Characterization of modified kaolin from the Ranong deposit Thailand by XRD, XRF, SEM, FTIR and EPR techniques

Characterization of modified kaolin fromthe Ranong deposit Thailand by XRD,XRF, SEM, FTIR and EPR techniques

N. WORASITH1 , { , B . A . GOODMAN2 , { , J . NEAMPAN3 , N . JEYACHOKE1AND

P . THIRAVETYAN1 ,*

1 School of Bioresources and Technology, King Mongkut’s University of Technology, Thonburi, Bangkhuntien,

Bangkok, Thailand, 2 Health and Environment Department, Environmental Resources &Technologies, Austrian

Institute of Technology, A-2444 Seibersdorf, Austria, and 3 Department of Geology, Chulalongkorn University,

Bangkok, Thailand

(Received 13 July 2010; revised 2 May 2011; Editor: Eric Ferrage)

ABSTRACT: Various physical and analytical techniques (XRD, XRF, SEM, FTIR and EPR) have

been used to investigate the effects of chemical and/or physical modification of Ranong kaolin,

which has been proposed as a potential bleaching clay for vegetable oils. Acid treatment after

grinding resulted in major changes compared with acid treatment of the original mineral sample or

mechanical treatment alone. Previous work has shown that the combined treatments produce

increases in surface area and new porous structures, and the present measurements show reductions

in Al:Si ratios. These are accompanied by a major reduction in O�H stretching vibrations as a result

of grinding, although acid treatment produced little subsequent effect on the O�H bands in the FTIR

spectra. However, acid treatment resulted in a reduction in the Al�OH�Al bending vibrations and

the appearance of Si�O bands associated with newly synthesized material; these effects were much

greater with samples that had been ground prior to the acid treatment. There were appreciable

qualitative differences in the way in which the EPR spectra of Fe and Mn were affected; the Fe signal

was sensitive to mechanical treatment, but little subsequent change was induced by acid extraction,

whereas the Mn peaks were sensitive to the both the pH and the chemical nature of the acid used.

These results therefore indicate that the Fe and Mn are in different types of site in the kaolin

structure. Little change was observed in the main oxygen-based free radical centre associated with Si

atoms, but that associated with Al was lost as a result of the treatments. Such mineral

characterization is of fundamental importance to understanding the modification of kaolins and

their uses as adsorbents in the food and environmental sciences.

KEYWORDS: kaolin, XRD, XRF, SEM, FTIR, EPR, grinding, acid activation, pH, bleaching clay, natural oildecolourization, Ranong deposit, Thailand.

The kaolin group minerals, kaolinite, dickite,

nacrite and halloysite, are 1:1 layer silicates, in

which each layer consists of a combination of one

sheet of tetrahedral and one sheet of octahedral

* E-mail: [email protected]{ On leave from Department of Chemistry, Faculty of Science and Technology, Rajamangala University ofTechnology Krungthep, 2 Nang Lin Chi Road, Soi Suan Plu, Sathorn, Bangkok, Thailand 10120.{ Current address: State Key Laboratory for Conservation and Utilization of Subtropical Agro-Bioresources,Guangxi University, Nanning, 530004 Guangxi, Peoples’ Republic of China.DOI: 10.1180/claymin.2011.046.4.539

ClayMinerals, (2011) 46, 539–559

# 2011 The Mineralogical Society

cations linked by bridging oxygen atoms (Murray,

2007). The surface of the tetrahedral sheet consists

of oxygen atoms, whereas that of the octahedral

sheet contains hydroxyl groups. In kaolinite, dickite

and nacrite the neighbouring layers are held

together by hydrogen bonds and these minerals

differ only in the stacking arrangements of the

layers. Halloysite can contain a variable number of

water molecules in this interlayer region; the d001peak of fully hydrated halloysite is at lower angle

(8.75º2y) than the other kaolin minerals (12.37º2y),but it collapses to about 12.28º2y when the

humidity of halloysite is below 30%. These

minerals are all dioctahedral, and thus nominally

have only two thirds of their octahedral cation sites

occupied.

With an ideal composition, the tetrahedral cations

are Si and the octahedral cations Al. Such a

structure is charge neutral, i.e. there is no net

layer charge but, in common with other alumino-

silicate minerals, various isomorphous substitutions

can occur. Substitution of Al3+ or Fe3+ for Si in the

tetrahedral sheet generates an electron surplus,

which may be compensated by either the generation

of a layer charge (which yields a cation exchange

capacity), or the protonation of oxygen atoms to

generate hydroxyl groups on the edge surfaces of

the tetrahedral sheet (e.g. Huertas et al., 1998).

Note that the possible substitution of tetravalent

ions, such as V4+ or Mn4+ for Si, would not

generate a surface charge but, like Fe, these ions

are capable of existing in more than one oxidation

state, so any process which alters their redox status

would be expected to influence the surface proper-

ties of minerals containing them. Various substitu-

tions in the octahedral sheet, such as V4+, Mn4+,

Fe3+, Fe2+, Mg2+, Mn2+, etc. may or may not lead

to the generation of a surface charge, or a change in

the degree of surface protonation; furthermore,

there is the possibility of maintaining charge

neutrality by substituting three divalent ions for

two trivalent ions, since one third of the octahedral

sites are vacant in an ideal structure.

Accurate knowledge of the detailed composition

of natural kaolin samples is difficult to obtain. X-

ray diffraction (XRD) is the fundamental method

for identifying the crystalline mineral phases, but it

requires careful work to discriminate between the

various kaolin group minerals and also some

impurity phases, such as illites or smectites.

Furthermore, XRD may not detect the presence of

poorly crystalline minerals such as oxides, even

when present in substantial quantities, since their

diffraction patterns are broad and weak. Such

phases, however, may have a major role in

determining surface properties, such as sorption

processes, because of their large surface areas.

The sensitivity of XRD is such that it often

cannot detect small variations in lattice parameters

that arise from isomorphous substitutions, especially

when they occur at relatively low levels, and their

presence is inferred from a combination of chemical

analyses and measurement of ion exchange capa-

cities. Thus, the presence of even small quantities

of impurity phases can result in misleading

conclusions concerning the compositions of

minerals, and this approach is of little value in

determining the chemical compositions of indivi-

dual components in mixed mineral specimens.

Consequently, a variety of chemical and spectro-

scopic methods has been developed for character-

izing aluminosilicate minerals (e.g. Hawthorne,

1988; Coyne et al., 1998). These include electron

microscopic techniques, transmission electron

microscopy (TEM) and scanning electron micro-

scopy (SEM), and various spectroscopic methods,

such as nuclear magnetic resonance (NMR),

infrared spectroscopy (IR), etc., which have

general applications to mineral characterization. In

addition, with samples such as kaolins that have

relatively low levels of paramagnetic ions, electron

paramagnetic resonance (EPR) spectroscopy (e.g.

Mabbs & Collison, 1992) can provide valuable

information on the nature and distribution of

paramagnetic transition metal ions and on free

radical defects (e.g. Calas, 1988), which might be

expected to play important roles in determining

sorption properties.

Because of its great abundance in the Earth’s

crust, kaolin has a long history of use by man, the

most important of these uses being in the ceramics,

paint and paper industries. However, there is

currently considerable interest in using kaolins as

the source material for the production of higher-

value products for a wide range of industrial uses,

and acid activation of kaolinites has been reported

to yield products for the removal of anions (Gogoi

& Baruah, 2008), cations (Bhattacharyya & Sen

Gupta, 2007) and dyes (Karaoglu et al., 2010) from

waters. Furthermore, there are also several recent

reports of kaolinites being used as the starting

materials in the production of materials with novel

physical and chemical properties (e.g. Frost et al.,

2001a,b, 2004; Temuujin et al., 2001; Belver et al.,

540 N. Worasith et al.

2002; Meenakshi et al., 2008; Vagvolgyi et al.,

2008; Steudel et al., 2009; Panda et al., 2010).

In the edible oil industry, it is often desirable to

decrease the concentrations of pigments, such as

chlorophyll and b-carotene, and acid-activated

bentonites are commonly used for this purpose

(e.g. Christidis et al., 1997; Wu et al., 2006). In a

search for lower cost bleaching clays, we have

investigated the performance of acid-activated

kaolin from the Ranong deposit in southern

Thailand, but it was found to be ineffective for

the removal of pigments from rice bran oil

(unpublished results). However, after modification

by a combination of physical and chemical

treatments, the sorption properties of this kaolin

improved to such an extent (~80% decolourization

of rice bran oil compared to ~82% with a

commercial bleaching clay under the same condi-

tions (Worasith et al., 2011)) that it has potential

use for bleaching vegetable oils. It was to develop a

detailed understanding of kaolin alteration that the

present investigations were undertaken. This was

deemed necessary because different samples of

bleaching clay activated by similar treatments

have been reported to have considerable differences

in performance (Woumfo et al., 2007), and it seems

likely that (perhaps subtle) variations in physical

properties and/or chemical composition can have a

significant impact on the bleaching performance of

clay minerals. Detailed physical and chemical

characterization is especially important for investi-

gations of reactions that occur on the surfaces of

natural mineral samples, since these are influenced

by factors that may vary appreciably from one

deposit to another. Also, natural mineral samples

often contain associated poorly crystalline phases

whose effects on adsorption reactions may be

relatively greater than their concentrations in the

specimens.

