Cement and Concrete Research 78 (2015) 3856

Contents lists available at ScienceDirect

Cement and Concrete Research

journal homepage: http://ees.elsevier.com/CEMCON/default.asp

Advances in understanding hydration of Portland cement

Karen L. Scrivener a,, Patrick Juilland b, Paulo J.M. Monteiro c
a
Laboratory of Construction Materials, Ecole Polytechnique Fdrale de Lausanne, 1015, Switzerland
b
Sika Technology AG, Zrich, Switzerland
c
Department of Civil and Environmental Engineering, University of California at Berkeley, United States

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

Article history: Progress in understanding hydration is summarized. Evidence supports the geochemistry dissolution theory as
Received 11 May 2015 an explanation for the induction period, in preference to the inhibiting layer theory. The growth of CSH is
Accepted 19 May 2015 the principal factor controlling the main heat evolution peak. Electron microscopy indicates that CSH needles
Available online 14 June 2015
grow from the surface of grains. At the peak, the surface is covered, but deceleration cannot be attributed to dif-
fusion control. The shoulder peak comes from renewed reaction of C3A after depletion of sulfate in solution, but
Keywords:
Hydration (A)
release of sulfate absorbed on CSH means that ettringite continues to form. After several days space becomes
Kinetics (A) the major factor controlling hydration.
Hydration products (B) The use of new analytical technique is improving our knowledge of the action of superplasticizers and leading to
Modeling (E) the design of molecules for different applications. Atomistic modeling is becoming a topic of increasing interest.
Admixture (D) Recent publications in this area are reviewed.
2015 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
1.1. Comparison of hydration kinetics between pure Portland systems and systems with SCMs . . . . . . . . . . . . . . . . . . . . . . . . 39
2. Hydration of C3S/alite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.1. Up to the end of the induction period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.2. The main heat evolution peak . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.2.1. Impact of SCMs on main heat evolution peak . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.3. Beyond one day . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3. Aluminates in main heat peak . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4. Perspectives on microstructural modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5. Effect of chemical admixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
6. Atomistic modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
7. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

1. Introduction area. In addition to published materials, we add input from discussions

In this paper we summarize progress in understanding the hydra-
tion of Portland based cements since the Madrid conference in 2011
[1]. The aim is to highlight the most important contributions to the
eld rather than give an exhaustive list of everything published in the
Corresponding author.
E-mail address: Karen.scrivener@ep.ch (K.L. Scrivener).
of a Nanocem workshop attended by many leading researchers in the
eld from Europe and North America [2].
Nowadays the overwhelming majority of cement sold outside
North America consists of Portland cement clinker blended with sup-
plementary cementitious materials (SCM). Nevertheless, early age
strength is still mainly due to the reaction of the clinker phases. It
is therefore important to understand this reaction, which may also
be modied by the physical presence of the SCM (so called ller ef-
fect discussed in more detail below). Later on the SCMs themselves

http://dx.doi.org/10.1016/j.cemconres.2015.05.025
0008-8846/ 2015 Elsevier Ltd. All rights reserved.
 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
Fig. 1. Rate of heat evolution normalized to clinker content for a Portland cement and the
same cement substituted 40% by quartz or by slag.
From [4].
react and contribute to the hydration process. Consideration of the
available options [3] shows that most cementitious materials used
are likely to continue to be based on Portland cement clinker for
the foreseeable future.
In this paper rst the reaction of plain Portland cement and blended
systems is compared. Then we discuss the hydration of the Portland ce-
ment clinker, focusing on the three main periods of reaction of alite/C3S:
Up to the end of the induction period (about 3 h1)
Main heat evolution peak (approx. 324 h), including impact of llers
Continuing reaction after 1 day (particularly from 1 to 28 days.)
This is followed by a discussion of the impact of the aluminate con-
taining phases and perspectives for microstructural modeling.
In the second part of the paper work on the impact of admixtures is
discussed, followed by a discussion on atomistic level modeling.
1.1. Comparison of hydration kinetics between pure Portland systems and
systems with SCMs
In blended (or composite) cements there is generally negligible reac-
tion of the SCM in the rst day or so. This is nicely illustrated in Fig. 1 from
Berodier [4]. Here the rate of heat evolution is shown as a function of the
Fig. 2. Schematic of the particle distribution after mixing for a 100% Portland cement system at a water to cement ratio of 0.4 (left) and for the same system but with 40% of the clinker
grains replaced by a ller (right). It is clear that replacement of cement grains by ller gives relatively more space for the hydrates of the clinker phases.
39
clinker phases for 3 systems: a pure Portland system, a blend of the same
Portland with 40% quartz (inert) and a blend with 40% slag. For the rst
28 h the curves for the blends with quartz and slag are essentially the
same, conrming that the reaction of slag itself is not occurring in this pe-
riod. Nevertheless, when the rate of heat evolution is normalized relative
to the clinker component (as here) it is clear that the slag and quartz en-
hance the reaction of the clinker component due to their physical pres-
ence; this is the so called ller effect and can be attributed to two main
factors. First, when SCM grains substitute clinker grains, as illustrated in
Fig. 2, there is relatively more space available for the hydrates of the clin-
ker phases to form in. Secondly, the surface of the SCM grains act as sites
for the heterogeneous precipitation and growth of hydrates.
In the example of Fig. 1, after 28 h the heat evolution of the blend
with slag is noticeably higher than that of the blend with quartz this
indicates the start of the reaction of the slag. Recent progress in under-
standing the impact of SCMs on hydration is discussed later, was
reviewed in [5] and also forms part of another article in this volume [6].
2. Hydration of C3S/alite
C3S/alite constitutes between 50 and 80% of Portland cement and
dominates the development of properties. So it is clear why many studies
have concentrated on the reactions of this phase. It is also worth noting,
that in a properly sulfated cement, in which the depletion of sulfate ions
in solution occurs shortly after the maximum of the heat main heat evo-
lution peak, the aluminate containing phases have little impact of the hy-
dration of C3S (see section 3). Furthermore the reaction of the C2S/belite
phase only really becomes signicant after about 10 days or so [1].
Before discussing progress in understanding hydration during the
different periods, it is worth noting a very important result published
by Thomas [7]. Through the use of a novel method with a scanning cal-
orimeter he could demonstrate that the activation energy of C3S had an
effectively constant value of 51 kJ/mol until well after the main heat
evolution peak up to a degree of hydration of at least 65%. This suggests
that the rate controlling step remains the same during this early hydra-
tion period, as different rate controlling mechanisms tend to have differ-
ent activation energies. The activation energy for cements was slightly
lower in the range 4449 kJ/mol for the rst 60 h or so of hydration.
The difference between pure C3S and cement can probably be attributed
to the presence of minor elements in solid solution in the alite of cement
and perhaps to the more rapid cooling conditions of industrial clinker.
2.1. Up to the end of the induction period
In the paper written for the Madrid conference [1], the mechanisms
governing the induction period, in which the initially rapid rate of hy-
dration decreases within the rst few minutes to remain at a low rate,
40 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
were discussed in detail. It was argued that the widely held theory of
the formation of a protective layer on the surface of cement grains
inhibiting further hydration was not in agreement with experimental
evidence. On the contrary the geochemical theory of dissolution,
in which the rate of dissolution is simply dependent on the
undersaturation of the solution with respect to the dissolving phase,
could explain all experimental observations [8]. Moreover there was ex-
cellent agreement between the dissolution behavior measured in dilute
solutions [9], that back calculated from calorimetry data [10] and theo-
retical calculations [8]. The essential aspect of this geochemical theory is
that at high degrees of undersaturation dissolution is rapid, as it is ener-
getically favorable for etch pits to form. At a critical degree of
undersaturation, which may be several orders of magnitude from the
equilibrium solubility, the rate of dissolution slows dramatically, as
etch pits can no longer form or do not contribute as signicantly, and
dissolution is mainly restricted to step retreat. Nevertheless, there is ev-
idence both from pure dissolution experiments [12] and also from con-
centrated pastes that etch pits are still visible for these undersaturation
levels [11].
Since Madrid, evidence has continued to accumulate to support
the geochemical dissolution theory. It was shown in [8] that a change
of dissolution process should occur at a critical undersaturation
(ln() = 64 on Fig. 3). Fig. 3 shows dissolution data for alite from
[9, 12]. From this data it is clear that the dissolution rate slows in the
range of calcium concentrations equivalent to the saturation of
Portlandite. Bullard et al. [13] have modeled the early hydration period
in a new version of the HydratiCA model and shown that the experi-
mental evolution of the solutions and solids can be well captured by
the dissolution theory, conrming the earlier publication of Kumar
et al. [10].
Despite the growing consensus on this subject some researchers
claim to have experimental evidence which cannot be explained by
the geochemical dissolution theory. Bellmann et al. studied the hydra-
tion of a C3S in paste and dilute suspension [14] and compare the de-
rived dissolution rates from these systems during the induction period
with the one obtained from pure dissolution experiments using the
same C3S sample. He found that the rates of dissolution obtained from
pure dissolution experiments were signicantly higher than the ones
determined on the dilute suspension and concentrated systems
100
10
-r ( mol.m-2.s-1)
0.1
9 ml.min-1.mm-2
18 ml.min-1.mm-2
36 ml.min-1.mm-2
0.01 300 ml.min-1.mm-2
1000 ml.min-1.mm-2
C3S MIII*
C3S TI*
C3S TII*
1E-3
-100 -95 -90
Fig. 3. Rates of dissolution obtained by Nicoleau et al. [9] and Juilland and Gallucci [12]. The blue box corresponds to concentration level close to CH saturation. corresponds to the ionic
product of the reaction of C3S with water: C3S + H2O 3Ca2+ + H4SiO4 + 6OH.
claiming therefore that dissolution could not be the rate limiting factor
at early age. However it should be noted that the C3S they studied was
somewhat unusual as it contained a signicant quantity of zinc in
solid solution and presented an induction period as long as 6 h.
The mechanisms controlling the induction period were debated in
depth at a recent workshop organized by the Nanocem consortium,
bringing together most researchers active in the domain including dissi-
dents of the geochemical theory [2]. There was general consensus on the
following points related to the initial slow down and induction period:
1. The surface becomes hydrated. This probably occurs even before ad-
dition of the mixing water due to inevitable exposure of the cement
grains to humidity in the environment [15]. The thickness of this hy-
drated layer is in the range 10100 nm. It is debatable whether this
altered surface layer can be regarded as a separate phase or not. In
any case the crystal structure is essentially a modication of that of
the anhydrous material rather than a hydrate phase such as CSH.
This hydration of the surface can explain why the surface energies
calculated from the geochemical theory [8] are much less than
those estimated from the anhydrous crystal structure [16].
2. This altered layer remains throughout the hydration process. Some of
the earliest work on 29Si NMR indicated the presence of hydrated
monomeric silicate groups [17]. Recent work by Pustovagar et al.
[15] showed that these species are present before hydration proper
(with liquid water) starts, remains nearly constant throughout the
acceleration period and that the amount depends only on the surface
area of the alite. This strongly supports the idea that these hydrated
monomers correspond to the surface hydration.
3. The primary mechanism controlling the kinetics up to the end of the
induction period is the undersaturation with respect to this surface
layer.
4. CSH nucleates quickly (on a scale of seconds to minutes). There are
several reasons why there is little growth of product during the in-
duction period, such that the degree of hydration at the end of the in-
duction period is only around 24%. As soon as the precipitation of C
SH occurs, the concentration of Ca in solution increases, eventually
up to around 35 to 40 mmol, at the end of the induction period [18].
In parallel the concentration of silicate drops dramatically to the
mol range. Consequently the rate of dissolution of C3S becomes
-85 -80 -75 -70 -65 -60 -55 -50
LN()
 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
Fig. 4. Hydrates growing around C3S at 10 h. Close to the surface there is a dense zone of
randomly oriented needles. Those oriented more perpendicularly to the surface then
seem to grow out.
From [20].
very slow as the solution is supersaturated with respect to calcium
hydroxide that is to say to the right of the gray zone in Fig. 3. In ad-
dition, as the amount of product is small it needs time to build up as
growth is proportional to the surface (or more correctly suitable
Fig. 5. SEM observation of C3S hydrated for a) 1 h 30 min, b) 3 h, c) 6 h and d) 9 h. Images are acquired on uncoated samples at 2 keV.
From [20].
41
growth sites on the surface). There may also be issues of orientation
TEM observations of Gallucci et al. [19] and more recently Bazzoni
et al. [20] suggest disorganized CSH starts to grow on the original
grains and then needles start to grow out from these regions at the
end of the induction period (Fig. 4). SEM pictures show the initial nu-
clei at random orientations on the surface and then also growth of
needle morphology at the end of the induction period (Fig. 5).
There is some conjecture about the composition of the CSH during
this early stage. If one considers the initial congruent dissolution path of
C3S this will rst cross the solubility curve of CSH at a low Ca to Si
ratio. By the end of the induction period the relatively high concentra-
tion of calcium in solution (above the equilibrium saturation of calcium
hydroxide) leads to the formation of a product with a high ratio of Ca/Si
as discussed in another publication in this proceedings [21]. The
amounts of early product are far too small for their composition to be
measured before the end of the induction period, so it is not clear if
the CSH changes in the way predicted by these thermodynamic argu-
ments or whether only the high Ca/Si forms for kinetic reasons.
2.2. The main heat evolution peak
At the end of the induction period CSH and Portlandite start to
grow rapidly. There is still considerable controversy about the trigger
for this increase in reaction rate. In pastes the weight of evidence indi-
cates that the precipitation of calcium hydroxide is the trigger. CH
does not precipitate during the induction period and the concentrations
in solution increase until a critical supersaturation of about 1.4 is
reached. Bullard and Flatt published a paper which considers two possi-
ble mechanisms [22]: either the presence of an inhibiting layer or the
geochemical approach discussed previously. Both hypotheses were
able to reproduce the hydration behavior at early age. However their
42 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856

