Beruflich Dokumente
Kultur Dokumente
Nicholas B Winter
NOTICE:
This e-book is sold for individual use only and may be stored on a single
computer. It MAY NOT be stored on any computer network or other
retrieval system that allows access by more than one person.
You MAY print a single copy for your own personal use.
You also MAY NOT give away, sell or share the content herein.
If you obtained this e-book from any source other than through
http://www.understanding-cement.com, you have a pirated copy. Please
help stop Internet crime by reporting this to: mailto:support@whd.co.uk
Introduction vi
7 Composite Cements 84
7.1 Introduction 84
7.1.1 Background 84
7.1.2 Difference between pozzolanic and latently hydraulic
mineral additions 85
7.1.3 Effects of mineral additions on hydration products 85
and paste microstructure 85
7.1.4 Summary of benefits of using mineral additions 87
7.1.5 Potential problems of using mineral additions 88
7.2 Blastfurnace slag 88
7.2.1 Blastfurnace slag composition 88
7.2.2 Blastfurnace slag as a cementitious material 91
7.2.3 Hydration products in mixes containing slag 93
7.3 Low-lime fly ash 93
7.3.1 Fly ash composition 93
7.3.2 Fly ash as a cementitious material 94
7.3.3 Hydration products in mixes containing fly ash 96
7.4 Microsilica (silica fume) 98
7.5 Limestone 99
This e-book is an informal introduction to cement. Perhaps you are just starting a
cement or concrete-related career; maybe you are a student, or perhaps you
already know quite a bit about cement but want to refresh your mind on some
cement chemistry. Whatever your interest in cement, if you are looking for a
basic guide to cement science, I believe that you will find this e-book helpful.
No one fully understands cement; the title of this e-book should really be
something like Understanding Cement a Bit Better. Cement is like any other
fascinating subject; the more you find out, the more you realise just how huge
the gap really is between what you know and what there is to be known.
Ive worked with cement for nearly thirty years, first at the research division of a
major cement producer, then as an independent consultant and Im only too
happy to talk about cement with anyone who is interested, so thank you for
reading this e-book. Much of my work has been to do with cement and concrete
microscopy and as a result, this e-book is very visual. They say that a picture is
worth a thousand words, and in the following pages there are a lot of images
taken using optical and scanning electron microscopes; pictures really do help in
gaining an understanding of how cement is made and how it works.
Until recently, cement generally meant Portland cement, the normal grey
powder, and Portland cement forms the main subject of this e-book as it is still
the principal cement used.
The material presented here is either mainstream cement science that has
undergone peer review before publication in journals or textbooks, or is industry
data regarded as non-controversial. Of course, if you should find something you
disagree with, please contact me via the Understanding Cement web site and I
would be delighted to discuss it.
Understanding Cement has its origins in a one-day seminar on cement that
Arthur Harrisson and I first gave back in 1991. Since then, the seminar content
and cement science have both evolved, and the technology for presenting it has
progressed hugely, from handwritten overhead projector slides to the internet.
This e-book is intended to be a fairly gentle introduction into the slightly arcane
world of cement science and I very much hope that the basic grounding it
provides will enable you to go on and read with confidence some of the many
other textbooks available on cement and concrete. Excellent though they mostly
are, they can be a little daunting at first, so Understanding Cement is here to
get you started.
References are provided at the end of each chapter; many are references to
more detailed treatments in some of the standard works on cement, particularly
Taylors Cement Chemistry and Leas The Chemistry of Cement. There are
also some references to individual papers where these are especially relevant.
However, I have limited the range of external references to some extent because
not all readers will have easy access to a technical library where these often
slightly obscure references may be obtained. It is the purpose of this e-book to
be helpful and encouraging, not to handicap anyone who does not have a
university library next door. In Appendix 1 are my suggestions for the nucleus of
a small cement library that will be of value for many years to anyone interested
in cement science. Together, these books contain more references than anyone is
likely to want to read in a lifetime of cement study.
I am very grateful to the many people who have helped me with this e-book,
especially the two groups of reviewers who so kindly read the draft version. The
expert reviewers were people who have spent their working lives in cement
manufacturing, concrete production, masonry, education or related fields; they
contributed greatly to the technical content. The readability reviewers had a
technical background but no specific knowledge of cement; they identified many
areas where I had not explained the subject matter sufficiently clearly. I am also
very grateful to those who have helped in other ways too numerous and diverse
to describe in detail. I particularly thank Don Ashcroft, Geoff Bowler, Tom
Burnham, Mike Burton, Josette Camilleri, Mike Connell, Ian Ferguson, Ron Green,
Arthur Harrisson, Paul Livesey, Robert Matthews, Kelly Park, Lindon Sear,
Anthony Tidder and Anna Wright.
NOTES
Spelling
The spelling used is standard British English, except for sulphur and related
words where I have adopted the International Union of Pure and Applied
Chemistry (IUPAC) standard spelling of sulfur.
Microscope images
This e-book contains many images taken using either a scanning electron
microscope (SEM) or petrographic microscope. SEM images are black-and-white
images as electrons dont have colour; petrographic microscope images are in
colour.
Microscope images have scale bars to indicate the size of features. These are
shown in microns, or micrometres (m). A micron is 10-6 metre, so there are
1,000 microns in a millimetre and 1,000,000 microns in a metre.
The microscope images do not show the magnification (eg: x1000) since this will
vary depending how the image is displayed. An image that is x100 displayed full-
screen on a 15-inch computer monitor will be x160 on a 24-inch monitor, so
indicating a fixed magnification is meaningless unless you know the size of the
original image. It is the scale bar that is important, not the magnification.
If you want to work out the magnification of an image, measure the scale bar as
displayed on your monitor, or on the page if you have printed it, and work it out.
For example, if a 1000 m (=1 mm) scale bar measures 10 mm on your monitor,
the magnification is x10. If a 100 m scale bar measures 10 mm (=10,000 m),
the magnification is 10,000/100 = x100. If a 50 m scale bar measures 20 mm,
the magnification is 20,000/50 = x400.
Glossary
You may find it helpful to download the cement glossary, available if you sign up
to the free Newsletter, from www.understanding-cement.com.
Contact details
Feedback
The best way to send feedback is to use the feedback form at:
www.understanding-cement.com/ucebookfeedback.html
If you have any technical problems with the e-book (eg: printing it out), contact
us using the Contact Form on the Understanding Cement web site and we will try
to help.
http://www.understanding-cement.com/contact.html
Consultancy
ceminfo@whd.co.uk
Nick Winter
December 2009
Cementing materials were used widely in the ancient world. The Egyptians used
calcined gypsum as a cement. The Greeks and Romans used lime made by
heating limestone, and then added sand to make mortar, with coarser gravel for
concrete.
The Romans found that a cement could be made which set under water and this
was used for the construction of harbours. The cement was made by adding
crushed volcanic ash to lime, and was later called a pozzolanic cement named
after the village of Pozzuoli near Vesuvius. Where volcanic ash was scarce,
crushed brick or tile was used instead. The Romans were therefore probably the
first to manipulate the properties of cementitious materials for specific
applications and situations.
Marcus Vitruvius Pollio, a Roman architect and engineer in the 1st century BCE
wrote in his Ten books of Architecture, a revealing insight into ancient
technology (1):
The Romans used pozzolanic material more widely than just in marine
construction. Vitruvious says:
First I shall begin with the concrete flooring, which is the most important
of the polished finishings, observing that great pains and the utmost
precaution must be taken to ensure its durability.
A brief history of cement
On this, lay the nucleus, consisting of pounded tile mixed with lime in
the proportions of three parts to one, and forming a layer not less than
six digits thick.
After the end of the Roman Empire, much of their building technology was lost -
certainly in Europe with regard to cement. Mortars hardened mainly by
carbonation of lime, a slow process, and it wasnt until the late Middle Ages that
the benefits of pozzolanic material mixed with lime were rediscovered. The great
European mediaeval cathedrals were clearly built by highly skilled masons but it
would probably be fair to say they did not have the technology to manipulate the
properties of cementitious materials in the way the Romans had done a thousand
years earlier.
The Renaissance and Age of Enlightenment brought new ways of thinking and
compelling new reasons to develop technology, including cement technology. For
example, maritime nations needed to build lighthouses on exposed rocks to
reduce shipping losses.
John Smeaton, the father of civil engineering in England, found that a mix of
hydraulic lime, produced by burning an impure limestone and then mixing the
lime with a natural pozzolan, produced a mortar that hardened under water (2).
Smeaton used his mortar with interlocking stone blocks in the construction of the
third Eddystone lighthouse (1759) off the coast of Cornwall in Southwestern
England. This mixture for a mortar of lime and pozzolan was specified for
government contracts until 1867, 43 years after Aspdins patent for Portland
cement.
www.whd.co.uk 2 www.understanding-cement.com
A brief history of cement
Joseph Aspdin took out a patent in 1824 for Portland Cement, a material he
produced by firing a mixture of finely-ground clay and limestone. What Aspdin
did that was different was to calcine the limestone, then mix the lime with clay
and fire it again. He called the product Portland Cement because the concrete
made from it looked like Portland stone, a widely-used building stone in England.
A ship carrying barrels of Williams cement went aground off the Isle of Sheppey
in Kent in 1848 and the barrels of set cement, minus the wooden staves, were
later incorporated into the wall of an inn in Sheerness. They are still there now
and the landlord occasionally receives requests from cement enthusiasts for small
pieces.
In 1845 Williams main competitor, Isaac Johnson, had also made an improved
cement, by firing a mixture of chalk and clay at higher temperatures (1400 C
1500 C), similar to those used today. Because both William Aspdin and Johnson
burnt at higher temperatures than had been used previously, minerals were
produced which were very reactive and more strongly cementitious.
While they both used similar materials and temperatures to make Portland
cement as we use now, three more important developments in the manufacturing
process lead to modern Portland cement:
A rotary kiln is a long, rotating, tube slightly tilted from the horizontal and will
described more fully in Chapter 3. Ransomes kiln did not work properly in trials
in 1887. Later, Hurry and Seaman in the USA resolved the problems and
produced the first modern rotary kiln. These gradually replaced the original
vertical shaft kilns from the early 1900s.
Rotary kilns heat the clinker mainly by radiative heat transfer and this is more
efficient at higher temperatures, enabling higher burning temperatures to be
achieved. Also, rotary kilns give a more consistent product because the clinker is
constantly moving inside the kiln; this gives a more uniform temperature in the
burning zone, the hottest part of the kiln.
www.whd.co.uk 3 www.understanding-cement.com
A brief history of cement
This brief and selective whirlwind tour of cement development sets Portland
cement in its historical context. If youre particularly interested in cement history,
the Blezard paper and his chapter in Lea referred to below, contain more
information with numerous references.
References, Chapter 1
Further reading
www.whd.co.uk 4 www.understanding-cement.com
2 Cement basics
Cement may be pure Portland cement, or, alternatively, it may be made from
Portland cement mixed with other materials that also have cementitious
properties, such as blastfurnace slag from iron smelting or fly ash from coal-fired
electricity power stations.
Pure Portland cement was the most widely used cement in most parts of the
world until recently. In some parts of the world, it still is but technical
developments as well as environmental concerns have led to the increased use of
other cements, particularly composite cements.
The meanings of the words cement and concrete are rather blurred in general
use, so lets start by defining these words more carefully, along with a few
others.
Cement basics
Portland cement itself can be divided into a number of different types, each
cement having different characteristics.
Normal, grey, cement for general-purpose use will be referred to in this e-book
as ordinary Portland cement. Other types of Portland cement include white
Portland cement and sulfate-resisting Portland cement, of which more later.
Lime concrete/mortar
Expansive cements
These cements used to be called high alumina cements. They are made from
lime or limestone mixed with bauxite (aluminium ore) or other high-alumina
material heated in a furnace until the raw materials have completely melted, then
cooled and the solid material is then ground to produce cement.
www.whd.co.uk 6 www.understanding-cement.com
Cement basics
Concrete made with CAC typically develops strength quickly and is resistant to
chemical attack. CACs have a wide range of compositions, mainly with different
ratios of lime to alumina, and they are generally brown, grey or black in colour.
One common type of CAC, Ciment Fondu, is grey but CACs can also be white if
made from pure alumina.
As well as being used in concrete, CACs are also used in grouts and other
specialised applications, often mixed with Portland cement and other materials
such as gypsum.
Lime mortar and concrete was used for thousands of years until eclipsed by other
cements particularly Portland cement. Its use declined to a very low level towards
the end of the 20th century, being used mainly in the rebuilding or repair of
historic or ancient buildings. More recently, the use of lime mortar and concrete
in the construction of new buildings has increased in some countries, the UK
being an example. Although it is not as strong, there are benefits in using lime
mortar instead of a mortar based on Portland cement, in particular:
Lime mortar is usually weaker than mortars made with Portland cement
and so can be removed from the brick or stone at the end of the useful life
of the building. Particularly in the case of bricks, this means that they can
be recycled, saving energy otherwise needed to make new bricks. If
mortar made with Portland cement is used, bricks generally cant be re-
used as it is almost impossible to remove the mortar.
www.whd.co.uk 7 www.understanding-cement.com
Cement basics
Expansive cements
www.whd.co.uk 8 www.understanding-cement.com
3 Portland cement composition and microstructure
When cool, the clinker is ground up to a fine powder and a small amount of
gypsum is added to control the setting properties of the cement. Here are a few
familiar, and perhaps less familiar, views of cement.
Figure 3.1 Bulk cement road tankers on a Figure 3.2 Cement bagging plant.
quay. (Photo courtesy Rugby Cement.) Bagged cement is sold mainly to small
builders and the do-it-yourself market.
(Photo courtesy Rugby Cement.)
Portland cement composition and microstructure
Figure 3.3 The familiar grey powder. The Figure 3.4 Cement particles viewed in a
coin is 23mm in diameter. scanning electron microscope.
www.whd.co.uk 10 www.understanding-cement.com
Portland cement composition and microstructure
Alite
Belite
A calcium aluminate phase
A calcium alumino-ferrite phase
Note that system includes temperature. Cooled clinker contains four main
phases the four principal clinker minerals but in the burning zone of the kiln,
it contains mainly alite, belite, free lime and the liquid phase. (Free lime is lime
that has not yet combined to form alite or belite or other minerals).
The compositions of the four main clinker minerals in any particular cement
clinker vary a little, depending on the compositions of the raw materials. Their
compositions are often simplified by approximating them to the following pure
compounds:
www.whd.co.uk 11 www.understanding-cement.com
Portland cement composition and microstructure
Table 3.1 Mineral names and approximate compositions of the four principal
clinker minerals.
Mineral name Approximate composition Formula
Alite: tricalcium silicate (Ca3SiO5)
Belite: dicalcium silicate (Ca2SiO4)
Calcium aluminate phase: tricalcium aluminate: (Ca3Al2O6)
Calcium alumino-ferrite phase: tetracalcium aluminoferrite: (Ca4Al2Fe2O10) or
(Ca2AlFeO5)
Cement analyses are usually shown as oxides, eg: CaO or SiO2. This is a
standard form of analysis in physical chemistry and geology. For anyone
unfamiliar with it, the following example may help:
www.whd.co.uk 12 www.understanding-cement.com
Portland cement composition and microstructure
However, the proportion of silicon to oxygen in quartz has a fixed ratio of 2:1,
determined by the valency of silicon and oxygen. Silicon has a valency of +4 and
oxygen -2; this means that a silicon atom can share four electrons with
neighbouring atoms and oxygen can share two. Since quartz is not electrically
charged, it follows that each silicon atom will bond with two oxygen atoms:
+4 + (2 x -2) = 0
This means that it isnt necessary to show the oxygen separately; the data in
Table 3.4 gives the same information as that in Table 3.3, but in a more
convenient form.
For more information, see Reference 2 at the end of this chapter, or try a basic
chemistry textbook.
Chemical compositions of cement clinker phases are expressed as oxides and this
is convenient in many ways but can lead to rather long-winded formulae. To
simplify them, cement chemists have adopted a form of notation that seems
strange to chemists not familiar with cement.
Using cement chemistry notation, the formulae are abbreviated. Remember that
these compositions are approximate because the minerals contain impurities.
(With names like tetracalcium aluminoferrite, the need for snappier names is
all-too clear.)
In other words, for each of the clinker main minerals, we now have at least three
possible descriptions, as below:
www.whd.co.uk 13 www.understanding-cement.com
Portland cement composition and microstructure
There are also the full chemical formulae for the pure compounds, eg: C3S or
Ca3SiO5.
Although strictly, these do not mean the same thing, they are frequently used
indiscriminately. This lax use means that names like C3A should not usually be
taken to signify a definite composition in the sense of a pure compound, unless
this is indicated by the context.
_ _
S and C are spoken as S-bar and C-bar respectively, but are less used now
than they once were. (Probably, they are a victim of technology - fine when notes
were hand-written or typed on a typewriter, but just too complicated to do on a
word-processor. Have a try and see why it isnt used so much these days!)
People unfamiliar with cement notation sometimes have trouble adjusting to the
notion of S representing SiO2, for example, but do stick with it it is usually a
great time saver and it does work.
www.whd.co.uk 14 www.understanding-cement.com
Portland cement composition and microstructure
Figure 3.6 Single clinker nodule and a Figure 3.7 Polished section containing
coin, 23mm in diameter, for scale. pieces of clinker nodules. Section diameter
is 30 mm.
Clinker nodules are porous and it is desirable to fill these pores with epoxy resin;
nodules are embedded in the resin using vacuum impregnation to force the resin
into as many small pores as possible. In Figure 3.7 the nodules have been
separated into different size fractions by sieving and the fragments held in place
during specimen preparation by card divisions, which remain visible.
When the epoxy resin had set, the hardened resin block was sawn to reveal the
nodules in section, then polished using diamond polishing compound in
successively finer grades.
For examination using an optical microscope, various etches liquids that react
with the surfaces of the crystals - may be used to highlight particular minerals.
Etches are not normally used for SEM examination.
www.whd.co.uk 15 www.understanding-cement.com
Portland cement composition and microstructure
Figure 3.8 Polished section of Portland Figure 3.9 Polished section of Portland
cement clinker, examined in reflected light cement clinker, backscattered SEM image.
using a petrographic microscope, showing Alite (a) and belite (b) can be clearly
crystals of individual minerals. Brown distinguished, alite being light grey and
crystals are alite and blue crystals are belite darker grey examples arrowed.
belite. The bright material between the Black areas are epoxy resin. Small bright,
alite and belite is a mixture of aluminate nearly white, crystals are ferrite and tiny
and ferrite. Grey areas are pores filled darker crystals close to the ferrite are
with epoxy resin. The colours are the mainly aluminate phase, but these are
product of etching; the section was etched hard to distinguish at this magnification.
with hydrofluoric acid vapour in order to SEM images are always black and white,
distinguish between the different unless they have had false colours added.
minerals. Belite is not actually blue and This is a different clinker to that in Figure
alite is not brown. 3.8 but the magnification is similar.
Some nodules are highly porous and others much denser. The small nodule in
Figure 3.10 is of fairly typical density. In this low-magnification view of most of a
nodule, alite is the mid-grey mineral comprising the bulk of the nodule. The
darker patches are belite clusters.
www.whd.co.uk 16 www.understanding-cement.com
Portland cement composition and microstructure
Ferrite and aluminate phase vary in appearance from one clinker to another and
within the same clinker. Sometimes ferrite and aluminate crystals are small and
intergrown but they can also be coarse separate blocky crystals (Figure 3.11).
www.whd.co.uk 17 www.understanding-cement.com
Portland cement composition and microstructure
When the clinker is ground up to produce cement, it is obvious that much of the
microstructure of the nodules is lost (Figure 3.12).
Alite (C3S)
Alite is very reactive in the presence of water. It is the main constituent of
Portland cement, typically between 45% and 70% of the clinker by weight. It is
the main strength-giving component of cement and is thus strongly hydraulic.
Belite (C2S)
Belite is reactive and hydraulic, but less so than alite. The proportion of belite in
clinker typically varies between 5% and 30%.
Aluminate (C3A)
The aluminate phase is the most reactive constituent of clinker. However, it is
not a major contributor to strength in Portland cement, in other words it is only
very weakly hydraulic. Although only weakly hydraulic, because it is very reactive
it has a strong influence on early setting properties of cement and it produces a
lot of heat. In order to control this reactivity, gypsum is added to cement to
control the aluminate phase hydration.
www.whd.co.uk 18 www.understanding-cement.com
Portland cement composition and microstructure
Ferrite (C4AF)
Ferrite phase is moderately reactive, but only weakly hydraulic. It reacts quickly
initially when mixed with water but the rate of reaction slows and unhydrated
ferrite can be found in concrete over a hundred years old. Ferrite is black and
gives cement its characteristic grey colour.
If alite and belite are the main hydraulic minerals in cement, the minerals that
give strength to concrete, what do the aluminate and ferrite phases do? Why not
just make cement containing alite and belite?
The short answer is that it would be very difficult to make cement that did not
contain at least some aluminate or ferrite. Well be looking at the reactions in
the cement kiln later, but broadly cement-making is a clinkering (sintering)
process. This means not all the material in the kiln melts at the hottest part of
the kiln, the burning zone, about three-quarters of the clinker remains solid and
only a quarter is liquid.
Remember that what we are trying to do in the kiln is make calcium silicates,
particularly alite. The solid material entering the burning zone of the kiln is,
roughly speaking, belite and free lime that we want to combine to make alite.
The liquid is mainly composed of oxides of calcium, iron and aluminium; when
cooled, this liquid crystallises into the aluminate and ferrite phases.
The rle of the liquid is to accelerate the reactions in the clinker. Ions are
transferred through the liquid much more easily than through a solid. All other
things being equal, the higher the proportion of liquid, the easier it is to combine
the belite and free lime to make alite. In cement-making terminology, the liquid
is also called the flux.
The liquid from which aluminate and ferrite crystallise is therefore critical to the
process of making cement. Without the liquid, ion transfer would be much
slower; belite and free lime would not combine adequately. The liquid is essential
and the aluminate and ferrite form from the liquid is it cools, even though they
do not contribute greatly to the strength of concrete.
www.whd.co.uk 19 www.understanding-cement.com
Portland cement composition and microstructure
The quantities of each of the main oxides (CaO, SiO2, Al2O3 and Fe2O3) in
the raw materials.
The extent to which they have combined to form the main clinker phases.
If we know the composition of the clinker from its oxide analysis, and if we also
know the compositions of the four main clinker minerals, we can calculate how
much of each mineral is present.
The calculation used will vary slightly, depending whether you are applying it to
cement or to clinker, and on what you are trying to achieve:
www.whd.co.uk 20 www.understanding-cement.com
Portland cement composition and microstructure
To apply the calculation to cement, you need to take into account any
material that was added to the clinker, such as gypsum and fine
limestone.
The following version of the Bogue calculation is applied to clinker and corrects
for free lime; it does not allow for added gypsum or fine limestone and so should
not be used for cement in this form. Well look at applying the calculation to
cement later (Chapter 5.9).
The calculation assumes that the four main clinker minerals have the actual
compositions given by:
(At the risk of becoming repetitive, these are only approximations of the true
compositions of the minerals.)
To make clinker, we are combining lime and silica and also lime with alumina and
iron. If some of the lime remains uncombined, we need to subtract this from the
total lime content before we do the calculation, or we will overestimate the actual
alite content. For this reason, a clinker analysis normally gives a figure for free
lime.