The current paper describes the characterization

of material produced by modifications of kaolin

from the Ranong deposit in Thailand, and which

was used in the rice bran oil bleaching studies of

Worasith et al. (2011). A mineral sample from this

deposit, which is thought to have been produced by

the hydrothermal alteration of granite (Kuentag &

Wasuwanich, 1978), has recently been the subject

of basic physical and chemical characterization

(Nuntiya & Prasanphan, 2006), using TEM, X-ray

fluorescence spectroscopy (XRF), and XRD. The

present work builds on these initial measurements,

but with an emphasis on characterizing products

obtained from this kaolin after physical and

chemical modification, and including additional

information that can be obtained with spectroscopic

techniques, such as FTIR and EPR.

MATER IALS AND METHODS

Sulphuric acid (H2SO4) and oxalic acid (H2C2O4)

used for chemical treatments were AR grade and

purchased from Merck, Germany, and Ajax

Finechem, Australia, respectively. The kaolin

sample from Ranong Province in southern

Thailand was supplied by Had Som Pan Co., Ltd.

This natural kaolin, pale yellow in colour, was

initially washed with distilled water and dried

overnight in an oven at 80ºC. Although this

temperature was chosen to minimize the possibility

of damaging halloysite-10 A, it is possible that this

phase was converted to halloysite-7 A, because of

the ease with which water can be removed from this

mineral. This sample was designated K.

Chemical treatments

Acid treatments were performed by adding dried

natural kaolin that had been gently crushed to pass

through a 0.074 mm sieve to 18% or 30% w/w

sulphuric acid; the ratio of clay:acid was 1 g:50 ml.

Samples were refluxed at 90ºC under mechanical

stirring (using an IKA hotplate stirrer model C-MAG

HS7) at about 250 rpm for 4 h. Then the super-

natants were removed and the residues washed with

two litres of distilled water; the washing was

repeating until the pH of the clay suspension was

53. The sample was then dried at 100ºC for 24 h,

this temperature being chosen because of the planned

use of the product as a vegetable oil bleaching agent.

These samples were designated KS18 and KS30. An

additional sample was prepared using 6% w/v oxalic

acid, but it was not fully investigated because it

became apparent early in the work that physical

treatment was an essential aspect of the kaolin

modification process to produce a bleaching clay

(which was the ultimate objective of this work).

Physical treatment

Dried natural kaolin, designated K, initially

crushed to 40.074 mm as for the chemical

treatments described above, was ground using a

planetary ball mill (Retsch model S 100). For this

physical treatment, 20 g of clay were added to a

Characterization of modified kaolinite 541

500 ml grinding jar containing twenty 20 mm

grinding balls (the weight ratio of balls to kaolin

was 30:1); both the pot and milling media were

stainless steel. The clay samples were ground for

1 h at 300 rpm, and the product designated as GK.

Combined physical and chemical treatments

Samples of ground kaolin GK were treated with

18% or 30% w/w sulphuric acid as described for the

preparation of the corresponding unground kaolin

samples, and designated GKS18 and GKS30. A

similar sample was prepared using oxalic acid

instead of sulphuric acid with acid concentrations

6% w/v and designated as GKO6. These acid

concentrations were used for the current investiga-

tions on the basis of their performance as rice bran

oil bleaching agents (Worasith et al., 2011).

Surface modification of the products from

combined physical and chemical treatments

It has been reported that kaolinite surface

properties are influenced by the synthesis pH

(Fialips et al., 2000; Worasith et al., 2011),

indicating the occurrence of protonation/deprotona-

tion reactions at surface oxygen atoms. Therefore,

the influence of pH on the physical/spectroscopic

properties of samples GKS18 and GKO6 was also

investigated by re-suspending the modified kaolin

samples in solutions of sulphuric or oxalic acids at

pH 2.0, 3.0, 3.5, 4.0 and 5.0. After separation by

centrifugation, these kaolin samples were dried at

100ºC for 24 h. These samples were designated

GKS18 P2.0, GKS18 P3.0, GKS18 P3.5, GKS18

P4.0, GKS18 P5.0, GKO6 P2.0, GKO6 P3.0, GKO6

P3.5, GKO6 P4.0 and GKO6 P5.0.

X-ray diffraction studies

A Bruker AXS model D8 Advance X-ray

diffractometer employing Ni-filtered Cu-Ka radia-

tion was used to investigate the mineralogy of the

natural kaolin sample and its modified products.

Kaolin powder was placed in a flat holder, with

20 mm diameter and 15620 mm irradiated area,

and diffraction patterns were then collected in the

range of 2y = 5º to 60º at a scanning speed 1º per

2y min�1, and working at 40 kV and 30 mA.

Mineral components in the X-ray diffractograms

were identified by comparison with standards in the

JCPDS Powder Diffraction File.

Additional measurements were made with

oriented samples of the natural kaolin in an

attempt to discriminate between halloysite-7 A,

kaolinite and illite components. These oriented

samples were subjected to four treatments: air

drying, glycolation, intercalation with formamide,

and heating to 550ºC for 1 h. The structural order

of natural kaolinite was estimated using the

Hinckley index (HI), and the crystallinities of the

kaolin minerals in the various treated products were

also estimated from the widths and heights of peaks

in the XRD patterns, since peak widths are

inversely related to the crystallinity of the sample.

X-ray fluorescence

A Bruker AXS model S4 Pioneer XRF spectro-

meter with a scintillator detector was used to

determine the chemical compositions of the kaolin

samples. Wax was added to the clay samples as a

binder before fusing into pellets (weight ratio of

clay:wax about 7:3), and compositions are

expressed as relative concentrations in the form of

oxides.

Scanning electron microscopy

The surface morphology of the kaolin samples

that had been dried at 80ºC and coated with gold to

enhance conductivity was investigated using a

scanning electron microscope (JEOL model JSM-

5410 LV) with an accelerating voltage of 15 kV

and a vacuum of 10�5 Pa. The samples were

inserted into the SEM chamber, transferred to the

path of the electron beam, and then scanned

automatically. Various magnifications were used

to compare the textures and shapes of modified

kaolins before and after treatments but only those of

620,000 are presented. In addition, chemical

analyses were performed by using a link to an

energy dispersive X-ray analysis system (EDX)

which incorporated ISIS series 300 software.

FTIR

Attenuated Total Reflectance Fourier Transform

Infrared Spectra (ATR-FTIR) were recorded on a

Bruker Tensor 27 FTIR spectrometer incorporating

a 1.8 mm Ge crystal. A sample of the clay powder

was placed on the Ge crystal and 1024 scans were

acquired for each spectrum at a resolution of

1 cm�1 in the mid-IR range (4,000�600 cm�1).


EPR spectroscopy

EPR spectra were recorded using various Bruker

and JEOL spectrometers operating at X-band

frequencies and using Gunn diodes as microwave

sources. All samples were studied at room

temperature (~22ºC) using 100 kHz modulation

frequency with other acquisition parameters deter-

mined by the line widths and saturation properties

of the component signals. Spectra were first

recorded as first derivatives of the microwave

absorption over the scan range 0�500 mT using

10 mW microwave power and 1 mT modulation

amplitude to obtain a general overview of the

signals. Separate spectra were then recorded using

parameters that were optimized for the individual

signals. Free radical components were recorded as

both first and second derivatives using microwave

powers in the range 2�10 mW and modulation

amplitudes in the range 0.2�1.0 mT over field

scans ranging between 5 and 40 mT. Receiver gain,

number of scans, conversion time and time constant

were adjusted individually for each spectrum

depending on the intensity of the signal being

characterized. Diphenylpicrylhydrazyl (DPPH) (g =

2.0036) was used as an external standard for the

determination of g-values.

