THE UNIVERSITY OF MICHIGAN DEPARTMENT OF CIVIL ENGINEERING Adsorption of Strong Acids, Phenol, and 4-Nitrophenol From Aqueous Solution by Active Carbon In Agitated Non-Flow Systems Vernon L. Snoeyink and Walter J. Weber, Jr. Federal Water Pollution Control Administration United States Department of Interior Grant No. WP-00706 Washington, D.C.

TABLE OF CONTENTS Page ACKNOWLEDGMENTS e o. e.. o. o, i i LIST OF TABLES.0 o o 0.. a4. a vi. LIST OF ILLUSTRATIONS o o.. 0.0 o 0. o a 0 i oi LIST OF APPENDICES. o O e C O o o X I INTRODUCTION o o o o 1 Plan of Research 0 o. o ^. 2 II. STRUCTURAL PROPERTIES AND SURFACE CHEMISTRY OF ACTIVE CARBON a o o 0. 4 Preparation of Active Carbon o o e 0 6 Structure and Structural Development. 8 Chemical Nature of the Active Carbon Surface 0o a e e 8 e 0 4 4 o 0 0 e e 13 Characterization of Oxygen Surface Structures. c o o 0 e o o o o0 0 0 0 0 19 Discussion o o o 0 o o 0 o 0 o o 0 o o o0 29 IIIe REACTION OF THE HYDRATED PROTON WITH ACTIVE CARBON I a O c 4 O O 0 o O 0 31 Experimental o o o a o a o o o o o ~ o o 31 Carbons o o o o 0 o o o o e 0 31 Experimental Systems o, a a, 32 Ash Content Analysis e c o e 34 Stoichiometry Studies e o c o o o c 35 Results and Discussion o o o o e0 e. o e 36 Rate-Controlling Step o 0 o o 0 a o 0 36 Diffusion Model o o c e e o e 0 e 0 0 * 37

Page Initial Hydronium-Ion Activity of Effect o o 0, 0. e c e 41 Salt Effects.o oe. d. 51 Effects of the Type of Acid e, e 60 Temperature Effects. o e o. d 6 64 Summary o n a,. y o o 68 IV. ADSORPTION OF PHENOL AND 4-NITROPHENOL. 71 Experimental c o.. o.., o 6 o 6 o6 71 Carbon 6 ~ o o o o 6 e 0 a. 71 Solute Properties and Analysis e o J 71 Equilibria Studies. o o e,, 72 Kinetic Studies o o 6 o o 73 Results and Discussion o o o o. o 74 Phenol and 4-Nitrophenol Adsorption Equilibria o 0. o. o 6 o o o a 74 Phenol Reversibility o0 o 0 o o o 6 84 Temperature Dependence of Equilibria 6 87 Phenol and 4-Nitrophenol Adsorption Rates o 6 e. 6 o o 6 o 6 0 0 6 o o 91 Temperature Dependence of Adsorption Rates o 0 6o I o c o o o0 98 Dependence of Adsorption Rate on Carbon Dosage o, 6 o o o o o o o O O O 107 Dosage 0 0 0 0 0 0 0 I 0 0 0 6 0 0 0 6 107 Dependence of Adsorption Rate on Initial Concentration of Solute o o o 0 6 6 109 Effect of Ash Reduction o o o O 6 0 6 6 111 Effect of pH on Adsorption, e o e o 113 Effect of pH on Equilibria o 6 6 6 113 Effect of pH on Rate. 6 6 6 o o 6 125 Effect of NaCl on Adsorption 6 e o 6. 132 iv

Page Effect of NaCl on Equilibria o 0 132 Effect of NaC1 on Rate..... 136 Effect of Type of Strong Acid on Adsorption...... o... o. 136 V. CONCLUSIONS. o. o o o o o. e o o o 137 Acid Adsorption. o c o o e e o o o o 137 Phenol and 4-Nitrophenol Adsorption... 139 Further Research e. 0 G. O o o 0 142 APPENDICES.. o o. o o c o 0 o o 0 O o. o o 143 REFERENCES.. e e o o o o o c. o o o 146 V

LIST OF TABLES Table Page III-1. Diffusion Coefficients as a Function of Initial pH for the HC1-NaCl System... 41 III-2. The Effect of NaC1 Concentration on Diffusion of HC1........... 51 III-3. The Effect of DaC104 Concentration on Diffusion of HC104........ 54 III-4. Diffusion Coefficients for Different Strong Acids....o. O.,...... 60 III-5 Diffusion Coefficients for the HCl-NaCl System at Different Temperatures... o e o 64 IV-1. Langmuir Parameters for Phenol and 4-Nitrophenol Adsorption at Different Temperatures... o e 89 IV-2. Diffusion Coefficients for Phenol and 4-Nitrophenol............ 95 IV-3. Diffusion Coefficients for Phenol and 4-Nitrophenol at Different Temperatures 98 IV-4. The Relative Rate Constant at Different Temperatures...... e.. 106 IV-5. Diffusion Coefficients at Different Carbon Dosages......... e 0 109 IV-6. Isosteric Heats of Adsorption at Different Phenol Surface Coverages... 124 IV-7. Diffusion Coefficients for Phenol at Different pH...... o..... 125 IV-8. Diffusion Coefficients for 4-Nitrophenol at Different pH o..e.. o..... 128 IV-9. Diffusion Coefficients at Different Temperatures for 4-Nitrophenol Anion... 132 vi

LIST OF ILLUSTRATIONS Figure Page II-1. The Arrangement of Carbon Atoms in the Graphite Crystal..... o o. 10 II-2. A Phenolic Surface Structure... o 22 II-3. A n-Lactone Surface Structure... 24 11-4. Some Reactions of Fluorescein-Type Lactones........ o o 25 II-5. The Chromene-Acid Reaction e. 27 II-6. Hydrolysis of the Carbonium Ion...0. 28 III-l1 Adsorption Isotherms for HC1 0 o o o o 42 III-2. Finite-Bath Rates of Adsorption for HC1 43 III-3. Infinite-Bath Rates of Adsorption for HC1.. 0 0 0 0. e. o. o o o o o 44 III-4. Finite-Bath Rates of Adsorption for HC1 on Different Carbons 0 0 0 0 0 0 o 47 III-5o The Effect of Ash Removal on FiniteBath Adsorption Rate...0 o 0. 49 III-6. The Effect of Salt Concentration on Adsorption Capacity for HC1..... 52 III-7, Finite-Bath Rate of Adsorption for HC1 at 0o02 M NaC1 e o o o0 0 53 III-8o The Effect of Salt Concentration on Adsorption Capacity for HC104 0 0 0 o 55 III-9, The Effect of Salt Concentration on Adsorption Capacity for Acid 0 0 0.. 56 III-10l Finite-Bath Rate of Adsorption for HC104 0 0. o o o o o. o 0 o. 0 57 III-11. Adsorption Isotherms for Different Strong Acids 0...... C 61 III-12. Finite-Bath Rate of Adsorption for HN03 o 62 vii

Figure Page III-13o The Effect of Temperature on Adsorption Capacity for HC1 0 e o O o l 65 III-14o Finite-Bath Rate of Adsorption for HC1 at 20 C o o o o c o 0 0 o o 66 III-15o Finite-Bath Rate of Adsorption for HC1 at 50~ C. o o o o. o 67 IV-lo Phenol Equil bria with Freundlich Equations... o o o o o o 0 o o 76 IV-20 Phenol Equilibria with Langmuir Equations o o o o o o o o o o 79 IV-30 4-Nitrophenol Equilibria with Freundlich Equations 0...0 o 0 80 IV-4o 4-Nitrophenol Equilibria with Langmuir Equation O0 O e e o 81 IV-5o Phenol and 4-Nitrophenol Equilibria as a Function of Reduced Concentration o 83 IV-6. Phenol Desorption Isotherms o o o o o o0 86 IV-7o Phenol Equilibria at Different Temperatures 0 0 0 0 0 88 IV-8. 4-Nitrophenol Equilibria at Different Temperatures 0 o 0 90 IV-9. Phenol Rate of Adsorption o 0o o o o o o 93 IV-l10 4-Nitrophenol Rate of Adsorption 0 0 o 94 IV-llo Phenol Rate of Adsorption at Different Temperatures o 0 0. o 0o o 99 IV-120 4-Nitrophenol Rate of Adsorption at Different Temperatures 0 0 0 0 0 0 0 0 100 IV-13o Phenol Rate of Adsorption. 0 0 0 0 0 0 0 102 IV-14o 4-Nitrophenol Rate of Adsorption.0 0 0 103 IV-150 Temperature Dependence of Phenol Rate of Adsorption 0 0 0 0 0 0 0 o 0 0 104 IV-16o Temperature Depends of 4-Nitrophenol Rate of Adsorption o o0 0 0 o 0 0 0 0 0 0 105 viii

Figure Page IV-17o 4-Nitrophenol Rate of Adsorption at Different Carbon Dosages.. 0 o 108 IV-18. Phenol Rate of Adsorption at Different Initial Concentration o c o 110 IV-19. Phenol Rate of Adsorption at Different iniitial Concentratiton o 112 IV-20o Phenol Equit bria at Different pH o. 115 IV-21, Phenol Equilibria at Different pH o o o C 116 IV-22o 4-Nitropheno' Equilibria at Different pH o o o o o o o o o o 118 IV-230 4-Nitrophenol Equilibrium as a Function of Reduced Concentration o o o o o o 120 IV-24, 4-Nitrophenol Anion Equilibria at Different Temperatures. o o o 122 IV-25o Phenol Equilibria at Different Temperatures for pH=2o0 o o o o o o o o 123 IV-26e Phenol Rate of Adsorption at Different pH 0 0 0 0 0 0 0 0 126 IV-27 4-Nitrophenol Rate of Adsorption at Different pH.......... 127 IV-28o Phenol Rate of Adsorption at Different pH o o o o o o o o o 130 IV-29o 4-Nitrophenol Rate of Adsorption at Different pH O o o o o o o 131 IV-30o 4-Nitrophenoi Anion Rate of Adsorption at Different Temperatures o o o 133 IV-31o 4-Nitrophenol Anion Equilibria at Different Salt Concentrations.. o 135 ix

LIST OF APPENDICES Appendix Page I, Computer Program for Solution of the Diffusion Equations o o o o o o o o o o 143 x

I INTRODUCTION Essentially complete removal of biologically-resistent organic pollutants is becoming imore and more important in water purification and waste treatment operations as the frequency of water reuse increases. In light of projected rates of population growth, it can be expected that expansion of the water reuse practice will continue in order to meet water supply demands in the future, It is the primary task of engineers and scientists to provide the processes and techniques which will make such reuse completely safe and aesthetically acceptable, at the lowest possible cost, One means for removal of organic materials from aqueous solution is by adsorption on active carbon. Active carbon has been used for some time in the water purification industry for taste and odor reduction, but its use in this regard has been rather inefficient, primarily because of the lack of fundamental information regarding the nature of the sorption process and optimum conditions for operation, Recently, however, much research has been done on definition of the nature of adsorption by active carbon (38, 68, 69, 70, 71) and on development of better methods for contacting carbon with the solutions to be treated (3, 15, 40, 67), However, there is *yeta definite need to know more about the basic mechanisms of adsorption on carbon0 Specifically, very little research has been done on the effect of 1

2 inorganic constituents on the adsorption of organic substances from aqueous solution. Weber and Morris (70) have found that the rate of uptake of alkylbenzenesulfonates increases with decreasing pH, while Gould (25) has noted that high pH has a significant effect on the rate of adsorption of substituted phenols. The author has been motivated by a desire to extend this previous research on the effect of pH to a more extensive study of the effects of inorganic constituents in general on the adsorption process. It has been hoped also that a study of this type would reveal useful information about the "active sites" at which adsorption on carbon takes place. PLAN OF RESEARCH It was decided to approach the problem of determining the effect of selected inorganic substances on the adsorption process by first studying the direct reaction between such inorganic species and carbon, The adsorption of organic molecules then could be investigated in the absence and presence of these inorganic species, and the results obtained in the former systems used for inte lrpretat ion of those obtained in the latter systems, Preliminary studies showed that, for the active carbons used, the reaction of strong acids with the carbon surface was quite significant and, therefore, hydronium ion concentration was selected as a major variable, In a? attempt to completely elucidate the hydronium ion-carbon reaction, it was studied as a function of pH, type of acid, salt concentration, temperature, ash content of the carbon and quantity of anion

3 removed from solution during the reaction. These variables were examined with respect to rate and extent of the reaction, Once the information f or the hydronium ion-active carbon reaction was obtained, the inorganic species could then be studied with respect to their effect on the adsorption of selected organic substa,ceso For this purpose, phenol and 4-nitrophenol were chosen as sorbates, Phenol was selected primarily because it is a common pollutant and presents considerable problems in waste treatment (49) o Also, it is sta;ile in solutions over long periods of time and its concentration is easily analyzed. In addition, it can be studied as both neutral and anionic species, a factor which is useful for interpretation of sorptive behavior under different conditionso Nitrophenol was selected because it is structurally similar to phenolo It is a common pesticide which persists in the natural aquatic environment because of its resistance to biological break down (68) Furthermore, it too is chemically stable in aqueous solution, can be easily analyzed, and can be studied as both the neutral and negative speciesc It was thought that studies on both of these solutes would be complementary0 Investigations of the kinetics and equilibria of adsorption of these two solutes were performed as a function of pH, type of acid, temperature, salt concentration and ash content of the carbono It is hoped that the results of this research will be useful in the further development and refinement of the process of adsorption on active carbon as a method for water purification and waste water treatmento

IIo STRUCrLTUPRAL PROPERTIES AND SU'RFACE CHEMISTRY OF ACTIVE CARBON Before proceedl..iing with.iiscussion of the effect of inorganic substances on t"a adsorption of organic molecules, it is important to carefully examine what is known about the structure and surface chemisteiry of active carbon in order to establish a basis for inte.pre.ytin:g experimental resultso It is of interest, for example, to examine how the extensive surfaces of carbons having the same surface areas can have totally different adsorptive properties, Part of such differences in adsorptive behavior can be explained in terms of relative pore size distributions, but part must also be attributed to differences between the respective surface properties of the carbonso The surface of any active carbon is comprised in part of residual electron- and ion-exchange functional groups, connected by electron-conducting bond systems (22) The nature of these functional groups is dete::mined to a large extent by the method of activation as well as by tihe type of raw material from which the carbon is prepared, With the existence of electron- and ion-exchange groups at the surface of carbon, it is reasonable to expect that electrolytes in solution may interact with the carbon to influence its behavior as an adsorbe..nt, Indeed, such interactions could markedly affect the over-all adsorption process under certain conditions; possibly effecting changes in the process which may be put to good advantage to achieve 4

5 higher efficiency or enhanced effectiveness of removal of pollutants. Weber and Morris (70), for example, have shown that the hydrogen ion activity of a solution affects both rate and capacity for adsorption of negatively charged species of sulfonated alkylbenzene from aqueous solution by carbonO These authors have interpretea their results as indicating that decreasing pH leads to an increasing association of the positively charged hydrated proton with negatively charged functional groups on the active carbon surface, thus neutralizing electrostatic repulsive force to allow the charged organic species to migrate more easily through the pore spaces and be adsorbed at inner surface sites. Garten and Weiss (22) have interpreted changes in hydrogen ion activity observed upon addition of active carbon to acidic or basic solutions in terms of interaction between the acid or base and a basic or acidic heterocyclic surface structure, respectively. It is highly probable that this interaction has a significant effect on the sorptive properties of active carbon Other inorganic species also seem to be important in their effects on the manner in which active carbon functions in aqueous solution as a sorbent for organic substancese Joyce and Sukenik (39) have observed a 50% increase in ash content (from 6 to 9% and from 12 to 18%) to be accompanied by a 50% decrease in the capacity of an active carbon after 10 cycles of sorption from a solution of the calcium salt of a sulfonated alkylbenzene followed by thermal regeneration~ Because the ash was soluble in acid solution, but insoluble in near neutral

6 solution, these authors suggest that acid treatment in conjunction with thermal regeneration may prove feasible as a means for restoring active carbon to near original capacity. Before such treatment is effected however, one should know how the acid may affect the active carbon surface, and therefore, how it may influence the adsorption process. Various chemical oxidizing agents can also interact with active carbon in aqueous solution to increase the amount of chemisorbed oxygen at the carbon surface (53) 3 Once more, however, the effect of the presence of this oxygen on the sorptive properties of active carbon has not yet been sufficiently defined. Many things are known, though, about the structure and surface properties which are invaluable, and this information will be used to develop a relatively unified concept or model of carbon which can be used as a basis for further research, PREPARATION OF ACTIVE CARBON The term active carbon actually implies a rather broad family of substances, with individual species being characterized by sorptive and catalytic properties rather than by definite structure or specific chemical composition (32), Particular properties may be imparted to an active carbon either by starting with different raw materials or by utilizing different preparative procedures. Substances from which active carbons are manufactured are carbonaceous in nature, usually carbohydrates; wood, coal,

7 sugar, nut shells, and vegetable matter are examples of typical raw materials. Differences in active carbons may exist as a result of the particular characteristics of the starting material, but many of the differences commonly ascribed to dissimilar starting materials may be due in fact to noncarbonaceous impurities. In addition, preparative conditions may be varied and selected additives employed to produce a particular active carbon of desired structure and surface properties, The production of active carbon involves first the formation of a char from the raw material. The starting material is heated —usually in the absence of air —to a temperature sufficient to effect drying and to drive off the volatile fraction~ Carbonization is then accomplished by elevating the temperature, producing a carbonaceous char (32, 38). In this process the temperature is increased slowly to allow time for each step to be completed before the next begins, The entire process is usually carried out below 6000 C., although exceptions do exist, Additives such as calcium chloride, magnesium chloride, zinc chloride, or any one of a number of similar materials may be used to catalyze the process and to impart certain characteristic properties to the finished product (32) The char developed in the carbonization process has relatively little internal surface area, and because a large surface is required for most adsorption applications, the char is generally "activated" by treatment with oxidizing gases under carefully controlled conditions0 The oxidizing gases —commonly carbon dioxide, steam, and air —attack the more readily

8 oxidizable portions of the char, resulting in the development of a porous structure and an extensive internal surface. The temperature of activation is a critical factor in the process, as will be discussed in more detail shortly, Treatment of the char with reagents suci as chlorine, dolomite sulfates, and several other substances to develop special properties in the finished product and to catalyze the action of the oxidizing gas is often effected prior to the activation process (32), According to Johnson et al. (38), carbonization and activation can sometimes be carried out in one step, rather than as separate operations. The total surface area of a carbon, by far tCe greatest percentage of which is internal, can be as great as 2,500 square meters per gram after activation. STRUCTURE AND STRUCTURAL DEVELOPMENT The molecular and crystalline structure of active carbon is an important determinant of the types of functional groups which can exist on the surface and is, therefore, an important consideration in any discussion of the surface chemistry of this materials Although there is little direct information available on the structure of active carbon, much can be derived from existing data on the structure of carbon black, Very little chemical difference exists between these two substances, and the only apparent physical difference is that the carbon black has much less internal surface area (22) As a basis for discussion of the active carbon structure, it is appropriate to consider first the structure of

