Cement and Concrete Research 186 (2024) 107688

Contents lists available at ScienceDirect

Cement and Concrete Research

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

Plugging effect of fine pore water in OPC and LC3 paste during accelerated
carbonation monitored via single-sided nuclear magnetic
resonance spectroscopy
Luge Cheng a,*, Ryo Kurihara a , Takahiro Ohkubo b, Ryoma Kitagaki c , Atsushi Teramoto d,
Yuya Suda e , Ippei Maruyama a,*
a
Department of Architecture, Graduate School of Engineering, The University of Tokyo, Tokyo 113-8656, Japan
b
Department of Applied Chemistry and Biotechnology, Graduate School of Engineering, Chiba University, Chiba 263-8522, Japan
c
Division of Human Environmental Systems, Graduate School of Engineering, Hokkaido University, Hokkaido 060-8628, Japan
d
Department of Architecture, Graduate School of Advanced Science and Engineering, Hiroshima University, Hiroshima 739-8527, Japan
e
Civil Engineering Program, Faculty of Engineering, University of the Ryukyus, Okinawa 903-0213, Japan

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

Keywords: This study investigates the influence of CO2 concentration on the carbonation process in cementitious paste,
Acceleration (A) focusing on water content distribution in ordinary Portlandite cement and limestone-calcined clay cement (LC3).
Carbonation (C) Employing single-sided nuclear magnetic resonance spectroscopy for water profiling, we revealed that under
Calcium-Silicate-Hydrate (C-S-H) (B)
accelerated carbonation of 5 % and 1 %, the water content in fine pores (interlayer space and gel pores) kept
Pore size distribution (B)
constant at the carbonation front, demonstrating the plugging effect where fine pore water removal governs
carbonation progress. This effect was absent under natural carbonation conditions because evaporation precedes
the carbonation process. This study emphasizes that to accurately characterize cementitious materials under
natural carbonation conditions, CO2 concentrations in accelerated methods should be constrained to prevent the
plugging effect.

1. Introduction method can be used to ascertain the potential for carbonation capture
during both the service life [3,4] and subsequent disposal [5].
Corrosion of the reinforcement is the predominant cause of degra­ The validity of the accelerated carbonation method to mirror the
dation in reinforced concrete structures [1] and is initiated by the characteristics of natural carbonation, which occurs in ambient air with
disruption of the passive film on the surface of the reinforcement [2]. CO2 concentration at approximately 400 ppm, through an increase in
This can result from the carbonation of the surrounding concrete, which the CO2 concentration is debated. CO2 concentration standards vary
reduces the pH value from approximately 13 to less than 9. To predict globally, from 3 %, as specified in the ISO standard [6], to 50 % in
the service life or evaluate the soundness of reinforcement structures, it France and other countries, such as 20 % in China [7] and 5 % in Japan
is imperative to determine the resistance of concrete against carbon­ [8].
ation. Acceleration of the carbonation test is used to evaluate the Many studies have shown differences in microstructural modifica­
carbonation resistance of cementitious materials, particularly supple­ tions and calcium carbonate (CC) between accelerated and natural
mentary cementitious materials (SCMs) with new compositions and carbonation conditions across a wide range of CO2 concentrations (from
limited previous applications. These tests were conducted using high 3 % to 100 %) [8,9] and tried to identify the most feasible CO2 con­
CO2 concentrations to shorten the duration of experimental procedures centration to represent natural carbonation. It was proposed [11] that a
in a laboratory setting. Conversely, carbonation capture is considered CO2 concentration below 20 % with 70 % RH at 20 ◦ C could represent
beneficial in cementitious materials as it potentially offsets CO2 emis­ natural conditions, considering the reduced permeability of concrete to
sions from cement production without the risk of CO2 escaping to the gas diffusion above this concentration. Other studies suggested that 3 %
atmosphere for a long period of time. The accelerated carbonation CO2 concentration at 65 % RH and 22 ◦ C was effectively represents

* Corresponding authors.
E-mail addresses: chengluge@g.ecc.u-tokyo.ac.jp (L. Cheng), i.maruyama@bme.arch.t.u-tokyo.ac.jp (I. Maruyama).

https://doi.org/10.1016/j.cemconres.2024.107688
Received 19 June 2024; Received in revised form 8 September 2024; Accepted 24 September 2024
Available online 8 October 2024
0008-8846/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

natural carbonation, maintaining similar remaining phases: ettringite conditions. The originality of this study lies in providing direct evidence
and calcium silicate hydrates (C-S-H) gel with similar Ca/Si ratios [12]. of the inhibitory effect of water content on the carbonation process
A similar carbonation condition under 3 % CO2 with 67 % RH and 20 ◦ C across varying CO2 concentrations. These findings were substantiated by
was proposed in [13] as it showed the similar changes in phase assem­ detailed mineralogical analyses, offering new insights into the dynamics
blage and microstructure under natural and accelerated carbonation (3 of cementitious materials under accelerated carbonation.
%) conditions. The condition of 10 % CO2 concentration at 53 % RH and
room temperature was proposed [14] based on the results showing 2. Experiment methods
equivalent carbonation degrees of portlandite (CH) and C-S-H under 10
% CO2 and natural carbonation conditions. Furthermore, with the index 2.1. Raw materials and mix design
of ranking the carbonation resistance of concrete mixtures, accelerated
carbonation with an increase of CO2 concentration up to 4 % with 57 % Two binder types were studied: OPC (Mitsubishi UBE Co., Ltd.) and
RH and 20 ◦ C was feasible to represent natural carbonation conditions LC3. LC3 was composed of 50 % OPC cement, 30 % calcined clay, 15 %
[15]. The proposed CO2 concentrations for the accelerated carbonation limestone, and 5 % gypsum. The calcined clay is made by mixing 75 %
method, intended to represent natural carbonation, were determined metakaolin (BASF MetaMax®) and 25 % quartz (UBE Co. Ltd., grade 7).
primarily based on comparisons between the resulting physical and The chemical and phase compositions of the raw materials, as measured
mineralogical characterizations of cementitious materials under by X-ray fluorescence (XRF) and XRD are detailed in Table 1. The par­
different accelerated carbonation conditions and natural carbonation. ticle size distributions of the raw materials are shown in Fig. 1.
However, during the carbonation process, different CO2 concentrations
led to different changes in water content, which affected the progression 2.2. Sample preparation
of carbonation. This effect is essential for a comprehensive under­
standing of carbonation progress; however, it has not been sufficiently A planetary mixer (ARE-500, THINKY) was used at 1000 rpm to
addressed for the evaluation of representative CO2 concentrations. achieve a water-to-cement ratio (W/C) of 0.55. The deionized water
As noted in [16], the carbonation depth remained constant with a used for mixing was maintained at 20 ◦ C for one day before use. For the
similar water loss. This observation led to the identification of a factor OPC paste, water was initially added to cement to achieve a W/B ratio of
that compromises the effectiveness of accelerated testing, known as the 0.30. The mixture was stirred at a rotation speed of 1000 rpm for 1.5 min
plugging effect, which means that abundant water hinders CO2 gas to ensure homogeneity. Subsequently, the remaining water was added to
penetration into deeper regions, and water distribution in the carbon­ achieve the target W/B ratio of 0.55, followed by an additional 1.5 min
ated zone dominates the overall carbonation progress. The effect of RH of mixing. After mixing, the fresh cement paste was placed in 500 mL
on carbonation differed between the powder and block samples. If the polypropylene bottles and rotated on a roller (IKA ROLLER 10 digital) at
sample size is larger, the behavior should resemble that of a block 60 rpm for 3.5 h to minimize bleeding and segregation. For the LC3
sample, whereas if the sample size is smaller and the moisture transport paste, to overcome the rapid setting gypsum was pre-dissolved in water
path from inside to outside is relatively short and quick, the dominant to induce retardation. The cement, limestone, metakaolin, and quartz
reaction carbonation kinetics are governed by factors other than water were mixed in advance for 10 min to achieve better homogeneity. All the
transport; consequently, the relationship between RH and the carbon­ materials were then added to water and mixed at a rotational speed of
ation rate of hardened cement particles or concrete fines is different 1000 rpm for 1.5 min. After confirming that the mixture was creamy, the
from that obtained by block samples. Studies have indicated that the OPC and LC3 pastes were poured into 50 mm × 50 mm × 30 mm pris­
degree of carbonation of powdered samples increases with increasing matic molds. The molds were then vibrated to remove entrapped air
RH [17]. However, for realistic block samples, the well-known bell- bubbles. Afterward, the casting surface was covered with polyethylene
curve relationship between RH and carbonation rate confirms that the wrap, and the molds were stored in a controlled environment at 20 ◦ C.
optimum RH [18–20] for the carbonation rate ranges from 50 to 80 %. The specimens were demolded after 24 h, sealed, and cured at 20 ◦ C for
Nevertheless, direct evidence detailing the role of water during the over three months to ensure sufficient hydration of the cement before
carbonation process, which is imperative for understanding the proceeding to the next phase of the study.
carbonation behavior of cementitious materials under various environ­
mental conditions, is lacking. 2.3. Exposure conditions
Based on the literature review above, there is an ongoing controversy
regarding whether the current accelerated carbonation methods can After the curing period, following the JIS standard accelerated
reflect the characterization of the natural carbonation of various
cementitious materials. Considering the important role of water in the Table 1
carbonation process, visualizing the water content distribution within XRF and XRD measurement results of OPC, metakaolin, limestone, quartz, and
cementitious materials during carbonation under different CO2 con­ gypsum.
centrations is an effective approach to address this question. Therefore, OPC Metakaolin (MK) Limestone (LS) Quartz Gypsum
this study was conducted to provide further evidence and improve the
SiO2 19.86 50.21 3.16 95.20 0.45
understanding of the effect of CO2 concentration on the carbonation Al2O3 5.55 47.01 0.74 2.08 0.28
process from the perspective of water content and mineralogical phases Fe2O3 2.80 0.47 0.74 1.01 0.49
based on single-sided proton nuclear magnetic resonance relaxometry MgO 1.41 0.00 0.65 0.26 0.38
(1H NMR relaxometry) and μ-X-ray diffraction (μ-XRD) mapping. Based K2O 0.41 0.13 0.07 0.49 0.02
Na2O 0.28 0.32 0.00 0.15 0.00
on the results of the water content distribution in the samples during
CaO 64.19 0.02 90.24 0.23 43.51
carbonation, the process and stagnation of the carbonation process SO3 2.70 0.04 0.04 0.28 54.38
under different CO2 concentrations were investigated. In addition, we Cl 0.015 0.00 0.00 0.00 0.18
included a newly developed cementitious material, limestone-calcined Alite 56.51 – – – –
clay cement (LC3), which comprises limestone and calcined kaolinitic Belite 18.27 – – – –
C3A 8.55 – – – –
clays and can replace up to 50 % of conventional ordinary Portland C4AF 7.07 – – – –
cement (OPC) while maintaining a comparable strength and exhibiting Calcite 4.43 – – – –
enhanced resistance to chloride penetration and alkali-silica reactions Bassanite 2.78 – – – –
[21,22]. Considering the potential for widespread application of LC3, Gypsum 1.52 – – – –
Periclase 0.85
this study investigates LC3 as well as OPC under different carbonation – – – –

