous shrinkage of cementitious materials is based on shrinkage (Category 2) in this study. Specifically, the autogenous
capillary theory. Specifically, the finer pore structure due to the pres­ shrinkage determined in Fig. 4 under completely sealed conditions can
ence of LC3 creates a lower equilibrium of internal relative humidity be expected higher than the autogenous shrinkage in unsealed shrinkage
(based on Kelvin-Laplace law) and a higher capillary pressure in the pore specimens. This can be explained by the water loss during the drying
fluid, leading to the higher autogenous shrinkage in LC3 concretes. process reducing the hydration process, resulting in lower autogenous
Moreover, LC3 paste after 90 days had a significantly higher cumulative deformation occurring in total shrinkage prisms. Both LC3 and GPC
pore volume in gel pores range (2.5–10 nm) and a lower volume in concrete specimens seem to experience this reduction of autogenous
medium capillary pores (10–50 nm) in comparison with 28 days of shrinkage. Nevertheless, Fig. 4 obviously exhibits a higher autogenous
water curing, indicating the continuous pore structure refinement and shrinkage for LC3 specimens than that obtained from GPC specimens but
hydration process in LC3 specimens up to 90 days. This is consistent with similar total shrinkage values were observed in Fig. 6 for LC3 and GPC
the continuous development of autogenous shrinkage LC3 concretes up concretes at the same water/binder ratio. It can be deduced that LC3
to 90 days as presented in Fig. 4. The GPC paste after 90-day water concretes generate lower drying shrinkage strains compared to the GPC
curing experienced only a slight increase in cumulative pore volume in concretes. It can be attributed to the finer pore structure of LC3 system in
the pore size range 2.5 nm to 20 nm compared to the specimens at 28 comparison with conventional concrete, leading to a denser micro­
days. structure and then reducing the water loss from LC3 specimens [56,57].
The finer pore structure of LC3 paste in comparison with GPC paste can
be well-observed in Fig. 5 as discussed in section 4.2.
The concretes drying shrinkage after 7 days immersed in lime water

Fig. 6. Total shrinkage development of concretes up to 100 days.

6
Q.D. Nguyen et al. Construction and Building Materials 314 (2022) 125720

(Category 3) is shown in Fig. 7. GPC-25 and LC3-25 exhibited indistin­ and b are constants relating to cement type and fc28 is the average
guishable shrinkage strains over the testing period. LC3-32 concrete compressive strength of concrete at 28 days. As reported in [58], a and b
showed higher shrinkage strains than that of GPC-32 concrete up to 30 can range from a = 0.05 to 9.25 and b = 0.67 to 0.98. The specific values
days, and then comparable values were obtained for the two concretes. of a = 4.0, b = 0.85 and a = 2.3, b = 0.92 were assigned for ASTM Type I
Drying shrinkage of 45 MPa concretes also showed a similar trend to cement (ordinary Portland cement) and Type III cement (high early
LC3-32 and GPC-32 concretes with a higher drying shrinkage for LC3-45 strength cement) respectively [58]. Furthermore, a and b coefficients,
up to 40 days. All concrete drying shrinkage stabilised after 90 days. considered as dependent on cement type must be determined for LC3
Considering the thickness of the concrete prisms, the shrinkage devel­ concrete. In this study, the 28-day compressive strength (fc28) was either
opment of the specimens became negligible after 90 days. Comparing measured (section 3.1) or “fitting parameter” determined from Eq. (1).
concretes with the same strength grade, both total and drying shrinkage In summary, two options were examined in the curve fitting of Eq. (1)
strains were similar values at the end of the testing period (100 days) and the results, including values of a, b and fc28, are shown in Table 4. In
regardless of the binder composition, indicating that the water/binder addition, the coefficient of determination (R2) and the root-mean-square
ratio is governing total and drying shrinkage rather than the binder deviation (RMSD) are also reported in Table 4.
compositions (plain cement or LC3). Results also revealed that LC3 As shown in Table 4, the values of a, b and fc28 are relatively similar
materials can produce slightly higher shrinkage at an early age but a within the 2 options. Values of a and b ranged from 3.53 to 4.57 and
similar value at a later age, which is consistent with the previous study
[27]. It should be noted that the drying shrinkage specimens (Category
3) in this study are in fact including the drying shrinkage and a part of Table 4
autogenous shrinkage occurring after 7 days. Overall, in this study, total Different options for fitting compressive strength of LC3 concretes based on ACI
shrinkage strains of LC3 concretes are similar to that of conventional 209R-92.
concretes, especially at a later age. LC3-25 LC3-32 LC3-45

