Chemically Activated Carbons from Olive Stones ...

651

Chemically Activated Carbons from Olive Stones — Peculiarities of Pore Structure and Interpretation of Nitrogen Adsorption Data M.A. Hourieh1, M.N. Alaya1*, F. El-Sejariah 1 and A.M. Youssef 2

(1) Faculty of Science, Aleppo

University, Aleppo, Syria. (2) Faculty of Science, Mansoura University, Mansoura, Egypt. (Received 23 June 2000; accepted 21 July 2000)

ABSTRACT: Chemically activated carbons were obtained from olive stones either by carbonization with H3PO4 at 300–600ºC or by carbonization with ZnCl2 at 600ºC. Nitrogen adsorption at 77 K was determined for all the activated carbons. The adsorption data were interpreted by considering some conventional adsorption models. Maximum activation with H3PO4 occurred at 450ºC. However, the adsorption capacities of the ZnCl2-activated carbons were far higher than those of carbons activated with H 3 PO 4 . Carbons activated with H 3 PO 4 or ZnCl 2 are mainly microporous with the non-micropores representing a small fraction of the total porosity. Although, the nitrogen isotherms are Langmuirian in shape, application of the Langmuir equation led to large monolayer capacities of uncertain confidence. The surface areas and micropore volumes determined by the application of the t-method of de Boer and the a S-method of Sing were comparable and were slightly higher than those determined by the application of the DR model based on micropore filling. The t-method and the a S-method are complementary to each other and would seem to give confident values because they are based on standard reference non-porous materials. The micropore region may be sub-divided into two sub-regions distinguished by the different filling mechanisms involved.

INTRODUCTION Activated carbons are excellent adsorbents because of their extended surface area, microporous structure, high adsorption capacity and high degree of reactivity (Asbak et al. 1981; Gregg and Sing 1982; Hourieh et al. 1997; Attia 1997). Their important applications relate to their use in removal of colour (McKay et al. 1998), odour (Youssef et al. 1990a,b) and undesirable organic impurities from potable water, in the treatment of domestic and industrial water (Mostafa et al. 1989; Singer and Yen 1980; Eltekova and Eltekov 1997; El-Nabarawy et al. 1997), solvent recovery, air purification (Youssef et al. 1997; Keaney and Alder 1998) and in a variety of gas-phase applications (Richter et al. 1987). Activated carbons are also used as catalysts and catalyst supports (Boehm et al. 1984; Zuckmantel and Eltekov 1979). Thus, the applications of activated carbons are of interest to most economic sectors and concern many diverse areas (Bansal et al. 1988). Any cheap material with a high carbon content and low in inorganics can be used as a precursor for the preparation of activated carbon. Coal is the most commonly used precursor (Razouk et al. 1968) due to its low cost and large supply. However, in some areas, agricultural wastes such as coconut shells (Puri et al. 1971), fruit stones (Rodriguez-Reinoso et al. 1985a,b), olive stones *Author to whom all correspondence should be addressed.

