J. Am. Ceram. Soc.

, 89 [12] 3701–3705 (2006)

DOI: 10.1111/j.1551-2916.2006.01279.x
r 2006 The American Ceramic Society

Journal
Morphology and Properties in Blended Cements with Ceramic Wastes
as a Pozzolanic Material
M. I. Sánchez de Rojas,*,w F. Marı́n,z J. Rivera,* and M. Frı́as*
*Eduardo Torroja Institute (CSIC), 28033 Madrid, Spain
z
División Tejas, Grupo Uralita, Spain

This paper describes the research run on ceramic materials, and discarded depends on the type of plant, product requirements,
more specifically fired clay roof tiles, ground to a fineness suitable and other technical considerations, but in any event amounts to
for use as an active replacement for portland cements (PCs). The a sizeable volume of waste. However, as these discards are made
utilization of waste products from calcined clay roof tiles as a from materials subject to quality standards of one kind or
pozzolanic material in the cement manufacture is studied. One another, their cost-efficient re-use constitutes a means of turning
of the first materials used as pozzolans in the history was thermal- them to some value.
treated clay, a material similar to the calcined clay precast During the manufacturing process, which involves dehydra-
elements. For this study, the materials selected are clay products tion, followed by firing at controlled temperatures ranging from
that have been eliminated for different reasons: a wrong tem- 7001 to 10001C, the clay minerals found in high proportions in
perature inside the kiln, dimensional and mechanical failures, and the natural materials used to make brick and similar products
durability loss. The results showed that these products have good acquire properties characteristic of ‘‘fired clay,’’ including poz-
pozzolanic properties; at an early age, the pozzolanic activity was zolanic properties. Historically, one of the first materials used as
found to be higher than fly ash, although it was lower than silica a pozzolan was thermally treated clay, a material similar to fired
fume. The morphology of clay tile–PC pastes are similar to the clay products.
morphology of the pastes containing other pozzolanic materials. Ceramic products such as tile and brick are made of natural
Hydrated calcium aluminate products are generated when the materials with a high clay mineral content. Clay is activated
preparation method yields less compact pastes with a greater firstly by dehydration, beginning at a temperature of about
voids’ volume to accommodate these products. The hydrated 5001C, followed by the separation of amorphous and highly
calcium aluminate phases were analyzed using scanning electro- active alumina, whose maximum concentration is reached at
nic microscopy and X-ray diffraction techniques. different temperatures depending on the type of mineral in-
volved, kaolinite- or montmorillonite (Mo)-type clay minerals
or a combination of the two acquire pozzolanic properties under
I. Introduction controlled calcination at temperatures from 5401 to 9801C.
Higher temperatures are required to activate clay with an illitic
A CLEAR correlation may be drawn among the richness of a
country and the generation of wastes. Developed countries
produce a large waste volume. The only positive fact is that their
base or clay shale containing large proportions of vermiculite,
chlorite, or mica. The constituent minerals of ceramic materials
own technological development offers several industrial tech- would be activated by the high firing temperatures used in their
nologies that allow the use of the wastes produced. This should manufacture, resulting in a product with pozzolanic properties.1
be the line to continue. In the industrial environment the Waste from clay brick plants, known as discards, is collected
problem of the generation of wastes is similar. It is necessary indiscriminately, irrespective of the reason for discarding the
to adapt and introduce those denominated ‘‘clean technologies’’ respective products. Discarding may be due to:
so that they minimize the generation of residuals and produce (1) Dimensional breakage and flaws that have no effect on
the same or improved products, with less raw materials and the intrinsic properties of the clay material.
minor consumption of energy. (2) Flaws, more typical of older kilns, caused by over- or
By-product deposit in controlled dumps is not the only underfiring that may affect the physical–chemical properties of
solution for its correct management. Moreover, its future use the product.
will be submitted to more strict restrictions, insufficient nowa- Earlier papers2 emphasized the pozzolanic properties of waste
days, which will lead to better economical solutions, such as clay discards. The kinetics of pozzolanic reaction were investigated
waste product reduction and recycling. following chemical methods and Langavant calorimetry in order
The utilization and re-evaluation of industrial by-products to establish the relationship between pozzolanic activity and heat
have to acquire greater priority in the future for different reasons development with time. The results obtained are compared with
such as energy cost, cost of raw materials and manual labor, those of other industrial by-products traditionally used in con-
necessity of controlled sinks, and mainly, environmental decay struction such as fly ash (FA) and silica fume (SF).
provoked by the progressive increase of all types of wastes. The impact of the temperature at which the waste is obtained
Over 26 million tons of ceramic materials (bricks, tiles, pan has likewise been studied. When the firing temperature is
forms, blocks, and so on) are manufactured every year in Spain. inappropriate (under- or overfired material), the chemical and
The percentage of product rejected for sale and therefore mineralogical composition of the waste varies significantly with
respect to the product obtained under optimum firing condi-
tions. The temperature used (around 9001C) is sufficient to
activate the clay and confers pozzolanic properties on the
discarded material. Note that hydrated products similar to those
C. Jantzen—contributing editor obtained with other pozzolanic materials are formed in the
respective pozzolanic reactions.3
The pozzolanic activity as a function of the specific surface
Manuscript No. 21281. Received December 22, 2005; approved July 25, 2006. area was previously studied using clay tiles and bricks discards.
w
Author to whom correspondence should be addressed. e-mail: [email protected] The results showed an increase of pozzolanic activity when the
3701
3702 Journal of the American Ceramic Society—Sánchez de Rojas et al. Vol. 89, No. 12

