Influence of chemisorbed oxygen in adsorption onto carbon from aqueous solution

Influence of Chemisorbed Oxygen in Adsorption onto Carbon from Aqueous Solution ROBERT W. COUGHLIN, FOUAD S. EZRA, 1 AND RICARDO N. TAN Department of Chemical Engineering, Whitaker Laboratory, Lehigh University, Bethlehem, Pennsylvania 18015 Received April 16, 1968 I s o t h e r m s are p r e s e n t e d for the adsorption of phenol, nitrobenzene, and sodium benzene sulfonate from aqueous solution on several different carbons, including a channel black and a d s o r b e n t carbons. T h e c o n t e n t of chemisorbed oxygen on these carbons was modified b y oxidation and reduction. The adsorption m e a s u r e m e n t s indicate t h a t chemisorbed oxygen s u b s t a n t i a l l y reduces the c a p a c i t y of the carbons for the adsorbates studied, and this b e h a v i o r appears to be reversible w i t h respect to t h e i n t r o d u c t i o n or removal of chemisorbed oxygen. D i l u t e phenol solutions display greater s e n s i t i v i t y to chemisorbed oxygen on the carbon t h a n do more c o n c e n t r a t e d solutions of phenol. I n t h e adsorption of sodium benzenesulfonate, the presence of chemisorbed oxygen does not appear to interfere w i t h t h e action of aqueous calcium ions in e n h a n c i n g the a m o u n t adsorbed.

INTRODUCTION Hydrophobic carbon surfaces have enjoyed widespread popularity as substrates for studies of adsorption from aqueous solution. Graphite and the graphitized carbon black, Graphon (Cabot Corporation), offer good examples of such surfaces which present high specificity for many organic solutes. The latter material has become a favorite for adsorption studies because it possesses a rather uniform surface with large specific area of about 100 m2/gm, made up mainly of basal graphitic planes. No attempt will be made here to catalogue these researches, but it is appropriate to refer to the use of Graphon in studies of adsorption of surfactants from aqueous solution (1) as well as its use in work concerned with heats of adsorption (2). Although a vast technical literature exists for adsorption onto less ideal, but more common, carbon surfaces (e.g., charcoals, carbon blocks, and active carbons) the heterogeneous nature of the surfaces of these 1 P r e s e n t address: D e p a r t m e n t of Chemical Engineering, U n i v e r s i t y of Wisconsin, Madison, Wisconsin.

carbons has often interfered with the drawing of fundamental inferences from the experimental results. Most commercial, adsorbent-grade carbons possess such heterogeneous surfaces and the heterogeneities stem largely from ehemisorbed oxygen in the form of acidic functional groups, although oxygen ehemisorbed on carbon can also exhibit basic properties. Graham (3) has studied the adsorptive capacity of Graphon and six different commercial active carbons and found that acidic chemisorbed oxygen can play a very significant role in adsorption from aqueous solution. This work revealed that acidic oxygen surface groups tend to reduce the capacity of carbon surfaces for adsorption of metanil yellow from aqueous solution whereas the adsorption of methylene blue did not appear to be materially affected. Graham attributed this behavior to a repulsive interaction between the anionic metanil yellow and the oxygen sites on the surface. He suggested that this kind of interaction might be expected for anionic adsorbates in general. These results are important considerations

Journal of Colloidand Interface Science,¥oi. 28, No. 3/4, November-December 1968 386

