Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Concentration of bleached-chemi-thermo-mechanical pulp effluent by propane hydrate formation Ngan, Yee Tak 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1996-0074.pdf [ 5.02MB ]
Metadata
JSON: 831-1.0058540.json
JSON-LD: 831-1.0058540-ld.json
RDF/XML (Pretty): 831-1.0058540-rdf.xml
RDF/JSON: 831-1.0058540-rdf.json
Turtle: 831-1.0058540-turtle.txt
N-Triples: 831-1.0058540-rdf-ntriples.txt
Original Record: 831-1.0058540-source.json
Full Text
831-1.0058540-fulltext.txt
Citation
831-1.0058540.ris

Full Text

C O N C E N T R A T I O N O F B L E A C H E D - C H E M I - T H E R M O - M E C H A N I C A L P U L P E F F L U E N T B Y PROPANE H Y D R A T E F O R M A T I O N by Y E E T A K N G A N B.ASc. University of British Columbia, 1992 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF T H E REQUIREMENTS FOR T H E D E G R E E OF M A S T E R OF APPLIED SCIENCE in T H E F A C U L T Y OF GRADUATE STUDIES (Department of Chemical Engineering) We accept this thesis as conforming to the required standard T H E UNIVERSITY OF BRITISH COLUMBIA December 1995 © YeeTakNgan, 1995 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholariy purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of ^Cht^^j y « c e . < / l » ^ The University of British Columbia Vancouver, Canada Date D e - c \ g A < DE-6 (2/88) ABSTRACT 11 During the past several years there is increasing public pressure for stringent environmental control for pulp mill water discharge to the receiving environment. The industry responded by developing zero liquid discharge technology that applies to mechanical pulp mills. The goal of these mills is to recover water suitable for reuse. That technology is based on evaporation. Another method of recovering water from a solution with impurities is through the formation of clathrate hydrate crystals. However, this method is not industrially proven. Only evaporation operates at an industrial scale. Clathrate hydrate crystals are inclusion compounds that can exist below or above the normal freezing point of water under suitable pressure and temperature. Effluent concentration through clathrate hydrate formation is based on the fact that impurities present in the effluent are not contained in the clathrate structure. The motivation for using clathrate hydrates for effluent concentration is the fact that the process can operate at low temperatures (1 to 5 °C for propane hydrate) and hence, the potential for scaling and corrosion is reduced. In this work, we chose to use vapour and liquid propane to form hydrate crystals for the purpose of concentrating BCTMP mill effluent, and recovering clean water. A new experimental apparatus was designed and built. The main characteristic of this apparatus is the fact that we are able to form the hydrate crystals, drain the effluent concentrate, wash and melt the hydrate crystals in one vessel. All unit operations were performed in a batch wise manner. We grew crystals at different conditions by varying the quantity of hydrate former, the impeller geometry, and varying the driving force. Experimental runs were performed with and without washing. The crystals were washed with either liquid propane or distilled water. Because BCTMP effluent is a dark colour solution that makes it difficult to observe the crystallization and separation process we worked first with a 2.5 wt % NaCl and then with the BCTMP effluent. I l l The effectiveness of separation for both the brine solution and the effluent was determined by analyzing the amount of impurities in the recovered water, the spent wash solutions, and the concentrate. We found that crystals grew the fastest with a large driving force while using liquid propane mixed in a rigorous manner during the crystallization process. The quality of the recovered water depended upon the amount of hydrate former used during the crystal growth, and the extent of drainage. Improved drainage was achieved by displacing concentrate from the crystal using liquid propane. Washing the crystal with water did not perform well because there was poor contacting between water and the crystals. It was also found that the rate of crystal growth did not affect the quality of recovered water. The average reductions in the conductivity of the brine solutions were 31.5%, 66.1%, 32.1% for the no washing, washing with water and displacement with propane experiments respectively. In the effluent experiments the corresponding reductions in conductivity and TOC (total organic carbon) were, 22.5%, 22.5%; 23.1%, 11.4%; 45.1%, 27.3% respectively. iv T A B L E O F C O N T E N T S ABSTRACT '. ii TABLE OF CONTENTS iv LIST OF TABLES vi LIST OF FIGURES vii LIST OF PHOTOGRAPHS ix ACKNOWLEDGMENTS x CHAPTER 1: Introduction 1 CHAPTER 2: Clathrate Hydrates 5 2.1 Structure & Properties 5 2.2 Clathrate Hydrate Concentration ;-. 10 2.3 Research Objectives 12 CHAPTER 3: Experimental Section 14 3.1 Experimental Apparatus 14 3.1.1 Reactor and wash column 14 3.1.2 Gas supply vessel 19 3.1.3 Wash water supply vessels 22 3.1.4 Temperature control system 22 3.1.5 Pressure control system 24 3.2 Effluent Tests 26 3.3 Experimental Procedures 27 3.3.1 Nucleation and Establishment of Hydrate-Propane-Water Equilibrium 27 3.3.2 Crystal Growth 29 3.3.3 Crystal Separation 32 CHAPTER 4: Results 39 V 4.1 Salt Solution Experiments 39 4.1.1 Experiments with No Washing 41 4.1.2 Experiments with Displacement with Propane 49 4.1.3 Experiments with Washing with Water 53 4.2.1 No Washing Experiment E l 58 4.2.2 No Washing Experiment E2 .-. 