Available online at www.sciencedirect.

com

Applied Clay Science 42 (2008) 125 – 129

www.elsevier.com/locate/clay

The use of Tunisian Barremian clay in the traditional ceramic industry:

Optimization of ceramic properties
Salah Mahmoudi a,b,⁎, Ezzeddine Srasra b , Fouad Zargouni a
a
Département de Géologie, Faculté des Sciences de Tunis, 1060, Belvédère, Tunis, Tunisia
b
Centre des matériaux, Technopole Borj Cedria, BP95, 2050, Hammam Lif, Tunisia
Received 12 July 2007; received in revised form 18 November 2007; accepted 15 December 2007
Available online 15 January 2008

Abstract

This paper shows the results of various analyses and ceramic aptitude tests made on two representative clay samples from Djebel Oust on
Northwest of Tunisia. Indeed, mineralogical, geochemical, physical and ceramic properties of crude samples are studied. The crystalline phases
presented in different clays are illite, kaolinite, illite–smectite mixed layers, chlorite, K-feldspar, quartz and calcite. This study reveals that an amount
of silica is relatively high. The content of CaO and the iron oxide should not exceed 6%. The grain size data indicates a clay-dominated assemblage.
The plasticity test shows a medium value, the firing shrinkage and the expansion are limited. These clays could be used in manufacturing of the
ceramic pieces.
© 2008 Elsevier B.V. All rights reserved.

Keywords: Clays; Mineralogy; Chemistry; Ceramic properties; Northwest Tunisia

1. Introduction The series of the lower-Cretaceous of Djebel Oust and namely

those of Barremian clays have been the focus of geologists for a
Clays and clay minerals are important in geology, agriculture, long time. They were the subject of various studies: age, depo-
construction, engineering, process industries and environmental sitional environment and tectonic settings. For example, Jauzein
application. For this reason, particular attention should be given (1962) studied the geological series of Northern Tunisia and in
to the exploitation of raw materials in Tunisia for industrial particular those of Djebel Oust. The geological map of Bir
application. The present paper aims at investigating a particular M'cherga to the 1/50,000 covers all series of Djebel Oust.
application of local clays collected from Djebel Oust to evaluate Maamouri et al. (1994) established a detailed zonation of the
their potential for traditional ceramic. series of the lower-Cretaceous by improving the former subdi-
The studies were carried out with two clay samples from visions. Mahmoudi (2004) studied the geological series of Djebel
Djebel Oust in Northwestern Tunisia (see the geological map of Oust and in particular the physico-chemical and geotechnical
Bir M'Cherga No. 28 at 1/50,000 scale), the most representative, characterization and application of Aptian–Albian clays.
taken from a clay-dominated formation that outcrops 35 km These different types of clay raw materials do not present
southwestern Tunisia and 20 km northwestern of Zaghouan. The great differences in mineralogical geochemical and physical
studied sector is composed of calcareous, extrusive, Jurassic, composition. Consequently, they show similar firing behaviour
anticline, surrounded by cretaceous formations. The argillaceous that are of decisive importance for the quality of the end pro-
formations are composed of grounds of Berrasian to Albian. The ducts (Fabri and Fiori, 1985; De Andres Gomes de Bared, 1990;
color of clay is green-grey becoming darker in depth. It contains Parras et al., 1996; Dondi et al., 1999; Gonzalez et al., 1999).
of the calcareous layer (1 to 2 m).
2. Materials and methods
⁎ Corresponding author. Département de Géologie, Faculté des Sciences de
Tunis, 1060, Belvédère, Tunis, Tunisia. Tel.: +216 97655581; fax: +216 71430934. Two representative clays used in the traditional ceramic have been con-
E-mail address: salahmahmoudii@yahoo.fr (S. Mahmoudi). sidered. The raw materials belong to Bir M'cherga from Barremian of Djebel

0169-1317/$ - see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.clay.2007.12.008
126 S. Mahmoudi et al. / Applied Clay Science 42 (2008) 125–129

