6480
Langmuir 1996, 12, 6480-6486
Effect of Surface Chemistry on Sorption of Water and Methanol on Activated Carbons Teresa J. Bandosz,* Jacek Jagiełło, and James A. Schwarz Department of Chemical Engineering and Materials Science, Syracuse University, Syracuse, New York 13244-1190
Andrzej Krzyzanowski Department of Fuels and Energy, University of Mining and Metallurgy, 30-059 Krakow, Poland Received April 10, 1996. In Final Form: September 27, 1996X Two activated carbons from different sources were oxidized with nitric acid and hydrogen peroxide and the effects of this on their porous structure and sorption of water and methanol were studied. The microstructural properties were derived from analysis of N2 isotherms. The surface acidity was assessed using Boehm titration and potentiometric titration. Changes in the surface chemistry were also monitored using difference reflectance IR Fourier transform. From the potentiometric titration the pKa distributions were calculated and the surface density of acidic groups was evaluated. The structural and surface chemical results demonstrate that carbons derived from different sources can have wide variation in their oxidation susceptibility. To further evaluate the impact of the recorded changes due to oxidation in the carbons studied, sorption uptake of water and methanol was measured and correlated with the acidity of the carbon surfaces.
Introduction Activated carbons are widely used as adsorbents in gaseous, aqueous, and nonaqueous streams, as electrode materials in fuel cells, as catalyst supports, and as fibers for structural reinforcement or filters.1,2 A particularly desirable property of activated carbons in their use as an adsorbent is the high surface area, which is the result of their microporosity. Another important consideration is the surface chemistry of these materials.3-5 It is well-known that the carbon matrix consists of heteroatoms; the main heteroatom is oxygen. Of particular importance are the heteroatoms located on the carbon surface. Different functional groups can be derived from these chemical centers. These groups are analogous to typical organic compounds.3-8 The most common oxygen functionalities are carboxyl, lactonic, carbonyl, and phenolic. The presence of these species can result in polar sorbates, like water or alcohols, to interact with the carbon surface in a specific way. Oxidation is one of the processes which can effectively change the oxygen content of the carbon and thus the properties of a carbonaceous material, specifically its affinity toward polar adsorbates. The mechanism of oxidation reactions was investigated by several authors.3,4,8 Among the methods leading to the formation of surface oxygen complexes, two main * Author to whom correspondence should be addressed. Present address: Department of Chemistry, The City College of New York, New York, NY 10031. E-mail:
[email protected]. X Abstract published in Advance ACS Abstracts, December 1, 1996. (1) Bansal, R. C.; Donnet, J. B.; Stoeckli, F. Active Carbon: Marcel Dekker: New York, 1988. (2) Smisek, M.; Cerny, S. Active Carbon: Elsevier: Amsterdam, 1970. (3) Puri, B. R. In Chemistry and Physics of Carbon; Walker, P. J., Jr., Ed.; M. Dekker: New York, 1970; Vol. 6. (4) Boehm, H. P. In Advances in Catalysis; Academic Press: New York, 1966; Vol. 16. (5) Boehm, H. P. Carbon 1994, 32, 759. (6) Bandosz, T. J.; Jagiełło, J.; Contescu, C.; Schwarz, J. A. Carbon 1993, 31, 1193. (7) Jagiełło, J.; Bandosz, T. J.; Schwarz, J. A. Carbon 1994, 32, 1026. (8) Leon y Leon, C. A.; Radovic, L. R. In Chemistry and Physics of Carbon; Thrower, P. A., Ed.; M. Dekker: New York, 1992; Vol. 24.