Option 1: a and b are fitting parameters, fc28 is the experimental value

5. Assessment of current standards and models predicting a 4.57 4.11 3.55
mechanical properties and shrinkage b 0.82 0.84 0.84
fc28 (MPa) 31.01 38.37 49.38
2
R 0.9977 0.9966 0.9794
5.1. Mechanical properties RMSD (MPa) 0.50 0.72 2.35

Option 2: a, b and fc28 are fitting parameters

The development of the compressive strength of concrete in design
a 4.49 4.13 3.53
standards is expressed as a function of time and the 28-day average b 0.80 0.85 0.84
concrete strength. Firstly, two well-established equations for the pre­ fc28 (MPa) 30.47 38.62 49.21
diction of the time-dependent development of concrete compressive R 2
0.9969 0.9955 0.9725
strength are considered in this study: RMSD (MPa) 0.57 0.84 2.71

Option 3: a = 4.0, b = 0.85 and fc28 is experimental value

a. ACI Committee 209 (ACI 209R-92) [58]: a 4.0 4.0 4.0
b 0.85 0.85 0.85
t fc28 (MPa) 31.01 38.37 49.38
fc (t) = × fc28 (1) 2
a+b×t R 0.9942 0.9971 0.9735
RMSD (MPa) 0.79 0.61 2.43
where fc(t) is the average compressive strength of concrete at t in days, a

Fig. 7. Drying shrinkage development of concretes up to 100 days (AS 1012.13).

7
Q.D. Nguyen et al. Construction and Building Materials 314 (2022) 125720

from 0.80 to 0.85 respectively. The estimated values of fc28 (Option 2) Table 6
are very close to the experimental value (Option 1), indicating that ACI Different options for fitting compressive strength of LC3 concretes based β-n
209R-92 can effectively predict concrete compressive strength from model.
early age values. In addition, a = 4.0 and b = 0.85 were used in Eq. (1) LC3-25 LC3-32 LC3-45
(denoted as Option 3 in Table 4) to determine R2 and RMSD values. R2 fc28 (MPa) 31.01 38.37 49.38
and RMSD values for Option 1, 2 and 3 exhibited very similar values in R2 0.88 0.90 0.92
three different concrete mixes. This indicated that a and b values from RMSD (MPa) 3.31 3.60 4.29
curve fitting (Option 1 and 2) provide no significant contribution for a
better prediction of compressive strength than that of common values
(Option 3). As a result, a = 4.0 and b = 0.85 are recommended for LC3
concretes in Eq. (1) to retain the simplicity and convenience of using ACI
209R in engineering practices.

b. β-n model (fib Model Code 2010 and Eurocode 2 [41,42]):

fc (t) = βcc (t) × fc28 (2)
[( )]
√̅̅̅̅
28
with: βcc (t) = exp s 1 − t and:

• βcc(t): a coefficient which depends on the age of the concrete (t)

• s: a coefficient which depends on the type of cement (s = 0.20 for
strength class of cement: 42.5R, 52.5 N and 52.5R (Class R); s = 0.25
for strength class of cement: 32.5R and 42.5 N (Class N); s = 0.38 for
strength class of cement: 32.5 N (Class S))

