C

1. Cement Chemistry
1.1

INTRODUCTION
1.2

R AW M A T E R I A L S

1.3

FUELS

1.4

CLINKER

1.5

GYPSUM

1.6

CEMENT
1.6.1

L S F, E T C

1.6.2

CLINKER COMPOUNDS

1.6.3

P R I N C I PA L C E M E N T C H A R A C T E R I S T I C S

1.7

contents

G Y P S U M - S O L U B L E C A L C I U M S U L P H AT E
1.7.1

INTRODUCTION

1.7.2

D I S S O L U T I O N O F C a S O4

1.7.3

O P T I M I S AT I O N O F S O L U B L E C A L C I U M S U L P H AT E

1.7.4

SLUMP LOSS/RETENTION

chapter 2

1.

CEMENT CHEMISTRY

in the tumbling action of the rotary kiln to form the well-known

characteristic clinker.

Compound

Formula Shorthand Molecular

Weight

Chemically, clinker consists of a mixture of compounds, which

are made up of various molecules and elements. The most
important elements in cement chemistry are shown in Figure 1
together with their atomic weights. The most relevant
compounds in Cement Chemistry are shown in Figure 2.

Water

H2O

18.02

Carbon Dioxide

CO2

44.01

Lime (Calcium Oxide)

CaO

56.08

Magnesia

MgO

40.31

The chemistry of raw materials, fuels and clinker are discussed

in more detail in the following sections.

Silica

SiO2

60.09

Titania

TiO2

79.90

Element

Symbol

Atomic Weight

Alumina

Al2O3

101.96

However, customers of cement place high demands on the

performance of the cement, expecting it to react in a predictable
manner with respect to its handling, workability, setting,
hardening and strength development.

Aluminium

A1

26.98

Calcium

Ca

40.08

Ferric Oxide

Fe2O3

159.70

Carbon

12.01

Phosphorus Pentoxide

P2O5

141.94

The first patent for Portland Cement was granted in 1824. Cements,
with a chemistry similar to today's cements, were not really
produced until the late 1800's. Shaft or bottle kilns were used.

Hydrogen

1.01

Sulphur Trioxide

SO3

80.06

Iron

Fe

55.85

Magnesium

Mg

24.31

Soda

Na2O

62.00

Manganese

Mn

54.94

Potash

K2O

94.20

Oxygen

16.00

Calcium Carbonate

CaCO3

CC

100.09

Phosphorus

30.97

Potassium

39.10

Magnesium Carbonate

MgCO3

MC

84.32

Silicon

Si

28.09

Sodium Carbonate

Na2CO3

NC

106.01

Sodium

Na

23.00

Calcium Sulphate

CaSO4

CS

136.14

Sulphur

32.06

Titanium

Ti

47.90

Potassium Sulphate

K2SO4

KS

174.26

1.1
INTRODUCTION
The production of cement can be considered as a chemical
process not so different to other chemicals such as Sodium
Hydroxide and Calcium Chloride. However there are two major
differences:
- Selling Price
- Product Complexity
Portland Cement can contain approximately 10 chemicals/minerals
and is relatively impure when compared to most "chemicals".
Part of this impurity arises from the "naturally occurring"
nature of the key raw materials, but also from the low selling
price, which is significantly lower than for most "Chemicals".
The impurities can vary quite considerably, both from one plant
to another but also within any one plant, and these can have an
important influence on the ultimate cement behaviour.

Rotary kilns were first used around 1900. These became

operated in a semi-dry manner around the 1950's and the
modern dry process kiln appeared in the 1960's and 1970's with
the more efficient pre-calciner process appearing a little later.
Wet, semi-wet, semi-dry, dry and pre-calciner process kilns all
remain in use throughout the world today. (See Section 2).
There are also a significant number of vertical shaft kilns still in
operation, particularly in China.
Portland Cement clinker is manufactured from a calcareous
material (e.g. limestone, chalk) and an argillaceous material (e.g.
shale, clay). The feed material is finely ground and carefully
mixed and heated to a very high temperature (~1500C). During
this heating some 25% of the mixture becomes liquid. This
assists in the chemical reactions and bonds the particles together

contents

chapter 1

chapter 2

For simplicity, cement chemists have traditionally used a "shorthand" for chemical symbols and these are also shown in Figure 2.

Figure 1. Principal Elements.

Figure 2. Principal Compounds.

1.2
RAW MATERIALS
We shall see later that the four most important oxides in
Portland Cement are lime (CaO), silica (SiO2), alumina (Al2O3)
and iron oxide (Fe2O3). That is, in short-hand C, S, A, and F.
The principal source of lime for cement manufacture is
limestone or chalk and this constitutes typically some 80% of
the raw material mix.
In practice the actual materials used for the source of CaO
cover the complete range of geological forms. However, the
differences of most practical relevance involve:
- Chemistry (including impurities)
- Hardness
- Porosity
- Crystal Size
- Moisture
- Location
and of course
- Cost

1.

When lower grade limestones or secondary materials are used

the associated levels of impurities, such as sulphur, alkalis and
magnesia can become important.

Figure 3. Raw Materials.

Raw Material

Source of

Thus, in terms of tonnages, the primary and secondary material

are of most importance with regard to location and cost.

Limestone or Chalk, CaCO3

CaO

Shale or Clay

Al2O3.Fe2O3.SiO2

However in addition to these, it maybe necessary to make

adjustments to the mix with sources of silica, alumina and iron
oxide. These will generally be at low addition levels, but can
significantly influence the cost of the final raw material mix.

Iron Oxide

Fe2O3

Bauxite

Al2O3.Fe2O3

Sand

SiO2

Slag

CaO (Al2O3.Fe2O3.SiO2)

Whilst the fuel is essentially there to provide the energy for

clinker formation, any associated ash (particularly since coal
remains the main source of fuel) must be taken into account
when designing the raw feed mix for a given clinker chemistry.
Some examples of raw materials are shown in Figure 3, whilst
typical chemical analyses for raw material, raw mix, coal ash
and clinker are shown in figure 4.

