See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/282633972

Early Carbonation Curing of Concrete Masonry Units with Portland Limestone

Cement

Article  in  Cement and Concrete Composites · July 2015

DOI: 10.1016/j.cemconcomp.2015.07.004

CITATIONS READS

19 186

2 authors, including:

Hilal El-Hassan
United Arab Emirates University
31 PUBLICATIONS   74 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Alkali-activated inorganic polymer concrete (AIPC) incorporating dune sand View project

All content following this page was uploaded by Hilal El-Hassan on 09 December 2017.

The user has requested enhancement of the downloaded file.

 Cement & Concrete Composites 62 (2015) 168–177

Contents lists available at ScienceDirect

Cement & Concrete Composites

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

Early carbonation curing of concrete masonry units with Portland

limestone cement
Hilal El-Hassan a, Yixin Shao b,⇑
a
Department of Civil Engineering, American University in Dubai, P.O. Box 28282, Dubai, United Arab Emirates
b
Department of Civil Engineering and Applied Mechanics, McGill University, 817 Sherbrooke Street West, Montreal, Quebec H3A 2K6, Canada

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

Article history: The effect of carbonation curing on the mechanical properties and microstructure of concrete masonry
Received 1 September 2014 units (CMU) with Portland limestone cement (PLC) as binder was examined. Slab samples, representing
Received in revised form 25 May 2015 the web of a CMU, were initially cured at 25 °C and 50% relative humidity for durations up to 18 h.
Accepted 21 July 2015
Carbonation was then carried out for 4 h in a chamber at a pressure of 0.1 MPa. Based on Portland lime-
Available online 21 July 2015
stone cement content, CO2 uptake of PLC concrete after 18 h of initial curing reached 18%. Carbonated and
hydrated concretes showed comparable compressive strength at both early and late ages due to the 18-h
Keywords:
initial curing. Carbonation reaction converted early hydration products to a crystalline microstructure
PLC
Curing
and subsequent hydration transformed amorphous carbonates into more crystalline calcite. Portland
CMU limestone cement could replace Ordinary Portland Cement (OPC) in making equivalent CMUs which have
CO2 uptake shown similar carbon sequestration potential.
Compressive strength Ó 2015 Elsevier Ltd. All rights reserved.
Microstructure

1. Introduction With similar major components as Ordinary Portland Cement

Portland limestone cement (PLC) is the product of intergrinding
Portland cement clinker with limestone and calcium sulfate. When
added to cement, limestone may affect hydration, carbonation and
compressive strength of concrete. Currently, major cement produc-
ers are marketing and investing in PLC with up to 20% cement
replacement by limestone. Such a replacement could reduce the
carbon footprint associated with cement production by 20%.
Early research believed that fine limestone particles acted as
nucleation sites, thus increasing the rate of hydration of the
calcium silicates at early age and possibly improving the
distribution of hydrates [1,2]. Ramachandran had reported that
the hydration rate of C3S was also increased with an increase in
the fineness and amount of CaCO3 [3]. The formation of hydration
products as ettringite had also been accelerated in the presence of
CaCO3, while C–A–H incorporated significant amount of calcium
carbonates into its structure and formed calcium silicocarbonated
hydrates [3]. Additionally, CaCO3 will react chemically with alumi-
nate phases to form carboaluminate phases [4]. Finally, because
limestone is softer than clinker, when interground, it will achieve
a finer particle size, thus refining particle packing and improving
particle size distribution [5].
⇑ Corresponding author.
E-mail address: yixin.shao@mcgill.ca (Y. Shao).
(OPC), PLC concrete can be carbonated to sequester CO2 to further
reduce the carbon footprint from cement production. This
early-age carbonation is the reaction between calcium silicates or
early hydration products with carbon dioxide producing a hybrid
binder structure of calcium silicate hydrates and calcium carbon-
ates [6]. It is different from weathering carbonation. Weathering
carbonation of limestone cement concrete was studied by Parrott
and Tsivilis et al. [7,8]. Results showed that the amount of
limestone present in the cement affected the degree of weathering
carbonation. When more than 19% of Portland cement clinker was
replaced with limestone, the microstructure analysis showed that
the depth of weathering carbonation of concrete increased during
18-month drying (at 20 °C and 60% relative humidity) [7].
However, with less than 15% replacement, weathering carbonation
degree was not intensified [7]. The effect of limestone addition of
up to 35% on the durability of concrete was apparent in other
research [8]. Within 15% addition, the limestone additive could
reduce water permeability and sorptivity of the concrete due to
improved cement particle size distribution [8].
The physical performance and carbonation behavior of Ordinary
Portland Cement (OPC) concrete was investigated in separate work
[9,10]. To promote high degree of carbonation, the procedure
necessitated the introduction of initial air curing prior to carbon
exposure. Carbonation curing was then performed by injecting
CO2 gas (99% purity) into a sealed chamber. Of the examined curing

http://dx.doi.org/10.1016/j.cemconcomp.2015.07.004
0958-9465/Ó 2015 Elsevier Ltd. All rights reserved.
 H. El-Hassan, Y. Shao / Cement & Concrete Composites 62 (2015) 168–177 169

schemes, the optimum combination appeared to be a procedure of Table 2

air curing of 18 h, followed by carbonation curing of 4 h, and Mixture proportion of PLC concretes.