CHEMISORBED OXYGEN IN ADS01~PTION ONTO CARBON

381

phenolic hydroxyl group, and (IV) a carbonyl group. These acidic functional groups can be identified by their reaction (or failure to react) with bases of different strength. Thus, Group (I) is neutralized by each of the bases NaHCO3, NaCO3, NaOH, and NaOC2Hs, Group (II) is neutralized by Na2CO3 or stronger bases but not by OXYGEN BOUND AT T H E SURFACE NaHCOa, etc. Accordingly, simple titration OF CARBON ADSORBENTS with different bases serves to identify the The "free valences" at the edges of the acidic surface oxides present on a given graphitic layer planes of mierocrystalline sample of carbon. It is possible to produce these acidic funccarbon are very reactive and form compounds with any suitable foreign atoms pres- tional groups by oxidation in air or pure ent. It follows that functional groups or oxygen or by mixing the carbon sample with surface compounds can be expected almost aqueous solutions of oxidizing agents like exclusively at the layer edges; foreign atoms NaOC1, KMnO4, or (NH4)2S2Os. It is also or molecules can be only weakly adsorbed on possible to partially remove the aeidie functhe basal faces by means of the graphitic 7r- tional groups by reduction or vacuum outelectron system, except where they are gassing at elevated temperatures. In addibound at lattice defects. Most important and tion, these groups can be made to react in best known among the surface compounds other ways such as esterification, formation of carbon are those with oxygen and sulfur, of acid chlorides, and aeetylation. although other elements such as chlorine MATERIALS AND METHODS and hydrogen can also combine with eleThree different mierocrystalline carbons mental carbon. Of these compounds the sur- were employed in this work: (1) an active face oxides of carbon have received the most carbon "Columbia Carbon LC325" (Union study, and it is the role of these oxides in Carbide Corporation), (2) an active carbon adsorption that is the principal topic of this "Darco $51" (Atlas Chemical Industries, report. In particular, the concern here is Inc.), and (3) a pelletized channel blaek, with the acidic surface oxides of carbon "Black Pearls 607" (P607) (Cabot Corporawhich are formed under the more usual con- tion). Initial preparation consisted of equilditions of treatment and manufacture of ibrating for 24 hours in 0.1 M HC1 in the microerystalline carbon products like active case of the active carbons (to remove alkacarbon. Basic surface oxides also occur, but line impurities) or mere crushing in the ease less frequently (they are formed only when of the P607. All carbon samples were thorthoroughly out-gassed carbon comes into oughly washed with doubly deionized water contact with oxygen after cooling to low and dried overnight at 110°C before use in temperatures), and their nature and strut- adsorption experiments. ture have not yet been elucidated very The surfaces of the carbons were modified thoroughly. by wet oxidation and reduction. Oxidation Of the many techniques for characterizing was earried out by stirring the carbon samthe acidic surface oxides of carbon we have ples in (NH4)2S2Os solution at room tememployed that used extensively by Boehm, perature and reduction followed by mixing DieM, Heek, and Sappok (4). In an extension the carbon and zinc amalgam, covering with of earlier work (5, 6), these investigators concentrated HC1, and Mlowing to stand for used typical identification reactions of or- one week. Subsequent to oxidation or reducganie chemistry to characterize oxygen tion the carbon samples were washed and chemisorbed on carbon as comprising four dried as described above. Additional experidifferent types of acidic surface groups: (I) mental details are given in Tables I and V. a strongly acidic carboxyl group, (II) a more The acid groups on the carbon surfaces weakly acidic carboxyl group, (III) a were determined by adding excess standard with regard to the use of carbon to adsorb organic molecules from aqueous solutions. The experimental results presented below are a contribution to knowledge in this area. However, it is appropriate first to discuss at least one model for the acidic, oxygen functional groups bound on carbon surfaces.

Journal of Colloid and Interface Science, Vol. 28, No. 3/4, N o v e m b e r - D e c e m b e r 1968

388

COUGHLIN, EZRA, AND TAN TABLE I SURFACE ACIDITIES AND SURFACE AREAS OF CARBON ABSORBENT SAMPLES Base consumption b (meq/gm)

B.E.T. (N2) specific surface areas (Sq.

Surface groups (meq/gm)

Producta

1)607 1)607 O 1)607 1% 1:)607 OR P607 OROROR LC325 LC325 O LC325 OR

NaOEt

NaOH

2.40 3.06 1.86 2.42 1.56 0.38 4.15 1.87

1.88 2.25 1.31 1.94 1.26 0.12 4.22 1.92

Na2CO3 NaI-ICO~

1.34 1.69 0.97 1.42 0.90 0.11 3.10 2.04

0.85 0.97 0.48 0.88 0.80 0.11 2.20 0.84

I

II

III

IV

meters~gram)

