CRITICAL

REVIEWS
Francesca Ridi
Dipartimento di Chimica
& CSGI
Università di Firenze
ridi@csgi.unifi.it

HYDRATION OF CEMENT:
STILL A LOT TO BE UNDERSTOOD
Despite the millenary history of the cement, several open questions still arise on the physico-chemical mechanisms underlying
its hydration process. This short review addresses some of the most interesting open issues, indicating the main developing
strategies for the investigation of this process.

A little bit of history themselves at the beginning were used to do) with volcanic ashes.
“There is […] a kind of powder which from natural causes produces According to the modern knowledge, a siliceous and aluminous mate-
astonishing results. It is found in the neighborhood of Baiae and in the rial in itself does not possess cementitious value, unless it is calcinated
country belonging to the towns round about Mount Vesuvius. This sub- and converted in an amorphous form. This calcination process, artifi-
stance, when mixed with lime and rubble, not only lends strength to cially produced in the modern times, was naturally made by the volcano
buildings of other kinds, but even when piers of it are constructed in the for the Romans.
sea, they set hard under water.” (Marcus Vitruvius Pollio, Liber II, De Nowadays, concrete is the synthetic material with the largest produc-
Architectura, ~25 BC). tion on Earth: more than 11 billion metric tons are consumed every year
In these words the Roman engineer Vitruvius firstly described the prop- all over the world. It is also one of the most complex inorganic systems.
erties of a material with surprising properties: a mixture of lime and After more than a century of systematic studies, basic questions are still
crushed volcanic ashes was able to set under water, the resistance unsolved regarding its internal structure over the nanometer to macro-
being increased along the time, in a way completely different to any scopic scale range, on its effects on concrete behavior, on the chemi-
other material. This “magic” material was called pozzolanic from Poz- cal and the physico-chemical mechanisms involved in the hydration
zuoli, the place near Vesuvio where the ashes were taken from. reaction, especially in the presence of organic polymers. Most of these
The reason of the success of the Roman concrete was the substitution questions pertain to the primary hydration product and binding phase
of the usual crushed stones (as Greeks, British people and the Romans of Portland cement paste, the calcium silicate hydrate (C–S–H) gel [1].

110 Aprile ‘10

What “cement” is? cate and aluminate phases with water is an exothermic process. In the
Cement is a complex mixture of inorganic phases, mainly constituted literature five stages of product formation in C3S and cement pastes
by calcium silicates and aluminates. It is produced by heating a mix- are identified based on the heat evolution: a) the initial reaction, due to
ture of limestone and clay above 1450 ºC. A few percent of calcium the early dissolution processes and to the very fast reaction of the alu-