RESULTS

X-ray diffraction

Composition of the original kaolin, K. The XRD

pattern from the original Ranong kaolin K is shown

in Fig. 1. As reported by Kuentag (2001) and

Nuntiya & Prasanphan (2006), the majority of the

peaks can be accounted for by the kaolin minerals;

the strong peaks at 7.14, 3.57 and 1.49A correspond

to the kaolinite d001, d002 and d060 reflections and

the peaks at 4.45, 2.56, 2.49, 2.34, 1.99, 1.66, and

1.49 are all consistent with kaolinite. Positive

identification of halloysite from these data is

difficult, because the main peaks from halloysite-

10 A at 10A and halloysite-7 A at 4.42 A both

overlap with peaks from illite, and the peak at

4.34 A can correspond to either halloysite or

kaolinite (Brindley & Brown, 1980). Illite, which

is a common impurity in kaolin-group minerals, is

present in appreciable quantities in the Ranong

kaolin, as seen by the presence of major peaks at

9.92 and 2.56 A in addition to those mentioned

above. Other peaks consistent with illite are

observed at 4.99 and 2.99 A. The presence of

quartz is seen by its principal reflection at 3.34 A,

and associated minor peaks at 4.25, 2.29, 1.82, and

1.54 A. In addition, weak peaks at 4.18 and 3.25 A

could correspond to the main reflections from

goethite and microcline, respectively, so these

minerals may be present as minor impurities, but

the absence of other diffraction peaks prevents a

positive identification.

The clay mineral groups were distinguished by

the behaviour of their characteristic basal reflec-

tions in response to various chemical and heat

treatments (Fig. 2). The insensitivity of the peaks at

~10.0 A to ethylene glycol, and heating to 550ºC

mean that they correspond at least mainly to the

d001 reflections from illite, and that halloysite-10 A

FIG. 1. Powder X-ray diffraction pattern of untreated kaolin, sample K, and the assignment to its component

minerals. K = kaolinite, H = halloysite, Q = quartz, I = illite, G = goethite and M = microcline


makes only a minor contribution (if any) to their

intensity. Halloysite-7 A and kaolinite are unaf-

fected by glycolation, but become amorphous to

X-rays after heating to 550ºC, and Fig. 2 shows the

conversion of peaks at d ~7 A and ~3.5 A to an

amorphous phase at ~12.38º2y at this temperature.

Finally, since intercalation with formamide results

in an increase in the d001 reflection of halloysite-7

A to ~10.0 A, the peak at 7.17 A in the original

kaolin corresponds primarily to halloysite. These

results are thus qualitatively similar to those

reported by Nuntiya & Prasanphan (2006) for

their kaolin sample from Ranong, but the sample

in the present work contained smaller amounts of

impurity phases. The Hinkley index (HI) (Hinckley,

1963) of about 0.32 suggests that the kaolinite is of

only moderate crystallinity, whereas both the FTIR

and EPR results (see below) indicate that the

mineral is well ordered. However, Brindley et al.

(1986) report that the HI is directly correlated with

kaolinite Fe content, so its relationship to crystal-

linity may be complicated by isomorphous substitu-

tions. Furthermore, the presence of overlapping

peaks from illite and quartz with those of kaolinite

used for measuring the HI complicates its accuracy

for kaolinite crystallinity determination in the

present sample.

Influence of individual chemical and physical

treatments on the kaolin crystallinity. XRD patterns

for Ranong kaolin before and after grinding or

sulphuric acid treatment are shown in Fig. 3. They

show that the mineral is resistant to the acid

treatment (Fig. 3b,c) since the d001 reflection at

7.14 A, the d002 reflection at 3.57 A, and the d060reflection at 1.48 A remained sharp, even under

strongly acidic conditions (Table 1). However,

grinding the clay resulted in a major decrease in

the relative intensities of the peaks from the kaolin

mineral and illite phases (Fig. 3d), consistent with

appreciable structural damage to these minerals.

Furthermore, there was also an increase in the

background in the 20�30º2y range which is typical

of the presence of amorphous phases.

Using the Scherrer formula (Langford & Wilson,

1978) the FWHM data for d001 indicates that the

mean number of layers in a kaolinite stack is ~90

for samples K, KS18 and KS30. This value

decreases to ~70 (i.e. a reduction of ~25%) after

grinding, but then increases again to ~90 for

GKS18 and GKO6. Grinding had very little effect

on the d060 peak and samples K and GK both have

~380 planes in a stack. However, it is of great

importance after acid attack, with the number of

layers in the b dimension decreasing to ~280 for

KS18 and KS30, and ~190 for GKS18 and GKO6.

Effects of combined physical and chemical

treatments. Further changes in the XRD patterns

were observed after treating the ground sample with

FIG. 2. Powder X-ray diffraction patterns of oriented kaolin, sample K, as a result of (a) air-drying,

(b) glycolation, (c) intercalation with formamide, and (d) heating to 550ºC.


either sulphuric or oxalic acids (Fig. 3e,f); apart

from the peak at ~26.74º2y, which probably

corresponds to quartz, there is an appreciable

reduction in intensity in all regions. Al compounds

resulting from grinding-induced kaolin breakdown

were removed by the acid treatments as indicated

by the XRF results, presumably as a result of the

formation of soluble Al salts. Although the peaks

corresponding to the kaolinite/halloysite minerals

were very weak in these XRD spectra, they were

still quite sharp (Table 1), suggesting that the

residual fraction of these minerals is relatively

unaltered.

X-ray fluorescence spectra

Table 2 shows the major elements determined by

XRF analyses of the original kaolin, the original

and ground kaolins treated with 18% w/w sulphuric

acid and the ground kaolin treated with 6% w/v

oxalic acid. For comparison, this table also shows

the results from analysis of a second set of samples

FIG. 3. Powder X-ray diffraction patterns for Ranong kaolin, sample K, and modified samples, KS18, KS30, GK,

GKS18 and GKO6. Peaks corresponding to the identified mineral phases are shown on the diffractogram of the

untreated sample.

TABLE 1. Variation of positions and line-widths (º2y) of the main kaolinite XRD peaks in the various Ranong

kaolin samples.

Sample d001 FWHM d002 FWHM d060 FWHM

KKS18KS30GKGKS18GKO6

7.097.117.127.157.157.14

0.270.290.270.370.270.28

3.563.563.573.573.573.57

0.290.320.280.410.400.31

1.491.491.491.491.491.49

0.320.440.420.310.600.69


obtained from this deposit. Si and Al are the

dominant elements in the original kaolin, along

with minor amounts of K and Fe; the concentrations

of Mn, Ti, Mg and Ca are all small. Decreases in

the Al:Si and Fe:Si ratios were observed as a result

of acid treatment of the unground clay, but much

larger decreases were observed after acid leaching

the ground kaolin. Since the main minerals in the

sample are halloysite, kaolinite, illite and quartz

with minor amounts of mica and orthoclase and

iron oxides, it is possible to use the data in Table 2

to get a rough idea of the composition of the

samples. By using the theoretical chemical compo-

sitions for these phases, and attributing all the Al to

kaolinite/halloysite, we find that this component

accounts for ~85% of the original kaolin sample

and this decreases to ~40% in GKS. In ground and

acid-treated products, the main phase is probably

amorphous silica (450%) which is consistent with

the strong development of Si�O bands in the FTIR

spectra (see below).

Scanning electron microscopy

Results from SEM/EDX studies of the Ranong

kaolin and its modified products are shown in

Fig. 4. The original clay sample K shows the

presence of both platy and tubular shapes that are

characteristic of kaolinite and halloysite respec-

tively. However, these structures were largely

destroyed by grinding (sample GK). Hot acid

treatments of the ground sample then resulted in

the formation of products with globular morphology

(samples GKS18 and GKO6). Chemical analysis of

sample surfaces performed using EDX indicate that

Si, Al, O are the major elements, and K and Fe are

minor components, as was found with the XRF

analyses of the bulk samples. Also as observed with

XRF, acid treatment of the ground samples resulted

in an appreciable decrease in the Al content.

FTIR spectroscopy

The main absorptions in the FTIR spectra of the

kaolin samples are shown in Fig. 5. In the original

kaolin sample, K, the O�H stretching vibrations

consist of two well defined peaks at 3694 and

3621 cm�1 flanking two weaker peaks at 3669 and

3652 cm�1. This spectrum is typical of kaolinite (e.g.

van der Marel & Krohmer, 1969, report bands at

3693, 3668, 3652 and 3620 cm�1). In contrast to the

HI from the XRD results, the good resolution of

these peaks suggests that the mineral is quite well

ordered, and provides support for the analysis of the

d001 and d060 peaks in the XRD results. As proposed

by Farmer (1974, 1998), and subsequently confirmed

by theoretical calculations (Balan et al., 2005), the

band at the lowest wave-number corresponds to

vibrations of the inner OH group, that at the highest

wave-number to the in-phase motion of the three

TABLE 2. Chemical composition (wt.%) of Ranong kaolin samples determined by X-ray fluorescence.

