1 THE EFFECTS OF DISSOLVED OXYGEN DURING THE REACTIVATION OF GRANULAR ACTIVATED CARBON AND LCA IN COMPARING REACTIVATION OF GAC TO VIRGIN GAC REPLACEMENT By THOMAS E CHESTNUTT JR A DISSERTATION PRESENTED TO THE GRADUATE SCH OOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2017
2 2017 Thomas E Chestnutt Jr.
3 This document is dedicated to my mom and dad
4 ACKNOWLEDGMENTS I want to thank everyone who has helped me along this journey. At the top of the list are my parents, Thomas and Mechthild Chestnutt. They have always encouraged me to improve myself, while stressing education throughout. Being the first person in and doctorate is truly humbling when one thinks of the struggles and hardships that my parents and others have gone through to afford me this op portunity. They did not have these opportunities and I am forever grateful to have them. There have been many influences through the years from teachers and coaches along the way that have helped to refocus and inspire me through time. I want to also th ank Dr. B.J. Fregly for advising me during my degree and also challenging me to excel. I owe large amount of gratitude to my advisor Dr. David Mazyck for taking me into his group in 2001 and giving me an opportunity to pursue this dream of mine. Some sixteen years later, he has still been encouraging and helpful in seeing this endeavor realized. I am evermore grateful to him for his belief, support and investment in my doctoral education. I am thank ful for my advisory committee, Dr. Jean Claud e Bonzongo, Dr. P aul Chadik, and Dr. Bin Gao for their important part in this process. I would like to thank Dr Chang Yu Wu for believing that I could come back to finish this journey. Additionally, I would like to thank Barbi Jackson and the rest of the D epartment of Environmental Eng ineering staff for their help through the administrative requirements that are there waiting along the way. Finally, I would like to thank some of the fine students that I have had a pleasure to meet during my time here in th e department. They are Matthew Tennant, Julee McKenzie, Ameena Kahn, Morgana Bach, Jack Drwiega, Jennifer Stokke, and Regina Rodriguez. I do not want to forget one student who was not only a
5 great friend but a tremendous person, who received his degree f rom the Center of Wetlands in 2001, Andres Buenfil. I will never forget you and the positive impact you had on so many people during your short time here on Earth.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 2 RESEARCH OBJE CTIVES AND HYPOTHESIS ................................ .................... 16 Research Objectives ................................ ................................ ............................... 16 Hypothesis ................................ ................................ ................................ .............. 16 3 THE EFFECTS OF DISSOLVED OXYGEN DURING THE REACTIVATION OF GRANULAR ACTIVATED CARBON ................................ ................................ ...... 18 Literature Review ................................ ................................ ................................ .... 18 GAC: Preferred Type of Activated Carbon ................................ ...................... 18 Reactivation ................................ ................................ ................................ ...... 18 Calcium Catalysis during Reactivation ................................ ............................. 20 Oxidants ................................ ................................ ................................ ........... 21 Adsorption ................................ ................................ ................................ ........ 22 Surface Chemistry ................................ ................................ ............................ 23 ................................ ................................ ......................... 25 Electron Donor Acceptor Theory ................................ ................................ 28 Surface functional groups ................................ ................................ .......... 29 Manipulation of Activated Carbon Surface ................................ ................. 30 MIB ................................ ................................ ................................ ............ 31 Rapid Small Scale Col umn Tests (RSSCTs) ................................ .................... 32 Experimental ................................ ................................ ................................ ........... 34 Spent Granular Activated Carbon ................................ ................................ ..... 34 Reactivation of GAC ................................ ................................ ......................... 34 Reactivation Protocols ................................ ................................ ...................... 36 Steam Curing Reactivation ................................ ................................ ........ 36 Steam Curing Plus Ramping Reactivation ................................ ................. 37 Steam Pyrolysis ................................ ................................ ......................... 38 Rapid Small Scale Column Tests (RSSCT s) ................................ .................... 38 GAC Particle Size for RSSCT ................................ ................................ .......... 39 RSSCT Water Source ................................ ................................ ...................... 40 Pore Size Distribution and Surface Area Analysis ................................ ............ 40
7 pH of Point of Zero Charge ................................ ................................ ............... 41 Bhm Titration ................................ ................................ ................................ .. 41 Results ................................ ................................ ................................ .................... 42 Review of Reactivation Protocols ................................ ................................ ..... 42 Dissolved Oxygen Content ................................ ................................ ............... 43 Mass Loss, Volume Loss, and D ensity ................................ ............................. 44 BET Surface Area, Mass and Volume Loss, and Apparent Density ................. 44 pH pzc (Point of Zero Charge) ................................ ................................ ............ 47 RSSCT Data ................................ ................................ ................................ ..... 48 Steam Curing Reactivated Carbons ................................ ........................... 49 Steam Curing Plus Ram ping Protocol ................................ ....................... 50 Steam Pyrolysis Protocol ................................ ................................ ........... 52 Comparison of Reactivated Carbons Using Varying Concentrations of DO to Conventional Reactivation and Virgin GAC ................................ ............... 53 Role of Surface Chemistry ................................ ................................ ................ 54 Conclusion ................................ ................................ ................................ ........ 56 4 A LIFE CYCLE ASSESSMENT AND COMPARISON OF GAC REACTIVATION TO VIRGIN GAC REPLACEMENT ................................ ................................ ......... 59 Literature Review ................................ ................................ ................................ .... 59 Experimental ................................ ................................ ................................ ........... 62 LCA ................................ ................................ ................................ .................. 62 Scope of Study ................................ ................................ ........................... 63 Study Assumptions ................................ ................................ .................... 65 Reactivation Protocols ................................ ................................ ...................... 67 Surface Area Analysis ................................ ................................ ...................... 68 Reactivation Cost Analysis ................................ ................................ ............... 68 Activation Cost Analysis ................................ ................................ ................... 69 Transportation Emissions ................................ ................................ ................. 70 Reactivation Emissions ................................ ................................ .................... 71 Activation Emissions. ................................ ................................ ....................... 72 Total Emissions. ................................ ................................ ............................... 72 Results ................................ ................................ ................................ .................... 73 LIST OF REFERENCES ................................ ................................ ............................... 81 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 89
8 LIST OF TABLES Ta ble page 3 1 Hillsborough Water Sample ................................ ................................ .................... 40 3 2 Reactivation Protocols and notation ................................ ................................ ........ 43 3 3 The dissolved oxygen content is shown for each reactivation ................................ 44 3 4 Mass and volume loss and apparent density for reactivations ................................ 46 3 5 Performance of Virgin and Low DO Reactivated Carbons A data from MacKenzie ................................ ................................ ................................ ......... 53 4 1 Reactivation Experimental Data and Functional Unit ................................ .............. 74 4 2 Operation and Maintenance Costs for Each Protocol ................................ ............. 75 4 3 Total Costs for Each Reactivation Protocol ................................ ............................. 75 4 4 Experimental Emission Values for Each Reactivation Protocol ............................... 76 4 5 Total Costs for GAC Use Without Reactivation (Scenario 1) ................................ .. 76 4 6 Total Emissions from Each Scenario ................................ ................................ ...... 79 4 7 Global Warming and Acid Rain Pollution Potentials ................................ ................ 80
9 LIST OF FIGURES Figure page 3 1 pinene molecular dimensions ................................ ................................ 26 3 2 Pendleton et al. diagra m of MIB adsorption process ................................ ............. 27 3 3 Possible structures of acidic surface oxygen groups ................................ .............. 29 3 4 1.Stainless steel pressure vessel. ................................ ................................ ........... 36 3 5 Point of zero charge for all reactivated carbons ................................ ...................... 47 3 7 RSSCT data for carbons produced via Steam Curing plus Ramping reactivation protocol ................................ ................................ ................................ ............... 51 4 1 Level 1 diagrams for scenari os 1 and 2. ................................ ................................ 64 4 3 Level 2 diagram for reactivation protocols ................................ .............................. 67 4 4 Impact Stressor Matrix for Impact Valuation ................................ .......................... 80
10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE EFFECTS OF DISSOLVED OXYGEN DURING THE R EACTIVATION OF GRANULAR ACTIVATED CARBON AND LCA IN COMPARING REACTIVATION OF GAC TO VIRGIN GAC REPLACEMENT By Thomas E Chestnutt Jr. December 2017 Chair: David W. Mazyck Major: Environmental Engineering Sciences Rapid small scale column tests (RSSCTs ) were employed to evaluate the impact of the dissolved oxygen (DO) concentration in the water used to create steam on the reactivation of spent granular activated carbon (GAC) T hree novel thermal reactivation procedures: steam curing, steam curing with r amped temperature, and steam pyrolysis reactivation were employed in this work Performance testing of the carbons for removal of the taste and odor causing compound 2 methylisoborneol (MIB) showed that MIB uptake generally increased as the DO concentrati on decreased. A decrease in MIB removal with an increase in surface acidity, a phenomenon found in the literature, may be responsible for the changes in adsorption performance, as the higher DO concentrations yielded carbons with higher total surface acidi ty In addition, the steam curing plus ramped temperature process, which was implemented at 375 (i.e. a bout 400 lower than typical reactivation temperatures) followed by ramped temperature pyrolysis with a low DO concentration (i.e < 4 mg/L) had com parable performance to the virgin carbon counterpart, which could manifest cost savings due to the steam curing associated with this protocol compared to conventional reactivation (i.e., less
11 energy use) Furthermore, since the mass loss associated with t his steam curing protocol was low, less virgin carbon make up would be required also improving the economic viability of this reactivation protocol A life cycle assessment framework was used to analyze the environmental impact associated with the reactiva tion of GAC used in potable water treatment. The layout and fundamental configuration of this research were based upon information from Cincinnati Water Works, a municipal water treatment facility in Cincinnati, Ohio. The t hree novel reactivation protocol s that were mentioned in the dissolved oxygen work were evaluated based on measurements of surface area, CO 2 emissions, performance of GAC to remove MIB and energy consumption during reaction The steam pyrolysis reactivation protocol was found to be an economically and environmentally favorable choice. Potable water treatment with reactivation onsite was shown to exhibit significant benefits regarding both economic and environmental standpoints.
12 CHAPTER 1 INTRODUCTION The use of activated carbon for air, water, and medicinal purposes extends so far back into history that its origin is difficult to document. Charcoal was used for drinking water filtration by ancient Hindus in India, and carbonized wood was used as a medical adsorbent and purifying agen t by the Egyptians as early as 1500 B.C.  Modern development and use has been documented more precisely. Activated carbon was first generated industrially at the firs t part of the twentieth century, when carbon activated from vegetable material was produced for use in sugar refining  Ostrejko wa s the first to patent a technique for activating carbon applying elevated temperatures and oxidation  The general techniques Ostrejko patented are still in use today. Activated carbon is widely used in numerous industrial processes, especially when separation of contaminants from fluids is the goal. As eutrophication continues throughout the world, algal blooms, which produce malodorous compounds, will similarly continue to be a problem. To provide quality drinking water, these taste and odor causing compounds should be removed to maintain customer confidence in their drinking water supply. In the United States, a survey of surface water treatment plants found that over 60% experienced taste and odor episodes. Recently, Akron, Ohio had to te mporarily cease operations because they could not properly maintain water quality because of the extent of a taste and odor episod e  The water treatment industry has found granular activated carbon (GAC) particularly effective in controlling taste an d odors adsorbing synthetic and natural organic compounds, and as an excellent substrate for supporting biofilms. It is anticipated that activated carbon use will continue to increase in the future. Although
13 GAC is very effective at purifying fluids, i t does have a finite adsorption capacity, and thus, must be replaced after it becomes exhausted (spent). A decision must then be made to either landfill or incinerate and replace it with virgin GAC or reactivate the spent GAC for reuse. The latter option eliminates solid and some of the air emissions that the other two options would create. The goal in reactivation is to return the spent GAC back to its original (virgin) capacity, and is typically performed in a rotary kiln, multiple hearth, or less frequ ently in a fluidized bed furnace. There are many protocols used to thermally reactivate spent GAC, but they are all very similar in process and design. Typically, reactivation consists of three ordered steps: (1) Drying the carbon at about ating in an oxygen deprived environment referred to as pyrolysis at about and (3) Oxidizing the charred organics via an oxidation step. Typically, H 2 O or CO 2 is employed as the oxidant during this third step, and the temperature is in excess of 700 process are temperature, time, steam to carbon ratio, and sometimes the order in which pyroly sis and oxidation are performed [5 ] Given a basic understanding of how activated carbon is manufactured and reactivate d, it is essential to understand how activated carbon adsorbs compounds. The main characteristics that allow GAC to remove molecules from water or other media are its physical characteristics (e.g., pore size distribution and surface area) and surface che mistry. The physical characteristics and surface chemistry are dependent on the starting material and the development of both is linked heavily to the activation process. Surface chemistry (dominated by oxygen functional groups) plays a major role
14 in the ability of activated carbon to operate efficiently, according to several researchers [6 14] Much of this literature focused on altering the surface chemistry of activated carbon and how these modifications inf contaminants. An intriguing possibility for changing the features of activated carbon could be as simple as changing a characteristic of the water used to produce the steam used for creation of pores during activatio n or for oxidizing char during reactivation. The proposed component to be changed is the dissolved oxygen concentration of the water used for steam production, and it was hypothesized that this modification would impact the types and/or quantities of oxyg en functional groups and perhaps the surface area/pore size distribution of the reactivated carbon. As stated above, the oxygen functional groups dominate the surface chemistry of activated carbon. Therefore, changing the concentration of oxygen function al groups or the type of oxygen functional activated carbon sorbs compounds. The ability to easily change the surface chemistry would allow for more flexibility in the way activated carbon is produced or reactivated. Moreover, this could also be considered a fundamental change in the way activated carbon is generated. Until now, an idea such as this has yet to be presented in the literature. One could envision this a s being a simplistic method for altering activated carbon production at an industrial level without the need to make costly changes to existing infrastructure. This proposal is built upon work done by Mazyck and Cannon [15, 16] while performing their research on methods to overcome the deleterious effects of calcium catalysis during thermal reactivation.
15 The proposed work will seek to achieve an understanding on a microscopic and macroscopic level. Herein, the microscopic level refers to the physical and surface characteristics of the carbon for MIB removal, while the macroscopic level is an evaluation of the larger effects of the reactivation process (e.g., energy consumption, emissions, acid rain potentials, etc.). It is essential t o have an understanding about the interactions occurring in the environment as a result of new technology. The ability to take a step back from a process and attempt to observe the broader scope of events is critical to the advancement of technology. One such tool that may aid in this evaluation is a life cycle assessment (LCA). An LCA is an evaluation of the environmental effects associated with any given activity from the gathering of raw material to the point at which all residuals are returned to the earth  Life cycle assessments are employed to identify and measure the direct and indirect impacts associated with an activity  The LCA will scrutinize three reactivation protocols and then, using the best one, compare reactivation of activated carbon with the process of using only virgin activated carbon at a water treatment facility. Thus, the thesis presented will show novel innovative, and practical methods for the production and assessment of activated carbon.