immediate surface water spray for water loss compensation. On
the other hand, the sealed hydration (0a) with no water loss during
entire curing served as the reference.
The purpose of this paper is to investigate if PLC can be used to
replace OPC in making equivalent concrete masonry units (CMU).
The effect of carbonation process parameters on the microstruc-
ture and characteristics of the reaction products of PLC concrete
was examined. A static carbonation setup developed in carbona-
tion of OPC concrete will be used to perform carbonation curing
of PLC concrete [10]. The reaction products of the concretes are
characterized utilizing X-ray Diffraction (XRD), Scanning
Electron Microscopy (SEM), and Thermogravimetry/Differential
Thermogravimetry (TG/DTG). A unique method was developed to
determine the cement content in concrete powder to allow a quan-
titative TG analysis for concrete [9]. The produced PLC concrete
blocks may serve as a potential carbon sink whilst using lesser
Portland cement in the production.
2. Experimental investigations
2.1. PLC concrete sample preparation and curing
To simulate a typical web or face shell of a 200-mm PLC con-
crete masonry unit, rectangular concrete samples 127 mm long,
76 mm wide, and 38 mm thick were cast. PLC used in this research
is a commercial product manufactured and marketed by Holcim
Canada. About 13–15% limestone is interground with OPC clinker.
The chemical compositions of both PLC and OPC are reported in
Table 1. The CO2 content determined by infrared carbon analyzer
in as-received PLC is about 7.0%, which is equivalent to a limestone
content of 15.9%. The cement is considered as high limestone
cement and is therefore labeled as Portland limestone cement. To
maintain comparable early compressive strength to OPC, PLC is
ground finer. The Blaine fineness of PLC is 500 m2/kg in comparison
to 390 m2/kg in general use OPC. The as-received lightweight
expanded slag aggregates had a water content of 5% by mass at sat-
urated surface dry (SSD). The sieve analysis of the same aggregate
blend was performed in previous work; the well graded aggregates
ranged in size between 0.2 and 6 mm [10]. The concrete mixture
proportion shown in Table 2 is based on a commercial CMU mix
design. The water-to-cement ratio (w/c) was 0.71 taking into
account the mixing water and the water in the aggregates (5%).
The raw materials were mixed in a pan mixer and samples were
compact formed using a vibrating hammer to simulate the indus-
try production of CMU. PLC concrete was then demolded right after
casting. Three curing schemes were investigated: (1) Sealed hydra-
tion in a plastic bag served as reference (0a) without water loss; (2)
Initial air curing of 18 h in an environmental chamber of 50% rela-
tive humidity (RH) and 25 °C followed by 4-h carbonation and sub-
sequent hydration (18a + 4c); (3) Initial air curing of 18 h in an
environmental chamber of 50% relative humidity (RH) and 25 °C,
followed by 4-h carbonation, water compensation and subsequent
hydration (18a + 4c + sp). Water compensation through surface
spray was introduced immediately after carbonation to restore
all the curing-induced water loss. The spraying process was ceased
Slab Mass Percent
(g) (kg/m3) (mass%)
Portland limestone cement 88 241 13
Mixing water 35 96 5
SSD expanded slag 554 1502 82
Concrete sample 677 1839 100
upon reaching surface saturation, and required a few days for com-
plete water compensation [9]. Subsequent hydration was carried
out in a sealed bag at a relative humidity of 80 ± 5% and room tem-
perature (24 ± 1 °C) until 24 h and 28 days for compressive
strength tests following ASTM C140 [11]. Three rectangular speci-
mens for each batch were tested and averaged with a compressive
area of 127 mm  38 mm.
The fractured concrete samples were preserved in an acetone
solution to stop hydration for microstructure analysis. Acetone
exchange with water to stop hydration was reported as the least
damaging method to preserve the microstructure [12]. Prior to
analysis, the samples were pre-dried overnight at 60 °C.
A schematic of the carbonation setup is shown in Fig. 1. Within
a few seconds, the pressure is reduced to 0.7 bars below atmo-
spheric prior to carbonation by means of a vacuum. A heater
installed at the inlet brings up the carbon dioxide gas of 99% purity
to room temperature as it flows into the sealed chamber. A pres-
sure regulator is connected to the system to retain the pressure
at 0.1 MPa. The whole system is placed on a digital scale to monitor
the mass increase due to the carbonation reaction.
2.2. Measurement of CO2 uptake in PLC concrete
CO2 uptake of carbonated PLC concrete was quantified by three
methods: mass gain, mass curve, and furnace decomposition.
Mass gain method estimates CO2 uptake in concrete by comparing
mass of samples before and after carbonation (Eq. (1)).
Carbonation-induced water loss was collected by absorbent paper
and added to the final mass. By treating the system as a closed sys-
tem, it was imperative to include the evaporated water, which was
initially inside the samples prior to carbonation.
CO2 uptakeð%Þ ¼ ðFinal mass þ Mass of water loss
 Initial massÞ=ðMass of cementÞ ð1Þ

Table 1
Chemical composition of PLC and OPC.