CRITICAL REVIEWS
sulfate are added to the obtained nodules of clinker and finely ground- minates with water; b) the induction period (or dormant period), last-
ed to produce the final cement powder. The major phases in a typical ing few hours, with a very low heat evolution; c) the acceleratory peri-
Portland cement are alite, belite, aluminate and ferrite. Alite is essen- od, in which the nucleation and growth of the calcium silicate hydrate
tially tricalcium silicate (Ca3SiO5, C3S) [2]: it generally constitutes 50- phase and of calcium hydroxide starts occurring rapidly and setting
70% of the total mass. In conventional cements it is the most impor- takes place; d) the deceleratory period; e) a period of slow, continued
tant phase, because it determines the setting time and the short-term reaction, called diffusional period, because the rate of hydration is
strength development. Belite is dicalcium silicate (Ca3SiO4, C2S, usu- determined by the diffusion of reacting species from the solution
ally the β polymorph). It constitutes 15-30% of the whole clinker; the through the hydrated phases, to reach the anhydrous grains.
rate of belite hydration is very slow, thus it substantially contributes to
the long-term strength development. Aluminate (Ca3Al2O6, C3A) and Experimental approaches
ferrite (Ca4Al2Fe2O10, C4AF) are respectively the 5-10% and 5-15% of The kinetics of ce-
typical Portland cements: they chemically react very rapidly with water, ment hydration can
forming a number of metastable hydrated species, that eventually be studied by follow-
transform over the time and convert to a final stable phase showing a ing the variation of
cubic structure (C3AH6). The rate of aluminate reaction is so high that the degree of hydra-
can cause a detrimentally rapid setting, unless a set-controlling agent, tion, α, as a function
as gypsum, is added. of the hydration time
t. A kinetic curve co-
The hydration reaction uld in principle be
When anhydrous cement is mixed with water, a number of exothermic obtained by sum-
chemical reactions take place both simultaneously and successively, ming the amounts of
commonly denoted with the term hydration (schematically shown in Fig. the individual phases
1). In the very first period after the adsorption of water on the surface of that have reacted,
Fig. 2 - Heat flow (blue curve, left axis) and degree of
the dry powder, the dissolution of part of the inorganic phases starts to reaction (green curve, right axis) as a function of time for example by
in the hydration of a C3S/water paste (w/c=0.4)
occur. Very soon, how- quantitative x-ray dif-
ever, new silicate and fraction analysis (QXDA), but, because of the experimental difficulties
aluminate hydrated pha- and the limited precision, other indirect methods are typically used.
ses begin to precipitate The most used method to study the cement hydration kinetics is prob-
from the solution on the ably isothermal calorimetry (IC) [3-4]. The sample is placed in a bath
existing grains, thus fa- with controlled temperature. The exothermic nature of the hydration
voring the further disso- process produces an increase in the sample temperature. The instru-
lution of the anhydrous ment records the temperature difference between the sample and the
phases through an in- bath: this temperature gradient corresponds to the rate of heat evolu-
congruent process. tion. The integration of this quantity gives the total evolved heat (Fig.
The hydrated phase 2). This technique allows both studying the hydration rate and calcu-
responsible for the bind- lating the activation energy. However, after a few days of hydration, the
ing characteristics of the rate of heat evolution becomes too low to be accurately separated
cement is an amor- from the instrumental noise. Furthermore, as this technique requires a
phous calcium silicate continuous use of the calorimeter, very long kinetics, such those pro-
hydrate, called C-S-H, duced by addition of some superplasticizers, can hardly be monitored
having the properties of for a time long enough to detect the acceleration period.
a rigid gel. A secondary Another technique used to study the cement hydration reaction is
product of the hydration quasi-elastic neutron scattering (QENS) [5-11]. A beam of neutrons
process is crystalline impinging on a hydrating cement paste is scattered both elastically
Fig. 1 - Schematic representation Ca(OH)2, portlandite. and inelastically. This scattering is almost totally due to the interaction
of the early stages of C3S hydration The reaction of the sili- of the neutrons with the hydrogens, because of its large incoherent