Thirdly, we assume all the silica is present as belite and calculate how
much lime is needed to form belite from the total silica content of the
clinker. There will be a surplus of lime.
www.whd.co.uk 21 www.understanding-cement.com
Portland cement composition and microstructure
In practice, the above process of allocating the oxides can be reduced to the
following equations:
C3S = 4.0710CaO-7.6024SiO2-1.4297Fe2O3-6.7187Al2O3
C2S = 8.6024SiO2+1.0785Fe2O3+5.0683Al2O3-3.0710CaO
C3A = 2.6504Al2O3-1.6920Fe2O3
C4AF = 3.0432Fe2O3
Table 3.6 An example of a typical clinker analysis (note that the analysis in Table 3.2
was of a cement, not a clinker).
SiO2 Al2O3 Fe2O3 CaO MgO K2O Na2O SO3 LOI IR Total
21.5 5.2 2.8 66.6 1.0 0.6 0.2 1.0 1.5 0.5 98.9
Free lime = 1.0% CaO
Combined CaO = (Total CaO Free CaO) = (66.6% - 1.0%) = 65.6% This is the
figure we use for CaO in the calculation.
C3S = 4.0710CaO-7.6024SiO2-1.4297Fe2O3-6.7187Al2O3
C2S = 8.6024SiO2+1.1Fe2O3+5.0683Al2O3-3.0710CaO
C3A = 2.6504Al2O3-1.6920Fe2O3
C4AF = 3.0432Fe2O3
Therefore:
www.whd.co.uk 22 www.understanding-cement.com
Portland cement composition and microstructure
So:
C3S = 64.7%
C2S = 12.9%
C3A = 9.0%
C4AF = 8.5%
Total = 95.1%
To the total, we can add the 1% free lime we deducted at the start to give
96.1%; note that the total still does not add up to 100%. This is because the
calculation neglects impurities in the four clinker minerals, and because of other
minor constituents in the clinker not accounted for, such as clinker sulfate and,
possibly, periclase (MgO).
It should be stressed that the Bogue calculation does not give the exact amounts
of the four main clinker phases present, although this is sometimes forgotten.
These differ from the true amounts mainly because the actual mineral
compositions differ a little (occasionally more than a little) from those assumed in
the calculation, as shown in Table 3.7.
Table 3.7 Comparison of clinker mineral compositions assumed by the standard Bogue
calculation, with typical compositions.
SiO2 Al2O3 Fe2O3 CaO MgO K2O Na2O SO3 Total
Bogue C3S 26.3 0.0 0.0 73.7 0.0 0.0 0.0 0.0 100
Bogue C2S 34.9 0.0 0.0 65.1 0.0 0.0 0.0 0.0 100
Bogue C3A 0.0 37.7 0.0 62.3 0.0 0.0 0.0 0.0 100
Bogue C4AF 0.0 21.0 32.9 46.1 0.0 0.0 0.0 0.0 100
Typical
25.2 1.0 0.7 71.6 1.1 0.1 0.1 0.1 99.9*
Alite
Typical
31.5 2.1 0.9 63.5 0.5 0.9 0.1 0.2 99.7*
Belite
Typical
3.7 31.3 5.1 56.6 1.4 0.7 1.0 0.0 99.8*
Aluminate
Typical
3.6 21.9 21.4 47.5 3.0 0.2 0.1 0.0 97.7*
Ferrite
*Balance is mainly P2O5, Mn2O3, TiO2, Cr2O3
Typical mineral compositions are from Reference 3 at the end of this chapter.
It is clear from Table 3.7 that there are appreciable differences between the
compositions assumed by the Bogue calculation and those typical of minerals in
cement clinker. For example, alite and belite both contain aluminium and iron,
and the aluminate phase (C3A) is not strictly tricalcium aluminate as it contains
some iron and silicon.
That said, the Bogue calculation gives a useful approximation and is widely used
within the cement industry. It provides a standardised format for estimating the
clinker mineral proportions and the approximations made are not normally
www.whd.co.uk 23 www.understanding-cement.com
Portland cement composition and microstructure
important.
Of much greater interest are relative changes over time in the indicated mineral
content rather than the absolute values. If concrete early strengths suddenly
drop and if alite content as calculated using the Bogue calculation also shows a
drop, the lower strength may well be due to the lower alite content.
The calculation can be expressed in other ways, of which perhaps the most useful
is as a set of four simultaneous equations. In the following, (A)= alite;
(B)=belite; (Al)=aluminate; (F)=ferrite and (bulk) refers to the oxide
composition of the clinker.
Note that C(bulk) should have the free lime subtracted before the calculation for the
best estimate of the alite content.
Similar calculations can be performed for cement, provided that gypsum and any
other added material, such as fine limestone, is taken into account (see Chapter
5.9).
In the above calculation, we subtracted the free lime from the total CaO content
in order to obtain the best estimate of the alite content. However, as mentioned
above, if used prescriptively in a standard specification the form of the calculation
specified may not require free lime to be deducted.
In the ASTM standard for Portland cement, ASTM C150-07, the free lime content
is not deducted from the total lime in calculating the phase composition; it isnt
even mentioned. This is because the objective of the calculation is different. The
specification defines limits to the phase composition of certain cement types to
ensure that the cements are suitable for specific purposes. Differences in free
lime content in individual cements are not an issue and would be an unnecessary
complication.
www.whd.co.uk 24 www.understanding-cement.com
Portland cement composition and microstructure
To see how alite content and free lime content are related, look at the effect of
free lime on the mineral content of a typical clinker, as shown by the Bogue
calculation (Table 3.8).
Applying the Bogue calculation gives an alite content of about 69% if no free lime
is assumed to be present (Clinker A) but only about 61% if 2% free lime is
assumed (Clinker C). This illustrates why a full clinker analysis should also show
the free lime content if the purpose of the analysis is to indicate the actual phase
proportions of the clinker.
Cements made from Clinkers A, B and C in Table 3.8 would probably perform
differently. If early strengths were dependent solely on alite content, Clinker A
with the highest alite content should give the best early strengths (from about 1
day to 7 days). With 2% free lime (Clinker C) the lower alite content should
result in the lowest early strengths. At later ages, the higher belite content of
Clinker C would start to partly close the gap with Clinker A, although this may
take several weeks or months.
However, the mineral proportions are not the only factors controlling the
performance of a cement, although they are very important. For example, as
mentioned previously a hard-burned clinker may contain calcium silicates which
are less reactive than those in a clinker burned with a lighter touch.
In other words, Clinker A with little or no free lime, was likely to have been very
hard-burned in order to get all the lime to combine. Clinker C would be much
lighter-burned (all other things being equal) and Clinker B somewhere in-
between. Clinker B might well have the best compromise between alite content
and silicate reactivity and cement made from it may out-perform cements made
from Clinkers A and C in terms of both early and late strengths.
www.whd.co.uk 25 www.understanding-cement.com
Portland cement composition and microstructure
lower density. Harder burning would cause the nodule to coalesce. Nodule
density is therefore, very approximately, a measure of how hard the clinker was
burned.
Indeed, the litre weight (weight of a known volume, ie: density - more strictly
the bulk density) of a clinker is often used as a rough measure of burning.
Usually, the clinker sample for calculating the litre weight has been sieved, so
that the nodule sizes are within defined upper and lower limits.
Figure 3.13a This clinker nodule is Figure 3.13b Detail of Figure 3.13a,
underburned; it was not at burning showing free lime (FL), alite (a) and
temperature for a sufficient time for adequate belite b). Harder burning would have
combination to take place. This image shows resulted in a densification of the
much belite (dark grey) in large, arcuate, microstructure and enabled more lime to
clusters. Some alite is present (light grey) combine with belite to produce more alite.
and much free lime (FL - white). This nodule
has an open, porous, structure and contains
more belite than alite. Compare with the well-
combined nodule in Figures 3.10-3.11 shown
at a similar magnification.
www.whd.co.uk 26 www.understanding-cement.com
Portland cement composition and microstructure
The LSF is a ratio of CaO to the other three main oxides. Applied to clinker, it is
calculated as:
The LSF controls the ratio of alite to belite in the clinker. A clinker with a higher
LSF will have a higher proportion of alite to belite than will a clinker with a low
LSF.
Values above 1.0 (100%) indicate that free lime is likely to be present in the
clinker. This is because, in principle, at LSF=1.0 all the free lime should have
combined with belite to form alite. If the LSF is higher than 1.0, there will be
excess lime that has nothing with which to combine.
In practice, there are always regions within the clinker where the LSF is locally a
little above, or a little below, the average for the clinker as a whole. This is partly
due to imperfect mixing and partly to particle size; a large particle will cause the
composition locally to deviate from the average composition. This means that
there is almost always some residual free lime, even where the LSF is
considerably below 1.0. It also means that to convert all the belite to alite, an
LSF slightly above 1.0 is needed; there will inevitably be free lime remaining.
SR = SiO2/(Al2O3 + Fe2O3)
A high silica ratio means that more calcium silicates are present in the clinker and
less aluminate and ferrite. SR is typically between 2.0 and 3.0.
AR=(Al2O3/(Fe2O3)
This determines the potential relative proportions of aluminate and ferrite phase
in the clinker.
www.whd.co.uk 27 www.understanding-cement.com
Portland cement composition and microstructure
A simple question, you may think, but there isnt a simple answer. It depends on
the requirements of the concrete to be made from the cement. For example, if
high early strengths were required, we would need:
A high LSF this means the bulk of the calcium silicates will be in
the form of alite.
Reactions in the kiln are harder to achieve with high LSF and SR mixes - they are
more difficult to burn, or to combine. This results in higher costs, and also in
increased CO2 emissions due to the higher CaO content of the cement and an
increased fuel requirement.
www.whd.co.uk 28 www.understanding-cement.com
Portland cement composition and microstructure
These different Portland cement variations are treated differently by the various
national standards; ASTM C 150 lists five different types of Portland cement, plus
three air-entrained sub-types, not all of which have equivalents in other
standards. Chapter 10 has more on standard specifications.
Concrete made with sulfate-resisting cement remains sounder for longer in more
extreme conditions of exposure to sulfate solution compared with ordinary
Portland cement. It can do this because it contains a minimal amount of
aluminate phase. There are several different forms of sulfate attack and these
will be considered later (Chapter 9.2).
So, how can we make cement that contains less aluminate phase?
In theory, we could have less flux overall by increasing the silica ratio.
In practice, if we increase the silica ratio too much, there will be insufficient liquid
at the burning temperature and good combination will be more difficult to
achieve, so this isnt a good option.
If we cant have much (or any) aluminate phase in the clinker, we will need to
lower the alumina ratio. This may mean that harder burning is necessary to
achieve combination (well see why in Chapter 5.3) but we can offset this by
having more liquid.
In other words, we can make sulfate-resisting cement by adjusting the silica ratio
downwards, as well as the alumina ratio. This means more liquid is formed
overall, but almost all of it will crystallise as ferrite. We cant take alumina out of
the raw materials, but we can add iron.
The ferrite that crystallises from the liquid in sulfate-resisting cement is more
iron-rich than the ferrite in ordinary Portland cement. Because of this, the normal
assumption made by the Bogue calculation that ferrite has the composition C4AF
is even more wrong than it is for normal clinker. ASTM C150 provides an
alternative calculation for use with Type V (sulfate-resisting) cement.
www.whd.co.uk 29 www.understanding-cement.com
Portland cement composition and microstructure
blastfurnace slag (gbs), or ordinary Portland cement and fly ash also provide
good resistance to sulfate solutions. These composite cements are increasingly
used to make sulfate-resisting concrete. In Europe, composite cements have
largely taken over from sulfate-resisting cement where sulfate resistance is
required; in the UK, for example, the demise of sulfate-resisting Portland cement
is almost complete, as it is no longer produced.
White Portland Cement is used for architectural purposes where a white concrete
finish is wanted. It is not treated as a separate type of cement in national
standards but is produced to meet the same criteria as normal grey cement.
Normal cement contains ferrite phase. Ferrite is dark grey or black, giving
cement its characteristic grey colour. So, to make white cement, we need to
make cement that contains little or no ferrite phase.
The principal element that gives the dark colour to ferrite is iron, at least in terms
of quantity. Other elements, particularly chromium and manganese are also
strong colorants. Titanium, copper and vanadium are also potential sources of
colour. In other words, we want as little of these elements as possible in the
clinker, achieved by careful selection of raw materials. Magnesium may also affect
colour.
Typical raw materials used for making white cement are pure limestone, such as
chalk, china clay and silica sand.
Of course, there is still likely to be a little iron, and other undesirable elements,
present in the raw materials. To prevent this from forming ferrite, white clinker
is usually burnt under slightly reducing conditions, converting Fe(III) to Fe(II).
Fe(II) substitutes for calcium and then resides in the clinker minerals, thus
avoiding ferrite formation and discolouration. The Bogue calculation will
overestimate the true ferrite content of a white clinker, as it assumes that all iron
is present in ferrite. White cement clinker is commonly water-quenched; the
rapid temperature decrease helps to prevent Fe(II) oxidising to Fe(III).
As well as being more finely ground, the cement also may have a higher alite
content than typical ordinary Portland cement.
www.whd.co.uk 30 www.understanding-cement.com
Portland cement composition and microstructure
References, Chapter 3
www.whd.co.uk 31 www.understanding-cement.com
4 Portland cement manufacturing - the main
components of a cement plant
If you happen to be a geologist, the quarry is probably the most interesting part
of a cement works (although the kiln might be if you view the clinkering process
as igneous rocks in the making).
Figure 4.1 Limestone quarry. (Picture courtesy Figure 4.2 Limestone blocks being
Castle Cement.) removed for crushing after blasting.
(Picture courtesy Castle Cement.)
Clay, marl or shale (supplies the bulk of the silica, alumina and ferric
oxide)
If the good stuff is all used up first, it may be difficult to make cement out of
what is left. Detailed forward planning is needed to make the best use of all the
materials available.
Raw materials are extracted from the quarry, then crushed and milled as
necessary to provide a fine material for blending in the required proportions.
They may also need to be dried. The raw materials may be milled together or
separately, depending on how hard they are and whether additional mills are
available.
The blended raw material, the raw meal, is stored in a silo before being fed to the
kiln. The silo provides a stock of raw meal, enabling production to be maintained
for a day or two in the event of failure of equipment or in the supply of materials.
www.whd.co.uk 33 www.understanding-cement.com
Portland cement manufacturing - the main components of a cement plant
In a modern works, the blended raw material passes from the silo (A in Figure
4.3) to the preheater tower (B) and then to the kiln (C). In the preheater tower,
hot gases from the kiln, and often the cooled clinker at the far end of the kiln,
are used to heat the raw meal. As a result, the raw meal is already hot before it
enters the kiln.
Well look first at the general principle of the rotary kiln (Figure 4.4). Ignore the
whimsical wheelbarrow not standard equipment in a modern kiln.
A rotary kiln is basically a long cylinder rotating about its axis about 1-3 times a
minute. At one end of the kiln is a flame. The kiln is inclined at a slight angle,
the end with the flame being lower. The rotation of the kiln causes the raw meal
to gradually pass along the kiln from where it enters at the cool end, to the hot
end where it eventually drops out as clinker and cools.
www.whd.co.uk 34 www.understanding-cement.com
Portland cement manufacturing - the main components of a cement plant
The original rotary cement kilns were called wet process kilns and in their basic
form, they were relatively simple compared with new modern kilns. The raw
meal was fed into the kiln at ambient temperature in the form of a slurry.
Clearly, slurry contains water that has to be evaporated, perhaps 35% or more of
the total feed. This takes a lot of energy and various developments of the wet
process were aimed at reducing the water content of the raw meal. An example
of this is the filter press. Imagine a musical accordion 10-20 metres long and
several metres across and youll get the concept. Such adaptions were described
as semi-wet.
The wet process has survived for over a century because many raw materials are
suited for blending as slurry. Also, for many years, it was technically difficult to
get dry powders to blend adequately.
Quite a few wet process kilns are still in operation, usually now with higher-tech
bits bolted on. However, most new cement kilns are of the dry process type and
we will focus on these.
www.whd.co.uk 35 www.understanding-cement.com
Portland cement manufacturing - the main components of a cement plant
A dry process kiln is much more thermally efficient than a wet process kiln.
Firstly, and most obviously, this is because the meal is a dry (or nearly dry)
powder and so there is no water that has to be evaporated.
Secondly, and less obviously, the process of transferring heat is much more
efficient in a dry process kiln.
The basic dry process system consists of the kiln and a suspension preheater.
The suspension preheater is a heat exchanger housed in a tall tower, consisting
of a series of cyclones in which fast-moving hot gases from the kiln keep the
meal powder suspended in air. All the time, the meal gets hotter and the gas gets
cooler until the meal is at almost the same temperature as the gas. The high
open structure made of steel girders (B) in Figure 4.3 is a preheater tower
containing the cyclones.
The raw meal is fed in at the top of the preheater tower and passes through the
series of cyclones in the tower before entering the kiln. Hot gas from the kiln
and, often, hot air from the clinker cooler are blown through the cyclones. Heat
is transferred efficiently from the hot gases to the raw meal.
The heating process is efficient because the meal particles have a very high
surface area in relation to their size and because of the large difference in
temperature between the hot gas and the cooler meal. Typically, 30%-40% of
the meal is decarbonated before entering the kiln.
www.whd.co.uk 36 www.understanding-cement.com
Portland cement manufacturing - the main components of a cement plant
A development of this process is the precalciner kiln (Figure 4.7). Most new
cement kilns are of this type. The principle is similar to that of the dry process
preheater kiln but with the addition of another burner, called the precalciner. This
is placed in the preheater tower, either as a separate unit or within the riser
duct, conveying kiln gas from the kiln to the last preheater cyclone. The meal,
which is already hot, reaches about 900 C in a few seconds and about 85%-95%
of the meal is decarbonated before it enters the kiln.
In practice, a preheater tower is likely to have four to six stages; only three are
shown in Figure 4.7 for simplicity. Some preheater designs are much more
complex but all share the principle of maintaining the feed particles in suspension
in a flow of hot gases, while transferring heat from the gases to the feed.
Optimising the efficiency of this process will entail some elegant mathematics and
applied physics to determine airflows and the best shape for the ducting. In a
preheater tower, just small parts of the total system are visible at each floor
level; only in drawings can they be seen in their entirety and their graceful form
appreciated.
Since meal enters the kiln at about 900 C, rather than 20 C in a wet process
kiln, a dry process kiln can be shorter and of smaller diameter for the same
output. This reduces the capital costs of a new cement plant. A dry process kiln
might be only 70 m long and 6 m wide but produce a similar quantity of clinker
(usually measured in tonnes per day) as a wet process kiln of the same diameter
but 200m in length.
www.whd.co.uk 37 www.understanding-cement.com
Portland cement manufacturing - the main components of a cement plant
At A in Figure 4.7, most of the meal is decarbonated and the feed enters the kiln
at about 950 C. At B the feed has reached about 1100 C 1200 C and
consists largely of belite, free lime and intermediate minerals. At C, the burning
zone, the feed temperature reaches 1400 C 1500 C; clinker nodules form and
most of the remaining free lime reacts with belite to form alite. The next chapter
will go into this in more detail.
The kiln is made of a steel casing lined with refractory bricks. There are many
different types of refractory brick and they have to withstand not only the high
temperatures in the kiln but reactions with the meal and gases in the kiln,
abrasion and mechanical stresses induced by deformation of the kiln shell as it
rotates.
Bricks in the burning zone are in a more aggressive environment compared with
those at the cooler end of the kiln (the back end), so different parts of the kiln
are lined with different types of brick.
Periodically, the brick lining, or part of it, has to be replaced. Refractory life is
reduced by severe changes in temperature, such as occur if the kiln has to be
stopped suddenly. As the cost of refractories is a major expense in operating a
cement plant, kiln stoppages are avoided as far as possible.
As the meal passes through the burning zone, it reaches clinkering temperatures
of about 1400 C 1500 C. Clinker nodules form as the burning zone is
approached; when the clinker has passed the burning zone, it starts to cool,
slowly at first, then much more quickly as it passes over the nose ring at the
end of the kiln and drops out into the cooler.
www.whd.co.uk 38 www.understanding-cement.com
Portland cement manufacturing - the main components of a cement plant
Fuels
The most common fuels used in a cement kiln are probably coal and petroleum
coke. Gas is also used.
Recently, particularly in the last ten years or so, many cement producers have
reduced their consumption of fossil fuels by supplementing them with other
greener types of fuel (Chapter 12). Examples of these substitute fuels include
bonemeal, car tyres (chopped into small pieces, or chipped) and waste paper.
The purpose of a cooler is, obviously, to cool the clinker. This is important for
several reasons:
The cooler also heats the air entering the kiln around the main burner; hot air
(800-1000 C) gives a better flame.
www.whd.co.uk 39 www.understanding-cement.com
Portland cement manufacturing - the main components of a cement plant
The cooled clinker is then conveyed either to the clinker store or directly to the
clinker mill. The clinker store holds several weeks supply of clinker (or more), so
that deliveries to customers can be maintained when the kiln is not operating.
A mill will usually have two or three chambers, with different size grinding media.
As the clinker particles are ground down, smaller media are more efficient at
further reducing particle size.
www.whd.co.uk 40 www.understanding-cement.com
Portland cement manufacturing - the main components of a cement plant
In an open circuit mill, the clinker passes through the mill and the feed
rate of clinker into the mill is adjusted to achieve the desired fineness of
the product.
Cement from an open circuit mill has a wider particle size range than cement
from a closed circuit mill; in an open circuit mill, some particles don't get ground
sufficiently while others are ground extremely finely. Closed circuit milling is more
efficient because it concentrates the milling effort more effectively on those
particles that require further grinding.
Recently, an alternative type of mill, the vertical roller mill (VRM) has also
become popular because of its lower power consumption.
www.whd.co.uk 41 www.understanding-cement.com
Portland cement manufacturing - the main components of a cement plant
www.whd.co.uk 42 www.understanding-cement.com
5 Portland cement manufacturing from raw
materials to cement
We can blend the limestone and clay in the correct proportions to give whatever
value for LSF we like, say 97%. However, the SR and AR will then be fixed by
whatever the composition of the raw materials determines them to be. Since
SiO2, Al2O3 and Fe2O3 are mainly contributed by the clay, it is the clay composition
that will largely determine these parameters.
In general terms, two types of raw material, such as limestone and clay, can be
proportioned to fix any one parameter only, say the LSF.
To fix x ratios, x+1 materials of suitable composition are needed, so to control all
three parameters, LSF, SR and AR, a cement plant needs to blend four different
materials of suitable composition. On a coal-fired plant, the composition of the
coal ash also needs to be allowed for, since the ash also ends up in the clinker
and combines with the other materials.
In practice, a cement plant may have five or six different raw materials in order
to control composition.
You cant use just any limestone or clay for cement-making - minor components
in the raw materials also need to be considered. For example, standard
specifications prescribe limits on the magnesium oxide that can be present in the
clinker, so a dolomitic limestone a limestone containing magnesium carbonate
as well as calcium carbonate may be unsuitable.
The most common minor constituents normally present in clinker and therefore in
the raw materials, are:
Sulfate (SO3) typically about 1%, usually between 0.5% and 1.5%
Well look at the first three of these, alkalis, sulfates and magnesia in more
detail.
