Applied Clay Science xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Applied Clay Science

journal homepage: www.elsevier.com/locate/clay

Research paper

Identifying the differences between clays used in the brick industry by

various methods: Iron extraction and NMR spectroscopy
⁎
Julie Peynea,b, Ameni Gharzounia, Isabel Sobradosc, Sylvie Rossignola,
a
Univ. Limoges, IRCER, UMR 7315, F-87000 Limoges, France
b
Bouyer Leroux, L'établère 44210, La Séguinière, France
c
Instituto de Ciencia de Materiales de Madrid, Consejo Superior de Investigaciones Científicas (CSIC), C/Sor Juana Inés de la Cruz, 3, 28049 Madrid, Spain

A R T I C LE I N FO A B S T R A C T

Keywords: This study focused on identifying the differences in a clay quarry that is used in brick production to understand
Clay minerals the variation between samples based on two kinds of clays: the first kind was suitable for brick production, but
Iron the second one did not show good production efficiency. Various samples were characterized by physical and
NMR spectroscopy chemical methods. In addition, free iron extraction by a CBD method and calcination at various temperatures
Calcination
were performed. The clay samples were then structurally characterized by XRD, FTIR and 27Al NMR spectro-
Bricks clay
scopy and thermally characterized using dilatometry. Finally, these results were correlated to the SEM in-
vestigation. These results revealed variations in the Si/Al molar ratio (1.4 to 3.3) due to the mineralogy variation
within the quarry. Indeed, the non-secondary phase and clay phase contents differed according to the location in
the quarry. NMR analysis also revealed that some iron atoms were located in a kaolinite structure, whereas other
iron phases, in goethite form, interacted with clay minerals only on the surface of, e.g., kaolinite and muscovite
and were not in their structures. From the thermal characterization, the influence of this structural iron in the
kaolinite was also evidenced by an increase in the total shrinkage. In conclusion, this work determined the
structural differences and their impact on brick production.

1. Introduction preparation (Jimenez-Quero et al., 2017). Even if it were possible to

Clay minerals are characterized by several physical and chemical
properties, such as their swelling behavior, cation exchange capacity
(CEC), cation adsorption, and high surface area (Bergaya and Lagaly,
2013) that cause these raw materials to be used strategically in nu-
merous industrial applications, such as brick production (Gualtieri
et al., 2015).
Conventionally, bricks are made from a mixture of numerous clays
extracted from different quarries (Kornmann, 2009). The final proper-
ties of the bricks (e.g., presence of failures, color, durability, compres-
sive strength) clearly depend on the raw materials used. Nonetheless,
the non-homogeneous mineralogy content of the brick clay mixture
creates difficulties in understanding the impact of each mineral from
each quarry. However, it is well known that an increase in the mon-
tmorillonite content leads to technical issues during the drying step due
to increases in the shrinkage and capillarity retention (Sigg, 1991). In
addition, the final color of the brick is clearly due to the initial iron
content (Li et al., 2016). The presence of kaolinite also increases the
plasticity of the brick mixture, thereby improving the mixture
separately determine the contribution of each mineral to the properties
of the mixture, the characterization techniques used in the industry
(mainly the particle size distribution, chemical analysis and miner-
alogy) would not allow the suitability of the clay for brick production to
be determined. The most advanced techniques, such as 27Al NMR and
SEM microscopy, are expensive and thus not yet in use. However,
several previous reports demonstrated that 27Al NMR spectroscopy is a
very powerful tool for deeply investigating the crystal chemistry of
layered silicates such as kaolinite, montmorillonite, and muscovite
(Thompson, 1984, Woessner, 1989, Takahashi et al., 2008). For ex-
ample, it is possible to evaluate the evolution of the 6-, 5- and 4-co-
ordinate aluminum sites with temperature during the dehydroxylation
reaction of kaolinite (Benharrats et al., 2003; Wanabe et al., 1987), thus
revealing the conversion of the 6-coordinate into 5- and 4-coordinate
aluminum. The structures of the clays considered for this application
are not often described in the literature. However, Mackenzie et al.
(1987) studied the thermal transformations of muscovite and used 27Al
and 29Si NMR to demonstrate that iron impurities lead to an increase in
recrystallization and favor the mullite to corundum conversion. In

⁎
Corresponding author.
E-mail address: Sylvie.rossignol@unilim.fr (S. Rossignol).

https://doi.org/10.1016/j.clay.2018.02.037
Received 12 October 2017; Received in revised form 23 February 2018; Accepted 25 February 2018
0169-1317/ © 2018 Elsevier B.V. All rights reserved.

