Anal Bioanal Chem (2003) 375 : 820–825
DOI 10.1007/s00216-003-1796-2
O R I G I N A L PA P E R
M. Luisa Cervera · M. Carmen Arnal ·
Miguel de la Guardia
Removal of heavy metals by using adsorption on alumina or chitosan
Received: 27 September 2002 / Revised: 19 December 2002 / Accepted: 20 December 2002 / Published online: 22 February 2003
© Springer-Verlag 2003
Abstract The removal of heavy metals from wastewater
by using activated alumina or chitosan as adsorbers was
evaluated. Cd(II) and Cr(III) were employed as models of
the behaviour of divalent and trivalent metal ions. The ad-
sorption of Cd(II) and Cr(III) onto the adsorbers evaluated
was studied as a function of pH, time, amount of adsorber,
concentration of metal ions and sample volume. A 0.4-g
portion of activated alumina can retain 0.6 mg Cr(III) and
0.2 mg Cd(II) from 20 mL sample adjusted at pH 4 and
stirred for 30 min. It is therefore possible to totally decon-
taminate 500 mL of a waste containing 5 mg L–1 Cd(II) and
Cr(III) with 10 g alumina. On the other hand, 0.4 g chi-
tosan can totally decontaminate 20 mL of a pH 5 solution
containing up to 50 mg L–1 Cd(II) and Cr(III). A 99.2±
0.1% retention of Cd(II) and 83±1% retention of Cr(III)
was obtained from 500 mL of a laboratory waste. The
aforementioned strategies were applied for the minimiza-
tion of analytical chemistry teaching laboratories and
atomic spectrometry laboratory wastes. On comparing
both adsorbers it can be concluded that chitosan is more
preferable than alumina due to the reduced price of chi-
tosan and the absence of side-pollution effects.
Keywords Metal ion removal · Chitosan · Alumina ·
Cadmium · Chromium
Introduction
Metal oxides can quantitatively adsorb heavy metals from
their aqueous solutions and are thus employed in metal
preconcentration [1, 2, 3] and metal removal [4].
In order to facilitate wastewater remediation and to re-
duce the amount of dangerous wastes containing metal
ions, different types of sorbents have been proposed [5],
M. Luisa Cervera (✉) · M. Carmen Arnal · M. de la Guardia
Department of Analytical Chemistry, Research Building,
University of Valencia,
50 Dr Moliner St. 46100 Burjassot, Valencia, Spain
e-mail: m.luisa.cervera@uv.es
like clays [6], zeolites [7], seaweed [8], dead biomass [9]
and solid wastes like coal fly ashes [10], bone charcoal
[11] and chitosan.
Chitosan is the deacylated product of chitin, a polysac-
charide consisting predominantly of unbranched chains of
β-(1→4)-2-acetamido-2-deoxy-D-glucose. Chitosan is found
naturally in fungi and arthropods in which it is the main
component of the exoskeleton, it can be prepared from
fishery wastes and is obtained from shrimp, lobster, or
crab shell and from cuttlebone [12].
Previous studies have demonstrated the ability of chi-
tosan to act as an adsorber of metal ions in aqueous solu-
tions [13] with evidence that sorption occurs through the
amine functional groups [14]. On the other hand, chitosan
impregnated with microemulsions [15], chitosan mem-
branes [16], crosslinked chitosan beads [17], chitosan al-
ginate encapsulated Escherichia coli [18] and chitosan
flakes [19, 20, 21] have been employed for metal ion re-
moval.
In the present study, we evaluate the use of chitosan
flakes for the removal of heavy metal ions in terms of the
effect of time, nature and amount of chitosan, pH of sam-
ples, concentration of metal ions and sample volume.
Cd(II) and Cr(III) were employed as models of the behav-
iour of divalent and trivalent ions and comparative studies
on the use of activated alumina were developed. Finally,
in the best conditions established, we evaluated the use of
both alumina and chitosan for the reduction of laboratory
wastes obtained from qualitative analysis and atomic
spectrometry laboratories.
Experimental
Apparatus
Atomic absorption measurements for Cd and Cr were made using
a Shimadzu, AA 660 V-3 (Kumamoto, Japan) flame atomic ab-
sorption spectrometer equipped with hollow cathode lamps of Cd
and Cr Photron (Victoria, Australia) at conditions indicated in
Table 1.
