Cement and Concrete Research 86 (2016) 1–11

Contents lists available at ScienceDirect

Cement and Concrete Research

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

A new view on the kinetics of tricalcium silicate hydration

L. Nicoleau a,⁎, A. Nonat b
a
BASF Research Construction Materials & Systems, BASF Construction Solutions GmbH, 83308 Trostberg, Germany
b
Laboratoire Interdisciplinaire Carnot de Bourgogne UMR6303 CNRS, Université de Bourgogne, 9 avenue Alain Savary, BP 47870, 21078 Dijon Cedex, France

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

Article history: C3S hydration is an interesting example of chemical coupling between C3S dissolution, C–S–H and portlandite
Received 18 June 2015 precipitation. It occurs because Ca2+, OH− and silicate ions are present in C3S, in both hydration products and
11 April 2016 in the surrounding solution. Various experimental data sets reveal that the undersaturation with respect to C3S
Accepted 27 April 2016
always increases when C3S hydration enters into the deceleratory phase, leading to the conclusion that C3S dis-
Available online 3 May 2016
solution is at the origin of this deceleration, not C–S–H growth. In addition, as soon as portlandite precipitates, the
Keywords:
dissolution limits the hydration already in the acceleratory hydration step. The evolution of the undersaturation
Hydration cannot account for the hydration peak. Rather, it results from an extension and subsequently a decrease of the C3S
Kinetics reactive surface area. The formation and coalescence of dissolution etch-pits can provide a reasonable explana-
Pore solution tion for the C3S hydration kinetics.
Ca3SiO5 © 2016 Elsevier Ltd. All rights reserved.
Cement

1. Introduction respect to portlandite to lead to the precipitation of portlandite. In

Tricalcium silicate (C3S1), the main component of ordinary Portland
cement, hydrates into calcium-silicate-hydrates (C–S–H) and then pos-
sibly into portlandite (Ca(OH)2), as follows:
C3S dissolution:
Ca3 SiO5 þ 5 H2 O→3Ca2þ þ 6OH− þ H4 SiO4
such conditions, the solution gets more and more concentrated in cal-
cium and hydroxide ions. If the maximal supersaturation with respect
to portlandite is exceeded, the latter precipitates leading to a new ki-
netic situation, hence resulting in the C3S dissolution and both C–S–H
and Ca(OH)2 precipitation. In all cases, the chemical equations impose
the following:
ð1Þ

xþyþz¼3
Cx-S-Hy precipitation:

x Ca2þ þ 2x OH− þ H4 SiO4 →ðCaOÞ x −ðSiO2 Þ−ðH2 OÞ b
þ ð2 þ x−bÞ H2 O ð2Þ
Ca(OH)2 precipitation:
y Ca2þ þ 2y OH− →CaðOHÞ 2 ð3Þ
with x + y = 3 and x the calcium to silicon molar ratio that is a variable
parameter of the C–S–H composition that strongly depends on the con-
centration of calcium hydroxide in solution. b, the amount of water mol-
ecules in C–S–H is also a variable close to the value of 1.8 [1,2] for C–S–H
with high Ca/Si ratios. C3S hydration consists of at least two of these re-
actions which are the dissolution of C3S and the precipitation of C–S–H.
Using the example of dilute suspensions, it is possible to hydrate C3S
under conditions undersaturated or insufficiently supersaturated with
⁎ Corresponding author.
E-mail address: luc.nicoleau@basf.com (L. Nicoleau).
1
In cement notation, C, S and H stand for CaO, SiO2 and H2O respectively. C3S = Ca3SiO5,
C–S–H = (CaO)x−(SiO2)−(H2O)y.
z being the number of moles of calcium in solution for 1 mol of C3S. The
authors chose H4SiO4 as silicate species for balancing the reactions but
are aware that the solution consists of a mix of H4SiO4, H3SiO− 4 and
H2SiO2−4 , for which the speciation depends on pH.
The hydration of cement today is firmly recognized to be a dissolu-
tion–precipitation process, and reaction kinetics has been a prominent
research topic for a century. Le Chatelier, being a fervent defender of
the “through-solution” theory, was the first to lay the foundations of hy-
dration chemistry [3]. Since then, even if our knowledge considerably
increased, the overall picture on hydration is still not complete and
some questions are still a matter of debate (see reviews in [4–6]).
More recently, our understanding on the dissolution reaction of
tricalcium silicate has been increased [7–9]. Under the new light of
these results, some critical points regarding alite hydration kinetics
are revisited in this paper (the tricalcium silicate phase in cement is
commonly called alite).
The hydration of alite is routinely measured by the well-established
technique of isothermal conduction calorimetry. Q̇ , the heat flow re-
leased during the hydration of alite is equal to the sum of heat flows

