You are on page 1of 38

Accepted Manuscript

Formation of nanocrystalline tobermorite in calcium silicate binders with low C/S ratio

P. Paradiso, R.L. Santos, R.B. Horta, J.N.C. Lopes, P.J. Ferreira, R. Colaço

PII: S1359-6454(18)30279-9
DOI: 10.1016/j.actamat.2018.04.006
Reference: AM 14493

To appear in: Acta Materialia

Received Date: 3 December 2017


Revised Date: 13 March 2018
Accepted Date: 4 April 2018

Please cite this article as: P. Paradiso, R.L. Santos, R.B. Horta, J.N.C. Lopes, P.J. Ferreira, R. Colaço,
Formation of nanocrystalline tobermorite in calcium silicate binders with low C/S ratio, Acta Materialia
(2018), doi: 10.1016/j.actamat.2018.04.006.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
Formation of nanocrystalline tobermorite in calcium silicate
binders with low C/S ratio

P. Paradiso1,2, R.L. Santos3, R. B. Horta1, J. N. C. Lopes1,4, P.J. Ferreira1,2,6, R.


Colaço1,2

1
Instituto Superior Técnico, University of Lisbon, Av. Rovisco Pais, 1049-001 Lisboa, Portugal
2

PT
IDMEC - Instituto de Engenharia Mecânica, University of Lisbon, Av. Rovisco Pais, 1049-001
Lisboa, Portugal
3
CIMPOR - Cimentos de Portugal, SGPS S.A., Rua Alexandre Herculano, 35, 1250-009 Lisboa,
Portugal.
4

RI
Centro de Química Estrutural, Av. Rovisco Pais, 1049-001 Lisboa, Portugal.
6
International Iberian Nanotechnology Laboratory, Av. da. Mestre José Veiga s/n, 4715 Braga,
Portugal

SC
Abstract

U
Pastes of calcium silicate binders with low C/S ratio with an overall CaO/SiO2
AN
(C/S) molar ratio of 1.1 were produced by mixing these materials with water or
with a aqueous solution of Na2SiO3. Both types of pastes (non-activated and
M

Na2SiO3 activated) were analysed using X-ray diffraction (XRD) and high-
resolution transmission electron microscopy (HR-TEM). The observations show
D

that the hydration process of this novel type of cements results in the formation
of tobermorite nanocrystals embedded in the amorphous calcium-silicate-
TE

hydrate (C-S-H) matrix, both in the case of water and of water+Na2SiO3


hydration. Selected area electron diffraction (SAED) patterns obtained in the
EP

HR-TEM observations enabled a full characterization of the tobermorite


crystalline structure (type, unit cell, space group, lattice angles and parameters).
C

Molecular dynamics (MD) simulations of the identified tobermorite structure


were performed at various temperatures, showing its viability and thereby
AC

supporting the experimental results. To the best of the authors’ knowledge, this
report is the first to describe the spontaneous formation of crystalline
tobermorite in the form of nanodomains dispersed in the C-S-H phase upon the
hydration of calcium-silicate binders.

Keywords: cement, hydraulic binders, calcium-silicate-hydrates, tobermorite.

1
ACCEPTED MANUSCRIPT
1 Introduction

Concrete is the most widely used material in the world, and cement is its major

chemically active component [1]. The success of cement as a structural

engineering material is observed because cement is a cost-effective, easy-to-

use, strong and durable material. Therefore, since the work of Louis Vicat [2]

PT
two centuries ago, the production of cement has been continuously increasing,

RI
and currently counts over 4 billion tons per year [3]. Nevertheless, with the

increasing use of cement, the concern of the environmental impact of its

SC
production also grows: presently, the cement industry is responsible for

approximately 5% to 7% of total anthropogenic CO2 emissions, since for every

U
ton of cement produced, approximately one ton of CO2 is emitted to the
AN
atmosphere [4]. Consequently, the reduction of the environmental impact of
M

cement production has become an increasing source of concern, further

emphasised by the signing of the Paris Agreement in 2016 [5].


D

This need to reduce carbon dioxide emissions in the cement industry has led to
TE

a remarkable process optimization, as well as to the development of several

eco-friendly solutions, such as energy recovery from waste valorisation and the
EP

incorporation of by-products generated by other industries. Several other


C

approaches focus on the reduction of CO2 directly by changing the chemistry of


AC

the clinker in such a way that lower amounts of carbonated raw-materials

(namely, limestone) are needed for its production. Examples are the

development of belite-rich clinkers or sulfoaluminate clinkers [6, 7]. Changes of

the cement mix, by decreasing the amount of alite-based clinker used and

replacing it by new or improved supplementary cementitious Examples are the

development of belite-rich clinkers or sulfoaluminate clinkers [6, 7]. Changes

2
ACCEPTED MANUSCRIPT
romising approaches that are currently being explored. Other approaches

include the development of a new type of cement based on calcium-

hydrosilicates or even the development of non-hydraulic systems such as

geopolymer-based binders (e.g., [10]) or CO2-curing cements ([11] and [12]).

All of these solutions present shortcomings, which is the reason why they have

PT
not been addressed yet by the cement industry. The solutions focusing on the

RI
substitution of Portland clinker by other SCMs have as major drawbacks the fact

that these materials (e.g., calcinated clays, fly ashes, slag, and recycled

SC
concrete) are not extensively available in a suitable quality for the production

U
needs of the cement industry. Meanwhile the solutions involving the
AN
development of new alternative binders present several drawbacks, namely, i)

the need of specific raw-materials which, due to their lack of abundance, are
M

much more expensive than traditional limestone; ii) their technical practicality

due to the very specific and necessary conditions of hydration, which require a
D

controlled environment and in some cases can only be applied in pre-


TE

fabrication; iii) the need of a significant technological change of the cement

production paradigm with large, even unaffordable, financial impacts for cement
EP

companies, and iv) the time needed for the acceptance of these binders as
C

construction materials, both by regulatory organisms and end users. Due to


AC

these issues, the search for alternative solutions to common ordinary Portland

cement (OPC), has not concluded to date.

