Reactive & Functional Polymers 48 (2001) 37–51
www.elsevier.com / locate / react
Expansion of adsorption isotherms into equilibrium surface
Case 1: solvent impregnated resins (SIR)
a,
a
b
Joan Serarols *, Jordi Poch , Isabel Villaescusa
a
`
`
`
´ Santalo´ , s /n,
i Matematica
Aplicada, Universitat de Girona, Escola Politecnica
Superior, Av. Lluıs
Departamento d’ Informatica
17071 Girona, Spain
b
´
`
´ Santalo´ , s /n, 17071 Girona, Spain
, Universitat de Girona, Escola Politecnica
Superior, Av. Lluıs
Departamento d’ Enginyeria Quımica
Received 1 July 2000; accepted 5 January 2001
Abstract
In this work, the adsorption of Au(III) and Zn(II) by impregnated resins has been studied in batch and column
experiments. In batch experiments successive contacts for each metal concentration were made until saturation was reached.
Equilibrium points corresponding to each contact brought a different isotherm. By introducing a new variable it was
observed that all the equilibrium points fitted a surface that we call the equilibrium surface. Once the equilibrium surface was
determined it was found that the different isotherms, obtained in batch and column experiments, were fitted on the surface.
Traditional isotherms and the equations of the equilibrium surfaces for zinc(II) and gold(III) adsorption with impregnated
resins have been compared and discussed. The equilibrium surface equation proposed in this work has been proven to be a
good tool for modelling adsorption by impregnated resins in batch and column experiments. 2001 Elsevier Science B.V.
All rights reserved.
Keywords: Gold(III) extraction; Zinc(II) extraction; Impregnated resins; Isotherms; Equilibrium surface
1. Introduction
Adsorption has been widely used to eliminate
metals from aqueous solutions. In an adsorption
process, the adsorption isotherm, the relation
between the amount of solute adsorbed (qe ) and
the remaining concentration in the aqueous
phase (Ce ) is basically important to describe
how adsorbates interact with adsorbents and so
*Corresponding author. Tel.: 134-972-418-413; fax: 134-972418-399.
E-mail address: serarols@ima.udg.es (J. Serarols).
it is critical in optimising and modelling the
process.
In order to obtain the adsorption isotherms,
researchers perform batch or column experiments. Batch experiments can be carried out in
two different ways: (a) a fixed amount of
adsorbent material (carbon, resin, biomaterial,
etc.) is put into contact with solutions containing different concentrations of adsorbate [1–
4]. (b) A fixed volume of a solution containing a
known concentration of adsorbate is put into
contact with different amounts of adsorbent
material [5,6]. In both cases the vessel containing adsorbent and adsorbate is continuously
1381-5148 / 01 / $ – see front matter 2001 Elsevier Science B.V. All rights reserved.
PII: S1381-5148( 01 )00045-1
38
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
agitated. After equilibrium is reached, the residual adsorbate concentration in the solution is
determined by the most suitable technique.
Normally, in this kind of experiment the adsorbent material is free of adsorbate and a
unique contact is performed for each different
adsorbate concentration or the amount of adsorbent material used, respectively. Nevertheless, Veeraraghavan [7] used a preloaded activated carbon for phenol adsorption and Zhou
[8] studied a second adsorption of aromatic
compounds on activated carbon and obtained a
new adsorption isotherm that he called ‘complementary adsorption’. Column experiments
can be performed in fixed or fluidised beds.
Fixed bed experiments consist in packing a
known amount of adsorbent material in a column. Solutions containing different adsorbate
concentrations flow through the column. The
system can be open, that is, fresh solution of a
known adsorbate concentration is flowed
through the column until final adsorbate concentration is equal to the initial concentration
[9,10]. When the system is closed, the initial
solution is recycled, that is, reintroduced into
the column until its concentration becomes
constant [7,11,12]. During the operation, samples are taken out and the adsorbate concentration is determined. Fluidised bed column
experiments consist in putting a known amount
of adsorbent material in a column and pass
upward through the bed of particles a solution
of known concentration at the fluidisation flow
rate. This methodology was used by Nakhla et
al. [13] to adsorb toxic compounds with activated carbon. From the results achieved with
any of the described experiments a set of
equilibrium points are obtained from which the
corresponding isotherms can be determined.
In the last few years, solvent impregnated
resins (SIR) introduced by Warshawsky et al.
[14,15], have been used in metal separation and
recovery processes. Several authors used Amberlite XAD-2 resins impregnated with different
extractants containing different types of functionalities for the selective extraction of metals
from aqueous solutions [16,17]. SIR can be
considered as a link between solvent extraction
and ion-exchange technologies [18], thus, impregnated resins containing this kind of extractant have a similar behaviour to ion-exchange resins [19].
In this work, the adsorption isotherms of
Au(III) and Zn(II) by Amberlite XAD-2 impregnated resins containing triisobutyl phosphine sulphide (TIBPS) and di(2-ethylhexyl)
phosphoric acid (DEHPA), respectively, have
been obtained in batch and column experiments.
In batch experiments, the adsorption was carried
out by successive contacts and in column experiments, by recycling the solution until
equilibrium was reached. The different isotherms obtained from both types of experiments
did not represent the global equilibrium of the
system. Thus, these results led us to study the
equilibrium of those systems by introducing a
new concept that we call equilibrium surface.
2. Experimental
2.1. Reagents and solutions
Amberlite XAD-2 purchased from Rohm and
Haas was washed as described elsewhere [16].
After washing, the resin was sieved to obtain
the particle size used in this work (840–630
mm).
Triisobutyl phosphine sulphide (TIBPS) provided as Cyanex 471 by Cyanamide Co. was
purified by recrystallisation from ethanol–water
mixture. Ethanol (Merck A.R.) was used without further purification.