S0743-7463(96)00340-X CCC: $12.00
categories can be distinguished: utilizing oxidizing gases such as oxygen, ozone, nitrous oxide, and nitric oxide and reactions in oxidizing solutions such as nitric acid, alkaline permanganate, hydrogen peroxide, acidic permanganate and acidic dichromate. The degree of surface oxidation, the nature of the resulting acidic centers, and the adsorption properties of oxidized carbon depend on several conditions which include chemical reagent used as oxidant, temperature of reaction, time of reaction, type of carbon, its ash content and microstructure.3 The objective of this paper is 2-fold: to continue our studies of the oxidation susceptibility of activated carbons from different sources (wood and peat moss);9,10 evaluate the performance of as received and oxidized carbons as sorbents for water and methanol. Boehm and potentiometric titration are applied to evaluate the nature of surface oxygen groups and the variations in their distributions imposed by different methods of oxidation. The surface structural and chemical characteristics are correlated with the sorption uptake of water and methanol as a function of the surface coverage of these adsorbates. Experimental Section Materials. Two activated carbons from different sources (Westvaco (B4×14)-W (wood) and Norit (Sorbonorit 2-A5998)-N (peat moss))9,10 were oxidized with 15 N (73%) HNO3 solution at 351 K for 2 h with continuous stirring. A vigorous oxidation reaction was indicated by an exothermic effect and release of brown fumes (nitric oxides). Samples were also treated with 30% H2O2 at 323 K for 2 h. After treatment, the carbons were washed with distilled water to zero acid removal and oven-dried at 383 K. Samples oxidized with nitric acid are designated as N1 and W1 and those oxidized with hydrogen peroxide as N2 and W2, respectively. Before experiments the initial carbon samples were washed out with distilled water to constant pH. Boehm Titration. The oxygenated surface groups were determined according to the method of Boehm.4 One gram of carbon sample was placed in 25 mL of the following 0.05 M solutions: sodium hydroxide, sodium carbonate, sodium bicar(9) Bandosz, T. J.; Jagiełło, J.; Schwarz, J. A. Anal. Chem. 1992, 64, 891. (10) Jagiełło, J.; Bandosz, T. J.; Schwarz, J. A. Carbon 1992, 30, 63.
© 1996 American Chemical Society
Sorption on Activated Carbons bonate, and hydrochloric acid.9,10 The vials were sealed and shaken for 24 h and then 5 mL of each filtrate was pipetted and the excess of base and acid was titrated with HCl and NaOH, respectively. The numbers of acidic sites of various types were calculated under the assumption that NaOH neutralizes carboxyl, phenolic, and lactonic groups, Na2CO3 neutralizes carboxyl and lactonic, and NaHCO3 neutralizes only carboxyl groups. The number of surface basic sites was calculated from the amount of hydrochloric acid which reacted with the carbon. Potentiometric Titration. Potentiometric titration measurements were performed with a 665 Dosimat (Brinkmann) combined with a Accumet pH meter model 50 equipped with a combination glass electrode (Corning). The microburet is able to dose the titrant in increments of 0.001 mL to the suspension of carbons. Samples of the carbons studied of 0.100-0.500 g in 50 mL of 0.01 N NaNO3 were placed in a container thermostated at 298 K and equilibrated for several hours with the electrolyte solution. To eliminate the influence of atmospheric CO2, the suspension was continuously saturated with N2. The carbon suspension was stirred throughout the measurements. Solutions of NaOH and HNO3 (0.1 N) were used as titrants. The titration curves, pH(V), consisting of about 400 experimental points were transformed into proton binding isotherms, Q(pH), by using the proton balance equation.6,11 The experiments were done in the pH range of 3-10.6,7 Each sample was titrated by both acid and base starting from the initial pH of the suspension.12 DRIFT. IR spectra were obtained on a Nicolet Impact 400 FT-IR spectrometer at a resolution of 4 cm-1 using a DTGS detector and SpectraTech Collector diffuse reflectance unit equipped with blocker. Sorption of Nitrogen. Nitrogen isotherms were measured on ASAP (Micromeritics) at 77 K. Before the experiment the samples were outgassed at 393 K under a vacuum of 10-3 mmHg. The isotherms were used to calculate the specific surface areas, micropore volumes, and pore size distributions (PSDs). Sorption of Water and Methyl Alcohol. Sorption isotherms of water and methyl alcohol were obtained using a home made microburet type of adsorption apparatus.13 Equilibrium adsorption amounts were measured at 278 K. Samples, before the measurements, were outgassed at 398 K to a constant vacuum of 10-4 mmHg.