simulations using HydratiCA showed that in the second case an inhibi-

tion of the nucleation or growth of Portlandite (without impacting the
other reactions) would produce retardation. Recent work by Juilland
et al. [23] conrms the inhibiting effect of sucrose on the precipitation
of CH without affecting much the other reactions which again strongly
supports the geochemical approach. Moreover in concentrated paste
systems, the end of the induction period is always associated with the
precipitation of Portlandite, as already noted nearly 40 years ago by
Young et al. [24] and most recently conrmed by Sowoidnich who
coupled isothermal calorimetry with electrical conductivity measure-
ments on a variety of paste systems [25].
In very dilute solutions it is possible that more rapid growth of CS
H starts to takes place before the critical supersaturation for Portlandite
precipitation has been reached, as indicated by the work of Damidot
et al. [26]. Indeed the Dijon group has made extensive studies of the hy-
dration of C3S in which the calcium ion concentration is controlled such
that no precipitation of Portlandite occurs and still see the take off in the
rate of growth of CSH. This would seem to suggest that it is also nec-
essary for a critical amount of CSH to form before its growth starts to
accelerate. In many more concentrated paste systems, particularly pure
C3S, it is often possible to see a small bump at the beginning of the in-
duction period. As discussed by Kumar et al. [10], this may be explained Fig. 6. Rate of heat evolution for pure C3S prepared at water to cement ratios of 0.4 and 0.8,
rst by the precipitation of Portlandite. Even though the precipitation of from [20]. The slight effect of increasing w/c on the shape of the peak is not in agreement
Portlandite is endothermic, the heat taken in by precipitation is with what would be predicted in the densifying growth model proposed by Bishnoi and
Scrivener [31].
outweighed by increase in the rate of dissolution of the C3S, due to the
drop in the calcium ion concentration and the level of undersaturation.
Shortly after the acceleration period proper begins as the rate of growth [28]). Several authors have tried to better capture the real physical pro-
of CSH increases. It is not clear whether this is take off in growth is cess. Garrault and Nonat proposed a simple model based on the forma-
purely a question of the amount of product surface on which growth tion of nuclei on a surface which grow outwards and laterally [29]. More
may occur or if there is also a change in the composition of the CSH recently they have proposed [30] that the rate of C3S dissolution plays a
or if the orientation of the CSH nuclei to allow growth away from role in the deceleration period.
the C3S surface is important. In the classical Avrami equation, deceleration occurs when the
The rate determining mechanism in the acceleration period has long growing regions of product start to impinge and the surface available
been attributed to the growth of CSH and there is considerable evi- for growth decreases. If the CSH product is assumed to have a massive
dence to support this both experimentally (as reviewed in [27] with fur- morphology with a density similar to that found in mature pastes there
ther evidence in [4] and [20]) and from models (reviewed in [28]). This is very little impingement of product at the peak [31]. This led Bishnoi
would appear to be at odds with the fact that the activation energy does and Scrivener to propose that the growth of CSH occurs rst in a dif-
not change at the end of the induction period as discussed above [7]. fuse manner and then densies [31]. While there is increasing experi-
Bullard et al. [13] note A near equality of the activation energies for mental evidence for the densication of CSH (as will be discussed
CSH growth and C3S dissolution would be a remarkable coincidence later), studies of hydration at different water to cement ratios [32, 20]
because the solids have such different structures and compositions.
However, such a coincidence should not be ruled out as the making
and breaking of bonds to silicate tetrahedra is highly likely to play a
role in both processes. Nevertheless, following the geochemical dissolu-
tion theory, there is a very close link between the growth of products,
which remove ions from the solution and the dissolution which replaces
those ions in solution, to maintain a solution concentration which
moves progressively to the invariant point determined by the presence
of calcium hydroxide and CSH. During the induction period the rate of
reaction is slow and the dissolution depends only on the solution con-
centration. After the end of the induction period the system moves to
more a dynamic equilibrium: as soon as ions are removed from the so-
lution by the precipitation of hydrates the local undersaturation with re-
spect to C3S will go up and so increase the dissolution rate until the ions
have been replaced in solution. For this reason it is possible to have
quite different dissolution rates for similar solution concentrations
an analogy is to imagine taking a snap shot of a motorway. One sees a
certain density of cars, but has no idea if these cars are moving slow or
fast. This inextricable link between precipitation and dissolution can ex-
plain why the activation energy does not change even when the rate de-
termining mechanism appears to change.
The acceleration period can be well captured by models which as-
sume that the rate of formation of CSH is proportional to its amount.
This is the essential mechanism behind the well known Avrami (or
JMAK) equation. However one must be cautious about simply tting Fig. 7. Impact of zinc and magnesium in alite solid solutions on hydration kinetics.
such laws unless the parameters can be explained (see discussion in From [33].
 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856 43

Fig. 8. Schematic of relationship between main heat evolution peak and growth of CSH needles.
From [20].

show that increasing the water to cement ratio does not much affect the
time to the peak (Fig. 6), as would be predicted by the mechanism pro-
posed by Bishnoi and Scrivener. The work by Bazzoni et al. on the kinet-
ics of hydration of alites with zinc and magnesium in solid solution [33]
provides further contrary evidence (Fig. 7).
Considering these studies, Bazzoni proposed [20, 33] that the kinet-
ics during the main heat evolution peak are controlled by the growth of
CSH needles. The needles nucleate on the surface of grains and,
those correctly oriented, grow outwards (Fig. 4). There appears to be a
certain length to which they grow quickly and thereafter their growth
is very slow. More and more needles nucleate and grow. When the sur-
face is completely covered with needles it is no longer possible for CS
H to grow in this manner and instead dense inner product starts to form.
This mechanism is shown schematically in Fig. 8.
A similar concept of a limitation to the extent of growth of hydrates
from the grain was proposed by Masoero et al. [34]. They proposed that
CSH only precipitates in an area close to the surface of C3S dened as a
reaction zone. As hydration proceeds C3S grains are reducing in size,
growth of CSH evolves mainly toward the center and gradually
stops its growth outward. With this idea, it was possible to capture
well the experimental kinetics, but there were no mechanisms pro-
posed as to why the zone of hydrate growth should be limited.
Electron microscopy was used by Bazzoni to study the impact of zinc
and magnesium on the nucleation density and the lengths of the CSH
needles [20, 33]. If these parameters were used in the ic modeling plat-
form good agreement could be obtained between the experiments and
the simulation, with only one tting parameter which was held con-
stant across all the simulations (Fig. 9). The discrepancy between the
simulation and the model after about 15 h is believed to be due partly
to consumption of the small particles not considered in this simple ver-
sion of the ic simulation.
This proposed mechanism leaves still several open questions: for ex-
ample, what are the parameters which limit the growth of the CSH
needles; are there less growth site available; is there a lack of reactant
(silicate ions) and more fundamentally how does CSH grow. Is it by
aggregation of nano-platelets as proposed by Nonat and co-workers
[35] and Labbez et al. [36] or is it by growth of sheet as originally pro-
posed by Gartner [37] and more recently modeled by Etzold et al. [38].
This growth process being still unclear, it is probably where more re-
search effort should be dedicated in the future.

Fig. 9. Comparison between experiment and simulation using parameters on needle growth and density from electron microscopy for hydration of alites doped with zinc or magnesium.
From [33].
44 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856

Fig. 10. a) Hydration of C3S in solutions of isopropanol and water, from [39]; b) renormalization of a).

Another study which poses more questions is the impact of
isopropanol on hydration kinetics [39], Fig. 10. Water and isopropanol
are fully miscible. Hydration with a solution containing a proportion of
isopropanol slows the hydration kinetics. The slowdown in the initial re-
action may be attributed to the lower activity of water, which will impact
the rate of dissolution of C3S and the growth of CSH. However it is inter-
esting to note that the curves for different concentrations of isopropanol
cannot be transformed by a maturity type transformation as is common
practice for hydration at different temperature. This suggests that differ-
ent mechanisms are operating in the deceleration period to the accelera-
tion period.
2.2.1. Impact of SCMs on main heat evolution peak
The study of Berodier and Scrivener provides new insights into the
impact of llers and SCMs on the main heat evolution peak [40]. As
can be seen in Fig. 1 both SCMs and llers enhance the rate of reaction
of alite in the acceleration period (steeper curve in acceleration period
when rate is normalized by clinker content). This effect, rst identied
by Gutteridge and Dalziel [41] has generally been attributed to the ller
surface providing extra nucleation sites for CSH. However, in [40] it
was shown that although the density of hydrate nucleation on the ller
surface was similar to that on the cement grains, the accelerating effect
was much smaller than could be explained by the provision of extra nu-
cleation surfaces. Rather, the acceleration can be attributed to the extra
shearing produced by the presence of the ller particles, Fig. 11. In fact
for llers and SCMs (e.g. slag and y ash) which have similar particle
size distributions of the clinker grains, the accelerating effect is very
modest and only the same as the difference between a cement paste
and a mortar, where the difference can also be attributed to extra
shear, this time produced by the sand grains. This is illustrated by the
calorimetry curves in Fig. 12 and the SEM pictures of the hydrate density
on the particle surfaces just after the beginning of the acceleration peri-
od in Fig. 13. It appears that the number of nuclei forming is probably
determined by the early dissolution period during the mixing step as
proposed by Garrault and Nonat [29]. The very early dissolution may
be increased by mixing as shown by Juilland and Gallucci [12]. It was al-
ready shown by Juilland et al. that mixing had an important effect on the
kinetics of the hydration of cement and alite [42].
One mineral addition, which clearly has an additional impact of the
nucleation of hydrates is limestone. In Fig. 11, it is clear that the accelera-
tion slopes with addition of limestone are higher than those for the other
SCMs (slag and y ash) and quartz. Scanning Electron Microscopy
(Fig. 14) shows the formation of hydrates on the surface of cement grains
(left) and limestone grains (right) after 1 h of hydration. Here it is clear
there is a much higher number of hydrate nucleation sites on the lime-
stone particles.
To summarize, the maximum in the rate of reaction seems to corre-
spond to a dramatic slowing in the growth of the CSH needles, which