Laboratory wastes, both untreated and treated by alumina and
chitosan, were analysed by ICP-MS using a Perkin–Elmer model
 821
Table 1 Instrumental parameters used in measures obtained with Table 2 Effect of the activation of alumina on the retention
flame atomic absorption spectrometer AA 660 V-3 Shimadzu process of Cd(II) and Cr(III)a
Cd Cr Activated Untreated Deactivated
Wavelength (nm) 228.8 357.9 Cd(II) 59±5 21±5 12±5
Lamp current (mA) 8 10 Cr(III) 97.5±0.6 65±3 64±3
C2H2 flow (L min–1) 2 2.6 aExperimental conditions: 20 mL solution of 5 mg L–1 Cd(II) and
Air flow (L min–1) 8 8
Cr(III) treated at pH 4 with 0.1 g alumina during 30 min. Data re-
Bandpass (nm) 0.25 0.25 ported correspond to the retention percentage of considered
Burner height (cm) 0.6 0.7 ions±the standard deviation of 3 independent assays
Standard interval (mg L–1) 0–2.0 0–2.5
Background corrector D2 –
Results and discussion
Elan 5000 Spectrometer (Connecticut, USA) in the TotalQuant Cd and Cr sorption on activated alumina
mode and an ICP-OES using a Perkin–Elmer model Optima 3200
RL (Connecticut, USA).
A Crison 2000 pH meter (Barcelona, Spain), equipped with a The effect of several experimental parameters on the metal
glass electrode and an AgCl/Ag reference electrode, was used for retention capacity of alumina, such as pH, alumina activa-
measuring the pH of solutions. tion grade, alumina mass, the concentration of Cd(II) and
A Selecta (Barcelona, Spain) oscillant shaker Vibromatic, with Cr(III), stirring system and the volume of sample were in-
oscillant regulation from 100 to 700 oscillations min–1, a 8-mm os-
cillation size and 60-min time programmer and a Jenway 1000 vestigated in order to establish the best conditions for
(Felsted, England) magnetic stirrer were employed for stirring the metal removal. The initial conditions of this study were
waste samples in the presence of the solid sorbent. An Heraeus 20 mL of a solution containing 5 mg L–1 Cd(II) and Cr(III)
Sepatech centrifuge was used for phase separations after sample at pH 4 and 0.1 g alumina. The mixture was shaken for
treatment. 30 min.

Reagents
Effect of activation process
Low (Mr≈150,000), medium (Mr≈400,000) and high molecular
weight (Mr≈2,000,000) chitosan (2-amino-2-deoxy-(1→4)-β-D-glu-
copyranan; poly(1,4-β-D-glucopyranosamine) were obtained from Preliminary studies on the effect of activation of alumina
Fluka (Buchs, Swiss) and low-cost chitosan flakes from Guinama were carried out using alumina activated at 500 °C over-
(Valencia, Spain). All these compounds were employed without night, untreated alumina and an alumina which was stored
any pre-treatment. overnight with 2 mL water; as can be seen in Table 2 the
Aluminium oxide, acid 504 C type from Fluka was activated by
heating at 500 °C for 12 h before being used. activation process drastically affects the sorption capacity
Standard solution of Cd(II) containing 1,000 mg L–1 was pre- of alumina and it is necessary to thermally treat the sor-
pared by dissolving 1.000 g of cadmium metal (reagent grade from bent before use and to remove water after operation in or-
Merck, Darmstadt, Germany) in 20 mL 5 M HCl, adding 2 drops of der to avoid deactivation processes. Additionally it can be
concentrated HNO3 and diluting to 1 L with deionized water. Stan- noticed that alumina retains better Cr(III) than Cd(II).
dard solution of Cr(III) was prepared by dissolving 7.6960 g
Cr(NO3)3·9 H2O (reagent grade from Merck, Darmstadt, Germany)
in 250 mL deionized water and diluting to 1 L adding HNO3 to ob-
tain a final concentration of 2% (v/v); this was checked against a Effect of pH
titrisol standard solution from Merck (Darmstadt, Germany).
Analytical reagent-grade water with a resistivity of 18.2 MΩ cm
was obtained with a Milli-Q water system from Millipore Inc. At mg L–1 levels, Cd2+ precipitates at around pH 9 and
(Paris, France) and all additional reagents employed were of ana- Cr3+ precipitates at a pH near to 5; the latter species is pre-
lytical grade. sent in low-acidity media as CrOH2+ and Cr(OH)2+.