http://dx.doi.org/10.1016/j.cemconres.2016.04.009
0008-8846/© 2016 Elsevier Ltd. All rights reserved.
2 L. Nicoleau, A. Nonat / Cement and Concrete Research 86 (2016) 1–11
released or stored by each reaction:
Q_ ¼ RC3 S diss: ΔHC3 S diss: þ RC–S–H prec: ΔHC–S–H prec: þ RCH prec: ΔHCH prec:
with ΔHi the reaction enthalpy of the reaction i. In paste, the 3 rates be-
come rapidly equal and then Q̇ is proportional to the hydration rate. The
characteristic heat evolution curve (Fig. 1) obtained during the hydra-
tion of C3S in paste presents different phases:
I The dissolution peak: this non-steady state step is difficult to interpret
because of the calorimeter inertia and often because of the frictional
heat due to the sample introduction.
II The so-called “dormant” period or sometimes called the induction pe-
riod. This period of low activity precedes the acceleration of hydra-
tion. It is generally very short in the case of pure C3S hydration.
III The acceleratory phase: this is a relatively short period of time but im-
portant for the early mechanical properties. During these first hours,
the rate increases until reaching a maximum value.
IV The deceleratory phase: with the acceleratory period it forms the
main hydration peak. The deceleratory phase continues until the
end of hydration.
Among the still-debated issues, two of them have received particular
attention during the last two or three decades. The first one is the occur-
rence of the dormant period. To explain this, two theories are often re-
ported in the literature. One is based on the precipitation of a transient
metastable C–S–H phase [10,11] and another one is based on C–S–H nu-
cleation kinetics, which may be hampered by aluminum ions, [12] for
instance. An alternative to these theories, recently proposed, suggests
that the dormant period results from underlying mechanisms
governing the opening of dissolution etch-pits [9].
The second most important issue is the reasons leading to the main
hydration peak. The acceleration is generally attributed to the increase
of the surface area of C–S–H in a kind of autocatalytic process. The
entry into the deceleratory regime following the maximum rate value
is more discussed. So far, most of the hypotheses advanced to rationalize
the deceleration also rely upon C–S–H growth. There are two assump-
tions, both of which postulated that C–S–H growth is constrained in
space. It is assumed that some constraints limit the growth rate and con-
sequently the hydration rate. They are derived from Avrami's model
[13–15] and have been adapted to C3S (or cement) hydration [16].
Some authors considered that the transformation takes place on the
surface of grains [17,18]. The theories arising in these articles are differ-
entiated by the origin of space constraints. The first proposes that C–S–H
growth is hindered by existing C–S–H that grows outward from the sur-
rounding cement grains [17], and the second proposes that the hin-
drance on growth is due to adjacent C–S–H growing on the same
Fig. 1. Very typical evolution of alite hydration at water-to-alite ratio of 0.5. The usual steps
are represented: I the pure dissolution peak, II the “dormant” period, III the acceleratory
period and IV the deceleratory period.
grain [18]. Note that in this latter, the consequence is not a decrease
but a constant hydration rate.
ð4Þ It is worthwhile mentioning that, to the best of the authors' knowl-
edge, these C–S–H growth-based hypotheses have never been experi-
mentally verified, but rather have been supported by self-consistent
fitting. Indeed, the typical hydration curve (as plotted in Fig. 1) has
been modeled or fitted using these theories. More generally, many dif-
ferent models (based on C–S–H) can be used to reproduce the sigmoidal
hydration curve [17–21] showing in some cases, excellent agreement
between experiments and fits. But even a perfect fit does not provide
validation of a theory [22]. The C3S hydration curve is the result of dis-
solution and precipitation reactions leading to possible dual interpreta-
tions if the theory used to produce a curve does not account for both
reactions. In addition to the possible space constraints on C–S–H
growth, C3S is depleting over time while C–S–H grows. After a certain
time, and as long as water is still available, C3S will not be able to provide
enough material for the C–S–H growth and the hydration rate will be
limited. This has been described by Nonat et al. [23] and appears to be
indisputable evidence. However, the corresponding hydration time
has never been experimentally identified, but has only been predicted
to occur.
In this article, a simple approach has been used that consists of the
examination of ion concentrations during hydration under various ex-
perimental conditions. The data are all extracted from the literature.
As explained in the next section, the evolution of ion concentrations al-
lows for the determination of which reaction drives the kinetics, for ex-
ample through the calculation of the undersaturation with respect to
C3S and the supersaturation with respect to hydration products. In
spite of the simplicity of the approach to monitor ion concentrations,
the evolution of the characteristics of the pore solution and the conse-
quences on hydration kinetics has never been thoroughly studied, ex-
cept during the very early period of hydration [24] and in a very
recent theoretical study [25]. The relation between the C3S dissolution
rate and its undersaturation has been recently described [7]. Unfortu-
nately, the same kind of work on the C–S–H precipitation is still pend-
ing. The role of the dissolution rate on hydration kinetics will be put
into perspective, and light will be shed on the phenomena taking
place during the main hydration peak shown by calorimetry. An alter-
native hydration scheme will be then proposed and meanwhile the rel-
evance of the existing theories based on C–S–H growth will be
discussed.
2. Notions of heterogeneous kinetics applied to C3S hydration
C3S hydration is an example of transformation of solid A into solid B
through the solution. As a matter of fact, during C3S hydration, C3S dis-
solves and releases ions into the solution at the flow Φdiss. At the same
time, hydrates, i.e. C–S–H and portlandite, precipitate and consume
ions from the solution at the flow Φprec. Because C3S, C–S–H and
portlandite are composed of the same species, dissolution and precipita-
tion rates, i.e., the appearance rate of C–S–H and portlandite and the dis-
appearance rate of C3S, are linked. Such a connection is the definition of
the so-called coupled reactions. In case of a steady state or a quasi-
steady state, both flows become equal:
φdiss: ðtÞ ¼ φprec: ðtÞ: ð5Þ
In the following, some known concepts about heterogeneous kinet-
ics are described to establish pertinent groundwork for C3S hydration
kinetics and to show that C3S dissolution and precipitation of C–S–H
and portlandite are coupled. Many interesting examples of coupling
can be found in geochemistry. Indeed coupling is typical in diagenesis,
metamorphism, metasomatism, chemical weathering, and other
water-rock reactions [26–28]. In these geochemical transformations,
mineral A transforms into mineral B through the surrounding fluid.
The dissolution rate of A is equal to the precipitation rate of B and the
 L. Nicoleau, A. Nonat / Cement and Concrete Research 86 (2016) 1–11 3

Fig. 2. (a) Evolution of the sodium concentration over the transformation of albite into sanidine in 0.1 M KHCO3 solution extracted from the Ref. [34]. According to the Eq. (8) the sodium
concentration is directly proportional to the degree of reaction. (b) Ratios of albite dissolution rates versus sanidine precipitation rates. (c) Evolution of undersaturation with respect to
albite and the supersaturation with respect to sanidine (SI = saturation index). The saturation index is defined as the logarithm of the supersaturation or undersaturation.
(Reproduced from Ref. [34] with permission from Elsevier, Copyright 2010). For (a), (b) and (c), the dots refer to experimental measurements and the lines are calculated from
numerical reaction path model simulation.

role of the fluid in the process is fundamental. Two ways to couple dis-
solution–precipitation reactions have been identified. On the one hand,
coupling is induced by the stress coming from crystallization as in pseu-
domorphic replacement [27,29]. In this transformation, the coupling of
dissolution–precipitation reactions allows the conservation of the vol-
ume and texture characteristics of solid A during its replacement by
solid B. On the other hand, as during C3S hydration, if some chemical el-
ements are common in A and B, coupling is ensured through the differ-
ence of chemical potentials between the solution and both solids for
elements common in A and B. Because hydration of C3S occurs by the
second process, chemical coupling will be developed in the next section.
The driving force for any chemical reaction is the deviation from
equilibrium [30]:
μ Ai Nμ solution
i Nμ Bi
with μi the chemical potential of the species i common in A and B. The
last equation means there is a gradient of chemical potential(s) in the
solution from the surface of A to the solution and from the solution to
the surface of B, which allows the transport of species from A to B.
When dissolution or precipitation of solid in water is concerned, the
equilibrium state is the solubility of solid in solution, and the deviations
from equilibrium are respectively called undersaturation and supersat-
uration for the dissolution and the precipitation reactions. At constant
temperature and pressure, the Gibbs free energy ΔG varies with the de-
viation from equilibrium. There are kinetic laws with different levels of
complexity, which link the dissolution or precipitation rates of minerals
to ΔG. A very generic law can be expressed by the following equation
[26,31]:
Rate ¼ k  Asurf  f ðΔGÞ  gðai Þ
where Rate is the overall rate of reaction, k the rate constant function of
the temperature, Asurf is the surface extent where reaction occurs
(called the reactive surface area), ΔG is the difference in Gibbs free en-
ergy between the solid and its molecules in solution and g(ai) a function
highlighting all the possible catalyzing and inhibiting effects of the spe-
cies i. Despite the multiplicity of the function f(ΔG) in nature, all rates
monotonically decrease with the deviation from equilibrium: in any
case, the solid A will dissolve more and more slowly when the solution
is less and less undersaturated with respect to A and the solid B will pre-
cipitate more and more slowly when the solution is less and less super-
saturated with respect to B. From this point, one may conclude that both
rates can be coupled if some chemical elements are common in A and B.
Indeed, the precipitation of solid B will contribute to increase the
undersaturation with respect to solid A, and, the dissolution of A will
ð6Þ contribute to increase the supersaturation with respect to B, and vice
versa. It means that during the transformation of solid A into solid B,
the composition of the solution is always between the solubility of A
and the solubility of B. The evolution of the solution upon transforma-
tion is called the kinetic pathway and is generally plotted on an activ-
ity/activity diagram.
A suitable example of coupled reactions known in geochemistry is
the transformation of albite into sanidine in KHCO3 solution:
NaAlSi 3 O8 þ Kþ →KAlSi 3 O8 þ Naþ : ð8Þ
Experiments carried out by Alekseyev et al. [32] and further proc-
essed and fitted by Zhu et al. [33,34] give an example of the coupling be-
tween albite dissolution and sanidine precipitation, which results in a
sigmoidal degree of reaction curve (Fig. 2a). Both reactions run at
ð7Þ equal rate from the 200th minute (Fig. 2b) and chemical balance is
4 L. Nicoleau, A. Nonat / Cement and Concrete Research 86 (2016) 1–11

