- Email: [email protected]
Textural properties of activated carbons from apricot stones
Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
Textural properties of activated carbons from apricot stones A.M. Youssefa , N.R.E. Radwanb,∗ , I. Abdel-Gawadb , G.A.A. Singerc b
a Department of Chemistry, Faculty of Science, Mansoura University, Mansora, Egypt Department of Chemistry, Faculty of Education, Suez Canal University, Suez 1262, Egypt c Nasr Petroleum Company, Suez, Egypt
Received 29 March 2004; accepted 13 September 2004 Available online 19 November 2004
Abstract Chemically activated carbons were prepared from apricot stones. Phosphoric acid (25–75 wt.%) was used as an activating agent at 400–600 ◦ C. Zinc chloride-activated carbon were also prepared at 600 and 700 ◦ C using three different zinc chloride:apricot stone ratios of 0.5, 1.0 and 2.0. Steam-activated carbons were obtained by gasifying un-activated carbon obtained by carbonizing apricot stones, at 900 ◦ C to burn-off 25 and 35%. The textural parameters were determined from the nitrogen adsorption data at 77 K. Different adsorption models were considered for the analysis of the adsorption results. Considerable differences between surface areas and pore volume as obtained by considering the different models, have been observed. However, the method of analysis based on comparing the isotherm determined for a given active carbon with the standard isotherm for a non-porous adsorbent seems to be trustworthy. © 2004 Elsevier B.V. All rights reserved. Keywords: Apricot stones; Activated carbon; Phosphoric acid; Zinc chloride; Steam activation
1. Introduction Activated carbons are non-specific adsorbents and therefore find wide application in the removal of colour [1–3], odour [4,5], toxic gases [6,7] etc. Activated carbons are now in use for the treatment of potable water [8,9] and waste water [10,11], particularly for the removal of heavy metals [12–15]. The adsorption properties of activated carbons are essentially attributed to their large surface area, large total pore volume, high degree of surface reactivity and favourable pore size distribution [16,17]. The texture (surface area and porosity) of activated carbons can be easily modified or even tailored to suit a specific application [18]. The chemistry of the surface of activated carbon also plays a dominant role in determining its adsorption properties and consequently its use [19,20]. It is also possible to modify the surface chemistry of activated carbons by controlling the amount and strength of the surface
∗
Corresponding author. Tel.: +20 48 2570685; fax: +20 62 664873. E-mail address: nagi r [email protected] (N.R.E. Radwan).
0927-7757/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2004.09.008
functional groups particularly those of the carbon–oxygen type [21,22]. Two conventional methods exist for the preparation of activated carbons from a carbonaceous material. These are: (i) chemical activation, which is based on the carbonization in a limited supply of air or in an inert atmosphere of a mixture of a carbonaceous material and an activating agent (zinc chloride or phosphoric acid) at some intermediate temperatures (400–700 ◦ C), followed by washing and drying [23,24]. (ii) Physical activation based on gasification at a relatively high temperature (900–1000 ◦ C) with an oxidizing gas (steam or carbon dioxide) of a non-activated carbon to certain percentage of burn-off [25]. Cheap raw materials have been recommended for the preparation of activated carbons. Agricultural by-products, which exist in large amounts, represent a solid pollutant to the environment. Many years ago these byproducts were used as a fuel in rural areas but now they do not find any application of commercial interest. The preparation of activated carbons from agricultural by-products has therefore been encouraged, since they are cheap precursors and their use in this manner would prevent their accumulation.
144
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
Among other countries, Egypt is rich in many agricultural by-products that can be used for the preparation of activated carbon, viz., cotton stalks and corn stalks, rice straw and rice husks, corncobs, date pits, some fruit stones, some nutshells and olive stones. Apricot stones have been used in the present study for the preparation of activated carbons, following impregnation with different percentage of phosphoric acid (25–75 wt.%) or zinc chloride (33–66 wt.%), activation was performed at different temperatures between 400 and 700 ◦ C. Two physically activated carbons were also prepared by gasifying non-activated carbon (prepared by carbonizing apricot stones at 600 ◦ C), with steam at 900 ◦ C to burn-off 25 and 35%. The textural properties of the resulting activated products were determined from nitrogen adsorption studies at 77 K. The adsorption data have been analyzed in terms of different theories and methods of textural analysis.
2. Experimental 2.1. Materials Dried ground apricot stones were first washed thoroughly with water and then dried again at 383 K. After cooling to room temperature, they were soaked for 72 h in the solution of the activating agent. Activation with phosphoric acid was carried out using various concentrations in the range 25–75 wt.% with occasional shaking. Activation with zinc chloride was carried out using different ratios of zinc chloride to apricot stones = 0.5, 1.0, 2.0 and the appropriate amount of zinc chloride was dissolved in the least amount of distilled water. The treated stones were then filtered and dried to constant weight at 343 K. The dried treated samples were then carbonized in absence of air at temperatures between 400 and 600 ◦ C for phosphoric acid-treated stones and at 600 and 700 ◦ C for zinc chloride-treated stones. The carbonized phosphoric acid-treated products were washed with distilled water until the pH of the resulting wash was ca. 6.0. In the designation adopted below for these carbons, “A” denotes apricot stones and “P” indicates treatment with phosphoric acid, the arabic number following the letter P gives the wt.% of phosphoric acid employed, while the arabic number following the dash gives the carbonization temperature. Thus, for example, AP50–600 stands for an activated carbon prepared from apricot stones by activation with 50 wt.% phosphoric acid followed by carbonization at 600 ◦ C. The carbonized zinc chloride-activated products were washed with 10% hydrochloric acid and then with distilled water until the resulting wash was Cl− -free. In the designation adopted for these carbons, the letter “Z” indicates treatment with zinc chloride, the arabic number following the letter Z gives the zinc chloride/apricot stone ratio, while the arabic number following the dash gives the carbonization temperatures. Thus, for example, AZ0.5–700 stands for an activated carbon prepared from apricot stones by activation with zinc chloride (zinc chloride/apricot stones = 0.5), followed by carbonization at
700 ◦ C. Steam-activated carbons were prepared by gasifying 600 ◦ C, precarbonized apricot stones A-600, with steam at 900 ◦ C to burn-off 25%, giving AS25 or to burn-off 35% giving AS35, where the letter “S” indicates steam activation and the arabic numbers following directly this letter give the percentage burn-off. 3. Techniques The adsorption of nitrogen at 77 K was determined using a conventional volumetric apparatus. Prior to such measurements, the solids were heated over night at 523 K under high vacuum (10−5 Torr). 4. Results and discussion The adsorption of nitrogen at 77 K proved to be rapid with the equilibrium being attained within less than 40 min at relative pressures less than 0.1 and in less than 20 min at higher relative pressures. This indicates that almost all the pores were accessible to nitrogen molecules at 77 K, this being true for all the carbons investigated. The desorption point were found to lie on the same isotherm as the adsorption data, indicating the absence of hysteresis characteristic of mesoporosity or specific interaction [26], however, AS35 is an exception where pronounced closed hysteresis loop was exhibited. Fig. 1 depicts representative nitrogen adsorption isotherms. With the exception of steam-activated carbons (AS25 and AS35), all the other isotherms are Langmuirian in shape being typical type I in the BDDT classification [27], which is characteristic of adsorption on a microporous adsorbent. On the other hand, the nitrogen adsorption isotherms of AS25 and AS35 are of type II in the same classification. Application of the Langmuir equation [28,29] was satisfactory and was found to cover a wide range of relative pressure, representative linear Langmuir plots are shown in Fig. 2. Such plots enable the monolayer capacity as well as the specific surface area, SL (m2 /g). The corresponding values of SL obtained are listed in column 2 of Table 1. The conventional BET equation [30] was also applied over the relative pressure range 0.02 ≤ p/p0 ≤ 0.2 to determine the monolayer capacity Vm and hence the specific surface area SBET (m2 /g) by adopting the value of 0.16 nm2 for the cross-sectional area of the nitrogen molecule at 77 K. Representative linear BET plots are depicted in Fig. 3, while the calculated SBET values for the investigated carbons are listed in column 3 of Table 1. The total pore volume VT (ml/g) expressed as the volume of liquid nitrogen adsorbed per gram carbon at relative pressure of ca. 0.98 p/p0 , is listed in column 4 of Table 1. Based on the assumption that the space in the micropores is similar to the space between two parallel plates, the average pore radius r¯ (nm) could be calculated from the relationship [31]: r¯ = 2VT
103 S BET
(1)
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
145
Fig. 1. Representative adsorption nitrogen isotherms at 77 K.