surface area increases, although the amount of lime fixed is

similar in the long term.3

Q, Mu
Specifically, the application targeted in the above studies was
the manufacture of concrete roof tiles. The authors showed that
waste clay brick can be used as a raw material in the manufac-

Q
ture of concrete roofing tiles, either as a replacement for cement,
exploiting its pozzolanic properties, or as part of the aggregate.4

Q
The morphology of cement pastes containing different poz-

Q, Mu
Q
zolanic replacements varies significantly with the preparation

Q, Mu
Mi

C
Mu
procedure used. Their evolution in cement pastes with replace-

C
C, Mu
Mu
ments depends on the type of replacement, the water/cement

Mu

Q
Mo
ratio, and the curing conditions.
The results indicate that the water/cement ratio is a determin-
ing factor in the formation of hydration products, as higher
water/cement ratios allow the formation of products such as 5 10 15 20 25 30 35 40 45 50 55 60
2 - Theta
hydrated calcium aluminates, which are not observed with lower
water/cement ratios or more compact pastes with smaller voids Fig. 1. 1: X-ray diffraction of clay tile.
content for hydration products. Hydrated calcium aluminate
formations were found in cement pastes and in pastes containing (2) Characterization of Materials Used as Pozzolans
FA made with a high water/cement ratio. Adding SF to the (A) Chemical Test: The activity of the pozzolanic mate-
paste yielded more compact and less porous pastes in which the rials was evaluated with a quick method based on lime uptake
formation of hydrated calcium aluminates was hindered regard- by the material when it was cured in a saturated lime solution.
less of the water/cement ratio used. The lime fixed by the sample (mol%) was found as the difference
The subject of the present paper is a study conducted with between the initial concentration of the saturated lime solution
industrial waste products. Despite improvement in manufactur- and the CaO present in the solution in contact with the sample
ing processes, industrial plants inevitably have to discard a at different reaction times.
certain amount of material regarded to be unfit for sale. The The results, shown in Fig. 2, indicate that the samples of
material chosen for this study was burned clay roofing tile, ground ceramic tile displayed acceptable pozzolanic activity,
which was crushed and ground to the necessary fineness. having been fixed at 68% after 7 days. The tile proved to be
more active than FA at early ages, although it was less than SF,
which continued to be the best-performing early age product,
II. Experimental Procedure fixing 80% of the total lime after 24 h. These results show that
the calcined temperature of these ceramic materials (by 9001C) is
(1) Materials enough to activate the clays, obtaining pozzolanic properties.
The materials used were cement and different pozzolanic (B) Mechanical Test: The mortar elaboration and dosage
materials. is performed according to EN 196-1: 1994 Standard.6 The mortar
(A) Reference Portland Cement (PC): The reference composition is 1/3 by mass (cement/sand) and the water/cement
cement used was a CEM I/42.5 type cement, according to EN ratio is 0.5. The cement-to-replacement proportions (by weight)
197-1: 2000 Standard,5 with a clinker content greater than or used were as follows: Cement/replacement: 100/0 and 85/15 (clay
equal to 95% and up to 5% in additional components. tile o45 mm).
(B) Clay Tile (T): The samples consisting of crushed The results of the compressive strength of the 4 cm