Cement Constituent (%)

CaO SiO2 Al2O3 Fe2O3 MgO Na2O K2O SO3 CO2
PLC 59.8 21.6 4.5 2.7 1.7 0.8 – 3.2 7.0
OPC 63.1 19.8 4.9 2.3 2.5 0.9 – 3.8 2.0
Fig. 1. Schematic of carbonation setup.
170 H. El-Hassan, Y. Shao / Cement & Concrete Composites 62 (2015) 168–177
The mass curve method was performed by placing the carbon-
ation setup on a digital balance, which was zeroed after vacuuming
the chamber. A mass curve was recorded as mass versus time until
the end of the process at which CO2 was released and the residual
mass, M, was measured. The system was calibrated by repeating
the tests using CO2-insensitive styrofoam samples of the same vol-
ume to obtain a second residual mass, m. The difference between M
and m represented the CO2 uptake by concrete (Eq. (2)). Data col-
lected by mass gain and mass curve methods are two simultaneous
measurements from the same process and therefore should be
comparable. They are independent from any CO2 content existing
before carbonation.
CO2 uptake ð%Þ ¼ ðM  mÞ=ðMass of cementÞ
As a third check, the concrete was subjected to an independent
furnace decomposition method. A furnace of maximum tempera-
ture of 1100 °C was employed to test large concrete samples with
mass range of 35–70 g. Separation of paste from concrete was thus
avoided. PLC concrete samples were ignited up to 105 °C, 470 °C,
and 950 °C to quantitatively measure the evaporable water, bound
water in hydration products, and carbon dioxide in carbonates
respectively [13].
CO2 content of powder samples was also determined by coulo-
metric titration in a hydrochloric acid (HCl) solution [14]. Two
samples were used to obtain an average. The powder sample was
immersed in the acid solution in a pipette, which was sealed from
the opening by a cork. A thin plastic tube was connected at one end
to the inside of the pipette and at the other end to a coulometer.
The reaction between hydrochloric acid (HCl) and
carbon-containing compounds such as calcium carbonate led to
the release of carbon dioxide gas. The coulometer measured the
amount of carbon released, and ultimately the carbon dioxide from
stoichiometric proportions.
2.3. Microstructure evaluation
The microstructure of the concrete specimens was character-
ized employing XRD, TG/DTG, and SEM. The samples were obtained
by hammer crushing of 5-mm thick concrete surface layer and
then sieving to pass a 125-lm sieve to remove coarse aggregates.
No attempts were made to separate the fine aggregates from the
cement. Therefore, powder samples were a mix of paste and fine
aggregates. XRD and TG analysis was performed on the same con-
crete powder to determine the reaction products.
The concrete powder was first subjected to XRD analysis. It was
analyzed by means of a Philips PW1710 Powder Diffractometer
(Cu Ka radiation, X’celerator proportional detector, scan interval
10–100°, 0.02°, and 0.5 s per step) to relatively quantify the
amount of calcium hydroxide, calcium silicates, and calcium car-
bonates in carbonated and hydrated samples.
The same concrete powder to that used in XRD analysis was
used in a thermogravimetry (TG/DTG) test. TG was utilized to esti-
mate the respective contents of C–S–H and ettringite (105–420 °C),
calcium hydroxide (420–470 °C), and calcium carbonate
(470–950 °C) [13]. A thermal analyzer (NETZSCH, TG 449 F3
Jupiter) with a resolution of 0.01 mg was employed with a sample
mass of 120 mg for each test. The TG/DTG curves were thus
obtained by plotting the mass loss between 25 °C and 950 °C at a
heating rate of 10 °C/min.
A Scanning Electron Microscope (Hitachi S4000) with an EDAX
(Phoenix EDS) microanalysis system was utilized to perform
high-vacuum SEM morphology analysis of fractured concrete in a
backscattered electron (BSE) mode. The air-exposed surface of a
concrete slab was chiseled to obtain gravel-size granules, and then
a thin gold palladium coating was applied to ensure conductivity
during analysis. The micrographs were used to identify the hydra-
tion and carbonation products in PLC concrete.
3. Experimental results and discussion
3.1. CO2 uptake and compressive strength of PLC concrete
The CO2 uptake and content calculated based on the mass gain,
mass curve, and furnace analysis are shown in Table 3. The first
two methods, which are insensitive to the initial carbon content,
were used to determine the carbon uptake of the 4-h carbonated
concrete. The results show a consistency among non-sprayed and
sprayed concrete at 18.3% by mass (mass%). As a third check and
ð2Þ
in order to determine the initial carbon content at the surface
and core, the furnace analysis method was employed. The hydrated
reference (0a) resulted in CO2 contents of 6.68 mass% and
6.54 mass% initially present prior carbonation at the surface and
core respectively. The equivalent limestone content is 15.20 mass%
and 14.86 mass%. Alternatively, the respective carbon content of
4-h carbonated concrete reached 26.14 mass% and 25.42 mass%
for the surface and core. The results show that the concrete sam-
ples were carbonated throughout the thickness with a higher
degree of reactivity at the exposed surface. The average carbon
uptake is then calculated based on that of the three methods and
is summarized in the last column of Table 3.
The cement content of concrete powder was determined using
past experimental work and equations [9]. All reaction products
are expressed in terms of dry cement content by mass (mass%).
The cement contents (CC) of different concrete batches are pre-
sented in Table 4 as correction factors. Their corresponding CO2
contents based on dry cement are summarized in Table 5. The TG
analysis of the hydrated reference detected approximately 7 mass%
of pre-existing carbonates, which were generally in good agree-
ment with furnace analysis. CO2 content determined by TG method
is compared with HCl titration method in Table 5. While TG and
furnace analyses are performed on a temperature range of
470–950 °C, HCl titration acts as a calibration technique indepen-
dent of temperature range selection. The obtained carbonates are
to be deducted from those of the carbonated samples to obtain
the CO2 uptake. It can be noted that 28 days after casting, the sam-
ples recorded similar CO2 uptake as to 1 day, confirming perma-
nent sequestration of CO2. In Table 4, the percent cement content
(CC) was used in calculating the mass of dry cement in powder
samples so that the comparison was on the same dry cement base.
Therefore, the furnace results in Tables 3 and 4 provide an addi-
tional comparison. The difference of the two tests is in that furnace
tests in Table 3 employed concrete sample mass, 35–70 g each,
representing a through-thickness average, while that in Table 4
used a much smaller mass, 10 g each, representing a 5-mm surface
layer. The CO2 content from both furnace tests appeared to be very
close.
Fig. 2 shows the 1-day and 28-day compressive strength of the
three concretes. The 24-h duration represents initial air curing fol-
lowed by 4-h carbonation and subsequent hydration to 24 h. For
Table 3
CO2 content and uptake in PLC concretes.
Sample CO2 uptake (mass%) Furnace CO2 Average CO2 uptake
content (mass%)
(mass%)
Mass gain Mass Surface Core
curve
0a (Ref) n/a n/a 6.68 6.54 0.00
18a + 4c 18.34 ± 1.6 18.23 26.14 25.42 18.45
18a + 4c + sp 18.29 ± 1.7 18.09 26.92 25.96 18.72
 H. El-Hassan, Y. Shao / Cement & Concrete Composites 62 (2015) 168–177 171