0.85 0.97 0.48 0.88 0.80 0.11 2.20 0.84

0.49 0.72 0.49 0.54 0.10 0 0.90 1.20

0.54 0.56 0.34 0.52 0.36 0.01 1.12 0

0.52 0.81 0.55 0.48 0.30 0.26 0 0

665 668 663 513 562 1200 556 400

a 1) = Black Pearls Carbon Black; LC = Columbia Carbon (Activated); O = oxidized; R = reduced; OR = oxidation followed by reduction. b Base Consumption was measured after equilibration periods of about 20 hours. Oxidation conditions were: 0.01M (NH~)~S208 solution, 15-20 hours for P607 and 0.1M (NHa)~S208 solution, two weeks for LC325. base and back-titrating with standard HC1 after equilibration. I n the case of NaOCH2CH3 and N a O H , phenolphthMein was used as indicator; in the case of Na2CO and NaHCO3 the indicator was methyl orange and heating was employed to drive off the CO2 formed upon base neutralization. Adsorption and desorption were studied b y equilibrating the carbon samples at 30°C for about one week with aqueous solutions of phenol and of nitrobenzene or for about 3 days with aqueous solutions of sodium benzene sulfonate. Concentrations were measured b y ultraviolet spectrophotometry at 270.0 m~ in the case of phenol, at 267.8 mu in the case of nitrobenzene, and 216.0 m~ in the case of sodium benzene sulfonate. B e c k m a n D K - 2 and Bausch & Lomb Spectronic-505 double-beam spectrophotometers were employed using quartz cells. RESULTS T h e quantities of base consumed and the concentrations of functional groups (comp u t e d from quantity of base consumed) are presented in Tables I and V for the various carbon samples. B.E.T. surface areas (N2) are also given for the samples in Table I but were not measured for the samples of Table V. P r e t r e a t m e n t of these samples is designated b y an appended " O " for oxidation or " R " for reduction; "O R " means oxidation followed b y reduction, etc. The treated

"Black Pearls 607" and "LC325" samples of Table V were different from the corresponding samples of Table I and are so designated b y the " A " in the labels P607OA, LC3250A, etc. U p o n equilibration of the carbon samples reproducibility of base consumption was within about 3 % to 4%. Prolonging the equilibration time from about 20 hours to 3 days caused some increase in total acidity with the surface concentrations of some acid groups apparently larger and some apparently lower. During equilibration, samples were not continuously agitated but were shaken occasionally. Continuous agitation caused a base consumption a few per cent larger t h a n in the nonagitated cases. Adsorption isotherms were obtained by equilibrating the samp]es for one week before analysis. When equilibration was carried out for two weeks the isotherms were 8 % to 10 % higher t h a n those obtained for one-week equilibration periods, except in the case of phenol on P607 (untreated), for which there was no detectable difference. In the case of sodium benzenesulfonate adsorption 3-day equilibration periods were employed and not varied. Figure 1 presents the adsorption isotherms for phenol on the various P607 samples of Table I, and Fig. 2 shows those for phenol on the LC325 samples of Table I. Figure 3 gives adsorption isotherms for nitrobenzene on two

Journal of Colloid and Interface Science, ¥oi. 28, No. 3/4, November-December 1968

CHEMISORBED OXYGEN IN ADSORPTION ONTO CARBON

•

300

389

.

.

~ 200

/ ,4,

~o I00

III//'" W V

. .

i

oR I

t00 200 Conc~rH-raYion, .#moles/h i-er

FIG. 1. Adsorption isotherms of phenol on Black Pearls 607 Carbon Black from aqueous solution (30°C, one week equilibrium time). @ LC525 • LC325 0 • LC525 OR

,~000~I ~.900

It

~ 2oo

~--@" -

i

--

c~

E I00

I

I00 200 Conc~.nPral-[on, Fmoles/hf-~r

FIG. 2. Adsorption isotherms of phenol on activated Columbi~ Carbon LC325 from aqueolls solution (30°C, one week equilibrium time). LC325 samples of Table I. Table I I summarizes the plateau values for the amount adsorbed on a per-gram and per-unit surface area (B.E.T.-N2) basis. Table I I I compares the specific surface area for the different samples (1) as obtained b y assuming the plateau of each adsorption isotherm corresponds to monolayer coverage and (2) as obtained b y N2 adsorption with the use of the B.E.T. method. For methods (1) and (2) the area occupied per phenol molecule was t a k e n as 16.1 A?/molecule for nitrogen, 41.2 A2/molecule for phenol, and 43.0 A2/mole -

cule for nitrobenzene. Figures 4, 5, and 6 are adsorption isotherms for sodium benzenesulfonate on the P607, LC325, and $51 samples described in Table V. The surface areas of most of these samples were not measured so it is impossible to present adsorption capacities on a unit surface area basis. DISCUSSION The isotherms in Figs. 1 through 3, although covering ranges of very low composition, appear to be of the T y p e I, or Lang-

Journal of Colloid and Interface ~cience, ¥ol. 28, No. 3/4, November-December 1968

390

COUGHLIN, EZRA, AND TAN

u

1800 m

• LC525 O, I wk.