In a paper published in 2014, Pellenq and colleagues formulate the theoretical

hypothesis that the C-S-H obtained by hydration of a clinker with a lower

amount of carbonated raw-materials (C/S ratio in the range of 1 to 1.1) should

present a well-defined tobermorite-like crystalline layered structure with minor

3
ACCEPTED MANUSCRIPT
structural defects [13]. This organized layered structure of tobermorite would be

progressively lost as the C/S ratio increases until it becomes amorphous for C/S

ratios larger than 1.5 [13]. The molecular dynamics (MD) simulations of Pellenq

and colleagues also showed that this low C/S tobermorite-like structure should

correspond to the best mechanical properties of C-S-H, specifically of its

PT
hardness and stiffness. Unfortunately, the C-S-H that forms upon hydration of

RI
ordinary Portland cement (the C/S ratio of which is typically above 3) presents a

C/S ratio of approximately 1.7; thus, its mechanical properties are expected to

SC
be lower than those of the layered C-S-H with C/S<1.5. Moreover, upon

hydration of the conventional alite-based cements, the precipitation of a

U
significant amount of portlandite-type calcium hydroxide (Ca(OH)2) occurs,
AN
which represents a misuse of the excess of calcium present in the OPC. In the
M

last several decades, since the first evidence of a similarity between the

structures of C-S-H and tobermorite [14], many computational models and


D

experimental approaches on the evaluation of C-S-H properties have been


TE

attempted [15]. However, the majority of the experimental studies conducted on

the analysis of low C/S C-S-H and its related naturally occurring mineral,
EP

tobermorite, used materials obtained by synthesis techniques, such as the

reaction of C2S in a solution containing a SiO2 source [16, 17], the reaction of
C

quartz or silicic acid with pure CaO or C2S at temperatures between 333 K [18,
AC

19] and 503 K [20], the reaction of silica-gel with Ca(OH)2 [21] or even by

decalcification of the hydration product obtained from C3S hydration [22].

To date, the reason that it is impossible to directly obtain C-S-H with a C/S of

approximately 1 by hydration of its corresponding calcium silicate mineral,

wollastonite (CaSiO3), is because wollastonite is not hydraulically active. Thus,

4
ACCEPTED MANUSCRIPT
the confirmation of the structural and mechanical properties of the low C/S C-S-

H obtained directly by hydration of a single calcium silicate phase, has been a

difficult issue to overcome. Therefore, the MD results, describing a low calcium

C-S-H with a tobermorite-like crystalline structure formed upon hydration [13],

have not been demonstrated to date.

PT
Recently, Santos et al. [23], proposed a method for the production of a new type

RI
of hydraulically active calcium-silicate (CS) binder, which contains 33% less

CaO in the raw mix than the typical ordinary Portland cement formulations. This

SC
binder is hydraulically active in the CaO/SiO2 range of 1.1 to 1.25. Upon

U
hydration, XRD analysis shows that a semi-crystalline form of C-S-H appears,
AN
whereas no portlandite could be detected [23]. In this paper, we focus on

studying the microstructure of these binders by high-resolution transmission


M

electron microscopy (HR-TEM). The observations show that the semi-crystalline

C-S-H formed upon hydration presents embedded nanodomains that were


D

identified as highly ordered tobermorite crystallites with diameters of


TE

approximately 10 to 20 nm. Surprisingly, the HR-TEM observations and the

corresponding selected area electron diffraction (SAED) patterns show that the
EP

tobermorite allotropic form that appears upon hydration of this C/S=1.1 binder is
C

the 9 Å tobermorite. The structure of this phase is fully characterized by the


AC

analysis and the viability of this phase is analysed by molecular dynamics

simulations. To the best of the authors’ knowledge, this report present the first

detailed characterization of tobermorite nanocrystals obtained directly upon

hydration of a hydraulic binder.

5
ACCEPTED MANUSCRIPT
2 Experimental Details

2.1 Materials Synthesis

Materials were produced by fully melting at 1773 K a mixture of typical raw

materials (limestone, sand, fly ash and electric furnace slag) with an overall C/S

PT
molar ratio of 1.1. After melting, a rapid cooling of the mixture was performed.

Once the material solidified in a glassy state, the clinker was ground in a ring

RI
mill for 180 s with propanol and was dried at 323 K in a stove for approximately

1 h. This powder was then used to produce two types of pastes: (1) a mixture

SC
with the addition of water and (2) a mixture with the addition of an alkaline

U
activator (Na2SiO3) in such a way that the water/binder weight ratio was kept
AN
constant at the value of 0.375. In the case of the alkali-activated samples, the

amount of activator added was defined to obtain a total Na2O content of 3% by


M

weight and a Si/Na molar ratio of 1.2 in the activator. The produced pastes were

poured into prismatic moulds with dimensions of 20x20x40 mm and cured in a


D

95% humidity controlled environment at 293 K. A more detailed description of


TE

the synthesis protocol can be observed in [23].


EP

2.2 Microstructural Characterization

The microstructure of the pastes was characterized using X-ray diffraction


C

(XRD) with Rietveld analysis and using HR-TEM. The XRD analysis was
AC

performed in a X’Pert Pro (PANalytical) diffractometer using monochromatic

CuKα1 radiation (λ=1.54059 Å) and working in a reflection geometry (θ/2θ). The

optics configuration was a fixed divergence slit (1/2°), a fixed incident antiscatter

slit (1°), a fixed diffracted anti-scatter slit (1/2°) and a X'Celerator RTMS (Real

Time Multiple Strip) detector working in the scanning mode with the maximum

active length. Data for each sample were collected from 5° to 60° (2θ). The

6
ACCEPTED MANUSCRIPT
samples were rotated during data collection at 16 rpm to enhance the particle

statistics. The X-ray tube worked at 45 kV and 40 mA. The amorphous content

of the samples, both anhydrous and hydrated, was determined by adding 20%

by weight of corundum (99,9% α-Al2O3 from Alfa Aesar) as an internal standard.

The phase determination and quantification was performed through Rietveld

PT
analysis using the Panalytical software Highscore Plus following the same

RI
procedure previously adopted by the authors [23].

All of the samples observed using HR-TEM have hydration ages superior to 90

SC
days. The samples were first pulverized using an agate mortar; next, the

U
particles were dispersed in pure ethanol and ultrasonically agitated at low power
AN
for ten minutes; finally, a drop of the suspension was transferred to lacey

carbon TEM grids. HR-TEM images of the samples were taken on a JEM-2100
M

JEOL TEM with an accelerating voltage of 200 kV using a double-tilt holder.

Selected area electron diffraction (SAED) patterns were acquired for


D

crystallographic diffraction studies. HR-TEM images were also used to obtain


TE

crystallographic information by analysing the fast Fourier transform (FFT) of the

images, using the digital micrograph software.