Di(2-ethylhexyl) phosphinic acid provided by
Laboratory Supplies Poole BH15 1TD was used
without further purification.
A solution of TIBPS of 1300 mg / dm 3 in 66%
ethanol–water and a solution of 12.5 v / v of
DEHPA in acetone were used as impregnation
solution for gold(III) and zinc(II) adsorption,
respectively. In both cases, the impregnation
procedure was the same used in the abovementioned work.
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
Stock solutions of Au(III) were prepared
from solid HAuCl 4 ? 3H 2 O (Aldrich) in a 0.5
mol / dm 3 NaCl solution. The pH in all gold
solutions was pH 2. NaCl and HCl (Merck
A.R.) were used to adjust the ionic strength and
the pH of gold solutions.
Stock solutions of Zn(II) were prepared by
dissolving ZnCl 2 (Merck A.R.) in MilliQ water.
Gold and zinc solutions of 1000 mg / dm 3
from Carlo Erba were used as standard solutions
for atomic absorption spectrophotometry determinations.
2.2. Experimental procedure
2.2.1. Batch experiments
Amounts of 0.2 g of impregnated resin were
put into contact with 20 cm 3 of metal solutions
of different concentrations (19.9–212.9 mg / dm 3
for Au(III) and 10.2–162.5 mg / dm 3 for Zn(II)),
and stirred for 2 h and 24 h for gold and zinc,
respectively. After filtration, the resin was put
again into contact with fresh gold or zinc
solution and the filtrate was kept for the determination of gold or zinc concentration by
atomic absorption spectrometry. For each initial
gold concentration this operation was repeated
eight times until the resin did not adsorb more
gold. For zinc, 17 contacts were performed for
each initial zinc concentration. The amount of
gold and zinc adsorbed was determined by mass
balance. All the experiments were performed at
room temperature twice.
39
2.2.2. Column experiments
The column consisted of a 10-cm glass tube
with an internal diameter of 4 mm. The same
experimental set-up was used in a previous
work [11]. For each experiment 0.2 g of impregnated resin was introduced into the column.
Two pieces of glasswool were used to keep the
resin packed. A volume of 100 cm 3 of metal
solution of different initial metal concentration
(39.3–196.0 mg / dm 3 or 9.4–248.2 mg / dm 3 )
for gold and zinc, respectively, contained in a
recipient continuously agitated was passed
through the column at a flow rate of 2 cm 3 / min.
The outflow metal solutions were recycled back
to the column until the reservoir metal concentration became almost constant. Samples of
2.0 cm 3 of the reservoir solution were taken
every 5 min and the metal concentration was
analysed by atomic absorption spectrometry.
The amount of metal adsorbed on the resin was
determined by mass balance. All the experiments were carried out twice at room temperature.
3. Results and discussion
3.1. Batch experiments
The results obtained for Au(III) and Zn(II)
adsorption are presented in Tables 1 and 2. In
these tables can be seen the initial metal concentration C0 , number of contact i, the final
Table 1
Gold adsorption by impregnated resin TIBPS / XAD-2
C0
19.97
i
Ce,i
qe,i
43.15
Ce,i
qe,i
Ce,i
75.97
qe,i
101.87
Ce,i
qe,i
134.79
Ce,i
qe,i
161.01
Ce,i
qe,i
193.07
Ce,i
qe,i
Ce,i
qe,i
1
2
3
4
5
6
7
8
0.96
1.55
5.23
10.46
12.77
15.85
17.77
19.68
1.90
3.74
5.22
6.17
6.89
7.30
7.52
7.53
1.31
12.07
19.34
28.57
31.37
31.79
34.95
41.12
4.18
7.29
9.67
11.13
12.31
13.44
14.26
14.28
8.47
36.55
50.95
52.35
64.69
66.49
69.31
73.12
6.75
10.69
13.19
15.56
16.68
17.63
18.30
18.32
23.12
66.49
78.54
81.87
91.62
94.95
101.10
101.47
7.88
11.41
13.75
15.75
16.77
17.46
17.54
17.55
41.02
78.03
107.77
106.23
126.05
134.79
134.21
134.78
9.38
15.05
17.76
20.61
21.49
21.49
21.54
21.54
54.73
114.98
131.88
135.96
155.77
157.52
156.35
160.86
10.63
15.23
18.14
20.65
21.17
21.52
21.99
22.00
84.69
139.45
160.43
166.26
177.33
186.66
193.07
193.06
10.84
16.20
19.46
22.14
23.72
24.36
24.36
24.36
104.44
159.27
180.24
186.07
197.14
206.47
212.30
212.62
10.84
16.21
19.47
22.15
23.72
24.36
24.42
24.43
C0 in mg Au(III) / dm 3 , Ce,i in mg Au(III) / dm 3 , qe,i in mg Au(III) / g XAD-2; i, number of contact.