Oxide ————————————— Sample —————————————(wt.%) K KS18 GKS18 GKO6

SiO2 43.30 (46.70) 55.35 (51.40) 74.80 (72.77) 70.70 (67.84)Al2O3 37.20 (37.90) 29.63 (33.63) 16.80 (13.50) 20.40 (16.13)K2O 1.36 (1.23) 1.46 (1.46) 1.81 (1.48) 2.01 (1.51)Fe2O3 0.66 (0.54) 0.49 (0.41) 0.44 (0.29) 0.50 (0.27)MnO 0.05 (0.04) 0.04 (0.03) 0.04 (0.03) 0.04 (0.03)TiO2 0.03 (0.03) 0.02 (0.02) 0.03 (0.03) 0.04 (0.03)CaO 0.01 (0.01) n.d. (n.d.) n.d. (n.d.) 0.03 (n.d.)MgO 0.04 (0.03) 0.03 (0.03) n.d. (0.03) n.d. (0.02)Na2O 0.04 (0.03) 0.03 (0.03) 0.04 (0.04) n.d. (0.04)LOI 17.31 (13.49) 12.94 (12.90) 6.08 (11.76) 6.08 (14.09)Al:Sia 1.01 (0.95) 0.63 (0.77) 0.26 (0.22) 0.34 (0.28)Fe:Sia 1.14% (0.85%) 0.66% (0.60%) 0.44% (0.27%) 0.53% (0.30%)

The numbers in brackets correspond to analyses of a second set of samples obtained from the same deposita atomic ratiosn.d. = not detected


surface OH groups, and those at intermediate

positions to out-of-phase motion of these groups.

The bands at ~935 and 913 cm�1 correspond to

the inner surface and inner Al�OH�Al bending

vibrations (Farmer & Russell, 1964) and those at

1116, 1030 and 1007 cm�1 to Si�O stretching

vibrations in kaolinite/halloysite, but there was only

a hint of a shoulder at ~1100 cm�1 where the Si�O

FIG. 4. Scanning electron microscope micrographs and representative energy dispersive X-ray analysis spectra

from small areas on the surface of Ranong kaolin samples K, GK, GKS18, and GKO6. Scale bars = 1 mm.


(apical) band is expected (e.g. Frost et al., 2002).

The band at 1030 cm�1 could also contain a

contribution from muscovite, which was detected

by Nuntiya & Prasanphan (2006) as an impurity in

their Ranong kaolin, but was not seen in the XRD

patterns from our present sample. The peak at

793 cm�1 is probably associated with quartz, which

was identified as an impurity in the XRD traces of

both Nuntiya & Prasanphan (2006) and the present

work. There was no evidence for a discrete

absorption at ~3425 cm�1, the frequency at which

the O�H stretching vibration occurs in smectites

which are also common impurities in kaolin

minerals. The bands at 751 and 687 cm�1

correspond to those assigned to Si�O stretching

vibrations (Ekosse, 2005).

Little change was observed in the spectra as a

result of sulphuric acid treatment, indicating that

the kaolin minerals are resistant to acid treatment,

although weak new peaks were seen at ~1220 and

828 cm�1 and a shoulder at ~1075 cm�1; since

peaks were observed in these positions, but with

higher intensity in the ground samples after acid

treatments, they indicate that the acid treatment did

induce some modification of the unground kaolin.

There was also some decrease in the relative

intensities of the bands at ~935 and 913 cm�1

from Al�OH�Al bending vibrations. In contrast,

FIG. 5. Fourier transform infrared spectra of the main absorption regions in the FTIR spectra of modified Ranong

kaolin samples.


grinding reduced the intensities of the kaolinite

O�H stretching vibration bands by ~65%, indi-

cating that physical modification of the mineral was

accompanied by extensive dehydroxylation as

observed by Miller & Oulton (1970). However,

the Al�OH�Al bending vibration bands were

reduced by a smaller amount, and the O�H and

Al�OH�Al peaks that remained after grinding

were unshifted. Thus, they indicate that the

kaolinite that remained was essentially unaltered

from that in the original sample.

Frost et al. (2001a,b) have also reported IR

results for a mechanochemically-treated kaolinite,

in which there were both similarities and differ-

ences from the measurements reported above. Both

our measurements and those of Frost et al. showed

decreases in the intensity of the OH stretching

vibrations (3695 and 3619 cm�1) and deformation

modes (935 and 913 cm�1), but Frost et al. also

observed an increase in the bending mode at

1650 cm�1 from water bound to the surface of the

modified kaolinite; this was extremely weak in the

present measurements (probably because of the

drying of the sample). Also, whereas we observed

little change in the Si�O stretching vibrations,

Frost et al. reported a significant reduction in

intensity that was accompanied by an increase in a

new band at 1113 cm�1, which was assigned to the

surface of the newly synthesized product. It should

also be noted that the Si�O stretching vibrations

reported by Frost et al. (2001b) were at 1034 and

1056 cm�1, which are considerably higher than the

values of 1011 and 1030 cm�1 that were observed

in the present work.

Acid treatment of the ground sample did not

produce any further changes in the OH stretching

frequencies, and the differences in the relative

intensities of the peaks are not considered

significant because of the low overall intensity in

this region of the spectra. There were, however,

appreciable changes in other regions of the spectra.

The intensities of the Al�OH�Al bending vibra-

tions were reduced, and are roughly correlated with

the reduction in the Al content (Table 2). There

were also significant differences in the relative

intensities of new bands in the sulphuric and oxalic

acid-treated samples, although their positions were

similar (Table 3). Intense broad bands generated at

~1205 cm�1 and 1067 or 1077 cm�1 (depending on

the acid used) probably correspond to Si�Ostretching bands associated with newly synthesized

material, since weak bands in similar positions were

TABLE 3. FTIR frequencies (cm�1) and assignments in Ranong kaolin samples.

NK KS18 KS30 GK GKS18 GKO6 Assignment

3694366936523621

~1650~1560

1116

10301007

935913

793751687648

3692367436493621

1654

~12201116~107510301007

935913828796751687647

3692367236503621

1650

~12201116~107510301007

935913828794751687647

3695367036463623

16501632

1116

10331011

935913

794751687647

36923669364836213397

(v broad)1646163312051116

106710361011940916828796754696649

36923669364836213346

(v broad)1650163412051116107710351013

937914828796754696642

O�H

O�H

Surface water

Si�OSi�O

Al�OH�Al

QuartzSi�O


observed in the spectra of the unground clay after

acid treatment. These are significantly different

from the results reported by Frost et al. (2001), who

described the formation of a new band at

1113 cm�1; the narrow peak at 1117 cm�1 in the

natural kaolin (and at 1121 cm�1 in the ground

sample) in the present work, was lost completely in

the acid treatment experiments of Frost et al.

(2001a,b, 2002, 2003, 2004). However these authors

found that the presence of quartz increased

appreciably the rate of grinding-induced breakdown

of the kaolinite, and this may also be a factor in

determining the physical characteristics of the

breakdown products.

There was a reduction in the relative intensities

of the Si�O deformation bands, which occurred at

1007 and 1030 cm�1 in the original kaolin sample,

and these were also shifted to slightly higher wave-

numbers after grinding and acid treatment, but

overall the changes in this region of the spectrum

were relatively minor. Larger changes were

observed with the Al�OH�Al bending vibrations

at 935 and 913 cm�1; although their intensities

were reduced after grinding, much greater addi-

tional reductions were apparent after the acid

treatment, consistent with dissolution of the

octahedral sheet which contains most of the O�Hgroups.

EPR spectroscopy

The wide scan EPR spectra of the Ranong kaolin

before and after grinding are shown in Fig. 6. As is

common for clay mineral samples with a relatively

low overall content of paramagnetic ions, these

spectra contain signals from four different sources;

(i) a group of peaks in the field range 50�250 mT

which originate from isolated Fe3+ ions, (ii) a broad

feature centred on 350 mT from magnetically

interacting ions, probably also Fe3+, (iii) a sextet

component with peak separation ~9.5 mT centred

on 350 mT from Mn (55Mn, I = 5/2), and (iv) a

narrow feature centred on 350 mT from free radical

defect centres in the mineral structure. Similar

features have been reported in previous studies of

kaolin samples (e.g. Goodman & Hall, 1994), and

are discussed individually below.

Only the low field signals were affected by

physical treatment of the original kaolin. These

signals consist of peaks at ~80, ~140, ~195 mT and

a cross-over at 163 mT; the last of these signals

increased and the others decreased as a result of

FIG. 6. Wide scan electron paramagnetic resonance spectra of unground and ground Ranong kaolin samples

before and after treatment with 18% w/w sulphuric or 6% w/v oxalic acid.


grinding. There was, however, little further change

in these signals when the ground samples were

extracted with sulphuric or oxalic acids. In contrast

to the low field Fe3+ signals, the Mn signal showed

little change as a result of grinding alone, but its

shape was modified as a result of acid treatment.

This resulted in a sharpening, but little change in

the separation of the Mn lines, and the effect was

greater with the ground than with the unground

kaolin; it was also much more pronounced with

oxalic than with sulphuric acid treatment.