The cement strength test mix design and results for GPC and LC3
according to EN 196-1 [59] are presented in Table 5. The compressive Fig. 8. Comparison of the development of compressive strength between
experimental data and β-n models.
strength of LC3 mortar at 2 days and 7 days was slightly lower than that
of GPC mortar but the value at 28 days was comparable between the two
mortars. From Table 5, GPC and LC3 can be classified as cement class N Fig. 8, the β-n model is suitable for LC3 concretes using the value s =
according to EN 197-1 [60]. The coefficient s was reported in previous 0.25.
studies considering different binder compositions. For example, the The β-n model was also utilized to estimate the time-dependent
values of s were in the range between 0.12 and 0.22 for blended cements development of tensile strength as follow:
with silica fume and fly ash or between 0.20 and 0.38 for blended ce­ fct (t) = (βcc (t))α × fct28 (3)
ments with slag and fly ash [61,62]. As a result, the value of coefficient s
must be recalibrated for new LC3 concretes. Based on the cement where:
strength test results presented in Table 5, LC3 is a class N cement. As a
result, the value s = 0.25 has been used for LC3 concrete prediction • α: aging parameter with α = 1 for t less than 28 days and α = 2/3 for t
model for both compressive strength and tensile strength time- >28 days
dependent development. fc28 is the experimental value as shown in • fct28: tensile strength of concrete at t = 28 days
Table 6. Fig. 8 shows the prediction of the compressive strength devel­
opment based on s = 0.25 for LC3 concretes (cement class N). The β-n Since the value of s coefficient depends on the binder composition, it
model can successfully predict the compressive strength of LC3 concrete, is feasible to consider only one value of s coefficient (s = 0.25) for both
with R2 ranging from 0.88 to 0.92 and small RMSD values. The exper­ compressive strength and tensile strength of LC3 concretes. The value of
imental values of LC3-25 and LC3-32 concrete compressive strength α can range from 0.5 to 1 as previously reported in standards and
were lower up to 7 days and then were comparable to the predicted research [41,42,61]. Therefore, the α value was considered as a fitting
values, which is expected for most of blended cement-based concretes. parameter and fct28 was taken as either an experimental result (Option 1)
This is also consistent with the low strength values of LC3 mortar after 2 or fitting parameter (Option 2) in Table 7. The two options led to rela­
days and 7 days compared to GPC (Table 5). In addition, LC3-45 pre­ tively similar values within the range 1.26 to 1.77. As a result, the value
sented similar compressive strength to β-n model from 3 days, indicating α = 1.4 is suggested for tensile strength prediction of LC3 concretes up to
that a low water/binder ratio could have a positive effect on early-age 28 days. However, it is recommended to carry out a tensile strength test
compressive strength development of LC3 concretes. From Table 6 and
Table 7
Table 5 Different options for fitting tensile strength of LC3 concretes based β-n model.
Mix design and standard strength of GPC and LC3 according to EN 196–1 [59]
LC3-25 LC3-32 LC3-45
Mix design (g) GPC LC3
Option 1: α is fitting parameter, fct28 is experimental value
CEN standard sand 1350 1350 α 1.70 1.27 1.26
GP cement 450 252 fct28 (MPa) 3.22 3.55 4.49
Calcined clay 0 135 R2 0.99 0.98 0.98
Limestone 0 63 RMSD (MPa) 0.12 0.21 0.19
Water 225 225
Option 2: α and fct28 are fitting parameters
Duration (days) Compressive strength (MPa) s 1.77 1.39 1.34
2 19.58 ± 1.91 15.72 ± 0.81 fct28 (MPa) 3.29 3.74 4.64
7 44.13 ± 1.55 39.15 ± 1.76 R2 0.99 0.98 0.98
28 53.60 ± 1.57 54.02 ± 2.25 RMSD (MPa) 0.12 0.18 0.18

8
Q.D. Nguyen et al. Construction and Building Materials 314 (2022) 125720

instead of using the predicted values by the model when the develop­
ment of tensile strength is an important factor for structural members, as
discussed in Eurocode 2 [41]. The α value for t >28 days was not
considered as there is no experimental data of tensile strength available
after 28 days for LC3 concretes.

5.2. Autogenous shrinkage

Regarding autogenous shrinkage of LC3 concretes, 4 empirical

models from standards and previous study were assessed in this study.

a. Eurocode 2 (EC2 [41]):

εca (t) = βca (t) × εca (∞) (4)

where εca(∞) is the ultimate autogenous shrinkage strain of concrete and

βca(t) is the time function of autogenous shrinkage increasing rate with
the expression can be found in Ref [41].

b. fib Model Code 2010 (MC 2010 [42]):

εca (t) = βca (t) × εca (fcm ) (5)
( )2.5
( ) fcm 6
with εca fcm = − α × 60+fcm × 10− [42]. The value of α in εca(fcm)

depends on the cement type and is presented in fib Model Code 2010
[42].

c. Australian Standard AS 3600:2018 (AS 3600 [43])

εca (t) = βca (t) × εca (∞) (6)

d. Bazant Model B4 (Bazant B4 [44])

[ (τ )α ]rt
εca (t) = εca (∞) × 1 + au (7)
t
In addition, parameters depending on different cement types are
showed in Ref [44].
According to experimental data in sections 4.2 and 4.3, cement class
N and regular cement type were considered for LC3 concrete shrinkage
in EC2, MC 2010 and Bazant model as LC3 concretes presented higher
autogenous shrinkage and comparable total shrinkage in comparison
with GPC concretes. The calculated autogenous shrinkage of the 4
different models for all LC3 concretes is presented in Fig. 9. All models
provided significantly lower values in comparison with experimental
data. It can be concluded that the three standards and one academic
model (Bazant B4) underestimated the autogenous shrinkage of LC3
concretes.