The values of r¯ as calculated for the carbons investigated are given in column 5 of Table 1. Two other independent methods were also applied to analyze the nitrogen adsorption isotherms, i.e. the t-method [32] and the αs -method [33]. In the first method the amount of ni-
trogen adsorbed (ml/g) is plotted versus the multilayer thick˚ as measured on a standard non-porous material ness t (A) of comparable BET-C constant (Fig. 4). The second method plots the volume of nitrogen adsorbed versus the reduced isotherm (αs ) determined on a standard non-porous material
Table 1 Textural parameters of carbons studied obtained from nitrogen adsorption at 77 K via application of models based on surface Sample
SL (m2 /g)
SBET (m2 /g)
VT (ml/g)
r¯ (nm)
St (m2 /g)
t Vmic (ml/g)
Snt (m2 /g)
Sα (m2 /g)
α Vmic (ml/g)
Snα (m2 /g)
A-600 AP25–400 AP25–500 AP25–600 AP50–400 AP50–500 AP50–600 AP75–400 AP75–500 AP75–600 AZ0.5–600 AZ0.5–700 AZ1.0–600 AZ1.0–700 AZ2.0–600 AZ2.0–700 AS25 AS35
270 431 530 576 594 759 759 700 874 1008 364 437 656 728 1017 1017 1008 1093
211 311 386 446 458 551 551 524 655 728 280 328 460 546 728 728 650 683
0.095 0.149 0.181 0.198 0.214 0.264 0.264 0.246 0.302 0.327 0.129 0.159 0.226 0.258 0.358 0.358 0.467 0.747
0.9 0.9 0.98 0.98 0.78 0.96 0.96 0.94 0.83 0.9 0.92 0.96 0.99 0.95 0.98 0.98 1.43 2.119
215 325 400 450 480 560 560 535 665 755 300 350 500 520 750 750 650 715
0.085 0.130 0.165 0.180 0.185 0.235 0.235 0.240 0.280 0.305 0.105 0.120 0.200 0.225 0.310 0.310 0.130 0.160
10 15 15 17 20 20 20 16 22 24 12 14 20 20 30 30 150 220
230 330 393 460 467 585 585 526 630 725 304 350 492 600 746 746 620 708
0.078 0.130 0.156 0.180 0.185 0.240 0.240 0.227 0.270 0.300 0.120 0.0132 0.210 0.235 0.320 0.320 0.134 0.160
9 12 13 15 16 18 18 15 20 21 12 13 18 20 27 27 148 230
146
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
Fig. 2. Representative linear Langmuir plots of nitrogen adsorption.
of the same chemical composition. The adsorption data reported by Sellez–Perez and Martin–Martinez [34] were used for such plots, representative α plots are shown in Fig. 5. The t-method allows the determination of the specific surface area St as determined from the slope of the solid line passing through the origin (Fig. 4). The micropore volume t is obtained from the intersection of the dotted line passVmic ing through the plateau of the V–t plot, zero thickness. The surface located in non-micropores Snt could also be calcut (ml/g) lated from the slope of the dotted line. St (m2 /g) Vmic t 2 Sn (m /g) are listed in columns 6–8 of Table 1, respectively. Similarly, the αs -method allows the determination of the same α and S α , respecthree parameters but with notations, Sα , Vmic n tively (columns 9–11 of Table 1). Inspection of the data recorded in Table 1 reveals the following: (1) Activation with phosphoric acid is associated with a considerable increase in the adsorption capacity compared with the non-activated carbon A-600. The degree of activation with phosphoric acid increased with the increase of the concentration of this activating agent from 25 to 75 wt.% and also with the rise of carbonization temperature from 400 to 600 ◦ C. This activating agent dehydrates
the carbonaceous material during carbonization and this leads to charring and aromatization of the carbon skeleton and the creation of a porous structure [35]. (2) Activation with zinc chloride should be undertaken at 700 ◦ C, particularly when the mass of zinc chloride used in activation is less than twice the mass of the carbonaceous precursor. (3) Gasification with steam at 900 ◦ C of A-600, to a burnoff equals 25% was found to be associated with more than three-fold increase in SL or SBET . However, further gasification to higher burn-off (35%) was found to be associated with a further but small increase in the surface area. Considering the total pore volume VT or the mean pore radius r, it is evident that the increase of burnoff from 25 to 35% was associated with more than 50% increase of these textural parameters. (4) The surface areas calculated through the application of the t-method and αs -method are comparable. These two methods are based on a standard non-porous material and each method can be used to complement the other. The data listed in Table 1 show that, in most cases the SBET values were lower than the St and Sα values, however, the difference was not high and may be attributed to the BET model being based on multilayer coverage, which
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
147
Fig. 3. Representative linear BET plots of nitrogen adsorption at 77 K.
is not precisely the case in adsorption on activated carbons where most of the pores accommodate only a limited number of layers as indicated by the values of their average radii (column 5 of Table 1). Some pores may accommodate only two layers, each one on the opposite walls. In addition, some pores may even be inaccessible to adsorbate molecules. It should be noted from the data in Table 1 that the calculated SL values are very high relative to the SBET , St and Sα values, an unsurprising fact since in most cases, the monolayer capacities determined by the application of Langmuir model lie beyond the equilibrium pressure employed. For this reasons, the surface areas of activated carbons based on Langmuir model should be taken with great reservation. (5) The surface area located in the micropore region represents a large fraction of the total surface area and consequently the surface located in the non-microporous region would be only a small fraction of the total surface area, however, steam-activated carbons AS25 and AS35 represent an exception. This refers to the importance of the method of activation. Steam-activation is frequently associated with pore widening, particularly at intermediate and high percentages of burn-off [36].