S
A
F
C

Some of these are discussed further in Section 2.

The proportion of non-calcareous material, i.e. any siliceous or
argillaceous material in the limestone will influence the use of
secondary materials.
The secondary material providing Al2O3, SiO2, Fe2O3 is
principally an argillaceous material like shale or clay. In general
this will contain clay minerals, as well as free silica.
However, sometimes, the principal material may contain
sufficient levels of S, A, F to produce the desired Clinker
Chemistry (origin of Ciments Naturel). More often though, the
desired mix maybe achieved with a mix of high and low grade
limestones. That is, high grade having a high level of CaO (say
50%) and low levels of Al2O3, SiO2, Fe22O3 and a low grade
material having a lower level of CaO but high levels of Al22O3,
SiO2 and Fe2O3.

contents

chapter 1

chapter 2

CEMENT CHEMISTRY

3.3
0.7
0.2
53.2

Limestone

Shale

S
A
F
C

52.8
14.2
8.7
1.0

Raw Mill
Raw
Meal

S
A
F
C

13.2
3.4
1.9
43.0

S
A
F
C

Coal
Ash

51.7
26.4
9.5
1.6

Kiln

S
A
F
C

20.9
5.6
3.0
65.7

Clinker

Figure 4. Typical Chemical Analyses of

Materials.

1.

CEMENT CHEMISTRY

Increasingly, the cement industry is considering the utilisation of

less traditional materials, which can often contain significant
levels of impurities (with respect to cement manufacture).

Figure 5. Limits for Use of Raw Materials and Fuels

Element

As

Limit g/t

Principal Reason for Limitation/Notes

For these (and also for the 'traditional' materials to some

extent) it is necessary to carefully assess their potential impact
on the industry. This may involve impact in the following areas:
- Raw Material Cost
- Availability, Quantities
- Capital Requirement
- Public Environmental Awareness
- Impact on Emission Limits
- Material Health and Safety
- Influence on Flame Characteristics
- Influence on Kiln Performance
- Influence on Clinker/Cement Properties
- Influence on Emissions
- Influence on Product Health and Safety

Antimony

Sb

3000

Environmental Emission Impact

Arsenic

As

270

Environmental Emission Impact

Beryllium

Be

90

Environmental Emission Impact

Cadmium

Cd

Environmental Emission Impact

Chrome

Cr

100
10

Product Health & Safety

Colour (White Cement)

Fluorine

500

Environmental Emission Impact. Increases strength at low levels (<0.5%) and

extends set time (esp. at low temp)

Cobalt

Co

100

Product Health & Safety

Mercury

Hg

1-2

Environmental Emission Impact

Lead

Pb

370

Environmental Emission Impact. Extends set time, reduces early strength

Nickel

Ni

130

Environmental Emission Impact

Selenium

Se

80

Environmental Emission Impact

Tellurium

Te

450

Environmental Emission Impact

Thallium

Th

0.02

Environmental Emission Impact

Vanadium

650

Environmental Emission Impact. Adverse effect on refractories

Zinc

ZnO

10000

Product Quality

Strontium

SrO

5000

Product Quality, produces alite instability

Barium

BaO

5000

Product Quality

Phosphorus

P205

10000

Product Quality. Extends set time, produces alite instability (high levels)

Titanium

TiO2

10000

Product Quality, reduces early hydraulic activity

Copper

Cu

1000

Product Quality, Reduces 'clinkerisation'

Sulphur

SO3

7500

Process, kiln & preheater deposits

Chloride

Cl

250

Process, kiln & preheater deposits

Alkali
(Eq.Na20)

Na20

10000

Process, kiln & preheater deposits

Increases early strength, decreases later strength

How individual materials or elements actually influence some of

these parameters can be quite complex, for example, involving
the degree to which elements are directly retained in the clinker.
A list of the potentially problem causing elements is shown in
Figure 5, together with their principal concerns (i.e.
environmental, health and safety, process operation or product
quality). Recommended maximum input levels in grams per
tonne of clinker equivalent are also shown.

contents

chapter 1

chapter 2

1.3
FUELS
The primary requirement for the fuel is to provide the necessary
energy involved in the clinker manufacturing process. This
typically amounts to between 700 and 1500 kcal/kg clinker,
with some 420 kcals/kg being the theoretical heat requirement,
(i.e. the heat required to convert the raw materials into the
clinker minerals). The remainder of the heat is essentially
accounted for by removal of moisture, sensible heat of exit gases
and the shell losses.

1.

Figure 6. Factors Influencing the Choice of Fuel.

Property

Influence on

Calorific Value

Fuel Consumption

Price

Fuel Cost/clinker tonne

Carbon/Hydrogen ratio

Flame characteristics

Ash content & chemistry

Mix proportioning, Clinker quality,

Refractories

Hardness

Preparation

Volatiles

Flame characteristics

Abrasivity

Preparation, Handling

The fuel consumption will of course be dependent on the energy

content of the fuel and the energy requirement of the kiln
process. For example, for a coal of 7300 kcals/kg and a modern
pre-calciner kiln requiring 760 kcals/kg the fuel consumption
will simply be 760/7300 = 10.4%, i.e. 1 tonne of clinker will
require 0.104 tonnes of coal. Similarly, for a less efficient wet
process of 1400 kcals/kg the fuel consumption would be 19.2%.