212.88
40
C0 (mg Zn(II)/dm 3 )
10.25
i
Ce,i
mg Zn(II)/
dm 3
qe,i
mg Zn(II)/g
XAD-2
Ce,i
mg Zn(II)/
dm 3
qe,i
mg Zn(II)/g
XAD-2
Ce,i
mg Zn(II)/
dm 3
qe,i
mg Zn(II)/g
XAD-2
Ce,i
mg Zn(II)/
dm 3
qe,i
mg Zn(II)/g
XAD-2
Ce,i
mg Zn(II)/
dm 3
qe,i
mg Zn(II)/g
XAD-2
Ce,i
mg Zn(II)/
dm 3
qe,i
mg Zn(II)/g
XAD-2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
1.25
2.12
3.05
3.98
4.86
5.55
6.21
6.97
7.05
7.91
8.23
9.05
9.96
10.07
10.18
10.20
10.20
0.90
1.71
2.43
3.06
3.60
4.07
4.47
4.80
5.12
5.36
5.56
5.68
5.71
5.72
5.73
5.74
5.74
9.00
19.00
22.16
29.51
31.94
33.68
35.42
34.72
36.11
37.64
38.19
38.89
39.72
40.15
40.25
40.30
40.32
3.13
5.27
7.09
8.17
9.01
9.68
10.17
10.73
11.15
11.42
11.64
11.78
11.84
11.86
11.87
11.88
11.88
25.71
42.14
53.68
57.58
61.36
65.15
66.67
68.18
69.70
70.45
71.50
71.57
71.57
71.57
71.57
71.57
71.57
4.59
7.53
9.32
10.73
11.75
12.39
12.88
13.23
13.41
13.53
13.54
13.54
13.54
13.54
13.54
13.54
13.54
47.22
74.32
85.14
89.19
92.78
94.44
96.11
100.00
101.71
102.20
102.23
102.23
102.23
102.23
102.23
102.23
102.23
5.51
8.30
10.01
11.32
12.27
13.05
13.67
13.90
13.95
13.96
13.96
13.96
13.96
13.96
13.96
13.96
13.96
51.43
107.14
113.57
125.71
127.71
128.57
129.14
131.43
132.86
132.86
132.86
132.86
132.86
132.86
132.86
132.86
132.86
8.15
10.73
12.67
13.40
13.92
14.36
14.74
14.89
14.90
14.91
14.91
14.91
14.91
14.91
14.91
14.91
14.91
80.11
139.77
143.18
146.59
151.14
153.41
158.64
160.23
160.68
162.50
161.36
162.50
162.50
162.50
162.50
162.50
162.50
8.24
10.51
12.44
14.04
15.17
16.08
16.47
16.69
16.88
16.88
16.99
16.99
16.99
16.99
16.99
16.99
16.99
40.34
71.59
102.27
132.95
162.50
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
Table 2
Zinc adsorption by impregnated resin DEHPA / XAD-2
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
concentration at equilibrium for each contact
Ce,i , and the total adsorbed metal amount at
each contact taking into account the accumulated amount of metal adsorbed in the previous
contact, qe,i . The qe,i were calculated with the
following expression:
FO
G
n
sC0 2 Ce,id .V
i 51
qe,i 5 ]]]]]]
WXAD-2
(1)
where, V is the volume of gold or zinc solution
(dm 3 ) and WXAD-2 the amount of impregnated
resin used (g).
In these tables, it can be observed that the
major percentage of adsorption takes place
during the first four and 10 contacts for gold
and zinc, respectively (see the increase of Ce,i ).
This percentage decreases with the successive
contacts until the solution concentration coincides with the initial concentration. This fact
indicates that for all the studied concentrations
the resin does not adsorb more metal.
Concerning the total metal amount adsorbed
it can be seen that for each contact at the
beginning qe,i, increases with the initial concentration. Up to a determined value of C, qe,i
becomes almost constant. This fact indicates
that the resin cannot adsorb more metal. The
41
total amount of metal adsorbed in the resin was
24.43 mg Au(III) / g XAD-2 and 16.99 mg
Zn(II) / g XAD-2.
3.2. Column experiments
The results obtained in column experiments
for Au(III) and Zn(II) adsorption are presented
in Tables 3 and 4.
In these tables, the time at which samples
were taken out for their analysis can be seen in
the first column. The total operation time was
60 and 80 min for gold and zinc, respectively.
The metal initial concentration is indicated in
the first row of these tables (t50). In these
tables, C shows the metal solution concentration
in the recycling recipient at different times and
q the amount of adsorbed metal on the resin.
When calculating the q, the sample volume used
for metal analysis by atomic absorption has
been taken into account (2 ml per sample).
As it can be seen in Table 3, for each initial
concentration used, the amount of gold adsorbed
on the resin increases during the adsorption
process up to a value of 24.30 mg Au(III) / g
XAD-2 for the initial concentration of 196 mg
3
Au(III) / dm . This value fairly matches the
obtained one in the batch experiments, 24.43
mg Au(III) / g XAD-2 (see Table 1).