and ground ceramic tile, with a BET specific surface area of 4 cm
16 cm specimens at 1 day and 28 days are shown in
3 m2/g, show a chemical composition highly acidic with a Fig. 3. The compressive strengths at 1 day do not vary signifi-
predominance of silica, aluminum, and iron oxide. The chemical cantly. At 28 days of curing, the value obtained is slightly lower
composition of clay tile is similar to the composition of the other than the reference mortar, but these are less than the percentage
types of industrial waste with pozzolanic properties (Table I). Its of cement substitution, indicating that these materials act as a
mineralogical composition is studied by X-ray diffraction pozzolanic material contributing to the compressive strength.
(XRD). The main crystalline compounds observed are quartz,
muscovite (Mu), calcite (C), microcline (Mi), and Mo (Fig. 1). (3) Morphology of Clay Tile–PC Paste
Two industrial by-products were also selected, FA and SF, in
The cement-to-replacement proportions (by weight) of the
comparison with the clay tile. Theirs chemical compositions and
mixtures used were as follows: cement/replacement: 100/0, 80/20
specific surface areas are given in Table I.
(FA and tile), 90/10 (SF).

Table I. Chemical Composition and Specific Surface Area

100
(BET)
Materials 80
Fixed Lime (%)

Chemical constituent (%) Cement Fly ash Silica fume Clay tile 60

SiO2 19.60 42.44 92.22 67.03 40

Al2O3 4.41 28.00 0.10 19.95
Fe2O3 3.30 17.38 0.62 6.29 20
CaO 63.21 7.40 0.54 0.11
MgO 4.20 1.38 0.27 1.37 0
Na2O 0.30 0.21 0.01 0.21 1 7 28 90
K2O 0.50 1.55 0.60 3.54 Age (Days)
SO3 3.00 0.05 0.18 0
LOI 1.40 1.23 5.50 0.47 Silica Fume Fly Ash Clay Tile
BET (m2/g) 0.60 1.40 22.20 3.00
Fig. 2. Pozzolanic activity: Fixed lime.
December 2006 Morphology and Properties in Blended Cements 3703

50

Compressive Strength
40
1 Day
(N/mm2)
30
28 Days
20

10
28 Days
0
1 Day
100/0
85/15
Samples
Fig. 3. Average compressive strength of three prisms (4 cm
4 cm

16 cm).
Photo 2. Scanning electronic microscopy micrograph of portland
Two different methods were used to prepare the cement pastes cement paste (procedure 2).
Procedure 1: Cement pastes prepared with a W/C ratio of 0.4.
The mold in which the paste was prepared was stored at a The ceramic material, from burned clay tile (T), was logically
temperature of 201711C and a relative humidity of 95%. found to have angled edges, as this material had to be ground to
Procedure 2: Cement pastes prepared, according to EN 196-5: obtain the particle size required. SF and FA, by contrast, are
1994 Standard,7 mixing 2070.01 g of cement with 100 mL of powdery products directly activated by high-temperature thermal
distilled water, stored in a closed beaker in a thermostatic closet treatment and require no prior grinding. A wide variety of
at a uniform temperature of 40170.51C. particle sizes were found. Microanalysis of these particles showed
The test age, for both procedures, was set at 90 days, when that they consisted of silicon, aluminum, potassium, and iron,
the activity of the materials analyzed had stabilized after the along with traces of clay mineral elements. Silica-rich particles
significant variation in lime fixation at earlier ages. from the quartz present in the samples were likewise found.