Fig. 3. Schematic representation of the evolution of the saturation degrees with respect to C3S and C–S–H from the surface to the bulk with and without stirring (represented, respectively,
by dotted and full lines).

achieved in solution which always remains undersaturated with respect equals the C–S–H precipitation rate as soon as C–S–H nucleates,
to albite and supersaturated with respect to sanidine (Fig. 2c). More i.e., after less than one hour in typical paste conditions [42]. As a further
generally, the position of the kinetic path with respect to both solubility evidence for the coupling of the C3S dissolution and C–S–H precipitation
products, and its evolution over time, are of primary importance to find rates, it is well known that the composition of the solution does not
out what reactions (dissolution or precipitation) drive the overall trans- evolve suddenly during the hydration except within the first minutes
formation kinetics. Indeed, if the kinetic path is close to the solubility [24]. This is called the quasi-steady state; the dissolution and precipita-
product of the dissolving solid A and far away from the solubility prod- tion flows can be therefore considered equal at any time as described by
uct of B, the transformation kinetic is driven by the early precipitation of Eq. (5). The flow ϕ, either of dissolution or of precipitation, depends on
B (as an example, the authors cite the hydration of calcium sulfate hemi- (1) I, the reactive interface extent and (2) the reaction rate at the inter-
hydrate [35–37]). Conversely, if the kinetic path is close to the solubility face, called the interfacial rate Ri which varies with ΔG, the deviation
product of the precipitating solid B and far away from the solubility of A, from equilibrium.
the transformation is driven by the dissolution of A (as an example, the
authors cite the carbonation of wollastonite [38,39]).
The diffusion of reactants and/or products also has an effect upon the φ ¼ I  Ri ðΔGÞ or φ ¼ I  Ri ðβÞ since ΔG ¼ RT  ln β ð9Þ
transformation kinetics and is an active phenomenon in all solid-
solution reactions [40,41]. Nevertheless, diffusion significantly influ-
ences the overall kinetics only when the diffusion of at least one species where β is the degree of undersaturation for dissolution or the degree of
is sensibly slower than the interfacial reaction rate. In this particular supersaturation for precipitation.
case, the concentration of species close to the reactive interface will
drastically differ from the bulk. As an example, when ions released by
dissolution are not expelled quickly enough from the surface, they are
stored close to the surface and the undersaturation and dissolution
rate are accordingly reduced. Both undersaturation and supersaturation
will change compared to the case of rapid diffusion and the position of
the kinetic pathway in the activity/activity diagram as well. Nonethe-
less, the evolution towards one of the solubility products always indi-
cates that the transformation is kinetically driven by one of the
reactions [34].
Interestingly, C3S hydration reveals many features similar to the al-
bite–sanidine transformation. Indeed, the C3S dissolution rate rapidly

Table 1
Solubility products of C3S, C–S–H and portlandite. The reported Ksp corresponds to the
equilibria Eqs. (1), (2) and (3). * from [7] and ** from [53].