16 CHAPTER 2 RESEARCH OBJECTIVES AND HYPOTHESIS Research Objectives Objective 1. Create reactivated bituminous coal based carbons with differing characteristics (i.e ., pore size distribution and surface chemistry) via a manipulation of the dissolved oxygen concentration in the water used to produce steam for reactivation. Objective 2. Investigate how varying the dissolved oxygen concentration in the water, which wa s used to create steam, changes the properties of the reactivated granular activated carbon. 1. Determine if steam produced from certain dissolved oxygen concentrations contributes to a change in physical and/ or surface characteristics of the activated carbon 2. If physical properties are being altered, then the physical attributes of the carbon will be examined and a possible mechanism will be discussed. Objective 3 adsorption of a taste and odor co mpound, namely 2 methylisoborneo l (MIB) Objective 4 Show a comparison of the role dissolved oxygen plays during the reactivation of spent GAC and on a virgin GAC. Objective 5 Assess the environmental and economic impact of reactivation and activatio n of GAC via a life cycle assessment (LCA) as a measuring tool. Hypothesis Altering the DO of the water subsequently heated for steam will modify the surface chemistry of the activated carbon. Indeed, introducing oxygen gas during reactivation could cause a slight enlargement of the carbon pore structure. For example, Wigmans reported that O 2 ca used an increase in the size of pores when compared to carbons activated with steam and /or carbon dioxide  Although slight
17 modifications of the pore size distribution ( PSD ) may occur, the primary manifestation from modified DO will change the oxygen functional group concentrat ion on the surface The DO will also prove to be important in the removal of contaminants (e.g. 2 methylisoborneol, MIB). In a rapid small scale column test (RSSCT), the ability of different carbons to remove MIB will be tested. It is hypothesized tha t an inverse relationship between the DO concentration used for reactivation and carbon efficiency for the removal of MIB will be observed. This will be influenced by the increase in Lastly, the L CA framework will be employed to analyze the environmental impact associated with the reactivati on of spent GAC. T hree different reactivation protocols will be evaluated (steam pyrolysis, conventional, and steam curing) based on measurements of surface ar ea, CO 2 emissions, performance of GAC, and energy use. It is believed that the LCA will s how that the reactivated carbon, rather than the virgin carbon, will be superior, as measured by the previously given metrics. This will prove to be a significant ben efit for potable water facilities to have reactivation onsite. This will result will also result in larger positive environmental implications. On the front end, potentially less GAC would be needed to be produced and on the back end, potentially less GAC will need to be landfilled.
18 CHAPTER 3 THE EFFECTS OF DISSOLVED OXYGEN DURING THE REACTIVATION OF GRANULAR ACTIVATED CARBON Literature Review GAC: Preferred Type of Activated Carbon In the water treatment industry, both powdered activated carbon (PAC) an d granular activated carbon (GAC) are used. Of these two forms, PAC has been more  PAC is favored for its ability to be variably dosed throughout the year depending on temp oral demand. This allows for flexibility and control when dealing with seasonal contaminants. Additionally, PAC is typically less expensive than GAC, and coupled with its small particle size, as compared to the 1 mm size of GAC, allows for a faster diffusi on rate of contaminants  However, PAC currently cannot be recovered from the sludge that settles during clarification. Hence, it cannot be thermally reactivated. Compared to GAC, the initial cost of PAC is less. However, if one takes into account that GAC has been shown to be effective after six reactivation cycles  the economic argument may be less valid. GAC is used as a solid medium in filterbed adsorbers, often replacing anthracite coal, or it can follow dual me dia filtration as a separate GAC contactor/adsorber. The particle size of GAC used commercially, typically ranges between 0.85 mm 1.15 mm (U.S. Standard 12 x 40 mesh). Due to the interest in reactivation, the remainder of this discussion will be focused on GAC. Reactivation Coal based carbons are thermally reactivated by exposing the spent GAC to three main steps: 1. Drying, which removes water and highly volatile organic
19 compounds at temperatures around 110 C; 2. Pyrolysis of nonvolatile adsorbates to form a carbonaceous char on the activated carbon surface. This is performed at temperatures between about 500 800 C and is carried out in an inert atmosphere or oxygen deprived environment; and 3. Oxidation of the pyrolyzed residual is achieved by emplo ying an oxidizing agent such as steam and/or CO 2 at a temperature usually above about 700 C, but less than 900 C [19, 23] Pyrolysis involves charring the precursor or organic matter in the case of sp ent GAC into a disordered carbon structure (char) by the addition of heat, without oxygen, to prevent the carbon from burning (oxidation). Oxidation is where the char is removed and where further pore development occurs. The elimination of carbon atoms f rom a char particle produces CO and/or CO 2 depending upon the reactivation temperature. thus generating porosity  This removal of carbon from the precursor is called [23, 24] gasification by steam is shown as: R C f + H 2 O = H 2 + R C(O) ( 3 1) R C f (3 2 ) CO + R C(O) = CO 2 C f ( 3 3) where, R C f = solid phase free active carbon sites, R C(O) = solid phase active carbon sites occupied by surface oxides, and n = an integer indicating the number o f active sites formed as a result of gasification. It can also be described by: 2C + 2H 2 2 ( 3 4) CO + H 2 2 + H 2 ( 3 5)
20 CO + 3H 2 4 + H 2 O ( 3 6) The above combination of equations can be simplified into the following ov erall equation: 2C + 2H 2 4 + CO 2 ( 3 7) Equa tion 3 4 yields production of methane via gasification of carbon by steam. The interest in this equation also lies in its thermo neutrality. The carbon gasification reaction, Equa tion 3 5, is endot hermic and is balanced out by the exothermic reactions (i.e., Equation 3 4 and 3 6). These equations are known as the water gas shift and methanation reactions, respectively. Since the gasification reaction requires external energy to proceed, this allows a certain amount of control when initiating the reaction. Equa tions 3 4 3 6 are relatively slow reactions, in the absence of a catalyst, but the rates can be somewhat increased by an increase in pressure. Furthermore, Walker et al.  pointed out that the rate of carbon gasification increases sharply with an increase in temperature; however, increasing temperature has a limiting effect on the equilibrium production of CH 4 In order to combat this, one should use a temperatu gasification reaction rates. Calcium Catalysis during Reactivation Cannon et al. and Knappe et al. demonstrated that calcium behaves catalytically during the thermal reactivation of spent GAC [26, 27] GAC pores accumulate calcium as it is processing water during treatment. This accumulation is augmented as calcium complexes with natural organic matter. Moreover, calcium is found naturally in GAC precursors. Regardless of its origin, calcium increases the rate of gasification during oxidation by a factor of 3 to 25  Thus using typical protocols on spent, calcium
21 enriched GAC could cause unfavorable results. Mazyck [ 28 ] elucidated this issue. Furthermore, it was shown that activated carbons produced under such conditions are of poorer quality with fewer micropores and a decreased apparent density. Co nsequently, Mazyck and Cannon  experimentally designed a reactivation p rotocol in response to the issue of activated carbon that is calcium enriched. The corresponding protocol called for an oxidation step (steam as oxidant) at a temperature tempe [15, 16] The protocol controlled the negative effects of calcium catalysis and produced carbons that also exhibited better removal of 2 methylisoborneol (MIB) compared to conventional reactivated and virgin GAC. Oxidants Oxidants are used to create pores in activated carbon or to gasify charred organic matter. Pore sizes are classified in the following categories: micropores (<20), mesopores (20 to 500 ), and macropores (>500 ). Of the three oxidants stated above, steam and carbon dioxide are most often used to activate and reactivate the  found that using steam, when activating brown coal chars, created an activated carbon where approximately a third of the total adsorptive pore volume consisted of micropores. The same researchers found that using carbon dioxide produced an activated carbon containing two thirds of the pore volume in the micropores  This is contrary to what Wigmans reported. Wigmans found steam produced more pores in the micropore region rather than the mesopore region  Other researchers have produced results that parallel the findings of Tomkow [30 33] Specifically, Molina Sabio et al. found that carbon dioxide and steam both produced micropores using olive pits. Yet, carbon dioxide exhibited a larger micropore volume
22 and a narrow micropore distribution, when compared to carbon that was activated by steam. Moreover, a widening of the pore struct ure was observed by steam activation  This controversy complicates the evaluation of the reactivated c arbons, for it is not clear what impact steam has on PSD. Furthermore, the steam herein will contain various concentrations of DO. Adsorption interface as compared with the ne ighboring phases is referred to as adsorption. The fundamental concept in adsorption science is the adsorption isotherm. The adsorption isotherm is the equilibrium relation between the quantity of the adsorbed material and the concentration or pressure i n the bulk fluid phase at constant temperature  Intertwining theory and experimental practices has developed adsorpt ion science. The interpretation of adsorption isotherms was elucidated by the Freundlich  Langmuir  and Polanyi  equations. It is possible to breakdown molecular adsorption into five basic steps: 1) Diffusion of reactants to the active surface; 2) Adsorption of one or more reactants onto the surface; 3) Surface reaction ; 4) Desorption of products from the surface; 5) and Diffusion of the products away from the surface  There exist two principal m odes of adsorption: chemical adsorption or chemisorption and physical adsorption or physisorption. The physical adsorption process has bonds, which are similar to those found in liquids, namely van der Waals forces. The energies of such associations are relatively weak and seldom exceed 5 kcal/mol. Physisorption generally occurs at low temperatures where adsorption is easily reversible  In physical adsorption, there is no significant redistribution of electron d ensity either in the adsorbate or substrate. It is worth noting that several layers may be physically
23 adsorbed on the surface of the substrate via physical adsorption. Due to the weak bonds formed between adsorbate and substrate, adsorbate adsorbate inter actions tend to dominate over adsorbate substrate interactions. Physical adsorption is characterized by a decrease in entropy and free energy of the adsorption system; therefore, this process is exothermic. The van der Waals forces give rise to a shallow minimum in the potential energy (PE) curve at a relatively large distance from the surface (typically d > 0.3 nm) before the strong repulsive forces arising from electron density overlap to cause a rapid increase in the total energy  Chemical adsorption involves a rearrangement of electron density between the substrate and adsorbate. While physical adsorption can have multiple layers of adsorbate, chemical adsorption may only occur as a monolayer  The range of these bonds is found between completely covalent or completely ionic bonding. The energies involved in chemisorption may be significant (10 to 150 kcal/mol). Indeed, the bond between adsorbate and substrate has a more permanent character than that of physical adsorption. Therefore, desorption requires high temperatures. A discrepancy in activation energies between adsorption and desorpti on can result in hystersis  its physical adsorption properties. However, recently researchers have shown the im portance of surface chemistry affecting the adsorption process [11, 13, 14, 41] Surface Chemistry In order to discuss surface chemistry, one needs to have some idea of the surface being observed. The structure of activated carbon is an essential determinant of functional groups that can exist on the surface. First, one should consider the structure of i deal graphite, because of its close resemblance to activated carbon at the
24 fundamental level. A graphite structure consists of a system of layers of fused carbon hexagons in a parallel arrangement that exists in a regular fashion (graphitic layers or basa l planes). The basal planes of graphite are flat and ordered parallel to and equidistant from one another to form crystallites. Activated carbon has a similar basis; however, it is irregular in comparison with the graphite structure. This irregularity i s due to impurities and vacancies in the carbon make up. The graphitic like layer of activated carbon can be thought of as a very large polynuclear aromatic molecule  The bonds to three adjacent carbon atoms and the fourth bonding or sp 2 hybridization. The layers are approximately 3.54 apart. This stacking distance becomes altered and perturbations occur in the carbon sequencing ( e.g., impurities and/or vacancies)  These impurities or vacancies cause layers to tilt with respect to one another. The layers are held together by weak van der Waals forces. Additional factors leading to a more disorganized matrix follow: tetrahedral bonding that cross link different layers; atoms other than carbon incorporated within the graphitic l ayers that form heterocyclic ring formations; and atoms that may bond at the basal plane edges creating functional groups [43, 44] It is the increased reactivity from free valences or unpaired electrons at the edge of the basal planes that permit suitable external atoms to form compounds onto the c arbon crystallite. Graham  in 1955, found that acidic oxygen functional surface groups tend to reduce the capacity of carbon surfaces to adsorb mantanil yellow from an aqueous solution. The Graham paper is an early discussion, in the literature, a ttempting to show oxygen functional groups as being pertinent in the discussion of adsorption or compound removal from a fluid state onto activated carbon. There exist
25 two major theories in what occurs on the surface of carbon during adsorption. The two Clearly, there exists merit in both approaches, hence neither can be ruled as absolute. Researchers continue to be split on exactly what is occurring, but to better u nderstand what is taking place one must understand the two approaches. interactions on surface interactions as they occur on activated carbon. Specifically, Coughlin et al. [43, 46] discussed the adsorption of aromatic compounds, phenol and nitrophenol, onto activated carbon. In the study, activated carbons were e ither oxidized with (NH 4 ) 2 S 2 O 8 or reduced by mixing carbon with a zinc amalgam and covering with concentrated HCl. The researchers used BET surface area to give some measure of comparison between the subsequent oxidized/reduced carbons. A result, in a di lute concentration equilibrium study, showed that even though the reduced carbon had a significantly lower surface area (150 m 2 /g); an increase of two times the phenol uptake when compared to the oxidized version was observed. However, when performed at a higher concentration (two orders of magnitude), both oxidized and reduced carbons achieved similar uptake. The conclusion drawn was that chemisorbed oxygen seemed to strongly influence phenol adsorption, in dilute concentrations, when the molecules are t hought to adsorb in the prone position on the basal planes, where attractive forces can operate over the entire phenolic nucleus. These same forces seem to have a less significant effect under conditions when the molecules are thought to adsorb in the ver tical or end on position. In the vertical case, the interaction between molecules of phenol is presumed to affect the process more than the interaction between the molecule and activated
26 carbon. One should consider the explanation put forth by the resear chers. The decreased performance exhibited by the oxidized carbon, in the dilute phenolic concentration, can be a result of oxygen functional groups bound on the edges causing electrons to localize in surface sta between phenol and activated carbon basal planes. Coughlin and Ezra  also pointed out that Dubinin theorized another contributing phenomen on. Dubinin  suggested that water molecules sorbed to the oxygen groups became secondary adsorption sites that retain additional water molecules by hydrogen bonding. The migration of organic molecules onto the activated carbon surface could be stericly hindered by the se complexes. This notion that hydrophilic or oxygen functional groups can attract water molecules has been reported by many researchers since Dubinin [11, 48, 49] Pendleton and coworkers  attempted to elucidate the issue (see Figure 3 1) by using two similar bicyclic pinene and 2 methylisoborneol (MIB). Figure 3 pinene molecular dimensions 
27 Pendleton concluded and illustrated (Figure 3 2) that the oxygen groups (hydrophilic sites) on the surface build up water clusters around the pores, thus making i Figure 3 2. Pendleton et al. diagram of MIB adsorption process  The issue with this approach lies in the number of hydrophilic sites on an activated carbon. As stated earlier, oxyge n functional groups can be found chemisorbed onto the edges of basal planes and where vacancies exist. Thus, it is difficult to fathom that steric hindrance could solely account for the differences. Moreover, one might also take issue with the explanatio pinene adsorption. Leng and Pinto [51 ] as well as Menendez  point out that water electron density on the basal planes, consequently weakening nonspecific dispersion interactions [43, 52] Because of the electron withdrawing effect of water adsorption and MIB being a nonpolar compound, it is likely that MIB a dsorption would be hindered  Mueller and Gubbins  contend that just a small amount of hydrophilic sites on the carbon surface can
28 have a remarkable effect on the amount of or site density of water adsorption  Following the Pendleton paper, Considine and Pendleton et al.  published results that showed perhaps the solvent is more of a factor. The study was performed with MIB in dichloromethane (DCM). A less polar solvent, DCM, produced similar MIB uptake results for a hydrophobic and a hydrophilic carbon. Electron Donor Acceptor Theory interactions papers discussed above, readers were presented an alternative theory, electron donor/acceptor theory. Mattson et al.  paid close attention to the pH, as it was asserted that not only does pH control the factors affect adsorption. The study again looked at adsorption of aromatic compounds. They found that nitrogen containing aromatics adsorbed more strongly than did phenol. Nitro groups are strong electron acceptors; therefore, create strong complexes with the oxygen groups as being attributed to strongly acidic carboxyl groups. It was also suggested that phenol forms a donor acceptor complex via a mechanism involving carbonyl oxygen of the carbon surface acting as the electron donor and the aromatic ring of the absorbate as the acceptor. Attent ion to this split in agreement over phenol adsorption was subsequently addressed by Radovic. Radovic outlined these theories in his comprehensive review   categorically ruled out the donor acceptor argument in their study of p nitrophenol adsorption in aqueous solution.
