Supercritical fluid extraction of organics from coal gasification wastewaters using carbon dioxide by Ambrey Gordon Gartner A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Chemical Engineering Montana State University © Copyright by Ambrey Gordon Gartner (1985) Abstract: Existing chemical extraction techniques for the removal of phenols from coal gasification wastewater require a moderately expensive solvent and additional distillation equipment for solvent recovery. This research was designed to determine if supercritical carbon dioxide, a non-toxic, non-flammable, and relatively inexpensive byproduct of coal gasification, could be used for phenol extraction. Experiments were run to investigate phase equilibria associated with the two phase system consisting of water, phenol, and carbon dioxide. System pressures varied between 1100 and 3000 psi at a temperature of 40 degrees Celsius. Phenol concentrations in the water phase varied from 600 to 8500 mg/l. Phase equilibrium data were first used to determine the ability of supercritical carbon dioxide to extract phenol from water, and later to determine the extraction column characteristics. Extraction equipment was built to test the feasibility of continuous extraction of phenol from water and wastewater. Twenty-three successful experiments were run, including one counter-current and seventeen co-current experiments using phenol/water standards of known initial concentration, and five co-current experiments using wastewater. The best extraction capability was obtained using counter-current extraction techniques. Extraction efficiencies, based on the approach to theoretical equilibrium, were determined for each of the co-current experiments. These efficiencies ranged from 81% to 145% for the phenol/water standard experiments, and from 76% to 163% for wastewater experiments. Relative extraction efficiencies were based on Run No. 5, which had the highest efficiency of a phenol/water standard experiment (145%). Extraction of phenol and phenol derivatives from coal gasification wastewater appeared to be higher than would be expected, based on equilibrium data. This could be due to additional experimental error involved in measuring the original phenol concentration in the wastewater. Also, the existence of co-solvents in the wastewater could enhance the solvent power of supercritical carbon dioxide.  SUPERCRITICAL FLUID EXTRACTION OF ORGANICS FROM COAL GASIFICATION WASTEWATERS USING CARBON DIOXIDE by Ambrey Gordon Gartner A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Chemical Engineering MONTANA STATE UNIVERSITY Bozeman, Montana October, 1985 N372om Oo-p" oV i i APPROVAL of a thesis submitted by Ambrey Gordon Gartner This thesis has been read by each member of the thesis committee and has been found to be satisfactory regarding content, English usage, format, citations, bibliographic style, and consistency, and is ready for submission to the college of Graduate Studies. I o / z.3 / <0 Date Chairperson, Graduate Committee Approved for the Major Department Date He?ad, Major Department Approved for the College of Graduate Studies //- /S' . Date Graduate Dean STATEMENT OF PERMISSION TO USE In presenting this thesis in partial fulfillment of the requirements for a master's degree at Montana State University, I agree that the Library shall make it available to borrowers under rules of the Library. Brief quotations from this thesis are allowable without special permission, provided that accurate acknowledgement of source is made. Permission for extensive quotation from or reproduction of this thesis may be granted by my major professor, or in his absence, by the Director of Libraries when, in the opinion of either, the proposed use of the material is for scholarly purposes. Any copying or use of the material in this thesis for financial gain shall not be allowed without my written permission. Signature VACKNOWLEDGEMENTS I would like to express my appreciation to the Chemical Engineering Department faculty at Montana State University for their guidance and help during the course of my research. Special appreciation.is extended to Dr. Warren P. Scarrah for his extra time and patience. In addition I would to thank Lyman Fellows for his craftsmanship and advice, Drs. Lloyd Berg and Phil McCandless for their gas chromatograph expertise, and Dr. Amirtharajah, Margie Reagen, and Barb Carlson for their assistance in sample analyses and the use of Environmental Engineering lab facilities. Deepest appreciation is extended to my wife, Sharon, whose support made all the work possible, and to Rachel and Benjamin, who gave the work meaning. vi TABLE OF CONTENTS PAGE LIST OF TABLES . . .................... viii LIST OF F I G U R E S ..................................... ix ABSTRACT . . . . . . . xii INTRODUCTION AND PREVIOUS RESEARCH ................ I Introduction ...................... I Supercritical Fluids ............................ 3 Extraction with Supercritical Fluids . . . . . . . . 7 Carbon Dioxide as a Solvent . .............. .. . 10 Previous Research............................... 14 RESEARCH OBJECTIVE ................................ 17 MATERIALS, EQUIPMENT, AND PROCEDURE . . . ........ 18 Materials............ 18 Equipment.............. '...................... 18 Equilibrium Data Equipment . . ................ 18 Extraction Apparatus .......................... 19 Analytical Equipment .......................... 27 Experimental Procedures ........ 28 Phase Equilibrium Run Procedures .............. 28 Extraction Run Preparations . ................ 29 Extraction Run Procedures .................... 31 Sample Analyses . ............................ 34 • EXPERIMENTAL RESULTS AND DATA ANALYSIS . ; ........ 38 Introduction .................................... 38 Reliability of D a t a ............................ 39 Phase Equilibria................................ 46 Counter-current Extraction ........ 54 Phenol/Water Co-current Extraction .............. 63 Wastewater Co-current Extraction . .............. 80 SUMMARY . ............................ 84 PAGE RECOMMENDATIONS FOR FUTURE STUDY . . . . . . . w ... . 88 NOTATIONS........................................ 91 REFERENCES CITED.................... '............. 9 5 APPENDICES.......................................... 100 Appendix A. Analytical Procedures . . ............ 101 Ammonia Analysis Using the Orion 901 Analyzer . 102 Chemical Oxygen Demand (COD) - Rapid Method . . 104 Appendix B. Sample Calculations ................ 107 Calculation of Phenol Mole Fractions . . . . . . . 108 Determination of Actual Distribution Coefficient.................................. HO Determination of Theoretical Distribution Coefficient.................................. Ill 'Determination of Extraction Efficiency, Relative Efficiency, and Relative Time . . . 112 Determination of Absolute Phenol Extraction . . 113 vii viii LIST OF TABLES TABLE PAGE 1. Comparison of Carbon Dioxide Fluid Properties . . 12 2. Heat of Vaporization of Carbon D i o x i d e ........ 12 3. References with Examples of Supercritical Fluid Extraction Using Carbon Dioxide............. 13 4. References with Solubility of Solids and Liquids in Supercritical Carbon Dioxide ................. 15 5. Raw Phase Equilibrium Data for the Phenol/Water/ Carbon Dioxide Ternary System . . .............. 46 6. Mole Fractions of Phenol in Water (x). Phenol in Carbon Dioxide (y), and Equilibrium K-values . . 47 7. Calculated Values of y (Mole Fraction of Phenol in Carbon Dioxide) at 0.33 and 0.803 g/cc . . . . 52 8. Values for Calculating the Number of Overall Gas-phase Transfer Units ....................... 61 '9. Summary of Co-current Extraction Run Data . . . . 64 10. Converted Co-current Extraction Run Data . . . . 66 11. Summary of Phenol Loading Capability of Carbon Dioxide at 40 Degrees Celsius . . . ............ 72 ix LIST OF FIGURES FIGURE PAGE 1. Reduced Pressure-Density Diagram Showing Supercritical Fluid (SCF) and Near-Critical Liquid (NCL) Regions . . . .......... . . . . . 5 2. Schematic of Counter-current Extraction Apparatus as Originally' Designed and B u i l t ......... 20 3. Schematic of Workable Counter-current Extraction Apparatus................................... . 24 4. Schematic of Concurrent Extraction Equipment . . 25 5. Phenol/Water/Carbon Dioxide Ternary Phase Diagram........................................ 40 6. Experimental Pressure and Density Diagram for Carbon Dioxide at 40 Degrees Celsius . . . . . . 42 7. Three Dimensional Pressure-composition Plot of Phenol/Water/Carbon Dioxide System at 40 Degrees Celsius, Rotated 45 Degrees . .................. 48 8. Three Dimensional Pressure-composition Plot of Phenol/Water/Carbon Dioxide System at 40 Degrees Celsius, Rotated 135 Degrees .................. 49 9. Phenol/Water/Carbon Dioxide Equilibria Measured Near.1200 and 2400 psi, at 40 Degrees Celsius . . 51 10. Calculated Values of x (Mole Fraction of Phenol in Water) and y (Mole Fraction of Phenol in Carbon Dioxide) for the Phenol/Water/Carbon Dioxide System, for Varying Densities, at 40 Degrees Celsius................................ . 53 11. Graphically Determined Values of Carbon Dioxide Phase Density and Composition for the Phenol/ Water/Carbon Dioxide System, for Varying Values of x (Mole Fraction of Phenol in Water), at 40 Degrees Celsius 55 X12. Calculated Distribution Coefficients for Varying Values of Carbon Dioxide Density, at 40 Degrees Celsius..................................... 56 13. Flow Schematic for Counter-current Extraction R u n ......................................... 57 14. Working Curves for Obtaining Gas-phase Composition Values .............................. 58 15. Graphical Integration of the Number of Overall Gas-Phase Transfer Units .............. . . . . 62 16. Normalized Extraction Efficiencies and Residence Times for Co-current Phenol/Water Extraction Experiments at 2400 psi and 40 Degrees Celsius . 68 17. Normalized Extraction Efficiencies and Residence Times for Co-current Phenol/Water Extraction Experiments at 2800 psi and 40 Degrees Celsius . . 69 18. Normalized Extraction Efficiencies and Residence Times for Co-current Phenol/Water Extraction Experiments at 1500, 1800, and 2000 psi and 40 Degrees C e l s i u s ......................... 70 19. Phenol Extraction Capabilities of Carbon Dioxide with' Respect to Pressure, at 40 Degrees Celsius . 74 20. Summary of Normalized Extraction Efficiencies and Residence Times for All Co-current Experiments Using Phenol/Water Standards ................... 75 21. Equations of Best Fit for Normalized Extraction Efficiencies and Residence Times for Co-current Phenol/Water Extraction Experiments at 2400 and 2800 psi and 40 Degrees Celsius . ................ 77 22. Equation of Best Fit for Normalized Extraction Efficiencies and Residence Times for All Co-current Experiments Using Phenol/Water Standards, at 40 Degrees Celsius . . . . FIGURE PAGE 78 xi 23. Plot Showing Values for Normalized Extraction Efficiencies and Residence Times for All Co-current Experiments Using Phenol/Water Standards, at 40 Degrees Celsius Differentiated by Test Pressure............... 79 24. Equation of Best Fit for Normalized Extraction Efficiencies and Residence Times for Experiments Using Coal Gasification Wastewater, at 40 Degrees Celsius................. 81 25. Plot Showing Values for Normalized Extraction Efficiencies and.Residence Times for All Co-current . Experiments, at Degrees Celsius, Differentiated by Phenol/Water Standards and Coal Gasification Wastewater ........................... 83 FIGURE . PAGE xii ABSTRACT Existing chemical extraction techniques for the removal of phenols from coal gasification wastewater require a moderately expensive solvent and additional distillation equipment for solvent recovery. This research was designed to determine if supercritical carbon dioxide, a non-toxic, non-flammable, and relatively inexpensive byproduct of coal gasification, could be used for. phenol extraction. Experiments were run to investigate phase equilibria associated with the two phase system consisting of water, phenol, and carbon dioxide. System pressures varied between 1100 and 3000 psi at a temperature of 40 degrees Celsius. Phenol concentrations in the water phase varied from 600 to 8500 mg/1. Phase equilibrium data were first used to determine the ability of supercritical carbon dioxide to extract phenol from water, and later to determine the extraction column characteristics. Extraction equipment was built to test the feasibility of continuous extraction of phenol from water and wastewater. Twenty-three successful experiments were run, including one counter-current and seventeen co-current experiments using phenol/water standards of known initial concentration, and five co-current experiments using wastewater. The best extraction capability was obtained using counter-current extraction techniques. Extraction efficiencies, based on the approach to theoretical equilibrium, were determined for each of the co-current experiments. These efficiencies ranged from 81% to 145% for the phenol/water standard experiments, and from 76% to 163% for wastewater experiments. Relative extraction efficiencies were based on Run No. 5, which had the highest efficiency of a phenol/water standard experiment (145%). Extraction of phenol and phenol derivatives from coal gasification wastewater appeared to be higher than would be expected, based on equilibrium data. This could be due to additional experimental error involved in measuring the original phenol concentration in the wastewater. Also, the existence of co-solvents in the wastewater could enhance the solvent power of supercritical carbon dioxide. IINTRODUCTION AND PREVIOUS RESEARCH Introduction One of the major engineering challenges of this century will be the conversion from an energy system based on naturally occuring crude oil to a system based on the production of synthetic fuels. Future fuels will consist of hydrocarbons, alcohols, and hydrogen produced mainly from coal, but also from lignite, peat, biomass, and eventually from genetically programmed organisms utilizing sunlight. In the production of synthetic fuels from coal and lignite, one of the main goals of most existing technology is the production of carbon monoxide and hydrogen, commonly called synthesis gas. ■ These chemical building blocks are then combined in a variety of ways to form a wide spectrum of hydrocarbons and alcohols. However, in the production of the synthesis gas, there are many side reactions involving steam, elemental carbon, and the complex mixture of mole­ cules that comprise coal and lignite. Given the appropriate reactor residence times, temperatures, and pressures, a family of prominent side reactions are associated with any particular type of conversion facility. 