You are on page 1of 13

31st Cement and Concrete Science Conference Paper Number 15

Novel Developments and Innovation in Cementitious Materials


12-13 September 2011
Imperial College London, United Kingdom

POTENTIAL IMPROVEMENTS IN CEMENT SUSTAINABILITY


E.M. Gartner
Lafarge Central Research

ABSTRACT

Although conventional Portland clinker (PC) based concretes have relatively low CO2 emissions compared to
most conceivable substitutes, the enormous volumes of concrete produced worldwide have nevertheless led
to public concerns about cement industry-generated CO2. Thus, most cement makers are currently trying to
replace PC to the greatest extent possible with supplementary cementitious materials (SCMs) such as coal-
derived fly ashes and granulated blast-furnace slags, which are considered to have low embodied energies
and low manufacturing CO2 emissions by virtue of the fact that they are industrial by-products. However, their
global supplies are limited; their quality tends to be very variable; and primary manufacture of equivalent
SCMs would, of course, be far more energy- and CO2-intensive. Thus, there is also an increased interest in
better use of natural SCMs, such as natural pozzolans and limestones. The low reactivity of common
pozzolanic minerals may to some extent be compensated by activation with concentrated basic alkali metal
solutions, as in “alkali-activated” binders. But low-CO2 clinkers, based on belite, calcium sulfoaluminate and
calcium aluminoferrite, also show promise, as they can be made with good quality control in conventional
cement kilns, which is advantageous in terms of capital investment. In the longer term, binders based on raw
materials with no fossil CO2 content, such as magnesium silicates, might become a promising alternative
approach as it can, in theory, produce truly “carbon neutral” binders.

1. INTRODUCTION materials except uncured softwood, stone, air-dried


clay or rammed-earth blocks. Recent UK data for
Concrete is by far the most widely used of man- specific embodied energy and specific embodied
made materials. It is difficult to obtain accurate carbon expressed as CO2 are shown in table 2.
figures for concrete per se, but, based on (Hammond and Jones, 2011)
reasonably reliable official statistics for cement
manufacture, it is estimated that current global Table 2. UK estimates of embodied primary energy and specific
CO2 emissions of construction materials. (Hammond and
concrete production must be of the order of 10
Jones)
cubic kilometers per year, (taking a broad view of
Primary Carbon,
what constitutes concrete). Recent approximate Density Carbon, as
Material Energy, as CO2
annual consumption estimates for the major t/m
3
kg.CO2/m3
classes of industrial materials are given in table 1. GJ/t mass %
Aggregate 2.5 0.083 0.5 11
Table 1. Estimated worldwide annual materials consumption
Rammed earth 1.5 0.45 2.3 34
Material Current Annual Consumption
(approx.) Common bricks 1.7 3.0 24 400
CEMENT > 3000 Mt Cement mortar 2.3 1.33 21 480
CONCRETE ≈ 10 km3 Structl. concrete
2.4 1.11 16 380
GYPSUM ≈ 250 Mt (of which more than 50 Mt are (excluding steel)
PRODUCTS from industrial by-products, e.g. FGD)
Ordinary concrete 1.45 0.67 7 105
LIME ≈ 130 Mt
PRODUCTS Aerated 0.75 3.50 30 225
> 3000 Mt (≈ 4 km3), of which < 1000 Mt
TIMBER
is for construction. (Most is for fuel!) Limestone block 2.18 0.85 5 110
≈ 1400 Mt (≈ 600 Mt is for construction,
Marble 2.50 2.00 12 290
STEEL of which ≈ 500 Mt is concrete
reinforcement) Plasterboard 0.8 6.75 38 300
ALUMINIUM ≈ 40 Mt Gypsum plaster 1.12 1.80 12 135
FIRED CLAY ≈ 4000 Mt (≈ 2 km3) - (mainly bricks) Clay tile 1.90 6.50 45 850
COAL ASHES ≈ 1000 Mt Timber 0.5-0.7 10.0 72 350 - 520
BF SLAG ≈ 120 Mt Hardboard 0.6-1.0 16.0 105 630 - 1050
POLYMERS ≈ 250 Mt Glass 2.50 15.0 85 2100
GLASS ≈ 100 Mt Steel (average,
7.80 20.1 140 11000
including recycled)
Concrete is also clearly a low energy material, with Aluminium (avg) 2.70 155 820 22000
both “embodied energy” and “embodied carbon” PVC (general) 1.38 77.2 2800 39000
values generally lower than all other construction
1
approaches to producing binders with inherently
These data come close to comparing construction lower CO2 emissions than the current generation of
materials at their point of use, although clearly the normalized Portland-based cements (Gartner,
actual numbers, (which are reportedly “cradle-to- 2004). Here I have updated this approach and
gate” estimates and thus allow for average extend it somewhat further towards concrete. In
transportation distances), will differ widely from addition, I list what I see as the principal scientific
location to location, especially for materials such as and technical challenges related to achieving major
concrete, which have a relatively low inherent CO2 emissions reductions in concrete construction
primary energy density. In reality, it only makes over the next few decades. However, I will not deal
sense to do this type of detailed comparison when in any depth with alternative technological solutions
comparing several different approaches to that involve the sequestration or chemical
constructing a structure with given performance conversion of the CO2 emitted from cement plants,
specifications in a specific location. However, table as these approaches involve complex process-
2 does clearly show that concrete is close to the related issues that clearly require a separate
bottom of the list in terms of embodied energy per treatment.
unit mass and also almost as low on a volume
basis, especially if one considers that timber is not Note also that the term “cost” as used throughout
often used as a volume-filling material. this paper is intended to incorporate the concepts
of energy and CO2 cost as well as financial cost.
The above global production statistics and
embodied energy and carbon estimates, 1.1 CO2 emissions associated with concrete
approximate though they may be, imply that the
modern world cannot exist without concrete, and The CO2 emissions associated with the
that it is extremely unlikely that any alternative manufacture of hardened concrete objects or
construction material will ever replace concrete in structures can be broken down into the following
most of its current applications. In fact, historically, three contributions:
concrete has supplanted other more ancient
construction materials, such as fired clay bricks, A: The CO2 emissions associated with the raw
because it is both more energy efficient and more materials (cement, aggregates, water, etc.)
economical to use in most applications. The basic B: The CO2 emissions associated with the
raw materials required for making conventional production of the energy used in concrete mixing.
Portland Clinker (PC) based cements, i.e. C: The CO2 emissions associated with the delivery
limestone, silica clay minerals and coal, as well as of both the raw materials and the fresh concrete
aggregates and water of the quality needed to and the placing and finishing of the concrete.
make ordinary concrete, are globally available in
vast quantities. Thus, one can consider concrete to In this paper, I will not attempt to deal with the
be highly “sustainable” in terms of raw materials logistics and processing issues involved, since they
supply. However, the enormous volumes of are very complex and clearly require a separate
concrete produced worldwide, coupled with a new detailed treatment. However, in most
appreciation of the role of CO2 in climate change circumstances, they are minor factors. I will thus
and ocean acidification, have recently led to public only deal with A: the raw material ingredients
concerns about total cement industry-generated required for making concrete. They are well known:
CO2 emissions, which are estimated to constitute
at least 5% of total anthropogenic CO2 emissions. 1. Aggregates (coarse and fine) – e.g.: gravel;
Note also that similar total figures and public crushed rock; sand; (and in some cases,
concerns also apply to steel, which is usually used lightweight aggregates)
synergistically with concrete in construction.
2. Binder (“hydraulic binder” or “cementitious”
Fortunately, there is still plenty of room for powders, plus any fine fillers that are not counted
improvement in the way we make cement and as aggregate)
concrete, and even apparently small improvements
in the technology can mean quite large energy 3. Water (sometimes partly substituted by air voids
savings and CO2 emissions reductions at the scale or bubbles on a volume basis)
of the planet. So the question that is currently of
most interest to the cement industry in this respect There are, in most cases, significant CO2
is: how can we continue to make high-quality emissions associated with the manufacture of all of
hydraulic binders, and use them to make cost- these ingredients except for water, although this
competitive concretes, while reducing our total may change as water resources become scarce
specific CO2 emissions to the minimal values and the need for water treatment more common.
theoretically feasible? The question of aggregate selection is another one
that I cannot deal with effectively in this paper,
Seven years ago I published a review of what I because it is usually a very local issue due to
referred to then as “industrially interesting” transportation costs. However, if one assumes that