The free radical signal in the original kaolin is

shown in Fig. 7. The main feature is highly

anisotropic with g// ~2.050 and g\ ~2.008, and is

similar to the A-centre described by Angel et al.

(1974). It corresponds to electron holes trapped on

oxygen atoms (Cuttler, 1981), and has been

associated with radiation-induced defects in

natural kaolinites (Clozel et al., 1994). The g\region in Fig. 7, however, clearly contains more

than one feature, and is consistent with the Q-band

EPR results of Clozel et al. (1994), who described

the presence of three distinct free radical centres in

kaolinite that had been subjected to g-irradiation.Furthermore several (at least six) peaks are evident

between the g// and g\ features of the second

derivative spectrum and correspond to hyperfine

structure from a radical centre involving interaction

FIG. 7. Expansion of the free radical region of the 1st and 2nd derivative electron paramagnetic resonance spectra

from the Ranong kaolin samples K, KS18, GK, GKS18 and GKO6.


of unpaired electrons with 27Al (I = 5/2) nuclei.

This is consistent with the B-centre described by

Clozel et al. (1995) as corresponding to O� centre

linking two Al atoms in octahedral sites, though

these authors were not able to determine whether

this was at a surface position, or associated with an

O atom linking the octahedral and tetrahedral

sheets. It should also be noted that the present

results cannot distinguish between a B-centre and

one with a single 27Al nucleus interacting with the

unpaired electron.

There was little influence of sulphuric acid

extraction on the A-centre free radical signal, but

there was an improvement in the resolution of the27Al hyperfine structure associated with the

B-centre. Since the most likely effect of the acid

treatments is removal of surface-adsorbed species,

this result suggests that the B-centre is associated

(at least partially) with a surface O atom on the

octahedral sheet. Furthermore, the absence of any1H hyperfine structure indicates that this oxygen is

not protonated.

Grinding resulted in a loss of the 27Al hyperfine

structure from the B-centre, although the signal from

the A-centre was little changed, either by grinding

alone, or subsequent acid treatment. However, as

seen by the variations in features labelled A and B

in the g\ region of the second derivative recording,

the combined treatments resulted in a change in the

relative amounts of the components that contributed

to this absorption. However, at least one of these

components probably corresponds to a defect centre

in quartz for which radiation-induced defects are

well known (e.g. Weeks, 1956).

There was little change in the low field Fe3+

signals after these kaolin samples had been treated

by re-suspension in either sulphuric or oxalic acid

at different pH values, but the widths of the Mn

FIG. 8. Effect of adjusting the surface pH on the electron paramagnetic resonance spectra from Ranong kaolin that

had been ground for 1 h in a ball mill and treated with either 18% w/w sulphuric acid or 6% w/v oxalic acid at

90ºC for 4 h, before being suspended in solutions with different pH values, then separating by centrifugation and

drying at 100ºC.


peaks decreased with increasing pH; also this effect

was far more pronounced with oxalic acid than with

sulphuric acid (Fig. 8). This result suggests that this

treatment is not simply a protonation/deprotonation

reaction, and that the chelating ability of oxalic acid

may also be an important factor. In the sample

treated with sulphuric acid at pH 3.5, a new feature

with g ~2.5 was also generated. Since it was

reproduced in replicate experiments, it corresponds

to a genuine product, but its interpretation is

unclear. However, under certain conditions, Fe3+

can produce an isotropic spectrum in this region

(Golding et al., 1977; Mabbs & Collison, 1992).

D I SCUSS ION

Although the principal objective of this paper was

the characterization of the products formed by

combined physical and chemical treatments of

Ranong kaolin, it was deemed necessary first of

all to investigate the unaltered material and the

products of individual chemical and physical

treatments. This is because the reaction pathways

of the mineral are likely to be influenced by both its

mineral components and its chemical composition.

The XRD pattern of the untreated kaolin sample

(Fig. 1) shows that it is composed primarily of

kaolinite and halloysite with some quartz and illite

or mica, along with possible minor amounts of

microcline and goethite, the latter probably being

responsible for the yellow colouration of the sample.

The methods that can be used to distinguish between

kaolinite and halloysite have been reviewed by

Joussein et al. (2005) and, by preparing oriented

samples intercalated with formamide, we were able

to show that halloysite was the dominant mineral in

the unaltered kaolin sample. Thus the present result

is qualitatively similar to that of Nuntiya &

Prasanphan (2006), although these latter authors

observed the additional presence of muscovite and

appreciably greater amounts of quartz. However,

local variations in the composition of kaolin deposits

are common as a result of variations in the facies of

the parent rock, the degree of weathering, and

variations in micro-environmental condition during

kaolin formation (Duzgoren-Aydin et al., 2002).

Systematic changes in composition with particle size

have been reported by Lombardi et al. (1987) for

kaolin deposits in Europe and North America, for

example. Thus it is not surprising that there should

be some differences between the samples of the

kaolin used in the present work and that investigated

by Nuntiya & Prasanphan (2006), even though they

were obtained from the same general geographical

region and are described under the same name.

The XRF results for the kaolin samples used in

the present work also indicate some differences in

chemical composition compared to the sample

investigated by Nuntiya & Prasanphan (2006);

although there were only minor differences in the

SiO2 and Al2O3 concentrations, and the MnO

concentrations are virtually identical, the Fe2O3

content is much smaller in the present sample.

Indeed, our analysis of a second set of samples

from this deposit (Table 2) showed greater varia-

tions for the Fe2O3 contents than for any other

elements. Thus there are probably considerable

variations in the concentrations of associated iron

oxide phases in different samples from the deposit.

This is a factor that must be considered when

developing novel uses for the mineral, because such

phases may make disproportionately large contribu-

tions to the sorption properties (as a result of their

often poorly crystalline nature); it should be noted

that deferration treatments may have an adverse

effect on the sorption properties. Furthermore, the

appreciable differences between the FTIR spectro-

scopic results from the present investigations and

those of the mechanochemically-treated kaolin of

Frost et al. (2001a,b) are also an indication of

different chemical behaviour of samples from

different origins; these could be the consequence

of different chemical compositions or structural

order, or simply the amount of quartz in the mineral

sample (Mako et al., 2001). Thus careful attention

must be paid to physical characterization when

developing samples for practical uses, such as

adsorbents in the food industry, and caution

should be exercised in extrapolating properties

from one specimen to another without detailed

physical characterization.

The tubular morphology observed in the SEM

result from the unaltered sample (Fig. 4) is

indicative of the presence of halloysite. However,

halloysite-10 A made only small contributions to

the XRD pattern, but measurements with inter-

calated formamide (Fig. 2) indicated that the

reflection at around 7.1�7.2 A corresponded

primarily to halloysite-7 A. The absence of any

tubular structures in the SEM images and the large

weakening of the peaks associated with kaolinite/

halloysite in the XRD patterns from the ground

kaolin samples (Fig. 3) indicate that the habits of

these minerals are largely destroyed by the physical


treatment. The SEM images show that grinding

produced a decrease in particle size (Fig. 4).

Increases in specific surface area and total pore

volume have been reported (Worasith et al., 2011),

and these observations are all consistent with the

formation of poorly crystalline or amorphous

phases. The FTIR spectroscopic results indicate

that �OH groups were lost as a result of grinding,

presumably as a result of dehydroxylation caused

by heat generated during the grinding process.

Since these groups are associated with the surface

of the octahedral sheet, there would inevitably be a

major change in the coordination environment of

the Al and a decrease in the number of hydrogen

bonds that hold together the individual alumino-

silicate layers. This result is also consistent with the

NMR study of Temuujin, et al. (2001), who

reported destruction of the octahedral layers of

kaolinite by prolonged grinding. However, the

present FTIR results indicated that the small

amount of residual kaolinite was largely unaffected

by the combined physical and chemical treatments.

The XRD, FTIR and EPR results all show only a

small effect of hot sulphuric acid treatment on the

unground kaolin (Figs 3, 5 and 6) and indicate that

the kaolinite and halloysite components have a high

degree of resistance to attack by mineral acids,

presumably because of the strong hydrogen bonds

in their structures. This result should be contrasted

with that of Steudel et al. (2009) who found that

activation of kaolinites by sulphuric acid at 80ºC

produced layered materials with a simple chemical

composition (silica) and large surface area as a

result of successive dissolution of octahedral sheets

by edge attack. However, in comparing the

behaviour of di- and trioctahedral minerals,

Steudel et al. (2009) also reported that this reaction

is aided by the presence of structural Mg or Fe,

which were removed faster than Al; thus the very

small Mg content of the Ranong kaolin (Table 2)

may be a reason for its resistance to acid.

