Yerbilimleri, 23 (2001), 53-60

Hacettepe †niversitesi Yerbilimleri Uygulama ve AraßtÝrma Merkezi BŸlteni

Bulletin of Earth Sciences Application and Research Centre of Hacettepe University

Desulphurisation Of Üvrindi Alunitic Kaolin

Üvrindi Alunitli Kaolininden KŸkŸrdŸn UzaklaßtÝrÝlmasÝ

Zafir EKMEK‚Ü, …zcan G†LSOY, Salih ERSAYIN , Ürfan BAYRAKTAR

Hacettepe †niversitesi, MŸhendislik FakŸltesi, Maden MŸhendisliÛi BšlŸmŸ, 06532 Beytepe, Ankara

ABSTRACT

In this paper, the results of degritting, classification, flotation, leaching and roasting tests carried out to produce a
final product with acceptable sulphur content (<0.5 % SO3) from Üvrindi (BalÝkesir-Turkey) alunitic kaolin are pre-
sented. Separation of alunite from kaolinite by physical separation methods was proved quite difficult, since aluni-
te grains were also disintegrated to ultrafine particle size range as kaolinite. Both acidic and alkaline leaching tests
were applied to reduce the sulphur content of the sample and a final product containing 0.73 % SO3 was obtained
by alkaline leaching. Although it was possible to obtain a final product with 0.48 % SO3 by roasting at 1000 °C,
due to conversion of kaolinite into metakaolinite, the casting property of the sample affected adversely.

Key words: Alunitic kaolin, classification, desulphurisation, flotation, leaching, roasting.

…Z

Bu yazÝda, Üvrindi (BalÝkesir-TŸrkiye) alunitli kaolininden kabul edilebilir kŸkŸrt i•eriÛine (<% 0.5 SO3) sahip bir son
ŸrŸn elde etmek amacÝyla yapÝlan kil a•ma, sÝnÝflandÝrma, flotasyon, li• ve kavurma deney sonu•larÝ verilmißtir. Üv-
rindi kaolin yataÛÝnda bulunan alunit tanelerinin a•ma ißlemi sonrasÝnda, kaolinit taneleri gibi •ok ince tane boyla-
rÝnda olmasÝ nedeniyle, fiziksel yšntemler kullanÝlarak alunit tanelerinin kaolinden ayrÝlmasÝnÝn zor olduÛu belirlen-
mißtir. …rneÛin kŸkŸrt i•eriÛinin azaltÝlmasÝ amacÝyla hem asidik, hem de alkali li• yšntemleri uygulanmÝß, ancak
alkali li• sonrasÝnda sadece % 0.73 SO3 i•eriÛine sahip bir son ŸrŸn elde edilebilmißtir. Kavurma deneylerinde
1000 °C sÝcaklÝkta % 0.48 SO3 i•erikli bir son ŸrŸnŸn elde edilmesine karßÝn, dškŸm šzelliklerini olumsuz yšnde
etkileyen meta-kaolinit fazÝnÝn olußmasÝ bu sÝcaklÝkta ger•ekleßmißtir.

Anahtar kelimeler: Alunitli kaolin, sÝnÝflandÝrma, kŸkŸrdŸn uzaklaßtÝrÝlmasÝ, flotasyon, li•, kavurma

INTRODUCTION
ron size impurities require special vigorous tre-
Kaolin is one of the most valuable industrial atment.
clays whose commercial value is determined by
its whiteness, chemical purity, particle size dist- Apart from the generally occurring impurities gi-
ribution, etc. The kaolin extracted from the com- ven above, there are more than 12 kaolin depo-
mercial deposits contains kaolinite as the major sits in Turkey where sulphur is the major impu-
component together with accessory minerals, rity (Alpar et al., 1973). The sulphur in these de-
such as quartz, muscovite, limonite, anatase, posits is generally associated with pyrite and/or
hematite, illite and organic matter. For industrial alunite. The physical and chemical characteris-
applications, kaolin must be extensively proces- tics (except sulphur content) of the kaolin ext-
sed and refined in order to be used as pigment, racted from these deposits are generally suitab-
filler, coater, extender and ceramic raw material, le for ceramic production. Since sulphur causes
etc. The partial or complete removal of these im- cracks and pores during firing at elevated tem-
purities in an economical manner has been the peratures, it is impossible to use such ores in
subject of many researches. The coarser impu- ceramic production directly. Therefore, SO3
rities, generally quartz, are quite easily separa- content of such raw materials must be reduced
ted by screening or classification, while the mic- to lower than 0.5 %.
54 Yerbilimleri