EP

2.3 Computational Simulations


C

VESTA software was used to model the crystalline structure, while the CaRIne
AC

Crystallography 3.1 software was used to simulate the diffraction patterns and

stereographic projections. The diffraction patterns of the structure obtained with

VESTA enabled the identification of the zone axis and interplanar spacings.

Molecular Dynamics (MD) simulations of C/S ratio of 1.1 were run employing

the appropriate Lennard-Jones interatomic parameters for the Si, O and Ca

7
ACCEPTED MANUSCRIPT
atoms. These parameters were taken from Freitas et al. [24] and fit the

Coulomb plus Lennard-Jones potential functional (1):

= +4 − (1)

where and are the partial charges of atoms of type i and j respectively,

PT
is the permittivity of vacuum, is the interatomic distance, and and are

RI
the Lennard-Jones pair interaction parameters. More details on this procedure

can be observed in a previous publication [24]. The Coulomb and Lennard

SC
Jones parameters are given in Table 1.

U
The Freitas et al. force field (FFF), used in the present work, is a simple,
AN
versatile and non-reactive potential that has been previously applied to the

study of amorphous and crystalline phases in the silica-lime-alumina ternary


M

diagram, including the estimation of the corresponding elastic constant and

moduli [24]. Comparisons between the FFF and others widely used force fields,
D

namely ClayFF, are given in [24]. Since the extension of the scope of the FFF to
TE

encompass tobermorite-like phases is warranted by its general character,


EP

previously demonstrated [24], in this work, the FFF has been extended to the

study of tobermorite by adding three-body potential functions, which were used


C

to model the presence of hydroxyl groups in the crystalline tobermorite


AC

structures. Related additional parameters, taken from Hassanali et al. [25], are

also given in Table 1. The different functional forms are also given in the same

reference (Equations. 1-4 of [25]).

All molecular dynamics (MD) simulations were performed with DL_POLY 2.20

under anisotropic isothermal-isobaric ensemble conditions. The initial

configuration of the MD simulation of the amorphous C/S 1.1 was generated as

8
ACCEPTED MANUSCRIPT
described in detail by Freitas et al. [24], by quenching a 5-nm-size cubic box to

300 K from 4000 K/2500 K. In the case of the tobermorite 9 Å crystal, the initial

simulation box was generated from information obtained from HR-TEM

observations, by replicating the crystalline unit cell in different directions to

obtain a triclinic simulation box with sides of at least 5.0 nm. The obtained box

PT
was then equilibrated at 300 K, 500 K and 800 K to study the stability of the

RI
crystalline structure. All following simulation runs are characterized by periodic

boundary conditions in all directions, cut-off distances of 3.45 nm and a time-

SC
step of 1 fs. All glass and crystal MD simulations were pre-equilibrated under

NVT (i.e. constant number of particles, volume and temperature) conditions; the

U
subsequent trajectories were run under NPT (i.e. constant number of particles,
AN
pressure and temperature) and NσT (i.e. constant number of particles, tensor
M

and temperature) conditions, respectively, and sampled every 100 fs.

2.4 Low-
Low-temperature differential scanning calorimetry (LT-
(LT-DSC)
D

Differential scanning calorimetry measurements were performed upon fully


TE

hydrated non-activated sample (more than one year of hydration) under


EP

nitrogen atmosphere. The samples, previously pulverized using an agate

mortar, were weighted and transferred in a sealed aluminium pan (diameter 5


C

mm, capacity 25 µL, from NETZSCH). A DSC 200F3 Maia model from
AC

NETZSCH was used and the data were elaborated with the Proteus software.

Each measurement was carried out with the following temperature program:

equilibrate at 20 °C for 10 minutes; cooling ramp from 20 to -80 °C at 10 °C/min;

equilibrate at -80 °C for 10 minutes and heating ramp from -80 °C to +120 at 10

°C/min. Three independent LT-DSC tests have been performed and the

experimental error was estimated to be lower than 5%.

9
ACCEPTED MANUSCRIPT
3 Results and Discussion

3.1 Microstructural Characterization

Figure 1 shows the XRD patterns of the anhydrous samples and those of the

pastes at the ages of 7 days and 90 days for the samples that were hydrated

PT
with water and with water + Na2SiO3. The initial anhydrous binder structure is

formed by an amorphous matrix with approximately 13% of crystalline

RI
pseudowollastonite, which resulted from the quenching process used to

SC
produce the binder as described in the experimental section. After hydration,

the total amount of the amorphous phase decreases with time, as well as the

U
proportion of pseudo-wollastonite, and 9 Å crystalline tobermorite peaks appear
AN
both in the activated and non-activated pastes after the 7th day of hydration

(insets of Figure 1.B and 1.C).


M

The summary of the evolution of the phases during hydration, as measured by


D

XRD-Rietveld analysis, is presented in Table 2. In the case of the non-activated


TE

paste, the amount of 9 Å crystalline tobermorite increases from 11.1% after 7

days of hydration up to 20.5% after 90 days of hydration. In the case of the


EP

alkaline activated paste, the amount of tobermorite further increases to 23.5%

after 7 days of hydration and to 28.3% after 90 days. This may be explained by
C

the fact that the presence of the alkali activator increases the pH of the paste
AC

and, thus, the presence of OH- groups in solution, which accelerates the

formation of CSH and of tobermorite [26].

To estimate the dimensions of the tobermorite crystals formed upon hydration,

we applied the Scherrer equation (2):


t= (2)
B cos θ B
10
ACCEPTED MANUSCRIPT

where t is the size of the crystalline domains, K is a dimensionless shape factor

(assumed to be the generally accepted default value of 0.9), λ is the X-ray

wavelength, B is the line broadening at half the maximum intensity, after

subtracting the instrumental line broadening, and θ is the Bragg angle (in

PT
degrees). When applied to the tobermorite peaks, the Scherrer equation

RI
produces the results shown in Table 3. From these data, it is clear that

tobermorite is formed after hydration in the form of 7 to 20 nm nanocrystallites.

