Adsorption from benzene-ethanol binary solutions on activated carbons with different contents of oxygen surface complexes

Adsorption from benzene-ethanol binary solutions on activated carbons with different contents of oxygen surface complexes

Carbon Vol 21. No. 2. pp. 117-120, 1983 Printed m Great Britain @XX-6223/83IMOI 174%03.0010 @ 1983 Pergamon Press Ltd. ADSORPTION FROM BENZENE-ETHAN...

442KB Sizes 0 Downloads 22 Views

Carbon Vol 21. No. 2. pp. 117-120, 1983 Printed m Great Britain

@XX-6223/83IMOI 174%03.0010 @ 1983 Pergamon Press Ltd.

ADSORPTION FROM BENZENE-ETHANOL BINARY SOLUTIONS ON ACTIVATED CARBONS WITH DIFFERENT CONTENTS OF OXYGEN SURFACE COMPLEXES H. JANKOWSKA and A. ~QTOWSKI MilitaryTechnicalAcademy,00-908Warsaw,Poland and J. OSCIKand R. KUSAK Departmentof PhysicalChemistry,Institute of Chemistry,M. Curie-SktodowskaUniversity,20-031Lublin, Poland (Received 11 Nooember 1981) Abstract-Adsorption from benzene-ethanol binary solutions on preparations of modified activated carbon CWZ3 containing diierent amounts of oxygen bound to the surface was investigated. The modification was carried out in such a way that the porous structure of the carbons was but slightly affected. However, significant differences were observed in the total contents of oxygen as well as in the particular types of oxygen surface functional groups (especially of strongly acidic character). It was found that the higher the contents of those groups the stronger is the preference to adsorb the more polar component of the solution. The most significant role should be ascribed here to carboxylic and phenolic groups which can give hydrogen bonds with alcohol molecules. 1. INTRODUCTION

The results of studies on the adsorption of aliphatic alcohols from benzene solutions[l-141 lead us to the conclusion that in the presence of surface bound oxygen adsorption of the more polar component of the solution is preferred. Puri et al.[1&14] have shown that for activated carbons and carbon black such a situation occurs in the case of acidic oxygen surface complexes, which release carbon dioxide during heating under vacuum. In distinction, surface complexes releasing carbon oxide on decomposition prefer adsorption of benzene. In the case of adsorption from methanol-benzene solutions on activated carbons containing varying amounts of oxygen superficially adsorbed it has been found[ 141 that the cause of specific adsorption of methanol at the active sites containing oxygen is the formation of hydrogen bonds. The preferred adsorption of benzene in the presence of non-acidic oxygen complexes at the surface is the result of interaction of benzene with the positive charges of those complexes. It can be believed that the problems considered are more complicated due to surface inhomogeneities of the activated carbons. The authors of Refs. [l&14], where chemical modification of the surfaces of activated carbons and carbon black was applied, did not pay attention to the possibility of the resulting, sometimes quite considerable[l2,14] changes of the specific surface area of the adsorbents affecting the adsorption from binary solutions. In the quoted works the authors determined the kind and amount of oxygen functional groups bound with the carbon surface only on the basis of analysis of the products of their thermal decomposition (CO*, CO, H20). As a result in the present work such methods of modifying activated carbons have been applied that allow us to vary in a large extent the contents of surface bound oxygen without affecting significantly their porous

structure. In the case considered the only parameter affecting adsorption is practically the chemical character of the activated carbons surface. In the present work a different method as compared with the quoted works [ lO141 was applied for determining the oxygen functional groups which allows us to distinguish four kinds of them, as well as an independent method of determining the total oxygen content. 2. EXPERJMENTAL Activated carbon preparation Hardwood activated carbon CWZ3, manufactured by the Hajn6wka Wood Dry Distillation Works at Hajmjwka, was used. The carbon was demineralized by treatement with concentrated hydrofluoric and hydrochloric acids according to the method described by Korver[ IS]. In this way the content of mineral admixtures was reduced to 0.2%. After the demineralization of carbon was completed, the fraction of grain size <0.075 mm was separated. In order to obtain carbon preparations with different amounts of surface bound oxygen, the carbon was subjected to modification by two different methods. One part of the carbon was oxidized with nitric acid[l6], and the second part was subjected to heating in a stream of deoxidized nitrogen in a fluid bed at 1370K. In order to determine the porous structure of the modified and only demineralized carbons the benzene vapour adsorption and desorption isotherms at 293 K were determined using the vacuum sorption McBainBakr balance. The isotherms obtained are presented in Fig. 1, and the parameters calculated on their basis[l71, characterizing the porous structure, are summarized in Table 1. In order to compare the obtained carbon preparations with greater precision, the specific surface areas were additionally determined on them by low117

118

il. JANKOWSKA et al,

Fig. t. The adsorption f-) and desorption (---) isotherms of benzene vapcm on CWZ-3 activated carbon: oxidizedfU)$W-

treated(U)andheattreated(@Iat 293K.