Aprile ‘10 111

 CRITICAL
REVIEWS
scattering cross section compared to the other elements present in one of the initial products. However it is a very simple and useful
the cement paste. The elastically scattered component has a Gauss- method as it provides a reliable way to determine the hydration curve
ian distribution in energy. The integrated intensity of this component is of a cement paste particularly useful for applied research. One of the
directly proportional to the number of non-mobile hydrogen nuclei, i.e. advantages in respect of IC is the intrinsic possibility to obtain the
to the water chemically bound within the hydrated phases. The inelas- whole hydration kinetics with a very limited use of the instrument: our
tic component exhibits a Lorentzian energy distribution, whose width approach allows having a complete hydration profile up to 28 days
is related to the state of diffusion (translational and rotational motions) using the calorimeter for 6 h only (30 measurements each lasting
of the hydrogen atoms. Thus, QENS enables to monitor simultaneous- about 12 min). Moreover the intrinsic limitation of IC due to the low
ly different water environments evolving during the hydration of a heat evolution after a few days is avoided and very long kinetics (for
cement paste. First, the bound water content vs hydration time is example those of pastes additivated with superplasticizing or retard-
obtained (Fig. 3A). If the stoichiometry of the reaction is known (as in ing polymers) can be easily monitored.
the case of pure tricalcium silicate), this quantity is directly proportion-
al to the hydration degree and thus to the kinetics of the process. The Open issues in the cement chemistry research
dynamics of the water confined in a solid matrix can be described Despite the millenary history, even after a century of systematic scien-
according to the so-called “relaxing-cage model” (a detailed descrip- tific research on the physical and chemical mechanisms underlying its
tion of this topic is beyond the scope of this short review and the hydration process and hundreds of engineering studies focused on
reader is invited to refer to references [12-14]). In this model the understanding the rheology of the pastes and the mechanical proper-
dynamics evolve during the hydration time according to a stretched ties of the cured cementitious samples, several questions are still open
exponential function: the values of the stretched exponent (β) and of in the cement research.
the relaxation time (τ) show that the confining properties of the matrix
increase in time (Fig. 3B-3C). C-S-H microstructure
The big disadvantage of neutron scattering measurements is the very One of the most important open issues concerns the formation and
limited access to neutron sources that makes this technique inapplic- the composition of the cement structure, from the nanoscale up to
able from an industrial point of view. Furthermore, a specific physico- the macroscopic scale. Most of the studies focus on the most impor-
chemical background is required for a correct interpretation of the tant binding phase of Portland cement, the C-S-H gel. Several mod-
experimental results. In the last years, our group proposed an alterna- els have been suggested in order to account for its properties. The
tive method, using Differential Scanning Calorimetry (DSC) to study the first models were based on the structural similarity with naturally
hydration process of cement. The progressive consumption of water occurring minerals, such as tobermorite and jennite [18]. All these
due to the formation of the hydrated phases is monitored: the unreact- models implicitly assumed that part of the silicate ions in the miner-
ed water is quantified, after different hydration times, through the inte- als were substituted by OH, in a random manner. Furthermore,
gration of its melting peak and expressed as Free Water Index (FWI) because of the layered structure of tobermorite and jennite, the exis-
(see Fig. 4) [15-17]. It is thus an indirect method, because it allows the tence of interlayer spaces containing strongly absorbed water, was
reconstruction of the kinetic process through the disappearance of postulated. These models have been fairly successful in qualitatively

exponent, β and (C) average relaxation time, τ, at 25 °C. Reproduced with permission from Reference 10. © Institute of Physics (the “Institute”) and IOP Publishing 2006
Fig. 3 - Time evolution of the QENS fitted parameters in the case of the four phases forming an ordinary cement powder (C3S, C2S, C3A, C4AF): (A) bound water (p), (B) stretch

112 Aprile ‘10

