Cement & Concrete Composites 62 (2015) 33–43

Contents lists available at ScienceDirect

Cement & Concrete Composites

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

Relation between carbonation resistance, mix design and exposure

of mortar and concrete
Andreas Leemann ⇑, Peter Nygaard, Josef Kaufmann, Roman Loser
Empa, Swiss Federal Laboratories for Materials Science and Technology, Überlandstr. 129, 8600 Dübendorf, Switzerland

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

Article history: When cement with mineral additions is employed, the carbonation resistance of mortar and concrete
Received 21 July 2014 may be decreased. In this study, mortars containing mineral additions are exposed both to accelerated
Received in revised form 8 April 2015 carbonation (1% and 4% CO2) and to natural carbonation. Additionally, concrete mixtures produced with
Accepted 23 April 2015
different cements, water-to-cement ratios and paste volumes are exposed to natural carbonation. The
Available online 14 June 2015
comparison of the carbonation coefficients determined in the different exposure conditions indicates that
mortar and concrete containing slag and microsilica underperform in the accelerated carbonation test
Keywords:
compared to field conditions. The carbonation resistance in mortar and concrete is mainly governed by
Carbonation
Accelerated testing
the CO2 buffer capacity per volume of cement paste. It can be expressed by the ratio between water added
Natural exposure during production and the amount of reactive CaO present in the binder (w/CaOreactive) resulting in a
Mix design novel parameter to assess carbonation resistance of mortar and concrete containing mineral additions.
Mortar Ó 2015 Elsevier Ltd. All rights reserved.
Concrete

1. Introduction binder-specific effects of curing on carbonation. Additionally, it

The increasing use of minerals additions, e.g., ground granu-
lated blast-furnace slag (GGBS), fly ash or limestone powder, to les-
sen the CO2 emission of cementitious binders affects the durability
of mortar and concrete. It can both improve and worsen their per-
formance depending on the type of interaction [e.g. 1–5]]. In regard
to carbonation resistance, a general decrease is observed when
cement clinker is replaced by mineral additions [6–11]. An
increased risk for rebar corrosion in reinforced concrete structures
due to carbonation of the concrete cover may result.
Accelerated tests can be used to assess the carbonation resis-
tance of mortar or concrete produced with cements containing
mineral additions. However, a wide range of CO2 concentrations,
different relative humidity conditions and temperatures are used
in such accelerated tests [8,10,12–14]. Therefore it is not clear,
whether the results obtained with one specific accelerated test
are transferable to another one and, even more important, to nat-
ural carbonation [15–20]. Another important point is the curing of
the samples before they are exposed to carbonation. Carbonation
resistance increases with increased time of curing [6,15,21,22], as
longer curing allows a higher degree of hydration of the binder.
However, the reactivity of cement clinker and mineral additions
like GGBS or fly ash can be quite different, leading to
⇑ Corresponding author. Tel.: +41 58 765 44 89.
E-mail address: andreas.leemann@empa.ch (A. Leemann).
has to be taken into account that the effect of curing is dependent
on the water-to-cement-ratio (w/c), with a stronger effect on mor-
tar and concrete produced with high w/c [23,24]. Such effects may
be the reason why no material parameter governing the carbona-
tion resistance of systems with different mineral additions has
been identified yet.
This study aims at empirically identifying a material parameter
(or a number of parameters) that defines the carbonation resis-
tance of mortar and concrete produced with cements containing
mineral additions. To ensure optimal curing, all samples were
moist cured for 28 days before starting exposure to CO2. In a first
series of experiments, mortars were produced. Their buffer capac-
ity and diffusivity were systematically changed by addition of
limestone powder, microsilica, portlandite and the use of GGBS
cement. The carbonation resistance was determined under acceler-
ated (1% and 4% CO2) and natural conditions (sheltered and unshel-
tered). Additionally, compressive strength and oxygen diffusion
were determined. Selected samples were further investigated with
mercury intrusion porosimetry (MIP) and scanning electron micro-
scopy (SEM), combined with energy dispersive X-ray spectroscopy
(EDX). In a second series of experiments, concrete mixtures pro-
duced with different w/c, cement type and paste volume were
exposed to natural carbonation for 2.5 years. Furthermore, com-
pressive strength and oxygen diffusion were measured.

http://dx.doi.org/10.1016/j.cemconcomp.2015.04.020
0958-9465/Ó 2015 Elsevier Ltd. All rights reserved.
34 A. Leemann et al. / Cement & Concrete Composites 62 (2015) 33–43
2. Materials and methods
2.1. Materials
2.1.1. Mortars
The composition of the ordinary Portland cement (OPC/CEM I
52.5 R), the GGBS cement (CEM III/B 42.5, GGBS content of 66–
80 mass-%) and the microsilica used for the mortars are given in
Table 1. The limestone powder used contains 92 mass-% of calcite,
5.5 mass-% of dolomite and minor amounts of quartz and mica. A
laboratory grade portlandite was used. The aggregate was standard
quartz sand with a grain size of 0–2 mm.
The mix design of the mortars (Table 2) was systematically var-
ied to study the effect of changing diffusivity and buffer capacity
on the carbonation resistance. By increasing the w/c of the mix-
tures produced solely with OPC and by replacing cement with
limestone powder, the diffusivity was increased and the buffer
capacity decreased. Cement was replaced with microsilica to
decrease both diffusivity and buffer capacity. The replacement of
cement with portlandite was expected to slightly increase the buf-
fer capacity and the diffusivity. The volume of paste was kept con-
stant for all mixtures and the water-to-binder ratio by mass (w/b)
was the same for all mortars containing mineral additions. The
names of the mortars (Table 2) and the concrete mixtures
(Tables 3 and 4) refer to the mineral additions used and their
approximate content in mass-%.
The mortars were produced according to EN 197-1 using a
Hobart mixer. From each mixture 15 mortar prisms
(40  40  160 mm3) and one cube (150  150  150 mm3) were
produced to measure compressive strength (3 prisms), carbonation
resistance (12 prisms) and oxygen diffusion (cube). All samples
were demolded after 24 h and subsequently moist cured (>95% rel-
ative humidity) until the age of 28 days.
After the moist curing, the mortar prisms used to test the car-
bonation resistance were conditioned at 20 ± 1 °C and 57 ± 2% rel-
ative humidity until the age of 56 days. Then, the upper and lower
sides of the prisms were coated with epoxy to allow CO2 ingress
only from the lateral cast surfaces. Subsequently, each of the four
sets consisting of three prisms were moved to one of the following
exposure conditions: carbonation chamber with 1% CO2 (20 °C, 57%
RH), carbonation chamber with 4% CO2 (20 ± 1 °C, 57 ± 2% RH),
sheltered outdoor exposure and unsheltered outdoor exposure.
The samples in the sheltered outdoor exposure were stored in a
shelter having one side entirely open. The unsheltered exposure
was on a roof where the samples were placed on steel profiles to
prevent capillary suction of water accumulated on the roof top.
Cores (diameter of 100 mm, height of 30 mm) were taken from
the cube after 28 days to measure oxygen diffusion. The cores were
conditioned at 20 ± 1 °C and 35 ± 2% RH for seven days and then
dried in an oven at 50 °C for 24 h before the measurement started.
The mass change during conditioning was recorded to obtain infor-
mation on the drying behavior of the mortars.
The samples for the investigation with the SEM were taken from
the mortar bars in the carbonation chamber after an exposure at 4%
CO2 for 35 days (mortars OPC-0.48, MS-15, P-15 and S-65). The
samples were dried in an oven at 50 °C for three days, epoxy
impregnated and polished.
Table 1
Composition of the two cements and the microsilica used for the mortars.
Material CaO SiO2 Al2O3 Fe2O3
CEM I 52.5 R 62.7 20.2 5.29 3.25
CEM III/B 42.5 L-LH HS 48.8 30.6 9.73 1.18
Microsilica 0.36 92.8 0.94 0.13
L.O.I. = loss on ignition.
n.a. = not analyzed.
The same mortars were used for the MIP analysis. Samples were
extracted with pincers from carbonated and non-carbonated areas
of mortar bars after an exposure in the carbonation chambers at 4%
CO2 for 35 or 91 days depending on the progress of carbonation.
Afterwards, the pieces were stored in isopropanol for 7 days to
remove the water and then dried at 50 °C for another 7 days.
The calculation of the reactive calcium oxide (CaOreactive) and
the reacted CaO (CaOreacted) in the mortars displayed in Table 2 is
explained in Section 4.
2.1.2. Concrete
Two ordinary Portland cements (CEM I 32.5 R, CEM I 42.5 N HS),
a cement containing 15 mass-% limestone powder (CEM II/A-LL), a
cement containing approximately 15 mass-% of limestone powder
and 20 mass-% low-calcium fly ash (CEM II/B-M (V-LL) 32.5 R) and
a GGBS cement (CEM III/B 32.5 N) were used to produce the con-
crete. The composition of the five cement types used is given in
Table 3. The mix design of the concrete is given in Table 4. The
cubes with a side length of 150 mm were demolded after 24 h
and subsequently moist cured (>95% relative humidity) until the
age of 28 days.
As the concrete mixtures were produced for another project
dealing with sulfate resistance [25], only one cube per mixture
was available for testing the carbonation resistance. At the age of
28 days, every cube was cut in the direction of casting in two
halves of identical dimensions. All sides except the cut surface
and the opposite cast surface were coated with epoxy. One half
of the cube was stored unsheltered on a roof and the other half
sheltered under a roof (identical storing conditions to the mortar
prisms).
At 28 days, cores (diameter of 100 mm, height of 50 mm) were
taken from one cube to determine oxygen diffusion. Conditioning
was identical to the mortar samples. The mass change during con-
ditioning was recorded to obtain information about the drying
behavior of the different concrete mixtures.
The calculation of the reactive calcium oxide (CaOreactive) and
the reacted CaO (CaOreacted) in the concrete mixtures shown in
Table 4 is explained in Section 4.
2.2. Methods
Compressive strength of the mortar prisms and the concrete
cubes were determined according to DIN EN 1015-11 [26] and
EN 12390-3 [27] respectively.
The procedure for the determination of carbonation coefficient
is based on SN 505 262/1 [28]. Before transferring the 12 mortar
prisms per mixture to the four different exposure conditions, the
initial carbonation depth was determined. An approximately
2.5 cm thick slice was split of the prisms. The freshly broken sur-
face was sprayed with phenolphthalein and photographed.
Carbonation depth was analyzed with image analysis on 10 points
per lateral cast surface. As three prisms were used for each of the
four different exposure conditions, the total number of measure-
ments per exposure condition, age and mixture was 60. The mea-
surements on the concrete were performed the same way. After
the initial measurement, the carbonation depth of the samples in
carbonation chambers was measured after an exposure of 7, 35
Cr2O3 MnO TiO2 P2O5 MgO K2O Na2O SO3 L.O.I.
0.01 0.06 0.29 0.22 1.65 0.90 <0.03 3.82 1.34
n.a. n.a. n.a. n.a. 5.19 0.59 0.29 1.43 n.a.
n.a. 0.03 0.01 0.05 0.74 1.09 0.39 0.54 2.66
 A. Leemann et al. / Cement & Concrete Composites 62 (2015) 33–43 35