2The Lurgi coal gasification reactor, one of the older and more proven technologies, produces a significant amount of water soluble chemicals, including phenols, phenol derivatives, and ammonia. Data obtained from the Grand Forks Slagging Gasifier at the Grand Forks Energy Research Center indicate that this type of reactor typically produces wastewater streams containing up to 8000 parts per million (ppm) phenol and phenol derivatives, and ammonia concentrations approaching 14,000 ppm. As a result considerable effort and expense are required to remove these components from process water. These concentrations are toxic to micro-organisms commonly used in wastewater treatment, resulting in designs relying on chemical treatment. Standard liquid-liquid extraction of phenols from water with isopropyl ether requires a moderately expensive solvent and the addition of extra distillation equipment for solvent recovery. There are distinct advantages for finding another solvent that would eliminate the need for the additional separations equipment. Carbon dioxide appears to be an economically attractive solvent because it is a readily available coal conversion byproduct and could be used on a once-through basis without recovery. Fortunately, carbon dioxide is also an effective solvent for supercritical fluid extraction (SCFE) and has been used to remove organics from 3many solid mixtures. Applications of SCFE using carbon dioxide include isolating valuable constituents from natural materials like oil seeds, hops, and spices and removing or reducing levels of undesirable substances such as nicotine or caffeine. However, there are few, if any, commercial applications where supercritical carbon dioxide is used to extract components from a liquid. Supercritical Fluids All substances possess a critical temperature, above which the liquid phase cannot exist. At the critical temperature the pressure that must be applied to cause condensation is called the critical pressure. Alter­ natively, the critical pressure can be regarded as the vapor pressure of the liquid at its critical temperature (I). When a liquid is heated in a sealed container, the properties of the liquid and vapor approach one another near the critical temperature. Below the critical temperature there are two distinct phases. As the sample is heated through the critical temperature, the meniscus separating the two phases disappears and one phase results. This phase is referred to as a supercritical fluid. Although the fluid has properties of both a liquid and a gas it is in reality neither a gas nor,a liquid. 4At supercritical temperatures a substance that is gaseous at ambient conditions will begin to behave like a liquid when the pressure is raised sufficiently. Below Sl0C, for example, liquid carbon dioxide can be used to dissolve natural oils and a wide range of other animal and vegetable materials. At. temperatures greater than 31°C carbon dioxide does not liquify, but it.still can be used as a solvent if the pressure is raised sufficiently for its density to become similar to that of a liquid (2). The supercritical fluid (SCF) region for a pure component is generally defined as that region of temperatures and pressures greater than or equal to the critical temperature and critical pressure of the pure component. For practical considerations, the SCF region of interest is defined less rigorously here to include the conditions bounded approximately by 0.9 1.0. In this region the SCF is highly compressible, as illustrated by de Filippi (3) in Figure I. For example, at a constant T of 1.10, increasing pressure from Pr < 1.0 to Pr > 1.0 significantly increases the density from relatively low values to liquid-like densities. At higher reduced temperatures, the pressure increase required to produce an equivalent density increase becomes greater. This practical consideration sets the upper bound on temperature. At a constant P of 1.50, Re du ce d Pr es su re 5 Normal liquid density Reduced Density Taken from de Philippi (3). Figure I. Reduced Pressure-Density Diagram Showing Supercritical Fluid (SCF) and Near Critical Liquid (NCL) Regions. 6decreasing, temperature has a similar effect on density, and at higher reduced pressures, the.density is less sensitive to temperature changes. In the vicinity of the critical point,, large density changes can be produced with either relatively small pressure or temperature changes (4). The use of supercritical fluids as solvents for extraction has unique advantages. Supercritical fluids combine the solvent power of a liquid with the transport properties of a gas. They can be used to selectively dissolve relatively involatile solutes.from a mixture while allowing the simple separation of solvent and solute by either a reduction in pressure or an increase in temperature. Pressure is the principal processing- variable used to alter the solvent power. In contrast, liquid extraction solvent power is varied by adjusting the temperature or the solvent composition; allowable temperatures are limited by the thermal characteristics of the solute while variations in solvent composition increase separation and purification costs. Significant reductions in energy requirements are also characteristics of SCFE separations (5). 7Extraction with Supercritical Fluids Supercritical fluid extraction (SCFE) is a technique ■ that exploits.the solvent power of supercritical fluids at temperatures and pressures near the critical point. By selectively choosing the appropriate solvent a wide range of operating temperatures and pressures can be utilized to tailor an extraction process design to a particular need. It is possible to dissolve both solids and liquids (including ones of low volatility) in compressed gas. The behavior of gas/solid systems is normally simpler than that of gas/liquid systems as the composition of the solid phase normally remains constant. When a liquid is being extrac­ ted, the situation is complicated by the influence of dissolved gas on the properties of the liquid phase (2). The advantages of using supercritical fluids under pressure include the following: High boiling components can be extracted at relatively low temperatures. Also, at the lower temperatures involved, heat sensitive compounds are undamaged during the extraction process. Separation of the solute and solvent is relatively easy at supercritical conditions, usually by relatively slight pressure changes. Separations can be achieved which are not possible by other techniques such as distillation or extraction. 8. If food components are to be separated, non-toxic solvents, such as carbon dioxide, may be used which leave no harmful residues. In general terms, compressed gases are relatively cheap solvents. The difficulties which have to be overcome include the following: High plant costs involving relatively high risk capital. Cost of maintenance normally required for high pressure plants. Current lack of process design data and lack of information on equipment reliability. In many unique and difficult processes there is a tendency to use patent laws as a protection against competition. The mass of patents in the field makes it extremely difficult to determine if infringement is likely in a particular application. Supercritical fluid extraction is particularly effective in the isolation of substances of medium molecular weight and relatively low polarity. Its principal advantage over distillation is that separation can be accomplished at moderate temperatures, and thus can be applied to the recovery of heat-labile substances of low volatility. It is therefore of interest in the food and petroleum industries. ) 9In comparison with liquid solvents, the supercritical fluid has high diffusivity and low density and viscosity, allowing rapid extraction and phase separation. The solvent power of the supercritical fluid can be varied over a wide range by varying pressure or temperature, and the fluid is easily recovered from the extract and from the extraction residue as a result of its high volatility. Again, the principal disadvantage of SCFE is the relatively high pressures required to achieve the necessary densities, generally in the range of 50 to 200 atmospheres. Another disadvantage is the complexity of the phase relationships under supercritical conditions. Theoretical techniques for predicting extraction behavior are new and relatively untested. An extensive bibliography assembled by McKugh and Krukonis (6) shows that most of the modeling work in this area has been developed in the last ten years (7-19). The SCFE concept was first recognized by Hannay and Hogarth (20) over a hundred years ago. They observed that potassium iodide was dissolved in supercritical ethanol and was precipitated on reducing the pressure. Later it was realized that the solvent power of supercritical fluids could be involved in geological processes through the influence of water on rock formation, and of methane in petroleum formation and migration. In power stations the 10 adoption of supercritical steam pressures led to the deposition of silica on the blades of steam turbines (21). The possibilities for substantial energy economies, relative to normal distillation processes, have made SCFE an area of intense interest. Arthur D. Little CS) probably triggered this interest with the claim that supercritical fluid extraction would make it possible to produce a gallon of ethanol using carbon dioxide as extractant with ah expenditure of 10,000 Btu of energy instead of the 25,000 to 50,000 Btu needed for conventional distillation. Carbon Dioxide as a Solvent Carbon dioxide has many characteristics which make it suitable for use as a solvent. It is completely miscible with low molecular weight hydrocarbons and oxygenated organics and is therefore a good solvent for many organics (3,22 ). It is relatively volatile, compared with most organics likely to be ,extracted, allowing easy separation. In addition, carbon dioxide has low viscosity and high diffusivity, and it is totally non-toxic, non-flammable, and relatively inexpensive (23). . The solubility of supercritical carbon dioxide in water is relatively small, generally less than seven-'weight percent, even at pressures exceeding 400 atmospheres. At near critical conditions, for example at 40°C and 80 11 atmospheres, the solubity is approximately 5.5 weight percent (24). This allows the use of carbon dioxide for a variety of extraction chores involving aqueous solutions. Furthermore, carbon dioxide can be removed easily by simple deaeration techniques. As a supercritical fluid carbon dioxide has additional properties that make it an excellent solvent. Its critical temperature (31°C) and pressure (72.8 atm) are readily accessible with well-established process technology, and equipment. Finally, its heat of vaporization is low, especially near the critical point, leading to low energy requirements in many processes (3). Table I shows a comparison of the fluid properties of carbon dioxide as a liquid, a supercritical fluid, and as a gas (3, 24). The ranges shown for the supercritical fluid are those in the near critical region. At sufficiently high pressures (approximately 500 atm at 40°C) the density of supercritical carbon dioxide can actually exceed I g/cc. Also, in the near critical region significant energy savings can result due to the very low heats of vaporization. Table 2 shows selected data relating temperature, pressure and heat of vaporization (3). The zero value listed at the critical point illustrates the unique quality that any fluid possesses when the distinction between a liquid and a gaseous phase disappears. 12 Table I. Comparison of Carbon Dioxide Fluid Properties. Liquid Supercritical Gas Density (g/cc) 0.5 - 1.2 0.2 - 0.7 0.001 • Viscosity (cP) 0.5 — 1.0 0.05 - 0.10 0.01 Diffusivity (cm^/s) 10™5 IO"4 - IO"3 10"1 Table 2. Heat of Vaporization of Carbon Dioxide. Saturation Heat of Temperature Pressure Vaporization F C (atm) (Btu/lb). 