See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/281735992

Portland cement and other calcareous hydraulic binders: History, production

and mineralogy

Article  in  European Mineralogical Union Notes in Mineralogy · August 2011

DOI: 10.1180/EMU-notes.9.11

CITATIONS READS

9 1,922

3 authors:

Jan Elsen Gilles Mertens

KU Leuven Qmineral
143 PUBLICATIONS   4,309 CITATIONS    41 PUBLICATIONS   1,702 CITATIONS   

SEE PROFILE SEE PROFILE

Ruben Snellings
Flemish Institute for Technological Research
135 PUBLICATIONS   5,243 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

EU H2020 MSCA-ITN-ETN SULTAN View project

Zeolites and Glass Alteration Products: Formation and Stability in Alkaline Environments View project

All content following this page was uploaded by Jan Elsen on 07 June 2016.

The user has requested enhancement of the downloaded file.

 Portland Cement and other Calcareous Hydraulic Binders
History, Production and Mineralogy

Jan Elsen, Gilles Mertens & Ruben Snellings

Katholieke Universiteit Leuven, Celestijnenlaan 200E, B-3001 Heverlee, Belgium

Introduction

Concrete is the most widely used construction material in the world and as a composite
material of cement and aggregates it is the second most widely consumed natural resource on
earth after water (Humphreys and Mahasenan, 2002). This success is largely due to the
abundance of raw materials for the cement production, the low cost of the material and the
ease of application and high versatility of concrete in construction. Total cement world
production in 2007 was estimated as 2,550, million tons by the U.S. Geological Survey
(January 2008 (http://minerals.usgs.gov/minerals/pubs/commodity /cement/mcs-2008)). The
current yearly output of hydraulic cement is sufficient to make about 2.5 metric tons per year
(t/yr) of concrete for every person on the planet. Interesting to note is that China’s percentage
of the world's production of cement in 2007 was 50 % compared to 42.5 % in 2004 and that
four out of the five largest actual cement producers are European companies. Currently a
number of environmental issues affect the cement industry, especially the carbon dioxide
emissions. The cement industry contributes actually about 5% to global anthropogenic CO2
emissions, CO2 is emitted from the calcination process of limestone and from the fuel
combustion in the cement kiln (Gartner, 2004). The biggest challenge in cement research the
coming years will be the reduction of carbon dioxide emissions.
Hydraulic binders are cementing materials which are distinguished from non-hydraulic
binders by their capability to harden under water. Cementing materials (cements) are calcined
materials that can be defined as adhesive substances capable of uniting fragments to form a
compact whole (Blezard, 1998). This review article is restricted to calcareous hydraulic
cements, being those with lime as one of their main constituents. The European standard EN
197 defines cement as ‘a hydraulic binder, i.e. a finely ground inorganic material which, when
mixed with water, forms a paste which sets and hardens by means of hydration reactions and
processes and which, after hardening, retains its strength and stability even under water’. The
most widely used cement type nowadays is Portland cement, invented and patented by Joseph
Aspdin in 1824 who produced it by burning impure limestones. The sintered product of
burning impure limestones or limestone and an aluminosilicate rock at temperatures up to
about 1450°C is defined as Portland cement ‘clinker’. The powder obtained by grinding the
clinker together with a few percent of gypsum, or with other types of calcium sulphate acting
as an early set retarder, is called ‘cement’. The cement chemist’s shorthand notation will be
used throughout this text for chemical formulae as follows: C=CaO, S=SiO2, A=Al2O3, F=
Fe2O3, M=MgO, K=K2O, N=Na2O, H=H2O. The cement mineral phases can be written as
follows:
Abbreviation Cement Mineral Phase Mineral Formula Mineral Name

C3 S alite Ca3SiO5 hatrurite

C2 S belite Ca2SiO4 larnite
C3 A aluminate phase Ca3Al2O6 /
C4AF ferrite phase Ca2(Al,Fe)2O5 brownmillerite
C2AS gehlenite Ca2Al2SiO7 gehlenite
C2ASH8 gehlenite hydrate Ca2Al2(SiO2)(OH)10·2.25H2O strätlingite
CH portlandite Ca(OH)2 portlandite
C-A-H calcium aluminate hydrate /
C-S-H calcium silicate hydrate /

The microscopical study of Portland cement to investigate the cement minerals started
at the end of the 19th century and Törnebohm (1887) gave the names of alite, belite, celite and
felite to four distinct phases he could observe microscopically (St John et al., 1998). Later
during the course of the 20th century with the use of X-Ray Diffraction techniques, these
different crystalline compounds could be identified and actually Portland cement clinker
contains the following four major mineral phases (idealised) in the following approximate
proportions :

• Alite = tricalcium silicate C3S 50-70 %

• Belite = dicalcium silicate C2S 15-30 %

• Aluminate phase = mainly tricalcium aluminate C3A 5-10%

• Ferrite phase = mainly C4AF 5-15 %

The classification of the different commercial cement types can be made on their
composition or on their performance related properties (Lea, 1998). In Europe, a large
number of countries have been involved in the development of the European standards for
cement at the end of the 20th century. They include the definition of wide range of different
compositions (Table 1). The other major international standardisation of cement type
classification is made through the American Society for Testing and Materials (ASTM).
Table 1- The 27 products in the family of common cements as defined by the European
Standard EN 197-1.

The mineralogy of Portland cement did not change much during the 20th century and
the mineral ratios in Portland cements have been mainly changed in response to a growing
demand for faster development of strength and greater strength overall (Bye,1999). The
principal change has been a gradual increase in the ratio of C3S to C2S. The cement industry
and the cement mineralogy will change more drastically the coming decennia because of the
environmental issues that will affect the cement industry, especially the carbon dioxide
emissions, as mentioned before. Future research for the developments of new, more
sustainable cement products can learn from the development and knowledge obtained during
the 19th century on a wide variety of binder products. Therefore the history and the evolution
of cement and other calcareous hydraulic binders are described in great detail in chapter 2,
especially the rapid evolution of these binders during the 19th century. The production process
of Portland cement clinker and the raw materials used are described in chapter 3, the
mineralogy of the Portland cement mineral phases in chapter 4 and the hydration of these
Portland cement mineral phases in chapter 5. Electron microscopical techniques and
petrographical techniques for investigating cement and concrete will not be described in detail
in this review article.
 History of cement and hydraulic binders

Mortars with different binder types have been used since ancient times for different
applications and their compositional variation is surprisingly large with great differences both
geographically and during different time periods, the evolution in time of the different types
of binders that have been used in Europe is illustrated in Figure 1 (modified from Furlan &
Bissegger, 1975). Initially, the only hydraulic binders made deliberately were those produced
by mixing lime with pozzolanas (Candlot, 1906). The hydraulic properties of these mortars
result from the reaction between the lime and the pozzolanic additives. This practice was
introduced by the Greeks, who used the volcanic tuff from the Island Thera (now called
Santorini) for this purpose (Blezard, 1998). The Romans found a similar material near
Vesuvius and later in other parts of their territories. The pozzolanas which were initially
natural, were partially replaced during Roman times by powdered tiles or pottery. Their
addition may result in similar reactions with lime to form a hydraulic binder.
Hydraulic binders can also be produced from the calcination of impure limestone. If
they are burnt without any additions, we call them natural hydraulic binders. If the limestone
is blended with other components before firing, the product is an artificial binder. Since the
Romans advised using pure limestone for the production of lime, the occurrence of hydraulic
binders is mainly related to the addition of pozzolanas. However, some exceptionally durable
early Greek mortars in which no pozzolanas could be identified, appear to have been
produced from a lesser purity limestone (Klemm, 2004). The hydraulic properties of limes
made from impure limestone seem to have been recognised from the 16th century (Klemm,
2004). However, mortars with definite hydraulic properties from earlier periods, in which no
evidence of pozzolanic additives could be found, have been recognised. It is suggested that
these mortars were made with limes produced from locally available impure limestone that
was burnt under specific conditions (Mertens et al., 2006).
The first important investigations carried out in the field of hydraulic limes made from
impure limestone were those of John Smeaton. His investigations in 1756 led to the discovery
that mortars with limes made from impure limestone gave the best results. By dissolving the
limestones in nitric acid he obtained an insoluble residue of quartz and clayey material to
which he attributed the hydraulicity. At about the same time in Sweden, Bergmann (1735-
1784) attempted to discover why some limes harden when immersed in water. The hydraulic
limes he analysed, all contained manganese. Therefore he attributed their hydraulicity to the
presence of this element. In France, at the beginning of the 19th century, Guyton de Morveau
analysed the properties of an artificial mixture with the same composition as the natural limes
studied by Bergmann. He calcined it and found that the lime obtained was an excellent ‘water
lime’. He also attributed this property to the manganese (4%) and not to the clay, which had
been added in a proportion of 6% to the lime carbonate (Pasley, 1997). In Switzerland,
Saussure (1740-1799) also dissolved ‘meagre limes’ in acid and found that the residue was
composed of quartz and clay. However, he adds that “the manganese would appear to have
greater influence than the siliceous content”. Vitalis (in 1807) among others, found that some
good hydraulic limes contained no manganese, and stated that “clay was the chief source of
their water setting properties” (Pasley, 1997).
The French engineer Collet Descotils was the first (in 1813) to relate clearly the properties of
the meagre limes to the presence of silica. He stated that an intimate combination of silica
with lime is produced when these limes are slaked. He found that the silica in the limestones
was insoluble in acid, whereas it becomes soluble in the lime derived from it. His conclusions
were correct in that a high quality meagre lime must contain a high quantity of finely
disseminated siliceous matter. The hydraulicity of binders is very variable. The first attempt to
classify hydraulic binders was made by Louis Vicat (1818), who introduced the Hydraulicity
Index (HI).

SiO2 + Al 2 O3
HI = (Eq. 1)
CaO

In this formula, Vicat compiled all knowledge then available and directly related the
hydraulicity to the SiO2 and Al2O3 contents. However, equal importance was incorrectly
attributed to the two constituents. Gradually it was found that Fe2O3 and MgO also had an
influence on the hydraulicity. An adapted formula was therefore developed about a century
later by E.C. Eckel (in Eckel, 2005). The formulation of this Cementation Index (CI) is the
following:

2.8 × SiO2 + 1.1 × Al 2 O3 + 0.7 × Fe 2 O3

CI = (Eq. 2)
CaO + 1.4 × MgO

It was conceived to be a direct expression of the quantity of CaO combined with the
other constituents to form hydraulic minerals. The use of this cementation index is based on a
number of assumptions. Firstly, it is assumed that all available SiO2 combines with CaO to
form C3S (Ca3SiO5) and that all Al2O3 combines to form C3A (Ca3Al2O6). MgO is considered
equivalent to CaO and Fe2O3 to Al2O3. This is an oversimplification, since the mineralogy of
hydraulic binders is more complex than assumed here. Eckel emphasized that the properties
of hydraulic binders depend not only on their composition (CI), but also on the conditions of
their manufacture. The hydraulic properties are indeed indirectly related to the burning
temperature and time, since these influence the mineralogy of the product (Mertens et al.,
2006). Mertens et al., (2008) compiled a number of chemical analysis results assembled from
the literature and these are represented graphically in Figure 2.