Table 4 and high heat release [15,16]. The hydrated reference does not dif-
Cement content (CC) in PLC concrete powder. fer between 1 and 28 days other than higher calcium hydroxide
Sample Test date Concrete chunk Concrete powder (TG CC peaks with 27 more days of hydration. Calcite peaks identified at
(Furnace method) method) 26.8° and 29.4° in the hydrated reference are associated to the
Mc (g) (DCO2)C (g) Mp (mg) (DCO2)P (mg) added limestone (15%) in the cement.
0a – Ref 1 day 9.21 0.08 120.60 3.14 0.38
Fig. 5 is a plot of the TG and DTG curves of the investigated
18a + 4c 1 day 10.54 0.37 119.90 8.32 0.26 batches at 1 day. According to Ramachandran and Beaudoin [13]
18a + 4c + sp 1 day 11.05 0.39 121.00 8.08 0.25 and the DTG curves obtained, the mass loss along TG curves can
0a – Ref 28 day 10.76 0.09 120.20 1.92 0.24 be classified into 6 categories: Below 105 °C, the mass loss is asso-
18a + 4c 28 day 10.32 0.35 121.50 7.20 0.23
ciated with the evaporable water and possibly the poorly formed
18a + 4c + sp 28 day 9.89 0.34 120.40 9.02 0.28
C–S–H. The water loss between 105 °C and 200 °C represents the
bound water in low temperature C–S–H and ettringite, while the
loss in the range of 200–420 °C is associated to the bound water
instance, sample 18a + 4c underwent 18 h of air curing followed by in well-formed hydration products including C–S–H and C–A–H.
4 h of carbonation and 2 h of subsequent hydration. At 1 day, it was Bound water in calcium hydroxide (CH) is decomposed between
obvious that carbonation was an accelerated hydration process 420 °C and 470 °C. While CO2 in poorly crystalline calcium carbon-
with 13.4 ± 0.9% higher strength in the carbonated concrete with ates designated by aragonite and vaterite are decomposed between
respect to its hydrated counterpart. Within 2 h of subsequent 470 °C and 680 °C, that in well crystalline calcite lies in the range of
hydration, water compensation improved the compressive 680–950 °C [13–17].
strength by 29.5 ± 2.1% in comparison to the non-sprayed carbon- The DTG curves of PLC concretes were compared: Drops
ated concrete, even though only part of the lost water was between 105 °C and 420 °C were common among all samples,
restored. while that in the range 420–470 °C was only present in the refer-
After 28 days, the carbonated concrete (18a + 4c) exhibited ence. Previous work suggested that this latter loss was due to the
lower compressive strength than the hydrated reference owing dehydration of CH [18], which is verified by the presence of CH
to the water loss in early air curing. Therefore, water spray after peaks in the XRD patterns. Between 470 °C and 680 °C, the carbon-
carbonation was essential to restore the lost water. Accordingly, ates formed in carbonated PLC concretes were progressively
a strength increase of 11.7 ± 0.8% was recorded after water com- decomposed as seen in the TG curves. However, the major drop
pensation. The results indicate that carbonation followed by water occurred in the well crystalline range beyond 680 °C. Calcium car-
compensation is beneficial. Early carbonation does not hinder sub- bonates generated by carbonation of calcium hydroxide were bet-
sequent hydration. It is the water loss during early air curing and ter crystalline and decomposed at temperature higher than 680 °C
carbonation curing that reduces degree of hydration. With water [19]. The hydration reference concrete also showed a similar, but
compensation through surface spray, it is possible to make carbon- less intense CaCO3 peak, due to added limestone.
ation strength comparable to hydration counterpart. Fig. 6 shows detailed mass loss based on TG/DTG analysis. It can
be noted, using the previously mentioned mass loss distribution,
that the hydrated reference had the lowest quantity of C–S–H
3.2. Microstructure analysis of PLC concrete
and C–A–H and exceeded by the carbonated counterparts. Due to
4-h carbonation, the anhydrous calcium silicates in PLC were con-
The 1-day XRD patterns for carbonated and hydrated PLC con-
sumed to form C–S–H, as carbonation is an accelerated hydration
cretes are plotted in Fig. 3. Calcium carbonate peaks including cal-
process [20]. This explains the increase in C–S–H and C–A–H con-
cite, vaterite, and aragonite are dominant and strong throughout
tents after carbonation. Further formation of C–S–H was identified
carbonated samples. In comparison to the hydrated reference
in the sprayed carbonated sample (18a + 4c + sp) after two hours of
(0a), higher carbonate peaks at 22°, 29.4° and 43° (2h) are present
subsequent hydration. CH formation was a direct indication of the
in the carbonated batches. A coexistence of hydration and carbon-
degree of hydration of any sample. The hydrated reference formed
ation products is noticed in the carbonated samples. Calcium
the highest amount of CH associated with no water loss during
hydroxide is distinctively identified at 18° in hydrated samples.
1-day curing. The sprayed carbonated sample (18a + 4c + sp)
However, with absence of this peak in the carbonated PLC concrete,
showed a comparable amount whilst being followed by its drier
it is conclusive that calcium hydroxide is consumed. In addition,
anhydrous calcium silicates are consumed, evidenced by the
decrease in peak intensity at 34.5° and 41.5°, indicating the exis-
tence of C3S. Ettringite has also been consumed verified by the lack
of the 52.5° peak after carbonation.
Fig. 4 shows the XRD patterns of the same PLC concretes after
28 days subsequent hydration. In comparison to the patterns
obtained after 1 day, the amount of vaterite and aragonite, shown
at 21.5° and 43°, have decreased. While aragonite is suggested to
form when the sample is allowed to dry out as in the case of
18a + 4c [6], vaterite forms due to the rapid carbonation reaction