C3S⁎ β–C–S–H⁎⁎ γ–C–S–H⁎⁎ Ca(OH)2 Fig. 4. Evolution of alite content determined by QXRD during the hydration at 20 °C of an
ordinary Portland cement mixed at a water to cement ratio of 0.5. The hydration rate of
ln Ksp −50 −32.7 −39.3 −11.8
alite is calculated after smoothing and derivatizing the alite content curve.
 L. Nicoleau, A. Nonat / Cement and Concrete Research 86 (2016) 1–11
Fig. 5. Evolution of the saturation indexes with respect to C3S solubility (undersaturation)
and with respect to C–S–H and Ca(OH)2 solubilities (supersaturations) during the
hydration of alite at w/c = 0.5 reported in [47] and shown in Fig. 4.
Then, the Eq. (5) is combined to the Eq. (9) to yield:
Rhyd: ¼ IC3 S  RiC3 S  βiC3 S ¼ ICSH  RiCSH  βiCSH
¼ ICaðOHÞ2  RiCaðOHÞ2  βiCaðOHÞ2
where IC3S, C–S–H, Ca(OH)2 denote the extent of C3S, C–S–H and portlandite
reactive interfaces and RiC3S, C–S–H, Ca(OH)2 the interfacial rates which de-
pend on βiC3S, CSH, Ca(OH)2 the under- and super-saturations at the re-
spective interfaces. Using this last equation could be risky without
further precaution. Indeed, serious complications may emerge because
different possible interfaces could arise from the complex heteroge-
neous hydration process. Some of these difficulties have been reported
in the literature. First, the precipitation of hydration products can con-
stitute a partial or complete layer that hampers the release of ions into
solution. In this case, the dissolving reactive interface is modified. The
interface extent between C3S and water is lower in the case of a perfect
contact between the anhydrous and hydrated phases. Second, a gap
might be assumed between the dissolving surface and the hydration
products. In this case, the saturation indexes would be significantly dif-
ferent from those in the bulk solution.
Nevertheless, there are strong arguments that indicate that the pre-
cipitation of hydration products or the diffusion of ions does not signif-
icantly influence how the hydration progresses, or in other words, it
does not change the overall hydration kinetics. As a matter of fact, C3S
hydration kinetics in paste or in stirred diluted suspensions are compa-
rable [43] or even alike when the hydration is started in saturated
portlandite solution [44]. In paste, the advection of species is only
Fig. 6. Evolution of the saturation indexes with respect to C3S, C–S–H and Ca(OH)2 calculated from the ion concentrations measured during the hydration of a white cement at w/c = 0.5
(A) and during the hydration of a gray cement at w/c = 0.35 (B) [48].
5
ensured by diffusion and in dilute suspensions by diffusion and stirring.
Secondly, the addition of a large amount of seeds prevents C–S–H from
precipitating onto C3S grains. The hydration is accelerated but the hy-
dration curve remains qualitatively the same [45,46]. This observation
will be further discussed in the next section. It seems that the coverage
of the C3S surface by C–S–H does not modify the hydration sequences.
On the basis of these arguments, the composition of the solution at
the interfaces is related to the bulk solution, but not equal (see Fig. 3).
Thus, the degrees of saturation calculated from the measured bulk solu-
tion composition are apparent and the following relations βiC3S b βbulkC3S,
βiC–S–H N βbulkC–S–H and βiCH N βbulkCH hold. Two cases are considered
and depicted in Fig. 3. In (a), no C–S–H layer onto the C3S surface is
formed. This is representative of the very early hydration or when a sig-
nificant amount of seeds is initially added. There is a strong effect of stir-
ring on both effective saturation degrees at the interface but no effect on
the measured one. In (b), a gap exists between the C3S surface and the
C–S–H layer. The stirring has a very limited effect on the effective satu-
ration degrees at both C3S and inner surfaces, but decreases the effective
saturation degree at the outer C–S–H surface. Note that in this case, the
absolute apparent lnβC3S values measured in the bulk solution are
higher than the effective value at the surface. On the contrary, the mea-
sured lnβC– S–H values are smaller than the effective ones.
3. Results from the literature
ð10Þ
The amount of published data that describe the evolution of ion con-
centrations during the hydration of cement (or C3S) is scarce, especially
in paste. In the following, a handful of papers has been carefully exam-
ined. Overall, this study covers the hydration of cement (or C3S) at a
water to cement ratio ranging from 0.35 to 250. Data related to the hy-
dration of cement in paste have been extracted from Lothenbach [47],
Rothstein [48], Deschner [49], Longuet [50], and data related to the hy-
dration of C3S in stirred diluted suspensions from Nicoleau [51] and
Garrault [52]. In this last experiment, the lime concentration in solution
was maintained constant at 11 mM. In order to evaluate how the inter-
facial rates vary during hydration, the undersaturation index with re-
spect to C3S solubility and the supersaturation indexes with respect to
C–S–H and portlandite solubilities have been calculated. The saturation
index, SI, is defined as the logarithm of the ratio between the ion activity
product Π and the solubility product Ksp, as follows:
Π
SI ¼ ln β ¼ ln ¼ lnΠ− ln Ksp : ð11Þ
Ksp
The logarithm of solubility products calculated according to Eq.(11)
at 20 °C can be found in Table 1. The value chosen for C3S is a reasonable
estimation taken from the kinetic study reported by Nicoleau et al. [7].
This estimated value held fixed constant, but should vary slightly
6 L. Nicoleau, A. Nonat / Cement and Concrete Research 86 (2016) 1–11
Fig. 7. Evolution of the saturation indexes with respect to C3S, C–S–H and Ca(OH)2
calculated from the ion concentrations measured during the hydration of a Portland
cement at w/c = 0.5. Data from [50].
according to pH. Yet, at any value of Ksp, the variations of SI and the in-
terfacial rate are only determined by the variation of lnΠ.
3.1. Data from Lothenbach — hydration of alite in ordinary Portland ce-
ment, w/c = 0.5
The first data set evaluated in this paper comes from Lothenbach
et al. [47] who kindly made available the numerical values. In this arti-
cle, the concentration of ions has been measured by ICP optical emission
spectrometry and the degree of alite hydration by QXRD. In Fig. 4, the
degree of reaction is plotted versus time. After smoothing and
derivatizing the curve, the hydration peak appears and shows a maxi-
mum at roughly 6 h. The relatively poor resolution of the degree of hy-
dration curve and the inherent inaccuracy of the QXRD technique do not
allow the precise resolution of the hydration rate evolution. The calcula-
tion of the saturation indexes (Fig. 5) shows that the composition of the
solution approaches the solubility equilibrium of C3S during the first
four hours and then deviates from this equilibrium until the end of ex-
periment. The opposite conclusion can be drawn for C–S–H: during
the first hours the composition of the solution deviates from the C–S–
H solubility and then subsequently approaches it. From the first hour
until the end of the experiment, the supersaturation with respect to
portlandite evolves only slightly.
Fig. 8. Calorimetry curve (A) and evolution of the saturation indexes (B) with respect to C3S, C–S–H and portlandite during the hydration of Portland cement (w/c = 0.5, T° = 23 °C), data
from [49].
3.2. Data from Rothstein — hydration of alite in ordinary Portland and
white cement pastes
The second data set was reported from Rothstein et al. [48]. This set
consists of two hydration experiments. The first experiment is the hy-
dration of a gray cement at w/c = 0.35 and the second one is a white ce-
ment hydrated at w/c = 0.5. Unfortunately, the degree of hydration
over time is not available in the paper. The evolution of the saturation
indexes is reported in Fig. 6A for the white cement and Fig. 6B for the
gray cement. Concerning the white cement, the saturation indexes
slowly increase during the first hours and then the ion concentrations
move towards C–S–H solubility and depart from C3S solubility. The
same tendencies can be highlighted during the hydration of the gray ce-
ment except that the solution gets supersaturated with respect to
portlandite after at least 8 h.
3.3. Data from Longuet — hydration of alite in Portland cement paste
In reference [50], Longuet was the first scientist to extract the pore
solution from cement pastes and to study its composition during the hy-
dration. The experiment, which is shown in Fig. 7, is the evolution of the
saturation indexes of silicate phases and portlandite during the hydra-
tion of Portland cement (noted CPA69 in the article) at w/c = 0.5 and
at 20 °C. The solubility of C3S has been reached at 5 h and from this
point the undersaturation with respect to C3S decreased. This remains
true until one year, the last experimental point of the study.
3.4. Data from Deschner — hydration of alite in Portland cement paste
This data set relates the hydration of ordinary Portland cement at a
water to cement ratio of 0.5 at 23 °C. In Deschner's article [49], a reason-
able estimation of the hydration rate of alite can be made with the cal-
orimetry curve. The evolution of the hydration rate is directly
correlated to the heat flow released and shown in Fig. 8A. The so-
called hydration peak is characterized by an increase of the
undersaturation with respect to C3S and a decrease of the saturation
with respect to C–S–H.
3.5. Nicoleau's data — hydration of C3S at w/c = 250 started in a saturated
Ca(OH)2 solution
This data set describes the hydration of pure C3S in dilute suspension
[51]. The hydration is started in a solution saturated with respect to
portlandite. During the experiment, the concentration of calcium and
silicate ions (Fig. 9) as well as the conductivity have been recorded
 L. Nicoleau, A. Nonat / Cement and Concrete Research 86 (2016) 1–11
Fig. 9. Evolution of calcium and silicate concentrations during the hydration of C3S started
in dilute solution (the liquid to solid ratio is 250) and saturated with respect to portlandite.
The hydration rate is calculated from the conductivity. Portlandite has not precipitated in
the experiment.
and the concentration of hydroxide ions calculated by charge balance.
The increase of conductivity is proportional to the increase of calcium
hydroxide concentration and thus proportional to the degree of hydra-
tion [54]. The hydration rate can therefore be calculated from the con-
ductivity (Fig. 9). It is worthwhile pointing out that during the entire
experiment, contrary to some other experiments presented in this
paper, portlandite does not precipitate. The evolution of saturation in-
dexes with respect to C3S and C–S–H is plotted in Fig. 10 but is less
marked than in the three previous reported experiments. A decrease
of C3S supersaturation is observed at approximately 8 h, which corre-
sponds to the time of the maximum hydration rate (Fig. 9). The super-
saturation with respect to C–S–H also decreases at that time.
3.6. Data from Garrault — hydration of C3S, water to solid ratio of 50 started
in a solution maintained at constant Ca(OH)2 concentration of 11 mmol/L
The last experiment from the literature is C3S hydration carried out
under controlled conditions reported by Garrault in her thesis [52].
The C3S sample is hydrated in 11 mmol/L of calcium hydroxide at a liq-
uid to solid ratio of 50. The concentration is maintained constant during
the entire experiment by means of a dedicated set-up. In particular, the
solution is aspirated through a filter and distilled water is injected when
the conductivity rises above the reference conductivity, i.e. the conduc-
tivity of the 11 mM calcium hydroxide solution. The amount of calcium
ions removed over time is monitored, which allows for calculation of
the degree of hydration. The silicate concentration has been measured
by spectrophotometry. The results are presented in Fig. 11. As the
Fig. 10. Evolution of the undersaturation with respect to C3S and the supersaturation with
respect to C–S–H and portlandite during the experiment shown in Fig. 9.
7
Fig. 11. Evolution of the degree of hydration and hydration rate of C3S. The silicate
concentration is also shown. The hydration is carried out in a solution regulated at
constant calcium hydroxide concentration of 11 mM. The liquid to solid ratio is 50.
calcium hydroxide concentration is fixed, the saturation indexes for
C3S and C–S–H, exhibited in Fig. 12, vary only with the silicate concen-
tration. Thus, the saturation index with respect to portlandite is not
shown. From the beginning until the end of experiment, the silicate con-
centration decreases, or in other words, the solution becomes more and
more undersaturated with respect to C3S, and respectively less and less
supersaturated with respect to C–S–H. At roughly the same time as the
hydration rate peak (at about 2.5 h), a more pronounced concentration
drop occurs.
The path followed over time by the activities of species involved in
the coupling of reactions is commonly called the kinetic path. For the
C3S hydration, this path is usually represented in the silica-lime dia-
gram. The kinetic paths of the reported experiments are plotted in
Fig. 13. This plot reveals the direction followed by the system,
i.e., which reaction in the coupling tends to drive the overall kinetics.
In addition, some considerations can be made on the precipitation of
portlandite. Indeed, the massive precipitation of portlandite is always
accompanied by a decrease of the concentration in calcium hydroxide.
At least, we can identify at which experimental point the portlandite
is certainly present when this data is not provided in the articles exam-
ined here.
All the experiments at low w/c reveal the same behavior: the kinetic
path moves from the left to the right-hand side as a result of the increase
of the activity product {Ca2+}{OH−}2, but only as long as portlandite
does not precipitate. Then it goes from the right to the left-hand side
after the precipitation of portlandite. In the same time, before the
portlandite precipitation, the supersaturation with respect to C–S–H in-
creases and decreases as soon as portlandite precipitates. In Deschner's
experiment, the left to right displacement is not observable because the
portlandite already precipitated before the first experimental point. This
is also what the evolution of the saturation degrees reveals in Figs. 5, 6,
and 8. In Garrault's experiment, the activity product {Ca2+}{OH−}2 is
experimentally fixed at a constant and only the move towards C–S–H
solubility is observed as conversely the increase of the deviation from
the C3S solubility (see Fig. 12). In Nicoleau's experiment, only the left
to right displacement of the kinetic path is observed because at such a
dilution (w/c = 250), the activity product {Ca2 +}{OH−}2 does not
reach the critical supersaturation with respect to portlandite. However,
in that case, the deviation from the C–S–H solubility first increases and
then decreases.
4. Discussion
The data reported above allow ascribing what reaction rate limits
the C3S hydration during the different identified steps and the conse-
quences on the evolution of its reactive surface area.
8 L. Nicoleau, A. Nonat / Cement and Concrete Research 86 (2016) 1–11