 packing densities, known as high density HD C-S-H, and low density
LD C-S-H [23]. The first version of the model (recently referred at as
Colloidal Model-I, CM-I) primarily focused on explaining how the prop-
erties of the material depend on the packing behavior of these basic

CRITICAL REVIEWS
globules [24]. In the description of CM-I model, the Jennings’ effort
was mainly directed to correlate the results obtained for the measure-
ment of cement specific surface area by means of different tech-
niques. As a matter of fact, because of the complexity of cement
microstructure, conflicting experimental results are obtained when dif-
ferent methods are used to probe this feature [25]. In particular, the
major challenge to the measurement of surface area arises from the
very wide length-scale range of the microstructure internal features,
from a few nanometers to tens of micrometers. Furthermore the
cement surface area is an evolving property, depending both on the
hydration time and on the curing conditions.
In the CM-I version of the Jennings’ model, the influence of the glob-
ules’ internal structure on the bulk properties was not explored in
detail. Recently, CM-I has been modified and extended to take into
Fig. 4 - Free Water Index (FWI) vs time curves of C3S/water (∆) and account the smallest porosity of the C-S-H phase associated to the
C3S/water/cellulose ether (◊) obtained by DSC. Water/cement=0.4, T=20 °C
basic globules internal structure [26]. The Colloidal Model-II, CM-II,
explaining the shrinkage behavior and gas sorption properties of represents a significant improvement in the description of cement
cement paste. However they are insufficient to fully address the C-S- microstructure, since it reconciles many controversial data reported in
H properties, especially regarding the viscoelastic response of the literature. In particular, this model gives an exhaustive interpreta-
C–S–H gel to mechanical loading (creep) and the relative humidity tion of the sorption isotherms experiments. According to CM-II, the
changes (drying shrinkage). microstructure of a cement paste can be schematically described as
Over the years the recognition of the colloidal and gel-like properties in Fig. 5: the basic globule is a disk-like object, whose thickness is
of C-S-H has progressively gained more and more importance. The around 4 nm, having a layered internal structure. The water inside the
first studies date back to the work of Powers and Brownyard, in the globule is located both in the interlamellar spaces and in very small
1950s [19, 20]. They described the broad structure of the material cavities (intraglobular pores, IGP), with dimensions around 1 nm. The
basing their model on evidences from total and non-evaporable water packing of these globules produces a porous structure, where two
contents and water vapour sorption isotherms. Their experimental other main populations of pores can be identified: the small gel pores
results suggested that the product of the cement hydration was com- (SGP), with dimensions of 1-3 nm; and the large gel pores (LGP), 3-
posed of solid units having a size of 14 nm, with nanometric gel 12 nm in size. The inclusion of the sub-nanometric porosity in the
pores. The colloidal description of C-S-H has been later included in model justifies most of the experimental evidences, representing a cru-
other models [21] and nowadays it is definitely the most accepted, cial step for the understanding and the control of the relationships
being the most appropriate to address the complex properties of cal- between structure and properties (Fig. 5).
cium silicate hydrate.
A milestone in the development of the colloidal model for the C-S-H
microstructure is the work published by Allen and coworkers in 1987,
in which the microstructure of a hydrating cement paste was moni-
tored by small angle neutron scattering. In that work the authors
assessed the presence of a growing population of 5 nm gel globules,
after the induction time. These globules successively were found to
aggregate in structures with correlation lengths around 40 nm [22]. In
recent years, based on the huge amount of data present in the litera-
ture, Hamlin M. Jennings in a series of papers formulated a clear and
coherent model for the calcium silicate hydrate microstructure. The
basic idea of the model is that the bulk microstructure is formed as a
consequence of the packing of basic globules having peculiar shape Fig. 5 - Shematic representation of the C-S-H microstructure according to the
and internal structure. Clusters of these particles group together in two Jennings’ Colloidal Model II (CM-II)