The physical adsorption of gases and vapours by microporous solids, in general, and by active carbons, in particular, may also be described by Dubinin’s theory as developed in successive stages since 1947 [37,38]. In the present formulation, the theory of micropore felling may be expressed by the Dubinin–Astakhov (DA) equation [39], which may be expressed as: A n V = V0 exp − (2) βE0 where V represents the volume filled at temperature T and relative pressure p/p0 , V0 is the total volume of micropores, the quantity A is equal to RT ln (p0 /p) and n, E0 and β are specific parameters of the system under investigation. The DA equation is applicable over the relative pressure range 0.05 < p/p0 < 0.1 where the influence of any non-microporous surface area is negligible. For typical active carbons, the exponent is equal to 2, which corresponds to the empirical equation postulated by Dubinin and Radushkevich in 1947 and known in the literature as the DR equation [37]: T 2 2 p0 V = V0 exp −B log (3) β p
148
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
Fig. 4. Representative V–t plots of nitrogen adsorption at 77 K.
The parameter B = (2.303 R/E0 )2 has the dimensions K−2 and is called the structural constant being related to the characteristics energy E0 by [40]: E0 (kJ/mol) =
0.01914 B0.5
(4)
The DR and DA equations are both based on the observation that a plot of log V versus (RT)2 ln2 (p0 /p) or An leads to an unique curve for a given adsorbate. The range of the classical DR equation has been verified for many-reported adsorption measurements [41,42]. The present authors are more inclined to the view that the state of the adsorbate in a microporous adsorbent is still a matter open for discussion, since evidence needs to be provided to give predominance for one postulate or another. For this reason, the nitrogen adsorption data of phosphoric acid-activated and zinc chlorideactivated carbons have been interpreted here by applying the DR equation and comparing some of the adsorption parameters thus obtained with those determined from the BET, αs and t-methods. Representative DR plots are depicted in Fig. 6 showing that linear plots covering a range of relative pressures were obtained. However, slight upward deviations from linearity at high relative pressure were observed, corresponding to micropores of size close to the mesoporosity region. Table 2 lists some of the adsorption parameters determined from the DR equation and includes parameters determined in other ways for the sake of comparison. It is seen from column 4 of Table 2 that the characteristic energies for adsorption ranged between 10.43 and 12.40 kJ/mol, i.e. they all lie in the range characterizing physical adsorption. This is expected for nitrogen in particular, since this molecule seldom undergoes specific interaction with any adsorbent at 77 K [20]. The values of X (nm), which give the pore radii are listed in column 6 of Table 2 and indicate micropore types, X = K/E0 where K = 13.03 − 1.53 × 10−5 E03.5 . However, it should be noted that in all cases these values are smaller than the values of r¯ listed in column 5 of Table 1. This may be attributed to the fact that r¯ values represent average pore radii, which contain a contribution from the non-micropores existing in a given sample. The consistency in the values of K listed in column 5 of Table 2 (12.93–12.97 nm kJ/mol)) may be attributed to the narrow range in which all E0 values exist, which is due to the location of the major fraction of the porosity in the micropore range. Evidently the micropore volumes as determined by the
Table 2 Application of the DR equation to nitrogen adsorption data at 77 K Sample
DR (ml/g) Vmic
D
E0 (kJ/mol)
K (nm kJ/mol)
X (nm)
SDR (m2 /g)
V0.1 (ml/g)
100 V0.1/VT
DR /V 100 Vmic T
A-600 AP25–400 AP25–500 AP25–600 AP50–400 AP50–500 AP50–600 AP75–400 AP75–500 AP75–600 AZ0.5–600 AZ0.5–700 AZ1.0–600 AZ1.0–700 AZ2.0–600 AZ2.0–700
0.09 0.148 0.169 0.191 0.204 0.252 0.252 0.238 0.295 0.316 0.127 0.158 0.221 0.251 0.353 0.353
0.06 0.07 0.053 0.06 0.06 0.06 0.06 0.06 0.07 0.07 0.072 0.072 0.07 0.075 0.072 0.072
11.66 10.8 12.4 11.66 11.66 11.66 11.66 11.66 10.8 10.8 10.65 10.65 10.8 10.43 10.65 10.65
12.95 12.97 12.93 12.95 12.95 12.95 12.95 12.95 12.97 12.97 12.97 12.97 12.97 12.97 12.97 12.97
1.11 1.2 1.04 1.11 1.11 1.11 1.11 1.11 1.2 1.2 1.22 1.22 1.2 1.24 1.22 1.22
254 418 476 538 575 710 710 671 832 890 364 445 623 724 995 995
0.08 0.129 0.156 0.174 0.186 0.234 0.234 0.212 0.265 0.298 0.11 0.134 0.198 0.226 0.318 0.318
84 91 86 88 87 89 89 86 88 91 85 95 87 88 88 88
95 99 93 96 95 95 95 97 98 91 99 99 97 97 99 99
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
149
Fig. 5. Representative α-plots of nitrogen adsorption at 77 K.