While alkalis and sulfate will combine where possible, it is unlikely that the
proportions of both will be present in just the right amounts to make, for
example, potassium sulfate. It is more likely that there will be an excess of
either alkali or sulfate in the clinker.
Normally, a small amount of alkali enters into the alite, belite and ferrite phases.
A slightly larger proportion of alkali is normally contained in the aluminate phase,
perhaps up to about 1%.
Separately, some sulfate also enters into the structures of the four main clinker
minerals, particularly belite, which may have up to about 0.5% SO3 within the
belite but not present as alkali sulfate. Belite sulfate contents generally increase
as the total clinker sulfate increases.
It is normally beneficial for the clinkering process, and for subsequent cement
hydration, that there is an excess of sulfate compared with alkali. The surplus
sulfate normally combines with lime to form a double-sulfate called calcium
langbeinite (2CaSO4.K2SO4). This is usually present intimately mixed with
arcanite.
www.whd.co.uk 44 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
On the other hand, if there is an excess of alkali compared with sulfate, the
excess alkali enters the belite and aluminate phases. This is detrimental and to
be avoided if possible. Alkalis and sulfates will be considered further when we
look at the reactions in the kiln.
Magnesia (MgO)
Small amounts of magnesium can replace calcium in the four main clinker
minerals. If the main clinker minerals cant accommodate all the magnesium
within their crystal structures, the excess forms crystals of periclase (magnesium
oxide).
Small amounts of periclase are not a problem but large amounts are undesirable,
as the periclase may hydrate within hardened concrete; the conversion of
magnesium oxide to magnesium hydroxide is an expansive reaction and can
cause damage. For this reason, cement standard specifications impose a
maximum limit on the MgO content of cement, typically 4%-6%.
Some heavy metals retard cement hydration (eg: lead, zinc); the presence of
these would obviously be undesirable from the viewpoint of the cement, as would
emission of toxic heavy metals during manufacture.
Fluoride can retard the setting of cement but it can also act as a mineraliser
together with sulfate, enabling alite formation in cement clinker at a lower
temperature, with consequent savings in fuel. Mineralised cements
characteristically have good early strengths.
There are different approaches to this, but the basic idea is to do a series of trial
mixes and assess how well they combine in a laboratory furnace.
www.whd.co.uk 45 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
For example, we might mix limestone and shale at an LSF of 0.94 and burn ten
samples of this mix at different temperatures for 30 minutes well call this Mix
A. When cool, we measure the free lime in each and draw a combinability curve
for Mix A by plotting free lime against burning temperature.
We then repeat this for other mixes, eg: the same limestone and shale mixed to
an LSF of 0.98; well call this Mix B. For two mixes, A and B, we now have data
relating burning temperatures and free lime (Figure 5.1). We would expect hotter
burning to give lower free lime contents.
The combinability temperature for each mix can be read from the plots at the
intersection of the curve with the line CaO=1%. In Figure 5.1, the combinability
temperatures, T(A) and T(B) for Mixes A and B are about 1430 C and 1560 C
respectively.
Combinability depends additionally on how well the raw materials react together;
a marl might be composed of small particles of calcium carbonate, silica and clay
all together in one rock type in approximately the desired proportions to make
cement. The small particles, already in close proximity with each other, will
probably react more easily than a mix of the same overall composition containing
a blend of three rock types say, chalk, silica and clay.
www.whd.co.uk 46 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
In order to find out how each of these parameters affects how particular raw
materials will combine, we need to do a large number of test burns. Assuming we
have a suitable range of raw materials to control all three parameters (at least
four materials) individual parameters can be changed to show how each
parameter affects the combinability temperature for that mix.
First, we produce a series of mixes with varying LSF but the same SR and AR and
fineness, we then burn each of these mixes at different temperatures to find the
combinability temperature for that mix (ie: the temperature that gives us 1%
free lime). We then repeat, but varying the SR and then the AR. Finally, we
repeat the whole exercise, this time using material of a different fineness.
(Fineness can mean the blended raw materials collectively, or we could test
variations in the fineness of each component).
It will be clear that combinability experiments may involve doing tens or even
hundreds of individual burns of test mixes and that this is not a trivial exercise.
We could expect that the combinability temperature would increase with LSF; to
produce a clinker in which the calcium silicates are all in the form of alite will
require the constituents to be evenly mixed. Any heterogeneities will result in
areas of excess lime or silica and these will be difficult to combine. With
decreasing LSF mix heterogeneities become less important.
Since it is the liquid flux that mainly facilitates combination, mixes of AR ratio of
1.38 will be easiest to burn. Strictly, this applies to the CaO-SiO2-Al2O3-Fe2O3
system. Minor constituents can alter the optimum AR; magnesium oxide, for
www.whd.co.uk 47 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
Combination is also affected by the viscosity of the liquid phase; a less viscous
liquid will be more mobile and this will aid combination. Liquid viscosity increases
with increasing AR and decreases with increasing temperature.
Finally, the coarseness of the reactants affect the rate of any chemical reaction,
so it is no surprise that coarser material is harder to combine than fine material.
Figure 5.2 Example of a typical curve Figure 5.3 Example of a typical curve
showing combinability temperature showing combinability temperature
variation with LSF. With increasing LSF, it variation with SR. With increasing silica
becomes progressively harder to achieve ratio, mixes become harder to burn, as
combination. Above 100% LSF, there will there is less liquid present for ion
always be some uncombined free lime. transport, without which lime cannot
combine with belite to form alite.
Figure 5.4 Example of a typical curve Figure 5.5 Example of a typical curve
showing combinability temperature showing combinability temperature
variation with AR at fixed LSF and SR. variation for a fixed composition with
Combinability temperatures are at a 90m sieve residue. Coarser material is
minimum at about AR=1.3-1.5. harder to combine.
www.whd.co.uk 48 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
So, returning to the question of what the best clinker composition might be, from
the previous section, suppose we want a high content of calcium silicates, with a
high alite content. This means both a high LSF and a high SR. However, the
curves in Figures 5.2 and 5.3 show that either will make the raw materials harder
to combine, so in a mix with both a high LSF and a high SR, good combination
might be difficult to achieve.
In practice, the main factor influencing the clinker composition will be the
compositions of the main raw materials to be used. For example, if a fairly pure
limestone and a clay are to be used, the clay composition will largely determine
the silica ratio and the alumina ratio, because the limestone wont contain much
silica and alumina. Much more control would be possible for the lime saturation
factor, by adjusting the ratio in which the limestone and clay are mixed.
Where coal is the fuel for the kiln, the calculation also needs to take into account
the effect of coal ash. Most of the ash will become incorporated into the clinker
and the quantity of ash is enough to have a significant effect on clinker
composition; ash may represent perhaps 2%-8% of the clinker.
To calculate the effect of coal ash, we would need to know: how much coal is
needed to produce a tonne of clinker; the proportion of ash retained in the clinker
(probably most of it); the proportion of ash in the coal and the composition of the
ash.
The principle of the proportioning calculation is simple but the actual calculation
is tedious when done manually. Computers can do it instantly by solving
simultaneous equations, so well save a few hours and move swiftly on.
www.whd.co.uk 49 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
In a dry-process kiln, some reactions will take place in the preheater tower
before the feed enters the kiln. Any moisture will evaporate, some or most of the
limestone will be calcined and the surfaces of siliceous particles in the feed will
start to combine with adjacent lime particles.
In a wet-process kiln, all the reactions obviously take place in the kiln.
In a wet-process kiln, water evaporation and calcination takes place in the kiln,
within the moving mass of feed. This situation is not ideal because heat transfer
has to take place through a large mass of material and water vapour and CO2
have to escape outwards, taking heat with them, when the objective is to
transfer heat inwards. Calcining takes place after the water has been driven off,
about a third of the way down the kiln.
In a dry process preheater kiln, any moisture in the feed particles is evaporated
and up to about half of the limestone is decarbonated before the feed enters the
kiln. Surface reactions on siliceous particles will occur to a limited extent.
In a dry process kiln with a precalciner, most of the limestone is calcined before
the feed enters the kiln. Calcination takes place in a few seconds, rather than the
half an hour or so inside a wet kiln at the same temperature. The surfaces of
siliceous particles will react with lime to a greater extent than in a kiln with a
preheater only, due to the higher temperatures.
www.whd.co.uk 50 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
As lime is calcined, it starts to react with other components of the raw feed. The
main initial silicate product is belite. Some calcium aluminate (CA and C12A7) and
ferrite phases also start to form.
Some minerals are formed in the kiln at relatively low temperatures compared
with the temperature in the burning zone. Most of these minerals dissociate in
the burning zone and are not therefore present in the clinker; these minerals are
termed intermediate compounds. Some intermediate compounds assist in
forming the final clinker minerals.
A sulfate melt phase may form, consisting mainly of alkali sulfates (Chapter 5.6);
this liquid phase can aid combination of lime and silica and is not miscible with
the main liquid phase. Alkali chlorides may be also present.
In the burning zone, above about 1350 C, reactions take place quickly. The
clinker is in the burning zone for perhaps 15-30 minutes but in this time a lot
happens:
www.whd.co.uk 51 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
The additional liquid causes feed particles to stick together, leading to the
formation of clinker nodules.
Alite formation
Alite forms mainly by the reaction of lime with belite. Some alite also forms
directly from free lime and silica. These reactions occur rapidly once the clinker
temperature is above about 1400 C.
Evaporation of volatiles
Alkalis, sulfates and chlorides are present in the burning zone chiefly as sulfate
melts and are liable to volatilise. The volatilised material passes down the kiln,
where it condenses on the cooler, part-reacted, feed. It then travels back towards
the burning zone, where it may volatilise again; it may follow this cycle many
times. These recycling volatiles are known as the recirculating volatile load.
www.whd.co.uk 52 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
Reducing conditions cause Iron (III) to be reduced to Iron (II). Iron (II) can
substitute for calcium and so can be present in all the clinker phases.
Less liquid in the clinker, because there is less Iron (III). Effectively, the
SR is increased, so alite formation is inhibited, necessitating harder
burning.
As the clinker cools, some or most of the iron (II) reoxidises to iron (III).
As it does so, some alite dissociates into belite and free lime because iron
that was substituting for calcium in the alite is no longer present; the
lower alite content results in some loss of cement hydraulicity.
Iron may reduce further to form metallic iron. Sulfides also form, altering
the balance between alkali and sulfate in the clinker (Chapter 5.6).
The effect on cement made from clinker produced under reducing conditions is
generally to reduce cement performance. Concrete strengths may be lower and
setting times erratic. Overall, reducing conditions in the kiln are undesirable.
The only exception to this is that of white cement. White cement clinker is
burned deliberately under slightly reducing conditions to reduce iron (III) to iron
(II) that way the cement is whiter. As there is little iron present anyway, the
negative aspects of reduction are minimal.
As soon as the clinker passes the burning zone, it begins to cool, even though it
is still inside the kiln. The cooling rate from the burning temperature down to
perhaps 1000 C is especially important; much of this cooling occurs before the
clinker reaches the cooler. Fast cooling of clinker is advantageous; it makes for
more hydraulically-reactive silicates and lots of small, intergrown, aluminate and
ferrite crystals.
Rapid cooling probably makes the clinker easier to grind by inducing thermal
cracking.
www.whd.co.uk 53 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
In the kiln, sulfate is beneficial as it forms a sulfate melt phase which promotes
reactions by acting as a flux, helping ion transfer in the same way as the main
liquid phase.
Although clinker leaving the kiln may typically contain perhaps 1% SO3, the
proportion of sulfate in the partly-combined material in the kiln will be higher,
possibly much higher; this is because of the recirculating volatiles, of which alkali
sulfates are likely to be the principal component. Within the kiln, the proportion
of sulfate melt phase could be 5% or more of the charge.
When clinker is ground to make cement, gypsum is added to control the setting
of the cement. However, the clinker alkali sulfate phases can also contribute to
controlling the set. Since sulfates are largely combined with alkalis, sodium and
potassium in the cement-making context, alkalis and sulfates have to be
considered together.
Clinker alkali sulfate forms mainly in pores in the clinker (Figure 5.6).
It is unlikely that the amount of available sulfate will exactly equate with the
amount of available alkali to make alkali sulfates; there will normally be an
excess of one over the other. An excess of available sulfate over available alkali
is the more desirable of the two alternatives.
www.whd.co.uk 54 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
If there is an excess of sulfate over alkali, calcium combines with the excess
sulfate to form another double sulfate known as calcium langbeinite
(K2O.2CaO.3SO3). Calcium langbeinite is an effective set retarder. Clinker alkalis
can enhance early strength development in concrete as they accelerate the rate
of alite hydration by increasing the alkalinity of the pore fluid,.
The types of clinker sulfate present are determined by the ratio of clinker alkali to
sulfate and by the types of alkali. The main forms of clinker alkali sulfate are:
While most alkali and sulfate in clinker is present as alkali sulfate inclusions in
pores in the clinker, some alkali and some sulfate are incorporated into the four
main clinker phases.
Alkali and sulfate, uncombined with each other, are generally present in belite;
typical levels are 0.2% SO3 and 0.9% K2O although this varies considerably.
Belite often also contains small inclusions of alkali sulfate.
Normal aluminate generally contains little sulfate but perhaps typically a total of
about 1% Na2O and K2O. Normal aluminate has a cubic crystal structure, but
another form of aluminate with a prismatic crystal structure also occurs in some
clinkers. Prismatic aluminate (or alkali aluminate) has a higher alkali content, up
to 4%-5% K2O and Na2O combined. It occurs where there is an excess of alkali
over sulfate in the clinker; it is detrimental in itself, leading to erratic setting
characteristics of cement, and is symptomatic of undesirable clinkering
conditions.
Suppose we have a clinker which has an excess of available alkali over sulfate.
For simplicity, suppose all the alkali is K2O.
Available means neglecting alkali and sulfate bound in the solid phases, alite and
belite, in the burning zone and not available to form alkali sulfates. (Free lime
does not normally contain much alkali or sulfate). Excess of alkali over sulfate
means that there is alkali left over after all the available sulfate has combined
with alkali. Potassium sulfate (K2SO4) consists of about 54.1% K2O and 45.9%
SO3.
www.whd.co.uk 55 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
For a clinker with 0.50% available sulfate, alkali and sulfate are in balance if the
clinker also contains about 0.59% available K2O.
A small excess of available alkali over sulfate is not usually important, unless it is
abnormal for clinker from that particular cement plant. If so, cement
characteristics may be altered and customers may notice. A large excess of
available alkali over sulfate is to be avoided if possible because:
In a clinker with excess alkali over sulfate, the excess alkali will be
present as an elevated alkali content in both the aluminate and belite
phases.
Excess alkali in belite will inhibit combination with free lime to make alite
in the burning zone.
Suppose now that the clinker sulfate content is gradually increased (perhaps by
burning a higher-sulfur fuel), while the alkali content is maintained. There will be
less alkali in the aluminate and belite and more alkali sulfate as arcanite. At
some point, the available alkali and sulfate will be balanced, with all alkali sulfate
present as arcanite. Combinability and cement performance should improve.
If the sulfate level is increased further, the excess sulfate combines with calcium
to form the double sulfate calcium langbeinite (K2SO4.2CaSO4). The clinker will
then contain a mixture of calcium langbeinite and arcanite. With yet more sulfate
addition, at some point all of the alkali sulfate will be present as calcium
langbeinite.
With still more sulfate, anhydrite will start to form, although this is unusual.
As the clinker sulfate increases, recirculating sulfate loads and deposit build-ups
may cause kiln operating problems.
www.whd.co.uk 56 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
The clinker sulfate melt phase contributes to ion transfer in the same way
as the main liquid phase - if there is less sulfate melt phase, its fluxing
contribution will be less and combination of the clinker minerals will be
more difficult to achieve.
If the recirculating volatile load is too high, production problems are likely
with sulfate-rich build-ups causing blockages.
The grinding of clinker requires a lot of energy. How easy a particular clinker is to
grind - grindability - is difficult to predict, but rapid cooling of the clinker is
thought to improve grindability due to the presence of microcracks in alite and to
the finer crystal size of the flux phases.
The fineness of cement is one of the main factors controlling its behaviour when
mixed with water. Cement ground to a finer particle size will clearly react more
quickly than the same cement milled more coarsely. Fineness is measured s the
specific surface area in square metres per kilogram (m2 kg-1). This is typically
determined by an air permeability test, eg: the Lea and Nurse method or the
Blaine method. These methods give similar results, typically 350-450 m2 kg-1;
rapid-hardening (high early strength) cements will be at the higher end of this
range. Other test methods may give different values.
www.whd.co.uk 57 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
Since the clinker gets hot in the mill due to the heat generated by grinding,
gypsum can be partly dehydrated. It then forms hemihydrate, or plaster of Paris
CaSO4.1/2H2O. On further heating, hemihydrate dehydrates further to a form of
calcium sulfate known as soluble anhydrite (CaSO4 with a trace of water); this
has a similar solubility in water to hemihydrate, which in turn has a higher
solubility than either gypsum or natural anhydrite.
Cement mills need to be cooled to limit the temperature rise of the cement. This
is done by air-cooling, or air-cooling and water-cooling.
variations in cooling rate of the clinker in the kiln and subsequent changes
in the proportions or size of the C3A crystals.
arcanite > calcium langbeinite > hemihydrate soluble anhydrite > gypsum >
natural anhydrite.
To summarise this complex issue of clinker sulfate and added gypsum and other
sulfate:
www.whd.co.uk 58 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
In a cement from any given source (ie: same kiln), the clinker sulfates
may vary in type and quantity over time.
At the clinker mill, sulfate will be added to a target figure, typically about
3% SO3, but because total clinker sulfate may vary, the proportion of
clinker sulfate to added sulfate may also vary.
Grinding aids
Limestone
Cynics may say that cement companies like to add limestone because limestone
is cheaper than cement. While this cost comparison is undeniable, there is a bit
more to it than that.
In the past, cement mills were open circuit mills and the cement produced by
them had a wide range of particle sizes, from very fine to over-coarse. More
recently, cement mills have tended to be closed circuit mills; cement from a
closed circuit mill tends to have a narrower particle size range compared with
cement from an open-circuit mill and this increases the water demand of the
cement.
www.whd.co.uk 59 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
Concrete made with cement that has a higher water demand will tend to be of
lower strength because the water/cement ratio is higher. By intergrinding
limestone, the very fine limestone particles act to reduce the water demand of
the cement because they occupy space between the cement particles that would
otherwise be occupied by water. The presence of a small amount of interground
limestone therefore can increase concrete strengths by reducing the
water/cement ratio required when mixing the concrete.
For related reasons, fine limestone has the positive effects of reducing bleeding
and/or allowing the cement to be ground slightly coarser for the same strength.
Reduces bleeding
Reduces the quantity of clinker in the cement, saving energy and raw
materials
Acts as nucleation sites, aiding early cement hydration and improving the
hydrate microstructure
Small amounts of fine limestone in cement can therefore reduce the energy
required to produce the same quantity of cement, and have beneficial effects on
concrete properties.
The added limestone acts as a filler, but a proportion of it will react with the
cement and contribute to the composition of the cement hydration products
(Chapters 6.3 and 11).
www.whd.co.uk 60 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
Reducing agents are added to control the Chromium VI content of the cement.
Typical reducing agents are ferrous sulfate, either the monohydrate (FeSO4.H2O)
or heptahydrate (FeSO4.7H2O) forms, or stannous sulfate (SnSO4). By mass of
cement, approximately 0.5% ferrous sulfate is typically added, and
approximately 0.05% stannous sulfate. At such low levels of addition, no
significant adverse effects on cement performance would normally be expected.
In the European Union, the Chromium VI Directive came into effect in 2005 and
prohibits the use or supply of cements containing more than 2 ppm water-soluble
hexavalent chromium by mass of cement.
Firstly, gypsum will be present, so we have to allow for the proportion of CaO
assumed to be present as calcium sulfate. One approach is to subtract 0.7 x
(SO3) from the CaO in the bulk composition of the cement, in the same way that
free CaO is subtracted. For many years, this was a standard procedure. Of
course, this introduces a slight error as it neglects any clinker sulfate, but if no
other material has been added, this is still the simplest approach. After making
the deduction for sulfate (and for free lime if required), the calculation can be
done in the usual way as outlined in Chapter 3.5.1.
To allow for limestone in the calculation, the CO2 content of the cement should be
determined and the limestone content estimated. If the purity of the limestone is
known, a more accurate determination can be made; for example, pure
limestone (100% calcium carbonate) contains 44% CO2 but if the limestone is
80% pure, it will contain only 35% CO2.
The American standard specification for Portland cement ASTM C 150-07 contains
a modified version of the Bogue calculation that allows for both added gypsum
and limestone:
www.whd.co.uk 61 www.understanding-cement.com
Portland cement manufacturing from raw materials to cement
C3S = 4.071CaO-7.600SiO2-1.430Fe2O3-6.718Al2O3-2.852SO3-5.188CO2
C2S = 2.867SiO2-0.7544C3S
C3A = 2.650Al2O3-1.692Fe2O3
C4AF = 3.043Fe2O3
www.whd.co.uk 62 www.understanding-cement.com
6 Hydration of cement chemical and physical
properties of cementitious materials
Cement hydration is the process by which cement and water react together.
When cement and water are mixed in suitable proportions, the result of the
reaction is a solid mass, composed of gel and crystalline material, which binds
together the constituents of a concrete mix.
From the clinker: alite, belite, aluminate and ferrite, plus alkali sulfates
and free lime
Water
Calcium hydroxide
A mix of cement and water forms a hard cement paste. Concrete is a mix of
cement, water and aggregate (ie: sand and gravel or crushed stone) in which the
aggregate is bonded in a matrix of cement paste. Because it is the cement paste
that binds all the constituents together as a single, solid, mass it follows that the
properties of the concrete (strength and durability for example) will be critically
influenced by the properties of the cement paste.
Hydration of cement chemical and physical properties of cementitious materials
Of the four main clinker phases, the aluminate phase is the most reactive,
followed by alite, then belite. Ferrite is initially very reactive but the hydration
subsequently slows appreciably.
When cement and water are mixed, exothermic reactions occur. The rate at which
heat is evolved varies for the first three days or so and then gradually declines.
The general shape of this plot is useful in understanding the reactions that occur
as the cement hydrates (Figure 6.1).
The shape of the plot shows a rapid exothermic reaction at first (I in Figure 6.1);
the * in Figure 6.1 symbolises a fast increase in heat evolution when water is
added, then a fast drop, too near the y-axis to be shown.
The initial reaction lasts only a few minutes, or less, and is followed by a
dormant, or induction, period (between I and II in Figure 6.1) of several hours
during which there is little further heat evolution. Subsequently, the rate of
hydration picks up again; during this second period of heat evolution, the
www.whd.co.uk 64 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
concrete sets and gains strength. The rate of heat evolution in this second period
reaches a peak (III in Figure 6.1) after 10-15 hours then, apart from possible
small increases at around 20 hours (IV in Figure 6.1) and 30-40 hours (V in
Figure 6.1), gradually tails off over the following days and weeks.
The main characteristics of the plot in Figure 6.1 can be interpreted as:
The initially high rate of heat evolution at I in Figure 6.1 is due mainly to
the hydration of the aluminate and alite phases and the hydration of
calcium sulfate hemihydrate to the dihydrate form (gypsum).