Fig. 1: The evolution in time since prehistorical times of the different types of binders that
have been used in Europe and their relative importance (modified from Furlan & Bissegger,
1975).
Natural Cement

Quick setting natural cement

At about the time of the first investigations related to the hydraulicity of limes, Parker
discovered that the calcination of septaria found in clay deposits on the Isle of Sheppey (UK),
gave a hydraulic binder. He called it ‘Roman cement’ and obtained a patent for it in 1796.
These septaria were calcined with “a heat nearly sufficient to vitrify them”, and “then reduced
to powder by any mechanical means”. This was the first natural cement. It is also called
‘Parker’s cement’. The name of ‘Roman Cement’ is misleading and was abusively used to
claim the rediscovery of the ancient Roman ‘secrets’ of lime production (Royer, 2004). In
1802, soon after these discoveries, septaria were also found near Boulogne-sur-mer in France
(Royer, 2004). They were used for the production of ‘Plâtre-ciment’ also named ‘Ciment de
Boulogne-sur-mer’ or ‘Ciment romain nouveau de Boulogne’. In other places in France,
analogous hydraulic binders have been produced such as ‘Ciment de Pouilly’ (Pouilly-en-
Auxois) and ‘Ciment de Vassy’ (Vassy-les-Avallon) in the Bourgogne region. The first of
these natural cements was produced from 1827, the second from 1835.
Soon after, other production sites arose, as in the area around Grenoble where ‘natural
cement’ was mainly produced from a homogeneous layer of argillaceous limestone called
‘Filon de Porte de France’. It was discovered in 1842 (Dumolard and Viallet, 1860) and used
from 1846 to produce ‘Ciment Prompt Naturel (CPN)’ or simply ‘Ciment Prompt’. This is a
natural cement where ‘Prompt’ refers to the quick-setting property of the binder. In the
literature (Talansier, 1885; Avenier, 2003) we find setting times of 5 to 10 minutes. The
‘Filon de Porte de France’ appears to be very homogeneous in composition. It is suitable for
the production of ‘ciment prompt’ because of the intimate mixture of argillaceous matter
(24%) and carbonates in the right proportion to give a good natural cement. One of the only
European natural cements still produced today is manufactured in this area. It is the ‘Ciment
Prompt Naturel’ of the company Vicat.
Documents from the 19th century seldom mention burning temperatures, but
temperatures of 1000 to 1100°C are suggested (Wolter, 2005). Most authors simply indicate
that the sintering temperature should not be reached (Dumolard and Viallet, 1860). However,
it is possible that the temperature in some parts of the kiln may reach the sintering point,
thereby producing overburnt lumps. This is not surprising since most of the kilns were filled
by alternating layers of limestone and layers of fuel. Those rocks located in the direct vicinity
of the fuel may reach temperatures high enough for vitrification. Similarly, other rocks can be
underburnt. Depending on the local practice, kilns of natural cement could be intermittent or,
more commonly continuous. The residence time of the limestone in the kiln could last up to
10 or even 15 days (Chantry, 1979). After the calcined product had been drawn from the kiln,
it was reduced mechanically to powder. Afterwards, the powder could be stored in silos. No
slaking by direct contact with water is necessary since only a minor part of these natural
cements consists of free CaO. The storage of the product in silos may be sufficient to slake
this small quantity. A general characteristic of these binders is that they are all quick setting,
which distinguishes them from other natural cements.
In the 19th century chemical analyses could reveal the composition of cements in
terms of their binary components, but not identify the individual minerals (Avenier, 2003).
This is probably the reason for the limited knowledge about cement mineralogy at that time.
According to Eckel (in Eckel, 2005), the only major minerals contributing actively to the
hydraulicity are C3S and C4AF (Ca4Al2Fe2O10). As mentioned previously, he also
erroneously considered MgO to be equivalent to CaO, even though he recognized that “if a
product contains much magnesia its CI may fall below 1 without demonstrable defects in the
cement”. Hereby, he indirectly admitted that MgO may not be identical to CaO in the mineral
formation at higher temperature. Vicat (1818) was the first to recognize that silica combines
with lime during calcination. He also showed that the presence of alumina accelerates the
hydration of a binder. Gradually, the knowledge about cement mineralogy refined. This is
partly due to the application of microscopy to hydraulic binders. In the late 19th century
(1882), Lechatelier considered C3S to be the main constituent of Portland cement clinkers
(Jana, 2005). Boero (1925) indicated that at lower temperatures, the alumina combines with
the decarbonated lime to form C3A. He also mentioned the formation of alumino-ferrites of
lime and the formation of C2S (Ca2SiO4) at higher temperature. In summary, this is probably
a good appreciation of the mineralogy when compared to the composition of presently
produced quick-setting natural cements. The main silicate is C2S. The main aluminates are
C3A and C4AF. Other main components are C3S, calcite and spurrite, but also C12A7 and
C2AS can be identified.

Slow setting natural cement

Like the quick setting natural cements, these slow or half-slow setting binders are produced
by calcining a naturally occurring argillaceous/siliceous limestone that is afterwards ground to
a fine powder. The difference is that these latter binders are burnt at a higher temperature.
Slow setting natural cement is produced from completely vitrified lumps, whereas quick
setting ‘naturel cements’ result from the grinding of non-vitrified but completely
decarbonated lumps. It has been mentioned earlier that the temperature is not the same in all
parts of the kiln and that some over- or underburnt lumps might be present. In some quick
setting natural cements, these lumps are invariably ground together with the main mass of the
lime. In others, the overburnt, vitrified lumps are removed by hand to make a separate cement
(Avenier, 2003). This will result in a slow setting natural cement. The overburnt lumps are
also called ‘grappiers’ and the resulting binder used to be sold as ‘Grappier cement’.
Overburnt lumps may also occur in kilns intended for the production of hydraulic lime. These
grappiers can also be used for the production of ‘grappier cement’. This type of cement was
invented at Teil (France) in 1869 (Boero, 1925). Recent analysis of dark and dense particles
obtained by traditional lime burning indicate the presence of C3S and other cement minerals
in these grappiers (Hughes et al., 2004). Lechatelier considered grappiers in a more general
sense as all particles that are not pulverized upon slaking. He included underburnt rocks,
hydrated and hardened particles, wollastonite, inert materials and also grains of cement
minerals (Candlot, 1906). Only the cement mineral grains actively give strength. However,
they cannot easily be separated from other, mainly inert particles making up the grappier. The
colour of the lumps is an indication of their nature. Black, hard and dense parts have
undergone sintering and are used for the slow setting binders, whereas less hard and less
dense yellow lumps are used for the quick setting binders. Half-slow setting cements are
produced from lumps heated just below the sintering temperature (Candlot, 1906).
This implies that rock from the same source can be used to make quick-, slow- or half-
slow setting cement. A slow setting natural cement can also be produced simply by burning
the right limestone at a higher temperature, which can be achieved by increasing the
proportion of fuel in the kiln. In the beginning, most kilns exclusively intended for the
production of these slow setting binders were intermittent (Candlot, 1906). The use of
continuous kilns was not common, since sintered lumps sticking to the refractory stones of the
inner surface of the kiln hindered a smooth transit of combustion air (Avenier, 2003). With
the advent of rotary kilns, this was no longer a problem. In some areas, rocks used for the
production of this type of binder were identical to those used for the production of quick
setting binders (Talansier, 1885; Boero, 1925), the only difference being the burning
temperature. Others indicate that limestones with a somewhat lower CI can also be used as
slow setting natural cement. Eckel (in Eckel, 2005) considered source rocks with a CI
between 1.00 and 1.15 to be suitable only for the production of slow and not for quick setting
natural cements. He argued that only a sufficiently high burning temperature can guarantee
an acceptable low level of free lime. The burning temperature of quick setting binders might
not be high enough to meet this condition. This practical knowledge is consistent with the
current mineralogical knowledge since more C3S is formed at higher temperatures, thereby
increasing the proportion of mineralogically bound CaO to SiO2. At lower temperature C2S,
with a lower ratio of CaO to SiO2, is dominant. Moreover, specific minerals as gehlenite
(C2AS) are less likely to occur in the slow setting binders, since they occur at lower
temperatures (Chantry, 1979, Varas et al., 2005).
For this type of natural cement too, there is a need to use homogeneous source rocks,
with fine and intimately distributed constituent phases, in the right proportions (Eckel, 2005).
The residence time of the rock in the kiln is in the same range as for the quick setting binders
(Chantry, 1979). The cementing properties of these materials have only been recognised
since the arrival of the quick setting natural cements and probably also since the advent of
artificial Portland cement, where the beneficial effect of sintering was first recognised. I.C.
Johnson proposed burning at temperatures sufficiently high to achieve vitrification (Klemm,
2004). He found that cement produced in this way was better, although setting times were
longer (Blezard, 1998).
Mineralogically, these binders are different from the quick setting natural cements in
that they probably contain less C2S and more C3S, due to the higher burning temperature.
Their composition is probably closer to, although not identical with that of artificial Portland
cement. Candlot (1906) mentioned that ‘natural Portland cement’ can show an early increase
in volume. This could be due to the presence of free CaO and would also explain why it is
recommended to store these natural cements in silos or expose them to air before use.
Gypsum can be added to these binders to retard setting times in the same way as in artificial
Portland cement. Also in this binder, the effects of gypsum will be directly proportional to
the percentage of Aluminium (Eckel, 2005). Because of their setting time, some hydraulic
binders are called ‘half-slow Portland cements’. Their setting times vary from 15 minutes to
2 hours (Boero, 1925). The setting time of ordinary ‘slow setting natural cement’ is in the
order of a few hours.

Hydraulic Lime
Hydraulic limes are limes containing enough free lime to be slaked with water and are capable
of setting under water. A minimum amount of lime must be present in the calcined product to
reduce the entire mass to a powder when it is slaked (Eckel, 2005). The combination of the
free lime with water induces an expansion that leads to the disintegration of the freshly burnt
limestone. The maximum amount of free lime is determined by the second condition. If too
much free lime is present, the product will not be able to harden under water. There is a large
range of products complying with this definition. They are most commonly classified
according their chemical composition and more specifically to their CI (equation 2) or HI
(equation 1). Nowadays, also the mechanical strength is used for classification. The term
‘hydraulic lime’ was first suggested by Vicat. ‘Lean lime’ and ‘water lime’ were terms used
previously.
The production of ‘hydraulic lime’ became well-established only in the 18th century.
At Tournai (Belgium) hydraulic limes have been produced at least since 1815. Le Cocq
(Chantry, 1979) mentions the production of lime that is petrified by water. Already in 1767,
Tournai was an important production centre of lime. Hydraulic limes from Europe have been
exported to the US, since their production was mainly focussed on the manufacture of natural
cement and artificial Portland cement in later times (Eckel, 2005). The advantage of
hydraulic lime as compared with Portland cement is their low cost. However, their slow
hardening and lower tensile strength drove back their production at the expense of Portland
cement. Initially, the lime was burnt in intermittent kilns. Continuous kilns have been used
extensively in later times (Chantry, 1979). According to Eckel (in Eckel, 2005), hydraulic
lime is burnt at a higher temperature than non hydraulic lime. However, since the main part
of the rocks is not sintered, the temperature is lower than that of slow setting natural cement.
Eckel (in Eckel, 2005), mainly inspired by Lechatelier, related the hydraulicity to the
formation of C3S. However, when combining lime to silica in the same 1:3 molar ratio, he
still identified some free lime. With current knowledge, this can be explained by the
formation of C2S rather than C3S in hydraulic limes, as suggested by Eckel.
Hydraulic limes can be produced from siliceous/argillaceous limestone. A
homogeneous and fine distribution of the clay or silicon is necessary (Boero, 1925;
Cammerman, 1919). However, it seems to be less important here than for the production of
natural cement (Candlot, 1906). An inhomogeneous distribution may lead to the formation of
free lime of which the presence is disadvantageous in a natural cement, but not (between
certain limits) for a hydraulic lime since it is slaked. Hydraulic limes are slow setting
compared to natural or artificial cements. Setting times range from 1 or 2 days for eminently
hydraulic limes to one month for feebly hydraulic limes (Boero, 1925). Hydraulic limes need
to be slaked. This was done on the production site in different ways. In most cases the lime
was not used directly after slaking but was stored in heaps or silos for a few days to assure
complete hydration of free lime. Hydraulic lime could also be sold as lumps that were to be
slaked by the customer (Eckel, 2005). Nowadays, slaking of hydraulic limes is performed
mechanically.
If natural hydraulic limes were already used for a certain period of time, artificial
hydraulic limes have only been produced since the early 19th century. In 1810, Edgar Dobbs
was granted a patent for burning a dried slurry. Guyton de Morveau too, experimented with
artificial mixtures, but it was Vicat (1818) who clearly described the process of firing an
artificial mixture of clay and lime in the right proportions. He also assisted Brillant and St.
Leger in setting up their artificial lime plant near Paris in 1926 (Vicat, 1837), where pure
chalk and clay are mixed and burnt in a proportion of 5 to 1. Artificial hydraulic lime was
also produced by John in 1818 by burning a mixture of clay and oyster shells (Pasley, 1838).
Also, quick setting natural cement has been produced artificially (Candlot, 1906).