9 ideal graphite in that it is closely approximated in the basic structural unit of both active carbon and carbon black (33, 63, 74) The structure of ideal graphite is given in Figure II-l1 According to Walker (6'), this structure is composed of a system of infinite layers of fused hexagons, Within each layer the carbon-carbon bond distance is 1415 A. indicating a onethird double bond character, Three of carbon's four electrons are engaged in forming regular covalent bonds with neighboring atoms and are localized, while the fourth resonates between several valence-bond structures, giving each carbon-carbon bond a one-third double bond character, Relatively weak van der Waals forces act between the parallel layers to hold the distance of separation to approximately 3,35 A. The carbon layers form an a-b-a-b-a-b stacking sequence in which one-half of the carbon atoms in any one plane lie above the center of the hexagons in the layer immediately below it. The carbon atoms are directly superimposed in alternate layerse X-ray diffraction spectra obtained for most natural graphites indicate that this structural configuration is in fact predominant (63) During the carbonization of the raw material in the preparation of active carbon, small aromatic nuclei, interpreted from x-ray spectrographs as microcrystallites consisting of places of fused hexagonal rings of carbon atoms having a structure similar to that of graphite, are formed (33, 74), The diameters of the planes making up the microcrystallite have

10 Figure II-1, The Arrangement of Carbon Atoms in the Graphite Crystal (63)

11 been estimated as 20 to 50 A. (33), The height of stacking is also in the range of 20 to 50 A,, indicating that each microcrystallite consists of 5 to 15 layers of aromatic planes. Although the structure of the microcrystallite is similar to that of graP:hite, it differs in many ways, During the formation of the microcrystallite, impurities should be expelled from its inte -ioro, However, Walker (63) states that interior vacancies exist in the microcrystallite and that their fo:lmation depends on the method of preparation, Possibly these vacancies contain some of the impurities; the presence of impurities may, in fact, influence the formation of such a vacancy. Sometimes there is a considerable content of disorganized, tetrahedrally bonded carbon, often cross-linking different layers (56). In addition, the ringed structures at the edges of the planes making up the microcrystallite are often heterocyclic owing to the nature of either the starting material or of the preparation process used (22) Heterocyclic groups would tend to affect both the distance of separation of adjacent planes and the sorp.tve properties of the carbon, The heteroatoms incorporated into the planes and the termination of the planes at the edges of the microcrystallite produce "free" valences which are very reactive (56). Usually these valences do not remain free for any length of time, but form compounds with any suitable element present. Needless to say, the "free" valences and the compounds formed as a result of them affect the nature of the carbon surface greatly. Also, orientation

12 between adjacent planes in the microcrystallite varies from that found in ideal graphitic structure to that of complete random orientation found in turbostratic carbons (63), The definite a-b-a-b-a-b stacking order, therefore, does not exist in most active carbons. Functional groups terminating the microcrystallite planes interconnect the microcrystallites, and are at least partially responsible for the turbostratic character in that they prevent orientation of the planes with respect to each othero Because of its special properties and the many differences between it and graphite, Garten and Weiss (22) have chosen to compare active carbon with a complex organic polymer rather than with a graphitic type particleo The structure of the char particle is developed further by treatment with oxidizing gases at high temperature during the activation processo Dubinin et alo (17), have noticed two stages of oxidation in the activation of carbonized sucroseo First, macropores were formed by the burnout of microcrystallite planeso Once a plane is attacked by an oxidizing gas, oxidation of that plane continues in preference to another area until a stable structure is developed (74) o Snow et al0 (56) have reported that a higher ash content causes faster pore development during activationo Apparently, inorganic matter causes stress localizations on the surface of the microcrystallite at which point, oxidation is initiated more easilyo The diameter of the macropore is generally taken to be in the range 5,000 to 20,000 Ao, while micropore diameters are usually

13 assumed to be less than 30 A0 (16) The macropores and micropores are connected by transitional pores The surface area of the interior walls of the macropores is less than 2 square meters per gram while the surface area for transitional pores can vary from 20 to 450 square meters per gram (16) However; the majority of surface area for carbons is usually found in the microporeo Transitional pores are often not well developed for carbons which are;used for gas adsorption since these molecules can be adsorbed in microporeso But large molecules can not enter the micropores; thus carbons used to adsorb large molecules usually have well-developed transitional poreso Carbonization and activation temperatures are significant considerations in the structural development of active carbonso Riley (19) has demonstrated that the size of the microcrystallites of carbons prepared from carbohydrates increases with increasing temperature of carbonization. In addition, the mechanical strength and electrical conductivity of a carbon increase when prepared at temperatures above 7000 Co The temperature at which activation is effected is important also because carbonization continues during this process via the mechanisms of actual removal, cracking, or cyclization of functional groups present at the edges of the microcrystallite (74) CHEMICAL NATURE OF THE ACTIVE CARBON SURFACE If we proceed on the basis of the structure outlined above, and neglect for the moment the effects of noncarbonaceous impurities, two major types of surfaces can be postulated for

14 active carbon: the basal planes of the microcrystallites and the edges of the carbon planes making up the sides of the microcrystallite. The basal planes are probably relatively uniform in natureo No attached functional groups should be present because the electrons of the carbon atoms are involved in covalent bonding with neigi-hboring carbon atoms Studies by Boehm (6) for example, show:ed that oxygen could only be chemisorbed on the periphery of the crystallite and not on the basal planeso Because oxidation occurs much more readily at the edge of the carbon plane than on the surface of the plane (35), few vacancies owing to oxidation during the activation process are expected in the basal planeo The majority of the sorptive processes occurring on this surface would be due then to the relatively weak van der Waals forces, although electrons are available for 7T-bonding if the sorbate is capable of entering into such a bond (24). Because most of the surface area of a carbon particle is found in the microporesp and because micropores are formed by the "eating out" of microcrystallite planes (74), the larger percentage of the total surface area is probably of the basal plane typeo The sides of the microcrystallites are more heterogeneous than the basal planes, and are characterized by various types of functional groups and vacancies owing to the action of oxidizing gaseso Corners or raised positions on the surface have the greatest tendency to participate in the electronsharing reactions which characterize chemisorption (75)

15 Sorption in depressions is probably physical in nature, but the bonds formed in this case should be much stronger than the bonds formed in physical sorption on the basal planes of the microcrystallite Although surfaces of the preceding types can be expected for a very pure active carbon, many carbons have large percentages of oxygen, hydrogen, and inorganic asho The concept of the type of surface available for sorption must be modified sommewhat when such substances are presento The presence of hydrogen on the surface of active carbon, for example lends quite definite properties to that surface. Certain sorption reactions are known to be more or less specific for surfaces comprised of hydrogen siteso Kipling and Shooter (42) reported that iodine vapor adsorption on Spheron 6 can be interpreted in terms of adsorption, on that part of the surface covered with hydrogen sites, to the exclusion of those portions of the surface covered with oxygen-containing functional groupso Hydrogen presumably exists either in the form of terminal groups on the fused aromatic planes of the microcrystallite or as part of the hydrocarbon functional groups attached to these planes (74)o Wolff (73) reported that hydrogen is present in amounts ranging from eight to 19 times that of oxygen, on a molar basis, for carbons containing 0,94 to 2o25% oxygen by weight Inorganic matter is present in most commercial active carbons to a very significant extent0 and its presence has

16 many noticeable effects (32, 73)o The total amount present consists of those inorganic constituents originally present in the starting material as well as various inorganic salts added to catalyze the carbonization and activation processes or to impart special sorptive properties such as selectivity to the active carbon. The fact that strong acid will remove almost all of the ash content of active carbon suggests that the inorganic matter exists primarily on; the microcrystallite surfaces (5, 73)o The special properties which the inorganic salts develop possibly are due to their effect on the active carbon surface and to their interactions with the sorbate or other solution constituents, Sorption on a surface covered with tightly held inorganic species would very likely differ in nature from sorption on an active carbon surface free of such specieso Pore structure would also be affected differently by different types of salts used to catalyze the activation process. Possible interactions between the inorganic salts on the active carbon surface and the sorbate or other solution constituents include complek formation, ionpair formation, precipitation reactions, and oxidation-reduction reactionso Basic studies of the sorptive properties of active carbon should include a careful study of the effect of the active carbon's inorganic constituentso Electron spin resonance studies of -ctive carbon have suggested free radical structures or structures with unpaired electrons in the microcrystallite (30)o According to Ingram (37), large number of unpaired electrons are trapped during

17 the carbonization period owing to bond breakage at the edges of the planar structureso These electrons are undoubtedly the "free" valences to which Boehm (6) refers and which are very reactive. Studies by Harker et alo (30) have indicated that oxygen interacts with these electrons to form oxygen complexes on the surface, The only means by which this oxygen can be removed is as oxides of carbon by degassing at high temperature, Oxygen, one of the major noncarbon constituents of active carbon, constitutes 2 to 25% by weight of this material, the exact amount present depending upon the temperature and method of activation (74)o The temperature of activation ranges from 4000 to 12000 Co, and the oxygen content decreases with increasing temperature of activation (20), Heating carbon at a temperature of 10000 to 12000 Co in a high vacuum removes nearly all of the oxygen from the carbon surface, usually as oxides of carbono For this reason oxygen is thought to exist in structures on the surface of the microcrystallite or between carbon planes near the surface of the microcrystallite (19) The presence of oxygen complexes on the surface of the carbon affects the sorptive properties of this material in that they tend to increase the polarity of the surface (4), In aqueous solution, therefore, a stronger solvent-active carbon bond must be broken before sorption of the sorbate can occur if oxygen complexes exist at the carbon surfacee Excluding any specific oxygen complex-sorbate interaction, increased

18 polarity decreases the quantity of nonpolar sorbate removed from solution by the carbon. For a polar sorbate the oxygen complex-sorbate interaction usually is stronger; compensating0 at least in part, for the added energy required to desorb water, Kipling (41) ciscovered that oxygen on the surface of carbon affects the specificity of sorption from an organic, binary liquid solutioiro As indicated previously, Kipling and Shooter (42) reported that iodine apparently does not absorb on that portion of the active carbon surface covered with oxygen sites, but does absorb on the part of the surface ~covered with hydrogen sites. These authors also indicate that the normally strong lateral interactions of absorbed iodine molecules appear to be lessened in the presence of, the oxygen$ thus preventing attainment of the extent of uptake realized in the absence of surface oxygeno An additional consideration is that sorbates which have a natural tendency to combine with oxygen probably sorb more easily on oxygenated surfaces than they do on non-oxygenated surfaces Oxygen can be added to an active carbon in one or more of three principal wayso If the starting material contains oxygen, this may become incorporated into the microcrystallite surface structure during the manufacture of the carbon (19)o If the starting material is a carbohydrate, for example heating will cause condensation and cyclization of the carbohydrate with the inclusion of an oxygen atom in the ringed structure

19 as an ether oxygeno As a second possibility, oxygen can be sorbed on active carbon at room temperatureo Part of the oxygen added in this manner can be evacuated at room temperature-this phenomenon being known as "the reversible effect" — while another part requires ev.acuation at temperatures in excess of 10000 Co (30) The latter phenomenon, which has been studied by electron spin resonance methods0 is known as "the surface oxide -ffecto" Thirdly, oxygen can be chemisorbed on the surface of the active carbon during treatment with oxidizing agentso As previously indicated, the principal oxidizing gases in the activation process are air, steam, and carbon dioxideo These three gases form oxygen complexes at the surface of the carbon which can be removed -as carbon monoxide, carbon dioxide, and water vapor —only by degassing at very high temperatures (32, 54)o Treatment of active carbon with chemical oxidizing agents in aqueous solution also leads to the chemisorption of oxygen on the surface of the carbon (32, 52, 54) o CHARACTERIZATION OF OXYGEN SURFACE STRUCTURES Because of their numbers and their probable effects on the sorption process; oxygen complexes on the surface of carbon merit careful considerationo An excellent review of this topic has been given by Boehm (6); anyone contemp:ating further work in this area should consult this review, One of the principal methods for study of the nature of oxygen structures on carbon has been by simple

20 acidimetric-alkalimetric titrations, and major interest has focused on the very important variable of temperature of activation (228 56)o Garten and Weiss (22) have provided an excellent review on the findings of such research, and these may be summarized briefly as followsCarbon activatedAl at 4000 Ce (L-carbon) will sorb base but not acid from aqueous solution while carbon activated at 800~ to 1000~ Co (H-cxrbon) will sorb acid but not base, Carbon activated at intermediate temperatures will sorb both acid and baseo Capacities for both acid and base are usually in the milli-equivalent per gram range, but this value depends on the substances added to the carbon during the activation processO Similar to ion exchange resins with basic and acidic functional groups, respectively, H-carbons produce a basic pH while L-carbons yield an acidic solution pH (22)~ Since no acid or base can be extracted from the solution after contact with the carbon, the conclusion is made that surface structures are responsible for this phenomenono Electrokinetic studies have indicated that H-carbons exhibit a positive surface potential, as opposed to a negative surface potential for L-carbonse However, unless care is taken to cool H-carbons in an inert atmosphere after activation, formation of structures which change the potential from positive to negative does occuro The presence of inorganic constituents at tie surface may also affect the potential to the point of reversing its sign from that which might be expected on the basis of the activation temperature~

21 A negative potential for L-carbons is consistent with the presence of surface acids, while the existence of a positive potential for H-carbons requires further considerationo Garten and Weiss (20, 22) using various techniques for the study of L-carbons, have concluded that surface acids responsible for the sorption of alkali from aqueous solution consist of at least three fundamental types: phenol; n-lactone; and f-lactone —the type present in fluorescein and phthalein dyestuffs (23), The relative proportions of these three groups vary with the method of preparation of the active carbono In determining the total number of acidic groups present on L-carbons by alkalimetric titrations, Garten et al1 (23) noted that the titration curves closely resembled the type obtained for weak -acids and were comparable with that for titration of a phenolic functional group ion exchange resin. These observations point to the possibility of phenolic surface structures, as illustrated in Figure 11-2o By way of additional evidence, a certain proportion of the acidic groups have resisted methylation with diazomethane, but themethoxyl group thus introduced resists hydrolysis when treated with boiling dilute acid, This behavior is characteristic of the phenolic structure (23, 53)o Conversely, a portion of the total number of acidic groups have resisted methylation with diazomethane, while supplementary infrared spectroscopy studies have produced data consistent with the postulation that these groups are of the n-lactone type given

22 OH Figure 11-2.7 A Phenolic Surface Figure Stru2 A Phenolic Surfae (23) Structure (23)

23 in Figure 11-3, with an absorption band at 1710 cmo A third fraction of the acidic groups on L-carbons was methylated with diazomethane, but with the methoxyl group thus formed being hydrolyzable in dilute boiling acide (23) Carbon compounds which have phenolic groups in association with lactone groups, as in fluorescein or phthalein dyestuffs, act similarly, Figure II-4 gives illustration of the structure of such lactones, calle.d f-lactones, as well as some of the reactions of this group (23)o The infrared data referred to previously have also revealed an absorption peak at 1750 cm', consistent with the presence of f-lactoneso While simple alkalimetric titration has been used rather extensively for measurement of the total number of acid groups present on carbon, some questions have been raised as to the quantitativeness of this procedure0 For example, Studebaker (62), using sodium aminoethoxide in ethylene diamine as a titrant and comparing his results with those obtained by NaOHtitration of active carbon in aqueous solution, has discovered that alkalimetric titration determines only about 68% of the total acidity as measured by the aminoethoxide titrationo Quinone groups are also known to be present on the surface of active carbon (21, 22, 62) ) An H-carbon cooled in nitrogen gas after activation does not sorb.base, but reduction of the H-carbon with sodium borohydride develops base properties (22, 62), This observation can be explained in terms of quinone reduction to hydroquinoneo Hallum and Drushel (26) also found polarographic evidence for quinoneso Reduced quinones,

24 H c=k FLc H Figure II-3. A n-Lactone Surface Structure (23)

25 O O ^ COOMsie - COO-Na+ ^K CH2N2 2 Dil. 7'NaOH Acid o OH 0 0-CO2 0 -DO'O.^.J-^^3 ^^< ~~~~ ^COOH Figure II-4. Some Reactions of Fluorescein-Type Lactones (23)

26 or hydroquinones, might be expected to behave as phenols when treated with diazomethane and thus would be included in the phenolic fraction given by the method of analysis used by Garten and Weisso Although a carbon activated at a temperature of 8000 to 10000 C. will sorb acid, evacuation at the temperature of activation followed by- cooling in an inert atmosphere will so affect the carbon that it will not sorb acid from a deaerated solution until oxygen is admitted (20)o Indeed, an exponential relationship exists between the acid sorbed and oxygen partial pressure for the range of 0 to 20 mm of Hgo Kolthoff (44) discovered that hydrogen peroxide is released by carbon to a solution from which acid is adsorbed (21)o Garten and Weiss (20), on the basis of this phenomenon, propose the presence of chromene (benzpyran) groups on the H-carbon surface, As illustrated in Figure II-5, this structure contains an activated >CH2 or >CHR group which can react with a strong acid and oxygen (65.) There is much difficulty in establishing the stoichiometry of this reaction, however, because active carbon also catalyzes the breakdown of the hydrogen peroxide (20)0 The carbonium ion formed by acid reaction with the chromene group, as shown in Figure 11-5, will hydrolyze readily in water to form the-chromenol group illustrated in Figure II-6, -11 a weakly basic group having a dissociation constant of 10 (58)o The fact that the carbonium ion tends to associate -so strongly with a negative ion could account for the fact that sorbed acid is very difficult to wash from carbon (20)o Boehm (6)

27 1'c \,R Y\ x/R C jNH \^ C Cl OR HC1 _ 02 +H202 /0 H R Figure II-5. The Chromene Acid Reaction (20)

0 0 C C1~~~~~~~~~ ~ clOH oO H20 +HC1 Figure 11-6. Hydrolysis of the Carbonium Ion (20)

29 stated that although the chromene reaction hypothesis explains many phenomena related to basic oxides on carbon, the evidence for it is rather circumstantial, and a more direct proof would be desirable. In addition to the evidence for chemisorption and the chromene groups, there is also strong evidence that some physical sorption of acid takes place. Garten and Weiss (22) have shown that a portion of the sorbed acid on an acid-saturated carbon can be displaced by strongly adsorbed organic moleculeso The assumption is made that the physically sorbed acid can be displaced and the chemisorbed acid can not. By this means, an approximation can be made of the number of chromene groups present on the carbon surface. The relative proportion of the quantity of acid physically sorbed to that chemically sorbed again depends on the methods of preparation, especially the length of time the carbon is activatedo DISCUSSION Pore surfaces of active carbon appear to consist principally of either the basal or the edge surfaces of the microcrystalliteo The planar surfaces are relatively homogeneous, but the edge surfaces are heterogeneous, quite probably covered with oxygen- and hydrogen-containing functional groups0 Some of the oxygen complexes appear to be highl- reactive with certain inorganic species in aqueous solution, particularly acids and baseso This interaction can be especially significant for active carbons having appreciable quantities of

30 surface oxygen, since it probably results in changes in the nature of the pore surface. Ash content, which is appreciable for most commercial carbons, might also be expected to affect markedly the nature of the active carbon surface if the inorganic species are distributed over the internal surface of the poreso If the ash is localized in specific vacancies rather -than distributed over the pore interior" however, the sorptive properties are likely to be affected very littleo The purpose of this review of structure and surface chemistry of carbon has been to set forth some of the more signifiCant interfacial factors and mechanisms to be consid'ered in interpretation of sorption phenomena associated with active carbon in aqueous solutions and to provide a basis for further research aimed at more thorough delineation of sorption processeso It will be used to this end in the following chapterso