2
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

12 70
Limestone
Incremental volume (%) 10 Metakaolin
Quartz 60
8

RH (%)
6
50
4
2 40
0
30
0.1 1 10 100 1000 0 4 8 12 16 20 24 28
Particle diameter (μm) Time (day)
Fig. 1. Particle size distributions of raw materials. (a)
1000
carbonation method [8], the samples were preconditioned to dry at the
800
optimum RH for the carbonation reaction, which was reported to be
approximately 60 % [17,18]. The samples were sealed on five sides

CO2 (ppm)
600
using two-sided tape and covered with parafilm, leaving one surface
exposed to a controlled environment. This setup was designed to allow 400
unidimensional transport of CO2 and moisture. The sealed samples were
then kept for 28 days in an airtight desiccator at 20 ◦ C containing 200
saturated salt solutions of NaBr. To prevent carbonation during the RH-
controlling process, soda lime particles were added to the desiccator to 0
absorb CO2. 0 4 8 12 16 20 24 28
After spending 28 days in the RH-controlled environment, the sam­ Time (day)
ples were transferred to the carbonation incubators (AS ONE) that (b)
maintained the temperature at 20 ◦ C and CO2 concentration at 5 % (5-
CO2) and 1 % (1-CO2). The natural carbonation environment (C-N) was Fig. 2. Example of the monitored RH values in carbonation incubator (a) and
set in a desiccator connected to an air pump to circulate the humidified CO2 concentration in desiccator (b) during the carbonation process with the
air and introduce external air. The relative humidity at each carbonation monitoring window of 28 days.
condition was controlled by a saturated salt solution of NaBr at around
60%RH. Since it is difficult to maintain the RH constant around 60 % installation of 10 mm spacers, and a resolution of 200 μm. The sensitive
with the partial pressure of CO2, periodical monitoring and treatment volume was extended horizontally by approximately 40 × 40 mm2.
(supplementation of the salt) of the salt solution is necessary. The RH Because the NMR signal originates from the mobile water in the cement
values were monitored using an RH sensor inside the carbonation in­ paste, the signal intensity can be correlated with the evaporable water
cubators and the CO2 concentration was monitored using a CO2 meter in content in the sample [23,24]. The specific parameters applied to the
a desiccator during the carbonation process. As shown in Fig. 2, the RH CPMG measurements of the OPC and LC3 pastes are listed in Table 2.
in the carbonation incubator fluctuated by approximately 60 %, and the Different parameters for OPC and LC3 were selected to cover the
CO2 concentration in the desiccator remained at approximately entire relaxation process, as shown in Fig. 3. To enhance the signal-to-
400–500 ppm. noise (S/N) ratio and reduce signal scatter, the number of scans was
For each carbonation condition, dummy samples of the same size set to 1024 for the OPC and 512 for the LC3. The S/N ratio was deter­
were prepared to measure the carbonation depth during each carbon­ mined by dividing the amplitude of the first echo by the standard de­
ation period. viation of the last 10 echoes for OPC and 20 echoes for LC3. During the
drying process, the S/N ratios for OPC and LC3 ranged from 92 to 213
2.4. Characterization methods and 94 to 195, respectively. During the carbonation process, the S/N
ratios for OPC and LC3 ranged from 73 to 161, and 76 to 85, respectively.
2.4.1. Single-sided 1H NMR The variability in the S/N ratios is attributed to the water loss in the pore
The relaxometry measurements were conducted using single-sided systems during these processes. The sensitive region for measurement
1
H NMR measurements by NMR MOUSE PM 25 (Magritek GmbH, Ger­ was progressively moved in 1 mm increments from the surface to the
many) with the Carr–Purcell–Meiboom–Gill (CPMG) technique [23], a interior of the sample. The CPMG decay at each depth was fitted to a
non-destructive measurement method, which has been applied to multi-exponential decay function to ascertain the associated signal in­
porous materials and yielded reliable results [21,22]. Measurements tensities. Another method for evaluating the data is the inverse Laplace
were performed before and after the predrying period and at 3, 7, 14, transformation, which requires high-quality of data with an S/N ratio
and 28-day intervals during the carbonation process. To prevent the exceeding 150, as noted in [24]. However, obtaining such high-quality
results from being influenced by potential large voids within the sample
that may have been caused during casting, two dummy samples were Table 2
monitored under the same determined intervals to verify their consis­ Parameters for the CPMG measurements undertaken on OPC and LC3 paste.
tency. For natural carbonation, additional measurements were per­ OPC paste LC3 paste
formed after 56 days. The device utilizes a permanent magnet that
Pulse length (μs) 17 17
generates a 0.32 T magnetic field B0, which corresponds to 1H Larmor Echo time (μs) 75 75
frequency of 13.12 MHz, and features a gradient strength of 300 kHz/ Number of echoes 50 160
mm (7.05 T/m). A surface radiofrequency coil, positioned atop the Repetition time (ms) 532 2128
magnet, is responsible for exciting and detecting the NMR signal within Number of scans 1024 512
Recorded time range (μs) 150–3,750 150–12,000
a sensitive range of 15 mm from the surface of the instrument, with the

3
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

Fig. 3. Signal decays of OPC (a) and LC3 (b) owning different relaxation time spans.

data can be time-consuming in an inhomogeneous magnetic field and Subsequently, the volumetric water content (w) within the sample at
may result in unacceptable water loss during measurement in this study. each depth was quantified based on the ratio of the total component of
For OPC, the signal was fitted using a bi-exponential function, while a the sample to the reference component of the copper sulfate solution:
tri-exponential function was used for LC3. This difference accounts for /
2∑
or 3
the different microstructures and corresponding water-containing pores w= Ai Aref (4)
with different relaxation times in the two materials. The decay function i=1
equations are given by Eqs. (1) and (2):
In addition, the capability of water capture by single-sided 1H NMR
S(t) = A1 e(− t/T2,1 ) + A2 e(− t/T2,2 ) (1) was confirmed through a comparative analysis with 1H NMR with a
homogeneous magnetic field, as shown in Appendix A. This analysis
revealed that single-sided 1H NMR could detect most of the free water in
S(t) = A1 e(− t/T2,1 ) + A2 e(− t/T2,2 ) + A3 e(− t/T2,3 ) (2)
OPC and LC3 with only a minimal amount of interlayer water eluding
where S(t) is the magnetization, t is the time, A1 , A2 and A3 are the detected.
magnetization components corresponding to the T2,1 , T2,2 and T2,3 life­
time constants. According to previous studies [28–30], each function 2.4.2. Carbonation depth measurement
corresponds to a pore. A1 , T2,1 and A2 , T2,2 in Eq. (1), associated with During carbonation, the carbonation depth was measured in the
water in the fine pores (interlayer space and gel pores) and coarse pores dummy samples at the same intervals as the single-sided 1H NMR
(interhydrate and/or capillary pores) for OPC, respectively. Similarly, in measurements for each carbonation period. A segment of the sample was
Eq. (2), A1 , T2,1 , A2 , T2,2 and A3 , T2,3 are associated with water in the sliced along the carbonation direction by using low-speed precision
interlayer space, gel pores, and coarse pores (interhydrate and/or cutter (IsoMet) with isopropanol (purity: 99.7 %) as the cooling and
capillary pores) for LC3, respectively. cutting medium, and then sprayed with a phenolphthalein pH indicator.
To avoid overfitting, specific ranges of T2 values were selected based The indicator used was a phenolphthalein 1 % ethanol solution with 1 g
on the T2 distribution analysis of the OPC and LC3 samples after hy­ phenolphthalein and 90 mL 95.0 V/V% ethanol diluted in water to 100
dration, as shown in Fig. 4. In this study, the ranges of T2,1 and T2,2 mL. The images were captured approximately 24 h after spraying
values were set as 0–0.5 ms and 0.5–2 ms for OPC, and the ranges of T2,1, phenolphthalein on the sectional area under a uniform lighting condi­
T2,2 and T2,3 were set as 0–0.5 ms, 0.5–2 ms and 2–6 ms for LC3. The tion (with an average illuminance of 2940 Lux). The conventional
functions demonstrate a high quality of fit, with coefficient of deter­ method of measuring the carbonation depth, quantifying the distance
mination R2 values exceeding 0.98 for OPC and 0.91 for LC3. from the exposed surface to the boundary between the purple (non­
A reference sample of copper sulfate solution (CuSO4⋅5H2O, 6 mM) carbonated) and colorless (carbonated) areas at five different points
was analyzed using the same parameters due to its shorter relaxation using Vernier calipers [8], has been criticized for its subjectivity and
time than that of water. The CPMG decay of the reference was fitted potential for low reproducibility [31]. In this study, a more objective
using a single exponential decay function as expressed in Eq. (3): approach was adopted by processing the images using the k-means
clustering method to detect color changes and subsequently converting
Sref (t) = Aref e(− t/T2 )
(3) these images into a binary format. Noise was eliminated using a

Fig. 4. Distribution of T2 relaxation for OPC (a) and LC3 (b) right after the hydration.

4
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

morphological closing operation that clearly delineated the boundary of based on Fick's first law:
the carbonation area. This enhanced method improves the accuracy of √̅̅
determining the carbonation front and involves the following steps [32]: Xc = k t (5)
1) converting the original image to a binary image by k-means clustering 3
The carbonation depths of OPC and LC under different carbonation
of two groups; 2) determining the carbonation front in the binary image; conditions conformed to Fick's first law. The influence of CO2 concen­
3) delineating the entire carbonated area; 4) measuring the distances tration on the carbonation rate coefficient (k) is shown in Fig. 7(b).
from the top side at 1 mm intervals in the carbonated area; and 5) The carbonation rate increases with increasing CO2 concentration.
calculating the median value as the carbonation depth to avoid the For OPC, the carbonation rates are approximately 22.4 (=1.12 mm/
dispersion of data due to crack formation. Fig. 5 illustrates the image day0.5) and 15.2 (=0.76 mm/day0.5) times higher when CO2 concen­
analysis procedure. tration is 125 (5 % CO2) and 25 (1%CO2) times higher, respectively,
compared to natural carbonation condition (0.04%CO2) (=0.05 mm/
2.4.3. Scanning electron microscopy day0.5). For LC3, the carbonation rate under 5 % and 1 % CO2 are
For the scanning electron microscopy (SEM) investigation coupled approximately 15.2 (=3.03 mm/day0.5) and 8.1 (=1.61 mm/day0.5)
with EDS mapping, fragment samples were obtained by cutting the OPC times higher, respectively, than that under natural carbonation condi­
and LC3 pastes by using low speed precision cutter (IsoMet) with iso­ tion (=0.2 mm/day0.5). A similar positive effect of the CO2 concentra­
propanol (purity: 99.7 %) as the cooling and cutting medium. These tion on the carbonation rate coefficient was also reported in [33,34].
fragments were impregnated in epoxy and polished using a cross- Furthermore, LC3 is more vulnerable to carbonation than OPC, with a
sectional polisher (JEOL SM-090010 090020). Subsequently, a thin carbonation rate approximately 2.1 to 4.0 times higher for LC3 under
conductive osmium layer was coated on the polished surface. different carbonation conditions. This increased vulnerability is pri­
A field-emission scanning electron microscope (FE-SEM, JSM7800F) marily due to the lower capacity of LC3 to buffer the pH of the pore
was used to capture Backscattered scanning electron (BSE) images of the solution. Specifically, the lower amount of CH and lower Ca/Si ratio in
polished surfaces, operating at an accelerating voltage of 15 kV, working the C-S-H of LC3 (discussion in Section 3.4), compared to OPC, result in
distance of 10 mm, and resolution of 1280 × 1024, with images captured less available calcium to maintain a high pH value in the pore solution.
at a magnification of 200 for observation of microstructure of cemen­ Similar results have been reported for other cementitious materials with
titious pastes, at a magnification of 500× for the calculation of Ca/Si cement replacement [35–37].
ratios, at a magnification of 20,000× for the observation of C-S-H gel.