Moisture

Flowability, Abrasivity, Preparation

Grading, Size

Many factors will influence the ultimate choice of fuel for a

given plant. These will include any potential influence on the
fuel preparation (e.g. coal grinding), flame characteristics, ash
chemistry (hence clinker quality), kiln build-ups (e.g. Sulphur
and Chloride contents), process emissions as well as, of course,
price, calorific value, supply and consistency. The choice will be
influenced by many of the parameters discussed for raw
materials (See Section 1.2 and Figure 5). Some of the relevant
parameters for fuels are shown in Figure 6.

The current primary fuels for the cement industry are

hydrocarbon fuels, their main constituents being carbon and
hydrogen. Coal, oil and natural gas are the most common in
use.

contents

chapter 1

chapter 2

CEMENT CHEMISTRY

In recent years there has been particular interest in petroleum

coke, waste solvents and used tyres.
However, as with the primary fuels, the use of alternatives will
depend on a range of factors concerning health and safety (both
the material itself and resultant influence on the cement), fuel
preparation, kiln operation, emissions and clinker/cement
quality and cost and availability. As an example, petroleum coke
(See Figure 8) often contains important levels of sulphur whilst
waste solvents often contain appreciable levels of chloride.
Figure 8. Petroleum Coke.
Most from United States

Coal replacement
limited by the impact
on:

Flowability, Preparation

High Sulphur (~5%, Coal <1%)

SO2

Sulphur, Chloride

Kiln build-ups, Corrosion

Harder than coal

Process

Nitrogen

Environment
Refractories

6 - 8mtpa (Prime fuel required

~100-150mtpa)

Burning conditions

Vanadium
Viscosity

Handling, atomisation, Flame

Increasingly, the traditional primary fuels are being replaced (at

least partially) by a large range of waste and by-product
materials, which have significant energy contents. A list,
showing examples of these, is shown in Figure 7.

Emission limits
Typically 60% of International
Coal Price e.g. Coal at 1.8 US$ per Mill output
GJ (say 55$/tonne)
Availability
Coke at 1.2 US$ per GJ (say
35$/tonne)
Price differential

Figure 7. Examples of some Alternative Fuels.

Comments:

Heavy Fuel Oil, Natural Gas

Petroleum Coke
Waste Solvents
Waste Oils, Orimulsion
Refuse Derived Fuel, Biogas
Tyres, Recycled Car parts
Raw Materials (e.g. Oil Shale)
Timber Waste
Straw
Impregnated Sawdust
Hazardous Waste
Water Purification Residues

Significant in short term

Less so in longer term
Ability to reduce clinker cost by ~2$/tonne

1.4
CLINKER
In general, Portland Cement Clinker contains four principal
chemical compounds, or clinker minerals as they are usually
referred to. These are:
Tricalcium Silicate
Dicalcium Silicate
Tricalcium Aluminate
Calcium Aluminoferrite
The composition of these together with their short-hand
notation mineral name and typical levels are shown in Figure 9.

1.

Figure 9. Clinker Minerals.

Mineral

Shorthand

Name

Formula

Formula

Typical

Range

Tricalcium Silicate

C3S

Alite

3CaO.SiO2

Ca3SiO5

60%

30-70%

Dicalcium Silicate

C2S

Belite

2CaO.SiO2

Ca2SiO4

20%

5-40%

Tricalcium Aluminate

C3A

Aluminate

3CaO.Al2O3

Ca3Al2O5

10%

5-15%

Calcium Aluminoferrite

C4AF

Ferrite

4CaO.Al2O3.Fe2O3

Ca4Al2Fe2O10

8%

5-15%

The silicate minerals are largely responsible for the strength

development characteristics of Portland Cement. However, the
reaction between lime and silica is difficult to achieve, even at
high temperatures. The reaction (Chemical Combination) is
facilitated by the presence of alumina and iron oxide, as these
assist in the formation of a molten flux through which the lime
and silica are able to partially dissolve and then react to form
C3S and C2S.
The sequence of reaction is illustrated in Figure 10. A typical
full clinker chemistry is shown in Figure 11.
Clinkers designed for sulphate resistance require materials with
low levels of Al2O3 so that the clinker contains a low (or zero)
level of C3A. In general the C3A content will be less then 1%
whilst the C4AF will be higher at 15-20%.
For White Cements materials with low levels of Fe2O3 are
required so that the clinker contains a low (or zero) level of
C4AF. As a consequence the C3A contents will usually be higher
at 12-16%, although it is possible to avoid high C3A by having
very high total silicates. In addition the clinker has to be
produced with materials low in Mn2O3 and Cr2O3.
The influence of chemistry is discussed in Section 1.6. However
clinker properties can also directly influence:
- Materials Handling
- Mill Performance
- Storage

contents

chapter 1

chapter 2

CEMENT CHEMISTRY

Figure 10. Sequence of

Formation of Calcium
Silicates in a Rotary Kiln.

It should also be noted that many plants use stored clinker in

addition to the 'fresh' kiln clinker. Sometimes these maybe from
external sources. Therefore the main characteristics influencing
the above can vary significantly.
Materials handling will most notably be influenced by the
clinker size grading and the temperature. For stock clinker, the
degree of degradation (i.e. ageing/hydration) will also be
important
Figure 11. Example of EN197 CEM I Clinker Chemistry.
SiO2

21.3

LSF

96.5

Al2O3

5.6

S/(A+F)

2.45

Fe2O3

3.1

A/F

1.81

CaO

66.0
C3S

60

MgO

1.3

C2S

16

Mn2O3

0.1

C3A

10

P2O5

0.2

C4AF

TiO2

0.3

LOI

0.1

Na2O

0.8

K2O

0.2

SO3

1.0

Free Lime

1.2

contents

chapter 1

Size grading will be a function of raw materials, kiln process

and kiln operation. In general hard burning conditions can be
expected to produce a dusty clinker (large proportion less than
1mm). In fact, it has been shown that free lime can be linked to
the clinker fines content, with low free limes (associated with
hard burning) leading to a high fines content. This in turn can
result in difficult handling and a hard grindability.