0 -17.8 20.8 120.1 20 4.4 38.6 95.0 60 15.6 50.9 76.6 70 21.1 58.1 63.8 80 26.7 65.9 44.8 87.8 31.0 72.7 0 Supercritical fluid extraction with carbon dioxide has been used extensively in the food industry. Here the properties of being non-toxic and non-flammable have significant advantages when extracting such valuable products as perfume oils, caffeine, hops, and vegetable 13 oils. It also is possible to remove.PCB1s from silicon- based transformer oils and to remove adsorbed materials from common adsorbents such as activated carbon and synthetic resins.. Table 3 shows a number of examples of the use of carbon dioxide for supercritical fluid extraction. Table 3. References with Examples Extraction Using Carbon of Supercritical Dioxide. Fluid Substance i-3 O O P, atm ref. Vegetable oil 20, 55 150-450 3 Polychlorinated biphenyls (PCB) 100 120-265 ■ 3 Silicone oil 100 160-250 3 Alachlor (pesticide) 50 - 120 88-272 3 Caffeine 90 160-220 21 Hops A O unidentified 25 Lilac oil 34 90 26 Lemon peel 40 300 27 Black pepper 60 100-400 27 Almond oil 40 600 27 14 Previous Research Research done previously at Montana State University has investigated the use of supercritical fluids for separating organics from the undesirable mineral portion of peat. Granlund (28) and Smith (29) investigated the semi-continuous flow extraction process where a variety of supercritical and near supercritical fluids were passed through a container containing a measured amount of peat.. Organics were extracted from the mineral portion of the peat and yields were measured for each type of solvent used. Earlier-work performed by Myklebust (30) investigated the potential for utilizing supercritical fluids for batch extraction of organics from peat. The University of Delaware Chemical Engineering Department has been active in SCFE research. Van Leer and Paulaitis investigated the solubilities of phenol and chlorinated phenols in supercritical carbon dioxide (31), and McHugh and Paulaitis investigated the solubilities of solid naphthalene and biphenyl in supercritical carbon dioxide (32). Kurnik, Holla, and Reid (33) of the Massachusetts Institute of Technology studied the solubility of several solids in supercritical carbon dioxide and ethylene. They 15 include a table of references with the solubilities of various solids in supercritical fluids. This data is shown in Table 4. Table 4. References with Solubility of Solids and Liquids in Supercritical Carbon Dioxide. Substance T, C P, atm ref. Naphthalene 35, 45, 55 100-300 23 Naphthalene 35-55 60-330 34 Ethanol/Water 40, 60 75-204 3 5 Isopropanol/Water 40 102 35 Phenanthrene 40 135-544 36 Diphenylamine 32-37 50-225 37 : Phenol 36, 60 78-239 31 Chlorinated Phenols 36 80-130 31 Napthalene 35-65 86-288 32 Biphenyl 36-58 105-484 32 Biphenyl 55 503-531 38 Naphthalene 58-80 1-250 38 16 Probably the most extensive research involving the solubility of various substances in carbon dioxide was published by Francis in 1954 (39). Mutual solubilities of liquid carbon dioxide with each of 261 other substances were reported. Nearly.half of these substances are miscible with carbon dioxide. Francis noted some relations to structure and plotted phase diagrams for 464 ternary systems involving carbon dioxide. He also pointed out that carbon dioxide has a strong homogenizing action upon pairs of other liquids at moderate concentrations, but a precipitating action at higher concentrations. In contrast to most solvents it has a selectivity against dicyclic hydrocarbons. Co-solvents were found necessary to make these unusual properties effective in the solvent extraction of hydrocarbon mixtures. While this paper does not include any research involving a supercritical phase, the measurement conditions ("liquid carbon dioxide at or near room temperature") are in the near-critical region, and the data provides an invaluable tool for roughly estimating the probable success one might have using supercritical carbon dioxide with any of the same combinations of substances. Using some of the groupings that Francis presents might also help predict the suitability of using supercritical carbon dioxide as an extractive agent for similar mixtures. 17 RESEARCH OBJECTIVE The purpose of this research is to investigate the feasibility of using supercritical carbon dioxide to extract phenols from coal gasification wastewater. In order to achieve this ultimate objective the work presented in this paper focuses on the development and testing of the necessary extraction equipment, investigation of water/ phenol/carbon dioxide equilibrium data, and the development of procedures for analyzing phenols in dilute aqueous solutions. . Extraction equipment was ordered, assembled, tested, and modified to obtain phenol extraction data at approxi­ mately 40°C and pressures ranging from 1100 to 2800 psig. Efforts were made to obtain data representing continuous flow conditions, both counter-current and co-current. These data are compared with the equilibrium . data, obtained as part of this research, to evaluate the effectiveness of each set of extraction conditions. 18 MATERIALS, EQUIPMENT, AND PROCEDURE Materials Carbon dioxide was obtained as a technical grade liquid, heated, and compressed into a special bottle rated at 6000 psi. A Haskel Model.AGD-62-C Gas Booster compressor was used to boost the CO^ from approximately 850 psi to 5000 psi. Original extraction data was obtained using a 10,000 mg/1 reagent grade phenol in distilled water standard solution. After the extraction equipment had been sufficiently tested and operating parameters better understood, another type of solution was tested for extraction. This solution was obtained from the Grand Forks Energy Research Facility and consisted of process water condensed from their experimental slagging gasifier. Equipment Equilibrium Data Equipment Equipment used for generating phase equilibrium data included a Parr Instrument Company Series 4000.pressure re­ action apparatus. This 500 milliliter bomb was constructed 19 of Inconel Ibl-S alloy. The bomb was specially fitted with a modified head to allow two valves for filling with the appropriate phenol/water solution and pressurizing with carbon dioxide. One of the valves was.also used for sampling the liquid phase. The apparatus also included a rocker equipped with an electrical heater which was used to keep the fluids thoroughly mixed at 40°C. The bomb temperature was monitored using a thermocouple located at the base of the bomb and connected to a 12 channel temper­ ature indicator. Temperature was manually controlled by a Fisher Scientific rheostat which supplied power to the rocker heating coils. Extraction Apparatus The equipment used in this research was designed to facilitate extracting phenol from water on a continuous basis. Figure 2 shows a schematic of the original extrac­ tion equipment. The carbon dioxide and water feed systems were essentially unchanged for the duration of all the extraction runs. All of the significant equipment changes involved the column itself and solution and solvent discharge lines. Essentially all the high pressure lines used in this research were located behind a three-sided. Wet test meter Discharge carbon dioxide Surge check valve Phenol accumulator Phenolic Measuring wastewater burrette Safety f]-- Carbon dioxide -.fillJ Porc stripped water vent Check valve Nitrogen fill port Nitrogen vent Extraction column, Heated enclosure Heated, high pressure carbon dioxide feed tank Water pump High pressure BPR - Back pressure regulator TI - Temperature indicator PI - Pressure indicator nitrogen storage FI - Flow MO Figure 2. Schematic of Counter-current Extraction Apparatus as Originally Designed and Built. 21 high-strength.steel shield. Control valves, gages, and electronic meters were accessible from the front of the shield. Carbon dioxide was stored under sufficient pressure to provide the appropriate test pressure for each run. The Haskel Gas Booster compressor was used to transfer. CC^ from a standard cylinder to a special high pressure cylinder rated at 6000 psi. The high pressure cylinder was warmed to supercritical temperatures (approximately 40°C) using an electrical heating blanket. The temperature was manually controlled with a Powerstat Type 3PN1168 variable transformer, and monitored using a thermocouple connected to a Cole-Palmer Model 8530 12 channel electronic indicator. Carbon dioxide pressure was regulated by a Tescom Model 44-1126-24-013 back pressure regulator. Heat was supplied with a 100 watt electrical heat tape as the pressure was reduced to provide heat of vaporization and to further raise the temperature to approximately 40°C. The temperature was measured with a thermocouple attached to the regulator surface and monitored using the Cole-Palmer indicator. From the regulator the solvent then entered a system of control and check valves which provided flow control to the extraction column. Once the carbon dioxide had passed through the column its pressure was reduced to atmospheric using a 22 Grove Mity-Mite Model S-91XW back pressure regulator. Back pressure was supplied by a Parr Instrument Company Series 4000 pressure reaction apparatus filled with nitrogen up to 5000 psi. This 500 milliliter bomb was constructed of Inconel 757-S alloy. Once again the.solvent stream was heated with electrical heat tape (200 watts) while its pressure was reduced. This was to prevent icing (both water and CO2) and to help prevent the deposition of solid phenol in the regulator and exiting tubing. Both heat tapes were manually controlled with a Powerstat Type 3PN1168 variable transformer. Atmospheric carbon dioxide was then passed through a Parkinson Cowan wet test meter to measure the actual solvent flow rate. The CO2 ports were designated as inlet/outlet ports since they were used as such depending upon whether the particular extraction was counter-current or co-current. The phenol/water solution was stored in a stainless steel Griffin beaker. Solution was gravity fed from the beaker into a Milton Roy Model HDB-1-30R pump capable of delivering a maximum of two milliliters per minute at 3000 psi. The flow rate was measured by timing the removal of solution from a burette and valve apparatus as shown in Figure 2. The solution was then pumped through 1/4 inch stainless steel tubing past a check valve, a rupture disc, two pressure indicators, and into the extraction 23 column. The pressure was measured by a Model DPS 0081 electronic pressure sensor made by Autoclave Engineers and by a 5000 psi gage used as a cross-check. After pressure ' measurement the solution lines entered an oven containing the extraction column. The 1/4 inch lines were then reduced to 1/8 inch before actually entering" the column through a side port located near the top of the cylindrical portion of the column. The water entry port was located approximately one inch below the COg inlet/outlet port. Water and COg exiting the column were originally reduced to atmospheric pressure using the same type of back pressure regulator. Both back pressure regulators were connected to the same nitrogen pressure system, as shown in Figure 2, so that the entire column and the fluids flowing through it would sense the same conditions. However, the slight differences in the regulators were enough to prohibit the use of this arrangement. The column length between the water inlet and the water outlet, determined by the regulator position, was limited.to approximately 20 inches, or a pressure head of less than one psi. At system pressures between 1100 and 3000 psi the back pressure regulators just could not provide the necessary control. The extraction equipment was then modified, as shown in Figure 3, so that water exiting the column would enter the bottom of a bomb that was pressurized by a "vapor" line 24 Pressure equalization vessel <■ Carbon dioxide at atmospheric — > pressure Water/phenol mixture Carbon dioxide Vv Extraction column (in heated enclosure) Pressure lock sampling system BPR - Back pressure regulator Figure 3. Schematic of Workable Counter-current Extraction Apparatus. from the top of the column. Extracted water accumulated in the bomb until the run was considered complete. A pressure lock system of valves allowed the removal of approximately four milliliters of extracted water for sampling purposes. A final major equipment modification was made to enable co-current extraction runs. A schematic of this arrangement is shown in Figure 4. Here, the pressure equalization vessel doubles as a phase separator and vessel to store 25 Discharge ^ carbon dioxide Carbon dioxide Water/phenol mixture Phase Extraction vessel separation (in heated enclosure) vessel Pressure lock sampling system BPR - Back pressure regulator Figure 4. Schematic of Co-current Extraction Equipment. extracted solution until the end of the run. Valves at the bottoms of the settler vessel and extraction column were added to enable each system to be isolated while lines were being purged and sampled. The extraction column itself was totally enclosed in a steel box which was fitted with two 500 watt electrical heaters manually controlled by a Powerstat Type 3PN1168 26 variable transformer. Oven temperature was measured by a thermocouple and monitored by the Cole-Palmer indicator. The column was 12 inches long with an inside diameter of 0.688 inches and outside diameter of one inch. The top end of the column was fitted with a custom-machined steel fitting having three ports. The top fitting had three ports, a carbon dioxide inlet/outlet port, a water solution inlet port, and an unused port. The bottom of the column was fitted with an Autoclave Engineers tee having two ports, one for the carbon dioxide inlet/outlet port, and the other for the water outlet. All four inlet and outlet ports were reduced to 1/4 inch NPT openings. Three of these openings were reduced further to 1/8 inch Swagelok fittings to facilitate tubing connections in the close quarters available in the oven. The fourth opening, located at the lowest point, was used for the extracted water solution outlet. This port provided a straight line path for the exiting water solution and was left as large as possible to reduce any flow restrictions. Three types of column packing were used in separate trials. These included 1/4 inch glass Raschig rings, 1/16 inch glass helices, and 1/16 inch steel helices., Support screens were cut to fit the column ends to prevent packing from entering any of the inlet/outlet ports. 