2
there will continue to be a significant demand for (1) Glassy basic calcium aluminosilicates (typically,
volume-filling construction materials regardless of granulated blast-furnace slags).
other factors, then there is little choice but to use
local aggregates to fill as much of the volume as is (2) “Pozzolans” (reactive aluminosilicates or silicas,
practical, since they are almost always cheaper with significantly lower basicity than most slags).
than all other suitable materials.
(3) “Limestones” (rocks rich in calcium carbonate).
Thus, the only ingredient that remains to be
discussed is the binder, which is usually by far the The “pozzolan” category is very broad, since it
most energy-intensive of the raw materials. It is includes many natural rocks as well as man-made
also assumed to include any chemical additives products. Natural pozzolans are usually of volcanic
that are necessary for its correct functioning. origin (e.g. tuffs, pumices, etc.) whereas man-
made pozzolans are usually products of fossil fuel
1.2 Other general considerations combustion (e.g. fly ashes). Metallurgical slags are
produced by reaction of limestone or dolomite with
It is generally assumed that most construction aluminosilicate minerals as part of the metal
concretes will continue to be manufactured using extraction and purification process, and so are
mineral-based hydraulic binders, since these are usually far more basic (calcium- and magnesium-
generally the least expensive and the most rich) than the natural metal ore. The calcium and
“sustainable” in terms of natural resources. The magnesium oxides serve as fluxes in the process
major sustainability issue with conventional and help remove undesirable minor components
hydraulic binders for construction applications, such as sulfur and phosphorus. It is thus clear that
which are based principally on PC clinker, is production of metallurgical slags usually involves
evidently that of high CO2 emissions. These very high CO2 emissions (both from fuel use and
emissions are due to the high basic calcium from mineral decarbonation) and that the only
content of PC clinkers. This basic calcium is reason that such materials are currently considered
generally derived from calcination of limestone, to be “low CO2” is by convention – because they
resulting in the release of one mole of “fossil CO2” are considered to be “by-products” of metal
for each mole of CaO in the clinker. Current global production. But this is an arguable distinction. The
cement and concrete norms allow for the use of same is also true for fly ashes, in the sense that
various “supplementary cementitious materials” they are in most cases by-products of electricity
(SCMs) as clinker substitutes in cement and generation from coal. Production of SCMs as
concrete, and, where such materials can be primary products usually involves high energy and
considered to involve low manufacturing CO2 economic costs and high CO2 emissions, which is
emissions, their use can reduce the embodied CO2 why it is rarely done. Even the activation of clay
per unit of concrete produced. For this reason, minerals at the relatively low temperatures (600–
most cement and concrete makers currently aim to 800°C) required to make amorphous reactive
use the highest SCM substitution rates that are pozzolans, (e.g. metakaolin), involves significant
possible under existing local market conditions. But fuel use for calcination, as well as capital
everyone is now competing for the limited existing investment in the plant.
supplies of usable (acceptable quality) low-CO2,
low-cost industrial by-product SCMs. This subject It should also be noted that the definition of
was discussed in [2] and there have been few “chemical admixtures” or “chemical additives” (as
significant changes since then. Natural pozzolans opposed to “binders”) is not always very clear cut.
can also be used as SCMs, but they are not Basically, added chemicals should only be
available everywhere and are in general of poor considered to be “admixtures” if their use is
quality (low hydraulic reactivity and high water essentially optional, i.e. as a means of modifying
demand) compared to the best man-made SCMs fresh concrete properties to suit very specific
(e.g. ground granulated blast furnace slags or needs such as workability or setting time.
carefully selected coal fly ashes). Limestones can Dispersants (e.g. “superplasticizers”) are the most
also be used, but only to a very limited extent (as widely used organic concrete admixtures, and salt-
they only react with basic calcium aluminates). based hydration accelerators are the most widely-
Thus, it seems that there are few realistic used inorganic admixtures. Whenever the dosages
alternative binder types that can be envisaged as are high enough one also has to consider the CO2
replacements for the current normalized set of emissions involved in making the additive.
Portland clinker based binders, due to the lack of However, since such additives are almost all
global availability of the raw materials. essentially “specialty chemicals”, they are typically
at least an order of magnitude more expensive
It is necessary to give an explanation of some of than the binder, so there is always a very strong
the basic terminology used here. Conventional incentive to minimize their dosage. One important
SCMs for use with PC-based cements fall into case in which the distinction between chemical
three clearly different chemical classes: additive and binder component becomes blurred is
that of alkali-activated pozzolans (e.g. alkali-
3
activated fly ash binders), in which the alkali Sensitivity of concrete durability to curing
“activator” (generally an alkali hydroxide or alkali conditions is also critical, as not all binders hydrate
silicate) is typically used at a dosage of the order of in the same way. Those that hydrate rapidly are
10-20% by dry mass of binder. At such dosages, likely to be less sensitive to water loss by
the activator is in fact far more than simply a evaporation that can otherwise reduce the overall
catalyst; it is actually an active binder component degree of hydration, but on the other hand they
that is consumed in the hardening reaction, and may be more sensitive to thermal cracking due to
should be considered as such.(Duxson et al, 2007) highly exothermic early-age reactions. Poor
concrete surface quality and early-age cracking are
2. TECHNICAL AND SCIENTIFIC ISSUES often associated with poor curing, and surface
RELEVANT TO ALL NEW BINDER SYSTEMS cracking also opens up the concrete to other more
serious longer-term durability problems.
I outline below a range of generic technical and
scientific problems that tend to limit the extent to We also clearly need to be able to predict the
which we can currently consider replacing Portland- tendency towards long-term dimensional changes
based cements with alternative binders having that can cause cracking of the concrete or other
lower associated CO2 emissions: damage to the integrity of the whole structure. This
can refer either to excessive expansions, which are
2.1 The prediction of concrete durability usually due to chemical processes such as crystal
Even if we can estimate total porosity of the matrix growth or swelling of gels; or else to excessive
with reasonable precision based on a knowledge of shrinkage, which is almost always due to
the degree of hydration of the binder phases at any dehydration of (porous) hydrated phases,
given age and the densities of the hydrates formed, sometimes coupled with the effect of large thermal
it is currently very difficult, for novel binder systems, gradients. Prediction of these effects for novel
to make accurate theoretical predictions about the binders is still very difficult, since the theoretical
permeability and the diffusivity of potentially framework is so poorly developed. Prediction of
aggressive species in the matrix, and the resultant creep, which can mitigate cracking, should also be
effect on durability. included under this heading.