Table 1 Table 2
The physico-chemical analyses of Barremian clays The geotechnical parameters of Barremian clays
Parameter Sample A Sample B Parameter (%) Sample A Sample B
Chemical SiO2 (%) 57.74 56.73 Liquid limit 36.1 34.6
compositions (%) Al2O3 (%) 17.01 15.68 Plastic limit 22.2 20.5
Fe2O3 (%) 5.62 4.71 Plastic index 13.9 14.1
CaO (%) 4.36 6.05 Fraction N 63 μm 6.0 0.8
MgO 0.83 0.48 Fraction 50–63 μm 3.5 1.1
Na2O 1.60 1.68 Fraction 50–20 μm 15.4 9.0
K2 O 3.10 3.04 Fraction 10–20 μm 13.5 18.0
TiO2 0.33 0.45 Fraction 10–2 μm 20.3 43.0
MnO 0.22 0.13 Fraction b 2 μm 41.1 27.8
L.I.O. 8.81 10.24
CaCO3 3.88 7.22
calculated by the equation: MF = 3FL / 2bh2, where: F = breaking load (kg), L =
b2 μm fraction (%) Illite 70 81
distance between supports (L = 29.67 mm), b = sample width (mm), h = sample
Kaolinite 15 10
thickness (mm).
Chlorite 09 05
The water absorption was determined in heated ceramic samples according to
Illite/smectite mixed layers 06 04
ISO-10543-3. The pieces were filled with water and heated until boiling. The
Whole sample (%) Clay minerals 60 50
ceramic bodies were introduced, separated from each other, after being weighed in
Quartz 35 32
water, so that they show a large free surface as possible. The ceramic bodies were
Calcite 03 13
kept in the boiling water for 2 h. When the water reached 35–40 °C, the ceramic
Feldspars 02 05
bodies were taken out, their surface was dried, and the bodies were weighed. The
Physical analyses Specific surface area (m2/g) 20 12
value of the water absorption (WA) is obtained as follows: WA = [(wf −wi)/wi] ⁎ 100.
Cation exchange capacity 14.5 16.5
(meq/100 of air-dried clay)
3. Results and discussion
Oust. Mineralogical analysis was performed by X-ray diffraction (XRD) (Philips
X' Pert Pro system, Cu Kα, 40 kV, 40 mA). For semi-quantitative analysis of the 3.1. Clay characterization
samples, the relative abundance of minerals was estimated from the heights of
the main reflections. The cation exchange capacity was measured with standard The two clays present the expected typical compositions
acetate ammonium procedure (Meir and Nuesch, 1999). The major element (Table 1), rich in silica, alumina and potash with minor contents
compositions were determined by atomic absorption. The grain size analysis
was carried out by sedimentation. The DTA/TG was obtained with a SETRAM
type 124. The expansion and firing shrinkage were measured using the dila-
tometer Adamel Lhomargy BI. The samples were heated from ambient tem-
perature to 900 °C with 10°/min. The liquid limit (LL), plastic limit (PL) and
plastic index (IP), IP = LL − PL) were determined in accordance with the French
Standard NF P 94-051.
To determine the linear firing shrinkage, the bending strength and the water
absorption, ten flat test pieces (100 mm × 50 mm × 10 mm, using 60 g of material
moistened up 7% water, per test piece) were prepared for from each mixture. The
compaction pressure was 250 bars. After compaction, the test pieces were oven-
dried at 105 °C until constant weight, fired at 1050 °C for 1 h.
The linear firing shrinkage was evaluated using the formula: [(lf −ld) /ld] × 100,
where ld and lf are the measured length of dried and fired samples (ISO 10545).
The geotechnical tests were performed on ten specimens of each sample.
The bending strength was measured with a three point flexural method ac-
cording to the norm ISO 10545-4. The average values of bending strength were

Fig. 1. Position of the studied clays on the Holtz and Kovacs diagram. Fig. 2. Granulometric analyses of samples A and B.
 S. Mahmoudi et al. / Applied Clay Science 42 (2008) 125–129 127