0.082 sample ~ I20C

• LC525, 1 wk. 0.0/g sample

m

•

LC325, 2 wks. O.02g sample

600 <

IO0 2~0 ConcenfraT'ion, ymoles / l l f e r

FIG. 3. Adsorption isotherms of nitrobenzene on activated Columbia Carbon LC325 from aqueous solution, 30°C. TABLE I I P L A T E A U VALUES FOR THE ADSORPTION OF P H E N O L AND N I T R O B E N Z E N E

Product

P607 P607 O P607 R P607 OR P607 OROROR LC325 LC325 O LC325 OR

Phenol adsorption

t~moles/ gm)

~moles/ m2)

290 142 330 250 312

0.478 0.213 0.498 0.488 0.555

1000 120 190

0.833 0.216 0.475

Nitrobenzene adsorption

~moles/ gm)

(izmoles/ m s)

m u i r v a r i e t y . T h e p l a t e a u s of t h e s e isot h e r m s suggest surface c o v e r a g e of t h e a d s o r b e n t in t h e r a n g e of 5 % to 40 %, well below t h e m o n o l a y e r region. T h i s is in agreem e n t w i t h t h e w o r k of M o r r i s a n d W e b e r (15), who o b s e r v e d t h a t t h e p h e n o l i s o t h e r m exhibits similar s h a p e a t high a n d a t low c o n c e n t r a t i o n s ; in t h e i r w o r k t h e c u r v a t u r e i

200

• P607 • P607-OA O P607-OA ~ f h O 2 m m o l e CaC~OOm]solb

1950 500

1.63 0.90

"~

A P607-ORA

.

TABLE I I I SURFACE A R E A OF CARBON SAMPLES

[O0

Surface area (m2/gm) Adsorbent

P607 P607 O P607 R P607 OR P607 OROROR LC325 LC325 O LC325 O k

Isotherm plateau for phenol

73.5 34.8 87-89.0 62.0 77.3 268.0 29.8 48.7

Isotherm plateau for nitrobenzene

504 196

B.E.T. (N2)

655 668 663 513 562 1200 556 400

50

100

Conc~nf-raf-ion , /Lt moles/lif-er

FIG. 4. Adsorption isotherms of sodium benzenesulfonate on Carbon Black P607 in aqueous solution (30°C, 3 days equilibrium time).

Journal of Colloid and Interface Science, ¥oi. 28, No. 3/4, November-December 1968

CHEMISORBED OXYGEN IN ADSORPTION ONTO CARBON

indicate any interrelationship as was reported by Boehm (4). This could be attributed s0tely to nonuniformity of oxidation and reduction treatments both with respect to reagents and treatment time in the work reported here. However, there are evident differences between the base consumptions by carbon of different treatment histories. It is quite clear from the tables that reduction markedly decreases surface acidity, whereas oxidation causes the surface concentration of acid groups to increase significantly. The effects of oxidation and reduction on the different functional groups appear not always to be similar in degree for all adsorbents, although the general trend is clearly evident from the data in Tables I and V. In particular, the untreated active carbons appear to possess a substantially smaller quantity of surface acidity when compared to the untreated Black Pearls 607. Upon oxidation, the active carbons develop rather large amounts of surface acidity and

300

200

(3 £

• LC325

6}

T LC325 wifh 0.2 mmole CaCl~ per I00 m/ soldfion

%.

•

~

391

LC325 wifh 0,4 mmole CaCI~ per" lOOm/ soluf ion

100

i

• S5I 5b Cone enfrafio n ,/gL mofe s / l i f e r

•

I00

• S51-O

FIG. 5. Adsorption isotherms of sodium ben-

zenesulfonate on activated Columbia Carbon LC325 in aqueous sohtion (30°C, 3 days equilibrium time).

S51-OR wifh 02 mmoles CaClJlOOmi sol'n

A SS/-OR

O

20C

e~

exhibited by data covering low concentrations was completely concealed when plotted with data obtained at high concentrations owing to compression within the first few divisions on the scale of concentration. They argued this sort of continuous curvature suggests multilayer adsorption with a range of activities for various portions of the adsorbent surface. A similar kind of argument has been invoked by Hansen, Fu, and Bartell (8), who found the number of adsorbed layers at Saturation to be on the order of three or four for alcohols possessing four or more carbon atoms. In Tables I and V neither the equilibrium consumption of the various bases nor the concentrations of the various surface functional groups deduced therefrom seem to

% u)

IO0

I-"

°

50 Concenfrafion , //A mole s / l i f e r

° I00

FIG. 6. Adsorption isotherms of sodium ben-

zenesulfonate on activated carbon "Dareo $51" in aqueous solution (30°C, 3 days equilibrium time).