SC
These nanocrystals tend to be slightly smaller in the case of the alkaline

U
activated samples than in the case of the non-activated ones.
AN
The nanostructural nature of C-S-H with C/S ratios in the range of 1.1 to 1.8 has

been discussed by Pellenq et al. [13] based on molecular dynamic (MD)


M

simulations. These researchers concluded that with the increase of the C/S
D

ratio, there is an increase in the disorder of the calcium-silicate hydrate


TE

structure, and the lamellae-like structure, typical for the tobermorite and present

until approximately C/S=1.5, tends to disappear, which agrees with the present
EP

experimental observations. Four types of tobermorites have been heavily

researched to date: 9 Å, 11 Å, 14 Å and clinotobermorite [27-31]. From these


C

types of tobermorite, the 11 Å type has been extensively used as model phase
AC

for the amorphous C-S-H [32-34]. However, as far as the authors know, its

direct formation of this or of any other type of tobermorite upon hydration has

never been reported. The C-S-H models, such as the one proposed by

Jennings et al. [35] and recently revisited by Gartner [36], are based on the

existence of <5 nm diameter building blocks, which match the tobermorite 11 Å

structure and would explain the C-S-H bulk properties found in hydrated cement

11
ACCEPTED MANUSCRIPT
pastes. Therefore, it is somewhat surprising that, in the present work, the best

fitting of the tobermorite peak at 30°, obtained by adjusting the scale factor,

lattice parameters (h, k, l, α, β and γ) as well as the Caglioti function W

parameter, is obtained for 9 Å tobermorite (insets in Figure 1). Moreover, at the

XRD level of resolution, no significant differences between the tobermorite peak

PT
of the non-activated and the alkali activated samples can be observed. Neither

RI
tobermorite 11 Å nor tobermorite 14 Å present adequate profiles for fitting the

acquired XRD data. Nevertheless, HR-TEM analysis was performed for a full

SC
characterization of the 9 Å tobermorite nanocrystals that were detected by XRD.

U
Figure 2 shows bright field HR-TEM images of the non-activated pastes (Figure
AN
2.A) and alkaline activated pastes (Figure 2.B). In both cases, the presence of

the tobermorite nanocrystallites (see as an example the nanocrystal highlighted


M

in the yellow area in Figure 2.A and Figure 2.B) that are randomly oriented and

dispersed in the amorphous matrix, resulting in well-defined diffraction rings,


D

which are observed in SAED (inset on the upper right corner), can be observed.
TE

In the two types of samples, it can be observed that the tobermorite


EP

nanocrystals have dimensions between 10 and 20 nm (in agreement with the

Scherrer equation’s results, provided in Table 3). Furthermore, comparing the


C

diffraction rings presented by the two types of hydrated samples (Figure 2.C),
AC

no significant difference in the rings distribution and dimensions were observed,

which confirms that the alkali activation does not affect the phase formation

during the hydration, as the XRD analysis results previously showed.

Figure 3.A shows the bright field image of the alkali activated sample, whereas

Figure 3.B shows the corresponding experimental SAED pattern (B). In Figure

3.B, the red circles represent the simulated diffraction pattern, which is

12
ACCEPTED MANUSCRIPT
superimposed on the experimental SAED pattern. The simulated diffraction

pattern was obtained through Vesta, which produced the best match for a

tobermorite crystallite along the [123] zone axis. The matching stereographic

projection was observed after successive adjustments of the 9 Å triclinic

tobermorite described by Merlino et al. [31] as space group P-1; cell parameters

PT
a=6.93 Å, b=6.91 Å c=9.52 Å, and lattice angles α=100.63o, β=96.13° and

RI
γ=114.76°. A validation of the structure was obtained by performing an inverse

Fourier transform of the inset in Figure 4.A. As a result, it was possible to

SC
identify the (0 1 3), (2 0 1) and (2 1 2) planes of tobermorite 9 Å and to obtain

their interplanar distance (Figure 4.B).

U
AN
The highly ordered structure of tobermorite observed in the present work using

HR-TEM agrees with the predictions made by Pellenq and colleagues [13],
M

strongly indicating that as the C/S ratio decreases, the calcium silicate layers

tend to become more ordered and less defective, i.e., with the formation of
D

crystalline tobermoritic regions in the C-S-H. These authors predicted that for
TE

C/S<1.5, the C-S-H that is formed should present a tobermoritic layered

structure, which has been previously observed by TEM in a C-S-H prepared


EP

using the reaction of a lime solution with amorphous silica with a C/S ratio of 0.9
C

[37]. In the simulations made by Pellenq et al., the layered structure of low
AC

calcium C-S-H corresponds to 11 Å tobermorite, with water molecules in the

interlayer positions, as described by Hamid [30] and Merlino [27]. However, in

the present work, the formation upon hydration of nanocrystals of tobermorite in

the 9 Å allotropic form with no bonded water molecules in its structure in the C-

S-H (C/S=1.1) is demonstrated.

13
ACCEPTED MANUSCRIPT
According to the Jennings colloidal model-II [38] the layered C-S-H
nanostructures are organized in packed clusters presenting zeolite-like cavities
which accommodate physiosorbed water molecules. The presence of these
type of water was evidenced by the LT-DSC analysis shown in

Figure 1. The cooling curve (blue curve

Figure 1) obtained by the cooling of the sample to -80 ºC, presents a

PT
characteristic peak centred at -45 ºC, which corresponds to the water frozen in

RI
small gel pores (SGP) [38], with entryways 1-3 nm. No other exothermic peak

was evidenced during the cooling, which demonstrates that no large or

SC
intermediate gel pores, and no free water, are present in the fully hydrated aged

U
sample, as we would expect. No other endothermic peaks were evidenced
AN
during the heating scan. These results are consistent with previous studies of

Bentz [39] and Baglioni et al. [40].


M

3.2 Computational Simulations


Simulations
D

The crystalline structure of the 9 Å tobermorite that was identified by HR-TEM


TE

SAED was computationally reconstructed using the VESTA software and is

presented in Figure 6. The solid skeleton of the 9 Å tobermorite, which


EP

characterizes the whole family of tobermorites [27-31], consists of three parts:

silica chains, intralayer calcium and interlayer calcium, as can be observed


C

along the (100) and (010) projections of Figure 5.A and Figure 5.B. In Figure
AC

5.C, single Si-O chains oriented along the [110] direction can be seen. The

same structure of silicate chains was described in the Wollastonite model

studied using an MD simulation in a previous work [24]. These Q2 chains are

connected to each other through the interlayer calcium ions and form a layer in

the (001) plane. Along the [001] direction these layers of Q2 single chains are

connected through the intralayer calcium ions, forming a double layered

14
ACCEPTED MANUSCRIPT
structure whose basal distance is shorter than that of two Q2 layers in

wollastonite [24]. In tobermorite 9 Å, eight of the ten water molecules present in

the unit cell of tobermorite 11 Å are lost; the two remaining water molecules are

present as hydroxyl groups SiOH at the positions (0.34 0.14 0.39) and (0.66

0.86 0.61).