27 Analytical Equipment Phenol concentrations were measured using a Varian 1800 Series Gas Chromatograph with a flame ionization detector. A new custom packed glass column, specifically designed for phenol and phenol derivatives analyses, was obtained and preconditioned according to the manufacturer's specifications. The column packing consisted of KG-02 on Uniport HP, with 80/100 mesh. Column length was ten feet with a two millimeter inner diameter and 0.125 inch outer diameter. All the phenol analyses, for this project used this single column at an oven temperature of 175°C, with a helium carrier gas and a hydrogen flame. While this research did not involve the removal of ammonia or chemical oxygen demand, it was anticipated that future work would. Equipment was therefore organized for the analysis of these two parameters. An Orion 901 Ion- alyzer with an ammonia ion-specific electrode was prepared to measure ammonia concentrations. The complete procedure and specific reagents used in the ammonia analyses are presented in Appendix A. Chemical oxygen demand standards were measured using a large hot plate, a thermometer reading at least 200°C, 500 milliter Erlenmeyer flasks, and reagent grade chemicals including mercuric sulfate, ferroin indicator, ferrous ammonium sulfate, concentrated sulfuric acid, and rapid dichromate acid solution. The special reagents and procedures used in this method are described in Appendix A. Experimental Procedures Phase Equilibrium Run Procedures A known volume, usually 100 milliters, of a phenol/ water standard was poured into the Parr bomb. The standard consisted of a known concentration of phenol, either 625, 1250, 2500, 5000, or 10,000 mg/1. Carbon dioxide was then compressed into the bomb to a pressure of about 1000 psi for the low pressure tests and to about 1500 psi for the high pressure tests. The bomb was then placed in the . rocker (but not rocked) and heated to 40°C to obtain a pressure reading. Carbon dioxide was added or vented as needed to attain the desired system pressure. The rocker was then set in motion and the heat adjusted so that the system would stay at 40°C for at least 10 minutes. After the system had stabilized the rocker was stopped and the bomb placed in an inverted position while the temperature was maintained. ' The liquid phase and the supercritical fluid phase were allowed to separate for about five minutes and a small amount of fluid was allowed to be vented. Immediately after the initial venting a three milliliter liquid sample was taken for analysis. 28 29 Extraction Run Preparations In preparation for each run the high pressure carbon dioxide cylinder was checked to see if there was sufficient solvent pressure for the particular run and if the bottle temperature was approximately 40°C. Powerstats were set to approximate settings to maintain the carbon dioxide temperature across both pressure drops and to maintain the oven temperature at 40°C. Experience was the best guide as to what settings would be required. If the high pressure cylinder needed to be recharged the Haskel compressor intake was connected to a liquid CO2 cylinder at room temperature, the discharge was connected to a port located on the front of the panel, the appropriate valves were opened and closed, and the compressor was started. When the high pressure cylinder was charged to about 4500 psi the valves, were reset and the compressor was disconnected. Also, before a run was begun, the back pressure regulator system was set to the desired system pressure and a check made for an essentially perfect seal. A nitrogen leak during a run of any reasonable length would cause the system pressure to change and ruin an otherwise good run. Once the original equipment was in place the nitrogen system was very reliable. Occasionally the high pressure 30 nitrogen bomb would also need to be recharged due to manual adjustments in the pressure settings. The Haskel compressor intake was then connected to a nitrogen cylinder and the discharge to a special nitrogen port on the front of the panel. Each back pressure regulator was isolated from the bomb, and the bomb was then charged with up to 5000 psi of nitrogen. The compressor was then disconnected and the back pressure regulator system was charged with whatever pressure was.desired. The phenol/water solution was checked before and during each run to ensure that the pump would not run dry and .to provide a continuous feed to the column. All but the last five runs were charged using a 10,000 mg/1 phenol standard prepared in- the lab. The last five runs were charged with actual coal gasification wastewater containing approximately 5000 mg/1 phenol, a variety of cresols, ammonia, and other organics. The wet test meter was checked before each run for the proper water level and proper gas seal. The carbon dioxide discharge line was vented into a 500 milliliter flask to allow any solids or liquids to drop out before entering the wet test meter. The flask was checked for a proper seal. 31 Extraction Run Procedures Although there were several different types of equipment arrangements, most of the experimental procedures were essentially the same. The principal differences involved the collection of extracted solution samples. At the start of each run the Tescom back pressure regulator was adjusted to approximately 50 to 100 psi higher than the system pressure. ■ The carbon dioxide feed control valves were then used to fine-tune the solvent flow rate for each particular run. This flow rate was one of the principle changeable parameters for each run and, as it turned out, was one of the biggest sources of problems, in producing reliable extraction data. Once solvent flow was reasonably established the water solution pump was started and the flow rate checked and adjusted if necessary. Usually the ■ solvent flow would then need to be readjusted to fine-tune the balance between solution and solvent flow. After both solvent and solution flows were established at reasonally constant rates the time and the wet test meter reading were recorded. Periodic flow checks were made to ensure that the flows remained constant. Adjustments were made as necessary to keep the system as close to a steady-state as possible. Almost always the only adjustment necessary was the adjustment to the carbon 32 dioxide flow. This was due to the constantly varying pressure on the high pressure side of the Tescom back pressure regulator, which changes the cooling effect across the regulator and the subsequent heat requirement of the manually controlled heat tape. Slight changes in original carbon dioxide pressure required periodic control valve adjustments as the run progressed. Once the run was considered to have reached steady- state (approximately two to six hours, depending on carbon dioxide mass flow rate) the solvent and solution flows were stopped and the extracted solution was sampled. This allowed a sufficient amount of solution to accumulate in the bottom of the collection bomb to enable a representa­ tive sample to be taken. Valves located at the bottom of the column and at the bottom of the collection bomb were closed immediately after the flows were stopped. The system was then allowed to rest, undisturbed, for approximately five minutes to allow any phenol-laden (pregnant) carbon dioxide that was undissolved in the water phase to float to the surface of the water phase. The sample line between the two valves was then drained via the lock valves. The valve below the collection bomb was then reopened to allow the liquid in the bottom of the bomb to flow into the sampling line. Liquid samples were then removed, approximately 3.5 milliliters at a time, and collected in labeled sample vials 33 Originally, each vial was analyzed to get some idea of whether the runs were indeed long enough and to check if the concentrations of phenol changed significantly with each sample. It appeared that the first flush, which included both solution and solvent, was not representative. Actually, that result was expected for co-current flow conditions since any phenol dissolved by the carbon dioxide would be redeposited when the combined mixture was reduced to atmospheric conditions. Also, the first sample after reopening the bomb bottom valve was questionable since carbon dioxide may have been trapped in some of the tubing immediately below the bomb. This sample analysis was normally discarded. Once several runs had been analyzed it appeared that the first line flush should be discarded along with at least the first three to four milliters of solution in the bomb. Then, approximately 10 milliters of extracted solution would be collected in a graduated cylinder and about three to four milliters transferred to a labeled vial for later analysis. These samples were considered representative of the steady state conditions during each run. The last bit of liquid to be sampled from the bomb was generally accompanied by a considerable amount of carbon dioxide and was considered to be non-representative of ( 34 steady-state conditions. This sample, was discarded, and generally the analysis of the one previous sample was considered questionable if it varied significantly from the average of the earlier samples. Sample Analyses Phenol concentrations were measured using a Varian 1800 Series Gas Chromatograph (GC) with a flame ionization detector. All the phenol analyses for this project used the packed glass column, specifically designed for phenol and phenol derivatives in a polar solvent (water). The helium flow rates were varied as needed to maintain reasonable peak sharpness, peak separation,and analysis time. Normal helium flow rates corresponded to regulated helium pressures of 25 to 40 psig supplied to the column. Air and hydrogen were adjusted to approximately a 10:1 flow ratio sufficient to maintain a stable flame. The. attenuation and range were adjusted to attain a peak height suitable for the particular phenol concentration range in question, ideally a peak height approximately 80% of full scale. A prepared phenol standard was then injected to determine a basis for unknown sample measurement. Typically, a sample size of approximately one microliter was used and enough repetitions were made of standards and samples to account for slight variations in sample sizes., 35 A concerted effort was made to use the same sample size for both standard and unknowns to eliminate any unnecessary measurement errors due to detector sensitivity, column overloading, or other parameters. Since even the highest phenol concentrations were only one percent (10,000 mg/1) phenol, very sensitive settings were required. Even so, the peaks were not ideal, often having wide bands with trailing tails. In order to be consistent the areas under the top half of each peak were measured. Unknown sample concentrations were calculated based on a directly proportional comparison of the average area of the standards. For example, an unknown sample peak with an area of 1.2 times the area of the standard's peak would have a concentration 1.2 times that of the standard, . assuming the sample sizes were the same and the attenuation, range, and all other parameters were constant. Helium flow rates were kept constant, as much as practical, during any particular analytical run, and only changed to adapt to the measurement of new phenol standards. A few problems were encountered due to sample sizes. Excessive sample sizes, generally greater than two microliters, sometimes caused wide peaks with excessively large tails, off-scale problems, and carrier gas pressure (flow) swings. Too small sample sizes created problems with getting consistent sample sizes, often resulting in 36 peaks with widely varying heights and areas. Also, if the unknown sample's concentration was significantly lower than the standard's concentration, a short, wide peak was produced and a reliable comparison was not possible. The operating parameters of the GC,. the sample sizes, and the phenol standard would then be.varied until a suitable combination was found, and analysis would be continued. All samples, both extraction run and equilibrium phase, were allowed to cool to approximately room temperature principally to allow carbon dioxide bubbles to separate from the liquid but also to allow the liquid density of each analyzed sub-sample to be virtually identical. Sub-samples were drawn from the sample vial with a 10 microliter syringe and injected into the GC injection port. Every effort was made to keep the detector clean and to keep the hydrogen, helium, and air flows as constant as possible.' Extreme care was taken when analyzing the equilibrium phase samples. Approximately, five GC analyses of each sample were averaged and two standards were run before and after each sample group. Standards were chosen so that phenol concentrations in both the standards and the samples would be nearly the same and would register between 50% and 95% of full scale on.the recorder chart. . Phenol concentations in the carbon dioxide- rich phase were calculated based on a mass balance of the 37 amount of phenol left in the water phase (measured), the original amount of phenol in the water, and a calculated amount Of carbon dioxide placed into the bomb. The actual concentration of phenol in the supercritical phase was not directly measured. Normally, duplicates of extraction run samples were analyzed and an average of the two was considered to be sufficiently accurate. For some of the later runs, when it was clear that several samples from the same run were all representative of the extraction run, individual analyses were grouped together and averaged for a nominal value. This helped reduce the excessive amount of time required for data acquisition and interpretation. 38 EXPERIMENTAL RESULTS AND DATA ANALYSIS Introduction Altogether, fifty-three extraction runs were attempted including forty-eight runs using 10,000 mg/1 phenol/water standard and five runs using actual coal gasification wastewater. Of these runs, twenty-three were successfully completed, one counter-current and the remainder co-current In addition there were seventeen successful phase equili­ brium runs at pressures ranging from 1100 to. 2 80 0. psi and initial concentrations ranging from 625 to 10,000 mg/1 phenol. / The initial intent of this research was to test the feasibility of using counter-current extraction techniques with the two phase, three component system of phenol, water and supercritical carbon dioxide. The extraction equipment was designed and built with the expectation that a number of different parameters could be tested to determine some optimum operating conditions. These parameters included pressure, relative flow rates of solvent and liquid solution, temperature, and column packing. Unfortunately, only one counter-current extraction run was successfully completed after twenty unsuccessful tries. Nine more 39 counter-current flow attempts failed, before it was decided to try co-current extraction tests. The most frequent cause of run failure was the inability to Control the carbon dioxide flow within the required degree of.accuracy. Reliability of Data Throughout this research there are two general categories, one relating to extraction runs and one relating to phase equilibrium data. A discussion of data reliability must necessarily separate the two, since there is a distinct difference in reliability. Originally, the phase equilibrium data was gathered to get some idea of the feasibility of extracting phenol from water using supercritical carbon dioxide. Very little information was available in the literature to indicate what would happen with this system or at what conditions an extraction might be optimal. The liquid- liquid-solid ternary phase diagram published by Francis (39), and enlarged in Figure 5, shows that the region of interest, for this research, lies entirely within one percent phenol of the water-carbon dioxide baseline. An attempt was made to experimentally develop a set of equilibrium data for this ternary system at the dilute conditions in question. 