Table 3
Results in column adsorption of gold by impregnated resin TIBPS / XAD-2
Time
(min)
C
mg Au(III)/
dm 3
q
mg Au(III)/g
XAD-2
C
mg Au(III)/
dm 3
q
mg Au(III)/g
XAD-2
C
mg Au(III)/
dm 3
q
mg Au(III)/g
XAD-2
C
mg Au(III)/
dm 3
q
mg Au(III)/g
XAD-2
C
mg Au(III)/
dm 3
q
mg Au(III)/g
XAD-2
0
5
10
15
20
25
30
35
40
45
50
55
60
39.32
36.32
35.71
33.92
32.01
28.74
26.64
24.43
22.72
22.21
21.50
21.22
21.21
0.00
1.51
1.82
2.70
3.61
5.12
6.02
7.06
7.75
7.92
8.22
8.31
8.32
78.62
69.42
65.34
64.65
63.10
59.81
56.90
52.84
51.75
48.82
46.97
46.62
46.55
0.00
3.22
5.71
7.02
7.73
9.36
10.74
12.69
13.12
14.41
15.36
15.48
15.50
118.01
113.93
108.14
98.34
98.32
92.18
85.35
83.68
80.91
79.50
77.81
76.54
76.44
0.00
2.92
5.43
8.87
9.81
12.79
15.94
16.72
18.04
18.67
19.48
20.08
20.10
157.03
143.02
138.20
133.44
131.54
123.92
122.42
121.04
116.76
112.47
113.46
112.42
112.50
0.00
5.65
8.28
10.85
12.73
16.42
17.12
17.77
19.84
21.82
21.49
21.88
21.90
196.03
187.35
182.48
174.83
165.23
159.65
157.69
155.55
153.40
149.33
147.95
147.32
147.30
0.00
6.15
8.52
11.63
15.31
18.06
18.07
20.08
21.01
23.08
23.78
24.29
24.30
42
Time
C
q
(min)
mg Zn(II)/
mg Zn(II)/g mg Zn(II)/ mg Zn(II)/g mg Zn(II)/l mg Zn(II)/g mg Zn(II)/ mg Zn(II)/g
C
q
C
q
C
q
dm 3
XAD-2
dm 3
XAD-2
0
9.43
0.00
36.96
0.00
5
7.39
1.05
33.93
1.52
10
5.94
1.76
31.96
15
5.29
2.07
30.71
20
4.79
2.31
25
4.13
30
C
q
C
C
q
mg Zn(II)/
mg Zn(II)/g
mg Zn(II)/ mg Zn(II)/g
q
C
mg Zn(II)/ mg Zn(II)/g
q
C
mg Zn(II)/ mg Zn(II)/g
q
mg Zn(II)/
mg Zn(II)/g
dm 3
XAD-2
dm 3
dm 3
dm 3
dm 3
XAD-2
XAD-2
dm 3
XAD-2
70.24
0.00
97.62
0.00
130.95
0.00
158.33
0.00
189.54
0.00
218.45
0.00
248.21
0.00
67.00
1.62
92.86
2.38
126.19
2.38
144.05
7.14
170.93
9.30
194.64
11.90
224.41
11.90
2.48
65.48
2.37
90.48
3.55
120.24
5.30
140.48
8.89
161.63
13.86
191.67
13.36
218.45
14.82
3.08
61.91
4.08
88.10
4.69
114.29
8.15
128.57
14.61
159.30
14.98
188.69
14.79
215.48
16.25
28.57
4.09
57.86
5.98
85.71
5.81
107.14
11.51
126.19
15.72
154.65
17.16
182.74
17.59
212.50
17.65
2.61
26.49
5.05
55.24
7.19
83.33
6.90
104.76
12.61
125.00
16.27
152.33
18.23
180.95
18.41
210.12
18.74
3.00
3.12
24.58
5.90
52.38
8.47
78.57
9.05
102.38
13.68
125.00
16.27
152.33
18.23
180.95
18.41
210.12
18.74
35
2.29
3.43
25.60
5.46
51.07
9.05
76.19
10.10
100.00
14.73
124.29
16.59
152.33
18.23
180.95
18.41
210.12
18.74
40
1.57
3.74
23.81
6.23
49.64
9.66
73.81
11.12
97.62
15.75
124.29
16.59
152.33
18.23
180.95
18.41
210.12
18.74
45
1.57
3.74
21.07
7.38
47.00
10.77
71.43
12.12
97.62
15.75
124.29
16.59
152.33
18.23
180.95
18.41
210.12
18.74
50
1.43
3.80
21.49
7.20
47.00
10.77
71.43
12.12
97.62
15.75
124.29
16.59
152.33
18.23
180.95
18.41
210.12
18.74
55
1.43
3.80
20.83
7.47
47.00
10.77
71.43
12.12
97.62
15.75
124.29
16.59
152.33
18.23
180.95
18.41
210.12
18.74
60
1.43
3.80
20.83
7.47
47.00
10.77
71.43
12.12
97.62
15.75
124.29
16.59
152.33
18.23
180.95
18.41
210.12
18.74
65
1.43
3.80
20.83
7.47
47.00
10.77
71.43
12.12
97.62
15.75
124.29
16.59
152.33
18.23
180.95
18.41
210.12
18.74
70
1.43
3.80
20.83
7.47
47.00
10.77
71.43
12.12
97.62
15.75
124.29
16.59
152.33
18.23
180.95
18.41
210.12
18.74
75
1.43
3.80
20.83
7.47
47.00
10.77
71.43
12.12
97.62
15.75
124.29
16.59
152.33
18.23
180.95
18.41
210.12
18.74
80
1.43
3.80
20.83
7.47
47.00
10.77
71.43
12.12
97.62
15.75
124.29
16.59
152.33
18.23
180.95
18.41
210.12
18.74
XAD-2
XAD-2
XAD-2
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
Table 4
Results in column adsorption of zinc by impregnated resin DEHPA / XAD-2
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
In Table 4 it can be seen that the amount of
zinc that the resin adsorbs for each initial
concentration is also increasing during the adsorption process up to 18.74 mg Zn(II) / g XAD2 corresponding to the initial concentration of
248.21 mg Zn(II) / dm 3 . This value is greater
than that obtained in batch experiments, 16.99
mg Zn(II) / g XAD-2 (see Table 2).
3.3. Isotherms
Previous studies demonstrated that the Langmuir isotherm fits the experimental data obtained for Au(III) and Zn(II) adsorption by
XAD-2 impregnated resins with TIBPS and
DEHPA, respectively [11,12]. The Langmuir
isotherm is represented by the equation:
bC
aC
qe 5 qmax ]]] 5 ]]].
1 1 bC 1 1 bC
(2)
In batch experiments, as said before, most of
the authors use one contact to determine the
adsorption isotherm. Thus, the first step was to
determine the isotherm that fitted the experimental data obtained in the first contact for
both systems studied. The Langmuir coefficients
43
Table 5
Langmuir coefficients of first contact isotherm
Metal
a
b
R2
Au(III)
Zn(II)
1.9670
0.8203
0.1720
0.1164
0.9954
0.9919
(a, b) as well as the correlation coefficient R 2
for gold and zinc are presented in Table 5. As
can be seen, the experimental data fit quite well
the Langmuir equation. These isotherms were
plotted in Fig. 1. In the same figure, the
equilibrium points obtained in the successive
batch contacts were also plotted. It can be
noticed that these isotherms fit the data corresponding to the first contact but do not fit the set
of equilibrium data obtained by the successive
contacts.