PT
Simulations equilibrating the tobermorite 9 Å box at 300 K, 500 K and at 800 K

RI
were performed. The tobermorite’s coefficient of thermal expansion that was

obtained is 3.5×10-5 K-1, which is close to the coefficient of thermal expansion

SC
previously calculated for C1.7SH1.9 using microthermoporomechanics

backanalysis (4.2 × 10−5 K−1) [41] and MD (4.5 × 10−5 K−1) [41]. The smaller

U
AN
value obtained for the tobermorite 9 Å may be attributed to the fact that the

C1.7SH1.9 has a higher C/S ratio and consequently has a higher content of the
M

chemically bound water, which leads to a greater thermal expansion [43]. The

MD theoretical results and the experimental results, concerning the volume of


D

the unit cell of tobermorite, agree. The deviation of the volumes, which is
TE

smaller than 2%, may indicate some heating of the sample under the electron

beam (e.g [44]), or may be caused by theoretical and experimental limitations.


EP

The oxygen and silicon connectivities of the tobermorite structure were


C

analysed using MD simulation analysis and are presented in Table 4. It was


AC

observed that the tobermorite structure is essentially formed by Q2 bonded Si,

being only 8.6% of Si atoms present as terminal groups in the Si chains.

Regarding oxygen, 62.5% is single bonded oxygen, 35.9% corresponds to

bridging oxygen and 1.6% is the oxygen present in the hydroxyl groups. For

comparison purposes, we present in the same table the Qn distribution of the

amorphous C/S 1.1, of the crystalline wollastonite (C/S=1) that was obtained

15
ACCEPTED MANUSCRIPT
using MD simulations [24] and the silicon Qn of the amorphous C/S 1.1 [23] and
29
of the respective C-S-H [45] measured by deconvolution of the Si MAS NMR

spectra. The comparison of the connectivity of these phases shows that there

are almost no differences in the type of bonded oxygens between 9Å

tobermorite, amorphous C/S 1.1 and crystalline wollastonite (the main

PT
difference being the absence of hydroxyl groups in the case of the crystalline

RI
wollastonite), the same cannot be said about the Si connectivity.

In fact, while the MD simulation of amorphous C/S 1.1 gives a Qn distribution

SC
around the Q2 (40%), but with the presence of Q1(27%) and Q3 (22%), in

U
agreement with 29Si NMR analysis [23], in the tobermorite 9 Å, Si atoms are

arranged by 91% in long parallel dreirerketten-type chains of {[SiO3]2-}n units,


AN
presenting a structure similar to that of the hydraulically non-reactive
M

wollastonite polymorph, which is formed by 100% of a stable Q2 silicate

structure, as described in a previous study [24]. The NMR analysis of the C-S-H
D

showed a predominance of Q2, which somehow is in accordance with the


TE

structure of tobermorite 9Å [45]. All these results are in agreement with

previously published studies showing that the tobermorite crystals (the Hamid
EP

tobermorite 11Å, tobermorite 14Å and tobermorite 9Å) are essentially formed by
C

separated Si-O chains, giving a predominant Q2 signal [46, 47], while in the
AC

Merlino tobermorite 11A, 2/3 of the connectivities are of the Q2 type [48].

In the case of glassy C/S 1.1, the amorphization process causes a large

distribution of Qn [23], as shown in Table 4. The clinker studied in the present

work is the product of a randomization resulting from the fact that a portion of

the single bonded O atoms either become detached from Si and become

isolated O atoms or form a bond to a second Si atom and become bridging

16
ACCEPTED MANUSCRIPT
oxygen atoms. These shifts of the O atoms, despite corresponding to only 1% in

the amorphization of the C/S 1.1, result in a considerable change of the

connectivity of the Si in the amorphous phase. The amorphization of the

material under study induces the ramification of Q2 chains, resulting in the

formation of Q4 and Q3, as well as breaking of Q2 chains, resulting in Q1 and Q0

PT
Si bonds. Figure 7, shows the MD boxes of the amorphous C/S 1.1, where the

different Qn are respectively highlighted and represented as red dots. The

RI
distribution of the Qn is observed as being homogeneous, suggesting that the

SC
Q0 represent favourable sites for the water molecules to enter the structure, and

consequently allow the hydration of the clinker.

U
AN
In this way, the present results strongly indicate that, in the presence of water,

this amorphous C/S=1.1 precursor with a wide range of Qn distributions,


M

rearranges to form the more stable Q2 type structure similar to that of the

equilibrium phase (wollastonite). This rearrangement results in the formation of


D

C-S-H with embedded 9Å tobermorite nanocrystals, with a wollastonite-type


TE

structure, i.e., a double layer formed of Q2 long silicate chains that are gathered

together by the iono-covalent attraction of calcium ions. Moreover, recent


EP

experimental results [45] show that, at least until 90 days of hydration, the
C

proportion of tobermorite increasingly grows, which cannot be currently


AC

confirmed by MD simulation experiments.

At this point, it is worth mentioning the existing models for the C-S-H that results

from the hydration of ordinary Portland cement (OPC), i.e., from the reaction of

C3S (alite) and C2S (belite) with water. Although these models cannot be

directly transferred to the present case, in which a C1.1S amorphous phase is

hydrated, they certainly must be compared to the results obtained in this study.

17
ACCEPTED MANUSCRIPT
Although the mechanisms of formation and the nano-to-micro structure of C-S-

H have been extensively studied and modelled [46, 49, 50], C-S-H remains a

poorly understood phase. The main cause for this fact is that C-S-H formed in

common OPC is a fully amorphous phase, and thus is difficult to characterize

using diffraction techniques. The O and H atoms, which play a key role, are

PT
difficult to quantify due to their light weight. Currently the most widely accepted

RI
model for the structure of C-S-H is the one proposed by Jennings [49].