Table 2
Mix design of the mortars.

Mortar Cement Mineral additions w/b Sand [kg/m3] CaOreactive [kg/m3] CaOreacted [kg/m3]
3
Type Mass [kg/m ] Type Mass
[kg/m3] [mass-%]
OPC-48 CEM I 518 – – – 0.48 1448 316 246
OPC-40 CEM I 575 – – – 0.40 1449 351 250
OPC-55 CEM I 471 – – – 0.56 1448 287 235
L-7.5 CEM I 480 Limestone 36.0 7.5 0.48 1447 293 228
L-15 CEM I 446 Limestone 66.9 15.0 0.48 1450 272 212
MS-7.5 CEM I 479 Microsilica 35.9 7.5 0.48 1448 292 228
MS-15 CEM I 445 Microsilica 66.8 15.0 0.48 1449 271 211
P-7.5 CEM I 476 Portlandite 35.7 7.5 0.48 1449 317 253
P-15 CEM I 441 Portlandite 66.1 15.0 0.48 1448 319 260
S-65 CEM III/B 506 – – – 0.48 1448 238 144

Table 3
Composition of the cements used for the concrete mixtures.

Cement Abbreviation used for concrete CaO SiO2 Al2O3 Fe2O3 MgO K2O Na2O SO3 L.O.I.a
CEM I 32.5 R OPC 64.1 20.7 4.7 3 2.2 0.9 0.2 2.4 0.9
CEM I 42.5 R HS OPC-HS 59.3 19.2 3.4 5.4 4 0.7 0.3 3.2 3.8
CEM II/A-LL 42.5 N L-15 61 17.8 4.3 2.7 1.7 1 0.2 2.8 7.8
CEM II/B-M 32.5 R L-FA-35 52.3 23.6 6.9 3.8 1.7 1.2 0.3 2.5 6.9
CEM III/B 32.5 N S-65 46.4 30.1 10 1.3 5.2 0.7 0.4 4.1 1.1
a
L.O.I. = loss on ignition.

Table 4
Mix design of the concrete.

Concrete /cement [–] Aggregate [kg/m3] ms/ga [–] Cement [kg/m3] Water [kg/m3] w/c [–] Addition [kg/m3] Paste volume [l/m3] CaOreactive [kg/m3] CaOreacted [kg/m3]
OPC 1929 0.54 335 150 0.45 1.3SPb 257 212 178
OPC 1929 0.54 315 157 0.50 0.3SP 257 199 175
OPC 1931 0.54 280 168 0.60 – 257 177 165
OPC 1799 0.54 400 180 0.45 – 307 253 212
OPC 1680 1.00 450 200 0.45 1.8SP + 1.1VMAc 343 284 239
OPC-HS 1927 0.54 335 150 0.45 1.3SP 257 189 159
OPC-HS 1927 0.54 315 157 0.50 0.6SP 257 178 157
OPC-HS 1929 0.54 280 168 0.60 – 257 158 147
OPC-HS 1678 1.00 450 200 0.45 2.3SP + 1.1VMA 343 254 214
L-15 1930 0.54 330 148 0.45 0.7SP 257 174 146
L-15 1931 0.54 310 155 0.50 0.3SP 257 163 143
L-15 1923 0.54 280 168 0.60 – 257 147 137
L-15 1679 1.00 445 197 0.45 1.8SP + 1.1VMA 343 234 197
L-FA-35 1924 0.54 325 146 0.45 1.0SP 257 143 120
L-FA-35 1927 0.54 305 152 0.50 0.3SP 257 134 118
L-FA-35 1922 0.54 275 165 0.60 – 257 212 112
L-FA-35 1677 1.00 435 193 0.45 2.2SP + 1.1VMA 343 191 160
S-65 1928 0.54 325 146 0.45 1.3SP 257 151 107
S-65 1930 0.54 305 152 0.50 0.6SP 257 142 106
S-65 1925 0.54 275 165 0.60 – 257 128 101
S-65 1794 0.54 390 174 0.45 – 307 181 129
S-65 1672 1.00 440 195 0.45 SP + VMA 343 204 145
a
ms/g = mass ratio between sand and gravel.
b
SP = polycarboxylate-based superplaticizer.
c
VMA = viscosity modifying agent.