63 4.2.3 No Washing Experiment E3 65 4.2.4 Displacement with Propane Experiment E4 67 4.2.5 Washing with Water Experiment E5 69 CHAPTER 5: Discussion 72 CHAPTER 6: Conclusion and Recommendations 77 REFERENCES 80 APPENDIX A: Procedure for Determining TS, VS, FS, and TSS 83 APPENDIX B: Set up Procedure 86 APPENDIX C: Calculation for Determining the quantity of propane inside the Gas Supply Vessel 89 APPENDIX D. Equilibrium Results 90 APPENDIX E: Propane Consumption Test for Vapour Propane in BCTMP Effluent 92 vi LIST OF TABLES TABLE 2.1 Hydrate Former Characteristic of Propane 13 TABLE 4.1 Operating Conditions (Salt Solutions) 40 TABLE 4.2 Results for Separation with No Washing 49 TABLE 4.3 Results for Propane Displacement 53 TABLE 4.4 Results for Experiments with Water Washing 54 TABLE 4.5 Operating Conditions (Effluent Experiments) 59 TABLE 4.6 Results for Run E l 61 TABLE 4.7 Results for E2 63 TABLE 4.8 Results for E3 67 TABLE 4.9 Results for Run E4 69 TABLE 4.10 Results for Run E5 70 TABLE 5.1 Factor Affecting Rate of Propane Consumption 72 TABLE 5.2 Comparing Results from Propane Displacement Experiment 73 TABLE 5.3 Summary of Salt Experiments 74 TABLE 5.4 Summary of Effluent Experiments 76 TABLE D . l Equilibrium Results for Salt Solutions 90 TABLE D.2 Equilibrium Results for Effluent Experiments........ 90 TABLE E. 1 Results for KE1 93 TABLE E.2 Results for KE2 93 vii LIST O F FIGURES FIGURE 2.1 Cavities and Unit Cells for Clathrate Hydrate Structures 6 FIGURE 2.2 Propane -Water Partial Phase Diagram 8 FIGURE 3.1 Experimental Apparatus 15 FIGURE 3.2 Crystallizer/Wash Column 16 FIGURE 3.3 Crystallizer/ Wash Column.... 17 FIGURE 3.4 Gas Supply Vessel 20 FIGURE 3.5 Wash Water Supply Vessel 23 FIGURE 3.6 Pressure Control System 25 FIGURE 3.7 Nucleation & Establishment of the Hydrate-Propane-Water Equilibrium 28 FIGURE 3.8 Procedure for No Washing Experiments 33 FIGURE 3.9 Procedure for Liquid Propane Displacement Experiments 34 FIGURE 3.10 Procedure for Washing with Water Experiments 35 FIGURE 4.1 Initial Condition during the Crystal Growth Period 42 FIGURE 4.2 Salt Experiments with No Washing 46 FIGURE 4.3 Brine Displacement with Liquid Propane 51 FIGURE 4.4 Salt Experiments with Liquid Propane Displacement 52 FIGURE 4.5 Salt Experiments Water Washing 56 FIGURE 4.6 Run E l No Washing (Results) 62 FIGURE 4.7 Run E2 No Washing (Results) 64 FIGURE 4.8 Run E3 No Washing (Results) 66 FIGURE 4.9 Run E4 Displacement with Liquid Propane (Results) 68 FIGURE 4.10 Run E5 Washing with Water (Results) 71 FIGURE D. 1 Partial Phase Diagram for Hydrate-Propane-Water Equilibrium (Equilibrium Results for Salt Solution and Effluent) ...91 FIGURE E. 1 Kinetic Experiment for Run KE1 94 FIGURE E .2 Kinetic Experiment for Run K E 2 IX LIST OF PHOTOGRAPHS PHOTOGRAPH 4.1 Hydrate Plug in Brine Solution 44 PHOTOGRAPH 4.2 Formation of a Bed of Crystals Bed Formation Procedure A) 47 PHOTOGRAPH 4.3 Floating Process 48 PHOTOGRAPH 4.4 A Bed of Crystals After Washing with a Non Ionic Dye Solution 55 PHOTOGRAPH 4.5 Hydrate Plug Formed in Experiment E l 60 ACKNOWLEDGMENTS I would like to express my sincere appreciation to the following persons and organizations: Prof. P. Englezos for his patience and all his support! The Natural Science and Engineering Research Council of Canada (NSERC) for their financial support. Louisiana Pacific Canada Ltd. for donating effluent samples. The staff at the Chemical Engineering Department and Pulp and Paper Center for all their help Tazim Rehmat, Margaret Chen, Alan Werker, and John Moritz, Belinda Larish, and Steve Helle for lending out laboratory equipment 1 Chapter 1: Introduction Effluents generated by thermo-mechanical pulp (TMP) mills and chemi-thermo-mechanical pulp (CTMP) mills contain various contaminates like lignin, inorganic salts, and phenolic constituents. Gaarder (1993) discussed the factors that determine the type and concentration of contaminants in a particular effluent and the impacts from the discharge into a particular ecosystem. During the past several years there is an increasing public pressure for stringent environmental control in the pulp and paper as well as in other industries. The pulp and paper industry responded by investing significantly on pollution prevention technology that is based predominantly on sedimentation and biological degradation of the contaminants. In addition, we saw the appearance of closed cycle or zero liquid discharge (ZLD) mechanical pulp mills as well as an intense effort to reduce water usage in Kraft mills. An essential part of a Z L D plant is technology that will be capable of removing dissolved substances or recovering clean water from the effluent. Processes for the recovery of water include evaporation, membrane separation, and crystallization (freeze concentration and clathrate hydrate concentration). At present, evaporation is the only process that may operate in industrial scale plants. Freeze concentration/crystallization refers to the concentration of a solution by generating ice crystals followed by the physical removal of the ice crystals from the solution, and the subsequent melting of these crystals (Heist, 1989; Englezos, 1994).' In clathrate concentration, the ice formation step is replaced by clathrate hydrate crystal formation. Both processes are based on the fact that the impurities present in the original solution are not contained within the ice or clathrate hydrate crystal structure. Freeze Concentration Freeze concentration is currently used in the food industry because it can separate water from a solution without the loss of flavor and aroma often associated with evaporation 2 (Chowdhury, 1988; Douglas, 1989; Fellows, 1990). Commercial freeze concentration systems process less than 190 L/min of solution. Freeze concentration of kraft pulping liquor and bleach plant effluent has been studied but neither was found to be economically competitive (Rousseau and Sharpe 1980; Rousseau, 1981; Coleman, 1986). A combined reverse osmosis/ freeze concentration system for