Aprile ‘10 113

 CRITICAL
REVIEWS
face- interacting” water. In particular, we have shown that during the
first hours of hydration the presence of the solid matrix does not affect
the freezing behavior of the water phase, which crystallizes as hexag-
onal ice, whereas for longer hydration times (more than 8 days) the
water confined inside the SGP porosity (1-3 nm) experiences impor-
tant structural changes, remaining in an amorphous phase also at the
lowest temperature investigated (i.e., -150 °C) (Fig. 6C).
This study provides a significant improvement in the comprehension of
the microstructural properties of cement. Furthermore these results
encourage a deeper investigation of the freezing-thawing effect on the
mechanical properties of the pastes.

Mechanisms regulating the early

hydration steps and the rate of the process
Another open issue in the field of cement chemistry is the description
of the specific mechanisms that control the rate of cement hydration.
Fig. 6 - A) DSC cooling scans of cement paste cured for 5 h and for 1, 5, 11, 28, This problem is particularly relevant for the interpretation of the influ-
45, and 60 days. The curves are offset for clarity. B) NIR spectra acquired from -
150 to +20 °C on cement paste cured for 1 day. C) NIR spectra acquired on ence of chemical agents, such as plasticizers, retarders, accelerators,
cement paste cured for 8 days
on the hydration rate of cement paste [28]. Pure C3S is often used as
An alternative approach to the study of the cement paste microstruc- a model system for Portland cement, because the hydration kinetics
ture evolution is to monitor the characteristics of the confined water and the properties of the hardened paste are quite similar. Several
throughout the hydration process. Very recently our group published a kinetic models are used in the literature to account for this process.
work in which the combined use of low temperature differential scan- There are numerous evidences providing that the early hydration peri-
ning calorimetry (LT-DSC) and low temperature near infrared spec- od occurs by a nucleation and growth process, and kinetic data from
troscopy (LT-NIR) enabled to investigate the properties of the water C3S hydration are frequently fitted to the Avrami-Erove’ev nucleation
confined in a hydrating white cement paste [27]. The progressive con- and growth equation:
finement of the water inside the paste has been followed from a few
hours after the mixing till two months of curing. LT-DSC thermograms α = 1 + αI - exp[-k(t-ti)M] (1)
show, upon cooling, several exothermic peaks in the temperature
range -10 to -42 °C, whose position and area depend on the hydra- where αi is the fraction of C3S reacted (degree of reaction) at the
tion process, as a consequence of the cement microstructure evolu- induction time ti, k is the rate constant and M is the exponent associ-
tion (Fig. 6A). The peaks have been interpreted in terms of Jennings ated with the nucleation type (dimensionality of the product phase,
CM-II for the hydrated calcium silicate (C-S-H) microstructure. In par- type of growth, and nucleation rate). In the literature several papers
ticular the peak around -40 °C is attributed to the water confined in the report the application of this kinetic model in order to describe the
Small Gel Pores (SGP), while the peaks in the region between -15 and effect of the water content [29], of the temperature [30, 15] and the
-30 °C are due to the water in the Large Gel Pores (LGP). After a few influence of organic polymers [16, 17] on the hydration reaction of tri-
days all the unreacted water is confined in the smallest porosity of the calcium silicate. As the nucleation-and-growth processes are temper-
structure, as only the peak at -40 ºC is detected. The LT-NIR spectra ature activated, the rate constants plotted against the inverse temper-
allowed the investigation of the physical state of this confined water ature follow an Arrhenius behavior:
during the evolution of the paste. NIR is an excellent tool to study solid
systems containing confined water, because it is very sensitive to small lnk = lnA - (Ea/RT) (2)
changes in the hydrogen bonds formed by water. The deconvolution
in terms of Gaussian components of the 7000 cm-1 FT-IR band pro- In this way the activation energy Ea of the process can be extracted
vided a detailed characterization of different O-H oscillators popula- and compared between different paste compositions.
tions and their temperature evolution. In the younger samples, where Recently a new model has been proposed to better address the
the water is still able to crystallize, the Gaussian “fingerprint” of the nucleation and growth kinetics of C3S [31]. In this work J.J. Thomas
hexagonal ice, centered at 6080 cm-1, has been evidenced upon cool- focuses on the fact that the nucleation of the hydrated phases in the
ing and compared to data from the literature (Fig. 6B). As the hydra- C3S hydration is essentially a boundary phenomenon. As the Avrami-
tion reaction proceeds, this feature disappears and the spectral Erove’ev equation was actually formulated to account for bulk
deconvolution of the water bands discriminated the fraction of “sur- processes, a new Boundary Nucleation and Growth Model (BNGM)

114 Aprile ‘10

 ics is sensibly retarded: in fact, the kinetics of this paste is well fitted
by the simple version of the model [33].

Modification of the hydration kinetics:

CRITICAL REVIEWS
addition of organic polymers
The modification of the concrete properties by adding small amounts
of additives was a common practice since Roman times. For exam-
ple, Romans used to add horsehair in order to enhance the mechan-
ical resistance of the concrete.
Nowadays several kinds of organic polymers are commonly intro-
duced in cement manufacturing to improve cement mechanical
properties (i.e. workability and durability). These products are classi-
fied according to specific purposes and uses (water content reduc-
tion, setting time acceleration or retardation, fluidity enhancement,
air entrapment, corrosion inhibition, water-proofing, modification of
the viscosity, etc.).
In cement paste casting, sufficient fluidity and workability is needed.
Both these properties can be improved simply by increasing the
amount of water in the paste. Nevertheless, this operation causes
two undesired side effects: a lower mechanical resistance in the
Fig. 7 - Some of the most used chemical additives.
A) Polynaphthalen sulphonic salts; B) Polymelamine sulphonic salts; C) Comb
hardened concrete and a lower resistance to meteorological alter-
polymer with acrylic (R=H) or methacrylic (R=CH3) backbone and polyethoxylic ations. In order to obtain fluid pastes, enhancing the workability and
side chains; D) Methyl HydroxyEthyl Cellulose
controlling the cement porosity, without increasing the water/cement
has been developed. This model was first elaborated for solid phas- ratio, chemical additives are used. These compounds are usually
es transformations occurring preferentially at grain boundary [32] and referred to as plasticizers and superplasticizers (SPs). The use of
has been demonstrated to describe very well the C3S hydration, these water-reducing additives in concrete began in the 1960s and
introducing less parameters than needed in the Avrami model. nowadays is essential in concrete technology and in the field of con-
According to the BNGM, the hydration process depends on two struction. They are organic polyelectrolytes, used to improve the
independent rate constants, associated with two distinct physical workability of mortar and concrete systems for demanding industrial
processes occurring during a boundary-nucleated transformation. applications. The first plasticizer used in industrial scale belonged to
Citing directly the Thomas’ paper “kB describes the rate at which the the class of lignin-sulphonated derivatives (LS) [34-37]. Very soon
nucleated boundary area transforms, while kG describes the rate at other chemical substances with better performances replaced LS
which the non-nucleated ‘‘grains’’ between the boundaries trans- compounds. The first-generation SPs were polynaphtalen-sulphonic
form”. The introduction of these two kinetic constants enables to salts (NSF) and polymelamin-sulphonic salts (MFS) [38-40] (Fig. 7A
include in the fitting also the induction period, demonstrating that this and 7B). The addition of low quantities of NSF or MSF produces high
is not a separate chemical process, but it is simply a period where the fluidity to the pastes: anyway an important drawback of these prod-
overall hydration rate is very low because few nuclei have formed. ucts was the fluidity decrease in time (slump loss).
The refinement of the mathematical methods used to analyze the Because of this effect, the technical advantage connected to their
hydration processes of additives-containing pastes represents utilization was gradually lost in time. In the last years, a second gen-
another interesting development of this topic. As previously said, the eration of SP has been developed. They are comb polymers (or co-
kinetic curves obtained by means of DSC can be monitored for very polymers) with acrylic or methacrylic backbones and polyetoxylic
long times. In this way, thanks to the calorimetric method developed side chains (Fig. 7C). These systems are able to maintain constant in
by our group, it is possible to study systems that are impossible to time the fluidity of the paste [41, 42].
be analyzed by IC. When BNGM model is applied to fit the long The effect of SPs on the hydration of Portland cement paste has
kinetics of pastes additivated with superplasticizers, any kB-kG com- been extensively studied, almost completely from an applicative
bination is found to describe the initial period of the process. For point of view [41-44]. In the last years F. Winnefeld and coworkers
these systems the boundary-nucleation model has to be modified in carried out a systematic work by on the effect of the molecular struc-
order to account for the presence of a real induction period. Interest- ture of comb-shaped SPs on the performance of cementitious sys-
ing enough, the addition of high amounts of a cellulosic polymer tems [41, 45, 46]. The effect of molecular weight, side chains length
does not require the use of a modified BNGM, even if the C3S kinet- and side chain density on workability and on the early hydration