The next step was the determination of the
Langmuir isotherms corresponding to the experimental data obtained in the successive batch
contacts and in column. As a result of this, for
each contact a different isotherm was reached.
As an example, in Fig. 2 the total experimental
data and the Langmuir isotherm 1 (1st batch
contact), Langmuir isotherm 2 (2nd contact)
and column Langmuir isotherm have been
Fig. 1. Gold and zinc Langmuir 1st contact isotherms with all batch equilibrium data.
44
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
Fig. 2. Gold and zinc 1st and 2nd contact and column Langmuir isotherms.
plotted. It can be observed that each isotherm
fits only the corresponding experimental data.
From these results, it can be concluded that:
• none of the isotherms fits the set of experimental data;
• the amount of metal adsorbed on the resin
(qe,i ) seems to be dependent on the number
of contacts; qe,i increases with the number of
contacts but the rate of this increase lowers
in each contact.
Observing these conclusions, we thought that
another variable related to the number of contacts could have influenced the process. By
introducing a new variable we may obtain a
new equation capable of describing the equilibrium. This new variable could be the time of the
experiment operation, the initial metal concentration or the previous metal preload on the
resin at the beginning of each contact. Considering that the time used was necessary for the
system to reach the equilibrium, it may be
considered that once the equilibrium is reached
the adsorption process is independent of the
operation time. On the other hand, the initial
concentration does not vary according to the
number of contacts and it is taken into account
for each series of data, so, the obtained data are
related to this variable.
Finally, the only possible variable is qp , that
is, the metal preload on the resin at the beginning of each contact. This variable can be easily
quantified and some other authors have considered it. Zhou studied a second adsorption of
phenol on activated carbon [8] and Veeraraghavan studied the phenol adsorption on a
25% preloaded activated carbon [7]. Taking into
account this third variable and plotting C, qp
and qe , we obtain a surface instead of a curve.
4. Equilibrium surface
In a metal adsorption process by impregnated
resins, we call the equilibrium surface, the
surface that contains all the possible equilibrium
points, at a constant temperature, obtained in
batch and column experiments with independence of the experimental procedure used.
The equation describing this surface should
be a function of C and qp :
qe 5 F(C, qp )
(3)
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
where qe (mg / g) is the amount of metal per
weight of impregnated resin, C (mg / dm 3 ) is the
metal concentration in the liquid phase and qp
(mg / g) the preloaded amount of metal per
weight of impregnated resin.
In order to determine the equilibrium surface
for a given system the following assumptions
were made.
• There is no desorption process, thus, if the
preloaded amount of metal qp is greater than
the maximum amount of metal that can be
adsorbed for a given concentration then the
amount of metal per weight of impregnated
resin must be equal to the preload (qe 5 qp ).
This maximum amount of metal adsorbed is
a function of the concentration C, h(C). In
particular, if the concentration is nil (C50),
h(C)50 and the adsorbed amount coincides
with the preload (qe 5 qp ).
• When the resin is not preloaded, qp 5 0, at
equilibrium, the amount of the adsorbed
metal qe must coincide with the value obtained from the intersection of function (3)
with the plane Cqe . This intersection curve
coincides with the isotherm of first contact in
batch.
Taking into account these assumptions, the
surface equation must be:
qe 5 F(C, qp ) 5
H
qp 1 f(C, qp ) if qp , h(C)
qp if qp $ h(C)
(4)
And, if qp 5 h(C) then f(C, qp ) 5 0.
In the studied systems, the isotherms of the
first contacts fit a Langmuir curve, thus, the
equation of the equilibrium surface (4) when
qp 5 0 must be:
aC
qe 5 F(C, 0) 5 ]]].
1 1 bC
(5)
Therefore, the intersection of the equilibrium
surface with the plane Cqe is a Langmuir curve.
And considering the experimental data obtained by successive contacts (all of them fit a
45
Langmuir curve type) an expression of Langmuir type is expected for h(C):
k5C
h(C) 5 ]]]
1 1 k6C
(6)
and the function f(C, qp ) will also be a Langmuir curve with correction factors
A(qp )(C 2 C*)
f(C, qp ) 5 ]]]]]]
1 1 B(qp )(C 2 C*)
(7)
where C* is the concentration value determined
by the equation qp 5 h(C*).
On the other hand, from one contact to
another the amount of adsorbed metal on the
resin diminishes until the resin does not adsorb
more metal thus, at the end of the process
Dq 5 0 that implies qe 5 qp . Then, the coefficients A(qp ) and B(qp ) of the expression (7) are
considered as:
A(qp ) 5 k 1 e 2k 3 q p
B(qp ) 5 k 2 e
2k 4 q p
(8)
.
(9)
If qp , h(C), then by isolating C of expression (6) the C* value can be calculated:
qp
k5C
qp 5 h(C) 5 ]]] ⇒ C* 5 ]]].
1 1 k6C
k 5 2 k 6 qp
(10)
By introducing the expressions (6)–(10) in
Eq. (4) a possible expression of the equilibrium
surface for the studied systems would be:
qe 5 F(C, qp )
5
S
S
D
D
qp
k 1 e 2k 3 q p C 2 ]]
k 5 2 k 6 qp
k5C
qp 1 ]]]]]]
if qp , ]]
qp
1 1 k6C
2k 4 q p
1 1 k2e
C 2 ]]
k 5 2 k 6 qp
5
(11)
k5C
qp if qp $ ]]
1 1 k6C
where qe is the amount of metal adsorbed in the
resin, mg Au(III) or Zn(II) / g XAD-2; qp is the
amount of metal preload in the resin, mg
Au(III) or Zn(II) / g XAD-2; C is the solution
concentration at equilibrium, mg Au(III) or
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
46
Zn(II) / dm 3 and k 1 , k 2 , k 3 , k 4 , k 5 and k 6 are
parameters to be determined.