bleach plant effluents was also found to be quite expensive (Wiley et al. 1978). A pilot plant freeze system was built at an Ontario BCTMP mill in 1991. Based on the limited results that were reported, it was found that the freeze process could reduce the concentration of low molecular weight organics, resin and fatty acids, the BOD and COD loads, and the electrolyte content in the recovered process water (Kenny et al. 1992). A freeze concentration system was built in 1991 at Louisiana-Pacific Canada Ltd.'s B C T M P mill near Chetwynd, B C to treat 4,000 L/min and to concentrate the effluent from 2% solids up to 10%. The freeze unit was a shell and tube heat exchanger. Ammonia flowed through the shell side while effluent flowed through the tube side. This unit would approximately result in an 80% recovery of the water from the effluent. The effluent was to be further concentrated up to 50% solids in an evaporation unit that was designed to use waste-heat from the refrigeration cycle. The freeze process was not found to be sufficiently reliable (Young, 1994) and the mill converted to evaporation. The following hampered the crystallization process. At low heat-transfer rates, the crystals grew to sufficient size for good washing characteristics and the crystallizer's tubes generally remained clean. But as the heat transfer rate was increased, the crystal size degraded and the tubes became coated. Layers of ice eventually reduced the heat transfer to zero and plugged the tubes. Freeze concentration has a potential energy advantage compared with evaporation because the heat of fusion of water (6.01 kJ/mol) is smaller than the heat of vaporization (40.66 kJ/mol). However, this advantage is reduced or even reversed with the use of multiple effect or vapor recompression evaporators. Because the freeze process operates below room-3 temperature the potential for scaling and corrosion problems is greatly reduced compared to evaporation. Clathrate Hydrate Concentration Clathrate hydrates are ice-like crystalline, non-stoichiometric compounds (Sloan, 1990; Englezos, 1993). They are formed from water and molecules such as carbon dioxide, and light hydrocarbons. Hydrate crystals can form at temperatures several degrees above the normal freezing point of water, thereby decreasing the energy requirements compared with freeze concentration. These temperatures, however, are not high enough to cause corrosion, scaling or loss of volatile substances, as is the case with evaporation. The concentration of aqueous streams by hydrate formation was patented by Glew (1962). Hydrate formation in seawater was considered as the basis of a process to recover pure water. The process was demonstrated at a pilot plant stage (Knox et al. 1961; Tleimat, 1980) but was never developed commercially because other technologies could achieve the same purity with lower cost. The process has recently become of interest in waste minimization, the food products industry and the concentration of other solutions (Heist, 1988; Douglas, 1989; Willson et al. 1990). Work has been initiated at the University of British Columbia to explore the use of hydrates in mechanical pulp mill effluent concentration (Gaarder, 1993). Clathrate hydrate concentration could potentially overcome the main problem that the freeze concentration process faced in Chetwynd because one can control the rate of crystal growth by operating at different pressures for a given temperature. The scope of this work was to study the separation of propane hydrate crystals from an effluent and determine the degree of water recovery. Past experience with the attempts to use clathrate crystals for the recovery of water from seawater were based on using three different pieces of equipment namely, the crystallizer, the separator (wash column or centrifiige) and the 4 melter. Because we want to reduce the capital and operating costs of clathrate hydrate concentration, we designed an apparatus that combines these three equipment into one. In the next chapter, a brief description of clathrate hydrates is given together with a review of previous clathrate hydrate concentration work. In chapter three, a description of the experimental apparatus is given. The experimental results are presented and discussed in the fourth and fifth chapter, respectively. Finally, conclusions and recommendations are given in the sixth chapter. 5 Chapter 2: Clathrate Hydrate Concentration The first section in this chapter will discuss how clathrate hydrates are formed. The second section will review the previous work on clathrate hydrate concentration. The last section presents the research objectives. 2.1 Structure & Properties Gas hydrates are non stoichiometric inclusion compounds. They have the appearance of wet snow. Hydrates will form under suitable temperature and pressures when water is brought in contact with sufficient amount of one or more hydrate forming substances. There are more than one hundred chemicals, which can form gas hydrates. Under suitable pressure-temperature conditions, a gas molecule can stabilize a lattice of water molecules forming a cage structure. The gas molecules fit inside the cage formed by hydrogen bonding of water molecules. The host gas helps to stabilize this structure. Without the host gas, the structure is unstable. There is no chemical interaction between the water and the guest molecules. Clathrate hydrate crystallizes in two structures, body centered cubic (si) and diamond lattice (SII). The size of the host gas determines the type of structure. There are two types of cavities in each structure, a large and a small cavity. The small cavity is denoted as 5*2 (pentagonal dodecahedron) because it consists of twelve pentagonal faces. The large cavities are the 5*262 and the 5*264 Twelve