Aprile ‘10 115

 CRITICAL
REVIEWS
mers used in the cement formulations. The use of cellulose-based
polymers is increasing in the cement industry because their use often
fixes many technological problems, mainly due to their interaction
with water. In particular, they have been used as anti-washout or
waterproofing admixtures, for the adhesive mortar production, and
finally as viscosity-modifiers. This latter application is valuable in the
extrusion technology. In the cement industry, the extrusion technique
is mainly used to produce flat shapes with improved resistance to
compression. Extrusion is a plastic-forming process that consists of
forcing a highly viscous plastic mixture through a shaped die. The
material should be fluid enough to be mixed and to pass through the
die, and on the other hand, the extruded specimen should be stiff
enough to be handled without changing in shape or cracking. These
characteristics are industrially obtained by adding cellulosic polymers
to the mixture.
In the last years we studied the effects of one of the best-performing
cellulose additives (methyl hydroxy-ethyl cellulose, MHEC: see Fig.
7D) on the setting process of C3S pastes. Water proton nuclear mag-
a) C3S in pure water, bar=1 µm; b) C3S in water with NSF, bar=3 µm; c) C3S in water
Fig. 8 - SEM micrographs of C3S suspensions (w/c=50), cured at 25 °C for 24h;
netic resonance relaxation experiments showed that cellulosic addi-
with polyacrylic SP, bar=2 µm; d) C3S in water with polycarboxylic SP,
bar=1 µm. Reproduced with permission from [16], 2003 American Chemical Society
tives strongly interact with water and determine its availability to be
consumed in the hydration process. The experimental results clearly
characteristics, as well as on the final microstructure and mechanical show that cellulose additives act as a regulator of water release dur-
properties was investigated in detail. Some important guidelines have ing the whole hydration process [48].
been finally drawn, very important from the industrial point of view, in With the aim to understand the action mechanism of MHEC on the
order to tailor particular molecular architectures for specific applica- major pure phases constituting a typical Portland cement we
tions. Nevertheless, one of the intrinsic difficulties in understanding obtained the hydration kinetics and compared the kinetic parameters
the influence of the SP on cement hydration is the difficulty to have (rate constants, activation energies, and diffusional constants) with
monodisperse polymers: very often their chemical formulae and their those obtained for un-additivated samples. We found that MHEC
molecular weight are not well known, leading sometimes to confus- addition in calcium silicate pastes produces an increase in the induc-
ing and contradictory results. tion time without affecting the nucleation-and-growth period. A less
Despite their extensive utilization, the comprehension of the physico- dense C-S-H gel was deduced from the diffusional constants in the
chemical mechanism of SPs interaction with the cement phases is presence of MHEC. Furthermore, we studied the hydration products
not yet complete. By means of calorimetry we studied the effect of by using thermogravimetry-differential thermal analysis (TG-DTA), X-
SP (NSF, polyacrylic, polycarboxylic polymers) on the kinetic parame- ray diffraction (XRD), and scanning electron microscopy (SEM). We
ters of the hydration process of tricalcium silicate [16]. The kinetic found that, in the case of the aluminous phases, the additive inhibits
analysis of these curves, by means of the Avrami-Erove’ev model, the growth of stable cubic hydrated phases (C3AH6), with the advan-
enabled us to obtain the activation energies for the nucleation-and- tage of the metastable hexagonal phases being formed in the earli-
growth process of the hydrated phase, the M parameter (related to est minutes of hydration.
the morphological characteristics of the growing crystals), and the Despite its millenary story, the interest for this peculiar material from
diffusional constants for pastes cured with and without SPs. The the scientific point of view is still growing: there’s no other inorganic
presence of additives produces a dramatic increment of activation system exhibiting similar evolving properties, resembling in some
energy value for the acceleration period, while the M parameter aspects those of biological systems. Several topics that are still in
remains roughly constant. These results indicate modifications in the progress in the field of the cement chemistry research are not cov-
hydration mechanism, probably related to the morphology of the ered in this contribute (for instance, the study of the forces responsi-
hydrated phase, as evidenced by SEM micrographs (Fig. 8) [47]. ble for the cohesion of cement pastes, or the identification of the rela-
Since the activation energies are directly related to the nucleation tionship between macroscopic properties and nanostructure).
process and, therefore, to the setting, their knowledge allows draw- Because of the complexity of the system, a multiplicity of approach-
ing a scale of additive efficacy, providing for the first time a quantita- es can be pursued. In this short review we selected only some of the
tive approach to the field. most relevant topics currently ongoing in the scientific community,
Cellulose ethers constitute another important class of organic poly- intentionally overlooking some other issues.

116 Aprile ‘10

References Francesca Ridi got the Ph.D in
[1] A.J. Allen et al., Nat. Mater., 2007, 6, 311. Chemistry at the University of Flo-
[2] According to the common cement chemistry notation each rence in 2004. Her present Post