and 91 days. The carbonation depth of the mortars exposed out-
doors was determined after an exposure of one year using the
same technique as for the samples stored in the carbonation cham-
bers. The measurements on the concrete were performed after
2.5 years (10 measurements per concrete). The carbonation coeffi-
cient K was determined by calculating the regression of the car-
bonation depth as a function of the square root of time (Eq. (1)):
ðdK  AÞ
K¼ p ð1Þ
t
where K is the carbonation coefficient, dK the carbonation depth in
mm, A the initial carbonation depth in mm after curing and t the
time in years.
The conditions during the entire outdoor exposure of the mor-
tars and part of the exposure of the concrete samples were
recorded in a meteorological station 50 m from the exposure site
(Fig. 1). The relative humidity was the highest during the winter
month, as there were several days per month with fog. During
these days the recorded relative humidity reached 100%. The low-
est amount of precipitation was recorded during the winter month.
36 A. Leemann et al. / Cement & Concrete Composites 62 (2015) 33–43
25
Temperature 160
Relative humidity
140
Relative humidity [%],
20 Precipitation
Temperature [°C]
precipitation [mm]
120
15 100
80
10
60
40
5 The cement replacement with limestone powder, microsilica and
20
0 0
Fig. 1. Monthly means of temperature, relative humidity and precipitation during
the outdoor exposure of the mortar and partly of the concrete samples (exposed
2.5 years).
The oxygen diffusion coefficient DO was measured as described
in [29–31] on two cores (diameter of 100 mm, height of 30 mm) for
the mortar mixtures and on three cores (diameter of 100 mm,
height of 50 mm) for the concrete mixtures. An oxygen flow was
applied on one side of the cores and a nitrogen flow on the other
side. The gas pressure on both sides of the cores was identical.
The oxygen content in the nitrogen flow was measured until equi-
librium was reached. Afterwards, the oxygen diffusion coefficient
was calculated according to [28].
MIP was performed with a maximum pressure of
pmax = 395 MPa. The pore structure derived from the first intrusion
was used to characterize the mortars [32].
The analysis of the hydrates in the non-carbonated and the car-
bonated areas of the mortars were conducted with an environmen-
tal scanning electron microscope (ESEM-FEG XL30). The samples
were studied in the high vacuum mode (2.0–6.0 106 Torr) with
an accelerating voltage of 12 kV and a beam current of 270–
280 lA. The composition of the hydrates was determined with
energy dispersive X-ray spectroscopy (EDX). An EDAX 194 UTW
detector, a Philips digital controller and Genesis Spectrum
Software (Version 4.6.1) with ZAF corrections was used. A total
of 160 points were analyzed with EDX in the non-carbonated
and carbonated area of each studied mortar.
3. Results
3.1. Compressive strength of mortar and concrete
The compressive strength increases with decreasing w/b of the
OPC mortars (Table 5). The replacement of cement with limestone
powder, portlandite and GGBS leads to decrease of compressive
strength at 28 days, while the replacement with microsilica
increases it. After 91 days, mortar S-65 reaches the same compres-
sive strength as mortar OPC-48.
The concrete mixtures show the usual trend of increasing com-
pressing strength with decreasing w/c (Table 6). The compressive
strength at a given w/c decreases a few MPa when cements con-
taining mineral additions are used.
3.2. Oxygen diffusion of mortar and concrete
The decrease of w/b of the mortars OPC leads to a decrease of
the oxygen diffusion coefficient DO2 (Table 5). The use of limestone
powder and portlandite increases the values, while the use of
microsilica and GGBS decreases them.
Oxygen diffusion of the concrete shows the same
cement-specific relation to w/c as compressive strength
(Table 6). Concrete S-65 displays significantly lower values com-
pared to the other mixtures at equal w/c.
3.3. Carbonation coefficients of mortar and concrete
Decreasing the w/b in the OPC mortars decreases the carbona-
tion coefficient K determined in accelerated and natural conditions.
GGBS generally leads to more pronounced increases of the carbon-
ation coefficient K than the use of portlandite.
Concrete OPC and OPC-HS show the lowest carbonation coeffi-
cients. The values increase with increasing clinker substitution by
mineral additions. Consequently, the highest carbonation coeffi-
cients KN are recorded for concrete S-65. The carbonation coeffi-
cients KN,S in the sheltered conditions are higher than KN,US in the
unsheltered conditions for the cast and the cut surfaces of the con-
crete mixtures. The values for the cut surface are lower than for the
cast ones. As the carbonation coefficients are based on only 10 mea-
surements per concrete mixture and condition, a certain variability
of the values has to be expected. The mean of the coefficient of vari-
ation for the carbonation coefficients is 14.7% (cast and cut surface
in sheltered and unsheltered exposure of all concrete mixtures).
3.4. Porosity of mortar
The total porosity and the median pore size of the mortar with
portlandite are the highest, while the use of microsilica and GGBS
lead to a decrease of both parameters (Table 7). Carbonation causes
a decrease in the total porosity of all mortars except mortar S-65,
where an increase is observed (Fig. 2). The median pore size
increases significantly with carbonation in mortars MS and S,
whereas only slight changes are evident in the other mortars.
However, when not only the median Md but the entire pore size
distribution is considered, a marked increase in the volume of
pores >100 nm is observed in all mortars (Table 7). Mortars
OPC-48 and S-65 clearly illustrate these effects of carbonation on
the pore structure (Fig. 2).
3.5. Microstructure of the mortars
When the microstructure of the mortars is assessed qualitatively
with SEM, a densification is observed, with decreased porosity due
to carbonation. An exception is mortar S-65, where an increase of
porosity is visible (Fig. 3). The EDX point analysis of the cement
paste shows the same trend in regard to the Si/Ca-ratio in all stud-
ied mortars. In the non-carbonated paste, C–S–H displays a molar
Si/Ca-ratio that ranges from 0.67 (mortar MS-15) to 0.37 (mortar
P-15), depending on the mineral addition used. Due to the size of
the interaction volume of the EDX point analyses (1–3 lm) at the
chosen acceleration voltage, there is some intermixing of the C–
S–H with the mineral additions. However, it is still possible to show
the effect of carbonation on the mineral assemblage. In the carbon-
ated areas of the mortars, a higher variation in the Si/Ca-ratio is
observed, with both a trend away from the initial C–S–H composi-
tion towards calcite and a trend towards a silicon-rich phase. This
can be illustrated by the example of mortar MS-15 (Fig. 4).
4. Discussion
4.1. Accelerated and natural carbonation
The carbonation coefficients KACC of the mortars determined in
the carbonation chambers with 1% and 4% CO2 show a linear
 A. Leemann et al. / Cement & Concrete Composites 62 (2015) 33–43 37

Table 5
Results of the mortar mixtures.

Mortar OPC-48 OPC-40 OPC-55 L-7.5 L-15 MS-7.5 MS-15 P-7.5 P-15 S-65
fc,7 [MPa] 37.8 51.8 26.3 33.3 30.7 36.6 35 33.5 31.3 14
fc,28 62.2 74.7 50.9 57.2 51.6 67.6 69.2 53.8 49 52.9
fc,91 69.7 78.8 54.9 62.6 57.2 77.2 79.0 60.6 54 69.1
DO2 [108 m/s2] 1.92 1.24 2.02 2.39 3.00 0.72 0.50 2.08 2.45 0.19
Dm [mass-%] 0.97 0.70 1.25 0.98 1.21 0.70 0.60 0.94 1.23 0.80
p
KACC,1% [mm/ y 6.44 4.29 10.43 8.37 9.37 9.03 10.29 7.39 7.85 15.71
KACC,4% 10.34 5.25 19.34 13.17 18.46 13.54 16.05 10.52 12.29 28.67
KN,S 0.87 0.49 1.63 1.42 1.69 1.24 1.35 0.91 1.22 1.46
KN,US 0.32 0.15 0.90 0.98 1.28 0.42 0.50 0.46 0.70 0.55

fc,X = compressive strength after X days.

DO2 = oxygen diffusion coefficient.
Dm = mass loss during conditioning for 7 days at 20 °C and 35% RH
KACC,X% = coefficient of carbonation in carbonation chamber with X% CO2
KN = coefficient of carbonation in sheltered (S) and unsheltered (US) natural conditions

Table 6
Results of the concrete mixtures.