Results and Discussion It is well-known that oxidation of carbons with strong oxidizers may result in the changes of their microstructure.3,10,14 In that regard, under similar thermal conditions, nitric acid is a stronger oxidant than hydrogen peroxide. Nitrogen isotherms obtained on carbons treated with H2O2 and HNO3 are shown in Figure 1. In the case of the Norit carbon, we observe a gradual decrease in the total sorption uptake with increasing severity of oxidation, but the shapes of the isotherms are similar. The changes recorded indicate that some structural changes occurred, but the microporous nature of the material is preserved even after aggressive oxidation (the N1 sample). It is interesting to note that the isotherm for the Westvaco carbon oxidized with hydrogen peroxide (W2) becomes more Langmuir-like when compared to the as-received sample, whereas oxidation with HNO3 completely destroys the porous nature of the material; the total sorption uptake is almost negligible in the case of the W1 sample. Structural parameters calculated from nitrogen sorption are collected in Table 1. Where results can be compared at comparable oxidation conditions, H2O2 oxidation results in about a 50% decrease in the surface area for Westvaco while for the Norit carbon only about a 10% decrease in the surface area is observed. The same relative changes occur in micropore volumes. Collectively these facts (11) Contescu, C.; Jagiełło, J.; Schwarz, J. A. Langmuir 1993, 9, 1754. (12) Bandosz, T. J.; Jagiełło, J.; Schwarz, J. A. J. Phys. Chem. 1995, 99, 13522. (13) Lason, M.; Zyla, M. Chem. Anal. 1963, 8, 282. (14) Bandosz, T. J.; Jagiełło, J.; Schwarz, J. A. Langmuir 1993, 9, 2528.
Langmuir, Vol. 12, No. 26, 1996 6481
Figure 1. Nitrogen adsorption isotherms at 77 K (solid symbols, adsorption branch; open symbols, desorption branch). Table 1. Structural Parameters Calculated from Nitrogen Adsorption Isotherms sample
SBET (m2/g)
Vmic (cm3/g)
Vmic* (cm3/g)
W W2 N N1 N2
1500 860 970 625 860
0.642 0.378 0.490 0.291 0.426
0.160 0.134 0.285 0.142 0.235
suggest that, in the case of Westvaco carbon, small pores collapsed as a result of oxidation by either oxidant while the Norit carbon structural properties were not affected as greatly. To study the changes in the pore structure in more detail, the pore size distributions (PSDs) from nitrogen adsorption isotherms were calculated using DFT15 and the classicalBJH method.16 Results are presented in Figures 2 and 3. In the case of Westvaco similar distributions are obtained for the W and W2 carbons; however, for the W2 sample the pore volume is smaller. The PSDs of Norit samples indicate that mainly the small pores are affected by acid treatment. It is beyond the scope of this paper to compare the distributions obtained by both methods. The distribution curves in Figures 2 and 3 show some differences related to the different approaches used; however, the overall effect is a decrease in pore volumes after oxidation which occurs for both carbons and is consistent with the changes in their adsorption isotherms. Our experimental range allows us to apply DFT only to pores larger than 11.8 Å. However, the volume in pores smaller than 11.8 Å can be also evaluated. We refer to this value as Vmic*. The values of the volumes in pores (15) Olivier, J. P.; Conklin, W. B. Presented at 7th International Conference on Surface and Colloid Science, Compiegne, France, 1991. (16) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc. 1951, 73, 373.
6482 Langmuir, Vol. 12, No. 26, 1996
Figure 2. PSDs calculated from DFT.