40 Carbon Dioxide One liquid phase Diagram taken from Francis (39), representing data at approximately 26 degrees Celsius and 65.1 atm. Solid, \ \ Liquid, \ Liquid Liquid, Liquid, Liquid Liquid,' \ Liquid One liquid phaseLiquid, LiquidOne liquid phase PhenolWater Figure 5. Phenol/Water/Carbon Dioxide Ternary Phase Diagram. 41 Since the. solubility of phenol in supercritical carbon dioxide is a strong function of density and the facilities to analyze high pressure carbon dioxide directly were not available, it was necessary to use the mass balance approach in measuring the phenol concentration in the supercritical fluid (SCF) phase. The actual mass of carbon dioxide can be determined by knowing the volume available to the SCF. phase and by knowing the density of carbon dioxide at 40°C and the test pressure. Density data used for these calculations were taken from the temperature-entropy diagram in Perry and Chilton (24). A plot of density versus pressure is shown in Figure 6. Water samples taken from each equilibrium data experiment were carefully analyzed by gas chromatograph to determine the concentration of phenol in the water rich phase. Using the published experimental density data and the GC phenol concentrations, a phenol concentration was calculated for the SCF phase and expressed as moles phenol per moles carbon dioxide. A number of possibilities existed for the introduction of error in this procedure. Major concerns included the following: I. GC concentrations were considered to be accurate to within 10% of the actual value. Any error in analysis was compounded by the use of mass balance for determining the mass of phenol in the carbon dioxide phase. Ca r bo n Di ox id e D en s it y (g /c c) 42 Data taken from Perry and Chilton (24) 300020001500 Pressure (psig) Figure 6. Experimental Pressure and Density Diagram for Carbon Dioxide at 40 Degrees Celsius. 43 2. The carbon dioxide density, and therefore its mass, was determined by interpolation of the data available from the literature. 3. Since the data was gathered near the critical temperature of carbon dioxide, slight changes in temperature would result in relatively large changes in density. The large amount of mass involved in the test equipment provided a considerable amount of lag time in heat input and temperature change and the temperature measuring apparatus provided a digital readout accurate only to the nearest degree Celsius. Therefore it is difficult to determine the actual margin of error due to measurements that were slightly higher or lower than 40 degrees Celsius. 4. Although a rocker was used to mix the phenol ■solution and the carbon dioxide, there was no definite way of knowing whether an equilibrium had been reached. As discussed below, residence time was a critical parameter in the extraction runs and could possibly have been a significant factor in the accuracy of the equilibrium data. The principal concern with the phase equilibrium data is the dependence of an uncertain mass balance to calculate the phenol concentration in the SCF phase. It would be far better to be able to actually measure the concentration in both phases.. The equilibrium data does, however, provide a general idea of what to expect during co-current extraction runs, where the approach to equilibrium should be the upper limit. For counter-current extraction the availability of equilibrium data allows the calculation of the number of theoretical trays in a column and the height of an equiva­ lent theoretical plate for a particular set of parameters. 44 Equilibrium data obtained from this research was used to roughly analyze both the co-current and the counter- current extractions. Several kinds of data were obtained from the extrac­ tion experiments. These included solution and solvent flow ratesf temperatures, pressures, and phenol concentrations determined by GC analysis. Solution flow rates were measured very accurately, to within 0.5% of the actual flow, and were subject to only slight variation during any of the valid runs. Solvent flows, on the other hand, were much harder to control, and varied during some runs by as much as 20% above and below the average flow. Since the carbon dioxide flow rates were measured by a wet test meter, which accumulated the total volume of CC^ during a timed run, an average flow rate was used for analyzing the extraction data. Oven (and column) temperatures were manually controlled to within one degree Celsius of the desired 4O0C. System pressure was generally kept within 25 psi of the run's nominal pressure. Extraction runs involving very high rates of carbon dioxide flow usually generated problems with pressure drifting due to excessive cooling near the back pressure regulator. On the last run this problem was virtually.eliminated by rearranging the electrical heat tape to more directly heat 45 the back pressure regulator itself. Much higher flow rates were then made possible without causing icing problems. Phenol/water standards were prepared in the lab and were accurate to within 0.02% of the nominal 10,000 mg/1. This is much more accurate than the estimated accuracy of any one GC analysis, considered to be within 10% of the actual concentration. However, known standards were analyzed before and after any unknowns, and unknown samples' were usually analyzed at least twice to minimize any analytical error. By using the GC peak area averages of the standards for reference, and the peak area averages of unknowns for the final concentration, it is believed that the actual concentrations measured and reported in this paper are accurate to within 10% of the true values. The final five extraction runs, using actual coal gasification wastewater, were analyzed on'a relative basis. Since there were other phenol derivatives present in this solution, there was some interferences present in the GC charts involving a widening of the peak and secondary peaks riding the main phenol peak's tail. Unknowns, that is, the extracted solution, were analyzed at the same GC setting as the original solution, and the results were expressed as a percentage of the original. Due to the wide differences in peak heights and widths, the accuracy of this data is probably on the order of plus or minus 20% of the actual 46 minus 20% of the actual measurement. In all extraction runs, if any data were considered to be out of the above ranges, the run was considered unusable and the data were not used. Phase Equilibria Raw phase equilibrium data are shown in Table 5. These data were then converted to convenient units for comparing concentrations, that is, moles phenol per mole of solvent. Table 5. Raw Phase Equilibrium Data Phenol/Water/Carbon Dioxide for the Ternary System. Pressure Phenol Concentrations.(mg/1) Run (psig) Initial Final I 1,210 625 597 2 1,205 1,250 1,180 3 1,205 2,500 2,380 4 1,205 5,000 4,460 5 1,180 10,000 8,500 6 2,330 625 345 7 2,400 1,250 735 8 2,400 2,500 1,320 9 2,355 5,000 2,380 10 2,400 10,000 4,850 11 2,650 1,250 1,135 12 1,880 2,500 1,724 13 1,570 5,000 3,611 14 2,850 5,0 00 4,154 15 1,100 10,000 9,500 16 1,600 10,000 8,440 17 1,600 10,000 5,940 47 A summary of these data and the respective distribution coefficientsf. or equilibrium K-values are shown in Table 6. Table 6. Mole Fractions of Phenol in Water (x), Mole Fractions of Phenol in Carbon Dioxide (y), and Equilibrium K-values. Run Density (g/cc) Mole X Fractions y K-values y/x I .345 .000114 .00000913 .0799 . 2 .340 .000226 .0000235 .104 3 .335 .000456 .0000403 .0884 4 .335 .000854 .000181 .212 5 • .300 .00163 .000581 .357 6 .794 .0000661 .0000402 .608 7 .803 .000141 .0000732 .520 8 .805 .000253 .000167 .662 9 .797 .000456 .000375 .822 10 .803 .000929 .000732 .789 11 .817 .000217 .0000275 .1,26 13 .664 .000692 .000239 .345 14 .830 .000795 .000458 .575 15 .222 .00182 .000849 .469 16 .670 .00162 .00105 .647 17 .670 .00114 .000690 .607 Figures 7 and 8 show three-dimensional plots of that data, viewed from 30 degrees above the pressure-composition plane and rotated horizontally 45 degrees and 135 degrees, respectively. These plots were generated using the SYMAP and SYMVU library routines on the MSU CP6 computer and drawn on a CALCOMP plotter. All seventeen data points were used to generate each plot, showing a general tendency for Figure 7. Three Dimensional Pressure-composition Plot of Phenol/ Water/Carbon Dioxide System at 40 Degrees Celsius, Rotated 45 Degrees. Mole fraction of phenol in carbon dioxide) 3000 2000 1000 Figure 8. Three Dimensional Pressure-composition Plot of Phenol/Water/Carbon Dioxide System at 40 Degrees Celsius, Rotated 135 Degrees. v: x HOO' 1 0 4 4 8 0 904 . 82 754 . 02 603-22 452 41 301 61 I 50 -80 0 00 Mole fraction of phenol in carbon dioxide) 50 a large increase in y (moles phenol per mole supercritical carbon dioxide) with increasing x (moles phenol per mole water) and a similar tendency for increasing y with increasing pressure. Since the solubility of phenol in supercritical carbon dioxide is mainly a function of density it is expected that the increase in y, at a constant x should level off at very high pressures, with most of the increase taking place between 1000 and 2000 psi. The unexpected waviness of the 3-D surface is most likely an indication of inherent experimental error. Figure 9. shows the equilibrium data measured near 1200 and 2400 psi. The correlation for each curve was determined using the quadratic form of the Method of Least Squares, as described by Neter and Waterman (40), are: At 1200 psi, y = 2.44*10-5 - 0.0883x + 285.6x2 At 2400 psi, y = -3.73*10-5 + 0'.9007x - 74.72x2 These equations were then used to calculate values for y at even intervals of x at 1200 and 2400 psi, corresponding to carbon dioxide densities of 0.33 and 0.803 g/cc, respectively. Use of the equations for values of x above 0.0016 was avoided because of the apparent discrepancy in the shapes of the two curves in that region, and also because it served no useful purpose in -( Mo le Fr ac ti on of P he no l in C ar bo n Di ox id e) 51 For 6 experimental points near 1200 psi y = 2.44 X IO-5- 0.0883x + 285.6x2 0.0012 I For 5 experimental points/near 2400 psi -3.73 X 10 + 0.9007x - 74.72x Range of extraction run experiments0.0010 - 0.0008 - 0.0006 - 0.0004 - 0.0002 - 0.0005 0.0010 0.0015 0.0020 x ( Mole Fraction of Phenol in Water) Figure 9. Phenol/Water/Carbon Dioxide Equilibria Measured Near 1200 and 2400 psi, at 40 Degrees Celsius. 52 later analyses of the extraction run data. The values are shown in Table 7 and plotted in Figure 10. The assumption of a linear curve on semi-log paper, between the densities of 0.33 (1200 psi) and 0.803 (2400 psi) was based on data obtained from Van Leer and Paulaitis (31). Their data, also plotted in Figure 10, shows the equilibrium concentrations of phenol in supercritical carbon dioxide at 360C above pure, solid phenol, resulting in a nearly linear curve for densities ranging from 0.84 to below 0.40 g/cc. Although the temperature is slightly lower than the 40°C used throughout this research., both temperatures are below the melting point of phenol, and the published data should give an excellent idea of the ultimate solubility of phenol in supercritical carbon dioxide. Table 7. Calculated Values Phenol in Carbon and 0.8 03. g/cc. of y (Mole Fraction of Dioxide) at 0.33 g/cc X Y at 0.33 g/cc y at 0.803 g/cc 0.0002 . 0.0000182 0.000140 0.0004 0.0000348 0.000311 0.0006 0.0000742 0.000476 0.0008 0.000137 0.000635 0.0010 0.000222 0.000789 0.0012 0.000330 0.000936 0.0014 0.000461 0.00108 0.0016 0.000614 0.00121 (M ol e Fr ac ti on of P he no l in C ar bo n Di ox id e) 53 <00 Pure Data taken from Van Leer and Paulaitis (31) .005- . 002- .001- .0005- . 0002- . 0001- .00005- .00002- .00001 Carbon Dioxide Density (g/cc) Figure 10. Calculated Values of x (Mole Fraction of Phenol in Water) and y (Mole Fraction of Phenol in Carbon Dioxide) for the Phenol/Water/Carbon Dioxide System, for Varying Densities, at 40 Degrees Celsius. 54 Points were taken from Figure 10 and reconstructed in Figure 11 to show the relationships of y versus x at varying densities. This family of curves is shown to get a perspective of the data from another dimension. Figure 12 shows a family of curves based on the calculated distribution coefficient, K, at varying values of carbon dioxide density and x. The curves show that K tends to converge near a value of 0.8 as the CC^ density approaches 0.8, at the experimental conditions of 40°C and 2400 psi. The curves shown in Figures 10, 11, and 12 should prove useful in predicting extraction capabilities for counter-current, co-current, and mixer-settler processes. Counter-current Extraction Equilibrium data gathered near 1200 psi provide an opportunity for analyzing the extraction data gathered in the counter-current run. The operating conditions and flow schematic are shown in Figure 13. By mass balance yg was calculated to be 0.000704 gmole phenol/gmole COg. The operating line obtained from this data and the equilibrium curve for the '1200 psig data are shown in Figure 14. The equilibrium curve was , determined using the six equilibrium data points nearest 1200 psi, including Runs numbered I, 2, 3, 4, 5, and 15. / (M ol e Fr ac ti on of Ph en ol in Ca rb on Di ox id e) 55 0.002 CO^ density = 0.803 g/cc Region consisting of pressures higher than 2400 psi. "0.001 0.0005 - 0.0002 0.0001 - CO9 density = 0.330 0.00005 - Region consisting of pressures lower than 1200 psi. Both carbon dioxide densities and y are extremely non-linear in this region.0.00002 - 0.00001 0.0004 0.0008 0.0012 0.0016 x (Mole Fraction of Phenol in Water) Figure 11. Graphically Determined Values of Carbon Dioxide Phase Density and Composition for the Phenol/Water/Carbon Dioxide System, for Varying Values of x (Mole Fraction of Phenol in Water), at 40 Degrees Celsius. Ca lc ul at ed Di st ri bu ti on Co ef fi ci en t Mole fraction of phenol in carbon dioxide Mole fraction of phenol in water y/x 0.4 - x=0.0002 0.2 - Carbon Dioxide Density (g/cc) Figure 12. Calculated Distribution Coefficients for Varying Values of Carbon Dioxide Density, at 40 Degrees Celsius. 