The rate of hydration then slows to a low level after a few minutes (II)
due to the deposition of hydration products on the surfaces of the cement
grains. During this dormant, or induction, period before the paste sets,
the concrete is plastic or flows freely.
Gradually, the rate of aluminate and alite hydration increases again and
belite also begins to hydrate (III).
Small subsequent increases in heat evolution (IV and V) are probably due
to ettringite formation and ferrite hydration respectively.
After a few days, the rate of heat evolution gradually tails off as the unhydrated
cement particles become scarcer, and as what cement remains is surrounded by
dense hydration product that slows any further reaction.
The cement paste sets when solid particles in the mix become tenuously
connected by the products of the hydration reactions, mainly ettringite and
calcium silicate hydrate. As time passes, the links between the solid particles -
the unhydrated cement particles, the hydration products and the aggregate -
become thicker and the concrete gains strength, or hardens, as hydration
www.whd.co.uk 65 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
products continue to form. These physical changes can be explained more fully
by first considering the hydration products that form as hydration proceeds.
The four main clinker minerals produce hydration products broadly as follows:
Alite: alite hydration produces calcium silicate hydrate gel plus a lot of
calcium hydroxide.
In contrast, alite (C3S in the pure form) has a Ca:Si ratio of approximately 3:1.
Alite hydration will therefore result in excess calcium if the C-S-H that forms has
a Ca:Si ratio of 2:1. The excess calcium oxide produces calcium hydroxide by
reacting with water. Calcium silicate hydrate is the main strength-giving
component of hydrated Portland cement and it is therefore usually advantageous
to maximise the proportion of C-S-H produced.
Ettringite: when cement and water are first mixed, the aluminate phase reacts
with dissolved calcium sulfate to form ettringite. Ettringite, in cement chemistry
notation, has the formula:
www.whd.co.uk 66 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
_
C6AS3H32 or C3A.3CaSO4.32H2O (mixing cement chemistry and normal notation)
Ettringite crystals are generally thought to form a coating over the surfaces of
hydrating crystals of aluminate phase, acting as a temporary barrier to further
hydration. However, more recently, this classical interpretation of the process by
which aluminate hydration is retarded has been questioned and the mechanism
may be more complex. Nevertheless, the concept of retardation due to a coating
of ettringite or other hydration products over the surface of the cement grains is
simple and more than adequate for most purposes.
www.whd.co.uk 67 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
_
C4ASH12 or C3A.CaSO4.12H2O
Subsequently, as more cement hydrates and the supply of gypsum and clinker
sulfate becomes exhausted, the ratio of available sulfate to aluminate decreases.
Monosulfate then forms instead of ettringite, and ettringite that has already
formed converts to monosulfate.
Fine limestone also has a physical effect on cement hydration through the
fine filler effect (Chapter 5.8).
www.whd.co.uk 68 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
C3(A,F).3CX.nH2O
X represents an anion with a single negative charge (eg: Cl- or OH-) or half an
anion with a double negative charge (eg: SO32-, CO22-); n indicates that the
amount of water is variable, eg: n=12 (usually) for monosulfate.
if X were chloride, the formula for that form of AFm would be:
C3A.CaCl2.10H2O
if X were sulfate, the formula for that form of AFm would be:
C3A.CaSO4.12H2O
The m - or mono - in AFm refers to the single unit of CX2 in the formula for
AFm phases.
The t - or tri - in AFt refers to the triple unit of 3CX in the formula for AFt
phases.
For example, the most significant AFt phase in cement hydration is ettringite,
where X is sulfate. Neglecting any iron substitution for aluminium, the formula for
ettringite can be written as:
C3A.3CaSO4.32H2O
There are other AFt phases, but ettringite is overwhelmingly the most important.
Unless you are doing a PhD in cement hydrate composition, just note that there
might be other AFt phases present in small amounts, but otherwise forget about
them.
www.whd.co.uk 69 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
AFm phases are numerous; in the good old days when cement was just clinker
and gypsum we could assume the main AFm phase was monosulfate. Where
interground limestone is present in the cement, as it commonly now is, that
assumption is no longer true.
The most significant AFm phases in the context of cement hydration are:
_
Monosulfoaluminate, or monosulfate: C3A.CaSO4.12H2O or C4ASH12
_
Monocarboaluminate, or monocarbonate: C3A.CaCO3.11H2O or C4ACH11
_
Hemicarboaluminate, or hemicarbonate: C4C0.5H12
Hydroxy-AFm C4AH13
Just to summarise where we have got to so far, the main products of cement
hydration are:
AFt phase
AFm phase(s)
Ettringite will be the main AFt phase, but which AFm phases are present in any
particular hydrated cement will depend on the balance between available alumina
and sulfate, and the availability of other ions, particularly carbonate. Temperature
may also have an effect.
www.whd.co.uk 70 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
Soon after mixing, if insufficient sulfate is available from the added gypsum and
the clinker sulfates to slow the hydration of C3A, the resulting rapid hydration of
the C3A produces AFm phase in the form of hydroxy-AFm. These crystals link up
the particles (cement, aggregate, hydrates) in the mix and quickly cause the mix
to become unworkable. This is termed a flash set. The rapid hydration of the
C3A results in the evolution of much heat. Concrete strengths where flash setting
has occurred may be lower than normal.
t=0: at the instant of mixing, the mix contains water and unhydrated particles of
cement.
www.whd.co.uk 71 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
t=a few minutes-a few hours: when mixed with water, there is a high initial
rate of reaction for a short period as aluminate and alite hydrate. Sulfate from
the gypsum and clinker sulfate starts to dissolve and an amorphous gel
containing alumina, lime, sulfate and silica forms. Small crystals of ettringite also
form. A period of lower activity follows (the dormant period). The initial set
typically occurs at about 2 hours and hydration restarts in earnest after 3-6
hours.
t=6-12 hours: Alite hydration produces C-S-H and large crystals of calcium
hydroxide (CH) up to 10m or more across. Continued aluminate hydration
produces larger rod-like ettringite crystals. C-S-H forms dense shells around
cement grains and a delicate sheet structure bridging the water-filled gaps
between the cement grains. By this time, the concrete has set but it has little
strength.
A small fluid-filled gap less than a micron across appears between the surface of
the unhydrated cement grain and the surrounding dense shell of C-S-H. The
cement reacts in two ways. Firstly, it reacts by a process of dissolution at the
surface of the reacting cement grain, with dissolved material passing through the
coating and reprecipitating in the large water-filled gaps between the cement
grains. Secondly, cement hydrates in-situ; this means that hydration product
(mainly C-S-H) forms within the volume occupied by the cement particle. Often,
the crystal shapes of former and alite and belite crystals can be seen in
microscopic examination of polished sections where the alite and belite have
been replaced entirely by C-S-H.
t=1-3 days: after a day or so, most of the available sulfate has been used, but
aluminate phase continues to hydrate. This causes a gradual change in the ratio
of available sulfate to aluminate, causing a change from ettringite formation to
monosulfate formation. In ettringite, the aluminium: sulfur ratio is 2:3 whereas
in monosulfate it is 2:1. With this decrease in sulfate availability, platey crystals
of monosulfate start to form instead of ettringite. After a few days, the ettringite
already present begins to decompose and is replaced by monosulfate. C-S-H and
CH continue to form in the large water-filled gaps between the cement grains;
the concrete gains strength as the C-S-H in the large water-filled gaps becomes
denser.
t=3 days onwards: the small gap between the anhydrous cement and the
dense C-S-H coating starts to fill. Smaller cement grains have fully reacted.
Once the small gap between the anhydrous cement grain and surrounding dense
shell has filled, reaction between the solid cement grain and the solid shell of
hydration product continues at a slower rate. Consequently, C-S-H and CH
formation in the large water-filled gaps between the cement grains continues also
at a slower rate and strength development continues at a slower rate reflecting
the slower C-S-H formation.
www.whd.co.uk 72 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
not quite the final products but are approaching them. After a year, any further
change will be very slow; most of the available sulfate is present in monosulfate
phase but a small excess of sulfate allows a little ettringite to persist.
With increasing water/cement ratios, the paste fraction of the concrete becomes
increasingly porous. This has two direct consequences:
These effects are apparent in the sequence of four images that follow, showing
cement pastes made using different water/cement ratios. However, in order to
understand these, it is useful to be able to identify the different features visible in
SEM images of polished sections of cement paste. Look first at Figure 6.3; this
image shows one of the images in the sequence that follows in Figures 6.4-6.7,
but with the different features indicated.
www.whd.co.uk 73 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
The cement paste shown in Figure 6.3 contains the following features:
The outer edges of these rims indicate the approximate size of the original
cement particle and the hydration products composing the rims have
therefore formed within the volume occupied by the original cement
particle. The hydrated material in these rims is therefore called inner
product or in-situ hydration product. Although it looks solid and dense
in the image, calcium silicate hydrate is a gel and contains huge numbers
of tiny pores, called gel pores, approximately 1 nm across (1 nm=10-9 m).
At this magnification, the gel pores are far too small to be distinguished.
Also visible are areas of calcium hydroxide (CH) these appear slightly
brighter than the C-S-H but darker than the unhydrated cement.
Circle A contains CH (just visible) and also C-S-H. AFt and AFm phases
(probably ettringite and monosulfate) are also likely to be present but are
too small to be seen at this magnification. The hydration products in Circle
A have formed in a region that was occupied by water when the cement
and water were mixed; these hydration products formed from the
precipitation of dissolved material are known as outer product or
undifferentiated hydration product.
www.whd.co.uk 74 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
Of particular importance in Figure 6.3 are the small black features; these
are capillary pores. Capillary pores generally range in size from a few tens
of microns down to less than 1 m; these are the remnants of the volume
originally occupied by water when the cement and water were first mixed,
most of this original water-filled volume having been filled by cement
hydration products. The pores became filled with epoxy resin when the
polished section was prepared and so the black features are actually
epoxy resin, but the extent of the resin shows the extent of the original
pore system.
The capillary pores in Figure 6.3 are generally small, up to about 5 m across,
and not very numerous. They appear not to be connected, but it is important to
remember when looking at polished sections that the third dimension is not
shown and that pores may be connected out of the plane of the section. In this
case, the pores clearly are interconnected, or they could not have become filled
with epoxy resin.
Now that you know what you are seeing in these images of cement paste, try
looking at the four images in Figures 6.4-6.7.
Changing the water/cement ratio has clearly altered the microstructure of the
pastes. Pastes made with different water/cement ratios contain similar features
to those shown in Figure 6.3, but the proportions are different.
Firstly, look at the image in Figure 6.5; this is familiar as the same image as in
Figure 6.3, but without all the labels getting in the way of parts of the picture.
Now look at the image in Figure 6.4, the paste with the lowest water/cement
ratio (0.33). As in Figure 6.3, the brightest features are unhydrated cement;
these particles contain all four main clinker minerals. The capillary pores (black,
as in Figure 6.3) are not numerous and are generally small (<2m-3m).
Now compare the pastes with the lowest and highest water/cement ratios to see
the greatest contrast (Figure 6.4, w/c = 0.33, and Figure 6.7, w/c = 0.60). Two
features are immediately clear: at the higher water/cement ratio, the capillary
pores are larger (up to about 10m) and very numerous, and there is less
unhydrated cement.
www.whd.co.uk 75 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
Figure 6.4 Polished section of cement Figure 6.5 Polished section of cement
paste, age 2 years, made using ordinary paste, age 2 years, made using ordinary
Portland cement, w/c=0.33. Portland cement, w/c=0.40.
Figure 6.6 Polished section of cement Figure 6.7 Polished section of cement
paste, age 2 years, made using ordinary paste, age 2 years, made using ordinary
Portland cement, w/c=0.50. Portland cement, w/c=0.60.
Most of the unhydrated cement in Figure 6.7 is ferrite phase; almost all of the
alite, belite and aluminate has hydrated. CH fills some of the pores and occupies
extensive areas; several regions of CH in Figure 6.7 are 50m long and about
5m wide.
Compared with the pastes made using the lowest and highest water/cement
ratios, the pastes shown in Figures 6.5 (w/c = 0.40) and Figure 6.6 (w/c = 0.50)
are intermediate in both porosity and proportion of residual unhydrated cement.
www.whd.co.uk 76 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
Additionally, the more porous pastes are likely to be more permeable, assuming
the pores are interconnected (which they mostly will be). Increased permeability
will mean that the concrete is less resistant to deleterious processes such as frost
damage, sulfate attack or chloride ingress.
Higher water/cement ratio concrete mixes are more convenient when placing the
concrete, as it flows more easily. However, the effects on the cement paste listed
above show that the consequences for the hardened concrete are strongly
negative.
The Powers-Brownyard model simplifies the cement paste into three components:
Unreacted cement
Hydration product
Capillary pores
Remember, this is a simplified model developed over 60 years ago. The hydration
product was referred to as cement gel and the model did not consider the
individual hydration products separately; neither did it consider the individual
clinker minerals. As the crystalline components of the hydration product (eg:
calcium hydroxide, ettringite) are clearly not a gel, to avoid confusion we will
follow Taylors suggestion and use the term hydration product where Powers and
Brownyard used cement gel.
www.whd.co.uk 77 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
The model assumes that, when water and cement is mixed, initially we have just
cement and water. Then, hydration products form and we have cement, hydration
product and capillary pores. As the cement continues to hydrate, the capillary
porosity decreases, the quantity of unhydrated cement decreases and the
amount of hydration product increases. As the quantity of hydration product
increases, the total gel porosity also increases.
Since the volume of cement paste does not change markedly during
hydration, and since the hydration product occupies more space than does
the cement that produced it, it follows that there is a minimum value of
w/c below which full hydration cannot occur. This value is approximately
w/c=0.38, assuming that additional water for curing is available. This is
known as the critical water-cement ratio, or w/c*.
At values below the critical water-cement ratio, complete hydration cannot occur
because there is no available space in which additional hydration product can
form. While a critical w/c value of about 0.38 is regarded as generally about
right, at least one team of researchers has shown complete hydration at a
significantly lower w/c ratio (5). Cement wouldn't be so interesting if it weren't
for occasional reminders that maybe we don't understand it as well as we
thought.
Above w/c=0.44, there is sufficient water for complete hydration and also enough
space in which the hydration products can form, so in principle, the cement can
hydrate fully.
www.whd.co.uk 78 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
Two properties are immediately obvious: how concrete behaves while it is being
mixed and placed, and the strength of the hardened concrete.
These three properties are huge subjects in their own right. What follows is the
briefest of introductions but references are given where more detail can be found
if needed.
In the slump test, a conical mould containing the concrete is emptied onto a flat
surface. The extent to which the concrete subsides is measured and is referred to
as the slump. Higher numbers mean the concrete is more workable.
www.whd.co.uk 79 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
Figure 6.8 Concrete slump test 1: the Figure 6.9 Concrete slump test 2: excess
slump cone is filled with concrete in three concrete is removed by rolling the tamping
stages, using a steel rod, the tamping rod, rod across the top of the cone. (Picture
to ensure good compaction. (Picture courtesy Kirton Concrete Services Ltd., UK).
courtesy Kirton Concrete Services Ltd., UK).
Figure 6.10 Concrete slump test 3: the Figure 6.11 Concrete slump test 4:
cone is lifted and the concrete slides out. measuring the slump the distance
Note the curved shape of the concrete between the centre of the concrete cone and
where it has slumped and the gap the tamping rod. (Picture courtesy Kirton
between the top of the concrete and the Concrete Services Ltd., UK).
tamping rod resting on the cone; the size of
this gap indicates the slump of the concrete.
(Picture courtesy Kirton Concrete Services
Ltd., UK).
The main factor affecting the workability of a concrete mix is the water content -
as more water is added, the mix becomes sloppier. Other factors include:
aggregate grading; the shape of the particles (rounded particles will obviously
move past each other more freely); the ratio of aggregate to cement (as the
proportion of cement paste increases, the contact between aggregate particles
will decrease) and the rheology of the cement paste component of the mix. In
turn, the rheology of the cement paste will depend on the water/cement ratio,
www.whd.co.uk 80 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
the fineness of the cement and whether any water-reducing admixtures are
present.
It is clear that, since the term workability refers to the mobility of the concrete
mix, workability is relevant from the time the concrete is mixed, through placing
to when it is compacted. Once the concrete starts to stiffen as it sets at the end
of the induction, or dormant, period, the workability will decrease. If more water
is added to increase workability again, concrete strengths will be reduced and the
concrete will be less durable.
The plot in Figure 6.12 is an illustrative example of concrete strength for the first
three months after curing.
50
40
)
-2
30
20
Strength (N mm
10
0
0 20 40 60 80 100
Time (days)
www.whd.co.uk 81 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
Many factors influence concrete strength. In the first few days after mixing, the
age of the concrete is one of the most critical. Other factors at all ages in
addition to water/cement ratio include: how well the concrete was compacted;
temperature of the concrete when placed and during early curing; subsequent
curing temperature; aggregate/cement ratio; factors relating to the aggregate
such as the aggregate grading, particle shape and strength; strength of the
aggregate-paste bond.
The water/cement ratio is normally considered the single most important factor.
Concrete strength increases as its density increases. The porosity of the concrete
is therefore of prime importance; as porosity increases, the density and strength
both decrease. Assuming the aggregate is of low porosity, the porosity of the
paste is the main remaining variable factor.
Assuming the aggregate to be impermeable and that no cracks are present, the
only route of ingress is the cement paste. Paste contains capillary pores and gel
pores but it is the system of capillary pores that controls the permeability of the
paste; while the gel pores represent about 30% of the volume of the hydration
product, the gel pores themselves are too small to affect permeability.
The size of the capillary pores, and the extent to which they are interconnected,
depends mainly on the water/cement ratio of the paste. At early ages of
hydration, it will also depend on the degree of cement hydration. From the
images of pastes in Figures 6.4-6.7 it is clear that capillary porosity is reduced as
the water/cement ratio decreases.
Minimising the porosity by minimising the water/cement ratio will help to improve
concrete durability because its permeability will be lower.
Three other factors affecting concrete dimensional stability are also important in
the context of cement, as they are influenced or controlled by processes
occurring within the cement paste. These are elasticity, shrinkage and shrinkage
cracking, and creep; all are affected by humidity or water movement within the
paste, and within the C-S-H in particular.
www.whd.co.uk 82 www.understanding-cement.com
Hydration of cement chemical and physical properties of cementitious materials
While important, these subjects are more within the domain of engineers and the
design of concrete structures. If you would like to read more on these specifically,
see Reference 10 below.
References, Chapter 6
5. M Rossler and I Odler, (1985) Cement and Concrete Research Vol 15,
p320.
8. Neville, Chapter 6.
www.whd.co.uk 83 www.understanding-cement.com
7 Composite Cements
7.1 Introduction
7.1.1 Background
Limestone
Mineral additions such as these have been incorporated in cement blends for
many years, initially for reasons of economy, then for the significant
improvements obtained in concrete performance in particular situations, such as
improved strength or durability.
Well look at limestone last, as it is rather different from other mineral additions,
not being either latently hydraulic or pozzolanic.
Some types of mineral addition are known as latently hydraulic; these are
slightly reactive in water but are more reactive when activated by an alkaline
material such as lime or cement. The most commonly-used mineral addition of
this type is slag.
Other materials are pozzolanic; these do not react with water but do react with
water and lime or cement. They are generally deficient in lime and so need the
addition of lime to form calcium silicate hydrate.
Crushed fired clay (brick, tile) and volcanic glass are materials known since
Roman times to benefit concrete and mortar. Volcanic glass is a natural pozzolan
and is widely used where it is available and is recognised by some national
standard specifications (eg: EN 196-1). Calcined clay and shale are used as
mineral additions, but crushed brick and similar materials have gone out of
general use.
Mineral additions can increase final concrete strengths compared with mixes
containing only Portland cement, when used in suitable proportions. The strength
increase is principally because the mineral additions increase the total amount of
calcium silicate hydrate (C-S-H) in the hydration product; since it is the C-S-H
that is mainly responsible for strength growth, if the proportion of C-S-H can be
increased it is likely that strengths will also increase, other factors being equal.
Curing temperature can also have a significant effect. Concrete cured at low
temperatures develops strength more slowly than concrete cured at higher
temperatures; however, the final strength of the concrete cured at lower
temperature is generally higher. The reason for this is not fully understood but
appears to be related to how the microstructure of the paste develops.
Large pours of concrete can reach high temperatures (60 C+) due to the heat of
hydration of the cement. By using mineral additions to replace some of the
Portland cement, the heat of hydration is reduced because the mineral additions
are less reactive than Portland cement. The concrete doesnt get as hot and the
final strengths are higher.
www.whd.co.uk 85 www.understanding-cement.com
Composite Cements
Compared with Portland cement, the ratio of Si/Ca in slag is higher; in low-lime
fly ash it is much higher. Microsilica and metakaolin contain little or no calcium.
Inclusion of any of these reactive mineral additions in cement therefore increases
the amount of silica available for the formation of hydration products compared
with a Portland cement-only mix. The additional silica combines with calcium
hydroxide (CH) in the paste, producing more C-S-H, and the Si/Ca ratio of the
C-S-H itself increases, permitting the formation of yet more C-S-H.
Comparison of porosities in pastes made with Portland cement only, and with
composite cements, is complex. Mineral additions generally react more slowly
than does Portland cement. At the same w/s ratio, the porosity of pastes made
with mineral additions is generally higher because a lower proportion of the
mineral addition will have reacted compared with the Portland cement it replaced.
This is particularly so at early ages. At later ages, enhanced C-S-H production in
composite cements will tend to partly compensate.
However, the permeability to water of pastes made with Portland cement only is
generally higher than those made with composite cements, provided the pastes
are old enough for an adequate proportion of the mineral addition to have
reacted.
How a paste made with a composite cement can have a higher porosity but a
lower permeability is not fully understood, but is probably related to the degree
of interconnection of the pores. Pores in Portland cement-only pastes appear to
be more continuous than pores in pastes made with composite cements. This
may be because the morphology of the C-S-H changes as the Si/Ca ratio
increases (1), with the formation of isolated voids.
Slag and fly ash, in addition to contributing silica to form additional C-S-H, also
contain a higher proportion of alumina compared with Portland cement. This
increases the proportion of hydrate phases containing aluminium, principally AFm
phase, but also hydrotalcite where slag is used due to the increased MgO content
of the slag. AFm phases may make some contribution to strength by reducing
paste porosity, but the extent to which they actually increase strength is unclear.
AFm phases are beneficial in that they can block capillary pores in the paste,
reducing permeability. As well as physically blocking pores, AFm phases also
chemically bind some ingressing dissolved anions, particularly chloride, and this
helps to reduce corrosion of steel reinforcement.
Increasing the AFm content of the paste will evidently bind some lime in the
additional AFm, which is then not available for C-S-H formation. However,
provided the ratio Si/Al > ~1 in the mineral addition, there should be a net gain
of C-S-H. In slag, the Si/Al atom ratio of the glassy fraction is typically between
2:1 and 3:1. In low-lime fly ash, the bulk Si/Al is typically approximately 1.5:1
but this includes alumina bound in unreactive crystals such as mullite; the Si/Al
ratio in the reactive glass will be higher.
www.whd.co.uk 86 www.understanding-cement.com
Composite Cements
Most of the above are self-evident. Figures for the CO2 emissions for Portland
cement production vary widely but are generally gradually declining as the
cement producers make efforts to lower them; somewhere between 800-1000 kg
tonne-1 probably covers the majority of production.