Solutions were stored into polyethylene flasks and all polyeth-
ylene and glassware material was soaked in 10% (v/v) HNO3 for
12 h and rinsed three times with deionized water before use.

Experimental procedure

Waste solution (500 mL) was adjusted to pH 4 and 5 with NH4OH

and HNO3 and then 10 g activated alumina or chitosan was added,
respectively. The system was shaken for 30 min and then both
phases were separated by filtration. The metal content of the fil-
trate was determined by atomic spectrometry.

Fig. 1 Effect of pH on the retention of Cd(II) and Cr(III) on acti-

vated alumina. Experimental conditions: 20 mL 5 mg l–1 Cd(II) and
Cr(III) treated 30 min with 0.1 g alumina
822
In order to evaluate the effect of sample pH on the ad-
sorption of Cd(II) and Cr(III) by alumina, a pH range
from 3 to 5 was assayed and the results are summarized in
Fig. 1. As can be seen, the retention of metal ions in-
creases on increasing the pH. However, in order to avoid
the influence of hydroxide precipitation and since non-
quantitative removal data for Cd were found, pH 4 was se-
lected, at which Cr(III) was partially retained at a level of
80.3±0.8% and Cd(II) at 27±9%.
Effect of the alumina mass
The increase of the alumina mass from 0.1 g to 0.4 g dra-
matically increases the retention of Cd(II) from 40% up to
100%; the affect on the retention of Cr(III) is less impres-
sive (see Fig. 2).
Effect of metal ion concentrations
Figure 3 depicts the effect of increasing concentrations of
Cd(II) and Cr(III) on the retention of these elements on
0.4 g activated alumina and, as can be seen, under the afore-
mentioned conditions 0.6 mg Cr(III) and 0.2 mg Cd(II)
can be retained quantitatively.
Study of different shaking systems
In order to provide a good contact between the metal ion
present in aqueous wastes and the sorbent, we evaluated
Fig. 2 Effect of alumina mass on the retention of Cd(II) and
Cr(III). Experimental conditions: 20 mL 5 mg L–1 Cd(II) and
Cr(III) treated at pH 4 for 30 min with different activated mass of
alumina
Fig. 3 Study of Cd(II) and Cr(III) retention capacity of 0.4 g acti-
vated alumina. Experimental conditions: 20 mL aqueous waste ad-
justed at pH 4 and stirred for 30 min with alumina
the use of different shaking systems: i) a vibrational shaker,
ii) a magnetic stirrer, and iii) air bubbling. A 75-mL
aliquot of a solution of 5 mg L–1 Cd(II) and Cr(III) was
treated with 1.5 g activated alumina for 30 min. Cr(III)
was quantitatively retained by using all the stirring sys-
tems and Cd(II) was retained at a level of 96.1±0.3%,
96±1% and 99±1% by using vibrational, magnetic stir-
ring, and air bubbling, respectively; thus, the shaking sys-
tem is not at all a critical parameter for metal adsorption
on activated alumina, and it is thus being possible to adapt
the treatment procedure to the available facilities and con-
venience of the laboratory.
Effect of the sample volume
Waste samples (500 mL) were treated with 10 g activated
alumina in order to verify the possibility of using the
aforementioned procedure for the reduction of waste vol-
umes in an applications laboratory scale. For Cd(II) we
found a retention percentage of 95.1±0.1% using a mag-
netic stirrer and 93±2% by using air bubbling; Cr(III) was
quantitatively retained in all cases.
Treatment of laboratory wastes
A 500-mL sample of qualitative analysis teaching labo-
ratory waste, containing 16 mg L–1 Cr, 866 mg L–1 Cu,
1,152 mg L–1 Hg, 193 mg L–1 Mn and 208 mg L–1 Zn, was
treated with 10 g alumina and, in this complex and highly
concentrated medium an average recovery of 53±2% for
Cr was found; the rest of elements were partially retained
at levels of 18±1% for Cu, 2±1% for Hg, 1±1% Mn and
11±2% Zn, evidencing that 106 mg of metal ions can be
removed with 10 g alumina (10.6 mg g–1 alumina).
On the other hand, the treatment of an atomic spec-
trometry laboratory waste, containing several elements at
µg L–1 concentration levels (see Table 3) was decontami-
nated to a large extent by alumina with only Ag and Ba
being recovered at a low percentages.