Sulphur content of alunitic kaolin is generally re- 50 % solid by weight and impeller speed of 1500
duced by thermo-chemical method in which the rpm for 10 minutes. The pulp was then sieved
raw material is subjected to temperatures of 900 through 300 µm and a 50 mm Mozley hydrocyc-
°C or higher (Can and Ündel,1988). The required lone with 6.4 mm vortex and 14.3 mm apex di-
roasting temperature can also be reduced to ameters was used for classification. The oversi-
600 °C by addition of 2-5 % of Na2CO3 or NaCl ze of the sieve was regarded as grit.
during roasting and leaching of the roasted ma-
terial by water (Girgin et al., 1993). There are al- Table 1. Chemical composition of Üvrindi alunitic ka-
so a few publications dealing with separation of olin sample
‚izelge 1. Üvrindi alunitli kaolin numunesinin kimyasal
alunite from kaolinite by selective flocculation bileßimi
and flotation (Koca and …zdaÛ, 1994; Abdel-
Khalek et al., 1996; Gebhardt, et al., 1998). Ho- Component %
wever, these studies are in laboratory scale and Al2O3 31.26
do not have any possibility for industrial applica- SiO2 52.41
tion due to the difficulties encountered mainly CaO 1.13
from ultrafine particle size of clays. Fe2O3 2.77
SO3 1.10
In this study, following degritting and classifica- K2O 0.17
tion stages, flotation, leaching and roasting tests Na2O 0.06
were applied to decrease the sulphur content of TiO2 0.76
alunitic kaolin sample taken from Üvrindi- Balike- LOI 10.34
sir (Turkey) and to produce a final product su-
itable for ceramic production.
Flotation test
MATERIAL AND METHODS
The flotation conditions applied in the flotation of
Ore Characterization Üvrindi alunitic kaolin was chosen based on the
flotation test results in the literature (Gebhadrt et
Alunitic kaolin sample was obtained from Üvrindi al., 1998). After degritting, the Ð38 µm material
(Balikesir) district in the western part of Turkey. obtained by wet sieving was used for the flotati-
Table 1 shows the main chemical composition on tests. AERO Promoter 845 and Na-Oleate
of the sample. As it can be seen from Table 1, were employed as promoter and collector res-
the sulphur and iron contents of the sample are pectively. Sodium silicate was used as a disper-
beyond the acceptable limits for ceramic pro- sant. The pH was adjusted to 6.5 using either
duction. NaOH or HCl. The flotation test was performed
on a 15 % pulp density in a 1 lt. Denver cell. Af-
Mineralogical studies indicated that the major ter pH adjustment, sodium silicate was added at
constituent was kaolinite. Quartz was the abun- dosage of 4 kg/t and conditioned for 5 minutes.
dant impurity. Alunite [KAl3(SO4)2(OH)6], being The pulp was re-conditioned for 10 minutes with
the source of sulphur in the sample, was detec- the collector and the promoter dosages of 1.7
ted by XRD (Figure 1). The iron contaminants and 0.8 g/t respectively. The flotation was per-
were identified mostly as staining on the kaolini- formed for 5 minutes for the first stage, and in
te grains, but free geothite grains were also re- the second stage the same dosages of collector
corded. and promoter were added again. After conditi-
oning for 5 minutes, flotation was further perfor-
Methods med for 5 minutes.