pentagonal faces and two hexagonal faces constitute the 5*262 cavity. The other large cavity, 5*264 is a polyhedron with twelve pentagonal faces and four hexagonal faces. Both structures si and sll contain the small cavity. Figure 2.1 illustrates the two structures The unit cell of si consists of six 5*262, and the two 5*2 cavities. Their respective mean free diameter is 5.1 A, and 5.8A. Pure gases with an effective diameter smaller than 5.1 A 6 Figure 2.1: Cavities and Unit cell for Clathrate Hydrate Structures (Sloan, 1990) 5l2 62 7 will form only si. Some gases that will form si are Ar, CH4, and H2S. The unit cell of sU consists of sixteen 5 ^ and eight 5^6^ cavities. Their respective mean diameter is between 5.1 and 6.7 A. Propane, isobutane, and sulfur dioxide are some gases that will form sU. Phase Diagram Figure 2.2 displays a partial phase diagram for a propane-water-sodium chloride system in the hydrate formation region. The experimental data were taken from Kubota et al. (1984). Using these data points, a method of least squares was used to generate the continuous lines shown in the graph. The line AQ represents the equilibrium condition of hydrate-propane (vapor)-water (liquid). In the region above line AQ, hydrates will exist. Below line AQ, no hydrates will exist. The line K L illustrates the vapour pressure of propane. The line segment CQ depicts the equilibrium condition of hydrate-propane (liquid)-water (liquid). Hydrate can exist only to the left of line CQ. In region right of curve CQ, hydrate will not be present. Therefore, propane hydrate is formed at pressure-temperature conditions that lie to the left of the lines A Q and QC. Joining these equilibrium curves produces four distinct regions. The upper left region is bounded by line segment KQC. Inside this region, two phases will exist liquid propane, Lp, and hydrate, Sh. The upper right region is confined by line segment CQL. Inside this area, liquid propane and liquid water, Lw, will exist. The lower right region is bounded by line segment AQL, and it contains liquid water, and vapour propane, Vp. The last region that is bounded by line segment A Q K contains vapour propane and hydrate. The driving force for hydrate nucleation may be defined as a pressure or as a temperature difference. For example, if an experiment is performed at T exp, Pexp then AP=Pexp-Peq and AT=Teq-TCxp where Peq and T e q are the equilibrium hydrate formation pressure at T e xp and the equilibrium hydrate formation temperature at P exp respectively. Certain compounds have the ability to depress the formation of hydrates, and they are referred to as inhibitors. Strong inhibitors are electrolytes, and alcohols. To illustrate the effect of a 8 hydrate inhibitor, line A1Q1 represents the equilibrium conditions for hydrate-propane (vapour)-water (liquid) with a sodium chloride concentration of 2.5 wt%. Hence, we see the curve shifted to the left when an inhibiting agent is added to the system. The formation of clathrate hydrate crystals in mechanical pulp mill effluents was studied by Gaarder (1993). Experiments were carried out with propane, carbon dioxide, and a 30-70 mol % propane-carbon dioxide mixture in effluent samples from four different high-yield mechanical pulp mills. The effluent samples were generated at four pulp mills, an unbleached and a bleached thermo-mechanical pulp mill (TMP1, TMP2), a bleached chemi-thermo-mechanical pulp mill (BCTMP), and a combined BCTMP/TMP mill (CTMP). Experiments were also conducted in effluent concentrate samples. The range of pressure-temperature hydrate crystal formation conditions was measured. It was found that hydrate crystals can form in these effluents several degrees above the normal freezing point of water. The presence of impurities in the effluents did not cause any appreciable change in the formation conditions with the TMP and CTMP effluent samples, but it did influence the formation conditions in the B C T M P effluent. This behavior was a result of higher electrolyte and organic content in the B C T M P effluent. Kinetics Hydrate crystal formation is a two step process. Nucleation is the first step in which the water molecules form a stable cage structure with the quest molecules. The time required for the stable nuclei to form is called induction time. The understanding of nuclei formation is not complete. However, several experimental studies revealed that the induction time depends mostly on the stirring rate, the history of the solution, and the degree of supercooling, i.e., the driving force (Natarajan et al, 1994). The growth period follows the nucleation process. The growth process is better understood (Englezos, 1993). Gaarder (1993), also, studied the induction and growth period of 10 hydrates in CTMP effluent. The results showed that the effluent history and the size of the driving force affected the crystal growth rate in a manner similar to hydrate formation in pure water. This means that the induction time is inversely proportional to the magnitude of the driving force while the growth rate is proportional. In addition, the way a hydrate former contacts a solution can effect the overall rate of formation. De Graauw and Rutten (1970) compared hydrate kinetic data on evaporating liquid propane with kinetic data collected from direct contact of gaseous propane. They determined that mass transfer can limit the rate of hydrate growth when the crystallization occurs under a small driving force using gaseous propane. They were not able to determine if mass transfer had a significant impact on the kinetic rate when liquid propane evaporated inside the crystallizer. From a water recovery point of view it is better to grow hydrate crystals using a liquid hydrate former because larger crystals are formed (Jeffrey and Saenger, 1991). 