Nevertheless, the kaolin was not completely inert

to acid attack. The XRF results (Table 2) show

respective reductions in the Al:Si and Fe:Si atomic

ratios from 1.01 to 0.63 and 1.14% to 0.66% for

treatment of the unground kaolin with 18%

sulphuric acid, and there was a reduction in the

relative intensity of the broad g = 2.0 component in

the EPR spectra, which is consistent with the

removal of associated poorly crystalline iron oxide

phases. Treatment of the original kaolin sample

with hot sulphuric acid had little or no influence on

the low field Fe3+ EPR signal (Fig. 6), indicating

that any removal of (surface-bound) material by

acid treatment had no significant effect on the local

environment of structural Fe. Also, although there

was a reduction in the relative intensity of the broad

g = 2.0 component, it retained appreciable intensity

after acid treatment, indicating the presence of

magnetically interacting ions in the aluminosilicate

structure; this could correspond either to illite, or an

Fe oxide phase. Finally, the FTIR spectra show the

formation of new features at ~1220, 1075 and 828

cm�1, which are probably associated with an acid-

modified component, and there was some improve-

ment in the resolution of the B-centre radical, which

suggests that this is associated with oxygen atoms

on the surface of the octahedral sheet.

Hot acid treatments of the ground kaolin sample

resulted in large increases in the specific surface

area and the total pore volume (Worasith et al.,

2011), and this was accompanied by a major

decrease in the Al:Si ratios to 0.22 and 0.29 for

the 18% w/w sulphuric acid and 6% w/v oxalic acid

treatments, respectively (see Table 2). This was also

accompanied by reductions in the Fe:Si ratios to

0.6% and 0.7% for the corresponding treatments

with 18% w/w sulphuric acid and 6% w/v oxalic

acid. Thus, after breakdown of the aluminosilicate

structures by physical treatment, the components of

the original octahedral sheet were selectively

extracted by acid, leaving a Si-rich phase derived

from the tetrahedral sheet. This latter phase was

characterized by a new, but as yet unidentified,

peak in the XRD diffractogram, and major new

Si�O stretching peaks in the FTIR spectra.

The EPR spectra provide an insight into the

structural changes that affect the Fe, Mn and free

radical components, all of which make relatively

minor contributions to the overall composition of

the kaolin, but which nevertheless represent

potential sites for chemical reactions. In the EPR

spectra of kaolinites, low field features from Fe3+

are common and correspond to two distinct

environments for the Fe3+ (Mestdagh et al., 1980;

Gaite et al., 1997). Balan et al. (1999) have shown

that both signals correspond to substitution of Fe3+

for Al in octahedral sites within the kaolinite

structure. The signal designated Fe(I) by Balan et

al. (1999), which corresponds to the central feature

at 163 mT in Fig. 6, has been assigned to a

disordered mineral structure. The remaining low-

field peaks in Fig. 6 are all derived from the signal

designated Fe(II) by Balan et al. (1999), and


correspond to Fe3+ in a well-ordered crystalline

phase with a small number of stacking defects in its

structure.

It is clear that grinding has a major influence on

the relative proportions of iron in the Fe(I) and Fe(II)sites. However, on the basis of the results in the

present work, it seems likely that the increased

fraction of the Fe3+ in the Fe(I) sites is the

consequence of dehydroxylation of the octahedral

sheet and not a change in the long-range order of

the kaolin-group minerals, which is the usual

assignment for such a component. The fraction of

the Fe3+ that remains in the Fe(II) sites is therefore

likely to correspond to the unaltered kaolinite

component that was seen in the XRD and FTIR

spectra. In contrast to the signals from the structural

Fe3+, the broad absorptions from magnetically-

interacting ions, along with those from the Mn

and free radical components, are not affected

significantly by grinding. The absence of a major

increase in the free radical signal as a result of the

grinding is somewhat surprising, since such

treatment is well-known to generate free radical

defects in mineral structures (e.g. Vallyathan et al.,

1988).

Although acid treatment of the original kaolin

had little effect on the distribution of Fe between

the Fe(I) and Fe(II) sites (see above), treatment of

the ground kaolin with either hot sulphuric or oxalic

acid (Fig. 6) increased the Fe(II) contribution to the

total Fe3+ signal, although it still remained much

lower than in the unground specimens. This

provides further evidence to support the conclusion

that these acid treatments result in selective

dissolution of the octahedral sheet from those

mineral components that were altered (dehydroxy-

lated) during the grinding process and as a result

the concentrations of the unaltered fractions were

increased by subsequent acid treatments.

All of the spectra contain features attributable to

interactions of unpaired electrons with 55Mn nuclei,

identifiable by the characteristic sextet structure

from the I = 5/2 nucleus. Although not a common

substituent of kaolinites, there have been previous

reports of EPR spectra of Mn in natural kaolin

specimens, but it is usually assigned either to the

presence of impurity phases containing Mn2+ (e.g.

Sengupta et al., 2006), or Mn2+ adsorbed on

kaolinite surfaces (e.g. McBride et al., 1975). The55Mn hyperfine splitting of ~9.3 mT in the original

kaolin is typical of that for Mn2+ octahedrally

coordinated to oxygen atoms, as for example in

Mn(H2O)62+ on the exchange sites in clay minerals

(McBride et al., 1975) or in oxide crystals, such as

CaO or CdO (Title, 1963). Unlike the low-field

Fe3+ components, little grinding-related change was

seen in the Mn signal, which would appear to

exclude their presence in the octahedral sheet,

although the 55Mn hyperfine splitting increased

slightly to ~9.5 mT. However, the shape of the Mn

signal was very sensitive to acid treatment and the

chemical nature of the acid used, and both the

unground and ground samples indicate that the Mn

is in sites that are accessible to acids. However, the

lack of any significant decrease in overall Mn

signal intensity and the similar concentrations in the

XRF measurements before and after acid treatment

would appear to exclude surface adsorption of Mn2+

as the explanation for this signal; thus the Mn is in

structural sites. Association of the Mn with an

impurity phase cannot be specifically excluded,

although XRD provided no evidence for the

presence of significant amounts of likely candidate

phases, such as muscovite which was seen in the

Ranong kaolin sample investigated by Nuntiya &

Prasanphan (2006). Furthermore, the relatively

sharp peaks indicate that any impurity phase

would need to have the Mn in a magnetically

dilute form (i.e. low concentration), and located

close to the surface (but not exchangeable) in order

to explain the effect of acid treatment on its spectral

shape.

Further changes in the Mn signal were observed

after the modified kaolin had been re-suspended in

solutions of different pH values. This adds further

support to the conclusion that the Mn is located in

positions in the lattice that are different from those

that are occupied by the Fe3+ ions, which were

unaffected by such treatment. Thus, by a process of

elimination we propose that the Mn is located in the

tetrahedral sheet and therefore is probably in the

Mn4+ oxidation state. The interpretation of a Mn

signal in the EPR spectra of aluminosilicate

minerals as corresponding to Mn4+ in tetrahedral

sites is unusual, and such signals are commonly

assigned to Mn2+ in octahedral sites (e.g. Martin et

al., 1999). However, the ionic radii for Mn4+ and

Al3+ are essentially identical (0.39 A for tetrahedral

coordination), somewhat larger than that of Si4+

(0.26 A) but smaller than that of tetrahedral Fe3+

(0.49 A) (Shannon, 1976). Thus, on size considera-

tions, Mn4+ can be accommodated in any site that

can accommodate Al3+, and on charge considera-

tions Mn4+ can replace Si4+ without creating any


charge imbalance. Furthermore, Mn4+ is more

easily accommodated in tetrahedral sites than

Fe3+, yet tetrahedral Fe3+ is an accepted substituent

in several aluminosilicate minerals (e.g. Goodman

et al., 1976; Rancourt et al., 1994)). In contrast, the

ionic radius of Mn2+ in octahedral coordination is

0.83 A, considerably larger than the values of

0.535 A and 0.645 A for octahedral Al3+ and Fe3+

respectively. It is also appreciably larger than the

value of 0.72 A for Mg2+ in octahedral coordina-

tion; thus there are difficulties in accommodating

the larger Mn2+ ion in aluminosilicate mineral

structures. However, there is a further complicating

factor of the possible presence of Mn3+, which is

not easily detected by EPR (see for example Mabbs

& Collison, 1992), especially at trace concentra-

tions, but whose ionic radius is very close to that of

Fe3+.

One may assume that the kaolinite structure

might contain trioctahedral regions in which Mn2+

could be located, but this interpretation is not

consistent with the presence of a well-resolved55Mn hyperfine structure. EPR spectra from

samples with clusters of Mn2+ ions would indeed

induce a broad peak as a result of exchange

interactions and would, therefore, contribute to the

broad g = 2.0 signal. It also seems unreasonable to

propose that Mn2+ is a minor component in

trioctahedral clusters in which the major divalent

ions are diamagnetic, because the analytical results

do not indicate the presence of sufficient quantities

of such ions that would be necessary for such a

structure.