The ability of the matrix to protect conventional 2.2 The control of fresh concrete properties and
(inexpensive mild steel) reinforcing bars is early-age curing procedures
especially hard to predict. The current approach The relationships between mix design and the
(i.e. maintaining a pH above about 12 to keep the rheological properties of fresh concrete are
steel passive, and keeping chloride ion reasonably well understood for Portland cement-
concentrations at the steel interface below a pre- based binders, but far less so for alternative
set “critical level”) is very conservative and is also systems with low or zero Portland clinker contents.
clearly very unfavourable to all non-Portland or Rheological properties of concern include both the
diluted-Portland binder systems, because, due to “yield stress” and the “viscosity” as functions of
lack of free portlandite, they all have less ability to time, temperature, shear rate, etc. A simple slump
buffer the local pH at a high value for a long period test is not sufficient to characterize fresh concrete
of time under atmospheric carbonation. In reality, properties for many applications. Pure Portland
the risk of damage from steel corrosion also cements are usually fairly robust with respect to
depends on many other factors, e.g. temperature; small errors in water dosage, but the same may not
relative humidity; the strength, electrical be true for all conceivable alternative binder
conductivity and gas permeability of the matrix; and systems. Even for Portland cements, the response
the concentrations of other chemical species which to chemical admixtures, especially dispersants
may influence the steel corrosion rate. used to reduce water-demand, (e.g.
Unfortunately, we are not yet able to make good “superplasticizers”), is very sensitive to small
predictions of the resultant of all these effects, so differences in cement composition. Since new
we still usually resort to the safety net of high pH admixtures can (in principle) be developed and
and low chloride. This is a subject that will require a tested rapidly, this should not be a major barrier to
much larger R&D effort in the near future if any of the introduction of alternative binders; but it must at
the alternative binder systems discussed here are least be considered by the manufacturer.
to become widely used.
Setting times and rates of strength development at
Resistance to freeze-thaw scaling is also very early ages can be critical for many applications,
difficult to predict, especially for non-air-entrained and alternative binders can be very different from
concretes. Air-entrainment seems to work well with Portland cements in this respect. For example,
a variety of binders, provided that a good air void high-volume fly ash (HVFA) concretes usually set
distribution can be obtained. But more work will be and gain strength very slowly compared to OPCs.
needed to confirm that this approach is valid for all So the concrete maker should properly take into
of the alternative binders of potential interest. account the differences in strength development

4
rate with respect to finishing and/or demoulding However, the need for very tight fly ash quality
times and also the curing needs of the concrete. control in such situations is obvious, and more
Although it is not difficult to predict these effects research is needed, both into the means of
semi-empirically based on simple lab testing, more measuring the quality of cementitious materials
comprehensive concrete formulation models that and into the means of producing and supplying
would allow a concrete maker to use the right mix consistently high-quality cementitious products.
design and curing procedures for any specific
application could be very helpful in reducing the 2.4 Concrete life-cycle and recycling issues
risk of costly errors.
Recycling is another important aspect of the
2.3 Variability of raw materials concrete life cycle. In addition to the energy and
CO2 emissions that are associated with the binder,
The responsibility for dealing with raw materials we must not forget that good quality aggregates
variability may rest more or less with the concrete constitute a natural resource which is likely to
maker, depending on “local tradition”. In some become increasingly scarce in developed areas,
countries a wide range of cements is available in especially considering the costs of transportation.
which supplementary cementitious materials are Recycling of as much as possible of old concrete,
already included, and in this case the concrete as close as possible to its source, is evidently
maker is less likely to blend in supplementary desirable. Local legislation can encourage more
cementitious materials himself. In such situations, recycling of coarse aggregates recovered from
the responsibility for quality control of the binder crushed concretes, but the fine aggregates and the
clearly lies with the cement producer. However, in cement paste matrix are usually not recycled
other countries it is more common for the concrete except as low-value fillers because transportation
maker to blend together a range of cementitious costs are too high to allow their return to a cement
ingredients, in which case the concrete maker has plant where they could be turned back into cement.
a much greater responsibility for the quality of the However, this is usually mainly a logistics problem,
binder phase in his concretes. The latter case in so changing the binder is unlikely to make much
principle should allow a much greater flexibility with difference to the situation.
respect to the types of concrete that can be
produced locally, and should therefore result in a The issue of CO2 uptake by existing concrete
better overall economy of raw materials (both in structures is also pertinent. In principle, hydrated
economic terms and in environmental terms) if Portland cement is capable of re-absorbing almost
managed correctly, (although it also increases the all of the fossil CO2 that was emitted from the
capital costs of the concrete plant due to the need limestone decarbonated to produce it in the first
for extra storage silos). However, it also requires a place. However, this is a very slow diffusion-limited
much greater technological understanding on the process, typically taking hundreds to thousands of
part of the concrete maker, and this in turn requires years (Gajda and Miller, 2000). The actual rate of
a much higher degree of technical training than is carbonation of hardened concrete objects is
currently the case in most places. The input of the sensitive to temperature and humidity, and also
local academic community can clearly be helpful in depends greatly on the size, shape and location of
this respect. the concrete object in question, which makes it
relatively impractical to use as a means of
With regard to supplementary cementitious defraying responsibility for the CO2 emitted in the
materials, and especially inexpensive “low CO2” manufacturing process of either the cement or the
industrial by-products such as fly ashes, adequate concrete. Moreover, it is important for the durability
quality control is often lacking. Their compositions of steel-reinforced concretes that they not
can often vary considerably on a daily basis. carbonate too rapidly; so we have conflicting
Existing product standards are rather limited and interests here, depending on which performance
designed to work in situations where the degree of aspect we wish to promote in any given application.
substitution for Portland cement is quite low. But This leads me to the conclusion that it would be
when one talks about HVFA binders, or other advantageous (especially in terms of global CO2
binder systems requiring a high volume fraction of emissions) for the concrete industry to choose their
industrial by-products, (e.g. alkali-activated fly ash binders on the basis of the final application, i.e.
concretes), the need for careful quality control of whether or not the concrete will be reinforced with
the by-product becomes far more important. For steel, and, if it will be, then what type of exposure
simple alkali-activated fly ash binders the fly ash conditions it will be subject to. But this approach
constitutes the only solid binder phase, since the would require a much better level of understanding
other reactive component is simply a concentrated than is currently available of the performance of
solution of sodium hydroxide or sodium silicate. In alternative binders in concrete. This is, therefore,
this case, fly ash “quality” must be the main factor another major R&D challenge.
dictating binder performance, although the factors
that actually make one fly ash better than another
in such systems are not yet completely understood.