Table 5
CO2 content and uptake in PLC concrete powder.

Sample CO2 content (mass%) Average CO2

uptake (mass%)
TG 1-day TG 28-day HCl 28-day
0a (Ref) 6.91 6.62 7.01 0.00
18a + 4c 26.71 26.06 25.98 19.40
18a + 4c + sp 26.92 26.54 26.55 19.82
Fig. 2. Compressive strength of PLC concretes.
172 H. El-Hassan, Y. Shao / Cement & Concrete Composites 62 (2015) 168–177

Fig. 3. XRD patterns of PLC concretes after 1 day.

Fig. 4. XRD patterns of PLC concretes after 28 days.

counterpart (18a + 4c). It is possible that, even with 0.8% loss in the
range of 420–470 °C, CH was not detected by XRD possibly due to
the presence of amorphous forms of CH [21]. Evidently, a sufficient
water supply promoted the hydration of unreacted calcium sili-
cates in the carbonated PLC concrete. Beyond 470 °C, the hydrated
reference showed some loss due to decarbonation of pre-existing
carbonates. The respective CO2 loss of carbonated concretes repre-
sented the CO2 content, and included the added limestone and pro-
duced carbonates. The mass loss between 470 and 680 °C,
representing the carbonates produced from carbonation of C–S–
H, showed the least thermal stability, suggesting poorly crystalline
calcium carbonate designated by aragonite and vaterite
[13–17,22]. The carbonates with better crystallinity, such as cal-
cite, required higher temperatures to decompose and were lost
above 680 °C [13].
Fig. 7 is a plot of the TG and DTG curves after 28 days subse-
quent hydration. The absence of the endotherm in the PLC con-
cretes between 120 and 130 °C was indicative of a probable
increase in stability of hydration products after 28 days. The mass
loss at 107 °C in the hydrated reference was possible evidence of
the formation of ettringite, which was also detected in the XRD
patterns. While the hydration reference concrete showed a
 H. El-Hassan, Y. Shao / Cement & Concrete Composites 62 (2015) 168–177 173

Fig. 5. DTG curves of 1-day PLC concretes.

Fig. 7. DTG curves of 28-day PLC concretes.

noticeable drop between 420 °C and 470 °C, the carbonated con-
cretes did not show similar results. This endotherm was associated
with the decomposition of CH.
Fig. 8 summarizes hydration and carbonation products of PLC
concretes at 28 days. The major change between 1 and 28 days
was the hydrated reference concrete, in which increase in mass
loss between 105 and 470 °C was indicative of increase in hydra-
tion products. Sealed curing conditions in hydration reference pre-
vented water evaporation, an essential factor for subsequent
hydration. The increase in C–S–H and C–A–H content led to an
increase in compressive strength as seen in Fig. 2. Similar behavior
was reported in the two carbonated concretes with higher losses
attributed to 28-day subsequent hydration. It was suggestive that
carbonation of PLC may slightly hinder the hydration of silicates
and aluminates, thus affecting the development of compressive
strength.