Conversely, underburned clinkers with a very high free lime can

also contain a high fines content.
Besides the degree of burning (and cooling), that influence the
microstructure, the grindability is strongly influenced by the
levels of C2S and SO3. The latter will affect the level of gypsum
required to produce a target SO3 content in the cement. Hence
the perceived grindability will be influenced by the proportion
of 'softer' gypsum present.
Clinkers with a wide size grading will be subject to more
segregation during storage and hence the material sizes in the
clinker feed to a mill can be variable. This in turn can result in a
variable mill performance.
Because of the natural size grading of clinkers, sampling is a
difficult task. For this reason extreme caution should be taken
when taking samples (See Section on Cement Grinding) of
clinker (and Gypsum).

chapter 2

1.

CEMENT CHEMISTRY

1.5
GYPSUM
Although clinker contains some SO3, arising from the raw
materials and fuel, additional SO3 is required to produce cement
from the clinker.
The SO3 in the clinker is usually combined with the alkalis to
form:Alkali sulphates
K2SO4, Na2SO4
Calcium Langbeinite
2 CaSO4. K2SO4
Calcium Sulphate
CaSO4
In discussions any additional SO3 is usually referred to as
'gypsum', but in reality the source of SO3 will contain gypsum,
anhydrite and other minerals such as clay, quartz and calcite as
well as free moisture.
A full analysis is required to precisely identify the minerals
present in any source of SO3. However the most important
parameters concern the total SO3 content and the level of
gypsum and anhydrite. The latter can be derived from the SO3
content and the loss on ignition at 50, 250 and 950C.
An example of calculations for gypsum and anhydrite is
provided in Technical Information Sheet (TIS) number MS001.
It has also been traditional to refer to the 'gypsum' as a 'set
regulator' or 'set retarder'. However the main role of the added
SO3 is to prevent rapid reaction of the aluminate which would
result in early stiffening, loss of workability and early set.
In addition to preventing the rapid reaction of C3A, the
presence of added SO3 also influences the wider hydration
process, which affects setting, workability and strength
development. These in turn affect the target SO3 content of the
cement. However this is typically in the range 2.5 - 3.5%.

contents

chapter 1

chapter 2

The level of SO3 additive ('Gypsum') required is simply

calculated by a mass balance using the 'gypsum', clinker and
cement SO3 contents. This is described in TIS No: MS002.
The above calculation often needs to be modified to allow for
non-clinker components, and this is also shown in TIS No:
MS002.
Combining the calculations of "gypsum" addition level and
gypsum/anhydrite contents provides the levels of added
anhydrite and gypsum in the cement.
Whilst the anhydrite (natural anhydrite) remains unaffected
during cement grinding, the gypsum largely dehydrates. The
resultant levels of anhydrite and gypsum dehydration products
(i.e. hemihydrate and soluble anhydrite) can have a marked
influence on cement performance, particularly regarding
concrete water demand and workability.
The sources of SO3 in cement, their dissolution rate, the
optimisation of soluble CaSO4 and concrete slump behaviour
are discussed in more detail in Section 1.7.

1.

CEMENT CHEMISTRY

1.6
CEMENT
1.6.1
LSF, etc
We have already seen that Clinker, and thus cement, principally
consists of the calcium silicate minerals, and these derive from
raw materials containing the four principal oxides, namely CaO,
SiO2, Al2O3 and Fe2O3.
To assist in the proportioning of the raw materials, the excess
CaO available in the limestone (or other), together with the
CaO required to saturate the oxides in the shale (or other) needs
to be known.
Lea and Parker derived a formula (1935) from the base of
100% lime saturation, which permits the calculation of the lime
required for saturation of the other oxides, i.e.
CaO

= 2.80 SiO2 + 1.18 Al2O3 + 0.65 Fe2O3

The lime saturation factor (LSF) for any mix of raw materials is
then given by:
LSF

= CaO / (2.80 SiO2 + 1.18 Al2O3 + 0.65 Fe2O3)

The ratio is usually expressed as a percentage. It is also applied

to the clinker and the cement, although the calculation can be
complicated by the presence of CaSO4, free lime, alkalis, etc.
For clinker the above formula is usually used, although this can
be misleading if there is a high SO3 content in which some of
the CaO has combined with the SO3 to leave calcium
langbeinite and calcium sulphate.

1.

Control of LSF will however not control the resultant levels of

aluminate or ferrite. To assist in their control, and targeting, it
is usual to use the parameters of silica ratio and alumina ratio.
Thus there can often be three target control parameters, i.e.

CONTROL PARAMETERS

NO. OF RAW
MATERIALS REQUIRED

LSF

LSF, SILICA RATIO

LSF, SILICA RATIO,

ALUMINA RATIO

contents

chapter 1

chapter 2

10

satisfy the iron oxide in this is assumed to be present as C3A. The

remaining lime is then proportioned between the C3S and C2S.
Again, for cement, the quantity of lime present as CaSO4 needs
to be taken into account. In addition, for cement, the lime
present which is uncombined (free lime) can also be deducted to
provide a more realistic calculation. However some estimations
are for potential Bogue composition (e.g. ASTM), in which the
total lime is used.
It is possible to refine these calculations further, for example by
allowance for the combination of some of the SO3 with alkalis
rather than CaO.

As you can see for 1 parameter, 2 materials are required, for 2

parameters, 3 materials, and for 3 parameters, 4 materials, and
so on.
Additional target control parameters, like alkalis, magnesia, will
therefore require more than 4 raw materials.
Typically, clinker will have:
MEAN

RANGE

LSF

95

90 - 98

SILICA RATIO

2.5

2-4

ALUMINA RATIO

1.7

1-3

For cement it is usual to deduct the lime present in the CaSO4,

before calculating the LSF (See TIS No. MS003).
In proportioning raw materials, at least two materials will be
required to achieve a target LSF. In 'Ciments Natural', a single
raw material chemistry will lead to variation in the clinker
chemistry (and hence LSF). Two materials permit control to a
single target LSF, by varying the component proportions.