Artificial Cement
The next step in the history of calcareous hydraulic binders was to grind a calcined mixture of
lime and clay. The patent granted to Joseph Aspdin in 1824 describes this process. He called
the product “Portland Cement”. The method involves ‘double kilning’ and was first described
by Vicat (1836). It consists of calcining the limestone/chalk before mixing it with clay and
afterwards burning the mixture. He probably used a low temperature and therefore his
product would have resembled more to a quick setting natural cement. The artificial cement
produced in this first stage of the development of Portland cement is called proto-Portland
cement. Gradually, it was found that the strength of Portland cement was increasing
compared with that of Roman cement (Klemm, 2004). It indicates that the “Portland cement
manufactured by William Aspdin was greatly improved over that first made by his father”
(Klemm, 2004). Blezard (1998) indicates that Aspdin attained vitrification and produced alite
crystals in his kiln. Johnson further developed this meso-Portland cement and was the first to
recognize the importance of clinker formation (Klemm, 2004). The subsequent developments
lead to the production of the normal Portland cement that we know today, with complete
vitrification, fast cooling and finer grinding due to the use of ball mills which had been
introduced around 1880 (Avenier, 2003). The mineralogy of the cements has evolved towards
an enrichment of the clinker in C3S with the addition of gypsum to retard the setting. The first
artificial cements were manufactured in intermittent kilns (Blezard, 1998). Since its
introduction around 1900, the rotary kiln gradually replaced them. Their combined use in the
double kilning process took place at least until 1953 in the Tournai region (Belgium)
(Chantry, 1979). The intermittent kilns were used for the production of lime and the rotary
kilns for a subsequent sintering of the mixture of lime and clay. Burning times have gradually
decreased through time. Quick setting binders have also been produced artificially. Some are
still produced today.

Fig. 2 : Composition of historic calcareous hydraulic binders in the system CaO-SiO2-Al2O3

(Mertens et al., 2008), (lower left section; each mark represents 10%).
 Portland cement clinker production and raw materials

Portland cement clinker is produced by burning a mix of calcium carbonate, for example
limestone or chalk, and an aluminosilicate rock, for example clay or shale, at about 1450 °C.
Portland cement clinker is mainly composed of four components; CaO, SiO2 ,Fe2O3 and
Al2O3, and the phase diagram of the CaO-SiO2-Al2O3 system (Figure 3), which was first
presented by Rankin and Wright (1912), provides the best basis for an understanding of the
formation of Portland cement clinker. This phase diagram relates to atmospheric pressure and
normal relative humidity. Two ternary compounds are stable in this system, gehlenite
(Ca2Al2Si2O7) belonging to the melilite family and probably forming an intermediate mineral
phase during Portland cement clinker production and anorthite (Ca2Al2Si2O8) as the second
one. A detail of the 1500oC isothermal section of this CaO-Al2O3-SiO2 system is presented in
Figure 4 (adapted from Glasser, 1970). An approximate typical bulk composition of raw
material is represented by the point X in this figure with C2S and C3S in equilibrium with a
liquid phase of composition Z (which contains CaO, Al2O3 and SiO2). During cooling C3S
will crystallize first and later on together with C2S with a changing liquid composition.
Equilibrium phase assemblage would consist of a liquid phase, C3S and C2S. However the
final mineral phase composition in Portland clinker production is more complex due to
fractionation processes leading to non-equilibrium conditions.
The production of white cement clinker requires raw materials with a very low iron
content as for example kaolinitic clays. Additional component such as MgO, K2O, Na2O and
SO3 can be present, but their amounts must be limited. The content of MgO is limited to 4-5
% because of the potential formation of periclase, the content of SO3 is limited because of the
potential formation of expansive delayed ettringite or other sulphate minerals and the content
of alkalis is limited because of the potential formation of the alcali-silica-reaction with certain
aggregates.
The selection and blending of raw materials for cement production is quite
straightforward from a geological viewpoint. Limestone/chalk and shale/clay are the major
raw materials and their proportions can be deduced from their chemical composition using the
concept of the Lime Saturation Factor (LSF);

CaO
LSF =
2.80 SiO2 + 1.18 Al2 O3 + 0.65 Fe2O3

The LSF defines the plane in the quaterny CaO-SiO2- Al2O3 - Fe2O3 system corresponding to
100 % lime saturation. If the LSF is higher than 1.0 then free lime CaO will form in the
Portland cement clinker, which must be avoided because through slow reaction with water
free lime causes destructive expansion of hardened concrete similar as with periclase (MgO).
Fig. 3 : Phase diagram of the CaO-SiO2-Al2O3 system originally determined by Rankin and
Wright (1915), modified from Taylor (1990), with encircled C2S, C3S and C3A cement
components.

Fig. 4: Detail of the 1500oC isothermal section of the phase diagram of the CaO-Al2O3-SiO2
system, modified from Glasser (1970), see text.
Fig. 5 : Illustration of the reactions taking place during the burning of the raw materials in the
Portland cement clinker production process, see text, modified from Taylor (1990).

The principal reactions taking place during the burning of the raw materials till about 1450°C
(Figure 5) are the following :
- decomposition of the carbonates i.e. calcination with the following endothermic reaction:
CaCO3 → CaO + CO2
- decomposition of the clay minerals,
- solid state reactions of calcite or the lime (CaO) formed from it with quartz and clay
minerals decomposition products forming belite, aluminate and ferrite. Belite forms from
900°C to 1200°C:

2 CaO + SiO2 → Ca2SiO4

The aluminate and ferrite phases formed may not be identical with the phases formed in the
final product. Liquid formation may start at temperatures above 1250°C or even lower
starting at about 800-900°C depending on the presence of halides or other fluxes. The liquid
phase promotes the reaction between lime and belite to form alite:
CaO + Ca2SiO4 → Ca3SiO5
Melting occurs above about 1338°C, which is the eutectic point in the C-A-S-F system.
During the cooling stage the molten phase will crystallize giving glass, aluminate and ferrite
and polymorphic transitions of alite and belite will occur.
Until recently the quality control of Portland cement clinker was performed mainly
using XRF-analysis techniques, but the last years, an increasing number of cement plants are
using also powder-XRD techniques for quality control (cf infra). When problems arise, other
techniques are used supplementary, mainly reflected light microscopy and BSE-SEM
techniques on polished sections. The identification of cement minerals in Portland cement
clinker using reflected light microscopy requires the use of selective etching techniques. A
review of these techniques can be found in Campbell (1999). The work of Yoshio Ono (1980)
demonstrated the practical use of optical light microscopy in the cement plant to determine
the quality of the clinker by measuring the size and birefringence of the alite crystals and the
size and colour of the belite crystals. Electron Microscopy is regarded as a very valuable
method to investigate the composition of Portland clinker, especially using BSE-imaging
techniques whereby the different mineral phases can be differentiated and quantified by their
grey level (Scrivener & Pratt, 1984; Zhao & Darwin, 1992). TEM, SEM, ESEM, EPMA,.. are
all valuable methods for research on Portland clinker, but these methods and results will not
be summarised in this review article.
In the cement industry, the usual way of quantifying clinker compositions was by
using the Bogue calculation or by means of microscopical methods (Füllmann et al., 2001).
In the Bogue calculation (Bogue, 1929) the mineral composition is indirectly calculated based
on the chemical composition of the mineral mixture. Since a fundamental assumption of the
method is the requirement of thermodynamic equilibrium conditions - that are not usually met
during clinker production (Midgley et al., 1960; Guirado et al., 2000), the results are not
always reliable. Furthermore, the Bogue calculation does not take into account the
incorporation of foreign ions within the structure of the main phases (Füllmann et al., 2001).
Microscopical methods are also used for the determination of cement composition. Both
optical and electron microscopy, often combined with Image Analysis techniques, have
shown to give good results (Scrivener et al., 1987). But microscopy might be more time-
consuming and the results more operator / equipment dependent (Westphal et al., 2002).
Given these shortcomings, the need for an adapted quantification method was obvious.
Early critical results concerning the quantification of mineral phases showed that ‘neither the
Bogue, nor X-ray diffraction methods gave accurate analyses of the cement’ (Aldridge, 1982).
However, efforts have been done since to make X-ray diffraction a suitable technique for the
quantification of cement phases (Redler, 1991; Scrivener et al., 2004; Peterson et al., 2006)
and also for the characterisation of the raw materials used in their production (Füllmann et al.,
2001). Beside the ability to quantify the components and the possibility to account for solid
solutions, one can also distinguish between different polymorphs using X-ray diffraction
(Gobbo et al., 2004). Although a matter of debate, quantitative phase analyses of Portland
cement using the Rietveld method have shown the presence of an amorphous component
(Suherman et al., 2002).
Powder-XRD-techniques are therefore becoming increasingly used to perform the
quality control of Portland cement clinker (Pöllman, 2002; De la Torre & Aranda, 2003;
Stutzman, 2005; Scrivener et al., 2004; Walenta & Füllmann, 2004; Pritula et al., 2004;
Peterson et al., 2002; Suherman et al., 2002) and hydraulic limes (Gualtieri et al., 2005;
Mertens et al., 2008). This is made possible by the recent development of ultra-high speed X-
ray detectors allowing truly interactive process during quality control in the cement plant and
by the use of robust Rietveld methods to quantify the different mineral phases (Figure 6). To
implement the Rietveld methodology (Young, 1993) it is essential to have the crystallographic
data for all phases present and these data will be described in the following chapter for the
main four mineral phases in Portland cement.
Fig. 6: Powder X-ray diffraction measurement and Rietveld refinement of a CEM-I Portland
cement. The difference between the calculated (not shown) and measured (upper black curve;
CuKα-radiation, graphite monochromator) patterns is shown by the light grey curve. Lower
small scores indicate the position of the reflections for each phase (same order as in legend).
Zincite was used as an internal standard.