Cd and Cr sorption on chitosan
A 20-mL aliquot of a solution containing 5 mg L–1 Cd(II)
and Cr(III), adjusted to pH 5 with HNO3 and NH4OH, was
treated with 0.1 g chitosan shaking mechanically for 5 min
before phase separation through filtration.
The effect of pH, shaking time, molecular weight and
particle size of chitosan, chitosan mass, shaking system and
volume and metal concentration of treated wastes, were
evaluated in order to find the best decontamination condi-
tions. Actual laboratory waste samples were treated and
data found are discussed critically.
 823
Table 3 Metal removal from an atomic spectrometry laboratory
waste by using alumina and chitosana
Element Initial Retention percentage
concentration
(µg L–1) Alumina Chitosan
Ag 32 32±2 97.0±0.2
Ba 271.7 50±4 –
Bi 1.5 91±3 88.3±0.7
Ce 2.7 98.6±0.2 98.5±0.4
Cd 107.9 99±1 99.75±0.04
Co 77.65 97.1±1.5 84.8±0.5 Fig. 5 Effect of the shaking time on the retention of Cd(II) and
Cr(III) by chitosan. Experimental conditions: 20 mL 5 mg L–1 solu-
Cr 545 99.5±0.1 97.4±0.1
tion with pH 5 treated with 0.1 g medium molecular weight chi-
Cu 206.3 98.4±0.3 95.3±0.3 tosan
Fe 2092 100 100
Hg 11,320 63.4±1.5 81±1
In 89.5 99.6±0.3 99.4±0.2
La 1.5 97.0±1.1 97.2±0.7 Effect of shaking time
Mn 110.6 84.5±1.8 87.2±0.4
Ni 164.1 100 100 On increasing the contact time between metal wastes and
Pb 232.2 99.3±0.1 93.4±0.1 chitosan we found an increase in the retention percentages
Sb 1.2 98.3±0.8 92±2 of both elements studied up to a maximum recovery of
Sn 13,320 96.4±0.9 95.8±0.4 77% for Cd(II) and 75% for Cr(III) at 30 min; the use of a
Te 0.6 100 98±4 shaking times of 60 or 90 min practically does not im-
U 1.8 96.3±0.6 87±1 prove the metal retention by chitosan (see Fig. 5).
Zn 755 100 100
aExperimental conditions: 500 mL waste were treated with 10 g ad-

sorber at pH 5 for chitosan and pH 4 for alumina, following the
procedure described in the text
Effect of pH
Figure 4 shows the influence of the waste pH, in the range
3–9, on the retention of Cd(II) and Cr(III) by medium mo-
lecular weight chitosan. As can be seen, the increase of
pH also increases the retention percentage of both studied
elements. However, it must be mentioned that the use of
pH values higher than 5 dramatically modifies the mecha-
nism of metal removal from simple adsorption to hydrox-
ide precipitation and because of that a pH of 5, for which
non-quantitative removal data were found, was selected
for the additional studies.
Fig. 4 Effect of pH on the retention percentage of Cd(II) and
Cr(III) using medium molecular weight chitosan. Experimental
conditions: 20 mL 5 mg L–1 solutions treated for 5 min with 0.1 g
chitosan
Effect of molecular weight and particle size of chitosan
In order to verify the influence of the nature and size of
the sorbent on the retention of Cd(II) and Cr(III), three
different chitosan product, from Fluka (Buch, Swiss) both
previously crushed in a water-refrigerated mill and used
in flakes without further modification, were assayed.
Table 4 summarizes data found in these experiments
and it can be seen that medium molecular weight chitosan
provides the best adsorption capacity for both Cd(II) and
Cr(III) with the additional advantage of not requiring any
size reduction treatment before use. The maximum aver-
age retentions of Cd(II), 66%, and Cr(III), 83%, corre-
spond to the use of 400,000 Da chitosan, employed in
flakes without a further reduction of the particle size and
are comparable with those found by using 2,000,000 Da
chitosan in a fine powder. Thus, medium molecular weight
chitosan can be recommended for metal removal.
Additional experiments, carried out using low-priced
chitosan obtained from the Spanish market, evidenced no
differences with those found with 400,000 Da chitosan. It
must be noticed that this product is directly produced as a
fine powder and probably is a mixture of medium and
high molecular weight chitosan. So, the product obtained
from Guinama can be recommended for application stud-
ies.