CEMENT CHEMISTRY

1.6.2
Clinker Compounds
The main clinker compounds can be determined by x-ray
diffraction but are typically estimated according to Bogue
(1929). These estimations are shown in TIS No. MS003.
In these calculations the iron oxide is assumed to be present in
the alumina-ferrite phase and all of the alumina not required to

Calculations according to Bogue should be treated only as the

estimation that they are. One area often giving rise to errors is
the assumption of an alumina/iron oxide ratio of 1 for the
ferrite phase. It is known that this can be variable in the
presence of some minor compounds in the clinker (e.g. fluorine).
1.6.3
Principal Cement Characteristics
[Link]
Introduction
The most important cement characteristics which are known to
have a significant influence on cement, mortar and concrete
performance include:
- Silicates
- Aluminate
- Ferrite
- Alkalis
- SO3, Clinker
- SO3, Cement
- Forms of SO3
- Free Lime
- Fineness (Blaine, PSD)
- Microstructure (Crystallography, Burning, Cooling)
- Surface Properties (Hydration, Carbonation)
- Composition (Non-Clinker components)
In addition, minor components, such as fluoride, chloride, heavy
metals can also be important (See Figure 5).

The following sections deal with these in more detail.

Additional information concerning cement performance is
provided in Section 7.
[Link]
Silicates
The total silicate level directly influences the strength potential.
A higher lime saturation factor (LSF) will result in a greater
ratio of C3S to C2S and this will tend to produce a higher early
strength (up to 7 days) for a given 28-day strength. As a guide,
a 1% change in C3S content will be equivalent to around 0.35
MPa at 2-days and 0.10 Mpa at 28-days (EN196 mortar).
Note should be taken of whether the data for C3S, etc is
"potential" Bogue composition or reflects the actual level of free
lime. Higher levels of free lime result in a lower C3S / C2S ratio.
[Link]
Aluminate
The principal influence of the C3A content (and its
crystallography) is on the water demand, workability and
setting behaviour. The interaction between C3A and SO3 is
important and directly influences the water demand and
workability characteristics.
In general high levels and/or reactive forms of C3A could lead to
a tendency for "flash set" behaviour. Both of these, which also
depend on the supply of soluble calcium sulfate, can have a
negative influence on the workability and slump retention
behaviour, which can often be more pronounced when concrete
admixtures are used.
Statistically C3A has been seen to influence strength. As a guide,
a 1% change in C3A content will be equivalent to around 0.5
MPa at 2-days and 1.0 Mpa at 28-days (EN196 mortar).
[Link]
Ferrite
In general, the influence of ferrite content will not be as
significant as other parameters. The most notable effect
concerns colour and cements are manufactured with low levels
for "off-white" cements and very low levels in white cements.

contents

chapter 1

chapter 2

The C4AF content can however create a barrier to silicate

hydration and hence reduce the strength development potential.
The effect depends on the burning, cooling and grinding history
of the clinker. Hence, in some cement, the enhancement of
ferrite hydration can contribute to increased silicate hydration.
The role of cement additives is known to be important in this
respect ("facilitated transport mechanism").
[Link]
Alkalis
The alkali oxides K2O and Na2O have a very strong influence
on cement properties, notably concerning early and late
strength, workability properties and bleeding.
Where there is sufficient or excess SO3 present in the clinker the
resultant alkali sulfates will be readily soluble and this has an
important influence on calcium ion solubility and hence on the
rate of hydration. (See TIS MS004). As a guide, the early
strength (1-3 days) will be increased by around 0.8 MPa
(EN196 Mortar) for every 0.1% increase in the equivalent
Na2O level (Na2O + 0.658K2O). At the same time the late
strength (28-days) will be decreased by around 1.7 MPa.
[Link]
SO3 Clinker (See Also Section 1.7)
As discussed above, the level of SO3 in the clinker is important
with respect to the alkali level and whether these are readily
soluble (as sulfate) or enter into the clinker minerals (and affect
reactivity).
The level of SO3 in the clinker also directly influences the
amount of "gypsum" required to achieve the target cement SO3
content. This in turn will directly influence the "grindability"
and hence the SSA (Blaine)/mill output relationship. As a guide
a 1% increase in "gypsum" addition equates to around 5%
increase in mill output at constant Blaine, (or 12m2/kg increase
in Blaine at constant mill output).
Since the level of clinker SO3 influences the amount of
"gypsum" added, the levels of the various forms of SO3 in the
cement can also be influenced (see below).

1.

CEMENT CHEMISTRY

11

[Link]
SO3 Cement (See Also Section 1.7)
The total SO3 content of the cement influences the hydration
process and thus effects the strength development, setting and
workability. Also, as discussed, the level of "gypsum" added
influences the Blaine. Hence for a constant clinker SO3 content,
a varying cement SO3 level indicates variations in added
"gypsum". Hence there will be a direct influence on the
Blaine/mill output relationship.
As a guide, a 1% increase in cement SO3 content will equate to
around a 30m2/kg increase in Blaine (constant mill output) or
10-15% increase in mill output (constant Blaine).
In general, higher levels of SO3 content will enhance the early
strength, extend the setting items and usually decrease the
concrete workability (increase the water demand). As a guide, a
0.1% increase in cement SO3 content will equate to 0.5 MPa at
2-days and 0.1 Mpa at 28-days (EN196 mortar).
There will be an optimum level of SO3 content with respect to
the 28-day strength. This will depend on many factors
including clinker chemistry and whether derived at constant
Blaine or constant mill output.
[Link]
SO3 - Forms (See Also Section 1.7)
The principal forms of SO3 in cement are:
a)

ex. clinker

- alkali sulfate, K2SO4, Na2SO4

- calcium langbeinite,
2CaSO4 .K2SO4
- calcium sulfate, CaSO4

b)

ex. added 'gypsum'

-

-natural anhydrite, CaSO4

gypsum, CaSO4 .2H2O
hemihydrate, CaSO4 .1/2H2O
soluble anhydrite, CaSO4

The major difference between these forms concerns their rate of

dissolution and hence the level of available soluble calcium
sulfate in the cement.