57 A y2 = 0.000704 x2 = 0.00191 T = 40° C P = 1130 psig G = 13.38 gmole/hr L = 7.83 gmole/hr V 1 = o.o X1 = 0.000709 > f X1 - Mole fraction of phenol in water leaving the extraction column (mole/mole) x2 - Mole fraction of phenol in water entering the extraction column (mole/mole) Y1 - Mole fraction of phenol in carbon dioxide entering the extraction column (mole/mole) y2 - Mole fraction of phenol in carbon dioxide leaving the extraction column (mole/mole) T - System temperature (degrees Celsius) P - System pressure (psig) G - Molar flow rate of carbon dioxide (gmole/hr) L - Molar flow rate of water (gmole/hr) Figure 13. Flow Schematic for Counter-current Extraction Run. (M ol e Fr ac ti on of Ph en ol in Ca rb on Di ox id e) 58 0.0010 y* - Gas-phase mole fraction corresponding to equilibrium with the bulk liquid composition. 0.0008 - T 0.0006 0.0004 - 0.0002 - 40° C 1130 psig Equilibrium curve y/ (Approximate)— Corresponding y.* 0.0005 0.0010 0.0015 0.0020 x (Mole Fraction of Phenol in Water) Figure 14. Working Curves for Obtaining Gas-phase Composition Values. 59 Since the equilibrium curve is nearly linear, one of the simplest means of calculating the number of overall gas phase transfer units, Nq ,^ is to use a graphical integration technique. The rationale for using overall transfer numbers, arid their derivation is explained in detail by Sherwood, et al.’(41), and is summarized below. Since such column characteristics as local gas and liquid phase mass transfer coefficients, fluid dynamics, and wetted interfacial areas are generally not known, and specifically unknown in this case, the analysis of the column is better accomplished by evaluating overall characteristics. By using hypothetical interfacial, compositions the overall gas-phase coefficient can be related to the mass flux, the pressure, (P) and the difference betweeri y and y* : . N = K P(y*-y), where y is a point on the operating line, and y* is the gas-phase mole fraction corresponding to equilibuium with the bulk liquid composition, x, as shown in Figure 14. The mass flux is N, and K is the overall gas-phase coefficient. In their analysis, Sherwood, et al. (41) also show that the differential section of a column is determined by: G dy dh = - -------- . , N a(l-y) where G is the molar flow rate of the gas phase, and the 60 constant, a, is the effective mass-transfer surface area . per unit of packed volume. By combining the two equations: h *2 K yI G dy P a(l-y) (y'*-y) r where h is the total height of a column. Multiplying and dividing by Y*BM gives: y*BM ay (1-y)(y*-y) 9 where y*BM (l-y)-(l-y*) ln[(l-y)/(l-y*) ] The left-hand term under the integral defines the overall gas-phase height of a transfer unit, Hq ,^ and the right-hand term defines the number of overall gas-phase transfer units, : y*BMdy (1-y)(y*-y) If an average value of H may be used then H3 og og can be calculated from the known column height and the evaluated . Also, if two runs had been possible the product, K a, could then be calculated by simultaneous solution. 61 The integral representing. Nq^ can be evaluated graphically by plotting BM versus (1-y)(y*-y) and calculating the area under the curve. For dilute solutions, as in this case, Y*gM is approximately equal to 1.0 and it will be treated as such here. The values of y and y* are obtained from the enlarged graph of the operating line and equilibrium curve shown in Figure 14. Table 8 shows the data obtained from Figure 14, and plotted in Figure 15. The area under the curve in Table 8. Values for Gas-phase Calculating the Number of Overall Transfer Units. Y (1-y) Y* (Y * —Y) 1/t(1-y)(y*-y)] 0.00000 1.00000 .000120 .000120 8,333 0.00005 0.99995 .000155 .000105 9,524 0.00010 0.99990 .000198 .000098 10,205 0.00015 0.99985 .000238 .000088 11,365 0.00020 0.99980 .000282 .000082 12,198 0.00025 0.99975 .000330 .0000080 12,503 0.00030 0.99970 .000378 .000078 12,824 0.00035 ' 0.99965 .000430 .000080 12,504 0.00040 0.99960 .00048.2 .000082 12,200 0.00045 0.99955 .000540 .000090 11,116 0.00050 0.99950 .000602 .000102 9,809 0.00055 0.99945 .000665 .000115 8,700 0.00060 0.99940 .000736 .000136 7,357 0.00065 0.99935 .000805 .000155 6,456 0.000704 0.999296 .000890 .000186 5,380 In te gr at io n fa ct or 62 14,000 12,000 - 10,000 - 8,000 - Area under curve = N Number of overall gas-phase transfer units 6,000 - 7.25 * - Gas-phase mole fraction corresponding to equilibrium with the bulk liquid composition. 4,000- 2,000 - ln[(1-y)/(1-y*)] 0.0004 0.0006 0.00080.0002 y (Mole fraction of phenol in carbon dioxide) Figure 15. Graphical Integration of the Number of Overall Gas-phase Transfer Units. 63 Figure 15 is N and is calculated to be 7.25. This°g indicates that the column contributes the equivalent of 7.25 equilibrium stages when operated at these specific, conditions where the equilibrium line is approximately linear. Since the column height is 12 inches, the overall height of a gas-phase transfer unit is 1.66 inches. This indicates that the column provides excellent exchange capabilities when operating under, counter-current conditions. Phenol/Water Co-current Extraction The twenty-two successful co-current extraction runs, including both phenol/water standards and wastewater, formed the bulk of the extraction data gathered during this research. These data were reduced to a common form, shown in Table 9. The seventeen extraction experiments using phenol/water standards were run to determine the effects of column packing, residence time, and system pressure on the extraction efficiency. The extraction efficiency, in this case, is defined as one hundred times the ratio of the distribution coefficient measured at the extraction conditions, K^, divided by Kt, the distribution coefficient as determined by the equilibrium experiments. The extraction run values of y (the mole fraction of phenol in the carbon dioxide phase) were calculated using a mass Tablei 9. Summary of Co-current Extraction Run Data Run Average Average Net C00/H0(D Total Final Percent No. Pressure Specific Flow Cone. of (psig) Gravity g/g mol/mol .ml/ml ml/min mg/1 Original PhenOl/Water Standard with 1/4 inch Raschig Rings I 1935 0.740 4.65 1.90 6.28 4.08 4370 43.7 2 2425 0.806 1.75 0.71 2.17 1.69 5700 57.0 3 2380 0.800 5.08 2.08 6.34 3.89 3850 38.5 4 2357 0.797 11.8 4.83 14.8 7.97 2420 24.2 5 2825 0.831 2.65 1.08 3.14 1.83 4170 42.9 6 2800 0.828 6.07 2.49 7.22 3.63 3600 36.0 Phenol/Water Standard with 1/16 inch Glass Helices I 1497 0.642 4.63 1.89 7.21 4.73 45 50 45.5 8 1517 0.647 2.9 2 1.19 4.51 3.11 5390 53.9 9 1810 0.720 2.39 0.98 3.3 2 2.40 5245 52.5 10 1825 0.722 5.80 . 2.37 8.03 4.92 4240 42.4 . 11 • 2008 0.752 3.26 1.34 4.34 2.86 4840 48.4 12 2010 0.752 6.84 2.80 9.10 5.43 3905 39.1 13 2410 0.804 3.47 1.42 4.32 2.78 4720 47.2 14 2400 0.803 6.80 2.78 8.47 4.95 2950 29.5 15 2375 0.800 14.0 5.72 17.5 9.63 2250 22.5 16 2816 0.830 3.02 1.24 3.58 2.30 4340 43.4 17 2815 0.830 8.88 3.63 10.6 5.69 2520 25.2 Coal Gasification Wastewater With 1/16 Inch Glass Helices 18 2390 0.800 7.63 3.12 9.54 5.37 1200 24.0 19 2400 0.800 6.89 2.82 8.61 4.97 1410 28.2 20 2450 0.802 8.19 3.35 10.2 5.27 1170 23.4 21 2470 0.804 2.89 1.18 3.59 . 2.26 2005 40.1 22 2425 0.801 21.13 8.65 .26.4 14.3 980 19.6 65 balance based on a known x (the mole fraction of phenol in the water phase) and the relative net flow rates of carbon dioxide and phenol/water solution. An experimental value for K (K= y/x) was then calculated for each run. To determine the theoretical K value it was necessary to use a trial and error approach based on an initial guess of an x value. A value for y was then calculated, again based on a mass balance. Then, by referring to Figure 10, the theoretical value of y was determined for the guessed value of x. A new value for x was then estimated, if necessary, and the process repeated until the calculated and plotted value of y were essentially equal. Respective K values were then calculated for each extraction run. Theoretical and actual x, y data and K values are listed in Table 10, and examples of the calculation procedures are shown in Appendix B. Upon analyzing the data it became apparent that some of the calculated K values, obtained during runs with relatively long residence times, were actually higher than the theoretical K values. Normally, the highest possible extraction efficiency for a co-current flow column would be 100%, when the two phases would be exactly in equilibrium. This indicated that there was probably some significant degree of error involved in the equilibrium experiments resulting in low K values. It was felt, however, that the Table 10. Converted but as the phenol concentration decreases in the exiting water the equilibrium concentration in the carbon dioxide phase also decreases. This is an expected result based on the equilibrium experiments. All 17 of the phenol/water extraction data points are shown in Figure 20, with the two types of packing identi­ fied. A plotted curve represents the Method of Least Squares equation of best fit for the entire set of points, and has an r-value of 0.77. Individual equations of best fit were calculated for the helices and the Raschig Rings. These equations are listed and plotted in Figure 20. Again, the type of packing appears to have little, if any, influence on the extraction efficiency at these . Ph en ol Re mo va l (m g Ph en ol /g Ca rb on Di ox id e) 74 2 . OOi (Run #16) 1.75- Relative mass flow rate of approximately 3.0 g CO9Zg water. (Run #8) 1.50- 1.25- (Run #14) 1 . 00- (Run #12) Relative mass flow rate of approximately 6.8 g CO9Zg water. 0.75- 0.50 1500 1750 2000 2250 Pressure (psig) • 2500 Figure 19. Phenol Extraction Capabilities of Carbon Dioxide with Respect to Pressure, at 40 Degrees Celsius. 75 Al I points*: Er= 50.24 tR + 50.11 (□ )- Experiments using glass helices: 55.01 t ( O ) - Experiments using Raschig Rings: 43.31 Points represent all 17 experiments using phenol/water standards, at system pressures varying from 1500 to 2800 psig. 0 0.20 0.40 0.60 0.80 1.00 tR (Residence Time Relative to Run #2) Figure 20. Summary of Normalized Extraction Efficiencies and Residence Times for All Co-current Experiments Using Phenol/Water Standards. 76 experimental conditions. However, increasing the residence time allows the system to more closely approach equilibrium conditions. Equations of best fit were plotted for data obtained using both types of packing at 2400 and 2800 p.si. These curves and their respective r-values are plotted in Figure 21. In addition, an overall equation of best fit for all 10 points is listed and plotted. With an r-value of 0.88, the overall equation has almost as good a fit as the individual curves. It is possible that increasing the system pressure from 2400 to 2800 psi. has only marginal benefit since the SCF density increase is only about three percent. The expected increase in solvent power could be hidden by relatively small experimental error. Figure 22 shows all 17 data points, undifferentiated with respect to packing or pressure. The same data points are shown in Figure 23, differentiated by experimental system pressure. The two figures show the overall dependence of the extraction efficiency upon residence time. For both types of packing and at all of the experimental test pressures an increase in residence time increased the extraction efficiency. Figures 22 and 23 indicate that varying the system pressure or the column packing has an insignificant affect on the ability of the system to approach equilibrium, at these test conditions. 77 All four experiments run at 2800 psig: En= 67.34 t r = 0.92 Equation of ^-best fit for all 10 points: D Er= 47.66 tR + 47.36 r = 0.88 All six experiments run at 2400 psig: Er= 38.75 tR + 51.78 r = 0.90 A - 2400 □ - 2400 O - 2 800 O - 2800 0 0.20 0.40 0.60 0.80 1.00 tR (Residence Time Relative to Run #2) Figure 21. Equations of Best Fit for Normalized Extraction Efficiencies and Residence Times for Co-current Phenol/Water Extraction Experiments, at 2400 and 2800 psi, and 40 Degrees Celsius. (P er ce nt Ef fi ci en cy Re la ti ve to Ru n #5 ) 78 100 80 - 60 - 50.24 t + 50.7740 - 0.77 20 - 0.20 0.40 0.60 0.80 1.00 (Residence Time Relative to Run #2) Figure 22. Equation of Best Fit for Normalized Extraction Efficiencies and Residence Times for All Co-current Experiments Using Phenol/Water Standards, at 40 Degrees Celsius. (P er ce nt Ef fi ci en cy Re la ti ve to Ru n #5 ) 79 40 - 0 - 1500 psi, Glass Helices o - 1800 psi, Glass Helices A — 2000 psi, Glass Helices and Raschig Rings Q - 2400 psi, Glass Helices and Raschig Rings □ - 2800 psi, Glass Helices and Raschig Rings T ™ 0.2() ~Y~ 0.40 "I 0.60 ~T~ 0.80 tR (Residence Time Relative to Run #2) Figure 23. Plot Showing Values for Normalized Extraction Efficiencies and Residence Times for All Co-current Experiments Using Phenol/Water Standards, at 40 Degrees Celsius, Differentiated by Test Pressure. 