Effect of chitosan mass
The increase of the chitosan mass employed for treating
20 mL of a 5 mg L–1 solution of Cd(II) and Cr(III) also in-
creases the retention efficiency; 0.4 g Cd(II) is retained at
824
Table 4 Effect of molecular
weight and particle size of chi- Chitosan
tosan on the retention of Cd(II)
and Cr(III)a Low
aExperimental conditions: Crushed Not crushed
20 mL 5 mg L–1 solution ad- Cd (II) 49.9±1.4% 36±6%
justed to pH 5 and treated with
0.1 g chitosan for 30 min Cr (III) 56±4% 71±3%
a level of 98.7±0.2% and Cr(III) at 97.2±0.5%. A chitosan
mass of 0.4 g was selected because an increase of the chi-
tosan mass to 0.6 g did not provide retention efficiencies
significantly higher (97.9±0.3% and 99.3±0.4% for Cd(II)
and Cr(III), respectively).
Effect of waste volume and metal concentration
The effect of Cd(II) and Cr(III) concentration on the re-
tention of these two metal ions on chitosan was studied in
the concentration range 0.1–50 mg L–1 and it was found
that 0.4 g of the sorbent retained 1 mg Cd(II) and 1 mg
Cr(III) at a level which varies from 97.4±0.5% to 100%.
In contrast, studies carried out with increasing volumes
of wastes, from 20 to 200 mL, containing in all cases 100 µg
Cd(II) and 100 µg Cr(III), evidenced that 0.4 g is not
enough to recover quantitatively these metal ions when
diluted aqueous wastes were treated with chitosan for
30 min. These experiments evidenced problems with ade-
quately mixing of the sorbent and the waste to be treated,
due to the flotation of chitosan flakes. Because of that, an
increasing mass of chitosan, of 1.5 g, was used for the
treatment of 75 mL waste, and recoveries of 97.4±0.7%
for Cd(II) and 97.4±0.2% for Cr(III) were found, which
compare favourably with data of 60% for Cd(II) and 35%
for Cr(III) using 0.4 g chitosan. So, from these results it
can be concluded that an appropriate relationship between
the amount of waste and chitosan must be maintained, this
aspect being more important than the metal concentration
or the chitosan mass effect on a reduced volume of treated
solution. Under the aforementioned conditions, 679.6 mg
of metal ions can be removed with 10 g chitosan.
Effect of shaking system
The flotation of chitosan creates big problems on using
magnetic stirring or air bubbling systems, especially when
relatively high volumes of wastes must be treated. A me-
chanical system of rotation was therefore selected and by
using this approach, the treatment of 500 mL wastes, con-
taining 5 mg L–1 Cd(II) and Cr(III), with 10 g chitosan was
possible; retention efficiencies of 99.2±0.1% and 83.4±
1%, respectively, were obtained.
Medium High
Crushed Not crushed Crushed Not crushed
58±9% 66±3% 66±13% 18±7%
65±5% 83±7% 81±8% 58±7%
Treatment of laboratory wastes
A qualitative analysis teaching laboratory waste (500 mL) with
118 mg L–1 Co, 1 mg L–1 Cr, 143 mg L–1 Cu, 1,165 mg L–1 Hg,
156 mg L–1 Mn, 141 mg L–1 Ni and 72 mg L–1 Zn was
treated with 10 g chitosan for 30 min and average removal
of the metal ions were 75±4%, 86±4%, 89±2%, 83±2%,
59±3%, 89±2% and 97.4±0.2%, respectively, for all the
elements present in the sample. This indicates the excel-
lent performance of the detoxification procedure based on
the use of chitosan and its ability to work in complex and
high-concentration solutions.
As compared with the efficiency reported for the use of
alumina in the metal removal from laboratory wastes, it
can be concluded that chitosan has a higher retention
power than alumina for all the considered elements, since
it was possible to remove 747 mg of heavy metals in a sin-
gle extraction step.
For comparative purposes, the same sample of an
atomic spectrometry waste (see Table 3) was treated alter-
natively with alumina and with chitosan and the results
confirm the efficiency of both assayed adsorbers for the
removal of heavy metals from diluted aqueous solutions.