Please cite this article as: Peyne, J., Applied Clay Science (2018), https://doi.org/10.1016/j.clay.2018.02.037
J. Peyne et al.
addition, Wei et al., 2012 highlighted the impact of goethite in a kao-
linite-goethite mixture or in a goethite interaction with kaolinite, re-
vealing a 6-coordinate aluminum contribution at a lower chemical shift
for the goethite-kaolinite mixture than for the goethite interaction with
kaolinite (3.04 and 5.76 ppm, respectively). It is also possible to detect
the kaolinite default in the structure and, more precisely, the iron in the
kaolinite structure, as described by Schroeder and Pruett, 1996. 27Al
NMR is thus an appropriate structural characterization method for
studying the impact of impurities on clay minerals. It is clear that the
iron oxide and (hydr)oxide contents can influence the clay structure.
Differences within a clay quarry could thus occur (Guzlena et al., 2017)
and lead to technical issues after numerous steps of brick production.
Thus, this study aims to understanding the differences in a clay quarry
and their influence on the quality of brick production. To achieve this
goal, two lots of clay with different impacts on brick production were
extracted in northwest France. Chemical treatments, such as a free iron
extraction and thermal treatment (with various calcinations at two
temperatures: 350 and 700 °C), were then performed to demonstrate
the influence of the iron and the thermal behavior of each clay. Next,
the physical and chemical features and structural information were
obtained by X-ray powder diffraction, FTIR and 27Al NMR spectroscopy.
These results were then correlated to the thermal behavior of the clays
before and after the chemical and thermal treatments.
2. Experimental
2.1. Raw materials and samples preparation
Numerous samples (18) from one quarry that is located in northwest
France and used in the brick industry were selected for this study. Two
different lots of samples (A and B) with two distinct sets of visual
characteristics, such as color and grain size, were observed. They were
also selected according to their influence on brick production, i.e., on
the scrap rate after the sintering of bricks. Indeed, A samples were
suitable for brick production, but B samples were not suitable because
they led to a significant increase in the scrap rates after the sintering
step. The sampling was performed based on the height of the rock face.1
The preparation of these samples was divided into two steps: (i) drying
for 24 h at 40 °C and (ii) coarse grinding (< 1 mm). The red/yellow/
brown color of these samples clearly demonstrates the presence of iron
oxohydroxide (Cornell and Schwertmann, 2002). Using the Mehra and
Jackson procedure (Mehra and Jackson, 1960), free iron oxide (organic
complexes, amorphous and highly crystallized forms) can be totally
dissolved. The main reactants used are a Na-citrate solution (0.3 M) and
a NaHCO3 solution (1 M), with sodium dithionite (Na2S2O4). A suitable
amount of the clay sample (4 g) was mixed with 40 mL of the Na citrate
solution and 5 mL of 1 M NaHCO3 in a water bath (80 °C). Then, 1 g of
Na2S2O4 was added, mixed continuously for 1 min and then mixed
occasionally for 15 min. Then, 10 mL of a saturated NaCl solution was
added and centrifuged for 5 min (at 2500 rpm/min). Finally, the su-
pernatant was decanted into a flask, and the residual solid was kept and
dried at 40 °C for 24 h. To characterize the temperature behavior, var-
ious calcinations in an electrical furnace for bulk samples before and
after the iron extraction were performed at 350 and 700 °C (T), which
corresponded to the temperatures after iron (hydr)oxide decomposition
and clay minerals dihydroxylation, respectively (Ruan et al., 2002;
Lopez et al., 2011). The following nomenclature is used: A/By-C-T,
where y specifies the number of the sample, C is the performed che-
mical treatment and T is the calcination temperature.
2.2. Characterization techniques
The mineral phases of the samples (either calcined or not) were
1
Bouyer Leroux, L'établère, 49,280 La Séguinière, France.
2
Applied Clay Science xxx (xxxx) xxx–xxx
identified by XRD with a BRUKER AXS D8 Advance powder dif-
fractometer using CuKα radiation (λα = 0.154186 nm). The in-
vestigated 2theta range used was between 2.5 and 65°, with a step size
of 0.012° (2theta) and an acquisition time of 2 s per step. JCPDS files
(Joint Committee Powder Diffraction Standard) were used for the
principal phase identification.
High-resolution MAS-NMR experiments were performed at room
temperature using a Bruker AVANCE-400 spectrometer operating at
104.26 MHz (27Al signal) on powder samples, which were spun at
10 KHz. The number of scans was 400(Autef et al., 2013). In order to
record the central transitions of aluminum (27Al) (I = 5/2), a π/8
(1.5 μs) pulse was applied to it and 1 MHz filter was used. A recycle
delay of 5 s was chosen to minimize saturation effects. Chemical shift
values were given with reference to an external aqueous AlCl3 solution.
The relative areas of the components were determined using DMFIT
program.
Diffuse reflectance spectra of the clay minerals (in powder form)
were obtained between 300 and 780 nm with a CARY 500 spectro-
photometer (cf. Supplementary files). A remission function (also called
the Kubelka Munk function, Kubelka and Munk, 1931; Barron and
Torrent, 1986) was applied to the experimental data according to the
(1 − R)2
literature, with f (R) = 2R , where R is the reflexion. Noise reduction
was then performed by fitting each Kubelka Munk spectrum with an
Adjacent Averaging (5 points) method. Finally, the second derivate was
calculated using Origin V8.
Dilatometric measurements were made under air by means of a
contact horizontal dilatometer (NETZSCH DIL 402C) on clay samples
pressed with a cylindrical geometry (H = 10 mm; Ø = 6.5 mm). Two
alumina (Al2O3) holders were placed at the surface end contacts of the
samples to prevent high-temperature diffusion between the samples and
the alumina rod. A calibration cycle (with a standard sample of Al2O3)
was performed and registered. The calibration data were then sub-
tracted from the data collected for each sample to eliminate the con-
tribution from the device. The thermal cycle used consisted of heating
at 10 °C/min to 1100 °C and cooling to 25 °C at a similar cooling rate.
3. Results and discussion
3.1. Physical and chemical characterization
The physical and chemical features of two sample sets (y = 1, 2, 3,
4) for the Ay-25 and By-25 lots are presented in Table 1. No significant
variation is observed in the weight percentage of the > 40 μm granu-