2.4.4. μ-XRD mapping 3.2. Mineralogical profile detected by μ-XRD mapping measurement
μ-XRD mapping measurements were performed using a PANalytical
Empyrean diffractometer equipped with Cu Kα radiation in Bragg- Fig. 8 shows the XRD patterns of the noncarbonated and carbonated
Brentano configuration. Part of the sample was sliced along the zones at the surfaces of OPC and LC3 after 28 days of carbonation.
carbonation direction from the dummy samples and positioned verti­ Qualitative analysis of the XRD patterns revealed distinct CH peaks in
cally on an X-Y-Z stage. This setup facilitated the measurement of XRD the noncarbonated OPC, whereas the noncarbonated LC3 exhibited no
patterns at 1 mm intervals from the carbonation surface up to a depth of obvious CH peaks; instead, the ettringite peaks were more pronounced
15 mm. The measurements for each sample were conducted three times than those in OPC. This can be explained by the pozzolanic reaction of
to ensure the accuracy and repeatability of the results. The scanning the metakaolin in the calcined clay, providing C-S-H and C-A-S-H and
range for the 2θ angle was between 10◦ and 70◦ . The scanning speed was ettringite [38,39]. The presence of the calcite peak in the noncarbonated
maintained at 1.18◦ /min with a collimator setting of 0.11◦ and Soller slit LC3 sample can be attributed to the addition of 5 % limestone to the
of 0.02 rad. The maximum irradiated area for a single scanning point mixture. Under natural carbonation conditions, CH peaks remained in
was 722 μm. Rietveld analysis was subsequently performed to quantify the OPC, with calcite being the primary carbonation product. However,
the crystalline phases using an external standard method with reference under accelerated carbonation conditions at 1 % and 5 % CO2 concen­
Si crystals. trations, the CH peak became minimal, with calcite and vaterite
Fig. 6 depicts a comprehensive experimental outline following the becoming the main carbonation products of the OPC. For carbonated
curing period, which clearly represents the experimental procedures LC3, calcite is the predominant carbonation product.
following the curing period. μ-XRD mapping measurements of the carbonated paste enable the
tracking of carbonation product distribution along the carbonation di­
3. Results rection under various carbonation conditions. Fig. 9 presents the
quantification results of the crystalline (portlandite, ettringite, calcite,
3.1. Carbonation depth vaterite, and aragonite) and amorphous components (primarily con­
sisting of C-S-H, calcium aluminum silicate hydrate (C-A-S-H), silica gel,
Fig. 7(a) illustrates the carbonation depth of the samples over the and calcined clay) from the XRD scans (shown in Fig. A2 in Appendix B)
carbonation duration, as determined by image analysis. The carbonation across the carbonation depths.
rate coefficient (k) can be derived from the relationship between the For the carbonated zone in OPC (Fig. 9(a)–(c)), under accelerated
carbonation depth (Xc ) and square root of the carbonation duration (t) carbonation conditions at CO2 concentrations of 5 % and 1 %, the pre­
dominant carbonation products were calcite and vaterite, along with a
notable reduction in the contents of amorphous substances, ettringite,
and CH, compared to the noncarbonated zone. This observation suggests
that the carbonation of the hydration products of C-S-H, ettringite, and
CH occurs within the carbonation duration of 28 days. Notably, the
amount of vaterite exceeded that of calcite at 1 % CO2, whereas sig­
nificant amounts of both calcite and vaterite were observed at 5 % CO2.
This finding aligns with the previous analyses indicating that the
mineralogy of the carbonation products is influenced by CO2 concen­
Fig. 5. Procedure of image analysis for carbonation depth: (a) original image tration [40–42]. This can be attributed to the higher CO2 concentration
after spaying phenolphthalein solution; (b) binary image determined by K- accelerating the formation rate of calcium carbonate, thereby promoting
means cluster; (c) carbonation depth measured every 1 mm from the top side. the transition of the crystal phases to a more stable state. Furthermore,

5
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

× ×

Fig. 6. The overall outline of the experiment.

OPC: 5-CO2 LC3: 5-CO2 3.5

Carbonation rate coefficient OPC LC3
Carbonation depth (mm)

20 OPC: 1-CO2 3
LC : 1-CO2 3.0
OPC˖C-N LC3: C-N 2.5
15 2.0

10 1.5
1.0
5 0.5
0.0
0
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5
0.5
Time (day ) CO2 concentration (%)
(a) (b)
Fig. 7. Carbonation depth determined by image analysis on OPC and LC3 under different CO2 concentration (a), and the relationship between carbonation rate
coefficient and CO2 concentration.

Fig. 8. XRD patterns of noncarbonated (NC) and carbonated samples under 5-CO2, 1-CO2, and C-N for (a) OPC and (b) LC3. (Ett: ettringite, Hc: hemicarboaluminate,
P: CH; C: calcite; V: vaterite; A: aragonite).

CH remained near the surface under 5 % CO2 conditions, unlike under 1 ettringite at the surface exhibited no significant change compared to
% CO2 conditions where a more intense carbonation degree was those in the noncarbonated zones, whereas the content of CH decreased.
observed. This is attributed to the reduction in the accessibility of the CH This observation suggests that the carbonation of CH takes precedence
crystals owing to the formation of a carbonation product layer on the CH over that of C-S-H, which is consistent with the different buffering ca­
crystals at higher CO2 concentrations (5 %) [42]. Under natural pacities of CH and C-S-H from a thermodynamic perspective [10,43].
carbonation conditions, the amounts of amorphous substances and For LC3 (Fig. 9(d)–(f)), the predominant carbonation products were

6
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

Fig. 9. Quantification results of XRD scans of OPC under 5-CO2 (a), 1-CO2 (b), and C-N (c), and LC3 under 5-CO2 (d), 1-CO2 (e), and C-N (f) after 28 days of
carbonation period.

calcite and vaterite, considering the existence of calcite in the non­

where m0CH , m0Ett , m0CC and m0CSH are the initial content of CH, Ett, CC, and
carbonated zone. This can be attributed to the relatively low CH content
C-S-H of the non-carbonated sample, respectively. mCH , mEtt , and mCC are
of LC3, which promotes extensive carbonation of C-S-H or C-A-S-H,
the content of CH, Ett, and CC detected at each depth. MCH , MEtt , MCC
resulting in the formation of calcite and metastable vaterite phases [43].
and MCSH are the molar mass of CH, Ett, CC and C-S-H. the C/S ratio(a) of
Furthermore, the carbonation degree of the main hydrates in OPC
C-S-H in OPC and LC3 was obtained in Section 3.4, which are 1.61 and
and LC3, ettringite (αEtt), portlandite (αCH) and C-S-H (αC-S-H), are
1.17, while the H2O/CaO ratio was assumed as 2.5. Fig. 10 shows the
calculated as follows:
carbonation degree of each hydrate product in OPC and LC3 under
( )/
αCH = m0CH − mCH m0CH (6) different CO2 concentrations.
For OPC under the 5-CO2 condition, the carbonation of Ett in the
( )/ 0
αEtt = m0Ett − mEtt mEtt (7) carbonated zone is almost complete, while the carbonation degrees of
CH and C-S-H are slightly lower, but still reach approximately 90 %.
m0CH − mCH m0Ett − mEtt Under the 1-CO2 condition, higher carbonation degrees of CH and C-S-H,
mCC − ⋅Mcc − ⋅Mcc − m0CC
αC− S− H =
MCH MEtt
(8) nearly 90 %, are observed at the surface 2 mm, gradually decreasing
m0
a⋅Mcc ⋅MCSH
CSH
until the carbonation front; however, Ett is completely carbonated at the

7
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

Fig. 10. Carbonation degrees of hydration products in OPC under 5-CO2 (a), 1-CO2 (b), and C-N(c), and LC3 under 5-CO2 (d), 1-CO2 (e), and C-N(f) at the carbonation
duration of 28 days, carbonation depth is represented by dashed line.

surface. Under C-N condition, the carbonation degrees of CH and C-S-H Below this threshold, the decomposition of the Ca–O sheets in C-S-H
are around 40 % at the surface and gradually diminish until 2 mm depth, leads to the precipitation and subsequent polymerization of silica gel
while the carbonation of ettringite is minimal. [45,46]. The heterogeneity of carbonated cementitious material can
For LC3 under the 5-CO2 condition, both the carbonation of the main result in the mixture of the high decalcified C-S-H and silica gel.
hydrate products, C-S-H and Ett, are nearly complete up to the carbon­ Therefore, the area where the Ca/Si ratio of C-S-H falls below 0.67
ation front at 15 mm, beyond which the carbonation degree drops (indicated by a dashed line in figures) can be considered as regions
significantly to around 40 %. Under the 1-CO2 condition, ettringite is where C-S-H decomposition occurs. For OPC, as shown in Fig. 11(a)–(c),
fully carbonated in the carbonated zone, while the carbonation degree of under 5-CO2 condition, the entire carbonated zone exhibits C-S-H
C-S-H is slightly lower at 70–80 %. Under C-N condition, the carbon­ decomposition due to the highest carbonation degree of hydration
ation degrees of C-S-H and ettringite at the surface are 56 % and 81 %, products than other conditions (Fig. 10(a)), indicated by a Ca/Si ratio
respectively. Notably, the carbonation degree of C-S-H in LC3 is lower than 0.67. Under the 1-CO2 condition, only the surface layer
consistently higher than in OPC across all carbonation conditions. This shows C-S-H decomposition, while the subsequent zone up to the
can be attributed to the lower content of CH and Ca/Si ratio in the C-S-H carbonation front shows C-S-H decalcification, with Ca/Si ratios higher
of LC3, which results in less calcium available to bind CO2, making it than 0.67. Under the C-N condition, no C-S-H decomposition occurs. For
more prone to decalcification and decomposition. LC3, as shown in Fig. 11(d)–(f), C-S-H decomposition occurs throughout
The Ca/Si ratio of C-S-H was obtained following the mass balance the entire carbonated zone due to the lower capacity to buffer the pH of
method described in [44], based on the XRD/Rietveld quantification the pore solution. Given that C-S-H decomposition and decalcification
results at each depth, as shown in Fig. 11. The Ca/Si ratios of C-S-H in result in different morphologies and subsequent precipitates, the zones
OPC and LC3 decrease in the carbonated zone due to the decalcification of C-S-H decomposition and decalcification under different carbonation
of C-S-H with the removal of the Ca. It has been reported that the conditions are highlighted by orange and green blocks, respectively, in
minimum allowable Ca/Si ratio for decalcified tobermorite is 0.67. the figures.