29 Surface functional groups Figure 3 3. Possible structures of ac idic surface oxygen groups: A) carboxyl groups; B ) phenolic groups; C) carboxylic anhydrides; D) lactols; E ) carbon y l groups; F) ether groups; G) lactone groups; H ) quinine groups  Long have the existence of surface functional groups such as carboxyls, phenols, lactones, and acidic anhydrides been shown as sources of surface acidity, but still today they are not completely understood. The basic character of the carbon surface is even less understood. It has not been clearly defined if oxide structures are responsible for basic character shown by some carbons. There exists a view th at progressive
30 localization of basal plane electrons by oxides formed on the basal edges causes Lewis type basic sites to form with the carbon structure itself  Some of the acidic groups 3 3  Manipulation of Activated Carbon Surface The ability to modify the activated carbon surface can prove advantageous when desiring a certain characteristic for a mass transfer process ( e. g., potable water by chemical or physical activation/reactivation. Alterations to activated carbon via chemical means are relatively new, when compared to physical means. Moreno were oxidized with different agents in an attempt to introduce various oxygen surfaces. The agents used were H 2 O 2 (NH 4 ) 2 S 2 O 8 and HNO 3 The results of the oxi dations showed: the highest total oxygen content was obtained after HNO 3 and the lowest by (NH 4 ) 2 S 2 O 8 ; (NH 4 ) 2 S 2 O 8 and H 2 O 2 provided the largest number on the external surface of the carbon and HNO 3 placed the highest number on the internal surface; the hi ghest total surface acidity was achieved from HNO 3 and lowest by (NH 4 ) 2 S 2 O 8 Despite having the lowest total surface acidity, (NH 4 ) 2 S 2 O 8 had the highest acid strength. This was attributed to closeness of carboxyl groups on the surface  Menendez et al.  used physical means to alter surface characteristics. Treatments using hydrogen and nitrogen gas were performed in an attempt to create bas ic carbons. They steadily increased the temperature at which the carbon was exposed to the two gases (500 to 950 C). The study provided interesting trends. Carbon exposed to nitrogen gas adsorbed more oxygen than did carbon exposed to the hydrogen gas treatment. Moreover, as the temperature was increased, the nitrogen
31 treated carbon adsorbed more oxygen. The inve rse was true of the hydrogen treated carbon. As the temperature increased, subsequent exposure to ambient conditions resulted in less and less oxygen being chemisorbed onto the carbon surface. They also found that high temperature treatment in an inert e nvironment effectively removed oxygen functional groups, but left behind very reactive carbon sites. These sites are compelled to adsorb oxygen at room temperature, thus becoming acidic once again. Hydrogen treatment, as the temperature increased, left a n increasingly stable basic surface structure, which is almost free of those reactive carbon sites capable of adsorbing oxygen at ambient conditions  Similar results were found by Nowack, Cannon, and Mazyck  g the oxidation step of reactivation. It is proposed that increased dissolved oxygen concentrations will interfere with the ability of hydrogen to anneal the reactive carbon sites created on the carbon surface during reactivation. In other words, dissolv ed oxygen interferes with the water gas shift reaction. Moreover, oxygen functional groups will form on the carbon surface; thus, stabilizing these reactive carbon sites. On the contrary, the annealing process is expected to occur at a higher rate when t he dissolved oxygen levels, which the carbon is exposed to, are low or close to zero. MIB MIB is an odorant that poses a significant issue for water treatment facilities. The structure of MIB (Figure 3 1 ) consists of a bicyclic hydrocarbon ring with one h ydroxyl group and three methyl groups. Pendleton et al. considered it to be hydrophobic in nature For purposes of adsorption, MIB is approximated as having a spherical shape,
32 and has a diameter of approximately 6 angstroms ()  The molec ular weight of MIB is 169 g/mol. Many researchers have studied methods for the removal of taste and odor causing compounds. Typically, the methods of choice for successfully removing these taste and odors to acceptable concentrations are the use of PA C or GAC as shown by several researchers [20, 63] Both are very effective for removing these compounds due to their high specific su rface areas (500 1000 m 2 /g), large adsorption capacities, high degree of surface reactivity, and their porosity. Rapid Small Scale Column Tests (RSSCTs) Generally, tests consist of batch equilibrium studies with PAC, full scale evaluation of GAC, and pil ot scale studies of GAC. RSSCTs deliver another way to test the effectiveness of GAC that is not frequently used. RSSCTs have an advantage over pilot studies because (1) RSSCTs can be done in a fraction of the time; (2) A smaller volume of water is requi red for testing; and (3) Extensive isotherm studies are not required to predict the performance in full scale  Although the RSSCT can significantly reduce the time and cost of the study, their limitations are that they can over predict GAC performance in full scale  This over predicting nature of the RSSCT is tolerable though since the relative comparison between carbons does not change. Crittenden  discussed the RSSCT, which incorporates scaling equations to assess the performance of large scale fixed bed adsorbers. Parameters required for the design are: the empty bed con tact time (EBCT), adsorbent particle size, and hydraulic loading rate. Another important factor is the relationship between intraparticle
33 if one does not use intrap article diffusivity with dependence on the particle size. Conversely, if the particle size is dependent on the intraparticle diffusivity, then proportional diffusivity parameters must be employed  The equations f or constant diffusivity are shown below where EBCT SC and EBCT LC are the empty bed contact times for the RSSCT and full scale design, respectively. R SC (RSSCT) and R LC (full scale design) are the adsorbent particle radius and V SC and V LC are the hydraulic loading rates for the RSSCT and full scale design respectively  SC ) an d full scale design (Re LC ) are also included in the equations to aid in scaling the length of the column. 3 8 3 9 The RSSCT column can be designed according to the assumption that the minimum column diameter to particle size rat io should be 50 or greater to avoid channeling  An example of the employment of this technology is work done by MacKenzie  GAC was looked at a plausible alternative to commonly used PAC to combat odor episodes associated with MIB. RSSCTs were used to compare a commercial, conventionally reactivated, steam cured reactivated, and steam pyroly sis reactivated GACs for their ability to remove MIB and its retention. MIB was spiked into two natural clarified waters, both dominated by <500 (MW) size fractions and no noticeable difference in other size fractions. The existence of an emphatic dispar ity in carbon
34 performance for MIB removal was discovered. Thus, a conclusion was formed that performance testing for the removal of a target compound is a must. The steam pyrolysis reactivation protocol, developed by MacKenzie, produced carbons that were better in removing the target compound, MIB, than the other protocols compared. Finally, no correlations were found to support conflicting literature of physical versus chemical interactions. Experimental Spent Granular Activated Carbon The spent GAC that was used for this research was provided by the Hop e well Water Treatment Facility in Hopewell, Virginia S pent GAC originated from a commercially available GAC precursor, which was a bituminous coal based carbon ( F 400 (Filtrasorb 400), a product of C algon Carbon Corporation ( Pittsburgh, PA )) and it had been reactivated six times upon reception of the material. T he spent GAC was received wet and was dried at 110 C for 24 hours. Dr ied carbon was sieved, producing a 16 x 20 (1.18 mm x 0.850 mm) mesh f raction The sieving step is done to remove variability between reactivation experiments. L astly the GAC was placed into a desiccator until needed. Reactivation of GAC A fluidized bed furnace similar to that used by Mazyck  was employed to reactivate the spent GAC. In preparing for a reactivation, the steam/gas preheating furnace (Barnstead Thermolyne) was set (Applied Test Systems) tube furnace was preheated to a desired temperature for the particular reactivation being performed. Deionized water used to form steam was contained in a stainless
35 steel pressure vessel and pressurized with nit rogen, oxygen, or air at 10 psi depending on the desired dissolved oxygen concentration. The DO was measured prior to each reactivation via a quasi inline method. A gravimetric flask ( 250 mL ) was connected to the out let hose of the pressure vessel used in the reactivation process mentioned above. A heavy duty dissolved oxygen meter (Extech Instruments model 407510) was used to measure the concentration in mg/L. The probe was placed inside of the flask hovering approximately 1 inch from the bottom of t he flask. A stir bar was also on the bottom of the flask, which rested on a magnetic stirr er The top of the flask was covered with parafilm, as to create some sort of seal from the outside environment. Samples were taken by turning on the desired gas ( O 2 N 2 or air), depending on the dissolved oxygen concentration desired, and because the gravimetric flask was open to the atmosphere a pressure drop was created forcing water from the vessel into the flask. Therefore, a dissolved oxygen measurement coul d be taken from a quasi inline sample. Measurement time (30 s) was from the time gas tanks were opened to the time a reading was taken. Measurements were made in duplicate. Upon taking the DO measurement, reactivation gases controlled by Scienceware Rit eflow rotameters were used to direct gas flow (1.0 cc/min) into the reactivation system. A type K thermocouple was used to measure the temperature inside of the one inch diameter fluidized bed contained in the tube furnace. Dried, spent GAC (10 g) was in troduced by pouring it down the top of the quartz column. GAC came to rest upon a quartz porous frit 2/3 of the way down the column; however, due to the gas flow the carbon became fluidized. At the completion of the reactivation process, the sample
36 was c ooled in nitrogen and the apparent density was calculated using an approach similar to ASTM D 2854 83 The depiction of the reactivation system is shown in Figure 3 4. Figure 3 4 1. Stainless steel pressure vessel; 1a. Gas inlet from tanks (not shown); 1b.Water outlet to flow meter; 2. Scienceware R i teflow flow meters control gas/water from gas tank or pressure vessel to the preheating furnace; 3. A quartz column containing porous frit (where the carbon becomes fluidized and reactivated); 4. The preh eating furnace changes the water to steam and sends it to the reactivation furnace; 5. Clam shell furnace is used to provide the desired temperature needed for the reactivation. Reactivation Protocols Steam Curing Reactivation The motivation for this reac tivation protocol was that Mazyck  reported that gasification for spent activated carbons proceeded optimally in the range of 350 1 1 1b 2 3 4 5
37 430 to avoid the formation of char, but rather the direction gasification of the natural organic matter that accumulated during the GAC service li fe The importance of temperature being that H 2 is optimally produced from the water gas shift reaction. H 2 can help anneal carbon active site s upon removal of functional groups from the carbon surface  The reactivation procedure mentioned above was followed and the temperature inside the tube furnace was placed at 375 Only an oxidation stage was performed. Steam was used as the oxidant and the time of exposure was one hour. Three reactivations at this temperature were performed using a different dissolved oxygen concentration for each experiment (i.e. <5.0, 6.0 8 .0, and >12.0 mg/L) Steam Curing Plus Ramping Reactivation This protocol was one of the major points found in the dissertation from Mazyck  directed at overcoming calcium catalysis during thermal reactivation (i.e., the carbons pore structure could be maintained even in the presence of calcium). Herein, t he spent carbon was fluidized under a flow of dissolved oxygen controlled steam for 2 was introduced as the temperature took 70 minutes wh ich was limited by the furnace This procedure was repeated three times one for each dissolved oxygen concentration (i.e. <5.0, 6.0 8.0, and >12.0 mg/L) The procedure was used to see how changing the dissolved oxygen concentration would alter the carbo functional group properties. Also, comparisons could be made between the ramping protocol and the steam curing protocol, because the only difference in the two protocols was the second phas e of the steam curing procedure (i.e., th e temperature ramping step)
38 Steam Pyrolysis MacKenzie  found th e steam pyrolys is procedure to out perform the steam curing procedure as well as comme rcially produced virgin carbons for the removal of MIB f r om two clarified waters. It was a shorter protocol than the other two in a temporal sense. The first phase of the process called for a reactivation temperature of 750 under a flow of steam for fifteen minutes. Immediately following, the second phase began with a flow of N 2 at 750 Rapid Small Scale Column Test s (RSSCT s ) RSSCT s were used to evaluate the performance of the reactivated carbons for their ability to remove MIB from clarified water. The RSSCT was designed to approximate full scale design parameters and sized using the scaling equations available in the literature ( [64, 6 8] ). The RSSCT column was manufactured from acrylic (1.5 cm x 0.41 cm), with inserts made of Teflon to avoid the sorption of MIB onto system components. It was previously determined that MIB does not sorb to PTFE, as it does to other materials (e.g., silicone). Also, a Master flex L/S rigid PTFE pump head, PTFE tubing, and PTFE fittings were used throughout the system to avoid sorption. A rate similar to that of MacKenzie  was used. The system was modeled in a range that could be conceivably used at a water treatment plant A proper flow rate ( 5 mL/min ) for the peristaltic pump was chosen The reactiva ted carbons were evaluated using the RSSCT by spiking aliquots of clarified water (stored in a 5 gallon high density p olyethylene Nalgene tank) with radio labeled MIB ( 80 ng/L ). The detection limit of the liquid scintillation counter was 3 ng/L. The radi o labeled 14 C MIB used in the study was purchased from American Radio labeled Chemicals, Inc. (St. Louis, MO). A Packard Tri Carb 4530 liquid scintillation
39 counter (LSC) (Meriden, CT) was used to analyze the MIB. The delivered 14 C MIB was in a methanol s olution, and had a specific activity of 55 mCi/mM. A stock solution ( 5 ) was made, and 105 14 C MIB/L of clarified water was the ratio used for the similar study by Mac Kenzie  methanol was predicted to cause pocketing of 14 C MIB in the solution that co uld have an effect on the reproducibility of the column tests. A filter of fiberglass wool, proven not to sorb MIB, was inserted at the exit of the reservoir in order to prevent particulate matter from clogging the mini column. Influent and effluent MIB concentrations were examined by Teflon two way stopcock sampling ports placed before and after the column. S amples (3mL) were extracted using a gas tight syringe ( 5 mL, Hamilton Series 1000) The sample was then added to Scinti Safe liquid scintillation cocktail ( 18 mL ) in liquid scintillation vials ( 20 mL ) Quantitative measurement of radioactivity in each sample was based on a standard curve of known concentrations (0, 12.5, 25, 50, 100, and 200 ng/L MIB). A linear standard curve and a correlation of 0.99 allowed the samples to be confidently converted from disintegrations per minute (DPM) to ng/L. GAC Particle Size for RSSCT The reactivated carbons used for these experiments were sieved down to meet the requirements of the RSSCT design of particle size range to column diameter. The determined GAC particle size required a 170 x 200 US Standard mesh or a mean particle size of 82 carbon was passed through the 170 mesh and collected on the 200 mesh sieve by manual shaking. The remaining carbon on the 200 mesh sieve was rinsed with distilled
