Journal of Colloid and Interface Science 247, 507–510 (2002)
doi:10.1006/jcis.2001.8204, available online at http://www.idealibrary.com on
LETTER TO THE EDITOR
Describing Adsorption of Paracetamol from Aqueous Solution on Carbons
While Utilizing the Most Widespread Isotherm Models—The Impact
of Surface Carbonyl and Basic Groups
2. MATERIALS AND PARACETAMOL ADSORPTION DATA
The 15 most widespread adsorption isotherm equations are applied for describing recently published paracetamol adsorption data
from aqueous solutions (pH 7). Twelve adsorption isotherms, measured at 300, 310, and 320 K, on the series of chemically modified
carbons D43/1 (Carbo-Tech, Essen, Germany) differing in surface
properties (from basic to strongly acidic) but possessing almost the
same porosity, are analyzed. The results of fitting theoretical models
to experimental data are arranged according to a decrease in the
average value of the determination coefficient. From the models
studied the best fit is obtained for Weber–Vliet, Dubinin–Astakhov,
and the model published by Jossens. The most important conclusion is that at the lowest temperature studied, where the effect of
carbon surface composition on adsorption properties is the most
strongly marked, the value of paracetamol maximal adsorption decreases as the amount of surface basic groups and carbonyls increases. C 2002 Elsevier Science (USA)
Acetaminophen for synthesis (Merck) containing more than 99% of the pure
compound was used for preparing the initial solution. As an adsorbent, the “commercial” activated carbon D43/1 (Carbo-Tech, Essen, Germany) was applied.
It was de-ashed using the procedure of Korver (the obtained carbon is called
D43/1-pure). This carbon was chemically modified, and concentrated (65%)
pure HNO3 (D43/1–HNO3 ), fuming H2 SO4 , (D43/1–H2 SO4 ), and gaseous ammonia (D43/1–NH3 ) were applied as the modifying reagents. The detailed analysis of carbon surface chemical composition using different methods mainly
FTIR, XPS, TG, and the enthalpy of immersion measurements, as well as the
results of porosity characterization were given previously (4–8). However, the
results of the determination of the surface concentration of carbonyls have not
been published yet. The concentration of these groups was determined following the method of Boehm (9, 10), applying titration with C2 H5 ONa. The results
are shown in Table 1, including also the concentration of basic surface groups,
determined previously (5). Paracetamol adsorption isotherms were measured at
300, 310, and 320 K; the detailed procedure as well as the results can be found
in (5).
3. THE EQUATIONS OF ADSORPTION ISOTHERM
APPLIED FOR DESCRIBING EXPERIMENTAL DATA
1. INTRODUCTION
It is well known that oral charcoal can be successfully applied as an antidote in
different intoxinations (1–3). The results of the detailed investigation into the influence of carbon surface chemical composition on adsorption (and the kinetics
of this process) of 4-hydroxyacetanilide (paracetamol), an analgestic/antipyretic
drug, from aqueous solution and at the neutral pH, were published recently
(4–8). It was shown that temperature as well as carbon surface composition
strongly influence adsorption properties (4, 5) (including the rate (6–8)) of activated carbons toward a paracetamol molecule in vitro. On the other hand, the differences in the porosity of chemically unmodified commercial carbons slightly
alter adsorption properties (4) which appear to be determined by the existence
of acidic and basic centers on the surface usually created by the chemisorption
of gases from the atmosphere.
It was also shown that, in the range of relative adsorption up to ca. 0.6,
the effective diffusion coefficient of paracetamol molecules increases with an
increase in the enthalpy of carbon immersion in water (6, 7).
The aim of the present study is dual: to check the applicability of the
most widespread models applied for describing the data of adsorption from
solution and, what is more important, to correlate the value of the maximum adsorption with the parameters characterizing carbon surface. Such results are very significant and explain changes, observed experimentally, in the
adsorption properties of carbon caused by the chemical modification of its
surface.