III REACTION OF THE HYDRATED PROTON WITH ACTIVE CARBON EXPERIMENTAL CARBONS The active carbon used for the majority of the experiments in this study was Columbia Activated Carbon, LC Grade, obtained from the National Carbon Company (13)o This carbon is a coconutshell carbon and was obtained in a size range which included particles passing a Uo S. Standard Sieve No. 20 and being retained on a No 40 sieve0 These particles were mechanically ground to a smaller size and were then carefully sieved to a size range which included particles passing a No0 50 sieve and being retained on a No, 60 sieve; the mean particle diameter for this size range is 273 micronse After sieving, the carbon was washed thoroughly with distilled water to remove dust and fines, and was then dried to a constant weight at 1050 Co, The inorganic content was 0,7% which is low for commercial carbons0 Primary reasons for choosing this carbon included its low ash content, its resistance to attrition in rapidly-stirred experimental reactors and the fact that several previous investigations to which the work relates were carried out using this same carbon (68, 69, 70, 71)o A coal-base carbon, Pittsburgh Activated Carbon obtained 31

32 from the Pittsburg Coke and Chemical Company (41), which was prepared in the same fashion and of the same mean particle size was also employed in several experiments. The maximum ash content for this material, as reported by the manufacturer, was 8%, This carbon, toot has been used in related previous investigations (39, 40, 67)o A pore size distribution was not available for the coconut-shell carbon, but, again, according to the manufacturer, approximately 55% of the intraparticle volume of the coal-base carbon was comprised of diameters between 15 and 20 Ao, and 33% between 20 and 500 A0 (51)o The coal-base carbon is designed primarily for adsorption from solution, and thus probably has a more highly developed transitional-pore surface area than does the coconut-shell carbon, which is intended primarily for application in gaseous systems (13, 51), EXPERIMENTAL SYSTEMS Rate-of-reaction studies for the hydronium ion-active carbon reaction were carried out utilizing both finite and infinite bath techniques. Test solutions were prepared at the desired ionic strength, temperature, and initial pHo These solutions were stirred rapidly with a motor-driven polythylene stirring blade, For each test, a carefully measured quantity of carbon was added in the dry form. The finite bath technique consisted of allowing the pH to vary and recording pH values as a function of time after carbon addition, All finite bath pH measurements were made with a Corning Model 12,

33 expanded-scale pH meter~ The infinite bath technique consisted of maintaining a constant pH throughout the reaction with a Sargent recording pH State As the reaction proceeded, the pH Stat continuously added and recorded the quantity of standard acid (0,1 N in this case) needed to maintain a constant pH, Stirring speeds in all cases were in excess of the minimum required to keep the carbon in suspension; separate experiments indicated independence of sorption rate on stirring speeds greater than this minimum which was about 500 RPM0 Except where specified, carbon dosages for all finite bath studies were lo5 grams per liters Equilibrium studies were performed to determine the extent of the hydronium ion-active carbon reaction. These studies were carried out in reaction vessels containing 1l5 liters of distilled water adjusted to the desired pH and ionic strength. Temperature control was provided by immersing the reactor in a water bath, except for tests at room temperature (25+3~C) At the start of each experiment a known quantity of carbon was added to the test solution, Stirring was again accomplished with a motor-driven polyethylene stirring blade at speed sufficient to keep the carbon in suspension at all times, After 2 to 3 days (experimentally determined as the time required for the system to come to equilibrium) the pH was measured and recorded, and an additional quantity of standard acid was added; 2 to 3 days later the equilibrium pH was again measured and another quantity of acid was added6 This procedure was

34 repeated several times, usually over a period of from 2 to 4 weeks, to provide a series of adsorption capacities with increasing equilibrium H30+ activity, Distilled water was added to the solutions periodically to correct the volume for evaporation, When evaporation was small, it was necessary to correct the data for increased volume owing to addition of acid, The data obtained could then be reduced to moles of acid reacted per gram:If carbon at a given hydronium-ion activity, using the Extended DeBye-Huckel Law for calculating activity coefficients, The activity coefficient for the 1 M NaCl solution, Y = 0~75, was obtained from Harned and Owen (31) Freundlich parameters were obtained from log-log plots of the experimental data6 Except when otherwise specified, the data reported in this paper have been obtained from experiments at 250C, ASH CONTENT ANALYSIS The ash content of an active carbon is an important consideration in adsorption studies6 Blackburn and Kipling (5) have demonstrated some of the effects of ash contest in the adsorption process6 To assess the effects of ash on the interaction of strong acid with active carbon, separate quantities of the experimental coconut carbon were washed with 1+1 hydrochloric acid and with 1+1 glacial acetic acid to reduce the ash content of the carbono Thirty- to forty-gram samples of carbon were shaken with the acids for about 5 days- The carbons were then washed continuously with distilled water for a

35 period of 3 months until the carbon could be contacted with distilled water for a few days without significantly reducing the pH. The carbons were then dried at 1050 C to a constant weight. The ash content was measured by burning a known weight of the carbon at 7000 C and weighing the residue, The ash content of the carbon washed with acetic acid was reduced from 0O7% to 0.6%, while that washed with hydrochloric acid was reduced from Oo7% to 0,3%o Adsorption studies could then be carried out on the treated and untreated carbon for purposes of comparison. STOICHIOMETRY STUDIES The HCl-NaCl system was studied to determine if C1 ion was removed stoichiometrically with H30+ as the reaction occurred, Two 1-liter solutions were prepared at PCH+ = 3o00 and 10-3 M NaCl; two 1-liter solutionsat PC + = 3o00 and 2 x 10 3 M NaCl; and two l-liter solutions at pC + = 2o70 and 10 M NaClh A five-gram quantity of coconut-shell carbon was added to each of the PCH+ = 3o00 solutions, and a ten-gram quantity to each of the pC H+ = 2o70 solutionso After one day of stirring, the H residual H30 concentration was measured using a pH meter and correcting for activity, and the residual C1- ion concentration was determined by means of the Mercuric Nitrate Method (59). The percent stoichiometry was then calculated from the data obtained~

36 RESULTS AND DISCUSSION StATE-CONTROLLING STEP The rate-limiting step for the hydronium ion-active carbon reaction appears to be intraparticle transport. From a phenomenological point of view, both the hydronium ion and the conjugate anion of the acid added are removed stoichimetrically. The studies made on the HCl-NaC1 systems described above show that the ration of C1l ion removed from solution to H30+ ion removed is in the range of 0o93il to 1:1o This corresponds to similar findings by Carr et alo (11) and Miller (16, 47). The data indicates that the electroneutrality requirement for sorption is satisfied in this reactiono Since the hydronium ion and anion are removed from solution stoichiometrically, it seems reasonable to expectthat they would also diffuse through the pore in pairs with the anion limiting the rate of diffusion0 The diffusion coefficients for different acids, calculated assuming intraparticle transport as ratelimited, are on the order of those expected for the anion0 thus giving support to the assumptiono Additionallyp the anion would probably be rate limiting because the aqueous solution conductivity is much less for an anion than for the hydronium ion0 Helfferich (34) states that film diffusion; the other likely possibility for being the rate-limiting ste, will control only under extreme conditionso A study of sorption rate vso stirring speed showed no increase in rate for stirring speeds above that required to keep the carbon in suspensiono Other evidence in

37 support of intraparticle transport as the rate controlling mechanism is the fact that the experimental data are described well by a diffusion model as will be illustrated shortlyo An activation engery of -1l8 to -2 5 kcal/mole -WK also discussed in more detail in a later section of this paper; falls in the range expected for a diffusion-controlled process (64) DIFFUSION MODEL Assuming that pore diffusion is rate limitinge a diffusion model based on Fick's second law can be utilized for calculation of diffusion coefficients from the experimental data. The model must take account of the simultaneous diffusion-reaction processo If the sorption reaction is not taken into account, the calculated diffusion coefficients will deviate considerably from the actual valueso Such a model has been developed by Crank (14) —and utilized later by Weber and Rumer (71) —by modifying Fick's second law to include a term which accounts for non-linear sorptiono The general form is ax = a J r2D a x as III-1 at i2 r ar) at Equation III-1 is given in spherical coordinates, thus assuming a spherical shape for the carbon particle, an assumption which accords reasonably well with microscopic observations of the geometry of particles of the experimental carbono In Equation III-1, X represents the H30+ activity in solution; t, time, r, the radial distance from the particle center; D, the diffusion coefficient; and S, the H30+ concentration at the surface of the carbono For the present experiments, the equilibrium

38 relationship between S and X is described in terms of the Freundlich expression S = FXN N>1o0 III-2 Equation 1 is subject to the boundary conditions X = 0 t = 0 for 0 < r ~ a and ra A(Xo-Xa) = 47TJ (X+ S) r2dr III-3 7 J 07 where A represents the volume of solution "served" by each particle; X0, the initial H30+ activity in bulk solution; Xap the H30I+ activity in bulk solution at any time; and a, the radius of the spherical particle. Equation III-1 is a non-linear, partial differential equation which can be solved by a finite difference technique after Crank (14) The solution is made easier by expressing the equation in terms of the dimensionless parameters: x X, T= Dt, P=r s S X a a X2 Before S can be divided by Xo to give the dimensionless quantity, s, it must be converted from units of moles per gram to moles per liter by multiplying by the specific weight of the carbono The specific weight for the 273-U coal-base carbon is 0,750 grams per cc as given by the manufacturer (51), and that for the coconut-shell carbon is assumed to be the same. The equation requiring solution, in dimensionless form, is then ax= 1 ia 2a) - ~s III-4 0a p2 rp apI

39 with the boundary conditions, x = 0 T= O for O pa 1 1 A (1 -x) = 4 (x + s) p2dp III-5 AYr^ J0o 7 and N s = F x N III-6 where F, the dimensionless Freundlich parameters is given by F = F x spo wto of carbon x XN-1 Assuming a spherical particle and using the specific weight, the term A can be calculated for the finite-bath systems. For the 15 grams per liter dosage used in these studies, A = 0,00533 cm3 for both the coconut-shell and the coal-base carbonso An IBM 7090 digital computer was used to calculate all rate curves; the program is given in Appendix Io Calculated curves for the finite bath studies were printed out in the form Xa/XO vs, T o These curves were then compared with the experimental Xa/Xo vso t curves, and D was calculated by a trial and error procedure, This approach was modified slightly in order to calculate the rate curves for the infinite bath studieso The value for "A" taken for calculation purposes, in this case 1000 cm3 was very much greater than the actual value, thus producing the effect of an essentially constant Xa with time, In this case, the integral value

40 a 4 T (X + S) r2dr III-7 which represents the total amount of solute in one carbon particle, was printed out as a function of the dimensionless time parameter, T This value, when multiplied by the total number of particles present in the solution volume, represents the number of moles of acid added to the solution to maintain a constant pHo By comparing the calculated data to the experimental curves for H30+ added vs0 time, the diffusion coefficient could be calculated by a trial and error procedureo Carbon dosages for the infinite bath experiments were 1.0 to 2.0 grams per litero The basic assumptions made in developing the model should be kept in mind when analyzing the resultso Sorption is assumed to be occurring at a rate much faster than pore diffusion, while pore diffusion is taken to be slower than film diffusion. The model assumes radial, flux only, and a homogeneous particle. Thus no differentiation is made between pore volume and solid carbon. Also', the assumption of spherical geometry neglects the fact that the particle surface is extremely irregular, The diffusion coefficient, D, is also assumed constant with varying acid concentration0 The inclusion of the non-linear isotherm effect requires that the isoth er*. parameters be known in order to calculate the rate curvesa Thus, for each system analyzed, both rate and equilibrium data were required.

41 EFFECT OF INITIAL HYDRONIUM-ION ACTIVITY The isotherm shown in Figure III-1 was determined by -2 the procedure described, for the coconut carbon in a 10 M NaCl system at 250Co The data shown area composite of several runs, The Freundlich equation, determined from a log-log plot of this data, is S = 6.5 x 10-4Xl27, The Freundlich equation was used because the log-log plot of the data was linear, Finite and infinite bath rate data were collected for the same system for different initial hydronium ion activities, Figure III-2 presents the calculated and experimental rate curves for the finite bath experiments and Figure III-3 those obtained by the infinite bath methodo Table III-1 is a summary of the results0 TABLE III-1 Diffusion Coefficients as a Function of Initial pH For the HCl-NaCl System 102 M NaCI, 250, S=6,5x10OX127 Diffusion Diffusion Coefficient, D, Coefficient, D, Initial pH 2 7 2 (HC1) (cm /sec)xl07 (cm2/sec) x07 (Finite Bath) (Infinite Bath) 3c50 6075 5o9 3.70 9o50 7o8 4000 13o00 8,6

dt 3 t J O 4 _ 0 r__ X-4.127 S =6.5x10 X En 2 0Q O....so -- c Coal-base carbon ]_~~1 /f^ A1~X-~ Coconut-shell carbon ia~~~~~~~~ _____ Calculated curve from Freundlich equation 0 __________ 0 4 8 12 16 20 24 28 X, H30 Activity, x 105 Figure III-1. Adsorption Isotherms for HC1

9..9b ~~~~~~\\ _'~ ~calculated curve.8 O —-\O experimental data.7 V'^^ ^-^ PH=3.50 7.6 \0 ~SSD^^>^^^ D=6 7 5xl0 cm Xo.5.4 -~ ^^^^^ —i D:=9'5X10 cm pH _ spH=370 ec |<2.~~ 2.00__~~~ 7D=9,15x10 cm pH 0 sec L,2~~~~~ ~D=13x10 cm sec.1 ____ 0 3.0 6,0 9.0 12.0 15.0 18.0 21.0 24.0 T = Dt a2 Figure III-2. Finite-Bath Rates of Adsorption for HC1

l I i' I 1 1! 16 16 -; pH=3.50 D=5.9xl-7 pH=3.70 12 O 8. X xlO_ pH —4. 0 0 _- ~| d0 6 7 - ^^" ___ calculated curve m 4~^ Jb^ ^t o__ ~ [ |experimental data 0 3 6 9 12 15 18 21 24 7= Dt a2 Figure III-3. Infinite-Bath Rates of Adsorption for HC1

45 The magnitude of the diffusion coefficients given in Table III-1 can be compared with a value of 3~3 x 10 cm2/sec determined experimentally by Stokes (61) for HC1 in bulk solution at infinite dilution~ The pore diffusion coefficients listed in Table III-1 for HC1 range from about 1/30 to 1/60 of that given by Stokes. McNeill and Weiss (45) have indicated that active carbon can be considered as a weak-base anion-exchange sorbentc According to Helfferich (34) diffusion coefficients in such resins can be several orders of magnitude less than the corresponding bulk solution coefficients, The C1 ion probably limits the rate of diffusion, since its mobility in aqueous solution is much less than that of the H30O iono Further evidence to support this conclusion has been obtained in the present work from determinations of pore diffusion coefficients for several different strong acids; these will be discussed shortly (see Table III-4) The large difference between the bulk solution coefficients and the pore coefficients may be attributable to one or more of a number of factors, the most likely of which is the size and geometry of the pores of the active carbon0 The smaller the pore, the more likely is the diffusing ion to interact with internal surfaces, both physically and electrostatically (34)o The relative effect of pore size is indicated by comparative studies carried out with 273-micron coal-base carbono Although there was no pore size distribution data available for the coconut carbon, the average size of the

46 micropores of this carbon can be assumed to be considerably smaller than that of the coal-base carbon, since the former was designed for adsorption from gas phase while the latter was designed for adsorption from solutiono The solution used for this equilibrium study had an initial pH of 3o50 (HC1) e =2 was 10 M in NaCl, and contained lo33 grams of the coalbase carbon per liter, the study was made at 25~Co The isotherm for this system is shown in Figure III-1, along with the isotherm for the coconut carbono Kinetic measurements were made on the same system using the finite-bath technique, and the calculated and experimental rate curves are shown in Figure III-4o The corresponding rate curve for the coconut carbon with this system is included in Figure III-40 The calculated diffusion coefficient for HC1 in the coal-base carbon is 1l4 x 105 cm /sec, which is nearly equal to that for HC1 in bulk solutiono Although many other factors might contribute to the large difference between the two carbons, relative pore size is probably a major one' Other factors which may also be responsible to some extent for the small coefficient obtained for the coconut carbon include the approximation of spherical particle geometry, the assumption of an isotropic medium. and the assumption of a radial diffusion path (34) o Specific chemical interaction between the chloride ion and metal ions present at the pore surfaces has also been considered as a possible factor contributing to retardation of

1.0 9 calculated curve.8 \ \ experimental data.7.6 X Xo.5 __ |'^^D -O~~-^-_ coconut-shell carbon 4 -7 2'*4 -'^^^ D=6.75x!0- cm sec.3.2 coal-base carbon -5 2 D=1.4xl0 cm5.1 1 0 6 12 18 24 30 36 42 48 T= Dt a2 Figure III-4. Finite-Bath Rates of Adsorption for HC1 on Different Carbons

48 HC1 diffusion, To evaluate this possibility, one portion of the coconut carbon was washed with acetic acid to reduce its ash content from O07% to 0i6%, and another portion with hydrochloric acid to reduce its ash content from 0,7% to 0,3% The finite bath technique was used to study these two carbons in otherwise identical systems consisting of 10- M NaCl, 1o5 grams carbon per liter, an initial pH of 3O50 (HC1) at a temperature of 25 Co The corresponding isotherms show no significant difference between these carbons and the untreated coconut carbono The experimental rate data and calculated curves for the two systems are plotted in Figure III-5, Because there was found to be no essential difference between the calculated curves for the acid treated and the untreated carbon, only one curve is shown, The experimental rate data for the acetic-acid-treated carbon are not described as well as are those for the hydrochloric-acid-treated carbon. nor as well as those for the untreated carbon fitted by the calculated curve shown in Figure III-lo The diffusion coefficient derived from the curve of best fit for the acetic-acid washed carbon -7 2 is 9 x 10 cm /sec, and that for the hydrocholoric-acidwashed carbon is 9o5 x 10-7 cm2/sec; this compares with 6075 x -7 2 10 cm /sec for the untreated carbon, The increase in the diffusion coefficient for the treated carbon over that for the untreated carbon indicates that interaction of the HC1 with the inorganic content of the carbon could be responsible to some extent for retarding diffusion, Analysis has shown that the ash is approximately 50% iron, which in ionic form tends

1.0.9.7 _ *6 8 Xo 5 calculated curve.4 O_ 0- CH3COOH Treated - 4 ~ D=9xl0-7 cm2 sec.3 HC1 Treated D=9.5x107 cm sec.2.1 I II| 0 3 6 9 12 15 18 21 24 T= Dt a2 Figure III-5. The Effect of Ash Removal on Finite-Bath Adsorption Rate

50 to form complexes with the C1 iono Diffusion coefficients for different initial pH calculated from data obtained by the finite bath technique compare favorably with those determined by the infinite bath technique as can be seen in Table III-l. The calculated rate curves fit the experimental data very well for initial pH values of 3a50 and 3.70, but for initial pH of 4o00, finite bath case, the fit is not as goodo For the latter case, the range of the diffusion -7 2 coefficient is 6-14x10 cm /sec for all of the experimental data to fall on the calculated curve Another effect shown by the data is that the diffusion coefficient increases with increasing initial pH, indicating that the coefficient is a function of activity. This is contrary to the assumption of a constant diffusion coefficient built into the rate modelo However, except for the case of an initial pH of 400, finite bath case, the effect is apparently not very great for the activity change occurring in a given rate study for those shown in Figure III-2; otherwise the experimental data would not be described as well by the calculated curveso The decrease in diffusion coefficient with increasing H30+ activity is probably due to hydration effectso The increased number of hydrated H30+ and C1- ions in the pore decreases the amount of free solvent available for diffusion of the ions, thus decreasing effective pore size and increasing retardation effects due to interaction of the diffusing species with the carbon framework0 A decrease in the diffusion coefficient is thereby brought about (34)