Concrete/cement w/c [–] Paste volume flow fc,28 DO2 Dm KN,S-CA KN,S-CUT KN,US-CA KN,US-CUT
p
[l/m3] [cm] [MPa] [108 m/s2] [mass-%] [mm/ y]
OPC 0.45 257 59 56.3 0.87 0.6 2.1 1.1 0.5 0.5
OPC 0.50 257 51 45.8 1.22 0.79 2.9 1.8 0.5 0.5
OPC 0.60 257 52 38.9 1.76 0.83 2.1 1.6 0.5 0.5
OPC 0.45 307 62 46.6 1.47 0.69 2.4 1.6 0.5 0.5
OPC 0.45 343 70 55.5 0.84 0.67 4.3 3.4 1.6 0.9
OPC-HS 0.45 257 42 58.7 1.24 0.62 4.4 3.0 1.0 0.6
OPC-HS 0.50 257 46 47.2 2.26 0.79 3.7 3.2 2.2 1.1
OPC-HS 0.60 257 49 34.0 4.03 1.10 6.2 4.8 3.6 2.6
OPC-HS 0.45 343 50 56.7 1.51 0.76 3.1 1.9 0.5 0.5
L-15 0.45 257 42 49.7 0.97 0.67 3.7 3.5 2.0 0.9
L-15 0.50 257 50 42.4 1.38 0.69 4.1 3.1 2.5 1.7
L-15 0.60 257 56 33.2 2.37 0.99 4.7 3.4 2.2 1.6
L-15 0.45 343 55 51.7 1.17 0.89 8.4 5.2 3.9 3.0
L-FA-35 0.45 257 49 44.3 1.20 0.71 5.4 4.0 2.4 1.4
L-FA-35 0.50 257 48 38.2 1.57 0.83 5.1 3.0 1.7 1.3
L-FA-35 0.60 257 54 28.2 3.01 0.91 6.1 4.1 3.4 2.2
L-FA-35 0.45 343 66 46.3 1.69 0.90 6.6 6.5 5.4 3.3
S-65 0.45 257 45 56.2 0.21 0.3 4.8 4.1 3.0 2.1
S-65 0.50 257 43 43.3 0.62 0.50 5.0 3.7 2.0 1.1
S-65 0.60 257 46 33.0 0.78 0.79 6.0 4.6 2.6 2.0
S-65 0.45 307 43 47.2 0.47 0.51 8.1 6.0 2.3 1.5
S-65 0.45 343 53 57.9 0.32 0.47 8.7 6.7 4.7 3.4

fc,28 = compressive strength after 28 days.

DO2 = oxygen diffusion coefficient.
Dm = mass loss during conditioning for 7 days at 20 °C and 35 RH.
KN = coefficient of natural carbonation in sheltered (S) and unsheltered (US) of the cast (CA) and cut (CUT) surfaces.

Table 7 45
Porosity determined with MIP of selected non-carbonated (nc) and carbonated (c)
OPC-48, nc
samples. 40 OPC-48, c
S-65, nc
Pore volume [mm3/g]

Mortar OPC-48 L-15 MS-15 P-15 S-65 35

S-65, c
Total porosity Unc 3
[mm /g] 44.3 46.5 39.7 56.5 38.3 30
Porosity > 100 nm Unc,100 [mm3/g] 2.9 3.8 5.4 4.7 2.3 25
Median Mdnc [nm] 31.0 30.7 15.4 34.2 7.1
Total porosity Uc [mm3/g] 27.4 38.3 27.3 37.0 45.1 20
Porosity > 100 nm Unc,100 [mm3/g] 8.8 10.7 14.1 10.7 20.2 15
Median Mdc [nm] 39.7 43.3 111.0 32.3 84.8
10
5
0
1 10 100 1000
relationship between each other (Fig. 5A). As a consequence, the
Pore radius [nm]
ranking of the different mortars in order of carbonation resistance
is practically the same at these two different CO2 concentrations. Fig. 2. Pore size distribution of the mortars OPC-48 and S-65 in the non-carbonated
There is still a reasonable correlation when KACC,1% is compared to (nc) and carbonated (c) state.
the carbonation coefficient KN,S (Fig. 5B). However, there is change
in the ranking of the mortars. The carbonation coefficient KN,S of conditions at 1% CO2. However, the most pronounced change con-
the mortars MS is lower than the one of the mortars L and mortar cerns mortar S-65. Its carbonation coefficients KACC at 1% and 4%
OPC-55. The ranking was the opposite under accelerated CO2 were considerably higher than the ones of the other mortars.
38 A. Leemann et al. / Cement & Concrete Composites 62 (2015) 33–43

A 30
OPC L

Carbonation coefficient K ACC,4%

25 MS P
S
20

[mm/y1/2]
15

10

0
0 5 10 15
Carbonation coefficient K ACC,1% [mm/y1/2]

B 18
OPC L
16

Carbonation coefficient K ACC,1%

MS P
14 S
12
10

[mm/y1/2]
8
6
4
2
0
0.0 0.5 1.0 1.5 2.0
Carbonation coefficient K N,S [mm/y1/2]

Fig. 5. (A) Carbonation coefficient KACC at 4% CO2 versus carbonation coefficient KACC
at 1% CO2 (R2 = 0.96). (B) Carbonation coefficient KACC at 1% CO2 versus carbonation
Fig. 3. Non-carbonated (A) and carbonated (B) cement paste in mortar S-65.
coefficient KN,S in the sheltered outdoor exposure (R2 = 0.69).

In the sheltered outdoor exposure its carbonation coefficient KN,S is
in the same range as the one of the mortars OPC-55 and L-15. The
reason for this change could be the lower CO2 content in natural
conditions or the higher relative humidity with yearly mean of
76% compared to 57% in the carbonation chambers. In general, car-
bonation is slowed down with increasing relative humidity at
levels above 60% [6,33–35]. To asses this, the data in the unshel-
tered conditions have to be considered (Fig. 6). Due to the exposure
to snow and rain, the moisture content of the mortars in the
unsheltered exposure has to be higher compared to the one in
the sheltered one. As a result, the carbonation coefficient KN of
the unsheltered mortar specimens is lower by a factor of about
1.5 (Table 5). More important, there is a further change in the rank-
ing of the mortars. The mortars MS now nearly reach the level of
mortar OPC-48. But again, the most marked change is observed
for mortar S-65. Its carbonation coefficient KN,US is now lower than
the values of mortars L, mortar P-15 and mortar OPC-55. It is clear
that the higher relative humidity slows down the rate of carbona-
tion [6,33–35]. But these data clearly indicate that it is additionally
the parameter responsible for the change in the rankings of the
mortars. It has a stronger impact on the carbonation of the mortars
MS and S-65 than on the other mortars. As a result of the increased
moisture content, the correlation between the carbonation coeffi-
cients KACC,1% and KN,US is worsening (Fig. 6).

0.3 18
non-carbonated OPC L
Carbonation coefficient K ACC,1%

carbonated 16
0.25 MS P
14 S
0.2 12
[mm/y1/2]
Al/Ca

10
0.15
8
0.1 6
4
0.05
2

0 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 0.0 0.2 0.4 0.6 0.8 1.0 1.2
Si/Ca Carbonation coefficient K N,US [mm/y1/2]

Fig. 4. Molar Al/Ca-ratio as a function of molar Si/Ca-ratio in non-carbonated area Fig. 6. Carbonation coefficient KACC at 1% CO2 versus carbonation coefficient KN,US in
and the carbonated area of mortar MS-15. the unsheltered outdoor exposure (R2 = 0.15).