lometry range (31–45 and 36–46 wt% for By and Ay, respectively).
Nonetheless, an increase in the total fraction of the 40–80 μm range can
be observed for the By lot (8–15 and 5–6 wt% for Ay and By, respec-
tively). The specific surface area (SBET) and cation exchange capacity
(CEC) values depend on the sampling lot, and they vary with the same
behavior. Indeed, higher SBET and CEC values are obtained for Ay than
By (27 m2/g − 11.6 cmol(+)/kg and 17 m2/g − 6.4 cmol(+)/kg, re-
spectively). The wettability values (measurement previously explained
in Peyne et al., 2017) also depend on the sampling lot, with higher
values for By-25 (162–235 and 129–146 μL/g for By-25 and Ay-25, re-
spectively). The chemical compositions are consistent and exhibit a
significant change, with higher Si/Al and Si/Fe molar ratios for the Ay
samples. Fig. 1 shows pictures of the samples before and after the
chemical treatment and calcination. The yellow color of the A1–25
sample is typical for a rich goethite soil (Cornell and Schwertmann,
2002). The calcination at 350 and 700 °C (A1–350 and A1–700) leads to a
change from yellow to a red color, assigned to the goethite to hematite
transformation (Carlos Duarte Cavalcante et al., 2017). The free iron
extraction induces a change in color from yellow/orange to gray/white,
in agreement with the protocol, which shows that a high amount of
goethite content was removed. However, calcination at 350 °C for this
sample does not lead to a significant change, with only a slightly darker
J. Peyne et al.
Table 1
Physical and chemical properties of various samples.
Samples Particle size distribution Physical and chemical features
[ > 40 μm] (%) ± 1 [40–80 μm] wt% ± 1 SBET (m2/g) ± 1
B1–25 45 15 17
B2–25 31 8 20
A3–25 46 5 27
A4–25 35 6 34
(A)
A1-T
(a) T=25 °C
(b) T=350 °C
(c) T=700 °C
Fig. 1. Pictures of (A) A1-T and A1-C-T; (B) B1-T and B1-C-T samples for various temperatures T = (a) 25, (b) 350 and (c) 700 °C.
gray color (A1-C-350). It is only after calcination at 700 °C that a red
coloration is observed for the sample (A1-C-700). Similar changes in color
were detected for the B1-T and B1-C-T samples. However, the A1–700
sample looks darker than the B1–700 sample. In addition, the B1-C-T
samples seem darker than the A1-C-T samples. All these differences in
color between the lots could be due to either a different type of iron
(hydr)oxide or different local coordination sphere for the iron (Her-
billon et al., 1976). For the following results description, the number of
the sampling will be deleted from the nomenclature to simplify the text.
UV spectroscopy was performed to characterize the type of (hydr)
oxides in the samples after iron extraction (C) and after thermal treat-
ment. The data obtained are presented in the supplementary file and
show that increasing the calcination temperature induces a decrease in
goethite and an increase in hematite. Moreover, the amount of goethite
is slightly lower for the A350 than for the B350 samples, suggesting a
higher transformation of goethite to hematite at this temperature, in
agreement with the darker color relative to the higher amount of he-
matite content (Table 1). After the free iron extraction for all samples
(AC,T or BC,T), the changes differ because goethite is not detected,
consistent with the chemical extraction (see Figs. 1 and 2 in the sup-
plementary file). A hematite contribution only appears at 700 °C, which
seems to be linked to clay mineral dehydroxylation (such as mica,
kaolinite). However, this characterization technique does not allow the
determination of the iron coordination sphere in this clay mineral.
3.2. Structural characterization
To evaluate the influence of the iron extraction on the clay minerals,
FTIR spectroscopy was performed. The FTIR spectra of the samples
before and after the free iron extraction are shown in Fig. 2 (a, b). The
A25 sample (Fig. 2a) exhibits three contributions at 3692, 3669 and
3642 cm−1, which can be attributed to the OH stretching of inner-
surface hydroxyl groups (Madejova, 2003; Madejova and Komadel,
Applied Clay Science xxx (xxxx) xxx–xxx
Chemical composition
CEC (cmol(+)/kg) ± 0.5 W (μL/g) ± 5 Si/Al molar ratio Si/Fe molar ratio
6.4 162 1.5 8.1
6.1 235 1.4 8.3
11.6 146 2.1 12.8
8.6 129 3.3 16.1
(B)
A1-C-T B1-T B1-C-T
2001; Vaculikova et al., 2011), and one at 3619 cm−1, which corre-
sponds to inner the Al2-OH stretching (Vaculikova et al., 2011) of
kaolinite minerals (Parker, 1969). The band at 1642 cm−1 is directly
linked to the OeH vibration band from the water and is mainly related
to the preparation of the KBr pellets. A broad band at 1172 cm−1 is
observed and is due to the SieOeSi from quartz (Ravinska et al., 2010).
The other contributions at 1106, 1031, and 1004 cm−1 are related to
kaolinite and to the SieOeSi stretching vibrations (Oikonolopoulis
et al., 2015). Additionally, the bending vibrations of the inner Al2-OH
and the SieOeAl vibration band from kaolinite are observed at 912 and
533 cm−1, respectively (Vaculikova et al., 2011). The SieO vibration
bands from quartz minerals can be seen at 796, 778 and 692 cm−1
(Ravinska et al., 2010). Similar contributions are observed for the AC/25
sample (Fig. 2a). An additional contribution is observed at 1402 cm−1,
which can be attributed to the sodium carbonate vibrations (Huang and
Kerr, 1960) arising from the iron extraction procedure. A weak con-
tribution can also be seen after the iron extraction at 933 cm−1, which
may be correlated to the inner surface Al2-OH. With the exception of
some contributions, the B25 raw clay sample shows a similar FTIR
spectrum (Fig. 2b). However, the two absorption bands at 3669 and
3642 cm−1 due to kaolinite are more defined, which clearly means the
kaolinite is highly crystalline (Felhi et al., 2008). In addition, the quartz
doublet at 792 and 778 cm−1 is less visible, which suggests a lower
quartz content. The free iron extraction does not lead to any change in
the FTIR spectrum (Fig. 2b).
The X-ray powder diffraction patterns of the clay samples at various
calcination temperatures and before/after the free iron extraction are
plotted in Fig. 3 (a, b). The A25 sample (Fig. 3a) is typical of a clay
mixture composed of quartz and kaolinite but also muscovite and
feldspar (albite and anorthite). The goethite content observed in the
UV–visible spectra (data shown in supplementary file) is not detected
due to its small amount (in comparison with quartz) and to the poor
crystallinity of this type of mineral, usually observed in soils
3
J. Peyne et al. Applied Clay Science xxx (xxxx) xxx–xxx