temperature nitrogen adsorption.The results were 1290, 1240and 1255m*g-’ for derni~~~zed carbon, oxidized carbon and heat treated carbon, respectively. For all three carbon preparations studied the contents of the particular kinds of surface oxygen functional groups were subsequently determined by neutralization with bases of different strength[H]. Besides the total content of oxygen was determinedaccordingto [191,and the results are given in Table 2, Measwnmentof adsorptionfrom binary solutions The adsorbates used-ethyl alcohol and benzene(analytical grade, manufactured by POCh at Gliwice) were dried over silica gel. The tested adsorbents were initially heated for 4h1 at 390K in a nitrogen atmosphere, The ethanol-~nzene solntions of different composition were subsequently added to the samplesof the tested carbon preparations weighedout in glass vessels. The vessels were then tightly closed and left to stand for 24hr in a thermostat set at 293K with periodicalshaking. After equ~ibriumis achieved, the compositions of the

Fig. 2. ~omp~son of the adsorptionof e~~~l~I~~~~ne~~~ mixture on CWZ3 activated carbon: oxidized Q), untreated &3) and heat treated (@) at 298K.

Adsorption from benzene-ethanol binary solutions on activated carbons

119

Table 2. Surface oxygen functions groups of the various sampIes of CWZ-3 activated carbon Functional

group

concentration

Smpl0

dominaralized

-oxidized

lac tone

phenolic

carbonyl

0.0183

0.0554

c.147

0.769

2.05

0.341

0.256

0.481

0.377

6.10

treated

nil

0.0122

G.0365

0.110

0.35

and HNO3carbon

and heat

activated

carbon

solutions were determined by gas chromatography [4]. The isotherms of excess adsorption of ethyl alcohol from ethanol-benzene solutions on all three carbon preparations obtained on this basis are presented in Fig. 2. 3. RESULTS ANDDISCUSSION All

content (96)

activated

activated

demineralized

Total oxyiJen

carboxylic

carbon dcminwalfzad

(r0eq.g”)

the isotherms determined in the present work are S-shaped, the positive values of excess ethanol adsorption appearing only in the range of small concentrations in the solution. The isotherms presented in Fig. 2 show considerable similarity as regards their shape, and the differences are only quantitative. Since the porous structure of the tested carbons is almost identical (the curves in Fig. 1 almost coincide, and the values of the porous structure parameters given in Table 1 are close as are the specific surface areas SBET determined from low-temperature nitrogen adso~tion), it can be believed that differences in the isotherms of adsorption from binary solutions are primarily due to differences in the chemical structure of the surface of the tested carbon preparations. Differentiation of the chemical character of the carbon surfaces has been achieved by the application of different methods of modi~cation. Considerable differences have been obtained as regards the contents of the particular kinds of surface oxygen funtional groups and the total amount of oxygen on the surface. Comparing the values given for the particular carbon preparations in Table 2 with the parameters describing the corresponding isotherms of excess adsorption of ethanol from binary solutions we can observe the occurrence of a distinct relationship. The increase of the amount of oxygen bound to the surface is accompanied both by an increase of the maximum ethanol adsorption: 0.115, 0.25 and 0.673 mmoleg-’ for heated carbon, demineralized carbon and oxidized carbon, respectively, and by the extension of the ethanol concentration range corresponding to its positive adsorption, x, =O.l, 0.14 and 0.245, respectively. If benzene is the adsorbate, we have an opposite relationship, i.e. the decrease of the oxygen content at the surface is accompanied by increased adsorption of benzene: 1.03, 1.62 and 1.99 mmole g-‘, respectively. A similar trend of changes has been eartier observed by other investigators[9-141 for adsorption from alcohol-benzene solutions on activated carbons and carbon black. It seems useful to determine the proportion of the surface occupied by the bound oxygen in order to allow