652

M.A. Hourieh et al./Adsorption Science & Technology Vol. 18 No. 7 2000

(Lopez-Gonzalez 1980; Alaya et al. 2000), nut shells (Ferro-Garcia et al. 1990) and sawdust (Yehia 1996) represent an abundant source. Two methods are known for the preparation of activated carbons. These are as follows. (i) Chemical activation in which the precursor is impregnated by an aqueous solution containing the activating agent (phosphoric acid or zinc chloride), usually by mixing and kneading. The chemically impregnated material is then extruded and pyrolyzed in a kiln at 400–600ºC in the absence of air or in an inert atmosphere. (ii) Physical activation which is normally conducted initially at 400–900ºC in an inert atmosphere followed by activation to a certain percentage burn-off carried out at 800–1100ºC in the presence of suitable oxidizing gases such as carbon dioxide or steam. Adsorption on microporous sorbents, in general, and on activated carbons, in particular, exhibits some peculiarities and its interpretation is still a subject of conflict among several investigators. It was the purpose of the present work to prepare activated carbons from olive stones by activation either with phosphoric acid or zinc chloride. The textual properties of the activated carbons were determined by nitrogen adsorption at 77 K and by applying the most conventional models used for the interpretation of adsorption data. EXPERIMENTAL Materials Phosphoric acid-activated carbons were obtained by impregnating dry crushed olive stones (1– 2 mm diameter) in 100 ml of phosphoric acid at a 50 wt% concentration. The resulting mass was heated gradually, in the absence of air, over 2 h up to 300ºC, 400ºC, 450ºC, 500ºC or 600ºC, respectively, then maintained at the maximum temperature for a further 3 h. After cooling, the carbonized mass was washed thoroughly with water until the pH of the latter attained a value of 6.0 and then dried. The activated carbons thus obtained were designated OP-300, OP-400, OP-450, OP-500 and OP-600 according to the carbonization temperature employed. More details regarding this method of preparation are given elsewhere (Khalil 1996; Hourieh et al. 1999). Four zinc chloride-activated carbons were obtained by impregnating 400 g of dry ground olive stones (1–2 mm diameter) in 200 ml of an aqueous solution containing 400 g anhydrous ZnCl2. Two samples were removed after 72 h impregnation, one being carbonized at 600ºC in the absence of air to give OZF-600 and the other also carbonized at 600ºC but in a nitrogen atmosphere to give OZFN-600. After removal from the impregnating solution, the other two samples were heated to complete dryness at 70ºC and the dried mixture was then carbonized at 600ºC either in a limited air supply or in a nitrogen atmosphere to give OZ-600 and OZN-600, respectively. All the carbonization products were washed thoroughly with distilled water until free from chloride ions and then dried. More details regarding this method of preparation are also given elsewhere (Youssef et al. 1978). Adsorption measurements The adsorption of nitrogen at 77 K was measured using a Micromeretics Gemini III 2375 surface area analyzer (Micromeretics, Norcross, GA, USA). Prior to all measurements, the solids were heated overnight at 548 K under high vacuum (10–4 Torr). RESULTS AND DISCUSSION The adsorption of nitrogen at 77 K proved to be rapid, with equilibrium being attained in less than 40 min at relative pressures < 0.1 and in less than 20 min at higher relative pressures. This indicated

Chemically Activated Carbons from Olive Stones

Figure 1. Representative nitrogen adsorption isotherms measured at 77 K.

653

654


Figure 2. Representative linear BET plots for nitrogen adsorption isotherms measured at 77 K.

655


that almost all the pores in the materials were accessible to nitrogen molecules at 77 K and that this was true for all the investigated carbons. The desorption points all lay on the same adsorption isotherm, indicating the absence of hysteresis characteristic of mesoporosity or specific interaction (Youssef 1976). The adsorption isotherms were typically type I in the BDDT classification which is characteristic of adsorption in microporous sorbents (Brunauer et al. 1940). Figure 1 depicts the nitrogen adsorption isotherms at 77 K on the carbons investigated. The conventional BET equation (Brunauer et al. 1938) was applied to determine the monolayer capacity and consequently the surface area SBET (m 2/g) by adopting the value of 0.162 nm2 for the cross-sectional area of the nitrogen molecule at 77 K. The linear BET plots thus obtained are depicted in Figure 2. The SBET values of the investigated carbons are listed in column 2 of Table 1 while the total pore volume VT (ml/g), expressed as the volume of liquid nitrogen taken up by 1 g of carbon at a relative pressure of ca. 0.95, is listed in column 3 of the same table. Based on the assumption that the space in the micropores is similar to the space between two parallel plates, the average pore radius (nm) was calculated from the relationship: = 2VT × 103/SBET. The values of thereby obtained for the carbons investigated are listed in column 4 of Table 1. It has been already mentioned that the nitrogen adsorption isotherms of the investigated carbons were all of type I, i.e., they were typical Langmuirian in shape. Application of the Langmuir equation (Langmuir 1916, 1918) was found to be satisfactory and covered a wide range of pressure (Figure 3). The monolayer capacities were determined and the surface areas SL calculated from these plots (column 5 of Table 1). Two other independent methods were also applied to analyze the nitrogen adsorption isotherms, i.e. the t-method (Lippens and de Boer 1965) and the aS-method (Sing 1968). The first method involves plotting the amount of gas adsorbed (ml/g) versus the multilayer thickness (t in Å) as measured on a non-porous carbon of comparable BET-C constant (Figure 4). The second method involves plotting the volume of gas adsorbed versus the reduced isotherm values (a) determined on a standard non-porous material of the same chemical composition. Adsorption data reported recently (Sellez-Perez and Martin-Martinez 1991) were used in the plots (Figure 5). The t-method allows the determination of the surface area St, the micropore volume (Vmic ) and t the non-microporous area (Snt ) (columns 6–8 of Table 1). Similarly the aS-method allows the determination of these three parameters with the notations Sa, Vamic and San (columns 9–11 of Table 1).