The surface defined in this way verifies the
above made assumptions.
• When qp $ h(C) 5 k 5 C /(1 1 k 6 C) and C →0,
high preload and low concentration, then,
qe 5 F(C, qp ) 5 qp , and in particular, if the
concentration C 5 0, qe 5 F(0, qp ) 5 qp . In
this case, the amount of metal on the resin is
equal to the preload (qp ), and the intersection
of the equilibrium surface with the plane
qp qe is the bisecting first quadrant.
• If the preload of metal on the resin is nil
(qp 5 0), the equilibrium surface equation is
the Langmuir isotherm obtained in the first
contact:
k1C
qe 5 F(C, 0) 5 ]]]
1 1 k2C
(12)
• For a given preload on the resin (qp 5
constant) and for high concentrations
(C → `), the following equation is obtained:
k 1 2(k 3 1k 4 )q p
e
sqed C →` 5 qp 1 ]
k
2
5 qp 1 qmax
(13)
where qmax 5 k 1 /k 2 e 2(k 3 1k 4 )q p .
The expression (13) indicates that the maximum amount of metal (gold or zinc) that the
resin can adsorb depends on the parameters k 1 ,
k 2 , k 3 , k 4 and on the initial preload (qp ). In
particular, if qp 5 0, the maximum adsorption
capacity would be attained for the first contact:
C →`
k1
5 qmax, 0 .
sqed q p 50 5 ]
k
(14)
2
In the expression (14) it can be observed that
the maximum amount of metal adsorbed on the
resin, without preload and for the first contact,
is given by the quotient of the parameters k 1 and
k 2 of the proposed equilibrium surface equation.
Therefore, in this case, the coefficients k 1 and
k 2 have the same meaning as the parameters a
and b of the Langmuir isotherm equation.
The six parameters k 1 , k 2 , k 3 , k 4 , k 5 and k 6 of
the expression (11) must always be positive in
order to avoid discontinuities in the surface.
From the chemical point of view: k 1 /k 2 is the
maximum amount of metal that the resin can
adsorb in the first contact (14). k 2 is the
adsorption affinity of the resin, similar to the
corresponding parameter (b) of the Langmuir
isotherm (12). k 3 , k 4 are correctors of the
maximum increase of the adsorption and the
affinity while the preload is increasing (13). k 5 ,
k 6 provide, for a given concentration, the maximum preload up to the one where there will not
be any adsorption (6).
4.1. Determination of the coefficients for the
two studied systems
In order to determine the coefficients of the
surface equation, only the equilibrium points
obtained in batch experiments were considered
(Tables 1 and 2). For each contact i, the used
points are the values (C, qp , qe ) where C 5 Ce,i
was considered as the equilibrium concentration, qp 5 qe,i21 as the amount of metal preload
on the resin and qe 5 qe,i as the amount of metal
adsorbed on the resin.
In order to obtain the coefficient values, the
sum square errors were minimised:
Os q
N
SSE 5
j 51
e, j
2 F(Cj , qp, j )d 2
(15)
where qe, j is the experimental data; F(Cj , qp, j )
are calculated values from the equilibrium surface Eq. (11) and N is the number of experiences.
From the batch experiments data indicated in
Tables 1 and 2, a non-lineal fit was done by
means of the program SPSS 7.5 for Windows in
order to obtain the values of the parameters k 1 ,
k 2 , k 3 , k 4 , k 5 and k 6 . These values are presented
in Table 6 for each system.
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
47
Table 6
Equilibrium surface parameters. Maximum amount of metal adsorbed (qmax,0 ) in the first contact without preload
Metal
k1
k2
k3
k4
k5
k6
qmax,0
mg (metal) / g XAD-2
R2
Au(III)
Zn(II)
1.3108
0.4638
0.1110
0.0469
0.2172
0.3000
0.1774
0.2411
0.9325
1.1519
0.0328
0.0625
11.8090
9.8891
0.9819
0.9942
4.2. Validation and fitness of the equilibrium
surface equation
In Fig. 3 the surface and the experimental
data (batch and column) are presented for gold
and zinc, respectively.
When plotting the equilibrium points corresponding to the batch experiments, (C, qp , qe ),
we considered C 5 Ce,i , qp 5 qe,i21 (amount of
adsorbed metal in previous contacts) and qe 5
qe,i .
As said before, the results obtained in column
experiments were not taken into account to
determine the surface equation. Nevertheless,
the column equilibrium points were included in
Fig. 3. When doing this, it was considered that
in a process with or without recycling for a time
t there is a concentration C(t) in the liquid phase
and a load q(t) in the resin that can be considered as the preload corresponding to the time t.
When time increases C(t) tends to Ce and q(t)
tends to qe . Hence, qp has been considered to be
equal to qe . Thus, the equilibrium points obtained in the column were represented as (C, qe ,
qe ).
When observing Fig. 3, all the experimental
data seem to be placed on the corresponding
surface. In order to confirm that these surfaces
describe the equilibrium with accuracy, we
plotted the qe,cal calculated vs. the experimental
qe,exp . As can be seen in Fig. 4 the points
(included the points obtained in column experiments) are distributed on the bisecting first
quadrant. The fact that the equilibrium points of
the column experiments fit properly the surface
may validate it. The corresponding regression
equations turned out to be qe,cal 5 0.9982
qe,exp 1 0.0016 (R 2 5 0.9916), and qe,cal 5
1.0095 qe,exp 2 0.0826 (R 2 5 0.9974) for gold
and zinc, respectively.