Journal of Colloid and Irderface Science,

Vol. 28~ No. 3/4, November-December1968

392

COUGtILIN, EZRA, AND TAN TABLE IV

1)ERCENTAGE OF ADSORBENT SURFACE A R E A (B.E.T.-N2) OCCUPIED BY 1)HENOL AND BY NITROBENZENE Adsorbent

1)607 1)607 O P607 1% P607 OR 1)607 OROROR LC325 LC325 O LC325 OR

Surface occupied by

11.2 5.2 13.2 12.1 13.8 22.3 5.4 12.2

Surface occupied by

42.1 35.2

their original low acidity is not nearly restored by subsequent reduction. This is in contrast to the oxidized P607, which appears to be returned to almost its originM acidity by reduction. The approximate locations of the apparent plateaus of the various isotherms in Figs. 1, 2, and 3 have been estimated visually, and these values have been assembled in Table II for the purpose of comparison. These "plateau-values" represent the adsorption capacities of the various carbons at the concentrations employed. Here these quantities are expressed in two ways: (1) ~moles adsorbed per gram of adsorbent and (2) ~moles adsorbed per square meter surface area of adsorbent; the latter quantities were computed by using the values of specific surface area obtained from N2 adsorption by the B.E.T. method. Whether the adsorptive capacities of the carbons given in Table II be compared on a per gram or per unit surface area basis, it is clear that the acid groups on the surface resulting from oxidation appear to have a significant effect on adsorption. In particular, the oxidized sample of P607 manifests adsorptive capacities for phenol about h~lf of those displayed by the reduced samples. In the case of the active carbon, LC325, the effect of oxidation appears to be even more pronounced. Oxidation reduces the adsorptive capacity of this carbon for phenol by a factor of about eight on a per gram basis or by a factor of about four on the basis of unit surface area. For the adsorption of

nitrobenzene on this carbon the effect of oxidation on adsorptive capacity is somewhat smaller: a decrease by a factor of about four on a per gram basis or by a factor of about two on a unit surface area basis. Examination of the B.E.T. (N2) surface areas for the adsorbent samples of Table I shows little material change upon oxidation and reduction treatment for the P607, whereas these treatments reduce the specific surface area of LC325 by a factor of two to three. This difference may be interpreted as resulting from the less severe oxidation conditions to which the channel black was subjected or it may be closely connected with the comp]ex structure of very fine pores in an active carbon compared to what is probably a more open network of larger pores in the case of the carbon black. The more delicate network of pores in an active carbon might be expected to suffer a more drastic decrease in surface area upon oxidation. The data of Table I also suggest that reduction of oxidized samples brings about small decreases in specific surface area. Reduction with hydrogen might be expected to split off parts of the oxygenated carbon structure, although experimental conditions were not those that usually favor hydrogenolysis. Capillary condensation of low molecular weight products of oxidation and reduction is another possible partial explanation for the observed losses in surface area. Table V lists the percentages of total adsorbent surface area (taken as the B.E.T.-N2 measurement) which is occupied by adsorbate in the plateau region. The adsorptive capacities of the carbons of Table I, expressed in this way in Table IV, also display the strong effects of treatment by oxidation or reduction. There is good correlation between total acidity (NaOCH2CH3 consumption) and the percentage of adsorbent surface covered by phenol. This is shown on the graph in Fig. 7. Similar correlations exist between the adsorptive capacities and the concentrations of weaker acidic sites on the surface but, in the latter instances, there appears to be a lesser dependence and the data scatter more. It is not yet possible to assess the possibly different degrees to which the different groups influence the adsorptive capacity.