CRITICAL REVIEWS
oxide is abbreviated with its first letter, C: CaO, S: SiO2, H: H2O, Doctoral fellowship focuses on the
A: Al2O3, F: Fe2O3 study of the hydration reaction of
[3] E. Knapen, D. Van Gemert, Cement and concrete Research, cement, on the effect of organic
2008, 39, 6. additives on it and on cement
[4] O. Bonneau et al., Cement and concrete Research, 2000, 30, microstructure, and on the physical
1861. and dynamical properties of water
[5] R. Berliner et al., Cement and concrete Research, 1998, 28, confined in solid matrices.
231.
[6] S. FitzGerald et al., Chem. Mater., 1998, 10, 397. [29] R. Berliner et al., Cement and concrete Research, 1998, 28, 231.
[7] E. Fratini et al., Phys. Rev. E., 2001, 64, 020201. [30] S. FitzGerald et al., Chem. Mater., 1998, 10, 397.
[8] E. Fratini et al., J. Phys. Chem. B., 2002, 106, 158. [31] J.J. Thomas, JACS, 2007, 90, 3282.
[9] E. Fratini et al., Phys. Rev. E., 2002, 65, 010201. [32] J. Cahn, Acta Metallurgica, 1956, 4, 449.
[10] E. Fratini et al., J. Phys.-Condens. Mat., 2006, 18, S2467. [33] F. Ridi et al., in preparation.
[11] S.A. FitzGerald et al., J. Mater. Res., 1999, 14, 1160. [34] O. Stránël, T. Sebök, Cement and concrete Research, 1997,
[12] A. Faraone et al., J. Chem. Phys., 2004, 121, 10843. 2, 185.
[13] L. Liu et al., J. Phys.-Condens. Mat., 2004, 16, S5403. [35] T. Sebök, O. Stránël, Cement and concrete Research, 1999,
[14] L. Liu et al., J. Phys.-Condens. Mat., 2006, 18, S2261-S2284. 29, 591.
[15] A. Damasceni et al., J. Phys. Chem. B., 2002, 106, 11572. [36] O. Stránël, T. Sebök, Cement and concrete Research, 1999,
[16] F. Ridi, et al., J. Phys. Chem. B., 2003, 107, 1056. 29, 1769.
[17] F. Ridi, et al., J. Phys. Chem. B., 2005, 109, 14727. [37] T. Sebök, O. Stránël, Cement and concrete Research, 2000,
[18] H. Taylor, Cement chemistry, Thomas Telford Publishing, 30, 511.
London, 1997. [38] P. Termkhajornkit, T. Nawa, Cement and concrete Research,
[19] T.C. Powers, PCA Bull., 1948, 22, 845-880. 2004, 34, 1017.
[20] T.C. Powers, Natl. Bur. Stds, Washington DC, Proc. 4th ISCC, [39] D. Bonen, S. Sarkar, Cement and concrete Research, 1995,
1962, 2, 577. 25, 1423.
[21] F. Wittmann, in Hydraulic Cement Pastes: Their Structure [40] I. Pirazzoli et al., Magn. Res. Imaging, 2005, 23, 277.
a Properties, Proceedings of a Conference Held at University [41] A. Zingg et al., J. Colloid. Interf. Sci., 2008, 323, 301.
of Sheffield 8.-9.4.1976, Cement And Concrete Association, [42] I. Papayianni et al., Cement and Concrete Composites, 2005,
1976, 96. 27, 217.
[22] A. Allen et al., Phylosophical Magazine B: Structural, [43] J. Gołaszewski, J. Szwabowski, Cement and concrete
Electronic, Optical, and Magnetic Properties, 1987, 56, 263. Research, 2004, 34, 235.
[23] J. Thomas et al.,Cement and concrete Research, 1998, 28, [44] H. Vikan et al., Cement Concrete Res., 2007, 37, 1502.
897. [45] F. Winnefeld et al., Cement Concrete Comp., 2007, 29, 251.
[24] H. Jennings, Cement Concrete Research, 2000, 30, 101. [46] A. Zingg et al., Cement Concrete Comp., 2009, 31, 153.
[25] J. Thomas et al., Concr. Sci. Eng., 1999, 1, 45. [47] F. Ridi, Study of the chemico-physical properties of
[26] H. Jennings, Cement Concrete Research, 2008, 38, 275. cementitious materials and their hydration processes. Effect
[27] F. Ridi et al., J. Phys. Chem. B. 2009, 113, 3080. of superplasticizing and cellulosic additives, Ph.D Thesis, 2004.
[28] J. Thomas et al., J. Phys. Chem. C, 2009, 113, 4327. [48] M. Alesiani et al., J. Phys. Chem. B., 2004, 108, 4869.

RIASSUNTO
Idratazione del cemento: molto ancora da capire
Con una produzione di più di 11 miliardi di tonnellate di calcestruzzo prodotte ogni anno, il cemento è uno dei materiali sintetici più ampiamente utilizzati
nel mondo. Nonostante il suo utilizzo sia noto fin dall’antichità molte domande sono tuttora aperte a riguardo della reazione di idratazione e costituiscono l’argomento
di numerosi progetti di ricerca sia nel campo della scienza dei materiali, che nel campo della chimica-fisica dei sistemi colloidali.

Aprile ‘10 117