Degritting and classification tests Leaching tests

The sample was crushed to -10 mm with a jaw The leaching tests were carried out in a mecha-
crusher and divided into representative samples nically stirred 1 lt glass vessel. The overflow
of approximately 2 kg. lots. Degritting tests we- product obtained from hydrocyclone separation
re carried out in a scrubber at a pulp density of was used as feed material and leached at 10 %
 Ekmek•i et al. 55

Figure 1: XRD pattern of Üvrindi alunitic kaolin

Þekil 1 : Üvrindi alunitli kaolinin XRD kÝrÝnÝmÝ

pulp density for 1 hour. Reagent grade H2SO4,

HCl and Na2CO3 were used for pH adjustment.

Roasting tests

The roasting tests were carried out in a muffle

furnace at temperatures ranging between 600-
1100 °C for 1 hour. The roasted sample was
then divided into two parts and one of them was
leached in water for 1 hour to dissolve any so-
luble sulphur compounds formed during roas-
ting. The structural changes in the roasted ma-
terial were determined by X-ray diffraction
analysis.

RESULTS AND DISCUSSION

Figure 2: Particle size distribution of Üvrindi alunitic
kaolin
Degritting and Classification Tests Þekil 2: Üvrindi alunitli kaolininin tane boyu daÛÝlÝmÝ

Following degritting of the original sample, its

particle size distribution was determined by wet
sieving down to 38 µm and by Coulter Counter
Industrial Model D for sub-sieve sizes. The par- Table 2. Iron and sulphur contents of different partic-
le size fractions of Üvrindi alunitic kaolin
ticle size distribution of the sample is given in Fi- ‚izelge 2. Üvrindi alunitli kaolininin farklÝ tane boyu
gure 2. fraksiyonlarÝnÝn demir ve kŸkŸrt i•erikleri

Moreover, iron and sulphur contents of the sieve Size Weight Fe2O3 SO3
fractions were determined to find out their distri- (microns) (%) (%) (%)
bution with respect to particle size (Table 2). Re- +300 27.18 5.12 0.90
sults of particle size analysis and of chemical -300+150 4.29 3.10 0.83
analysis of the sieve fractions revealed that al- -150+75 5.09 2.65 0.78
most 60 % of the original sample was finer than
-75+45 4.26 2.18 0.78
38 µm, at finer sizes the sulphur content incre-
ased while the iron content decreased conside- -45+38 1.19 1.72 0.83
rably. -38 57.99 1.92 1.28
56 Yerbilimleri

Kaolins for ceramic productions are controlled In order to determine whether the sulphur con-
generally in terms of iron contents, particle size tent was due to alunite or adsorbed SO42- ions
distribution, strength and rheological properties on kaolinite particles at sub-sieve sizes, XRD
(Harben, 1992; Patterson and Murray, 1983). analysis were done on +5.6 mm, -0.212 + 0.106
Therefore, in the first step kaolin with suitable mm and Ð0.038 mm fractions. As can be seen
particle size distribution (<20 µm) should be pro- from the XRD patterns given in Figure 4, alunite
duced. The original sample was degritted at 50 was identified only in Ð0.038 mm fraction, cle-
% pulp density for 10 minutes and screened arly proving that the origin of sulphur at sub-si-
through 300 µm sieve. The undersize material eve sizes was alunite. Hence, it was concluded
was then diluted to 20 % pulp density and fed to that most of the alunite grains were passed to
the hydrocyclone at 3.5 bar inlet pressure. The the overflow product and it was impossible to
particle size distribution of the overflow product
obtain a final product with low sulphur content
is given in Figure 3.
by only degritting and classification.

Figure 3: Particle size distribution of the cyclone

overflow product
Þekil 3: Hidrosiklon Ÿst akÝmÝnÝn tane boyu daÛÝlÝmÝ

Approximately 44 % of the hydrocyclone feed

were taken as overflow product with particle si-
ze finer than 20 µm. Iron and sulphur contents
of the hydrocyclone products showed that while
iron content of the overflow decreased, its sulp-
hur content increased with respect to the feed
grade (Table 3).