5
3. ALTERNATIVE BINDER SYSTEMS applications. The advantage of using Portland
clinker as the activator is that its hydration is
The following five binder systems will be discussed sufficiently rapid to give reasonable early concrete
in detail, based on a thorough review of published strengths, while the slower pozzolanic reaction can
information (Gartner and Macphee, 2011). For all then be used to consume the excess free lime
of these systems the raw materials are sufficiently produced. In many pozzolanic cements the ultimate
abundant for the binder to be of global interest as a C-S-H may have a lower C/S ratio than the average
Portland clinker substitute; but the practical value of about 1.7 typical of pure Portland cements,
performance and manufacturing costs are not yet which allows for even more C-S-H to form (i.e., the
known with much confidence. high C/S C-S-H formed initially can contribute
some of its own lime to a further pozzolanic
1. Pozzolans activated by lime and/or alkalis, e.g. reaction).
high-volume fly ash cements and “geopolymers.”
Even making the most conservative assumption
2. Reactive calcium silicates produced by “non- that the C-S-H maintains the same high C/S ratio
Portland” processes, e.g. “Celitement®” as in pure Portland cements, the overall CO2-
(KIT/Schwenk, Germany). efficiency of C-S-H production is greatly improved
by using pozzolans as additives to Portland
3. Systems based on MgO hydration and/or cements, because the pozzolans themselves
carbonation, e.g. binders proposed by Tec-Eco typically have very low associated CO2 emissions
(AU) and Novacem (UK). relative to PC. The practical value of this depends
on the widely-accepted assumption that C-S-H is
4. Systems based on precipitation of calcium the major binding phase in conventional concrete,
carbonates, e.g. binders recently proposed by and that CH (portlandite) is a far less effective
Calera Corp. (USA). binder. Thus, every mole of CH persisting in
concrete can be considered to represent 1 mole of
5. Belite-rich clinkers containing calcium sulfo- fossil CO2 emitted unnecessarily, which should
aluminate and ferrite, e.g. Lafarge’s “Aether®.” ideally be replaced by either C-S-H or C-A-H. This
argument has its limitations, however, because
3.1 Pozzolans activated by lime and/or alkalis some excess CH can be useful in certain cases
Pozzolans are reactive aluminosilicates, (AxS), or in (e.g., to ensure that the pH remains high enough to
some cases relatively pure silicas (x→0), which avoid steel corrosion in reinforced concrete). Apart
can be either glassy (such as natural volcanic from its lower pH buffering capacity, however, C-S-
glasses or man-made glasses such as fly ash or H is preferable to CH in terms of its contribution not
silica fume) or poorly-crystalline (such as only to strength but also to durability, because it
metakaolin). The best known and best understood has a much lower solubility and also a lower
class of pozzolan-based cements is that based on permeability due to finer pores.
lime activation, which includes activation by
Portland cement. This involves the following In addition to lime activation, certain pozzolans can
generalized reaction: also be activated by reaction with other bases. A
particularly relevant case is the reaction between
alumina-rich pozzolans (such as metakaolins or
AxS + mC + nH ⇒ C-S-H + C-A-H (1)
coal fly ashes) and basic alkali metal silicate
(pozzolan + lime + water ⇒ calcium silicate
solutions [RSz (aq.)], where z can also be zero (i.e.
hydrates + calcium aluminate hydrates)
for pure alkali hydroxide solutions):
The compositions of the C-S-H and C-A-H phases
ASy + RSz(aq) + nH ⇒ RAS(y+z)Hn (2)
(and also mixed C-A-S-H phases) formed vary with
the reaction conditions and also depend on the
This reaction is referred to by many people as
presence of other components such as sulfates
‘‘geopolymerization,’’ following the terminology of
and carbonates, which can give anion-substituted
Davidovits, (2009), but the product is no more or
AFm or AFt phases.
less a “polymer” than is C-S-H, and it is considered
by most authors to be a poorly crystalline zeolite
Eq. (1) is classically referred to as the ‘‘pozzolanic
precursor “gel” (Provis et al, 2009). Since, for cost
reaction,’’ and it consumes lime, pozzolans and
reasons, R is usually sodium oxide (N), such
water to make new hydrated phases that occupy
products are commonly referred to as “N-A-S-H
more volume that the original solid phases. The
gels.” Note that the main amorphous products
lime itself can also be obtained from an indirect
formed during the hydration of Portland-pozzolan
source, which is usually the hydration of the alite
cements are often referred to as “C-A-S-H gels,”
(and to a lesser extent, the belite) in a Portland
rather than C-S-H gels, because of their significant
cement. Traditional pozzolanic cements, as
alumina contents. The main difference between the
developed by the Romans, used lime as the
two gels is that C-A-S-H gels are based on a two-
activator, but this reaction is too slow at room
dimensional (lamellar) nanostructure built from
temperature to be of use for most modern concrete
6
calcium hydroxide sheets, (related to tobermorite or Moreover, the kinetics of geopolymerization are
jennite), whereas N-A-S-H has a nanostructure strongly dependent both on alkali concentration
related to zeolites, based on a three-dimensional and on the temperature, i.e. highly concentrated
aluminosilicate network. Thus, if one likes to think alkali solutions coupled with elevated curing
in terms of polymerization networks, C-A-S-H can temperatures are preferable for rapid hardening.
be considered to be essentially a 2-D calcium (Provis et al, 2009; Gartner and Macphee, 2011).
hydroxide polymer stabilized by silica and some This sensitivity to dilution with excess water makes
alumina, and N-A-S-H can be considered to be a 3- geopolymerization difficult to use in normal
D silica framework polymer modified by equivalent concrete applications, especially since concrete
substitutions of alumina and alkali cations. The aggregates are often wet. On the other hand, the
most important quantitative aspect of this structural inherently low bound water content of fully-reacted
difference lies in the base/acid ratio, (C+R)/(S+A), geopolymer pastes gives them excellent fire
which is much higher for C-A-S-H gels than for N- resistance and relatively low drying shrinkage.
A-S-H gels. This means that more basic Thus, “pure” geopolymers appear better suited for
components are required for C-A-S-H-based “low-temperature ceramics” applications than as an
binders than for N-A-S-H-based binders. Since the alternative to Portland clinker-based cements for
two basic components needed, i.e. either CaO or ordinary concrete construction applications.
Na2O, cannot easily be obtained without very
significant CO2 emissions, (of the order of 1000 However, in order to benefit from the relatively low
kg.CO2 per tonne of metal oxide in both cases), it basic oxide content of these alkali-activated
can be seen that binders based on N-A-S-H gels pozzolanic systems in terms of reduced net CO2
tend to have an inherent advantage as low-CO2 emissions, while nevertheless being able to
binders, provided that the reactive aluminosilicates produce binders suitable for use in ordinary
required come from low-energy/low-CO2 sources. concretes cured at ambient temperatures,
researchers are currently attempting to develop
Another important difference between “pozzolanic optimized combinations of alkali- and lime-
hydration” (reaction (1)) and “geopolymerization” activated binders (Shi et al, 2011). In this case the
(reaction (2)) lies in the role of water. Reaction (1) alkali source can vary more widely, because the
is a true hydration reaction, in which anhydrous reaction between certain alkali salts, such as the
oxide phases are converted to hydrated phases sulfate or carbonates, and lime (which may itself
incorporating significant amounts of chemically come either from PC or from a slag), can result in
bound water; whereas reaction (2) is essentially a the “in-situ” production of a concentrated alkali
coupled dissolution/condensation process in which metal hydroxide solution, (because calcium
the final product incorporates the alkali metal oxide sulfates and carbonates are usually relatively
into its structure but in principle contains very little insoluble under these conditions). Complex
chemically bound water – in fact, the theoretical formulations of pozzolans with slags, PC, alkali
value of n in equation (2) may well be zero. In salts and alkali hydroxides or silicates are currently
terms of kinetics there is also an important being investigated by many workers in an attempt
difference between the two reactions. In normal to develop novel binder systems that can be used
pozzolanic hydration the basic oxide, CaO, even if as substitutes for more conventional PC-pozzolan
present initially, converts almost immediately to its cements in concrete construction applications.
hydrated form, portlandite (Ca(OH)2), in contact Unfortunately, as the number of ingredients
with water, and so for most purposes one can increases, so does the complexity of the system.
consider the reaction to be between portlandite, Thus, quality control may well be the overriding
which is sparingly soluble in water, and the problem in practice. Moreover, due to the need for
pozzolan, which is extremely poorly soluble in significant amounts of basic calcium in order to
water. Thus, the concentrations of the principal obtain quick setting and hardening at ambient
reactants in the aqueous phase remain relatively temperatures, many of these complex binder
small but also relatively constant throughout the formulations may not end up being much better
reaction; and the reaction rate is mainly controlled than conventional PC-pozzolan cements in terms
by phase boundary reactions (dissolution of the of their overall CO2 emissions.
anhydrous phases and growth of hydrates) which
depend mainly on the surface areas of the solids 3.2 Reactive calcium silicates produced by
involved. This means that changes in water/binder “non-Portland” processes, e.g. “Celitement®”
ratio have little effect on the overall kinetics, and
mainly affect the properties of the final product as a It is generally accepted that C-S-H is the most
result of the changes in porosity resulting from the important binding phase in hydrated Portland
presence of excess water beyond that needed for cement pastes. Researchers at the Karlsrühe
the stoichiometry of the hydration reaction. On the Institute of Technology (KIT) have recently
other hand, geopolymerization involves the developed a new class of binder, called
presence of a concentrated alkali solution because “Celitement®,” that reportedly makes use of these
essentially all of the alkali metal required is initially binding properties in a more energy-efficient way,
present in solution as alkali hydroxide or silicate. avoiding the need to produce Portland clinkers, and