(a) 80 and 40 ppm and another in the range of 20 to −20 ppm, which are
attributed to four- and six-coordinate aluminum, respectively (Bardy
et al., 2007). The 27Al NMR spectra of the A25 clay sample (Fig. 4a)
shows two contributions at 68 and 56 ppm, due to four-coordinate
aluminum (Al1IV IV
68ppm and Al256ppm) and two contributions related to six-
coordinate aluminum (Al11ppm and Al2VI
VI
−6ppm) (Sanchez-Munoz et al.,
2013). The signal at 68 ppm may be related to Al in Q3(3Si) sites,
whereas the signal at 56 ppm may be related to Al in Q4(4Si) sites, as
suggested by Takahashi et al. (2008), in the structural framework of
montmorillonite.
A similar spectrum is obtained after calcination at 350 °C (A350 -
Fig. 4a). However, the intensities of the Al1IV IV
68ppm and Al256ppm peaks
decrease slightly. Calcination at 700 °C (A700 – Fig. 4a) leads to the
appearance of a five-coordinate aluminum contribution at 30 ppm
(AlV), linked to the dehydroxylation of clay-type minerals such as
kaolinite and muscovite (Massiot et al., 1995; Carroll et al., 2005). In
addition, it appears that the Al2IV 56 ppm contribution increases but that
the Al1IV68 ppm contribution is not detected. Iron extraction (AC-25–
Fig. 4a) leads to similar spectra to A25. However, an important increase
4000 3500 3000 1500 1000 500 in the Al1IV IV
68 ppm and Al256 ppm contributions at the expense of the six-
coordinate aluminum contributions (Al1VI VI
1 ppm and Al2-6 ppm) is observed
Wavenumber (cm-1) after calcination at 350 °C (AC-350 – Fig. 4a). Comparable behavior is
observed and is even accentuated after calcination at 700 °C (AC-700–
(b) Fig. 4a). In addition, a new contribution at 40 ppm is observed, which
could be attributed to another four-coordinate aluminum (Al3IV). It also
seems that the appearance of the five-coordinate aluminum contribu-
tion is more pronounced for this sample. The Al1VI contribution cannot
be detected after calcination at 700 °C and free iron extraction. The
major contributions observed for the B25 clay mineral (Fig. 4b) are due
to six-coordinate aluminum environments (at 1 and −6 ppm, Al1VI and
Al2VI, respectively), whereas those at 68 and 56 ppm for four-coordinate
aluminum are minor (Al1IV IV
68ppm and Al256ppm). Calcination at 350 °C
leads to a weak increase in these contributions (B350 - Fig. 4b). The
main differences between the A700 and B700 samples occurs in the in-
tense bands of the 5-coordinate aluminum (AlV) at 30 ppm and of the 4-
coordinate aluminum (Al1IV IV
68ppm and Al256ppm), to the detriment of the
contributions of the 6-coordinate aluminum (Al1VI VI
1ppm and Al2-6ppm).
Equivalent behaviors are observed for the BC-25 and BC-350 samples
(Fig. 4b). However, B700 and BC-700 exhibit important differences, par-
ticularly an important diminution of the AlV contribution and a higher
transformation rate of the 6-coordinate aluminum to the 4-coordinate
aluminum (Al1VI VI IV IV
1ppm and Al2-6ppm in Al168ppm and Al256ppm). In addition,
4000 3500 3000 1500 1000 500
the contribution of the newly formed 4-coordinate aluminum is more
-1
Wavenumber (cm ) pronounced. A more detailed description, with the FWMH and areas of
the various contributions for each NMR spectrum, is provided in
Fig. 2. FTIR spectra of the (a) e A25 and AC-25 (b) e B25 and BC-25 samples. Table 2. It seems that at all temperatures studied, the AlVI -6ppm con-
tribution areas are higher for AT and AT/iron. This can be attributed to a
(Vodyanitskii, 2010). The free iron extraction (AC-25) is characterized local environment of kaolinite or of Al-substituted goethite, which may
by similar reflections. The 350 °C calcination (A350 and AC-350) does not be observed in soil (Kim et al., 2015). In addition, the Al1IV 68ppm and