80 Wastewater Co-current Extraction Data gathered in the extraction experiments using coal gasification wastewater were reduced and normalized in the same manner as the data relating to the phenol/water standards and are shown in Tables 9 and 10. Note that the wastewater extraction efficiencies are relative to the standards, and that one of the five runs has an efficiency' greater than 100%. Four of the five runs have efficiencies greater than or equal to 90% indicating that there are possibly some components in the wastewater that contribute to the ability of carbon dioxide to extract phenol. Figure 24 shows a plot of the five wastewater extraction effi­ ciency data points. Several types of equations were tried to determine the type of curve that would best describe the experimental data. A linearized version of a non-linear reciprocal equation, proposed by Heller (42), was chosen.to obtain a reasonably close fit of the points. This equation has the form: y = x/(a + bx). This reciprocal equation was linearized, and solved with the least squares program, by taking y equal to 1/y and x equal to 1/x and solving for the equation: (1/y) = b + (a/x). The resulting equation, having an r-Value, of 0.98, is ■ 81 5 80 o 40 tD/(0.001542 + 0.006068 t_) 0.98 ^ 20 0.800.20 0.60 (Residence Time Relative to Run #2) Figure 24. Equation of Best Fit for Normalized Extraction Efficiencies and Residence Times for Experiments Using Coal Gasification Wastewater, at 40 Degrees Celsius. 82 plotted in Figure 24, along with a listing of the actual equation. Other non-linear equation forms that were tried, but discarded, included a. simple quadratic (which yielded a concave downward parabola) and an exponential curve of the form, y = a(xD) (having an r-value of 0.92). Finally, Figure 25 shows all 22 co-current extraction data points, differentiated by points representing phenol/ water standards and wastewater experiments. The wastewater points show clearly higher extraction efficiencies, relative to the standards. (P er ce nt Ef fi ci en cy Re la ti ve to Ru n #5 ) 83 120 -i □ 40 - 20 □ - Coal Gasification Wastewater Experiments O - Phenol/Water Standard Experiments CdW 0 0.20 0.40 0.60 0.80 1.00 tR (Residence Time Relative to Run #2) Figure 25. Plot Showing Values for Normalized Extraction Efficiencies and Residence Times for All Co-current Experiments, at 40 Degrees Celsius, Differentiated by Phenol/Water Standards and Coal Gasification Wastewater. 84 SUMMARY The principal objective of this research was to investigate the feasibility of using supercritical carbon dioxide to extract phenols and phenol derivatives, from coal gasification wastewater. Five major steps were undertaken to achieve that goal: 1. Measurement of phase equilibria for the phenol/ water/ carbon dioxide ternary system. 2. Design, construction, and modification of the extraction equipment. 3. Measurement of column effectiveness using counter- current extraction techniques. 4. Measurement of extraction efficiencies using co-current extraction of phenol from phenol/water standards. 5. Measurement of extraction efficiencies using co-current extraction phenol and phenol derivatives from coal gasification wastewater. Seventeen phase equilibrium experiments were run to obtain the basic information needed to analyze the effectiveness of operating the extraction column at various flow rates and pressures. Three-dimensional plots of the data were generated to provide visual perspectives of the effects of varying system pressure and composition. The data show the expected general tendency for increased 85 solubility of phenol in carbon dioxide as the pressure is increased and as the concentration of phenol in the water phase is increased. The equilibrium data is unfortunately subject to a number of possible errors due to the inability to directly measure the concentration of phenol in the carbon dioxide phase. However, it is believed that the results are reasonably indicative of actual values and that these data can be used only for general interpretation of extraction efficiencies. The successful operation of the extraction .equipment was preceded by an extensive period of equipment design, construction, trial runs, and modifications. A number of temperature and pressure controls were needed to properly maintain the desired system parameters during the course of an extraction run. Water and carbon dioxide flow rates were measured as accurately as possible and all systems were designed and built to provide for operator safety at pressures up to 5000 psi. Major obstacles requiring equipment design modifications included carbon dioxide flow control and measurement, sample collection, extraction column hydraulics, and proper heating of the carbon dioxide/ phenol stream as it was depressurized to atmospheric pressure. 86 Approximately 30 attempts were made to counter- currently extract phenol from a phenol/water standard. Many of the equipment modifications mentioned above were required to achieve the single successful experiment at 1130' psig and 40°C . Counter-current runs attempted at higher pressures resulted in severe column flooding due to relatively poor carbon dioxide flow control and very narrow density differentials. Although limited in number, the one counter-current extraction experiment shows that supercritical carbon dioxide can be very effective in removing phenol from water, at easily obtained pressures, using commonly available industrial equipment. Seventeen successful co-current extraction experiments were run at pressures varying from 1500 to 2800 psi, using two types of packing. Each experiment's data set was analyzed to determine the approach to theoretical equilibrium, or the relative efficiency. The type of packing had little or no effect on column effectiveness. However, the amount of residence time available for mass transfer significantly affected relative efficiencies. Five co-current extraction experiments were run using coal gasification wastewater. Phenol extraction was measured as a percentage of the original phenol concen­ tration in the wastewater. Extraction efficiencies were calculated based on the approach to theoretical equilibrium 87 as determined by the phenol/water/carbon dioxide phase equilibrium data gathered as part of this research. Four of the five runs indicated that extraction efficiencies of 90% or greater had been achieved. One of those four was greater than 100%, indicating that there could be some component in the wastewater that enhances the solubility of phenol in supercritical carbon dioxide, or that there is additional experimental error due to unknown phenol concentrations in beginning and final water samples. 88 'RECOMMENDATIONS FOR FUTURE STUDY There is a great deal of potential for improving the steady-state counter-current extraction process. Additional effort should be made to provide a workable column that can operate dependably between 1000 and 1500 psi. This column should have automatic temperature control, precise feed volume control of both carbon dioxide and wastewater, proper feed distribution, and level control for exiting liquids. A packed column, approximately two to four inches in diameter and ten to twenty feet tall would provide a safe, relatively inexpensive means of investigating supercritical continuous counter-current extraction. The liquid level should be controlled by a dump valve operation, either automatic or manual, but based on an observed level. A sight glass suitable for the desired pressure range, or a level indicator could be used. Liquid samples can then be taken during the entire course of the run via the dump line. Column internals could be obtained as standard industrial equipment. Ongoing research in the area of supercritical fluid extraction (SCFE) of phenols from wastewater should attempt to measure concentrations of extracted materials in both 89 the SCF phase and the liquid phase. This provides a means of cross-checking measurements using material balances and is inherently more accurate. Concentrations in the SCF phase can be measured using thin layer chromatography (43) or supercritical fluid chromatography (44). Either method would require a direct hookup to a slip stream of the SCF exiting the column. The ability to analyze the SCF phase would be par­ ticularly helpful during phase equilibrium investigations. Continued research in this area is necessary to refine the data already gathered, and to properly analyze extraction column effectiveness. Additional research in co-current extraction should seriously consider a mixer-settler approach. This should reduce the time required for the SCF and liquid phases to reach equilibrium, and eliminate the unknowns of column hydraulics and distribution. Proper control of feed rates and feed temperatures could make this method of extraction a convenient tool for obtaining equilibrium data, especially if a method of recycle were built in. The effects of varying system temperatures should be investigated. Generally, as system temperatures are , increased, a substantial increase in pressure is required to achieve what is believed to be desirable carbon dioxide densities. On the other hand, as SCF densities approach 90 normal liquid densities, the increases in temperature tend to increase solvent power, mimicking the characteristics of liquid/liquid extraction. However, little is known about the actual effects of temperature for this particular SCF/ liquid system arid an optimum temperature should be determined. Finally, further research should investigate the ability of supercritical carbon dioxide to extract other major components of coal gasification wastewater. These include ammonia, COD (of which phenol is a constituent), hydantoins (45), ketones, aldehydes, and cyanide. Proof of these extraction capabilities could well make supercritical fluid extraction a viable industrial process, especially at the near critical conditions of carbon dioxide. 91 NOTATIONS 92 NOTATIONS a - Effective mass transfer surface area per unit of packed volume BPR - Back pressure regulator CO2 - Carbon dioxide CP - Critical point E - Efficiency (KA/KT times 100) ER - Relative efficiency FI - Flow indicator G - Molar flow rate of carbon dioxide (gmole/hr) GC - Gas chromatograph - Height of an overall gas-phase transfer unit H2O - Water K - Distribution coefficient (y/x) K - Overall gas-phase coefficient - Actual measured distribution coefficient CI - Theoretical distribution coefficient L - Molar flow rate of water (gmole/hr) N - Mass flux NCL - Near critical liquid. Nq^ - Number of overall gas-phase transfer units P — System pressure Pc - Critical pressure 93 PI - Pressure indicator Pr - Reduced pressure, P/Pc r - Correlation coefficient for Method of Least Squares SCF - Supercritical fluid SCFE - Supercritical fluid extraction T - System temperature Tc - Critical temperature TI - Temperature indicator Tj. - Reduced temperature, T/Tc tR - Relative residence time x - Mole fraction of phenol in the liquid phase x, - Mole fraction of phenol in water leaving counter-current extraction column x„ - Mole fraction of phenol in water entering counter-current extraction column y - Mole fraction of phenol in the carbon dioxide phase y -.Mole fraction of phenol in carbon dioxide entering counter-current extraction column y - Mole fraction of phenol in carbon dioxide leaving counter-current extraction column y* - Gas-phase mole fraction of phenol corresponding to equilibrium with the bulk liquid composition y*BM - Log mean average of (1-y) and (1-y*) 94 REFERENCES CITED 95 REFERENCES CITED 1. Mastertonz W. , and E. Slowinskiz Chemical Principlesz Saunders, Philadelphia, 1977, 265-266. 2. Bott, T. R., Fundamentals of Carbon Dioxide in Solvent Extraction, Chemistry and Industry, London, Issue 12, 1982, 394-396. 3. de Filippi, R. P., COg as a Solvent: Application to Fats, Oils and Other Materials, Chemistry and Industry, London, Issue 12, 1982,. 390-394. 4. Paulaitis, M. E., V. J. Kfukonis, R. T. Kurnik, and R. C. Reid, "Supercritical Fluid Extraction", Rev. Chem. Eng., 1982, I, 179. 5. Little, A.D. Chemical Week, 9 July 1980, 31. 6. McKugh, M., and V. J. Krukonis, "Processing with Supercritical Fluids", AIChE Advanced Seminar, September, 1983. 7. Scott, R. L., Thermodynamics of Critical Phenomena in Fluid Mixtures, Ber. Bunsenges. Phys. Chem., 1972, 76, 296. 8. Scott, R. L. and P. H. van Konynenburg, State Properties of Solutions - Van der Waals and Related Models for Hydrocarbon Mixtures, Discuss. Faraday Soc., 1970, 49, 87. 9. Balder, J. R. and J. M. Prausnitz, Thermodynamics of Ternary, Liquid-Supercritical Gas Systems with Applications for High Pressure Vapor Extraction, Ind. Eng. Chem. Fund., 1966, 5, 449. 10. Wenzel, H. and W. Rupp, Calculation of Phase Equilibria in Systems Containing Water and Supercritical Components, Chem. Eng. Sci., 1978, 33, 683. 96 11. Peng, D. Y. and D. B. Robinson, Two and Three Phase Equilibrium Calculations for Systems Containing Water, Can. J. Chem. Eng., 1976, 54, 595. 12. Deiters, U. and G. M. Schneider, Fluid Mixtures at High Pressures. Computer Calculations of the Phase Equilibria and the Critical Phenomena in Fluid Binary Mixtures form the Redlich-Kwong Equation of State, Berichte Bun. Gesselschaft, 1976, 12, 1316. 13. Chai, C. P., "Phase Equilibrium Behavior for Carbon Dioxide and Heavy Hydrocarbons", Ph.D. Thesis, University of Delaware, Newark, Delaware, 1981. 14. Prausnitz, J. M., "Phase Equilibria for Complex Mixtures",, presented at the Third Int. conference on Fluid Properties and Phase Equilibria for Chemical Process Design, Callaway Gardens, Georgia, April 10-15, 1983. 15. Mathias, P. M. and T. W. Copeman, "Extension of the Peng-Robinson Equation of State to Complex Mixtures: Evaluation of the Various Forms of the Local Composition Models", presented at the Third Int. Conference of Fluid Properties and Phase Equilibria for Chemical Process Design, Callaway Gardens, Georgia, April 10-15, 1983. 16. Wong, J. M. and K. P. Johnston, Thermodynamic Models . for Nonrandom and Strongly-Nonideal Liquid Mixtures, Ind. Eng'. Chem. Fund., 1984, 23, 320-326. 17. Johnston, K. P. and C. A. Eckert, An Analytical Carnahan-Starling van der Waals Model for Solubility of Hydrocarbon Solids in Supercritical Fluids, AIChE Journal, 1981, 27, 773. 