However, this model presents some inconsistencies with experimental data,

SC
namely, in what concerns the kinetics of growth of C-S-H during hydration,

which should be the rate limiting step in the hydration of C3S [36]. Recently,

U
Gartner, Maruyama and Chen [3] presented a new model for the formation and
AN
structure of high C/S (1.7) C-S-H. This last model relies on two main
M

hypotheses, namely: (1) C-S-H grows with basal sheets of tobermorite-like

structures; (2) the basal tobermorite sheets of this C-S-H are essentially
D

anhydrous. Up to now, the C-S-H structure and reaction mechanisms have


TE

continually escaped detailed and direct experimental analysis for two main

reasons: i) difficulty in separating it from other phases and ii) its broad diffraction
EP

signal caused by its amorphous nature, which is why, although the neutron

scattering, electron microscopy, and computer simulation studies have


C

suggested the existence of such a nano building block, a number of doubts


AC

remain [48]. The present results, despite the fact that the hydration occurs in a

much lower C/S ratio structure than that of C3S, seem to be in accordance with

both the predictions of the MD simulations of Pellenq et al. [13] and with the

empirical models for the formation of high C/S C-S-H [3], thereby contributing to

the clarification of the setting mechanisms of hydraulic binders.

18
ACCEPTED MANUSCRIPT

4. Conclusions

In the present study, a detailed analysis of the hydration product on a new

calcium-silicate amorphous hydraulic binder with a C/S ratio of 1.1 is performed.

PT
After hydration, both Rietveld XRD and HR-TEM results show the presence of

amorphous C-S-H and nanocrystalline tobermorite. The HR-TEM observations

RI
show that the nanocrystals are of tobermorite 9 Å with dimensions (10-20 nm).

SC
Molecular dynamic simulations enabled a full characterization and validation of

the crystalline structure of the crystalline phase. To the best of the authors’

U
knowledge, this report is the first to describe the spontaneous formation of
AN
tobermorite in early-age C-S-H cement pastes.
M

Acknowledgements

The authors would like to acknowledge CIMPOR - Cimentos de Portugal, SGPS


D

S.A. for providing financial support (Industrial Contract CIMPOR/ADIST


TE

Nanocement 2016/2020).
EP

References
[1] K. Scrivener, V. John, E. Gartner, Eco-efficient cements: Potential
C

economically viable solutions for a low-CO2 cement-based materials


industry, 2016.
AC

[2] L.J. Vicat, Recherches expérimentales sur les chaux de construction, les
bétons et les mortiers ordinaires, chez Goujon, Paris, 1818.
[3] E. Gartner, I. Maruyama, J. Chen, A new model for the C-S-H phase formed
during the hydration of Portland cements, Cement and Concrete Research
97(Supplement C) (2017) 95-106.
[4] A. Hasanbeigi, L. Price, E. Lin, Emerging energy-efficiency and CO2
emission-reduction technologies for cement and concrete production: A
technical review, Renewable and Sustainable Energy Reviews 16(8)
(2012) 6220-6238.

19
ACCEPTED MANUSCRIPT
[5] A. Ghezloun, A. Saidane, H. Merabet, The COP 22 New commitments in
support of the Paris Agreement, Energy Procedia 119(Supplement C)
(2017) 10-16.
[6] A.K. Chatterjee, High belite cements—Present status and future
technological options: Part I, Cement and Concrete Research 26(8) (1996)
1213-1225.
[7] E. Gartner, G. Li, High belite-containing sulfoaluminous clinker, method for

PT
the production and the use thereof for preparing hydraulic binders, Google
Patents, 2006.
[8] M.S. Imbabi, C. Carrigan, S. McKenna, Trends and developments in green

RI
cement and concrete technology, International Journal of Sustainable Built
Environment 1(2) (2012) 194-216.

SC
[9] S. Sánchez Berriel, A. Favier, E. Rosa Domínguez, I.R. Sánchez Machado,
U. Heierli, K. Scrivener, F. Martirena Hernández, G. Habert, Assessing the
environmental and economic potential of Limestone Calcined Clay

U
Cement in Cuba, Journal of Cleaner Production 124(Supplement C)
(2016) 361-369.
AN
[10] P. Duxson, J.L. Provis, Designing Precursors for Geopolymer Cements,
Journal of the American Ceramic Society 91(12) (2008) 3864-3869.
M

[11] R.E. Riman, S. Gupta, V. Atakan, Q. Li, Bonding element, bonding matrix
and composite material having the bonding element, and method of
manufacturing thereof, Patents, 2013.
D

[12] R.E. Riman, T.E. NYE, V. Atakan, C. VAKIFAHMETOGLU, Q. Li, L. Tang,


Synthetic formulations and methods of manufacturing and using thereof,
TE

2012.
[13] M.J. Abdolhosseini Qomi, K.J. Krakowiak, M. Bauchy, K.L. Stewart, R.
EP

Shahsavari, D. Jagannathan, D.B. Brommer, A. Baronnet, M.J. Buehler, S.


Yip, F.J. Ulm, K.J. Van Vliet, R.J.M. Pellenq, Combinatorial molecular
optimization of cement hydrates, Nat Commun 5 (2014).
C

[14] R.D. Komarneni S, Tobermorites: a new family of cation exchangers,


Science 221(4611) (1983) 647-648.
AC

[15] S. Papatzani, K. Paine, J. Calabria-Holley, A comprehensive review of the


models on the nanostructure of calcium silicate hydrates, Construction and
Building Materials 74 (2015) 219-234.
[16] X. Cong, R.J. Kirkpatrick, 29Si MAS NMR study of the structure of calcium
silicate hydrate, Advanced Cement Based Materials 3(3–4) (1996) 144-
156.
[17] R.J. Kirkpatrick, J.L. Yarger, P.F. McMillan, Y. Ping, X. Cong, Raman
spectroscopy of C-S-H, tobermorite, and jennite, Advanced Cement Based
Materials 5(3) (1997) 93-99.

20
ACCEPTED MANUSCRIPT
[18] A. Gmira, R. Pellenq, I. Rannou, L. Duclaux, C. Clinard, T. Cacciaguerra,
N. Lequeux, H. Van Damme, A Structural Study of
Dehydration/Rehydration of Tobermorite, a Model Cement Compound,
STUDIES IN SURFACE SCIENCE AND CATALYSIS, KK Unger,
Marseilles, France, 2002, pp. 601-608.
[19] A. Gmira, M. Zabat, R.J.-M. Pellenq, H. Van Damme, Microscopic physical
basis of the poromechanical behavior of cement-based materials,
Materials and Structures 37(1) (2003) 3-14.

PT
[20] X. Cong, R.J. Kirkpatrick, X. Cong, R.J. Kirkpatrick, 29Si and 17O NMR
investigation of the structure of some crystalline calcium silicate hydrates,
Advanced Cement Based Materials 3(3) (1996) 133-143.