The lower carbonation coefficient KN of the cut compared to the
cast concrete surfaces (Table 6) can be likely attributed to optimal
conditions for hydration of the cut surfaces whereas the cast sur-
faces may dry to a certain degree. Moreover, the probability for
segregation and the accumulation of bleeding water leading to a
local increase of w/c is higher on the cast surfaces. Therefore, the
values from the cut surfaces are used in the following comparisons.
The coefficients of variation calculated for the carbonation coeffi-
cients of the concrete mixtures as described in paragraph 3.3 seem
to be in a reasonable range for the number of measurements allow-
ing such comparisons.
Similar to the mortars there is a change in the ranking of the
concrete mixtures going from sheltered to unsheltered exposure.
The concrete mixtures S-65 show a more pronounced decrease of
the carbonation coefficient KN,US than the majority of the other con-
crete mixtures (Fig. 7). It is less pronounced than in the mortars but
still obvious. In the case of concrete L-FA-35 the relatively low fly
ash content of 15 mass-% seems not to be sufficient to cause the
same behavior as observed in the other mixtures containing silic-
eous mineral additions. One possible reason for the behavior of
mixtures containing GGBS and microsilica is their finer pore struc-
ture (Table 7) [36,37]. Capillary condensation depends on relative
humidity and pore size: the smaller the pore size, the lower the rel-
ative humidity at which condensation occurs. As a result, a mate-
rial with a relatively high amount of fine pores may have a
higher amount of water filled pores due to condensation than a
material with a coarser pore structure at identical relative humid-
ity. Transferring this to the mortar and concrete mixtures S-65 and
MS, a higher amount of water filled pores could slow down carbon-
ation in these mixtures. The second possible reason for the
A 1.4
OPC L 4.2. Relation of carbonation with physical concrete parameters
1.2 MS P
Carbonation coefficient K N,US
S The carbonation coefficient appears to be correlated to the com-
1.0
0.8
[mm/y1/2]
0.6
0.4
0.2
0.0
0.0 0.5
Carbonation coefficient K
B 4
OPC
Carbonation coefficient K N,US-CUT
OPC HS
3 L-15
L-FA-35
[mm/y1/2]
S-65
2 sion coefficients show the same relative changes as a function of
1 transformation from portlandite to calcite accompanied by decalci-
0
0 2
Carbonation coefficient K N,S-CUT [mm/y1/2]
Fig. 7. Carbonation coefficient KN of the mortar (A) and the concrete mixtures (B) in
the sheltered and unsheltered outdoor exposure (A: R2 = 0.65, B: 0.83). The line
indicates a ratio between x and y of 1:1.
A. Leemann et al. / Cement & Concrete Composites 62 (2015) 33–43 39
comparatively high decrease of the carbonation coefficient in the
sheltered and unsheltered exposure could be the drying behavior
after a period of high humidity caused by fog or rain. The mortars
S-65 and MS and concrete S-65 dry relatively slowly during the
conditioning for the oxygen diffusion and additionally have very
low oxygen diffusion coefficients DO2 (Tables 6 and 7). Therefore,
it can be expected that the moisture level at the depth of the car-
bonation front remains high for a longer time and as a consequence
carbonation would become slower than in a faster drying mortar or
concrete.
Moreover, it has to be mentioned that the preconditioning for
oxygen diffusion and the following diffusivity measurement was
started at an age of 28 days. It can be expected that the diffusivity
of mortar and concrete S-65 decreases more with time than the
one of the other mixtures. This likely increases the difference to
the other mixtures. A slower water loss due to evaporation of con-
crete containing GGBS and concrete containing fly ash compared to
pure OPC concrete was observed as well by Parrot [38,39]. It is
expected that higher amounts of fly ash than those used in this
study may cause a similar behavior to the one observed for the
mixtures S-65 and MS.
These results have implication for the testing of accelerated car-
bonation. Rebar corrosion due to carbonation is a problem in
unsheltered exposure, where there is a sufficient level of humidity
to support the reaction [40,41]. These are exactly the conditions
where the correlation between carbonation coefficients KACC and
KN,US is poor (R2 = 0.15 for the mortars/Fig. 7), mainly due to the
behavior of mortars S-65 and MS. It is likely that the performance
of such mortar and concrete in real structures will be underesti-
mated based on results obtained in accelerated carbonation tests
conducted at a relatively low relative humidity.
pressive strength when the mortars containing siliceous additions
are not taken into account (Fig. 8). However, when all mortars are
considered, there is no clear relation. The same applies to the con-
crete mixtures where a cement-specific relation to compressive
strength and w/c is observed (Fig. 9). On the contrary, no correla-
tion is evident when all mixtures are considered.
Oxygen diffusion shows a comparable relation to carbonation as
compressive strength. Although there is a cement-specific relation,
the carbonation coefficient KN and the oxygen diffusion coefficient
1.0 1.5 DO2 of mortar and concrete (Fig. 10) as a whole do not correlate.
[mm/y1/2] Concrete S-65 for example exhibits the lowest oxygen diffusion
N,S
coefficient DO2, but its carbonation coefficients KN,S are the highest.
On the other side, concrete OPC and OPC-HS have the highest oxy-
gen diffusion coefficients DO2 for a given carbonation coefficient.
Unfortunately, no published data are available on the relation-
ship between concrete mix design and CO2-diffusion. It is only
clear that CO2 diffusion in cement-based material is slower than
oxygen diffusion [42]. But it is not clear, whether O2 and CO2 diffu-
concrete mix design.
In the carbonated layer, a coarsening of the pore structure and a
fication of C–S–H is observed in the studied mortar samples. This
process goes together with a decrease of total porosity in all mor-
4 6 tars except for mortar S-65, where an increase is observed. The
coarsening of the pore structure seems to dominate the effect on
diffusivity: the oxygen diffusion coefficient DO2 is increased by car-
bonation in cement pastes produced with pure OPC, with OPC
combined with fly ash and OPC combined with GGBS [43]. Such
an acceleration in transport due to carbonation is observed as well
40 A. Leemann et al. / Cement & Concrete Composites 62 (2015) 33–43

18 The CaOreactive of the binders used for the mortar and concrete
OPC L
Carbonation coefficient KACC,1%

16 mixtures is given in Tables 2 and 3. In the case of CEM I 52.5 used

MS P for the mortars, 1.7 mass-% CaO were subtracted, taking into
14 S account a limestone powder content of 3 mass-% (based on infor-
12 mation from cement producer). Concerning the cements used for
[mm/y1/2]

10 the concrete mixtures, 0.6 mass-% of CaO was subtracted for a

8
6 in the case of CEM I 42.5 HS (information from the cement pro-
4 ducer). In the concrete mixtures produced with CEM II/A-LL and
2
0 Now, the w/CaOreactive can be compared with the determined
50 60 70
Compressive strength 91 Tage [MPa]
Fig. 8. Carbonation coefficient KACC at 1% CO2 as a function of compressive strength
of the mortar at 91 days. The ellipsoid marks the mortars with siliceous mineral
additions.
in the water permeability of pastes produced with blended
cements [44].
Service life modelling in regard to carbonation has to take into
account various parameters [45–48], including concrete quality. It
is obvious from the results of this study, that compressive strength
and w/c alone are inadequate to characterize the behavior of con-
crete containing mineral additions.
4.3. Relation of carbonation with the CO2 buffer capacity
The carbonation of mortar and concrete can be regarded as an
attack by CO2. The intruding CO2 reacts with the CaO of the
hydrates to form calcite and silica-rich gels [49]. The higher the
reactive CaO content (CaOreactive) per volume of paste is, the more
CO2 per volume of paste can be captured and the slower the car-
bonation front should progress into the material. The CaOreactive
per volume of cement paste can be calculated from the composi-
tion of the binder and the w/c. The CaO content of cement and min-
eral additions is measured by X-ray fluorescence, a standard
procedure in the characterization of these materials. If cement con-
tains limestone powder, the amount of CaO present in the lime-
stone powder has to be subtracted as it is not reactive.
Consequently, the CO2 buffer capacity per volume of paste and
with it the carbonation resistance of mortar and concrete can be
expressed by the ratio between water and CaOreactive
(w/CaOreactive). Of course, the CO2 can only react with CaO that is
present in hydrates and not the one present in anhydrous cement
or mineral additions. This effect of the degree of hydration will be
discussed later in this section.
7
Carbonation coefficient K N,S-CUT
limestone powder content of 1.0 mass-% in the case of CEM I
32.5 and 2.8 mass-% of CaO for a limestone content of 5.0 mass
CEM II/B-M (V-LL) corrections were made for the 15 mass-% of
limestone powder they contain.
80
carbonation coefficients. There is a clear correlation between the
w/CaOreactive and the carbonation coefficients KACC determined at
1% and 4% CO2 (R2 = 0.89 and 0.91). Both, the carbonation coeffi-
cient KN,S of the mortar and the concrete mixtures exposed to nat-
ural carbonation in the sheltered exposure correlate with the
w/CaOreactive (R2 = 0.83 and 0.87/Fig. 11). In the unsheltered expo-
sure with an increased moisture content in the material the corre-
lation is getting worse for the mortars (R2 = 0.29) and stays about
the same for the concrete mixtures (R2 = 0.84).
So far, the degree of hydration has not been taken into account.
The w/CaOreactive implies that clinker and mineral additions have
hydrated completely, with no anhydrous material left. Of course,
this is not the case for the studied mortar and concrete mixtures
even if they have been cured for 28 days and the hydration has
likely proceeded further afterwards. As shown by the effect of cur-
ing [6,15,21,22], the degree of hydration plays a major role in the
carbonation resistance. However, it is difficult to assess it for mor-
tar and concrete exposed to carbonation. First, there are data in lit-
erature for the degree of hydration of cement pastes but only little
of mortar and especially of concrete [50–54]. Secondly, the degree
of hydration is likely changing during exposure [20], since suffi-
cient moisture is provided especially in the unsheltered exposure
conditions. Assessing the degree of hydration is especially difficult,
when two or more components like cement clinker, GGBS and fly
ash with specific hydration kinetics are present. Thirdly, when con-
crete is cured for a short time and then exposed to drying, a pro-
nounced gradient of the degree of hydration can be expected
with increasing depth. This makes it impossible to accurately
assess the w/CaOreacted at the carbonation front.
In spite of these uncertainties, the degree hydration is esti-
mated here based on data from literature and applying some sim-
plifications. This at least allows judging its principle effects. Data
from Parrot and Killoh [50] are used for the cement clinker. For
the mortar used in the accelerated carbonation tests, a degree of
hydration of 75% (w/c = 0.40), 82% (w/b = 0.48) and 86%
7
Carbonation coefficient K N,S-CUT