DR-method represented 94–98% of the total pore volumes. If the micropore values as determined by the DR equation are converted to equivalent surface areas, it is possible to obtain a set of SDR (m2 /g) values (column 7 of Table 2). It will be noted that in most cases the SDR values listed are smaller than the comparable SL values. It has been mentioned already that the Langmuir model on which SL values are based gives monolayer capacities beyond the equilibrium pressure and hence conversion of monolayer capacities into specific surface areas is of limited significance. Equally for chemically activated carbons, the SDR values are always greater than the SBET values, unsurprising to the authors because each set of values is based on a different model. The pore diameters of the chemically activated carbons investigated as determined by the DR methods are centered between 1.04 and 1.24 nm. On the other hand these pore diameter were in the range 1.56–1.98 nm when the BET model was considered. This means that although the two model have predicted the micropore structure, agreement was not found to exist concerning the appropriate range in which these pores were
existing. The textural parameters as determined from the DRmodel and from the BET model should be taken with great reservation. The textural properties determined from the tmethod and αs -method, both based on a standard nonporous reference, remain more trustworthy. Column 8 of Table 2 lists the volume of adsorbate taken up at p/p0 = 0.1, a relative pressure at which micropores are filled with adsorbate molecules (i.e. nitrogen molecules at 77 K). All the chemically activated carbons investigated exhibit a microporosity representing 84–91% of the total porosity in the sample (column 9 of Table 2). The same conclusion may be reached by dividing the micropore volumes as estimated from the DR equation by the total pore volumes, the corresponding values listed in column 10 of Table 2 indicating that for all chemically activated carbons investigated, the micropore fraction ranged between 91 and 99% of the total porosity. To summarize the observed difference, in the determination of surface area and porosity of chemically activated carbons, by different models refer to the heterogeneity of these sorbents and indicates that the determination of their textural
150
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
Fig. 6. Representative DR plots of nitrogen adsorption at 77 K.
parameters is a complex problem. The conventional classification of pore, which consider micropores as these with diameter < 2.0 nm needs to be reconsidered. In the pores with diameters < 2.0 nm, more than one model of adsorption can be considered; i.e. activated diffusion in micropores, micropore filling and surface coverage. Each of these model probably predominates in a certain narrow range of microporosity, which should be subdivided into sub-micropore ranges. However, on can count on the methods based on comparing the isotherm determined for the given active carbon with the standard isotherm for nonporous adsorbents, which acts as a reference system.
References [1] G. Mckay, J. Chem. Technol. Biotechnol. 40 (1982) 759. [2] A.M. Youssef, A.A. EL-Khouly, A.I. Ahmed, E.T. EL-Shafy, Adsorpt. Sci. Technol. 12 (1994) 211. [3] A.M. Youssef, A.A. EL-Khouly, A.I. Ahmed, E.I. EL-Shafey, J. Environ. Sci. Mansoura 10 (1995) 195. [4] R.C. Bansal, J. Donnet, F. Stoeckli, Active Carbon, Marcel Dekker Inc., New York, 1988, p. 335.
[5] A.M. Youssef, A.M. El-Wakily, K. Tollan, Th. El-Nabarawy, Afinidad 47 (1990) 336. [6] F. Derbyshire, M. Jagtoyen, M. Thwaites, in: J.W. Patrick (Ed.), Porosity in Carbons, Edward Arnold, London, 1995, p. 228. [7] P. Davini, Carbon 39 (2001) 1387. [8] S.H. Abdel-Halim, A.M.A. Shehata, M.E. El-Shahat, Water Res. 37 (2003) 1678. [9] C.O. Ania, J.B. Parra, J.J. Pis, Fuel Process. Technol. 77–78 (2002) 337. [10] N. Klimenko, M. Winter-Nielsen, S. Smolin, L. Nevynna, J. Sydorenko, Water Res. 36 (2002) 5132. [11] A.M. Youssef, M.R. Mostafa, E.M. Dorgham, Bull. Soc. Chim. Fr. (1989) 741. [12] A.M. Youssef, A.M. El-Wakil, E.A. El-Sharkawy, A.B. Farag, Kh. Tollan, Adsorpt. Sci. Technol. 13 (1996) 115. [13] V. Strelko, D.J. Malik, J. Colloid Interface Sci. 250 (2002) 213. [14] A. Zhang, G. Vchiyama, T. Asakura, Adsorpt. Sci. Technol. 21 (2003) 761. [15] A.M. Youssef, Th. El-Nabarawy, S.E. Samra, Colloids Surf. 235 (2004) 153–163. [16] B. Cagnon, X. Py, A. Gusllot, F. Stoeckli, Microporous Mesoporous Mater. 57 (2003) 273. [17] O.N. Baklanova, G.V. Plaksin, V.A. Drozdov, V.K. Duplyakin, N.V. Chesnokov, B.N. Kuznetsov, Carbon 41 (2003) 1793.
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151 [18] Th. El-Nabarawy, M.R. Mostafa, A.M. Youssef, Adsorpt. Sci. Technol. 15 (1997) 59. [19] I.I. Salame, T.J. Bandosz, J. Colloid Interface Sci. 210 (1999) 367. [20] A.M. Youssef, T.M. Ghazy, Th. El-Nabarawy, Carbon 20 (1982) 248. [21] A. Gil, G. dela Puente, P. Grange, Microporous Mater. 12 (1997) 51. [22] J. Chorna, W. Burakiewicz-Mortka, M. Jaroniec, Z. Li, J. Klinik, J. Colloid Interface Sci. 214 (1999) 438. [23] L.B. Khalil, Adsorp. Sci. Technol. 13 (1996) 317. [24] A.M. Youssef, Carbon 13 (1975) 1. [25] A.M. Youssef, G.A. El-Shobaky, Th. El-Nabarawy, Surf. Technol. 7 (1978) 469. [26] A.M. Youssef, J. Colloid Surf. Sci. 55 (1976) 447. [27] S. Brunauer, L.S. Deming, W.E. Deming, E. Teller, J. Am. Chem. Soc. 62 (1940) 1723. [28] I. Langmuir, J. Am. Chem. Soc. 38 (1916) 2219. [29] I. Langmuir, J. Am. Chem. Soc. 40 (1918) 1368. [30] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309.
151
[31] P. Podko´scielny, A. Dabrowski, M. B¨ulow, Appl. Surf. Sci. 196 (2002) 312. [32] B.C. Lippens, J.H. deBoer, J. Catal. 4 (1965) 319. [33] K.S.W. Sing, Chem. Ind. (Lond.) (1968) 1520. [34] M.J. Sellez-Perez, J.M. Martin-Martinez, J. Chem. Soc. Faraday Trans. 87 (1991) 1237. [35] M.N. Alaya, M.R. Hourieh, A.M. Youssef, F. El-Sejariah, Adsorpt. Sci. Technol. 18 (2000) 27. [36] A.M. Youssef, M.N. Alaya, N. Nawar, Adsorpt. Sci. Technol. 11 (1994) 225. [37] M.M. Dubinin, E.D. Zaverina, L.V. Radushkevich, Zh. Fiz. Khim. 21 (1947) 1351. [38] M.M. Dubinin, Carbon 21 (1985) 359. [39] M.M. Dubinin, V.A. Astakhov, Izv. Akad. Nauk SSSR, Ser. Khim. (1971) 5. [40] R.C. Bansal, J. Donnet, F. Stoeckli, Active carbon, Marcel Dekker Inc., New York, 1988, p. 132. [41] A.M. Youssef, T.M. Ghazy, Th. El-Nabarawy, Carbon 20 (1982) 218. [42] A.M. Youssef, A.M. El-Wakil, K. Tollan, Sepn. Sci. Technol. 26 (1990) 445.