being much higher than the latter in the very beginning of hydration. In
this short time window, the hydration is limited by the precipitation
rate. In the more peculiar Garrault's experiment, because the sample
was hydrated in low calcium hydroxide concentration during the entire
hydration, a lot of C–S–H nuclei [18] have been nucleated which pro-
vides enough early CSH surface area to enable the system to immedi-
ately enter a regime limited by the dissolution.

4.3. The acceleratory period (step III)

In all the experiments during which the acceleration of hydration oc-

Fig. 12. Evolution of saturation indexes calculated with respect to C3S and C–S–H
solubilities during the experiment shown in Fig. 11 and reported in [52].
4.1. The deceleratory period (step IV)
In all the experiments, whatever the w/c ratio, for pure C3S or ce-
ment, or whatever the initial lime concentration, the supersaturation
with respect to C–S–H always decreases during the deceleratory period
(step IV). Conversely, the undersaturation with respect to C3S increases.
Clearly, during the step IV the kinetic path is always moving towards the
C–S–H solubility and departs from the C3S solubility. That means the
precipitation instantaneously consumes more ions than the dissolution
releases. In other words, the hydration rate is limited by the dissolution
rate in the deceleratory step.
curs without increase of the calcium hydroxide concentration, the
undersaturation with respect to C3S always increases (hand-in-hand
with a decrease of supersaturation with respect to C–S–H) indicating a
system getting limited by the dissolution reaction. It applies to all the
experiments in paste where the precipitation of portlandite occurs typ-
ically in the beginning of the acceleration and also in dilute suspensions
when the calcium concentration is maintained constant (Garrault's ex-
periment). In Nicoleau's experiment, the undersaturation briefly de-
creases during the first hour and then increases during the
acceleratory step until the peak maximum. Started in saturated lime so-
lution, this experiment does not lead to a large enough C–S–H/solution
interface at the acceleration onset to consume all the ions released by
the dissolution of C3S. This is probably because of the very low amount
of initial C–S–H nuclei in saturated lime solution [52].
The conclusion of these experimental observations is that, in cement
pastes, the dissolution is the limiting reaction during the major part of
the hydration peak. In these circumstances, the hydration peak has to
be shaped by the dissolution process.
4.4. Evolution of the dissolution surface area during steps III and IV