Journal of Colloid and Interface Science, Vol. 28, No. 3/4, November-December 1968

CHEMISORBED OXYGEN IN ADSORPTION ONTO CARBON

393

TABLE V SURFACE ACIDITIES OF CARBON ADSORBENT SAMPLES Base consumptiona (meq/g) Product

P607 P607 CAb P607 OI~Ab LC325 LC325 OAb LC325 ORAb $51 $51 O $51 OR

NaOEt

NaOH

2.65 3.50 2.97 0.59 3.10 2.46 1.08 4.21 3.17

2.03 2.67 2.23 0.48 2.98 2.09 0.71 3.57 2.37

Na2CO~

1.32 1.85 1.42 0.28 2.05 1.29 0.47 2.45 1.58

Surface groups (meq/gm) NaIIC0~

I

II

1.09 1.52 1.07 0.30 1.50 0.92 0.40 1.85 1.20

1.09 1.52 1.07 0.30 1.50 0.92 0.40 1.85 1.20

0.23 0.33 0.35 0 0.55 0.37 0.07 0.60 0.38

III

IV

0.71 0.82 0.81 0.20 0.93 0.80 0.24 1.12 0.79

0.62 0.83 0.74 0.11 0.12 0.37 0.37 0.64 0.80

a Base consumption was measured after equilibration periods of about 3 days. Oxidation conditions were : 0.5/1// (NH,)2S2Os solution, 2 days for P607A and saturated solution, 2 days for LC325 and $51. b Samples labeled with " A " were from batches of treated P697 and LC325 different from those described in Table I. I

• P607 Samples • LC325 Samples 2C

:n/5

c3 s.

5

I 2

I 3

Tof-al AcMif 9, meq/g

FIG. 7. Percentage area occupied by phenol vs. total amount of acidity. Isotherms for sodium benzenesulfonate adsorption are plotted in Figs. 4, 5, and 6 for the three different ldnds of carbon. The carbon adsorbents and their pretreatments are described in Table V. Note that this table describes carbon samples used to study sodium benzenesulfonate adsorption; Table I describes the carbons used in the other experiments. As in the case of phenol and nitrobenzene adsorptive capacity is markedly lowered b y oxidation of the carbon but restored upon subsequent reduction. Some of the isotherms in these figures have been determined in the presence of added CaC12, which greatly increases the adsorp-

tion of sodium benzenesulfonate. Similar behavior has been noted by Zettlemoyer, Skewis, and Chessick (1) and interpreted as due to a decrease in mutual repulsion between surfactant head groups brought about by interaction with divalent calcium ions (1, 9). I t is interesting to note that in the present work the capacities of the carbons for sodium benzenesulfonate adsorption appear to be considerably smaller than for the adsorption of phenol and nitrobenzene. These differences are in excess of what might be expected from a consideration of the relative molecular size of the individual adsorbates. This appears to be in accord with

Journal of Colloid and Interface Science, Vol. 28, No. 3/4, November-December 1968

394

COUGHLIN, EZRA, AND TAN

the concept of mutual repulsion of sulfonate "head groups." It is clear from the graphs that the presence of acidic groups on the carbon surface does not appear to interfere with the action of the calcium ions in reducing the repulsive interaction among head groups. Dependence of adsorption on the nature of the carbon surface has been reported before (3, 9-12), but the present authors are not aware of any work in which the same carbon has had its adsorptive capacity lessened and subsequently regenerated by chemical treatment. In fact, from the data reported here, the treatment of carbon black seems to be reversible through up to three cycles of oxidation and reduction at least in so far as the phenomenon of phenol adsorption is concerned. Most previous work reported about adsorption onto carbon surfaces has not concerned aqueous solutions. Perhaps the two papers most relevant to the present work are those by Graham (3) and Clauss, Boehm, and Hofmann (11). Graham (3) used six different active carbons and a graphitized carbon black (Graphon) and found that acid groups on the carbon surface tend to reduce the capacity for adsorption of metanil yellow from aqueous solution although not for adsorption of methylene blue, which has about the same size molecule. On the other hand, Clauss et al. (11) showed that a B.E.T. analysis of phenol adsorption data from aqueous solution on several different carbon blacks treated in different ways resulted in values of specific surface area in substantial agreement with those obtained by B.E.T.(N2) adsorption. In this ease then, phenol adsorption from aqueous solution did not appear to be affected by acidic surface groups in different concentrations on the carbons (as a result of graphitizing, outgassing at high temperature, activation, etc.). The apparent tack of agreement between these results and the present work can be attributed to the high phenol concentrations used by Clauss and co-workers. This means that much of their data fell in the multilayer region of surface coverage whereas t h e results reported here pertain to coverages well below the monolayer region. The present results, obtained at the low aqueous phenol con-