40 water to remove any excess carbon fines, and then dried in an oven overnight at 105 o C. The carbon was stored in glass vials and placed in a dessicator when not in use. RSSCT Water Source Hillsborough County Water Treatment Plant, located in Tampa, Florida, provided the water used in the performance testing. The water used for this experiment was collected from the clarifiers on October 15, 2001. This plant also treats their influent water by conventional treatment. Water characteristics are shown below. Table 3 1. Hillsboro ugh Water Sample Date of Collection 10/15/2001 Location of Sample Plant Influent Plant Effluent pH 6.75 8.23 TOC (mg/L) 34.1 4.5 Alkalinity (mg/L as CaCO 3 ) 79 80 Color (pcu) 300 5 Turbidity (NTU) 1.4 0.1 Pore Size Distribution and Surface Area Anal ysis A Quantachrome gas sorption analyzer (Autosorb 1) was employed to determine the pore structure of the GAC samples. GAC ( 15 mg ) was outgassed at 11 0 o C for 24 hours, and then encased in liquid nitrogen at 196 o C (77 K). Subsequently, known volumes o f N 2 gas were dosed into the test container at approximately 90 equilibrium pressures ranging from 10 6 atm (10 4 kPa) to 1 atm (101 kPa). The density functional theory (DFT) (< 20 ) and Barret, Joyner, and Halenda (BJH) (> 20 1000 ) theories were appl ied to the measured relative pressure data to compute a pore size distribution for each sample. This combination of models was employed as a response to well known literature stating that the BJH model does not reliably predict data under the 20
41 size po re or micropore region  and the DFT model does not accurately predict pore volume in pores greater than about 70 [6 9 ]. pH of Point of Zero Charge The method used to carry out the pH pzc analysis was similar to the method used by Menendez et al.  The method is referred to a s a reverse mass titration usi ng the 10% by weight GAC fraction to do the analysis. In previous studies, the 10% by weight GAC fraction was observed to be the optimal point or where the zero slope occurred consistently; therefore it was considered the pH pzc for this study. GAC was dri ed at deionized (nanopure) water was purged under N 2 gas for 20 minutes. GAC (0.5 g) was measured out and placed into an Erlenmeyer flask (25mL) along with a stir bar. Next, N 2 purged nanopure water (4.5 mL) wa s drawn up using a pipette and expressed into the Erlenmeyer flask. The stopper was immediately placed on top and then sealed with para film. The flask was placed on a magnetic stirrer and mixe d for 24 hours. A pH probe was then used to determine the pH of the GAC and in turn the pH pzc Bhm Titration The concentration of acidic oxygen complexes created on the surface of the reactivated carbon samples was determined according to the B hm titration method  An allotment of 0.5 g of carbon and 25 mL o f 0.05 N NaOH were shaken in sealed vials for three days. The suspensions were then filtered using a 0.45 mm filter and 20 mL of the filtrate was titrated with 0.1 N H2SO4 to a pH of 4.5. The amount of base consumed by each sample was determined by compari ng each titration to that of a blank. Any base consumed by the sample results from the neutralization of acidic functional groups.
42 Results Review of Reactivation Protocols The rmal reactivations focused on the variation of the dissolved oxygen content in th e water that was used to make steam. The dissolved oxygen concentration levels examined were approximately <4.5 mg/L, 6.0 8.0 mg/L, and 12 .0 14.0 mg/L. These levels were attained, as previously stated, by using nitrogen, air or oxygen to pressurize the stainless steel vessel containing nanopure water. Nitrogen gas was used to reach the low concentration level ( < 4. 5 mg/L) of dissolved oxygen, air was used to attain the middle value range ( 6.0 8.0 mg/L), and oxygen gas was used when attaining the higher end of the values ( > 12 mg/L) used in this study. Those three distinct DO concentrations were used for three reactivation protocols ( steam curing steam curing plus ramping and steam pyrolysis). Steam (oxidation) and nitrogen (pyrolysis) gas flow rates were the same for all three reactivation protocols. The first protocol was to 1.0 mL/min. The second protocol was the same as the first, with an additional step. In this added step, the temperatur flow of nitrogen gas, while the steam was turned off. This is the protocol Mazyck and Cannon  developed in response to calcium catalysis called the steam curing plus ramping protocol. The third protocol performed was the protocol designed by McKenzie  attempt to design a protocol t hat is similar to the conventional methods of industry, with respect to reactivation time and temperature, while trying to attain similar characteristics as the one designed by Mazyck. The protocol called for a reactivation temperature of en minutes of steam followed by fifteen minutes under nitrogen gas at the
43 same temperature. This protocol was named the st eam pyrolysis reactivation The three reactivation proto cols are summarized in Table 3 2 Table 3 2 Reactivation Protocols and n otation Reactivation Protocol Description Notation (relative to DO) 1. Steam C uring Oxidant: Steam (variable DO) Pyrolysis: none 1N= low DO 1A= medium DO 1OX= high DO 2. Steam Curing Plus Ramping min. steam oxidant (variable DO) 2. Immediately ramped under pyrolytic conditions (nitrogen gas) 2N= low DO 2A= medium DO 2OX= high DO 3. S team P yrolysis min. under steam oxidant (variable DO) min. under pyrolytic conditions (nitrog en gas) 3N= low DO 3A= medium DO 3OX= high DO The aim of producing these nine reactivated carbons was to discern if a difference, with respect to surface chemistry and pore size distribution, in the resulting carbons, within each protocol existed. Be cause only the dissolved oxygen was changed within each protocol, this could present the dissolved oxygen content as being an important factor to address when considering reactivation. Dissolved Oxygen Content Dissolved oxygen was measured from the wate r in the pressure vessel. It was this water that was converted to steam and used to oxidize the spent GAC in the fluidized bed. The temperature of the water being used monitored and was constant in a small range (21 24 ). The measurement was taken as described in the methods section. Dissolved oxygen content is li sted for each carbon in Table 3 3
44 Table 3 3 The dissolved oxygen content is shown for each reactivation Carbon Dissolved Oxygen Temperature mg/L 1N 4.6 21.5 1A 7.7 24.1 1OX 13 22.4 2N 4.2 21.4 2A 7.5 21.8 2OX 13.1 23.8 3N 4.9 21.3 3A 7.4 21.5 3OX 12.2 23.6 Mass Loss, Volume Loss, and Density Mass loss is a characteristic of the process and nature of the spent carbon. Mass loss is neces sary but large mass losses should be avoided to minimize the loss of activated carbon. Mass loss for each carbon was measured and presented below. BET Surface Area, M ass and Volume Loss, and Apparent D ensity Of most importance to the thermal reactivation industry are the mass loss, volume loss, apparent density (AD), and iodine number associated with the thermal reactivation of a spent GAC. Herein, BET surface area was used in lieu of iodine number. The objective is to minimize mass and volume loss, retur n the spent apparent density back to the virgin AD, and create a BET surface area similar to the virgin carbon; although typically the surface area of the reactivated carbon is less than the virgin carbon because of calcium catalysis [15,67] When comparin g the three reactivation protocols (Table 3 4 ), it is observed that the stea m pyrolysis reactivation had the highest mass losses while the steam curing protocol had the lowest mass losses. Since the steam curing protocol did not include exposure of the sp ent GAC to temperatures greater than 375 it was expected to
45 exhibit the lowest mass loss, when compared to the other two reactivation protocols. The mass loss caused by the steam pyrolysis and steam curing plus ramping protocols were similar within st andard variance between experiment s. Generally as the DO concen tration of the water increased for each reactivation protocol, mass loss increased. The volume loss for the thermal reactivation protocols essentially mimicked the mass loss data. With respect to the AD of the reactivated carbons, the steam pyrolysis protocol was the nearest to returning the spent GAC back to the virgin value, while the steam curing plus ramping protocol either had an identical AD or slightly over reactivated the spent GAC, dep ending on the DO concentration, where over reactivation is partially defined as a reactivated AD lower than the virgin AD. If the reactivated GAC has an AD similar to the virgin AD, it can be assumed that the mass loss is related to desorption of natural o rganic matter. The steam curing protocol had the highest AD of the three reactivation protoc ols (AD 0.52 g/mL for 1A and 1OX ) indicating that this protocol did not fully reactivate the spent GAC. The final physical characterization was BET surface area. Th ere was no consistent trend between either the different thermal reactivation protocols or the different DO concentrations used in the reactivation protocols and overall surface area. In some cases, the BET surface area of the reactivated carbons was equal to or greater than the virgin carbon. The occurrence of BET surface area greater than virgin F400 may be explained by calcium catalysis or a similar catalysis mechanism. The increase in reactivation due to calcium catalysis could yield BET surface area of a reactivated carbon greater than its virgin counterp art. The BET surface area of 3OX is above the
46 typical range of F400 BET surface area (900 1000m 2 /g), which is encouraging since it is common practice in the thermal reactivation industry to blend virgin carbon, which is more expensive, with reactivated carbon to achieve a surface area desired by the customer ( e.g., 800m 2 /g). Calcium catalysis would potentially have a greater effect on the steam pyrolysis carbon, which is in line with what was seen regard ing BET surface area. As an aside, the BET surface area for the conventionally reactivated GAC (a process defined previously), which employed the same source of spent activated carbon, was 750 m 2 /g (MacKenzie  ). Table 3 4 Mass and volume loss and app arent density for reactivations Carbon Mass Loss Vol. Loss Apparent Density BET surface area (%) (%) (g/mL) (m 2 /g) Low 1N 6.10 1.5 0.49 641 Temp 1A 6.36 1.6 0.52 664 1OX 6.5 2.4 0.52 750 Steam 2N 13.5 3. 4 0.45 995 Curing 2A 20.8 3.6 0.46 884 2OX 28.7 4.0 0.48 949 Steam 3N 14.3 3.3 0.49 875 Pyrolysis 3A 21.5 3. 7 0.47 882 3OX 30.3 4.1 0.49 1187 F400 NA NA 0.48 900 Spent GAC NA NA 0.56 600 Comparing the three protocols, it is observed that the increase dissolve d oxygen content in the steam correlated to the increase in mass loss within the three protocols Volume loss for all carbons was low. It is expected that the majority of the volume loss occurred internally because this is where the majority of the con taminants will adsorb. The virgin apparent density of F400 is 0.48 g/mL. The spent GAC had an apparent
47 density of 0.56 g/mL. Typically, the goal of reactivation is to return the spent carbon to the apparent density of its virgin status. As Table 3 4 po ints out, the steam curing reactivation did not achieve this result. The two other protocols, Steam curing plus ramping and steam pyrolysis did return the carbon back to or near its virgin apparent density value. pH pzc (Point of Zero Charge) As previou sly stated, pH pzc can indicate qualitatively the amount of oxygen pzc (basic carbon), then it will have a low acidic surface oxygen content. Conversel y, if a carbon is a cidic (low pH), then it will have a high surface oxygen content. The results of pH pzc experiments are illustrated on the chart below. Recall, from the Experimental section, the experiment was run in duplicate. Figure 3 5 Point of zero charge for all re activated carbons
48 The spent carbon which was denoted as Raw in Figure 3 5, was included as a reference point This graph shows the carbons reactivated via the steam curing protocol have pH pzc values about 6. 5 Both the steam curing plus ramping and the steam pyrolysis carbons have similar pH pzc values in comparison with one another, with the steam pyrolysis carbons having a slightly larger value. Only the carbon reactivated at low temperature had a slightly acidic to neutral character, while the others approached a strong basic character. The similarity in pH can be attributed to similar surface oxygen content. Nonetheless, it is unclear from this analysis what type of functional groups ( e.g., carboxylic, lactonic,) may exist on the GAC surface. To d ate, there exists no clear methods with which to adequately characterize the specific surface functional groups that may be present in the reactivated carbon examined. RSSCT D ata Taste and odors in the raw water supply can be an annoyance to water treatm ent plant customers This is due to their low odor threshold concentration (OTC), which is in the nanogram per liter range [ 71, 72, 73] The low concentration allows these compounds to evade capture by conventional water treatment  The odor thresh old concentration is defined as the lowest concentration that a compound, in this case MIB, is found to be perceptible to humans. The OTC for MIB has been found to be in the range of 6 10 ng/L  The focus of this work was not on creating or tailoring a carbon for the best removal of a contaminant; however, it was another test that could show order to test the set of carbons on their performance capabilities. Figures 3 6 through 3 8 in the following sections, will show the MIB effluent concentration on the y axis and the x axis will be the bed volumes processed. The ordinate will be in terms of ng/L and
49 the abscissa in bed volumes. Experiments proceeded un til sometime after the OTC was reached. For each DO concentration the RSSCT was done in triplicate and are presented in the graphs that follow. Steam C uring Reactivated Carbons The motivation for this protocol was, as discussed earlier, that most oxygen f unctional groups remain on the carbon surface at low temperatures. The 1N produced nearly 6000 bed volumes when it finally broke through the OTC. This amount, when compared to the 1A (3600 bed volumes) and 1OX (Figure 3 6 ) (1440 bed volumes), is substant ially more than the 1A processed and more than three times as many as the 1OX. The carbon produced with the lowest DO steam performed the best, followed by the carbon produced with the middle range DO concentration, and lastly the carbon produced from the highest DO steam. This trend in MIB removal can be related to the amount of oxygen functional groups on the carbons surface as supported by Pendeleton [ 50]. Although the pHpzc values for these reactivated carbons are not very different, this macroscopic measurement is likely not taking into account the functional group differences that exist on these carbons. This result may suggest that surface chemistry plays an important role in MIB removal, a hypothesis supported by the work of Pend leton et al.  They also, along with C onsid in e , showed that a reactivated carbon with a limited amount of oxygen (acidic) functional groups was desired for enhanced MIB re moval. It could be hypothesized that as the DO concentration of the water decreased, the resultant reactivated carbon would have the least amount of oxygen functional groups, a hypothesis supported later in this work
50 Figure 3 6 Average RSSCT data for carbons produced via steam curing reactivation protocol Steam Curing Plus Ramping Protocol Thi s is the protocol Mazyck and Cannon  devised to overcome calcium cataly sis. The carbons, subsequently produced by Mazyck, proved to have a superior removal of MIB when compared with conventionally reactivated carbon. This reactivation protocol was the longest in duration, lasting just over two hours. The temporal length of the protocol is dependent on the furnace being used. The time needed to ramp to the desired 850 C is a function of furnace heating power; hence the time taken to do this re activation will vary depending upo n the furnace used. Mazyck reported the ramping procedure took just 20 minutes  T he RSSCT column data for the steam curing plus ramping reactivated carbons ( Fig 3 7 ) sho wed that the 2OX carbon processed the least amount of BVs (6500 BV) before re aching the OTC, compared to 2N and 2A The 2A carbon processed approximat ely 7400 BV followed by
51 the 2N carbon, which reached the OTC at 8200 BV. The trends for this reactivated carbon were the same as those for the steam curing reactivation whereby the carbon reactivated with the least amount of DO performed the best followed by the medium DO concentration and finally the highest DO concentration. Furthermore, the BET surface are a data could not explain the trends, which was the sa me conclusion of previous work [68 ] It can be noted that the steam curing plus ramping protocol showed more similar performance across the DO range than did the steam curing protocol of Fig. 3 6 One co uld deduce that the SCR protocol lessened the effect of surface chemistry. While some of the oxygen present in the reactivation furnace participates in gasification of the adsorbed organic compounds, it should be expected, based on the work of P uri [76 ] t hat some of the oxygen also chemisorbs to the carbon. However, Figure 3 7 RSSCT data for carbons produced via Steam Curing plus Ramping reactivation protocol