6
6
5
5
[mm/y1/2]

4
[mm/y1/2]

4
3
3 OPC
OPC
OPC HS 2 OPC HS
2
L-15 L-15
1 L-FA-35 1 L-FA-35
S-65 S-65
0 0
0.40 0.45 0.50 0.55 0.60 0.65 0 1 2 3 4
w/c Oxygen diffusion coefficient D O [10-8 m2/s]

Fig. 9. Carbonation coefficient KN of the cut concrete surfaces in the sheltered Fig. 10. Carbonation coefficient KN of the cut concrete surfaces in the sheltered
exposure as a function of w/c. exposure as a function of the oxygen diffusion coefficient DO2.
 A. Leemann et al. / Cement & Concrete Composites 62 (2015) 33–43 41

A 1.6 (R2 = 0.87 and 0.85), it worsens in the sheltered (Fig. 12A) and more
OPC L so in the unsheltered exposure (R2 = 0.46 and 0.06). The concrete
1.4
mixtures show a better correlation in the sheltered exposure
Carbonation coefficient KN,S

MS P
1.2 S (Fig. 12B) than in the unsheltered one (R2 = 0.79 and 0.61).
Compared to the w/CaOreactive, the correlation of the w/CaOreacted
1.0
with the carbonation coefficients K is the same in the accelerated
[mm/y1/2]

0.8 testing, but it is worse in the sheltered and unsheltered outdoor

exposure. Assuming complete hydration, the w/CaOreactive in the
0.6
mixtures containing GGBS deviates the strongest from reality.
0.4 However, the good correlation indicates that this inaccuracy is
0.2
partly compensated by the effect of the increased moisture level
in the outdoor exposure. The influence of the CO2 buffer capacity
0.0 on carbonation is possibly lessened with increasing moisture con-
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
tent when mixtures with GGBS and microsilica are used.
w/CaOreactive [-] Nevertheless, the CO2 buffer capacity appears to be the govern-
ing parameter for the carbonation resistance of mortar and con-
B 8
crete considering the entire data set. Because of the uncertainties
OPC
related to the assessment of the degree of hydration, it seems to
Carbonation coefficient KN,S-CUT

OPC HS
L-15
be preferable to express the buffer capacity by the w/CaOreactive.
6
L-FA-35 This parameter can be a useful tool to assess the carbonation resis-
S-65 tance of different mix designs with variations in cement type, paste
[mm/y1/2]

volume and w/c. Its reliability can be worsened at a high moisture

4
level of mortar or concrete stored in unsheltered exposure, because
capillary condensation and drying behavior as physical parameters
are of increased importance under these conditions like discussed
2
in Section 4.1. Additionally, it has to be pointed out that the mortar
and concrete samples in this study have been moist cured for
28 days. The correlations shown above may become less accurate
0
0.6 0.8 1.0 1.2 1.4 in the case of shorter curing.
w/CaOreactive [-] The generally good correlation of both the w/CaOreactive and the
w/CaOreacted with the carbonation coefficients KACC and KN,S gives
Fig. 11. Carbonation coefficient KN,S in the sheltered outdoor exposure as a function
of the w/CaOreactive of the mortar (A: R2 = 0.79) and of the concrete mixtures (B: A 1.6
R2 = 0.87).
1.4
Carbonation coefficient KN,S

(w/c = 0.55) is assumed. For the mortars with mineral additions 1.2
(w/b = 0.48) the same degree of clinker hydration is assumed as
1.0
for the pure OPC sample. This is a simplification as the degree of
[mm/y1/2]

clinker hydration can increase with the addition of 0.8

slower-reacting mineral additions [50]. Because the CaO content
0.6
of the microsilica is very low, it was not considered. The degree
of hydration of GGBS is difficult to assess, as it depends on the ratio 0.4 OPC L
of cement clinker to GGBS, on the specific batch of GGBS and on MS P
0.2
age [51–54]. Here, it was assumed that the degree of reaction of S
GGBS is lower than the one of the clinker by a factor of 2.0 [51– 0.0
54]. Although the degree of hydration in the mortars exposed out- 0.9 1.1 1.3 1.5 1.7
doors can be expected to be higher than in the accelerated test, the w/CaOreacted [-]
same values were assumed to simplify.
B 8
In the case of concrete, average degrees of hydration for the
OPC
cement clinker during the natural exposure of 2.5 years of 84%
Carbonation coefficient KN,S-CUT

OPC HS
(w/b = 0.45), 88% (w/b = 0.50) and 93% (w/c = 0.60) were assumed,
6 L-15
independently of the amount of mineral additions. The average rel-
L-FA-35
ative humidity at the exposure site is about 76% (Fig. 1). The higher
S-65
moisture content in the unsheltered exposure likely leads to a
[mm/y1/2]