2.2 Clathrate Hydrate Concentration In all previous attempts to exploit the phenomenon of hydrate concentration for desalination or concentration of organic solutions, the process necessitated the following unit operations: crystallization, crystal/concentrate separation, and crystal decomposition. The decomposition process does not affect the quality of the product water. The quality of the product water is only a function of crystallization and separation. Research should be towards growth of hydrate crystals and separation of crystals from the mother effluent with the objective to grow large crystals to minimize surface area that attracts impurities (Barduhn, 1968; Rautenbach and Seide, 1978). Barduhn (1968) found propane hydrate crystals being 25 to 50 microns in size and dendritic. Such crystals produced a bed of hydrates that is low in permeability. Rautenbach and 11 Seide (1978) improved the size and shape of the propane hydrate crystals by varying the stirring intensity and residence time. After the crystallization process, the hydrate crystals must be efficiently separated from the concentrated solution. Separation involves drainage of concentrate from the crystals, and washing the concentrate off the crystals. Knox et al. (1961) conducted preliminary washing tests on propane hydrates in brine solution. They employed a bench scale filtration unit. They found that the salt content of the washed crystals was under 500 ppm. Pavlov and Medvedev (1965) found the separation of propane hydrates and brine is best accomplished with a squeezing action. Sugi and Sato (1967) used a pressurized wash column operating at 1.5 bars to wash R-21 hydrates from brine. Two U.S. companies, Mason-Rust and Sweet Water, have each built a hydrate process pilot plant for desalination of sea water (Barduhn,1968). Mason-Rust used F-12 as the hydrating agent while Sweet Water used propane. Both of these companies felt washing was the major problem in the process. Mason-Rust tried drum filters and found them inadequate because of channeling. Sweet Water attempted to use a system of liquid-liquid cyclones to wash the hydrate crystals. The system of cyclones used liquid propane to initially displace the brine from the crystals. The liquid propane, and hydrate crystals flowed through a system of cyclones where it enriched each phase. A small water stream washes the final enriched hydrate stream. Sweet Water reported operational problems with this system of cyclone separation. Both Mason-Rust, and Sweet Water decided to switch to a pressurized wash column. Mason-Rust reported success with the wash column. They reported a product water content of 250 ppm. Sweet Water reported unsuccessful results because the washing produced a product water of 2000 ppm. Although both companies made progress, in the fall of 1968 their pilot plants were decommissioned because the Office of Saline Water (OSW) withdrew financial support. 12 Even though clathrate hydrate concentration did not succeed to create an industrial scale desalination plant, in the concentration of mechanical pulp mill effluents, it has a better chance to succeed because the recovered water requirements are not as stringent as with potable water. In addition the use of a single peice of equipment instead of three separate ones (crystallizer, separator, melter) is expected to improve the process because it avoids the pumping of slurry. 2.3 Research Objectives The scope of the work undertaken in the present study was to experimentally investigate the recovery of water for reuse from BCTMP effluents via clathrate hydrate formation. The specific objectives were the following: (a) To design an apparatus with the main characteristic that uses a single vessel as a crystallizer, wash column, and crystal melter; (b) Determine the effectiveness of separation of hydrate crystals from effluent concentrate by washing under pressure; and (c) Determine the quality of the recovered water from the melted crystals. Among mechanical pulp mill effluents, BCTMP contains the highest concentration of total solids, volatile solids, suspended solids, conductivity, total organic carbon, and inorganic carbon. In this work, we chose to work with BCTMP effluent. We conducted our experiments with propane. Even though R-12 performed well in desalination experiments, it is not a good choice because it has a high ozone depletion potential. Current and future regulations and environmental pressure will not permit their use. A suitable hydrate former should have the following characteristics: nontoxic, low ozone depletion potential, low global warming potential, low explosion and flammability. Furthermore, the hydrate former must have low hydrate formation pressures, and low solubility in water for easy 13 recovery. Based on the above criteria, we chose propane. Table 2.1 gives propane's characteristics. Table 2.1: Hydrate Former Characteristics of Propane (Gaarder, 1993) Characteristics Propane Maximum equilibrium T & P in pure water (quadruple point) 5.3 ° C at 0.542 MPa Ozone Depletion Potential (ODP) 0 Global Wanning Potential (GWP) -Explosion Potential Lower levels 2.1% Upper level 9.5% Flammability Auto-ignition Temp 432 ° C Toxicity Asphyxiant (No threshold given) Solubility in water at maximum temperature (auadruple point) 0.06 wt% Unit Cell Structure II Maximum number of water molecules per hvdrate former molecule in hydrate phase 17.95 Hydrate Density (g/cm3) 0.88 Heat of Hydrate Formation (KJ/mol hydrate) -134.2 Chapter 3: Experimental Section 14 In order to study the formation of propane hydrate crystals in a B C T M P effluent and their subsequent separation from the mother liquor we designed and constructed an apparatus, the unique feature of which is that hydrate formation and separation takes places in one vessel. This apparatus and the experimental procedure are described next. 3.1 Experimental Apparatus Figure 3.1 illustrates the overall schematic representation of the experimental apparatus. The main pieces of equipment are as follows: crystallizer /wash column, gas supply vessel, and wash water supply vessel. The gas storage vessel supplies the hydrating agent to the crystallizer/wash column. The quantity of gas used in each experiment is measured. The wash water supply vessel enables us to inject wash water into the crystallizer/wash column during the washing process. The crystallizer/wash column does two jobs. It forms the hydrate crystals while concentrating the mother liquor, and it washes the hydrate crystals. Control systems for temperature and pressure are necessary to keep these two variables at desired levels. 