(2) Studies Conducted on Cement Pastes

III. Results and Discussion
The reference material for the studies was a PC paste manufac-
(1) Studies Conducted on Materials Used as Active Additions tured under the same test conditions.
Scanning electronic microscopy (SEM) revealed that FA is SEM micrograph of the fracture surface at 90 days of PC
constituted for numerous cenospheres and irregularly shaped paste, after hydration under procedure 1, show poorly crystal-
particles. Most of these cenospheres had a smooth surface, line or amorphous calcium silicate hydrates (CSH) with a C/S
characteristic of FA with low lime content, a composite of ratio of 1.5, also known as plombierite.8 This CSH might be
silicoaluminates with a low calcium content. Other very porous classed as type I CSH formations.9 Other more compact,
particles, constituting unburned coal, were also observed. calcium-rich silicates with a C/S ratio of 2.1 also appeared
SF is made up of small spheres, varying in size from 1 to 50 adjacent to CH plates (Photo 1). Under the conditions described
mm. At higher magnification, these spheres are seen to consist of above for procedure 2, the cement paste had CSH filaments with
small particles, o1 mm across, indicative of a certain degree of a C/S ratio of 1.3, as well as portlandite and hydrated calcium
condensation, even though this product is marketed as undensi- aluminate (prismatic species) with a C/A ratio of 2 (Photo 2).
fied fume. The microanalysis of these particles shows that they In general, the FA–PC paste shows the surface of the ceno-
are made of silica. In addition to these sphere particles, other spheres with a thin CH film and CSH gel cover,10 similar to the
angular formations with a high Si content were also observed. double layers that envelope the inert component in cement paste
Such particles might well comprise the silicon carbide detected or concrete. This film eventually gives rise to strong, dense layers
by XRD. Silicon carbide may be responsible for the high loss on under the continuous deposition of CSH gel and calcium hydro-
ignition found for this product.

Photo 1. Scanning electronic microscopy micrograph of portland Photo 3. Scanning electronic microscopy micrograph of fly ash–port-
cement paste (procedure 1). land cement paste (procedure 2).
3704 Journal of the American Ceramic Society—Sánchez de Rojas et al. Vol. 89, No. 12

Photo 6. Scanning electronic microscopy micrograph of T-portland

cement paste (procedure 2). (A) Hydrated calcium aluminate. (B)
Photo 4. Scanning electronic microscopy micrograph of fly ash–port-
Hydrated calcium silicate CSH.
land cement paste (procedure 2).