Taking the United Kingdom as an example, figures for the indicative embodied
CO2 (ECO2) for the main cementitious materials in concrete have been agreed by
the cement, slag and ash industries (2), based on data from 2007 for the
calculations (Table 7.1).
Table 7.1 Indicative embodied CO2 for cementitious constituents of concrete (source UK
Quality Ash Association).
Material Embodied CO2 (kg CO2 tonne-1)
Portland cement (CEM I) 930
Ground granulated blastfurnace slag 52
Fly ash (from coal burning power generation) 4
Limestone 32
www.whd.co.uk 87 www.understanding-cement.com
Composite Cements
Although the final strength of mixes containing composite cements can be higher
than equivalent Portland cement-only mixes, the rate of strength growth is
usually lower, as the rate of reaction of the mineral addition is less than that of
Portland cement. Possible exceptions are microsilica and metakaolin, both of
which are very reactive.
More care is needed in concrete curing, and the overall curing time will be
longer.
A higher total cementitious content (ie: Portland cement plus slag, flyash etc.)
would probably be needed for equivalent 28-day strength if fly ash were used, or
if slag replaced more than 30%-50% of the Portland cement. However, many
variables can have an effect, including the total cementitious content of the
concretes being compared, aggregate type and any admixtures used.
Much may also depend on the concrete curing temperature. Test data on cubes,
cylinders and prisms obtained at ambient temperatures (usually 20 C) may not
accurately reflect strengths in large pours of concrete where the heat of
hydration has raised the temperature of the concrete.
Using mineral additions may make the concrete more susceptible to carbonation
(Chapter 9.3) because depletion of calcium hydroxide from the cement paste to
produce additional C-S-H lowers the alkalinity of the pore fluid. To some extent,
this susceptibility is offset by the lower permeability to CO2 of concretes
containing mineral additions.
www.whd.co.uk 88 www.understanding-cement.com
Composite Cements
When allowed to cool slowly the crystalline content of the slag increases, with a
consequent decrease in hydraulicity. Rapid cooling is generally achieved by one
of two methods:
Pelletised slag is produced by pouring the liquid slag onto a rotating drum
cooled by water (Figure 7.2). The slag breaks into pellets and is propelled
through the air in a chamber full of water spray. Pelletised slag used in
concrete generally has a lower glass content than granulated slag,
typically 70%-90%.
Chemical moduli have been used to evaluate slags. For example, British Standard
BS EN 15167-1: 2006 specifies that (CaO+MgO)/(SiO2)>=1 and that
CaO+MgO+SiO2 must be at least 2/3 the total mass. An earlier version, BS6699:
1986 used the modulus (CaO+MgO+Al2O3)/(SiO2) >=1.
www.whd.co.uk 89 www.understanding-cement.com
Composite Cements
Figure 7.3 Granulated blastfurnace slag, Figure 7.4 SEM image of granulated
before grinding. blastfurnace slag (gbs) after grinding.
www.whd.co.uk 90 www.understanding-cement.com
Composite Cements
Table 7.1 Chemical composition of a typical blastfurnace slag used in concrete in the UK,
(source: Hanson Cement, UK).
SiO2 Al2O3 FeO CaO MgO K2O Na2O S2- LOI Total
36.0 13.0 0.5 40.0 8.0 0.5 0.3 0.8 0.5 99.6
Balance is due to minor oxides, typically P2O5, MnO, TiO2, and possibly a trace of SO3.
Because of the need for precise control of the iron production process, the slag
from a given source tends to have a consistent composition. As used in cement,
the material is essentially a crushed glass and so the composition varies little
between individual particles.
For example, a fast-cooled slag with a high glass content will have the same bulk
composition as a slag from the same liquid that was cooled more slowly.
However, the composition of the glassy fraction will be different because the
composition of the liquid will have changed during cooling as crystals formed
from the melt. The glassy fractions of the two slags cooled at different rates may
therefore have different reactivities.
When used as a mineral addition, slag typically comprises 20%-70% of the total
cementitious material, although higher levels are used for specific applications.
The main technical benefits of using slag in a composite cement are:
Will give higher later strengths, but generally lower early strengths.
www.whd.co.uk 91 www.understanding-cement.com
Composite Cements
Using slag reduces the rate of depletion of raw materials used to produce
Portland cement.
Using slag reduces the total carbon dioxide emissions per unit volume of
concrete.
Strengths
Partial replacement of Portland cement by slag generally results in lower early
strengths, due to the slower rate of reaction of the slag. Later strengths should
be higher in a suitably-designed mix. In mature concrete, more residual
unhydrated slag is likely to persist than unhydrated Portland cement in an
equivalent mix. After a year, perhaps 50%-75% may have reacted, but the actual
figure will be dependent on factors including the w/s ratio, the intrinsic reactivity
of the slag, the proportion of slag in the mix and the fineness of the slag. Higher
later strengths, in a properly designed and executed mix, are mainly due to
enhanced C-S-H formation.
Heat of hydration
Compared with Portland cement, slag typically hydrates more slowly as it is a
less-reactive material. The heat of hydration of a composite cement containing
slag is therefore less than that of a comparable PC-only mix. Both the rate of
heat evolution and the total heat evolved are lower in mixes where Portland
cement is partially replaced with slag.
Colour
The grey colour of Portland cement is due to the ferrite phase. Slag does not
normally contain ferrite or other strongly-coloured minerals so partial
replacement of Portland cement with slag should result in a lighter colour of
concrete. Although slag powder is almost colourless, fresh fracture surfaces of
concrete containing slag are usually a blue-green colour. This is thought to be
due to sulfide from the slag entering the hydration products, probably the AFm
phase. After a few days, the blue-green colour fades as the sulfide oxidises.
www.whd.co.uk 92 www.understanding-cement.com
Composite Cements
Low-lime fly ash used in Europe typically contains high proportions of silica and
alumina (Table 7.2) and contains a high proportion of glass. This broadly
corresponds to a Class F ash in the USA.
Table 7.2 Chemical composition of a low-lime fly ash used in concrete (UK-sourced
ash).
SiO2 Al2O3 Fe2O3 CaO MgO K2O Na2O SO3 Total
Balance is mainly carbon, with minor oxides including P2O5 and Mn2O3.
The glass content of fly ash is typically 70%-90%, with the bulk of the crystalline
material consisting of quartz (SiO2) and mullite (an aluminium silicate, Al6Si2O13
or Al6Si3O15)
The bulk of most fly ash is composed of spherical particles, ranging in size from
under 1m to about 50m, although some particles may be up to 100m in
www.whd.co.uk 93 www.understanding-cement.com
Composite Cements
diameter. Most particles are solid (with some air bubbles) although some are
hollow shells.
Figure 7.5 Coal fired power station, source Figure 7.6 SEM image of fly ash particles,
of fly ash used in concrete. (Photo courtesy as used in concrete.
UK Quality Ash Association.)
Most fly ash particles are mainly composed of an impure glassy aluminosilicate.
However, because they are formed individually from mineral assemblages in coal,
the composition of each particle is slightly different, or sometimes very different.
Some fly ash is calcareous; these ashes are typically higher in CaO, SO3 and MgO
and contain less SiO2 and Al2O3 compared with siliceous fly ash. These ashes are
described as Class C fly ash in the USA if the sum of the SiO2 and Al2O3 contents
is less than 70%. The following focuses on siliceous fly ash.
When used as a mineral addition, fly ash typically comprises up to 30% of the
total cementitious material, sometimes more. The main technical benefits of
using fly ash in a composite cement are:
Because fly ash particles are spherical, mixes containing fly ash require
less water for the same workability compared with a mix containing
Portland cement only; this reduces permeability.
www.whd.co.uk 94 www.understanding-cement.com
Composite Cements
Less efflorescence.
The main properties determining whether an ash is suitable for use in concrete
are its pozzolanic reactivity and the ability of the ash to reduce the concrete
water demand. Too much carbon in the ash increases the water demand, affects
the colour of the concrete and can affect the action of concrete admixtures.
Carbon in ash can usually be accommodated in the mix design, provided the
carbon content does not change; a variable carbon content is more difficult to
deal with.
Strengths
Concrete in which fly ash has replaced some of the Portland cement would be
expected to have lower early strengths due to the slower rate of reaction of the
ash particles compared with Portland cement. However, the basis of any
comparisons has to be considered carefully as one of the principle benefits of the
use of fly ash is that the w/s ratio can be lowered for the same workability, due
to the particle shape of the ash. Other adjustments to the mix design may also
be made.
In large pours of concrete where the heat of hydration raises the concrete
temperature appreciably, the mix containing fly ash will still be weaker than the
Portland cement Control mix at early ages, but will tend to overtake the Control
later; the larger the pour and higher the temperature, the sooner the fly ash
concrete is likely to become the stronger of the two mixes. (Note that the curing
temperatures will not be the same; for equivalent pours, the mix containing fly
ash will have a lower heat of hydration, see below).
Cement and fly ash finenesses will affect the rates of strength growth in these
comparisons.
www.whd.co.uk 95 www.understanding-cement.com
Composite Cements
Heat
Low-lime fly ash reacts more slowly than ordinary Portland cement and
consequently it is widely used in large concrete pours, such as dams, where high
fly ash contents are often used. The maximum temperature reached by the
concrete is much reduced, lessening the chances of thermal fracturing and
increasing the final concrete strength.
Other benefits
Benefits cited by the ash industry, in addition to the above, include a reduction in
efflorescence and in shrinkage and creep. Efflorescence is caused by soluble salts
reaching the surface (particularly calcium hydroxide); these salts are usually
derived from the cement. If the cement content of a concrete is reduced by
partial replacement of Portland cement with fly ash, it follows that less soluble
material will be in the mix and so the likelihood of efflorescence occurring will be
lessened. Additionally, if fly ash reduces the permeability of the paste, less
soluble material will reach the surface, again reducing efflorescence.
The magnitude of any drying shrinkage and creep in concrete is related to the
w/c or w/s ratio of the mix. Since using fly ash as a mineral addition in cement
enables the use of a mix of lower w/s ratio for the same workability, shrinkage
and creep should be reduced.
Silica released from fly ash hydration forms C-S-H, the necessary lime being
derived from CH from cement hydration and by increasing the Si/Ca ratio of the
C-S-H.
Alumina is also released; some enters the C-S-H structure, increasing the Al/Ca
ratio of the C-S-H. The remaining alumina forms other compounds, including AFm
phase, hydrogarnet (mainly C3AH6) and stratlingite (C2ASH8).
Typically, between a quarter and a half of the ash may eventually react (this
might take months or years), depending on factors including ash glass content,
hydration temperature, percentage of cement replacement and the alkali content
of the Portland cement. It is primarily the glassy phase that reacts; the
crystalline inclusions, mainly quartz and mullite, are generally inert. Mullite and
quartz crystals can occasionally be seen in SEM images of fracture surfaces,
protruding from part-reacted ash particles where glass has been extracted by
reaction with the cement pore fluid.
While fly ash contributes to later strength gain through the pozzolanic reaction,
the slow rate of reaction means that, initially, fly ash is almost chemically inert in
the early stages of Portland cement hydration. Before any significant fly ash
reaction has occurred, as with fine limestone particles, the ash particles provide
www.whd.co.uk 96 www.understanding-cement.com
Composite Cements
nucleation sites on which hydration products can form from solution, increasing
the rate of Portland cement hydration at early ages. This partly offsets the slower
rate of strength growth due to the lower fly ash reactivity compared with the
Portland cement it has replaced.
The pozzolanic reaction process by which fly ash contributes to concrete strength,
is believed to be that sodium and potassium in the pore fluid reacts with silicate
material extracted from the fly ash by the action of hydroxyl ions. An alkali-
silicate gel is formed, which is unstable in the presence of calcium and is rapidly
converted to C-S-H.
The presence of sodium and potassium hydroxide in the cement pore fluid will
increase the pore fluid alkalinity and accelerate fly ash reaction. However, lime
(calcium hydroxide) does not contain sodium or potassium hydroxide although a
lime/fly ash mix produces an hydraulic binder. This shows that sodium and
potassium hydroxide are not essential to fly ash reaction; the alkalinity produced
by the lime is sufficient, although the reaction rate will be slower.
Figure 7.7 Polished section of concrete Figure 7.8 Polished section of concrete
made using Portland cement (70%) and fly made using Portland cement (30%) and
ash (30%). Key: c-(and other similar bright slag (70%). Key: circled features are relicts
features) are relicts of partly-hydrated of partly-hydrated cement particles; s-sand
cement particles; s-silica sand; arrowed particles, mainly silica and feldspar;
features are examples of fly ash particles, arrowed, light grey particles are slag
but there are many more. Note that the fly particles. A dark reaction rim can just be
ash particles are not all the same grey seen around some of the slag particles. In
level; this is because each particle has a contrast with the fly ash particles in Figure
different composition, usually only slightly. 7.7, all the slag particles here are of the
Most are medium-grey and composed of same grey level, as they are all of similar
aluminosilicates; a few are bright, iron-rich, composition having been quenched from
particles. (NB: in backscattered electron the same liquid.
images, grey level is approximately
proportional to mean atomic number:
Fe2O3, for example, will appear brighter
than SiO2.).
www.whd.co.uk 97 www.understanding-cement.com
Composite Cements
Because it has a high surface area, it has a high water demand and this limits the
level of cement replacement by microsilica unless water-reducing admixtures are
used.
For more information on microsilica in concrete, see Lea (5) or Taylor (6).
www.whd.co.uk 98 www.understanding-cement.com
Composite Cements
7.5 Limestone
Limestone is a mineral addition widely used in cement. It is neither pozzolanic
nor latently hydraulic.
Small additions of limestone (up to 5%) have been considered in Chapter 5.8. In
the context of mineral additions, the proportions of limestone in the cement are
much higher. For example, European Standard EN 197-1 specifies cements
containing 6%-20% limestone (CEM II/A-LL)1 and 21%-35% limestone
(CEM II/B-LL).
PLC has been used in Europe for many years and is now becoming more widely
used elsewhere. To take an example of how PLC is used, in the UK it is widely
used in both precast and ready-mixed concrete; in Europe as a whole, PLC
represents about a quarter of the cement market.
Cost
Shrinkage control
Reduction in bleeding
Fine limestone in cement is often regarded just as a filler and indeed it does fill
gaps between clinker particles. However, as already discussed (Chapter 6.3),
limestone is not inert, as some of it takes part in the hydration process.
The dilution effect of replacing some of the clinker with limestone can be offset
by grinding the clinker more finely. For the same water/cement ratio, PLC
concrete may have a lower 28-day strength than a concrete containing Portland
cement only, and a higher cement content may be needed to produce similar
strengths.
For the same concrete strength, PLC has similar durability characteristics to
Portland cement in most respects, although its resistance to sulfate attack is still
being evaluated, particularly with regard to thaumasite formation. (Sulfate attack
and thaumasite are discussed in Chapter 9).
1
Or CEM II/A-L; CEM II/B-L for lower purity limestone, see Chapter 10.
www.whd.co.uk 99 www.understanding-cement.com
Composite Cements
In PLC concrete, the rate of hydration of alite is increased by the presence of fine
limestone, giving good early strengths. Hemicarbonate and monocarbonate are
formed from the reaction of aluminate phase with some of the limestone; most of
the sulfate in the cement is likely to be present as ettringite. Only a small
proportion of the limestone will react, although the proportion will increase in
concrete containing Portland cement with a higher aluminate content, or if
additional alumina is available from slag or fly ash in the mix.
Monocarbonate and hemicarbonate formation fill pore space within the concrete,
reducing permeability and thereby increasing durability. They may also make
some contribution to strength.
The increased use of PLC is highly likely if the proportion of Portland cement in
concrete is to be reduced for environmental reasons. Limestone is readily
available, but other mineral additions (particularly slag and fly ash) are in limited
supply. Also, it is undesirable to transport bulk materials further than necessary.
The availability of fly ash and slag depends on coal-fired power generation or iron
smelting; they are not available everywhere without being transported over long
distances but limestone is readily available wherever cement is produced.
References, Chapter 7
Additionally:
Some excellent datasheets (downloadable pdf files) on the use of fly ash in
concrete are available from the UK Quality Ash Association web site at:
http://www.ukqaa.org.uk/Publications.html#TechnicalDatasheets
http://www.heidelbergcement.com/uk/en/hanson/products/cements/ggbs_and_r
elated_products/ggbs_.htm
There is more on slag at the UK cementitious slag makers association web site:
http://www.ukcsma.co.uk/index.html
(Web sites have an irritating habit of changing, so if these pages change and the
links do not work, look around the web site as the pdf files will probably still be
available somewhere.)
Here, we use cement variability to mean the extent to which cement from the
same source, and produced to the same specification, varies over time. In other
words, if you buy a bag of cement produced from clinker made on Monday and
then another bag from clinker made on Friday, will the cement in the two bags
behave in exactly the same way when mixed with water?
On a larger scale, if you receive a delivery of bulk cement to a silo today, will the
cement in this delivery behave identically to that delivered last week or last
month?
The short answer is: Yes, probably. For most practical purposes, they should all
behave in such a similar way that it wont be easy to distinguish cement from the
two bags or deliveries. However, you might just possibly notice differences in
cement characteristics. These differences may be so slight as to make no
practical difference but occasionally they may cause problems.
The main differences that may be noticed typically relate to colour, setting times,
early strengths, later strengths or how the cement responds to admixtures. If
you are manufacturing pre-cast concrete products, early strengths are likely to
be important as they will control how quickly you can de-mould the product and
re-use the mould. If you are an engineer requiring a minimum 28-day strength,
then later strengths will probably be your prime concern.
used to make cement may include waste paper, car tyres or solvents in addition
to the normal gas, oil or coal. The availability and composition of each of these
may vary on a daily basis, requiring constant adjustments to the raw meal
composition and fuel mix. Given all this variability, it would be surprising if the
behaviour of the product never fluctuated just a little.
Despite all of the above problems, in most places modern cement is more
consistent and of a higher quality than ever before. Improvements have been
made possible through developments in the major components of the cement
plant - the kiln and the raw materials and cement mills - but especially in
computer-controlled operations and in on-line analysis.
Prior to a few decades ago, analyses of everything - raw materials, blended raw
materials, clinker and cement - would have been carried out by taking samples,
dissolving them in suitable reagents, then quantifying compositions by titration or
gravimetrically, followed by calculation using pencil and paper. This might all take
several hours, or even up to 24 hours. Today, online sampling and analysis is
carried out automatically in minutes using X-ray fluorescence and X-ray
diffraction.
If, for example, the raw feed composition starts to drift from the target values,
the change should be noticed quickly and corrective action taken, again probably
automatically. Such early recognition of deviations from target compositions,
followed by rapid corrective action, should minimise variability in the composition
of the cement.
Tests carried out on the cement by the manufacturer will depend on the standard
to which the cement is produced, but will typically include tests on composition,
tests for expansion on hydration and tests on strength in standard mixes. These
tests should ensure that the cement conforms to the appropriate standard.
It is simply not possible for a cement producer to test his cement with all of the
materials with which it may be used. These could include fly ash, slag, microsilica
and any of hundreds of admixtures. Even a few basic permutations of all of these
would produce thousands of mixes to be tested - not just once, but every day, or
hour or whatever analysis interval was deemed appropriate. This would clearly be
unrealistic and anyway would add greatly to the cost of the cement.
The price of cement is often a key factor in making purchasing decisions. A major
cement user will often contract with a particular supplier for a period of time (eg:
a year) and a pound or dollar or two either way in the cost per tonne may make
all the difference in deals worth millions. Adding an expensive battery of tests
that will be largely unnecessary wont appeal to either the cement producer or
purchaser.
The cement makers lot is not, then, always a happy one. He has to contend with
variable raw materials and variable fuels. Increasingly, he also has to comply with
environmental legislation that requires him to reduce CO2 emissions or pay heavy
penalties; in turn this may involve the use of alternative fuels that may
themselves have the potential to cause more variability in the product.
He has to comply with other environmental laws limiting what can be emitted
from his site. In more densely-populated areas he has the general public living
nearby, for whom the cement producer is invariably responsible for dust from any
source falling onto their house or car. He has customers for whom price often
appears to be a more important factor than the supply of a consistent product or
good customer service - at least until something goes wrong.
On top of all that, he has shareholders looking for a return on the tens or
hundreds of millions that they invested to build the cement plant in the first
place. If he fails to deliver, the shareholders will sell him to another cement
producer who will demolish his kiln and turn his quarry into a shopping mall. Its
enough to make anyone wish they had listened to mum and dad and become a
...lawyer or a doctor or a civil engineer.
Obviously, the purchaser can expect that all cement will conform to the standard
to which it was produced. However, the standard is likely to permit a range of
physical and compositional characteristics. Usually, a purchaser has the choice of
several different sources of cement, all nominally similar and produced to the
same standard, and he will probably have a preferred source. This preference
may be based on price or on technical considerations.
Perhaps one source has a particularly dark or light colour that suits the
application better than other cements, or maybe it has a particularly good early
strength, or it just works better in his process than other alternatives.
Suppose you are a producer of concrete products who has been using cement
from a particular source for some months without any problems; suddenly, you
notice a change in the strength of the concrete, or that the cement has a
different colour. What may have happened?
Firstly, it might be a different cement, or cement made from a blend of the usual
clinker with other imported clinker. This might happen, for example, when a
cement kiln is shut down for periodic maintenance or because of a technical
problem. The cement producer may bring clinker in from elsewhere to
supplement the lost production. Usually, he will advise his larger customers in
advance that this will happen, or he should do. If a change in cement
characteristics coincides with a kiln shutdown, this may well be the explanation.
With luck, normal cement characteristics will be restored when the kiln restarts.
However, suppose you notice changes in behaviour over time in cement from a
single source, even though no major alterations in the raw materials or clinker
composition were apparently made over this period. Surely something must have
changed? Well, yes it might have.
Provided that test concrete (or mortar) mixes are prepared at a constant
water-cement ratio, minor imbalances in the rate of supply or
consumption of sulfate should not make much difference to strengths.
Specifications for cement in national standards normally specify testing at
particular water-cement ratios. However, the ready-mixed concrete
industry, and others, generally prepares concrete to a particular slump,
not to a particular water-cement ratio.
This situation could arise, for example, where two cements have a similar
oxide composition but one has been milled hotter than the other,
decomposing the gypsum to hemihydrate. As hemihydrate is more soluble
Anything that affects the rate at which the C3A requires sulfate to control
setting, or the rate at which sulfate becomes available from all sources,
may have a similar effect. This could be a change in the total cement
sulfate content, a change in the proportion of clinker sulfate to total
sulfate, a change in cement fineness, a change in the C3A content of the
cement, a change in C3A reactivity or a change in the extent to which C3A
and ferrite crystals are intergrown in the clinker.
Optimum gypsum: the gypsum content of the cement affects the rate of
strength development and the volume stability of the concrete.
Given all the things that could go wrong, of which the above is really just an
introduction, you might feel that maybe cement isnt quite so reliable after all.
However, the extent to which that impression is wrong is really a tribute to the
evolution of the manufacturing process over many years. Of course, things can
occasionally go badly wrong, but, fortunately, that is now unusual. Or, at least, it
should be.