4.2. The induction period (step II)
This period does not necessarily occur, it is not the case in Garrault's
experiment for instance. But if such a period appears, the supersatura-
tion with respect to C–S–H increases and the undersaturation with re-
spect to C3S decreases, because the dissolution releases more ions
than the precipitation can consume. It is certainly due to the difference
in the dissolution and precipitation interfacial surface areas, the former
According to Eq. (11), the hydration rate is equal to the product of Ri
(β), the interfacial rate and I, the reactive interface area for both reac-
tions, i.e. the C3S dissolution and the C–S–H precipitation. Note that it
is also true for portlandite if it precipitates. In the previous paragraph,
the role of C3S dissolution on shaping the hydration peak has been
highlighted. When the dissolution becomes limiting, the
undersaturation and consequently the interfacial dissolution rate con-
tinuously increase, that implies a non-monotonic variation of the C3S/

Fig. 13. Kinetic paths of the various experiments shown in this article are plotted in the Ca(OH)2-SiO2 solubility diagram. The solubilities of the relevant phases, i.e. C3S, C–S–H and
portlandite are represented. Within each path, the arrows indicate the direction of the time evolution. The concentration in calcium hydroxide cannot exceed about 33 mM without
instantaneous precipitation of portlandite. This limit defines a critical Ca(OH)2 activity product range, symbolized by a dashed box, where portlandite must precipitate. 33 mM
corresponds to the right-hand side of the box and to an immediate precipitation, the left-hand side corresponds to a lower metastability of the solution with respect to portlandite and
this latter should precipitate only after a couple of minutes. Within a single experimental data set, the identification of portlandite precipitation is pinned with a bigger dot. This point
corresponds to an inversion of the path direction.
 L. Nicoleau, A. Nonat / Cement and Concrete Research 86 (2016) 1–11 9

Fig. 14. On the top-left corner, evolution of the heat flow during the hydration of alite in the presence of 2% magnesium chloride at a water to solid ratio of 0.5. Cryo-SEM investigation has
been performed and samples studied after 30 min ①, 2 h ② and finally 4 h ③. Some representative images ①, ② and ③ highlight the precipitation of C–S–H on alite grains. In ① some C–S–
H is already visible on the surface of the grain; amorphous brucite Mg(OH)2 has also precipitated. In ②, the height of the C–S–H layer is at least 180 nm, and in ③ about 600 nm.