induce any change in the X ray powder patterns (Fig. 5a). However, the Al2IV
56ppm contributions are significantly higher. This can be attributed to

700 °C calcination (A700 and AC-700) leads to disappearance of the
kaolinite reflections, in agreement with kaolinite dehydroxylation
(Lopez et al., 2011) and the additional appearance of a dome between
20 and 30°(2θ). The characterization of the BT and BC-T samples is si-
milar to that of the AT and AC-T samples, respectively (Fig. 5b), although
the feldspar content, in agreement with the X-ray diffraction patterns
(Fig. 3). For all samples, it appears that the iron extraction enhances the
transformation of the six coordinate aluminum to the 4-coordinate
aluminum (from 26.2 to 65.5% and from 6.9 to 76.8% for A700 and AC-
700, respectively). The new 4-coordinate aluminum that appears after

the feldspar reflections are less apparent. Moreover, the kaolinite and
muscovite reflections of BT and BC-T are more intense than they are for
AT and AC-T. These qualitative data from X-ray diffraction do not pro-
vide a final conclusion about the iron coordination sphere in the clay
minerals.
To better characterize the iron coordination sphere and especially
the iron position in the clay structure, a spectroscopic study was per-
formed using 27Al NMR. The 27Al NMR spectra of the clay minerals after
calcination and after free iron extraction are presented in Fig. 4 (a, b).
All the clay samples are characterized by a broad contribution between
calcination at 700 °C may be due to secondary or intermediate phases
such as Al-O-Ca (Pardal et al., 2012; Wel-Lin Shao, 1989). The calcium
could come from calcite decomposition, which was not detected by X-
ray powder diffraction due its low amount (Stanmore and Gilot, 2005;
Pardal et al., 2012; Wel-Lin Shao, 1989), and could then react with
other compounds. In the same way, the Al2IV contribution, mainly from
muscovite, is greater for A350 than it is for B350. An increase from 8.4 to
23.4% for A350 and AC-350 is observed and may indicate that the free
iron inhibits this local environment for the A samples, but this is not
observed for the B samples. For the previous samples at 700 °C, the
decrease in the Al1VI 1 ppm contribution and the appearance of the Al
V

4
J. Peyne et al. Applied Clay Science xxx (xxxx) xxx–xxx

(a) (b)

Q
Q Q
Q
I Q Q I I
I I I I I I I I I I I I Q
I A I I I I IQ
I Kaol
I I
I Kaol I A
AC-700 BC-700

A700 B700

AC-350 BC-350

A350 B350

AC-25 BC-25

A25 B25

3 5 10 15 20 25 30 35 40 3 5 10 15 20 25 30 35 40
2θ (°) 2θ (°)

Fig. 3. XRPD patterns of (a) AT, AC-T and (b) BT, BC-T samples (I: Illite, Kaol: Kaolinite, Q: Quartz, A: Albite) (PDF files of I: 00-003-0849, Q: 01-078-1252, Kaol: 00-003-0052, A: 04-014-
0490).

contribution are strong and are linked to kaolinite dehydroxylation. In 3.3. Thermal behavior
addition, this observation is related to the increase in the AlIV con-
tributions, which are consistent with iron that could be linked to the To observe the influence of the iron chemical speciation in the two
kaolinite sheet. clays, the A25, AC-25, B25 and BC-25 samples were characterized by

(A) (B)

Al2(IV)
Al2(IV) (IV) Al3(IV)
Al1
Al3(IV)
Al1(IV)
Al1 (VI)
Al(V) Al1(VI)
Al(V) Al2(VI) Al2(VI)
AC-700 BC-700

A700
B700

AC-350
BC-350

A350
B350

AC-25
BC-25

A25
B25

100 80 60 40 20 0 -20 -40 100 80 60 40 20 0 -20 -40

(ppm) (ppm)

27
Fig. 4. Al NMR spectra of the samples (A) AC-T, AT and (B) BC and BC-T for T = 25, 350 and 700 °C.

5
J. Peyne et al. Applied Clay Science xxx (xxxx) xxx–xxx

Table 2
27
FWHM and area of the various contributions from Al MAS-NMR spectra.

Sample Al1 IV
68 ppm Al2IV 56 ppm Al3IV 40 ppm AlV 30 ppm Al1VI 1 ppm Al2VI −6 ppm

FWHM (ppm) Area (%) FWHM (ppm) Area (%) FWHM (ppm) Area (%) FWHM (ppm) Area (%) FWHM (ppm) Area (%) FWHM (ppm) Area (%)

A25 14 7.2 10 12.4 13 72.7 12 7.8

AC-25 12 5.8 9 11.8 12 68.6 12 13.8
A350 14 5.1 9 8.4 12.5 80.7 12 5.8
AC-350 12 10.1 11 23.4 12 58.2 15 8.3
A700 10 65.5 30 6,9 12 26.2 5.2 1.3
AC-700 14 76.8 14 6.1 33 10.3 10 6.9
B25 18 5.6 12 1.1 13 88.0 12 5.2
BC-25 14 5.3 18 3.9 13 86.0 12 4.6
B350 16 7.2 12 2.2 13 88.8 12 1.8
BC-350 12 5.7 16 3.6 13 89.8 12 1.2
B700 12 6.0 12 11.5 45 37.2 14 35.4 12 1.1
BC-700 14 19.1 14 36.3 12 9.8 30 28.7 14 6

(a) (b)
4 4

2 2
Dilatation (%)

0 Dilatation (%) 0
30 200 400 600 800 1000 1100 30 200 400 600 800 1000 1100

-2 Temperature (°C) -2 Temperature (°C)

-4 -4

-6 -6

-8 -8

Fig. 5. Dilatometry measurements with heating of samples before and after the iron removal for (a) e A25, AC,25 and (b) e B25, BC,25.