8
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

Fig. 11. Ca/Si ratio of C-S-H in OPC under 5-CO2 (a), 1-CO2 (b), and C-N(c), and LC3 under 5-CO2 (d), 1-CO2 (e), and C-N(f) at the carbonation duration of 28 days;
Ca/Si ratio of 0.67 is represented by a dashed line; zones of C-S-H decomposition and decalcification are highlighted by orange and green blocks, respectively. (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.3. Water content distribution detected by single-sided 1H NMR OPC and LC3 along the penetration direction from the exposed surface
during predrying. A continuous decline in water content was observed
Fig. 12 illustrates the distribution of the volumetric water content of along the penetration depth, with a more pronounced reduction near the

Fig. 12. Volumetric water content distribution during the RH controlling process.

9
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

surface owing to greater evaporation. Notably, a steeper gradient in the measured at each carbonation age, the colored bands indicate the
water content distribution near the surface was observed for LC3. standard deviation of each carbonation depth, and the red circles mark
Fig. 13 shows the volumetric water content distributions in OPC and the intersection points of the water content and carbonation depth.
LC3 along the depth from the exposed surface during carbonation under Additionally, the water content distribution of the reference samples
various conditions. The dashed lines represent the carbonation depth maintained at 60 % RH without CO2 exposure was continuously

Fig. 13. Volumetric water content distribution of OPC under 5-CO2 (a), 1-CO2 (b), and C-N (c), and LC3 under 5-CO2 (d), 1-CO2 (e), and C-N (f) during the
carbonation process.

10
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

monitored (Fig. A3 (Appendix C)). The water content of the non­ 1.61 for OPC and 1.17 for LC3.
carbonated pastes in OPC and LC3 remained identical after the 28-day Fig. 15 presents the BSE images of the C-S-H gel in noncarbonated
predrying process. Therefore, the significant decrease in the water OPC and LC3 captured at a magnification of 20,000×. The appearance of
content in the carbonated zone of both OPC and LC3 shown in Fig. 13 can the C-S-H in OPC shows fibrillar morphology, while the C-S-H in LC3
be attributed solely to the carbonation process. shows a foil-like and honeycomb-like morphology.
For OPC (Fig. 13(a)–(c)), under accelerated carbonation conditions
with 5 % and 1 % CO2 conditions, the water content at each carbonation
3.5. Microstructural change due to carbonation in OPC and LC3
front remains consistent, indicating that the carbonation progress stag­
nated at a similar water content over the 28-day carbonation duration.
Fig. 16 shows BSE images of non-carbonated and carbonated surface
This phenomenon was absent under natural carbonation, in which the
layer for OPC and LC3 under 5-CO2 condition at 28 days of carbonation
water content values at the carbonation fronts varied and were much
duration captured at a magnification of 200×. In BSE imaging, phase
lower than those under other conditions.
brightness is primarily determined by the atomic number of the inter­
For LC3 (Fig. 13(d)–(f)), the stagnation of the carbonation front at a
acting elements, with black areas representing voids. In the non-
similar water content was evident under 5 % CO2 conditions after three
carbonated samples, distinct phase contrasts are observed, with the
days of carbonation, characterized by stagnation of the carbonation
brightest regions indicating unreacted cement grains, and brighter grey
front at a consistent water content. The figure does not depict the
regions indicating hydration products. In the carbonated images, the
carbonation depth at 28 days because of the limitations of the single-
dark grey areas, which are rich in silicon, correspond to silica gel or
sided 1H NMR measurement range. Under the 1 % CO2 condition, the
decalcified C-S-H, while the brighter grey areas likely represent a
water content at each carbonation front varied, and the water content
mixture of CC polymorphs. Notably, shrinkage is observed in the
values at each carbonation front differed. These variations in the water
carbonated OPC and LC3 within the decalcified C-S-H and silica gel. For
content were similar to those observed under natural carbonation
non-carbonated LC3, the microstructure is more porous compared to
conditions.
OPC. After carbonation, LC3 exhibits even greater porosity due to the
negligible content of CH, resulting in a continuous reduction in solid
3.4. C-S-H analysis by SEM volume from the carbonation of C-S-H or C-A-S-H. The results indicate
that, upon complete carbonation, both OPC and LC3 exhibit a porous
Fig. 14(a) and (c) show BSE images of noncarbonated OPC and LC3 microstructure due to the reduction in total solid volume. Cementitious
captured at a magnification of 500×. The Ca/Si ratios of the non­ materials with lower CH content and Ca/Si ratio display a more pro­
carbonated OPC and LC3 were derived using the Edxia analysis per­ nounced porous microstructure after carbonation.
formed by Anaconda3 Glueviz [47]. The BSE images and corresponding
2D kernel density plots of Al/Ca versus Si/Ca are shown in Fig. 14(b) and 4. Discussion
(d). The core of the density plot is denoted by a darker color, repre­
senting the C-S-H cloud. Accordingly, the Ca/Si ratios are calculated as Understanding the influence of the CO2 concentration on the

Unreacted cement grain

C-S-H

Metakaolin

Calcite
Portlandite

(a) (c)

(b) (d)

Fig. 14. BSE images of noncarbonated OPC (a) and LC3 (c), and corresponding Al/Ca versus Si/Ca plot for OPC (b) and LC3 (d).

11
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

Outer C-S-H
Outer C-S-H Inner C-S-H

Cement grains

(a) (b)

Fig. 15. BSE images of C-S-H gel in noncarbonated OPC (a) and LC3 (b).

Unreacted cement grain CC amorphous

C-S-H

Decalcified C-S-H or silica gel

Portlandite

(a) (b)

C-(A)-S-H

Decalcified C-S-H or silica gel

Calcite Metakaolin

(c) (d)
3
Fig. 16. BSE images of OPC sample before (a) and after carbonation (b); LC sample before (c) and after carbonation (b).

carbonation process is fundamental for developing an accelerated content distribution near the surface was observed in LC3, indicating
carbonation method for cementitious materials. This paper provides differences in the water movement between OPC and LC3, which follows
insights into the water content and mineralogical phases of OPC and LC3 the nonlinear diffusion equation [48]. In this study, the Matano method
during the carbonation process using single-sided 1H NMR, a nonde­ [49] based on the Boltzmann transformation was employed to deter­
structive technique, complemented by μ-XRD mapping measurement. mine the water diffusion coefficient, with respect to the Boltzmann
variable λ as follows:
x
4.1. Water content evolution during the predrying process λ = √̅̅ (9)
t

According to the water content distributions of OPC and LC3 during where x is the distance from the exposed surface (mm), and t is the
the predrying process shown in Fig. 12, a steeper gradient of water

12
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

drying duration (days). 1E-8

OPC
Based on the experimental results in Fig. 12, Fig. 17 plots the rela­
LC3
tionship between water content w and the Boltzmann variable λ in OPC 1E-9
and LC3, and the curves fitted using the hyperbolic relationship equation
proposed in [50]: 1E-10

D (m2/s)
a
w = 1+f − ( )2 (10) 1E-11
λ
2
+b
1E-12

where a, b, and f are constants determined from the shapes of the curves, 1E-13
as shown in Fig. 17. 0.20 0.25 0.30 0.35 0.40 0.45 0.50
Accordingly, the water diffusion coefficient D (m2/s) corresponding Water content (g/cm3)
to each w can be derived using the following function:
Fig. 18. Water diffusion coefficient in OPC and LC3 during predrying process.
∫ /
1 0 ∂w
D(w) = ⋅ λdw (11)
2 w ∂λ pores (interhydrate and/or capillary pores) for LC3. The results reveal
that the predominant reduction in water content occurs in coarse pores
Fig. 18 shows the different water diffusion coefficients of OPC and
owing to evaporation, whereas the water content in fine pores is
LC3 during the predrying process. It is clear to see that at similar water
retained, as explained by the Kelvin equation [56]. For OPC, the water
contents, the D of LC3 is one to two orders of magnitude lower than that
content in fine pores decreases from approximately 0.26 to 0.22 g/cm3
of OPC, leading to a steeper gradient of water content distribution in
after the drying process. Similarly, for LC3, the water content in fine
LC3.
pores decreases from approximately 0.35 to 0.30 g/cm3 after drying.
The lower water diffusion coefficient can be explained by the foil-like
This indicated that the water content in the fine pores was reduced to
and honeycomb-like morphology C-S-H [51,52] with a lower Ca/Si ratio
approximately 85 %–86 % of the original water content after the drying
in LC3 (shown in Fig. 15(b)) which makes water transport more difficult.
process.
In contrast, the C-S-H in OPC has a higher Ca/Si ratio and fibrillar
Fig. 20 illustrates the evolution of T2 relaxation times for different
morphology [53,54] (shown in Fig. 15(a)), facilitating easier water
pore classes in OPC and LC3 during the drying process. Because the T2
transport. Similar phenomena have been detected in other SCMs with a
values correspond to the volume-to-surface ratio [57], the observed
lower Ca/Si C-S-H ratio, which is related to the lower connectivity or
decreases in the T2 values for both OPC and LC3 indicate significant
tortuosity of the pore system [54,55].
water removal from the coarse pores and an approximately 25 %
Fig. 19 illustrates the water content distribution before and after a
reduction in the water content in the fine pores. Additionally, T2 values
28-day drying process, categorized into different pore classes: fine pores
associated with the minimal water content were omitted from the
(interlay space and gel pores) and coarse pores (interhydrate and/or
Fig. 20 to prevent misleading interpretations. Notably, the decrease in
capillary pores) for OPC, and interlayer space, gel pores, and coarse
T2,2 from the exposed surface of the gel pore water in LC3 suggests a
transformation from gel pores to interlayer spaces. This transformation
0.50
is due to the stacking and agglomeration of the C-S-H sheets driven by
capillary tension or the attraction force between the interlayer counter
0.45 ions of Ca2+ and the negatively charged C-S-H sheet surfaces as water
Water content (g/cm3)

molecules are removed from the space between C-S-H sheets (interlayer
0.40
a 0.2014 space and gel pores) [58–60], thus expanding the coarse pore outside of
the C-S-H agglomeration [61,62]. This phenomenon was further visu­
0.35 b 0.9993
f -0.572
R2 0.85 alized in the water content distribution near the surface (Fig. 19(d)),
0.30
3d where severe drying resulted in a decreased gel pore water and an
0.25 7d increased water content in the interlayer space. The expansion of the
14d coarse pore volume facilitates CO2 penetration from external conditions
0.20
28d during the subsequent carbonation process.
0 2 4 6 8 10
λ 4.2. Relationship between calcium carbonate and water content
(a)
0.50
The amount of water present in the pores influences the formation of
the carbonation product polymorphs. Fig. 21 shows the relationship
0.45 between the quantities of calcite, vaterite, and aragonite and the water
content in the completely carbonated zone. For both OPC and LC3, an
Water content (g/cm3)

0.40
a 0.0127 increase in the water content favors the formation of calcite, whereas a
b 0.1293
0.35
f -0.543 decrease in the water content leads to a higher amount of vaterite. This
R2 0.8 phenomenon was attributed to the inhibition of the polymorphic
0.30
3d transformation from metastable vaterite to thermodynamically stable
0.25 7d calcite in the absence of sufficient water [18,63]. Lack of water inhibits
14d the dissolution of metastable polymorphs and the growth of stable
0.20
28d polymorphs [63]. In the case of LC3, calcite consistently appeared in
0 2 4 6 8 10 greater amounts than vaterite. This can be explained by the seeding
λ effect of calcite, which was already present in the LC3 paste (Fig. 9). The
(b) original calcite acts as a nucleation seed during the recrystallization
process, promoting secondary nucleation and accelerating the trans­
Fig. 17. Relationship between water content and Boltzmann variable λ for OPC formation of vaterite into calcite [64]. The presence of aragonite did not
(a) and LC3 (b). show any indicative change with varying water content, likely due to its

13
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

Fig. 19. Water content in different classes in OPC before (a) and after (b) the pre-drying process, and in LC3 before (c) and after (d) the pre-drying process.