The current study is based on the paper published in this journal by Khan
and coworkers (11). They collected almost all the most widespread adsorption
isotherm equations applied for describing adsorption from solution. Thus, it was
decided to use exactly the same notation as the mentioned above authors and the
same equations, i.e., BET (Eq. [1] in (11)), Radke and Prausnitz (R-Pr) (Eq. [2]
in (11)), Toth (T) (Eq. [3] in (11)), Fritz and Schluender (F-S) (Eq. [4] in (11)),
Holl and Kirch (H-K) (Eq. [5] in (11)), Marczewski and Jaroniec (M-J) (Eq. [6]
in (11)), polynomial–Freundlich (pF) (Eq. [7] in (11)), Jossens (J), (Eq. [8] in
(11), note that in the cited paper in this equation q0 should be replaced by qe
(12)), Dubinin and Astakhov (D–A) (Eq. [10] in (11)), Fukuchi (F) (Eq. [11] in
(11)), the model they called ideal adsorbed solution (IAS) (Eq. [12] in (11)), and
generalized model (GM) (Eq. [13a] in (11); note that, in the cited paper, in this
equation exp should be added before the terms in brackets both in the numerator
and in the denominator).
Moreover, other models were applied, as follows.
Redlich–Peterson (R-Pe) (13, 14):
q e = a R ce
1 + bR ceβ ,
[1]
where qe is adsorption, ce is equilibrium concentration, and aR , bR , and β are
constants where the last one lies between 0 and 1.
507
0021-9797/02 $35.00
C 2002 Elsevier Science (USA)
All rights reserved.
508
LETTER TO THE EDITOR
TABLE 1
The Concentration of Surface “Carbonyls” cCO and Surface
Basic Groups cb for Carbons Studied
Carbon
cCO
(mmol/g)
cb
(mmol/g)
D43/1-pure
D43/1–HNO3
D43/1–H2 SO4
D43/1–NH3
0.191
0.494
0.228
0.000
0.175
0.088
0.071
0.564
Note. Other characteristics were given previously (4–8).
Newman (N) (12):
−1 −1 −2
qe = (Ace )−1 + l Bcel
,
× (Ace )−1 + Bcel
[2]
where A, B, and 1 are constants.
Weber–Vliet (W-V) (15):
ce = W1 × qe × exp W2 × qeW3 + W4 ,
[3]
where W1 –W4 are constants.
All 15 models were fitted to paracetamol experimental data using the procedure
described recently (20). The goodness of the fit is expressed by the determination
coefficient (DC).
FIG. 1. Graphical comparison of experimental (squares, dashed line) and
theoretical paracetamol adsorption data for five arbitrarily chosen models.
Adsorption on carbon D43/1-pure at T = 320 K.
4. RESULTS AND DISCUSSION
To perform the calculations, the values of solubility determined previously (5)
were used. The models were usually minimized in the range of an equilibrium
mole fraction up to 5 × 10−4 ; i.e., the area of polymolecular adsorption or/and
that of the formation of crystalline state (16) was not analyzed.
The results are shown in Tables 2 and 3. The models are arranged according to
the decreasing value of the average DC (see Tables 1 and 2). For three best fitting
models the parameters are also included in these tables. It should be pointed out
that the physical importance of the obtained parameters will be discussed in the
future (17), and in the current study the author pays attention to the comparison
of the values of maximum adsorption (q0 ). This is because the main purpose of
this study is to correlate these values with carbon surface characteristics in order
to gain information about the role surface groups play in the mechanism of
paracetamol adsorption.
It can be noticed that the best average fit is observed for Weber–Vliet and the
worst for BET and Redlich–Peterson models. The drastic decrease from 90.61%
for pF to 87.32% in DC value is observed for the H-K model. Figure 1 shows
some arbitrarily chosen results of the fitting.