dilatometry (Fig. 5 a, b).
For all the samples, two shrinkages are observed, the first at ap-
proximately 900 °C due to mullite formation and the other at 1050 °C
due to the free silica produced by the thermal decomposition of the clay
minerals, which delays the viscous flow (Castelein et al., 2000). The A25
clay (Fig. 5a) is characterized by a quasi-null total shrinkage, and the
B25 clay (Fig. 5b) shrinkage is more pronounced (up to 2%). After the
iron extraction, the dimensional variation of the AC-25 sample is similar
to the A25 sample, while the BC-25 sample (Fig. 5b) is characterized by
an increase in the total shrinkage (6%). This reveals that the free iron in
the AC-25 sample does not perturb the sintering phenomena of the clay
phases (Fig. 5a). However, this chemical treatment leads to the ap-
pearance of these two shrinkages at slightly lower temperatures for BC-
25. It seems that during the dehydroxylation of the kaolinite, the pre-
sence of iron in the clay structure (observed by the X-ray mapping, see
Supplementary files) leads to the formation of some iron oxide, which
favors sintering (Soro, 2003). This results in a lower sintering tem-
perature, consequently increasing the total shrinkage. Finally, the
presence of some structural iron inside the kaolinite structure can be
linked to the increase in the scrap rate during brick production. Indeed,
the increase of the shrinkage at high temperature may induce higher
stress, leading to scrap in the final shape of the bricks.
4. Discussion
The 27Al NMR results reported above showed some differences,
depending on the temperature, for the 6-coordinate aluminum (Al1VI
and Al2VI). Based on this result, the ratios between the contributions
before and after the iron extraction were calculated for the 4-, 5- and 6-
coordinate aluminum for AT/AC-T and BT/BC-T samples (Fig. 6A,B – a,b).
A ratio higher than 1 is characteristic of an AlX contribution that is less
important after the iron extraction than before the extraction. A ratio
smaller than 1 indicates an AlX contribution that is greater after the iron
extraction than before. The contribution ratios of the Al2VI for A25/AC/
25 and A350/AC/350 (▼- Fig. 6A–a) are constant and < 1. However, an
increase in this ratio is observed between 25 and 350 °C for the BC/BC-T
contributions (Fig. 6A–b). The decrease in the Al2VI contribution of the
BC,350 sample could be linked either to the some aluminum that is in-
itially in a highly distorted environment in the sample (Newman et al.,
1994) or to the presence of some gibbsite minerals in the initial sample
(Ashbroook et al., 2000). In addition, the increase in the Al2VI con-
tribution after the iron extraction for the AC/T sample (ratio < 1) may
mean that the free iron was more extractable. This is consistent with the
darker color of the A350 clay compared to B350, where the free iron was
more susceptible to transformation into hematite at this temperature.
Indeed, the lighter color could be due to some iron phases. For all
samples, the ratio of the Al1VI contributions (▲- Fig. 6A), attributed to
the 6-coordinate aluminum of kaolinite, remains constant between 25
and 350 °C and increases after the 700 °C calcination. At 700 °C, this
ratio is slightly higher for the B700/BC,700 ratio than for A700/AC,700 (5.9
and 3.8, respectively), which suggests a more significant dehydroxyla-
tion process for the B clay sample after the iron extraction. The Al1VI
transformation into Al1,2IV aluminum is more favorable in case of the B
sample, which seems to reveal an interaction between the free iron and
the kaolinite content. The ratio of the AlV contributions at 700 °C ( -
Fig. 6B – a,b), higher for B700/BC-700 than for A700/AC-700 (1.36 and
0.65, respectively), is characteristic of more rapid kinetics of the kao-
linite dehydroxylation (from a 6-coordinate Al to a 4-coordinate alu-
minum) after the iron extraction for the B sample. This is consistent
with the previous conclusion. The variation with temperature of the

6
J. Peyne et al. Applied Clay Science xxx (xxxx) xxx–xxx

(A)
(a) (b)

7 7

6 6

5 5

4 4
AlVI(%) (AT)

AlVI(%) (AC/T)

AlVI(%) (BT)

AlVI(%) (BC/T)
3 3

2 2

1 1

0 0
0 100 200 300 400 500 600 700 800 0 100 200 300 400 500 600 700 800

Temperature (°C) Temperature (°C)

(B)

7 7

6 6

5 5
AlIV-V(%) (AC/T)
AlIV-V(%) (AT)

4 4
AlIV-V(%) (BC/T)
AlIV-V(%) (BT)

3 3

2 2

1 1

0 0
0 100 200 300 400 500 600 700 800 0 100 200 300 400 500 600 700 800

Temperature (°C) Temperature (°C)

Fig. 6. Ratio of the (A) AlVI contribution and (B) AlIV and AlV, before and after the iron oxide removal for (a) AT, AC,T and (b) BT, BC,T with ▼ Al2VI; ▲Al1VI; ●Al2IV; ■ Al1IV and AlV.

ratio of the Al1IV contributions (●- Fig. 6B) differs in the two samples.
As a function of temperature, the ratio for the B samples remains con-
stant at approximately 0.3 (Fig. 6Bb), but for the A samples, a small
decrease is observed from 1.12 to 0.35 for 25 and 350 °C, respectively
(Fig. 6Ba). This shows that iron extraction favors the transformation of
6-coordinate aluminum to 4- and 5-coordinate aluminum. The change
in the ratio of the Al2IV contributions with temperature is different for
the A and B samples (Fig. 6B). Between 25 and 350 °C, the ratio de-
creases from 1.3 to 0.5, respectively, for the A samples, but for the B
samples, a slight increase from 1.02 to 1.31 occurs at 25 and 350 °C,
respectively. At T = 700 °C, only the contributions for B700 and BC-700
are observed, which may be consistent with an easier kaolinite dehy-
droxylation at this temperature for this sample, which is even easier
when the free iron is extracted, as observed by a strong increase in the
contribution intensity for BC-700. The many changes in these contribu-
tions seem to indicate that the free iron in the samples may interact
with some clay minerals, such as kaolinite and muscovite, only on the
surface of the clay minerals, such as the surrounding kaolinite sheets.
Finally, two hypotheses can be determined from these results: (i) the
A25 sample contains more free iron phases, easily accessible to the
chemical treatment, and has iron phases that interact with the musco-
vite minerals, and (ii) the B25 sample is more characterized by kaolinite
in interaction with some iron phases, which may decrease the efficiency
of the kaolinite dehydroxylation.
5. Conclusions and theoretical model
Based on all these characterization results, it is possible to establish
a theoretical model of the two clays (A and B) from the same quarry
(Fig. 7). Calcination at 350 and 700 °C leads to different colors for the
two lots of clays. However, UV spectroscopy only confirms the presence
of goethite and its transformation into hematite after calcination for
both clays. This technique did not yield any conclusions about the co-
ordination sphere of the iron (hydr)oxides. Structural characterization
using 27Al NMR or X ray powder diffraction and thermal character-
ization were therefore mandatory to gain this goal.
(i) First, a higher content of secondary phases such as quartz and
feldspar was detected for the A samples, while higher amounts of
kaolinite and muscovite were observed for the B samples. For a
similar iron content, the darker color of the A samples after cal-
cination at 350 °C means that the iron in these samples is more
accessible.
(ii) A higher total shrinkage with temperature is observed for the B