Fig. 8. Mass loss due to dehydration and decarbonation in 28-day PLC concretes.

Table 6
Percentage of carbonation-induced poorly and well crystalline calcium carbonates in
PLC concretes.

Sample Poorly crystalline Well crystalline

1-day 28-day 1-day 28-day
18a + 4c 35.4 23.0 64.6 77.0
18a + 4c + sp 33.6 28.0 66.4 72.0

In order to study the effect of subsequent hydration on the sta-

Fig. 6. Mass loss due to dehydration and decarbonation in 1-day PLC concretes.
bility of carbonates, the percentage of poorly crystalline and well
crystalline carbonates were calculated. Table 6 summarizes the
poorly crystalline (470–680 °C) and well crystalline (680–950 °C)
proportions of carbonation-induced carbonates in each batch.
After 1 day, 33–35% of the total 4-h reaction carbonates were
174 H. El-Hassan, Y. Shao / Cement & Concrete Composites 62 (2015) 168–177

Table 7
Total calcium hydroxide and calcium carbonate content (mass%) by TG in PLC
concretes.

Sample Calcium Calcium Calcium

hydroxide carbonatea carbonateb
1-day 28-day 1-day 28-day 1-day 28-day
0a (Ref) 3.81 10.08 4.02 3.34 11.68 11.71
18a + 4c 3.10 7.42 19.93 13.52 40.75 45.70
18a + 4c + sp 3.30 7.75 19.30 16.02 41.89 44.32
a
Poorly crystalline.
b
Well crystalline.

Fig. 11. SEM micrograph of sprayed PLC carbonated concrete (18a + 4c + sp) after
28 days.

Fig. 9. SEM micrograph of PLC hydrated reference (0a) after 28 days.

Fig. 10. SEM micrograph of PLC carbonated concrete (18a + 4c) after 28 days.

poorly crystalline, while 65–67% were well crystalline. After

28 days, the percent of well crystalline increased to 72–77%. It
was conclusive that unstable carbonates (aragonite and vaterite)
transformed into more stable polymorphs (calcite) in PLC
concretes.
Table 7 shows the CH and CaCO3 content in terms of cement
mass (mass%). The mass loss between 420 °C and 470 °C was asso-
ciated to the water in CH, while that between 470 °C and 950 °C
was attributed to the CO2 in calcium carbonate in all polymorphic
states. Accordingly, the hydrated reference produced the largest
amount of calcium hydroxide during 1 day and 28 days. Water
presence played a critical role in the development of CH.
However, carbonated concrete lost a significant amount of CH
through a constant mass loss shown in the TG curves. The CaCO3 Fig. 12. SEM micrograph of PLC powder.
 H. El-Hassan, Y. Shao / Cement & Concrete Composites 62 (2015) 168–177 175

Table 8
CO2 content and uptake in OPC concretes.

Sample CO2 uptake (mass%) Furnace CO2 content Average CO2 uptake
(mass%) (mass%)
Mass gain Mass
curve
0a (Ref) n/a n/a 1.85 0
18a + 4c 24.21 ± 1.8 24.44 27.17 24.63
18a + 4c + sp 25.73 ± 1.8 24.19 27.53 25.21

Table 9
Total calcium hydroxide and calcium carbonate content (mass%) by TG in OPC
concretes.