TABLE 2
Three Models That Describe the Experimental Data with the Highest Average Determination Coefficients (DC)
and the Parameters Obtained from Approximation
W–V
Carbon
D–A
W1 [(mole/l)/
W2
T [K] DC [%] (mmole/g)] [(g/mmole)W3 ]
W3
−0.406
−7.061
−6.526
D43/1-pure
300
310
320
91.97
98.36
97.87
1.66 10−3
3.59 10−3
8.34 10−6
−6.084
−39.208
−170.488
D43/1–H2 SO4
300
310
320
96.53
89.65
98.21
8.49 10−2
1.02 102
1.38
−7629.97
−1.40 106
−3718.72
D43/1–HNO3
300
310
320
92.70
96.99
98.69
4.88 10−4
3.41 10−3
1.57 10−3
1.614
−26.508
−13.131
0.511
−6.381
−6.959
D43/1–NH3
300
310
320
92.56
90.60
97.93
1.11 10−3
6.42 10−2
1.95 10−5
−2.471
21.812
−113.804
Average DC (%)
95.17
W4
q0
a
DC [%] [mmole/g] [(mole/J)b ]
J
b
DC [%]
a1
a2
[(l/mole)b ]
b
7.329 94.82
1.696 95.65
7.381 88.54
1.477
1.764
2.250
2.15 10−6 6.094 86.63 1.22 104
1.16 10−6 5.528 96.68 1.0 105
2.6 10−6 4.090 93.67 1.6 106
4.376
3.935
5.389
0.395
0.970
0.637
−17.490 −1.847 98.85
−22.095 −8.347 99.26
−12.810 −4.934 99.05
1.955
2.021
3.201
5.56 10−3 1.820 96.86
1.49 10−4 3.541 96.70
8.61 10−2 0.822 96.96
1 105
1.7 105
1.9 106
3.867
4.603
6.407
0.954
0.723
0.524
2.415 92.57
1.990 94.45
2.207 96.46
0.913
2.243
4.094
3.45 10−4 4.031 90.78
1.32 10−2 1.748 95.53
3.58 10−1 0.487 94.78
1 103
1 105
2.1 106
3.349
4.817
8.145
0.270
0.656
0.368
−3.637
2.990 89.30
0.208 −26.138 93.88
−5.957
6.538 93.72
1.414
2.031
2.553
9.01 10−7 6.421 89.81
1.70 10−3 2.678 96.59
6.10 10−3 1.510 95.47
1 104
8.5 104
4.9 105
3.661
4.994
4.108
0.694
0.561
0.753
94.71
94.21
509
LETTER TO THE EDITOR
TABLE 3
The Remaining Models Arranged According to the Decrease in the Average DC Value
DC (%)
Carbon
T (K)
F–S
IAS
R–Pr
M–J
F
pF
H–K
GM
N
T
BET
R–Pe
D43/1-pure
300
310
320
92.74
95.59
92.26
97.40
96.40
89.77
88.04
94.03
95.80
94.21
89.97
91.58
91.48
95.89
89.24
88.72
90.94
89.92
92.54
93.98
75.98
80.58
83.64
85.37
81.03
84.35
91.07
86.51
92.99
85.43
87.76
94.66
90.39
86.16
93.53
79.53
D43/1–H2 SO4
300
310
320
98.89
96.47
98.91
97.03
91.87
92.92
92.60
93.38
99.27
98.68
93.72
99.01
86.27
95.61
90.02
98.27
82.51
95.20
73.88
95.72
71.90
86.16
99.79
98.78
73.25
83.55
97.89
87.39
94.88
74.35
89.69
86.33
80.99
92.53
95.47
80.26
D43/1–HNO3
300
310
320
85.41
97.07
89.82
92.88
98.75
76.78
89.60
95.29
95.22
93.05
96.49
96.07
96.33
96.32
85.44
91.46
92.59
93.42
88.45
94.63
93.92
89.87
84.36
89.51
91.35
95.27
96.01
88.27
94.54
80.86
88.98
95.41
59.13
80.54
95.13
57.71
D43/1–NH3
300
310
320
83.08
90.42
95.98
86.88
94.39
97.77
81.56
89.66
96.84
74.42
85.45
94.86
89.35
93.03
87.29
82.37
89.11
92.81
89.06
94.43
83.33
76.64
79.82
90.98
80.01
78.04
90.93
78.09
88.02
91.10
84.07
86.74
94.62
78.68
90.17
91.82
93.05
92.74
92.61
92.29
91.36
90.61
87.32
87.13
86.90
86.87
86.56
85.13
Average DC (%)
The similarity between the values of maximum adsorption (q0 ), obtained from
different models, depends not only on the type of carbon studied but also on
temperature. Thus, for example, for adsorption on D43/1–H2 SO4 at T = 310 K,
almost all studied models provide similar values of maximum adsorption (only
F-S and M-J give smaller ones by about 30%), while at 320 K the value of this
parameter increases rapidly for DA and R-Pr models. At the lowest temperature,
the largest differences between q0 values are observed. The F-S model usually
shows that q0 values approach the average value taken from other models. The
comparison of q0 values lead to the conclusion that, in most cases, DA and
R-Pr models generate the largest q0 values. On the other hand, the T and G-M
equations lead to the smallest values. The most representative case is the adsorption on carbon D43/1–HNO3 , shown in Figs. 2 and 3.