5.3. Total shrinkage

As mentioned previously, the total shrinkage strain is calculated as

the combination of autogenous shrinkage strain and drying shrinkage
strains as follows [41–44]:
εcs = εca + εcd (8)

where: Fig. 9. Comparison between experimental data of autogenous shrinkage and

different models. a) LC3-25, b) LC3-32, c) LC3-45.
• εcs: the total shrinkage strain
• εca: the autogenous shrinkage strain (the equation is presented in where εcd,∞ is the ultimate drying shrinkage strain of concrete and f is
section 5.2) the increasing rate function of drying shrinkage. Overall, as presented in
• εcd: the drying shrinkage strain Eq. (9), the drying shrinkage strain at t in days depends on the exposure
environment including temperature (T), relative humidity (RH), the age
The time-dependent development of concrete drying shrinkage of concrete (t), the age of concrete at the starting of drying process (ts)
strain can be generally expressed as follows: and the size (in mm) of the cross-section. Similarly to section 5.2, 4
εcd (t) = εcd,∞ (T, RH) × f (t, ts , h0 ) (9) models including Eurocode 2, Model Code 2010, Australian Standard AS

9
Q.D. Nguyen et al. Construction and Building Materials 314 (2022) 125720

3600 and Bazant Model B4 were considered for total shrinkage strain.
Only the drying shrinkage strain equation is presented in these models as
the autogenous shrinkage component is previously discussed in section
5.2.

a. Eurocode 2 (EC2 [41]):

εcd (t) = βds (t, ts ) × εcd,∞ (10)
[ (
where: εcd,∞ = kh × εcd,0 and εcd,0 = 1.0 × (220 + 110 × αds1 ) × exp −
)]
αds2 × ffcmo
cm
× 10− 6 × βRH . αds1 and αds2 are cement type dependent

coefficients from Eurocode 2 [41].

b. Model Code 2010 (MC 2010 [42])

εcd (t) = εcd0 (fcm ) × βRH (RH) × βcd (t − ts ) (11)

where:

εcd0 (fcm ) = [(220 + 110 × αds1 ) × exp(− αds2 × fcm ) ] × 10− 6

(12)

with αds1 and αds2 also depending on the cement type from fib Model
Code 2010 [42].

c. Australian Standard AS 3600:2018 (AS 3600 [43])

εcd (t) = k1 × k4 × εcd.b (13)
The expression of parameter was reported in Ref [43] and there is no
parameter which is dependent to cement type.

d. Bazant B4 Model (Bazant B4 [44])

( )
εcd t , t s = εcd,∞ (t s ) × kh × S(t ) (14)

With t̃ and t̃s are utilized to capture the effect of temperature and
ageing on shrinkage development. t̃ s is temperature corrected age at the
starting of the drying process and t̃ is exposure duration with temper­
ature corrected. When the temperature is 23 ◦ C as used in this study, the
values of t̃ s and t̃ can be expressed as: t̃ s = ts and t̃ = t − ts . The values
of depending parameters are presented in Bazant et al. [44].
The development of total shrinkage in LC3 concretes calculated from
the four models in comparison with the experimental results is presented
in Fig. 10. Table 8 shows the coefficient of determination (R2) and RMSD
of all models for total drying shrinkage developments of LC3 concretes.
Eurocode 2 and fib Model Code 2010 considered cement class N. The
values of Bazant B4 model were calculated based on the assumption of
regular cement type [44]. Eurocode 2 could effectively predict LC3-45
concrete shrinkage with a R2 of 0.98 whilst LC3-32 and LC3-45 con­
crete were overestimated by Eurocode 2 formulas. fib Model Code 2010
only provided good predictions of LC3-32 concrete shrinkage. Austra­
lian Standard AS 3600 exhibited overestimation and poor correlation for
all LC3 concretes. On the other hand, Bazant B4 model (academic
model) could successfully predict the developments of LC3 concretes for
all compressive strength grades. In addition, Bazant B4 model is
capturing the increase in total shrinkage relating to the paste content Fig. 10. Comparison between experimental data of total shrinkage and
different models. a) LC3-25, b) LC3-32, c) LC3-45.
through the ratio total aggregate/paste content and paste content/con­
crete density. Considering Table 8, Bazant B4 model produced the best
predictions for LC3 concretes (44 wt% binder replaced by calcined clay 6. Conclusion
and limestone) with the highest values of R2 and the lowest RMSD
among the four models. As previously mentioned in section 4.3, the This study aims to evaluate the effect of calcined clay limestone
drying shrinkage measured using Australian Standard AS 1012.13 cement (LC3) on the development of autogenous, drying and total
(Category 3) contained drying shrinkage and a part of autogenous shrinkage of concrete up to 100 days. The 44 wt% of GPC was
shrinkage and was not considered for the assessment of Section 5. substituted by the combination of calcined clay and limestone with the
approximate weight ratio 2:1. Three different concrete grades including
25 MPa, 32 MPa and 45 MPa of both GPC and LC3 concrete were