Fig. 11. Relationship between slope of the rate of heat release curve during the acceleration period and the estimated shear rate between particles during mixing.
From [40].
 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
Fig. 12. Comparison of calorimetry curves for plain cement paste, the same paste with ad-
dition of quartz ller and in a mortar at a sand to cement ratio of 3.
From [40].
initially constitutes the outer CSH, but the reasons for this are still not
clear. Some potential hypotheses are:
Limitation on CSH growth:
Supply of ions, e.g. silicate
Reduction of growth sites available
Limitation of dissolution of alite
Partial coverage, etch pits coalescence
Limitation of CH growth
Reduction of growth site available
Connement of growth
2.3. Beyond one day
As discussed in Madrid [1] there is a real paucity of knowledge on the
mechanisms controlling hydration kinetics beyond one day. Especially
the period 128 days is of great practical importance, as 75% of the de-
sign strength may develop during this period. Unfortunately, this situa-
tion has not improved much in the last 4 years. The work of Berodier [4]
highlights the importance of the water to cement ratio in this later peri-
od of hydration and indicates that there are perhaps two later regimes,
Fig. 15. From 1 to 6 days the amount of hydration seems to be mainly de-
termined by the water to cement ratio or the space available for the ce-
ment hydrates. After about 6 days all the systems seem to have very
similar kinetics.
Fig. 13. Comparison of hydrates forming on the surface of cement grains for the three systems in Fig. 12.
From [40].
45
Further insight on later kinetics was obtained from studies of the
state of water by 1H NMR [43]. Fig. 16 shows the evolution of different
states of water in a white cement paste: water in capillary pores;
water in gel pores; water in CSH interlayer and water in crystalline
hydrates (Portlandite and ettringite). Up to around 23 days the
amount of water in the crystalline hydrates, in the CSH interlayer
and in the CSH gel pore increase in proportion. Around 23 days
the amount of water in the gel pore stops increasing, even though the
water in CSH interlayer and crystalline hydrates continues to in-
crease. At the same time the size of the capillary pores which still con-
tain water stabilize at a width of about 8 nm. Taking into account the
differences in neness between the white cement studied in [43] and
the gray cement in [4] and additional (unpublished) information on
the impact of water to cement ratio on the time at which the plateau
in gel porosity occurs; this seems to correspond to the transition be-
tween periods Va and Vb shown in Fig. 15. Data for mercury intrusion
porosimetry (Fig. 17) [4, 44] also indicates that the pore entry size
does not decrease further in regime Vb, even though the total amount
of pores intruded continues to decrease. The critical pore entry radius
(rst peak in the differentiated curve) does not decrease further than
around 8 nm. NMR measures a pore thickness assuming a slit shape
for the pores, while MIP measures a pore radius assuming a cylindrical
pipe model. Nevertheless, given uncertainties in the model for MIP
and the impact of the drying needed for this technique, these values al-
most certainly relate to the same phenomena. Muller et al. [43] called
these pores interhydrate and they can probably be related to the
spaces between the CSH needles as illustrated in Fig. 18.
In period Vb hydration continues with the formation of CSH, but
no more gel porosity forms, the average density of the bulk CSH in-
crease from a value of 1.71.8 at around 1 day to 2.1 as the system ap-
proaches full hydration.
The lack of long term data is in no small part due to the difculty in
carrying out long term measurements. For pure C3S, the particle size of
the grains is quite small and hydration is virtually complete after a few
days. For a typical cement the degree of hydration is around 80% at
28 days. This underlines the fact that the complete consumption of ce-
ment grains less than 20 m or so plays a large part in the slowing
down of reaction rate in the rst days and week.
Discrete measurements at set times, by X-ray diffraction/Rietveld or
by BSE image analysis are time consuming and do not have the precision
to give good data at later ages. Fig. 15 shows calorimetry data, which can
be obtained continuously and with high precision up to about 28 days.
However, beyond this point, the rate of heat evolution is so low that sig-
nicant errors can arise in the cumulative heat evolution data due to
small uctuation in the base line, etc. A method better suited to look
at long term kinetics is chemical shrinkage [45]. This has two main ad-
vantages, rst, the method intrinsically measures the cumulative
amount of reaction, and second, water is continuously supplied to the
sample so reaction is not limited by lack of water.
46 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
Fig. 14. Precipitation of hydrates in cementlimestone blend after 1 h of hydration. a) On cement grains and b) on limestone grains.
From [40].
Fig. 19 shows data for the same three systems as in Fig. 1, namely a
plain Portland cement; the same cement with 40% substitution by
quartz and the same cement with 40% substitution by slag; all at a
water to solid ratio of 0.4. The noise in the data underlines the problem
of obtaining accurate long term data, but it is not too bad. If we consider
rst the curve for the cement with 40% quartz, here the chemical shrink-
age attens out at around 40 days, this is due to the exhaustion of
hydratable material. Here the effective water to cement ratio is around
0.67, there is little impact of lack of space on the hydration kinetics.
For the plain cement the rate of reaction becomes slower than that of
the cement quartz system after a day or so. From around 1 day the
curve for the blend with slag is higher than that for the plain cement
as the slag hydrates. At around 25 days the chemical shrinkage from
the slag is exactly 1.67 times higher than that from the plain cement,
which corresponds to the increase in water to cement ratio in other
words at this point the slag hydrates have lled all the extra space
which cannot be lled by the cement hydrates due to the lack of clinker.
After this point the plain cement and the slag blend have very similar ki-
netics. Berodier looked at these long term kinetics by plotting the data
on a semi log plot, Fig. 20. The slope of the curves after 6 days is = .12
and 0.09 for the plain and slag blends respectively. This slope is the
time exponent in the degree of reaction equation. These values around
0.1 are very, very much lower than would be expected for a diffusion
controlled reaction, which would have a time exponent of 0.5.
At the present time we can only speculate on the mechanisms oper-
ating in this long term regime. Hydration in this period will be mainly of
the C2S phase and perhaps ferrite phase. Nevertheless, this may lead to
Fig. 15. Cumulative heat of hydration for a plain Portland cement at different water to ce-
ment ratios.
From [4].
increases in strength in the long term compared to 28 day strength of
between 20 and 120% [46]. The 1H NMR data [43] suggests that there
is an increase in the number sheets in the nanocrystalline regions of
CSH. Following this observation, Fig. 21 shows schematically a possi-
ble mechanism whereby growth of extra layers on the nanocrystalline
regions of CSH effectively lls in the gel pores.
Berodier and Scrivener showed [44] that for blended cements con-
taining slag or y ash, the minimum critical pore entry radius reached
at high degrees of hydration was always similar to that of the plain
Portland cement (around 8 nm). It was also observed that the degree
of reaction of the slag and or the y ash could be limited by lack of
space, as explained above for plain Portland cement. Moreover, the re-
action of slag could suppress the reaction of the clinker phases, probably
due to a competition for space.
We hope in four years' time at the next ICCC there will be more data
on long term hydration to improve our knowledge of the mechanisms
operating after 1 day. Nevertheless, it is becoming clear that the amount
of space available is a critical factor.
3. Aluminates in main heat peak
In [27], the state of knowledge about the reactions of the aluminate
containing phases was reviewed. It was concluded that the early inhibi-
tion of the reaction of C3A in the presence of gypsum could NOT be at-
tributed directly to the presence of a surface coating of ettringite. A
more likely mechanism for the slowdown of the reaction of the alumi-
nate phase is the absorption of sulfate ions on reactive dissolution
Fig. 16. Evolution of states of water in white cement paste at w/c = 0.4. Blue/circles, water
in capillary pores; green/triangles, water in gel pores (as seen by triangles in inset sche-
matic); red/squares, water in CSH interlayer (squares in inset schematic); orange/dia-
monds, water in crystalline hydrates (Portlandite and ettringite).
Adapted from [43].
 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856 47

Fig. 17. Evolution of pore structure as measured by mercury intrusion porosimetry.

From [4, 44].

sites as proposed by Minard et al. [47]. Due to this retardation of the C3A
reaction by sulfate the main reaction of this phase should occur after the
main peak of reaction of alite in a properly sulfated Portland cement. At
sulfate/C3A ratios typically used in Portland cements, the concentration
of sulfate in solution is exhausted at around 1518 h and the reaction of
C3A accelerates. However, in Portland cements the formation of
ettringite continues after the exhaustion of sulfate in solution, Fig. 22.
Quennoz and Scrivener [48] conrmed the formation of ettringite dur-
ing the shoulder peak around 15 h by in situ XRD, Fig. 23. These obser-
vations were only semi-quantitative, but were later conrmed by more
quantitative measurements in a commercial OPC [49]. Interestingly
both studies show that the amount of crystalline ettringite detected
by XRD continues to increase during the rst 15 h or so of hydration be-
fore the depletion of gypsum even though [49] indicates there is very lit-
tle consumption of C3A after the rst hour. However, the amount of C3A
dissolving during mixing [49] was sufcient to form all the ettringite
formed up to the sulfate depletion peak. This suggests that some amor-
phous alumina containing phase may form rst and ettringite progres-
sively crystallizes from this source prior to the renewed rapid reaction
of C3A after sulfate depletion.
Calcium monosulfoaluminate (monosulfo or AFm) only forms
later in a low broad peak, usually after 24 h and this may often not be
well visible in calorimetry curves of typical cements. Quennoz and
Scrivener showed that the sulfate needed for the continued formation
of ettringite came from the CSH [48] conrming an earlier observa-
tion of Gallucci et al. [19]. While solid gypsum is present and the sulfate
concentration in solution is high, sulfate is absorbed on the growing C
SH [50, 51]. When the concentration of sulfate ions in solution drops,
sulfate is desorbed from the CSH and is available to form ettringite.
In the presence of SCMs or llers this shoulder peak of the reaction of
C3A is usually enhanced as seen in Fig. 1. Some researchers have attrib-
uted this to the reaction of SCMs themselves (e.g. [52]). However, Fig. 1
proves that this is not the case as the shoulder peak in enhanced in the
presence of quartz as well as with slag. Kocaba [53] also showed that
even though the height of the shoulder peak was enhanced in the pres-
ence of slag, the heat evolved could all be accounted for by the amount
of C3A reacting. Quennoz and Scrivener [48] showed that the height of
this peak was increased just by increasing the water to cement ratio.
Thus the dilution effect of llers and SCMs may be a factor in this peak
enhancement. However it is also seen that the neness of the addition