Sample Calcium Calcium Calcium

hydroxide carbonatea carbonateb
1-day 28-day 1-day 28-day 1-day 28-day
0a (Ref) 5.51 13.18 7.05 7.95 1.23 1.32 Fig. 13. One-day and 28-day compressive strength of OPC concretes.
18a + 4c 4.14 12.19 42.95 32.50 14.98 25.92
18a + 4c + sp 10.02 23.45 43.86 39.55 14.99 22.23
a
Poorly crystalline.
b
Well crystalline. development, and microstructure changes [9,10]. The CO2 uptake
of carbonated OPC is presented in Table 8. Carbonated OPC con-
crete (18a + 4c) had shown an uptake of 24.6 mass%, exceeding
Table 10 that of PLC by up to 6 mass%. The 15% cement substitution with
Percentage of carbonation-induced poorly and well crystalline calcium carbonates in limestone decreased the amount of carbonation reactants, includ-
OPC concretes.
ing C3S and C2S, and ultimately decreased the rate of reaction.
Sample Poorly Crystalline Well Crystalline With the exclusion of the added limestone and based on Portland
1-day 28-day 1-day 28-day cement content only, the uptake of carbonated concretes was
comparable at 22 mass%. Additionally and from Tables 7 and 9,
18a + 4c 73.1 48.4 26.9 51.6
18a + 4c + sp 72.9 57.2 27.1 42.8 the total amount of carbonates, pre-existing and
carbonation-induced, reached around 60% in both OPC and PLC
concretes respectively [9]. However, as seen in Table 6 for PLC
content in carbonated concrete in Table 7 included the originally and Table 10 for OPC, both carbonated concretes have experienced
existed and the carbonation-produced. It was 15–16 mass% in the phase transformation of carbonates, producing carbonates with
hydrated reference and 59–62 mass% in carbonation after 18-h higher degree of crystallinity (higher decomposition temperature)
air curing. Carbonation after air curing (18a + 4c and after 28 days subsequent hydration. Even with lower reactivity of
18a + 4c + sp) allowed the reaction to occur not only to anhydrous PLC with CO2 in early carbonation curing, the overall formation
silicates but to hydration products, mainly CH as well as C–S–H of carbonates in both concretes was similar. Carbonation of con-
[6,20]. crete masonry units has a significant environmental impact in
Fig. 9 shows the microstructure of a hydrated reference after terms of carbon emission reduction. Each 200-mm lightweight
28 days. Ettringite are incorporated alongside C–S–H. The XRD pat- concrete block weighs 15 kg with a cement content of 13% by mass.
terns of the hydrated reference provided supplementary evidence For OPC concrete, the total carbon content could reach 25% by
to this finding. Previous research resulted in similar observations cement mass and each block could sequester 0.49 kg of CO2. For
[3]. The microstructure of carbonated sample (18a + 4c) after PLC concrete, it was 0.36 kg of CO2 from carbonation curing in
28 days is presented in Fig. 10. It was observed that crystals have addition to the emission reduction due to the replacement of 15%
formed around ball-like shapes with dimensions on the order of clinker by limestone.
1 lm and 5 lm respectively. The compressive strengths of PLC and OPC concretes are shown
Water compensation after carbonation (18a + 4c + sp) aimed to in Figs. 2 and 13 respectively [9]. The compressive strength of
maximize CO2 uptake while maintaining high compressive hydrated OPC exceeded that of its PLC counterpart by 29 ± 5%.
strength. Fig. 11 shows the relatively denser microstructure after Although the mix design for both concretes was similar in propor-
28-day subsequent hydration with the formation of sharp, closely tions, 15% of cement in PLC was replaced by limestone, thus reduc-
packed crystals. As a result, the compressive strength exceeded ing the amount of hydration reactions. Accordingly, the degree of
that of the hydrated and other carbonated concrete at both subse- hydration and, more specifically, the compressive strength could
quent hydration durations. not compare to OPC. Similar results were seen in the carbonated
Examination of powder PLC in Fig. 12 showed that 10 lm parti- concretes. Smaller amounts of carbonation and hydration reactants
cles represent the ground limestone added to the cement in pro- led to a lower compressive strength.
duction phase. Larger irregular shapes (15–20 lm) were also The microstructure changes were also compared. Figs. 3 and 14
inspected and identified as a form of anhydrous calcium silicates, present the respective XRD patterns of PLC and OPC concretes after
either C3S or C2S. 1 day [9]. The patterns for both carbonated concretes identified
vaterite, aragonite, and calcite as carbonation products with the
consumption of calcium hydroxide. Furthermore, a comparison of
3.3. Comparison with ordinary Portland cement TG/DTG analysis of Figs. 6 and 8 with Fig. 15 showed that in both
concretes C–S–H and C–A–H developed after 28 days of hydration
It is interesting to compare PLC carbonated concretes with OPC [9]. The recorded mass loss at 420–470 °C, which was associated to
counterpart. The comparison includes CO2 uptake, strength CH decomposition, increased from 1 to 28 days, proving that
176 H. El-Hassan, Y. Shao / Cement & Concrete Composites 62 (2015) 168–177
Fig. 14. XRD patterns of OPC concretes after 1 day.
Fig. 15. Mass loss due to dehydration in 1-day and 28-day OPC concretes.
carbonation did not hinder the formation of CH. However, from
Tables 7 and 9, it was apparent that the amount of CH produced
in OPC exceeded that of PLC due to additional hydration reactants
in the former concrete.
Carbonated PLC concrete presented a crystalline microstructure
upon using SEM as shown in Fig. 11. On the other hand, OPC con-
cretes (Fig. 16) were more poorly crystalline, which is believed to
have contributed to higher compressive strength [9]. The lime-
stone particles ranging in size between 9 and 11 lm (Fig. 12)
may serve as nuclei for calcium carbonate growth. The size of car-
bonates in PLC concrete is in the range of 2–5 lm. They are well