10
Q.D. Nguyen et al. Construction and Building Materials 314 (2022) 125720

Table 8 Validation, Resources, Writing - Review & Editing, Supervision. Wengui

R2 and RMSD of different models for total shrinkage. Li: Writing – review & editing. Arnaud Castel: Methodology, Valida­
LC3-25 LC3-32 LC3-45 tion, Resources, Writing – review & editing, Supervision, Project
2 2 administration, Funding acquisition.
R RMSD R RMSD R2 RMSD
(µm/m) (µm/m) (µm/m)
Declaration of Competing Interest
Eurocode 2 0 155.4 0.619 94.85 0.977 26.09
fib MC 0.726 69.30 0.945 35.92 0.634 102.4
2010 The authors declare that they have no known competing financial
AS 3600 0 203.4 0.018 152.2 0.898 54.08 interests or personal relationships that could have appeared to influence
Bazant B4 0.953 28.62 0.944 36.35 0.965 31.53 the work reported in this paper

Acknowledgement
considered. Nitrogen adsorption technique was conducted to evaluate
the pore size distribution of pastes. The mechanical properties including
This research project was funded by the Australian Research Council
compressive strength and tensile strength until 28 days were also
(ARC) and Cement Concrete & Aggregates Australia (ARC Linkage
investigated. Several models and standards predicting mechanical
Project No. LP170100912). The assistance of the laboratory staff is
properties and shrinkage development were used to assess their appli­
acknowledged here.
cability for LC3 concretes. The conclusions are summarized as follows:

References
(1) LC3 concretes can achieve a comparable or higher 28 days
compressive strength than that of conventional concrete with [1] A.B. Hossain, J. Weiss, Assessing residual stress development and stress relaxation
100 wt% of OPC. The development of LC3 concrete tensile in restrained concrete ring specimens, Cement Concrete Comp 26 (2004) 531–540.
strength was similar to the compressive strength. Eurocode 2 and [2] I. Khan, A. Castel, R.I. Gilbert, Tensile creep and early-age concrete cracking due to
restrained shrinkage, Construction and Building Materials 149 (2017) 705–715.
ACI 209R-92 can effectively describe the development of
[3] V. Baroghel-Bouny, P. Mounanga, A. Khelidj, A. Loukili, N. Rafaï, Autogenous
compressive strength and tensile strength of LC3 concretes using deformations of cement pastes: Part II. W/C effects, micro–macro correlations, and
current standard parameters. threshold values, Cement and Concrete Research 36 (2006) 123–136.
(2) Minor swelling peaks in autogenous deformation were observed [4] S.A. Altoubat, D.A. Lange, Creep, Shrinkage, and Cracking of Restrained Concrete
at Early Age, ACI Materials Journal 98 (2001).
in both GPC and LC3 concretes at the early age. The presence of [5] A.M. Neville, Properties of Concrete, Fifth Edition, Pearson Education Ltd England,
ettringite and portlandite may cause expansion in GPC concretes 2011.
whilst the autogenous swelling mechanism of LC3 concretes has [6] K. Scrivener, F. Martirena, S. Bishnoi, S. Maity, Calcined clay limestone cements
(LC3), Cement and Concrete Research 114 (2018) 49–56.
been unresolved. LC3 concretes presented higher autogenous [7] M. Antoni, J. Rossen, F. Martirena, K. Scrivener, Cement substitution by a
shrinkage at a later age up to 100 days due to a refinement of the combination of metakaolin and limestone, Cement and Concrete Research 42 (12)
pore structure. The nitrogen absorption test presented a contin­ (2012) 1579–1589.
[8] F. Avet, L. Sofia, K. Scrivener, Concrete Performance of Limestone Calcined Clay
uous refinement of the pore structure of LC3 pastes which was not Cement (LC3) Compared with Conventional Cements, Advances in Civil
observed with GPC pastes. Engineering Materials 8 (3) (2019) 20190052, https://doi.org/10.1520/
(3) The development of total shrinkage was similar between GPC and ACEM1903-EB10.1520/ACEM20190052.
[9] S. Nickovic, A. Vukovic, M. Vujadinovic, V. Djurdjevic, G. Pejanovic, Technical
LC3 concretes. The water/binder ratio was the major factor Note: High-resolution mineralogical database of dust-productive soils for
governing the total shrinkage strains of all concretes. Assuming atmospheric dust modeling, Atmos. Chem. Phys. 12 (2) (2012) 845–855.
that total shrinkage is a combination of autogenous shrinkage and [10] K. Scrivener, F. Avet, H. Maraghechi, F. Zunino, J. Ston, W. Hanpongpun,
A. Favier, Impacting factors and properties of limestone calcined clay cements
drying shrinkage, LC3 concretes drying shrinkage was lower than
(LC3), Green Materials 7 (2019) 3–14.
that of GPC concretes. This can be attributed to the finer pore [11] Q.D. Nguyen, M.S.H. Khan, A. Castel, Engineering Properties of Limestone Calcined
structure of LC3 concretes. Clay Concrete, Journal of Advanced Concrete Technology 16 (2018) 343–357.
(4) Eurocode 2, fib Model Code 2010, Australia Standard AS 3600 [12] Q.D. Nguyen, M.S.H. Khan, A. Castel, Chloride Diffusion in Limestone Flash
Calcined Clay Cement Concrete, ACI Materials Journal 117 (2020).
and Bazant B4 model (academic model) were utilized to model [13] M.S.H. Khan, Q.D. Nguyen, A. Castel, Performance of limestone calcined clay
the shrinkage development of LC3 concretes. All models under­ blended cement-based concrete against carbonation, Advances in Cement Research
estimated the development of LC3 concrete autogenous 32 (2020) 481–491.
[14] Q.D. Nguyen, T. Kim, A. Castel, Mitigation of alkali-silica reaction by limestone
shrinkage. The recalibration of these models is required when calcined clay cement (LC3), Cement and Concrete Research 137 (2020), 106176.
more data for LC3 concretes becomes available. [15] Z. Shi, S. Ferreiro, B. Lothenbach, M.R. Geiker, W. Kunther, J. Kaufmann,
(5) The four different models exhibited different predictions for the D. Herfort, J. Skibsted, Sulfate resistance of calcined clay – Limestone – Portland
cements, Cement and Concrete Research 116 (2019) 238–251.
development of the total shrinkage of LC3 concretes. The Bazant [16] Q.D. Nguyen, A. Castel, Reinforcement corrosion in limestone flash calcined clay
B4 model assuming regular cement type presented the best pre­ cement-based concrete, Cement and Concrete Research 132 (2020), 106051.
diction among these models. Australian Standard AS 3600 over­ [17] Y. Akkaya, C. Ouyang, S.P. Shah, Effect of supplementary cementitious materials
on shrinkage and crack development in concrete, Cement and Concrete Composites
estimated the total shrinkage of all LC3 concretes. 29 (2007) 117–123.
(6) A broader database of LC3 concrete shrinkage values is recom­ [18] A. Darquennes, S. Staquet, M.-P. Delplancke-Ogletree, B. Espion, Effect of
mended to improve and establish optimal parameters for the autogenous deformation on the cracking risk of slag cement concretes, Cement and
Concrete Composites 33 (2011) 368–379.
current standards and empirical models.
[19] Z. Liu, W. Hansen, Aggregate and slag cement effects on autogenous shrinkage in
cementitious materials, Construction and Building Materials 121 (2016) 429–436.
[20] M. Şahmaran, Ö. Yaman, M. Tokyay, Development of high-volume low-lime and
high-lime fly-ash-incorporated self-consolidating concrete, Magazine of Concrete
CRediT authorship contribution statement Research 59 (2007) 97–106.
[21] J. Li, Y. Yao, A study on creep and drying shrinkage of high performance concrete,
Cement and Concrete Research 31 (2001) 1203–1206.
Quang Dieu Nguyen: Conceptualization, Methodology, Software, [22] D.W. Mokarem, R.E. Weyers, D.S. Lane, Development of a shrinkage performance
Validation, Formal analysis, Investigation, Data curation, Writing – specifications and prediction model analysis for supplemental cementitious
material concrete mixtures, Cement and Concrete Research 35 (2005) 918–925.
original draft, Supervision, Visualization. Sumaiya Afroz: Methodol­ [23] Z. Giergiczny, Fly ash and slag, Cement and Concrete Research 124 (2019) 105826,
ogy, Validation, Investigation, Data curation, Writing – review & edit­ https://doi.org/10.1016/j.cemconres.2019.105826.
ing. Yingda Zhang: Methodology, Validation, Investigation, Data
curation, Writing – review & editing. Taehwan Kim: Methodology,