7
reportedly resulting in significantly lower CO2 are rather similar to gypsum plasters. The
emissions for equivalent concrete performance performance of Sorel cements is also strongly
(Stemmermann et al, 2010). Their approach is to dependent on the ratios of the main ingredients, as
produce a coating of a C-S-H precursor phase on well as on the reactivity of the MgO used (Li and
the surface of particles of a relatively unreactive but Chau, 2007). However, they generally set and
hard substrate such as quartz. The precursor harden very rapidly, can give considerably higher
phase is typically formed from α-C2SH, a calcium mechanical strengths than gypsum plasters, and
silicate hydrate produced by reaction of lime and also have good fire resistance. So, despite their
silica in an autoclave. Crystalline α-C2SH is not relatively high cost, they have some good specialty
hydraulically reactive, but it can reportedly be applications, mainly for use in dry environments.
activated by intergrinding it with a hard and Despite their high basicity, Sorel cements usually
relatively unreactive silica-rich filler such as quartz. carbonate only very slowly in use, probably due in
Apparently it becomes amorphous (and probably part to the formation of complex hydroxy-carbonate
also dehydrates to some extent) as it is ground, hydrates at the surface. It is also reported that the
and thus forms a reactive coating on the filler water-resistance of hardened Sorel cement
particles that can hydrate in water at ambient products can be improved significantly by treating
temperatures, leading to C-S-H-mediated bonding them with soluble phosphate salts (Kurdowski and
between the filler particles. Sorrentino, 1983).