Fig. 3. Dilatometry curves of clay samples.

of magnesium, titan and carbonates, accompanied by a signi- is dominated by illite and quartz with lesser amounts of kao-
ficant amount of iron oxides, which is responsible for the dark linite, chlorite, illite/smectite mixed layers, feldspar and calcite.
coloring of the sintered bodies. The mineralogical composition The illite present favorable properties for ceramic use. The
content of quartz is tolerable since it can be easily digested by
vitreous flow during firing. The two samples contain a high
quantity of potash, feldspar is always low or very low, so that
potash is indicative of the amount of illite. The calcite content is
lower, the samples also present the lowest loss on heating,
which is basically related to the decomposition of calcite. The
cation exchange capacity and specific surface area are also
listed in Table 1. The low values indicate that the clays contain a
low amount of illite/smectite mixed layers, which is accordance
with mineralogical analysis.

Table 3
Composition of mixtures and corresponding measured values of flexural strength,
water absorption and the linear firing shrinkage
Mixtures Sample Sand (S) Linear firing Water Flexural strength
B (%) (%) shrinkage absorption (FS) (N/mm2)
(LF) (%) (WA) (%)
M1 100 00 1.95 8.34 17.54
M2 95 05 1.67 8.98 17.32
M3 90 10 1.62 10.33 15.69
M4 85 15 1.20 14.15 15.38
M5 80 20 0.94 16.42 15.02
M6 75 25 0.70 20.15 10.41
M7 70 30 0.57 23.47 7.02
M8 65 35 0.47 25.08 6.38
M9 60 40 0.39 25.34 6.42
Fig. 4. DTA and TG curves of samples A and B.
128 S. Mahmoudi et al. / Applied Clay Science 42 (2008) 125–129

3.2. Ceramic properties

3.2.1. Plasticity
Fig. 1 and Table 2 shows the position of these clays on the
Holtz and Kovacs Diagram (Holtz and Kovacs, 1981) the plas-
ticity index was near 14%. The position of these clays on the
Holtz and Kovacs Diagram, show that they are illitic clays. On
the other hand, the high content of quartz and the lack of
expandable clay minerals are the main factor for the low plas-
ticity. The specific surface area is fairly consistent with the
plasticity data, in particular with the plastic limit (22.2% for
sample A, 20.5% for sample B). These mechanical properties
agree for drying (negligible contraction).
Fig. 6. Variation of water absorption WA (%) with sand content.
3.2.2. Granulometric analysis
For ceramic products, particular attention should be given
to the finer fraction (lower than 2 μm). The fraction below 2 μm
is very high (41% for sample A, 28% for the sample B) (Table 2, At low temperature range of 117 to 120 °C an endothermic
Fig. 2). This particle size distribution is excellent for uses in reaction occurs due to the removal physically bound water
ceramics. The problem arising by the presence of particles N 50 μm (Fig. 4). As the temperature increases, an endothermic effect
can be solved by a simply crushing and sifting. Thus, due to the occurs due to dehydroxylation (530–540 °C) of the clay mineral
size distribution the raw materials are classified as silty clays and structure and formation of quasi-amorphous materials in both
be usable for ceramic bodies. samples the decarbonation reaction of calcite (716–770 °C) is
detected.
3.2.3. Thermal analysis
Transformation of α quartz into to β quartz between 500 and 3.2.4. Fired products and simultaneous optimization
600 °C of quartz appears in both samples. As typical of clays As sample B has a high percentage of illite, alumina, a large
(Santos, 1997) an expansion occurs at 700 °C, which is likely amount of the finer fraction and a low degree of plasticity and
due to microstructural rearrangement accruing during carbonate linear firing shrinkage, sample B was selected for the firing studies.
decomposition (Assal et al., 1999; Thomas and Peer, 2001). The The manufacture of ceramic products such as floor and wall
expansion values are 0.95% for sample A and 0.93% for sample tiles depends much on the raw material (Carretero et al., 2002;
B. The expansion is accompanied by a rapid shrinkage due to Correi et al., 2004). The water absorption, linear firing shrink-
vitrification. The presence of Fe2O3 and K2O accelerates the age and the flexural strength are essential to judge the quality of
vitrification. At about 851 °C, sample B undergoes an additional ceramic product. With regard to the ceramic mixture, we con-
expansion (Fig. 3). This is attributed to the entrapment of CO2 sider a system (Correi et al., 2004) formed by an inert com-
resulting from the dissociation of calcite (Peters and Iberg, ponent (the sand) and a plastic component (clays).
1978). It is in agreement with the mineralogical composition Having measured the values of the flexural strength (FS), the
which shows more carbonate in sample B than in sample A. The water absorption (WA) and the linear firing shrinkage (LF)
final shrinkage varies between 1.9% to 2%. The rate of shrink- (Table 3), a regression equation of the shape
age in the two samples is relatively low because these samples
consist mainly of illite and quartz. X
3
f ðS Þ ¼ ai S i ð1Þ
i¼0