52 during t he ramped temperature stage of the protocol it is likely that some of the oxygen fun ctional groups that were added during the steam curing step decomposed, thus minimizing the difference in surface chemistry Indeed, the pHpzc for this protocol exhibited basic nature (greater than pH 9.5). Steam Pyrolysis Protocol The steam pyrolysis re activation procedure was chosen because MacKenzie  found it to perform well for MIB r emoval. From this protocol, the 3N performed the best followed by 3A and lastly 3OX (Figure 3 8 ) Initial breakthrough with 3N occurred after 4000 BV and the OTC was not surpassed until approximately 9200 BV. When comparing the performance of this carbon to the other thermal reactivation protocols, this protocol achieved the best results with respect to MIB removal. Again, it is likely the importance of surface chemistry since the 3OX reactivated carbon had the highest BET surface area of any carbon (1187 m 2 /g) and yet was out performed by 3N (875m 2 /g). Figure 3 8 RSSCT data for c arbons produced via Steam Pyrolysis reactivation protocol
53 The RSSCT data for all three protocols produced similar trends. The overall trend found was that the greater the conce ntration of DO in the water used to produce steam for the reactivation process, the fewer bed volumes the corresponding carbon would process, with respect to MIB. Comparison of Reactivated Carbons Using Varying Concentrations of DO to Conventional Reacti vation and V irgin GAC MacKenzie  also studied MIB removal using the clarified water and conditions discussed herein, including the same source of spent activated carbon, namely the Hopewell Water Treatment Plant. The time to reach breakthrough and the OTC for virgin GAC and conventionally reac ti vated GAC are shown in Table 3 5 For comparison, the corresponding BVs for the best reactivated carbon from the data shown in Figs. 3 6 3 8 are listed. The steam pyrolysis protocol with low DO was able to sign ificantly outperform a virgin coal based carbon with an approximately 50% increase Table 3 5 Performance of Virgin and Low DO Reactivated Carbons A data from M acK enzie  Protocol BV to initial Breakthrough BV to OTC Steam Curing 4200 6500 Steam Cu rin g plus ramped temperature 0 8050 Steam Pyrolysis 6400 9000 Conventional reactivation A 2300 4500 Virgin bituminous coal based carbon A 2700 6100 in BV to breakthrough and OTC. While the steam curing plus ramping protocol did not improve upon BV to breakthrough, the BV to OTC with this protocol and a low DO was
54 also an improvement over the virgin carbon. Both the steam pyrolysis and steam curing plus ramping protocols employ high temperatures (750 C/850 C), while the steam curing protocol employs the more modest temperature of 375 and shows comparable BV to breakthrough as the conventional reactivation and an improvement of approximately 45% on the BV to OTC. By employing the steam curing react ivation protocol with low DO, significant energy cost savings may result with similar if not improved performance compared to conventional reactivation. Furthermore, as virgin carbon is often blended with reactivated carbon to improve performance and compe nsate for mass lost during reactivation, a high performing reactivated carbon like 1N would reduce the amount of virgin carbon addition, resulting in decreased cost. Therefore, by employing the steam curing protocol with low DO the carbon user could have d ecreased cost of reactivation (lower reactivation temperature required) and operation (reduced virgin carbon addition). Interestingly, at medium and high DO levels, the steam curing protocol shows a performance below that of the conventional reactivated ca rbon, so that only by altering the DO level does the steam curing protocol become competitive. Role of S urface C hemistry Based on the work of Pendleton  and Considine  water adsorption onto activated carbon is deleterious to MIB adsorption and this phenomenon is more important than the physical attributes of the carbon, though the physical attribute s cannot be neglected. Tennant [79 ] argued that the physical attributes of the carbon (i.e. the pore size distribution) dominated over any impacts surfa ce chemistry, as measured by total surface acidity, had on MIB removal for powdered activated carbon MacKenzie et al. supported this finding in that MIB adsorption was shown to increase and correlate
55 extremely well with an increase in BET surface area for three commercially available virgin activated carbons. However, the current study showed no apparent trend when MIB uptake was compared to the BET surface areas for the reactivated carbons; which does not concur with MacKenzie  findings for reactivate d carbons. There are various methods to quantify the surface chemistry of an activated carbon and one such process is the Bhm titration method. When the total acidity of the reactivated carbons crea ted herein was measured (Fig. 3 .6 ), the carbons that were exposed to high temperatures (e.g. greater than or equal to 750 C) had less acidic surfaces while the steam curing reactivated carbon had the highest concentration of acidic functional groups. There are several possible conclusions that could be made fr om this data. First, the carbons that had higher surface acidity would attract more water molecules than those with surfaces that are less acidic (Kipling [80 ]) With this said, one would predict that steam curing plus ramping and steam curing would have s imilar performance. Indeed, steam curing plus ramping and steam pyrolysis did yield better performance than that demonstrated by steam curing ; however, differences in BET surface area and pore size distribution could also be responsible for this trend. Of particular interest, aside from the difference in surface acidity based on the reactivation protocol, is the change in surface acidity manifested as a result of changes in the DO concentration used within the reactivation protocols. In all three reactivati on protocols, the samples created using the highest DO concentration also had the highest total surface acidity and the samples created with the lowest DO concentration had the lowest total surface acidity. An increase in total surface acidity correspondi ng to a decrease in MIB adsorption found in this work
56 supports the conclusions of the works of Pendleton  Figure 3 6. Comparison of total acidity for the reactivated carbons. and Considine  namely the importance of surface chemistry on the adso rption of MIB. It should be noted that, although surface chemistry is likely a key component in the adsorption of MIB, there is likely to be some dependency on the pore size distribution of the carbon Conclusion The carbons produced using the steam curing reactivation possessed slightly acidic pH pzc values (~6.6), while the steam curing and Steam pyrolysis had basic values (~10). Also, the apparent density of the steam curing reactivated carbons did not attain the value of the virgin carbon. The RSSCT dat a showed that the three steam curing carbons processed the fewest number of bed volumes. This is believed to be attributed to the pH pzc and apparent density values. The higher density values show that possibly there was less unrestricted pore volume avai lable to adsorb MIB. This would agree with the RSSCT data. Also, the fact that pH pzc value was slightly acidic could suggest the
57 presence of oxygen functional groups, which has been shown to be detrimental to MIB adsorption [50, 55, 75] The steam curing protocol and steam pyrolysis protocols had similar pH pzc values (~10). The apparent density of the steam curing carbons was a bit lower th an the virg in value, whereas steam pyrolysis carbons had apparent densities very close to the virgin apparent density. The RSSCT data showed a difference between the two protocols in performance. The 3N processed approximately 1100 more bed volumes than the 2N. Th e 3A and 3OX processed approximately 700 bed volumes more than did the 2A and 2OX. In agreement with MacKenzie  the steam pyrolysis protocol outperformed the steam curing carbons. The app arent density of the steam curing plus ramping to the increas ed time of reactivation (~2hr versus 0.5 hr ). Thus, it remains unclear what role the dissolved oxygen concentration has on the carbon surface and pore structure. The goal of this study was to determine if changes in the DO concentration of the water use d to create the steam in reactivation protocols could be used to tailor the reactivation process. The data shown here confirms that changes in the DO concentration of the water used to create steam manifest differences in physical properties (e.g. mass lo ss) as well as performance. The performance of the carbon, with respect to MIB removal, generally was found to be inversely proportional to the DO concentration of the water used to create the steam for reactivation; the lower the DO, typically the greater the MIB removal. All three reactivation protocols investigated performed better than conventional reactivation when the DO concentration of the water used to make the steam was low.
58 The use of the steam curing plus ramping process with low DO showed compa rable performance to a virgin carbon and improved performance compared to conventional reactivation, which can yield cost savings in two ways: (1) the steam curing process utilizes a lower temperature tha n conventional reactivation, steam curing plus rampi ng and steam pyrolysis, thus lowering the cost of energy; (2) addition of virgin carbon to reactivated carbon to improve performance would be unnecessary or significantly reduced. An analysis of the total acidity of each carbon showed an increase in total surface acidity with an increase in the DO concentration used in the water used to create steam within each reactivation protocol. Therefore, the decrease in performance for the removal of MIB can be related to an increase in water adsorption, and a resul ting decrease in MIB adsorption, relating to an increase in the total surface acidity of the carbon.
59 CHAPTER 4 A LIFE CYCLE ASSESSMENT AND COMPARISON OF GAC REACTIVATION TO VIRGIN GAC REPLACEMENT Literature Review Life cycle assessment (LCA) methods for identifying, evaluating, and choosing from alternative opportunities which will improve energy, resource, and environmental release profiles of a product or process are still evolving. LCA is a tool that allows one to assess the economic and environmental costs associated with a product or a process in LCA have been reported [81 83 ] An LCA theoretically chronicles the life of a product considered and to focus on the main issues of concern. The three components o f an LCA are the inventory analysis, impact analysis, and the improvement analysis. The inventory analysis component is comprised of two subsets: scoping and data collection. Impact analysis is a method used to evaluate environmental issues that may ari se from the processes under consideration (e.g., global warming, acid rain, etc.). Suggestions for the improvement of the process, in regards to the environmental impacts found in the impact analysis step, are detailed in the improvement analysis. Some r esearchers have offered an economic evaluation on the use of GAC in water treatment facilities [8 1] Adams found that with an increa se in size of the water treatment plant (gallons treated), the cost of GAC gradually decreased. Moore et al. reported the changes in GAC pore structure in a study that compared virgin and thermally reactivated GAC for use in water treatment. He found tha t a carbon that had experienced six cycles of water treatment and thermal reactivation outperformed a virgin (F 400) GAC for the removal of total organic carbon (TOC) in the first half year of service. It was shown that the pore
60 size distribution of the c arbons was vastly different  Thus, they showed the effectiveness of thermal reactivation of GAC. A life cycle assessment of the activation or reactivation of GAC has not appeared in the literature. For this study, an LCA i s used to understand the benefits and/or drawbacks to reactivation of granular activated carbon GAC, as well as compare three reactivation protocols for spent activated carbons used in water treatment plants. GACs may be produced from any carbon based ma terial, but bituminous coal is most commonly used. The activation of bituminous coal to produce GAC is most typically accomplished through thermal activation. The precursor carbon is first pulverized to promote macroscopic homogeneity and then mixed with a binder (i.e., agglomerated activation versus direct activation where a binder is not used). The carbon is then compressed to form hard briquettes. The next step is carbonization/pyrolysis whereby the carbon is heated in the absence of air to free the carbon from its associated atoms. Mild oxidative gases, such as steam, are then passed through the carbon at temperatures between 315 C and 925 C. This step approxi m ately 0.85 to 0.95 mL/g  The use of GAC for the removal of priority organic pollutants is very popular. GAC is effective at removing both soluble or ganic and inorganic compounds  The largest market for activated carbon is currently the municipal w ater purification industry, where activated carbon beds are used for the dual purpose of ph ysical filtration and sorption [ 5 ] Water treatment facilities typically employ GAC for the removal of organic compounds, namely synthetic organic compounds (SOCs) and taste and odor causing
61 compounds. Once the capacity of the GAC has been exhausted through service, spent carbons can be landfilled, incinerated, or reactivated for reuse. R eactivation has been widely employed due to its cost effectiveness [90,91] T he four common thermal reactivation systems are rotary kilns, multi hearth furnaces, fluidized bed furnaces, and infrared furnaces. Thermal reactivation consists of three stages: drying, pyrolysis, and gasification  The pyrolysis step promotes the ev olution of volatile organic compounds (VOCs) as well as water, and the thermal decomposition of less volatile organic molecules. Thermal decomposition is responsible generall y employing steam at a similar temperature scheme. This creates an environment that allows gasification to occur, thus resulting in the removal of the char and re exposing the pores. This study is based largely upon data from Cincinnati Water Works (CWW ), a drinking water treatment facility with on site GAC reactivation. CWW processes approximately 107 to 136 MGD, requiring 12 filter beds run in parallel/series, each containing 60,000 lbs of GAC, for a total of 7.2 million lbs of GAC. The facility is re cognized as one of the largest municipal potable water trea tment systems employing GAC  approximately 40,000 lbs of spent GAC per day. The total organic carbon (TOC) entering the GA C beds averages 1.5 mg/L. The contactors at CWW are deemed spent with water to create a slurry and pumped into a holding tank, where it remains until
62 being reactivated. The reactivation process results in carbon losses on the order of 10 15%, which are replaced with virgin carbon  The LCA presented herein is comprised of two parts. F irst three reactivation protocols were compared : conventional reactivation, steam c uring plus ramping temperature, and steam pyrolysis reactivation. Using the most desirable reactivation protocol determined from part 1, a comparison was then made between a facility that uses virgin carbon exclusively (scenario 1) and one with an on site reactivation plant (scenario 2). Together, these two parts serve to elucidate the environmental and economic impacts associated with each reactivation protocol, as well as GAC usage with or without reactivation. Experimental LCA For this study, an L CA was used to evaluate three reactivation protocols. Then the best determined reactivation protocol was compared, over a twenty year period, to a system where reactivation was not employed. The three reactivation protocols were: Steam Pyrolysis Reacti vation (SPR), Steam curing reactivation, and conventional reactivation. The SPR and Steam curing reactivation methods are described above. The conventional reactivation process is when pyrolysis is done first followed by the oxidation step the most comm on reactivation process currently used The pyrolysis step was performed under a nitrogen flow and a controlled temperature ( 750 ) and time (15 min) This step was foll owed by an oxidation step where steam was allowed to flow through the carbon at a constant flow rate ( 1 mL/min ) a nd temperature ( ) for fifteen minute s The best protocol was then determined by employment of a functional unit. Where the functional unit was defined by,
63 ( 4 1 ) Surface area was chosen as it can be a surrogate for performance. Surface area is a parameter used to determine the effectiveness of the reactivation. Mass loss is used as a surrogate for the air pollution produced by reactivation. Energy input i s used to get a feeling for the magnitud e of energy consumption of a reactivation process Therefore, the functional unit suggests that the high er the value of the functional unit, the better the reactivation protocol, in terms of environmental impacts identified, its impacts were computed over a twenty year period. This was then compared to those of a process using GAC without reactivation. Scope of Study The first step in the LCA process is to define the scope of the project by creating a material and energy flo w diagram  These diagrams begin from a macroscopic view of the system (level 1), which places boundaries on the project scope. Level 1 diagrams for scenario 1 and 2 are depicted in Figures 4 1a and 4 1b, respectively.