11
Q.D. Nguyen et al. Construction and Building Materials 314 (2022) 125720

[24] M. Rezvani, T. Proske, Influence of chemical-mineralogical properties of limestone [45] R. Snellings, J. Chwast, Ö. Cizer, N. De Belie, Y. Dhandapani, P. Durdzinski,
on the shrinkage behaviour of cement paste and concrete made of limestone-rich J. Elsen, J. Haufe, D. Hooton, C. Patapy, M. Santhanam, K. Scrivener, D. Snoeck,
cements, Construction and Building Materials 157 (2017) 818–828. L. Steger, S. Tongbo, A. Vollpracht, F. Winnefeld, B. Lothenbach, RILEM TC-238
[25] M. Rezvani, T. Proske, C.-A. Graubner, Modelling the drying shrinkage of concrete SCM recommendation on hydration stoppage by solvent exchange for the study of
made with limestone-rich cements, Cement and Concrete Research 115 (2019) hydrate assemblages, Materials and Structures 51 (2018) 172.
160–175. [46] K. Scrivener, R. Snellings, B. Lothenbach, A Practical Guide to Microstructural
[26] M. Bouasker, P. Mounanga, P. Turcry, A. Loukili, A. Khelidj, Chemical shrinkage of Analysis of Cementitious Materials, CRC Press, 2016.
cement pastes and mortars at very early age: Effect of limestone filler and granular [47] E.P. Barrett, L.G. Joyner, P.P. Halenda, The Determination of Pore Volume and
inclusions, Cement and Concrete Composites 30 (1) (2008) 13–22. Area Distributions in Porous Substances. I. Computations from Nitrogen Isotherms,
[27] Y. Dhandapani, T. Sakthivel, M. Santhanam, R. Gettu, R.G. Pillai, Mechanical Journal of the American Chemical Society 73 (1951) 373–380.
properties and durability performance of concretes with Limestone Calcined Clay [48] N. De Belie, J. Kratky, S. Van Vlierberghe, Influence of pozzolans and slag on the
Cement (LC3), Cement and Concrete Research 107 (2018) 136–151. microstructure of partially carbonated cement paste by means of water vapour and
[28] AS 3972: General purpose and blended cements, Standards Australia, Sydney, nitrogen sorption experiments and BET calculations, Cement and Concrete
Australia 2010. Research 40 (2010) 1723–1733.
[29] Quang Dieu Nguyen, M.S.H. Khan, Arnaud Castel, Taehwan Kim, Durability and [49] Q. Zeng, K. Li, T. Fen-Chong, P. Dangla, Pore structure characterization of cement
Microstructure Properties of Low-Carbon Concrete Incorporating Ferronickel Slag pastes blended with high-volume fly-ash, Cement and Concrete Research 42 (2012)
Sand and Fly Ash, Journal of Materials in Civil Engineering 31 (8) (2019) 194–204.
04019152, https://doi.org/10.1061/(ASCE)MT.1943-5533.0002797. [50] K.S.W. Sing, Reporting physisorption data for gas/solid systems with special
[30] H.M. Rietveld, A Profile Refinement Method for Nuclear and Magnetic Structures, reference to the determination of surface area and porosity (Recommendations
Journal of Applied Crystallography 2 (2) (1969) 65–71. 1984), Pure and Applied Chemistry 57 (1985) 603–619.
[31] Q.D. Nguyen, S. Afroz, A. Castel, Influence of Calcined Clay Reactivity on the [51] C. Van Bunderen, R. Snellings, L. Vandewalle, Ö. Cizer, Early-age hydration and
Mechanical Properties and Chloride Diffusion Resistance of Limestone Calcined autogenous deformation of cement paste containing flash calcined dredging
Clay Cement (LC3) Concrete, Journal of Marine Science and Engineering 8 (2020) sediments, Construction and Building Materials 200 (2019) 104–115.
301. [52] P. Lura, K. Van Breugel, I. Maruyama, Effect of curing temperature and type of
[32] Q.D. Nguyen, A. Castel, T. Kim, M.S.H. Khan, Performance of fly ash concrete with cement on early-age shrinkage of high-performance concrete, Cement and
ferronickel slag fine aggregate against alkali-silica reaction and chloride diffusion, Concrete Research 31 (2001) 1867–1872.