can be thought for each property, where i ∈ [0;3], f presents a

third-degree polynomial for the water absorption, linear firing
shrinkage and the flexural strength, S; sand and αi; the coeffi-
cient determined from experiment results. The response func-
tion f can be expressed in its canonical form as a third-degree:

LF ¼ x1 S 3 þ x2 S 2 þ x3 S þ x4 ð2Þ

WA ¼ y1 S 3 þ y2 S 2 þ y3 S þ y4 ð3Þ

FS ¼ z1 S 3 þ z2 S 2 þ z3 S þ z4 : ð4Þ

The experiment values of the flexural strength (FS), the

Fig. 5. Variation of flexural strength FS (N/mm2) with sand content. water absorption (WA) and the linear firing shrinkage (LF) were
 S. Mahmoudi et al. / Applied Clay Science 42 (2008) 125–129 129

sence of expandable minerals. This agrees with mineralogical

analysis and is favorable for uses in ceramic. The studied raw
materials are classified as silty clays and be used as ceramic
bodies. Samples A and B show little plasticity, and drying is fast.
After firing at 1050 °C, there seems to be a clear relation
between flexural strength (FS), water absorption (WA), linear
firing shrinkage (LF) and percentage of sand. The addition of
sand decreases the bending strength, in such a way that sand
addition can be controlled by standards norm applied to ceramic
bodies.