51 SALT EFFECTS The HCl-NaCl system has been studied further to determine the effect of NaCi concentration on the acid-carbon reaction, The carbon dosage was varied from 0O33 to 2o0 grams of 273-micron coconut carbon per liter for the equilibrium studies, and the temperature was 25 0Co NaCl concentration was studied over a range from 5x10-3 to 160 Me The effect of salt concentration on the equilibrium position of the reaction is shown in Figures III-6 and 1II-9o Typical rate data, along with the corresponding calculated curve, are shown in Figure -2 III-7 for a 2x10 M concentration of NaClo The finite bath method was used for these rate studies. A summary of the results is given in Table III-2o It should be noted that the range of the Freundlich parameters used for the rate curve calculations produced very slight changes in the shapes of the resulting curveso TABLE III-2 The Effect of NaCl Concentration on Diffusion of HC1 Initial pH =.350, Temperature = 25~ Freundlich NaCl D Parameters Concentration, (moles/liter) xl02 (cm2/sec) x07 F. x104N ______ 5,6 0,125 0,5 4o00 6.5 0,127 1.0 6o75 6~9 0o133 2o0 7650 7ol 0O120 0lo0 1510 7~6 0 o0105 100 O068600

3 0 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~ i X f 4 - F H~~~~~~~ 2l 0~~~~~~~~ se (A3^^"^ " ~~ ~~~V ^n3!^ / rri ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~5l M~~~~~~~~~~~~~~x. jili, _J~^ v xu _ IfT"^?^"^ ~~~~~~~~~~~~~~~~NaC1, M cny~~~ o-o5x10~~~ F2x1o -3 T —-y 5x10 0 4 8 12 16 20 24 28 X, H30+ Activity, x105 Figure III-6. The Effect of Salt Concentration on Adsorption Capacity for HC1

1.0.9 —.8 ~.7.6 0 0 Xo I 7 9 -. 2 D=7.5x10-7 cm2.4 sec.3 calculated curve.2 O —0 experimental data.1 LI I! I J..._ l...I 0 3 6 9 12 15 18 21 24 T= Dt a2 Figure III-7. Finite-Bath Rate of Adsorption for HC1 at 0.02 M NaCl

54 Similar studies have been carried out on the HClO4-NaClO4 system. All conditions were the same as those for the HC1-NaC1 tests, except that the carbon dosage was 1o33 grams per liter for the equilibrium study and the concentration of NaClO4 ranged from 5x10 3 to 10-1 Mo The effect on the extent of reaction is shown in Figures III-8 and III-9, and typical rate data along with the corresponding calculated curve for a concentration of 10- M NaC104 are given in Figure III-lO The results are summarized in Table III-3o TABLE III-3 The Effect of NaC104 Concentration on Diffusion of HC104 Initial pH = 3~50, Temperature = 250 i Freundlich NaC104 D, Parameters Concentration - F. x14 N(moles/liter)x102 (cm2/sec)x106 F,xlO N_______________________ 7,6 0O1050 0,5 4~7 802 0,1030 1o0 5~3 8.4 Ool000 2~0 680 10.0 0o1065 10o0 100l The equilibrium studies were performed with only sodium salts. Rate studies, however, were performed with LiCl and KC1 as well as with NaCl, and no significant difference was found. There is a marked increase in the quantity of acid sorbed with increasing salt concentration for both the HCl-NaCl and

4 r1Ixlg) M2xl0 M ^ n/^ NaC104, M~~~2x11" Nt 0 2 A__A lx(m -3 3 ro m 1x10 1 -h'~ 3 5xl03 - ____ calculated curve NaClPO mJl -~~~~~~~~~~z 4~~~~~~~~~~~4 M 2 0 I-2 I 04 8 12 16 20 24 28 + 5X10 -3 czalculated curveary 0 4 8 12 16 20 24 28 X, H30 Activity, x105 Figure III-8. The Effect of Salt Concentration on Adsorption Capacity for HC104

56 I —---------------— 4 1 |~~~~~i I4.4 Xeq =2.4x10 4.0 HC104- NaC104 3.6 3,2 HC 1- NaC l ^ /U L // 3 2 1 2.8 HC1- NaCl 2.4 2.0 -log Salt Concentration Figure III-9. The Effect of Salt Concentration on Adsorption Capacity for Acid

1.0 _____________________________________________ -2.-~~9 -\ 1~~~Ix10 M NaC104.8 \ *i calculated curve 0_ —O experimental data.7.6 - X.5 - ^ ^^-^.O D=5.3xl-6 CM X.5 0 Xo 0 -- O.4.3.2.1 Ii 0 3 6 9 12 15 18 21 24 7 Dt Figure III-10. Finite-Bath Rate of Adsorption for HC104

58 the HC104-NaClO4 systems, A two-hundred-fold increase in NaCl concentration produces a 60% increase in HC1 sorption at pH 3,40, while a twenty-fold increase in NaC104 concentration causes a 30% increase in HC104 sorption at the same pHo This observation is consistent with a physical sorption modelO Electrophoretic mobility measurements have shown that the active carbon has a negative surface potential~ It is possible, therefore, that the proton is primarily adsorbed while the anion is secondarily adsorbed in the double layero If the proton is adsorbed readily but held back because the anion is not easily taken into the double layer, then an increase in salt concentration would have the effect of increasing the anion pressure, and more acid would tend to be adsorbedo The effect is consistent with that noted by Steenberg (22), and Parks and Bartlett (50) Indeed, Boehm (6) found evidence that physical sorption of acid took place on the basal planes of the microcrystalliteo Other sorption mechanisms can not be ruled out, however, because reverisble chemisorption could show a similar effect within a given range of salt concentrationso Also, pores which might be inaccessable at low salt concentrations could become of importance at the higher anion pressures0 The increase in the diffusion coefficient for the HC1NaCl system with increased chloride concentration is especially noteworthy0 This effect best can be explained by the fact that the driving force for the rate-limiting C1- ion is increased, with a resulting increase in the HC1 fluxo Since the effect of anion concentration is not built into the mathematical

59 diffusion model, higher diffusion coefficients results The reason for the effect being greater for the HCl-NaCi system than for the HCO04-NaClO4 system may be related to relative ionic size and to the fact that the diffusion coefficient for the HC1 is much further from its bulk solution value than is that for HC104o Because of hydration effects, the C1l ion is larger in solution than is the C104 ion (12) As a result, there are probably many more retardation effects due to ionic size which can be overcome by the increased in Cl" ion activity, The fact that the diffusion coefficient for HCO1 is ten times 4 as large as that for HC1 could also be due to the relative sizes e The much higher capacity of the active carbon for HC104 then for HC1 has been observed by others for ion exchange resins (12), The observation is in keeping with the smaller hydrated radius of the ClO4 ion as noted aboveo Chu et al, (12) have claimed that an additional effect derives from the inability of the C104 ion to orient surrounding water molecules in bulk solution as effectively as does the Cl~ ion, The water molecules adjacent to the ClO4 ion are attracted more strongly to other water molecules than to the C104-; thus a disrupture zone is created in the structure of the bulk-solution watero Because the structure of the water in the resin (active carbon) is already highly disrupted by the charged internal surfaces, the bulk-phase water tends to "push" the C104 ion into the resin0 The Cl ion, however, tends to be held back in bulk phase because its attraction for the water is stronger than

60 is that of the COI ion, EFFECTS OF THE TYPE OF ACID Additional kinetic and equilibrium studies have been performed with HN03 and H2SO40 The results of the equilibrium studies are given in Figure III-11; comparable isotherms for HC1 and HC104 are shown also for comparison, These studies were carried out at 2'~C in 10-2 N solutions of the sodium salt of the conjugate base, Carbon dosages were again 1,33 grams of 273-micron coconut carbon per liter for the equilibrium studieso Typical rate data for HN03 in a 10-2 N NaNO3 solution are shown in Figure III-12, Table III-4 is a tabulation of the results TABLE III-4 Diffusion Coefficients for Different Strong Acids Initial pH = 3750 2- __ —--- Sodium Salt Concentration = 10 N _____ Temperature = 250~C Freund l3. C h Parameters D Acid Fx104 N (cm2/sec)xl107 H2SO4 14o5 1970 2o00 HC1 6o5 o1270 6o75 HN03 6,6 1065 24~50 HC104 82.1030 53,00 The equilibrium curve for the HN03 falls between those for the HC1 and HC1040 as expected on the basis of the explanation given previously for the difference between HCl and HC104 capacities6 The N03- ion would be less hydrated than

4 HCl04-NaC104 r 3 - H2S04-Na2S04 -- ^ A fiji I --— ~~~~~~-i-'- R03 -NN03 w 2 oQHC 1 -NaC 1 calculated curve HNO3- NaN03. — A H2S04-Na2S04 0 I I I I I 1 0 4 8 12 16 20 24 28 + 5 X, H30 Activity, x10 Figure III-11. Adsorption Isotherms for Different Strong Acids

1.0......9 lx10-2 M NaN03.8 calculated curve,7 L a O —-Q experimental data.6 x Xo.5.4 6 ~ - D=2.45x10 cm2 sec.3.2.1... I........, I I. I I 3 6 9 12 15 18 21 24 T= Dt aa Figure III-12. Finite-Bath Rate of Adsorption for HNO3

63 the Cl ion but more hydrated than the C104 ion. The disruptive effect on the bulk water structure would thus be expected to be intermediate, and the capacity sequence to be HC104:HN03 >HCl, as observedo This corresponds to the affinity scale given by Kitchner for ion exchange resins (43)~ Because of the valence effect on the interaction of the S042 ion with the active carbon and on the relative extent of hydration, prediction of 12S04 capacity relative to the monoprotonic acids is difficult. It is interesting to note that the active carbon has a higher capacity for HC104 than for H2S04. Thus it is probably true that the more extensive hydration of the SO42- prevents it from entering some of the smaller pores. The diffusion coefficient for HN03 is intermediate between those for HC1 and HClO4, in keeping with the relative effects of hydration and disruption of water structure0 The diffusion coefficient for HSO04, however, is lower than those of the monoprotonic acids, According to Helfferich (34), the change in the magnitude of the diffusion coefficient between bulk solution and resin phase is much greater for divalent ions than for monovalent ions. This effect could be the result of a combination of increased electrostatic interaction with internal surfaces and of increased physical resistance attributable to the size of the hydrated ion. The sequence of the diffusion coefficients, S042-(2x10-7)< Cl-(6o75x10-7)" NO3-(24.5x107)<Cl04-(53x10-7), compares favorably with that found by Soldano and Boyd (57, 60),

63 the C1 ion but more hydrated than the C104' ion. The disruptive effect on the bulk water structure would thus be expected to be intermediate, and the capacity sequence to be HC104 >HN03 - HC1, as observed. This corresponds to the affinity scale given by Kitchner for ion exchange resins (43)~ Because of the valence effect on the interaction of the S042 ion with the active carbon and on the relative extent of hydration, prediction of H2S04 capacity relative to the monoprotonic acids is difficult0 It is interesting to note that the active carbon has a higher capacity for HC104 than for H2S04O Thus it is probably true that the more extensive hydration of the S042- prevents it from entering some of the smaller pores. The diffusion coefficient for HN03 is intermediate between those for HC1 and HC104, in keeping with the relative effects of hydration and disruption of water structuree The diffusion coefficient for HSO04, however, is lower than those of the monoprotonic acids. According to Helfferich (34), the change in the magnitude of the diffusion coefficient between bulk solution and resin phase is much greater for divalent ions than for monovalent ionso This effect could be the result of a combination of increased electrostatic interaction with internal surfaces and of increased physical resistance attributable to the size of the hydrated ion. The sequence of the diffusion coefficients, S042 (2x10-7)< Cl-(6.75x10-7)< NO3 (24.5x10-7)<C104-(53x107), compares favorably with that found by Soldano and Boyd (57, 60),

64 PO43 (0.57x10 7)<cWO4- (180x10-7)<Cl-(3,54x10-7)<Br03 (4.56xlO07) The hydrated radii probably follow the same sequence in each case. In addition, the diffusion coefficients for Cl are of nearly the same magnitude. TEMPERATURE EFFECTS 0-2 Equilibrium studies have been performed on the HC1, 10 M NaCl system at temperatures at 2~C, 25~, and 50 C. The results are shown in Figure III-13 and the equilibrium parameters are tabulated in Table III-5. The results of rate studies on the same systems at the three different temperatures are shown in Figures III-14 and III-15 for 2 C and 50~C, and in Figure III-2 for 25~C, while the results are tabulated in Table III-5, TABLE III-5 Diffusion Coefficients for the HCl-NaCl System at Different Temperatures Initial pH = 3.50, 10"2 M NaCl - Freundlich Parameters Temperature D ~C F,xl04 N (cm2/sec)x107 2 4.4 0.093 4.0 25 6.5 00127 6.75 50 9.3 0.17 8.50 The diffusion coefficients, as expected, increased with increasing temperature. Variation of the diffusion coefficient as a function of temperature can be expressed in terms of the

3.0 2.4 O 0 EQ 2 0 4 8 12 16 20 24 28 Figure II-13. The Effect of Temperuature on Adsorption Capacity for HC 0 4 8 12 16 20 24 28 + 5 X,H30 Activity, x10 Figure III-13. The Effect of Temperature on Adsorption Capacity, for HC1

1.0.9.8.7 0.6 O0 D=4x10-7 cm2 -- O — ^_sec X.5 Xo ~~0~'-" —--—._ o.4 ---- calculated curve.3 O --- experimental data.2 Temperature = 2~C *~~1,-.......I....I I...!.... I 0 3 6 9 12 15 18 21 24 T= Dt aaa Figure III-14. Finite-Bath Rate of Adsorption for HC1 at 2~C.

1.0 9o.8.7O.6 0 D=8.5Sx1O7 cm2 O sec Xo.5 --.4 calculated curve.3 0 —0 experimental data ^.2 -Temperature = 50~C..1 0 3 6 9 12 15 18 21 24 -T Dt a2 Figure III-15. Finite-Bath Rate of Adsorption for HC1 at 500C.

68 Arrhenius equation, which, in logarithmic form is log = E T2 - T1 III-8 Di 230R T1 T2 In Equation III-8, D is the diffusion coefficient in cm2/sec; Ea is the activation energy in cal/mole- K; R is the ideal gas constant (1987 cal/moie-~K); and, T is the temperature on the Kelvin scaleO A value of Ea = -(1 8 to 2,5) kcal/mole-~K, was calculated from the diffusion coefficients listed in Table III-5, in line with values expected for electrolyte diffusion (34) o The data shown in Figure III-13 for the equilibrium studijes suggest that the process goes from exothermic at low equilibrium activities to endothermic at high equilibrium activitieso The differences between isotherms is not very significant and without further research to verify the suggested trends, an elaborate explanation is not warrantedo Zettlemoyer and Narayan (76) have shown, though, that the heats of adsorption of electrical double layer formation are endothermico An endothermic reaction is indicated by the higher equilibrium activity data, and this indication can be taken in support of physical adsorption of acid on the carbon surfaceo SUMMARY Garten and Weiss (20) have postulated that strong acids react with chromene functional groups on the active carbon surface as shown in Figure II-5o This reaction can be used to explain some of the

69 observations of the present worko The carbonium ion has a high affinity for the OH ion, as illustrated by a pKb = 11 (58)a The carbonium salt will thus hydrolyze readily as given in Figure II-6c This reaction would account for the fact that carbon which had been contacted with 1+1 HC1 for five days could be washed continuously with distilled water over a period of three months to remove essentially all of the sorbed acid, This is illustrated by the fact that essentially the same isotherm was obtained for 1+1 HCl-treated carbon as for untreated carbon. The possibility that diffusion of oxygen into the pore is rate limiting has been ruled out because solutions used were exposed to the atmosphere, and oxygen was found to be rate limiting only when the partial pressure is less than 20 mm Hgo (8) It has been noticed also that only a small fraction of the total amount of sorbed acid could be removed by washing with water which contained the salt of the acid in the same concentration as was present in the solution with which the carbon was equilibrated originally, This observation suggests that the conjugate anion is tightly held, which is consistent with the findings of others (22, 48, 72) It would appear, however, that not all of the acid is taken up by means of the chromene reaction, since 50% to 60% of the acid sorbed at an equilibrium pH of approximately 3,5 can be displaced by adding phenol to the solution in sufficient quantity to bring the phenol concentration to approximately

70 0.1 M at equilibrium~ This percentage, at least, of the acid originally sorbed is evidently taken up as a displaceable entity Garten and Weiss (22) have show that the amount of acid which can be displaced by a strongly adsorbing organic molecule is dependent upon the extent of oxidation' of the active carbon during manufactureo The nature of reactions at the surface therefore depends upon the method of preparation It becomes difficult to iiiake any generalized conclusions regarding the displaceable character of the sorbed acidso It should be noted that some of the sorbed acid may be in pores which are not readily accessible to the organic molecule, and thus are not displaced by the latter, There is not sufficient evidence to conclusively prove or disprove one or two definite mechanisms for reaction of acid with active carbon, However, the chromene-acid reaction as herein described appears to be a logical and significant part of the overall reaction. Physical, electrostatics or other mechanisms for acid sorption probably account for the remained of the reactiono Regardless of the actual mechanism, the rate and extent of the acid reaction with active carbon are important factors, and the information obtained about these factors should prove useful for interpreting sorption data of various kinds0 In addition, systematic sorption studies should also lead ultimately to a better understanding of the nature of the acidcarbon reactiono

IV, ADSORPTION OF PHENOL AND 4-NITROPHENOL EXPERI MENTAL CARBON The active carbon used for these studies was the 273-El coconut-shell carbon# the same type as was used for the majority of the studies reported in Chapter III, The carbon was prepared in the same manner as described in Chapter IIIo SOLUTE PROPERTIES AND ANALYSIS The solutes used for these studies, phenol and 4nitrophenol (PNP) were selected because they are typical pollutants which can be studied as both neutral and negative species, and can be easily analyzed in aqueous solution6 Phenol is found in many wastes, especially steel mill wastes (49), and 4-nitrophenol is a typical pesticide (67)o The phenol used was obtained as analytical reagent grade in the crystalline form, Solution concentration analyses were performed on a Beckman DU Spectrophotometer at the wave length of maximum adsorption, 270 mV The molar adsorptivity at this wave length was experimentally determined to be 1540 liter/ moletcrm Since the pKa for phenol is 9o9 (8) all concentrations were measured at a pH below 6, thereby assuring that only the neutral species was present Glass-distilled water was used for all experiments and for preparation of all stock solutions0 71

72 The solubility limit of the neutral phenol species was determined to be about 08 M. The anion solubility was not determined, but it is very soluble, The 4-nitrophenol was obtained as reagent grade in powdered formo Spectrophotometric analyses for concentration were performed at a wave length of 226 rrSL The molar absorptivity at this wave length of peak absorbance is 6580 liters/molecme. This molar adsorptivity applies only to the neutral species; thus all analyses were performed at a pH less than 4, assuring that essentially all species were present in the neutral form since the pKa is 700 (8) The solubility of the neutral species is 0.1 M while that for the anionic species is 0,2 M, as determined experimentally, Glass-distilled water was again used for all experimentso EQUILIBRIA STUDIES Equilibria studies were performed by first preparing 3 liters of solution at the desired pH, salt concentration, solute concentration and temperature0 Then, 100 ml portions of this solution were pipeted into a 125-ml, French-Square, wide-mouth bottle which contained a carefully measured quantity of carbon0 The usual procedure was to fill 20 bottles in this manner, along with 4 bottles which contained no carbon for use as blanks0 The weights of carbon added to these bottles were carefully calculated to give a3 wide a range of equilibria concentrations as possible for a given initial concentrationo The bottles were then shaken on either a