solution reactive interface which has first to increase and then to de-
crease as a result of the increase and decrease of the hydration rate.
The increase of the dissolution surface area can appear surprising at
first sight, since a monotonic decrease of the surface area is expected
when an isotropic dissolution of grains is assumed. Yet, the dissolution
of C3S is characterized by the formation of etch-pits and it has been pro-
posed that the extension of pits and their subsequent coalescence leads
to non-monotonic variation of the surface area of dissolution [9]. A re-
cent NMR study supports the non-monotonic variation of the C3S sur-
face area and the role of etching [55]. The decrease of the dissolution
reactive interface can also result from the expected decrease of the over-
all grain surface area. It could be argued too that the decrease of the C3S/
solution interface results from the precipitation of C–S–H onto the C3S
surface as postulated in [23]. However, some studies show that alite
grains are nearly covered with hydration products before the decelera-
tion [20]. Another convincing example is shown in Fig. 14 which reports
C3S hydration accelerated by the addition of soluble magnesium salt
(leading to the precipitation of brucite). Cryo-SEM pictures taken during
the acceleratory step III indicate that alite grains are covered by signifi-
cant amounts of C–S–H already at the beginning of hydration, and that a
thick C–S–H layer is even formed before the deceleratory step.
On the contrary, when C3S grain surfaces are quasi-devoid of C–S–H
as it is the case in the presence of large amounts of C–S–H nuclei [9,56]
the typical shape of the hydration rate curve is maintained (Fig. 15)
whatever their amount. The nuclei lead to an increase of the C–S–H/so-
lution interface area allowing the decrease of the supersaturation with
respect to C–S–H and the increase of the undersaturation with respect
to C3S according to scheme (a) in Fig. 3.
4.5. Late hydration
The various elements raised previously converge towards the fact
that alite dissolution is of paramount importance for shaping the hydra-
tion curve, and that C–S–H precipitation is never the limiting step right
after the end of the acceleratory step. The relative importance of both
reactions is still to be determined for the late hydration. It could be con-
ceivable that, at relatively low w/c, C–S–H could be constrained in space,
and then its precipitation would be even more limiting than the C3S dis-
solution rate. In such a case, due to the crystallization pressure, the su-
persaturation would be expected to increase. Yet, Rothstein's,
Lothenbach's and Longuet's data do not show any move back towards
C3S solubility or conversely a deviation from C–S–H solubility even
after days or years of hydration. It indicates that C–S–H precipitation
would never be limiting the hydration, and that the C3S dissolution
would be always the limiting kinetic step in the coupling. This conclu-
sion is not contradicted by the similar apparent activation energies re-
cently measured for the sole alite dissolution (~ 48 kJ/mol) [57], or
calculated for the sole alite dissolution (~53 kJ/mol) [58] and measured
for the overall hydration (~51 kJ/mol) [59].
It has to be also mentioned that Thomas clearly demonstrated in this
last article that the apparent activation energy of C3S hydration is con-
stant from the earliest age of hydration until at least 70% of hydration,
which tends to indicate that the same process always kinetically con-
trols the hydration. The fact that Thomas' value is very close to Juilland
et al.'s value, which was obtained for pure dissolution experiments fur-
ther supports our conclusions that the C3S dissolution becomes very
early the slowest step of hydration. Finally, it seems appropriate to
10 L. Nicoleau, A. Nonat / Cement and Concrete Research 86 (2016) 1–11
Fig. 15. Hydration of alite at water to solid ratio of 0.5 with additions of calcium silicate
hydrate nuclei. ξat peak, the degree of hydration reached at the maximum rate is
indicated beside the legend.
highlight another conclusion that arose in Thomas' article. Indeed, as
diffusion processes are rather insensitive to temperature, the activation
energies for diffusion are normally low. Since no significant decrease of
the apparent activation energy is seen, at least until 70% of degree of hy-
dration, it shows that diffusion is not controlling the hydration rate.
5. Conclusion
In this paper, the evaluation of various pieces of published data en-
abled a kinetic analysis of the chemical coupling established between
C3S dissolution and C–S–H precipitation during C3S hydration. In partic-
ular, based upon well-established principles, the evolution of ion con-
centrations indicates what limits the hydration kinetically. In the case
of C3S hydration, the deceleration start delineating the transition be-
tween hydration steps III and IV, is always accompanied by a move of
ion concentrations towards C–S–H solubility and a deviation from C3S
solubility. Thus, the deceleration is caused by dissolution process.
During the acceleratory step III, the kinetic analysis also indicates
that the dissolution is limiting and it highlights the necessary develop-
ment of the dissolution reactive interface to explain the acceleration of
hydration. Both the increase of dissolution interface, as well as the sub-
sequent deceleration, can be rationally explained by the formation and
coalescence of etch-pits. This new approach is in stark contrast with
the usual models based on C–S–H growth. Furthermore, experimental
evidence invalidates the role of C–S–H growth in the deceleration of
the hydration. The conclusions drawn in this examination of literature
data seem to be valid for a large binder composition range varying
from pure C3S to gray cement, and for various dilution conditions. In ad-
dition, the precipitation of portlandite is neither required for the accel-
eration nor for the deceleration but always leads to the immediate
limitation by the dissolution reaction. Finally, there is no indication
that the hydration process becomes controlled by any other processes
than C3S dissolution, even at later ages, neither by C–S–H growth nor
by diffusion.
Some new perspectives arise from those statements. First, as the sur-
face dissolution topography contributes to the acceleratory step, the
prominent role of the C–S–H growth process becomes somewhat un-
clear. Indeed, the concept of the autocatalytic growth [60], which stipu-
lates that the C–S–H surface catalyzes precipitation, has never been
experimentally demonstrated and should be a future exciting topic. Sec-
ond, many interpretations about the effects of a large variety of addi-
tives, which influences the hydration, for instance, were mostly
grounded on modifications of the nucleation and growth of C–S–H.
For sure, the effects of these additives on C3S surface dissolution deserve
a closer look.
Acknowledgment
Luc Nicoleau is grateful to Stephen Farrington (BASF Corp. Cleve-
land) and Bruce Christensen (BASF SE) for the careful reading of this
paper. Luc Nicoleau thanks Sebastian Bergold for the stimulating discus-
sion and both authors acknowledge BASF Construction Solutions and
CNRS for the support necessary to carry out this analysis.
References
[1] A.J. Allen, J. Thomas, H.M. Jennings, Composition and density of nanoscale calcium-
silicate-hydrate in cement, Nat. Mater. 6 (2007) 311–316.
[2] A.C.A. Muller, K.L. Scrivener, A.M. Gajewicz, P.J. McDonald, Densification of C–S–H
measured by 1H NMR relaxometry, J. Phys. Chem. C 117 (2013) 403–412.
[3] H. Le Chatelier, Experimental Researches on the Constitution of Hydraulic Mortars,
McGraw Editions, New York, 1905.
[4] J.J. Thomas, J.J. Biernacki, J.W. Bullard, S. Bishnoi, J.S. Dolado, G.W. Scherer, A. Lüttge,
Modeling and simulation of cement hydration kinetics and microstructure develop-
ment, Cem. Concr. Res. 41 (12) (2011) 1257–1278.
[5] 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
(12) (2011) 1208–1223.
[6] K.L. Scrivener, A. Nonat, Hydration of cementitious materials, present and future,
Cem. Concr. Res. 41(12), 651–665.
[7] L. Nicoleau, A. Nonat, D. Perrey, The di- and tricalcium silicate dissolutions, Cem.
Concr. Res. 47 (2013) 14–30.
[8] L. Nicoleau, E. Schreiner, A. Nonat, Ion-specific effects influencing the dissolution of
tricalcium silicate, Cem. Concr. Res. 59 (2014) 118–138L.
[9] M.A.B. Nicoleau, Analytical model for the alite (C3S) dissolution topography, J. Am.
Ceram. Soc. (2015), http://dx.doi.org/10.1111/jace.13647 (published online).