The comparison of the isotherms with the
equilibrium surface can be done in two ways.
Fig. 3. Equilibrium surface and experimental data (batch and column) for: (a) gold, (b) zinc.
48
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
Fig. 4. Calculated vs. experimental qe values: (a) gold, (b) zinc.
1. Taking into account the isotherms obtained
in Section 3.3 and the introduction of the
new variable (qp , preload) it is possible to
get a curve in the space (C, qp , qe ) whose
projection on the plane Cqe is the corresponding isotherm.
For each batch contact i the points (Ci ,
qe,i 21 , qe,i ) were fitted to a space curve that
has the form si (C) 5 (C, qp,i (C), qe,i (C))
where qp,i (C) 5 a 9i C /(1 1 b i9 C) and qe,i (C) 5
a i C /(1 1 b i C) (ith contact isotherm) (see
Fig. 5). The projection of this si (C) on the
plane Cqe corresponds to the isotherm of
contact i. In this sense, si (C) is the plotting
of the isotherm i on the equilibrium surface.
In the case of the column the same procedure
was followed. The points (C, qe , qe ) were
fitted to a space curve sc (C) 5 (C, qp,c (C),
qe,c (C)) where qp,c (C) 5 qe,c (C) 5 aC /(1 1
bC) (column isotherm). In Fig. 5, the surface
and the space curves si (C) and sc (C) have
been plotted. It can be observed that the set
of curves is properly placed on the respective
equilibrium surfaces (gold and zinc). It can
also be observed that in both systems the
isotherm corresponding to the first contact
Fig. 5. Gold and zinc isotherms on the equilibrium surface: (a) gold, (b) zinc.
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
practically coincides with the intersection of
the surface and the plane Cqe .
2. It is also possible, from the equilibrium
surface, to obtain the curve on the surface
that corresponds to a given experimental
process (i.e. the curve of the 3rd contact in
batch) and project it on the plane Cqe for its
comparison to the isotherm obtained with the
experimental data.
In order to obtain the curve corresponding
to the contact in batch i the procedure is the
following:
for
i51
qp,1 (C) 5 0 qe,1 (C) 5 F(C, 0)
i52
qp,2 (C) 5 qe,1 (C) qe,2 (C) 5 F(C, qp,2 )
i53
qp,3 (C) 5 qe,2(C ) qe,3 (C) 5 F(C, qp,3 )
:
i5n
qp,n (C) 5 qe,n 21 (C) qe,n (C) 5 F(C, qp,n )
And the curve on the surface for a contact
i is:
si (C) 5 (C, qp,i (C), qe,i (C))
(16)
where qp,i (C) 5 qe,i21 (C) and qe,i (C) 5 F(C,
qp,i ), called the theoretical isotherm (see Fig.
6a.1 and b.1).
When this curve si (C) is projected on the
plane Cqe , the following curve is obtained:
di (C) 5 (C, qe,i (C))
(17)
that must be interpreted as the isotherm of
contact i obtained from the equilibrium
surface equation, called the calculated isotherm (see Fig. 6a.1 and b.1).
In the case of column experiments, taking
into account that qe,c 5 qp,c , the curve on the
surface is
sc (C) 5 (C, qp,c (C), qe,c (C))
(18)
where qe,c (C) 5 qp,c (C) 5 h(C) is called the
theoretical isotherm (see Fig. 6a.2 and b.2).
49
When the curve sc (C) is projected on the
plane Cqe , the following curve is obtained:
dc (C) 5 (C, h(C))
(19)
that must be interpreted as the column
isotherm obtained from the equilibrium surface equation, called the calculated isotherm
(see Fig. 6a.2 and b.2).
In Fig. 6 the isotherms (calculated and
experimental) corresponding to the 3rd contact and the column are presented. In both
cases experimental and theoretical isotherms
almost coincide. Similar results were found
for the rest of the contacts.
5. Conclusions
From the results, it has been demonstrated
that the traditional isotherms do not permit us to
represent all the equilibrium points obtained in
the adsorption of gold and zinc by the impregnated resins used in this work. In this study,
the concept of equilibrium surface has been
defined and developed. The equilibrium surface
equation has been obtained for the two studied
systems and the corresponding parameters determined. The two calculated equilibrium surfaces for gold and zinc have successfully fitted
all the equilibrium points obtained for each
system. Their fitness has been demonstrated by
comparison of both theoretical and calculated
isotherms. The equilibrium surface has been
validated with the equilibrium points obtained
in column experiments. Thus, the equilibrium
surface equation proposed in this work has been
proven to be a good tool for modelling adsorption by impregnated resins in batch and
column experiments.
As the general equation has been proposed
from very general assumptions, it could be valid
for other adsorption systems. The significance
of this study is to provide a tool to be applied in
the modelling of adsorption processes.
50
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
Fig. 6. Theoretical and calculated isotherms on equilibrium surface. (a.1) Gold 3rd contact isotherm, (a.2) gold column isotherm; (b.1) zinc
3rd contact isotherm, (b.2) zinc column isotherm.