In alkali-silica reaction, alkalis in the cement paste fraction of the concrete react
with with certain types of reactive silica to form an expansive gel.
Background
Alkali-silica reaction (ASR) can cause serious expansion and cracking in concrete,
resulting in major structural problems and sometimes necessitating demolition.
Although first recognised in the 1940s in the USA, it wasnt until the 1970s that
ASR really became the focus of attention of concrete science worldwide. It also
came to the attention of the media, who promptly dubbed it concrete cancer.
Extensive research was undertaken, which later led to better practise in the
production of concrete to minimise the risk of ASR in the future. Of course, it
wasnt purely scientific interest that had led to research funding suddenly being
available; rebuilding major structures damaged by ASR was becoming expensive.
The reaction
ASR is caused by a reaction between the hydroxyl ions in the alkaline cement
Deleterious processes in concrete
pore solution in the concrete and reactive forms of silica in the aggregate. These
reactive forms of silica are typically chert, quartzite, opal and strained quartz
crystals. Other materials, such as glass can also produce ASR. Pyrex glass and
calcined flint have been used as test reactive aggregates in experimental
concrete mixes. Ordinary glass (as in household jars and bottles) is also reactive.
Undispersed agglomerations of microsilica can also cause ASR.
Silicate anions are detached from the reactive aggregate by hydroxyl ions in the
cement pore fluid; sodium and potassium ions are the ions most readily-available
to balance the silicate anions and an alkali-silicate gel is formed. This can take up
(imbibe) water and is mobile. The gel is unstable in pore fluid containing
dissolved calcium, and calcium silicate hydrate (C-S-H) is produced. This releases
sodium and potassium from the gel into the pore fluid, increasing the pore fluid
hydroxyl ion concentration, with the potential for continued reaction.
Figure 9.4 SEM image of ASR in chert Figure 9.5 Detail of Figure 9.4 showing gel
particle (centre); internal microcracking is extruded from chert particle into a crack in
widespread and some cracks extend into the concrete. Ettringite is also present
the surrounding concrete. (lower right) in a gap between the cement
paste and a sand particle.
Usually, the best method to confirm that ASR has occurred is to examine the
concrete in either thin section using a petrographic microscope, or in polished
section using a scanning electron microscope (SEM). Gel may be seen in cracks
and in aggregate particles.
In the case of alkali-silica reaction, the reaction usually occurs much later,
possibly years after the concrete was placed. Large aggregate particles (large,
that is, compared with cement-sized pozzolan) generate a significant volume of
gel, which then takes up water and expands within the hardened, mature,
concrete.
Figure 9.6 This is the same image as in Figure 9.5, with X-ray spectra superimposed
showing how alkali-silica gel composition changes with time to become more like that of
the surrounding calcium silicate hydrates. At A the gel spectrum shows large peaks due to
silicon and potassium and only a very weak peak due to calcium. At B the calcium peak
has become much stronger and the potassium peak much weaker. At C the potassium
peak has disappeared entirely and the gel has approximately the same composition as the
normal calcium silicate hydrate comprising the bulk of the cement paste. Clearly, the gel is
older with increasing distance from the aggregate particle in which it originated - the
'oldest' gel has had more time in which to take up calcium from the surrounding paste,
and has now become calcium silicate hydrate.
Expansion of the gel as water is taken up, is likely to result in damage to the
surrounding concrete. Over time, the gel will slowly take up calcium and
eventually the composition of the gel may become similar to that of the calcium
silicate hydrate in the cement paste (see Figure 9.6). By then, though, the
concrete may already be severely damaged.
Water - needed for the gel to expand; expansion due to ASR will not occur
if there is no available water in the concrete.
The cement is usually the main source of alkali in the pore fluid in concrete,
derived from alkali sulfate and from alkalis in the main clinker minerals, mainly in
belite and aluminate. So if we restrict the alkali content of the cement, wont that
limit the alkali in the pore fluid? Yes, it will, but this does not allow for different
cement contents in concrete, so simply imposing a limit on cement alkalis is not a
complete answer to the problem. Also, we need to remember that higher alkali
cements tend to give better early concrete strengths; alite and belite hydrate
faster as pore fluid alkalinity increases.
Imposing limits on cement alkali content that were too low (however defined)
may therefore affect cement performance. It would also increase the cost of
cement production, as bypass dust and precipitator dust high in alkalis could not
be returned to the kiln and would have to go to landfill. Additionally, some raw
materials high in alkali could not be used and would have to be replaced by
possibly more expensive alternatives. So, although restricting the cement alkali
content makes an important contribution to controlling ASR in concrete where
potentially reactive aggregate is to be used, it is not necessarily a complete
answer to the problem.
Upper limits for sodium oxide equivalent on low alkali cement are typically of the
order of 0.60%-0.75%; check the standards or codes of practice applicable to
your location.
An alternative approach is to limit the alkali content of the concrete. This would
then take into account that different concretes have different cement contents.
Limiting the alkali content can be achieved by using lower alkali Portland cement,
by using a lower cement content in the concrete or by partially replacing the
Portland cement with slag, fly ash or microsilica.
For example, using some typical Figures from Taylor, fly ash typically contains
1.5% Na2O and 4.2% K2O (1) and slag contains approximately 0.4% Na2O and
0.7% K2O (2). Compared with the alkali levels in Portland cement, slag is
therefore broadly similar and fly ash is considerably higher. However, both are
less reactive than Portland cement and even in mature concrete some slag or fly
ash will remain unreacted.
Only the reacted fraction of the slag or fly ash will have released alkali into the
pore fluid. As this proportion is not known precisely, and anyway will vary
between different concrete mixes, it is difficult to calculate just how much alkali
from the slag or fly ash has become available for reaction. National Rules will
determine what substitution levels are effective and what allowance to make for
slag or PFA alkalis.
Can we restrict water availability? In most cases, not really, unless the concrete
is entirely under cover and will never get wet. While expansion due to ASR is
unlikely in concrete that remains dry inside buildings, even building interiors can
become saturated in flood conditions. In any case, most concrete is used in
foundations, roads, bridges or other open structures where it will be exposed to
water.
1
The constant 0.658 is derived from the ratios of atomic weights of sodium and
potassium oxides: Na 22.99; K 39.10; O 15.99; therefore Na2O (22.99 x
2)+15.99=61.97; K2O (39.10 x 2)+15.99=94.19; 61.97/94.19 = 0.658.
If concrete does not contain reactive aggregate, it follows that there will be no
expansion due to ASR. So, can we prevent expansion by not using reactive
aggregate in concrete? Yes, we can, but we need to decide what we mean by
reactive and we also need some method by which we can determine whether or
not a particular aggregate is reactive.
In some areas, chert represents the bulk of aggregate used in concrete, with no
resulting expansion due to ASR because the proportion of chert in the concrete is
well above the pessimum. If we excluded all potentially reactive aggregate, we
would not be able to use chert, but such a restriction is unnecessary since it can
clearly be used safely. We just need to avoid aggregate that contains chert near
the pessimum proportion.
There are broadly three types of test for alkali reactivity of aggregate:
Chemical tests include ASTM C289 - 07 Standard Test Method for Potential Alkali-
Silica Reactivity of Aggregates (Chemical Method).
Expansion tests include ASTM C227-03 Standard Test Method for Potential Alkali
Reactivity of Cement-Aggregate Combinations (Mortar-Bar Method) and British
Standard BS 812: Part 123: 1999 Testing aggregates - method for determination
of alkali-silica reactivity concrete prism method.
Conventional expansion tests with mortar bars or concrete prisms are the closest
simulation to real concrete but may require two years or more to show results;
this may not always be convenient. Accelerated expansion tests can shorten this
timescale but do not simulate real conditions as closely.
In summary, all these tests for aggregate reactivity can be very useful indeed
essential - but all have potential problems and the perfect test for ASR
susceptibility has yet to be developed.
Yes, definitely. Despite the limitations weve looked at above, the incidence of
expansion due to ASR has declined markedly over the last 10-20 years where
these different approaches to limiting expansion have been applied. For example,
in the UK, there have been no reported cases since the current rules were
introduced in the 1980s. Generally, codes of practice and national standards
define how these approaches are used in practice in different parts of the world.
Usually, several strategies to limit expansion are applied at the same time.
For example, aggregate producers will routinely have their aggregate examined
petrographically to determine the rock types and their potential reactivity. These
aggregates then go into concrete that contains cement that has limits on alkali
content specifically to minimise the risk of ASR; the concrete may well also
contain fly ash or slag or other mineral addition to limit the available alkali.
Of course, these relatively recent ways of dealing with the problem only control
the problem in concrete produced using these methods. Instances of ASR still
occur in some older concrete because there were fewer controls over what was
used to produce it. Even where good controls are apparently in place, there is
always the potential for things to go badly wrong.
The dissolved carbonate can then react with calcium hydroxide to form calcium
carbonate, releasing hydroxyl ions into the pore fluid, which can then react with
more dolomite:
ACR was first identified in Canada and has also been confirmed in China and the
Middle East (and probably other places too) but it is rare overall in comparison
with ASR.
In any case, the most reliable method of identifying ASR or ACR is microscopic
examination of the concrete. Typically, cores are taken from the concrete for
petrographic examination using either a petrographic microscope to examine thin
sections of the concrete, or by SEM examination of polished sections of pieces of
the concrete.
The majority of instances of sulfate attack have two main effects on the cement
paste:
The need for additional calcium (and water) to form ettringite from monosulfate
is shown by the following equation:
We therefore typically have two processes occurring at the same time. Cracking
is generally due to the expansive pressure caused by the growth of ettringite
within the cement paste. Weakening of the paste by decalcification of the calcium
silicate hydrate (C-S-H) that gives concrete its strength, causes further damage.
Eventually, the concrete disintegrates.
Sulfate attack causes expansion and cracking, and a general loss of concrete
strength, with ettringite often widespread within the cement paste in affected
areas. However, the quantity of ettringite that can form is limited by the available
alumina from the cement. When this is exhausted, gypsum may form if there is a
continued supply of sulfate; gypsum formation also requires additional calcium,
obtained from CH or C-S-H decalcification, and the deterioration of the cement
paste continues.
This is the more common type and typically occurs where water containing
dissolved sulfate penetrates the concrete. A fairly well-defined reaction front can
often be seen in polished sections; ahead of the reaction front the concrete is
normal, or near normal. Behind the reaction front, the composition and
microstructure of the concrete will have changed. These changes may vary in
type or severity but commonly include:
Extensive cracking
Expansion
The effects described so far are typical of attack by solutions of sodium sulfate or
potassium sulfate. Solutions containing magnesium sulfate are generally more
aggressive, for the same concentration. This is because the magnesium also
takes part in the reactions, replacing calcium in the solid phases with the
formation of brucite (magnesium hydroxide) and magnesium silicate hydrates.
The displaced calcium precipitates mainly as gypsum.
Sources of sulfate, other than groundwater already mentioned, which can cause
sulfate attack include:
Figure 9.7 SEM image of sulfate attack in Figure 9.8 SEM image of sulfate attack in
concrete; ettringite has replaced C-S-H concrete; here, ettringite (examples - e) has
within the marked areas. largely replaced C-S-H throughout this field of
view.
This can occur when a source of sulfate in sufficient quantity is incorporated into
the concrete when mixed. Examples include the use of sulfate-rich aggregate,
excess of added gypsum in the cement, oxidised sulfide minerals in the
aggregate, or contamination. Proper screening and testing procedures should
generally avoid internal sulfate attack.
normally arises from the sulfate in the cement and may occur some time
months or years after the concrete was placed.
In normal concrete where no sulfate ingress has occurred, the total amount of
ettringite that forms as the result of DEF is evidently limited by the sulfate
contributed by the cement initially. It follows that the quantity of ettringite
forming due to DEF is relatively small. Ettringite crystals form widely-dispersed
throughout the paste. If expansion due to DEF causes cracking, this is
particularly likely around aggregate particles; ettringite may subsequently form
in the cracks but this does not necessarily mean the ettringite in the cracks was
the cause of the initial cracking.
DEF causes a characteristic form of damage to the concrete. While the paste
expands because ettringite crystals are forming within it, the aggregate does not
expand. Consequently, gaps form around these non-expanding 'islands' of
aggregate within the paste (Figures 9.9-9.11). Often, these peripheral gaps
become filled with ettringite (Figure 9.11).
Once gaps have formed at the periphery of aggregate particles, any further
ettringite formation is likely to occur within these gaps, where it can form freely,
rather than within the paste where the crystals would have to exert an expansive
force in order to create space in which to crystallise.
Figure 9.10 DEF expansion: SEM image Figure 9.11 DEF expansion: SEM image
showing a limestone particle with a gap showing ettringite (e) filling a gap between
between the limestone and the cement paste a limestone aggregate particle (L) and the
(arrowed). cement paste. Small silica sand grains (s)
remain bonded to the cement paste and do
not show peripheral gaps.
The images in Figures 9.10 and 9.11 are characteristic of damage to concrete
due to DEF. Where peripheral gaps form around aggregate particles, the
aggregate is no longer contributing to concrete strength, since it is effectively
detached from the cement paste. Often, these gaps become filled with ettringite;
the ettringite does not form a strong bond between the paste and aggregate and
so the concrete is still weakened.
The effect of cement composition on DEF is not well understood. Some factors
correlate strongly but the causes are not clear. In laboratory tests, DEF expansion
has been shown to correlate positively with cement-related factors, including:
high sulfate
high alkali
high MgO
cement fineness
high C3A
high C3S
Expansion due to DEF and ASR have been linked; in one study, initial expansion
due to ASR was enhanced by subsequent expansion due to DEF (4).
From the above, you will probably have guessed that delayed ettringite formation
is not fully understood.
Thaumasite can form in both concrete and mortar. The normal cement hydration
products, mainly calcium silicate hydrate and calcium hydroxide, are decomposed
as a result of both sulfate attack and of carbonation. Thaumasite itself is a weak
and friable mineral, so when it replaces the calcium silicate hydrate, the concrete
is severely weakened.
[Ca3Si(OH)6.12H2O] (SO4)(CO3)
or
CaSiO3.CaCO3.CaSO4.15H2O
Normal sulfate attack usually results in the formation of ettringite. This uses
aluminium provided by the cement and clearly this is limited in quantity in
normal concrete. However, thaumasite formation does not involve aluminium;
given an adequate supply of sulfate and carbonate, thaumasite can continue to
form until the calcium silicate hydrate is completely decomposed. Consequently,
while the use of sulfate-resisting Portland cement provides some defence against
normal sulfate attack, it does not give any particular protection against
thaumasite formation.
Efflorescence (Chapter 9.7) may be associated with sulfate attack, with white
crystalline growth on the surface of the damaged masonry or render.
DEF would not normally be expected to occur in mortar unless the mortar had
been heated by an external source; the heat of hydration of the cement would
normally be dissipated by the bricks or stone.
Severe damage to concrete and mortar can be readily identified from visual
inspection. A range of different analytical techniques can be used to confirm that
sulfate attack has occurred. A calculation of the cement content by chemical
analysis, together with a determination of the sulfate content, can indicate that it
may have occurred if the sulfate content is appreciably higher than would be
expected from sulfate normally present in the cement. This is perhaps the
simplest test, although it is not conclusive - for example it would not normally
indicate that DEF had occurred, since sulfate will probably not be present in
excess.
9.3 Carbonation
Although included in this chapter on deleterious processes, not all of the effects
of carbonation on concrete are bad. It increases concrete density and so reduces
permeability. Carbonation is associated with the corrosion of steel reinforcement
and with shrinkage but it also increases both the compressive and tensile
strength of concrete.
Carbonation occurs when carbon dioxide dissolves in the pore fluid of the
concrete to produce carbonate ions. These react with calcium to produce calcium
carbonate, usually in the form of calcite, although aragonite can form in hot
conditions.
Because the process involves dissolved carbon dioxide, a film of pore fluid on the
hydration products in the concrete is necessary for carbonation to occur. As
carbonation proceeds, calcium is supplied by the calcium hydroxide in the cement
paste, and also by the calcium silicate hydrate (C-S-H). Calcium hydroxide is
depleted and the ratio of calcium to silicon in the C-S-H decreases. In fully
carbonated cement paste, most of the calcium is present as calcium carbonate.
The C-S-H has lost its calcium and is now an impure silica gel, also containing
aluminium from other cement hydration products and sodium and potassium
from the pore fluid (Figure 9.13).
Saturated concrete carbonates slowly, because the capillary pores in the concrete
are full of water and this inhibits further ingress of water or gas. In permanently
dry conditions, carbonation proceeds slowly or not at all, as there is not enough
water in contact with the hydration products.
The surface of fresh concrete carbonates very quickly, within a few hours or days,
but the depth to which the concrete is carbonated is, at first, very limited -
perhaps only a few microns.
Over time, the carbonation front gradually moves further into the concrete.
After a year, the distance from the surface that the carbonation front has reached
may be 1 mm to perhaps 5 mm, depending on the permeability of the concrete.
Dense, impermeable, concrete made with a low water/cement ratio will carbonate
much more slowly than porous concrete made using a high water/cement ratio.
Carbonate ions have to travel a greater distance from the surface to the
carbonation front as the carbonation depth increases. Consequently, the rate of
progress into the concrete of the carbonation front gradually slows.
If you apply phenolphthalein indicator to concrete and it quickly turns purple, the
concrete pore fluid is strongly alkaline. In Figure 9.14, the carbonation depth
shown by this rapid test is clearly about 5mm-7mm. The only tools needed are
the indicator liquid and a hammer and perhaps a cold chisel to fracture the
concrete.
Obviously, the phenolphthalein test is rather crude, but its simplicity makes it
very useful. It is important to make sure the concrete surface being tested is a
freshly-fractured surface; testing an old surface wont tell you anything as it will
have started to carbonate. Ideally, you would break the concrete (or take mini-
cores or use drilling dust) then test it immediately.
d=kt
Occasionally, this can be quite useful. Suppose you find that a piece of concrete
that you know to be nine years old has a carbonation depth of 12 mm; for that
piece of concrete in that particular environment, k is therefore 4.
You can then work out the likely carbonation depth at any other time; after three
months it was only 2 mm and after 100 years it will be 40 mm. Dont expect this
relationship between time and carbonation depth to work perfectly. All things
being equal, it should work well. However, in cement and concrete all things are
never equal. Measurement of carbonation depth is imprecise because of
distortions in the carbonation front due to factors such as aggregate, microcracks
and local variations in concrete porosity. Also, the carbonation depth may be
affected locally by wind direction, nearby sources of carbon dioxide (eg: road
traffic) and probably a myriad of other factors.
Normal concrete pore solution is saturated with calcium hydroxide and also
contains sodium and potassium hydroxide; the pH is typically pH 13-14. Concrete
with a pore solution of pH 10-12 is less alkaline than sound concrete, but would
still produce a strong colour change with phenolphthalein indicator. It therefore
follows that the indicator test is likely to underestimate the depth to which
carbonation has occurred, as will be seen below.
Below about 4 mm depth, unhydrated cement particles remain and the paste
microstructure has not obviously altered greatly from its normal, high-alkaline,
state (Inset B, 7 mm depth). However, even here, the paste appears slightly
altered, becoming a little denser and with a loss of calcium hydroxide (compare
with the image in Figure 6.5).
The concrete mid-way through the slab is shown in Figure 9.16. Even here, there
are signs of slight carbonation. X-ray microanalysis of the C-S-H would show a
slight increase in the Si/Ca ratio.
Nevertheless, despite its limitations, the indicator test is very useful as a means
of making a quick and easy initial assessment and it is very widely used. Use it,
but be aware of its limitations.
Highly carbonated
~4mm depth
Less carbonated
Figure 9.15 Polished section of concrete from the same slab as in Figure 9.14.
Steel that has sufficient cover of dense, impermeable, concrete (the cover
zone) should last almost indefinitely, as the alkaline conditions in the concrete
protect it from corrosion. However, the importance of quality control to ensure
adequate cover cannot be overemphasised; porous and permeable concrete,
made with a high water/cement ratio and containing steel near the surface, is
particularly vulnerable to damage.
Normally, a thin layer of iron oxide called the passivation layer around the steel
prevents it from oxidising further. Provided the pH of the pore fluid remains high
(above approximately 11.5), the steel should remain in good condition. However,
if the alkalinity of the surrounding concrete pore fluid falls, as occurs when
concrete carbonates, the passivation layer is lost and the steel corrodes.
The corrosion process is expansive due to the formation of rust. This causes
pressure to be exerted on the surrounding concrete, resulting in cracking and
spalling. Also, the bond between the steel and the surrounding concrete is
weakened or broken and the steel is weakened through becoming thinner. In
addition to damage to the concrete, the benefits of the presence of the steel
reinforcement are therefore also lost.
For corrosion to occur, the steel must be in contact with the pore solution, which
acts as the electrolyte. Oxygen must be also present. This need for oxygen
means that corrosion may not occur in carbonated concrete provided it is
sufficiently dense to restrict the supply of oxygen.
For corrosion to be induced by chloride, the chloride ions have first to penetrate
the concrete. The permeability of the concrete will be important in controlling
how quickly chloride can diffuse through the microstructure and reach the steel.
An additional factor affecting the rate of corrosion is the extent to which the
cement hydration products can bind chloride ions; AFm phase can bind chloride
to form Friedel's salt (C3A.CaCl2.10H2O). Friedels salt is frequently seen when
using scanning electron microscopy and X-ray microanalysis to examine concrete
in polished section that has been exposed to road de-icing salt.
9.5 Leaching
Leaching in concrete is the process by which soluble material is extracted from
the concrete by flowing water. The water may flow over the concrete surface, or
percolate through cracks. Water readily extracts sodium and potassium hydroxide
from the concrete pore fluid, lowering the pH; it also progressively dissolves
calcium hydroxide and decomposes calcium silicate hydrate (C-S-H), ettringite
and AFm phase.
Since it is C-S-H that gives concrete its strength, C-S-H decomposition gradually
reduces the strength of the affected concrete until it disintegrates. Calcium is
leached from the C-S-H, leaving an impure hydrous silica gel. As leaching
progresses, the concrete becomes more porous and permeable (Figure 9.17) and
may also be susceptible to damage by other mechanisms such as frost attack.
Dense concrete, made using a low water/cement ratio mix, is obviously likely to
be less affected by leaching than porous concrete made using a high
water/cement ratio mix. Other factors affecting how quickly leaching can damage
concrete include whether the water is hard or soft, water temperature and what
is already dissolved in the water. Concrete exposed to acidic water, such as from
peaty moorland, is particularly at risk.
The affected concrete surface typically has a rough appearance, with flakes of
concrete gradually becoming detached.
Air entrainment of concrete can greatly improve the resistance of concrete and
mortar to frost damage. A surfactant admixture is used to lower the surface
tension of the pore fluid, resulting in the entrapped air in the concrete being
distributed between myriads of tiny bubbles. These small voids are typically
10m-200m in diameter.
The voids prevent or limit damage by providing space into which the pore fluid
can expand on freezing. To be effective, the voids need to be closely spaced. The
void spacing factor is the average value of the maximum distance from any
point in the paste to the nearest void. Typical values range from 100m -200m
but the void spacing factor necessary to give good frost protection varies from
one concrete to another.
The problem is more likely to occur where bricks have been left unprotected and
are saturated before construction.