[10] H.N. Stein, J.M. levels, , Influence of silica on the hydration of 3CaO, SiO2, J. Appl.
Chem. 14 (1964) 338–346.
[11] E.M. Gartner, H.M. Jennings, Thermodynamics of calcium silicate hydrates and their
solutions, J. Am. Ceram. Soc. 70 (10) (1987) 743–749.
[12] F. Begarin, G. Garrault, A. Nonat, L. Nicoleau, Hydration of lite containing aluminum,
Adv. Appl. Ceram. 110 (2011) 127–130.
[13] M. Avrami, Kinetics of phase change. I. General theory, J. Chem. Phys. 7 (1939)
1103–1112.
[14] M. Avrami, Kinetics of phase change. II. Transformation—time relations for random
distribution of nuclei, J. Chem. Phys. 8 (1940) 212–224.
[15] M. Avrami, Kinetics of phase change. III. Granulation, phase change, and microstruc-
ture, J. Chem. Phys. 9 (1941) 177–184.
[16] J.J. Thomas, A new approach to modeling the nucleation and growth kinetics of
tricalcium silicate hydration, J. Am. Ceram. Soc. 90 (10) (2007) 3282–3288.
[17] S. Bishnoi, K.L. Scrivener, Studying nucleation and growth kinetics of alite hydration
using μic, Cem. Concr. Res. 39 (2009) 849–860.
[18] S. Garrault, A. Nonat, Hydrated layer formation on tricalcium and dicalcium silicate
surfaces: experimental study and numerical simulations, Langmuir 17 (2001)
8131–8138.
[19] K. Ioannidou, R.J.M. Pellenq, E. Del Gado, Controlling local packing and growth in
calcium-silicate-hydrate gels, Soft Matter 20 (2014) 1121–1133.
[20] A. Bazzoni, Study of Early Hydration Mechanisms of Cement by Means of Electron
Microscopy, Thesis of the EPFL (Lausanne), Switzerland, 2014.
[21] M. Etzold, P.J. McDonald, A.F. Routh, Growth of sheets in 3D confinements — a
model for the C–S–H mesostructure, Cem. Concr. Res. 63 (2014) 137–142.
[22] G.W. Scherer, Models of confined growth, Cem. Concr. Res. 42 (2012) 1252–1260.
[23] S. Garrault, L. Nicoleau, A. Nonat, Tricalcium silicate hydration modeling and numer-
ical simulations, Proceedings of the RILEM International Symposium on Concrete
Modeling CONMOD'10, Lausanne, Switzerland June 2010, pp. 22–25, , http://dx.
doi.org/10.13140/RG.2.1.3622.2480.
[24] P. Barret, D. Bertrandie, Fundamental hydration kinetic features of the major cement
constituents: Ca3SiO5 and βCa2SiO4, J. Chim. Phys. 83 (1986) 765–775.
[25] J.W. Bullard, G.W. Scherer, J.J. Thomas, Time dependent driving forces and the kinet-
ics of tricalcium silicate hydration, Cem. Concr. Res. 74 (2015) 26–34.
[26] A.C. Lasaga, J.M. Soler, J. Ganor, T.E. Burch, K.L. Nagy, Chemical weathering rate laws
and global cycles, Geochim. Cosmochim. Acta 58 (10) (1994) 2361–2386.
[27] D. Nahon, E. Merino, Pseudomorphic replacement in tropical weathering: evidence,
geochemical consequences, and kinetic-rheological origin, Am. J. Sci. 297 (1997)
393–417.
[28] C.M. Pina, L. Fernández-Díaz, M. Prieto, A. Putnis, In situ atomic force microscope ob-
servations of a dissolution-crystallisation reaction: the phosgenite-cerussite trans-
formation, Geochim. Cosmochim. Acta 64 (2) (2000) 215–221.
[29] E. Merino, T. Dewers, Implications of replacement for reaction-transport modeling, J.
Hydrol. 209 (1998) 137–146.
[30] I. Prigogine, R. Defay, Chemical Thermodynamics, Longmans, Green and Co., London,
1965.
[31] A.C. Lasaga, Kinetic Theory in Earth Sciences, Princeton Series in Geochem, Princeton
Univ. Press, Princeton, New Jersey, 1998.
[32] V.A. Alekseyev, L.S. Medvedeva, N.I. Prisyagina, S.S. Meshalkin, A.I. Balabin, Change
in the dissolution rates of alkali feldspars as a result of secondary mineral precipita-
tion and approach to equilibrium, Geochim. Cosmochim. Acta 61 (6) (1997)
1125–1142.
 L. Nicoleau, A. Nonat / Cement and Concrete Research 86 (2016) 1–11
[33] C. Zhu, P. Lu, Alkali feldspar dissolution and secondary mineral precipitation in batch
systems: 3. Saturation states of product minerals and reaction paths, Geochim.
Cosmochim. Acta 73 (2009) 3171–3200.
[34] C. Zhu, P. Lu, Z. Zheng, J. Ganor, Alkali feldspar dissolution and secondary mineral
precipitation in batch systems: 4. Numerical modeling of kinetic reaction paths,
Geochim. Cosmochim. Acta 74 (2010) 3963–3983.
[35] U. Ludwig, N.B. Singh, Hydration of hemihydrate of gypsum and its supersaturation,
Cem. Concr. Res. 8 (1978) 291–300.
[36] F. Brandt, D. Borsbach, Bassanite (CaSO4, 0.5 H2O) dissolution and gypsum (CaSO4, 2
H2O) precipitation in the presence of cellulose ethers, J. Cryst. Growth 233 (2001)
837–845.
[37] G. Dumazer, V. Narayan, A. Smith, A. Lemarchand, Modeling gypsum crystallization
on a submicrometric scale, J. Phys. Chem. C 113 (2009) 1189–1195.
[38] D. Daval, I. Martinez, J. Corvisier, N. Findling, B. Goffé, F. Guyot, Carbonation of Ca-
bearing silicates, the case of wollastonite: experimental investigations and kinetic
modeling.
[39] D. Daval, I. Martinez, J.-M. Guigner, R. Hellmann, J. Corvisier, N. Findling, C. Dominici,
B. Goffé, F. Guyot, Mechanism of wollastonite carbonation deduced from micro- to
nanometer length scale observations, Am. Mineral. 94 (2009) 1707–1726.
[40] A.C. Lasaga, Metamorphic reaction rate laws and development of isograds, Mineral.
Mag. 50 (1986) 359–373.
[41] W.M. Murphy, E.H. Oelkers, P.C. Lichtner, Surface reaction versus diffusion control of
mineral dissolution and growth rates in geochemical processes, Chem. Geol. 78
(1989) 357–380.
[42] S.T. Bergold, F. Goetz-Neuenhoeffer, J. Neubauer, Quantitative analysis of C–S–H in
hydrating alite pastes by in-situ XRD, Cem. Concr. Res. 53 (2013) 119–126.
[43] D.M. Kirby, J.J. Biernacki, The effect of water-to-cement ratio on the hydration kinet-
ics of tricalcium silicate cements: testing the two-step hydration hypothesis, Cem.
Concr. Res. 42 (2012) 1147–1156.
[44] D. Damidot, A. Nonat, P. Barret, Kinetics of tricalcium silicate hydration in diluted
suspensions by microcalorimetric measurements, J. Am. Ceram. Soc. 73 (11)
(1990) 3319–3322.
[45] J.J. Thomas, H.M. Jennings, J.J. Chen, Influence of nucleation seeding on the hydration
mechanisms of tricalcium silicate and cement, J. Phys. Chem. C 113 (11) (2009)
4327–4334.
11
[46] L. Nicoleau, T. Gädt, L. Chitu, G. Maier, O. Paris, Oriented aggregation of calcium silicate
hydrate platelets by the use of comb-like copolymers, Soft Matter 9 (2013) 4864–4874.
[47] B. Lothenbach, F. Winnefeld, Thermodynamic modelling of the hydration of
Portland cement, Cem. Concr. Res. 36 (2006) 209–226.
[48] D. Rothstein, J.J. Thomas, B.J. Christensen, H.M. Jennings, Solubility behavior of Ca-,
S-, Al- and Si-bearing solid phases in Portland cement pore solutions as a function
of hydration time, Cem. Concr. Res. 32 (2002) 1663–1671.
[49] F. Deschner, F. Winnefeld, B. Lothenbach, S. Seufert, P. Schwesig, S. Dittrich, F. Goetz-
Neuenhoeffer, J. Neubauer, Hydration of Portland cement with high replacement by
fly ash, Cem. Concr. Res. 42 (10) (2012) 1389–1400.
[50] P. Longuet, L. Burglen, A. Zelwer, La phase liquide du ciment hydraté, Ciments et
bétons, Rev. Mater. Constr. (1973).
[51] L. Nicoleau, Physico-chemical interactions between latex dispersions and cement
phases, Thesis of the University of Burgundy (Dijon), France, 2004.
[52] S. Garrault-Gauffinet, Etude expérimentale et par simulation numérique de la
cinétique de croissance et de la structure des hydrosilicates de calcium, produits
d'hydratation des silicates tricalcique et dicalcique, Thesis of the University of Bur-
gundy (Dijon), France, 1998.
[53] J. Haas, A. Nonat, From C–S–H to C–A–S–H: experimental study and thermodynamic
modelling, Cem. Concr. Res. 68 (2015) 124–138.
[54] P.W. Brown, E. Franz, G. Frohnsdorff, H.F.W. Taylor, Analyses of the aqueous phase
during the early C3S hydration, Cem. Concr. Res. 14 (1984) 257–262.
[55] E. Pustovgar, R.P. Sangodkar, A.S. Andreev, M. Palacios, B.F. Chmelka, R.J. Flatt, J.-B.
d.'’E. de Lacaillerie, Nat. Commun. 7 (2016) 10952.
[56] L. Nicoleau, New calcium silicate hydrate network, Transp. Res. Rec. 2 (2010) 42–51.
[57] P. Juilland, E. Gallucci, Morpho-topological investigation of the mechanisms and ki-
netic regimes of alite dissolution, Cem. Concr. Res. 76 (2015) 180–191.
[58] S.A. Grant, G.E. Boitnott, C.J. Korhonen, R.S. Sletten, Effet of temperature on hydra-
tion kinetics and polymerization of tricalcium silicate in stirred suspensions of
CaO-saturated solutions, Cem. Concr. Res. 36 (2006) 671–677.
[59] J.J. Thomas, The instantaneous apparent activation energy of cement hydration mea-
sured using a novel calorimetry-based method, J. Am. Ceram. Soc. 95 (10) (2012)
3291–3296.
[60] E.M. Gartner, J.M. Gaidis, Hydration Mechanisms I, Materials Science of Concrete,
The American Ceramic Society, Westerville, 1989.