higher degree of hydration in the samples stored in these condi- 4

tions. However, due to the lack of any data the same degree of
hydration is assumed. For the fly ash a degree of hydration lower
by a factor of 2.0 than the one of the clinker was assumed 2
[51,55]. The degree of hydration of GGBS (approximate content
of 66 mass-% in the CEM III/B) was assumed to be lower than the
one of the clinker by a factor of 1.4. The higher degree of hydration 0
0.6 0.8 1.0 1.2 1.4 1.6
and smaller difference between cement clinker and GGBS com-
w/CaOreacted [-]
pared to the mortars was chosen due to longer duration of the nat-
ural exposure (2.5 years versus 1 year). Fig. 12. Carbonation coefficient KN,S in the sheltered outdoor exposure as a function
While the correlation of the w/CaOreacted with the carbonation of the w/CaOreacted of the mortar (A: R2 = 0.46) and the concrete mixtures (B:
coefficients KACC of the mortars is good at 1% and 4% CO2 R2 = 0.79).
42 A. Leemann et al. / Cement & Concrete Composites 62 (2015) 33–43
some indications about the behavior of C–S–H and portlandite in
regard to carbonation. Due to the use of mineral additions, their
relative amount is different in both mortar and concrete. For exam-
ple, mortar P has a higher ratio of portlandite to C–S–H due to the
portlandite addition compared to the mortars produced with OPC.
Mortars MS and S have a lower one due to the C–S–H formed by
these siliceous mineral additions. In the first one, portlandite is
consumed due to the hydration of the microsilica, resulting in
additional C–S–H formation [e.g. 56,57]]. In the second one, there
is little portlandite due to the low clinker content and additional
C–S–H formation by the hydraulic GGBS reaction. However, there
is no trend in the carbonation coefficients indicating a slower or
faster reaction with CO2 based on the different C–S–H to port-
landite ratios. Consequently, the contribution of C–S–H and port-
landite to the carbonation resistance seems to be only dependent
on their respective CaO-contents. It has been postulated that port-
landite carbonates more slowly than C–S–H [58,59]. This seems to
be a physical effect: a carbonated shell can form around large port-
landite agglomerates, slowing down carbonation [49,60–62].
Therefore, portlandite can still be present in the zone identified
as carbonated by phenolphthalein [58]. However, this seems to
have negligible effect on the carbonation depth of the studied sam-
ples as determined with the phenolphthalein test.
5. Conclusions
Mortars produced with pure OPC and additions of limestone
powder, microsilica, portlandite and GGBS were exposed to accel-
erated carbonation with 1% and 4% CO2 and to sheltered and
unsheltered outdoor conditions. Additionally, concrete mixtures
produced with different cement types, w/c and paste volume
were stored outdoors for 2.5 years in sheltered and unsheltered
outdoor exposure. All samples were moist cured for 28 days
before being exposed to carbonation. Apart of the carbonation
coefficient, compressive strength, oxygen diffusion and pore char-
acteristics were determined. The results allow drawing the fol-
lowing conclusion for the range of mix designs and the curing
used in this study:
 There is an excellent correlation between the carbonation
coefficients KACC,4% and KACC,1% of the mortars. Their ranking
in regard to carbonation resistance is essentially the same.
 Going from the accelerated carbonation to the carbonation
in the sheltered outdoor exposure, mortars MS and espe-
cially mortar S-65 show a larger decrease of the carbona-
tion coefficient KN,S than the other mortars.
 Going from the sheltered to the unsheltered outdoor expo-
sure, the decrease of the carbonation coefficient KN,US is
again more pronounced for mortars MS and especially mor-
tar S-65. The same applies to the concrete mixtures S-65.
There are two possible reasons for the comparatively large
decrease of the carbonation coefficient KN compared to the
other mixtures. First, the higher amount of small pores in
the mixtures MS and S-65 could lead to a higher amount
of water filled pores due capillary condensation at a given
relative humidity, slowing down carbonation. Secondly, a
high moisture level at the carbonation front remains longer
after a high humidity event, due to the low diffusivity and
slow drying of these mixtures. This would decrease the car-
bonation rate as well.
 There is a cement-specific correlation of the carbonation
coefficient K with w/c, compressive strength and oxygen
diffusion DO2 of the mortar and concrete mixtures.
However, if all mixtures are taken into account, no correla-
tion is evident.
 The parameter governing the carbonation coefficient K of
the mortar and concrete mixtures is the CO2 buffer capacity
per volume of cement paste. It can be expressed by the
w/CaOreactive or w/CaOreacted. However, both parameters
have restrictions. The first parameter assumes complete
hydration of cement and mineral additions, which is cer-
tainly not the case for the studied mixtures. The later
parameter is limited by the uncertainty related to the
assessment of the degree of hydration. The w/CaOreactive
and w/CaOreacted are still valuable to assess the carbonation
resistance of mortar and concrete with various mix designs.
 The correlation of the CO2 buffer capacity with the carbon-
ation coefficient K varies with the relative humidity and is
becoming worse in the unsheltered outdoor exposure likely
due to the increasing importance of capillary condensation
and drying behavior.
In this study the buffer capacity per volume cement paste
expressed by the w/CaOreactive or w/CaOreacted has been identified
as a novel parameter to assess the carbonation resistance of mortar
and concrete containing mineral additions. Based on these findings
further research is needed to clarify the influence of curing and
exposure conditions on the carbonation resistance of such systems
in more detail. As an example, the purpose for the water storage of
all specimens during the first 28 days in this study was to mini-
mize the effect of curing on the different cements and mineral
additions used. The relation between this type of curing and the
typical curing as applied on construction sites and specified in
standards on the carbonation resistance of mortar and concrete
produced with various mineral additions has to be investigated.
Furthermore, the mechanisms leading to the decreased carbona-
tion coefficient KN of concrete produced with GGBS, microsilica
and likely other siliceous mineral additions especially in the
unsheltered outdoor exposure should be further investigated.
This is of particular interest, because concrete with these types of
mineral additions might perform significantly better in conditions
permitting rebar corrosion than indicated by accelerated carbona-
tion tests.
Acknowledgements
cemsuisse is acknowledged for the financial support of this pro-
ject and P. Lura for the careful review of the manuscript.
References
[1] Mangat PS, Khatib JM. Influence of fly ash, silica fume, and slag on sulfate
resistance of concrete. ACI Mater J 1993;92:542–52.
[2] Thomas MDA, Shehata MH, Shashiprakash SG, Hopkins DS, Cail K. Use of
ternary cementitious systems containing silica fume and fly ash in concrete.
Cem Concr Res 1999;29:1207–14.
[3] Bleszynski R, Hooton RD, Thomas MDA, Rogers CA. Durability of ternary blend
concrete with silica fume and blast-furnace slag: laboratory and outdoor
exposure site studies. ACI Mater J 2002;99(5):499–508.
[4] Rozière E, Loukili A, El Hachem R, Grondin F. Durability of concrete exposed to
leaching and external sulphate attacks. Cem Concr Res 2009;39:1188–98.
[5] Leemann A, Lothenbach B, Hoffmann C. Biologically induced concrete
deterioration in a wastewater treatment plant assessed by combining
microstructural analysis with thermodynamic modeling. Cem Concr Res
2010;40:1157–64.
[6] Wierig HJ. Longtime studies on the carbonation of concrete under normal
outdoor exposure. In: Proceedings RILEM seminar on the durability of concrete
structures under normal outdoor exposure; 1984. p. 239–49.
[7] Thomas MDA, Matthews JD. Performance of fly ash concrete in UK structures.
ACI Mat J 1993;90:586–93.
[8] Papadakis VG. Effect of supplementary cementing materials on concrete
resistance against carbonation and chloride ingress. Cem Concr Res