3.1.1 Crystallizer/Wash Column Figures 3.2 and 3.3 display the cross sectional views of the crystallizer. The views are 9 0 ° from each other. The crystallizer/wash column produces hydrate crystals in a semi-batch manner. Liquid effluent feeds into the crystallizer from the top by gravity. The gas storage vessel supplies the required quantity of propane. The hydrate former enters the crystallizer/wash column from the bottom through a 1/8" orifice opening. Just before the hydrate former enters the crystallizer, a shell and tube heat exchanger cools the propane to the operational temperature. Propane flows in the tube side of the exchanger. The tube side is constructed from a three feet long quarter inch translucent polyethylene tubing. SI 1 6 Figure 3.2: Crystallizer and Wash Column WASH WATER INLET 13/4" EFFLUENT INLET COOLANT INTLET O-RTNG COOLANT OUTLET INNER PLEXIGLASS TUBE OUTER PLEXIGLASS TUBE READY ROD IMPELLER SHAFT END CAP THERMOCOUPLE PROPANE INLET 13/4" 1 7 Figure 3 . 3 : Crystallizer and Wash Column .13/4' THERMOCOUPLE PROPANE OUTLET Mechanical Seal Assemby LIQUTD OUTLET WITH SCREEN -5 5/8* 13/4* 18 The tube is coiled inside a thirteen inch long by two inch diameter plexiglass cylindrical tube. A cooling jacket maintains the temperature inside the crystallizer constant. Copper constant thermocouples from Omega measure the temperature within the vessel. The accuracy of the temperature readings is reported to be ± 0.1 K. Thermocouples are situated at the top of the vessel, at the bottom of the crystallizer, at the wash water inlet, and at the propane inlet. A pressure control system can keep the pressure at a constant at the desired level. A Bourdon Heise test pressure gauge with a span of 300 psi and a pressure transducer measures the pressure inside the vessel. Brian engineering supplies this pressure gauge, and they report an accuracy for this pressure gauge to be ±0.5% of the span. Another Bourdon Heise pressure gauge from Marsh measures the pressure at the propane inlet. The accuracy for this gauge is ± 2 psi. A mechanical stirrer mixes the contents within the vessel. Two mechanical assemblies seal the mixing shaft. The top assembly is a housing that holds a bearing and an oil seal. The bottom assembly holds two bearings and an oil seal. The vessel was leak proof up to 105 psi. Two stirring geometries were implemented. Initially, we used three four inch diameter marine impellers that were equally spaced on a 30 inch long shaft. This configuration is referred to as stirring geometry I. Later, the mixing assembly was modified by increasing the length of impeller shaft by 4 inches, and adding two four inch diameter six blade Rushin turbine impellers. Again the impellers were equally spaced. This stirring configuration is referred to as stirring geometry II. The main body of the crystallizer/wash column is constructed from a 5 1/2" outside diameter acrylic tube with a wall thickness of 3/8". A 9" outside acrylic tube serves as cooling jacket for the crystallizer. Two end caps enclose both of these tubes together. Twelve ready rods fix the two end caps in place. O-rings seal the two end caps. The inside dimensions is 36 inch high, and 5 1/2" wide. The maximum volume is 13.6 L. Once the crystals have formed, 1 9 they must be separated and washed. The mother liquor drains from the crystallizer through two outlets. Each outlet is 3/4" diameter with screens containing a mesh size of 1 mm. Temperature profile inside the Crystallizer The temperature was measured in the vapour phase and the bulk liquid phase which consisted of a brine solution with a liquid propane layer on top of the brine solution. In the bulk liquid the temperature of the liquid phase was uniform when mixing was accomplished at 100 rpm with a torque of 50 oz-in. Both the brine solution and the liquid propane layer did not exhibit a temperature gradient at various heights. However, a temperature gradient was found in the vapour phase. At the beginning of the experiment when there is a large quantity of liquid propane, the gradient was small. The average temperature difference was +0.6 ° C from the top of the liquid propane layer to the tip of the thermocouple located in the vapour phase. The distance from the thermocouple to the liquid layer was 8 cm. The temperature gradient increased when no liquid propane was inside the crystallizer. In that case the temperature difference increased to +2 .5° C from the top of the brine solution to the tip of the thermocouple. The distance from the thermocouple to the brine solution was 6 cm because more brine solution was used in this experiment. Free convention currents cause by the boiling of liquid propane provides better mixing in the vapour phase. This results in a lower temperature gradient when liquid propane is present. 