73 reciprocating or a tumbling shaker for a period of 3 to 4 weeks which was experimentally determined as the time required for equilibrium. After the shaking period, the concentration of each bottle was analyzedo The points in the isotherm could then be calculated from the equation, S= (C-eq) V m IV-1 where C is the initial concentration in moles per liter and o is taken to be the average concentration of the 4 blanks, Ceq is the equilibrium concentration in moles per liter for a given bottle, V is the volume of solution contained in the bottle in liters, m is the weight of carbon contained in the bottle, and S is the solute removed by the carbon in moles per gram, S could then be plotted against Ceq to yield what is called the adsorption isotherm. If the data showed too much scatter, the procedure was repeated0 The final pH of each bottle was also measured and compared with the initial pH in order to determine if too large a pH change had occurred for the data to be valid, The data was discarded if a change greater than 0^2 of a pH unit occurred in the range of pH=4 to 10, and 0.1 of a unit for pH less than 4 and greater than 10 KINETIC STUDIES Rate studies were performed by preparing 3 liters of solution at the desired initial solute concentration, pH, salt concentration and temperatureo Temperature control of the system was accomplished with a water bath except at room temperature, The system was stirred with a motor-driven,

74 polyethylene stirring bladeo After carefully measuring the initial concentration of solute, Co, a known quantity of carbon was added to the system, The solute concentration was then measured several times with respect to time after addition of the carbono The stirring rate was such that the carbon was well-suspended at all times, The concentration of the system was analyzed by turning off the stirring motor to allow the carbon to settle just below the surface of the water, pipetting out the sample, diluting and making the necessary pH adjustment, and measuring the absorbance spectrophotometricallyo The stirring motor was turned off for 15 to 20 seconds while each sample was being removed, A study was carried out to determine the effect of this shut-off time on the rate curve, and it was found that essentially identical rate curves resulted for 2 systems, one of which was analyzed more times than is necessary to determine a good rate curve0 The pH of the system was analyzed after the test was over to determine if an excessive pH change had occurred during the run, and if it had, the data was discardedo The rate data taken in this manner was found to be very reproducible, The data was generally analyzed by means of a St vs0 t~ plot where St is given by Equation IV-1 with Ceq replaced by Ct, the concentration at time, t. RESULTS AND DISCUSSION PHENOL AND 4-N'ITROPHENOL ADSORPTION EQUILIBRIA The first system studied in depth consisted of distilled

75 water, phenol and carbon. The initial pH of this system was 5,6 and the deviation from neutrality was probably due to atmospheric C02 dissolved in the distilled water. The results of this study are shown in a log-log plot in Figure IV-1o An attempt was made to formulate these data in terms of the Brunauer-Emmett -Teller (BET) (10) and Langmuir (1) equilibrium adsorption equations. It was impossible, however, to get either of these equations to fit any appreciable range of these data, but approximations could be made to individual sections of these dataO The problem seemed to be caused by a heterogeneous surface which resulted in a constantly changing heat of adsorption (2, 27, 28, 55) This was shown by an attempt to plot these data according to the linear relationship, C = 1 + C IV-2 S Smb S derived from the Langmuir equation, S = SmbC IV-3 where C is the equilibrium concentration in moles per liter, S is the equilibrium concentration of the adsorbed phase in moles per gram, Sm is the maximum value which S can attain, and b is a constant in liters per mole indicative of the energy of adsorptiono The constant, b, is defined by the equation, b = b e Q/RT IV-4 where b is a constant and Q the heat of adsorption, is constanto When equation IV-2 is used, a straight line plot of

ra 2.01 38 0 021.c 11C.117 CIOQ 6 046C 4 calcL11atecl lo ft4 e-xxpe e rimen t~ o~ ~ Sl~ 3~~~~dPj- r Ur V.Ceq (rnol es/, 0"7; -1. er) 3 mol gu11bri.Tatons

77 the data does not result as it should if the assumptions of the model apply, but a curved line does indicating a changing value for b, and therefore for Q, and Smo It is possible, though, to fit major concentration ranges of the data with the empirical Freundlich equation, S = FCN, N< 1 IV-5 The constants F and N can be determined by log-log plots of the data, It is interesting to note that if an exponential distribution is assumed for the heat of adsorption, f(Q) a e-Q/RT IV-6 and if the variation in'b with surface coverage is attributed entirely to the variation in heat of adsorption, then S(C,TT) = S(b,C,T ) (bdb I-7 -0 where S(C,T) is the adsorption function S(bC,T) is the Langmuir equation and f(b) is the distribution function for b, becomes S(C,T) = FCN, N<1 IV-8 which is equation IV-5, the Freundlich equation (1)e F in this case is given by F = (CRT 1 b IV-9 N There is no assurance that the above derivation is uniqueo Therefore, that the surface is heterogeneous with respect to site energy is only suggested by the fact that experimental data conforms to the Freundlich equation, but is by no means proven (1)e One of the problems involved in using the Freundlich equation is that it does not predict an ultimate capacity for

78 the adsorbento It is apparent upon examining Figure IV-1 that the slope approaches the horizontal, indicating that such an ultimate capacity, approximately 4x103 moles per gram in this case, does existo A complicating factor which also could cause a changing slope is that the distribution function for the heat of adsorption is different than the exponential function postulated. For purposes of comparison, the phenol equilibrium data was approximated with the Langmuir equation, Equation IV-3, for a "high" and a "low" concentration rangeo The constants for the Langmuir equation were determined with Equation IV-30 These equations are plotted together with the experimental data in Figure IV-2. The increase in Sm,. 1.15xl0 3 to 2.84x10-3 moles per gram, and decrease in b, 6.7x104 to 2e93x103 liters per mole, upon going from the low concentration to the high concentration approximation is especially noteworthyo The equilibria data shown in Figure IV-3 are for the 4-nitrophenol-distilled water-carbon system. The pH of this system was lowered slightly to 4o0 with HC1 in order to insure that essentially all solute species were present in the neutral form. This system is assumed to be comparable to the phenol system studied aboveo The two Freundlich equation approximations to the data are shown in Figure IV-3o The Langmuir equation seemed to fit tne 4-nitrophenol adsorption data much better than the phenolo This equation, as determined from Equation IV-2, is shown together with the experimental data in Figure IV-4o Evidently the heat of

Ceq (moles/liter) x104 0 4 8 12 16 20 24 28 4 S = 8.32C (upper scale) X 3 1+2930C 0 A / 77C| ~~~~~1 /e~~-A / -s~~~~~~ 1+6.7x104C / | —-— ~ — ( lower scale) 0 8 16 24 32 40 48 56 Ceq (moles/liter) x10 Figure IV-2. Phenol Equilibria with Langmuir Equations

10-2 8 6 3 10k 8 o o 4 2 _ 10-4 I I I I II I [l 10 I4 06 2 34510- 2 3 4 5 810 2 3 4 5 6 8 10Ceq (moles/liter) Figure IV-3. 4-Nitrophenol Equilibria with Freundlich Equations

32 to 24 ~'0'0 us0 I / S - 71.4C 16 0 1 2 3 4 5 6 7 Ceq (moles/liter) x10 Figure IV-4. 4-Nitrophenol Equilibria with Langmuir Isotherms

82 adsorption does not change as much with surface coverage for 4-nitrophenol adsorption at this pH as it does for phenolo This could be due to the fact that 4-nitrophenol is adsorbed at different sites, sites of more uniform energy than phenol; or that the effects of an increasing heat of interaction between sorbate molecules with surface coverage is compensating for the effect of a decreasing heat of adsorption as described by Brunauer (9) It is informative also to compare the adsorption isotherms of phenol and 4-nitrophenol on the same ploto This can best be done by plotting the moles of sorbate removed per gram of carbon, S, vs. the reduced equilibrium concentration, C/Cs, after Hansen and Craig (29). Cs is the maximum possible solution concentration as limited by solubilityo This reduced concentration has the effect of eliminating differences in equilibrium adsorption characteristics due to solubilityo Differences in equilibrium adsorption curves which are plotted in this manner should then be due to the nature of the adsorption bond formed, or to the fact that adsorption occurs at different types of sites on the adsorbent surface The quantity of phenol and 4-nitrophenol is plotted as a function of C/Cs in Figure IV-50 The experimental data is not shown; rather, the normalized curves of the data shown in Figures IV-1 and IV-3 are used for this purpose. It can be seen from this plot that in the lower C/Cs range, more 4-nitrophenol is adsorbed than phenol. A determination of Langmuir "b" values for the same -4 -3 concentration range, lx10l to lx10 M, revealed that b for

10'go' | 4-Nitrophenol _ 4-4 a 10 ox~~ ^"^ ~~~~Phenol 10-4 J 10-6 10-5 10-4 10-3 10-2 101 Ceq/Cs Figure IV-5. Phenol and 4-Nitrophenol Equilibria as a Function of Reduced Concentration

84 phenol is 15,600 liters per mole while b for 4-nitrophenol is 24,700 liters per mole, indicating that phenol has a lower heat of adsorption. This higher heat of adsorption for 4-nitrophenol could possibly be due to adsorption taking place at sites where the nitro group can contribute significantly to the heat of adsorption. At high C/Cs, the curves show the same capacity for each solute; the solute molecules are evidently packed unto the surface in such a wiay that the nitro group cannot contribute to the energy of adsorption, and in such a way that the larger molecular volume of the nitrophenol does not affect the number of molecules adsorbedo Alternatively, the equilibrium concentration is high enough to cause adsorption at every site, regardless of the adsorption energy for the adsorbateo PHENOL REVERSIBILITY Reversibility studies were performed on the phenol system in order to characterize more fully the nature of adsorption by active carbono In one 3-liter, distilled water-phenol system, 5 grams of carbon were brought to equilibrium at a phenol concentration of 5x103 Mo In another 3-liter, distilled water-phenol system, 1 gram of carbon was brought to equilibrium as a concentration of 4x10-3 M" The carbon was kept in suspension at all times with a motor driven, polyethylene stirring blade, Every 3 to 5 days the carbon was allowed to settle to the bottom of the reaction vessel and 2e5 liters of the equilibrated solution was decanted and replaced with distilled watero Each time the new surface concentration could be calculated and plotted as a function of equilibrium concentration0 This

85 procedure was repeated for a period of 5 months. The results are shown in Figure IV-6 with the normalized adsorption curve reproduced from Figure IV-1 for comparison, Figure IV-6 shows the hysteresis effect which Adamson (1) describes as being due to "ink bottle" pores, These pores have a narrow entrance but a large interioro A high concentration of adsorbate is required to force phenol molecule into the pore, but once the pore is filled, the solute is "trapped" inside the pore as solution concentration is decreased in the manner described above The difference between the two desorption curves can be interpreted as being due to the fact that the higher initial equilibrium concentration for the 5 gram system results in more of these types of pores being filledo In each case, only about 50% of the phenol can be desorbed from the carbono An alternative explanation is that the "ink bottle" pores are not entirely responsible for the hysteresis effect, but that the carbon surface catalyzes the breakdown of at least some of the phenolo After 5 months of performing the desorption experiment, the uVo spectrum of the desorbed species was compared with that of a freshly prepared phenol solution and no significant differences were noted. It should be noted, though, that if the decomposition product was C02, it would not have been observed by this meanse The small phenol heat of adsorption reported below does not support the decomposition hypothesis, howevero There does not seem to be sufficient evidence to prove either explanation, but the former is more

10-2 2 -3 t 10 <~r-fl~~ ^~,~o0 —- Desorption Curve #1 5 usJ^ |I ~ ~~ ~ / ^^-Y ~ Desorption Curve #2 Adsorption Curve 2 10-4. I I _ 10-6 2 5 10-5 2 5 10-4 2 5 10-3-210 Ceq (moles/liter) Figure IV-6. Phenol Desorption Isotherms

87 strongly indicated than the latter~ TEMPERATURE DEPENDENCE OF EQUILIBRIA Equilibria studies were performed on the phenol-distilled water-carbon system, pH 506, at temperatures of 110 and 370 C., in addition to the room temperature studies at 250 C, already reported. The results are illustrated by means of a log-log plot in Figure IV-7o The isosteric heat of adsorption, qst4 and the heat of adsorption as given by the Langmuir adsorption model, Q, were both calculated from the data showno The isosteric heat of adsorption is defined (55) as, qst = RT2 alnC IV-10 aT- s The isosteric heat is actually a differential heat of adsorption and can be described as the energy necessary to remove an adsorbed molecule from its adsorbed state, including the attractive forces of its adsorbed neighbors, to an infinite distance from the surface, plus energy equivalent to the degrees of freedom of the "free" molecule in excess of those in the adsorbed state (55)o At a surface coverage of 2,0x103 moles of phenol per gram of carbon, qst is calculated as l 8 kcal./mole-~K for the data shown in Figure IV-70 The heat of adsorption, Q, can also be calculated from the Langmuir equations by first determining the Langmuir equation parameters for the low concentration range of the experimental data and then by using the b values in Equation IV-4.

10I-3I 10 ---- --- -— j-, —----— |o o 11~C 7 3u ^^^^^ 7 — 37~C 10-5 2 3 5 7 10-4 2 3 5 7 10-3 Ceq (moles/liter) Figure IV-7. Phenol Equilibria at Different Temperatures

89 The Langmuir equation constants are given in Table IV-10 Theheat of adsorption, Q, using only the b values for 110 and 370 C. is 0,8 kcal./mole-~K, TABLE IV-1 Langmuir Parameters for Phenol and 4-Nitrophenol Adsorption at Different Temperatures Temperature.Ljangmuir Parameters Solute OCo -.S.xlO 13bx10-3 11 175 22o7 phenol 25 lo70 21,0 phenol 37 lo65 20.2 phenol 11 3 07 26.3 4-nitrophenol 25 2,89 24,7 4-nitrophenol 37 2,76 21,3 4-nitrophenol The Langmuir adsorption model is based on the assumption that Q is constant throughout the adsorption process, and thus there is no valid point of comparison between Q and qsto It is interesting to note, though, the values which the data predict for each, especially since the Langmuir equations are so often usedo Equilibrium studies were performed at the same temperatures for the 4-nitrophenol-distilled water-carbon system, and the data for 110 and 37~C is given in Figure IV-8 along with the Freundlich equations for the datao The isosteric heat of a-3 adsorption, calculated this time at S = 2~0 x 10 moles per

5 2 o7 X 1 a -o-.. ~~Col~O 10 ~" —-3 L_ 00c 0-3 ~ W7 0v -OIc. 1 Equi 1 ib a d3a 7Oc 8-10" —.-no..._ Equilibria at Diff

91 gram, is q 506 kcal/mole-0~K For comparison; the Langmuir St equations were again determined, and the constants are shown in Table IV-1. Q, as calculated from these constants, is 1,5 kcal/mole-~Ke By comparing the values of Q and qst for both phenol and 4-nitrophenol, it can be seen that the heat of adsorption for 4-nitrophenol is about twice as large as that for phenol regardless of whether Q or qst is used for the comparison. These heats of adsorption all fall within the range indicative of physical adsorption (75) o PHENOL AND 4-NITROPHENOL ADSORPTION RATES The rate-limiting step for adsorption of phenol and 4nitrophenol was assumed to be intraparticle pore diffusion as it was for the acid adsorption study reported in Chapter IIIo The batch systems used for the rate studies were well mixed and the carbon was kept in suspension at all timeso The stirring rate was well above the 500 RPM Weber (66) found was necessary to make the rate of adsorption independent of the rate of stirring. Based on the assumption that pore diffusion is limiting, the diffusion model described and used in Chapter III is applicableo It should be noted, however, that in this case, concentration and activity are equivalento The computer program is also the same and is given in Appendix Io The equilibrium equation used for all rate curve calculations was the Freundlich question, which:fit the experimental data best for equilibrium concentrations less than the initial concentration of the rate study0 The carbon dosage used for all phenol studies

92 was 0.135-grams per liter and the initial concentrations Cop was 6~9 x 10-5 M unless otherwise noted, The carbon dosage for the 4-nitrophenol system was 0320 grams per liter and the -4 initial concentration was 1 x 10 M unless otherwise noted, The carbon dosage of 0,135 grams per liter corresponds to an "A", or volume served per particle, of 0.0593 cm, and 0,320 grams per liter to an "A" of 0~025 cm 0 The calculated rate curves and experimental data for phenol and 4-nitrophenol are shown in Figures IV-9 and IV-10 respectively. In each case, the experimental data shown are a composite of two or more experiments, The ordinate for these plots is C/Co, or the solute concentration at any time, t, divided by the initial concentration and the abcissa is the dimensionless time parameter, T, which is given by Dt/a2 D in this case is the pore diffusion coefficient and a is the particle radius, 0001365 cm. The D used to plot the experimental data was arbitrarily selected as the one which made the data correspond to the calculated curve at the time and C/Co corresponding to one-half completion of the adsorption process6 The system parameters and the calculated diffusion coefficients are given in Table IV-2, Included also in this table is the range of diffusion coefficients which would make all data correspond to the calculated curve0

1.0.9 Co=6.9xl15 M Wt. carbon=0.135 grams/liter.8 _ _ —-- calculated curve O — o experimental data.7.6- D=5.7xl06 Cm2.5 -0 - -- sec C Coo —-.4 0.3.2.1 0 20 40 60 80 100 120 140 T = Dt a2 Figure IV-9. Phenol Rate of Adsorption

1.0 o9.-4;o=lxl0 - M Wt. carbon=0.32 grams/liter.8.7 - \O calculated curve.6 \ 0 —O experimental curve C Co.5.4 _ O.3 -- O ^^"^^ D=1. 2x10O6 cm ~~~~~~~~^-<~.2 - ~ —sec.1 L 0 8 16 24 32 40 48 56 T = Dt a2 Figure IV-10. 4-Nitrophenol Rate of Adsorption

95 TABLE IV-2 Diffusion Coefficients for Phenol and 4-Nitrophenol Freundlich D 6 Solute Parameters (cm /sec)xlO F N __________:- -1 _-F.......iii,.. Phenol 8046.40 5o7 (2 0 - 10l6) 4-Nitrophenol,0175.223 1,2 (0.8 - 2a0) As can be noted from Figures IV-9 and IV-10e the calculated rate curves do not correspond very well with the experimental datao The diffusion coefficient range calculated by making all data Correspond to the theoretical curve increases with increasing time of reaction, with the effect being much more significant for the phenol system than for the 4-nitrophenol system. The fact that a constant diffusion coefficient was not found is contrary to the assumption made in developing the model, and is also contrary to the results obtained in Chapter III for acid adsorption rates. The effect of surface diffusion on the walls of the pores was examined in an attempt to determine the reason for the differences between experimental and calculated curves6 This was done by first assuming that only surface diffusion was important and that the mass transfer contribution from pore diffusion was negligible0 This could easily be done by replacing the dimensionless concentration term "x" in Equation III-4 by the dimensionless surface concentration term "s" to yield,