7
J. Peyne et al.
Fig. 7. Theoretical model of the differences for (a, e) A25 and (b,
clays before and after the iron extraction. This result is accentuated
after the iron extraction. This arises from the presence of structural
iron in the kaolinite phases, which favors sintering.
(iii) 27Al NMR showed that some iron phases, in goethite form, inter-
acted with the clay minerals muscovite and kaolinite for the A and
B samples, respectively.
References
Ashbroook, S.E., McManus, J., MacKenzie, K.J.D., Wimperis, S., 2000. Multiple Quantum
and Cross Polized 27Al NMR of mechanically treated mixtures of kaolinite and
gibbsite. J. Phys. Chem. B 104, 6408–6446.
Autef, A., Joussein, E., Poulesquen, A., Gasgnier, G., Pronier, S., Sobrados, I., 2013. Role
of metakaolin dehydroxylation in geopolymer synthesis. Powder Technol. 250,
33–39.
Bardy, M., Bonhomme, C., Fritsch, E., Maquet, J., Hajjar, R., Allard, T., Derenne, S., Calas,
G., 2007. Al speciation in tropical podzols of the Upper Amazon Basin: a solid state
27
Al MAS and MQMAS NMR study. Geochim. Cosmichim. Ac. 71, 3211–3222.
Barron, V., Torrent, J., 1986. Use of the Kubelka Munk theory to study the influence of
iron oxides on soil colour. Eur. J. Soil Sci. 37, 499–510.
Benharrats, N., Belbachir, M., Legrand, A.P., D'espinose de la Caillerie, J.B., 2003. 29Si
and 27Al MAS NMR study of the zeolitization of kaolin by alkali leaching. Clay Miner.
38, 49–61.
Bergaya, F., Lagaly, G., 2013. Chapter 1 - General Introduction: Clays, Clay Minerals, and
Clay Science, Developments in Clay Science. vol. 5. pp. 1–19.
Carlos Duarte Cavalcante, L., Sousa Bezera da Silva, H.L., Domingos Fabris, J., Domingos
Ardisson, J., 2017. Red and Yellow Ocres from the Archaleogical Site Pedra to
Cantagalo I, in Piripiri, Piaui, Brazil. In: Hyperfine Interact. 2017. pp. 222–238.
Carroll, D.L., Kemp, T.F., Bastow, T.J., Smith, M.E., 2005. Solid state NMR character-
ization of the thermal transformation of a Hungarian white illite. Solid State Nucl.
Mag. 28, 31–43.
Castelein, O., Soulestin, G., Bonnet, J.P., 2000. Influence of heating rate on the thermal
comportment and the mullite formation from a kaolin raw material. Ceram. Int.
27 (5).
Cornell, R.M., Schwertmann, U., 2002. The Iron Oxides, Structure, Properties, Reactions,
Occurrences and Uses. Wiley VCH Verlag GmbH & Co, KGa, Weinheim.
Felhi, M., Tlili, A., Gaied, M.E., Montacer, M., 2008. Mineralogical study of kaolinitic
clays from Sidi El Bader in the far North of Tunisia. Appl. Clay Sci. 39, 208–217.
Gualtieri, A.F., Ricchi, A., Lassinantti, M., Gualtieri, S., Maretti, M., 2015. Kinetic study of
the drying process of clay bricks. J. Therm. Anal. Calorim. 153–167.
Guzlena, S., Sakale, G., Certoks, S., 2017. Clayey material analysis for assessment to be
used in ceramic building materials. Modern Building Materials, Structures and
Techniques. MBMST 2016. Process. Eng. 172, 333–337.
Huang, C.K., Kerr, P.F., 1960. Infrared study of the carbonate minerals. Am. Mineral. 45,
311–324.
Jimenez-Quero, V., Maza-Ignacio, O.T., Guerrero-Paz, J., Campos-Venegas, K., 2017.
Industrial wastes as alternative raw materials to produce ecofriendly fired bricks. J.
Phys. 792 (012065).
Kim, J., Illot, A.J., Middlemiss, D.S., Chernova, N.A., Pinney, N., Morgan, D., Grey, C.P.,
2015. 2H and 27Al solid state NMR study of the local environments in Al doped 2-line
ferrihydrite, goethite and lépidocrocite. Chem. Mater. 27, 3966–3978.
Kornmann, M., 2009. Matériaux de terre cuite – Propriétés et produits, c906. Techniques
de l'Ingénieur.
8
Applied Clay Science xxx (xxxx) xxx–xxx
) B25 clays used in the bricks industry.
Kubelka, P., Munk, F., 1931. Ein Beitrag Zur Optik Der Farbanstriche. Z. Tech. Phys. 12,
593–601.
Li, H., Dong, L., Jiang, Z., Yang, X., Yang, Z., 2016. Study on utilization of red brick waste
powder in the production of cement based red decorative plaster for walls. J. Clean.
Prod. 133, 1017–1026.
Lopez, F., Martirena Fernandez, R., Scrivener, J.F., 2011. The origin of the pozzolanic
activity of calcined clay minerals: a comparision between kaolinite, illite and mon-
tmorillonite. Cem. Concr. Res. 45, 113–222.