Fig. 18. TEM images of hydrated C3S paste at 7 days showing ne structure of CSH. The size of the small bright areas within the hydrates corresponds to the size of 816 nm seen for the
connected capillary pores from NMR and MIP.
From [20].
48 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
Fig. 19. Chemical shrinkage data (normalized by cement content) for the same three sys-
tems in Fig. 1: plain Portland cement, Portland cement with 40% substitution by quartz and
Portland cement with 40% substitution by slag.
From [4].
has an important impact indicating that provision of more surface for
nucleation of hydrates is also important [4, 54].
The impact of SCMs on the aluminate reaction means that the level
of calcium sulfate addition should be re-optimized for a blended cement
rather than using the same calcium sulfate clinker ratio as in the corre-
sponding plain Portland cement. This is also why adding SCMs at the
concrete stage may lead to problems of undersulfation, particularly in
the presence of admixtures [55]. Antoni et al. [56] showed that optimi-
zation of the gypsum addition could increase the one day strength in a
blend of clinker with calcined clay and limestone by more than 50%
compared to just having the same sulfate clinker ratio as in the base
Portland cement. This emphasizes that the reactions of the aluminate
containing phases make an important contribution to early strength,
which is important practically as early strength is usually the main lim-
itation to levels of SCM addition.
C3A is generally acknowledged to be more reactive than the ferrite
phases (C2(A,F) solid solution usually referred to as C4AF). However, re-
sults shown in ([1], Figs. 18 and 19), indicate that up to 50% or the ferrite
phase may have reacted in the rst day. An extensive study of the reac-
tions of the ferrite phase by Dilnesa et al. [57, 58] used EXAFS spectros-
copy to study the hydration products of this phase. In the rst hours
ferrihydrite (FHx) formed , but already by one day Al- and Fe-siliceous
hydrogarnet is formed. Ferrihydrite is thermodynamically less stable
than Fe-siliceous hydrogarnet and therefore destabilized with time.
Fig. 20. Same data as in Fig. 19 on a semi log plot with linear t to regime beyond 6 days.
4. Perspectives on microstructural modeling
The main objective of modeling at the microstructural level is to fa-
cilitate study of the links between the hydration reaction and property
development. The hydration process is simulated in the computer to
generate a microstructure, from which properties are then computed.
Early microstructural models such as CEMHYD3D and HYMOSTRUC
made a valuable contribution in facilitating visualization of these links
(see [28] for detailed discussion of the development of microstructural
models and further references). However, in both these models the hy-
dration kinetics had to be input externally via (for example) a calorim-
etry curve. The ic platform developed by Bishnoi [59, 31] developed the
idea of a model with integrated kinetics rst proposed by Pignat and
Navi [60]. The main strength of the ic platform is that there are no in-
herent assumptions about the hydration mechanisms, these all have
to be input by the user. As discussed in section 2, this platform has
been very helpful in testing hypotheses on the hydration mechanisms.
As indicated in this paper there are still several open questions about
the mechanism controlling hydration. It is expected that progress on
these will be made in the next years. Up to the end of the induction pe-
riod, the degree of reaction is very small, the microstructural aspect
that is to say the arrangement of grains and hydration product in
space is relatively unimportant and the hydration kinetics can be well
captured by considering the alite dissolution curve and the thermody-
namics of the solution [10, 13]. Through the main hydration peak it is
clear that the outward growth of the early CSH is the most important
aspect and here the use of electron microscopy observations can enable
fairly good simulations to be made. The main area where work is still
needed is the long term period after 12 days. Here the experimental
evidence indicates that the lling of space is a critical factor, so micro-
structural models must be able to consider this explicitly as in the ic
platform.
Work on the prediction of properties from a simulated microstruc-
ture has revealed further issues related to the difference in scale be-
tween the size of features in the microstructure and the
representative volume element needed to compute properties. For a ce-
ment paste the representative volume element is on the order of 100
200 m. (For mortar or concrete it will, of course, be much larger, de-
pending on the aggregate size, but these materials can then be modeled
on a larger scale, considering the paste as a homogeneous component).
As discussed in Section 2.3 the smallest capillary pores are around 8 nm
in size, and pores of this size dominate the capillary pore structure after
only a few days. To take account of such features, explicitly in a cube of
microstructure of 100 m in dimension, would require 20,000
50,000 voxels along each edge, 10131014 voxels in total. This is still
well beyond the capacity modern computers to make meaningful calcu-
lations of hydration processes on a useful timescale. The so-called vec-
tor representation used in ic aims to avoid these problems of scale, by
using simple shapes such as spheres, but these are also not adequate to
describe the needle like morphologies of early CSH. A study by Do
et al. succeeded in resolving pores in microstructures generated by
ic down to a size of 10 nm [61] (for a static microstructure (no hydra-
tion) and using parallelization). However the resulting pore structure
was connected at a much coarser level than real microstructures as in-
dicated by mercury intrusion porosimetry measurements.
Numerous studies have shown that stiffness (which correlates
strongly with strength) can be well predicted just from the overall
amounts of the phases present. This is the basis of the gel space ratio ap-
proach of Powers and Brownyard [62] and rened by Termkhajornkit
et al. [63] Such approaches seem to work well at higher degrees of hy-
dration (N 5070%) and are coherent with the idea (Section 2.3) that
by this stage all the space is lled by hydrates plus ne pores ~8 nm.
At earlier degrees of hydration, numerical approaches which consider
the connection of the phases also work well [64].
Transport properties, important for durability, are another question.
Clearly here the size of the connected pores are important. However, as
 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
Fig. 21. Schematic of a possible mechanism for long term hydration with densication of the CSH. Extra layers are added to the nanocrystalline regions, effectively lling in the gel pores.
Black lines represent sheets of calcium, oxygen and silicon; red/squares are interlayer water and green/triangles are gel water.
also originally suggested by Powers and Brownyard [62], cementitious
materials at a water/binder ratio below about 0.4 will almost certainly
have a pore structure which is only connected at a scale of around 8
10 nm. This is supported by MIP measurements discussed in
Section 2.3. This is on the same scale as the gel porosity so, for the pur-
poses of modeling transport properties, it may be sufcient to consider
these ne interhydrate pores as part of the CSH and use the effec-
tive properties of the CSH gel homogenized at a scale below
10 nm. This was done by Zalzale et al. [65] to look at the effect of satu-
ration on the gas and water permeability of cement pastes with a lattice
Boltzmann approach and showed considerable promise.
So despite the fact that it is not practical to represent every last detail
of the microstructure, microstructural models may be quite successful
for modeling properties, particularly at high degrees of hydration. For
mechanical properties even just the overall volumes of phases may be
enough.
However, another area which needs further study, from the perspec-
tive of microstructural modeling, is the effect of temperature. Quite
modest increases in temperature, to the range 4060 C, will affect the
nature of the aluminate containing hydrates [66] and also have an im-
portant inuence on the bulk density and space lling ability of C
SH [67].
5. Effect of chemical admixtures
In a comprehensive review paper, [55] presented a critical evalua-
tion of the existing models and experimental data on setting retarders,
accelerators, and water reducing dispersants used in cementitious
Fig. 22. Calorimetry curve of modern Portland cement, showing typical shoulder peak
where a secondary formation of ettringite occurs and subsequent broad peak correspond-
ing to the formation of AFM phase.
From [27].
49
materials. A book chapter [68] by Flatt and Schober provides an exten-
sive literature review and a fundamental introduction to the adsorption
of superplasticizers on the cement minerals, polymer structural param-
eter they inuence, and the rheology, mechanical properties and dura-
bility of cementitious materials. The reader is encouraged to study
these reviews, as the present work will concentrate on more recent
publications. It is necessary to judicially analyze the details of the exper-
imental program when comparing results from different researchers.
For instance, a large number of experiments use a xed dosage when
comparing the effect of polycarboxylate superplasticizers on the hydra-
tion of cements. However, if PCEs with longer side chains are used at a
constant mass, less charges are available; this makes it difcult to de-
couple the decrease of charge and role of the side chains. In addition,
some polymers adsorb strongly when at a constant dosage while
other do not, and the difference in surface coverage can have a major ef-
fect on the rheology of the fresh paste.
The interaction of cement and chemical admixtures have been stud-
ied by atomic force microscopy [6973], zeta potential [71, 70, 74, 75],
small-angle neutron and X-ray scattering [76], calorimetry [77, 73, 78],
in situ synchrotron X-ray diffraction [79], adsorption tests [80, 81, 73,
70], X-ray photoelectron spectroscopy [82], and small-angle X-ray scat-
tering analysis [83].
The competitive adsorption between polycarboxylate super-
plasticizers and sulfate ions in the interstitial solution of cement paste
may lead to a loss of workability of the concrete [8486]. Therefore,
there is great interest in improving the composition of superplasticizers
to make them more resilient against variations in the concrete mix pro-
portions. With this objective in mind, research has been conducted in
Fig. 23. Evolution of phase assemblage for a system of a laboratory clinker containing 92%
alite and 8% C3A with addition of 4.1% gypsum.
From [48].
50 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856

layered calcium organosilicate hybrids with covalently linked organic
functionalities [87, 88]. The original observation that silylated
hydrosoluble polymers can be covalently bonded to calcium silicate hy-
drate [89] led to a detailed study of polycarboxylate-based super-
plasticizers modied with trialkoxysilane functions synthesized by
radical copolymerization [90]. The authors observed that that a partial
substitution of carboxyl by trialkoxysilane groups enhances the adsorp-
tion; however, an excess of silylated functions can be detrimental for
adsorption. Recently, Dalas et al. [91] reported that comb-like polymers
with dicarboxylate or phosphate function are less sensitive or insensitive
to the adsorption competition with sulfate ions. Cheung et al. [55]
carefully discussed the complex interactions of chemical admixtures on
the aluminatesulfate balance, particularly for high-calcium y ash
systems.
With the goal of reducing the percentage of clinker in blended
Portland cements, there has been intensive research to study the inter-
action of superplasticizers with limestone [9296, 70], y ash [93,
97100], silica fume [101, 95, 102, 103] slag [104, 78, 105], nanosilica
[106, 100, 107109], and metakaolin [105]. Recently, Burgos-Montes
et al. [110] studied the compatibility between superplasticizer admix-
tures and cements with mineral additions. Figure 24 shows that,
compared to neat cement pastes, although the incorporation of 30%
limestone increased the adsorption of melamine and lignosulfonate
superplasticizers, it did not signicantly change the behavior of the
naphthalene or polycarboxylate-based admixtures. The adsorption of
lignosulfonate in y ash pastes was smaller compared to that in the ref-
erence paste, but the consumption of melamine-based superplasticizer
was greater in the y ash pastes. For silica fume pastes, the adsorption
plateau was not observed due to large specic surface of silica fume,
(Fig. 24d) and the consumption of LS and PMS was greater than PNS.
Plank et al. [111] measured the thermodynamic parameters H,
S and G affecting the complex adsorption mechanism of PCE
superplasticizers on limestone. The adsorption is driven by an enthalpic
contribution caused by the electrostatic attraction, and an entropic con-
tribution originated by the release of counter ions and water molecules
into the pore solution. The authors reported that all PCE molecules used
in their research had negative values of G, implying that the polymer
adsorptions are energetically favorable. Electrostatic attraction and re-
lease of enthalpy are the driving mechanism for the adsorption of
PCEs containing short-sided chains, while large gains in entropy are
the main mechanism for the adsorption of PCEs containing long-sided
chains. More recently Dalas et al. [89] discussed the applications and
limitations of using the Langmuir model for polymer adsorption in ce-
mentitious systems.

Fig. 24. Superplasticizer adsorption isotherms: (a) Portland cement; (b) Portland cement with 30% wt. limestone; (c) Portland cement with 30% wt. y ash; (d) Portland cement with 10%
wt. silica fume with naphthalene (PNS)-, melamine (PMS)-, lignosulfonate (LS)-, and polycarboxylate (PCE)-based admixtures.
 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
Giraudeau et al. [112] presented an incisive discussion on the impor-
tance of the surface and intercalation chemistry of polycarboxylate copol-
ymers for cementitious systems. Plank et al. [113] made relevant
contributions for the understanding of polycarboxylate intercalation
into C3A hydrate phases, and how this intercalation is affected by the
amount of sulfate in the cement. Without sulfate, polycarboxylate interca-
lates during the C3A hydration (unless there is an excessive steric size of
the polycarboxylate). For low-sulfate concentrations, the polycarboxylate
intercalates with intercalated alkali sulfate; however, no intercalation is
observed for high-sulfate concentrations. In a more recent publication,
Habbaba et al. [114] used this concept to analyze why the delayed addi-
tion of superplasticizers creates concrete with greater owability than
when superplasticizers are added with the mixing water [115]. Their re-
sults showed that for cements with a low alkalisulfate/C3A molar ratio,
there were large differences in the rheology of the paste if early or delayed
additions were used; this is due to the intercalation of the superplasticizer
into the C3A hydrates. However, for cements with a high alkalisulfate/C3A
molar ratio, there was little differences in the cement rheology when ei-
ther early or delayed additions were used; it is assumed this because
AFt is formed and intercalation of the superplasticizer is not possible. Ng
et al. [81] studied the intercalation of polycarboxylate superplasticizers
in calcium aluminate cement with small-angle X-ray scattering, and
their results indicated that the intercalation depended on the grafting
density and anionicity, and was independent of the molecular weight of
the polycarboxylate.
The use of alkaline activators to increase the reactivity of y ash at
early ages is becoming more prevalent [116, 117]. Note that hydroxides
as an activator may have negative interactions with polycarboxylate
ether (PCE) superplasticizers [118]. Marchon et al. [119] showed that
the hydroxides did modify the PCE performance, but did so by reducing
its adsorption and not the results of chemical changes. It is assumed that
this is probably due to a competitive adsorption with the hydroxyls, al-
ready suggested by Flatt et al. [120]. The authors then proposed an ad-
sorption equilibrium constant KA,1 to rank the PCEs' susceptibility to
hydroxides. For a polymer containing n segments with N monomers
in the backbone, each segment having one side chain containing P
monomers, the adsorption equilibrium constant is given by
KA,1 (N 1)2z2P9/5N3/5n1 105, where z is the number of charges
carried by each monomer. The inuence of hydroxides on the rheology
of cement paste containing a PCE-based superplasticizer can be mea-
sured by computing the relative change in yield stress rel