18. Johnston, K. P., D. H. Zieger, and C. A. Eckert, Solubilities of Hydrocarbon Solids in Supercritical Fluids: The Augmented van der Waals Treatment, Ind. Eng. Chem. Fund., 1982, 21, 191. 19. MacKay, M. E. and M. E. Paulaitis, Solid Solubilities of Heavy Hydrocarbons in Supercritical Solvents, Ind. Eng. Chem. Fund., 1979, 18, 149. 20. Hannay J. B. and J. Hogarth, Proc. Royal Society, London, Ser. A, 1879, 29, 454. 97 21. Williams, D. F., Extraction with Supercritical Gases, Review Article Number 5, Chemical Engineering Science, Vol. 36, No. 11, pp. 1769-1788, 1981. Printed in Great Britain, Pergamon Press Ltd. D. F. Williams, National Coal Board, Coal Research Establishment, Stoke Orchard, Cheltenham, Gloucestershire, GL524RZ, England. 22. Gardner, D. S., Industrial Scale Hop Extraction with Liquid COg, Chemistry and Industry, London, Issue no. 12, 1982, 402-406. 23. Brogle, H., COg as a Solvent: Its Properties and Applications, Chemistry and Industry, London, Issue no. 12, 1982, 385-390, 24. Perry, R. H., and C. H. Chilton, Chemical Engineers' Handbook, Fifth Edition. McGraw-Hill Co., New York, 1969, Chapter 3. 25. Vollbrecht, R., Fundamentals of Carbon Dioxide in Solvent Extraction, Chemistry and Industry, London, Issue . 12, 1982, 397-399. 26. Mack, H., and M. Koepsel, Parfumerie und Kosmetik, 1973, 54, 233-237. 27. Calame, J. P., and R. Steiner, Fundamentals of Carbon Dioxide in Solvent Extraction, Chemistry and Industry, London, Issue 12, 1982, 399-402. 28. Granlund, C. R., "Semicontinuous Supercritical Fluid Extraction of Peat", Master of Science Thesis/ Montana State University, 1982. 29. Smith, S. S., "Semicontinuous Supercritical Extraction", Master of Science Thesis, Montana State University, 1983. 30. Myklebust, K. L., Jr., "Supercritical Fluid Extraction of Peat", Master of Science Thesis, Montana State University, 1982. 31. Van Leer, R. A. and M. E. Paulaitis, Solubilities of Phenol and Chlorinated Phenols in Supercritical Carbon Dioxide, J. Chem. Eng. Data, 1980, 25, 257-259. 98 32. McHugh, M. and M. E. Paulaitis, Solid Solubilities of Naphthalene and Biphenyl in Supercritical Carbon Dioxide, J. Chem. Eng. Data, 1980, 25, 326-329. 33. Kurnik, R. T., S. J. Holla, and R. C. Reid, Solubility of Solids in Supercritical Carbon Dioxide and Ethylene, J. Chem. Eng. Data, 1981, 26, 47-51. 34. Tsekhanskaya, Y. V., M. B. Iomtev, and E. V. Muskina, Zh. Fiz. Khim.,1964, 38, 2166.. 35. Kuk, M. S., and J. C. Montagna, Solubility of Oxygenated Hydrocarbons in Supercritical Carbon Dioxide, Chemical Engineering at Supercritical Fluid Conditions, Edited by M. E. Paulaitis, J. M. L. Penninger, R. D. Gray, Jr., and P. Davidson, Ann Arbor Science Publishers, Ann Arbor Michigan, 1983, Chapter 4. 36. Elsenbeiss, J., San Antonio, TX, Aug. 1964, Southwest Research Institute Final Report Contract No. DA 18-108-AMC-244(A). 37. Tsekhanskaya, Y. V., M. B. Iomtev, and E. V. Muskina, Russ. J. Phys. Chem. (Engl. Transl.), 1962, 36, 1177. 38. Paulaitis, M. E., M. A. McHugh, and C. P.-Chai, Solid Solubilities in supercritical Fluids at Elevated Pressures, Chemical Engineering at Supercritical Fluid Conditions, Edited by M. E. Paulaitis, J. M. L. Penninger, R. D. Gray, Jr., and.P. Davidson, Ann Arbor Science Publishers, Ann Arbor Michigan, 1983, Chapter 6. 39. Francis, A. W., Ternary Systems of Liquid Carbon Dioxide, J. Phys. Chem., 58, 1099, 1954. 40. Neter, J., and W. Wasserman, Applied Linear Statistical Models, Richard D. Irwin, Inc., Homewood, Illinois, 1974. 41. Sherwood, T. K., R. L. Pigford, and C- R- Wilke, Mass Transfer, McGraw-Hill Co., New York, 1975. 42. Heller, H., Choosing the Right Formula for Calculator Curve Fitting, Chemical Engineering, February 13, 1978, 119-120. 99 43. Stahl, E., W. Schilz, E. Schutz, and E. Willing, A Quick Method for the Microanalytical Evaluation of the Dissolving Power Supercritical Gases, Angew. Chem. Int. ' Ed. Enlg., 1978, 17, 731-738. 44. Gere, D. R., R. Board, and D. McManigill, Supercritical Fluid Chromatography with Small Particle Diameter Packed Columns, Anal. Chem., 1982, 54, 736-740. 45. Olson, E. S. , J. Wi Diehl, and D. J . Miller, "Hyda.ntoins in the Condensate Water from the Slagging Fixed Bed Gasification of Lignite", U.S. Dept, of Energy, Grand Forks Energy Technology Center, Grand Forks, North Dakota, 1982. 46. Jeris, J. S. A Rapid COD Test, Water and Wastes Engineering, 1967, 4, 89-91. 47. Standard Methods, 14th Ed., 1976, 650. APPENDICES 101 . APPENDIX A ANALYTICAL PROCEDURES AMMONIA ANALYSIS USING THE ORION 901 IONALYZER 102 1. Before the analysis period starts, the instrument will have been on and the electrode slope will have been determined. The electrode will be in a 10 ppm solution with added NaOH. 2. Prepare a fresh 10 ppm N standard by adding I ml of 1000 ppm standard to 98 ml of distilled water in a 150 ml beaker. 3. Prepare a blank by placing 100 ml of distilled water in a 150 ml beaker. Place both the standard and the blank in the tray of 23°C water. After these samples have set for 15-20 minutes, the instrument may be calibrated. 4. Set the instrument mode to CONCN. Set the STD VALUE switches to 10,000. Set the SLOPE switches to the value given the instructor (sign must be minus). 5. Rinse the electrode in distilled water by dipping it into a beaker of distilled water two or three times. Place the electrode in the 10 ppm standard prepared above. Add I ml of 10 M NaOH to the solution. Use the magnetic stirrer at all times. Press READ/CLEAR MV and wait for a stable reading. Press SET CONCN. 6. Rinse the electrode well in distilled water. 7. Place the electrode in the blank and add I ml NaOH stirring constantly. Wait 4 minutes and press SET BLANK. Rinse the electrode in distilled water and place it in the 10 ppm solution where it was originally. 8. The instrument is now calibrated and is ready for unknown samples. To read sample concentrations rinse the electrode, place it in the sample, add I ml 10 M NaOH and wait for a stable reading. Be sure to rinse the electrode between samples. When the electrode is not in use, it should be in the 10 ppm solution. 9. The concentration readings will be in ppm. N in the sample in contact with the electrode. 103 10. Caution: If the READ/CLEAR MV button should be inadvertently pushed, the calibration should be repeated (steps 2 through 7). 11. SHUT DOWN PROCEDURE: . a) Disconnect the electrode from the instrument and place it in the solution labeled for overnight storage. b) Connect the shorting wire across the instrument where the electrode was disconnected. c) Set the instrument on STDBY mode. d) Unplug the instrument. 104 CHEMICAL OXYGEN DEMAND (COD) - RAPID METHOD (46, 47) Method Titrimetry Apparatus Large hot plate preheated for one hour, thermometer reading at least 200°C, 500 ml Erlenmeyer flasks, spoon to hold 0.3 grams of HgSO^, glass beads. Principle " See "Chemical Oxygen Demand - Standard Method" (47). In the rapid method the digesting .liquor contains both concentrated HgSO^ and H^PO^. The sample is heated to 180OC and immediately cooled, so the test is run in about 20 minutes as compared to 3 hours for the standard method. Interferences and Accuracy In general, results are usually 2% to 10% lower than the standard method. A substance which is difficult to oxidize by the standard method (such as acetic acid, glycine, pyridine) is less completely oxidized by the rapid method.. For high accuracy, compare 10 replicates of the wastewater analyzed by each method and apply a correction factor to the rapid method. Procedure 1. Add approximately 0.3 g of HgSO^ to a 500 ml Erlenmeyer flask with a spoon. (With practice the tip of a spatula can be used.) Add a few boiling beads. 2. Pipet 10 ml or a suitable aliquot of.the sample diluted to 10 ml into the flask (approximately 2 to 8 mg COD in sample - 400 to 1600 mg/1 - for . best results). Mix well for 15 seconds or if Cl is present mix I minute. IMPORTANT I 105 3. Two blanks are run by substituting 10 ml distilled water for the sample and continuing with Steps 4 . through 9. The average of the two titrations is used in the calculations. 4. Carefully add 25 ml of the RAPID DICHROMATE ACID solution to the flask and swirl the contents. Do not use either "Dichromate Standard" nor COD Standard Method sulfuric acid reagent. 5. Handling the flask by the neck, place the flask on a preheated hot plate and heat to approximately 180°C. A thermometer should be immersed and frequent swirling should be employed. The temperature is read by tilting the flask to immerse thethermometer sufficiently. Approximately 5 to 8 minutes is required to reach 180°C, at which time the flask is removed from the hot plate. With experience, 4 or 5 flasks may be heated at one time. 6. Cautiously add approximately 300 ml of distilled water. (The addition of water to the very hot acid mixture can induce vigorous boiling and spattering. Be careful.) 7. Cool the solution in a water bath to ambient temperature. 8. Add 5 drops of ferroin indicator and titrate to an orange end point with ferrous ammonium sulfate, 0.05 N. 9. Two standards are run by adding 25 ml of 0.050 N- dichromate standard, acidifying with 20 ml concentrated sulfuric acid and continuing with steps 5, 6, and 7. 10. Discharge into sink with copious running water and continue to run water for at least 10 minutes after last discharge to prevent corrosion of drain pipes. Calculations The COD is calculated as follows: (25) (0.05) ml from Step 9 Where N equals the normality of Fe(NH^)2(SO^)2 N x 8000 x (ml from Step 3 - ml from Step 8) COD (mg/1): ml sample Reagents 1. Rapid Dichromate Acid.Solution. Dissolve 5 g K2Cr3O7 and 20 g Ag3SO4 in a solution . consisting of I liter each of concentrated H3SO4 and H3PO4. Label "RAPID DICHROMATE ACID SOLUTION (for Rapid COD only)". 2. Ferrous ammonium'sulfate. Dissolve 20 g of Fe(NH4)3(SO4)3-S H3O in distilled water, add 20 ml concentrated H3SO4 and dilute to I liter. This solution is standardized against 0.050 N K3Cr3O7 prior to use. 3. Ferroin indicator solution. Dissolve 1.485 g I,10-phenanthroline monohydrate, together with 695 mg FeSO4*'7 H3O in water and dilute to 100 ml. 4. Mercuric sulfate. 5. Dichromate standard, 0.05 N. Dissolve 2.4518 g K3Cr3O7 dried, primary standard grade in distilled water and dilute to I liter. All of above are ACS reagent grade chemicals. APPENDIX B SAMPLE CALCULATIONS 108 CALCULATION OF PHENOL MOLE FRACTIONS Calculation of Mole Fraction of Phenol in Water (x). I mg phenol I g I mole phenol 18 g water ----------- x ---- x -----;-------- x ----------- = 1000 g water 1000 mg 94 g phenol I mole water I mg phenol - ----------- = 1.915x10 mole phenol/mole water 1000 g water For dilute solutions (less than 1%) the mole fraction of phenol in water is essentially equivalent to the mole ratio. Example (Run #1): 4370 mg/1.x 1.915 x 10 ^= 0.0008369 mole phenol/mole water 109 Calculation of y (Mole Fraction of Phenol in Carbon Dioxide) by Mass Balance. Example (Run #2): Let x^ = Initial mole fraction of phenol in water Xg = Final mole fraction of phenol in water Xg = 10000 mg/1 x 1.915 x 10 ^= 0.001915 mole/mole Xg = 5700 mg/1 x 1.915 x 10 ^= 0.001092 mole/mole The relative flow of carbon dioxide to water is 0.71 mole carbon dioxide to I mole water. For every mole of water that enters and exits the column (0.001915 -0.001092) moles of phenol are transfered to 0.71 moles of carbon dioxide. 0.001915 - 0.001092 Therefore: y = ------------------- 0.71 = 0.001160 moles phenol/mole CC^ H O DETERMINATION OF ACTUAL DISTRIBUTION COEFFICIENT Let CT = actual measured distribution coefficient. mole fraction phenol in carbon dioxide phase K = ----------------------------------------------------------------------------------------------- ------------------------------------------ = mole fraction phenol in water phase Example (Run #2): • x = 0.001092 mole/mole y = 0.001160 mole/mole (by mass balance) CT A LFLLDDXLOLFLLDLNG A DFLXG IDETERMINATION OF THEORETICAL DISTRIBUTION COEFFICIENT Let K1J1 = the theoretical distribution coefficient. Example (Data from Run #2): Carbon dioxide density = 0.806 g/cc Estimate x^ = 0.0012 mole/mole Therefore y = (0.001915 - 0.0012)/0.71 = 0.00101 By referring to Figure 10, y should equal 0.00094 when x = 0.0012. Pick a new estimate for x^ = 0.0013 Now y .= (0.001915 - 0.0013)/0.71 = 0.000866 From Figure 10, y should now equal 0.00098. Try x^ = 0.00122 Then y = (0.001915 - 0.00122)/0.71 = 0.000979 But Figure 10 shows y = 0.00096 for x = 0.00122 Try x^ = 0.00123 Then y = (0.001915 - 0.00123)/0.71 = 0.000965 And Figure 10 shows y = 0.00097 Since the two values of y correspond within the accuracy of Figure 10, the theoretical x and y values are taken to be: x = 0.00123 and y = 0.000965 and CI A LFLLLNXMOLFLLDGW A LFBbM 112 DETERMINATION OF EXTRACTION EFFICIENCY, RELATIVE EFFICIENCY AND RELATIVE TIME. Extraction■efficiency (E). E A CTOCI y 100 Example (Run #2): CT = 1.062 and CI A LFBbM Therefore., E = KA/KT x 100 = (1.062/0.785) x 100 = 135% Relative Efficiency (Er). Each run's extraction efficiency was normalized relative to the most efficient extraction run which used the phenol/water standard. The most efficient run, Run #5, had an extraction efficiency of 145%. For Run #2., where E = 135%, ER = (135/145) x 100 = 93% Relative Time. Total column flow rates were calculated for each run and normalized relative to the run with the slowest flow rate (longest residence time). Run #2. For Run #2, the combined carbon dioxide and water flow rate was 1.69 ml/min. For Run #1, the same value was 4.08 ml/min. Therefore, the relative residence time is: tR = 1.69/4.08 = 0.41 113 DETERMINATION OF ABSOLUTE PHENOL EXTRACTION Example (Run #1): Relative mass flow rate = 4.65 g carbon dioxide/g water Initial phenol concentration in water = 10000 mg/1 Final phenol concentration in water = 4370 mg/1 10000 mg/1 - 4370 mg/1 I g water Phenol extracted = --------=— :-------'---- x --------- 1000 g water/1 4.65 g CO^ = 1.21 mg phenol/g CO2 om)We)e sr»-rc — _ 1762 10013799 Main N378 Gartner, Ambrey Gordon G197 Supercritical fluid cop.2 extraction of... DATE ISSUED TO Main H378 G197 cop. 2