Table: Mineral composition of the cement refined in Figure 6

Calculated
(based on
Refined
internal Normalised
standard) to 100%
Zincite 9,93
C3A cub 7,35 8,2 8,2
C4AF 9,74 10,9 10,8
C3S 54,03 60,5 60,0
Β-C2S 12,99 14,5 14,4
Gypsum 0,65 0,7 0,7
Anhydrite 3,34 3,7 3,7
Calcite 1,98 2,2 2,2
Total 100,8 100,0
 Portland Cement Minerals

Alite

The crystal structure of alite exhibits seven polymorphs depending on temperature and
impurities, which have been identified by a combination of DTA, high temperature XRD and
high-temperature light microscopy studies. The following notations were used by Taylor
(1990); T1, T2, T3 for the three triclinic forms, M1, M2 and M3 for the monoclinic and R for
the rhombohedral one. On being heated C3S undergoes the following series of phase
transitions:

620°C 920°C 980°C 990°C 1060°C 1070°C

T1 ↔ T2 ↔ T3 ↔ M1 ↔ M2 ↔ M3 ↔ R

The M1 or M3 form is normally present in Portland cement clinker and T2 rarely due
to the incorporation of impurities. The first determination of the crystal structure was made
by Jeffrey (1952) using rotation, Laue and Weissenberg photographs. The forms now known
as the R, T1 and the M3 appeared to have very similar structures. Later on the structures of
the T1 form (Golovastikov et al., 1975), the M3 form (Nishi et al., 1985) and the R form
(Nishi & Takéuchi, 1984; Il’inets & Malinovski, 1985) were determined more in detail.
Mumme (1995) determined the crystal structure of alite out of a Portland cement clinker and
described a structure similar to the M1 unit cell of Nishi et al., (1985).

- Jeffrey (1952)
R3m; a=0.70, c=2.50; Z= 9.
- Nishi & Takéuchi (1984)
R3m; a=0.7135, c=2.5586; Z= 9.
- Il’inets & Malinovski (1985)
R3m; a=0.70567, c=2.4974; Z= 9.
- Golovastikov et al., (1975)
P-1; a= 1.167, b = 1.424, c = 1.372, α = 105.5°, β = 94.3, γ = 90°, Z= 18.
- Nishi et al., (1985)
Cm ; a = 3.3083, b = 0.7027, c = 1.8499, β = 94.12°, Z = 36.
- Mumme (1995)
Cm ; a = 1.2235, b = 0.7073, c = 0.9298, β = 116.31°, Z = 6.
No structural model is actually available for the T2, T3 and M2 polymorphs. The
known structures are closely similar as regards the positions of the calcium silicon and
oxygen ions but differ markedly in the orientations of the SiO4 tetrahedra. Isolated SiO4
tetrahedra are linked with five- and six-coordinated calcium ions (Figure 7), or seven and
eight-coordinated calcium ions depending on the . A good overview and discussion about the
crystal structure of the known polymorphs of C3S can be found in Dunstetter et al., (2006)
where the T1, M1, M3 and R polymorphs are described in terms of superstructures of three
related elementary blocks; a rhombohedral, a monoclinic and a triclinic unit cell. The C3S
alite corresponds to the nesosilicate mineral hatrurite (Gross, 1977) named after the Hatrurim
Formation in Israel.
Fig. 7. Projection along (001) of the hatrurite crystal structure, corresponding to alite
(Mumme, 1995). Isolated SiO4 tetrahedra are linked with six-coordinated calcium ions. Ca
atoms in light-blue, silicate tetrahedra in dark-blue and oxygen atoms in red.

Optical data for the monoclinic form (Ono et al., 1968);

α = 1.716-1.720 ; γ = 1.722 – 1.724; 2V = 20-60° with negative optical sign;
Birefringence = 0.002-0.010.
The crystals are often idiomorphic approaching hexagons in outline. They are mostly
colourless in transmitted parallel light, a zonation can often be observed but polysynthetic
twinning is rare. In transmitted cross-polarized light, the crystals have a low first-order grey
colour with a negative elongation.

Belite

The crystal structure of belite exhibits five polymorphs at normal pressure, which have been
identified by a combination of DTA and high temperature XRD-studies. The following
notations were used by Taylor (1990) with H = high and L = low;

1425°C 1160°C 630-680°C < 500°C

α ↔ α’H ↔ α’L ↔ β ↔ γ
690°C

780-860°C

As with C3S the higher temperature polymorphs cannot be preserved on cooling unless
stabilized with foreign ions. The structures of α’H, α’L and β polymorphs are derived from
that of α-C2S by progressive decreases in symmetry which take place due to changes in the
orientation of the SiO4 tetrahedra and small movements of the calcium ions. The γ-form is
stable at room temperature but has a much lower density than the other polymorphs and this
causes the clinker to crack and to fall to a more voluminous powder on cooling, which is
known as ‘dusting’. Normally the belite contains enough stabilizing ions to prevent the
transformation to the γ-form during the clinker cooling process. The β-C2S form corresponds
to the nesosilicate mineral larnite, which was first described by Tilley (1929) at Scawt Hill
near Larne in Ireland.
The crystal structure of β-C2S was first determined by Midgley (1952) using
Weissenberg photographs and using the structure of β-K2SO4 as a starting point, which was
suggested by Bredig (1950) using XRD-powder data. The structure of β-C2S was refined by
Cruickshrank (1964) and redetermined by Jost et al., (1977) giving following cell data;

Space Group: P21/n ; Z = 4;

a = 5.502(1) b = 6.745(1) c = 9.297(1) β = 94:59(2)°.

Seven and eight-coordinated calcium atoms (bonds < 0.288 nm) share faces to form
chains parallel to b and c (Jost et al., 1977). These chains are linked by SiO4 tetrahedra
sharing one edge with a calcium polyhedron (Figure 8 with 6-coordination of calcium (bonds
< 0.27nm)).

Fig. 8. Projection along (100) of the larnite crystal structure, corresponding to β-C2S, belite
(Tsurumi et al., 1994). Six-coordinated calcium atoms (bonds < 0.27nm) share faces to form
chains parallel to b and c. These chains are linked by isolated SiO4 tetrahedra sharing one
edge with a calcium polyhedron. Ca atoms in light-blue, silicate tetrahedra in dark-blue and
oxygen atoms in red

The crystal structure of the γ-form is of the olivine type, with octahedrally coordinated
calcium ions.

Optical data for the β-C2S- polymorph (Gille et al., 1965);

α = 1.717 ; β = 1.722 ; γ = 1.736 ; 2V(+) = 64-69 ° .
Using transmitted parallel light, the belite grains are often rounded particles with
marked lamellar texture and with a typical amber-yellow colour. Belite grains with two or
more sets of parallel striations are called type I belite and with only one set of striations are
called type II belite (more rare). The crystals have in transmitted cross-polarized light a first
order white and yellow interference colours. The lamellae of each set of striations in type I
belite grains extinguish simultaneously and are thus not twin lamellae, this in contrast with the
type II belite striations being twin lamellae extinguishing alternately at different angles.

Ferrite

Ferrite was first described by Törnebohm (1987) as one of the major constituents of Portland
cement and was named ‘celite’. Hanson et al., (1928) reported that celite corresponded to the
solid solution Ca2(AlxFe1-x)2O5. This phase can be prepared with any composition where 0 <
x < 0.70. The end member C2A with x = 1 has been prepared but only at a pressure of 2500
MPa (Aggarwal et al., 1972). The natural mineral was first described by Hentschel (1964)
and was named brownmillerite Ca2(Al,Fe)2O5 after Thomas Brownmiller, Chief Chemist of
the Alpha Portland Cement Company-US. The pure iron end member Ca2Fe2O5 was first
described by Chesnokov & Bazhenova (1985) and was named srebodolskite.
Srebodolskite and the compositions up to x ≈ 0.28 crystallize in the space group Pnma
whereas the samples with x > 0.28 crystallize in the space group Ibm2 (Krüger & Kahlenberg,
2005). The basic building units of the two structures are the same (Figure 9, Jupe et al., 2001)
and can be described as perovskite-like layers of corner-sharing (Al,Fe)O6 octahedra
perpendicular to b and chains of corner-sharing (Al,Fe)O4 tetrahedra parallel to c. A 3D
framework is formed by alternating stacking of octahedral layers and sheets of tetrahedral
layers. The basic difference between the two structural types results from different
orientations of the tetrahedral chains. Kahlenberg et al., (1998) pointed to a phase transition at
961 K from an orthorhombic primitive to a body-centred Bravais lattice using high
temperature powder XRD studies. A high temperature single-crystal XRD study of Krüger
and Kahlenberg (2005) shows that Ca2Fe2O5 forms an incommensurable modulated structure
adopting the superspace group Imma(00 γ)s00, with γ = 0.588, with disordered tetrahedral
chains.

Fig. 9. Projection along (100) of the brownmillerite crystal structure, corresponding to ferrite
(Jupe et al. , 2001). The structure can be described as perovskite-like layers of corner-sharing
(Al,Fe)O6 octahedra perpendicular to b and chains of corner-sharing (Al,Fe)O4 tetrahedra
parallel to c. A 3D framework is formed by alternating stacking of octahedral layers and
sheets of tetrahedral layers. Ca atoms in light-blue, FeO4 tetrahedra and FeO6 octahedra in
yellow-ochre and oxygen atoms in red.
Optical data (Ono et al., 1968):
α = 1.96-1.98 ; β = 2.01 -2.05 ; γ = 2.04 – 2.08;
2V = moderate with negative optical sign;
Birefringence = 0.010.
Positive elongation and pleochroic with strongest absorption for γ.

The crystals are coloured normal yellow brown in transmitted parallel light and the
colour changes to green brown with increasing magnesium content. Ferrite occurs as
interstitial material between the silicates and idiomorphic crystals are rare.

Aluminate phase

The structure of C3A was determined by Mondal and Jeffery (1975). The cubic structure with
a = 1.5263 nm, space group Pa3 and Z = 24, consists of rings of six AlO4 tetrahedra (Figure
10). C3A can incorporate Na+ by substitution of Ca2+ with an upper limit of 5.7 %, thus
giving solid solutions with general formula: Na2xCa3-xAl2O6. The substitution occurs without
change in the crystal structure up to a limit of ≈ 1 % Na2O, higher degrees leads to an
orthorhombic structure for 0.16 < x < 0.20 with spacegroup Pbca and to a monoclinic
structure for 0.20< x < 0.25 with space group P21/a (Takéuchi et al., 1980).

A B
Fig. 10. A. Projection along (100) of the crystal structure of the aluminate mineral phase
(Mondal & Jeffery, 1975). The structure consists of rings of six AlO4 tetrahedra, B shows a
projection along (-1-1-1) with only aluminium atoms showing rings of six AlO4 tetrahedra.
Calcium atoms in light-blue, AlO4 tetrahedra in blue and oxygen atoms in red.

Optically, C3A typically forms one component of the interstitial material between the
C3S and C2S crystals. In thin section using transmitted parallel light, crystals are colourless to
brown and mostly isotropic is cross-polarized light.
 Hydration of the Portland cement mineral phases
The term ‘hydration’ in cement and concrete research denotes the totality of changes that
occur when anhydrous cement is mixed with water. This mixture is called a paste.