SP;OH SP = SP;OH , where SP is the yield stress of the paste contain-
ing only the superplasticizer, and SP;OH is the yield stress of the paste
containing both the superplasticizer and hydroxides. The relative yield
Fig. 25. Relative yield stress change coefcient rel as a function of the adsorption equi-
librium constant KA,1 calculated with the number of charges z = 1 for all types of mono-
mers in the backbone.
From [117].
51
strength approaches to 1 when the hydroxides exert a very strong inu-
ence but tends to 0 when this inuence is small. Fig. 25 shows a strong
correlation between the relative yield strength and the adsorption equi-
librium constant KA,1. The sensitivity of the cement paste to hydroxides
is small for relative yield strengths lower than 0.6, which corresponds to
superplasticizers having KA,1 greater than 175. Superplasticizers with
KA,1 smaller than 70 are highly sensitive to large amounts of hydroxides.
It is difcult, however, to predict the sensitivity to hydroxides for
superplasticizers with KA,1 in the range of 70 to 175.
Because of their negative backbones, polycarboxylate super-
plasticizers tend to be adsorbed on positively charged surfaces. Using
AFM experiments, however, Ferrari et al. [71] showed that repulsive
forces can develop among low adsorbing materials. Using XPS, Kalina
et al. [80] identied that the mechanochemical activation of monocalcium
aluminate and polyvinyl alcohol causes the creation of new COAl
chemical bonds.
The optimization of superplasticizers has been critical for the devel-
opment of high-performance concrete [121124]. Schr et al. [101]
provide insightful observations on the preferential adsorption of
methacrylic acid ester- and allyl ether-based polycarboxylate
superplasticizers on cement and silica fume in ultra-high-performance
concrete. Their results indicate that a combination of adsorbing
polycarboxylate superplasticizers is more effective than individual poly-
mers to improve the dispersion of the cement/silica blend present in
ultra-high-performance concrete.
Superabsorbent polymers (SAPs) have been used for internal curing
of high-performance concrete [125, 126] shotcrete applications [127],
improvement of freezethaw resistance of concrete [128, 129]. A 2012
Rilem state-of-the-art report [130] summarizes the developments and
applications of superabsorbent polymers in concrete. Mechtcherine
et al. [131] studied the absorption and desorption of water by various
superabsorbent polymers in fresh mortar and their effect on the rheol-
ogy of the mixtures. The results showed that increasing the water-to-
binder ratio led to higher water absorption by the polymers and, conse-
quently, greater changes in the Bingham parameters of the mixtures as
compared to the reference mix. The authors point out the competition
for absorbing water when silica fume is added to the mixtures.
6. Atomistic modeling
Using density functional theory (DFT) calculations, Durgun et al.
[132] were able to describe at a fundamental level the slower hydration
rate of C2S compared to C3S. The results indicate that the higher reactiv-
ity of C3S is due to the reactive sites around its more ionic oxygen atoms.
Their study demonstrated that the inuence of many impurities on
structural properties is similar for C3S and C2S but different in terms of
the electronic structure. Based on their model, the authors suggest po-
tential candidates to increase the reactivity of C2S. In a subsequent
paper, Durgun et al. [133] used rst-principles computational methods
to calculate surface energies () for all lower-index orientations and
to determine the most stable surfaces as well as the equilibrium Wulff
structures. The energy required to remove Ca atoms from the surface
(average redox potential), which is linked to dissolution rates, was
found to be inversely proportional to , conrming that it is more dif-
cult to extract Ca atoms from lower surface energy surfaces. Combining
classical and quantum mechanical methods, Manzano et al. [134] stud-
ied the inuence of Mg2+, Al3+, and Fe3+ substitutions on the structure,
elastic properties, and electronic structures of alite and belite. Huang
et al. [135] used DFT calculations to analyze the inuence of (Al, Fe,
Mg) impurities on triclinic C3S. The authors determined the elastic ten-
sor for triclinic C3S and noted that a signicant sensitivity to the type,
nature, and concentration of the impurity substitution. Spin polarization
was identied as the dominant factor.
In an important contribution, Mishra et al. [136] developed force-
eld parameters for tricalcium silicate to quantify cleavage energies, an-
isotropic mechanical properties, and the adsorption of organic additives,
52 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
Fig. 26. Molecular dynamics simulation of organic molecules adsorbed on hydroxylated and dry (040) surfaces of C3S. Circular highlights indicate surface reconstruction: (a) TEA on the
hydroxylated tricalcium silicate surface, (b) TIPA on the hydroxylated surface at 383 and 298 K, and (c) glycerine on the hydroxylated and dry surfaces at 383 K.
such as triisopropanol amine (TIPA), triethanolamine (TEA), MDIPA (N-
methyl-diisopropanolamine), and glycerine. The hydroxylated C3S sur-
faces are disordered with supercial hydroxide, rotation of the silicate
anions, and partial pullout of individual silicate ions. Fig. 26 shows the
contacts of hydroxyl groups in the organic molecules with supercial
calcium ions, the hydrogen bonds, and other polar interactions. The re-
sults indicate that: (a) the adsorption mechanism of these additives in-
cluded coordination of Ca2 + ions, hydrogen bonds, and other polar
interactions with the C3S surface; (b) adsorption is weaker on hydroxyl-
ated C3S surfaces; and (c) adsorption strength decreases in the order
glycerine N MDIPA N TIPA ~ TEA. In a follow-up paper [137], the authors
used this formulation to study tricalcium aluminate, and the results
show that a supercial hydration introduces an amorphous calcium hy-
droxide surface layer that reduces the agglomeration energy from ap-
proximately 850 mJ m2 to 500 mJ m 2. The presence of adsorbed
organic molecules can signicantly reduce this agglomeration energy.
Predictions of the elastic properties of tricalcium aluminate deter-
mined from the density functional theory simulations were validated
with experimental results from synchrotron high-pressure experiments
[138]. First-principles calculations were used to study tobermorite
[139], jennite, [137] tricalcium aluminate [140], ettringite [141], and
monocarboaluminate hydrates [136].
Moon et al. [142] reported experimentally that the bulk modulus of
monocarboaluminate (54 GPa) is signicantly higher than any other hy-
dration products in concrete. In addition, Moon et al. [143] provide a de-
tailed analysis of the full elastic tensors and mechanical properties of the
crystal structure computed by rst-principles methods based on density
functional theory. Since the monocarboaluminate is an important hy-
dration product when limestone is added to cement, this high elastic
modulus could inuence mechanical properties of the cement paste
matrix. Using force-eld atomistic methods, Manzano et al. [144] com-
puted the elastic properties of alite, belite, Portlandite, and CSH, and
Shahsavari et al. [145] developed empirical force elds for complex hy-
drated calcio-silicate layered materials.
Calcium silicate hydrates (CSH), due to its importance and com-
plexity, have been the subject of much research. The pioneering work
by Pellenq et al. [146] presented a molecular model for CSH which
was able to explain many experimental results, attracting a signicant
number of citations. Richardson [147] criticized many of CaO distances
predicted by the model, and that more than half the Ca atoms in the
model are coordinated to fewer than six O atoms. Qomi et al. [148]
used the model to create a database of atomic congurations for a
broad range of CSH chemical compositions, and determined their cor-
responding defect attributes and mechanical properties. The authors
then proposed a combinatorial approach to optimize the properties of
CSH considering the following defect attributes: calcium-to-silicon
ratio and two correlation distances describing medium-range silicon
oxygen and calciumoxygen environments. In a study to quantify the
order and disorder of the molecular structure, Bauchy et al. [149]
showed that while CSH retains some signatures of a tobermorite-
like layered structure, hydrated species are completely amorphous.
Their modeling captured the main features of pair distribution function
results [150] of synthetic CSH.
Recently, intensive research has been focused on the advanced un-
derstanding of the calcium silicates/water interface. Youssef et al.
[151] showed that the non-bridging oxygen atoms on the disordered
silicate chains orient the hydrogen atoms of the interfacial water mole-
cules toward the calciumsilicate layers, making the CSH nano-pore
 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
space hydrophilic. The orientation of water molecules in the interlayer
of 14 tobermorite was obtained by molecular dynamics [152]. The
characterization of water conned at ambient temperature inside a C
SH grain (intragranular) and between CSH grains (intergranular)
for different relative humidities at 300 K was done by Bonnaud et al.
[153]. Using density functional theory (DFT), Ebbert et al. [154] showed
that CSH phase in the presence of water will change its chemical com-
position to reach the thermodynamic ground state of the system. Qomi
et al. [155] studied how substrate composition affects chemo-physical
properties of water in ultra-conned hydrophilic media by performing
molecular dynamics simulations on a set of 150 models with different
compositions of calcium and silicon contents. They reported that in C
SH the self-diffusivity increases with increasing conned water densi-
ty. Using a reactive force eld, Manzano et al. [156] documented the dis-
sociation and polarization of water in the micropores CSH. Their work
indicated that the short-range order and elastic properties of CSH are
not signicantly affected by water dissociation reactions but that the
shear strength is. CSH under shear loading localizes the shear strain
preferentially at nanoscale sites that are rich in highly conned water
[157].
Gonzalez-Teresa et al. [158] developed a computational model for
the development of CSH gel using a random, sequential addition of
nano-colloid nuclei. Even with the existing approximations, the model
captured the growth of CSH, and the results were consistent with
the degree of hydration and calorimetric results. Molecular dynamics
simulations developed by Dolado et al. [159] were able to predict a
branched three-dimensional network of CSH, where the segmental
branches were 3 3 6 nm. The authors suggest that the resulting
branched structure results from the interweaving and restructuring
process during the growth of the CSH segments. No separate CSH
brick can be observed at the scale, and the branched structure resem-
bles the model proposed by Feldman and Sereda [160]. A recent polydis-
perse colloidal packing model [161, 162] provides new insights on the
nanostructure and nanomechanics of CSH. Using the Grand Canonical
Monte Carlo method to mimic precipitation events during molecular
dynamics simulations, Ioannidou et al. [163] proposed that the acceler-
ation of the early hydration reactions is associated with the formation of
an optimal local crystalline packing that favors the growth of large and
elongated aggregates. The deceleration of the reactions is associated
with the defects produced during precipitation that favor branching
and gelation. An alternative numerical model for the early growth of
CSH has been proposed by Etzold et al. [164]: the sheet nuclei form
immediately on the cement surface before the growth acceleration.
The cement dissolves, and CSH precipitates at active growth sites on