Fig. 20. Distribution of T2 relaxation time during the drying process for OPC at the duration of 3 days (a), 7 days (b), 14 days (c) and 28 days (d), and LC3 at the
duration of 3 days (e), 7 days (f), 14 days (g) and 28 days (h). T2 values, corresponding to the minimal water content, were omitted from the figure.

minimal amount. indicate the standard deviation of each carbonation depth, and the red
circles mark the intersection points of the water content and carbonation
depth. The water content distribution in OPC and LC3 under other
4.3. Water content evolution and dynamic change of C-S-H during the
carbonation conditions is detailed in Appendix D. For both OPC and LC3,
carbonation process
the total water content was released as carbonation progressed. Notably,
within the carbonated zone in LC3, interlayer water content increased,
During the carbonation process, Fig. 22 shows the representative
while gel pore water content decreases within the carbonated zone.
distribution of water content in different pore classes and carbonation
Further insights into the dynamic microstructural changes are pro­
depth in OPC and LC3 under 1-CO2 condition. The dashed lines represent
vided by the T2 distribution along the carbonation direction in OPC and
the carbonation depth measured at each carbonation age, the grey bands

14
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

Fig. 21. Relationship between phase amount of carbonation products (calcite, vaterite, aragonite) and water content in OPC (a) and in LC3 (b).

Fig. 22. Water content distribution in different pore classes in OPC at 7 days (a), 14 days (b) and 28 days (c), and LC3 at 7 days (d), 14 days (e) and 28 days (f) during
the carbonation process under 1-CO2.

LC3, which corresponds to the pores size where water is present. Fig. 23 carbonation duration. The C-S-H decomposition zone and the C-S-H
shows the representative T2 distribution for OPC under the 5-CO2 and 1- decalcification zone, with exhibit different dynamic changes and mor­
CO2 conditions, and for LC3 under 1-CO2 condition at 28 days of phologies of C-S-H, are highlighted by an orange block and green block

15
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

2.0 2.0 6
T2,1 T2,2 T2,1 T2,2 T2,1 T2,2 T2,3
5

Relaxation time (ms)

Relaxation time (ms)
Relaxation time (ms)
1.5 1.5
4

1.0 1.0 3

2
0.5 0.5
1
0.0 0.0 0
0 3 6 9 12 15 0 3 6 9 12 15 0 3 6 9 12 15
Depth (mm) Depth (mm) Depth (mm)
(a) (b) (c)

Fig. 23. Distribution of T2 relaxation time for OPC under 5-CO2 (a) and 1-CO2 (b), and LC3 under 1-CO2 (c) at the 28 days of carbonation duration; T2 values
associated with the minimal water content were omitted from the figure.

in the figure, respectively. T2 values associated with minimal water carbonated cementitious material can result in a mixture of highly
content were omitted from the figure. decalcified C-S-H and silica gel. As shown in in Fig. 23(a), the T2,1 values
As the decalcification of the C-S-H progresses, as observed in in the C-S-H decomposition zone in OPC are slightly lower than those in
carbonated zone in OPC under 1-CO2 and C-N conditions (Fig. 11(b), C-S-H decalcification zone (Fig. 23(b)). This indicates that the fine pore
(c)), the basal space increases with a deceasing Ca/Si ratio [65]. This size, where water is present, is smaller in the C-S-H decomposition zone.
increase in basal spacing can be attributed to the charge-balancing effect This trend is more pronounced in the C-S-H decomposition zone in LC3,
of Ca2+ in the C-S-H interlayer space, which reduces the attractive force as shown in Fig. 23(c), where both T2,2 and T2,3 values decrease in the
and increase the mobility between the negatively charged C-S-H layers carbonated zone, reflecting a reduction in fine pore size. Notably, the
[66]. Consequently, the increased basal spacing and the enlarged water content in the interlayer space increases, while the water content
necking pores facilitate the rapid removal of water from the interlayer in gel pores decreases (Fig. 22(d)–(e)). This suggests that when the
space and gel pore. With promoting further agglomeration and stacking tobermorite structure of C-S-H decomposed, with a formation of the
of C-S-H, and creating the space outside of C-S-H agglomeration en­ mixture of highly decalcified C-S-H, silica gel, and CC phases, the water
hances the interconnections among coarse pores during the carbonation content is presented in the porous silica gel, the narrow spaces between
process [67]. This phenomenon is also evidenced by the reduction in T2,1 the silica gel and/or CC phases.
values in the carbonated zone, which reflect densification of fine pores, Furthermore, it has been reported that the volumetric contraction of
as shown in Fig. 23(b). silica gel is greater than that of decalcified C-S-H because the silica gel
At extensive levels of decalcification, where the Ca/Si ratio drops can polymerize freely in three dimensions, whereas C-S-H which is
below 0.67, the decomposition of the Ca–O sheets in C-S-H leads to the constrained to two dimensions [68], thus generating more significant
precipitation and subsequent polymerization of silica gel. This phe­ shrinkage of silica gel. Overall, both the decomposition and decalcifi­
nomenon is observed in OPC under 5-CO2 and in LC3 under the all cation of C-S-H lead to increased volume and interconnectivity of coarse
carbonation conditions (Fig. 11(a), (d)–(f)). The heterogeneity of the pores, facilitating gas penetration.

Fig. 24. Dynamic changes of nanostructure of C-S-H and water molecules during drying, decalcification, and decomposition processes through the carbonation
duration in OPC and LC3 (dark blue squares represent the water molecules in coarse pores; light blue squares represent the water molecules in fine pores). (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

16
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

The dynamic changes in the nanoscale pore structure of C-S-H and in Fig. 25, the plugging effect is associated with the water content of the
the behavior of water molecules during drying, decalcification, and fine pores. When the water content in the fine pores exceeds 0.81–0.85
decomposition processes through the carbonation duration, are illus­ times the original water content, the carbonation process stagnates
trated in Fig. 24. under accelerated carbonation conditions. Conversely, when the water
content in the fine pores is below this threshold, carbonation can pro­
4.4. Water content at the carbonation front in OPC and LC3 ceed under natural conditions. This limitation of the water content in the
fine pores is related to the connectivity and space of the coarse pores
Based on the water content distribution in different pore classes and between the C-S-H layers, which affects CO2 penetration. Specifically,
carbonation depths at each carbonation age in OPC and LC3 (Figs. 22, Fig. 26 presents a schematic illustrating the carbonation process under
A4, and A5), the water contents in each pore class at the carbonation accelerated carbonation conditions, where the plugging effect is present,
front are summarized in Fig. 25. For OPC under 5 % CO2 condition, as and natural carbonation conditions, where it is absent.
shown in Fig. 25(a), the carbonation front exhibited stagnation at As discussed in Section 4.2, during the predrying process, the
minimal water content in the coarse pores while maintaining a similar removal of water molecules from coarse pores by evaporation and the
water content of approximately 0.21 g/cm3 in the fine pores, which is increased space in coarse pores owing to the agglomeration and stacking
81 % of the original water content in fine pores. For LC3 under 5 % CO2 of C-S-H sheets enhance the interconnectivity of the pore structure in the
condition, as illustrated in Fig. 25(b), a similar stagnation in water levels paste. Given that the diffusion coefficient of CO2 in air is over 6000 times
was observed at ages of 7 and 14 days, with a consistent total water higher in liquid, this increased empty space in coarse pores allows CO2
content of about 0.30 g/cm3 in the fine pores, which is 85 % of the to rapidly penetrate from the external environment, initiating the initial
original water content in fine pores. A similar phenomenon was carbonation (Stage I). Once CO2 gas penetrates the gaps between the C-
observed for OPC under the 1 % CO2 carbonation condition, where the S-H agglomerations and silica gel, it rapidly dissolves into the interhy­
water content in fine pores stagnated at approximately 0.20 g/cm3, in­ drated water and hydrolyzes to form bicarbonate (HCO−3 ) and carbonate
dependent of the water content in coarse pores. For LC3, although the ions (CO2− 2+
3 ). Calcium ions (Ca ) then migrate from C-S-H and CH and
total water content at the carbonation front varied (Fig. 13(e)), the react with dissolved CO2 to precipitate calcium carbonates with various
water content in the fine pores remained constant at approximately 0.30 polymorphs at locations with abundant water. During the carbonation
g/cm3. The average water content in fine pores is denoted by a dashed process (Stage II), under natural carbonation conditions, a low con­
line. centration of CO2 allowed fewer CO2 molecules to penetrate, preventing
Under natural carbonation conditions, Fig. 25 illustrates varying the water content from governing the carbonation process. Conversely,
levels of water content at the carbonation front for both OPC and LC3. under accelerated carbonation conditions, a higher degree of carbon­
Notably, the water content in the fine pores at the carbonation front was ation in the carbonated zone induces severe decalcification and
lower than that observed under the other carbonation conditions. decomposition of the C-S-H, leading to significant dynamic changes in
its nanostructure. As carbonation progressed, C-S-H underwent densi­
fication and further polymerization of the silica gel, resulting in the
4.5. Mechanism of plugging effect during carbonation process
shrinkage of fine pores and loss of water content in the coarse pores.
Subsequently, the interconnectivity of the coarse pores increased and
As mentioned in Section 4.3, the consistency of the water content in
calcium carbonates precipitated at the tips of the C-S-H fibrils [69,70].
the fine pores of OPC and LC3 at the carbonation front under 5 % and 1 %
Despite the increase of solid phase due to carbonation, as shown in the
CO2 concentrations indicates that water in the fine pores plays a
quantitative result of μ-XRD (Fig. 9) the increased space and more
dominant role in the carbonation process, where the plugging effect is
interconnected channels in the carbonated zone facilitate the penetra­
evident. Conversely, the significantly lower water content at the
tion of CO2 gas molecules [71,72]. This continues until the depth where
carbonation front indicates that the carbonation process under natural
the water content in the fine pores is higher than the threshold value of
carbonation conditions is not predominantly governed by the water
0.81–0.85 times the original water content. At this point, the remaining
content, where such a plugging effect is absent. The plugging effect can be
space in the coarse pores is insufficient for large amounts of CO2 mol­
explained by the dynamic change in the nanostructure of C-S-H during
ecules to penetrate, leading to the stagnation of the carbonation process
the drying and carbonation processes.
at the same water content limitation in the fine pores. This results in the
As noted in [61], a gap appears in fine pores only when the RH is
plugging effect observed in Stage III.
below 40 %, indicating that in our study, the water content in the fine
Under accelerated carbonation conditions with higher CO2 concen­
pores is saturated at 60 % RH. Therefore, the decreased water content in
trations (5 % and 1 %), the water content in the fine pores of C-S-H
the fine pores during the drying and carbonation processes, owing to the
predominantly influenced carbonation progress, resulting in a notice­
agglomeration and stacking of C-S-H sheets, corresponds to an increase
able plugging effect. This plugging effect was absent under natural
in the number of coarse pores. As discussed in Section 4.3 and illustrated

Fig. 25. Water content and its composition at the carbonation front of OPC (a) and LC3 (b); dashed lines represent the average level of water content in fine pores
under accelerated carbonation conditions.