TABLE 1. Some Textural Parameters for Carbons Investigated Sample

SBET (m2/g)

VT (ml/g)

(nm)

SL (m2/g)

St (m2/g)

Vmic t (ml/g)

Snt (m2/g)

77 619 660 651 616

0.0345 0.2754 0.2993 0.2922 0.2779

0.90 0.89 0.91 0.90 0.90

97 780 856 838 711

72 699 758 727 656

0.0325 0.2552 0.2672 0.2785 0.2429

OZ-600 1310 OZN-600 1236 OZF-600 620 OZFN-600 646

0.6751 0.6232 0.2885 0.3023

1.03 1.01 0.93 0.94

2013 1832 829 868

1314 1207 678 719

0.6405 0.5840 0.2746 0.2823

OP-300 OP-400 OP-450 OP-500 OP-600

Sa (m2/g)

Vamic (ml/g)

San (m2/g)

2 18 30 16 17

71 704 777 737 704

0.0325 0.2522 0.2661 0.2754 0.2591

2 21 32 17 18

55 57 13 12

1320 1246 675 730

0.6149 0.5616 0.2723 0.2831

58 59 16 14

656


Figure 3. Representative linear Langmuir plots for nitrogen adsorption isotherms measured at 77 K.

Inspection of the data listed in Table 1 reveals the following: (i) Activation with phosphoric acid should be undertaken at 400–500ºC, the optimum temperature for activation with phosphoric acid being 450ºC. Activation with phosphoric acid at 300ºC appears to have been incomplete as predicted from the low surface area of sample OP-300 which suggests that the porous structure of this sample was poorly developed. During carbonization, the impregnated chemicals dehydrate the carbonaceous material, resulting in charring and aromatization of the carbon skeleton and the creation of a porous structure. It seems that at 300ºC phosphoric acid fails to generate these textural variations. Increasing the activation temperature with phosphoric acid to 400ºC was found to be associated with a ca. eightfold increase in the surface area compared to activation at 300ºC. A further increase in the activation temperature to 450ºC resulted in a ca. 10% increase in the surface area. However, above 450ºC, a decrease in the surface area of the product was found, with this decrease becoming more pronounced at higher temperatures. (ii) Activation with zinc chloride at 600ºC gave high surface area products compared to activation with phosphoric acid at the optimum temperature (450ºC). Thus, the surface area of OZ-600 was approximately twice that of OP-450. The surface areas of zinc chloride-activated samples carbonized in the absence of air were very close to those of samples carbonized in a nitrogen


Figure 4. Representative Vl versus t plots for nitrogen adsorption measured at 77 K.

657

658


Figure 5. Representative aS-plots for nitrogen adsorption measured at 77 K.