RI
[21] W. Kunther, S. Ferreiro, J. Skibsted, Influence of the Ca/Si ratio on the
compressive strength of cementitious calcium-silicate-hydrate binders,

SC
Journal of Materials Chemistry A 5(33) (2017) 17401-17412.
[22] J.J. Chen, J.J. Thomas, H.F.W. Taylor, H.M. Jennings, Solubility and

U
structure of calcium silicate hydrate, Cement and Concrete Research
34(9) (2004) 1499-1519.
AN
[23] R.L. Santos, R.B. Horta, J. Pereira, T.G. Nunes, P. Rocha, J.N.C. Lopes,
R. Colaço, Novel high-resistance clinkers with 1.10 < CaO/SiO2 < 1.25:
production route and preliminary hydration characterization, Cement and
M

Concrete Research 85 (2016) 39-47.


[24] A.A. Freitas, R.L. Santos, R. Colaco, R. Bayao Horta, J.N. Canongia
D

Lopes, From Lime to Silica and Alumina: Systematic Modeling of Cement


Clinkers using a General Force-Field, Physical Chemistry Chemical
TE

Physics 17(28) (2015) 18477-18494.


[25] A.A. Hassanali, S.J. Singer, Model for the Water−Amorphous Silica
Interface:  The Undissociated Surface, The Journal of Physical Chemistry
EP

B 111(38) (2007) 11181-11193.


[26] H.F.W. Taylor, Cement Chemistry, 2nd ed., Thomas Telford 1997.
C

[27] S. Merlino, E. Bonaccorsi, T. Armbruster, The real structure of tobermorite


11Å: normal and anomalous forms, OD character and polytypic
AC

modifications, European Journal of Mineralogy 13(3) (2001) 577-590.


[28] E. Bonaccorsi, S. Merlino, A.R. Kampf, The Crystal Structure of
Tobermorite 14 Å (Plombierite), a C–S–H Phase, Journal of the American
Ceramic Society 88(3) (2005) 505-512.
[29] S. Merlino, E. Bonaccorsi, T. Armbruster, The Real Structures of
Clinotobermorite and Tobermorite 9 Å: OD Character, Polytypes, and
Structural Relationships, 2000.

21
ACCEPTED MANUSCRIPT
[30] S. Hamid, The crystal structure of the 11A natural tobermorite
Ca2.25[Si3=7.5(OH)1.5] 1H2O, Zeitschrift fur Kristallographie 154 (1981)
189-198.
[31] S. Merlino, E. Bonaccorsi, T. Armbruster, Tobermorites; their real structure
and order-disorder (OD) character, American Mineralogist 84(10) (1999)
1613-1621.
[32] R.J.-M. Pellenq, A. Kushima, R. Shahsavari, K.J. Van Vliet, M.J. Buehler,

PT
S. Yip, F.-J. Ulm, A realistic molecular model of cement hydrates,
Proceedings of the National Academy of Sciences 106(38) (2009) 16102-
16107.

RI
[33] S. Grangeon, F. Claret, C. Lerouge, F. Warmont, T. Sato, S. Anraku, C.
Numako, Y. Linard, B. Lanson, On the nature of structural disorder in
calcium silicate hydrates with a calcium/silicon ratio similar to tobermorite,

SC
Cement and Concrete Research 52(0) (2013) 31-37.
[34] A.J. Allen, J.J. Thomas, H.M. Jennings, Composition and density of

U
nanoscale calcium–silicate–hydrate in cement, Nature Materials 6(4)
(2007) 311-316.
AN
[35] J.W. Bullard, H.M. Jennings, R.A. Livingston, A. Nonat, G.W. Scherer, J.S.
Schweitzer, K.L. Scrivener, J.J. Thomas, Mechanisms of cement
hydration, Cement and Concrete Research 41(12) (2011) 1208-1223.
M

[36] E.M. Gartner, A proposed mechanism for the growth of CSH during the
hydration of tricalcium silicate, Cement and Concrete Research 27(5)
D

(1997) 665-672.
[37] A. Gmira, M. Zabat, R.J.-M. Pellenq, H. Van Damme, Microscopic physical
TE

basis of the poromechanical behavior of cement-based materials,


Materials and Structures 37(1) (2004) 3-14.
EP

[38] H. M. Jennings, Refinements to colloid model of C-S-H in cement: CM-II,


2008.
[39] K.A.B. Snyder, D. P., Suspended hydration, freezable water consumption,
C

and moisture transport in cement pastes exposed to 90% relative


humidity, Cement and concrete research 34 (2004) 2045-2056.
AC

[40] F. Ridi, P. Luciani, E. Fratini, P. Baglioni, Water Confined in Cement Pastes


as a Probe of Cement Microstructure Evolution, The Journal of Physical
Chemistry B 113(10) (2009) 3080-3087.
[41] S. Ghabezloo, Micromechanics analysis of thermal expansion and thermal
pressurization of a hardened cement paste, Cement and Concrete
Research 41(5) (2011) 520-532.
[42] M.J.A. Qomi, F. Ulm, R.J.-M. Pellenq, Physical Origins of Thermal
Properties of Cement Paste, Phys. Rev. Applied 3(6) (2015) 064010.00-
064010.17.

22
ACCEPTED MANUSCRIPT
[43] N.M.A. Krishnan, B. Wang, G. Falzone, Y. Le Pape, N. Neithalath, L. Pilon,
M. Bauchy, G. Sant, Confined Water in Layered Silicates: The Origin of
Anomalous Thermal Expansion Behavior in Calcium-Silicate-Hydrates,
ACS Applied Materials & Interfaces 8(51) (2016) 35621-35627.
[44] B. Viguier, A. Mortensen, Heating of TEM specimens during ion milling,
Ultramicroscopy 87(3) (2001) 123-133.
[45] R.L. Santos, R.B. Horta, J. Pereira, T.G. Nunes, P. Rocha, R. Colaço,

PT
Alkali activation of a novel calcium-silicate hydraulic binder with CaO/SiO2
= 1.1, Journal of the American Ceramic Society, in press (2018).
[46] I.G. Richardson, The calcium silicate hydrates, Cement and Concrete

RI
Research 38(2) (2008) 137-158.
[47] T. Maeshima, H. Noma, M. Sakiyama, T. Mitsuda, Natural 1.1 and 1.4 nm

SC
tobermorites from Fuka, Okayama, Japan: Chemical analysis, cell
dimensions, 29Si NMR and thermal behavior, Cement and Concrete
Research 33(10) (2003) 1515-1523.