17
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

Fig. 26. Plugging effect and CO2 penetration during the carbonation process. Dark blue lines represent monolayers of C-S-H, the light blue area represents the water
content, and the penetration of CO2 is represented as the orange arrows. (For interpretation of the references to color in this figure legend, the reader is referred to the
web version of this article.)

carbonation conditions, where the process was instead limited by the 5. Conclusions
low CO2 concentration. This allowed sufficient space for the few CO2
molecules to penetrate without being impacted by the plugging effect of This study aims to understand the impact of CO2 concentration on
water content. the carbonation process by investigating the water content distribution
The occurrence of this plugging effect under varying CO2 concen­ and mineralogical phases using single-sided 1H NMR and μ-XRD map­
trations is likely related to difference in Ca/Si ratios and the resulting C- ping. This provides a comprehensive analysis of two cementitious ma­
S-H morphology in cementitious materials. As described in Section 4.4, terials, OPC and LC3 pastes, under different CO2 concentrations, offering
the consistent water content in fine pores under accelerated carbonation direct evidence of the inhibitory effect of water content on the carbon­
conditions for OPC and LC3 where the plugging effect occurs can be ation process.
treated as the reference value: approximately 0.20 g/cm3 for OPC and
0.30 g/cm3 for LC3, as illustrated in Fig. 25. The relative ratios of water 1) Water content and carbonation product polymorphs: The amount of
content at the carbonation front under C-N condition for OPC and LC3 water present in the pores influenced the formation of carbonation
compared to these reference values indicate the occurrence of plugging product polymorphs. An increase in the water content favors the
effect. These relative ratios plotted against CO2 concentration are shown formation of calcite, whereas a decrease in the water content leads to
in Fig. 27. It is observed that when CO2 concentrations are below 1 %, a higher amount of vaterite. In LC3, calcite predominates over
the relative ratio indicating the occurrence of plugging effect is consis­ vaterite owing to the seeding effect of the pre-existing calcite, which
tently higher than in LC3, suggesting a greater likelihood of plugging accelerates the transformation of the metastable phase to the stable
effect in cementitious material with a higher Ca/Si ratio. This finding phase.
implies that similar effects may occur in other supplementary materials 2) The dynamic change of C-S-H during decalcification and decompo­
with varying Ca/Si ratios, further investigation is warranted. sition process: During the carbonation process, the decalcification
and decomposition of C-S-H leads to extensive agglomeration and
dense stacking of C-S-H and the shrinkage in the formed silica gel in
OPC and LC3, causing an increase in the space and interconnectivity
Relative ratios of water content at the carbonation front

of coarse pores in the carbonated zone.

3) Plugging effect under accelerated carbonation: Under accelerated
carbonation conditions (5 % CO2, 1 % CO2), the carbonation progress
1.0 was primarily governed by the water content in the fine pores
(interlayer space and gel pores). The plugging effect is evident for both
0.8 OPC and LC3 with a threshold value of water content in fine pores,
0.81–0.85 times the original water content, affecting further CO2
0.6 penetration. Conversely, under natural carbonation conditions, the
plugging effect was absent, which can be attributed to the slower pace
0.4
of carbonation compared to the drying process. Such a plugging effect
0.2
OPC (Ca/Si=1.61) is more likely to occur in cementitious material with a higher Ca/Si
LC3(Ca/Si=1.17) ratio.
0.0 4) Future studies on the accelerated carbonation method: To accurately
0 1 5 characterize cementitious materials under natural carbonation con­
CO2 concentration (%) ditions, the CO2 concentration used in accelerated carbonation
methods should be constrained to prevent the plugging effect. Further
Fig. 27. Trend of plugging effect occurrence at different C/S ratios across
studies should explore the use of lower CO2 concentrations and
various CO2 concentrations. various pretreatments to refine the accelerated carbonation method.

18
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

CRediT authorship contribution statement interests or personal relationships that could have appeared to influence
the work reported in this paper.
Luge Cheng: Writing – original draft, Visualization, Investigation,
Formal analysis. Ryo Kurihara: Writing – review & editing, Method­ Data availability
ology, Investigation. Takahiro Ohkubo: Writing – review & editing,
Methodology, Formal analysis. Ryoma Kitagaki: Writing – review & Data will be made available on request.
editing, Funding acquisition. Atsushi Teramoto: Writing – review &
editing, Funding acquisition. Yuya Suda: Methodology, Funding Acknowledgments
acquisition. Ippei Maruyama: Writing – review & editing, Visualiza­
tion, Supervision, Project administration, Methodology, Funding This study was based on the results obtained from a project
acquisition, Formal analysis, Data curation, Conceptualization. (JPNP21023) commissioned by the New Energy and Industrial Tech­
nology Development Organization (NEDO).
Declaration of competing interest

The authors declare that they have no known competing financial

Appendix A. A comparative analysis of the NMR signals by single-sided 1H NMR and 1H NMR with homogeneous magnetic field

Single-sided 1H-NMR Single-sided 1H-NMR

Signal Intensity (Norm.)

Signal Intensity (Norm.)

Conventional 1H-NMR Conventional 1H-NMR

0.01 1 100 10000 0.01 1 100 10000

T2 Relaxiation time (ms) T2 Relaxiation time (ms)
(a) (b)

Fig. A1. T2 distribution of OPC (a) and LC3 (b) obtained from single-sided H-NMR and 1H NMR with homogeneous magnetic field.

1
H NMR relaxometry, employing a strong homogeneous magnetic field, has been extensively used to identify water reservoirs in hardened cement
paste [30,58] using the CPMG technique [23,72] and inverse Laplace transform (ILT) calculations [73]. An open-source 1D/2D Inverse Laplace
program [74] was used for ILT calculations. The validity of the single-sided 1H NMR measurement was confirmed through a comparative analysis of
the NMR signals from the OPC and LC3 paste after hydration, utilizing both single-sided 1H NMR and 1H NMR with a homogeneous magnetic field
using an Oxford Benchtop NMR Analyzer (MQC+). This device operates at a resonance frequency of 23.4 MHz (0.47 T), and the π/2 pulse length is
calibrated before each measurement. A total of 300 logarithmically spaced echoes, ranging from 0.04 to 6000 ms, were recorded using 256 scans to
achieve an adequate S/N ratio of 6.01 × 108 for OPC and 9.02 × 105 for LC3. For single-sided 1H NMR measurement, a significantly high scan number
of 4096 was implemented to ensure the acquisition of high-quality data necessary for ILT calculation with an S/N ratio of 278 for OPC and 441 for LC3.
Fig. A1 shows the normalized quasicontinuous T2 distribution for both OPC (a) and LC3 (b) paste, as determined by homogeneous and single-sided 1H
NMR measurements. While both measurements demonstrate comparable effectiveness in water capture, the single-sided 1H NMR profile shows a
minor peak shift in OPC, indicating a limitation in detecting minimal water components with short relaxation time, approximately ranging from 0.01
to 0.1 ms, typically associated with interlayer water. This inconsistency in the T2 relaxation distribution between single-sided 1H NMR and 1H NMR
with a homogeneous magnetic field is due to the differences in the magnetic fields.

Appendix B. XRD scan patterns of OPC and LC3 paste across the carbonation depth

19
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

P C
C

20
15

15

)
h(mm
V

m)
10

t h(m
Q

Intensity (Counts)
Intensity (Counts)

10

Dept
Dep
5
V 5

20 30 40 50 60 70 10 20 30 40 50 60 70
theta (°) theta (°)

(a) (d)
P

15
15
C
C
10

m)
m)

t h(m
10
t h(m

Intensity (Counts)

Q
Intensity (Counts)

Dep
Dep

V
5
5 V

20 30 40 50 60 70 10 20 30 40 50 60 70
theta (°) theta (°)

(b) (e)
P

15 15

C
m)

10
)
10
h(mm

C Q
t h(m
Intensity (Counts)

Intensity (Counts)

Dept
Dep

5 5
V

10 20 30 40 50 60 70 10 20 30 40 50 60 70
theta (°) theta (°)

(c) (f)
Fig. A2. 3D plot of XRD patterns distribution along the carbonation depth of OPC under 5-CO2 (a), 1-CO2 (b), and C-N (c), and LC3 under 5-CO2 (d), 1-CO2 (e), and C-
N (f). (P: Portlandite; C: calcite; V: vaterite; Q: quartz).
Fig. A2 presents the XRD patterns of the OPC and LC3 paste at 1 mm intervals from the carbonation surface to a depth of 15 mm under 5 % and 1 %
CO2 and natural carbonation conditions after 28 days. For the OPC, under accelerated carbonation conditions, prominent peaks of calcium carbonate
(calcite and vaterite) were observed, with no obvious peaks of ettringite. In contrast, the CH peaks were more pronounced and became increasingly
apparent as depth increased. Under natural carbonation, a negligible vaterite peak was observed, and the CH peaks were more distinct, even within the
carbonated area, compared to the other conditions. For LC3, the intensity of the calcite peak at 29◦ was enhanced in the carbonated zone relative to the
noncarbonated zone, along with the appearance of peaks at approximately 47◦ –48◦ , whereas ettringite peaks vanished in the carbonated zone.

20
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

Appendix C. Water content distribution in reference OPC and LC3 paste

0.5
0.5

0.4

Water content (g/cm3)

Water content (g/cm3)
0.4

0.3 0.3

0.2 0.2
28d 28d
0.1 35d 0.1 35d
42d 42d
56d 56d
0.0 0.0
0 3 6 9 12 15 0 3 6 9 12 15
Depth (mm) Depth (mm)

(a) (b) 3

Fig. A3. Water content distribution of OPC (a) and LC3 (c) controlled under 60 % RH condition after the drying pretreatment.

The reference samples of the OPC and LC3 pastes were controlled under 60 % RH conditions without CO2 after the pretreatment of the drying
process, the water content of which was monitored continuously, as shown in Fig. A3.

Appendix D. Water content distribution in different pore classes in OPC and during LC3 during carbonation progress

Figs. A4 and A5 show the water content distribution in different pore classes in OPC and LC3 during carbonation under 5 % CO2 and natural
carbonation conditions, respectively.