References
Fig. 7. Variation of linear firing shrinkage LF (%) with sand content.
Assal, H.H., El- Didamony, H., Ramez, M., Mossalamy, F.H., 1999. The role of
lime inclusions on the properties of fired clay articles. Ind. Ceram. 19, 82–92.
used to iteratively calculate the coefficients (xi, yi and zi) of Carretero, M.I., Dondi, M., Fabri, M., Raimondo, M., 2002. The influence of
Eqs. (2), (3) and (4). The calculations were carried out with shaping and firing technology on ceramic properties of calcareous and non-
MATLAB 7.0.1. The model of Eqs. (5), (6) and (7) so obtained calareous illitic–chloritic clays. Appl. Clay Sci. 20, 301–306.
Correi, S.L., Hotza, D., Segadae, S.A., 2004. Simultaneous optimization of
can then be used to select the best mixture to produce a ceramic
linear firing shrinkage and water absorption of triaxial ceramic bodies using
body with specified properties of the flexural strength (FS), the experiments design. Ceram. Int. 30, 917–922.
water absorption (WA) and the linear firing shrinkage (LF). De Andres Gomes de Bared, A., Garcia Ramos, G., Raigon Pichardo, M.,
Sanchez Soto, P., 1990. Propiedades ceramics de arcillas procedentes de
LFðkÞ ¼ 0:00051429 S 2  0:062038 S þ 2:006 ð5Þ Guadix (Granada). Bol. Soc. Esp. Ceram. 29, 37–40.
Dondi, M., Ecrolani, G., Guarini, G., Marsigli, M., Mingazzini, C., Ventiri, I.,
WAðkÞ ¼ 0:00067017 S 3 þ 0:040191 S 2  0:11047 S þ 8:4154 ð6Þ 1999. Major deposit of brick clays in Italy: part 2. Technological properties
and uses. Tile Brick Int. 15, 360–370.
FSðkÞ ¼ 0:00052754 S 3  0:035277 S 2 þ 0:29038 S þ 17:04: ð7Þ Fabri, B., Fiori, C., 1985. Clays and complementary raw materials for stoneware
tiles. Mineral. Petrogr. Acta 29, 535–545.
Figs. 5, 6 and 7 show the flexural strength (FS), the water Gonzales, I., Galan, E., Fabbri, E., 1999. Probllematica de las emisiones de fluo,
absorption (WA) and the linear firing shrinkage (LF) of nine coloro y azufre durante la cossion de materiales de ladrilla. Bol. Soc. Esp.
Ceram. 37, 307–313.
mixtures obtained from the Eqs. (5), (6) and (7). The use of a Holtz, Kovacs, 1981. Kansas Geotechnical Survey, current research in Earth
high amount of sand provokes a decrease of the flexural Science. Bulletin 244, part 3, The Relationship between Geology and
strength (FS) and the linear firing shrinkage (LF), on the other Landslide Hazards of Atchison, Kansas and Vicinity.
hand an increase of the water absorption (WA). Indeed, a pro- Jauzein, A., 1962. Contribution à l'étude géologique des confins de la dorsale
portion of sand N 25% shows bad water absorption (WA) and tunisienne (Tunisie septentrionale), Thèse d'Etat. Faculté des Sciences de
Paris. 475 pp.
flexural strength (FS) according to ISO-10543-3 and ISO 10545-4, Maamouri, A.L., Salaj, J., Maamouri, M., Matmati, F., Zargouni, F., 1994. Le
respectively for commercial use of the ceramic bodies. The al- Crétacé inférieur du DJebel Oust (Tunisie Nord – Oriental) Microstratigraphie –
lowed maximum water absorption is 20% and the allowed mini- Biozonation – Aperçu sédimentologique. Zemny Plyn Nafta 39, 73–105.
mum flexural strength is 15 N/mm2. Table 3 and Fig. 7 show a Mahmoudi, S., 2004. Etude physico- chimique des argiles apto-albiennes de
decrease of the linear firing shrinkage when the amount of sand Djebel Oust et leur application industrielle, Mastère, Université de Tunis El
manar. Faculté des Sciences de Tunis. 76 pp.
increases. Meir, L.P., Nuesch, R., 1999. The lower cation exchange capacity limit
montmorillonite. J. Colloid Interface Sci. 217, 77–85.
4. Conclusion Parras, J., Sanchez-Jimenez, C., Rodas, M., luque, F.G., 1996. Ceramic
application of Middle Ordovician shales from central Spain. Appl. Clay Sci.
11, 25–41.
The two samples A and B collected from Barremian of Djebel
Peters, T., Iberg, R., 1978. Mineralogical changes during firing of calcium-rich
Oust indicate as dominant minerals illite and quartz, with acces- brick clays. Am. Ceram. Soc. Bull. 57 (504 pp.).
sory phases of kaolinite, illite–smectite mixed layers, chlorite, Thomas, H., Peer, G., 2001. Mineral raw materials in the brick and tiles industry —
calcite and feldspar. Chemical analysis shows that these clays are important parameters in the daily practice of geoscientist, part 2. Z. I. 12,
rich in silica, alumina and potash with minor contents of mag- 20–26.
Santos, P.S., 1997. Ciencia e tenologia de argilas, 3rd edition. Edgard blucher
nesium, titan and carbonates, accompanied by a significant
ltda, S. Paulo.
amount of iron oxides. Samples A and B reveal a low cation
exchange capacity and specific surface area, indicating the ab-