The values of calculated maximum adsorption help to explain the impact
of carbon surface chemical composition on the adsorption of the paracetamol
molecule. This is a very important problem since the effect of carbon surface
chemical composition on adsorption from solution is well documented, however,
not necessarily well understood, as was pointed out by Radovič (18). Moreover,
in the case of N-containing functional groups (as occurring on carbon D43/1–
NH3 surface), this effect has not received adequate attention in literature (18).
To determine factors influencing the value of paracetamol adsorption the
author tried to correlate the maximum adsorption values (q0 ) with the parameters
characterizing the chemical composition of carbon surface. As mentioned above,
the F-S model usually shows that q0 values approach the average value taken
from other models. Figure 4 shows that the values of q0 obtained from this model
decrease as the total amount of “carbonyl” and basic surface groups increases. It
will be shown that the q0 data calculated for two Polish carbons lie on the same
line (17). The results are astonishing; especially taking into account the fact that
it is widely accepted that the surface carbonyls increase the adsorption of phenols
by the interactions with the benzene ring (19). From our previous calorimetric as
well as kinetic study it arises that the role of surface basic groups in paracetamol
adsorption is unquestionable (4, 5, 7) and these groups play a very important role
in adsorption on modified as well as on nonmodified carbons. The appearance of
these groups on carbon surface leads to the increase in surface hydrophobicity
and the simultaneous decrease in paracetamol effective diffusion coefficient
(comparing to other nonmodified as well as modified carbons) is observed.
Thus, those groups rather poorly interact with the solvent. They, in the case of
paracetamol adsorption, evidently do not enlarge the adsorption value; however,
the relative enthalpy of immersion of paracetamol (21 h w ) increases with the
FIG. 2.
1–HNO3 .
FIG. 3. The same as in Fig. 2, but the values obtained from the other models
are compared.
The comparison of the values of q0 for adsorption on carbon D43/
510
LETTER TO THE EDITOR
of zero charge of carbons will be published in the near future. It will be shown
that some additional parameters characterizing a carbon surface correlate with
the kinetic and adsorption data of a paracetamol molecule.
ACKNOWLEDGMENTS
Author gratefully acknowledges the financial support from KBN Grant 3T09A
005 18. The help of Dr. S. Biniak in measurements of carbonyl surface concentrations is gratefully acknowledged.
REFERENCES
FIG. 4. The correlation between the values of q0 and the total amount of
“carbonyl” and basic surface groups. Adsorption at T = 300 K. It has been
shown (17) that the data measured for adsorption on two Polish carbons lie on
the same line.
increase in the content of basic surface groups (cb ) as (4)
21 h w = a/(1 + be−ccb ),
[4]
where a, b, and c are constants.
The role of carbonyls is very interesting and the reduction of paracetamol
adsorption occurs probably by the repulsion effect with the same groups of a
paracetamol molecule. It is postulated that a paracetamol molecule interacts by
the OH group with carbon basic surface sites, and the repulsion effect occurs
between the CO group of this molecule and similar groups attached to the
surface. The increase in the relative enthalpy of adsorption with the increase
in the content of basic surface groups leads to the decrease in the mobility of
adsorbed molecules and the effect of the increase in adsorption, comparing to
a nonmodified carbon, is not visible, as it was suggested previously (5). On the
other hand, the increase in the content of surface carbonyls leads to the increase
in repulsive forces and both effects lead to the decrease in the value of q0 .