smaller than 11.8 Å, Vmic*, collected in Table 1 clearly show that the volume of very small pores decreases twice for the N1 carbon compared to N which supports our hypothesis about destruction of small pores due to severe oxidation conditions. The results obtained also show a systematic decrease in Vmic* for both carbons with increasing severity of oxidation. Structural changes revealed by changes in sorption of nitrogen occur together with changes in surface chemistry. The Boehm method provides an indication of changes in the inventory of surface species4 as a function of oxidation conditions. The results are collected in Table 2. Significant differences between Norit and Westvaco are observed. The Westvaco samples as received and after mild oxidation with hydrogen peroxide are characterized by a more acidic surface than the corresponding Norit samples. Strong oxidation with nitric acid significantly increases the number of acidic groups for both carbons; however, the W1 sample has a greater abundancy of strongly acidic carboxylic species. When carbons are oxidized with hydrogen peroxide, different distributions are obtained.3-9,17 The results obtained indicate that Norit is much more resistant to oxidation with H2O2 than Westvaco and only a small increase in the number of acidic groups is observed. Additional information about the distribution of oxygen groups on the surface of our carbons can be obtained from potentiometric titration experiments6,7,17 assuming that the system under study consists of acidic sites characterized by their acidity constants, Ka. It is also assumed that the population of sites can be described by the a continuous pKa distribution f(pKa). The experimental data can be transformed into a proton binding isotherm, Q, (17) Jagiełło, J.; Bandosz, T. J.; Putyera, K.; Schwarz, J. A. J. Colloid Interface Sci., 1995, 172, 341.
Bandosz et al.
Figure 3. Mesopore size distributions calculated from the BJH method.
representing the total amount of protonated sites which is related to the pKa distribution by the following integral equation
Q(pH) )
∫-∞∞q(pH, pKa) f(pKa) d(pKa)
(1)
The solution of the equation is obtained using the numerical procedure SAIEUS,18 which applies regularization combined with non-negativity constraints. The choice of the degree of regularization/smoothing is based on the analysis of a measure of the effective bias introduced by the regularization and a measure of uncertainty of the solution. SAIEUS was tested using simulated data and experimental titration data of organic standards and it was demonstrated that this method can completely resolve peaks which are less than 1 pKa unit apart.17 Figure 4 shows the proton binding isotherms obtained for the carbons studied. The solid lines represent the fit of the experimental points to eq 2. The changing shape of the isotherms with oxidation treatment indicates the existence of different degrees of surface heterogeneity, which is reflected in the calculated pKa distributions (Figure 5). The surfaces of Westvaco samples are characterized by the presence of four well-resolved peaks with various numbers of species depending upon the method of oxidation (Table 3). These groups can be classified as strong acids (carboxylic) at pKa below 8 and weak acids (phenols) at pKa over 8.6,7,19,20 It is noteworthy that oxidation with nitric acid results in a high contribution to the total acidity (18) Jagiełło, J. Langmuir 1994, 10, 2778. (19) Kortum, G.; Vogel, W.; Andrussow, K. Dissociation Constants of Organic Acids in Aqueous Solutions: Butterworth; London, 1961. (20) Perdue, E. M.; Reuter, J. H.; Parrish, R. S. Geochim. Cosmochim. Acta 1984, 48, 1257.
Sorption on Activated Carbons
Langmuir, Vol. 12, No. 26, 1996 6483 Table 2. Boehm Titration Results
sample
basic groups (mequiv/g)
acidic groups (mequiv/g)
carboxylic (mequiv/g)
lactonic (mequiv/g)
phenolic (mequiv/g)
W W1 W2 N N1 N2
0.14
0.75 4.40 2.36 0.24 2.05 0.25
0.27 3.50 1.31 0.02 1.02 0.03
0.14 0.12 0.21 0.07 0.15 0.12
0.33 0.78 0.84 0.14 0.87 0.09
0.68 0.10 0.75
Figure 4. Proton consumption isotherms. Solid lines indicate the goodness of fit to eq 2.