centrations of interest from the standpoint of water treatment (18), are a rather more sensitive measure of the interaction between the phenol solution and the carbon surface in contrast to the work of Clauss et al. Previous experiments on the phenol water system have shown that the phenol isotherm displays two plateaus in adsorption on carbon (8), whereas adsorption on a polar substrate such as alumina (16) leads to an stype isotherm with a single plateau. The l a t t e r behavior has been interpreted (17) as a monofuneti0nal attraction toward the polar substrate arising from the hydroxyl group of the phenol. Here the phenol molecules are viewed a s adsorbed end-on with larger eoneentrati0ns bringing about cooperative adsorption owing to interaction of neighboring adsorbate molecules. This accounts for the s-shaped isotherm. However, in adsorption of phenol on carbon from aqueous solutions of low concentration (first plateau or step of the isotherm) the initial portion of the isotherm is not s-shaped but appears to possess normal Langmuirian character. This suggests that the attraction of phenol for the carbon substrate lies probably in nonpolar forces operating over the entire phenol nucleus. This led Giles et al. to suggest that the second step in the isotherm may represent the uncovering of a portion of the original surface of the carbon substrate by a reorientation of the phenol molecules from the prostrate to the end-on position. The approximate positions of the two steps in the isotherm are consistent with surface coverage in flat orientation at the first plateau and complete coverage in vertical orientation at the second. Alternatively, the second plateau may represent a second condensed monolayer formed on top of the first. The difference between the role of the carbon surface in adsorption equilibrium at low phenol concentrations as compared to large concentrations is graphically illustrated by Fig. 8. Here some additional data for phenol adsorption on carbon at high concentrations are plotted together with the data discussed above. It is clear from Fig. 8 that the chemically bound oxygen on the carbon adsorbents appears to influence the position of the first plateau of the phenol isotherm but not the second or high-eoncen-

Journal of Colloid and Interface Science, Vol. 28, No. 3/4, November-December 1968

CHEMISORBED OXYGEN IN ADSORPTION OI~TO CARBON ,

2.2

,

l

I-

S

• Reducedcarbon & Omgmalcarbon • Oxid,zedcarbon

.

,

,

I

395

[

i I ]!

~

& ~ 0,2 OI o,,

0.2

o.3

....

,b

2;o

Phenolconcenfratlon,rnmoles/hter

.~b

FIG. 8. Adsorption of phenol on Black Pearls 607 Carbon Black from aqueous solutions, 30°C. tration plateau. Thus oxygen chemisorbed on carbon appears to strongly influence phenol adsorption only under conditions when the molecules are thought to be adsorbed in the prone position on the graphitic basal planes with the attractive forces operating over the entire phenol nucleus. This oxygen appears to exert very little influence on phenol adsorption under conditions where the molecules are thought to be adsorbed in the vertical or "end-on" position with forces of interaction between phenol molecules themselves presumably affecting the process more than forces between the phenol and the substrate. It is interesting to consider possible explanations for the role of the acidic surface oxygen groups in inhibiting adsorption of phenol and nitrobenzene molecules from aqueous solution. The major portion of these oxygen groups on carbon are presumably located at the edges of the layer planes where they would not be expected to markedly interfere, in a steric sense, with adsorption of organic molecules onto the basal planes. It is conceivable that oxygen chemically bound on the edges localizes electrons in surface states thereby removing them from the 7r-electron system of the basal planes. Walker, Austin, and Teitjen (13) have explained the effect of oxygen chemisorption on the thermoelectric power of carbon by similar reasoning. Increases in thermoelectric power of carbon due to oxygen cheInisorption were attributed by these

workers to a depletion in the number of electrons and an increase in the population of positive holes in the conduction band of the 7r-electron system. These considerations appear consistent with the concept of dispersion forces between the phenol 7r-electron system and the 7r band of the graphitic planes of the carbon as responsible for adsorption. Removai of electrons from the ~r band of the carbon by chemisorbed oxygen might be expected to interfere with and weaken these forces. Another phenomenon that could be expeered to play a role in the adsorption studies reported here is the bonding of water molecules to the oxide functional groups. According to Dubinin (14) the water molecules adsorbed to the oxygen groups become secondary adsorption centers which retain other water molecules by means of hydrogen bonds. As a result, complexes of associated water form within the pores of a carbon adsorbent. These complexes could prevent the migration of organic molecules to a large portion of the active surface area within a particle of adsorbent. However, if this were the case the chemisorbed oxygen would be expected to exert an influence on the second phenol adsorption plateau similar to that evident for the first plateau. Another possible explanation of the observed phenomena may reside in a model based on capillary condensation in which the lowconcentration part of the isotherm is affected by oxygen complexes in the micropores