The manufacture of Celitement thus typically Another important class of MgO cements is based
requires three fundamental steps: on magnesium phosphates (Wagh, 2004).
Magnesium forms fairly insoluble hexahydrated
(1) the calcination of limestone to make quicklime; binary orthophosphates with either ammonium or
potassium. The ammonium salt is called “struvite”
(2) reaction of lime, silica and water at about 200°C (MgNH4PO4.6H2O) and it has an isomorphous
and 15 bars pressure to produce α-C2SH; potassium-substituted equivalent. The hydration
reaction is usually accomplished by adding solid
(3) drying and intergrinding of α-C2SH with a hard MgO to a concentrated aqueous solution of the di-
siliceous filler such as quartz. hydrogen phosphate salt of the monovalent cation,
(e.g. a solution of KH2PO4). It is a rapid exothermic
It is claimed that the net CO2 emissions may be as acid-base reaction in which the MgO dissolves in
low as 50% of those of CEM I OPCs. This is mainly the acidic solution from which the hydrated double
because the binder contains a high content of a salt precipitates. This type of cement can set and
relatively inert filler such as quartz. It is also harden very quickly, and can also give very high
reported that certain Celitement compositions can strengths. Moreover, since struvite is an order of
give mortar compressive strengths of up to 80MPa magnitude less soluble than gypsum, magnesium
after 28 days of humid curing, and also can give a phosphate cements have fairly good water
low permeability matrix which might be helpful in resistance, especially under neutral or basic
retarding steel corrosion. On the other hand, the conditions.
high filler content of most Celitement binders
implies that they will probably require significant Both Sorel cements and magnesium phosphate
dosages of dispersants (superplasticizers) in order cements generally make use of MgO (periclase)
to achieve such performance. powders made by calcination of magnesite
(MgCO3). The reactivity of the MgO powder is an
3.3 Systems based on MgO hydration and/or important aspect of the quality control of such
carbonation cements, and is mainly controlled by the calcination
temperature used. Higher calcination temperatures
MgO-based cements are not new. Sorel cements, and longer times allow the periclase crystals to
invented over 140 years ago, (Sorel, 1867), are grow larger and thus lead to reduced reactivity.
based on the hydration of MgO in concentrated Magnesium phosphate cements usually require
solutions of either magnesium chloride or very unreactive “dead burned” MgO powders in
magnesium sulfate, forming respectively hydrated order to give acceptably long working times, and it
magnesium hydroxy-chlorides or hydroxy-sulfates is often also necessary to add set retarders such
(Kurdowski and Sorrentino, 1983). There are as soluble borates.
several well-defined binary hydrate phases, such
as 5Mg(OH)2.MgCl2.8H2O, that can occur in these Despite the fact that magnesium is one of the most
systems, but all of them have appreciable water common elements on the surface of the earth, the
solubilities, with a strong tendency to lose MgCl2 or two classes of MgO-based cements described
MgSO4 to the solution by leaching, leaving brucite above are restricted to specialty applications mainly
(Mg(OH)2) as the main residual product, because it because of the relatively high cost of pure MgO.
is very insoluble under neutral or basic conditions. This is due partly to the scarcity of the main raw
This makes such cements too water-sensitive for material, magnesite, which is far less common than
most outdoor applications, in which respect they limestone or dolomite. Note also that (a)
8
magnesium phosphate cements suffer from a more as “nesquehonite” (MgCO3.3H2O) (Harrison, 2007).
severe sustainability problem due to the global But it is already known that one can produce strong
scarcity of phosphate ores and their priority precast products by the carbonation of Portland
application in fertilizers, and (b) it is possible to cement concretes, and some conventional
make Sorel cements from “soft-burned dolime” concrete block makers already do this, so it is still
(partially-calcined dolomite), which is an intimate not clear why carbonated precast products made
mixture of CaCO3 and MgO and is less expensive from MgO that is itself made by calcining MgCO3
than pure MgO; but the performance of such should be any more sustainable than those made
cements is relatively poor due to their dilution with from lime or OPC.
calcite (Kurdowski and Sorrentino, 1983).
However, the above situation has changed
It is also well known that magnesium is abundant in significantly over the last three years, due to the
the oceans. Average seawater contains about 1290 research of Vlasopoulos and Cheeseman (2009) at
++
ppm of Mg , which is equivalent to about 2.2 g/l of Imperial College, London. They developed and
MgO, but it is present in an essentially fully- patented a new class of MgO-based hydraulic
neutralized form (the present-day average surface binders which also include additions of crystalline
pH being about 8.07 due to saturation with respect magnesium hydroxy-carbonate hydrates such as
to atmospheric CO2). Thus, MgO can only be hydromagnesite (Mg(OH)2.4MgCO3.4H2O). But,
separated from seawater by the addition of a more importantly, they also noted that it should in
strong base, e.g. lime (CaO or Ca(OH)2) or “hard- theory be possible to extract reactive MgO from
burned dolime” (fully-calcined dolomite, an intimate common magnesium silicate rocks with acceptably
mixture of CaO and MgO), to precipitate brucite, low energy requirements and CO2 emissions. This
which is then dried and calcined to give periclase. approach is the basis of a start-up company called
But this is, overall, a more energy- and CO2- “Novacem” which is currently conducting the R&D
intensive and also more expensive process than needed in order to determine how to make such
the calcination of magnesite. cements in industrial quantities. Evidently, much of
this work is still confidential, but the summary given
However, the average content of MgO in surface here is based mainly on a recent public
rocks is very high. Thus, over the last decade, presentation (Vlasopoulos, 2010).
several fresh attempts have been made to try and
develop more sustainable cement technologies As noted above, the important difference between
based on magnesium oxide. An Australian the Novacem approach and previous approaches
company called “Tec-Eco” has developed and is the idea that MgO might be extracted from
patented a series of binders incorporating various common basic magnesium silicate rocks, such as
amounts of reactive MgO produced by calcining peridotites or serpentinites, with CO2 emissions low
magnesite at low temperatures (Harrison, 2008). enough to render the overall process sustainable.
Harrison (2007) also claimed that the low- The advantages of using basic magnesium
temperature calcination of magnesite required less silicates as raw materials are twofold: firstly, they
energy and emitted less CO2 than the manufacture are very abundant on a global basis, despite the
of Portland cements, and that adding reactive MgO fact that they are not as well-distributed over the
to Portland-pozzolan cements could improve their Earth’s surface as limestones or clays; and,
performance and also increase their capacity to secondly, they contain essentially no fossil CO2.
absorb atmospheric CO2. Other researchers The Earth’s mantle is roughly 70% magnesium
(Vandeperre et al, 2008) have since investigated orthosilicate (Mg2SiO4), and this occurs as surface
the details of such systems but have not observed deposits in the form of olivine. Partial hydration of
any remarkable benefits due to the incorporation of such rocks near the Earth’s surface usually leads
MgO. In any case, Harrison’s proposition that MgO- to the formation of serpentine (Mg3Si2O5(OH)4).
based cements were inherently more sustainable Since neither of these minerals contains fossil CO2,
than CaO-based cements was not supported by the main industrial problem is how to efficiently
the evidence, because (a) MgO derived from extract reasonably pure MgO from them with
natural magnesite or from seawater releases more acceptably low costs and CO2 emissions.
fossil CO2 per unit mass than calcination of Novacem’s approach is to make use of
limestone to make lime (or OPC clinker), and must supercritical carbonation in a high-pressure reactor.
also accomplish this highly endothermic reaction by This decomposes magnesium silicate into
the combustion of fossil fuels; and, (b) brucite-rich magnesite and amorphous silica, which can, if
pastes carbonate even less rapidly than hydrated necessary, be separated. The magnesite can be
Portland cement pastes under most exposure calcined in the usual way (at low temperatures) to
conditions. The possibility certainly exists that one give reactive MgO; and the resulting CO2 gas can
could make prefabricated articles by deliberately be recaptured and returned to the pressure reactor,
carbonating wet MgO under an elevated partial so it essentially serves only as a “catalyst” in the
pressure of CO2, and that this might allow one to process.
develop strong microstructures based on
metastable hydrated magnesium carbonates such
9
A second important feature of Novacem’s approach chemically with the aluminates in cement to form
(Vlasopoulos and Cheeseman, 2009) is the calcium carboaluminate hydrates which are
deliberate inclusion of certain types of hydrated classed as “AFm” phases, and may also contain
3+ 3+
magnesium carbonate in the cement formulation some Fe ions substituting for Al (Lothenbach et
itself, in order to control the setting and hardening al, 2008). The more reactive alumina available, the
properties. If these magnesium carbonates are more carboaluminate hydrates can form, and some
obtained by consumption of CO2 from the of this reactive alumina can come from SCMs such
atmosphere or from other industrial processes as slags (Hoshino et al, 2006) or fly ashes (De
during the manufacture of the Novacem cement, Weerdt et al, 2011). However, the amount of
and are included in sufficient amounts in the calcium carbonate that can react in this way in
cement composition, the manufacture and use of conventional PC-based cements is generally less
such cements could in theory even be net CO2- than about 10%. So it is interesting to note that
negative. Thus, the Novacem approach does not “pure” calcium carbonate-based cements have also
specifically require that the binder absorb any been shown to be feasible on a laboratory scale.
atmospheric CO2 during its working life (e.g. in Such binders were initially developed primarily for
hardened concrete structures) in order for it to be potential use as bone-repair cements (Fontaine et
considered a “low-CO2” cement. However, any al, 2004; Combes et al, 2006). The approach used
such absorption would be an additional bonus. was to precipitate reactive amorphous calcium
carbonate (ACC) powders by mixing solutions of
Recent results suggest that Novacem binders are calcium chloride with solutions of sodium
capable of giving good-enough compressive bicarbonate, followed by rapid drying. In order to
strength development for a wide range of potential stabilize ACC, magnesium or strontium ions were
concrete applications. However, it is currently too included in the chloride solutions. It was shown that
early to determine whether or not the manufacture mixtures of ACC and vaterite powders would react
of Novacem will be industrially feasible at an in water to precipitate calcite or aragonite, and that
acceptable cost and whether or not an acceptable this reaction gave a hardened product. In one case,
net CO2 balance can be achieved in practice. a paste compressive strength of 13 MPa was
Moreover, even if the approach is demonstrated to obtained at a water/solid ratio of 0.4.
be CO2-neutral, there still remains much work to be
done to establish the quality and reliability of such The importance of the above work is that it shows
cements for major construction applications. One clearly that hydraulic binders can be made from
important issue will presumably be that of calcium carbonates without any decarbonation. It is
reinforced concretes, as magnesium hydroxides especially interesting to note that the “hydration”
are far less basic than calcium hydroxides, implying reaction essentially involves no binding of water,
that MgO-based cementitious matrixes will since both the initial binder phases (ACC and
probably only be capable of buffering the pH of vaterite) and the final products (aragonite or
their pore solutions at values close to 10, which is calcite) are ostensibly anhydrous forms of CaCO3.
too low for the passivation of mild steel. However, This appears to confirm the mechanism proposed
this problem is not unique to Novacem. Many of the by Shchukin and Amelina (1979) for the
other low-CaO binder systems discussed in this development of strength in crystalline binders, i.e.
paper are likely to have difficulty in maintaining a the formation of “bridging” contacts when crystals
high-enough pH for the passivation of steel over growing from a supersaturated solution come into
long periods of exposure to the atmosphere, contact with one another.
because of their tendency to carbonate rapidly. In
this respect, the high CaO content of Portland The work of Combes et al (2006) was aimed at
cements is still a major durability advantage in biological applications, rather than binders that
most structural applications unless the permeability might be more suitable for construction
of the paste matrix of low-CaO binders can be applications. However, the Calera Corporation of
made so low as to limit steel corrosion rates to California has recently proposed (Constantz, 2010)
acceptable levels by restricting transport of ions that carbonate-based binders or aggregates could
and/or oxygen. be produced in large quantities and at an
acceptable cost by the carbonation of sea water or
3.4 Systems based on precipitation of CaCO3 various brines in the presence of suitable sources
of alkalinity. They claim that alkalinity can be
Calcium carbonate, in the form of limestone, is one produced efficiently (in the form of NaOH) by a
of the most globally-abundant raw materials for the proprietary new low-voltage “electrodialysis”
manufacture of hydraulic binders. Clearly, it would process in which salt water can be separated into
be of great interest to be able to use it as a major two solutions, rich in NaOH and HCl, respectively,
binder ingredient without the need to decarbonate it without producing significant amounts of unwanted
first, thus avoiding the release of its fossil CO2 into gaseous products such as hydrogen or chlorine as
the atmosphere. Currently, limestone powders are occurs in the conventional “Chlor-Alkali” process
used in various classes of Portland cement widely used in the industrial production of NaOH
concretes, and it is known that calcite can react and chlorine. The excess HCl produced by this new
10
process could presumably be neutralized by give an AFm phase (monosulfoaluminate) plus
dissolving basic minerals, as previously suggested additional AH3:
by House et al (2007). This could constitute a novel
approach to the sustainable capture of C4A3$ + 18H ⇒ C4A$H12 + 2AH3 (4)
anthropogenic CO2 emissions and their conversion
into industrially useful products, although little However, after the first day, most of the additional
detailed technical information on such products or strength development is due to hydration of the
the manufacturing processes involved has yet been belite and ferrite phases. But their initial hydration
published. products are very different to those produced by
hydration of ostensibly the same phases in OPC.
3.5 Belite-rich cements containing CSA and The latest evidence (Wang et al, 2010) suggests
ferrite phases: Lafarge’s “Aether®” technology that belite initially hydrates mainly together with the
consumption of AH3, as shown below, to give
Many types of calcium sulphoaluminate-based crystalline strätlingite (C2ASH8) rather than C-S-H
cement have been developed over the past few as usually formed in OPCs.
decades for a wide range of possible applications,
especially in China, although relatively few are C2S + AH3 + 5H ⇒ C2ASH8 (5)
actually produced industrially (Sui and Yao, 2003).
There have also been many attempts to develop The ferrite phase also apparently hydrates with
reactive belite cements in order to reduce the consumption of some belite to produce katoites
temperature and energy costs of clinkering in (siliceous hydrogarnets) covered by the general
Portland cement kilns (Gies et al, 1986; Mielke et formula [C3(A,F)SxH6-2x] where the A/F ratio is
al, 1992; Chatterjee, 2003). One approach unknown but x is probably not greater than 1. Thus,
currently being investigated by Lafarge (Gartner by 14 days the main hydrates all appear to be
and Li, 2006) is to combine the above approaches crystalline. This implies that it is not necessary to
to make intermediate clinkers containing belite, rely on amorphous phases such as C-S-H or AH3
calcium sulpho-aluminate (ye’elimite) and calcium for strength generation in such cements. However,
alumino-ferrite as the three principal phases, (given C-S-H does begin to form later on, as belite
in decreasing order of content), and also including continues to hydrate after all of the AH3 has been
certain combinations of minor elements in order to consumed, at which point strätlingite also begins to
stabilize the belite in the more reactive ά form. be consumed by reaction with lime produced by
Such “BCSAF” clinkers can produce cements with belite hydration, giving more katoite (Wang, 2010;
similar performance to Portland cements (OPCs) Morin et al, 2011).
and can in principle be manufactured in
conventional Portland cement plants with the It is interesting to note (Quillin, 2007) that moist-
emission of 20-30% less CO2 than the equivalent cured concrete specimens made from BCSAF
OPCs (Li et al, 2007). More recently, Lafarge has cements in 2006 continued to increase in strength
adopted the name “Aether®” for cements produced steadily over at least several months, and did not
using this technology, which is now the subject of exhibit major bulk volume changes, despite the
an EC “Life+” development project (www.aether- major changes in the hydrate phase assemblage
cement.eu). that we may now assume to have taken place.