64 Figure 4 1 Level 1 diagra ms for scenarios 1 and 2. In order to further break the process down, level 2 diagrams are constructed. A generic level 2 diagram ( Figure 4 2 a ) begins with the extraction of raw materials, includes their conversion to activated carbons, proceeds with thei r use in a treatment plant, and ends with their disposal. The extraction of raw materials was excluded from this study. For scenario 1, the scope begins with GAC activation and ends with landfill disposal of spent GAC after a 2 year period ( Figure 4 2 b ). The project scope for scenario 2 begins with GAC activation, includes a yearly reactivation process, and ends with the combustion of GAC to CO 2 ( Figure 4 2 c) Comparison of the three reactivation protocols requires a separate level diagram ( Figure 4 3 ) b ecause it looks exclusively at the process of reactivation. Boundaries are placed on the process that begins with the introduction of spent GAC and ends with the reactivated product.
65 Study Assumptions The assumptions for the reactivation protocols we re as follows: each carbon has the same water treatment performance at the full scale; all spent GAC was reactivated every year; make up carbon purchases included cost and impacts; and no environmental impacts associated with the construction of the reacti vation plant were considered. The following three assumptions were made for GAC activation: the carbon has equal water treatment ability; the virgin GAC has a two year life span; all spent carbon is landfilled; and there is no change in the price of carbo n over a twenty year period. Additionally, all emissions from the generation of electricity were excluded based on data limitations.
66 Figure 4 2b (Scenario 1) Mixed Solid Waste Transportatio n of spent GAC to landfill Bituminous coal GAC Activation 2.1 Emissions Transportation of GAC to treatment plant Drinking Water Treatment Plant 2.2 Energy Landfill 2.3 Mixed Solid Waste Reactivated GAC returned to carbon beds Emissions Excess heat Energy GAC Activation 2.1 Drinking Water Treatment Plant 2.2 Bituminous coal Emissions Transport of GAC to treatment plant Onsite Reactivation 2.3 Energy Spent GAC Steam Figure 2c (Scenario 2) Raw material extraction 2.1 Activation of GAC 2.2 Drinking water treatment plant 2.3 Disposal 2.4 GAC reactivation 2.5 Not included in LCA Figure 4 2a (Generic) Figure 4 2. Level 2 diagrams Figure 4 2c. (Scenario 2 )
67 Figure 4 3. Level 2 diagram for reactivation protocols Per the experimental section, comparison between reactivat ion schemes was made by us e of a functional unit, defined as the surface area divided by the product of energy required and percent mass loss. The functional unit was chosen for the following reasons: surface area upon reactivation is a typical measure use d to describe the quality of the reactivation and provides an estimate as to how the reactivated carbon will perform. Energy consumption for each reactivation protocol was used to calculate the relative amounts of energy that would be required. Mass loss was included in the functional unit, because it is a relative measure of the emissions that evolve from the process. The functional unit indirectly accounts for performance (surface area) and environmental costs (emissions and energy). Reactivation Proto cols The conventional reactivation protocol utilizes a pyrolysis (heating in an inert or oxygen devoid environment) step followed by an oxidative step using steam. The time (15 minutes) and temperature (750 C) regimes used are the same for both steps. T wo other thermal reactivation protocols; steam curing plus ramping temperature (SCR) and Emissions Heat Reactivation Furnace 2.2 Steam generator 2.1 Stea m Energy Water Emissions After Burn er 2.3 Spent GAC Energy Reactivated GAC Venturi & Tray Scrubbers 2.4 Off gas Energy Water Energy Off gas Water
68 steam pyrolysis reactivation (SPR), were also investigated. The SCR protocol was dev eloped by Mazyck and Cannon  as a response to the problems posed by calcium ca t alysis during reactivation  The protocol consists of oxidation for 1 hour at 375 C in steam followed by pyrolysis in N 2 During the pyrolysis step, which lasts about one hour, the temperature is ramped from 375 to 850 C. MacKenzie et al.  d eveloped the SPR to mimic the results obtained by the SCR protocol under conditions more appropriate for industry scale. The protocol consists of gasification for 10 to 30 minutes at temperatures in the range of 650 to 850 C in steam followed by pyrolysi s at similar times and temperatures in N 2 MacKenzie et al. found the optimum SPR time to be 5 minutes for gasification followed by 15 minutes for pyrolysis  In this study, the optimum conditions found by MacKenzie et al. were employed using both ox idation and pyrolysis at 750 C for 15 minutes each. The reactivations were performed in a fluidized bed using 10g of spent F 400 carbon from the Hopewell, Virginia water treatment facility. Surface Area Analysis A gas sorption analyzer (NOVA 1200, Qua ntachrome) was employed to determine the surface area of the GAC samples. Samples were outgassed at 110 C for 24 hours followed by N 2 adsorption at 77 K. The Brunauer Emmett Teller (BET) model was applied to the resulting isotherm to determine the surfa c e area of each carbon  Samples were run in duplicate, with error being negligible. Reactivation Cost Analysis Total cost of a reactivation protocol over a 20 year period is based upon capital costs for building construction, operation and maintenan ce (O&M) costs and the cost of the initial virgin carbon. Construction cost data are based on 1979 and 1980 studi es of
69 the CWW plant  This data were converted to current day values using the Bureau of Labor Sta tistics Inflation Calculator  A 6 % interest rate was applied to the total capital costs of the GAC reactivation facility, and the value was amor tized over a 20 year period  The financial data do not include labor costs. Operation and maintenance costs were obtained directly from t he CWW facility  Laboratory measurements were used to adjust make up carbon values for the three protocols. A weighting factor was developed based on the ratio of carbon mass loss (percent) from CWW to carbon mass loss (percent) from the lab scale p rotocols. These weighting factors were then applied to the make up carbon cost provided by CWW to calculate the cost of make up carbon for the different protocols. Maintenance costs were assumed to be the same for all protocols. Electricity and natural gas costs for the three protocols are based on the energy required for reactivation. Conventional and steam pyrolysis processes use the same amount of energy, but the SCR protocol requires additional energy input because of the additional time involved. Bec ause the conventional process is similar to the procedure in use at CWW, electricity and natural gas usage can be assumed the same. Normalization of the energy usage for the lab scale processes provided a scaling factor for the energy used by the SCR prot ocol. This factor was applied to natural gas and electricity costs provided by CWW to calculate costs for the SCR protocol. Costs for the operation of the afterburner were based on energ y input per pound of carbon  the current cost of natural gas [1 00] and the number o f pounds of carbon in a bed  Activation Cost Analysis The following costs were considered when determining the scenario 1 total costs: 1.) purchase of virgin F 400; 2.) transportation of virgin GAC to the water treatment
70 fac ility; 3.) transportation of spent GAC to a local landfill; and 4.) landfill tipping fee for spent GAC disposal. The cost for F 400 GAC was assumed to be the same cost for reactivation. The transportation costs are b ased upon literature values  and ad justed for mileage. Landfill tipping fees were obtained directly from the Rumpke Sanitary L andfill in Cincinnati, Ohio  Solid waste volumes from the activation process were estimated using a waste repo rt from the Pennsylvania DEP  and density value s for solid waste components ). Solid waste from activation is primarily composed of emission control sludge, mixed paper and cardboar d, and other generic wastes  Volume of spent GAC was calculated from the bulk density of F 400 carbon. Rea ctivation solid wastes are generated solely from the activation of makeup carbon. Transportation Emissions Diesel open top dump trucks with a 12,000 ft 3 (340m 3 ) capacity were assumed to be the pr imary mode of GAC transport  Transportation emissions from diesel trucks were determined using three sources. Carbon dioxide and PM 10 emissions were determined assuming the model year of the truc k was between 1991 and 1994  42 Emission Factor Index was used to determine CO, NOx, and CH 4 emis sions assuming a 1995 model year with emissions recorded in 2000  Other assumptions embedded in the AP 42 emission factors were that vehicle operation was at low altitude and 72F, with an average speed of 19.6 MPH. This average speed may be too lo w for the transport of virgin carbon unless it accounts for idling; however, it is a realistic assumption for disposal of spent carbon to the landfill which is approximately 15 miles from the treatment fa cility. GREET 1.6 software  was used to determ ine SO 2 VOC, and N 2 O emissions. These emissions are likely to be underestimated
71 because they are based upon a light duty truck, which was undefined in the model. Nonetheless, they still provide relative numbers for comparison between the activation/disp osal and reactivation schemes. The three sources for emission factors each reported the results as emissions per mile which were converted to total emissions based upon the distances to the treatment facility and the landfill. Reactivation Emissions In order to quantify the impacts of the three reactivation protocols, the emissions per mass of GAC were calculated. Current air control devices at CWW include an afterburner followed by a venturi s crubber and a tray scrubber  Because the reactivation process consists of gasification of GAC and adsorbed pollutants, the afterburner sufficiently converts all air emissions to CO 2 [94,99] Additional CO 2 emissions are produced by natural gas combustion for heating and afterburner operation. Carbon dioxid e emissions from GAC mass loss during reactivation were determined from the mass loss experiments performed for each protocol. Calculations were based upon the assumptions that 100% of the mass loss is converted to CO 2 and that the carbon composition of s pent GAC is 80%, equal to that of bituminous coal  The CO 2 emissions from the combustion of natural gas were calculated based upon the amount of natural gas needed to satisfy the energy requirements of each reactivation protocol and for the operatio n of the afterburner. Knowing the volume of natural gas required and the products of combustion per unit volume for natural gas  the total mass of CO 2 was determined. Operation of the venturi scrubber was assumed to be negligible from both an econo mic and environmental viewpoint. Water is the main input and output from the scrubber. Owing to its placement at a water
72 treatment facility, a water supply is readily available and the wastewater is simply recycled to the headworks of the water treatment plant  Activation Emissions. Pollutant data  total annual company sales, and the price of Calgon F 400  f rom Calgon Carbon Company in Neville Island, Pennsylvania were used to calculate emissions fro m activation of Calgon F 400  Two assumptions were required to generate emissions using this information: 1.) All Calgon Carbon products result in equal emissions and 2.) All Calgon Carbon products have the same price per mass. These assumptions allow for the determination of pollu tant emissions on a per mass basis of F 400. The pollutant data did not include CO 2 emissions; therefore, the following assumptions were made: bituminous coa l is composed of 80% carbon  ; thirty percent of the bituminous coal is converted to GAC produ ct  ; ninety nine percent of mass loss during activation is converted to CO 2 ; and natural gas is the only fuel used in the manufacturing process. The total CO 2 production is the sum of the CO 2 emissions during activation and the combustion of natural gas. The emissions from natural gas were determined using a similar approach as that for reactivation. Total Emissions. The total amount of pollutant emission s from GAC usage without reactivation rio 2) were evaluated over a 20 year period. Total emissions from scenario 1 are comprised of those from transport (to the treatment facility and the landfill) and from the activation process. The t otal emissions from scenario 2 are produced from the act ivation and transport of initial and makeup carbon, as well as emissions from the reactivation process. Although emissions from electricity generation could be quantified for scenario 2, they were not included in
73 this comparison because electricity use for scenario 1 could not be calculated from the data available. Only including electricity usage for one scenario would not provide a valid comparison. Results The SCR protocol experienced a mass loss of 8.7%, the SPR protocol 8.0%, and the conventional pr otocol 9.1%. SCR produced a greater surface area, 693 m 2 /g followed by SPR at 663 m 2 /g and the conventional protocol at 644 m 2 /g. The energy requirements for the SCR reactivation, SPR, and conventional were 8.9, 2.73, and 2.73 KWh, respectively. These w ere calculated using data from lab scale fluidized bed experiments and the following equation: (4 1) The comparison of the three reactivation protocols was made using the functional unit equation: (4 2 ) The functional unit value for SPR was 30.4 m 2 /g KWh, followed by the conventional protocol at 25.9 m 2 /g KWh and finally by the SCR protocol at 9.0 m 2 /g KWh.