crystalline, formed right after carbonation, and remain the same
size during subsequent hydration. Nevertheless, this
network-type distribution of carbonate crystals results in a more
porous microstructure. On the contrary, carbonated OPC concrete
(18a + 4c + sp) as in Fig. 16 has shown lower amount of crystalline
phases after 28 days subsequent hydration along with smaller
sized crystalline carbonates in general [9]. Even though the CO2
Fig. 16. SEM micrograph of OPC carbonated concrete (18a + 4c + sp) after 28 days.
uptake was similar, because of the much smaller carbonate parti-
cles with the dominant amorphous microstructure, carbonated
OPC concrete has gained more strength. The effect of cement par-
ticle size on carbonation degree was not apparent.
4. Conclusions
The carbonation curing of PLC concrete was studied. Its environ-
mental impact as well as its effect on the compressive strength and
microstructure of PLC concrete was examined. A comparison of
carbonated PLC and OPC concrete was also performed. Based on
the findings of this study, the following conclusions may be drawn:
1. In the presence of initial air curing, a CO2 uptake of about 19% in
terms of cement mass was achieved in a lightweight PLC con-
crete with expanded slag aggregates. It was corresponding to
a carbonation degree of 38%. Compared to 24% in the carbona-
tion of OPC concrete, a reaction efficiency of 48%, PLC showed
lower degree of reactivity due to lesser amounts of reactants.
 H. El-Hassan, Y. Shao / Cement & Concrete Composites 62 (2015) 168–177
2. The carbonated PLC has less carbonation reactants and the
strength was lower than OPC counterparts in this particular
concrete masonry mix with PLC.
3. Carbonation consumed calcium hydroxide, calcium silicates,
and ettringite to produce calcium carbonates in its three poly-
morphs. The XRD peaks showed evidence of this phenomenon.
Even though a mass loss between 420 and 470 °C was detected
in the carbonated PLC concrete, calcium hydroxide was not
identified by XRD.
4. The carbonation-induced calcium carbonates are distinguished
between poorly and well crystalline. For 18-h air cured and
4-h carbonated PLC concrete, 33–35% of the
carbonation-induced carbonates were poorly crystalline, while
65–67% remained well crystalline after 24 h curing. After
28 days, the well crystalline component was increased to
72–77%. It was conclusive that phase transformation of carbon-
ates occurred in PLC concrete during subsequent hydration. A
similar phenomenon was observed in carbonated OPC concrete.
Acknowledgements
The authors gratefully acknowledge the support by the Natural
Science and Engineering Research Council (NSERC) of Canada,
Canadian Concrete Masonry Producers Association (CCMPA), and
Lafarge Canada.
References
[1] I. Soroka, N. Setter, The effect of fillers on strength of cement mortars, Cem.
Concr. Res. 7 (4) (1977) 449–456.
[2] V. Bonavetti, H. Donza, G. Menéndez, O. Cabrera, E.F. Irassar, Limestone filler
cement in low w/c concrete: a rational use of energy, Cem. Concr. Res. 33 (6)
(2003) 865–871.
View publication stats
177
[3] V. Ramachandran, C. Zhang, Influence of CaCO3 on hydration and
microstructural characteristics of tricalcium silicate, Cemento 83 (3) (1986)
129–152.
[4] T. Matschei, B. Lothenbach, F.P. Glasser, The role of calcium carbonate in
cement hydration, Cem. Concr. Res. 37 (4) (2007) 551–558.
[5] S. Tsivilis, E. Chaniotakis, G. Kakali, G. Batis, An analysis of the properties of
Portland limestone cements and concretes, Cem. Concr. Compos. 24 (3) (2002)
371–378.
[6] C.J. Goodbrake, Reaction of beta-dicalcium silicate and tricalcium silicate with
carbon dioxide and water, J. Am. Ceram. Soc. (1979) 168–171.
[7] L.J. Parrott, Some effects of cement and curing upon carbonation and
reinforcement corrosion in concrete, Mater. Struct. 29 (1996) 164–173.
[8] S. Tsivilis, J. Tsantilas, The permeability of Portland limestone cement concrete,
Cem. Concr. Res. 33 (9) (2003) 1465–1471.
[9] H. El-Hassan, Y. Shao, Z. Ghouleh, Reaction products in carbonation cured
lightweight concrete, J. Mater. Civil Eng. 25 (6) (2013) 799–809.
[10] H. El-Hassan, Y. Shao, Z. Ghouleh, Effect of initial curing on carbonation of
lightweight concrete masonry units, ACI Mater. J. 110 (4) (2013) 441–450.
[11] ASTM C140, Standard Test Methods for Sampling and Testing Concrete
Masonry Units and Related Units, ASTM International, 2004.
[12] N.C. Collier, J.H. Sharp, N.B. Milestone, J. Hill, I.H. Godfrey, The influence of
water removal techniques on the composition and microstructure of hardened
cement pastes, Cem. Concr. Res. 38 (2008) 737–744.
[13] V.S. Ramachandran, J.J. Beaudoin, Handbook of Analytical Techniques in
Concrete Science and Technology, William Andre Publishing/Noyes, 2001.
[14] E.W.D. Huffman, Performance of a new automatic carbon dioxide coulometer,
Microchem. J. 22 (1977) 567–573.
[15] J.M. Bukowski, Reactivity and strength development of CO2 activated non-
hydraulic calcium silicates, Cem. Concr. Res. 9 (1979) 57–68.
[16] S. Goto, Calcium silicate carbonation products, J. Am. Ceram. Soc. (1995) 2867–
2872.
[17] D. Li, H. Shi, J. Deng, Y. Xu, Study on the thermal decomposition kinetics of
nano-sized calcium carbonates, J. Zhejiang Univ. 4 (3) (2003) 363–368.
[18] C. Chang, J. Chen, The experimental investigation of concrete carbonation
depth, Cem. Concr. Res. 36 (2006) 1760–1767.
[19] D.R. Moorehead, Cementation by the carbonation of hydrated lime, Cem.
Concr. Res. 16 (1986) 700–708.
[20] J.F. Young, R.L. Berger, J. Breese, Accelerated curing of compacted calcium
silicate mortars on exposure to CO2, J. Am. Ceram. Soc. 57 (9) (1974) 394–397.
[21] R.L. Berger, J.D. McGregor, Effect of temperature and water-solid ratio on
growth of Ca(OH)2 crystals formed during hydration of Ca3SiO5, J. Am. Ceram.
Soc. 56 (2) (1973) 73–79.
[22] G. Villain, G. Platret, Two experimental methods to determine carbonation
profiles in concrete, ACI Mater. J. 103 (4) (2006) 265–271.