Table 3. Iron and sulphur contents of cyclone pro-

ducts.
‚izelge 3. Hidrosiklon ŸrŸnlerinin demir ve kŸkŸrt
i•erikleri.
Product Fe2O3 SO3
(%) (%)
Figure 4: XRD patterns of some particle size fracti-
Overflow 1.82 1.53 ons of Üvrindi alunitic kaolin
Underflow 2.68 0.85 Þekil 4: Üvrindi alunitli kaolininin bazÝ tane boyu frak-
Feed 2.31 1.14 siyonlarÝnÝn XRD kÝrÝnÝmlarÝ
 Ekmek•i et al. 57

Flotation Test Table 5. Results of leaching tests carried out under

different pH and temperatures
The results of flotation test performed on the ‚izelge 5. FarklÝ pH ve sÝcaklÝklarda ger•ekleßtirilen
li• deneylerinin sonu•larÝ
Ð0.038 µm fraction are reported in Table 4. As
mentioned above, the flotation conditions opti- pH SO3
mised by Gebhardt et al. (1998) to obtain an alu- (%)
nite concentrate from low grade alunite disper- Cold Leaching 7-7.5 1.20
sed very finely in the matrix of a quartz Ðkaolini- (15-20 °C) 2-2.5 1.65
te ore were employed to remove alunite from Üv- Hot Leaching 7-7.5 1.12
rindi alunitic kaolin. According to the results re- (60-65 °C) 2-2.5 1.73
ported by Gebhardt et al. (1998), alunite could Feed 1.53
be removed with a recovery of 42 % with one
flotation stage from an ore containing 1.6 %
SO3. The cumulative recovery after two conse- Sulphur content of the sample was slightly redu-
cutive flotation stages was increased to approxi- ced at neutral pH and the effect of hot leaching
mately 65 %. Although, Üvrindi alunitic kaolin has was negligible. However, the sulphur content
similar mineralogical and chemical compositi- was increased to higher values than that of feed
ons, only 18.24 % of SO3 could be removed af- sample at acidic pH, in spite of successive was-
ter two stages of flotation (Table 4). The SO3 hings with clean water. This unexpected result
content of kaolinite could only be reduced from was attributed to reprecipitation of varieties of
1.09 to 1.06 %. hydroxysulphates containing K, Al and Fe ions
dissolved from the sample during sulphuric acid
leaching (Figure 5). The Eh-pH diagram of Al-K-
Table 4. Results of alunite flotation test. S-H2O system at 25 °C drawn by using Outo-
‚izelge 4. Alunit flotasyonu deney sonu•larÝ. kumpu HSC Chemistry software showed that
Product Weight SO3 Recovery precipitation of alunite [KAl3(OH)6(SO4)2] is fa-
(%) (%) (%) voured between pH 2.3-7. When the temperatu-
Float 1 4.87 1.40 6.27 re of the solution is increased to 65 °C, the sta-
Float 2 11.23 1.16 11.97 bility region of alunite broadens down to pH 0
Concentrate 83.90 1.06 81.76
(Figure 6). The slight increase in the sulphur
Feed 100.00 1.09 100.00
content of the sample leached at 65 °C was in
agreement with broadening of the stability regi-
on of alunite.
Comparison of the particle size analysis of both
samples showed that Üvrindi alunitic kaolin was
finer than that of used by Gebhardt, et al.
(1998). Hence, the difference in the flotation re-
sults of these two studies may be attributed to
the difference in the fineness between the two
samples and inefficient flotation of ultrafine par-
ticles by classical froth flotation method.

Leaching Tests

The leaching tests were executed to assess the

influence of pH and pulp temperature on the dis-
solution behaviour of alunite in kaolin. The natu-
Figure 5: Eh-pH diagram of Al-K-S-H2O system at 25
ral pH of the sample was around 7 - 7.5 with tap
°C ( dashed lines show the aqueous pha-
water. In the first step, sulphuric acid was used se)
for pH regulation. The results of leaching tests Þekil 5: Al-K-S-H2O sisteminin 25 °C sÝcaklÝkta Eh-
carried out under different pH and temperatures pH diagramÝ ( kesikli •izgi sulu fazÝ gšs-
are summarised in Table 5. termektedir)
58 Yerbilimleri