Mackenzie, K.J.D., Brown, I.W.M., Cardile, C.M., Meinhold, R.H., 1987. The thermal
reactions of muscovite studied by high resolution solid state 29Si and 27Al NMR. J.
Mater. Sci. 22, 2645–2654.
Madejova, J., 2003. FTIR techniques in clay mineral studies. Vib. Spectrosc. 31, 1–10.
Madejova, J., Komadel, P., 2001. Baseline studies of the clay minerals society source
clays: infrared methods. Clay Clay Miner. 49, 410–432.
Massiot, D., Dion, P., Alcover, K.F., Bergaya, F., 1995. 27Al and 29Si MAS NMR study of
kaolinite thermal decomposition by controlled rate thermal analysis. J. Am. Ceram.
Soc. (11), 2940–2945.
Mehra, O.P., Jackson, M.L., 1960. Iron oxide removal from soils and clays by a dithionite-
citrate system buffered with sodium bicarbonate. Clay Clay Miner. 317–327.
Newman, R.H., Childs, C.W., Churchman, G.J., 1994. Aluminum coordination and
structural disorderin halloysite and kaolinite by 27Al NMR spectroscopy. Clay Miner.
29, 305–312.
Oikonolopoulis, I.K., Perraki, M., Tougiannidis, N., Perraki, T., Kasper, H.U., Gurk, M.,
2015. Clays from Neogene Achlada lignite deposits in Florina basin (Wester
Macedonia, N. Greece): a prospective resource for the ceramics industry. Appl. Clay
Sci. 103, 1–9.
Pardal, X., Brunet, F., Charpentier, T., Pochard, I., Nonat, A., 2012. 27 Al and 29Si Solid
State NMR characterization of calcium aluminosilicate hydrate. Inorg. Chem. 5,
1827–1836.
Parker, T.W., 1969. A classification of kaolinites by infrared spectroscopy. Clay Miner. 8,
135–147.
Peyne, J., Gautron, J., Doudeau, J., Joussein, E., Rossignol, S., 2017. Influence of calcium
addition on calcined brick clay based geopolymers: a thermal and FTIR spectroscopy
study. Constr. Build. Mater. 152, 794–803.
Ravinska, R., et al., 2010. Mineralogical characterization studies of ancient potteries of
Tamilnadu, India by FTIR spectroscopy. J. Chem-NY 7, 185–190.
Ruan, H.D., Frost, R.L., Kloprogge, J.L., Duong, L., 2002. Infrared spectroscopy of goethite
dehydroxylation: FTIR microscopy of in situ study of the thermal transformation of
goethite to hematite. Spectrochim. Acta A 58, 967–981.
Sanchez-Munoz, L., Sanz, J., Sobrados, I., Gan, Z., 2013. Medium range order in dis-
ordered K feldspars by multinuclear NMR. Am. Mineral. 98, 2115–2131.
Schroeder, P.A., Pruett, R.J., 1996. Fe ordering in kaolinite: insights from 29Si and 27Al
MAS NMR spectroscopy. Am. Mineral. 87, 26–38.
Sigg, J., 1991. Les produits de terre cuite. pp. 135 (Edition Septima – Paris).
Soro, N.S., 2003. Influence des ions fer sur les transformations thermiques de la kaolinite.
In: Thèse de l'Université de Limoges, Spécialité: Sciences des Matériaux Céramiques
et Traitements de Surface.
Stanmore, B.R., Gilot, P., 2005. Review-calcination and carbonation of limestone during
thermal cycling for CO2 sequestration. Fuel Process. Technol. 86, 1707–1746.
Takahashi, T., Kanehashi, K., Saito, K., 2008. First evidence of multiple octahedral Al sites
in Na-montmorillonite by 27Al multiple quantum MAS NMR. Clay Clay Miner. 56 (5),
520–525.
Thompson, J.G., 1984. 29Si and 27Al nuclear magnetic resonance spectroscopy of 2:1 clays
minerals. Clay Miner. 19, 229–236.
Vaculikova, L., Plevova, E., Vallova, S., Koutnik, I., 2011. Characterization and
J. Peyne et al. Applied Clay Science xxx (xxxx) xxx–xxx

differentiation of kaolinites from selected Czech deposits using infrared spectroscopy
and differential thermal analysis. Acta Geodyn. Geomater. 18, 59–67.
Vodyanitskii, Y.N., 2010. Iron hydroxides in soils: a review of publications. Eurasian Soil
Sci. 43, 1244–1254.
Wanabe, T., Shimiz, H., Nagasawa, K., Masuda, A., Saito, H., 1987. 29Si and 27Al MAS
NMR study of the thermal transformation of kaolinite. Clay Miner. 22, 37–48.
Wei, S., Tan, W., Zhao, W., Zu, Y., Liu, F., Koopal, K., 2012. Microstructure, interaction
mechanisms and stability of binary systems containing goethite and kaolinite. Soil
Sci. Soc. Am. J. 76, 389–398.
Wel-Lin Shao, W., 1989. NMR Studies of Amorphous Silicon and Aluminosilicate Glass.
Iowa State University (PhD thesis).
Woessner, D.E., 1989. Characterization of clay minerals by 27Al nuclear magnetic re-
sonance spectroscopy. Am. Mineral. 74, 203–215.