659

atmosphere. Thus the SBET values of OZ-600 and of OZN-600 were 1310 m2/g and 1236 m2/g, respectively, i.e. exhibiting a difference of ca. 5%. Also, the difference between the SBET values of OZF-600 and OZFN-600 amounted to only ca. 4%. On the other hand, the SBET value of OZF-600 was ca. one-half that of OZ-600. This suggests that the amount of zinc chloride used in the activation process is an important factor in determining the textural properties of the activated carbonization product. It also predicts that the function of zinc chloride as an activating agent predominates during carbonization. On this basis, it could be recommended that the mixture (carbonaceous material + zinc chloride) should be transferred without separation to the activation reactor. (iii) The surface areas calculated by the application of the t-method and the aS-method are comparable, with the difference being less than 7%. These two methods are based on a standard non-porous material and each method can be used in a complementary manner to the other. In most cases, the SBET values were lower than the St and Sa values. The BET model assumes multilayer formation which is not actually the case in activated carbons. Thus, in activated carbons, most pores accommodate a very limited number of layers, some pores accommodating only two layers each on the opposite pore wall and in some fine pores only one layer may be accommodated shared between the opposite walls. In addition, some pores may even be inaccessible to adsorbate molecules. On these bases, the SBET values measured for the activated carbons should only be accepted with some reservation. Again, the SL values were very high in comparison to the St and Sa values. The Langmuir model is based on monolayer adsorption and in most cases the monolayer capacities as determined by the application of the Langmuir equation lie beyond the equilibrium pressures. Hence, the Langmuir model cannot be used for the interpretation of the adsorption data for activated carbons. Similar arguments have been reported concerning chars derived from Saran (Culver and Heath 1955). Thus, for Saran chars, a surface area of 3000 m2/g was calculated by applying the Langmuir equation, a figure which is actually higher than the area (2630 m2/g) which could be provided by 1 g carbon if this were present solely as graphite layers of only one atom thickness and accessible on both sides to gas molecules. A tenuous structure of this kind would be very difficult to reconcile with the mechanical strength of the material. (iv) The surface areas located in the micropore region represent a large fraction of the total surface areas and consequently the surface located in the non-microporous region must be a very small fraction of the total surface of the activated carbon sorbents. This is predicted from columns 8 and 11 of Table 1 listing Snt and San, from which a good agreement evidently existed between these two parameters. It is also evident that good agreement existed between Vmic and Vmic , a t which provides additional evidence that the t-method and aS-method are complementary to each other. The average pore radius, , of most of the phosphoric acid-activated carbons was less than 1.0 nm, again demonstrating the predominance of microporosity in the materials studied. The same is also true for zinc chloride-activated carbons where = 1.0 ± 0.07 nm. The physical adsorption of gases and vapours by microporous solids in general, and by activated carbons in particular, may also be described by Dubinin’s theory as developed in successive stages since 1947 (Dubinin et al. 1947; Dubinin and Astakhov 1971; Dubinin and Stoeckli 1980; Dubinin 1985). In its present formulation, the theory of micropore filling may be expressed by the equation of Dubinin and Astakhov (DA equation), i.e. W = W0 exp[– (A/b E0)n]

(1)

where W depicts the volume filled at a temperature T and a relative pressure P/P0, W0 is the total

660


volume of micropores, the quantities A = RT ln(P0/P), n, E0 and b being specific parameters of the system under investigation. The DA equation is applicable over the relative pressure range 0.05– 0.1 where the influence of the non-microporous surface area is negligible. For typical active carbons the exponent n is equal to 2, which corresponds to the original empirical equation postulated by Dubinin and Radushkevich in 1947 and known in the literature as the DR equation, i.e. W = W0 exp[–B(T/b )2 log2(P0/P)]

(2)