21
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

0.5 0.7
Fine pore Interlayer space
Coarse pore 0.6 Gel pore
0.4

Water content (g/cm3)

Coarse pore

Water content (g/cm3)

0.5
0.3 0.4

0.2 0.3

0.2
0.1
0.1

0.0 0.0
0 3 6 9 12 15 0 3 6 9 12 15
Depth (mm) Depth (mm)

(a) (d)
0.5 0.7
Fine pore Interlayer space
Coarse pore 0.6 Gel pore
0.4

Water content (g/cm3)

Water content (g/cm3)

Coarse pore
0.5
0.3 0.4

0.2 0.3

0.2
0.1
0.1

0.0 0.0
0 3 6 9 12 15 0 3 6 9 12 15
Depth (mm) Depth (mm)

(b) (e)

0.5 Fine pore

Coarse pore
Water content (g/cm3)

0.4

0.3

0.2

0.1

0.0
0 3 6 9 12 15
Depth (mm)

(c)

Fig. A4. Water content in different pore classes in OPC at 7 days (a), 14 days (b) and 28 days (c), and in LC3 at 7 days (d) and 14 days (e) during the carbonation
process under 5 % CO2.

22
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

0.5 0.7
Fine pore Interlayer space
Coarse pore 0.6 Gel pore
0.4

Water content (g/cm3)

Water content (g/cm3)
Coarse pore
0.5
0.3 0.4

0.2 0.3

0.2
0.1
0.1
0.0 0.0
0 3 6 9 12 15 0 3 6 9 12 15
Depth (mm) Depth (mm)

(a) (c)
0.5 0.7
Fine pore Interlayer space
Coarse pore 0.6 Gel pore
0.4
Water content (g/cm3)

Water content (g/cm3)

Coarse pore
0.5
0.3 0.4

0.2 0.3

0.2
0.1
0.1

0.0 0.0
0 3 6 9 12 15 0 3 6 9 12 15
Depth (mm) Depth (mm)

(b) (d)
0.7
Interlayer space
0.6 Gel pore
Water content (g/cm3)

Coarse pore
0.5

0.4

0.3

0.2

0.1

0.0
0 3 6 9 12 15
Depth (mm)

(e)
3
Fig. A5. Water content distribution in different pore classes in OPC at 28 days (a) and 56 days (b) in LC at 14 days (c), 28 days (d), and 56 days (e) during the
carbonation process under 0.04 % CO2.

References [9] S.A. Bernal, J.L. Provis, D.G. Brice, A. Kilcullen, P. Duxson, J.S.J. Van Deventer,
Accelerated carbonation testing of alkali-activated binders significantly
underestimates service life: the role of pore solution chemistry, Cem. Concr. Res. 42
[1] Tuutti, Corrosion of steel in concrete, in: Dissertation, Lund University, Sweden:
(10) (2012) 1317–1326, https://doi.org/10.1016/j.cem-conres.2012.07.002.
Swedish Cement and Concrete Research Institute, 1982.
[10] J.H.M. Visser, Influence of the carbon dioxide concentration on the resistance to
[2] L. Cheng, I. Maruyama, A prediction method for the corrosion rate of steel rebar in
carbonation of concrete, Construct. Build Mater. 67 (2014) 8–13, https://doi.org/
carbonated mortar under variable environmental conditions, J. Adv. Concrete
10.1016/j.conbuildmat.2013.11.005.
Technol. 21 (8) (2023) 611–630, https://doi.org/10.3151/jact.21.611.
[11] H. Cui, W. Tang, W. Liu, Z. Dong, F. Xing, Experimental study on effects of CO2
[3] W. Ashraf, Carbonation of cement-based materials: challenges and opportunities,
concentrations on concrete carbonation and diffusion mechanisms, Construct.
Construct. Build Mater. 120 (2016) 558–570, https://doi.org/10.1016/j.
Build Mater. 93 (2015) 522–527, https://doi.org/10.1016/j.
conbuildmat.2016.05.080.
conbuildmat.2015.06.007.
[4] I.-S. Yoon, O. Çopuroğlu, K.-B. Park, Effect of global climatic change on
[12] M. Castellote, L. Fernandez, C. Andrade, C. Alonso, Chemical changes and phase
carbonation progress of concrete, Atmos. Environ. 41 (34) (2007) 7274–7285,
analysis of OPC pastes carbonated at different CO2 concentrations, Mater. Struct.
https://doi.org/10.1016/j.atmosenv.2007.05.028.
Constr. 42 (4) (2009) 515–525, https://doi.org/10.1617/s11527-008-9399-1.
[5] L. Li, M. Wu, An overview of utilizing CO2 for accelerated carbonation treatment in
[13] V. Shah, K. Scrivener, B. Bhattacharjee, S. Bishnoi, Changes in microstructure
the concrete industry, J. CO2 Util. 60 (2021) 102000, https://doi.org/10.1016/j.
characteristics of cement paste on carbonation, Cem. Concr. Res. 109 (2018)
jcou.2022.102000.
184–197, https://doi.org/10.1016/j.cemconres.2018.04.016.
[6] I. O. for Standardization, ISO 1920-12-2015 Testing of Concrete. Part 12:
[14] M. Thiéry, P. Faure, A. Morandeau, G. Platret, J.F. Bouteloup, P. Dangla,
Determination of the Carbonation Resistance of Concrete, 2015.
V. Baroghel-Bouny, Effect of carbonation on the microstructure and the moisture
[7] GB/T 50082-2009, Standard for Test Methods of Long-Term Perfornance and
properties of cement-based materials, Int. Conf. Durab. Build. Mater. Components
Durability of Ordinary Concrete, 2009.
3 (2011) 1–8 [Online]. Available: http://scholar.google.com/scholar?
[8] J. Japanese Industrial Standard Committee: Tokyo, JIS A 1153; 2012, Method of
hl=en&btnG=Search&q=intitle:Effect+of+carbonation+on+the+microstructure
Accelerated Carbonation Test for Concrete, 2012, p. 2012.
+and+the+moisture+properties+of+cement-based+materials#2.