Following this mechanism, the adsorption of paracetamol at the acidic pH
should probably lead to the increase in adsorption value after the basic groups
are neutralized. Thus, for carbon possessing the smallest total amount of surface
basic groups as well as “carbonyls” (D43/1–H2 SO4 ), following the postulated
mechanism the effect of the increase in solution acidity on paracetamol, adsorption should be the slightest. On the other hand, these are only speculations since
at the acidic pH values the process of the protonization of the π electrons from
basal planes of carbon layers occurs (18) and it is hard to predict now how it
influences the adsorption properties toward a paracetamol molecule (the surface charge of carbon can be positive or equal to zero at acidic pH (20)). Thus,
further experimental studies are necessary to be carried out. The determination
of the distribution of surface functional groups, TPD (temperature-programmed
desorption) results, and the results of the measurement of the pH of the point
1. Pond, S., Jacos, M., Marks, J., Garner, J., Goldschlager, N., and Hansen,
D., Lancet, 1177 (Nov. 21, 1981).
2. Neuvonen, P. J., Clin. Pharmacokinet. 7, 465 (1982).
3. Neuvonen, P. J., Vartiainen, M., and Tokola, O., Eur. J. Clin. Pharm. 24,
557 (1983).
4. Terzyk, A. P., Adsorpt. Sci. Technol. 17, 441 (1999).
5. Terzyk, A. P., Rychlicki, G., Colloids Surf. A 163, 135 (2000).
6. Terzyk, A. P., Adsorpt. Sci. Technol. 18, 477 (2000).
7. Terzyk, A. P., J. Colloid Interface Sci. 230, 219 (2000).
8. Terzyk, A. P., Colloids Surf. A 177, 23 (2001).
9. Boehm, H. P., Diehl, E., Heck, W., and Sappok, R., Angew. Chem. 76, 742
(1964).
10. Jankowska, H., Świ tkowski, A., Starostin, and L., L awrinienkoOmiecynska, J., “Adsorption of Ions onto Activated Carbon” [in Polish],
PWN, Warsaw, 1991.
11. Khan, A. R., Ataullah, R., and Al.-Haddad, A., J. Colloid Interface Sci.
194, 154 (1997).
12. Jossens, L., Prausnitz, J. M., Fritz, W., Schlunder, E. U., and Myers, A. L.,
Chem. Eng. Sci. 33, 1097 (1978).
13. Redlich, O., and Peterson, D. L. A., J. Phys. Chem. 63, 1024 (1959).
14. Juang, R. S., Wu, F. C., and Tseng, R. L., J. Chem. Eng. Data 41, 487 (1996)
15. Vliet, B. M., Weber, W. J., and Hozumi, H., Water Res. 14, 1719 (1980).
16. Kipling, J. J., “Adsorption from Solutions of Non-electrolytes,” Academic
Press, London, 1965.
17. Terzyk, A. P., Adsorpt. Sci. Technol., in press.
18. Radovič, L. R., in “International Symposium on Carbon, Tokyo, 1998,”
p. 380.
19. Puri, B. R., Bhardwaj, S. S., Kumar, V., and Mahajan, O. P., J. Ind. Chem.
Soc. 52, 26 (1975).
20. Radovič, L. R., Moreno-Castilla, C., and Rivera-Utrilla, J., in “Chemistry
and Physics of Carbon” (L. R. Radovic, Ed.), Vol. 27, p. 227, 2001.
Artur P. Terzyk1
Department of Chemistry
Physicochemistry of Carbon Materials Research Group
N. Copernicus University
Gagarina 7
87-100 Toruń, Poland
Received May 17, 2001; accepted December 27, 2001
1
To whom correspondence should be addressed. Fax: (48–56) 654–2477.
E-mail: [email protected].