the CSH structure.
7. Concluding remarks
As the majority of cements will be based on Portland clinker for the
foreseeable future, understanding the hydration mechanisms is very
important. Over the past decade or so, due to developments in analytical
techniques and trans-institutional organizations, such as the Nanocem
network (www.Nanocem.org) our knowledge is advancing steadily.
Concerning the induction period: it is becoming increasingly clear that
this is NOT caused by an inhibiting layer of product, which subse-
quently disappears. However hydration of the surface layer will cer-
tainly affect the thermodynamics of the surface and so its dissolution.
Evidence to support the geochemical dissolution hypothesis is grow-
ing. But work is still needed in this area to better understand interfa-
cial reactions and particularly the impact of admixtures on these.
Concerning the main hydration peak: this is dominated by the growth
of hydrates CSH and Portlandite. The intimate link between precip-
itation of products and alite dissolution makes it difcult to disentan-
gle the rate controlling mechanisms. However it is clear that the peak
in the rate of the reaction is NOT due to the onset of diffusion control.
53
To move forward it is essential to consider the impact of a wide range
of parameters particle size, water to cement ratio, doping ions,
water activity (isopropanol mixes), etc. It is very easy to t a narrow
range of data with rate equations relating to several different mecha-
nisms. To move forward we need to be able to derive parameters for
simulations relating to the reaction rate, directly from experiment,
rather than by tting. Recent studies by electron microscopy hold
promise here.
Concerning the long term hydration: the rate here is also quite different
to what might be expected from diffusion control. The lling of space
by hydration products seems to be the main limiting factor. During
this period SCMs commonly present in blended cement contribute
to the lling of space. The capillary pore space is rened and is soon
dominated by pores connected at a size of around 8 to 16 nm.
Concerning the reactions of the aluminate containing phases: the control
of the reaction in the early stages by addition of calcium sulfate (e.g.
gypsum) is NOT directly due to formation of an inhibiting ettringite
layer. In real cements ettringite formation continues after the deple-
tion of sulfate ions in solutions, due to the desorption of sulfate ions
from the CSH phase. The ferrite phase also reacts signicantly to
produce a siliceous hydrogarnet solid solution.
The mechanism behind the action of superplasticizers starts to be well
understood, it is becoming clear that in most cases steric rather than
electrostatic effects dominate. This understanding has led to the de-
sign of a vast number of different SP molecules optimized for different
purposes.
More researchers are exploring the potential of atomistic modeling
tools to understand the mechanisms operating at this level. Such ap-
proaches are promising, but careful validation against experiment is
essential. It is also most important to realize that in real cementitious
materials water is omnipresent. The very large surface of the main hy-
dration phase, CSH is all in contact with water. Consequently ap-
proaches which only consider phases in Vacuum may be misleading.
However, including water in models is challenging due to the varied
and complex interactions of this molecule e.g. high polarizability,
ease of dissociation, proton hoping, etc.
Microstructural level modeling can be of great value in linking the mi-
crostructure formed during hydration to the development of proper-
ties. However (as for all modeling approaches), we need to be
realistic about the limitations and approximations needed. It is also
important to give more consideration to the real thermal history ex-
perienced by concrete in the eld, rather than only considering hydra-
tion at 20 C.
It should be noted that many of the statements above highlight that
established ideas, repeated in numerous text books, may NOT be correct
and researchers need to keep an open mind to advancing knowledge.
Acknowledgments
Karen Scrivener would like to thank the many friends and colleagues
who have contributed to the discussion on hydration over the years. In
particularly three PhD students, Amelie Bazzoni, Elise Berodier and
Arnaud Muller have made outstanding contribution to understanding
the hydration process in recent years and have contributed directly to
this paper by providing gures and diagrams.
Paulo Monteiro acknowledges the nancial support given by the Na-
tional Research Foundation-Prime Minister's ofce, Republic of
Singapore through a grant to the Berkeley Education Alliance for Re-
search in Singapore (BEARS) for the Singapore-Berkeley Building Ef-
ciency and Sustainability in the Tropics (SinBerBEST) Program.
References
[1] K.L. Scrivener, A. Nonat, Hydration of cementitious materials, present and future,
Cem. Concr. Res. 41 (2011) 651665.
54 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
[2] Cement Hydration Workshop, Organised by Nanocem Consortium. Villars, CH, June
2628, 2014.
[3] K.L. Scrivener, Options for the future of cement, Indian Concr. J. 88 (2014) 1121.
[4] E. Berodier, Impact of the supplementary cementitious materials on the kinetics
and microstructural development of cement hydration, Thse EPFL, n 6417,
2015 (http://infoscience.ep.ch/record/204690/les/EPFL_TH6417.pdf).
[5] B. Lothenbach, K. Scrivener, R.D. Hooton, Supplementary cementitious materials,
Cem. Concr. Res. 41 (2011) 12441256.
[6] M.C.G. Juenger, R. Siddique, Recent advances in understanding the role of supple-
mentary cementitious materials in concrete, Cem. Concr. Res. (2015) 7180 (in
this volume).
[7] J.J. Thomas, The instantaneous apparent activation energy of cement hydration
measured using a novel calorimetry-based method, J. Am. Ceram. Soc. 95 (10)
(2012) 32913296.
[8] P. Juilland, E. Gallucci, R. Flatt, K.L. Scrivener, Dissolution theory applied to the in-
duction period in alite hydration, Cem. Concr. Res. 40 (2010) 831844.
[9] L. Nicoleau, A. Nonat, D. Perrey, The di- and tricalcium silicate dissolutions, Cem.
Concr. Res. 47 (2013) 1430.
[10] A. Kumar, S. Bishnoi, K.L. Scrivener, Modelling early age hydration kinetics of alite,
Cem. Concr. Res. 42 (7) (2012) 903918.
[11] J.M. Makar, G.W. Chan, End of induction period in ordinary Portland cement as ex-
amined by high-resolution scanning electron microscopy, J. Am. Ceram. Soc. 91
(2008) 12921299.
[12] P. Juilland, E. Gallucci, Morpho-topological investigation of the mechanisms and kinet-
ic regimes of alite dissolution, Cem. Conc. Res. 76 (2015) 180191.
[13] J.W. Bullard, G.W. Scherer, J.J. Thomas, Time dependent driving forces and the ki-
netics of tricalcium silicate hydration, Cem. Concr. Res. 74 (2015) 2634.
[14] Frank Bellmann, Thomas Sowoidnich, Horst-Michael Ludwig, Denis Damidot,
Dissolution rates during the early hydration of tricalcium silicate, Cem. Conc.
Res 72 (2015) 108116.
[15] E. Pustovgar, J.-B. d'Espinose, M. Palacios, A. Andreev, R. Sangodkar, B. Chmelka, R.J.
Flatt, Revealing the steps of tricalcium silicate hydration, Proceedings 14th ICCC
Beijing, 2015.
[16] E. Gartner, Discussion of the paper dissolution theory applied to the induction pe-
riod in alite hydration by P. Juilland et al., Cem. Concr. Res. 40 (2010) 831-844,
Cem. Concr. Res. 41 (2011) 565567.
[17] A.S. Rogers, Hydration of tricalcium silicate followed y 29Si NMR with cross-
polarisation, J. Am. Ceram. Soc. 71 (1988) 9196.
[18] P.W. Brown, E. Franz, G. Frohnsdorff, H.F.W. Taylor, Analyses of the aqueous phase
during early C3S hydration, Cem. Concr. Res. 14 (1984) 257262.
[19] E. Gallucci, M. Mathur, K. Scrivener, Microstructural development of early age hy-
dration shells around cement grains, Cem. Concr. Res. 40 (2010) 413.
[20] A. Bazzoni, Study of early hydration mechanisms of cement by means of electron
microscopy, Thse EPFL n6296, 2014 (http://infoscience.ep.ch/record/200217/
les/EPFL_TH6296.pdf).
[21] B. Lothenbach, A. Nonat, Calcium silicate hydrates: solid and liquid phase compo-
sition, Cem. Concr. Res. (2015) 5770 (in this volume).
[22] J.W. Bullard, R.J. Flatt, New insights into the effect of calcium hydroxide precipita-
tion on the kinetics of tricalcium silicate hydration, J. Am. Ceram. Soc. 93 (2010)
18941903.
[23] P. Juilland, E. Gallucci, Proceedings of 14th ICCC Beijing, 2015.
[24] J.F. Young, H.S. Tong, R.L. Berger, Compositions of solutions in contact with hydrat-
ing tricalcium silicate pastes, Journal of the American Ceramic Society 60 (56)
(1977) 193198.
[25] T. Sowoidnich, A Study of Retarding Effects on Cement and Tricalcium Silicate Hy-
dration Induced by Superplasticizers, University of Weimar, 2015 (Thesis).
[26] D. Damidot, A. Nonat, P. Barret, Kinetics of tricalcium silicate hydration in diluted
suspensions, J. Am. Ceram. Soc. 73 (1990) 33193322.
[27] 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, Cem. Concr. Res. 41
(2011) 12081223.
[28] J.J. Thomas, J.J. Biernacki, J.W. Bullard, S. Bishnoi, J.S. Dolado, G.W. Scherer, A. Luttge,
Modeling and simulation of cement hydration kinetics and microstructure devel-
opment, Cem. Concr. Res. 41 (2011) 12571278.
[29] S. Garrault, A. Nonat, Hydrated layer formation on tricalcium and dicalcium silicate
surfaces: experimental study and numerical simulations, Langmuir 17 (2001)
81318138.
[30] S. Gaurrault, L. Nicoleu, A Nonat, Tricalcium Silicate Hydration Modelling and Nu-
merical Simulations in Proceedings of CONMOD 10, Lausanne, Switzerland, 2010.
9194.
[31] S. Bishnoi, K.L. Scrivener, Studying nucleation and growth kinetics of alite hydra-
tion using ic, Cem. Concr. Res. 39 (2009) 849860.
[32] D.M. Kirby, J.J. Biernacki, The effect of water-to-cement ratio on the hydration ki-
netics of tricalcium silicate cements: testing the two-step hydration hypothesis,
Cem. Concr. Res. 42 (2012) 11471156.
[33] A. Bazzoni, S. Ma, Q. Wang, X. Shen, M. Cantoni, K.L. Scrivener, The effect of magne-
sium and zinc ions on the hydration kinetics of C3S, J. Am. Ceram. Soc. 97 (2014)
36843693.
[34] E. Masoero, J.J. Thomas, H.M. Jennings, A reaction zone hypothesis for the effects of
particle size and water-to-cement ratio on the early hydration kinetics of C3S, J.
Am. Ceram. Soc. 97 (2014) 967975.
[35] S. Gaufnet, . Finot, E. Lesniewska, A. Nonat, Observation directe de la croissance
d'hydrosilicate de calcium sur des surfaces d'alit et de silice par microscopie
force atomique: Direct observation of the growth of calcium silicate hydrate on
alite and silica surfaces by atomic force microscopy, C.R. Acad. Sci., Ser. IIA: Earth
Planet. Sci. 327 (1998) 231236.
[36] C. Labbez, B. Jnsson, C. Woodward, A. Nonat, M. Delhormea, The growth of
charged platelets, Phys. Chem. Chem. Phys. 16 (2014) 2380023808.
[37] E.M. Gartner, A proposed mechanism for the growth of CSH during the hydra-
tion of tricalcium silicate, Cem. Concr. Res. 27 (1997) 665672.
[38] M.A. Etzold, P.J. McDonald, A.F. Routh, Growth of sheets in 3D connements a
model for the CSH meso structure, Cem. Concr. Res. 63 (2014) 137142.
[39] E. Boehm-Courjault, J.E. Rossen, K. Scrivener, Composition and morphology of CS
H in C3S pastes hydrated by water/isopropanol mixes using electron microscopy,
Proceedings 14th ICCC, Beijing, 2015.
[40] E. Berodier, K.L. Scrivener, Understanding the ller effect on the nucleation and
growth of CSH, J. Am. Ceram. Soc. 97 (2014) 37643773.
[41] W.A. Gutteridge, J.A. Dalziel, Filler cement: the effect of the secondary component
on the hydration of Portland cement: part I. A ne non-hydraulic ller, Cem. Concr.
Res. 20 (1990) 778782.
[42] P. Juilland, A. Kumar, E. Gallucci, K. Scrivener, Effect of mixing on the early hydra-
tion of alite and OPC systems, Cem. Concr. Res. 42 (2012) 11751188.
[43] A.C.A. Muller, K.L. Scrivener, A.M. Gajewicz, P.M. McDonald, Densication of CSH