References, Chapter 9
Alkali-silica reaction:
BRE Digest 330, 2004: Alkali-silica reaction in concrete: 2004 edition - four part
set.
Leas Chemistry of Cement and Concrete, 4th edition, ed. Peter Hewlett,
Elsevier, 1998, p964.
Steel Corrosion
Leaching
Air entrainment
For example, one of the first specifications drawn up by the British Standards
Institute after it was formed in 1901 was a standard specification for Portland
cement (BS 12, 1904). Other early specifications were for electric power cables
(BS 7), steel for use in railway lines or railroad tracks (BS 11), tramway and dock
rails and fishplates (BS 2) and steel for steam boilers (BS 14). Such materials
were fundamental to the brave new world of industrial society and so these were
the priorities for quality control.
Cement was fundamental then and it still is, of course. This e-book is not really
about standards but about the principles of how cement works but well just have
a quick look at two important cement standards.
While Portland cement is basically limestone and clay (or other suitable
materials) heated at around 1450 C, standards for Portland cement differ in
detail. The compositional requirements differ to some extent and the methods of
testing cements for compliance with the standards are also different.
The two principal worldwide standards for Portland cement are: EN 197-1
(Europe) and ASTM C 150 (USA). Additionally, although no longer used in the
United Kingdom, the British Standard BS12: 1996 is still often specified in
international cement trading.
EN 197-1 covers not only Portland cement but also cements containing Portland
cement and other materials. ASTM C150 applies to Portland cement only.
Standard Specifications for Portland Cement
Five types of Portland cement are described (Table 10.1), with three more air-
entrained sub-types.
Limestone, up to 5% by mass.
Compositional and physical requirements are given for each of the cement types,
with references to appropriate ASTM test methods.
Annex A1 of the standard shows how to calculate the potential cement phase
composition. This calculation allows for the presence of calcium sulfate and
interground limestone by using a modified form of the traditional Bogue
calculation.
Cements containing other constituents such as fly ash and slag are specified in:
and
Cements are specified by both their constituents and by their strength class.
Initially, this looks a bit complicated but all will become clear.
First, we need to look at the cement constituents; these dictate the basic cement
type (Table 10.2).
CEM II, CEM III, CEM IV and CEM V cements all contain Portland cement
mixed with other materials.
Macs
All types of cement may contain up to 5% macs. Macs are minor additional
constituents. Typical examples of macs are fine limestone, kiln dust or fly ash. A
cement containing a mac is considered for specification purposes to be the same
as if the mac were not present and the cement has to meet the same
performance criteria.
Codes
Other than for CEM I, the following codes are necessary in order to interpret the
cement types:
The suffix A means a low level of addition of the mineral addition and the
suffix B means a high level of mineral addition (Table 10.3). For CEM III
blastfurnace cements, suffixes A, B and C indicate increasing proportions of
slag.
For the CEM II cement type, codes for the second main constituent are
specified as follows:
S blastfurnace slag
D silica fume
P natural pozzolana
Q natural calcined pozzolana
V siliceous fly ash
W calcareous fly ash
L or LL limestone (LL means high-purity limestone)
T burnt shale
M two or more of the above
Table 10.3 EN 197-1 CEM II cements: these are some of the more common
cement types.
Designation Name Second main % Second % Clinker
constituent main
constituent
CEM II/A-S Portland-slag Blastfurnace 6-20 80-94
cement slag
CEM II/B-S Portland-slag Blastfurnace 21-35 65-79
cement slag
CEM II/A-D Portland-silica Silica fume 6-10 90-94
fume cement
CEM II/A-V Portland fly- Fly ash 6-20 80-94
ash cement
CEM II/B-V Portland fly- Fly ash 21-35 65-79
ash cement
CEM II/A-L (or Portland- Limestone 6-20 80-94
-LL) limestone
cement
CEM II/B-L (or Portland- Limestone 21-35 65-79
LL) limestone
cement
As will be inferred from the list of codes for the second CEM II constituent, the
list of cement types in Table 10.3 is not complete; the cements shown are the
more common types but there are others specified in the standard, containing for
example, calcareous fly ash or burnt shale.
Cements are also specified according to their strength class. Three standard
strength classes are defined, based on the minimum 28-day mortar prism
strength:
32,5
42,5
52,5
(NB: The use of the comma is part of the specification; units are MPa.)
The minimum and maximum strengths of the three strength classes are shown
below (Table 10.4).
Table 10.4 EN 197-1 cement strength classes for all CEM cements.
Strength Minimum Minimum Minimum Maximum
class 2-day strength 7-day 28-day 28-day
strength strength strength
32,5 N - 16 32,5 52,5
32,5 R 10 - 32,5 52,5
42,5 N 10 - 42,5 62,5
42,5 R 20 - 42,5 62,5
52,5 N 20 - 52,5 -
52,5 R 30 - 52,5 -
CEM I 52,5 R
is a high early strength Portland cement producing at least 52.5 MPa at 28 days
and at least 30 MPa at 2 days, based on mortar prism strength tests.
CEM II/A-LL 42,5N: this is a Portland composite cement with a low proportion of
the second constituent (the A), the second constituent being high-purity
limestone (the LL). The proportions would be 80%-94% Portland cement clinker
and 6%-20% high-purity limestone. The cement is classed as producing normal
strength development (the N): at least 42.5 MPa at 28 days and at least 10
MPa at 2 days, based on mortar prism strength tests.
(CEM II/B-V 32,5R): this is a Portland composite cement with a high proportion
of the second constituent (the B), the second constituent being siliceous fly
ash. The proportions would be 65%-79% clinker and 21%-35% fly ash. The
cement is classed as producing rapid strength development, at least 10 MPa at 2
days, and at least 32.5 MPa at 28 days, based on mortar prism strength tests.
There are additional low early strength classes for CEM III cements covered by
EN 197-4:2004.
Although the system may appear complex, the combination of the cement
designation and the strength class conveys a lot of information very succinctly. It
tells you what the cement contains, and in roughly what proportions; it also
indicates the early and the 28-day strengths based on the standard mortar prism
test.
Cements with a very low heat of hydration are specified as Special and are
included in EN 14216; these have a characteristic heat of hydration of not more
than 220 J/g.
Also:
1
EN 196-4 is absent because of practical difficulties in distinguishing some constituents chemically in
cement, such as microsilica. A draft text for Part 4 is in Technical Report CEN/TR 196-4: 2007
Methods of testing cement - quantitative determination of constituents. This gives a useful guide to
techniques.
While the compositional criteria for Portland cements under EN-197 and C 150
Portland cements may be broadly similar, they are not identical and the testing
procedures are quite different. For example:
The upper limit for loss on ignition for CEM I cements under EN-197 is
5%; under ASTM C 150-07 it is 3.0% for all types except Type IV, where
the limit is 2.5%.
The upper limit for Insoluble Residue under EN-197 for CEM I is 5%;
under ASTM C 150-07 it is 0.75% for all cement types.
If you were just laying a base for your garden shed, you might not worry too
much about the details of the specification process. However, suppose you were
an engineer on a major construction project who authorised the use of a cement
other than that specified. Faced with newspaper headlines like:
you would not be starting from a strong position in trying to argue that the
cement had not contributed to the problem, whether or not it actually had. So,
dont mix the standards.
The most useful further reading would clearly be to obtain copies of the
standard(s) most relevant to your location. Note that all standards are copyright.
C 150-07
ASTM Standard C 150-07, 2007, "Standard Specification for Portland Cement,"
ASTM International, West Conshohocken, PA. Available for download from:
www.astm.org
Price $37.00.
EN 197-1
BS EN 197-1:2000 Cement. Composition, specifications and conformity criteria
for common cements. Available for download from the British Standards
Institute shop at:
www.bsigroup.com/en/
Price 168.00 (UK Pounds).
(No, that isnt a typo EN 197-1 will cost you 168, or very roughly $275 US at
the exchange rate at the time of writing. However, you do get 52 pages,
compared with 8 pages for ASTM C 150-07, and EN 197-1 covers a wide range of
cements not just Portland cement.)
In the first part of this chapter, we are going to look at the two principal ways
that cements for different applications can be produced:
Combine Portland cement with other materials, eg: slag, fly ash
In the second part of this chapter, we dig deeper. The aim is to develop an
intuitive understanding, a minds eye image in order to visualise both clinker
and hydrating systems.
The requirements of cement differ depending how concrete (or mortar) made
from it is to be used. For example, pre-cast products usually need a cement that
gives high early strength so that the concrete can be de-moulded and the moulds
re-used. Large pours of concrete need a cement with low heat of hydration to
minimise thermal cracking. Where sulfate is present in groundwater, the concrete
needs to be resistant to deterioration through sulfate attack.
These different demands need different cements with which to make the
concrete.
The characteristics of Portland cement can be altered in two principal ways. The
Cement concepts
first is to modify the cement clinker to produce the desired characteristics and
the second is to mix Portland cement with other materials such as slag or fly ash.
The second approach can be seen in the European EN-197 standard specification,
which describes only one Portland cement, CEM I, and a long list of other
cements containing mixtures of Portland cement and other materials.
Of course, blended cements are available in the USA, and some concrete in
Europe is still made with CEM I only, although the proportion has been
decreasing for some years. In the UK, CEM I is still available in bulk, but in
bagged form it is scarce, except for white cement. Typically, a bag of cement at a
builders merchant is CEM II/A-LL, a Portland-limestone cement, or CEM II/B-V, a
Portland-fly ash cement.
Suppose you wanted to make a high early strength Portland cement. You would
want to produce a clinker that had a high alite content and a low belite content,
since alite is more reactive than belite. You might also want to increase the alkali
content, as this will increase the alkalinity of the cement pore fluid and accelerate
hydration; of course, this might lead to problems with alkali-silica reaction if
reactive aggregates are present in the concrete, but here we are just talking
about principles.
You would also grind the cement finer as smaller particles will react more quickly.
You might also produce a mineralised cement; alite in mineralised cement is
more reactive than alite in normal Portland cement.
Of the above approaches, the two most commonly used are finer grinding of the
cement and increasing the alite content.
In ASTM C150-7, there are no particular limits on C3S and C2S contents for Type
III cement but the limits for fineness applicable to other cements are removed,
and minimum one-day compressive strengths are specified together with higher
minimum compressive strengths at 3 days than for other cements.
In EN 197, the specification for high early strength CEM I is similarly empirical
the cement has to meet the required strengths at 2 days and 28 days for that
strength class.
If you wanted to make a cement that had a low heat of hydration, you would
want to limit the proportions of the most reactive clinker minerals, the alite and
aluminate. The cement would then have a substantial belite content.
To illustrate this, the ASTM C 150-7 standard specification prescribes limits for
the clinker phases in cements of low or moderate heat of hydration:
For a Type IV cement (low heat of hydration), the maximum C3S content
is limited to 35% and there should be at least 40% C2S. Also, the
maximum permitted C3A content is 7%.
(NB: phase proportions here are calculated using the prescribed Bogue-type
formula; these limits do not apply in C 150-07 if optional heat of hydration limits
are to be applied).
Sulfate-resisting:
Instead of producing C3A, most of the aluminium will be then present in the
ferrite phase, which will have a higher ratio of iron to aluminium compared with
the ferrite in normal Portland cements. The objective in minimising the aluminate
content is obviously to limit the potential for the formation of ettringite in sulfate-
rich environments.
Weve just seen that one way of varying cement properties is to adjust the
composition of the Portland cement clinker. Another way is to blend Portland
cement with other cementitious materials, just as the Romans did when they
blended lime mixed with volcanic glass or ground brick and tile to produce
pozzolanic cements.
Here, we will use composite cements to mean blended cements with at least
two cementitious components, regardless of how or where they were blended.
To make a high early strength cement, the simplest approach to achieve the
highest early strengths would be not to blend anything with Portland cement;
most materials you might add, such as slag or fly ash, react more slowly than
Portland cement. However, microsilica and metakaolin are highly reactive and
may improve early strengths. Sprayed concrete mixes may contain metakaolin or
microsilica. Mixes of Portland cement and CAC can give early strength due to
flash setting.
1
Intergrinding means that the additional material (eg: slag or fly ash) is added to the clinker in the
cement mill. Blending means that the materials are mixed in some other way, usually using
specialised dry-blending equipment.
Sulfate-resisting:
Composite cements made with Portland cement and either fly ash or slag can
give at least equivalent sulfate resistance characteristics to that obtained when
using sulfate-resisting Portland cement.
Suitably-designed mixes containing a Portland fly ash cement with, for example,
70% Portland cement and 30% fly ash should have good sulfate-resisting
qualities. Alternatively, a Portland blastfurnace cement could be used with a high
slag content, for example 30% Portland cement and 70% slag.
However, a cement labelled: CEM II/B-V 32,5R would have a fly ash content of
26% 35% and such cements are sold as having equivalent sulfate resistance to
that of sulfate-resisting Portland cement, if the cement has been shown to meet
additional criteria relating specifically to sulfate resistance. These additional
criteria are not specified in any European standard but in individual national
standards and the cement is labelled with an additional suffix, eg: +SR.
However, these bureaucratic niceties are not what we are focussing on here; the
main point is that composite cements containing slag or fly ash can offer
equivalent, or even better, performance for sulfate-resistance compared with
sulfate-resisting Portland cement.
If you drive a car, it helps if you know what the gearbox and clutch do in order to
drive it well. You dont have to be able to strip down the engine, but you do need
to have some minds eye image of each of the main parts of the car informing
you what they do and, roughly, how they work in relation to the car controls.
Have you tried to teach an aged relative how to use a computer? They might be
an absolute whizz, of course, but quite often older people have difficulties with
computers because they lack an understanding of them at a conceptual level. You
cant use Microsoft Word by trying to memorise a series of instructions; rote
learning for computers, or other complex systems, just doesnt work.
Similarly, rote learning isnt much good with cement. Most people interested in
cement are interested because they make it or they use it or they are
researching it; whatever your interest or level of understanding, it helps to have
a conceptual model of cement appropriate to the situation.
The model can be simple or complex, depending on what is most useful. A simple
model might be:
Cement is grey powder you use to make concrete that gives you nasty burns if
you get it on your skin.
This is a very useful and practical model that even the most distinguished
professor researching cement should have uppermost in his mind when
concreting his driveway. (If he does concrete his driveway, that is.)
However, having come as far as this, we might perhaps try to come up with
something a bit more illuminating. You should have the essential components of
a better model now, from the previous chapters. All you need to do is to practise
using your minds eye image.
Ive been using scanning electron microscopy and optical microscopy to examine
cement and concrete since 1981, during which time I must have looked at
thousands of samples of clinker, cement and concrete. My experience of
cementitious materials is therefore primarily visual and so Ive developed very
simple visual models of clinker, cement and concrete, based on standard
images. I use these models automatically, not just when examining material
I call them models - you could just as well call them techniques or methods. The
point is that they are structured ways of conceptualising cementitious material,
be they clinker or concrete or anything else. They help to understand how
cement made from a particular clinker might behave, and they help to
understand the hydration products that are likely to form from a Portland
cement, or from a composite cement.
There is one model for clinker and one for hydrating cement. Depending whether
you are a producer of cement or a user, one may be of more general use for you
than the other, but either way, have a go at using both.
Clinker composition: get a minds eye view of Portland cement clinker and
how differences in composition affect cement properties.
In other words, you use your minds eye image to view an archetype
that you understand, and then you alter it by gradually adding in new
conditions.
By changing the inputs, you can adjust the archetype in your mind using
thought experiments to allow for changes in compositions or other
circumstances so that your model is flexible and applicable to a wide range of
materials and situations.
Thought experiments are experiments you dont actually do, of course; instead
you explore ideas. Some thought experiments, like Schrodingers Cat, cant be
done in practice at all; those described here are far less obscure and could be
done in practice, but thought experiments do save a lot of time.
If you can use your minds eye image to work through each of the thought
experiments in the following two sections, you will have working conceptual
models that you can apply to cement clinker and to hydrating systems. You
might develop the models further over the years by refining them or adding new
pieces, and that would be very valuable, but the basic models, as described here,
should serve most people well for a long time.
Some of these thought experiments are easier than others. If you find any of
them difficult, dont worry. Come back to them again later and they should all
make sense eventually.
Focus on it, remember it, visualise it and get to know it so you can recognise the
clinker minerals. Once it is firmly fixed in your mind, you can start to play with it
and adjust it using thought experiments based on the clinker compositional
parameters for Lime Saturation Factor, Silica Ratio and Alumina Ratio.
Figure 11.1 Clinker archetype, SEM image Figure 11.2 Clinker archetype, same image
of polished section. Key: a-alite; b-belite; as in Figure 11.1 but without the distraction
c-aluminate (C3A); f-ferrite. Black areas of the labels.
are pores filled with epoxy resin.
Whats special about this clinker? you may ask. Well, the point about this
clinker is that there is nothing special about it. It is just a normal, well-produced
clinker although perhaps a little slow-cooled. It is a typical clinker and we'll
assume it could equally well be used to produce cement compliant with either
ASTM C 150 Type I, or EN-197 CEM I.
Look at the individual crystals of clinker minerals. Identify the alite crystals. Now
look at the belite, the aluminate and the ferrite. Figures 11.1 and 11.2 both show
our standard image; Figure 11.2 is the same image without the labels showing
examples of the four main clinker minerals. You dont want your standard clinker
image to be cluttered with labels - just use the image to get to know what each
of the clinker minerals looks like, then focus on the image without the labels.
To help with adjusting this archetypal image in your minds eye in the thought
experiments that follow, the clinker compositional parameters for Lime Saturation
Factor, Silica Ratio and Alumina Ratio, are shown below for convenience. These
formulae govern the proportions of the clinker minerals and are a big help in
visualising the clinker as its composition changes.
(There are other variations of this formula - use one of those if you are more
familiar with it.)
Suppose that this image is typical of the clinker produced by a cement plant on a
particular day and that the Lime Saturation Factor (LSF) is 94%, the Silica Ratio
(SR) is 2.5 and the Alumina Ratio (AR) is 1.6. The exact numbers dont matter
too much, we are just going to look at what happens as we change them.
Imagine you alter the raw materials going into the kiln by increasing the
proportion of limestone (assume it is pure calcium carbonate).
The clinker CaO content will increase, and so the LSF will also clearly increase;
the other oxides will decrease. The SR and AR will remain the same.
Now imagine increase the LSF to something a little over 100% and visualise free
lime crystals forming - see Figure 3.13 for a picture of free lime in clinker. (Of
course, remember that in reality a little free lime will always occur below an LSF
of 100% due to clinker heterogeneities, especially coarse particles of limestone or
silica in the feed, or if the clinker is underburned, but dont worry about that
now).
Now picture the LSF slowly decreasing from over 100% to about 90%; first,
imagine the free lime disappearing and then the proportion of alite diminishing
and the proportion of belite increasing. This clinker will be much easier to burn,
but cement made from it will have lower early strengths than cement made from
the standard clinker (again, all other things being equal). Later strengths should
largely catch up, though it might take a year or so. The cement will have a lower
heat of hydration.
Now imagine you increase the Silica Ratio from 2.5 to 2.7 by increasing the
proportion of silica in the raw feed. As the SR increases, the LSF decreases and
the AR stays the same.
same, however, the LSF is reduced, so picture a litle more belite and less alite.
The increased SR means that there will be less liquid, making the clinker harder
to burn; at the same time, though, the LSF is lower, making it easier to burn so
these effects will tend to cancel each other. Overall the clinker burnability may be
broadly similar to our standard clinker. All other things being equal, cement made
from this clinker would have lower early strengths than cement made from the
standard clinker because it contains less alite. However, the later strengths
should be at least as good and possibly better as the total silicate content is
slightly higher.
Now think of reducing the silica in the raw feed and the SR decreasing from 2.7
back through 2.5 down to 2.4. The LSF increases to about 100% and the AR
stays the same. Alite increases markedly and belite becomes scarce, as there is
now sufficient lime to form alite from all the available silica.
There is a little more ferrite and aluminate, in the same ratio as previously, and
fractionally less calcium silicates in total. All other things being equal, cement
made from this clinker would have higher early strengths than cement made
from the standard clinker because it contains more alite. The later strengths may
be fractionally lower as the total calcium silicate content is lower, although the
difference will be small.
Now imagine increasing the Alumina Ratio to 2.0. You could do this by increasing
the Al2O3 content or by decreasing the Fe2O3. For this thought experiment, just
add some alumina and decrease the SiO2, CaO and Fe2O3 proportionally, don't
think about what raw materials you change to do this it might get complicated.
The LSF and SR will both decrease.
The increased AR will make this clinker harder to burn (maximum liquid at the
lowest temperature is at an approximate AR of 1.38) but this will be offset by the
lower LSF and SR, both of which will make it easier to burn. Which effect would
win out is unclear, so lets suppose burnability will be similar to the standard
clinker. Cement made from this clinker will have a lower alite content and lower
silicates in total, so early and later strengths would probably be lower compared
with the cement made from the reference clinker (again, all other things being
equal). Although there will be less alite, there will be more aluminate and, since
aluminate is the most reactive of the four main clinker minerals, the heat of
hydration is likely to be higher compared with cement made from the reference
clinker.
Starting with the standard clinker, try thinking of decreasing the AR to 1.4 but
this time by adding iron oxide. LSF and SR will decrease, so the total silicates will
decrease and there will be proportionally less alite and more belite.
Compared with the standard clinker, this clinker will be easier to burn as there
will be maximum liquid at minimum temperature and the SR will be lower, so
there will be more liquid overall. Also, the LSF will be lower, making combination
easier.
There will be less alite and less aluminate, so the heat of hydration will be lower.
Early strengths of cement made from this clinker will be a little lower; later
strengths may also be a bit lower as the total proportion of silicates will be less
and so there will be less hydraulic material present in the cement.
Note that if you lowered the AR by reducing the Al2O3 content, the SR and LSF
would increase.
Our archetypal image in Figure 11.1 doesnt show any free lime, although there
was some elsewhere in the nodule, say about 1.5% in the clinker overall.
Suppose we alter the burning conditions, while keeping the kiln feed composition
constant.
Suppose we initially had 56% alite and 18% belite (using the Bogue calculation),
and then we burned the clinker harder. The free lime content will decrease and
we will have more alite and less belite.
If that gave us 0.5% free lime instead of the original 1.5%, we might expect to
have around 60% alite and 15% belite. More alite is good, so we would have
better cement strengths, wouldnt we?
Yes, in principle, we want to make alite. However, there comes a point where
trying to get every bit of alite that you can becomes counter-productive. As we
said in Chapter 3, overburning produces less-reactive alite.
Suppose we now burn less hard and the free lime rises from 0.5% to maybe
2.5% or 3%. The alite should become more reactive and give better cement
strengths but there will be less of it; at some point, the optimum trade-off
between alite content and alite reactivity will be passed and cement strengths will
start to decrease again. If we allowed the free lime to rise further, we would
eventually start to have problems with expansion due to the hydration of free
lime within hardened concrete or mortar.
Where that optimum free lime content lies is likely to depend on a range of
factors, including the clinker LSF, feed combinability and the related issue of raw
material fineness (especially fineness of silica particles).
The optimum may also depend on how the cement is to be used. In a composite
cement, the presence of free lime, when hydrated, provides a ready source of
immediately-available CH to produce additional C-S-H or other hydration
products as the slag or fly ash, or other mineral addition, reacts. Of course, the
bulk of the CH will still be provided by alite hydration and the clinker LSF will
have a stronger effect in determining the available CH. The practical effects of
clinker LSF and free lime variability on paste microstructure in composite
cements are unclear and could make an interesting research project.