by carboxylic groups whereas oxidation with hydrogen peroxide results in a proportionally larger number of groups at pKa > 8. This is consistent with the results of the Boehm titration reported above (Table 2) and the results of other studies.3-5,8 Oxidation of the Norit carbon leads to changes not only in the number but also in the nature of surface species. For the N sample two peaks with the indication of peak at pKa below 3 are obtained within the experimental window. Mild oxidation with hydrogen peroxide only slightly changes the distribution. On the other hand, aggressive nitric acid oxidation results in a creation of new species at pKa ) 6.84 and a larger number of phenolic groups20 (Table 2). Table 3 collects the results obtained from our analysis. A peak position is given as a value of its first-order moment; the number of species is given in parentheses as a value of the integral of the peak. In all cases the root mean square error of fitting the data to eq 2 was relatively small. It is seen that different methods of oxidation of different carbons result in different distributions of acidic groups but the peak positions are similar, which suggests that similar surface organic acid/base functionalities are created in each case. Potentiometric titration data are limited to the experimental window between 3 and 10; thus it is difficult to directly compare the results obtained with those from the Boehm titration; the latter covers a broader spectrum of pKa values.4,5 An attempt is made here to compare the number of groups obtained from integration of peaks at
Figure 5. pKa distributions for carbons with different degrees of surface oxidation.
pKa < 8 with the number of carboxylic groups obtained from the Boehm method. The results are presented in Figure 6. It is seen that in the case of carbons with a negligible amount of basic groups or lack of these groups (W, W1, W2), the agreement is relatively good. When carbons with a significant number of basic groups, such as found on N or N2, are considered, the number of groups obtained from potentiometric titration is much higher. Also in the case of Norit oxidized with nitric acid the agreement is not as good as in the case of Westvaco carbons. IR spectra of Westvaco samples presented in Figure 7 support our earlier observation about the structural changes and the disappearance of the features of typical activated carbon in the case of the W1 sample. A band at 1750 cm-1 due to CdO stretching vibrations together with the C-O stretching vibrations band at 1260 cm-1 indicate the presence of carboxyl structures.21 The intensity of the former bands increases more significantly than the latter with the more aggressive oxidation. In the case of the sample oxidized with nitric acid the absorption bands of nitro groups at 1550 and 1350 cm-1 (21) Zawadzki J. In Chemistry and Physics of Carbon; Thrower, P. A., Ed.; M. Dekker: New York, 1989; Vol. 21.
6484 Langmuir, Vol. 12, No. 26, 1996
Bandosz et al.
Table 3. Peak Positions and the Number of Groups (in Parentheses; [mmol/g]) for Carbons Studied sample W W1 W2 N N1 N2
pK 2-3
pK 3-4
pK 4-5
3.94 (0.076)
4.98 (0.055) 4.43 (1.282) 4.43 (0.362)
a a a 3.44 (0.811) 3.31 (0.451)
4.84 (0.270)
pK 5-6 5.67 (1.288) 5.60 (0.361) 5.18 (0.423) 5.22 (0.711) b
pK 6-7
pK 7-8
0.06 (0.075)
7.58 (0.140) 7.04 (0.725) b 7.70 (0.287) b 7.42 (0.209)
6.82 (0.362) 6.87 (0.612)
pK 8-9
pK 9-11
8.42 (0.647) 8.82 (0.415)
a a
8.57 (0.618)
a a
All (mmol/g)
rms error
0.346 3.942 1.500 0.710 2.752 0.930
0.002 0.008 0.002 0.006 0.011 0.009
a Uncertainty of the peak position and the number of groups due to the limit of the experimental window. b Possible shift of the peak position.
Figure 6. Comparison of the number of carboxylic groups obtained from Boehm titration and potentiometric titration using the SAIEUS procedure.
Figure 8. IR spectra for Norit samples.
cm-1 due to C‚‚O structures. On the other hand, when the IR spectra of Norit samples are considered, a different surface inventory is revealed (Figure 8). The initial N sample and oxidized with hydrogen peroxide (N2) show an almost featureless spectrum. The strong oxidation with nitric acid results in bands at 1770, 1600, and 1300 cm-1. They are due to vibrations of CdO, C‚‚O, and C-O groups, respectively.21 The presence of bands at 1770 and 1300 cm-1, as observed in the case of the Westvaco samples, indicates the surface is rich in carboxylic groups. The changes in surface chemical groups together with microstructural changes can influence the sorption of polar molecules such as water and methanol.22-30 The surface
Figure 7. IR spectra for Westvaco samples.
are observed.22 For all Westvaco samples we also distinguish the presence of an intense band at about 1600 (22) Maddox, M.; Ulberg, D.; Gubbins, K. E. Fluid Phase Equlibria 1995, 104, 145.