3.1.2 Gas Supply Vessel Figure 3.4 displays the gas storage vessel. The purpose of this vessel is to store liquid propane, and measure the quantity of propane used during each experiment. Each experiment requires propane to be injected into the crystallizer. The supply vessel is used to condense and store propane. Figure 3.4: Gas Suppply Vessel PROPANE r —H \ ID' 21 The operating temperature inside the storage vessel varied in the range of 0 to 5°C. Since propane is stored in the liquid state, the pressure must be above the saturation pressure. The operating pressure range in the supply vessel is between 0.62 and 1.02 MPa. With these restrictions, we can charge the supply vessel up to a maximum of 1025g. A copper constant thermocouple measures the temperature in the vessel. Omega supplies both the thermocouples and the meter, and they reported the accuracy for the combined system to be ±0.1 K A Bourdon Heise pressure gauge with a span of 200 psig measures the pressure inside the vessel. Brian engineering supplies this device, and they report an accuracy for this pressure gauge to be ± 2 . 0 % of the span. The supply vessel has two chambers, an upper and lower one. The upper chamber holds nitrogen gas, and the lower chamber stores the liquid propane. Pressure above the saturation point is required to maintain propane in the liquid phase. The storage vessel uses a piston to maintain a pressure above saturation. This piston is free to move up and down, and a rod guides the path of this piston. Two hydraulic rings form a seal between the upper and lower chamber. Pressurized nitrogen gas in the upper chamber keeps propane fluid above saturation pressure. Knowing the pressure and temperature within the vessel, published equilibrium data from Thomas and Harrison (1982), give the molar volume of the liquid. By measuring the displacement volume of the propane, it is possible to calculate the quantity of propane used in each experiment. The displacement length in the guiding rod multiplied by the cross sectional area is equal to the displacement volume. The error in reading the displacement length is ±0.1 cm, which equates to ± 4 g of liquid propane. The construction material for the storage vessel is a 316 stainless steel hollow cylindrical bar. The vessel's end caps are made from stainless steel circular plates. The piston is machined from a cylindrical aluminum bar. Twelve bolts fasten the two end caps in place. O-rings seal the piston and the end caps. The vessel is pressure tested to 1.2 MPa. 22 3.1.3 Wash Water Supply Vessel Figure 3.5 displays the water supply vessel. The purpose of this equipment is to supply cold wash water at pressures slightly greater than the operating pressure of the wash column. Water enters from the top of the vessel. The vessel uses nitrogen gas to pressurize the water to operating condition. Typical operating conditions are between 0.2 MPa and 1.0 MPa and always above that in the crystallizer. Nitrogen gas enters the vessel from the top. The increased pressure drives the wash water downwards in a manner analogous to a piston. The wash water exits at the bottom of the vessel and into the wash column. A cooling jacket maintains the temperature at a constant level. The construction material of the wash water column is a 316 stainless steel cylindrical hollow bar with a cooling jacket welded on to the vessel's body. A circular end plate is welded to the bottom of the vessel to enclose one end of the vessel. Another circular end plate encloses the other end. This plate is fastened on by six stainless steel bolts. An O-ring seals this plates. The vessel's inside diameter is 3.0", and the height is 10". The total volume of this vessel is 1158 ml. A copper constant thermocouple measures the temperature. A Bourdon pressure gauge measures the pressure inside the vessel. 3.1. 4 Temperature Control System Figure 3.1 also describes the cooling system. The system has two cooling circuits. Each circuit utilizes two Forma Scientific refrigeration units, model 2095, to cool each process vessel. One of the cooling bath, Bath 1, cools the crystallizer/wash column, and the heat exchanger. The other bath, Bath 2, cools the wash water supply column, and the gas supply vessel. It was observed that under extreme temperature fluctuations in the surroundings the baths can hold the temperature at a constant value within + 0.7 °C. 23 Figure 3.5: Wash Water Supply Vessel 24 Both cooling circuits are connected in parallel. Each process vessel has a cooling jacket. The jackets are tubular in shape. Fiber glass insulation is used for each jacket. The refrigerant is a 50/50 wt% solution of ethylene glycol and water. 3.1.5 Pressure Control System This work used liquid and vapour propane to grow hydrate crystals. When vapour propane was used to form clathrate crystals, pressure is always changing. A pressure control system was implemented in the crystallizer in order to conduct experiments under constant pressure. The experiments conducted with liquid propane did not require a pressure control system. Liquid propane was injected into the crystallizer in a batch manner. The pressure remained constant during the growth period because the liquid propane was at saturated conditions. Figure 3.6 depicts the control system. The pressure transducer sends a signal to the controller. The controller compares the signal to the set point value, and it opens or closes the control valve accordingly. The control valve is an E.V. A.