Hydration of the calcium silicate phases, Nanostructure of C-S-H in hardened Portland

cement

The addition of water to the clinker phases initiates a complex scheme of hydration reactions
to form a hardened cement paste, the essence of the modern construction industry. The main
component of hardened cement pastes is a C-S-H (calcium silicate hydrate) phase, resulting
from the hydration of C3S and C2S. C-S-H constitutes 60-70 % of the hardened cement pastes,
acts as the principal binder component and as such controls the development of strength of the
paste. Early studies by Le Chatelier (1919) and Michaëlis (1909) sparked an intense and long-
lasting debate on the nature and structure of the binder phase in hydrated Portland cement
(PC). Le Chatelier proposed that the main binder product consisted out of very fine
crystalloids of C-S-H phase, while Michaëlis advocated that it was composed of small
colloids. More than a century of productive research later, the advent of increasingly powerful
analytical techniques, most notably solid-state NMR and TEM, significantly enlarged the
knowledge about the nature and nanostructure of the C-S-H phase and prompted the
development of more elaborate structural models. However, several fundamental aspects still
remain a subject of debate, mainly due to the amorphous character, the large variability and
heterogeneity in composition and silicate anion structure of the C-S-H. The amorphous or
semi-amorphous character of the C-S-H in hydrated Portland cement largely prevents
structural studies by regular X-ray diffraction methods. A combination of electron microscopy
and various spectroscopic techniques showed that the average Ca/Si ratio in C-S-H in PC is
approximately 1.75, but can vary significantly within the same paste, even on a µm-scale. The
compositional range in PC is situated between 1.2 and 2.1, but can be lower when
supplementary cementitious materials are added (Richardson, 2008). The hyphens in the C-S-
H structural formula denote the compositional heterogeneity. The compositional variability
shows a bimodal distribution in younger pastes and evolves into a unimodal distribution in
mature pastes. Silicate anions are conceived to be arranged in dimers in young pastes, in older
pastes the paired dimers are observed to be connected by a bridging tetrahedron resulting in
mean chain lengths of 5 (pentamer). Though mean chain length can vary continuously,
individual chains display lengths of 2, 5, 8,… (3n-1). This sequence is typical for finite
“dreierkette” or wollastonite-like chains. It should be noted that minor tetrahedral substituent
ions such as Al3+ can be incorporated into C-S-H only in the bridging tetrahedron (Richardson
and Groves, 1993; Andersen et al., 2006). The absolute prerequisite of a proposed
nanostructure model of C-S-H is that it must encompass all experimental observations. In the
following, the development of the currently most widely accepted models is approached from
a historical perspective.
As early as the ‘50s Bernal (1954) related the structure of C-S-H in PC with that of
tobermorite based on resemblances in the X-ray diffractograms. Tobermorite is a rare,
crystalline, natural mineral which has several subspecies, most notably 1.1 nm tobermorite
(Ca5Si6O16(OH)2·4H2O) and 1.4 nm tobermorite or plombierite (Ca5Si6O16(OH)2·7H2O),
which are distinguished by the differing basal spacing. The structure of 1.1 nm tobermorite
was determined by Megaw and Kelsey (1956), the structure of the 1.4 nm tobermorite was
determined only recently by Bonaccorsi et al. (2005). Both structures consist of a main layer
of edge-sharing calcium polyhedra (6 or 7 fold coordination). On either side of the Ca-O layer
infinite “dreierkette” silicate chains are running parallel with the b-axis. Both paired
tetrahedra share two oxygens with the main layer, the bridging tetrahedron protrudes into the
interlayer space and shares only one oxygen with the main layer. Additional Ca2+ ions and
water molecules are situated in the interlayer. In the 1.1 nm form the bridging tetrahedra link
over the interlayer to form double chains, in 1.4 nm tobermorite the main layers are shifted
over a distance of b/2 and are somewhat further apart, in consequence only single silicate
chains occur (Fig. 11). The 1.4 nm tobermorite species is more similar to C-S-H in terms of
water content and silicate anion structure and was therefore utilized as a starting point of
many later models.

A B

Fig. 11. 1.4 nm tobermorite crystal structure (Bonaccorsi et al., 2005). A. Note the dreierkette
pattern of paired and bridging tetrahedra oriented parallel to the b-axis. B. Projection along
(010), perpendicular to the silicate chains, showing the sharing of oxygens by paired and
bridging tetrahedra with the Ca-O layer. Ca polyhedra in light-blue, silicate tetrahedra in dark-
blue, oxygen atoms in red, water molecules in black, hydroxyls in orange.

Tobermorite shows a Ca/Si ratio of 0.83, inconsistent with the observed ratios in PC
C-S-H, furthermore the silicate anion structure of infinite dreierkette chains did not
correspond to experimental results. Therefore, Taylor and Howison (1956) proposed to raise
the Ca/Si ratio by removing bridging tetrahedra and replacing them by charge balancing Ca2+
ions and protons. In this way, according to their view, a Ca/Si ratio of 1.25 could be reached
when all bridging tetrahedra were omitted and only paired tetrahedra or dimers were left. The
persisting deficiency in Ca2+ led to development of new models where Ca2+ is not only
introduced to charge balance the removal of bridging tetrahedra, yet additional Ca2+ is
introduced in the interlayer coordinated by charge balancing hydroxyls and additional water
molecules (Stade & Wieker, 1980; Stade, 1980), thus raising the Ca/Si ratio up to the level
encountered in PC C-S-H. This model is known as the tobermorite/portlandite (T/CH) model
(Richardson, 2008).
 A different model was proposed by Taylor (1986, 1993). In this tobermorite/jennite
(T/J) model, structurally imperfect layers of 1.4 nm tobermorite are stacked in disorderedly
fashion with structurally imperfect layers of jennite (Ca9Si6O18(OH)6·8H2O; Ca/Si ratio of
1.5). Both structures would be modified by the omission of many bridging tetrahedra.
Furthermore, regions of different nanostructure would be poorly defined and could merge into
each other within individual layers. The essential features of the jennite structure were
originally conceived by Taylor (1968), however, the absence of reliable diffraction data
allowed the structure to be refined only recently (Bonaccorsi et al., 2004). The jennite
structure (Fig. 12) is broadly similar to tobermorite in that it has a layered structure and
contains Ca2+ ions and water molecules in the interlayer. The main layer is composed of
ribbons of edge-sharing calcium octahedra and “dreierkette” silicate chains. The ribbons of
calcium octahedra share vertices to form a corrugated layer in which the silicate chains are
set. In contrast to tobermorite, only half of the oxygens of the calcium octahedra are shared
with the silicate chains, the others form hydroxyl groups. Again, the removal of bridging
tetrahedra and replacement by Ca2+ ions for charge balance would allow for raising the Ca/Si
ratio up to 2.25 for a defect jennite phase consisting solely of dimers of paired tetrahedra.

A B

Fig. 12. Jennite crystal structure (Bonaccorsi et al., 2004). A. Projection along (100), showing
dreierkette silicate chains. B. Projection along (010), perpendicular to the silicate chains. Note
the corrugate Ca-octahedra which constitute the main layers together with the “embedded”
silicate chains. Ca polyhedra in light-blue, silicate tetrahedra in dark-blue, oxygen atoms in
red, water molecules in black. Each oxygen of the main layer not bonded to Si4+ is a hydroxyl.

Richardson and Groves (1992, 1993) put forward a more general model encompassing
both the T/CH and the T/J model by adopting a constitutional approach. The existing models
were extended by allowing for more variable levels of protonation of the silicate chains (i.e.
formation of Si-OH or Si-O-Ca linkages at the extremities of the finite silicate chains). The
incorporation of substituent atoms in the bridging silicate tetrahedra was considered. Most
recent models can be broadly categorized as building on either the T/CH viewpoint (Cong &
Kirkpatrick, 1996a, 1996b; Nonat & Lecoq, 1998; Nonat, 2004) or the T/J viewpoint (Chen et
al., 2004). Despite significant progress in recent years, a consensus is still not reached on the
issue of the existence of one or more distinct nanostructural domains (tobermorite and/or
jennite).
Crystal structures of calcium aluminate hydrates in hardened Portland Cement

Calcium aluminate hydrates are the main hydration products of the C3A clinker phase. In
combination with gypsum, present in Portland cements (PC) as a setting retarder, first an
ettringite phase forms, which is later partly consumed in the formation of a so-called AFm
(Al2O3-Fe2O3-mono) phase. In the following section the compositional and structural features
of these hydrate phases are introduced.

Ettringite
The formation of ettringite hardening and hardened PC is of significant importance. On the
one hand, ettringite will crystallize as primary hydration product of the highly reactive C3A
phase when gypsum is present. The early ettringite formation occurs before the actual
hardening and precipitation of C-S-H phases initiate and in consequence controls the early
strength development or setting of PC. On the other hand, the formation of secondary
ettringite in hardened cement pastes exposed to sulphate attack results in expansion and
deterioration of the concrete.
Ettringite displays hexagonal prisms, often highly elongated along the c-axis. The
structural formula is Ca6[Al(OH)6]2(SO4)3·26H2O. Ettringite is trigonal, the space group is
P31c. The crystal structure is composed of columns of Al(OH)6 octahedra alternating with
triangular groups of edge-sharing 8-fold coordinated CaO8 polyhedra (Moore & Taylor, 1970;
Taylor, 1997). On the outward side of the column, the Ca atoms are each coordinated by four
H2O molecules (Fig.13a). The columns of Ca6[Al(OH)6]2·12H2O are oriented parallel to the c-
axis and set in a hexagonal pattern. The interstitial channels are occupied by sulphate groups
and additional water molecules. The crystal structure of ettringite is relatively flexible,
allowing extensive exchange of the “channel” anions and substitution of Al3+ by similar
trivalent cations. Recent work elucidated the important role of the hydrogen bonding network
in improving the structural cohesion and stability of both the columns in se as the linkages
between the columns (Fig. 13b) (Hartman & Berliner, 2006).

A B
Fig. 13. The ettringite crystal structure (Hartman & Berliner, 2006). A. Projection illustrating
the central column of CaO8 and AlO6 polyhedra. B. Projection along (0001) presenting the
hydrogen bonding network between the Al-Ca columns and the sulphate tetrahedra in the
interstitial space. Ca dodecahedra in light-blue, Al octahedra in blue, sulphate tetrahedra in
yellow, oxygen atoms are in red, hydrogen atoms in dark red.
Sulphate attack on hardened cement pastes may also lead to the formation of expansive
thaumasite, Ca3Si(OH)6(SO4)(CO3)·12H2O, when the C-S-H phase reacts with aqueous
carbonate and sulphate ions. The thaumasite crystal structure is similar to the ettringite
structure, with Si4+ substituting for Al3+ in octahedral coordination and both CO32- and SO42-
anions in the channel sites. Both phases show significant solid solution between the end
members, though a miscibility gap exists between Si:Al 1:1 and 8:1 due to the change in the
space group symmetry from P31c for ettringite to P63 for thaumasite (Barnett et al., 2002).

AFm phases
AFm phases constitute a significant part of the reaction products in a mature PC paste, where
they can be either forming relatively large hexagonal, platy crystals or weakly crystalline, ill-
defined particles intimately mixed with C-S-H phases (Taylor, 1997). The AFm phases are
part of the group of anionic clays or layered double hydroxides (LDHs), the general structure
n-
formula is [Ca 2 (Al, Fe)(OH)6 ]·X1/n ·xH2O , where X is an n-valent anion, the number of water
molecules x is depends on the nature of the X anion. The principal layers are related to the
brucite structure with an ordered arrangement of Ca2+ and Al3+ in a 3:1 ratio. The smaller
ionic radius of Al3+ with respect to Ca2+ distorts the main layer, forcing Ca2+ ions to leave the
central plane and bond to an additional H2O molecule in the interlayer (Fig. 14a). The layered
structure is built by periodical stacking of the positively charged main layers with negatively
charged interlayers comprising anion complexes and water molecules (Fig. 14b) (Rousselot et
al., 2002). The layer thickness, c’, depends on the anion type, in some cases an additional
layer of H2O molecules can be present between the main layers. Also the space group is
determined by the interlayer anion species and moreover by the periodical stacking of the
layers, effectuating polytypism and symmetry lowering. In PC, the interlayer mainly contains
hydroxyls (varying layer thickness and space groups depending on water content; Fischer &
Kuzel, 1982), sulphate (Fig. 14) (layer thickness of 0.893 nm, space group R-3; Allmann,
1977), and carbonate (layer thickness of 0.755 nm, space group P1; Francois et al., 1998).
Many AFm phases may contain more than one anion type, and variations in water content or
anion type are easily accomplished.