In this equation, the parameter B has the dimensions of K–2 and is called the structural constant. It is related to the characteristic energy, E0, via the equation E0 (kJ/mol) = 0.01914/B0.5. The DR and DA equations are based on the observation that a plot of W versus (RT)2 log2(P0/P) or An leads to an unique curve for a given adsorbate. The range of applicability of the classical DR equation has been verified for many reported adsorption measurements (Youssef and El-Wakil 1980; Youssef et al. 1990a,b). The present authors are more inclined to consider the state of the adsorbate in microporous sorbents as a matter still open for discussion. Evidence is required to allow predominance for one postulate or the other. For this reason, the nitrogen adsorption data are interpreted here by applying the DR equation and some of the adsorption parameters thereby obtained compared with those determined from the BET, aS- and t-methods. The corresponding DR plots for the samples studied in the present work are depicted in Figure 6. Linear plots covering a range of relative pressures were obtained, although upward deviations from linearity were observed at relatively high pressures P/P0 > 0.10. It is predicted that adsorption in micropores may be more or less complete at P/P0 = 0.10; above this value of the relative pressure, the adsorption of nitrogen may take place in nonmicropores which probably only comprise a small fraction of the total porosity. Table 2 lists some of the adsorption parameters determined from the DR equation. Also included for comparative purposes are other parameters determined by other models. It is evident from the data listed in column 4 of Table 2 that the characteristic energies for adsorption lie in the range corresponding to physical adsorption. This is not surprising, particularly for nitrogen which seldom undergoes specific interaction with any sorbent at 77 K (Youssef 1980). The values of x (nm), which give the pore radii, are listed in column 5 of Table 2 and indicate a micropore type. However, in most cases, these values are smaller than the values, suggesting that the latter values contain a contribution from non-micropores existing in the samples. If the micropore volumes as determined from the DR equation are converted to equivalent surface areas, a set of areas SDR (m2/g) are obtained (column 3 of Table 2). Such SDR areas were always lower than the SL values, the difference reaching ca. 29% for sample OZ-600. It has been mentioned already that the Langmuir model gives monolayer capacities beyond the equilibrium pressure and that the conversion of this monolayer capacity into a specific surface area is not sound. The values of SDR were always higher than the SBET values, the difference amounting to 5–16% in this case. It is obvious that this difference increased with an increase in the upward deviation observed in the DR plots (Figure 5). This point will be discussed in more detail below. Fair agreement is observed when the SDR and Sa values are compared. For some samples, SDR and Sa values are comparable (OP-400, OP-450, OP-500, OZF-600 and OZFN-600) with a difference of less than 3%. However, higher differences were found for other samples investigated, with the maximum being 10% for sample OZ-600. The same trend but with slightly less agreement was observed when the SDR areas were compared with the St values. Column 9 of Table 2 lists the values of VDmicR /Vmic for all the samples investigated. Evidently, this t ratio ranged between 0.79 and 1.01. Similarly, the VDmicR /Vamic ratio listed in column 10 of Table 2 displays a range of 0.86–1.02.


661

Figure 6. Representative Dubinin–Radushkevich (DR) plots for nitrogen adsorption measured at 77 K.

We now discuss the agreement and disagreement between the surface areas as determined from a consideration of the different models and also the correlation between the micropore volumes as calculated by the application of these models. If we consider two different viewpoints, namely the surface coverage and micropore filling, it appears as if an overlap exists between more than one group of pores in the samples studied. Activated carbons with relatively wide micropore distributions cannot readily be characterized in such a way as to obtain confident values of the pore volume and adsorption capacity (Rodriguez-Reinoso et al. 1985a,b; Garrido et al. 1987). However,

662


TABLE 2. DR Textural Parameters and their Correlation with those Determined by the t-Method and the aS-Method Sample

VDmicR (ml/g)

SD R (m2/g)

E0 (kJ/mol)

x (nm)

V0.1 (ml/g)