2000;30:291–9.
[9] Sisomphon K, Franke L. Carbonation rates of concretes containing high volume
of pozzolanic materials. Cem Concr Res 2007;37:1647–53.
[10] Rozière E, Loukili A, Cussigh F. A performance based approach for durability of
concrete exposed to carbonation. Constr Build Mater 2009;23:190–9.
 A. Leemann et al. / Cement & Concrete Composites 62 (2015) 33–43
[11] Gruyaert E, Van den Heede P, De Belie N. Carbonation of slag concrete: effect of
the cement replacement level and curing on the carbonation coefficient –
effect of carbonation on the pore structure. Cem Concr Compos
2013;35:39–48.
[12] Thomas MDA, Matthews JD. Carbonation of fly ash concrete. Mag Concr Res
1992;44:217–28.
[13] De Ceukelaire L, Van Nieuwenburg D. Accelerated carbonation of a blast-
furnace cement concrete. Cem Concr Res 1993;23(2):442–52.
[14] Chen CT, Ho CW. Influence of cyclic humidity on carbonation of concrete. J
Mater Civ Eng 2013;25:1929–35.
[15] Thomas MDA, Matthews JD, Haynes CA. Carbonation of fly ash concrete, vol.
19. ACI Spec Publ; 1992. p. 539–56.
[16] Sanjuán MA, Andrade C, Cheyrezy M. Concrete carbonation tests in natural and
accelerated conditions. Adv Cem Res 2003;15:171–80.
[17] Hyvert N, Sellier A, Duprat F, Rougeau P, Francisco P. Dependency of C–S–H
carbonation rate on CO2 pressure to explain transition from accelerated tests
to natural carbonation. Cem Concr Res 2010;40:1582–9.
[18] Bernal SA, Provis JL, Brice DG, Kilcullen A, Duxson P, van Deventer JSJ.
Accelerated carbonation testing of alkali-activated binders significantly
underestimates service life: the role of pore solution chemistry. Cem Concr
Res 2012;42:1317–26.
[19] Neves R, Branco F, de Brito J. Field assessment of the relationship between
natural and accelerated concrete carbonation resistance. Cem Concr Compos
2013;41:9–15.
[20] Younsi A, Turcry Ph, Aït-Mokhtar A, Staquet S. Accelerated carbonation of
concrete with high content of mineral additions: effect of interactions
between hydration and drying. Cem Concr Res 2013;43:25–33.
[21] Osborne GJ. Carbonation and permeability of blastfurnace slag cement
concretes from field structures, vol. 114. ACI Spec Pub; 1992. p. 1209–38.
[22] Fatthudi NI. Concrete carbonation as influenced by curing regime. Cem Concr
Res 1988;18:426–30.
[23] Geiker MR, Laugesen P. On the effect of laboratory conditioning and freeze/
thaw exposure on moisture profiles in HPC. Cem Concr Res 2001;31:1831–6.
[24] Hooton RD, Geiker MR, Bentz EC. Effects of curing on chloride ingress and
implications on service life. ACI Mater J 2002;99:201–6.
[25] Loser R, Leemann A. Sulfatwiderstand von Beton: verbessertes Verfahren
basierend auf der Prüfung nach SIA 262/1, Anhang D, UVEK, Bundesamt für
Strassen, Bericht Nr. 1416, Bern; 2013.
[26] DIN EN 1015-11. Methods of test for mortar for masonry Determination of
flexural and compressive strength of hardened mortar. 2007.
[27] EN 12390-3. Testing hardened concrete – Part 3: compressive strength of test
specimens; 2002.
[28] SN 505 262/1, Betonbau – Ergänzende Festlegungen, Anhang I,
Karbonatisierungswiderstand, SIA Zürich; 2013.
[29] Lawrence CD. Transport of oxygen through concrete. In: Glasser FP, editor. The
chemistry and chemically-related properties of cement. British Ceramic
Society Proceedings, vol. 35; 1984. p. 277–93.
[30] Buenfeld NR, Okundi E. Effect of cement content on transport in concrete. Mag
Concr Res 1998;50:339–51.
[31] Villani C, Loser R, West MJ, Di Bella C, Lura P, Weiss JW. An inter lab
comparison of gas transport testing procedures: oxygen permeability and
oxygen diffusivity. Cem Concr Compos 2014;53:357–66.
[32] Kaufmann J, Loser R, Leemann A. Analysis of cement-bonded materials by
multi-cycle mercury intrusion and nitrogen sorption. J Colloid Interface Sci
2009;336:730–7.
[33] De Ceukelaire L, Van Nieuwenburg D. Accelerated carbonation of a blast-
furnace cement concrete. Cem Concr Res 1993;23:442–52.
[34] Verbeck G. Carbonation of hydrated Portland cement. In: Crawford PH,
editor. Cement and Concrete. Baltimore: American Society for testing
Material; 1958.
[35] Russell D, Basheer PAM, Rankin GIB, Long AE. Effect of relative humidity and
air permeability on prediction of the rate of carbonation of concrete. Proc Inst
Civ Eng Struct Build 2001;146:319–26.
[36] Mehta PK, Gjørv OE. Properties of Portland cement concrete containing fly ash
and condensed silica-fume. Cem Concr Res 1982;12:587–95.
43
[37] Li S, Roy DM. Investigation of relations between porosity, pore structure, and
C1 diffusion of fly ash and blended cement pastes. Cem Concr Res
1986;16:749–59.
[38] Parrot LJ. Moisture profiles in drying concrete. Adv Cem Res 1988;1:164–70.
[39] Parrot LJ. Factors influencing relative humidity in concrete. Mag Concr Res
1991;43:45–52.
[40] Tuutti K. Corrosion of steel in concrete. Stockholm: Svenska
Forskningsinstitutet för cement och betong; 1982.
[41] González JA, Andrade C. Effect of carbonation, chlorides and relative ambient
humidity on the corrosion of galvanized rebars embedded in concrete. Br
Corros J 1982;17:21–8.
[42] Houst YF, Wittmann FH. Influence of porosity and water content on the
diffusivity of CO2 and O2 through hydrated cement paste. Cem Concr Res
1994;24:1165–76.
[43] Ngala VT, Page CL. Effects of carbonation on the pore structure and diffusional
properties of hydrated cement paste. Cem Concr Res 1997;27:995–1007.
[44] Auroy M, Poyet S, Le Bescop P, Torrenti JM. Impact of carbonation on water
transport properties of cement-based materials. In: NUWCEM, 2nd
international symposium on cement-based materials for nuclear waste,
Avignon, France; 2014.
[45] Papadakis VG, Vayenas CG, Fardis MN. Fundamental modeling and
experimental investigation of concrete carbonation. ACI Mater J
1991;88:363–73.
[46] Steffens A, Dinkler D, Ahrens H. Modeling carbonation for corrosion risk
prediction of concrete structures. Cem Concr Res 2002;32:935–41.
[47] Saetta AV, Vitaliani RV. Experimental investigation and numerical modeling of
carbonation process in reinforced concrete structures: part I: theoretical
formulation. Cem Concr Res 2004;34:571–9.
[48] Marques PF, Chastr C, Nunes Â. Carbonation service life modelling of RC
structures for concrete with Portland and blended cements. Cem Concr
Compos 2013;37:171–84.
[49] Groves GW, Brough A, Richardson IG, Dobson CM. Progressive changes in the
structure of hardened C3S cement pastes due to carbonation. J Am Ceram Soc
1991;74:2891–6.
[50] Parrot LJ, Killoh DC. Prediction of cement hydration. Proc Br Ceram Proc
1984;35:41–53.
[51] Feng X, Garboczi EJ, Bentz DP, Stutzman PE, Mason TO. Estimation of the
degree of hydration of blended cement pastes by a scanning electron
microscope point-counting procedure. Cem Concr Res 2004;34:1787–93.
[52] Lumley JS, Gollop RS, Moir GK, Taylor HFW. Degrees of reaction of the slag in
some blends with Portland cements. Cem Concr Res 1996;26:139–51.
[53] Pane I, Hansen W. Investigation of blended cement hydration by isothermal
calorimetry and thermal analysis. Cem Concr Res 2005;35:1155–64.
[54] Kocaba V, Gallucci E, Scrivener KL. Methods for determination of degree of
reaction of slag in blended cement pastes. Cem Concr Res 2012;42:511–25.
[55] Ben Haha M, De Weerdt K, Lothenbach B. Quantification of the degree of
reaction of fly ash. Cem Concr Res 2010;40:1620–9.
[56] Mitchell DRG, Hinczak I, Day RA. Interaction of silica fume with calcium
hydroxide solutions and hydrated cement pastes. Cem Concr Res
1998;28:1571–84.
[57] Lothenbach B, Scrivener K, Hooton RD. Supplementary cementitious materials.
Cem Concr Res 2011;41:1244–56.
[58] Parrott LJ, Killoh DC. Carbonation in a 36 year old, in-situ concrete. Cem Concr
Res 1989;19:649–56.
[59] Castellote M, Andrade C, Turrillas X, Campo J, Cuello GJ. Accelerated
carbonation of cement pastes in situ monitored by neutron diffraction. Cem
Concr Res 2008;38:1365–73.
[60] Van Balen K. Carbonation reaction of lime, kinetics at ambient temperature.
Cem Concr Res 2005;35:647–57.
[61] Thiery M, Villain G, Dangla P, Platret G. Investigation of the carbonation front
shape on cementitious materials: effects of the chemical kinetics. Cem Concr
Res 2007;37:1047–58.
[62] Morandeau A, Thiéry M, Dangla P. Investigation of the carbonation mechanism
of CH and C–S–H in terms of kinetics, microstructure changes and moisture
properties. Cem Concr Res 2014;56:153–70.