Textural properties of activated carbons from apricot stones A.M. Youssefa , N.R.E. Radwanb,∗ , I. Abdel-Gawadb , G.A.A. Singerc b
a Department of Chemistry, Faculty of Science, Mansoura University, Mansora, Egypt Department of Chemistry, Faculty of Education, Suez Canal University, Suez 1262, Egypt c Nasr Petroleum Company, Suez, Egypt
Received 29 March 2004; accepted 13 September 2004 Available online 19 November 2004
Abstract Chemically activated carbons were prepared from apricot stones. Phosphoric acid (25–75 wt.%) was used as an activating agent at 400–600 ◦ C. Zinc chloride-activated carbon were also prepared at 600 and 700 ◦ C using three different zinc chloride:apricot stone ratios of 0.5, 1.0 and 2.0. Steam-activated carbons were obtained by gasifying un-activated carbon obtained by carbonizing apricot stones, at 900 ◦ C to burn-off 25 and 35%. The textural parameters were determined from the nitrogen adsorption data at 77 K. Different adsorption models were considered for the analysis of the adsorption results. Considerable differences between surface areas and pore volume as obtained by considering the different models, have been observed. However, the method of analysis based on comparing the isotherm determined for a given active carbon with the standard isotherm for a non-porous adsorbent seems to be trustworthy. © 2004 Elsevier B.V. All rights reserved. Keywords: Apricot stones; Activated carbon; Phosphoric acid; Zinc chloride; Steam activation
1. Introduction Activated carbons are non-specific adsorbents and therefore find wide application in the removal of colour [1–3], odour [4,5], toxic gases [6,7] etc. Activated carbons are now in use for the treatment of potable water [8,9] and waste water [10,11], particularly for the removal of heavy metals [12–15]. The adsorption properties of activated carbons are essentially attributed to their large surface area, large total pore volume, high degree of surface reactivity and favourable pore size distribution [16,17]. The texture (surface area and porosity) of activated carbons can be easily modified or even tailored to suit a specific application [18]. The chemistry of the surface of activated carbon also plays a dominant role in determining its adsorption properties and consequently its use [19,20]. It is also possible to modify the surface chemistry of activated carbons by controlling the amount and strength of the surface
∗
Corresponding author. Tel.: +20 48 2570685; fax: +20 62 664873. E-mail address: nagi r [email protected] (N.R.E. Radwan).
0927-7757/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2004.09.008
functional groups particularly those of the carbon–oxygen type [21,22]. Two conventional methods exist for the preparation of activated carbons from a carbonaceous material. These are: (i) chemical activation, which is based on the carbonization in a limited supply of air or in an inert atmosphere of a mixture of a carbonaceous material and an activating agent (zinc chloride or phosphoric acid) at some intermediate temperatures (400–700 ◦ C), followed by washing and drying [23,24]. (ii) Physical activation based on gasification at a relatively high temperature (900–1000 ◦ C) with an oxidizing gas (steam or carbon dioxide) of a non-activated carbon to certain percentage of burn-off [25]. Cheap raw materials have been recommended for the preparation of activated carbons. Agricultural by-products, which exist in large amounts, represent a solid pollutant to the environment. Many years ago these byproducts were used as a fuel in rural areas but now they do not find any application of commercial interest. The preparation of activated carbons from agricultural by-products has therefore been encouraged, since they are cheap precursors and their use in this manner would prevent their accumulation.
144
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
Among other countries, Egypt is rich in many agricultural by-products that can be used for the preparation of activated carbon, viz., cotton stalks and corn stalks, rice straw and rice husks, corncobs, date pits, some fruit stones, some nutshells and olive stones. Apricot stones have been used in the present study for the preparation of activated carbons, following impregnation with different percentage of phosphoric acid (25–75 wt.%) or zinc chloride (33–66 wt.%), activation was performed at different temperatures between 400 and 700 ◦ C. Two physically activated carbons were also prepared by gasifying non-activated carbon (prepared by carbonizing apricot stones at 600 ◦ C), with steam at 900 ◦ C to burn-off 25 and 35%. The textural properties of the resulting activated products were determined from nitrogen adsorption studies at 77 K. The adsorption data have been analyzed in terms of different theories and methods of textural analysis.
2. Experimental 2.1. Materials Dried ground apricot stones were first washed thoroughly with water and then dried again at 383 K. After cooling to room temperature, they were soaked for 72 h in the solution of the activating agent. Activation with phosphoric acid was carried out using various concentrations in the range 25–75 wt.% with occasional shaking. Activation with zinc chloride was carried out using different ratios of zinc chloride to apricot stones = 0.5, 1.0, 2.0 and the appropriate amount of zinc chloride was dissolved in the least amount of distilled water. The treated stones were then filtered and dried to constant weight at 343 K. The dried treated samples were then carbonized in absence of air at temperatures between 400 and 600 ◦ C for phosphoric acid-treated stones and at 600 and 700 ◦ C for zinc chloride-treated stones. The carbonized phosphoric acid-treated products were washed with distilled water until the pH of the resulting wash was ca. 6.0. In the designation adopted below for these carbons, “A” denotes apricot stones and “P” indicates treatment with phosphoric acid, the arabic number following the letter P gives the wt.% of phosphoric acid employed, while the arabic number following the dash gives the carbonization temperature. Thus, for example, AP50–600 stands for an activated carbon prepared from apricot stones by activation with 50 wt.% phosphoric acid followed by carbonization at 600 ◦ C. The carbonized zinc chloride-activated products were washed with 10% hydrochloric acid and then with distilled water until the resulting wash was Cl− -free. In the designation adopted for these carbons, the letter “Z” indicates treatment with zinc chloride, the arabic number following the letter Z gives the zinc chloride/apricot stone ratio, while the arabic number following the dash gives the carbonization temperatures. Thus, for example, AZ0.5–700 stands for an activated carbon prepared from apricot stones by activation with zinc chloride (zinc chloride/apricot stones = 0.5), followed by carbonization at
700 ◦ C. Steam-activated carbons were prepared by gasifying 600 ◦ C, precarbonized apricot stones A-600, with steam at 900 ◦ C to burn-off 25%, giving AS25 or to burn-off 35% giving AS35, where the letter “S” indicates steam activation and the arabic numbers following directly this letter give the percentage burn-off. 3. Techniques The adsorption of nitrogen at 77 K was determined using a conventional volumetric apparatus. Prior to such measurements, the solids were heated over night at 523 K under high vacuum (10−5 Torr). 4. Results and discussion The adsorption of nitrogen at 77 K proved to be rapid with the equilibrium being attained within less than 40 min at relative pressures less than 0.1 and in less than 20 min at higher relative pressures. This indicates that almost all the pores were accessible to nitrogen molecules at 77 K, this being true for all the carbons investigated. The desorption point were found to lie on the same isotherm as the adsorption data, indicating the absence of hysteresis characteristic of mesoporosity or specific interaction [26], however, AS35 is an exception where pronounced closed hysteresis loop was exhibited. Fig. 1 depicts representative nitrogen adsorption isotherms. With the exception of steam-activated carbons (AS25 and AS35), all the other isotherms are Langmuirian in shape being typical type I in the BDDT classification [27], which is characteristic of adsorption on a microporous adsorbent. On the other hand, the nitrogen adsorption isotherms of AS25 and AS35 are of type II in the same classification. Application of the Langmuir equation [28,29] was satisfactory and was found to cover a wide range of relative pressure, representative linear Langmuir plots are shown in Fig. 2. Such plots enable the monolayer capacity as well as the specific surface area, SL (m2 /g). The corresponding values of SL obtained are listed in column 2 of Table 1. The conventional BET equation [30] was also applied over the relative pressure range 0.02 ≤ p/p0 ≤ 0.2 to determine the monolayer capacity Vm and hence the specific surface area SBET (m2 /g) by adopting the value of 0.16 nm2 for the cross-sectional area of the nitrogen molecule at 77 K. Representative linear BET plots are depicted in Fig. 3, while the calculated SBET values for the investigated carbons are listed in column 3 of Table 1. The total pore volume VT (ml/g) expressed as the volume of liquid nitrogen adsorbed per gram carbon at relative pressure of ca. 0.98 p/p0 , is listed in column 4 of Table 1. Based on the assumption that the space in the micropores is similar to the space between two parallel plates, the average pore radius r¯ (nm) could be calculated from the relationship [31]: r¯ = 2VT
103 S BET
(1)
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
145
Fig. 1. Representative adsorption nitrogen isotherms at 77 K.