The level of soluble calcium sulfate principally influences the

following:
- concrete water demand (concrete slump,
workability)
- slump retention or rate of slump loss
- silo set or pack set
- rate of strength development
[Link]
Free Lime
The level of free lime itself can influence the initial reactivity
and setting behaviour, but more importantly, is a guide to the
level of burning (see below). It can also alleviate problems of
silo set (acts as a desiccant). (See also section [Link] - Silicates).
Free lime target is typically between 1 and 2%. Higher levels
represent incomplete reaction between lime and silica (hence less
silicates for strength development), whilst lower levels indicate a
more complete reaction, but usually at the expense of hard
burning conditions and reduced strength development potential.
(See also section [Link]).
[Link] Fineness - Blaine, PSD
The cement fineness is naturally influenced by the mill output
but, as already discussed, can be influenced by other
parameters, such as SO3 level, filler level, etc. The principal
measure is the Blaine (m2/kg) but a more important one is the
particle size distribution (PSD). Often, this is not available but
the Blaine, in combination with a residue level (e.g. Alpine 45micron), can provide a good estimate of the PSD. (See TIS no.
MS005 and Section 3.3)
In general, high levels of fineness will result in enhanced
strengths. The Blaine is a better indication of early strength,
whilst the residue is a good guide to the 28-day strength.
As a general guide, for OPC type cements (Type I), a 10m2/kg
increase will be equivalent to around 0.30 Mpa (all ages) and a
1% increase in 45-micron residue will be equivalent to around 0.35 Mpa at 2-days and -0.40 Mpa at 28-days. (See also TIS no.
MS006).

contents

chapter 1

chapter 2

1.

It should be noted that if influenced by parameters such as SO3,

prehydration, filler, etc, than these guidelines will not be valid.
It has also been noted that the more efficient milling systems
will result in a narrower PSD, i.e. lower 45-micron residue for a
given Blaine. (See also section 3).
[Link] Microstructure
The clinker microstructure can have a very strong influence on
cement performance but is less well defined on a routine basis.
It is influenced principally by:- raw material mineralogy
- raw feed fineness
- burning regime and temperature
- fuel type and fineness
- cooling
The microstructure is studied by microscopic examination.
As a guide, hard burning and/or slow cooling will lead to large
crystal sizes which can adversely influence both strength
development and grindability. In the extreme, for clinkers with a
high MgO content, the burning and cooling regime could lead
to expansive properties, due to large periclase crystals.
Clinker Microstructure Characteristics:
Alite Sizes, Typically 15100 microns

Smaller (<40) - better strength

Larger (>60) - less reactive

Burning Temperature

Lower - better crystal sizes

Cooling Regime

Faster - smaller crystal size

Crystal Impurities

Purer crystals - less reactive

Reduction

Loss of SO3, Flowability problems

Raw Feed Fineness

Coarser - difficult combinability

Raw Feed homogeneity

Siliceous - lower strength

Calcareous - better strength
Heterogeneity - difficult
combinability

Porosity

Lower - difficult grindability

CEMENT CHEMISTRY

12

[Link] Surface Properties

The cement surface characteristics can be expected to influence
the early age performance, for example, water demand,
workability, setting and early strength.
Significant levels of prehydration or carbonation, e.g. due to the
utilisation of outside stock clinker or high water injection rates
can result in a high fineness, but with a misleading influence on
strengths.
The only guide to surface properties readily available is the loss
on ignition (LOI). In general a 1% increase in the LOI will lead
to around a 3 Mpa reduction in 28-day strength and a 0.50
Mpa reduction at 2-days (EN 196 mortar).
The absence of any moisture can lead to high initial reactivity (a
possible contribution to the high reactivity of cements produced
in a finish roll press) and resultant reduced workability.
[Link] Cement Composition
Naturally, besides the clinker, other components can have a
significant influence on the final cement properties. Variations in
level and quality of non-clinker components could directly
influence setting, workability and strength performance. For
example, as a guide, a 1% change in limestone can influence 28
day strength by around 1%. See also section 9.
Non-clinker components can also effect the fineness/mill output
relationship and hence influence the evaluation of mill
performance.
Insoluble residue (IR) can provide a guide to the presence of
some non-clinker components.

[Link] Other Properties

Other parameters that can also have an influence on cement
performance include:
MgO
Fluoride
Chloride
P2O5
Strontium, Barium
Heavy Metals
Transition Metals

(e.g.. an expansive behaviour)

(e.g.. on setting and late strength)
(e.g.. on setting, early strength and
corrosion)
(e.g.. on setting behaviour)
(e.g.. on C2S stabilisation)
(e.g.. Pb, Zn) (e.g. on setting
behaviour)
(e.g. Cr, V, Mn) (e.g.. on colour)

(See also Figure 5).

Apart from MgO (and Mn2O3 and P2O5 where XRF analysis is
available) it is not common to have this data readily available.
However there are specific examples where trace levels can be
important and relevant.
[Link] Conclusion
It will not always be possible to acquire all of this data, but the
following should usually be readily available:
- Silicates
(either as C3S, C2S or calculated from
SiO2, Al2O3, Fe2O3 and CaO, SO3
and free lime)
- C3A
(either as C3A or calculated as above)
- C4AF
(either as C4AF or calculated as
above)
- Alkalis
(Na2O and K2O)
- SO3
(for the cement and possibly the
clinker)
- Free Lime
- Fineness
(Blaine and possibly the Alpine, or
other, residue)
- Composition (i.e. non-clinker components)
The alkalis, SO3, fineness, silicates and composition are likely to
be the most relevant parameters in understanding relative
cement performance. A table of the typical relationships for
some of these parameters is shown in Figure 12.

contents

chapter 1

chapter 2

1.