In order to prevent re-precipitation of alunite, re-

dox potential of the pulp was decreased to redu-
cing potential region (< 0 mV) by using a strong
reducing agent, sodium dithionite. Since, the
stable form of sulphur is H2S(a) in highly acidic
and reducing conditions (Pourbaix, 1966), for-
mation of hydroxysulphates in the solution was
not expected. When the sample was leached in
the presence of sodium dithionite at pH 2-2.5
and 15-20 °C temperature, the SO3 content was
decreased to 1.32 % after and to 1.24 % after 60
minutes leaching. The results of the tests sho-
wed that the increase in the sulphur content of
Figure 6: Eh-pH diagram of Al-K-S-H2O system at 65
°C the sample was prevented in reducing conditi-
Þekil 6: Al-K-S-H2O sisteminin 65 °C sÝcaklÝkta Eh- ons, but negligible amount of sulphur could be
pH diagramÝ removed even in highly acidic solutions.

When H2SO4 was replaced by HCl, sulphur con-

Dissolved iron ions in the solution may be found tent of the sample was decreased from 1.53 to
in aqueous form of FeSO4(a), Fe2+(a) and Fe- 1.04 % SO3 at a pH of 2-2.5 and 60-65 °C pulp
SO+4 (a) or in solid precipitate form of FeSO4, temperature. This was due to formation of
Fe2O3 and FeS2 depending on pH and Eh of the AlCl3(a), K+(a) and Fe2+(a) in acidic solutions
solution and the amount of iron ions dissolved rather than hydroxysulphate precipitation. Altho-
(Figure 7). Precipitation of jarosite ugh this value is lower than that obtained by hot
[KFe3(SO4)2(OH)6] may also be possible. Howe- leaching with H2SO4 (1.24 % SO3), it is still far
ver, considering slow kinetics of jarosite precipi-
from the acceptable value (0.5 % SO3).
tation and requirement of several hours and
Alkaline leaching was carried out as an alterna-
high temperatures (100 °C) for complete preci-
tive to acid leaching. Pulp pH was adjusted to
pitation (Das et al., 1996), jarosite precipitation
12-13 by using Na2CO3 and the temperature
may be in negligible amount in the experimental
was maintained at 60-65 °C. Since, the sulphur
conditions of this work (65 °C and 1 hour leac-
dissolved is in the form of SO4-2 and precipitati-
hing time). Detailed leaching tests should be un-
on of alunite is no longer possible in alkaline so-
dertaken to reach certain conclusions about re-
precipitation of alunite and jarosite in the soluti- lutions (see Figure 6), the sulphur content of the
on. sample was decreased to 0.75 % SO3 by leac-
hing with Na2CO3.

Roasting Tests

The effect of roasting and leaching after roasting

on the reduction of the sulphur content of the
sample was illustrated in Figure 8. The sulphur
content was decreased slightly between 600
and 800 °C. However, it was rapidly decreased
after 800 °C and a product with 0.48 % SO3 was
obtained at 1000 °C. The sulphur was comple-
tely removed from the sample probably in the
form of SO2 and/or SO3 gas. Leaching of roas-
Figure 7: Eh-pH diagram of Fe-S-H2O system at 65
°C
ted sample slightly reduced the sulphur content
Þekil 7: Fe-S-H2O sisteminin 65 °C sÝcaklÝkta Eh-pH and this reduction was negligible at high tempe-
diagramÝ ratures.
 Ekmek•i et al. 59

Structural variations of kaolinite were determi-

ned by X-ray diffraction. The XRD patterns of
the roasted samples are given in Figure 9. Ka-
olinite structure is destroyed starting from 600
°C and changes to metakaolin. Between 700
and 1000 °C temperatures, only quartz peaks
were determined. With this change to metaka-
olin the sample also loses its plastic properties
when mixed with water. However, it is known
that metakaolin can be rehydrated by extended
exposure to water to again form kaolinite and
thus, to regain its plastic property (Lawrence,
1972). Hence, the roasted samples were leac-
hed with water both to remove any water solub-
Figure 8: Effects of roasting and leaching after roas-
ting on reduction of sulphur content le sulphur species formed during roasting and to
Þekil 8: Kavurma ve kavurma sonrasÝnda li• ißlemle- regain their plastic property.
rinin kŸkŸrt i•eriÛindeki azalmaya etkileri