(23) Ulberg, D. E.; Gubbins, K. E. Mol. Phys. 1995, 84, 1139. (24) Buczek, B.; Grzybek, T.; Bernasik, A. In Fundamentals of Adsorption V; M. D. LeVan, Ed.; Kluwer: Boston, MA, 1996; p 109. (25) Rodriguez-Reinoso, F.; Molina-Sabio, M.; Munecas, M. A. J. Phys. Chem. 1992, 96, 2707. (26) Dubinin, M. M.; Serpinsky, V. V. Carbon 1981, 19, 402. (27) Youssef, A. M.; Ghazy, T. M.; El-Nabarawy, Th. Carbon 1982, 20, 113. (28) Evans, M. J. B. Carbon 1987, 25, 81. (29) D’Arcy, R. L.; Watt, I. C. J. Chem. Soc., Faraday Trans. 1970, 66, 1236. (30) Dubinin, M. M. In Chemistry and Physics of Carbon; Walker, P. L., Ed.; M. Dekker: New York, 1966; Vol. 2.
Sorption on Activated Carbons
Langmuir, Vol. 12, No. 26, 1996 6485 Table 5. Numbers of Adsorbed Molecules (molecule/nm2) of Water and Methanol at Chosen p/p0 H2O
Figure 9. Water adsorption at 298 K.
Figure 10. Methanol adsorption at 298 K. Table 4. Surface Density of Oxygen Groups (group/nm2) sample
Boehm titration
potentiometric titration
W W2 N N1 N2
0.36 1.65 0.57 2.07 0.70
0.14 1.05 0.44 2.65 0.65
density of species determined by analysis of Boehm titration and potentiometric titration data is given in Table 4. We report values now on an aerial basis to account for structural changes leading to changes in surface area during oxidation. Similar values in the range between 0.14 and 2.65 groups/nm2 are obtained from both methods.24,25 The surface density of groups for the W1 sample is not reported since this carbon undergoes significant changes in its carbonaceous structure,14 and after strong oxidation it rather remains a mixture of high molecular weight organic compounds with negligible surface area when compared to a typical activated carbon. The isotherms of water and methanol adsorption are collected in Figures 9 and 10. These molecules, due to the presence of a strong dipole moment, are adsorbed in a specific way on a carbon surface rich in oxygen groups.25,26 Surface oxides of carbons are considered as primary, highenergy adsorption sites.26 These primary sites act as secondary sites for further adsorption via hydrogen bonding. The methods to estimate the number of primary sites on a carbon surface have been proposed by Dubinin and Serpinsky (DS)26 and Evans28 (equation derived by DArcy and Watt29). The Dubinin and Serpinsky method was used to interpret those isotherms corresponding to type V isotherms based on Brunauer-Emmett-Teller classification.26 When this approach was applied to an isotherm with an initial Langmuir type section, it failed to properly fit the initial adsorption points.24,28 Due to the observed poor performance of the DS method, Evans proposed the use of a combination of Langmuir isotherms describing adsorption on high-energy sites (the D’Arcy
sample
A1 (p/po ) 0.3)
A2 (p/po ) 0.9)
CH3OH A1 (p/po ) 0.3)
RB
RPT
N N2 N1 W W2
1.08 1.32 6.20 1.31 4.21
14.4 16.7 14.9 14.8 16.5
4.96 6.20 6.82 4.20 5.89
1.9 1.9 3.0 3.6 2.5
2.4 2.0 2.3 9.3 4.0
and Watt approach29). Using this method he obtained a much better fit to the experimental points of non-type V isotherms. To describe our experimental results, we tried to use both the DS and the Evans approaches; however, they failed for the majority of cases. A relatively good fit was obtained using the DS equation only for the N1 and N2 samples when type V isotherms were obtained. Water adsorption isotherms obtained in the cases of samples with a high density of oxygen groups (more than 1 group/ nm2) are hybrids of different isotherms with high uptake capacity at low relative pressure. According to RodriguesReinoso and co-workers,25 the region of low relative pressure, below the “plateau” of the water isotherm, is controlled by the chemical nature of the carbon surface, whereas starting from a relative pressure above the point of the plateau beginning, the microporosity of the carbon is a critical factor which determines the total sorption uptake. Since for the majority of our water adsorption isotherms it was very difficult to clearly distinguish the beginning of a plateau, we decided to use a different approach. It is assumed that the amount adsorbed (A1 (molecules/ nm2)) at p/p0 ) 0.3 is related to direct interactions of water molecules with surface oxygen (primary centers) whereas the sorption uptake