The hydration path of Aether cements is very 4. CONCLUDING REMARKS


different to that of OPCs. It is the hydration of
ye’elimite together with added calcium sulfate, In developing new hydraulic binders for large scale
(preferably as anhydrite) to give ettringite plus an use, a number of key objectives must be met:
amorphous aluminium hydroxide gel (AH3) that
accounts for most of the strength development (1) CO2 emissions and energy consumption from
during the first day (Li et al, 2007). This can be binder manufacture should be as low as possible.
represented by following net transformation: Of course, ordinary pollutant emissions (e.g. NOx,
SOx, Hg, etc.) must also be at least within
C4A3$ + 2C$ + 38H ⇒ C6A$3H32 + 2AH3 (3) regulatory limits defined for PC manufacture.

For this reaction it had previously been shown (2) The binder must be fit for the purposes for
(Sahu et al, 1991) that the rate of ettringite which it will be used (concretes, mortars, etc.)
formation is dependent on the rate of solubility of
the calcium sulfate phase used. It appears that the (3) Raw materials should be inexpensive and
use of anhydrite rather than gypsum is preferable readily available locally. Wastes and by-product
for long-term strength development. materials from other sectors should be used to the
extent that good quality control can be maintained
Once the readily soluble sulfate has been and that transportation costs remain acceptable.
exhausted, ye'elimite can continue to hydrate to