SD R/S t

SD R/S a

VDmicR /Vmic t

VDmicR /Vamic

OP-300 OP-400 OP-450 OP-500 OP-600

0.0316 0.2480 0.2688 0.2609 0.2288

89 700 759 737 646

8.79 18.07 17.39 17.79 19.27

1.48 0.70 0.73 0.71 0.65

0.0181 0.2123 0.2312 0.2249 0.2053

1.23 1.00 1.00 1.01 0.98

1.25 0.99 0.98 1.00 0.92

0.97 0.97 1.01 0.94 0.94

0.97 0.98 1.01 0.95 0.88

OZ-600 OZN-600 OZF-600 OZFN-600

0.5064 0.4587 0.2452 0.2658

1420 1296 693 751

14.29 15.54 17.08 17.39

0.90 0.82 0.74 0.73

0.4926 0.4648 0.2084 0.2289

1.09 1.07 1.02 1.04

1.08 1.04 1.03 1.03

0.79 0.79 0.89 0.94

0.82 0.82 0.90 0.93

the present authors believe that confident values may be obtained when the adsorption data are analyzed on the basis of reference data reported for non-porous solids, such as the aS-method or the t-method. Comparable surface areas and micropore volumes were obtained when these two models were considered for the analysis of the nitrogen adsorption data for the investigated carbons. The difference between Sa and St on the one hand and SDR on the other hand and also between Vamic and Vmic on the one hand and VDmicR on the other hand can be explained as follows. Micropores t are conventionally classified as pores with diameters less than or equal to 20 Å (2 nm). This original concept must be reformulated since the micropore region may be subdivided into two sub-regions distinguished by the different adsorption mechanisms involved. In both regions, the uptake at a given relative pressure is higher than it would be on a corresponding open surface. In the lower sub-region, pore filling predominates while the mechanism of surface coverage operates in the higher sub-region, the pore diameter range within each sub-region depending on the geometry and the nature of the adsorbate molecule. When small mesopores exist, the surface coverage mechanism operates over the entire range of relative pressure. Overlap between the existing pores may explain the differences between the textural parameters determined by the different adsorption models. Models based on a reference non-porous solid remain the most successful in determining the textural parameters of activated carbons with confidence. CONCLUSIONS Activation with phosphoric acid yields a maximum porosity at 450ºC and is associated with dehydration of the carbonaceous material and charring and aromatization of the carbon skeleton to create a porous structure. Activation with zinc chloride is more efficient relative to activation with phosphoric acid. Zinc chloride activation gave a product with a surface area approximately twice the area of phosphoric acid-activated carbons. The calculated surface area as determined by nitrogen adsorption varied with the model used in the interpretation of the adsorption data. The Langmuir model gave high surface areas, the BET model gave minimum values whilst the t-method and the aS-method gave comparable surface areas lying between those determined by the Langmuir and BET models.