74 Table 4 1. Reactivation Experimental Data and Functional Unit Protocols Surface Area Mass Loss Apparent Density Energy Functio nal Unit ( m 2 /g ) (%) (g/mL) (k wh ) ( m 2 /g k wh ) SCR 693 8.7 0.500 8.9 9.0 SPR 663 8.0 0.500 2.73 30.4 Conventional 644 9.1 0.488 2.73 25.9 The operation and maintenance costs reported in Table 4 2 were based upon values provided from CWW and adju sted using laboratory measurements. Total costs for each reactivation protocol are summarized in Table 5 3. Purchase of the initial 7.2 million lbs of virgin carbon to fill the 12 filter beds costs $3.96 million and does not change for the different prot ocols. Total costs associated with the construction of the reactivation facility are approximately $ 1.76 million for a 6% interest rate over a 20 year period. Operation and maintenance costs present the biggest difference in cost between the three protoc ols. Energy use for the SCR protocol results in nearly twice as much cost over the 20 year period. Reactivation costs for the three protocols are as follows: conventional, 20Â¢/lb GAC reactivated; SCR, 37Â¢/lb GAC reactivated; and steam pyrolysis, 19Â¢/lb G AC reactivated. From a purely economic point of view, steam pyrolysis is the best protocol for reactivation. \
75 Table 4 2 Operation and Maintenance Costs for Each Protocol Item Conventional SCR SPR Make up carbon ($/bed) Maintenance ($/bed) Electr icity ($/bed) Natural gas ($/bed) Afterburner gas ($/bed) Total cost ($/bed) Total cost ($/yr) Total cost ($/20 yr) 30,021 6,200 a 6,800 a 39,000 a 13,635 95,656 1,147,872 22,957,440 28,701 6,200 a 22,168 127,140 13,635 197,844 2,374,128 47,482,560 26,392 6,2 00 a 6,800 a 39,000 a 13,635 92,027 1,104,324 22,086,480 a These values wer e obtained directly from CWW  Table 4 3 Total Costs for Each Reactivation Protocol Itemized Cost Conventional SCR SPR Capital costs ($/20 yr) 1,764,068 1,764,068 1,764,068 Virgin GAC ($) 3,960,000 3,960,000 3,960,000 O&M ($/20 yr) 22,957,440 47,482,560 22,086,480 Total Cost ($/20 yr) 28,681,508 53,206,628 27,810,548 Cost ($/lb GAC) 0.20 0.37 0.19 Table 4 4 displays the CO 2 emissions for e ach GAC reactivation protocol. Steam pyrolysis was found to emit the least amount of CO 2 because it results in the smallest amount of mass loss and has the lowest heating energy requirement among the three protocols. Steam pyrolysis exhibited the most de sirable performance from both an economic and environmental point of view, which led to its selection as the best
76 protocol. It was used as the reactivation protocol for the comparison of scenario 1 to scenario 2. Table 4 4 Experimental Emission Values for Each Reactivation Protocol Protocol Mass Loss Heating Energy Afterburner Total Emission lb CO 2 /lb F 400 lb CO 2 /lb F 400 lb CO 2 /lb F 400 lb CO 2 /lb F 400 SCR 0.255 3.23 0.35 3.83 SPR 0.235 0.99 0.35 1.57 Conventional 0.267 0.99 0.35 1.61 Tabl e 4 5 summarized the costs for scenario 1 over a 20 year period. The most salient conclusion from this data is that the cost per mass of GAC used for scenario 1 is approximately 94 Â¢. This price is considerably higher than that of reactivation (see Table 4 3), regardless of which reactivation protocol is employed. Table 4 5 Total Costs for GAC Use Without Reactivation (Scenario 1) Cost ($/20 yr) Purchase of virgin F 400 carbon Transport of virgin carbon to treatment facility 39,600,000 23,807,896 Tr ansport of spent carbon to landfill 2,506,094 Landfill disposal fee 1,881,235 Total 67,795,225 Cost ($/lb GAC) 0.94 Total emissions for the 20 year operation period are presented in Table 4 6. Scenario 2 produces lower amounts of each pollutant t han scenario 1 because the majority of the pollutants, except for CO 2 are assumed to be generated from activation and transportation. Given that reactivation requires less production of virgin carbon,
77 emissions from transport and activation are lowered. Pollutant emissions from scenario 2 excluding CO 2 were 23 to 26% of those from scenario 1, with CO2 emissions being 37% less as well. Considering this significant reduction in pollutant emissions, GAC treatment using reactivation provides a more environme ntally responsible option. The impacts associated with scenarios 1 and 2 are based upon three sources: 1.) transportation emissions, 2.) point source emissions from the activation and reactivation processes, and 3.) solid waste generation. Using the 20 y ear emission data presented in Table 5 6 global warming and acid rain potentials were calculated  Potentials for ozone depletion and smog formation were also considered, but the processes considered did not emit constituents responsible for these i mpacts. A comparison of the global warming potential (GWP) and acid rain potential (ARP) for scenario 1 and scenario 2 may be seen in Table 4 7. Higher potential values represent greater risk to the environment. The total solid waste generated from acti vation and reactivation over a twenty year period was calculated as 3.7 million yd 3 and 171,000 yd 3 respectively; therefore, scenario 1 requires nearly 22 times more landfill volume than scenario 2. The global warming, acid rain, and landfill usage impa cts for scenarios 1 and 2 are summarized in Figure 4 4 using an impact s tressor matrix from Graedel  Figure 4 4 illustrates that each of the impacts from activation are greater than those from reactivation. Acid rain impacts were assumed to be of m oderate reliability since 42 Index, CA Air Resources Board, and the GREET software) and from activation emissions, which are less reliable since they incorporate two main assum ptions, as
78 stated in the Methods section. Global warming data had a moderate reliability because many assumptions were made regarding the emissions from the activation process. Landfill exhaustion data for reactivation has the lowest reliability because of the uncertain solid waste composition. Although similar solid waste assumptions were made for activation, the majority of activation waste is due to spent GAC disposal. Therefore, data reliability for activation is considered moderate. Overall, is has been shown that GAC treatment with reactivation results in significant benefits from both environmental and economic standpoints when compared to GAC treatment using only virgin carbon.
79 Table 4 6. Total Emissions from Each Scenario Scenario 1: No reac tivation Scenario 2: Reactivation Pollutant Production Transport Total Production (make up) Transport (make up) Reactivation Total lb/20 years lb/20 years lb/20 years lb/20 years lb/20 years lb/20 years lb/20 years CO 2 6.10E+08 2.42E+06 6.13E+08 1.59E +08 5.69E+05 2.27E+08 3.86E+08 CO 2.78E+04 1.21E+04 2.46E+04 7.22E+03 2.85E+03 6.10E+03 NO 2 3.01E+04 9.10E+03 2.27E+04 7.82E+03 2.14E+03 5.67E+03 PM 10 2.83E+04 7.22E+02 1.35E+04 7.37E+03 1.70E+02 3.49E+03 SO 2 3.45E+04 6.59E+02 1.62E+04 8.96E+03 1.55 E+02 4.19E+03 VOC 5.31E+03 1.21E+04 1.45E+04 1.38E+03 2.85E+03 3.47E+03 CH 4 1.11E+02 1.11E+02 2.61E+01 2.61E+01 N 2 O 7.22E+02 7.22E+02 1.70E+02 1.70E+02 NH 3 3.88E+05 3. 88E+05 1.01E+05 3.86E+08 Cu compounds 1.80E+02 1.80E+02 4.67E+01 4.67E+01 Cr compounds 1.76E+02 1.76E+02 4.58E+01 4.58E+01 HCl 1.41E+02 1.41E+02 3.66E+01 3.66E+01 HF 1.41E+01 1.41E+01 3.66E+00 3.66E+00 Pb 1.41E+01 1.41E+01 3.66E+00 3.66E+00 Zn compounds 1.76E+02 1.76E+02 4.58E+01 4.58E+01
80 Table 4 7 Global Warming and Acid Rain Pollution Potentials GW P ARP Activation 6. 14 E+08 3.6 9 E+0 5 Reactivation 3.86 E+08 9. 5 7 E+0 4 Figure 4 4 Impact Stressor Matrix for Impact Valuation
81 LIST OF REFERENCES 1. Cheremisinoff, Paul N., Angelo C. Morresi, Carbon Adsorption Applicati ons. carbon, 1980. 2. Bansal, Roop C., Jean Baptiste Donnet, Fritz Stoeckli, Active Carbon 1988, New York, NY: Marcel Dekker, Inc. 3. Ostrejko, German patent number 136 792 1901. 4. McGuire M, Suffet I., Aqueous chlorine/ activated carbon interactions. J ournal of Environmental Engineering, 1984. 110 (3): p. 629 643. 5 Waer, M.A., Snoeyink, V.L., Mallon K.L., Carbon Regeneration: Dependence on time and temperature. JAWWA, 1992. 84 : p. 82. 6 Michael Gundry, P.Eng., Totten Sims Hubicki Associates, Activate d carbon use in Ontario drinking water treatment plants. Environmental Science & Engineering, 2000. 13 (1). 7. Boehm, H. P., Surface oxides on carbon and their analysis: a critical assessment. Carbon, 2002. 40 (2): p. 145 149. 8. de la Puente, G., J. J. Pis, J. A. Menendez, and P. Grange, Thermal stability of oxygenated functions in activated carbons. Journal of Analytical and Applied Pyrolysis, 1997. 43 (2): p. 125 138. 9. Salame, Issa I. and Teresa J. Bandosz, Surface Chemistry of Activated Carbons: Combinin g the Results of Temperature Programmed Desorption, Boehm, and Potentiometric Titrations. Journal of Colloid and Interface Science, 2001. 240 (1): p. 252 258. 10. Bandosz, T. J. and Q. Le, Evaluation of surface properties of exhausted carbons used as H2S ad sorbents in sewage treatment plants. Carbon, 1998. 36 (1 2): p. 39 44. 11. Bandosz, Teresa J., Effect of pore structure and surface chemistry of virgin activated carbons on removal of hydrogen sulfide. Carbon, 1999. 37 (3): p. 483 491. 12. Phillips, Jonathan Bo Xia, and J. Angel Menendez, Calorimetric study of oxygen adsorption on activated carbon. Thermochimica Acta, 1998. 312 (1 2): p. 87 93. 13. Bandosz, T. J., B. Buczek, T. Grzybek, and J. Jagiello, The determination of surface changes in active carbons b y potentiometric titration and water vapour adsorption. Fuel, 1997. 76 (14 15): p. 1409 1416. 14. Menendez, J. A., E. M. Menendez, M. J. Iglesias, A. Garcia, and J. J. Pis, Modification of the surface chemistry of active carbons by means of microwave induce d treatments. Carbon, 1999. 37 (7): p. 1115 1121.
82 15. Mazyck, David W. and Fred S. Cannon, Overcoming calcium catalysis during the thermal reactivation of granular activated carbon: Part I. Steam curing plus ramped temperature N2 treatment. Carbon, 2000. 3 8 (13): p. 1785 1799. 16. Mazyck, David W. and Fred S. Cannon, Overcoming calcium catalysis during the thermal reactivation of granular activated carbon: Part II. Variation of steam curing reactivation parameters. Carbon, 2002. 40 (3): p. 241 252. 17. Vigon B.W., Tolle, D.A., Cornaby, B.W., Latham, H.C., Harrison, C.L., Boguski, T.L., Hunt, R.G.,and Sellers, J.D., Life Cycle Assessment: Inventory Guidelines and Principles 1993, U.S. EPA: Cincinnati, OH. 18. Bishop, P.L., Pollution Prevention: Fundamental s and Practice 2000, Boston: McGraw Hill. 251 293. 19. Wigmans, T., Industrial aspects of production and use of activated carbon. Carbon, 1989. 27 (1): p. 13 22. 20. Lalezary S, Pirbazari M, McGuire M, Evaluating activated carbons for removing low concentr ations of taste and odor producing organics. JAWWA, 1986. 78 (11): p. 76 82. 21. Snoeyink, VL, Adsorption of Organic Compounds Water Qualtiy and Treatment, ed. P. F. 1990, New York: McGraw Hill. 781 867. 22. Moore, Brian C., Fred S. Cannon, Judy A. Westric k, Deborah H. Metz, Clifford A. Shrive, Jack DeMarco, and David J. Hartman, Changes in GAC pore structure during full scale water treatment at Cincinnati: a comparison between virgin and thermally reactivated GAC. Carbon, 2001. 39 (6): p. 789 807. 23. Mazyc k, David W., Steam curing plus ramped temperature treatment: A novel approach to thermal reactivation in Civil and Environmental Engineering 2000, Pennsylvania State University. p. 195. 24. Ergun S, Menster M., Reactivations of carbon with carbon dioxide and steam. Chemistry and Physics of Carbon, ed. W.J. P. Vol. 1. 1965, New York: Marcel Dekker. 204 261. 25. Walker, Phillip L. Jr., Shiro Matsumoto, Tamotsu Hanzawa, Takatoshi Muira, and Ismail M.K. Ismail, Catalysis of gasification of coal derived cokes and chars. Fuel, 1983. 62 : p. 140 149. 26. Cannon, F., V. Snoeyink, R. Lee, G. Dagois, and J. DeWolfe, Effect of calcium in field spent GACs on pore development during regeneration. JAWWA, 1993. 85 (3): p. 76 89. 27. Knappe, D., V. Snoeyink, G. Dagois, and J. DeWolfe, Effect of calcium on thermal regeneration of GAC. JAWWA, 1992. 84 (8): p. 73 80.
83 28. Mazyck, David W., Steam curing plus ramped temperature treatment: A novel approach to thermal reactivation in Civil and Environmental Engineering Ph.D., 2000, Pennsylvania State University. p. 195. 29. Tomkow, K., T. Siemieniewska, F. Czechowski, and A. Jankowska, Formation of porous structures in activated brown coals chars using O2, CO2, and H2O as activating agents. Fuel, 1977. 56 (2): p. 121 124. 30. Rodrig uez Reinoso, F., The use of steam and CO2 as activating agents in the preparation of activated carbons. Fuel and Energy Abstracts, 1995. 36 (6): p. 427. 31. Rodriguez Reinoso, F., M. Molina Sabio, and M. T. Gonzalez, The use of steam and CO2 as activating a gents in the preparation of activated carbons. Carbon, 1995. 33 (1): p. 15 23. 32. Rodriguez Valero, M. A., M. Martinez Escandell, M. Molina Sabio, and F. Rodriguez Reinoso, CO2 activation of olive stones carbonized under pressure. Carbon, 2001. 39 (2): p. 3 20 323. 33. Arriagada, R., R. Garcia, M. Molina Sabio, and F. Rodriguez Reinoso, Effect of steam activation on the porosity and chemical nature of activated carbons from Eucalyptus globulus and peach stones. Microporous Materials, 1997. 8 (3 4): p. 123 130. 34. Molina Sabio, M., M. T. Gonzalez, F. Rodriguez Reinoso, and A. Sepulveda Escribano, Effect of steam and carbon dioxide activation in the micropore size distribution of activated carbon. Carbon, 1996. 34 (4): p. 505 509. 35. Dabrowski, A., Adsorption from theory to practice. Advances in Colloid and Interface Science, 2001. 93 : p. 135 224. 36. Freudlich, Herbert, The elements of colloidal chemistry 3rd Ed. ed. 1925, New York: E.P. Dutton and Co. 37. Langmuir, Irving, The collected works of Irving Langm uir, with contributions in memoriam, including a complete bibliography of his works ed. G. Suits. Vol. 9. 1962, New York: Pergamon Press. 38. Somorjai, G.A., Introduction to Surface Chemistry and Catalysis 1994, Boston: John Wiley & Sons Inc. 39. Adamson, A.W., Physical Chemistry of Surfaces 1990: Wiley. 40. Oudar, J., Physics and Chemistry of Surfaces 1975: Blackie and Sons Ltd. 41. Frederick, Holly T., Fred S. Cannon, and Brian A. Dempsey, Calcium and TOC loading: effect of hydroxyl and car boxyl substituents. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 2001. 191 (1 2): p. 161 177.