Figure 12. "Rules of Thumb" Relationships.

Effect on
Incremental strength
change
2d
28d
MPa
MPa

Parameter

Unit

SSA

m2/kg +10

0.30

0.30

45-micron residue %

+1

-0.35

-0.40

Eq. Na2O

+0.1

0.80

-1.70

LOI

+1

-0.50

-3.00

C3S

+1

0.35

0.10

Free Lime

+1

0.50

-1.50

SO3

+0.1

0.50

0.10

C3A

+1

0.50

1.00

Data concerning important properties, such as the clinker

microstructure, surface properties and the forms of SO3 in the
cement will be virtually impossible to readily acquire.
It is therefore recommended, where possible, that when samples
are received and during plant trials data including the full
chemical analysis is recorded. This will usually be in the form:SiO2
Al2O3
Fe2O3
CaO

LSF
S/(A+F)
A/F

MgO
Mn2O3
TiO2
P2O5

C3S
C2S
C3A
C4AF

SO3
Na2O
K2O
LOI

Free Lime

Insoluble Residue
Blaine (SSA)
Residue (e.g.. 45 micron)

Composition (e.g. non clinker levels)

Gypsum details (e.g. Gypsum, anhydrite)

CEMENT CHEMISTRY

13

1.7
'GYPSUM' - SOLUBLE CALCIUM SULPHATE
1.7.1
Introduction
The principal sources of SO3 in cement are:
a)
ex clinker, i.e.
alkali sulphates
K2SO4, Na2SO4
calcium langbeinite 2Ca SO4 .K2SO4
calcium sulphate CaSO4
Note: The calcium sulphate present in clinker, formed where
there is a large excess of SO3 over alkalis, will be equivalent to
anhydrite (anhydrite I - slowly soluble).
b)

ex the added calcium sulphate ('gypsum'), which can be

present as:
natural anhydrite CaSO4
gypsum
CaSO4 .2H2O
hemihydrate
CaSO4 .1/2H2O
soluble anhydrite CaSO4

Note:

(1)Natural anhydrite or anhydrite II.

(2)Hemihydrate and soluble anhydrite (anhydrite III)
can be present as and forms. The forms are
more reactive (in plaster properties terms).
The forms predominate where gypsum is processed
in a less than water saturated atmosphere at
atmospheric pressure, although some forms will
also be present.
To produce only the forms, processing has to be
carried out at elevated pressures or at atmospheric
pressure in a water-saturated atmosphere.
Hence it follows that in cement milling the forms
will be preferentially produced.

Natural anhydrite is unaffected by the temperatures involved in

cement milling and storage.
Gypsum will dehydrate to hemihydrate and soluble anhydrite.
The degree of this conversion will principally depend on
temperature, residence time and atmosphere.

contents

chapter 1

chapter 2

1.

It is difficult to distinguish between soluble anhydrite and natural

anhydrite by analysis. (XRD can provide some indication).
As a guide, the following is likely:
In cement milling, below 80C, little gypsum will dehydrate.
Between 80-100C, up to around 50% of the gypsum will
dehydrate. Between 100-120C, more than 50% will dehydrate
and some soluble anhydrite will be formed. Above 120C, little
or no gypsum will remain and a greater proportion of soluble
anhydrite and hemihydrate will be found.
In storage, any retained gypsum (ex the cement mill) can slowly
dehydrate at temperatures above around 70C.
Thus a combination of relatively low milling temperatures but
moderately high storage temperatures (e.g. milling at 100C,
storage above 70C) can result in silo problems as a result of
gypsum dehydration, water migration and cement hydration
(i.e. silo set and lump formation).

CEMENT CHEMISTRY

14

Cement B: No anhydrite, moderate milling temperature produces a moderate level of D.SO3.

Cement C: Anhydrite present, high mill temperature - produces
a moderate level of D.SO3.
Cement D: Anhydrite present, low milling temperature produces a low level of D.SO3.
1.7.2
Dissolution of CaSO4
The principal differences between the various forms of CaSO4
present in cement concern their rate of solubility.
Gypsum gradually dissolves up to the level that represents a
saturated solution of CaSO4. Natural anhydrite dissolves more
slowly. Hemihydrate or soluble anhydrite dissolve very rapidly
and form a super-saturated solution with respect to CaSO4. The
rates of solubility are shown in Figure 13b.
Figure 13b. Dissolution Rates for Calcium Sulfate Forms

Figure 13a provides a range of scenarios for a cement SO3

content of 3.0% and a clinker SO3 content of around 1.0%
(assuming a typical level of alkalis of around 0.6 eq. Na2O).
Figure 13a. Examples of Cement Sulfate Forms.
Cement Type:

Cement

3.0

3.0

3.0

3.0

Clinker

1.0

1.0

1.0

1.0

Natural Anhydrite

0.0

0.0

1.0

1.0

Gypsum

0.1

1.0

0.1

1.0

Hemihydrate + Soluble Anhydrite 1.9

1.0

0.9

0.0

Sources of SO3:

Cement A: No anhydrite, high mill temperature - produces high

level of D.SO3 (i.e. hemihydrate + soluble
anhydrite).

Dissolution rates can be influenced by the presence of certain

concrete admixtures and by other sources of SO4, e.g. from the
alkalis.