96 ac = 2 as a s v-1l aT p2 apr aT This equation could then be solved by the computer in the same manner as before, subject to the same boundary conditions as before. A curve similar to that shown in Figure IV-9 was calculated for the phenol system and from it the surface diffusion coefficient could be calculated. Ds for the same experimental data shown in Figure IV-9 was determined to be 5 to 15 x o-101 cm2/sec, Because this variation is about the same as the variation in the pore diffusion coefficient calculated above, it was decided that this procedure was not a good alternative to calculating the pore diffusion coefficient. Another alternative was to combine the effects of both pore diffusion and surface diffusion in one model, This has been done by others (7), but the solution arrived at for the phenol system following their procedure did not yield a unique set of diffusion coefficients This alternative can not be discarded, however, until a more extensive study is undertaken, Other possibilities which could be causing the observed discrepancies in addition to surface diffusion include the possibility that D is a function of concentrations or that adsorption from solution significantly affects the rates in addition to diffusion, The possibility that the variation is due to equilibrium curve equations predicting too high a value for surface concentration at very low solution concentrations is not likelye Wide changes in the isotherms equation for low concentration were found to make very little difference

97 in the calculated curve, The best possibility at present seems to be that surface diffusion is playing a significant role. This is especially indicated by heat of adsorption data which will be presented below. Since a constant D could not be determined for these systems, it was decided to use the D calculated at the time the adsorption process for the system was one-half completed for comparison. For phenol, this corresponds to a value of 5.7 x 106 cm2/sec, which is a reasonable valueo This value for D can be compared to an aqueous solution value of 8,9 to 9.6 x 10-6 cm2/sec (18), The value for pore diffusion should be somewhat lower because both porosity and tortuosity of path effects which are not included in the diffusion model would tend to make the diffusion coefficient smaller, The value of D for 4-nitrophenol calculated in the same manner is 1,2 x 10-6 cm2/sec, This coefficient is expected to be somewhat less than that for phenol because of the larger molecular size, It should also be noted that the variation in D for these two solutes is much greater for phenol than for 4-nitrophenol0 If surface diffusion of the solute is an important effect; it does not appear to be as important for 4-nitrophenol as for phenol0 This observation is in keeping with the fact that the heat of adsorption for phenol is one-third that for 4nitrophenolO The phenol molecules would not be held as tightly on the surface, and surface diffusion could thus play a larger role o

98 TEMPERATURE DEPENDENCE OF ADSORPTION RATES The rates of adsorption of phenol and 4-nitrophenol (PNP) have been studied as a function of temperatures The system in each case consisted of solute, distilled water and carbon, except that the pH of the 4-nitrophenol system was adjusted slightly to 4o0 with HC1o The calculated rate curves together with the experimental data are shown in Figures IV-11 and IV-12 for temperatures of 11~ and 37~Co The calculated diffusion coefficients together with the parameters used for the calculations are shown in Table IV-3 The activation energy can then be calculated by means of Equation III-8, Using only the 11~ and 37 C diffusion coefficients, the activation energy for phenol can be calculated as -5o3 kcal/mole-~K and as -508 kcal/mole-~K for 4-nitrophenolo TABLE IV-3 Diffusion Coefficients for Phenol and 4-Nitrophenol at Different Temperatures Freundlich D Solute Temperature Parameters (cm2/sec)xl06 PeoF N____ Phenol 110 o014 o273 2o8 (lol - 4~2) Phenol 25~ o014 o28 4o7 (2o6 - 7,3) Phenol 37~ o014 o29 6o2 (2,3 - 9O0) PNP 11~ o0073 o127 1oO (05 - lo5) PNP 250 o0175 o223 lo2 (Oo8 - 2o0) PNP 37~ o0068 o136 2o4 (lo3 - 3o8)

1.0O pH=5.6.9 - 0Co=6.9x10-5 M Wt. carbon=0.135 grams/liter.8 AO ^ Z0 —-037~C *.~7 - ^S>^ A 0 ~~~-6 2/se D=6.2xl cm /sec ^^^*>^^ ~A —A 110C C.6 O D=2,8xl0"'6 cm2/sec Co.5 ___ A ^~~^ —-- — ^ _. ___ I ^ 4.3.2.1 0 20 40 60 80 100 120 140 T =Dt Figure IV-11. Phenol Rate of Adsorption at Different Temperatures

1.0 pH=4.0 ~9 \Co=lxlO-4 M 9i\^,~~~~~ - Wt. carbon=0.32 grams/liter *8 -- S.7 O__D=2.4xl0_ ~ 37 c - D=2. 4x0 cm2/sec.6 n 1 — Co D=lOxlO cm2/sec.5 -.4 3 -AL 0^.2 -.1 0 8 16 24 32 40 48 56 T Dt a2 Figure IV-12. 4-Nitrophenol Adsorption at Different Temperatures

101 An alternative means for calculating the activation energy involves the use of the slope, k, of the St vso t plots of the rate datao This methods yields activation energies which compare favorably with those calculated from the diffusion coefficientso Such a plot is shown for the phenol system in Figure IV-13 and for the 4-nitrophenol in Figure IV-140 These plots show that the straight line portion of the curve intercepts the abcissa of a positive value of t~0 This same intercept has been noted for all rate studies of strong acid adsorption, and phenol and 4-nitrophenol adsorptiono The possibility that this intercept was due to the time required to "wet" the carbon at the beginning of a rate study was checked by wetting the carbon thoroughly with distilled water before adding it to the system. The rate curve obtained using the wetted carbon did not differ at all from the rate curves obtained using the dry carbon. Some other phenomenon which is not readily apparent to the author is evidently causing this intercepto Regardless of the intercept, however, the slope of the straight line portion of this plot, k, can be used to good advantage, Weber and Gould (68) found that by using the square of this slope, k2, as a relative rate constant, that Equation III-8 gives reasonable values for the activation energy of intraparticle diffusion. A plot of Equation III-8 for the phenol system at temperatures of 20, 11 250, 37~ and 50~C is shown in Figure IV-15 and for 4-nitrophenol at the same temperatures in Figure IV-16, Ea as calculated from the slopes of these plots is -3.6 kcal/mole-~K for phenol and -3o9 kcal/mole-~K for

I I..!I -. I — I -- I.....!I 4 0 or 3 ffiES~~/ | slope, k, = 4.4xl-5 k $4 moles/gram-min2 o Co=6. 9x105 u- | 0DWt. carbon0. 135 grams/liter |cnV]~~~~~ I/^~ pH=5.6 -- L -- I L I B l 0 2 4 6 8 10 12 14 Time2 (min.) Figure IV-13. Phenol Rate of Adsorption

4-4 Co=lxlO M Wt. carbon=0.32 grams/liter pH=4.0 ^^^^ ~slope,k, = 4.22x10 moles/gram-min~ o 3 0 1 2 3 4 5 6 7 Time (min.) Figure IV-14. 4-Nitrophenol Rate of Adsorption

.8 Co=6.9x10 5 M 7 Wt. carbon-0.135 grams/liter pH=5.6.6 Ea=-3.6 kcal,/mole- K ~ 5 4 0.2.2 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 T-1 (OK)x103 Figure IV-15. Temperature Dependence of Phenol Rate of Adsorption

.8 Co=lxlO4 M Wt. carbon=0.32 grams/liter pH=4.0.6.5 0Ea=-3.9 kcal/mole-~K + 4 CN 3~~~~~~~~~~~~~~~~~~~~~ 0' o~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 T~1 (OK)xl03 Figure IV-16. Temperature Dependence of 4-Nitrophenol Rate of Adsorption

106 4-nitrophenol The value of Ea = -3o9 kcal/mole-OK for 4-nitrophenol can be compared with a value of -8.9 kcal/mole-K found by Weber and Gould (68) This difference is apparently due to the different systems which were used —Co = lx104M and 0,32 -5 grams per liter of carbon in this case, and Co = lx10 M and 0.025 grams per liter used by Weber and Gould, The values of k2 used for these plots are tabulated in Table IV-4o The fact that the S vs, th is linear for some time after the carbon is added to the system also supports the assumption that diffusion is rate controllingo Crank (14) illustrates many processes for which this occurs. TABLE IV-4 The Relative Rate Constant at Different Temperatures Temperature Phenol 4-Nitrophenol 0 C kx105 k2xlO10 kx105 k2x100 moles moles ~. — 5 gm-mn 5 gm-mln 2 3,5 11o2 3o29 10o8 11 4,0 16o0 3~78 14o2 25 4,4 19,4 4.22 17o8 37 4,8 23,0 5~55 30,8 50 5,6 31,3 5~66 32.0

107 DEPENDENCE OF ADSORPTION RATE ON CARBON DOSAGE The effect of carbon dosage on the rate of adsorption of 4-nitrophenole The system chosen for the study was the distilled water 4-nitrophenol system at pH = 4.0. The carbon dosages studies were 0,032, 0.064, 0o16 and 0,32 grams per liter. The results are shown in Figure IV-17 in an St vSo t plot, The slopes of these rate curves vary from 4.2x10 5 moles/gm-min~ for a dosage of 032 grams per liter to 11o8x 10m5 moles/gm-minm for a dosage of 00032 grams per liter Thus, decreasing carbon dosage causes the rate of adsorption to increase significantlyo The difference in rate seems to be due to a more rapid decrease in driving force in the systems with the higher dosage. For example, at a surface coverage of -4 2.5 x 10 moles per gram, the solution concentration in the 0.032 grams per liter system is 9.2 x 10" M, while for the 0.32 grams per liter system the solution concentration is 2.5 x 10 Mo As the carbon dosage approaches zero, the infinite bath case would be approached, and the rate of removal would no longer be a function of carbon dosageo In spite of the fact that the rate of uptake as indicated by the St vso t~ plot would be expected to change with carbon dosage, the diffusion coefficient should not change if the assumptions included in the model are valid. However D was found to increase with decreasing carbon dosage, contrary to the assumptions in the model, and contrary to the results of the acid adsorption datao The problem again seems to be that the model inadequately describes the data as discussed

8 pH=4.0 k=1 -5 -4 / k=lOxl Co=lxlO M k118 5 / 0.064 grams/liter k=11.8k=6.8 x10-5 0.032 grams/liter k=6.4xgr /l D 6p_/.....0.16 grams/liter 0 6 k=4. 2x105 0.32 grams/liter 2 0 2 4 6 8 10 12 14 Time (min.) Figure IV-17. 4-Nitrophenol Rate of Adsorption at Different Carbon Dosages

109 before, and a model which describes the data better is now being soughto The D chosen for comparison was arbitrarily selected as the one for which the data and calculated curve corresponded when the adsorption process was one-third completed~ The results of this calculation are given in Table IV-5o It should be noted, however, that D increases not only with decreasing carbon dosage, but also with increasing surface coverageo This point supports the hypothesis that surface diffusion is a significant part of the rate phenomenono TABLE IV-5 Diffusion Coefficients at Different Carbon Dosages o223 -4 4 Nitrophenol, S = o0175C, Co = 1 x 10 M Carbon Dosage A D (gms/liter) cm3 (cm2/sec) x 10 0o032 0025 5o4 0064 0125 4ol 0,160 005 lo5 0o320 0o025 0o8 DEPENDENCE OF ADSORPTION RATE ON INITIAL CONCENTRATION OF SOLUTE Rate studies were performed on the phenol-distilled watercarbon systems at initial phenol concentrations of 6o9 x 105 -3 and 1 x 10 Mo The calculated rate curves together with experimental data are shown in Figure IV-18o The diffusion coefficient calculated at the point at which the adsorption

1.0,. 9 A Co=lxl0-3 Co=6.9x10-5 Wt. carbon=0.32 grams/liter Wt. carbon=0.135 grams/liter pH=5.6 pH=5.6.8 A 0^^D -6 2/sec D=58x10- c 2/sec D=5.7x D=8xlO cm cm.7 \ ~ —.6 - ~. ~~3~ — 2 _.1 0 20 40 60 80 100 120 140 T = Dt a2 Figure IV-18. Phenol Rate of Adsorption at Different Initial Concentrations

1ll1 process is 50% completed is 5.7 x 10"6 cm2/sec for Co = 609 x 10 5 M, and 8.0 x 10-6 cm2/sec for Co = 1 x 10-3 M. The fact that the diffusion coefficient changes with initial concentration again is not in accord with the assumption of a constant diffusion coefficient which is inherent in the model, but the fact that the diffusion coefficient increases with surface coverage again indicates that surface diffusion could be an important consideration, It is also interesting to look at the St vSo t~ plot of this rate data in Figure IV-19, The slope of this plot, -4 -3 k, is 4.7 x 10 moles/gm-min for Co = 1 x 10 M and 4o4 x 10 5 moles/gm-min for Co = 6.9 x 10 5 M. Assuming a straight line relationship for dependence of k2 on Co after Weber and Gould (68), k2 is increasing at the rateof 2~34 x 10 molesliter/gm2-min. This compares with a value of 1.97 x 105 moles-liter/gm2-min found for 2, 4-dinito-o-sec-butylphenol (68) EFFECT OF ASH REDUCTION The effect of reduction of ash content of the carbon on adsorption of phenol was also studiedo The carbon was the same as that described in Chapter III which was treated with 1+1 HC1 to reduce the ash content from 0,7% to 0o3%0 Both rate and equilibrium studies were performed on the phenol-distilled water-carbon system at pH = 5.o6 No significant effect was noted either on the equilibrium isotherm or on the rate of phenol adsorptiono

20~! - 4 I l l 20 e Co=lx10-3 k=4.7x10-4 - Wt. carbon=0.32 grams/literk / pH=5.6 / H o15 /-8" 3X Time k=4.4x10-( Initial Co=6.9xncentratons Wt. carbon=0.135 grams/liter pH=5.6 5 1 2 4 6 8 10 12 i4 Time2 (min) Figure IV-19. Phenol Rate of Adsorption at Different Initial Concentrations

113 The reason that no effect of treatment of the carbon for reduction of ash content was noticed probably was due to the fact that the original -ash content was small. If the ash content had been larger, the effect on adsorption could well have been more significant, and this should be further studied using carbons of high ash contento EFFECT OF pH ON ADSORPTION The effect of solution pH on adsorption of phenol and 4-nitrophenol was also studied extensively, The pH effect on adsorption can be attributed, in general, to either effects on the sorbate molecule or on the sorbent surfaceo The adsorption characteristics of the sorbate in its neutral form will be different than those for the anion, at least because the anionic species is more soluble and thus does not tend to be adsorbed as readily from solution. However, repulsive forces between the anionic sorbate and the sorbent could also become more significant, especially if the surface has a net negative charge. The effect of different pH on the adsorbent surface could take place in the form of a destruction or alteration of surface sites, or in the alteration of the surface potential so that attractive and repulsive forces are affected0 In order to examine these effects more closely, rate and equilibrium studies were performed on both phenol and 4-nitrophenol over a range of pH'so EEffect of pH.o Equilibria Phenol equilibria studies were performed on the

114 phenol-distilled water system at different pH with only HC1 or NaOH used for pH adjustment, The results of these studies are shown in Figure IV-20, and the Freundlich constants for these curves are shown in Table IV-7, It should be kept in mind that the PKa of phenol is 9o9 so that the principal adsorbing species above this pH is most likely anionic, A summary plot of these isotherms is shown in Figure IV-21, A significant result is the reduction in amount adsorbed with pH decreased from 7 a,5 The solubility of phenol is constant at 008 M in the pH range 2.0 to 5.6 as determined experimentally and it is assumed that no change in solubility occurs between pH = 5~6 and 7,5~ Apparently the nature of the surface is being altered in some manner by the acid used to adjust the pHo The mechanism of phenol adsorption on oxide surfaces such as activated Alumina has been determined as adsorption with the -OH group forming a bond with the surface oxide (36). Assuming that at least some of the phenol is being adsorbed in this manner on the carbon oxide groups, then some of these sites would be altered as acids react with the chromene functional groupso Also, if acid is being adsorbed on the microcrystallite basal planes inside the carbon pores (6), the phenol which would adsorb "flat" on these basal planes would have to compete with the acid0 Equilibrium capacity also decreases with increasing pH above 7,5. (Capacity measurements were made at only one pH above the pK because crystals appeared in the solution at higher pHe Apparently, the carbon was catalyzing some reaction which produced the crystals since phenol solutions at

2.0 -yi 1.5 0 — 2H V' 7. 5 0 5H 6 HrS ~~~r /^ _____ ~~~~calculated curve ~~~0~~~~~~~~0 — Ceq (moles/liter) x410 Figure IV-20. Phenol Equilibria at Different pH

2.5 i 2.0 0 r4 rd t12O -Z" ~~~~-4 Ceq=7x10 M 1.0 0.5 0 2 4 6 8 10 12 14 pH Figure IV-21. Phenol Equilibria at Different pH

117 high pH are usually quite stableo Further research will be required to find out more about this reaction.) Sodium phenolate is very soluble in solution and as a consequence the extent to which it is soluble was not determined, Experiments with 4-nitrophenol indicate, however, that the high pH is affecting the surface of the carbon in addition to increasing the solubility of the sorbate (see Figure IV-23). The Langmuir parameters were calculated for the same phenol concentration range for the pH = 7.5 data and the pH = 10.6 data, and it was found that the b value for the neutral species was about 75% larger than for the anionic species, (34,000 vso 19,000 liters/mole), thus indicating a lower heat of adsorption for the anionic species. Equilibrium studies were made to determine the effect of pH on 4-nitrophenol adsorption in the same manner as for phenol. The results of these studies are shown in Figure IV-22, and the Freundlich equation parameters for these curves are shown in Table IV-8o The equilibrium capacities in this case did not decrease with decreasing pH from the pK = 7~0o This would indicate that some of the sites at which phenol and 4-nitrophenol are adsorbed are not the same, at least for low concentration. It was shown previously (see Figure IV-5) that at high equilibrium concentration, the surface concentrations are the same for both phenol and 4-nitrophenol, and in this case it is likely that the same sites are covered with sorbateo Specifically, in the lower equilibrium concentration region, 4nitrophenol is apparently not adsorbed at the specific surface

3.2: - pH=6. 5 3.2~ ~ o H 2.4 8 Ad A_ V pH=10.0| 0 1 2 3 4 5_ 6 7 ~~~O 0'~~~~~~~~4 0 1 2 3 4 5 6 7 8 Ceq (moles/liter) xlO Figure IV-22. 4-Nitrophenol Equilibria at Different pH

119 oxides which are affected by pH and at which phenol absorbs, Alternatively, 4-nitrophenol may displace the physically adsorbed acid more easily than phenol. The latter mechanism is indicated by the fact that the heat of adsorption of 4-nitrophenol is about three times that of phenol (qst, phenol = 18 kcal/moleOK; qst 4-nitrophenol = 5.6 kcal/mole-OK) The capacity of carbon for 4-nitrophenol decreases with increasing pH above 700 as it did for phenol. In order to determine if the solubility differences between the neutral and anionic species were the principal cause of this decrease in capacity, a plot of reduced concentration (C/Cs) was made as a function of S for the pH = 4 and pH = 10 data, This plot is shown in Figure IV-23; only the normalized data curves are shown. The solubility limit of the anion was experimentally determined as 0.2 M while that for the neutral species was Ool Mo It can be seen from this plot that solubility does not play a major role, but that the effect of high pH on the carbon surface is more significant. Calculation of the Langmuir parameters revealed b for the pH = 10 data was 14,000 liters/mole compared with 50,000 liters/mole for the pH = 4 datao This indicates a lower heat of adsorption for the anionic species, and therefore it is possible that increased repulsive forces between the negative molecule and the negative surface are causing the lower capacity. The negative surface charge of the carbon is possibly being increased by physical adsorption of OH"ions on the surface, or by ionization of very weak, acidic functional groups on the surface such as the phenolic functional groupo