84 42. Snoeyink, V.L. and W.J. Weber Jr., The surface chemistry of active carbon. A discussion of structure and surface functional groups. Enviro nmental Science & Engineering, 1967. 1 (3): p. 228 234. 43. Coughlin, R.W. and F.S. Ezra, Role of surface acidity in the adsorption of organic pollutants on the surface of carbon. Environmental Science & Engineering, 1968. 2 (4): p. 291 297. 44. Garten, V.A. D.E. Weiss, and J.B. Willis, A new interpretation of the acidic and basic structures in carbons. I. Lactone groups of the ordinary and fluorescein types in carbons. Australian Journal of Chemistry, 1957. 10 : p. 295 308. 45. Graham, D., Characterization o f physical adsorption systems. III. The separate effects of pore size and surface acidity upon the adsorbent capacities of activated carbons. Journal of Physical Chemistry, 1955. 59 (9): p. 896 900. 46. Coughlin, R.W., F.S. Ezra, and R.N. Tan, Influence of Chemisorbed Oxygen in Adsorption onto Carbon from Aqueous Solution. 1968. 28 (3): p. 386 396. 47. Dubinin, M.M. Chemistry and Physics of Carbon, ed. P.L.J. Walker. Vol. 2. 1966, New York: Marcel Dekker. 51. 48. Mahajan, O. P., C. Moreno Castilla, and P. L. Jr. Walker, Surface treated activated carbon for removal of phenol from water. Separation Science and Technology, 1980. 15 (10): p. 1733 1752. 49. Salame, Issa I. and Teresa J. Bandosz, Study of Water Adsorption on Activated Carbons with Different Degrees of Surface Oxidation. Journal of Colloid and Interface Science, 1999. 210 (2): p. 367 374. 50. Pendleton, P., S. H. Wong, G. Levay, R. Denoyel, and J. Rouquerol, Properties of activated carbon controlling 2 methylisoborneal adsorption. Carbon, 1997. 35 (8): p. 1141 1149. 51. Leng, C.C. and N.G. Pinto, Effects of surface properties of activated carbons on adsorption behavior of selected aromatics. Carbon, 1997. 35 (9): p. 1375 1385. 52. Radovic, L. R., I. F. Silva, J. I. Ume, J. A. Menendez, C. A. Leon y Leon, and A. W. Scaroni, An experimental and theoretical study of the adsorption of aromatics possessing electron withdrawing and electron donating functional groups by chemically modified activated carbons. Carbon, 1997. 35 (9): p. 1339 1348. 53. Nowack, K. O., F. Cannon, and D. W. Mazyck, Enhancing activated carbon adsorption of 2 methylioborneol Part 1. Submitted to ES&T, 2002. 54. Mueller, E.A. and Gubbins K.E., Molecular simulation study of hydrophilic and hydrophobic behavior of activated carbon surfaces. Carbon, 1998. 36 (10): p. 1433 1438.
85 55. Considine, Robert, Renaud Denoyel, Phillip Pendleton, Russell Schumann, and Shiaw Hui Wong, The influence of surface chemistry on activated carbon adsorption of 2 methylisoborneol from aqueous solution. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 2001. 179 (2 3): p. 271 280. 56. Mattson, J.S., H.B. Mark Jr, M.D. Malbin, W.J. Weber Jr, and J.C. Crittenden, Surface chemistry of active carbon: specific adsorption of phenols. Journal of Colloid and In terface Science, 1969. 31 (1): p. 116 130. 57. Radovic, L.R., C. Moreno Castilla, and J. Rivera Utrilla, Carbon materials as adsorbents in aqueous solutions Chemistry and physics of carbon, ed. L.R. Radovic. Vol. 27. 2000, New York: Marcel Dekker. 227 405. 58. Haydar, S., M. A. Ferro Garcia, J. Rivera Utrilla, and J. P. Joly, Adsorption of p nitrophenol on an activated carbon with different oxidations. Carbon, 2003. 41 (3): p. 387 395. 59. Barton, S.S., M.J.B. Evans, E Halliop, and J.A.F. MacDonald, Acidic a nd basic sites on the surface of porous carbon. Carbon, 1997. 35 (9): p. 1361 1366. 60. Layahe, Jacques, The chemistry of carbon surfaces. Fuel, 1998. 77 (6): p. 543 547. 61. Moreno Castilla, C., M.V. Lopez Ramon, and F. Carrasco Marin, Changes in surface ch emistry of activated carbons by wet oxidation. Carbon, 2000. 38 : p. 1995 2001. 62. Menendez, J. A., J. Phillips, B. Xia, and L. R. Radovic, On the Modification and characterization of chemical surface properties of activated carbon: In the search of carbo ns with stable basic properties. Langmuir, 1996. 12 (18): p. 4404 4410. 63. Herzing, D.R., V.L. Snoeyink, and N.F. Wood, Activated Carbon Adsorption of the Odorous Compounds 2 Methylisoborneol and Geosmin. JAWWA, 1977. 69 (4): p. 223. 64. Crittenden, J.C., P redicting GAC Performance With Rapid Small Scale Column Tests. JAWWA, 1991. 83 (1): p. 77. 65. Pirbazari, M., H.S. Borow, S. Craig, V. Ravindran, and M.J. McGuire, Physical Chemical Characterization of Five Earthy Musty Smelling Compounds. Water Science & T echnology, 1992. 25 (2): p. 81 88. 66. Lang, J.S., J.J. Giron, A.T. Hansen, R.R. Trussel, and W.W. Hodges Jr., Investigating Filter Perfor mance as a Function of the Ratio of Filter Size to Media Size. JAWWA, 1993: p. 122 130. 67. MacKenzie, J., M.F. Tennant D.W. Mazyck Tailored GAC for the Effective Control of 2 Methylisoborneol 2002, JAWWA, 2005 p. 76 87
86 68. Crittenden, J.C., J.K. Berrigan, H.W. Hand, Design of Rapid Small Scale Adsorption Tests for a Constant Diffusivity. J. Water Pollut., Control Fede ration 58 (4), p. 312 319 69. Tennant, M.F. and D.W. Mazyck, Steam Pyrolysis activation of Wood Char for Superior odorant removal Carbon, 2003. 41 : p. 2195 2202 70 Boehm, H. P., Some Aspects of the Surface Chemistry of Carbon Blacks and other Carbons. Ca rbon, 1994. 12 (5): p. 759 769 71. Adams, J.Q. and R.M. Clark, Cost Estimates for GAC Treatment Systems. JAWWA, 1989. 81 : p. 35 42. 72. Young, W.F., H. Horth, R. Crane, T. Ogden, and M. Arnott, Taste and Odour Threshold Concentrations of Potential Potable W ater Contaminants. Water Research, 1996. 30 (2): p. 331 340. 73. Medsker, L.L., D. Jenkins, J.F. Thomas, and C. Koch, Odorous Compounds in Natural Waters. 2 Exo Hydroxy 2 Methylbornane, the Major Odorous Compound Produced by Several Actinomycetes. Env. Sci. Tech., 1969. 3 (5): p. 476 477. 74. Gillogly, T.E.T., V. Snoeyink, J.R. Elarde, C.M. Wilson, and E.P. Royal, C14 MIB Adsorption on PAC in Natural Water. JAWWA, 1998. 90 (1): p. 98 108. 75. Nowack, K. O., F. Cannon, and D. W. Mazyck, Enhancing activated carb on adsorption of 2 methylioborneol Part 2. Submitted to ES&T, 2002. 76. Puri, B.R., 1970. Surface complexes on carbon. In: Walker, Jr., P.L.(Ed.), Chemistry and Physics of Carbon, vol. 6. Marcel Dekker, New York, NY. 77 Boehm, H. P., Functional Groups on the Surfaces of Solids. Angewandte Chemie, 1966. 5 (6): p. 533 543. 78 Mawhinney, D.B. and J.T. Yates Jr., FTIR study of the Oxidation of Amorphous Carbon by Ozone at 300 K Direct COOH Formation. Carbon, 2001. 39 : p. 1167 1173. 79. Tennant, M.F. and D.W. Mazyck, The Role of Surface Acidity and Pore Size Distribution in the Adsorption of 2 methylisoborneol via Powdered Activated Carbon Carbon, 2007. 45 : p. 858 864 80. Kipling, J.J., C.G. Gasser, Adsorption from liquid mixtures at solid surfaces. J. P hys. Chem. 64 (6), p. 710 81 Curran, M.A, Environmental Life Cycle Assessment ed. McGraw Hill. 1996, New York. 82 Graedel, T.E. and B.R. Allenby, Industrial Ecology ed. P. Hall. 1995, New Jersey.
87 83 Vigon, B.W., Tolle, D.A., Cornaby, B.W., Latham, H.C. Harrison, C.L., Boguski, T.L., Hunt, R.G.,and Sellers, J.D., Life Cycle Assessment: Inventory Guidelines and Principles in U.S. Governmental Printing Office U.S.E.P. Agency, Editor. 1993: Washingtion D.C. 84. Graedel, T.E.; Allenby, B.R. Industrial E cology; Prentice Hall: New Jersey, 1995. 85. Vigon, B.W.; et al. Life Cycle Assessment: Inventory Guidelines and Principles; U.S. Environmental Protection Agency; U.S. Governmental Printing Office: Washington, DC, 1993; EPA 600 R 92 245. 86. Curran, M.A. Environmental Life Cycle Assessment; McGraw Hill: New York, 1996. 87. Wastewater Technology Fact Sheet Granular Activated Carbon Absorption and Regeneration : 2000. U.S. Environmental Protection Agency, Office of Water; EPA 832 F 00 017; http://www.epa.gov /owm/mtb/carbon_absorption.pdf (accessed Oct 2002). 88. Mantell, Charles L. Carbon and Graphite Handbook ; John Wiley and Sons, Inc: New York, 1968. 89. Droste, R.L. Theory and Practice of Water and Wastewater Treatment ; John Wiley & Sons, Inc: New Y ork, 1997. 90. Roskill The Economics of Activated Carbon 5 th Ed.; Roskill Information Services: London, 1994. 91. Cannon F.S.; Knappe D.R.U.; Snoeyink V.L.; Lee R.G.; DeWolfe J.R.; Dagois G. The Effect of Metals on Thermal Regeneration of Granular Activa ted Carbon; AWWA Research Foundation and AWWA Denver: 1994. 92. Lykins B.W.; Clark R. M. Granular Activated Carbon: Design, Operation and Cost ; Lewis Publishers, Inc: Michigan, 1991. 93. Rodriquez Reinoso F. In Introduction to Carbon Technologies; Marsh, H.; Heintz E.A.; Rodriguez Reinoso F, Ed.; Alicante, 1997; pp. 35 101. 94. Kashinkunti, R. Water Quality & Treatment Cincinnati Water Works, OH, personal communication, 2002. 95. Brunauer, S.; Emmett, P.H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309 316 96. Adams, J.Q.; Clark, R.M.; Lykins, B.W.; DeMarco, J.; Kittredge, D. J. Environ. Eng. 1988 114.4, 944 961. 97. U.S. Department of Labor, Bureau of Labor Statistics Inflation Calculator. http://www.bls.gov (accessed Oct 2002).
88 98. Cooper, S.P. Huck Fastners, an ALCOA business, Tucson, AZ, personal communication, 2002. 99. DeMarco, J.; Miller, R.; Hartman, D. JAWMA 1988, 80.3, 66 72. 100. U.S. Department of Energy. http://www.eia.doe.gov (accessed Nov 2002). 101. Rumpke Sanitary Landfill, Customer Representative, Cincinnati, OH, personal communication, 2002. 102. Pennsylvania Department of Environmental Protection, Residual Waste Report 2 002 ; http://www.dep.state.pa.us/wm_apps/residualwaste/default.asp (accessed Oct 2002). 103. Townsend, T.G. University of Florida, Gainesville, 104. California Air Resources Board (ARB), 2002. Mobile Source Emissions Inventory Program, Heavy Duty Diesel Trucks Emission Factors. 105. http://www.arb.ca.gov/msei/on road/downloads/tsd/HDT_Emissions_ New.pdf (Accessed Dec 2002). 106. U.S. EPA, 1995. AP 42: Compilation of Air Pollution Emission Factors. http://www.epa.gov/otaq/models/ap42/ap42 h7.pdf (Accessed Dec 2002). 107. Argonne National Laboratories and the University of Chicago. The Greenhouse Gases, Regulated Emission, and Energy Use in Transportation (GREET) Model. Version 1.6 Software downloaded from http://www.transportation.anl.gov/ttrdc/greet/ (accessed Nov 2002). 108. Golder Associates. Survival Reference Guide 109. Neville Island Air Pollution, Neville Island Good Neighbor Committee, http://trfn.clpgh.org/nignc/emit2.htm (accessed Oct 2002). 110. Calgon Carbon, Co rp. Sales Representative, Pittsburgh, PA, personal communication, 2002. 111. Calg on Carbon, Corp. http://www.calgoncarbon.com/corporate/investor/financial.php (accessed Oct 2 002). 112. Bansal, R. C.; Donnet, J B ; Stoeckli, F. Active Carbon ; Marcel Dekker, Inc:New York, 1988. 113. Allen, D.T.; Shonnard, D.R.; Green Engineering: Environmentally Conscious Design of Chemical Processes ; Prentice Hall: New Jersey, 2002. 114. Grae del, T.E. Streamlined Life Cycle Assessment ; Prentice Hall: New Jersey, 1998.
89 BIOGRAPHICAL SKETCH T homas E Chestnutt Jr. was born in the state of Bavaria West Germany, to Mr. Thomas Elmore Chestnutt Sr. and Mechthild Chestnutt. Upon graduating from Gul f High School in New Port Richey, FL in 1991, he went onto receive a scholarship to attend the University of Dayton. He graduated in 1995 with a degree in biochemistry and minors in German, Chemistry, and biology. He also played Divis i on I soccer at the University of Dayton. After undergraduate studies, he went to play professional soccer in Tampa, FL and finished in Engineering Mechanics, studying biomechanics in the Aerospace Department at the University of Florida, under the supervision of Dr. B.J. Fregly in 2001. He then realized his passion for the environment, the importance of drinking water and sustainability in every facet of life, thus began his doctoral work with Dr. David W. Mazyck in Environmental Engineering. After obtaining his Ph.D. in December 2017, he will begin working with young people to help create sustainable practices that will positively impact our community.