11
(4) Other processing costs also need to be Gajda, J., Miller, F. M., 2000. PCA R&D Bulletin
acceptable, such that the final cost of such Serial No. 2255, PCA, Skokie, IL, USA,
cements is low enough to be competitive as a Gartner, E., 2004. Industrially interesting
means of reducing cement-related industrial CO2 approaches to ‘‘low-CO2’’ cements. Cement and
emissions, as opposed, for example, to the Concrete Research, 34: 1489–1498
proposed separation of CO2 from cement plant Gartner, E. M., Li, G. S., 2006. High belite
exhaust gases followed by pressurization for sulfoaluminate clinker: fabrication process and
transport and underground storage. binder preparation, World Patent
WO2006018569.
It will clearly be difficult to change global cement Gartner, E., Macphee, D., 2011. A physico-
manufacturing so as to greatly reduce CO2 chemical basis for novel cementitious binders,
emissions in a short time frame. Even if suitable Cement and Concrete Research, 41: 736-749
raw materials and production processes are Gies, A. Knofel, D., Bujan, M., 1986. Influence of
available globally, transportation costs (and their alkalis on the composition of belite rich cement
associated CO2 emissions) may be high, which clinkers and the technological properties of the
suggests that no single solution to the problem is resulting cement, Cement and Concrete
likely, and thus that a wide variety of local solutions Research, 16: 411-422
are likely to co-exist. This will require a good supply Hammond, G., Jones, C., 2011. Inventory of
of well-trained technologists! Carbon & Energy (ICE) V2.0, University of Bath,
(available from www.bath.ac.uk/mech-
In looking at the known alternative systems, it eng/sert/embodied/)
appears that those closest to industrialization are Harrison, A. J. W. 2008. Reactive Magnesium
binders based mainly on pozzolans. This includes Oxide Cements, US Patent 7,347,896.
not only PC-activated pozzolans but also alkali- Harrison, J., 2007. Sustainability for the cement
activated pozzolans, or combinations of both. Next and concrete industry, Part 2, ZKG International,
in line are binders based on alternative clinkers rich Knowledge, 60(2): 1-6
in belite and calcium sulfoaluminate, and which can Hoshino, S., Yamada, K., Hirao, H., 2006.
be made in existing cement plants. Finally, in the XRD/Rietveld analysis of the hydration and
very long term, other very different binders, such as strength development of slag and limestone
binders based on magnesium silicate raw blended cement, Journal of Advanced Concrete
materials, appear to hold promise for extremely low Technology, 4: 357-367
CO2 emissions. House, K. Z., House, C. H., Schrag, D. P., 2007.
Electrochemical Acceleration of Chemical
References Weathering as an Energetically Feasible
Approach to Mitigating Anthropogenic Climate
Chatterjee, A. K., 2003. High belite Portland Change, Environmental Science and Technology,
cement – an update on development, 41: 8464-8470
characterization and applications, Proceedings of Kurdowski, W., Sorrentino, F., 1983. Special
the 11th International Congress on the Chemistry Cements, in "Structure and Performance of
of Cements, Durban, South Africa, 1: 31-40 Cements,” (ed. P. Barnes), Applied Science
Combes, C., Miao, B., Bareille, R., Rey, C., 2006. Publishers, UK.
Preparation, physical–chemical characterisation Li, G. S., Walenta, G., Gartner, E. M., 2007.
and cytocompatibility of calcium carbonate Formation and Hydration of Low-CO2 Cement
cements, Biomaterials, 27: 1945-1954 based on Belite, Calcium Sulfoaluminate and
Constantz, B., 2010. Calera - using CO2 to make Calcium Aluminoferrite, Proceedings of the 12th
useful materials, Carbon Capture Journal, 16: International Congress on the Chemistry of
23-25 Cements, Montreal, Canada.
Davidovits, J., 2009. Geopolymer chemistry and Li, Z., Chau, C. K., 2007. Influence of molar ratios
applications, Geopolymer Institute, St Quentin, on properties of magnesium oxychloride cement,
France, Cement and Concrete Research, 37: 866-870.
De Weerdt, K., Kjellsen, K. O., Sellevold, E., Lothenbach, B., Matschei, T., Möschner, G.,
Justnes, H., 2011. Synergy between fly ash and Glasser, F. P., 2008. Thermodynamic modelling
limestone powder in ternary cements, Cement of the effect of temperature on the hydration and
and Concrete Composites, 33: 30-38 porosity of Portland cement, Cement and
Duxson, P., Fernandez-Jimenez, A., Provis, J. L., Concrete Research, 38: 1-18.
Lukey, G. C., Palomo, A., Van Deventer, J. S. J., Mielke, I., Muller, A., Stark, J., 1992. Active belite
2007. Geopolymer technology: the current state cement, Proceedings of the 9th International
of the art, J. Materials Science, 20(42): 2917– Congress on the Chemistry of Cements, New
2933 Delhi, India, 2: 399
Fontaine, M. L., Combes, C., Mounic, S., Rey, C., Morin, V., Walenta, G., Gartner, E. M.,
2004. Hydraulic cements based on calcium Termkhajornkit, P., Baco, I., Casabonne, J. M.,
carbonates, French Patent FR2830249B1 2011. Hydration of a Belite-Calcium
Sulfoaluminate-Ferrite cement, Proceedings of
12
the 13th International Congress on the Chemistry
of Cements, Madrid, Spain.
Provis, J. L., Van Deventer, J. S. J., 2009.
Geopolymers: structure, processing, properties
and industrial applications, CRC Press, Boca
Raton, Florida, USA,
Sahu, S., Havlica, J., Tomkova, V., Majling, J.,
1991. Hydration behaviour of sulphoaluminate
belite cement in the presence of various calcium
sulfates, Thermochimica Acta, 175: 45-52
Shchukin, E. D., Amelina, E. A., 1979. Contact
Interactions in Disperse Systems, Advances in
Colloid and Interface Science, 11: 235-287
Shi, C., Fernandez-Jimenez, A., Palomo, A., 2011.
New cements for the 21st century: The pursuit of
an alternative to Portland cement, Cement and
Concrete Research, 41: 750-763
Sorel, S., Sur un nouveau ciment magnésien (On a
new magnesium cement), 1867. Comptes
Rendus Hebdomadaires des Séances de
l'Académie des Sciences, France, 65: 102-104.
Stemmermann, P., Schweicke, U., Garbev, K.,
Beuchle, G., and Möller, H., 2010. Celitement - a
sustainable prospect for the cement industry,
Cement International, 8: 52-66.
Sui, T., Yao, Y., 2003. Recent progress in special
cements in China, Proceedings of the 11th
International Congress on the Chemistry of
Cements, Durban, South Africa, 4: 2028-2032
Vandeperre, L. J., Liska, M., Al-Tabaa, A., 2008.
Hydration and Mechanical Properties of
Magnesia, Pulverized Fuel Ash, and Portland
Cement Blends, Journal of Materials in Civil
Engineering, 20: 375-383
Vlasopoulos, N., Cheeseman, C. R., 2009. World
Patent Application WO2009156740.
Vlasopoulos, N., 2010. Novacem Carbon Negative
Cement, Presentation at the Society of
Chemistry and Industry Conference on Low
Carbon Cements, London, UK.
Wagh, A. S., 2004. Chemically Bonded Phosphate
Ceramics, Elsevier, Oxford, UK.
Wang, J., Baco, I., Morin, V., Walenta, G., Damidot,
D., Gartner, E. M., 2010. Hydration mechanisms
of cement based on low CO2 clinkers containing
belite, ye’elimite and calcium alumino-ferrite,
Proceedings of the 7th International Symposium
on Cement & Concrete, Jinan, China.
Wang, J., 2010. Hydration mechanism of cements
based on low-CO2 clinkers containing belite,
ye’elimite and calcium alumino-ferrite, PhD
Thesis, University of Lille, France.
Quillin, K., 2007. Calcium Sulfoaluminate Cements
– CO2 reduction, concrete properties and
applications, Report BR496, BRE, Garston, UK.

13