at p/p0 ) 0.9 (A2) represents the combined effects of surface chemistry and porosity (secondary centers). The A values referred to are collected in Table 5. The isotherm obtained for the W1 carbon is not analyzed in these terms due to the different nature of the solid and uncertainty in its surface area. Analysis of A1 values indicates an increase in the number of adsorbed water molecules with an increase in the degree of carbon surface oxidation. In order to further check the consistency of our approach, the ratios of A1 values to surface densities of oxygen groups calculated from Boehm titration (RB) and potentiometric titration (RPT) were calculated and the numbers obtained are collected in Table 5. In all cases the values are greater than 1. In proposing the use of the above approach, it is not the intention of the authors to establish a method to evaluate the exact number of primary adsorption centers, and a choice of p/p0 ) 0.3 was only arbitrary. The R values suggest that, at the chosen pressure, the adsorption on secondary sites via hydrogen bonding has already started. It follows that, if our densities of oxygen groups are not underestimated, adsorption on secondary sites starts at a lower pressure than p/p0 ) 0.5, which was suggested previously.30 Comparison of the numbers of molecules adsorbed at p/p0 ) 0.9 (A2) does not lead to any apparent relationship between porosity or surface area. In fact, for all analyzed carbons the number of adsorbed molecules of water per square nanometer of surface was similar. In is noteworthy that two carbons of different origin, N2 and W2, having the same surface area (860 m2/g), also have almost the same number of water molecules, A2, adsorbed per unit area. On the other hand, they differ significantly in their
6486 Langmuir, Vol. 12, No. 26, 1996
A1 values, which is governed by their different values for the surface density of oxygen groups. The methanol adsorption isotherms obtained for N and N2 carbons (low density of oxygen groups) are representative of type I, which is characteristic for microporous materials. For all carbons the amount adsorbed at low relative pressure is much higher compared to water (Table 5). This is attributed to a larger carbon-methanol interaction through the oxygen surface groups.25 At low pressure an increase in their density, as in the case of water adsorption, results in an increase in the number of adsorbed methanol molecules; however, in this case the relationship is not well-defined. This could be due to the significant influence of micropores which, in the case of microporous carbon, Norit, enhance the adsorption potential. Similar shapes of the isotherms for pairs of N/N2 and W/W2 samples indicate that similar changes in their structure had occurred due to oxidation, which is in agreement with the results obtained from nitrogen adsorption. Conclusions Results presented demonstrate the effect of different degrees of carbon surface oxidation on their structural and chemical properties. Carbons from different origins
Bandosz et al.
have different oxidation susceptibilities; strong oxidation in some cases (Westvaco) leads to the destruction of typical activated carbon features and changes in its chemical and physical nature. The application of two methods to assess surface functionalities resulted in similar estimates of the surface density of oxygen groups. The application of the method of Dubinin-Serpinsky to the hybrid-shaped water adsorption isotherms did not result in a good fit, and thus it failed to obtain the exact number of primary adsorption centers as proposed. Comparison of the amount adsorbed at an arbitrarily chosen p/p0 ) 0.3 as the limiting pressure below which sorption of water is governed by the density of oxygen groups with the surface density of these groups suggests that adsorption on secondary sites started at lower relative pressure. A direct relationship between porosity and total uptake of water was not established. However, in the case of methanol adsorption the presence of small pores, which could enhance the adsorption potential at low relative pressure, and the presence of surface oxygen groups lead to a significant increase in the amount adsorbed at p/p0 ) 0.3 when compared to water. LA960340R