A B
Fig. 14. Crystal structure of AFm(SO4) (Allmann, 1977). A. View along (0001), displaying
the characteristics of the principal layer of composition [Ca2Al(OH)6·2H2O]+. Ca2+ ions are
displaced and protrude slightly into the main interlayer to accommodate the bonding of a H2O
molecule. B. View along the c-axis, portraying the layered nature of the AFm(SO4) phase.
The setting and hardening of cement is largely controlled by the hydration reactions of the
calcium-silicates and -aluminates present in the Portland clinker. However, these hydration
mechanisms are still poorly understood (Damasceni et al., 2002; Castellote et al., 2002). The
main difficulties are the limited crystallinity of some hydration products and the fast evolution
of the reactions immediately after the addition of water to the anhydrous cement.

Recently, the latter restraint can be overcome by the use of new techniques or novel
technological developments. Efficient X-ray detectors and/or powerful X-ray sources such as
synchrotron radiation for instance, permit to rapidly generate high accuracy X-ray diffraction
patterns. Particularly, the quick hydration of C3A and C4AF, with or without the addition of
(hydrated) calcium sulphate has been extensively studied using synchrotron radiation (Merlini
et al., 2008; Meller et al., 2004; Jupe et al., 1996). The hydration of both C3A and C4AF in
the absence of calcium sulphate at ambient temperature, proceeds through the formation of
metastable hexagonal lamellar hydrates (C2AH8 and/or C4AH13 or C4AH19) that transform into
the more stable hydrogarnet C3AH6 at longer times of reaction or at higher temperatures (Jupe
et al., 1996; Meller et al., 2004). In the case of C4AF, it is unclear whether the iron is
incorporated into the hexagonal hydrates or if it is only present in an iron-rich residual phase.
Nevertheless, the stable hydrogarnet C3(A,F)H6, may contain considerably amounts of iron
(Meller et al., 2004). A remarkable difference between the hydration of C3A and C4AF is the
rate of C3AH6 formation. It typically forms within minutes in C3A pastes (Jupe et al., 1996),
whereas it may take days in pastes of C4AF. Synchrotron X-ray diffraction data revealed that
the hydration of C3A and C4AF in the presence of calcium sulphate immediately starts with
the formation of ettringite which transforms into Afm at longer reaction times (in the order of
hours), higher temperatures or when the calcium sulphate concentration has decreased
considerably (Christensen et al., 2004; Meller et al., 2004). Recent studies of calcium
aluminate hydration using neutron powder diffraction (Hartman and Berliner, 2005) yield
qualitatively comparable results.
The hydration of the calcium silicate-phases is slower and yields a poorly crystalline C-S-H
phase and portlandite as hydration products. The kinetics of the hydration are generally
determined by measuring the decrease in the water content, the decrease in the amount of
calcium silicate and/or the increase in the portlandite concentration. Differential Scanning
Calorimetry (DSC) for instance was used to determine the fraction of water in a C3S-water
paste that could solidify and melt at different reaction times (Damasceni et al., 2002). An
activation energy of about 30.5kJ/mol was obtained from these data and a two step reaction
mechanism could be identified. The first kinetic stage, after the pre-induction and induction
period, obeys the Avrami-Erofeev law for nucleation and growth (accelaration period),
whereas the second stage is consistent with a three-dimensional diffusion process
(deceleration period). Recently mainly in situ measurement techniques, as quasi-elastic
neutron scattering (Thomas et al., 2001; Fratini et al., 2002) or Raman spectroscopy (Tarrida
et al., 1995) for instance, are applied to determine the amounts of reacting material and hence
the kinetics. Nevertheless, reaction rates and activation energies for C3S measured by other
techniques, whether in situ or not, are all in the range from 30 to 50kJ/mol (Thomas and
Jennings, 1999). In situ techniques are also used to monitor the evolution of the reaction
products. As these are generally of low crystallinity, spectroscopic techniques, focussing on
the short range, as solid-state Nuclear Magnetic Resonance (NMR) generally yield most
interesting results. An in-situ NMR-study of the hydration of 29Si-enriched C3S pointed out
that the bulk hydrate initially formed is mainly dimeric. At later stages of the reaction,
polymerisation occurs, most likely by the insertion of bridging tetrahedra (Brough et al.,
1994).
The hydration of C2S is comparable to that of C3S and yields similar hydration products
(Odler, 2001). Nevertheless, the process progresses more slowly and the induction period is
largely extended compared to that in C3S-water systems. Moreover, compositional and
structural differences of the C-S-H phase formed in C2S-water pastes compared to that in C3S-
water pastes are reported (Tong et al., 1991; Mohan & Taylor, 1981).

Not only the hydration of individual clinker phases, but also that of the hydraulic binders
which are mixtures of calcium-silicates and -aluminates, are studied using fast in-situ
measurement techniques. Neutron diffraction for instance is one of these non-disturbing
analytical methods with short acquisition times (Castellote et al., 2002). The advantage of the
technique is the use of large samples, which are considered to be more representative
compared to the small samples used in X-ray diffraction measurements (Clark and Barnes,
1995). However, a clear disadvantage is the prescribed use of D2O instead of H2O, which is
known to have an influence on the hydration kinetics (Thomas and Jennings, 1999).
Moreover, in a comparative study of neutron diffraction and synchrotron X-ray powder
diffraction, the latter method turned out to have a higher precision (Clark and Barnes, 1995).
Recent synchrotron X-ray powder diffraction data of hydrating Ordinary Portland
Cement (OPC) indicate that ettringite forms immediately after the addition of water to the
cement (Merlini et al., 2007a). A continuous increase in the ettringite content is noted during
the overall OPC's induction period. During the acceleration stage, the rate of C3A dissolution
and hence ettringite formation increases; C3S dissolution becomes evident and noticeable
amounts of C-S-H start to form. As for the pure phase, initial C3S hydration in OPC can be
fitted using the Avrami nucleation and growth model (from 2 to 7 hours), whereas diffusion is
rate-limiting subsequently. In the study of oil-well cements, hydration reactions at higher
temperature were followed up successfully with synchrotron X-ray powder diffraction
(Colston et al., 1998). Moreover, the technique is increasingly used to identify the role of
various admixtures (Merlini et al., 2007a; 2007b; Jupe et al., 2007). Also for the study of
other binders than OPC, synchrotron X-ray radiation is used (De la Torre et al., 2009;
Snellings et al., 2009). However, as laboratory X-ray diffractometers are getting increasingly
efficient, in-situ studies on cement hydration become feasible. This is confirmed by recent
research (Hesse et al., 2009; Scrivener et al., 2004; Christensen et al., 2003), although it is
known that counting statistics of the synchrotron data are far superior (Christensen et al.,
2003).

Conclusions
The success of concrete, which is the most widely used construction material in the
world, is largely due to the abundance of raw materials for the cement production, the low
cost of the material and the ease of application and high versatility of concrete in
construction. Total cement world production in 2007 was estimated as 2,550, million tons by
the U.S. Geological Survey. Our knowledge of the mineralogy of Portland cement has
expanded considerably during the last decennia, mostly by the use of advanced instrumental
methods. The knowledge of the hydration mechanisms and the hydration products of the
Portland cement mineral phases also showed a significant progress in recent years but a
generally accepted view of hydration is not reached yet. The Portland cement composition did
not change much during the 20th century, the principal change has been a gradual increase in
the ratio of C3S to C2S. The cement industry and the cement mineralogy will change more
drastically the coming decennia because of the environmental issues that will affect the
cement industry, especially the carbon dioxide emissions. Future research for the
developments of new, more sustainable cement products can learn from the development and
knowledge obtained during the 19th century on a wide variety of binder products. The biggest
challenge in cement research the coming years will be the reduction of carbon dioxide
emissions.