1.7.3
Optimisation of Soluble Calcium Sulphate
The availability of soluble CaSO4 is required to 'satisfy' the
initial reactivity of the C3A. The 'optimum' supply of soluble
CaSO4 will therefore depend on the initial 'reactivity' of the
C3A. This will principally depend on:
-

A schematic of the above scenarios is shown in Figure 14.

level of C3A (actual, rather than just according to Bogue)

fineness
surface freshness (e.g. presence of prehydration)
alkali/sulphate balance
crystal size (burning, cooling, etc.)

An excess of alkalis in the clinker can result in alkali modified

C3A which is more reactive.
Balanced retardation of the C3A will result in the formation of
fine grained ettringite (C3A.3CaSO4.32H2O). This provides
maximum mobility in fresh concrete thereby optimising
workability:
- BALANCED RETARDATION
An excess of soluble CaSO4 over that necessary will provide a
supersaturated solution of CaSO4 and resultant precipitation of
gypsum crystals. The morphology of these will not provide
maximum mobility and hence workability (initial) will not be
optimised. This can be more pronounced where the cement
content is high. Prolonged mixing is generally accepted as a
means to counter this loss of workability:
- FALSE SET
Insufficient soluble CaSO4 will allow the formation of
monosulphate (as opposed to tri-sulphate [= ettringite]) or even
C3A hydrates (4CaO.Al2O3.13H2O). Again, the morphology of
these will prevent maximum mobility being obtained. More
importantly, the rate of slump loss will be severe:
- FLASH SET

1.

CEMENT CHEMISTRY

15

In general:
- the optimum level of D.SO3 will be around 1.0% SO3
(i.e. typically in the range 0.7 - 1.3% SO3).
- for more reactive clinkers the optimum D.SO3 will be higher.
- for less reactive clinker the optimum D.SO3 will be lower.
- for less reactive clinkers the curve will be flatter in the
region below the optimum D.SO3 but steeper in the
region above the optimum (i.e. tendency for false set
behaviour more important).
- for more reactive clinkers the curve will be flatter in the
region above the optimum D.SO3, but steeper in the
region below the optimum (i.e. tendency for flash set
behaviour more important).
1.7.4
Slump Loss/Retention
The preceding points mainly refer to the initial slump or
workability or water demand in concrete. However a consistent,
and predictable, slump loss behaviour over the initial period of,
say, up to one hour is more important.

Figure 14. Optimisation of Cement Sulfate.

Therefore, for any clinker, there is an optimum level of soluble
CaSO4 (i.e. D.SO3) which provides maximum initial workability
(or minimum water demand). This relationship is illustrated in
Figure 15 showing concrete water demand versus D.SO3.
Figure 15. Concrete Water Demand versus Soluble Calcium Sulfate.

In general, low levels of D.SO3 and/or reactive clinkers (e.g.

excess alkalis over SO3) will result in a moderate to low initial
slump and severe rate of slump loss.
Conversely high levels of D.SO3 and/or less reactive clinkers,
whilst also producing a moderate to low initial slump, can
provide an overall better retention of slump behaviour.
With very high levels of D.SO3 (i.e. false setting) the initial
slump can be very low. But after mixing the slump can be
substantially improved.
In general, high w/c ratios (e.g. as a result of poor initial slump)
result in less of a slump loss. Conversely low w/c ratios can
produce a more severe slump loss.
Figure 16 shows slump behaviour over 60 minutes for two
cements. Slumps were determined at 6, 20, 40 and 60 minutes,
in each case after a short period of re-mixing.

contents

chapter 1

chapter 2

Figure 16. Slump Loss.

slump of only 16mm. However, the slump at 60 minutes is now

18mm, i.e. back to the level of the first cement.
A summary of the principal points concerning cement SO3 are
shown in Figure 17.
Figure 17. Key Points concerning Soluble Calcium Sulfate.

A - Cement with Mill temperature ~100C D.SO3 ~ 1.0%

B - Cement with Mill temperature ~80C D.SO3 ~ 0.4%
C - Laboratory heat treated Cement B
D.SO3 ~ 1.6%
The cement milled at 95-100C had a level of D.SO3 of around
1.0% and gave slumps of 40 and 17mm at 6 and 60 minutes
respectively. In order to achieve slumps of 40mm, w/c ratios of
0.60 and 0.64 respectively would be necessary. Therefore the
slump loss was equivalent to an increase in water demand of
7%.
The cement milled at 80-85C had a lower level of D.SO3 of
around 0.4% and gave slumps of 25 and 7mm respectively.
Again, to achieve slumps of 40mm, w/c ratios of 0.62 and 0.69
respectively would be necessary. Therefore the slump loss was
equivalent to an increase in water demand of 11%.
The third curve shows the results of this latter cement after
laboratory heat treatment. The D.SO3 was raised to around
1.7%. The cement now false sets and there is a very low initial

contents

chapter 1

chapter 2

SO3 present in cement as:

Ex. clinker alkali sulfate
calcium langbeinite
calcium sulfate
Ex. added "gypsum" natural anhydrite
gypsum
hemihydrate
soluble anhydrite

Gypsum dissolves slowly

Natural anhydrite dissolves very slowly
Hemihydrate & soluble anhydrite dissolve very rapidly

SO3 ex. hemihydrate & soluble anhydrite has been referred as "D.SO3"

Clinker reactivity is strongly influenced by:

C3A level
Cement fineness
Surface "freshness"
Alkali:sulfate balance
Thermal history (e.g. crystal sizes)

Excess alkalis over sulfate results in a more reactive clinker

Need to consider availability of soluble calcium sulfate

Need to consider clinker "reactivity"
Need to match sulfate availability to clinker reactivity

Balanced retardation of C3A provides optimum water demand

Excess D.SO3 produces "FALSE SET" tendency
Insufficient D.SO3 produces "FLASH SET" tendency
"FLASH SET" tendency probably more negative

1.

CEMENT CHEMISTRY

16