This hypothesis of the location of Mn4+ in

tetrahedral sites is in contrast to the location of

V4+ in Georgia kaolin, for which Gehring et al.

(1993) found strong evidence for it being located

primarily in the octahedral sheet. However the

Jahn-Teller effect (Jahn & Teller, 1937) in V4+

makes its accommodation in the symmetrical

tetrahedral sites difficult, and its ionic radius in

octahedral coordination (0.58 A) (Shannon, 1976) is

intermediate between those of Al3+ and Fe3+, thus

supporting its easy substitution in the octahedral

sheet.

The main free radical signal corresponds to the

A-centre reported previously for defects induced by

exposure to g-irradiation (Clozel et al., 1994). Its

relative insensitivity to prolonged grinding is

consistent with the assignment to an electron hole

in an O atom associated with Si, and its stability

adds support to the proposed use of this resonance

for measuring the radiation history of the sample

(Clozel et al., 1990). The 27Al hyperfine structure

that is associated with the B-centre was not resolved

in the ground samples, consistent with grinding-

induced damage to the octahedral sheet.

In recent years there has been growing interest in

the modification of kaolinite or kaolin minerals to

produce new materials with specific value-added

properties. These have involved one or a combina-

tion of acid, alkali, heat or mechanical treatments,

and the relative stability of these non-swelling

minerals means that it is necessary to use more

severe activation treatments than with swelling

minerals, such as bentonites. Nevertheless the

relatively low cost of kaolins means that even the

need for more complex modification procedures can

still be cost effective. The material described in the

present paper is an example where there is now

good evidence that it can be a low-cost alternative

to bentonite bleaching clay for the decolourization

of rice bran oil (Worasith et al., 2011).

CONCLUS IONS

Because of their potential for use as bleaching clays

in the food industry, the products from physical and

chemical treatments of the Ranong kaolin have

been characterized using a range of physical and

chemical techniques. Acid treatments alone had

little effect on the aluminosilicate mineral struc-

tures, but (dry) grinding produced a major reduction

in �OH content, as a result of dehydroxylation of

the surface of the octahedral sheet. Subsequent acid

treatments then resulted in the production of

material with greatly increased surface area, low

Al:Si ratios, and amorphous structure, which is

presumably responsible for the enhanced sorption

properties of the product.

We were also able to obtain a detailed picture of

the paramagnetic ions and free radical defects

within the Ranong kaolin, and the extent to which

their environments are influenced by physical and

chemical treatments. Grinding-induced changes

primarily affected the Fe3+ (i.e. octahedral cation

sites), but not the Mn EPR signal. On the other

hand, acid-induced changes (i.e. in the degree of

protonation of surface oxygen atoms) were seen in

the shape of the Mn EPR signal, a result which may

suggest that the Mn is located in the tetrahedral

sheet, and hence probably is in the Mn4+ state.

Further investigations including X-ray absorption

spectroscopy in order to probe coordination of Mn


atoms would be needed to fully validate such an

assumption. Finally, the major free radical centre,

the so-called A-centre which corresponds to defects

caused by exposure to low levels of g-irradiation,was little affected by either the physical or

chemical treatments. However, despite all of this

information on the kaolin composition and beha-

viour, we still do not know to what extent these

various paramagnetic components contribute to the

chemical behaviour of the products from the

physical and chemical treatments of the kaolin.

ACKNOWLEDGMENTS

Special thanks are due to Arag Vitittheeranon, of the

Office of Atoms for Peace (OAP), Thailand, Dr.

Sakorn Suwan, Chulalongkorn University, Thailand

for use of their EPR spectrometers, and Sirikan

Noonpui, Pilot Plant Development and Training

Institute, King Mongkut’s University of Technology,

Thonburi, Thailand for assistance with the FTIR

spectroscopy. We are also grateful to Drs. Katja

Emmerich, Claude Fontaine and an anonymous

reviewer for their thorough reviews and detailed

comments that have helped to improve the presentation

of this work.

REFERENCES

Angel B.R., Jones J.P.E. & Hall P.L. (1974) Studies of

doped synthetic kaolinite I. Clay Minerals, 10,

247�256.Balan E., Allard T., Boizot B., Morin G. & Muller J.P.

(1999) Structural Fe3+ in natural kaolinites: new

insights from electron paramagnetic resonance

fitting at X- and Q-band frequencies. Clays and

Clay Minerals, 47, 605�616.Balan E., Lazzeri M., Saitta A.M., Allard T., Fuchs Y. &

Mauri F. (2005) First-principles study of OH-

stretching modes in kaolinite, dickite, and nacrite.

American Mineralogist, 90, 50�60.Belver C., Munoz M.A.B. & Vicente M.A. (2002)

Chemical activation of kaolinite under acid and

alkaline conditions. Chemistry of Materials, 14,

2033�2043.Bhattacharyya K.G. & Sen Gupta S. (2007) Influence of

activation of kaolinite and montmorillonite on

adsorptive removal of Cd(II) from water. Industrial

& Engineering Chemistry Research, 46, 3734�3742.Brindley G.W. & Brown G. (1980) Crystal Structures of

Clay Minerals and their Ident i f ica t ion .

Mineralogical Society Monograph no. 5.

Mineralogical Society, London.

Brindley G.W., Cao C.C., Harrison J.L., Lipsicas M. &

Raythatha R. (1986) Relation between structural

disorder and other characteristics of kaolinites and

dickites. Clays and Clay Minerals, 34, 239�249.Calas G. (1988) Electron paramagnetic resonance. Pp.

513�571 in: Spectroscopic Methods in Mineralogy

and Geology, (F.C. Hawthorne, editor). Reviews in

Mineralogy, Vol. 18, Mineralogical Society of

America, Washington, DC.

Christidis G.E., Scott P.W. & Dunham A.C. (1997) Acid

activation and bleaching capacity of bentonites from

the islands of Milos and Chios, Aegean, Greece.

Applied Clay Science, 12, 329�347.Clozel B., Calas G., Muller J.-P., Dran J.-C. & Herve A.

(1990) Kaolinites as dosimeters: a new possibility of

tracing radionuclides migration. Chemical Geology,

84, 259�260.Clozel B., Allard T. & Muller J.-P. (1994) Nature and

stability of radiation-induced defects in natural

kaolinites: new results and a reappraisal of published

works. Clays and Clay Minerals, 42, 657�666.Clozel B., Gaite J.-M. & Muller J.-P. (1995) Al-O-Al

paramagnetic defects in kaolinite. Physics and

Chemistry of Minerals, 22, 351�356.Coyne L.M., Blake D.F., McKeever S.W. & McKeever

S .W .S . , e d i t o r s ( 1 9 9 8 ) S p e c t r o s c o p i c

Characterization of Minerals and their Surfaces.

ACS Symposium Series no. 415, American

Chemical Society, Washington D.C.

Cuttler A.H. (1981) Further studies of a ferrous iron

doped synthetic kaolin: dosimetry of X-ray induced

defects. Clay Minerals, 16, 69�80.Duzgoren-Aydin N.S., Aydin A. & Malpas J. (2002)

Distribution of clay minerals along a weathered

pyroclastic profile, Hong Kong. Catena, 50, 17�41.Ekosse G.-I. (2005) Fourier transform infrared spectro-

photometry and X-ray powder diffractometry as

complementary techniques in characterizing clay

size fraction of kaolin. Journal of Applied Science &

Environmental Management, 9, 43�48.Farmer V.C. (1974) The Infrared Spectra of Minerals.

Mineralogical Society, London.

Farmer V.C. (1998) Differing effects of particle size and

shape in the infrared and Raman spectra of kaolinite.

Clay Minerals, 33, 601�604.Farmer V.C. & Russell J.D. (1964) The infrared spectra

of layered silicates. Spectrochimica Acta, 20,

1149�1173.Fialips C.-I., Petit S., Decarreau A. & Beaufort D.I

(2000) Influence of synthesis pH on kaolinite

crystallinity and surface properties. Clays and Clay

Minerals, 48, 173�184.Frost R.L., Mako E., Kristof J., Horvath E. & Kloprogge

J.T. (2001a) Mechanochemical treatment of kaoli-

nite. Journal of Colloid and Interface Science, 239,

458�466.Frost R.L., Mako E., Kristof J., Horvath E. & J.T.

Kloprogge E. (2001b) Modification of kaolinite

surfaces by mechanochemical treatment. Langmuir,


17, 4731�4738.Frost R.L., Mako E., Kristof J. & Kloprogge J.T. (2002)

Modification of kaolinite surfaces through mechan-

ichemical treatment � a mid-IR and near-IR

spectroscopic study. Spectrochimica Acta, 58A,

2849�2859.Frost R.L., Horvath E., Mako E., Kristof J. & Redey A.