The values of r¯ as calculated for the carbons investigated are given in column 5 of Table 1. Two other independent methods were also applied to analyze the nitrogen adsorption isotherms, i.e. the t-method [32] and the αs -method [33]. In the first method the amount of ni-
trogen adsorbed (ml/g) is plotted versus the multilayer thick˚ as measured on a standard non-porous material ness t (A) of comparable BET-C constant (Fig. 4). The second method plots the volume of nitrogen adsorbed versus the reduced isotherm (αs ) determined on a standard non-porous material
Table 1 Textural parameters of carbons studied obtained from nitrogen adsorption at 77 K via application of models based on surface Sample
SL (m2 /g)
SBET (m2 /g)
VT (ml/g)
r¯ (nm)
St (m2 /g)
t Vmic (ml/g)
Snt (m2 /g)
Sα (m2 /g)
α Vmic (ml/g)
Snα (m2 /g)
A-600 AP25–400 AP25–500 AP25–600 AP50–400 AP50–500 AP50–600 AP75–400 AP75–500 AP75–600 AZ0.5–600 AZ0.5–700 AZ1.0–600 AZ1.0–700 AZ2.0–600 AZ2.0–700 AS25 AS35
270 431 530 576 594 759 759 700 874 1008 364 437 656 728 1017 1017 1008 1093
211 311 386 446 458 551 551 524 655 728 280 328 460 546 728 728 650 683
0.095 0.149 0.181 0.198 0.214 0.264 0.264 0.246 0.302 0.327 0.129 0.159 0.226 0.258 0.358 0.358 0.467 0.747
0.9 0.9 0.98 0.98 0.78 0.96 0.96 0.94 0.83 0.9 0.92 0.96 0.99 0.95 0.98 0.98 1.43 2.119
215 325 400 450 480 560 560 535 665 755 300 350 500 520 750 750 650 715
0.085 0.130 0.165 0.180 0.185 0.235 0.235 0.240 0.280 0.305 0.105 0.120 0.200 0.225 0.310 0.310 0.130 0.160
10 15 15 17 20 20 20 16 22 24 12 14 20 20 30 30 150 220
230 330 393 460 467 585 585 526 630 725 304 350 492 600 746 746 620 708
0.078 0.130 0.156 0.180 0.185 0.240 0.240 0.227 0.270 0.300 0.120 0.0132 0.210 0.235 0.320 0.320 0.134 0.160
9 12 13 15 16 18 18 15 20 21 12 13 18 20 27 27 148 230
146
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
Fig. 2. Representative linear Langmuir plots of nitrogen adsorption.
of the same chemical composition. The adsorption data reported by Sellez–Perez and Martin–Martinez [34] were used for such plots, representative α plots are shown in Fig. 5. The t-method allows the determination of the specific surface area St as determined from the slope of the solid line passing through the origin (Fig. 4). The micropore volume t is obtained from the intersection of the dotted line passVmic ing through the plateau of the V–t plot, zero thickness. The surface located in non-micropores Snt could also be calcut (ml/g) lated from the slope of the dotted line. St (m2 /g) Vmic t 2 Sn (m /g) are listed in columns 6–8 of Table 1, respectively. Similarly, the αs -method allows the determination of the same α and S α , respecthree parameters but with notations, Sα , Vmic n tively (columns 9–11 of Table 1). Inspection of the data recorded in Table 1 reveals the following: (1) Activation with phosphoric acid is associated with a considerable increase in the adsorption capacity compared with the non-activated carbon A-600. The degree of activation with phosphoric acid increased with the increase of the concentration of this activating agent from 25 to 75 wt.% and also with the rise of carbonization temperature from 400 to 600 ◦ C. This activating agent dehydrates
the carbonaceous material during carbonization and this leads to charring and aromatization of the carbon skeleton and the creation of a porous structure [35]. (2) Activation with zinc chloride should be undertaken at 700 ◦ C, particularly when the mass of zinc chloride used in activation is less than twice the mass of the carbonaceous precursor. (3) Gasification with steam at 900 ◦ C of A-600, to a burnoff equals 25% was found to be associated with more than three-fold increase in SL or SBET . However, further gasification to higher burn-off (35%) was found to be associated with a further but small increase in the surface area. Considering the total pore volume VT or the mean pore radius r, it is evident that the increase of burnoff from 25 to 35% was associated with more than 50% increase of these textural parameters. (4) The surface areas calculated through the application of the t-method and αs -method are comparable. These two methods are based on a standard non-porous material and each method can be used to complement the other. The data listed in Table 1 show that, in most cases the SBET values were lower than the St and Sα values, however, the difference was not high and may be attributed to the BET model being based on multilayer coverage, which
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
147
Fig. 3. Representative linear BET plots of nitrogen adsorption at 77 K.
is not precisely the case in adsorption on activated carbons where most of the pores accommodate only a limited number of layers as indicated by the values of their average radii (column 5 of Table 1). Some pores may accommodate only two layers, each one on the opposite walls. In addition, some pores may even be inaccessible to adsorbate molecules. It should be noted from the data in Table 1 that the calculated SL values are very high relative to the SBET , St and Sα values, an unsurprising fact since in most cases, the monolayer capacities determined by the application of Langmuir model lie beyond the equilibrium pressure employed. For this reasons, the surface areas of activated carbons based on Langmuir model should be taken with great reservation. (5) The surface area located in the micropore region represents a large fraction of the total surface area and consequently the surface located in the non-microporous region would be only a small fraction of the total surface area, however, steam-activated carbons AS25 and AS35 represent an exception. This refers to the importance of the method of activation. Steam-activation is frequently associated with pore widening, particularly at intermediate and high percentages of burn-off [36].