References
Aggarwal, P.S., Gard, J. A., Glasser, F. P. & Biggar, G. M. (1972): Synthesis and properties
of dicalcium aluminate, 2CaO·Al2O3. Cement and Concrete Research, 2: 291-297.
Aldridge, L.P. (1982): Accuracy and precision of phase analysis in portland cement by Bogue,
microscopic and X-ray diffraction methods. Cement and Concrete Research, 12, 381-
398.
Allmann, R. (1977): Refinement of the hybrid layer structure (Ca2Al(OH)6)+(0.5SO4·H2O)-.
Neues Jahrb. Mineral. Monatsh., 136-144.
Andersen, M.D., Jakobsen, H.J. & Skibsted, J. (2006): A new aluminium-hydrate species in
hydrated Portland cements characterized by 27Al and 29Si MAS NMR spectroscopy.
Cement and Concrete Research, 36:3-17.
Avenier, C. (2003): Fabrication et utilisation des ciments en Isère au XIXème siècle. Cercle
des Partenaires du Patrimoine.
Barnett, S.J., Macphee, D.E., Lachowski, E.E., Crammond, N.J. (2002): XRD, EDX and IR
analysis of solid solutions between thaumasite and ettringite. Cem. Concr. Res., 32:
719-730.
Berggren, J. (1971): Refinement of the crystal structure of dicalciumferrite, Ca2Fe2O5. Acta
Chem. Scand., 25: 3616-3624.
Bernal, J.D. (1954): The structure of cement hydration compounds. Proc. 3rd Int. Symp.
Chem. Cem., London, 1952, 216-236.
Blezard, R.G. (1998): The history of calcareous cements. In Lea’s chemistry of cement and
concrete. Fourth edition. (eds) Hewlett P.C., Butterworth Heinemann.
Boero, J. (1925): Fabrication et emploi des chaux hydrauliques et ciments. Le matériel des
cimenteries. Deuxième édition. (eds) Béranger, Ch., Librairie polytechnique,
Paris/Liège.
Bogue, R.H. (1929): Calculation of compounds in portland cement. Ind. Eng. Chem. Anal., 1:
192-197.
Bonaccorsi E., Merlino, S. & Taylor, H.F.W. (2004): The crystal structure of jennite,
Ca9Si6O18(OH)6·8H2O. Cement and Concrete Research, 34:1481-1488.
Bonaccorsi, E., Merlino, S. & Kampf, A.R. (2005): The crystal structure of tobermorite 14 Å
(plombierite), a C-S-H phase. J. Am. Ceram. Soc., 88:505-512.
Bredig, M.A. (1950): Polymorphism of calcium orthosilicate. J. Am. Ceram. Soc. 33: 188-
192.
Brough, A.R., Dobson, C.M., Richardson, I.G. & Groves, G.W. (1994): In situ solid-state
NMR studies of Ca3SiO5: hydration at room temperature and at elevated temperature
using 29Si enrichment. Journal of Materials Science, 29, 3926-3940.
Bye, G.C. (1999) Portland Cement (2nd edition), Thomas Telford Publishing, London.
Cammerman, C. (1919): Le gisement calcaire et l’industrie Chaufournière du Tournaisis.
Revue Universelle des Mines 6 (Tome II): 1-61.
Campbell, D.H. (1999): Microscopical Examination and Interpretation of Portland Cement
and Clinker. Portland Cement Association, Skokie, Illinois, USA, 224p.
Candlot, E. (1891): Ciments et chaux hydrauliques. Fabrication, propriétés, emplois. París,
Librairie Polytechnique, Baudry Et Cie, Éditeurs.
Candlot, E. (1906): Ciments et chaux hydrauliques: Fabrication - Propriétés – Emploi.
Troisième édition. (ed) Béranger Ch., Librairie polytechnique, Paris/Liège.
Castellote, M., Alonso, C., Andrade, C., Campo, J. & Turrillas, X. (2002): In situ hydration of
Portland cement monitored by neutron diffraction. Applied Physics A, 74, s1224-
s1226.
Chantry, F. (1979): Les cent chaufours d’Antoing à Tournai. Section archéologie industrielle
de la SRHAT., 320 p.
Chesnokov, B.V. & Bazhenova, L.F. (1985): Srebrodolskite Ca2Fe2O5 – a new mineral. Zap.
Vses. Mineral. Obshch., 114: 195–199 (in Russian).
Chen, J.J., Thomas, J.J., Taylor, H.F.W. & Jennings, H.M. (2004): Solubility and structure of
calcium silicate hydrate. Cement and Concrete Research, 34: 1499-1519.
Christensen, A.N., Jensen, T.R. & Hanson, J.C. (2004): Formation of ettringite,
Ca6Al2(SO4)3(OH)12.26H2O, AFt, and monosulfate, Ca4Al2O6(SO4).14H2O, AFm-14,
in hydrothermal hydration of Portland cement and of calcium aluminum oxide -
calcium sulfate dihydrate mixtures studied by in situ synchrotron X-ray powder
diffraction. Journal of Solid State Chemistry, 177: 1944-1951.
Christensen, A.N., Scarlett, N.V.Y., Madsen, I.C., Jensen, T.R. & Hanson, J.C. (2003): Real
time study of cement and clinker phases hydration. Dalton Transaction 2003, 1529-
1536.
Clark, S.M. & Barnes, P. (1995): A comparison of laboratory, synchrotron and neutron
diffraction for the real time study of cement hydration. Cement and Concrete
Research, 25, 639-646.
Colston, S.L., Jacques, S.D.M., Barnes, P., Jupe, A.C. & Hall, C. (1998): In-situ hydration
studies using multi-angle energy-dispersive diffraction. Journal of Synchrotron
Radiation, 5, 112-117.
Cong, X. & Kirkpatrick, R.J. (1996b): 29Si MAS NMR study of the structure of calcium
silicate hydrate. Adv. Cem. Based Mater., 3:144-156.
Cong, X. & Kirkpatrick, R.J. (1996a): 29Si and 17O NMR investigation of the structure of
some crystalline calcium silicate hydrates. Adv. Cem.Based Mater., 3:133-143.
Damasceni, A., Dei, L., Fratini, E., Ridi, F., Chen, S.-H. & Baglioni, P. (2002): A novel
approach based on differential scanning calorimetry applied to the study of tricalcium
silicate hydration kinetics, Journal of Physical Chemistry B, 106, 11572-11578.
De la Torre, A.G. & Aranda, M.A.G. (2003): Accuracy in Rietveld quantitative phase
analysis of Portland cements, J. Appl. Crystallogr. 36: 1169-1176.
De la Torre, A.G., Aranda, M.A.G., Cuberos, A.J.M., Martin-Sedeño, M.-C. & Merlini, M.
(2009): In-situ hydration of activated belite cements studied by synchrotron X-ray
 powder diffraction. Book of abstracts, 11th European Powder Diffraction Conference,
Microsymposium 15, 19-22 September 2009, Warsaw, Poland.
Dumolard, F. & Viallet, C. (1860): Ciment grenoblois exploité et fabriqué à la porte de
France, Grenoble, Imp. Allier.
Dunstetter, F., de Noirfontaine, M.N. & Courtial, M. (2006): Polymorphism of tricalcium
silicate, the major compound of Portland cement clinker 1. Structural data: review and
unified analysis, Cement and Concrete Research, 36 : 39–53.
Eckel, E. (2005): A facsimile of the third (1928) edition. Cements, limes and plasters.
Donhead Publishing Ltd.
Fischer, R. & Kuzel, H.J. (1982): Reinvestigation of the system C4A·nH2O-C4A·CO2·nH2O.
Cement and Concrete Research, 12: 517–526.
Francois, M., Renaudin, G. & Evrard, O. (1998): A Cementitious Compound with
Composition 3CaO.Al2O3.CaCO3.11H2O. Acta Cryst. C54 : 1214-1217.
Fratini, E., Chen, S.-H., Baglioni, P. & Bellissent-Funel, M.C. (2002): Quasi-elestic neutron
scattering study of translational dynamics of hydration water in tricalcium silicate.
Journal of Physical Chemistry B, 106, 158-166.
Füllmann, T., Pöllmann, H., Walenta, G., Gimenez, M., Lauzon, C., Hagopian-Babikian, S.,
Dalrymple, T., & Noon, P. (2001): Analytical methods. International Cement Review
(2001, January): 41-43.
Furlan, V. & Bissegger, P. (1975): Les mortiers anciens, histoire et essais d' analyse
scientifique. Revue suisse d'art et d'archéologie, 32: 2–14.
Gartner, E. (2004): Industrially Interesting Approaches to 'Low-CO2' Cements. Cement and
Concrete Research, 34(9): 1489-1498.
Gille F., Dreizler I., Grade, K., Kämer, H. & Woermann, E. (1965): Mikroskopie des
Zementklinkers, Bilderatlas, Verein Deutscher Zementwerke e. V. (Hrsg.), Beton-
Verlag, Düsseldorf, 75p.
Glasser F.P (1970): Application of the phase rule to cement chemistry. In: Alper A.M. (ed):
Chapter 5 in Refractory Materials, Volume 6-II, Academic Press, New York, 147–190.
Gobbo, L., Sant’Agostino, L. & Garcez, L. (2004): C3A polymorphs related to industrial
clinker alkalies content. Cement and Concrete Research, 34, 657-664.
Golovastikov, N.I., Matveeva, R.G. & Belov, N.V. (1975): Crystal structure of the
tricalcium silicate 3CaO·SiO2=C3S. Sov. Phys. Crystallogr., 20 (4): 441.
Gross, S. (1977): The mineralogy of the Hatrurim Formation. Israel. Geol. Sur. Israel Bull.,
70: 35-36.
Gualtieri, A.F., Viani, A. & Montanari, C. (2006): Quantitative phase analysis of hydraulic
limes using the Rietveld method. Cement and Concrete Research, 36: 401–406.
Guirado, F., Galí, S. & Chinchón, S. (2000): Quantitative Rietveld analysis of aluminous
cement clinker phases. Cement and Concrete Research, 30: 1023-1029.
Hartman, M.R. & Berliner, R.(2006): Investigation of the structure of ettringite by time-of-
flight neutron powder diffraction techniques. Cement and Concrete Research, 36:
364–370
Hartman, M.R. & Berliner, R. (2005): In situ powder diffraction investigation of the hydration
of tricalcium aluminate in the presence of gypsum. Journal of Solid State Chemistry,
178, 3256-3264.
Hentschel, G. (1964): Mayenit, 12CaO 7Al2O3, und Brownmillerit, 2CaO.(Al,Fe)2O3, zwei
neue Minerale in den Kalksteineinschlüssen der Lava des Ettringer Bellerberges.
Neues Jahrb. Mineral., 22-29.
Hesse, C., Goetz-Neunhoeffer, F., Neubauer, J., Braeu, M. & Gaeberlein, P. (2009):
Quantitative in situ X-ray diffraction analysis of early hydration of Portland cement at
defined temperatures. Powder Diffraction, 24, 112-115.
Hughes, J.J., Banfill, Ph., Forster, A., Livesey, P., Nisbet, S., Sagar, J., Swift, D. & Taylor, A.
(2004): Small-scale traditional lime binder and traditional mortar production for
conservation of historic masonry buildings. Proceedings of the International Building
Lime Symposium, Orlando (US), 1-13.
Humphreys, K. & Mahasenan, M. (2002): Climate Change. (Toward a Sustainable Cement
Industry, Substudy 8). Battelle - World Business Council for Sustainable
Development, March.
Il'inets, A.M. & Malinovskii, Y.A. (1985): Crystal structure of the rhombohedral
modification of tricalcium silicate Ca3SiO5. Sov. Phys. Dokl., 30: 191.
Jana, D. (2005): Concrete petrography – past, present and future. Proceedings of the 11th
Euroseminar on Microscopy applied to building materials, Paisly (Scotland), 1-17.
Jeffery, J.W. (1952): The crystal structure of tricalcium silicate. Acta Crystallogr., 5: 26–35.
Jost, K.H., Ziemer, B. & Seydel, R. (1977): Redetermination of the structure of dicalcium
silicate. Acta Crystallogr., B33: 1696-1700.
Jupe, A. C., Cockcroft, J. K., Barnes, P., Colston, S. L., Sankarb, G. &. Hall C. (2001): The
site occupancy of Mg in the brownmillerite structure and its effect on hydration
properties: an X-ray/neutron diffraction and EXAFS study. J. Appl. Cryst., 34: 55–61.
Jupe, A.C., Turrillas, X., Barnes, P., Colston, S.L., Hall, C., Häusermann, D. & Hanfland, M.
(1996): Fast in-situ x-ray-diffraction studies of chemical reactions: A synchrotron
view of the hydration of tricalcium aluminate. Physical Review B, 53, R14 697-R14
700.
Jupe, A.C., Wilkinson, A. P., Luke, K. & Funkhouser, G.P. (2007): Slurry consistency and in
situ synchrotron x-ray diffraction during the early hydration of Portland cements with
calcium chloride. Journal of the American Ceramic Society, 90, 2595-2602.
Klemm, W.A. (2004) : Cement Manufacturing – A historical perspective. In Batthy, J.I.,
Miller, F.M. and Kosmatka, S.H., 2004. Innovations in Portland cement
manufacturing: Skokie, IL, Portland Cement Association, 1-35.
Krüger, H. & Kahlenberg V. (2005): Incommensurately modulated ordering of tetrahedral
chains in Ca2Fe2O5 at elevated temperatures. Acta Crystallogr., B61: 656-662.
Lea, F.M. (1998) Lea’s chemistry of cement and concrete. Fourth edition. (eds) Hewlett
P.C., Butterworth Heinemann.
Le Chatelier, H. (1919): Crystalloids against colloids in the theory of cements. Trans.
Faraday Soc., 14, 8-11.
Megaw, H.D. & Kelsey, C.H. (1956): Crystal structure of tobermorite. Nature, 177:390-391.
Meller, N., Hall, Ch., Jupe, A.C., Colston, S.L., Jacques, S.D.M., Barnes, P. & Phipps, J.
(2004): The paste hydration of brownmillerite with and without gypsum: a time
resolved synchrotron diffraction study at 30, 70, 100 and 150°C. Journal of Materials
Chemistry, 14, 428-435.
Merlini, M., Artioli, G., Cerulli, T., Cella, F. & Bravo, A. (2008): Tricalcium aluminate
hydration in additivated systems. A crystallographic study by SR-XRPD. Cement and
Concrete Research, 38, 477-486.
Merlini, M., Artioli, G., Meneghini, C., Cerulli, T., Bravo, A. & Cella, F. (2007a): The early
hydration and the set of Portland cements: In situ X-ray powder diffraction studies.
Powder Diffraction, 22, 201-208.
Merlini, M., Menehini, C., Artioli, G. & Cerulli, T. (2007b): Synchrotron radiation XRPD
study on the early hydration of cements. Zeitschrift für Kristallographie Supplement,
26, 411-416.
Mertens, G., Elsen, J., Laduron, D. and Brulet, R. (2006): Mineralogy of the calcium-silicate
phases present in ancient mortars from Tournai. Archéometrie, 30: 61-65.
Mertens, G., Madau, P., Durinck, D., Blanpain, B. & Elsen, J. (2008): Quantitative
mineralogical analysis of hydraulic limes by X-ray diffraction. Cement and Concrete
Research, 37: 1524-1530.
Mertens, G., Lindqvist, J.E., Sommain, D. & Elsen J., (2008) Calcareous Hydraulic Binders
from a Historical Perspective. Proc. of the 1-st Historical Mortars Conference,
Characterization, Diagnosis, Conservation, Repair and Compatibility. Lissabon, 1-
15.
Michaëlis, W.S. (1909): the setting of calcareous hydraulic cements. Z. Chem. Ind. Kolloide,
5:9-22.
Midgley, H.G., Rosaman, D. & Fletcher, K.E. (1960): X-ray diffraction examination of
Portland cement clinker. Proceedings of the fourth international symposium on the
chemistry of cement, 2 (Tome 2), Washington, 69-74.