more accurate analysis of the relationship observed in the present work. Such calculations can, however, be carried out only with a certain approximation. Since oxygen in the surface functional groups accounts for about 90% of the total oxygen occurring at the surface of activated carbon[20], in the calculations use can be made of the total oxygen content as given in Table 2. If we assume that the surface area per one oxygen atom is 7.8 A*[20], then for the heated carbon the surface occupied by oxygen bound to it is 10.3 rn’g-‘, for demineralized carbon it is 60.2m2g-“, and for oxidized carbon -179m2g-‘. As it is seen from those data, the carbon preparations tested differ largely as regards the surfaces occupied by oxygen. In the case of oxidized carbon the surface occupied by oxygen bound chemically is comparable with that of mesopores (Table 1). For demineralized and heat treated carbons those surfaces are 2.5 and I.5 times, respectively, smaller than those of the mesopores. An analogous comparison with the total surface area S,,, shows that the percentage of the surface area occupied by oxygen is many, many times smaller. On the basis of the considerations given above we can assume that the major part of the inner surface area of the tested carbon preparations is not covered with oxygen. Thus, preference for benzene adsorption is observed in a wide range of ethanol concentrations in the solution. For the heated carbon preparation (of minimum oxygen content) the maximum ethanol adsorption is many times smaller than that of benzene, viz. 0.115 mmole g-’ as compared with 1.99mmole-‘. The shape of the isotherm approximates best that of U, what is characteristic of adsorption from alcohol-benzene solutions on Graphon graphitized carbon black surfaces[9]. The increase of oxygen content produces on the other hand an approximation of the isotherm shape (Fig. 2) to that of type IV according to Schay and Nagy classification[21]. The same shape of the relationship was observed by Gasser and Kipling[9] for Spheron 6 carbon black of heterogeneous surface containing a large amount of bound oxygen. It can be believed that the differences in the courses of the isotherms obt~ned in the present work are surely due to changes in the chemical character of the carbon surfaces resulting from their modification. The increase of the total oxygen content on the surface is related to the increase of the number of functional groups, especially of strongly acidic character, e.g. carboxylic or

120

H. JANKQW~KA et al.

phenolic ones (Table Z), Those groups contribute to the preference of the more polar component of the solution, i.e. ethanol, in view of the possibility of formation of hydrogen bonds according to the model described earl~er~4J.In distinction, the part of the carbon surface area that is free from oxygen adsorbs benzene, as it has been suggested by Gasser and Kipling[9], due to the interactions of T-electrons. Benzene can also be adsorbed on the part of oxygen groups that show a weakly acidic or non-acidic character owing to the interactions of the benzene ring a-electrons with the positive charges of those groups. It seems that the mentioned interactions are of the greatest but not the only importance. Some other interactions related to the energetic non-homogeneity of the carbon surface as well as to the different defects, unsaturated sites and free radicals may also add in some extent to the effects. They may be, for instance, the cause of partial chemisorption of alcoholI221. The results obtained in the present work on an independent path fully confirm the observed earlier [ lO-141 significant effect of the form in which oxygen is bound with the carbon surface (surface functional groups) on adsorption of ethanol from its benzene solutions. Furthermore, it can be concluded that in tests involving a series of carbon preparations with surfaces modified by various methods it is indispensable when comparing isotherms of adsorption from binary solutions that possible changes in the specific surface area of the carbon preparations (or even better the changes of parameters of their porous structure) be taken into account.

REFERENCES

1. D. C. Jones and L. Outridge, J. Chem. Sot. 1574(1930). 2. J. I. Kipling and D. A. Tester, J. Chem. Sot. 4123(1952). 3, J. J. K$ling and D. A. Tester, Nature 167,612 (1951). 4. J. OS&. J, Goworek and R. Kusak. J. Cofloid interface Sci. 79, 308 (1981). 5. F. E. Bartell, G. H. Scheffler and C. K. Sloan, J. Am. Chem. sot. 53,2501(1931). 6. J. J. Kipling and D. B. Peakall, J. Chem. Sot. 4828(1956). 7. C. G. Gasser and J. J. Kipling, I. Phys. Chem. 64,710 (1960). 8. J. J. Kipling and D. B. Peakall, J. Chem. Sot. 4054 (1957). 9. C. G. Gasser and J. J, Kipling, Proc. 4th Carbon Cot& Buffalo, p. 57. Pergamon Press, Oxford (1960). 10. B. R. Puri, S. Kumar and N. K. Sandle, Indian J. Chem. 1, 418 (1%3). 11. B. R. Puri, Carbon 4,391(1966). 12. B. R. Puri, D. D. Singh and B. C. Kaistha, Carbon 10,481 (1972). 13. B. R. Puri, 0. P. Mahajan and B. C. Kaistha, Indian J. Chem. 12, 161(1974). 14. R. C. Bansal and T. L. Dhami, Carbon IS, 153(1977). 15. J. A. Korver, Chem. Weekblad 46,301(1950). 16. I. A. Kuzin and B. K. Strashko. Zh. Prikf. Khim. 39. 603 (1966). 17. M, M. Dubinin, Chemistry and Physics of Carbon (Edited by P. L. Walker Jr.), Vol. 2, pp. U-120. Marcel Dekker, New York (1%6). 18. H. P. Boehm, E. Diehl, W. Heck and R. Sappok, Angew. Chem. 76,742 (1964). 19. W. Kopycki, D. Fraisse nad J. Binkowski, Chem. Anal. 25, 829 (1980). 20. Th. van der Plas, Phys. Chem. Aspects of Adsorbenfs, p. 425. Academic Press, London (1970). 21. G. Shay and L. G. Nagy, Acta Chimica Acad. Sci. Hung. 50, 207 (1966). 22. M. Rozwadows~ and H. G&ten, Carbon 14, 169(1976).