xide. The latter is sometimes deposited in a particular stacked containing aluminum, iron, and potassium, with an indetermi-
plate morphology around the FA particles.11 This fact, in some nate morphology. Quartz particles from the initial ground tile
cases, produces voids where the cenospheres once were, which were also observed. Hydrated calcium silicates with C/S ratios of
are not totally filled with the pozzolanic reaction products.12 1.6–2 are also visible, along with aluminum-containing forma-
In the present study, the reaction between the FA and tions with Al/S/C ratios of 1.5/1/5, perhaps as a result of the
the calcium hydroxide in the cement was clearly visible after uptake of aluminum in the CSH gel. When procedure 2 was used
90 days. In procedure 1, the calcium/silica ratio of hydration to prepare the T–PC paste, the samples were very homogeneous
products in FA–PC pastes was found to be higher than in the FA in appearance but with different CSH morphologies. The
particles. In the reaction products, the C/S/A ratio was 4/2/1, smooth surface prismatic formations in Photo 6 correspond to
whereas in the cenospheres the ratio was 1/2/1. According to the hydrated calcium aluminates with a C/A ratio of 2 (A) and CSH
X-ray energy-dispersive spectrometric analyses conducted, the with a C/S ratio of 1.5 (B) and some aluminum in their
aluminum from the FA was taken up into the CSH gel. composition. CH crystals (A) and hydrated calcium silicates
However, no hydrated calcium aluminate was found. can be seen in Photo 7. No ettringite or monosulfate formations
In procedure 2 curing conditions, the FA–PC paste contained were observed in any of the clay tile paste samples, regardless of
laminar CH formations adjacent to fully attacked cenospheres the preparation procedure used.
covered with hydration products (Photo 3). The aluminum The different addition-PC pastes were also studied by using
in the FA was taken up in the CSH, which had a C/S/A ratio XRD technique (Figs. 4 and 5). Under the procedure 1 curing
of 1/1/0.4 and hydrated calcium aluminate was found (Photo 4). condition, the results show that no hydrated calcium aluminate
The SF–PC pastes were generally more compact and denser was found (Fig. 4), but when procedure 2 was used (Fig. 5),
than the samples with FA. After 90 days (with procedure 1 and hydrated calcium aluminates were found.
procedure 2), the morphology of the hydrated calcium silicates Therefore, a different morphology of pastes with tile is ob-
was particularly dense (Photo 5), a finding that concurs with tained according to the process of preparation. SEM reveals that
other authors’ observations.13,14 hydrated calcium aluminates are formed in pastes with a higher
The procedure 1, T–PC paste was less dense than the SF–PC water/cement ratio. This fact is also confirmed by XRD. Figure 4
paste. The different components are embedded in a uniform shows the absence of hydrated calcium aluminates in the pastes
mass, and also observed were unreacted tile particles, CH plates, obtained with procedure 1, whereas with procedure 2 (Fig. 5)
and hydrated calcium silicates, with C/S ratios of 1.4–2.3 hydrated calcium aluminate is formed (C4AH13–C4AcH11).

Photo 5. Scanning electronic microscopy micrograph of silica fume– Photo 7. Scanning electronic microscopy micrograph of T–PC paste
portland cement paste (procedure 2). (Procedure 2). (A) Portlandite.
December 2006 Morphology and Properties in Blended Cements 3705

The more compact and less porous structure of SF–PC pastes

hindered the formation of hydrated calcium aluminates in both
P
procedure 1 (lower water/cement ratio) and procedure 2 (higher
SF
water/cement ratio).
There are no morphological differences between cement
E HCA E
pastes made with clay tile and those made with other pozzolanic
T materials. Other elements, namely aluminum and iron, were
found in the CSH structure. Hydrated calcium aluminates
FA formed when procedure 2 (higher water/cement ratio) was
used, but not with procedure 1, providing further support for
PC the preceding conclusions.
The content of hydrated calcium aluminates and ettringite in
PC pastes without replacement and with FA is verified by XRD
(Fig. 5). In the SF–PC, the XRD reveals the content of small
5 10 15 20 25 30 35 40 45 50 55 60 quantities of hydrated calcium aluminates. This confirms that
2 - Theta aluminates are better identified by XRD when they are scarce or
Fig. 4. Procedure 1: X-ray diffraction patterns at 90 days of portland they have crystals of small size; when hydrated calcium alumi-
cement (PC) pastes containing fly ash (FA), clay tile (T), silica fume (SF), nates are in a larger quantity, such as in T–PC paste, the
and without replacement (PC). microscopy technique allows their identification.

P
Acknowledgments
These studies, conducted under a Research Project (AMB 96-1095) funded by
SF HCA the Interministerial Commission for Technology and Science (CICYT), also
E HCA E benefited from the co-operation of the Uralita Group Tile Division.