3.2 m 2.4. -- ^ \_ pH=4. 0 0 x 1l.6 ci) 0~~~~~~~~~~~~~~~~~~~~~.8 m I^ -/. ^ / 1 p 0 1 2 3 4 5 6 7 CeqJCs x103 Figure IV-23. 4-Nitrophenol Equilibria as a Function of Reduced Concentration

121 Equilibrium data at different temperatures were taken for both phenol and 4-nitrophenol0 The data for the 4-nitrophenol system at pH = 10 are shown in Figure IV-24 for temperatures of 11~ and 37~Co The isosteric heat of adsorption; qst, calculated for this data at S = 6 x 10-4 moles/gram is 0,75 kcal/mole-~K as compared with 6 kcal/mole-~K for the neutral species at S = 2.5 kcal/mole- ~K The implications of this low heat of adsorption are that the anionic species will not be bound as tightly at the surface and that surface diffusion could be more prevalent for the anionic species than for the neutral speciesO The temperature dependence of the 4-nitrophenol adsorption process at pH = 2 was found to be essentially the same as at pH = 4o0. No difference in temperature dependence was expected because the capacity for 4-nitrophenol is the same at these two pH'so The temperature dependence of phenol adsorption at pH = 2 was found to be quite different than that at pH = 5o6, however, in keeping with the effect of pH on capacity for the neutral speciesO The data for pH = 2 at temperatures of 110 and 37~C are shown in Figure IV-25o The systems used for this study contained 1 x 10 2 M NaCl, but it should be noted that at all pH's studied, there was no effect on the capacity of carbon for either phenol or 4-nitrophenol due to the presence of 1 x 10-2 M NaCl as compared with no NaC1i The temperature effect for phenol adsorption at pH = 2 was also found to vary more extensively with surface concentration than was previously

10-3 c i) 0 %, 3 I~~~~~ tf^^^~o~~~ ~pH=10.0O O —— ~ 370 2 6 _ 11~ 10-4 i.... —a..........-,,I 10-5 7 2 4 5 7 1 10Ceq (moles/liter) Figure IV-24. 4-Nitrophenol Anion Equilibria at Different Temperatures

2 v 5 = A d 37~C ~ w 4 L^ -lx10-2 M NaCi | ~ pH=2.0 10 2 3 4 5 2 3 4 5 7 104 2 37 10Ceq (moles/liter) Figure IV-25. Phenol Equilibria at Different Temperatures

124 observed at pH = 5.6~ In order to express this effect quantitatively, the isosteric heat of adsorption, qt was calculated for pH = 2 and pH = 5.6 at different surface coverages. The data for the pH = 5.6 calculations was taken from Figure IV-70 The results are tabulated in Table IV-6o TABLE IV-6 Isosteric Heats of Adsorption at Different Phenol Surface Coverages S(mole/gram)x 103 1.0 1.25 1o50 1,75 2.0 qst, pH = 2.0 5.4 4e8 3.7 0o7 (kcal/mole-OK) q t, pH = 5.6 2.7 2~2 2.0 2c5 1.8 ( cal/mole-OK) The greater effect of temperature on phenol adsorption at pH = 2 is quite possibly related to the effect of temperature on the quantity of acid adsorbed, Although there is much scatter in the data, the acid adsorption isotherms shown in Figure III-13 indicate that the process is endothermic at high acid concentrations. Since the phenol adsorption process is exothermic, an increase in temperature would cause less phenol to be adsorbede In addition, however, the increase in temperature at the same time would cause more acid to be adsorbed on the carbon surface, the presence of which would further decrease the amount of phenol which could be adsorbed. The acid, then, would tend to magnify the effect of temperature.

125 Effect of pH on Rate Kinetic studies were performed to determine the effect of pH on rate of phenol and 4-nitrophenol adsorption. The theoretical rate curves were again calculated and the diffusion coefficients were determined. The diffusion coefficients for phenol together with parameters used to calculate the rate curves are given in Table IV-7, and for 4-nitrophenol in Table IV-80 Typical calculated rate curves and experimental data are shown in Figure IV-26 for phenol at pH = 7o5 and pH =- 10 6 and in Figure IV-27 for 4-nitrophenol at pH = 4~0 and pH = 10 0O TABLE IV-7 Diffusion Coefficients for Phenol at Different pH ji;~ Freundlich D pH Parameters (Cm2/sec) x 10 ___ |_ F N2,0 0016 o316 5o7 (14 - 7o4) 4~0 o012 o260 3o5 (O9 - 5ol) 5.6.046.40 5,7 (2o0 - 10o6) 7.5 o020.306 4o0 (1.6 - 5o7) 10o6 o041.453 2.2 (lo5 - 2~4)..-:..........,......L.......... _..... _: —.

1 0.9.8 - DD=2.2xl06 cm2/sec pH=10.6.7 C.6~ ~ - " - - - __ _ Co.5 -.5.4- Co=6.xl0""Co D=4.0xl0l6 cm2/ s,- _ Wt. carbon=0.135 grams/liter pH=7.5.3 -.2 * i I__!_________ __i __I______ 0 20 40 60 80 100 120 140 T = Dt a2 Figure IV-26. Phenol Rate of Adsorption at Different pH

,o - Co=lxlO M \Wt. carbon=0.32 grams/liter.78 A 3 *6 ~- \0 p02"- P =10.0 O - Co co.5 -- D=3.8x10-6 cm2/sec 6.4 O *.~3 ~ < ^->^ pH=4.0.2 - D=1.2x106 cm2/sec. 1 I..-, - _ i - I - -I 1 0 8 16 24 32 40 48 56 T =Dt aa Figure IV-27. 4-Nitrophenol Rate of Adsorption at Different pH

128 TABLE IV-8 Diffusion Coefficients for 4-Nitrophenol at Different pH Freundlich D 6 pH Parameters (cm2/sec) x 10 F N ___________F N__ 2,0.0088.160 1.1 (0,6 - 1.6) 4 0 e0175 o223 1,2 (0o8 - 1.75) 6.5.019 o230 1.3 (0,5 - 2o0) 10.0 0011 ~333 3.8 (1o6 - 5.3) 12 0,0045.27 3.2 (14 - 4,8) The diffusion coefficients for phenol appear relatively constant for the neutral specieso The coefficient used for comparison again was arbitrarily selected as the one for which the calculated curve and experimental data correspond when the adsorption process is 50% completed. The diffusion coefficient is smaller, though, for the anionic specieso A smaller coefficient is expected because of the additional repulsive forces as the negative molecule passes by the negative surfaces but it is difficult to place any real significance on the difference noted because of the arbitrary means of selecting the particular value of the coefficient for comparisono The range of diffusion coefficients, however, also indicate that the diffusion coefficient for the anionic species is smaller than the neutral species. The 4-nitrophenol diffusion coefficient is also relatively constant over the pH range corresponding to the neutral species. The higher diffusion coefficient for the anion

129 is unexpected, however One possible expianation is that because the heat of adsorption is lower han for the neutral species (qst = 0o75 vso 6 0 kcal/mole-~K at S = 2 5 x 10C3 moles/gram) surface diffusion plays a more prominent role making the diffusion coefficient largero This effect would not be as great for phenol since the heat of adsorption of the neutral phenol species is much less than that for the neutral nitrophenol speciesO However; this can not be asserted with any degree of certainty until a better model is developed to interpret the rate datao Summary plots of the relative rate constant; k, as a function of pH are given in Figure IV-28 for phenol and Figure IV-29 for 4-nitrophenolo These constants, toot are quite uniform over the pH range of the neutral species, but drop off sharply for the anionic specieso The decrease in k is probably due to the fact that the capacity of the carbon for the anion has decreased, and to the increased resistance to movement inside the pore due to repulsive forceso It is important to note that the relative rate constants for both phenol and nitrophenol decrease significantly for anionic species as compared with the neutral species, while D for the nitrophenol anion increased and D for phenol decreases slightly. The reason for this occurrence is that the relative rate constant includes the effect of pH on capacity, while the model used to calculate D separates out the capacity effect, The temperature effect on the rate of 4-nitrophenol anion adsorption was studied at pH = lOoOo The calculated

0 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _________ ANN F- 5 4 g 0 U) ci) 4 _- 0 i-i - -26 Co=6.9x102 M 4 rcd 3 W ^t. carboneO.135 grams/liter o - U Cd 0 24 6 8 10 12 14 pH Figure IV-28. Phenol Rate of Adsorption at Different pH

C) 0 ------------------ ---- ----- --. --------. ---- AN S^V 5 g44 gE~~~~~~~~~ ~~~~Co=lxlO M Wt. carbon=0.32 grams/liter 4~~~~ 0 r-4 P"I 4J 3 -P 2 pH \ H C4) - (U 0) 4)~ 1 0246 8 10 12 14 PH Figure IV-29. 4-Nitrophenol Rate of Adsorption at Different pH

132 rate curve and experimental data are shown in Figure IV-30 for temperatures of 110 and 37~ CO Only one calculated rate curve is shown because the curves calculated for the two temperatures are nearly identical, in keeping with the very small temperature effect on equilibrium at this pHo The parr ameters used for the calculation in addition to the calculated diffusion coefficients are shown in Table IV-9o The activation energy calculated from the 110 and 370 C diffusion coefficients is -3.8 kcal/mole- K6 This value can be compared with a value of Ea =-2.2 kcal/mole- K calculated previously by Weber and Gould (68) using the relative rate constant, ko It should be noted that again, as for the neutral species, different initial concentrations and different carbon dosages were used which could cause the difference in activation energieso TABLE IV-9 Diffusion Coefficients at Different Temperatures for 4-Nitrophenol Anion Temperatures Freundlich D ~C Parameters (cm2 sec)x106 F N____ 11 o0094 o330 3o2 (0,7-307) 25 o0110 O333 3.8 (1,6-5o3) 37.0080 o313 5o7 (lo5-6o7) EFFECT OF NaCl ON ADSORPTION Effect of NaCl on Equilibria Equilibrium studies were performed for phenol and

1.0.9 _~..8 \ _.7 A 8O.6 C o o.5 _ Co=lx10-4 M H.4 Wt. carbon=0.32 grams/liter 0_3 0 11~C, D=3.2cm2/sec _3 37~C, D=5.7cm2/sec.2.1 L 1 E!... 0 8 16 24 32 40 48 56 7-= Dt a2 Figure IV-30. 4-Nitrophenol Anion Rate of Adsorption at Different Temperatures

134 4-nitrophenol systems at pH = 2.0 and NaCI concentrations of 0, -2 lx10 and 1 M. No significant effects of NaCl on capacity were noted for any of these systems, Studies on the 4-nitrophenol system at pH = 10,0, however, show that the effect of 1 M NaCI on the isotherm is quite significant when compared with the isotherms for systems contain-2 ing 0 and lx10 M NaCl. This comparison is shown in Figure IV-31. The data for 110-2 M NaCl is not shown because it is essentially the same as the 0 M NaCl isotherm. The concentration measurements made on the 1 M NaC1 - 4-nitrophenol system were changed to activity, X, by means of an activity coefficient of y=.75 given by Harned (31)o Comparison of the Langmuir equation b values calculated for the lower concentration regions of the two isotherms indicate that the heat of adsorption is greater for the 1 M system than for the O M system, A b value of 36,000 liters/mole was calculated for the 1 M system compared with 15,000 liters/mole for the O M systemR A distinct possibility for the cause of the difference between the two isotherms is the high salt concentration is compressing the electrical double layer, thereby reducing the repulsive forces between the organic anion and the surface and allowing more adsorption to occur at lower equilibrium concentrations. No change would be expected in the amount of sorbate which would be taken up at high equilibrium concentration since new adsorption sites are not being created, and this is verified by the fact that the isotherms are rapidly converging to the same surface coverage.

I. I I 20 Co=lx104 pH=10.0 Wt, carbon=0.32 grams/liter 15 o rd R 1M NaC OM NaC 1 tI 10 o UJ 5 1 2oe45rd O 1 2 3 4 5 6 7 Ceq (moles/liter) x104 Figure IV-31. 4-Nitrophenol Anion Equilibria at Different Salt Concentration

136 Effect of NaCl on Rate A study of the rate of adsorption as a function of salt concentration showed that salt has little effect on the kinetics of adsorption of the neutral species of both phenol and 4-nitrophenol, as expected, Also, at pH = 10.0, the 4-nitrophenol anion has essentially the same diffusion coefficient in the 1 M NaCl system as in the O M NaCI system. The high salt concentration apparently does not reduce the repulsive forces in the pore enough to allow the anion to pass more easily. EFFECT OF TYPE OF STRONG ACID ON ADSORPTION Kinetic and equilibrium studies were made for both phenol and 4-nitrophenol at pH = 2.0 using HC104, H3P04, H2SO4 and HNO3 for pH adjustment. The results, however, differed very little from those reported above for HCle It is felt that some differences would occur if the study were performed at much lower surface coverages, but their effect appears to be too small to be observed for the surface coverages used in these studies.

V. CONCLUSIONS ACID ADSORPTION Strong acids were found to react extensively with both the coconut-shell and coal-base carbons studied, and this reaction apparently has significant effect on the adsorption of specific organic molecules from solution by active carbon, Specific information was found about the strong acid-carbon interaction which proved useful in interpreting the results of phenol and nitrophenol adsorption on carbon and should prove useful in interpreting results of adsorption of other organic molecules on carbon. The kinetic data for this reaction, for example, was found to conform to an intraparticle diffusion model with the rate of conjugate anion movement in the pore appearing to limit the rate of diffusion of the acid. Diffusion coefficients were calculated which compared favorably with those found for anions in ion-exchange resinso Physical and chemical interactions between the diffusing species and the pore wall apparently act to reduce the rate of diffusion as shown by studies on carbons with different ash content and different pore size. The rate of pore diffusion of acids was found to increase significantly with increasing salt concentration in keeping 137

138 with higher anion activity gradients in the pore, The temperature dependence of acid diffusion is cha0 racterized by an activation energy of -2 to -3 kcal/mole- K, in keeping with values for electrolyte diffusion, The isotherm for acid sorption is best described by the Fruendlich equation, thus indicating a heterogeneous surface with respectto site energy, The capacity of carbon for strong acids increases significantly with increasing salt concentration, The capacity of carbon for different types of acids varies; the coconut-shell carbon capacity for HC104 is twice that for HCl for example. The capacity of carbon for acid depends upon the type of carbon as shown by the fact that the coal-base carbon has twice the capacity for HC1 as coconut-shell carbon per unit weight of adsorbent and per unit surface area since the coconutshell carbon has a surface area of 800 sq. meters per gram (71), while that for the coal-base carbon is about 1000 sq, meters per gram (51)o The reduction of the ash content of the coconut-shell carbon had no effect on its capacity for adsorbing acid. This was probably due to the fact that the original ash content was very low and the conclusion may not be valid for carbons of higher ash content. The mechanism of acid adsorption was not uniquely defined, but the data obtained can be explained in terms of both physical adsorption and chemical reaction with the chromene functional

139 group. PHENOL AND 4-NITROPHENOL ADSORPTION Phenol and 4-nitrophenol adsorption was studied both in the absence and presence of strong acids and salts in order to learn more about the nature of the active carbon surface and the mechanism of adsorption on the surface. A study of phenol and 4-nitrophenol adsorption in distilled water in the absence of inorganic acids and salts strongly indicates that the active carbon surface is very heterogeneous with respect to the energy of the sites at which adsorption occurs. Heterogeneity is indicated by the fact that the adsorption isotherms for both phenol and 4-nitrophenol can best be described by the Freundlich equations Reversibility studies on phenol, also in the absence of inorganic acids and salts, indicate that a large percentage of adsorption occurs in "ink bottle" pores for the coconutbase carbon thus making it difficult to desbrb the phenoln This finding is important because it could mean that the oxidizing agents used in the regeneration process would not be able to easily remove the adsorbed organic molecules from the carbon surface. 4-Nitrophenol was found to be adsorbed more extensively, and with a greater heat of adsorption, than phenol in the lower, reduced equilibrium concentration range. Adsorption evidently becomes non-specific at higher surface coverages, however, since the ultimate capacity for both solutes appears to be about the

140 -3 same at 4x10 moles per gram, This is also shown by the fact that heats of adsorption decrease with increasing surface coverage A percentage of the sites at which phenol adsorption takes place are evidently different than those at which 4-nitrophenol adsorption occurso The capacity of carbon for phenol. is reduced by 25% when adsorption takes place in a pH = 2 system as compared with a pH - 506 system, while the capacity for 4-nitrophenol is approximately constant over the same pH range, This occurs even though the two solutes have similar acid-base characteristicso The acid is apparently destroying, or competing with phenol for, some of the sites at which phenol adsorption occurso This finding illustrates the importance of adjusting to near neutrality the pH of solutions from which phenol is to be adsorbed on carbon: The capacity of carbon for phenol is also decreased by 30 to 50% over the equilibrium concentration range of 1 to 10x10-4 M as the solution pH changes from 7 5 to 10O6, while the capacity for 4-nitrophenol is reduced by about 67% for the pH change from 4,0 to lO0.0 This decrease in capacity can best be attributed to an increase in repulsive forces between the anionic sorbate and the negative carbon surface., The isosteric heat of adsorption decreases from 6 kcal/mole-~K for the neutral species to I kcal/mole-OK for the anionic species, thus indicating the size of these repulsive forces A 1 M NaCI concentration at pH = 10 appears to compress the electrical double layer, thereby reducing repulsive forces

141 and allowing the 4-nitrophenol anion to adsorb more easily. This NaCl concentration had no significant effect on the neutral species of either solute. Reduction in ash content of the coconut-shell carbon from 0.7% to 0,3% did not affect the carbon's capacity for phenol, This again is probably due to the fact that the original ash content is very small. The kinetic data for both solutes did not conform well to the pore diffusion model used for strong acid diffusion, Surface diffusion is strongly indicated as being of importance in addition to pore diffusion because 1) A larger variation in the diffusion coefficient calculated from the pore diffusion model occurs for the neutral phenol molecule than for the neutral 4-nitrophenol species in keeping with the lower heat of adsorption for phenolo The phenol molecule would thus be more likely to diffuse on the surface, 2) The variation in the 4-nitrophenol diffusion coefficient increases with increasing surface coverage as shown in the effect of carbon dosage study, Surface diffusion would be more prevalent at higher surface coverages again because of the lower heat of adsorption. 3) The diffusion coefficient for the 4-nitrophenol anion is larger and varies more than that for the neutral species, again in keeping with the lower heat of adsorption for the aniono The temperature effect on rate of adsorption is in keeping with the assumption of diffusion since the activation energy is in the range of -3 to -6 kcal/mole- Ko The rate of solute removal from solution per unit weight

142 of adsorbent increases with decreasing carbon dosage for a given initial concentrationO Thus, more efficient use is made of carbon at smaller dosageso For a given carbon dosage, the rate of solute removal from solution per unit weight of adsorbent increases with increasing initial concentration, FURTHER RESEARCH A logical extension of this research would involve the development of a diffusion model which would include the effects of surface diffusion and/or the actual adsorption step as well as pore diffusion0 Such a model could be tested with the data presented in this study. Further studies should also be performed on the effect of the ash content of carbon on adsorption, For this study, carbons with a much higher ash content than the coconut-shell carbon should be usedo Included could be carbons which have been through one or more regeneration cycles and which have a significant ash build-upo The results obtained in this study on the effect of pH and salt concentration on adsorption should be extended to other carbons and other soluteso Especially, the effect of pH in systems which have a low solute concentration should be examined since the effect could well be rore important than in the concentration range examined in this study,