23
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

[15] A. Leemann, F. Moro, Carbonation of concrete: the role of CO2 concentration, study. Part II: carbonated phases, J. Am. Ceram. Soc. 90 (3) (2007) 908–917,
relative humidity and CO2 buffer capacity, Mater. Struct. 50 (1) (2017) 30, https:// https://doi.org/10.1111/j.1551-2916.2006.01429.x.
doi.org/10.1617/s11527-016-0917-2. [44] Y. Suda, T. Saeki, T. Saito, Relation between chemical composition and physical
[16] T.T. Saeki, H. Ohga, S. Nagataki, Mechanism of carbonation and prediction of properties of C-S-H generated from cementitious materials, J. Adv. Concrete
carbonation process of concrete, Doboku Gakkai Ronbunshu 1990 (414) (1990) Technol. 13 (5) (2015) 275–290, https://doi.org/10.3151/jact.13.275.
99–108, https://doi.org/10.2208/jscej.1990.414_99. [45] G.W. Groves, A. Brough, I.G. Richardson, C.M. Dobson, Progressive changes in the
[17] S. Steiner, B. Lothenbach, T. Proske, A. Borgschulte, F. Winnefeld, Effect of relative structure of hardened C 3 S cement pastes due to carbonation, J. Am. Ceram. Soc.
humidity on the carbonation rate of portlandite, calcium silicate hydrates and 74 (11) (1991) 2891–2896, https://doi.org/10.1111/j.1151-2916.1991.tb06859.
ettringite, Cem. Concr. Res. 135 (2020) 106116, https://doi.org/10.1016/j. x.
cemconres.2020.106116. [46] J.J. Chen, J.J. Thomas, H.M. Jennings, Decalcification shrinkage of cement paste,
[18] E. Drouet, S. Poyet, P. Le Bescop, J.M. Torrenti, X. Bourbon, Carbonation of Cem. Concr. Res. 36 (5) (2006) 801–809, https://doi.org/10.1016/j.
hardened cement pastes: influence of temperature, Cem. Concr. Res. 115 (2019) cemconres.2005.11.003.
445–459, https://doi.org/10.1016/j.cem-conres.2018.09.019. [47] F. Georget, W. Wilson, K.L. Scrivener, Edxia: microstructure characterisation from
[19] I. Galan, C. Andrade, M. Castellote, Natural and accelerated CO2 binding kinetics in quantified SEM-EDS hypermaps, Cem. Concr. Res. 141 (2021) 106327, https://doi.
cement paste at different relative humidities, Cem. Concr. Res. 49 (2013) 21–28, org/10.1016/j.cemconres.2020.106327.
https://doi.org/10.1016/j.cemconres.2013.03.009. [48] K. Sakata, A study on moisture diffusion in drying and drying shrinkage of
[20] L. Cheng, N. Saeki, R. Kurihara, I. Maruyama, Characterization of capillary pores of concrete, Cem. Concr. Res. 13 (2) (1983) 216–224, https://doi.org/10.1016/0008-
carbonated cement apste under different relative humidity conditions by SEM 8846(83)90104-7.
image analysis, Cem. Sci. Concr. Technol. 77 (1) (2024) 308–316, https://doi.org/ [49] C. Matano, On the relation between diffusion-coefficients and concentrations of
10.14250/cement.77.308. solid metals, Jpn. J. Phys. 8 (1933) 109–113.
[21] F. Zunino, K. Scrivener, Microstructural developments of limestone calcined clay [50] H. Akita, T. Fujiwara, Y. Ozaka, Water movement within mortar due to drying and
cement (LC3) pastes after long-term (3 years) hydration, Cem. Concr. Res. 153 wetting (in Japanese), J. JSCE 13 (420) (1990) 61–69 ([Online]. Available: http://
(2022) 106693, https://doi.org/10.1016/j.cemconres.2021.106693. link.springer.com/10.1007/978-3-319-59379-1%0Adoi:10.1016/B978-0-12-
[22] K. Scrivener, F. Martirena, S. Bishnoi, S. Maity, Calcined clay limestone cements 420070-8.00002-7%0Adoi:10.1016/j.ab.2015. 03.024%0Adoi:10.1080/
(LC3), Cem. Concr. Res. 114 (2017) 49–56, https://doi.org/10.1016/j. 07352689.2018.1441103%0Ahttp://www.chile.bmw-motorrad.cl/sync/
cemconres.2017.08.017. showroom/lam/es/).
[23] S. Meiboom, D. Gill, Modified spin-echo method for measuring nuclear relaxation [51] E.T. Rodriguez, I.G. Richardson, L. Black, E. Boehm-Courjault, A. Nonat,
times, Rev. Sci. Instrum. 29 (8) (1958) 688–691, https://doi.org/10.1063/ J. Skibsted, Composition, silicate anion structure and morphology of calcium
1.1716296. silicate hydrates (C-S-H) synthesised by silica-lime reaction and by controlled
[24] C. Glawe, F. Georget, M. Raupach, T. Matschei, Multi technique characterization of hydration of tricalcium silicate (C3S), Adv. Appl. Ceram. 114 (7) (2015) 362–371,
the carbonation affected zone including non-destructive single sided 1H NMR, https://doi.org/10.1179/1743676115Y.0000000038.
Cem. Concr. Res. 178 (2024) 107438, https://doi.org/10.1016/j. [52] J.J. Thomas, J.J. Chen, A.J. Allen, H.M. Jennings, Effects of decalcification on the
cemconres.2024.107438. microstructure and surface area of cement and tricalcium silicate pastes, Cem.
[28] A.C.A. Muller, K.L. Scrivener, A.M. Gajewicz, P.J. McDonald, Densification of Concr. Res. 34 (12) (2004) 2297–2307, https://doi.org/10.1016/j.
C–S–H measured by 1H NMR relaxometry, J. Phys. Chem. C 117 (1) (2013) cemconres.2004.04.007.
403–412, https://doi.org/10.1021/jp3102964. [53] Y. Ma, et al., Influences of leaching on the composition, structure and morphology
[29] A.M. Gajewicz-Jaromin, P.J. McDonald, A.C.A. Muller, K.L. Scrivener, Influence of of calcium silicate hydrate (C–S–H) with different Ca/Si ratios, J. Build. Eng. 58
curing temperature on cement paste microstructure measured by 1H NMR (2022) 105017, https://doi.org/10.1016/j.jobe.2022.105017.
relaxometry, Cem. Concr. Res. 122 (2019) 147–156, https://doi.org/10.1016/j. [54] G.L. Kalousek, A.F. Prebus, Crystal chemistry of hydrous calcium silicates: III,
cemconres.2019.05.002. morphology and other properties of tobermorite and related phases, J. Am. Ceram.
[30] R. Kurihara, I. Maruyama, Surface area development of Portland cement paste Soc. 41 (4) (1958) 124–132, https://doi.org/10.1111/j.1151-2916.1958.tb13525.
during hydration: direct comparison with 1H NMR relaxometry and water vapor/ x.
nitrogen sorption, Cem. Concr. Res. 157 (2022) 106805, https://doi.org/10.1016/ [55] K. Sasaki, R. Yokozawa, T. Saeki, The effect of carbonation on chloride ion
j.cemconres.2022.106805. diffusivity in cementitious hardened cement paste (in Japanese), Jpn. Concr. Inst.
[31] J. Il Choi, Y. Lee, Y.Y. Kim, B.Y. Lee, Image-processing technique to detect 29 (1) (2007) 1035–1040.
carbonation regions of concrete sprayed with a phenolphthalein solution, [56] R. von Helmholtz, Investigations of vapors and mists, especially of such things from
Construct. Build Mater. 154 (2017) 451–461, https://doi.org/10.1016/j. solutions, Ann. der Phys. und Chem. 263 (4) (1886) 508–543, https://doi.org/
conbuildmat.2017.07.205. 10.1002/andp.18862630403.
[32] R.G. da Silveira, G.C. Isaia, Proposal of a digital image analysis method for [57] F. D’Orazio, S. Bhattacharja, W.P. Halperin, K. Eguchi, T. Mizusaki, Molecular
determining the carbonated front in concretes, Ambient. Construído 22 (3) (2022) diffusion and nuclear-magnetic-resonance relaxation of water in unsaturated
275–298, https://doi.org/10.1590/s1678-86212022000300619. porous silica glass, Phys. Rev. B 42 (16) (1990) 9810–9818, https://doi.org/
[33] R.V. Silva, A. Silva, R. Neves, J. de Brito, Statistical modeling of carbonation in 10.1103/PhysRevB.42.9810.
concrete incorporating recycled aggregates, J. Mater. Civ. Eng. 28 (1) (2016) 1–7, [58] A.M. Gajewicz, Characterisation of Cement Microstructure and Pore - Water
https://doi.org/10.1061/(asce)mt.1943-5533.0001366. Interaction by H Nuclear Magnetic Resonance Relaxometry, University of Surrey,
[34] B.I.E. Kraft, M. Müller, R. Achenbach, H.M. Ludwig, M. Raupach, B. Kraft, 2014. https://openresearch.surrey.ac.uk/esploro/outputs/99513994102346.
Influence of CO2 concentration during accelerated carbonation tests with [59] I. Maruyama, N. Sakamoto, K. Matsui, G. Igarashi, Microstructural changes in
alternative binders, NanoWorld J. 9 (2023) 195–200, https://doi.org/10.17756/ white Portland cement paste under the first drying process evaluated by WAXS,
nwj.2023-s2-034. SAXS, and USAXS, Cem. Concr. Res. 91 (2017) 24–32, https://doi.org/10.1016/j.
[35] V.G. Papadakis, Effect of supplementary cementing materials on concrete cemconres.2016.10.002.
resistance against carbonation and chloride ingress, Cem. Concr. Res. 30 (2) (2000) [60] S. Gaboreau, et al., Hydration properties and interlayer organization in synthetic C-
291–299, https://doi.org/10.1016/S0008-8846(99)00249-5. S-H, Langmuir 36 (32) (2020) 9449–9464, https://doi.org/10.1021/acs.
[36] T. Bakharev, J.G. Sanjayan, Y.-B. Cheng, Resistance of alkali-activated slag langmuir.0c01335.
concrete to carbonation, Cem. Concr. Res. 31 (9) (2001) 1277–1283, https://doi. [61] I. Maruyama, T. Ohkubo, T. Haji, R. Kurihara, Dynamic microstructural evolution
org/10.1016/S0008-8846(01)00574-9. of hardened cement paste during first drying monitored by 1H NMR relaxometry,
[37] C. Andrade, R. Buják, Effects of some mineral additions to Portland cement on Cem. Concr. Res. 122 (2019) 107–117, https://doi.org/10.1016/j.
reinforcement corrosion, Cem. Concr. Res. 53 (2013) 59–67, https://doi.org/ cemconres.2019.04.017.
10.1016/j.cemconres.2013.06.004. [62] X. Xu, J.T. Han, D.H. Kim, K. Cho, Two modes of transformation of amorphous
[38] B.B. Sabir, S. Wild, J. Bai, Metakaolin and calcined clays as pozzolans for concrete: calcium carbonate films in air, J. Phys. Chem. B 110 (6) (2006) 2764–2770,
a review, Cem. Concr. Compos. 23 (6) (2001) 441–454, https://doi.org/10.1016/ https://doi.org/10.1021/jp055712w.
S0958-9465(00)00092-5. [63] T. Ogino, T. Suzuki, K. Sawada, The rate and mechanism of polymorphic
[39] K. Weise, N. Ukrainczyk, E. Koenders, Pozzolanic reactions of metakaolin with transformation of calcium carbonate in water, J. Cryst. Growth 100 (1–2) (1990)
calcium hydroxide: review on hydrate phase formations and effect of alkali 159–167, https://doi.org/10.1016/0022-0248(90)90618-U.
hydroxides, carbonates and sulfates, Mater. Des. 231 (2023) 112062, https://doi. [64] D. Zhao, J.M. Williams, P. Hou, A.J. Moment, S. Kawashima, Stabilizing
org/10.1016/j.matdes.2023.112062. mechanisms of metastable vaterite in cement systems, Cem. Concr. Res. 178 (2024)
[40] Z. Šauman, Carbonization of porous concrete and its main binding components, 107441, https://doi.org/10.1016/j.cemconres.2024.107441.
Cem. Concr. Res. 1 (6) (1971) 645–662, https://doi.org/10.1016/0008-8846(71) [65] G. Renaudin, J. Russias, F. Leroux, F. Frizon, C. Cau-dit-Coumes, Structural
90019-6. characterization of C–S–H and C–A–S–H samples—part i: long-range order
[41] N. Hyvert, A. Sellier, F. Duprat, P. Rougeau, P. Francisco, Dependency of C–S–H investigated by Rietveld analyses, J. Solid State Chem. 182 (12) (2009)
carbonation rate on CO2 pressure to explain transition from accelerated tests to 3312–3319, https://doi.org/10.1016/j.jssc.2009.09.026.
natural carbonation, Cem. Concr. Res. 40(11) 1582–1589. doi:https://doi.org/10 [66] Y. Yan, et al., Effect of alkali hydroxide on calcium silicate hydrate (C-S-H), Cem.
.1016/j.cemconres.2010.06.010. Concr. Res. 151 (2022) 106636, https://doi.org/10.1016/j.
[42] I. Galan, F.P. Glasser, D. Baza, C. Andrade, Assessment of the protective effect of cemconres.2021.106636.
carbonation on portlandite crystals, Cem. Concr. Res. 74 (2015) 68–77, https:// [67] G. Geng, R.J. Myers, M.J.A. Qomi, P.J.M. Monteiro, Densification of the interlayer
doi.org/10.1016/j.cemconres.2015.04.001. spacing governs the nanomechanical properties of calcium-silicate-hydrate, Sci.
[43] L. Black, C. Breen, J. Yarwood, K. Garbev, P. Stemmermann, B. Gasharova, Rep. 7 (1) (2017) 10986, https://doi.org/10.1038/s41598-017-11146-8.
Structural features of C-S-H(I) and its carbonation in air—a Raman spectroscopic [68] G. W. S. C.J. Brinker, Sol-Gel Science, No. Chapter 6. New York, 1990.

24
L. Cheng et al. Cement and Concrete Research 186 (2024) 107688

[69] S. von Greve-Dierfeld, et al., Understanding the carbonation of concrete with Further Reading
supplementary cementitious materials: a critical review by RILEM TC 281-CCC,
Mater. Struct. 53 (6) (2020) 136, https://doi.org/10.1617/s11527-020-01558-w.
[25] F. Díaz-Díaz, P.F.D.J. Cano-Barrita, F.M. León-Martínez, F. Castellanos, Unilateral
[70] V.T. Ngala, C.L. Page, Effects of carbonation on pore structure and diffusional
low-field magnetic resonance measurements of carbonation depth in unsaturated
properties of hydrated cement pastes, Cem. Concr. Res. 27 (7) (1997) 995–1007.
and saturated Portland cement mortars, Cem. Concr. Res. 138 (2020) 106237,
[71] C.-H. Li, K. Nakarai, Y. Ishii, K. Yokozuka, Experimental study on effect of
https://doi.org/10.1016/j.cemconres.2020.106237.
carbonation of early aged cement paste on micro-pore structure and effective
[26] M. Van Landeghem, J.B. D’Espinose De Lacaillerie, B. Blümich, J.P. Korb,
diffusion coefficient of oxygen, Cem. Sci. Concr. Technol. 63 (1) (2009) 99–106,
B. Bresson, The roles of hydration and evaporation during the drying of a cement
https://doi.org/10.14250/cement.63.99.
paste by localized NMR, Cem. Concr. Res. 48 (2013) 86–96, https://doi.org/
[72] H.Y. Carr, E.M. Purcell, Effects of diffusion on free precession in nuclear magnetic
10.1016/j.cemconres.2013.01.012.
resonance experiments, Phys. Rev. 94 (3) (1954) 630–638, https://doi.org/
[27] T. Colinart, P. Glouannec, Investigation of drying of building materials by single-
10.1103/PhysRev.94.630.
sided NMR, Energy Procedia 78 (2015) 1484–1489, https://doi.org/10.1016/j.
[73] T. Ohkubo, Open source 1D/2D inverse Laplace program, Bull. Nucl. Magn. Reson.
egypro.2015.11.174.
Soc. Jpn. 9 (2018) 92–98, https://doi.org/10.1044/leader.PPL.23122018.24.
[74] T. Ohkubo, Open source 1D/2D inverse Laplace program, Retrieved from, https
://amorphous.tf.chiba-u.jp/memo.files/osilap/osilap.html, 2023.

25