Journal of Colloid and[nterface Sciefwe, Vol. 28, No. 3/4, November-December 1968

396

COUGHLIN, EZRA, AND TAN

whereas condensation at higher concentration in the wider pores might not b e so affected. T h e evidence is not overwhelming, however, and it is hoped t h a t continuing research will shed more light on this question. I t is possible to conclude from the work presented above t h a t chemical t r e a t m e n t of a carbon surface can strongly alter its adsorptive capacity and t h a t these effects are due to more than mere changes in surface area. T h e results reported here for the effects of chemisorbed oxygen are significant in view of the fact t h a t regeneration of carbon adsorbents frequently invoNes some degree of oxidation. The addition of a reduction step to the regeneration process would appear to have the obvious technical advantage of increased adsorptive capacity for molecules like nitrobenzene, benzenesulfonate, and phenol at low concentrations. Further, in the manufacture of adsorptive carbon, control of chemical and physical process conditions might be harnessed to produce carbon surfaces suitable for particular adsorption applications. ACKNOWLEDGMENTS The authors would like to thank 1~. C. Trivedi for assistance in determining B.E.T. (N2) surface areas, and the Federal Water Pollution Control Administration for support under Grant No. WP 00969-01. REFERENCES 1. ZETTLEMO~ER, A. C., SKEWIS, J. D., ANI) CHESSICK, J. J., J. Am. Oil Chemists Soc. 39, 280 (1962). 2. ZETTLEMOYER, A. C., AND CHESSICK, J. J., Advan. Chem. Ser. 43, 88 (1964). 3. GRAHAM,D., J. Phys. Chem. 59, 896 (1955). 4. BOEHM, H. P., DIEHL, E., HECK, W., AND

SAPPOK, R., Angew. Chem. Intern Ed. •ngl.

3(10), 669-677 (1964). 5. STUDEBAKER, M. L., I-IUFFMAN, K. W. D., WOLFE, A. C., AND NABORS, L. G., Ind. Eng. Chem. 48, 162 (1956). 6. GARTEN, V. A., WEISS, D. E., AND WILLIS, J. B., Australian J. Chem. 10, 295 (1957). 7. WEBER, W. J., AND MORRIS, J. C., J. Sanit. Eng. Div. ASCE 90 (SA3), 79 (1964). 8. HANSEN, R. S., Fu, Y., AND BARTELL, F. E., J. Phys. Chem. 53, 796 (1949). 9. ZETTLEMOYER, h. C., AND NARAYON, K. S.,

"Adsorption From Solution By Graphite Surfaces," in P. L. Walker, Jr., ed., "Chemistry And Physics of Carbon," Vol. 2, p. 197. Dekker, New York, (1965). 10. PARFITT, G. D., AND WILLIS, E., J. Phys. Chem. 68(7), 1780 (1964).

11. CLAuss, A., BOEJZM, H. P., AND HOFMANN, U., Z. Anorg. Allgem. Chem. 290, 35-51 (1957). 12. KIPLING, J. J., AND GASSER, C. G., J. Phys. Chem. 64, 710 (1960). 13. WALKER, P. L., AUSTIN, L. G., AND TIETJEN,

J. J., "Oxygen Chemisorption Effects On Graphite Thermoelectric Power," in P. J. Walker, ed., "Chemistry and Physics of Carbon," Vol. 1, p. 328. Dekker, New York, 1965. 14. DUBININ, M. M. "Porous Structure And Adsorption Properties Of Active Carbons," in P. J. Walker, ed., "Chemistry and Physics of Carbon," Vol. 2, p. 51. Dekker, New York, 1965. 15. MORRIS, J. C., AND W. J. WEBER, Proc. 1st Intern. Conf. Water Pollution Res., Pergamon Press, Oxford, 1962. 16. CUMMINGS,T., GRAVEN,H. C., GILES, G. H., RAHMAN, S. M. K., SNEDDEN, J. G., AND STEWART, C. E., J. Chem. Soc. 1959, 535.

17. GILES, C. H., MACEWAN, W. H., NAKHWA, S. N., SMITH,D., J. Chem. Soc. 1960, 3973. 18. COUGIILIN, R. W., AND EZRA, F. S., d. Environmental Sci. & Technology 2(4), 291

(1968).

Journal of Colloid and Interface Scien~, Vol.28, No. 3/4, November-December1968