measured by H-1 NMR relaxometry, J. Phys. Chem. C 117 (2013) 403412.
[44] E. Berodier, K.L. Scrivener, Evolution of pore structure in blended systems, Cem.
Concr. Res. 73 (2015) 2535.
[45] M. Geiker, Characterisation of development of cement hydration using chemical
shrinkage, Chapter 3 in A practical guide to microstructural analysis of cementi-
tious materials, eds: K.Scrivener, R. Snelling and B. Lothenbach, to be published
CRC press October 2015
[46] S. Popovics, Fig. 3.12, pg 198 in Strength and Strength related Properties of Con-
crete: A Quantitative Approach, John Wiley and Sons, 1998.
[47] H. Minard, S. Garrault, L. Regnaud, A. Nonat, Mechanisms and parameters control-
ling the tricalcium aluminate reactivity in the presence of gypsum, Cem. Concr.
Res. 37 (2007) 14181426.
[48] A. Quennoz, K.L. Scrivener, Interactions between alite and C3Agypsum hydrations
in model cements, Cem. Concr. Res. 44 (2013) 4654.
[49] D. Jansen, F. Goetz-Neunhoeffer, B. Lothenbach, J. Neubauer, The early hydration of
ordinary Portland cement (OPC): an approach comparing measured heat ow
with calculated heat ow from QXRD, Cem. Concr. Res. 42 (2012) 134138.
[50] R. Barbarulo, Comportement des matriaux cimentaires: actions des sulfates et de
la temprature, Effects of Sulfate and Temperature on the Chemical Behavior of
Cement-Based Materials, in French, Ecole Normale Suprieure de Cachan, France-
UniversitUniversite Laval, Canada, 2002 (PhD thesis).
[51] R. Barbarulo, H. Peycelon, S. Prene, Experimental study and modelling of sulfate
sorption on calcium silicate hydrates, Ann. Chim. Sci. Mater. 28 (2003) S5S10.
[52] E.H. Kadri, S. Aggoun, G. Schutter, K. Ezziane, Combined effect of chemical nature
and neness of mineral powders on Portland cement hydration, Mater. Struct. 43
(5) (2010) 665673.
[53] V. Kocaba, Development and evaluation of methods to follow microstructural de-
velopment of cementitious systems including slags, Thse EPFL, n 4523, 2009.
[54] S. Dittrich, J. Neubauer, F. Goetz-Neunhoeffer, The inuence of y ash on the hydra-
tion of OPC within the rst 44 ha quantitative in situ XRD and heat ow calorim-
etry study, Cem. Concr. Res. 56 (2014) 129138.
[55] J. Cheung, A. Jeknavorian, L. Roberts, D. Silva, Impact of admixtures on the hydra-
tion kinetics of Portland cement, Cem. Concr. Res. 41 (2011) 12891309.
[56] M. Antoni, J. Rossen, F. Martirena, K. Scrivener, Cement substitution by a combina-
tion of metakaolin and limestone, Cem. Concr. Res. 42 (2012) 15791589.
[57] B.Z. Dilnesa, Fe-containing hydrates and their fate during cement hydration: ther-
modynamic data and experimental study, Thse EPFL, n 5262, 2012.
[58] B.Z. Dilnesa, E. Wieland, B. Lothenbach, R. Dhn, K.L. Scrivener, Fe-containing
phases in hydrated cements, Cem. Concr. Res. 4555 (2014).
[59] S. Bishnoi, K.L. Scrivener, ic: a new platform for modelling the hydration of ce-
ments, Cem. Concr. Res. 39 (2009) 266274.
[60] P. Navi, C. Pignat, Simulation of cement hydration and the connectivity of the cap-
illary pore space, Adv. Cem. Based Mater. 4 (1996) 5867.
[61] Q.H. Do, S. Bishnoi, K.L. Scrivener, Numerical simulation of porosity in cements,
Transp. Porous Media 99 (2013) 101117.
[62] T.C. Powers, T.L. Brownyard, Studies of the physical properties of hardened
Portland cement paste, J. Am. Concr. Inst. 43 (Nine parts) (19461947).
[63] P. Termkhajornkit, Q.H. Vu, R. Barbarulo, S. Daronnat, G. Chanvillard, Dependence
of compressive strength on phase assemblage in cement pastes: beyond gel
space ratio experimental evidence and micromechanical modelling, Cem.
Concr. Res. 56 (2014) 111.
[64] Q.H. Do, Modelling properties of cement paste from microstructure: porosity, me-
chanical properties, creep and shrinkage, Thse EPFL, n 5881, 2013.
[65] M. Zalzale, P.J. McDonald, K.L. Scrivener, A 3D lattice Boltzmann effective media
study: understanding the role of CSH and water saturation on the permeability
of cement paste, Model. Simul. Mater. Sci. Eng. 21 (2013) 085016, http://dx.doi.
org/10.1088/0965-0393/21/8/085016 (16 pp.).
[66] B. Lothenbach, T. Matschei, G. Mschner, F.P. Glasser, 2008, thermodynamic
modelling of the effect of temperature on the hydration and porosity of Portland
cement, Cem. Concr. Res. 38 (2008) 118.
[67] E. Gallucci, X. Zhang, K. Scrivener, Effect of temperature on the microstructure of
calcium silicate hydrate (CSH), Cem. Concr. Res. 53 (2013) 185195.
[68] R.J. Flatt, I. Schober, Superplasticizers, in: N. Roussel (Ed.), Understanding the Rhe-
ology of Concrete, Woodhead publishing 2012, pp. 144208.
[69] A. Kauppi, K.M. Andersson, L. Bergstrm, Probing the effect of superplasticizer ad-
sorption on the surface forces using the colloidal probe AFM technique, Cem.
Concr. Res. 35 (1) (January 2005) 133140.
[70] R.J. Flatt, I. Schober, E. Raphael, C. Plassard, E. Lesniewska, Conformation of
adsorbed comb copolymer dispersants, Langmuir 25 (2009) 845.
 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
[71] L. Ferrari, J. Kaufmann, F. Winnefeld, J. Plank, Interaction of cement model systems
with superplasticizers investigated by atomic force microscopy, zeta potential, and
adsorption measurements, J. Colloid Interface Sci. 347 (2010) 1524.
[72] L. Ferrari, J. Kaufmann, F. Winnefeld, J. Plank, Multi-method approach to study in-
uence of superplasticizers on cement suspensions, Cem. Concr. Res. 41 (2011)
10581066.
[73] L. Ferrari, J. Kaufmann, F. Winnefeld, J. Plank, Reaction of clinker surfaces investi-
gated with atomic force microscopy, Constr. Build. Mater. 35 (2012) 9296.
[74] S. Srinivasan, S.A. Barbhuiya, D. Charan, S.P. Pandey, Characterizing cement
superplasticiser interaction using zeta potential measurements, Constr. Build.
Mater. 24 (2010) 25172521.
[75] M.M. Alonso, M. Palacios, F. Puertas, Compatibility between polycarboxylate-
based admixtures and blended cement pastes, Cem. Concr. Compos. 35
(2013) 151162.
[76] W.-S. Chiang, E. Fratini, F. Ridi, S.-H. Lim, Y.-Qi. Yeh, P. Baglioni, S.-M. Choi, U.-S.
Jeng, S.-H. Chen, Microstructural changes of globules in calciumsilicatehydrate
gels with and without additives determined by small-angle neutron and X-ray
scattering, J. Colloid Interface Sci. 398 (2013) 6773.
[77] F. Ridia, E. Fratinia, P. Luciania, F. Winnefeld, P. Baglionia, Hydration kinetics of
tricalcium silicate by calorimetric methods, J. Colloid Interface Sci. 364 (2011)
118124.
[78] P. iler, I. Kolov, J. Krtk, J. Havlica, J. Brandtetr, Inuence of superplasticizers [107] I. Navarro-Blasco, M. Perez-Nicol.s, J.M. Fernandez, A. Duran, R. Sirera, J.I. Alvarez,
on the course of Portland cement hydration, Chem. Pap. 68 (2014) 9097.
[79] M.-C. Schlegel, A. Sarfraz, U. Mller, U. Panne, F. Emmerling, First seconds in a
building's lifein situ synchrotron X-ray diffraction study of cement hydra-
tion on the millisecond timescale, Angew. Chem. Int. Ed. 51 (2012)
49934996.
[80] M. Palacios, P. Bowen, M. Kappl, H.-J. Butt, M. Stuer, C. Pecharromn, U. Aschauer, F.
Puertas, Repulsion forces of superplasticizers on ground granulated blast furnace
slag in alkaline media, from AFM measurements to rheological properties, Mater.
Constr. 62 (2012) 489513.
[81] L. Lei, J.A. Plank, Concept for a polycarboxylate superplasticizer possessing en-
hanced clay tolerance, Cem. Concr. Res. 42 (2012) 12991306.
[82] L. Kalina, J. Msilko, J. Koplk, F. oukal, XPS characterization of polymer
monocalcium aluminate interface, Cem. Concr. Res. 66 (2014) 110114.
[83] S. Ng, E. Metwalli, P. Mller-Buschbaum, J. Plank, Occurrence of intercalation of PCE
superplasticizers in calcium aluminate cement under actual application conditions,
as evidenced by SAXS analysis, Cem. Concr. Res. 54 (2013) 191198.
[84] J. Zimmermann, C. Hampel, C. Kurz, L. Frunz, R.J. Flatt, Effect of polymer structure
on the sulfatepolycarboxylate competition, in: T.C. Holland, P. Gupta, V.M.
Malhotra (Eds.),Ninth ACI International Conference on Superplasticizers and
Other Chemical Admixtures in Concrete, Seville 2009, pp. 165175.
[85] R.J. Flatt, J. Zimmermann, C. Hampel, C. Kurz, L. Frunz, C. Plassard, E. Lesniewska, R.J.
Flatt, J. Zimmermann, C. Hampel, C. Kurz, L. Frunz, C. Plassard, E. Lesniewska, The
role of adsorption energy in the sulfatepolycarboxylate competition, in: T.C.
Holland, P. Gupta, V.M. Malhotra (Eds.),Ninth ACI International Conference on
Superplasticizers and Other Chemical Admixtures in Concrete, Seville 2009,
pp. 153164.
[86] K. Yamada, A summary of important characteristics of cement and
superplasticizers, in: T.C. Holland, P. Gupta, V.M. Malhotra (Eds.),Ninth ACI Inter-
national Conference on Superplasticizers and Other Chemical Admixtures in Con-
crete, Seville 2009, pp. 153164.
[87] J. Minet, S. Abramson, B. Bresson, C. Sanchez, V. Montouillout, N. Lequeux, New lay-
ered calcium organosilicate hybrids with covalently linked organic functionalities,
Chem. Mater. 16 (2004) 39553962.
[88] J. Minet, S. Abramson, B. Bresson, A. Franceschini, H. Van Damme, N. Lequeux, Or-
ganic calcium silicate hydrate hybrids: a new approach to cement based nanocom-
posites, J. Mater. Chem. 16 (2006) 13791383.
[89] A. Franceschini, S. Abramson, B. Bresson, H. VanDamme, N. Lequeux, Cement
silylated polymers nanocomposites, International Congress on the Chemistry of
Cement, 2007.
[90] W. Fan, F. Stoffelbach, J. Rieger, L. Regnaud, A. Vichot, B. Bresson, N. Lequeux, A new
class of organosilane-modied polycarboxylate superplasticizers with low sulfate
sensitivity, Cem. Concr. Res. 42 (2012) 166172.
[91] F. Dalas, A. Nonat, S. Pourchet, M. Mosquet, D. Rinaldi, S. Sabio, Tailoring the anionic
function and the side chains of comb-like superplasticizers to improve their ad-
sorption, Cem. Concr. Res. 67 (2015) 2130.
[92] M. Sahmaran, H.A. Christianti, I.O. Yaman, The effect of chemicals admixtures and
mineral additives on the properties of self-compacting mortars, Cem. Concr.
Compos. 28 (2006) 432440.
[93] R. Magarotto, I. Torresan, N. Zeminian, Inuence of the molecular weight of
policarboxylate superplasticizers on the rheological properties of fresh cement
pastes, mortar and concrete, Proceedings of the 11th ICCC, 2 2003, p. 514.
[94] P.F.G. Banll, Additivity effects in the rheology of fresh concrete containing water-
reducing admixtures, Constr. Build. Mater. 25 (2011) 642.
[95] C. Artelt, E. Garcia, Impact of superplasticizer concentration and of ultra-ne parti-
cles on the rheological behaviour of dense mortar suspensions, Cem. Concr. Res. 38
(2008) 633642.
[96] N. Mikanovic, C. Jolicoeur, Inuence of superplasticizers on the rheology and stabil-
ity of limestone and cement pastes, Cem. Concr. Res. 38 (2008) 907.
[97] S. Erdogdu, C. Arslantrk, S. Kurbetci, Inuence of y ash and silica fume on the
consistency retention and compressive strength of concrete subjected to
prolonged agitating, Constr. Build. Mater. 25 (2011) 12771281.
[98] S. Kawashima, J.-W.T. Seo, D. Corr, M.C. Hersam, S.P. Shah, Dispersion of CaCO3
nanoparticles by sonication and surfactant treatment for application in y ashce-
ment systems, Mater. Struct. 47 (2014) 10111023.
55
[99] C. Wei, S. Peiliang, S. Zhonghe, F. Jianfeng, Adsorption of superplasticizers in y ash
blended cement pastes and its rheological effects, J. Wuhan Univ. Technol.-Mater.
Sci. 27 (2012) 773778.
[100] M. Toledano-Pradosa, M. Lorenzo-Pesqueiraa, B. Gonzlez-Fonteboab, S. Seara-
Pazc, Constr. Build. Mater. 48 (2013) 628635.
[101] J. Plank, C. Schroe, M. Gruber, M. Lesti, R. Sieber, Effectiveness of polycarboxylate
superplasticizers in ultra-high strength concrete: the importance of PCE compati-
bility with silica fume, J. Adv. Concr. Technol. 7 (2009) 512.
[102] L. Senff, D. Hotza, Mortars with nano-SiO2 and micro-SiO2 investigated by exper-
imental design, Constr. Build. Mater. 24 (2010) 14321437.
[103] Ch. Schr, M. Gruber, J. Plank, Preferential adsorption of polycarboxylate
superplasticizers on cement and silica fume in ultra-high performance concrete
(UHPC), Cem. Concr. Res. 42 (2012) 14011408.