Figure 9: XRD patterns of roasted samples at different temperatures

Þekil 9 : FarklÝ sÝcaklÝklarda kavrulmuß numunelerin X-ÝßÝnÝ kÝrÝnÝmlarÝ
60 Yerbilimleri

However, XRD patterns of the leached samples REFERENCES

after roasting were the same as the only roasted
samples. Hence, it was concluded that kaolini- Abdel-Khalek, N.A., Arafa, M.A., and Hassan, F.,
te structure could not be restored even after 1 1996. Froth flotation of ultrafine Egyptian
hour leaching of the roasted sample in water. At kaolin ore. Changing Scopes in Mineral
temperatures higher than 1000 °C, kaolinite was Processing, M. Kemal, V. Arslan, A. Akar,
and M. CanbazoÛlu (eds.), A.A. Balkema,
transformed into mullite and cristoballite pha- Rotterdam, 395-400.
ses.
Alpar, S.R., GŸrgey, Ü., Rodopman, K. ve Ustaer, C.,
1973. SŸlfat ve pirit ihtiva eden kaolin mine-
Although sulphur content and brightness of the
rallerinin arÝtÝlmasÝ. T†BÜTAK Proje Raporu
sample was increased by roasting at 1000 °C, No: MAG-246, Ankara, 30s.
the plasticity in casting, which is one of the most
important controlling parameter in ceramic pro- Can, M.S. ve Ündel, Ü., 1988. Alunitli kaolenler ve ref-
rakter sanayiinde kullanÝmlarÝ. M.T.A. Ra-
duction, disappeared owing to the fact that por No: 242, Ankara, 13s (yayÝmlanma-
transformation of kaolinite structure into other mÝß).
phases.
Das, G.K., Acharya, S., Anand, S., and Das, R.P.,
1996. Jarosites: A review. Mineral Proces-
sing and Extractive Metallurgy Review, 16,
CONCLUSIONS 185-210.
Gebhardt, J.E., Piga, L., and Schena, G., 1998. Floc-
Separation of alunite from kaolinite by degritting culation and flotation behavior of a low-
and classification was not possible since most of grade alunite ore. Minerals and Metallurgi-
alunite grains were also merged to the overflow cal Processing, 15(4), 48-52.
product together with kaolinite grains. This se- Girgin, Ü., Ekmek•i, Z., ve Erkal, F. 1993. SÝndÝrgÝ alu-
paration was also not possible by froth flotation nitli kaolini zenginleßtirme •alÝßmalarÝ. TŸr-
due to inefficiency of this method at ultrafine kiye 13. Madencilik Kongresi, Üstanbul,
particle sizes. (Abstract in English), 549-560.
Harben, P.W., 1992. The Industrial Minerals Handy-
Sulphur content of the alunitic kaolin sample co- book. Industrial Minerals Division Metal
uld be decreased down to 0.75 % SO3 by leac- Bulletin PLC, London, UK, 148p.
hing with Na2CO3. Leaching in acidic solutions, Koca, S., and …zdaÛ, H., 1994. Flotation of alunite
even in strongly reducing potentials, was not from kaolin. Progress in Mineral Proces-
successful due to formation of solid phases sing Technology, H. Demirel, and S. Er-
[KAl3(OH)6(SO4)2 and KAl(SO4)2.12H2O] at the sayÝn(eds.), A.A. Balkema, Rotterdam,
experimental conditions. 135-140.
Lawrence, W.G., 1972. Ceramic Science for the Pot-
Roasting was considered to be an alternative ter. Chilton Book Company, New York,
method to decrease sulphur content of the 239p.
sample. However, the sulphur content could be Patterson, S.H., and Murray, H.H., 1983. Clays. In-
reduced down to 0.5 % SO3 by raosting at tem- dustrial Minerals and Rocks. S.J. Lefond
peratures as high as 1000 °C. Since kaolinite (ed.), Port City Press, Bultimore, Maryland,
was decomposed and its casting property was 585-652.
diminished, roasting at high temperatures was Pourbaix, M., 1966. Atlas of Electrochemical Equilib-
considered to be inapplicable. ria. Pergamon Press, London, 545-553.