The physical adsorption of gases and vapours by microporous solids, in general, and by active carbons, in particular, may also be described by Dubinin’s theory as developed in successive stages since 1947 [37,38]. In the present formulation, the theory of micropore felling may be expressed by the Dubinin–Astakhov (DA) equation [39], which may be expressed as: A n V = V0 exp − (2) βE0 where V represents the volume filled at temperature T and relative pressure p/p0 , V0 is the total volume of micropores, the quantity A is equal to RT ln (p0 /p) and n, E0 and β are specific parameters of the system under investigation. The DA equation is applicable over the relative pressure range 0.05 < p/p0 < 0.1 where the influence of any non-microporous surface area is negligible. For typical active carbons, the exponent is equal to 2, which corresponds to the empirical equation postulated by Dubinin and Radushkevich in 1947 and known in the literature as the DR equation [37]: T 2 2 p0 V = V0 exp −B log (3) β p
148
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
Fig. 4. Representative V–t plots of nitrogen adsorption at 77 K.
The parameter B = (2.303 R/E0 )2 has the dimensions K−2 and is called the structural constant being related to the characteristics energy E0 by [40]: E0 (kJ/mol) =
0.01914 B0.5
(4)
The DR and DA equations are both based on the observation that a plot of log V versus (RT)2 ln2 (p0 /p) or An leads to an unique curve for a given adsorbate. The range of the classical DR equation has been verified for many-reported adsorption measurements [41,42]. The present authors are more inclined to the view that the state of the adsorbate in a microporous adsorbent is still a matter open for discussion, since evidence needs to be provided to give predominance for one postulate or another. For this reason, the nitrogen adsorption data of phosphoric acid-activated and zinc chlorideactivated carbons have been interpreted here by applying the DR equation and comparing some of the adsorption parameters thus obtained with those determined from the BET, αs and t-methods. Representative DR plots are depicted in Fig. 6 showing that linear plots covering a range of relative pressures were obtained. However, slight upward deviations from linearity at high relative pressure were observed, corresponding to micropores of size close to the mesoporosity region. Table 2 lists some of the adsorption parameters determined from the DR equation and includes parameters determined in other ways for the sake of comparison. It is seen from column 4 of Table 2 that the characteristic energies for adsorption ranged between 10.43 and 12.40 kJ/mol, i.e. they all lie in the range characterizing physical adsorption. This is expected for nitrogen in particular, since this molecule seldom undergoes specific interaction with any adsorbent at 77 K [20]. The values of X (nm), which give the pore radii are listed in column 6 of Table 2 and indicate micropore types, X = K/E0 where K = 13.03 − 1.53 × 10−5 E03.5 . However, it should be noted that in all cases these values are smaller than the values of r¯ listed in column 5 of Table 1. This may be attributed to the fact that r¯ values represent average pore radii, which contain a contribution from the non-micropores existing in a given sample. The consistency in the values of K listed in column 5 of Table 2 (12.93–12.97 nm kJ/mol)) may be attributed to the narrow range in which all E0 values exist, which is due to the location of the major fraction of the porosity in the micropore range. Evidently the micropore volumes as determined by the
Table 2 Application of the DR equation to nitrogen adsorption data at 77 K Sample
DR (ml/g) Vmic
D
E0 (kJ/mol)
K (nm kJ/mol)
X (nm)
SDR (m2 /g)
V0.1 (ml/g)
100 V0.1/VT
DR /V 100 Vmic T
A-600 AP25–400 AP25–500 AP25–600 AP50–400 AP50–500 AP50–600 AP75–400 AP75–500 AP75–600 AZ0.5–600 AZ0.5–700 AZ1.0–600 AZ1.0–700 AZ2.0–600 AZ2.0–700
0.09 0.148 0.169 0.191 0.204 0.252 0.252 0.238 0.295 0.316 0.127 0.158 0.221 0.251 0.353 0.353
0.06 0.07 0.053 0.06 0.06 0.06 0.06 0.06 0.07 0.07 0.072 0.072 0.07 0.075 0.072 0.072
11.66 10.8 12.4 11.66 11.66 11.66 11.66 11.66 10.8 10.8 10.65 10.65 10.8 10.43 10.65 10.65
12.95 12.97 12.93 12.95 12.95 12.95 12.95 12.95 12.97 12.97 12.97 12.97 12.97 12.97 12.97 12.97
1.11 1.2 1.04 1.11 1.11 1.11 1.11 1.11 1.2 1.2 1.22 1.22 1.2 1.24 1.22 1.22
254 418 476 538 575 710 710 671 832 890 364 445 623 724 995 995
0.08 0.129 0.156 0.174 0.186 0.234 0.234 0.212 0.265 0.298 0.11 0.134 0.198 0.226 0.318 0.318
84 91 86 88 87 89 89 86 88 91 85 95 87 88 88 88
95 99 93 96 95 95 95 97 98 91 99 99 97 97 99 99
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
149
Fig. 5. Representative α-plots of nitrogen adsorption at 77 K.
DR-method represented 94–98% of the total pore volumes. If the micropore values as determined by the DR equation are converted to equivalent surface areas, it is possible to obtain a set of SDR (m2 /g) values (column 7 of Table 2). It will be noted that in most cases the SDR values listed are smaller than the comparable SL values. It has been mentioned already that the Langmuir model on which SL values are based gives monolayer capacities beyond the equilibrium pressure and hence conversion of monolayer capacities into specific surface areas is of limited significance. Equally for chemically activated carbons, the SDR values are always greater than the SBET values, unsurprising to the authors because each set of values is based on a different model. The pore diameters of the chemically activated carbons investigated as determined by the DR methods are centered between 1.04 and 1.24 nm. On the other hand these pore diameter were in the range 1.56–1.98 nm when the BET model was considered. This means that although the two model have predicted the micropore structure, agreement was not found to exist concerning the appropriate range in which these pores were
existing. The textural parameters as determined from the DRmodel and from the BET model should be taken with great reservation. The textural properties determined from the tmethod and αs -method, both based on a standard nonporous reference, remain more trustworthy. Column 8 of Table 2 lists the volume of adsorbate taken up at p/p0 = 0.1, a relative pressure at which micropores are filled with adsorbate molecules (i.e. nitrogen molecules at 77 K). All the chemically activated carbons investigated exhibit a microporosity representing 84–91% of the total porosity in the sample (column 9 of Table 2). The same conclusion may be reached by dividing the micropore volumes as estimated from the DR equation by the total pore volumes, the corresponding values listed in column 10 of Table 2 indicating that for all chemically activated carbons investigated, the micropore fraction ranged between 91 and 99% of the total porosity. To summarize the observed difference, in the determination of surface area and porosity of chemically activated carbons, by different models refer to the heterogeneity of these sorbents and indicates that the determination of their textural
150
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151
Fig. 6. Representative DR plots of nitrogen adsorption at 77 K.