(2003) Slow transformation of mechanically dehy-

droxylated kaolinite to kaolinite � an aged mechan-

ochemically activated formamide-intercalated

kaolinite study. Thermochimica Acta, 408, 103�113.Frost R.L., Horvath E., Mako E. & Kristof J. (2004)

Modification of low and high defect kaolinite

surfaces: implications for kaolinite mineral proces-

sing. Journal of Colloid and Interface Science, 270,

337�346.Gaite J.-M., Ermakoff P., Allard T. & Muller J.-P.

(1997) Paramagnetic Fe3+ � a sensitive probe for

disorder in kaolinite. Clays and Clay Minerals, 45,

496�505.Gehring A.U., Fry I.V., Luster J. & Sposito G. (1993)

The chemical form of vanadium (IV) in kaolinite.

Clays and Clay Minerals, 41, 662�667.Gogoi P.K. & Baruah R. (2008) Fluoride removal from

water by adsorption on acid activated kaolinite clay.

Indian Journal of Chemical Technology, 15,

500�503.Golding R.M., Singhasuwich T. & Tennant W.C. (1977)

An analysis of the conditions for an isotropic g-

tensor in high-spin d5 systems. Molecular Physics,

34, 1343�1350.Goodman B.A. & Hall P.L. (1994) Electron paramag-

netic resonance spectroscopy. Pp. 173�225 in: ClayMineralogy: Physical Determinative Methods (M.J.

Wilson, editor), Chapman & Hall, London.

Goodman B.A., Russell J.D., Fraser A.R. & Woodhams

F.W.D. (1976) A Mossbauer and infra-red spectro-

scopic study of the structure of nontronite. Clays and

Clay Minerals, 24, 53�59.Hawthorne F.C., editor (1988) Spectroscopic Methods in

Mineralogy and Geology. Reviews in Mineralogy,

18, Mineralogical Society of America. Washington

D.C.

Hinckley D.N. (1963) Variability in ‘‘crystallinity’’values among the kaolin deposits of the coastal

plain of Georgia and South Carolina. Proceedings of

the 11th National Conference on Clays and Clay

Minerals, 229�235.Huertas F.J., Chou L. & Wollast R. (1998) Mechanism

of kaolinite dissolution at room temperature and

pressure: Part 1. Surface speciation. Geochimica et

Cosmochimica Acta, 62, 417�431.Jahn H. & Teller E. (1937). Stability of polyatomic

molecules in degenerate electronic states. I. Orbital

degeneracy. Proceedings of the Royal Society of

London. Series A, Mathematical and Physical

Sciences, 161, 220�235.

Joussein E., Petit S., Churchman J., Theng B., Righi D.

& Delvaux B. (2005) Halloysite clay minerals � a

review. Clay Minerals, 40, 383�426.Karaoglu M.H., Dogan M. & Alkan M. (2010) Removal

of reactive blue 221 by kaolinite from aqueous

solutions. Industrial & Engineering Chemistry

Research, 49, 1534�1540.Kuentag C. (2001) Clay deposits in Thailand. Pp. 28�43

in: The 2nd Workshop on Clays and their

Applications, Department of Mineral Resources,

Thailand.

Kuentag C. & Wasuwanich P. (1978) Clay. Economic

Geology Bulletin No. 19, Economic Geology

Division, Department of Mineral Resources,

Thailand.

Langford J.I. & Wilson A.J.C. (1978). Scherrer after

sixty years. A survey and some new results in the

determination of crystallite size. Journal of Applied

Crystallography, 11, 102-113.

Lombardi G., Russell J.D. & Keller W.D. (1987)

Compositional and structural variations in the size

fractions of a sedimentary and a hydrothermal

kaolin. Clays and Clay Minerals, 35, 321�335.Mabbs F.E. & Collison D. (1992) Electron paramag-

netic resonance of d- transition metal compounds.

Elsevier, Amsterdam.

McBride M.B., Pinnavaia T.J. & Mortland M.M. (1975)

Electron spin relaxation and the mobility of

manganese(II) exchange sites in smectites.

American Mineralogist, 60, 66�72.Mako E., Frost R.L., Kristof J. & Horvath E. (2001) The

effect of quartz content on the mechanochemical

activation of kaolinite. Journal of Colloid and

Interface Science, 244, 359�364Martin F., Micoud P., Delmotte L., Marichal C., Le Dred

R., De Parseval P., Mari A., Fortune J.-P., Salvi S.,

Beziat D., Grauby O. & Ferret J. (1999) The

structural formula of talc from the Trimouns deposit,

Pyrenees, France. The Canadian Mineralogist, 37,

997�1006.Meenakshi S., Sairam Sundaram C. & Sukumar R.

(2008) Enhanced fluoride sorption by mechano-

chemically-activated kaolinites. Journal of

Hazardous Materials, 153, 164�172.Mestagh M.M., Vielvoye L. & Herbillon A.J. (1980)

Iron in kaolinite: II. The relationship between

kaolinite crystallinity and iron content. Clay

Minerals, 15, 1�13.Miller J.G. & Oulton T.D. (1970) Prototrophy in

kaolinite during percussive grinding. Clays and

Clay Minerals, 18, 313�323.Murray H.H. (2007) Applied Clay Mineralogy:

Occurrences, Processing, and Application of

Kaolins, Bentonites, Palygorskite-Sepiolite, and

Common Clays. Elsevier, Amsterdam.

Nuntiya A. & Prasanphan S. (2006) The rheological

behavior of kaolin suspensions. Chiang Mai Journal


of Science, 33, 271�281.Panda A.K., Mishra B.G., Mishra D.K. & Singh R.K.

(2010) Effect of sulphuric acid treatment on physico-

chemical characteristics of kaolin clay. Colloids and

Surfaces A, 363, 98�104.Rancourt D.G., Christie I.A.D., Royer M., Kodama H.,

Robert J.-L., Lalonde A.E. & Murad E. (1994)

Determination of accurate [4]Fe3+, [6]Fe3+ and [6]Fe2+

site populations in synthetic annite by Mossbauer

spectroscopy. American Mineralogist, 79, 51�62.Sengupta P., Saikia N.J., Bharali D.J., Saikia P.C. &

Borthakur P.C. (2006) ESR investigation of deferra-

tion treatment of iron-rich kaolinite clay from

Deopani, Assam, India. Current Science, 91, 86-90.

Shannon R.D. (1976) Revised effective ionic radii and

systematic studies of interatomic distances in halides

and chalcogenides, Acta Crystallogriphica A, 32,

751�767.Steudel A., Batenburg L.F., Fischer H.R., Weidler P.G.

& Emmerich K. (2009) Alteration of non-swelling

clay minerals and magadiite by acid activation.

Applied Clay Science, 44, 95�104.Temuujin J., Okada K., MacKenzie K.J.D. & Jadambaa

Ts. (2001) Characterization of porous silica prepared

from mechanically amorphized kaolinite by selective

leaching. Powder Technology, 121, 259�262.Title R.S. (1963) Electron paramagnetic resonance

spectra of Cr+, Mn++, and Fe3+ in cubic ZnS.

Physical Review, 131, 623�627.

Vagvolgyi V., Kovacs J., Horvath E., Kristof J. & Mako

E. (2008) Investigation of mechanochemically mod-

ified kaolinite surfaces by thermoanalytical and

spectroscopic methods. Journal of Colloid and

Interfac Science, 317, 523�529.Vallyathan V., Shi X., Dalal N.S., Irr W. & Castranova

V. (1988) Generation of free radicals from freshly

fractured silica dust. American Review of

Respiration Disease, 138, 1213�1219.van der Marel H.W. & Krohmer P. (1969) O-H

stretching vibrations in kaolinite and related miner-

als. Contributions to Mineralogy and Petrology, 22,

73�82.Weeks R.A. (1956) Paramagnetic resonance of lattice

defects in quartz. Journal of Applied Physics, 27,

1376�1381.Worasith N., Goodman B.A., Jeyashoke N. &

Thiravetyan P. (2011) Decolorization of rice bran

oil using modified kaolin. Journal of the American

Oil Chemists’ Society. <doi:10.1007/s11746-011-

1872-2>

Woumfo D., Kamga R., Figueras F. & Njopwouo D.

(2007) Acid activation and bleaching capacity of

some Cameroonian smectite soil clays. Applied Clay

Science, 37, 149�156.Wu Z., Li C., Sun X., Xu X., Dai B., Li J. & Zhao H.

(2006) Characterization, acid activation and bleach-

ing performance of bentonite from Xinjiang. Chinese

Journal of Chemical Engineering, 14, 253�258.


Characterization of modified kaolin from the Ranong deposit Thailand by XRD, XRF, SEM, FTIR and EPR techniques

Documents

acid treatment aftergrinding

acid extraction

acid activation

guangxi university

epr spectra of fe

chulalongkorn university

epr havebeen

epr techniquesn