APPENDIX I $COMPILE, MAD, EXECUTE, DUMP, PRINT OBJECT, PUNCH OBJECT DIMENSION C(i0), S(50) CNEW(50, SNEW(50), 1 CC(50), CTEMP(50) STEMP(50) INTEGER IPM,IP START READ AND PRINT DATA PARADIUS DELTERRTEND, 1 CO; RHO, COUt ^iT DELR=1 /P IP=P+1 EXECUTE ZEROo (CC(1) ooCC(IP C (1),o aC(IP) I CNEW(1) oooCNEW(IP),S(1) o S(IP) SNEW(1),, 1 SNEW(IP),TQ) ARR=4 *3 14159*(RADIUSoP 3 )/A R=R*RHO*(CO) o Po (N-) NEW INDEX=0 THROUGH DIFF, FOR M=P, -1,MoL,2o EM=M-1 COEF=DELT/( 2 *EM*DELR*DELR) TERM= (EMlo) * (C (M-1) +CNEW (M-1)) TERMS=(El4+1) * (C (M+1) +CNEW (M+1) ) 1 *2o*EM*(C(M)+CNEW(M)) TERM WHENEVER CNEW(M) G SNEW(M) CC(M)=C(M) + COEF*TERMS - (SNEW(M)-S(M)) CC(M) =oABSoCC(M) WHENEVER oABSo (CC(M)-CNEW(M)) oGoERR*CNEW(M) 1 INDEX=1 CNEW (M)=CC (M) SNEW (M)=R* (CNEW (M) o P oN) OTHERWISE SS=S (M) +COEF*TERMS- (CNEW (M) -C (M)) SNEW (M) =oABSo SS WHENEVER SNEW(M) o LTEST CC(M) =0O OTHERWISE CC (M)= SNEW (M)/R) oPo (1o/N) - END OF CONDITIONAL WHENEVERoABS (CC(M)-CNEW(M),GoERR*CNEW(M) CNEW (M) CC (M) DIFF END OF CONDITIONAL WHENEVER CNEW(1) oGoSNEW(1) CC(1)=C(1) + (3o*DELT/(DELR*DELR))*((CNEW 1 (2)+C(2) )(CNEW(1)+C(1)))-(SNEW(1)-S(1i)) CC ( 1 )=oABS oCC ( 1 ) WHENEVER oABSo (CC(1)-CNEW(1)) oGoERR*CNEW(1), 1 INDEX= 1 cNEwt ( 1) =CC(1) SNEW () =R* (CNEW (1) oPoN) 143

144 OTHERWISE SS=S(1) +(3,*DELT/(DELR*DELR))*((CNEW(2) C(2)) 1 -(CNEW (1)+C ))) (CNEW(1) (C (-C(1) SNEW(l ) =ABSoSS WHENEVER SNEW( (, LTEST CC(1) =0 OTHERWI SE CC (1)= (SNEW l) /R) Po (1 o/N) END OF CONDI7TIONAL END OF CONDI2tIONAL WHENEVER oABS (CC(1)-CNEW(1)) oGoERR*CNEW(1) 1 INDEX=1 CNEW(1) =CC(1 ) SUMi=0 SUM2=0 THROUGH SUM, FOR I=P,-2,IoLo2 SUMI = SUM1 + (I-1) oP,2,* (CNEW(I)+SNEW(I)) SUM2 = SUM2 + (I-2).Po2,2*(CNEW(I-1)+SNEW(I-1)) PROD1=4 *SUM1 PROD2 =2*SUM2 COE=DELR, P 3,/3 GRAL=COE* (P*P* (CNEW (IP) +SNEW (IP) ) +PROD1+PROD2) CC (IP) =1 - GRAL*ARR CC(IP) =ABS CCC(IP) WHENEVER oABS(CC(IP)-CNEW(IP)) GeERR*CNEW(IP), 1 INDEX=1 CNEW(IP)=CC(IP) SNEW(IP) =R* (CNEW(IP) oPoN) WHENEVER INDEX o Elo, TRANSFER TO NEW T=T+DELT Q=Q+L WHENEVER Q GE COUNT PRINT RESULTS T, CNEW (IP) Q=0 o END OF CONDITIONAL WHENEVER To G, TEND-PROF PRINT RESULTS T,CNEW(IP), SNEW(1) oooSNEW(IP) END OF CONDITIONAL WHENEVER T Go SUB DELT=NEXT COUNT-CHANG SUBSUBS 1 SUB1=SUB2 SUB2=SUB3 NEXT=NEXT 1 NEXTi=NEXT2 NEXT2=NEXT3 END OF CONDITIONAL WHENEVER To Go TEND, TRANSFER TO START THROUGH NEWVR, FOR I=1,1,IoGoIP CTEMP (I) =C(I) STEMP (I) =S (I)

145 C(I)=CNEW(I) S(I)=SNEW(I) CNEW (I) =CNEW(I)'+(CNEW (I) -CTEMP (I) ) NEWVR SNEW((I) =SNEW (I) + (SNEW (I) -STEMP (I) ) TRANSFER TO NEW END OF PROGRAM

REFERENCES (1) Adamson, Ao Wo, Physical Chemistry of Surfaces, ppo 400-402, 548; Interscience Pubiiers New York, London, 1967, (2) Adamson, Ao W., Ling, I,, "Effect of Radiation on the Surfaces of Solids. I. Analysis of Adsorption Isotherms to Obtain Site Energy Distributions," in Solid Surfaces and the Gas-Solid Interface, Advances in C emistry Series T33,7 Ro Fo Gould, Edo0 Amo Chemo Soco Publications, 1961l (3) Allen, Jo Bo, Joyce, Ro S., "Column Calculations for Intraparticle Diffusion Controlled Adsorption," Proco Fifty-Ninth Annual Meeting, American Institute i Detroit, Dece er 4 0 of ChemicalEngner s, Detrit, De c-emb. -6'- (4) Barrer, R. Mo, "Specificity in Physical Sorption," Jo Colloid Interface Scio 21 (4), 415 (1966) o (5) Blackburn, So, Kipling, Jo Jo, "Adsorption from Binary Liquid Mixtures: Some Effects of Ash in Commercial Charcoal," JO Chemo Soco, London, 1955, 4103o (6) Boehm, Ho PO, "Chemical Identification of Surface Groups" in Advances in Catalysis, Do Do Eley, H Pines, PO B30 Wisz-IS',E-ds-, Vo~.6, Academic Press, New York and London, 1966o (7) Brecher, Lo E,, Kostecki, J0 Ao, Camp, Do To, "Combined Diffusion in Batch Adsorption Systems Displaying Bo Eo To Isotherms: Part I," in Chemical En ineering Pro ress, Symposium Series 74, VoTo T3 19 (8) Brewster, R0 Qo Organic Chemistry p~ 6220 PrenticeHall, Inco, Englew7ood Clif-s', New Jersey, 1953o (9) Brunauer, S., "Solid Surfaces and the Gas-Solid Interface," in Solid Surfaces and the Gas-Solid Interface, Advances in Chemistry Series-3 RRo Fo Gould, Ed, Amo Chemo Soco Publications, 19610 (10) Brunauer, So, Emmett, P Ho, Teller, Eo, "Adsorption of Gases in Multimolecular Layers," J. Amo Chem SOco 60, 309 (1938), 146

147 (11) Carr, Co W,, Fruendlich, Ho Sollner, K. o "The Influence of Salts on Adsorption of Strong Acids," JO Amo Chemo Soco 63, 693 (1941) (12) Chu, B,, Whitney, Do Co, Diamond- R0 Mo, "On IonExchanges Resin Selectivities " Jo Inorg, Nucl Chem0 24, 1405 (1962), (13) "Columbia Activated Carbon,)" circular, National Carbon Co0o New York, 962o (14) Crank, Jo The Mathematics of Diffusion, Clarendon Press London, _9-~. (15) Dostal, Ko Ao, ilarrington, JO J, Robeck Go G, Clark, Ro Mo "Development of Optimization Modeis for Carbon Bed Design,' Jo Am Water Werorks Assoco 58 (9), 1170 (1966) (16) Dubinin, M0 Mo, "Porous Structure and Adsorption Properties of Active Carbon," in Chemistry and Physics of Carbon, Volo 2, P o Lo Walker, Edo0, Marcel Dekk1er, Inc,,, New York, 1966, (17) Dubinin, Mo M p Plavnik, Go Mo, Zaverina, E, Dc. "integrated Study of the Porous Structure of Active Carbons from Carbonized Sucrose," Carbon 2, 261 (1964) (18) Edeskuty, Fo J,, Amundson, No Ro "Effect of Intraparticle Diffusion: Io Agitated Non-Flow Adsorption Systems," Indo Engo Chemo 44, 1698 (1952) (19) Garten, V0 Ao Weiss, Do Eo "The Quinone-Hydroquinone Character of Activated Carbon and Carbon Black," Australian Jo Chem. 8, 68 (1955) (20) Garten, Vo A,, Weiss, Do Eo "A New Interpretation of the Acidic and Basic Structures in Carbons IIo The Chromene-Carbonium Ion Couple in Carbon," Australian J. Chemo 10, 309 (1957) o (21) Garten, Vo Ao Weiss, Do Eo, "The Reaction of Carbon Blacks with Sodium Borohydride " Australian Jo Chem, 14 155 (1961)o (22) Garten, Vo A,, Weiss, Do E,, "The Ior-and ElectronExchange Properties of Activated Carbon in Relation to its Behavior as a Catalyst and Adsorbent," Rev, Pure Appl% Chem, 7, 69, (1957),

148 (23) Garten, Vo Ao. Weiss, Do Eo Willis, Jo Bo0 "A New Interpretation of the Acidic and Basic Structures in Carbon Io Lactone Groups of the Ordinary and Fluorescein Types in Carbons," Australian J0 Chemo 10, 295 (1957)c (24) Gasser, C, Go, Kipling, Jo J,, "The Effect of Surface Complexes on Adsorption from the Liquid Phase by Carbon Black,""'roco Confo Carbon, 4th, Buffalo,? No Yo 1959, ppoP —- Pergamon Press New York, 1960, (25) Gould, J. P., "The Influence of Structure and pH on the Kinetics of Adsorption of Substituted Phenols from Aqueous Solution by Active Carbon," Masters Thesis, Univo of Micho, 1967o (26) Hallum, Jo V.,, Drushel, Ho V,, "The Organic Nature of Carbon Black Surfaces," J Physo Chemo 62, 110 (1958) (27) Halsey Go "Physical Adsorption on Non-Uniform Surfaces," Jo Chem. Physo 16 (10); 931 (1948)o (28) Halsey, Go Taylor, Ho So,, "The Adsorption of Hydrogen on Tungsten Powders," J Chem, PhySo 15 (9) 624 (1947)o (29) Hansen, Ro So Craig, Ro Po "The Adsorption of Aliphatic Alcohols From Aqueous Solutions by Non-Porous Carbons," J, Physo Chemo 58, 211 (1954)o (30) Harker, Ho, Jackson, C,, Wynne-Jones, Wo Fo Ko, "Electron Spin Resonance in Carbon," Proco Royo Soc O, London A262, 328 (1961) o (31) Harned, Ho So Owen, B Bo The Physical Chemistry of Electrolyte Solutions, po 593, Tehnhodi PublishMhg Corpo, New York, 19580 (32) Hassler, Jc Wo Activated Carbon, Chemical Publishing Co, New York, 19630 (33) Heckman, Fo Ao, "Microstructure of Carbon Black," Rubber Chem. Technolo 37, 1245 (1964) o (34) Helfferich, Fo, Ion Exchane ppo 234-309, McGrawHill Book COo, Inc., New York, 1962o (35) Hennig, Go Ro, "Surface Oxides on Graphite Single Crystals," Proco Conf. Carbon, 5th, Penn State Univ,, Pao, 1961 1 ppo 14 3-4 6 Peramgamon Press, New York, 19620 (36) Higgins, Bo Richards, Jo Ho, "The Effect of Aprotic Solvents on the Adsorption of Phenol by Alumina," Jo Pharmo Parhmacolo 19 (10), 641 (1967) o

149 (37) Ingram, D. J. E,, "Electron Resonance Studies of HeatTreated Organic Compounds," Proc. Con'f Carbon, 3rd, Buffalo, N. Y., 1957, ppo 93- 10T ergamon Press New York, 1959, (38) Johnson, R. L., Lowes, F,, Jr., Smith, Re M, Powers, Te J, "Evaluation of the Use of Activated Carbons and Chemical Regenerants in Treatment of Waste Water,r" Public Health Service Publication Noe 999-WP-13, May 1964o (39) Joyce, R. S., Sukenik, V. A., "Feasibility of Granular Activated Carbon Adsorption for Waste Water Renovation," Public Health Service Publication No, 999-WP-12, May 1964, (40) Keinath, T. M., Weber, W, J,, Jr., "A Predictive Model for the Design of Fluid-Bed Adsorbers," presented at 40th Annual Conference of the Water P6llution Control Federation, New York, October 8-13, 1967. (41) Kipling, J, J,, "Adsorption on Solids from Binary Liquid MiXtures —The Importance of the Surface Structure of the Adsorbent," Proc. Intern. Congr. Surface Activity, 2nd, London III, p. 462 (1957). (42) Kipling, J, J,, Shooter, P. V., "The Adsorption of Iodine Vapor by Graphon and Spheron 6," J. Colloid Interface Sci. 21 (2), 238 (1966). (43) Kitchner, Je A., "Physical Chemistry of Ion-Exchange Resins," in Modern Aspects of Electrochemistry, Jc Bockris, Ed., ButtEerworths Scientific-PUbliations, London, 1959. (44) Kolthoff, Ie M., "Properties of Activated Charcoal Reactiviated in Oxygen at 400 ~Ce " J. Am, Chem, Soc. 54, 4473 (1923). (45) McNeil, R., Weiss, D, Eo, "Some Novel Carbonaceous Polymers," Proc. Conf, Carbon, 4th, 1958, pp, 281-290, Pergamon Press, New YorkT 1960. (46) Miller, E. J,, "Adsorption by Activated Charcoal. I, Proof of Hydrolytic Adsorption," J. Am, Chem, Soc. 46, 1150 (1924) (47) Miller, E, Jo, "Adsorption by Activated Sugar Charcoal. Adsorbability of Hydrogen and Hydroxyl Ions," J. Amc Chem, Soc. 47, 1270 (1925), (48) Miller, E" J,, "The Adsorption of Electrolytes by AshFree Charcoal," J. Phys Chem. 36, 2967 (1932).

150 (49) Nemerow, Ne L,, Theories and Practices of Industrial Waste Treatment, Addison Wes ley Publishing CoT Inc, Reading, Mass, e 1963. (50) Parks, L, R., Bartlett, P. G(, "The Effect of Inorganic Salts on the Adsorption of Inorganic Acids and Bases," J. Am. Chemr Soc. 49, 1698 (1927) (51) "Pittsburg Activated Carbon," circular #AC7-556, Pittsburg Coke and Chemical Co., Pittsburg, Pa.(52) Puri, B. R., Mahajan, 0. P., Singh, D. D.,, "Interaction of Carbon with Chlorine Water. II, ^" J. Indian Chem. Soc. 37, 171 (1960). (53) Puri, B. R., Mahajan, Oe P., Singh, D D., "Chemisorption of Oxygen by Charcoal on Treatment with Some Mild Oxidizing Agents," J. Indian Chem. Soc. 38, 135 (1961). (54) Puri, B, Ro, Singh, D, D., Nath, J o, Sharma, L. R., "Chemisorption of Oxygen on Charcoal and Sorption of Acids and Bases," Ind. Eng. Chemo 50, 1071 (1958). (55) Ross, S., Olivier, Jo P., On Physical Adsorption, Ch. 4 and 5, Interscience Pubishers, New York, 1964e (56) Snow, C, W,, Wallace, D. R., Lyon, L. LE, Crocker, G, R., "Reaction of Carbon Blacks with Oxygen," Proc. Conf. Carbon, 3rd., Buffalo, N. Y,, 1957, pp. 279-87, Pergamon Press, New York, 1959. (57) Soldano, B. A., Boyd, G. E., "Self-Diffusion of Anions in Strong-Base Anion Exchangers," J, Am, Chem. Soc. 75, 6099 (1953). (58) Sondheimer, E,, "On the Relation Between Spectral Changes and pH of the Anthocyanin Pelargonidin 3-Monoglucoside," J. Am. Chem. Soc. 75, 1507 (1953) (59) Standard Methods, 1lth edition, p. 79, American Public HealtFh Association, inc,, New York, 1960. (60) Stern, K. H., Amis, E. S., "Ionic Size," Chem, Rev's. 59, 1 (1959) (61) Stokes, R. H. "The Diffusion Coefficients of Eight Uni-univalent Electrolytes in Aqueous Solution at 25~," J. Am, Chem. Soc. 72, 2243 (1950) (62) Studebaker, Me L., "Direct Titration of Carbon Blacks," Proc. Conf. Carbon, 5th, Penn State Univ.., Pa,, 1961, T8pp.'189-97, Pergamon Press, New York, 1963o

151 (63) Walker, P. L., Jr. "Carbon —An Old but New Material," Am. Scientist 50, 259 (1962) (64) Walker, P. L., Jr., Austin, Lo G., Nandi, Se P., "Activated Diffusion of Gases in Molecular-Sieve Materials," in Chemistry and Physics of Carbon, Vol. 2, P. L. Walker, Jr., Ed, Marcel Dekker,T Inc, New York, 1966. (65) Wawzonek, S., "Coumarins," in Heterocyclic Compounds, R. C. Elderfield, Ed., Vole 2, Wiley, New York, 1950, (66) Weber, W. J., Jr., "Adsorption of Alkylbenzenesulfonates on Carbon," Ph.D, Thesis, Harvard Univ., 1962a (67) Weber, W, J., Jr., "Fluid-Carbon Columns for Sorption of Persistent Organic Pollutants," Paper 1-12, Proc. Interno Conf. Water Pollution Res 3rd, Munich, Germany, 1966. (68) Weber, W. J.,Jr., Gould, J. P., "Sorption of Organic Pesticides from Aqueous Solution," Advances in Chemistry Series, No. 60 Chem. Soc. Publications, 1966, (69) Weber, W. J. Jr., Morris, J. C., "Kinetics of Adsorption on Carbon from Solution," J. Sanit, Engc Div. Am, Soc. Civil Engrs 89, SA2, 31 (t193) (70) Weber, W. J., Jr., Rumer, R. R., Jr., "Intraparticle Transport of Sulfonated Alkylbenzenes in a Porous Solid: Diffusion with nonlinear Adsorption," Water Resources Research 1 (3), 361 (1965). (71) Weber, W. J., Jr., Morris, J. C., "Equilibria and Capacities for Adsorption on Carbon," J. Sanit. Eng. Div. Am. Soc. Civil Engrs. 90, SA3, 79 (1964) (72) Weller, S., Young, T. F., "Oxygen Complexes on Charcoal," J. Am. Chem. Soc. 70, 4155 (1948), (73) Wolff, W, F., "The Structure of Gas-Adsorbent Carbons," J. Phys., Chem. 62, 829 (1958)e (74) Wolff, W, F,, "A Model of Active Carbon," J. Phys. Chem. 63, 653 (1959). (75) Zettlemoyer, A. C., "Modern Techniques for Investigating Interactions with Surfaces," Chem. Rev'so 59, 937 (1959). (76) Zettlemoyer, A. Co, Narayan, K. S., "Adsorption from Solution on Graphite Surfaces," in Physics and Chemistry of Carbon, Vol. 2, P. L. Walker, Jr. Ed., Marcel Dekker, Inc., New York, T966,