Cement and Concrete Research 139 (2021), 106265. [53] Alejandra Tironi, Alberto N. Scian, Edgardo F. Irassar, Blended Cements with
[33] AS 1012.8.1: Methods of testing concrete - Method for making and curing concrete Limestone Filler and Kaolinitic Calcined Clay: Filler and Pozzolanic Effects, Journal
- Compression and indirect tensile test specimens, Standards Australia, Sydney, of Materials in Civil Engineering 29 (9) (2017) 04017116, https://doi.org/
Australia, 2014. 10.1061/(ASCE)MT.1943-5533.0001965.
[34] ASTM C39/C39M-18: Standard Test Method for Compressive Strength of [54] M. Sidney, Y. J. Francis, D. David, Concrete 2nd Edition. 2003.
Cylindrical Concrete Specimens, ASTM International, West Conshohocken, PA, [55] J. Seo, S. Park, H.N. Yoon, H.K. Lee, Effect of CaO incorporation on the
2018. microstructure and autogenous shrinkage of ternary blend Portland cement-slag-
[35] E.-I. Tazawa, S. Miyazawa, Influence of cement and admixture on autogenous silica fume, Construction and Building Materials 249 (2020), 118691.
shrinkage of cement paste, Cement and Concrete Research 25 (1995) 281–287. [56] Zhengqi Li, Drying shrinkage prediction of paste containing meta-kaolin and
[36] P.J.P. Gleize, M. Cyr, G. Escadeillas, Effects of metakaolin on autogenous shrinkage ultrafine fly ash for developing ultra-high performance concrete, Materials Today,
of cement pastes, Cement and Concrete Composites 29 (2007) 80–87. Communications 6 (2016) 74–80.
[37] İ.B. Topçu, T. Bilir, Experimental investigation of drying shrinkage cracking of [57] Erhan Güneyisi, Mehmet Gesoğlu, Erdoğan Özbay, Strength and drying shrinkage
composite mortars incorporating crushed tile fine aggregate, Materials & Design 31 properties of self-compacting concretes incorporating multi-system blended
(2010) 4088–4097. mineral admixtures, Construction and Building Materials 24 (10) (2010)
[38] ASTM C496/C496M-17: Standard Test Method for Splitting Tensile Strength of 1878–1887.
Cylindrical Concrete Specimens, ASTM International, West Conshohocken, PA, [58] ACI 209R-92 (Reapproved 2008): Prediction of creep, shrinkage, and temperature
2017. effects in concrete structures, ACI Committee 209 - American Concrete Institute,
[39] AS 1012.13: Methods of testing concrete: Determination of the drying shrinkage of Farmington Hills, MI, 2008.
concrete for samples prepared in the field or in the laboratory, Standards Australia, [59] EN 196-1: Methods of testing cement Part 1: Determination of strength, European
Sydney, Australia, 2015. Standard, Brussels, 2016.
[40] EN 12390‑16: Testing hardened concrete - Part 16: Determination of the shrinkage [60] EN 197-1: Cement Part 1: Composition, specifications and conformity criteria for
of concrete, European Standard, Brussels, 2019. common cements, European Standard, Brussels, 2011.
[41] EN 1992-1-1: Eurocode 2: Design of concrete structures, Part 1-1: General rules and [61] T. Kanstad, T.A. Hammer, Ø. Bjøntegaard, E.J. Sellevold, Mechanical properties of
rules for buildings, European Standard, Brussels, 2015. young concrete: Part II: Determination of model parameters and test program
[42] fib - International Federation for Structural Concrete, Fib Model Code for Concrete proposals, Materials and Structures 36 (2003) 226–230.
Structures 2010, Wilhelm Ernst & Sohn, Berlin, Germany, 2013. [62] B. Klemczak, M. Batog, M. Pilch, Assessment of concrete strength development
[43] AS 3600: Concrete structures Standards Australia, Sydney, Australia, 2018. models with regard to concretes with low clinker cements, Archives of Civil and
[44] R.T.C. TC-242-MDC, RILEM draft recommendation: TC-242-MDC multi-decade Mechanical, Engineering 16 (2016) 235–247.
creep and shrinkage of concrete: material model and structural analysis*, Materials
and Structures, 48 (2015) 753-770.

12