Softer burning has other benefits as well, saving fuel and prolonging the life of
the kiln lining.
If the visualised images get a bit hazy and you arent quite sure what should be
happening, go back and look at Chapter 5 on clinker, especially the plots of
combinability curves (Figures 5.1-5.5). All should become clear.
One main reason why all things might not be equal in the real world is that raw
materials are not pure oxides; it is difficult to add just one oxide, as most raw
materials contain more than one, and this would cause unintended secondary
effects. For example, if you added clay or fly ash to raise the Alumina Ratio, you
would also be adding more silica.
Your freedom to compensate for this would depend what other raw materials you
were using; if you were also using milled silica sand, you could reduce the
proportion of sand to compensate for the additional silica in the fly ash. Even
then, other minor oxides in the fly ash may also have an effect; if the fly ash
contained magnesium, the temperature at which the melt formed on heating
might change. However, practical difficulties like this can be ignored in thought
experiments if appropriate - that is one of their great advantages. If you later
wanted to do one of the experiments for real, you would first identify these
secondary effects and apply the conceptual model to them also to predict the
results.
In other words, think what you want to achieve first, then work out how close
you can get to it in reality. Thought experiments can help with both.
Finally, if you have a cement made on Thursday that seems different from
cement made on Monday, but seems similar in compositional terms, the chapter
on cement variability looks at other reasons why this could happen.
With clinker, the formulae for LSF, SR and AR help to visualise how clinker will
change as its composition changes. Unfortunately, as yet we have no such helpful
formulae for hydrating systems, so we adopt as our archetype the process of
hydration of the simplest hydrated Portland cement system, containing clinker
and interground gypsum only.
2. Aluminate and ferrite hydration release alumina (also lime and a little
iron); the alumina is of prime importance as it is required to form
ettringite and AFm phases.
3. The ratio of available alumina to sulfate in the cement paste controls the
relative proportions of monosulfate (AFm) phase and ettringite, but see
also Principle 4.
It also helps to keep in mind the compositions of the reacting materials and of
the principal hydration products:
and
_
Hemicarbonate: C4AC0.5H12
To start with, think of a Portland cement that is composed only of ground clinker
and gypsum added at the cement mill. It has no interground limestone, or
grinding aid that affects cement hydration. It contains no carbonate and no
carbonate is allowed to penetrate from outside. (In Europe, this almost has to be
a thought experiment because most CEM I contains interground limestone,
unless you use white cement.) The image in Figure 11.3 below, reproduced from
Figure 6.3 for convenience, is a good archetype and we'll use it for the thought
experiments that follow.
In most of what follows, we will be thinking of the cement paste fraction of the
concrete (or mortar etc.). Concrete is not just cement paste with stones in it,
because the presence of aggregate introduces other important physical factors
such as the water demand of the aggregate and the interface between the
aggregate and the paste. However, we want to concentrate on the hydration
process and this is most simply done by thinking of pastes - just cement and
water.
You may first want to review the cement hydration process outlined in Chapter
6.3.4.
The next thought experiment is to visualise some cement pastes with different
water/cement ratios. It should be really easy, because there are some pictures
showing exactly this in Chapter 6; for convenience, here they are again (Figures
11.4-11.7).
Figure 11.4 Polished section of cement paste, Figure 11.5 Archetypal image: polished
age 2 years, made using ordinary Portland section of cement paste, age 2 years, made
cement, w/c=0.33. using ordinary Portland cement, w/c=0.40.
Figure 11.6 Polished section of cement paste, Figure 11.7 Polished section of cement paste,
age 2 years, made using ordinary Portland age 2 years, made using ordinary Portland
cement, w/c=0.50. cement, w/c=0.60.
In this particular experiment, dont think about individual hydration products, just
concentrate on three components: the cement hydration products as a whole, the
pores and the remaining cement particles that have not hydrated. Implicitly,
there is also a fourth component, water, filling the pores.
Think firstly of mixing 1 kg cement with 400 g water to give a water/cement ratio
of 0.4, then curing the paste under water for a year or two, until it is mature and
little further hydration will occur.
By 28 days, perhaps about 80%-90% of the alite has hydrated, together with
about half of the belite, almost all of the aluminate and perhaps half the ferrite.
Some of what remains unhydrated will probably never hydrate, although
hydration will continue at a slow, and ever slower, rate almost indefinitely. After a
year or two, hydration will have mostly ceased.
The archetypal image in Figure 11.3 (and in Figure 11.5) is our archetype; it
shows paste made with a water/cement ratio of 0.4 after 2 years. It is composed
largely of cement hydration products, with a little unhydrated calcium silicate
(top left) and some ferrite (bright, in larger relict cement grains). There are some
pores visible; these are capillary pores and would have been water-filled. In the
polished section, water in capillary pores has been replaced by epoxy resin.
Next, visualise how a mature paste made with w/c=0.5 might appear. (Don't
cheat and look at the images, just think about it!) There will be more water and
less cement in the paste, so we would expect a higher porosity and fewer
remaining unhydrated cement grains.
Now have a look at Figure 11.6, which shows exactly this; the paste
microstructure is considerably more porous than the paste made with w/c=0.4 in
Figure 11.5. There is still some residual ferrite; some of this may have hydrated,
we can't really tell from the images, but what is there now will probably persist
almost indefinitely.
While thinking again of the paste made with w/c=0.4, now picture a mature
paste made with w/c=0.6. There will be much less cement and much more fluid-
filled pore space; the hydration products will be more porous, with larger
capillary pores; again, this is just what Figure 11.7 shows. There is a clear trend
in the characteristics of the images in terms of porosity and residual unhydrated
cement as we go from w/c=0.4 through w/c=0.5 to wc=0.6.
Finally, consider a paste made with w/c approximately 0.3. The trend continues
in the opposite direction; we would expect a lot of unhydrated cement and little
porosity. According to the Powers-Brownyard model, not all the cement can
hydrate because there is not enough volume to accommodate all of the hydration
product. That is apparent from Figure 11.4; very little porosity is visible but
unhydrated cement grains are numerous. By extension, since there is little
capillary porosity visible, the available water is also very limited. Assuming
hydration product to occupy 2.2 times the volume of the unhydrated cement, it is
clear that most of the remaining cement cannot hydrate; there is not enough
space for the hydration product, or sufficient available water.
This expands Principle 3 above. Although they represent only a small proportion
of the total paste, ettringite and AFm phases are important; they affect concrete
setting properties and can have an effect on long-term durability.
The balance between available alumina and sulfate in the cement paste controls
the relative proportions of AFm phases and ettringite. In this thought experiment,
think of three cements, all made from the same clinker as used in our cement
archetype above, but containing different amounts of gypsum added when the
cement was milled.
Well call these low sulfate, medium sulfate and high sulfate mixes; the
actual numbers arent important here but think of them as, say, 1.5%, 3.0% and
4.5% SO3 if you like. The cements are used to make pastes, all at the same
water/cement ratio and cured until they are mature, say for a year or more.
Assume there is sufficient sulfate to prevent a flash set at the lower sulfate
contents and that any false-set that might occur is mixed through at the higher
sulfate contents. Dont worry about the exact numbers shown, they are only
intended to be illustrative, and just follow the arguments for each paste.
Medium sulfate mix: suppose the cement in the archetype mix above contained
a medium level of sulfate. In our mature paste, hydration has virtually
completed and the principal hydration product is C-S-H. Some CH is present, and
there is sufficient available sulfate to combine with all the available alumina to
form monosulfate phase. A small surplus of available sulfate produces a little
ettringite from some of the monosulfate. This is our reference paste to compare
with the following pastes as we change the amount of sulfate.
This paste contains: C-S-H; CH; AFm as monosulfate and a small amount of
ettringite.
Low sulfate mix: in this paste, the available alumina is about the same as in the
medium sulfate mix; there is less available sulfate, all of which is present in AFm
phase as monosulfate, but there is still an excess of available alumina.
Consequently, there is no ettringite at all, because this requires a higher ratio of
sulfate to alumina. The small deficit of sulfate is accommodated by the formation
of hydroxy-AFm phase, in which sulfate is partly replaced by hydroxide.
High sulfate mix: compared with our medium sulfate reference paste, this
paste with a higher sulfate content has more available sulfate and there is now
sufficient to convert almost all the available alumina to ettringite, with a little
residual monosulfate.
At higher still levels of sulfate, all the available alumina will have formed
ettringite; there will be some residual gypsum as no other sulfate-containing
phases can form.
This experiment has considered what might be present in mature pastes because
it is easier to do this when the reactions have largely stopped. At early ages, the
reactions are more complicated and there will be much local variability. Cement
particle size and crystal microstructure, to name just two factors, affect the rate
at which reactants become available. Also, not all the sulfate in the cement is
ultimately present in ettringite or monosulfate; some is taken up by the C-S-H,
as is some alumina. The use of the term available is used to allow for these
uncertainties.
Imagine mixing cement containing fine limestone with water. At the instant of
mixing, before the cement and water react, think of the fine limestone dispersed
between the coarser cement grains. The fine limestone occupies space that would
otherwise be occupied by water. The mix will be more cohesive, reducing
bleeding. The cement may also be ground coarser, giving better control over
early hydration, strength growth and heat evolution and will also save energy.
The first effects of limestone intergrinding are therefore physical and are in some
ways beneficial. The water demand might increase, depending on the particle size
of cement and limestone; this would reduce strength and so not be beneficial.
Imagine the fine particles of limestone being engulfed by C-S-H and ettringite
precipitating around them. This is Principle 5, the fine filler effect. The fine
limestone will act as nucleation sites for the formation of hydration products; this
will tend to reduce setting times and increase early strengths by increasing the
rate at which the cement hydrates. As noted in Principle 5 above, it may also aid
later strengths.
Where monosulfate has formed, carbonate from the limestone displaces the
sulfate. The sulfate forms ettringite from existing monosulfate and the carbonate
forms AFm phase as hemicarbonate. Eventually, all the monosulfate disappears,
replaced by hemicarbonate and ettringite.
The hemicarbonate may then react with more carbonate, if available, to produce
monocarbonate. (Monocarbonate contains twice as much carbonate compared
with hemicarbonate, so hemicarbonate will tend to form if carbonate is scarce in
relation to the total amount of AFm phase, with monocarbonate forming as more
carbonate becomes available).
Think of our archetypal Portland cement paste use the image in Figure 6.3,
reproduced yet again for convenience in Figure 11.8 below. Then imagine adding
to it a small amount of microsilica, say 3% by weight of cement, and mixing to a
paste with w/c=0.5. How will the microsilica affect the properties of the paste?
We have added very fine and reactive material composed largely of silica;
assuming the microsilica to be properly dispersed, the fineness should have the
physical effect of improving the cohesiveness of the paste. In concrete, this
should reduce bleeding.
Principles 5 and 6 will apply; think of the numerous small particles acting as
nucleation sites for the formation of hydration products. This should improve
early concrete strengths and probably also reduce setting times. As the
microsilica itself then starts to react, it will contribute silica to the cement
hydration products, forming additional C-S-H from the CH in the paste. Look
particularly at the CH in the paste without microsilica in Figure 11.8, and then
visualise the extent of CH in the paste diminishing as the C-S-H increases with
microsilica addition.
The Si/Ca ratio of the C-S-H will increase from approximately 0.5 to 0.6 or more.
Blastfurnace slag is a by-product of iron smelting. When the molten slag is cooled
rapidly using water from about 1500 C to 800 C, it forms a granular, latent
hydraulic, material composed of about 95%, or more, glass. The three principal
oxides are those of calcium, silicon and aluminium. Compared with Portland
cement, slag used in concrete contains less CaO and more SiO2 and Al2O3. Slag is
not as reactive as Portland cement and ideally is ground finer.
The proportions of Portland cement and slag in a composite cement vary widely,
typically from 30% slag to 70% or more. Suppose we have a mix containing 50%
of our archetypal Portland cement and 50% ggbs at w/c=0.5.
When first mixed with water, the Portland cement fraction of the mix reacts more
quickly than the slag and so has a higher effective w/c ratio than if the mix were
100% Portland cement. The paste microstructure for the first few days or weeks
is therefore more porous compared with a 100% Portland cement mix. As the
slag hydrates, the porosity is reduced and eventually the paste should become
less porous compared with a 100% Portland cement mix.
The finest of the slag particles probably provide nucleation sites on which
hydration products form, especially if the slag has been ground finer than the
clinker1. Principle 4 will therefore apply to some extent. (If the Portland cement
also contained interground limestone, the limestone may act to accelerate both
cement and slag hydration; however, our archetypal Portland cement contains no
limestone so we wont consider this further).
The main effect of the slag is on the proportions of the hydration products and
their composition, and on the paste microstructure.
In our mix of Portland cement and slag, at an early age (eg: one day) the
Portland cement fraction has hydrated more than the slag component, and the
hydration products are broadly those of the Portland cement archetype.
Gradually, the slag reacts.
1
Slag is harder to grind than clinker, so intergrinding slag and clinker produces cement in which the
slag is coarser than the clinker. Preferably, the slag should be finer than the clinker, so ideally the two
are milled separately and then blended.
Think of the composition of the slag; compared with Portland cement, the slag
has a higher ratio of Si/Ca and also of Al/Ca (compare the typical analyses in
Tables 3.2 and 7.1). The main effect on the hydrating system will be to increase
the ratio of total available Si/Ca compared with that from the Portland cement
alone and Principle 6 will apply. The proportion of CH in the paste will decrease,
the proportion of C-S-H will increase and the Si/Ca ratio of the C-S-H will
increase.
Principle 3 will also apply. The slag contributes more alumina than sulfate, so the
ratio of alumina to sulfate in the paste will increase. Any ettringite that had
formed earlier from Portland cement hydration will be lost and virtually all of the
available sulfate will be present in the form of monosulfate. Hydroxy-AFm will
form if insufficient sulfate is present to combine with the available alumina to
produce monosulfate.
Comparing the Portland cement paste in Figure 11.8 with the image in Figure
11.9 containing Portland cement and slag, there are two main differences visible
in the hydration products: in the paste containing slag, there is more C-S-H in
relation to CH, in fact, there is very little CH visible in Figure 11.9; most has been
converted to C-S-H.
The second difference is that, in Figure 11.8, there isn't any AFm phase visible.
Some will be present, but the crystals are too small and dispersed to be
apparent. In contrast, in Figure 11.9, platy crystals of AFm are clearly visible and
widespread. This is because alumina from slag hydration has produced additional
AFm. In the absence of any available carbonate, the AFm phases will probably be
a mixture of monosulfate and hydroxy-AFm.
Most sulfur in slag is present as sulfide, not as sulfate, although a small amount
of sulfate may be present. Sulfide released by slag hydration probably becomes
incorporated into AFm phase. Other minor hydrated phases will be present in
composite cements containing slag, in particular, a phase similar to hydrotalcite.
In summary, compared with a paste made using Portland cement alone, a paste
made from a composite cement containing slag would be expected to contain
more C-S-H, less CH, more monosulfate and less ettringite probably no
ettringite at all if there is sufficient slag. The increased proportion of C-S-H may
improve strength (depending on slag content, fineness and reactivity) and reduce
the paste permeability. The infilling of pore space by the platy monosulfate
crystals may add to strength and will certainly contribute to lowering permeability
(Figure 11.9).
Think of the spherical shape of fly ash particles aiding concrete workability by
acting as small ball bearings in the mix. This means that lower water/cement
ratios may be used in mixes containing fly ash compared with mixes containing
only Portland cement, or Portland cement and other angular particles such as
slag.
The finest fly ash particles are less than a micron across - imagine them also
contributing to some extent to providing nucleation sites for the formation of
hydration products (Principal 5).
Fly ash reaction also releases alumina, altering the balance between available
sulfate and alumina. This results in an increase in the proportion of monosulfate
and, probably, the disappearance of any ettringite that had formed initially before
the fly ash reacted significantly. As with the slag mix above, hydroxy-AFm will
form if insufficient sulfate is present to combine with all the available alumina to
produce monosulfate.
Fly ash is less reactive than either Portland cement or slag and so early strengths
will be lower. However, later strengths will be enhanced by the formation of
additional C-S-H, reducing porosity and permeability. Strengths will also be
enhanced by the reduction in water requirement due to the lubricating action of
the fly ash spheres. Comparing the strengths of fly ash and Portland cement
mixes, the relative strengths will depend on curing temperature.
If the mixes are cured at ambient temperature (nominally 20 C), and assuming
the continued availability of water for curing, after several months the strength of
the fly ash mix may approach that of the Portland cement mix, although it may
not quite catch up.
(NB: in two otherwise similar large pours of a Portland cement mix and a
Portland cement/fly ash composite mix, the Portland cement mix will reach a
higher temperature due to higher heat of hydration. This is likely to have an
adverse effect on strength and is another reason why the fly ash mix will be
stronger).
After a month or so, mixes containing fly ash may well show better strengths
than otherwise similar mixes containing Portland cement only. AFm formation in
pores will further reduce permeability and will tend to bind ingressing chloride.
As with the slag mix in Thought Experiment 9, if the Portland cement contained
interground limestone, carbonate would be available for AFm formation.
With cement or cement clinker, we can apply the Bogue calculation, or modified
versions of it, to get a good idea of the proportions of the phases present. With
hydrated systems it is more difficult, mainly because we don't know how much of
the mix constituents have actually reacted. For example, looking at Figure 11.9,
there was originally 50% slag and 50% Portland cement. Unless we know how
much slag and cement have actually reacted, we don't know what is available to
form the hydration products.
Simple. We could build structures that were better, stronger and more durable.
Probably cheaper, too. We could maximise strengths and produce just enough
pore filling phases, particularly AFm, to minimise permeability. (If we have more
AFm than we need, lime in AFm might be better used producing C-S-H and
additional strength; this we could do by supplying more reactive silica and less
alumina.)
In all three papers, the authors anticipate the prospect of quantifying paste
mineralogy from the bulk composition and in these publications they have made
a major contribution towards achieving that objective. These papers are pretty
technical but highly recommended if you wish to read more about cement
hydration products.
Portland cement production leaves large holes in the ground where raw materials
were extracted, it uses fossil fuels and it emits a lot of carbon dioxide. It may
also produce a lot of dust unless constant efforts are made to minimise it.
Traditionally, the cement manufacturers green credentials have been regarded by
some as a bit dubious, to put it mildly, but in recent years cement producers
have cleaned up their act a lot.
Holes in the ground, especially near cities, have become valuable for use as
landfill sites to dispose of rubbish or for redevelopment. Old quarries, for years
fenced off with Danger, Keep Out signs, have been redeveloped as massive
shopping centres, now with much smarter signs encouraging people to enter.
Many old quarries have been landscaped or returned to nature and in many
parts of the world, dust emissions from cement works are a small fraction of
what they were thirty or forty years ago.
Significant though these achievements are, the problems they addressed are
more straightforward than those of fossil fuel use and CO2 production. Even in
these areas, though, many cement producers have made major progress.
Improvements in plant efficiency, mainly by phasing out old wet process kilns
in favour of dry process kilns.
Using alternative fuels, such as waste solvents, waste paper, old car tyres and
bonemeal.
Together, these changes mean that less fuel per tonne of cement is used overall,
and a smaller proportion of the fuel that is used is fossil fuel.
The use of alternative fuels to partly replace fossil fuels means that a scarce
resource is conserved and also that nuisance wastes are disposed of productively.
Some hazardous wastes can be disposed of more effectively by burning in a
cement kiln than by other means, due to the high burning temperature and
longer residence time compared with an incinerator.
At least one UK cement works now produces clinker using 70% alternative fuels
and only 30% coal. This significant achievement is all the more impressive
considering that some of the non-combustible residues in these fuels affect
clinker composition. Car tyres contain steel, so the iron content of the clinker will
change unless the raw feed composition is adjusted; they also contain zinc, a
retarder of cement hydration. Bonemeal contains calcium phosphate which is
likely to inhibit alite formation unless well-dispersed throughout the clinker;
bonemeal is therefore only used in limited quantities in order control the clinker
phosphate content.
The bulk of the other 43% of the total CO2 will be due to burning fuel in the kiln,
with the remainder due to CO2 from electricity generation used in grinding the
raw materials, milling the clinker and transporting raw materials and the finished
product.
The bulk of the CO2 emitted by the burning of alternative fuels (paper, tyres,
bonemeal) is recycled CO2 that was only recently taken from the atmosphere;
unlike the burning of fossil fuels, continued burning of these fuels should not
produce a cumulative increase in atmospheric CO2.
Alternative fuels can therefore make a significant reduction in the net CO2
emissions related to cement production. However, the bulk of the CO2 emitted is
still due to limestone calcination. Since this is an integral part of the process of
Portland cement production, it seems there are only three ways in which this can
be significantly reduced:
Capturing CO2 from cement plants is also being examined but will be expensive
and seems unlikely to happen to any great extent in the short term.
Using less Portland cement is a very viable option and is already being done. For
example, for a 50/50 mix of Portland cement and slag compared with 100%
Portland cement, CO2 emissions are reduced considerably. Using a figure of 52 kg
CO2 per tonne slag (Table 7.1), instead of 900 kg CO2 emitted per tonne of
Portland cement, we have 450+26=476 kg CO2 per tonne cement.
If we factor in that perhaps only one-third of the fuel used to make the Portland
cement need be fossil fuel, the total CO2 produced from limestone calcination and
the burning of fossil fuel could be as low as about 350 kg tonne-1 cement. This is
only about 40% of the initial 900 kg tonne-1 cement - quite a saving.
Of course, this assumes that there is sufficient slag (or fly ash) available, without
the need for transporting it over long distances, and that, from a global
perspective CO2 reductions are accounted for consistently. The steel and
electricity industries may want to include these savings in calculations for their
own activities and the savings cant be counted twice.
References, Chapter 12
1. "The Cement Industry and Global Climate Change: Current and Potential
Future Cement Industry CO2 Emissions". M Natesan; S Smith, K Humphreys,
(2003). Greenhouse Gas Control Technologies - 6th International Conference.
Oxford: Pergamon. pp. 9951000. ISBN 9780080442761.
2. http://www.cementindustry.co.uk/PDF/MPAC_Performance_2008.pdf
Having now got to grips with some of the basics, I very much hope that you will
continue to explore cement science with some of the excellent textbooks
available.
If you have access to a technical library, the following books will doubtless be
available. However, if your work involves cement in any significant aspect, you
should probably get your own copy of some of the more important books anyway.
If you are an engineer rather than a chemist, the first of the books on the short
list below would be an excellent purchase, especially if you can wait for the new
edition to come out. However, if you are involved in cement or concrete
production from a technical or scientific perspective, or if you are an academic
with an interest in either cement or concrete, you really need the first three
books: Bye, Taylor and Lea. These three are the essentials of any cement
reference library.
This is a superb little book, the small size of which belies the huge quantity of
concentrated information that it contains. A third edition is in preparation.
This book is focussed more closely on concrete from an engineering viewpoint but
it contains some accessible cement and concrete chemistry. It is widely regarded
as a standard text on concrete.
The standard text for anyone interested in concrete petrography, this book
contains a mass of information on petrographic techniques and also some very
accessible cement and concrete chemistry.