6. Notation
a, b
C
C0
Ce,i
qe
qe,i
qp
isotherm parameters
gold or zinc concentration in the
liquid phase (mg Au(III) or Zn(II) /
dm 3 )
initial concentration in the liquid
phase (mg Au(III) or Zn(II) / dm 3 )
equilibrium concentration for a contact i in liquid phase (mg Au(III) or
Zn(II) / dm 3 )
gold or zinc concentration in the
resin (mg Au(III) or Zn(II) / g)
gold or zinc concentration in the
resin for a contact i (mg Au(III) or
Zn(II) / g)
gold or zinc concentration pre-
qe,cal
qe,exp
qmax
ki
N
t
V
WXAD-2
loaded in the resin (mg Au(III) or
Zn(II) / g)
gold or zinc concentration calculated by the surface equilibrium
equation (mg Au(III) or Zn(II) / g)
gold or zinc experimental concentration (mg Au(III) or Zn(II) / g)
gold or zinc maximum concentration in the resin (mg Au(III) or
Zn(II) / g)
equilibrium surface parameters, i5
1, 2, 3, 5, 6
number of experiences
time (s)
sample volume (dm 3 )
resin weight (g)
J. Serarols et al. / Reactive & Functional Polymers 48 (2001) 37 – 51
Acknowledgements
This work was supported by the University of
Girona project (UdG-9100087). The assistance
´
of Ms. Montserrat Caceres
with the laboratory
work is gratefully acknowledged and Mr.
Christian Serarols for English language revision.
References
[1] E.G. Furuya, H.T. Chang, Y. Miura, H. Yokomura, S. Tajima,
S. Yamashita, K.E. Noll, Intraparticle mass transport mechanism in activated carbon adsorption of phenols, J. Environ.
Eng. 122 (10) (1996) 909–916.
[2] H. Tamura, M. Kudo, R. Furuichi, Polyfunctionality of
carboxyl sites on IRC-50, a MR-type ion exchange resin,
evaluated by modelling with the Frumkin isotherm, React.
Funct. Polym. 38 (1998) 177–181.
[3] E. Guibal, C. Milot, J.M. Tobin, Metal-anion sorption by
chitosan beads: equilibrium and kinetic studies, Ind. Eng.
Chem. Res. 37 (1998) 1454–1463.
[4] R.-L. Tseng, F.-C. Wu, R.-S. Juang, Effect of complexing
agents on liquid-phase adsorption and desorption of copper(II) using chitosan, J. Chem. Tech. Biotechnol. 74 (1999)
533–538.
[5] R. Ganho, H. Gibert, H. Angelino, Cinetique de l’Adsortion
du Phenol en Couche Fluidisee´ de Charbon Actif, Chem.
Eng. Sci. 30 (1975) 1231–1238.
¨
[6] B. Eichenmuller,
G. Bunke, K. Behrend, R. Buchholz, P.
¨ Adsorption of acenaphthene on porous organic polyGotz,
mers, J. Environ. Eng. 123 (9) (1997) 836–841.
[7] S. Veeraraghavan, Modelling and Experimental Verification
of Models of Liquid–Solid Fluidised-Bed and SemifluidizedBed Adsorbers: a Reaction Engineering Approach, Kansas
State University, Manhattan, KS, 1989, Ph.D. Dissertation.
´
[8] M.L. Zhou, Modelisation
de l’Adsorption sur Charbon Actif,
` Docteur, No. 724.
Universite´ de Rennes I, 1992, These
51
[9] R.-T. Huang, T.-L. Chen, H.-S. Weng, Binary adsorption in a
fixed-bed column packed with an ion-exchange resin, Sep.
Sci. Technol. 30 (13) (1995) 2731–2746.
[10] J. Ho, P.C. Wankat, N.-H.L. Wang, Pore and surface diffusion
and bulk-phase mass transfer in packed and fluidised beds,
Ind. Eng. Chem. Res. 37 (1) (1998) 228–239.
[11] J. Serarols, J. Poch, M.F. Llop, I. Villaescusa, Determination
of effective diffusion coefficient for gold(III) on a macroporous resin XAD-2 impregnated with triisobutyl phosphine
sulphide, React. Funct. Polym. 41 (1999) 27–35.
[12] J. Serarols, J. Poch, I. Villaescusa, Determination of the
effective diffusion coefficient of Zn(II) on a macroporous
resin XAD-2 impregnated with di-2-ethylhexyl phosphoric
acid (DEHPA). Influence of metal concentration and particle
size, in: Poster presented POC’2000, Tianjin, China, 2000.
[13] G.F. Nakhla, M.T. Suidan, Modelling of toxic wastewater
treatment by expanded-bed anaerobic GAC reactors, J.
Environ. Eng. 118 (4) (1992) 495–511.
[14] A. Warshawsky, A. Patchornik, Recent developments in
metal extraction by solvent impregnated resins, in: M. Streat
(Ed.), The Theory and Practice of Ion Exchange, SCI,
London, 1976, p. 38.1.
[15] A. Warshawsky, Extraction with solvent impregnated resins,
in: J.A. Marinsky, Y. Marcus (Eds.), Ion Exchange and
Solvent Extraction, Vol. 8, Dekker, New York, 1981, pp.
230–310.
´ J. De Pablo, M. Valiente, M.
[16] I. Villaescusa, V. Salvado,
Aguilar, Liquid–solid extraction of gold(III) from aqueous
chloride solutions by macroporous resins impregnated with
triisobutyl phosphine sulfide (CYANEX-471), React. Polym.
17 (1992) 69–73.
[17] J.L. Cortina, R. Arad-Yellin, N. Miralles, A.M. Sastre, A.
Warshawsky, Kinetics studies on heavy metal ions extraction
by Amberlite XAD-2 impregnated resins containing a
bifunctional organophosphorous extractant, React. Funct.
Polym. 38 (1998) 269–278.
[18] J.L. Cortina, A. Warshawsky, in: J.A. Marinsky, Y. Marcus
(Eds.), Ion Exchange and Solvent Extraction, Vol. 13,
Dekker, New York, 1997, p. 195.
[19] F. Helfferich, in: Ion Exchange, McGraw-Hill, New York,
1962, pp. 234–319.