U
[48] C.S. Skinner LB, Benmore CJ, Wenk HR, Monteiro PJM, Nanostructure of
Calcium Silicate Hydrates in Cements, Physical Review Letters 104 (2010)
AN
195502-4.
[49] H.M. Jennings, A model for the microstructure of calcium silicate hydrate in
M

cement paste, Cement and Concrete Research 30(1) (2000) 101-116.


[50] H.F.W. Taylor, Proposed Structure for Calcium Silicate Hydrate Gel,
Journal of the American Ceramic Society 69(6) (1986) 464-467.
D
TE
C EP
AC

23
ACCEPTED MANUSCRIPT
Figure Captions

Figure 1. XRD results showing the phase development with hydration. (A) anhydrous sample;
(B) hydrated non-activated sample, (C) hydrated sample activated with Na2SiO3. Legends are
as follows: P – Pseudowollastonite; T - Tobermorite; * - internal standard Al2O3. In the inset, a
higher magnification of the tobermorite peak after 7 days of hydration in the non-activated and
Na2SiO3 activated samples is shown.

PT
Figure 2. Morphology of the hydrated sample, non-activated (A) and alkali-activated (B) C/S 1.1
samples were observed using HR-TEM. The insets show the electron diffraction data taken from

RI
the same areas and compared in (C). Nanocrystals are highlighted by the areas circled in
yellow.

SC
Figure 3. A) HR-TEM image of alkali activated paste and B) experimental SAED pattern along
the [123] zone axis. The diffraction spots were indexed using a computer simulation of the 9 Å

U
tobermorite structure, using the CaRIne software. The deviation between the plane spacing and
angles of the experimental SAED and the computer simulation are reported in Table A in the
AN
additional data.
M

Figure 4. A) HR-TEM image of tobermorite nanocrystals in the alkali activated paste. The inset
shows the corresponding FFT of the crystallite within the purple square. B) Inverse Fourier
transform of the FFT in A) showing the interplanar spacing of planes (013), (2 12) and (201).
D
TE

Figure 1. DSC thermograms of a cement paste. In red the heating scan and in blue the cooling

scan.
EP

Figure 6. MD simulation of the crystal structure of the triclinic tobermorite 9 Å experimentally


C

observed in the present work, as seen down [010] (A) and along [100] (B) and [001] (C),
represented by VESTA software. The unit cell of tobermorite 9 Å is enclosed by black lines. The
AC

red and yellow spheres represent, respectively, Si and O, which form the silicate chains, the
green spheres represent the hydrogens of the hydroxyl groups SiOH. The blue and the grey
spheres represent, respectively, the interlayer and intralayer calcium.

n
Figure 7. Distribution of the silicon connectivities Q in the amorphous wollastonite obtained
using an MD simulation. The red dots in each image indicate the silicon tetrahedra linked
through Q0, Q1, Q2, Q3 or Q4.

24
ACCEPTED MANUSCRIPT
Table Captions

Table 1. Interatomic potential parameters employed in the MD simulation of the amorphous C/S
1.1 and crystalline tobermorite.

Table 2. Quantitative results obtained by XRD-Rietveld analysis for the weight percentages of

PT
the phases present in the hydraulic binder produced and in the pastes prepared from it, with
and without alkaline activation (note: the amorphous phase is accounted as the sum of both the
contributions from the amorphous component of the hydration products and the anhydrous

RI
amorphous phase).

SC
Table 3. Application of the Scherrer equation to the tobermorite peaks at 2Θ= 30° and 50°.

Table 4. Analysis of the silicate and oxygens connectivities of the tobermorite 9 Å, the

U
amorphous C/S 1.1 structure and of the crystalline wollastonite [24] using MD simulations, and
AN
Si connectivities of the amorphous C/S 1.1 [23] and of the respective C-S-H [45] obtained by
29
deconvolution of the Si MAS NMR spectra.
M
D
TE
C EP
AC

25
ACCEPTED MANUSCRIPT
Tables

PT
Table 1

Coulomb term Lennard-Jones parameters (eqn (1)) Three-Body Potential Parameters [23]

/#$ /Å

RI
Atomi qi/e ij pair Atom triplet K (eV) Ɵ0 (deg) a ρ (Å) D(eV)

Ca 1.2 Ca-Ca 0.00011 4.1978 Si-O-H 0.16482 64.2835 -0.42372 1.85 1.24-1.5

O -1.2 Ca-O 0.08404 2.4203 H-O-H 13.0000 0.00 0.00 1.4 0

SC
Si 2.4 O-Si 1.16692 1.4988 O-H-O 13.0000 0.00 0.00 1.4 0

H 0.6 O-O 0.01407 3.2528 Si-O-Si 13.0000 0.00 0.00 1.6 0

U
O-H 0.63524 0.9518 AN
M
D
TE
C EP
AC

26
ACCEPTED MANUSCRIPT

PT
Table 2
Weight %

RI
Amorphous Pseudo-
(anhydrous and Tobermorite
hydrated)
wollastonite

Anhydrous -- 86.9 13.1 ---

SC
7 days 78.6 10.4 11.1
Non-activated
90 days 70.9 8.5 20.5

U
7 days 70.4 6.2 23.5
Na2SiO3
90 days 64.9 6.7 28.3
AN
M
D
TE
C EP
AC

27
ACCEPTED MANUSCRIPT

PT
Table 3

RI
Sample 2Ɵ cos(Ɵ) B(Ɵ) B (rad) λ CuKa (nm) t (nm)

30 0.967 0.669 0.012 0.154 12.3


Non-activated

SC
50 0.907 0.450 0.008 0.154 19.5

30 0.967 1.088 0.019 0.154 7.6


Na2SiO3
50 0.907 0.611 0.011 0.154 14.4

U
AN
M
D
TE
C EP
AC

28
ACCEPTED MANUSCRIPT

Table 4.

PT
0 1 2 3 4

RI
OI OS OB Q Q Q Q Q

Oxygen Oxygen
Oxygen

SC
single- bridging
isolated
bonded

T9 1.6% 62.5% 35.9% 0% 8.6% 91.3% 0.1% 0%

U
W Cryst [24] 0% 67% 33% 0% 0% 100% 0% 0%
AN
C/S 1.1 amorph 1% 64% 35% 6% 27% 40% 22% 4%

C/S 1.1 amorph [23] -- -- -- 17.2% 28.3% 43.6% 10.9% --


C-S-H [45] -- -- -- -- 26% 74% -- --
M
D
TE
C EP
AC

29
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC

You might also like