[104] M. Palacios, Y.F. Houst, P. Bowen, F. Puertas, Adsorption of superplasticizer admix-
tures on alkali-activated slag pastes, Cem. Concr. Res. 39 (2009) 670677.
[105] A. Kashani, J.L. Provis, J. Xu, A.R. Kilcullen, G.G. Qiao, J.S.J. van Deventer, Effect of
molecular architecture of polycarboxylate ethers on plasticizing performance in
alkali-activated slag paste, J. Mater. Sci. 49 (2014) 27612772.
[106] S. Abd.El.Aleem, M. Heikal, W.M. Morsi, Hydration characteristic, thermal expan-
sion and microstructure of cement containing nano-silica, Constr. Build. Mater.
59 (2014) 151160.
Assessment of the interaction of polycarboxylate superplasticizers in hydrated
lime pastes modied with nanosilica or metakaolin as pozzolanic reactives, Constr.
Build. Mater. 73 (2014) 112.
[108] J.M. Fernndez, A. Duran, I. Navarro-Blasco, J. Lanas, R. Sirera, J.I. Alvarez, Inuence
of nanosilica and a polycarboxylate ether superplasticizer on the performance of
lime mortars, Cem. Concr. Res. 43 (2013) 1224.
[109] M.-H. Zhang, J. Islam, Use of nano-silica to reduce setting time and increase early
strength of concretes with high volumes of y ash or slag, 29 (2012) 573580.
[110] O. Burgos-Montes, M. Palacios, P. Rivilla, F. Puertas, Compatibility between
superplasticizer admixtures and cements with mineral additions, Constr. Build.
Mater. 31 (2012) 300309.
[111] J. Plank, B. Sachsenhauser, J. de Reese, Experimental determination of the thermo-
dynamic parameters affecting the adsorption behaviour and dispersion effective-
ness of PCE superplasticizers, Cem. Concr. Res. 40 (2010) 699709.
[112] C. Giraudeau, J.B.E. Lacaillerie, Z. Souguir, A. Nonat, R. Flatt, Surface and intercala-
tion chemistry of polycarboxylate copolymers in cementitious systems, J. Am.
Ceram. Soc. 92 (2009) 24712488.
[113] J. Plank, D. Zhimin, H. Keller, F. Hossle, W. Seidl, Fundamental mechanisms for
polycarboxylate intercalation into C3A hydrate phases and the role of sulfate pres-
ent in cement, Cem. Concr. Res. 40 (2010) 4557.
[114] A. Habbaba, Z. Dai, J. Plank, Formation of organo-mineral phases at early addition of
superplasticizers: the role of alkali sulfates and C3A content, Cem. Concr. Res. 59
(2014) 112117.
[115] R.J. Flatt, Y.F. Houst, A simplied view on chemical effects perturbing the action of
superplasticizers, Cem. Concr. Res. 31 (2001) 11691176.
[116] A. Palomo, A. Fernandez Jimenez, G. Kovalchuk, L.M. Ordonez, M.C. Naranjo, OPC-
y ash cementitious system: study of the gel binders produced during alkaline hy-
dration, J. Mater. Sci. 42 (2007) 29582966.
[117] C. Shi, A. Fernandez Jimenez, A. Palomo, New cements for the 21st century: the
pursuit of an alternative to Portland cement, Cem. Concr. Res. 41 (2011) 750763.
[118] M. Palacios, F. Puertas, Effect of superplasticizer and shrinkage-reducing admixtures
on alkali-activated slag pastes and mortars, Cem. Concr. Res. 35 (2005) 13581367.
[119] D. Marchon, U. Sulser, A. Eberhardt, R.J. Flatt, Molecular design of comb-shaped
polycarboxylate dispersants for environmentally friendly concrete, Soft Matter 9
(2013) 10719.
[120] R.J. Flatt, Y.F. Houst, P. Bowen, H. Hofmann, J. Widmer, U. Sulser, U. Maeder, T.A.
Burge, Interaction of superplasticizers with model powders in a highly alkaline
medium, ACI Special Publication, v. 173 1997, pp. 743762.
[121] Y.F. Houst, P. Bowen, F. Perche, A. Kauppi, P. Borget, L. Galmiche, J.-F. Le Meins, F.
Lafuma, R.J. Flatt, I. Schober, P.F.G. Banll, D.S. Swift, B.O. Myrvold, B.G. Petersen,
K. Reknes, Design and function of novel superplasticizers for more durable high
performance concrete (superplast project), Cem. Concr. Res. 38 (2008) 11971209.
[122] J. Goaszewski, Inuence of cement properties on new generation superplasticizers
performance, Constr. Build. Mater. 35 (2012) 586596.
[123] P. Pereira, L. Evangelista, J. de Brito, The effect of superplasticizers on the mechan-
ical performance of concrete made with ne recycled concrete aggregates, Cem.
Concr. Compos. 34 (2012) 10441052.
[124] R.J. Flatt, N. Roussel, C.R. Cheeseman, Concrete: an eco material that needs to be
improved, J. Eur. Ceram. Soc. 32 (2012) 27872798.
[125] V. Mechtcherine, M. Gorges, C. Schroe, A. Assmann, W. Brameshuber, A.B. Ribeiro,
D. Cusson, J. Custdio, E.F. da Silva, K. Ichimiya, S. Igarashi, A. Klemm, K. Kovler,
A.N.M. Lopes, P. Lura, H.-W. Reinhardt, R.D.T. Filho, J. Weiss, M. Wyrzykowski, G.
Ye, S. Zhutovsky, Effect of internal curing by using superabsorbent polymers
(SAP) on autogenous shrinkage and other properties of a high-performance ne-
grained concrete: results of a RILEM round-robin test, Mater. Struct. 47 (3)
(2014) 541562.
[126] J. Pirard, V. Pollet, N. Cauberg, Mitigating autogenous shrinkage in HPC by internal
curing using superabsorbent polymers, Int. RILEM Conf. Vol. Chang. Hardening
Concr. Test. Mitigation 2006, pp. 97106.
[127] O.M. Jensen, Use of superabsorbent polymers in construction materials, 1st Int.
Conf. Microstruct. Relat. Durab. Cem. Compos. Nanjing 2008, pp. 757764.
[128] A.-E. Brdern, V. Mechtcherine, Multifunctional use of SAP in strain-hardening
cement-based composites, Int. RILEM Conf. Use Superabsorbent Polym. Other
New Addit. Concr. August, Lyngby 2010, pp. 1122.
56 K.L. Scrivener et al. / Cement and Concrete Research 78 (2015) 3856
[129] H.W. Reinhardt, A. Assmann, S. Mnnig, Superabsorbent polymers (SAP) an ad-
mixture to increase the durability of concrete, 1st Int. Conf. Microstruct. Relat.
Durab. Cem. Compos. 1315 Oct., Nanjing 2008, pp. 313322.
[130] V. Mechtcherine, H. Reinhardt, Application of super absorbent polymers (SAP) in
concrete construction, Rilem State-of-the-Art Report, Springer, 2012.
[131] V. Mechtcherine, E. Secrieru, C. Schr, Effect of superabsorbent polymers (SAPs)
on rheological properties of fresh cement-based mortars development of yield
stress and plastic viscosity over time, Cem. Concr. Res. 67 (2015) 5265.
[132] E. Durgun, H. Manzano, R.J.M. Pellenq, Jeffrey C. Grossman, Understanding and
controlling the reactivity of the calcium silicate phases from rst principles,
Chem. Mater. 24 (2012) (12621267).
[133] E. Durgun, H. Manzano, P.V. Kumar, J.C. Grossman, The characterization, stability,
and reactivity of synthetic calcium silicate surfaces from rst principles, J. Phys.
Chem. C 118 (2014) 1521415219.
[134] H. Manzano, E. Durgun, M.J.A. Qomi, F.-J. Ulm, R.J.M. Pellenq, J.C. Grossman, Impact
of chemical impurities on the crystalline cement clinker phases determined by at-
omistic simulations, Cryst. Growth Des. 11 (2011) (29642972).
[135] J. Huang, L. Valenzano, T.V. Singh, R. Pandey, G. Sant, Inuence of (Al, Fe, Mg) im-
purities on triclinic Ca3SiO5: interpretations from DFT calculations, Cryst. Growth
Des. 14 (2014) 21582171.
[136] R.K. Mishra, R.J. Flatt, H. Heinz, Force eld for tricalcium silicate and insight into
nanoscale properties: cleavage, initial hydration, and adsorption of organic mole-
cules, J. Phys. Chem. C 117 (2013) 1041710432.
[137] R.K. Mishra, L. Fernndez-Carrasco, R.J. Flatt, Hendrik Heinz, A force eld for
tricalcium aluminate to characterize surface properties, initial hydration, and or-
ganically modied interfaces in atomic resolution, Dalton Trans. 43 (2014) 10602.
[138] J. Moon, S. Yoon, R.M. Wentzcovitch, S.M. Clark, P.J.M. Monteiro, Elastic properties
of tricalcium aluminate from high-pressure experiments and rst-principles calcu-
lations, Am. Ceram. Soc. 95 (2012) 29722978.
[139] R. Shahsavari, R.J.M. Pellenq, F. Ulm, First-principles study of elastic constants and
interlayer interactions of complex hydrated oxides: case study of tobermorite and
jennite, J. Am. Ceram. Soc. 92 (2009) 23232330.
[140] H. Manzano, J.S. Dolado, A. Ayuela, Structural, mechanical, and reactivity properties
of tricalcium aluminate using rst-principles calculations, J. Am. Ceram. Soc. 92
(2009) 897902.
[141] H. Manzano, A. Ayuela, A. Telesca, P.J.M. Monteiro, J. Dolado, Ettringite strengthen-
ing at high pressures induced by the densication of the hydrogen bond network, J.
Phys. Chem. 116 (2012) 1613816143.
[142] J. Moon, J.E. Oh, M. Balonis, F.P. Glasser, S.M. Clark, P.J.M. Monteiro, High pressure
study of low compressibility tetracalcium aluminum carbonate hydrates
3CaOAl2O3CaCO311H2O, Cem. Concr. Res. 42 (2012) 105110.
[143] J. Moon, S. Yoon, R. Wentzcovitch, P.J.M. Monteiro, First principles elasticity of
monocarboaluminate hydrates, Am. Mineral. 99 (2014) 13601368.
[144] H. Manzano, J.S. Dolado, A. Ayuela, Elastic properties of the main species present in
Portland cement pastes, Acta Mater. 57 (2009) 16661674.
[145] R. Shahsavari, R.J. Pellenq, F.-J. Ulm, Empirical force elds for complex hydrated
calcio-silicate layered materials, Phys. Chem. Chem. Phys. 13 (2011) 10021011.
[146] R.J. Pellenq, A. Kushima, R. Shahsavari, K.J. Van Vliet, M.J. Buehler, S. Yip, F.J. Ulm, A
realistic molecular model of cement hydrates, Proc. Natl. Acad. Sci. U. S. A. 106
(2009) 1610216107.
[147] I.G. Richardson, The importance of proper crystal-chemical and geometrical rea-
soning demonstrated using layered single and double hydroxides, Acta Crystallogr.
B69 (2013) 150162.
[148] M.J.A. Qomi, K.J. Krakowiak, M. Bauchy, K.L. Stewart, R. 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) 4960, http://dx.doi.org/10.1038/ncomms5960.
[149] M. Bauchy, M.J.A. Qomi, F.-J. Ulm, R.J.-M. Pellenq, Order and disorder in calcium
silicatehydrate, J. Chem. Phys. 140 (2014) 214503.
[150] S. Soyer-Uzun, S.R. Chae, C.J. Benmore, H.-R. Wenk, P.J.M. Monteiro, Compositional
evolution of calcium silicate hydrate (CSH) structures by total X-ray scattering, J.
Am. Ceram. Soc. 95 (2012) 793798.
[151] M. Youssef, R.J.-M. Pellenq, B. Yildiz, Glassy nature of water in an ultraconning
disordered material: the case of calciumsilicatehydrate, J. Am. Chem. Soc. 133
(2011) 24992510.
[152] S. Yoon, P.J.M. Monteiro, Molecular dynamics study of water molecules in interlay-
er of 14 tobermorite, J. Adv. Concr. Technol. 11 (2013) 180188.
[153] P.A. Bonnaud, Q. Ji, B. Coasne, R.J.-M. Pellenq, K.J. Van Vliet, Thermodynamics of
water conned in porous calciumsilicatehydrates, Langmuir 28 (2012)
1142211432.
[154] C. Ebbert, G. Grundmeier, N. Buitkamp, A. Krger, F. Messerschmidt, P. Thissen, To-
ward a microscopic understanding of the calciumsilicatehydrates/water inter-
face, Appl. Surf. Sci. 290 (2014) 207214.
[155] M.J.A. Qomi, M. Bauchy, F.-J. Ulm, R.J.-M. Pellenq, Anomalous composition-
dependent dynamics of nanoconned water in the interlayer of disordered calci-
umsilicates, J. Chem. Phys. 140 (2014) 054515.
[156] H. Manzano, S. Moeini, F. Marinelli, A.C.T. van Duin, F.-J. Ulm, R.J.-M. Pellenq, Con-
ned water dissociation in microporous defective silicates: mechanism, dipole dis-
tribution, and impact on substrate properties, J. Am. Chem. Soc. 134 (2012)
22082215.
[157] H. Manzano, E. Masoero, I. Lopez-Arbeloa, H.M. Jennings, Shear deformations in
calcium silicate hydrates, Soft Matter 9 (2013) 73337341.
[158] R. Gonzalez-Teresa, J.S. Dolado, A. Ayuela, J.C. Gimel, Nanoscale texture develop-
ment of CSH gel: a computational model for nucleation and growth, Appl.
Phys. Lett. 103 (2013) 234105.
[159] J.S. Dolado, M. Griebel, J. Hamaekers, F. Heber, The nano-branched structure of ce-
mentitious calciumsilicatehydrate gel, J. Mater. Chem. 21 (2011) 44454449.
[160] R.F. Feldman, P. Sereda, A new model for hydrated Portland cement and its practi-
cal implications, Eng. J. Can. 53 (1970) 53.
[161] E. Masoero, E. Del Gado, R.J.-M. Pellenq, F.-J. Ulm, S. Yip, Nanostructure and
nanomechanics of cement: polydisperse colloidal packing, PRL 109 (2012)
155503.
[162] E. Masoero, E. Del Gado, R.J.-M. Pellenq, S. Yipe, F.-J. Ulm, Nano-scale mechanics of
colloidal CSH gels, Soft Matter 10 (2014) 491.
[163] K. Ioannidou, R.J.-M. Pellenq, E. Del Gado, Controlling local packing and growth in
calciumsilicatehydrate gels, Soft Matter 10 (2014) 1121.
[164] M.A. Etzold, P.J. McDonald, A.F. Routh, Growth of sheets in 3D connements a
model for the CSH meso structure, Cem. Concr. Res. 63 (2014) 137142.