663

In general, chemical activation gave microporous carbons with a very small fraction of surface area located in the non-micropores. Again the t-method and the aS-method may be used successfully to determine the micro- and non-microporous structures. It is more convenient to consider that the microporosity includes two sub-regions. Pore filling may predominate in the lower subregion while surface coverage predominates in the higher sub-region. REFERENCES Alaya, M.N., Hourieh, M.A., Youssef, A.M. and El-Sejariah, F. (2000) Adsorp. Sci. Technol. 18, 27. Asbak, T.M.W., Gouwerok, M. and Polman, E. (1981) Stakach/Strake 33, 378. Attia, A.A. (1997) Adsorp. Sci. Technol. 15, 707. Bansal, R.C., Donnet, J. and Stoeckli, F. (1988) Active Carbon, Marcel Dekker, New York, p. 335. Boehm, H.P., Muir, G., Stoehr, T., de Rincon, A.R. and Tereczki, B. (1984) Fuel 63, 1061. Brunauer, S., Emmett, P.H. and Teller, E. (1938) J. Am. Chem. Soc. 60, 309. Brunauer, S., Deming, L.S., Deming, W.E. and Teller, E. (1940) J. Am. Chem. Soc. 62, 1723. Culver, R.U. and Heath, N.S. (1955) Trans. Faraday Soc. 51, 1596. Dubinin, M.M. (1985) Carbon 21, 359. Dubinin, M.M. and Astakhov, V.A. (1971) Izv. Akad. Nauk SSSR, Ser. Khim. 5. Dubinin, M.M. and Stoeckli, H.F. (1980) J. Colloid Interface Sci. 75, 34. Dubinin, M.M., Zaverina, E.D. and Radushkevich, L.V. (1947) Zh. Fiz. Khim. 21, 1351. El-Nabarawy, Th., Mostafa, M.R. and Youssef, A.M. (1997) Adsorp. Sci. Technol. 15, 59. Eltekova, N.A. and Eltekov, Yu.A. (1997) Adsorp. Sci. Technol. 15, 109. Ferro-Garcia, M.A., Carrasco-Martin, F., Rivero-Utrilla, J., Utera-Hidalgo, E. and Moreno Castilla, C. (1990) Carbon 28, 91. Garrido, J., Linares-Solano, A., Martin-Martinez, J.M., Molino-Sabio, M., Rodriguez-Reinoso, F. and Torregrosa, R. (1987) Langmuir 3, 76. Gregg, S.J. and Sing, K.S.W. (1982) Adsorption, Surface Area and Porosity, Academic Press, London. Hourieh, M.A., Alaya, M.N. and Youssef, A.M. (1997) Adsorp. Sci. Technol. 15, 419. Hourieh, M.A., Alaya, M.N., Youssef, A.M. and El-Sejariah, F. (1999) Adsorp. Sci. Technol. 17, 675. Keaney, A. and Alder, J.F. (1998) Adsorp. Sci. Technol. 16, 101. Khalil, L.B. (1996) Adsorp. Sci. Technol. 13, 317. Langmuir, I. (1916) J. Am. Chem. Soc. 38, 2219. Langmuir, I. (1918) J. Am. Chem. Soc. 40, 1368. Lippens, B.C. and de Boer, J.H. (1965) J. Catal. 4, 319. Lopez-Gonzales, J.D., Martinez Vilchez, F. and Rodriguez-Reinoso, F. (1980) Carbon 18, 413. McKay, G., Yee, F.F., Nassar, M.M. and Magdy, Y. (1998) Adsorp. Sci. Technol. 16, 623. Mostafa, M.R., Samra, S.E. and Youssef, A.M. (1989) Indian J. Chem. 28A, 946. Puri, B.R., Kumar, B. and Kalra, K.C. (1971) Indian J. Chem. 9A, 970. Razouk, R.I., Saleeb, F.Z. and Youssef, A.M. (1968) Carbon 6, 325. Richter, E., Knoblauch, K. and Juntgen, H. (1987) Gas. Sep. Purif. 1, 35. Rodriguez-Reinoso, F., Martin-Martinez, J.M., Peter Lodo, I. and Prado-Burgude, M. (1985a) Carbon 23, 19. Rodriguez-Reinoso, F., Martin-Martinez, J.M., Molina-Sabio, M., Torregrosa, R. and Garrido, J. (1985b) J. Colloid Interface Sci. 106, 315. Selles-Perez, M.J. and Martin-Martinez, J.M. (1991) J. Chem. Soc., Faraday Trans. 87, 1237. Sing, K.S.W. (1968) Chem. Ind. (London) 1520. Singer, P.C. and Yen, C.Y. (1980) Activated Carbon Adsorption of Organics from the Aqueous Phase, Suffet, J.H., McGuire, M.J., Eds, Ann Arbor Science, Ann Arbor, MI, Vol. 1, p. 167. Yehia, H.M. (1996) Adsorp. Sci. Technol. 13, 367. Youssef, A.M. (1976) J. Colloid Interface Sci. 55, 447. Youssef, A.M. (1980) J. Res. Inst. Catal. Hokkaido Univ. 28, 89.

664


Youssef, A.M. and El-Wakil, A.M. (1980) Surf. Technol. 10, 303. Youssef, A.M., El-Shobaky, G.A. and El-Nabarawy, Th. (1978) Surf. Technol. 7, 451. Youssef, A.M., El-Wakil, A.M. and Tollan, K. (1990a) Sep. Sci. Technol. 26, 445. Youssef, A.M., Mostafa, M.R. and Dorgham, E.M. (1997) Adsorp. Sci. Technol. 15, 803. Youssef, A.M., El-Wakil, A.M., Tollan, K. and El-Nabarawy, Th. (1990b) Afinidad XLVII, 336. Zuckmantel, N.A. and Eltekov, Yu.A. (1979) Z. Naturforsch. 34b, 188.