Conclusions
Studies developed by using Cr(III) and Cd(II) as ion mod-
els for the removal of heavy metals from aqueous solu-
tions by means adsorption onto alumina and chitosan have
evidenced that 10 g of the solid and 30 min shaking could
be enough for the strong reduction of polluted wastes
from 500 g up to approximately 11 g of wet residues, thus
improving by a factor of 50 the capacity of stocking and
reducing practically in the same way the cost of waste dis-
posal.
On comparing the efficiency of the two assayed adsor-
bers it must be taken into consideration that the organic
nature of chitosan and its easy availability from fisher fac-
tories as a residue, which provides an added value of this
adsorbent as a recycled product and, additionally, avoids
the side-effect of the use of alumina relating to the possi-
bilities of mobilization of alumina in acidic conditions.
On the other hand, chitosan can be used without any pre-
treatment, thus reducing the power consumption required
by the activation of alumina.
In short, the use of chitosan as an adsorber for heavy
metal removal from aqueous contaminated wastes is an
environmentally friendly, highly efficient alternative.

Compared with other adsorbents proposed in the liter-
ature, chitosan is more effective for the removal of Cr(III)
than bone charcoal [11] and as efficient for Cd removal as
coated coffee beans [22]. Under the experimental condi-
tions established in the present study, we obtained better
results than those found in a previous study [20].
However, on considering the adsorption process, the role
played by the presence of chloride, in the case of Cd(II),
and that of partially hydroxylated species of Cr(III) must
also be recognized.
Acknowledgements Authors acknowledge the financial support
of the Dirección General de Enseñanza Superior e Investigación
Científica Project No. PB98–0947-C02–02 and of the Ministerio
de Ciencia y Tecnología Project No. AGL2002–00729. M.C.A. ac-
knowledges the grant from the Delegación de Medioambiente of
the University of Valencia for studies on academic residue mini-
mization. We also thank Miss I.E. Martí-Herrero and J. López-
Burguete for their collaborative work.
References
1. Carbonell V, Salvador A, de la Guardia M (1992) Fresenius J
Anal Chem 342:529–537
2. Marqués MJ, Morales Rubio A, Salvador A, de la Guardia M
(2001) Talanta 53:1229–1239
3. Pannain MC, Santelli RE (1995) Talanta 42:1609
4. Cooper C, Jiang JQ, Ouki S (2002) J Chem Technol Biotech-
nol 77:546–551
825
5. Bailey SE, Olin TJ, Bricka RM, Adrian DD (1999) Water Res
33:2469–2479
6. Kapoor A, Viraraghavan T (1998) J Environ Eng Asce 124:
1020
7. Jeong SY, Lee JM (1998) Bull Korean Chem Soc 19:218
8. Stirk WA, Van Staden J (2000) Botanica Marina 43:467
9. Zouboulis AI, Matis KA, Lazaridis NK (2001) Sep Sci Technol
36:349
10. Ricon P, Lecuyer J, LeCloirec P (1998) Environ Technol 19:
1005–1016
11. Dahbi S, Azzi M, de la Guardia M (1999) Fresenius J Anal
Chem 363:404–407
12. Tseng RL, Wu FC, Juang RS (1999) J Environ Sci Health Part
A 34:1815–1828
13. Schmuhl R, Krieg HM, Keizer K (2001) Water SA 27:1–7
14. Danubies L, Guimon C, Yiaconmi S, Guibol E (2001) Colloids
Surf A 177:2–3
15. Dantas TND, Neto AAD, Moura MCP, Neto ELB, Telemaco
ED (2001) Langmuir 17:4256–4260
16. Kaminski W, Modrzejewska Z (1997) Sep Sci Technol 32:
2659–2668
17. Kawamura Y, Yoshida H, Asai S, Kurahashi I, Tanibe H
(1997) Sep Sci Technol 32:1959–1974
18. Bang SS, Pazirandeh M (1999) J Microencap 16:489–499
19. Bassi R, Prasher SO, Simpson BK (1999) Environ Technol 20:
1177–1182
20. Bassi R, Prasher SO, Simpson BK (2000) Sep Sci Technol 35:
547
21. Jansson Charrier M, Guibal E, Roussy J, Surjous R (1996) Wa-
ter Sci Technol 34:169
22. Minamisawa M, Nakajima S, Mitsue Y, Miyazawa K, Ueno K,
Hatori A, Miyajima S, Hoshino M, Yoshida S, Takai N (2002)
Nippon Kagaka Kaishi 459–461