FA
References
PC 1
B S. Johansson and P. J. Andresen, ‘‘Pozzolanic Activity of Calcined Moler
Clay,’’ Cem. Concr. Res., 20, 447–52 (1990).
2
M. I. Sánchez de Rojas, M. Frı́as, J. Rivera, M. J. Escorihuela, and F. Marı́n,
‘‘Research About the Pozzolanic Activity of Waste Materials from Calcined
Clay,’’ Mater. Constr., 51 [261] 45–52 (2001).
5 10 15 20 25 30 35 40 45 50 55 60 3
M. I. Sánchez de Rojas, M. Frı́as, J. Rivera, and F. Marı́n, ‘‘Waste Products
2 - Theta from Prefabricated Ceramic Materials as Pozzolanic Addition’’; 11th International
Congress on the Chemistry of Cement Durban, South Africa, 11–6, 2003.
Fig. 5. Procedure 2: X-ray diffraction patterns at 90 days of portland 4
M. I. Sánchez de Rojas, F. Marı́n, M. Frı́as, and J. Rivera, ‘‘Viability of
cement (PC) pastes containing fly ash (FA), clay tile (T), silica fume (SF), Utilization of Waste Materials from Ceramic Products in Precast Concretes,’’
and without replacement (PC). Mater. Constr., 51 [263] 149–61 (2001).
5
. EN 197-1:2000 Standards. Cement. Part 1: Composition, Specifications and
Conformity Criteria for Common Cements.
IV. Conclusions 6
. EN 196-1:1994 Standards: Methods of testing cement. Part 1: Determination
of Strength.
The studies carried out indicate the feasibility of using waste 7
. EN 196-5:1994 Standards: Methods of testing cement. Part 5: Pozzolanicity
materials from calcined clay as pozzolans. These results indicate test for pozzolanic cement.
8
that the firing temperature of these materials (around 9001C) is E. Bonaccorsi, S. Merlino, and A. R. Kampf, ‘‘The Crystal Structure of Tober-
morite 14 Å (Plombierite), a CSH Phase,’’ J. Am. Ceram. Soc., 88 [3] 505–12
enough to activate the clays, giving them pozzolanic properties. (2005).
Crushed and ground-burned clay tile exhibit acceptable pozzo- 9
S. Diamond, ‘‘Cement Paste Microstructure: An Overview at Several Levels’’;
lanic activity, which is higher than in FA. pp. 2–30 in Proceedings of Conference on Hydraulic Cement pastes: Their Structure
The morphology of cement pastes containing different poz- and Properties. Cement and Concrete Association, Sheffield, 1976.
10
S. Diamond, D. Ravina, and J. Lovell, ‘‘The Ocurrence of Duplex Films on
zolanic additions varies significantly with the preparation pro- Flyash Surfaces,’’ Cem. Concr. Res., 10, 297–300 (1980).
cedure used. According to the results obtained in this study, the 11
Y. Halse, P. L. Pratt, J. A. Dalziel, and W. A. Gutteridge, ‘‘Development of
water/cement ratio is a determining factor in the formation of Microstructure and Other Properties in Flyash Opc Systems,’’ Cem. Concr. Res.,
hydration products: higher ratios accommodate the formation 14, 491–8 (1984).
12
M. Frı́as and M. I. Sánchez de Rojas, ‘‘Microstructural Alterations in Fly Ash
of substances such as hydrated calcium aluminates, which are Mortars: Study on Phenomena Affecting Particle and Pore Size,’’ Cem. Concr.
not found in pastes with lower water/cement ratios or more Res., 27, 619–28 (1997).
13
compact pastes with smaller void contents for hydration pro- M. Regourd, B. Mortureux, P. C. Aitcin, and P. Pinsonneault, ‘‘Microstruc-
duct formation. ture of Concretes Containing Silica Fume’’; pp. 249–60 in Proceedings of 4th
Conference Cement Microscopy Edited by J. Bayles. International Cement Micro-
Hydrated calcium aluminate formations were found in PC scopy Association, Las Vegas, 1982.
pastes with no additions and in pastes with FA when procedure 14
M. I. Sánchez de Rojas, J. Rivera, and M. Frı́as, ‘‘Influence of the Microsilica
2 was used. State on Pozzolanic Reaction Rate,’’ Cem. Concr. Res., 29, 945–9 (1999). &