Mohan, K. & H.F.W. Taylor (1981): Analytical Electron Microscopy of Cement Pastes: IV,
β-Dicalcium Silicate Pastes. Journal of the American Ceramic Society, 64, 717-719.
Mondal, P. & Jeffery J. W. (1975): The Crystal Structure of Tricalcium Aluminate,
Ca3Al206 . Acta Crystallogr., B31: 689-697.
Moore, A. E. & Taylor, H. F. W. (1970): Crystal structure of ettringite. Acta Crystal., B26:
386–393
Mumme, W.G. (1995): Crystal structure of tricalcium silicate from a Portland cement clinker
and its application to quantitative XRD analysis. Neues Jahrb. Mineral., Mh. 4: 145–
160.
Nishi, F. & Takéuchi, Y. (1984): The rhombohedral structure of tricalcium silicate at 1200°C.
Zeitschrift fur Kristallographie, 168:197-212.
Nishi, F., Takéuchi, Y. & Maki, I. (1985): The tricalcium silicate Ca3O[SiO4]: the monoclinic
superstructure. Zeitschrift fur Kristallographie, 172: 297–314.
Nonat, A. (2004): The structure and stoichiometry of C-S-H. Cement and Concrete Research,
34:1521-1528.
Nonat, A. & Lecoq, X. (1998): The structure, stoichiometry and properties of C-S-H prepared
by C3S hydration under controlled conditions. In: Grimmer, P., Grimmer, A.-R.,
Zanni, H., Sozzani, P. (Eds.): Nuclear Magnetic Resonance Spectroscopy of Cement-
Based Materials, Springer, Berlin, 197-207.
Odler, I. (2001): Hydration, setting and hardening of Portland cement In: Hewlett, P.C.
(2001): Lea's chemistry of cement and concrete, Butterworth, Heinemann, 241-298.
Ono, Y. (1980): Microscopical estimation of burning condition and quality of clinker, Proc of
the 7th Int. Congr. of Chem. of Cement, Paris, 2, 206-211.
Ono, Y., Kawamura, S. and Soda, Y. (1968): Microscopic Observations of Alite and Belite
and Hydraulic Strength of Cement, Proc of the 5th Int. Symp. on Chem. of Cement,
Tokyo, 1: 275-280.
Palladio (1570) : I quattro libri dell’Architettura, Venise, Dominico de’ Franceschi: cited in
Sbordoni-Mora, L., 1982. Les matériaux des enduits tradtionnels: in Mortars, cements
and grouts used in the conservation of historic buidings. ICCROM Symposium 1981,
Rome.
Pasley, C.W. (1997): Observations on lime (First published in 1838). Ed. Michael Wingate.
Donhead publishing.
Peterson, V.K., Ray, A.S. and Hunter, B.A. (2006): A comparative study of Rietveld phase
analysis of cement clinker using neutron, laboratory X-ray, and synchrotron data.
Powder Diffraction, 21, 12-18.
Pöllman H. (2002) Composition of cement phases. In Structure and Performance of Cements,
(eds) Bensted J. &, Barnes, P., Spon Press, London and New York, 25-57.
Pritula, O., Smrcok, L., Többens, D.M. & Langer, V. (2004): X-ray and neutron Rietveld
quantitative phase analysis of industrial Portland cement clinkers. Powder Diffr., 19:
232–239.
Rankin, G.A. & Wright F.E. (1915): The ternary system CaO-Al2O3-SiO2 with optical
study. Am. Jour. Sci., 4: 1-79.
Redler, L. (1991): Quantitative X-ray diffraction analysis of high alumina cements. Cement
and Concrete Research, 21, 873-884.
Regourd, M., Chromy, S., Hjorth, L., Mortureux, B. & Guinier, A. (1973): Polymorphisme
des Solutions Solides du Sodium dans l'Aluminate Tricalcique. J. Appl. Cryst., 6 :
355-364.
Richardson, I.G. (2008): The calcium silicate hydrates. Cement and Concrete Research,
38;137-158.
Richardson, I.G. & Groves, G.W. (1992): Models for the composition and structure of
calcium silicate hydrate (C-S-H) gel in hardened tricalcium silicate pastes, Cement and
Concrete Research, 22:1001-1010.
Richardson, I.G. & Groves, G.W. (1993): The incorporation of minor and trace elements into
calcium silicate hydrate (C-S-H) gel in hardened cement pastes. Cement and Concrete
Research, 23:131-138.
Rousselot, C., Taviot-Guého, F., Leroux, P., Léone, P., Palvadeau, P. & Besse, J. (2002):
Insights on the Structural Chemistry of Hydrocalumite and Hydrotalcite-like
Materials: Investigation of the Series Ca2M3+(OH)6Cl·2H2O (M3+: Al3+, Ga3+, Fe3+,
and Sc3+) by X-Ray Powder Diffraction. Solid State Chem. 167: 137.
Royer, A. (2004) : Le ciment romain. Etude d’un matériau et son utilisation dans les
restaurations de monuments historiques au XIXe siècle: les cathédrales d’Amiens et
de Bourges. Mémoire de muséologie à l’Ecole du Louvre.
Scrivener, K.L., Füllmann, T., Gallucci, E., Walenta, G. & Bermejo, E. (2004): Quantitative
study of Portland cement hydration by X-ray diffraction/Rietveld analysis and
independent methods. Cement and Concrete Research, 34: 1541–1547.
Scrivener, K.L., Patel, H.H., Pratt, P.L. & Parrott, L.J. (1987): Analysis of phases in cement
paste using backscattered electron images, methanol adsorption and thermogravimetric
analysis. Microstructural Development During the hydration of Cement. Proc. Mater.
Res. Soc. Symp, 85: 67-76.
Scrivener, K.L., Pratt, P.L. (1984), Back scattered electron images of polished cement
sections in the scanning electron microscope, Proceedings of the 6th International
Conference on Cement Microscopy, USA, 145-55.
Smith D.K. (1962): Crystallographic changes with the substitution of aluminum for iron in
dicalcium ferrite. Acta Crystallogr., 15: 1146-1152.
Snellings, R., Mertens, G., Hertsens, S. & Elsen, J. (2009): The zeolite-lime pozzolanic
reaction: Reaction kinetics and products by in situ synchrotron X-ray powder
diffraction. Microporous and Mesoporous Materials, 10p., In press:
doi:10.1016/j.micromeso.2009.05.017
St John, D.A., Poole, A.W. & Sims, I. (1998): Concrete Petrography, A handbook of
investigative techniques, Arnold Publishers, London-Sydney-Auckland.
Stade, H. (1980): Zum Aufbau schlecht geordneter Calciumhydrogensilicate. II. Über eine aus
Poly- und Disilicat bestehende Phase. Z. anorg. allg. Chem., 470:69-83.
Stade, H. & Wieker, W. (1980): Zum Aufbau schlecht geordneter Calciumhydrogensilicate. I.
Bildung und Eigenschaften einer schlecht geordneten Calciumhydrogendisilicatphase.
Z. anorg. allg. Chem., 466:55-70.
Stutzman, P. (2005): Powder diffraction analysis of hydraulic cements: ASTM Rietveld
round-robin results on precision. Powder Diffr., 20: 97–100.
Suherman, P.M., Riessen, A.V., Oconnor, B., Li, D., Bolton, D. & Fairhurst, H. (2002):
Determination of amorphous phase levels in Portland cement clinker. Powder Diffr.,
17: 178-185.
Takéuchi, Y., Nishi, F. & Maki, I. (1980): Crystal-chemical characterization of the (CaO)3 *
(Al2O3) - (Na2O) solid-solution series. Zeitschrift für Kristallographie, 152: 259-307.
Talansier, (1885): Les ciments de la porte de France. Le Génie Civil 24 (Tome VII), 3-7.
Tarrida, M., Madon, M., Le Rolland, B. & Colombet, P. (1995): An in-situ Raman
spectroscopy study of the hydration of tricalcium silicate. Advanced Cement Based
Marerials, 2, 15-20.
Taylor, H.F.W. (1968): The calcium silicate hydrates. Proc. 5th Int. Symp. Chem. Cement,
Tokyo, 2.
Taylor, H.F.W. (1986): Proposed structure for calcium silicate hydrate gel. J. Am. Ceram.
Soc., 69:464-467.
Taylor, H.F.W. (1990): Cement chemistry, Academic Press, New York.
 Taylor, H.F.W. (1993): Nanostructure of C-S-H: current status. Adv. Cem. Based Mater.,
1:38-46.
Taylor, H.F.W. (1997) Cement chemistry (2nd ed.), Thomas Telford Edition, London.
Taylor, H.F.W. & Howison, J.W. (1956): Relationships between calcium silicates and clay
minerals. Clay Miner. Bull., 3:98-111.
Thomas, J.J., Fitzgerald, S.A., Neumann, D.A. and Livingston, R.A. (2001): State of water in
hydrating tricalcium silicate and Portland cement pastes as measured by quasi-elastic
neutron scattering. Journal of the American Ceramic Society, 85, 1811-1816.
Thomas, J.J. & Jennings, H.M. (1999): Effects of D2O and mixing on the early hydration
kinetics of tricalcium silicate. Chemistry of Materials, 11, 1907-1914.
Tilley, C.E. (1929): On larnite (calcium orthosilicate, a new mineral) and its associated
minerals from the limestone contact-zone of Scawt Hill, Co. Antrim. Mineralogical
Magazine, 22: 77-86.
Tong, Y., Du, H. & Fei, L. (1991): Comparison between the hydration processes of tricalcium
silicate and β-dicalcium silicate. Cement and Concrete Research, 21, 509-514.
Törnebohm, A. (1897): The petrography of Portland cement. Tonindustrie Zeitung, 21:1148-
1150, 1157-1159.
Tsurumi, T., Hirano, Y., Kato, H., Kamiya, T & Daimon M. (1994): Crystal structure and
hydration of belite. Ceramic Transactions, 40: 19-25.
Varas, M.J., Alvarez de Buergo, M., Fort, R. (2005): Natural cement as a precursor of
Portland cement : Methodology for its identification. Cement and Concrete Research,
35: 2055-2065.
Vicat, L.J. (1818) : Recherches expérimentales sur les chaux de construction, les bétons et les
mortiers ordinaires. Goujon, Paris.
Vicat, L.J. (1997) : Mortars and cements (first published in 1837). (ed). Michael Wingate.
Donhead publishing.
Walenta, G. & Füllmann, T. (2004): Advances in quantitative XRD analysis for clinkers,
cements and cementitious additions. Powder Diffr., 19: 40-44.
Westphal, T., Walenta, G., Füllmann, T., Gimenez, M., Bermejo, E., Scrivener, K. &
Pöllmann, H. (2002): Characterisation of cementitious materials. International Cement
Review (July 2002), 47-51.
Wolter, A. (2005): Belite cements and low-energy clinker. Cement International, 3:107-117.
Young R.A. (1993): The Rietveld Method, IUCr Monographs on Crystallography-5, Oxford
University Press.
Zhao, H., Darwin, D. (1992), Quantitative backs scattered electron analysis of cement paste.
Cement and Concrete Research, 22 (4): 695-706.

View publication stats