parameters is a complex problem. The conventional classification of pore, which consider micropores as these with diameter < 2.0 nm needs to be reconsidered. In the pores with diameters < 2.0 nm, more than one model of adsorption can be considered; i.e. activated diffusion in micropores, micropore filling and surface coverage. Each of these model probably predominates in a certain narrow range of microporosity, which should be subdivided into sub-micropore ranges. However, on can count on the methods based on comparing the isotherm determined for the given active carbon with the standard isotherm for nonporous adsorbents, which acts as a reference system.
References [1] G. Mckay, J. Chem. Technol. Biotechnol. 40 (1982) 759. [2] A.M. Youssef, A.A. EL-Khouly, A.I. Ahmed, E.T. EL-Shafy, Adsorpt. Sci. Technol. 12 (1994) 211. [3] A.M. Youssef, A.A. EL-Khouly, A.I. Ahmed, E.I. EL-Shafey, J. Environ. Sci. Mansoura 10 (1995) 195. [4] R.C. Bansal, J. Donnet, F. Stoeckli, Active Carbon, Marcel Dekker Inc., New York, 1988, p. 335.
[5] A.M. Youssef, A.M. El-Wakily, K. Tollan, Th. El-Nabarawy, Afinidad 47 (1990) 336. [6] F. Derbyshire, M. Jagtoyen, M. Thwaites, in: J.W. Patrick (Ed.), Porosity in Carbons, Edward Arnold, London, 1995, p. 228. [7] P. Davini, Carbon 39 (2001) 1387. [8] S.H. Abdel-Halim, A.M.A. Shehata, M.E. El-Shahat, Water Res. 37 (2003) 1678. [9] C.O. Ania, J.B. Parra, J.J. Pis, Fuel Process. Technol. 77–78 (2002) 337. [10] N. Klimenko, M. Winter-Nielsen, S. Smolin, L. Nevynna, J. Sydorenko, Water Res. 36 (2002) 5132. [11] A.M. Youssef, M.R. Mostafa, E.M. Dorgham, Bull. Soc. Chim. Fr. (1989) 741. [12] A.M. Youssef, A.M. El-Wakil, E.A. El-Sharkawy, A.B. Farag, Kh. Tollan, Adsorpt. Sci. Technol. 13 (1996) 115. [13] V. Strelko, D.J. Malik, J. Colloid Interface Sci. 250 (2002) 213. [14] A. Zhang, G. Vchiyama, T. Asakura, Adsorpt. Sci. Technol. 21 (2003) 761. [15] A.M. Youssef, Th. El-Nabarawy, S.E. Samra, Colloids Surf. 235 (2004) 153–163. [16] B. Cagnon, X. Py, A. Gusllot, F. Stoeckli, Microporous Mesoporous Mater. 57 (2003) 273. [17] O.N. Baklanova, G.V. Plaksin, V.A. Drozdov, V.K. Duplyakin, N.V. Chesnokov, B.N. Kuznetsov, Carbon 41 (2003) 1793.
A.M. Youssef et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 252 (2005) 143–151 [18] Th. El-Nabarawy, M.R. Mostafa, A.M. Youssef, Adsorpt. Sci. Technol. 15 (1997) 59. [19] I.I. Salame, T.J. Bandosz, J. Colloid Interface Sci. 210 (1999) 367. [20] A.M. Youssef, T.M. Ghazy, Th. El-Nabarawy, Carbon 20 (1982) 248. [21] A. Gil, G. dela Puente, P. Grange, Microporous Mater. 12 (1997) 51. [22] J. Chorna, W. Burakiewicz-Mortka, M. Jaroniec, Z. Li, J. Klinik, J. Colloid Interface Sci. 214 (1999) 438. [23] L.B. Khalil, Adsorp. Sci. Technol. 13 (1996) 317. [24] A.M. Youssef, Carbon 13 (1975) 1. [25] A.M. Youssef, G.A. El-Shobaky, Th. El-Nabarawy, Surf. Technol. 7 (1978) 469. [26] A.M. Youssef, J. Colloid Surf. Sci. 55 (1976) 447. [27] S. Brunauer, L.S. Deming, W.E. Deming, E. Teller, J. Am. Chem. Soc. 62 (1940) 1723. [28] I. Langmuir, J. Am. Chem. Soc. 38 (1916) 2219. [29] I. Langmuir, J. Am. Chem. Soc. 40 (1918) 1368. [30] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309.
151
[31] P. Podko´scielny, A. Dabrowski, M. B¨ulow, Appl. Surf. Sci. 196 (2002) 312. [32] B.C. Lippens, J.H. deBoer, J. Catal. 4 (1965) 319. [33] K.S.W. Sing, Chem. Ind. (Lond.) (1968) 1520. [34] M.J. Sellez-Perez, J.M. Martin-Martinez, J. Chem. Soc. Faraday Trans. 87 (1991) 1237. [35] M.N. Alaya, M.R. Hourieh, A.M. Youssef, F. El-Sejariah, Adsorpt. Sci. Technol. 18 (2000) 27. [36] A.M. Youssef, M.N. Alaya, N. Nawar, Adsorpt. Sci. Technol. 11 (1994) 225. [37] M.M. Dubinin, E.D. Zaverina, L.V. Radushkevich, Zh. Fiz. Khim. 21 (1947) 1351. [38] M.M. Dubinin, Carbon 21 (1985) 359. [39] M.M. Dubinin, V.A. Astakhov, Izv. Akad. Nauk SSSR, Ser. Khim. (1971) 5. [40] R.C. Bansal, J. Donnet, F. Stoeckli, Active carbon, Marcel Dekker Inc., New York, 1988, p. 132. [41] A.M. Youssef, T.M. Ghazy, Th. El-Nabarawy, Carbon 20 (1982) 218. [42] A.M. Youssef, A.M. El-Wakil, K. Tollan, Sepn. Sci. Technol. 26 (1990) 445.