-1 electronic valve from Badger Meter Incorporated. It is a digitally controlled valve. The response time is almost instantaneous. The controller is a PJD microprocessor based controller, model CN2001(*)-F2, from Omega. The controller's settings are as follows: proportional band =5%, Reset 0.25 repeats/min, and rate= 1.0 min. The controller could keep the pressure constant within ± 0.01 MPa. The pressure transducer is a standard diaphragm pressure transducer made by Omega. Its part model number isPX605. sz 26 3.2 Effluent Tests Chemical analysis of the various compounds in the effluent would be costly and unpractical. Testing the effluent for its physical characteristics would be more practical. Electrolyte content, carbon content, solid content, pH, and the amount of light transmittanced through the solution give the physical characteristics of the effluent. Each solution was weighed with a Mettler PE 16 scale. The accuracy of this scale is within 0.1 g. Conductivity is a measure of the amount of ions. In this work an Orion Model 160 Conductivity meter was used. The accuracy for this instrument is ± 0 . 5 % for solutions up to 199.99 mS. The meter is capable of automatically compensating for temperature variance. A Shimadza TOC 500 analyzed the carbon content in the solution. This analyzer measured the total organic carbon (TOC) and the total inorganic carbon (TIC) content. This test indicated the concentration of organic components in our samples. Solid content was obtained by measuring the total solid content (TS), fixed solids (FS), the volatile solids (VS), and the total suspended solids (TSS). After evaporation (at 105 °C) and drying, the TS value was obtained by gravimetric analysis. The residual weight of the effluent samples after thermo treatment at 550 ° C represents the FS value. The difference between the TS and the FS values gives the VS values. The TSS values were obtained from gravimetric analysis after the sample had been filtered through a selected filter medium. In Appendix A the procedures used in determining TS, VS, FS, and TSS are described. Testing for solids content helped evaluate the overall concentration of the samples. A FJACH (model DR/200) spectrometer determined the quantity of light absorbed through the each effluent sample at a wave lenth of 455 nm (maxium absorbance). The photometric reproducibility of this instrument is ±0 .005 . 27 A Cole Palmer OS669-20 pH meter was used to measure the pH. Buffer solutions of pH 4.01, 7.01, and 10.01 calibrated the meter. The reproducibility of this meter is ± 0.01. 3.3 Experimental Procedure We conducted experiments with salt solution and BCTMP mill effluent. For each type of solution, the procedure can be divided into four steps: (i) preliminary set up which is described in Appendix B; (ii) nucleation, (iii) crystal growth; and (iv) crystal separation. 3.3.1 Nucleation and Establishment of Hydrate-Propane-Water Equilibrium Two hydrate nucleation procedures were used. The first procedure was referred to as Nucleation Procedure One. For this procedure, the injection pressure was well below the saturation pressure. Thus vapour propane was used only to nucleate the hydrate crystals. This technique required a considerable amount of nucleation time. The second procedure was referred to as Nucleation Procedure Two. Propane was injected into the crystallizer at saturated conditions until a layer of liquid propane sat on top of the feed solution. A distinctive interface was observed between the feed solution and the liquid propane layer. Approximately one quarter of an inch of liquid propane was allowed to accumulate inside the crystallizer. This translated to approximately 50 g of liquid propane. By quickly venting propane gas out of the system, a rapid decrease in temperature occurred in the gas phase due to the heat of vaporization. As a result, this temperature drop creates a considerable driving force, AT, which aids in the nucleation process. Figure 3.7 illustrates the cooling process at the vapour and liquid interface. As seen in the figure, the nucleation process moves along the propane vapour and liquid equilibrium line (line I). Hydrate crystals quickly nucleated using this procedure. Once hydrate crystals were formed, the residue liquid propane was vented until the pressure dropped below saturated pressure. 29 After venting out the residual liquid propane, the gas phase was allowed to stabilize to the experimental temperature (point A). This process operated along line II. In each experiment, the quantity of hydrate crystals after the nucleation process was quite different. To standardize the starting point of these experiments, a common reference point was needed. Our experiments chose the starting point at the hydrate-propane-water incipient equilibrium condition. Propane gas was vented out of the vessel until it is observed that a small amount of hydrate crystals remained on top of the liquid solution (point B). This process operated along line III. As the pressured dropped below the hydrate equilibrium pressure, hydrate crystals decomposed, releasing propane gas that elevated the pressure. This process operated along line IV. The pressure was allowed to stabilize to the hydrate-vapour propane-feed solution (2.5 wt% NaCl or BCTMP effluent) equilibrium value (point C). The system was monitored for five hours. If the pressure and temperature readings were constant (within ±0 . IK, and ±0.004 MPa), the system was assumed to be at its hydrate-water-vapour equilibrium. If the conditions were not stable, the system was monitored until it was stable for a five hour period. 3.3.2 Crystal Growth Before the crystal growth stage can begin, the quantity of propane inside the gas supply vessel must be measured. The supply vessel was compressed to an approximate pressure of 0.24 MPa above the saturation pressure corresponding to the temperature inside the gas supply vessel. This excess pressure was arbitrarily chosen to be adequately above saturated pressure. This procedure ensures that all the propane in the supply vessel was in the liquid state. By recording the temperature, pressure, and the height of the piston, all the state variables are known. The quantity of propane inside the supply vessel before propane is injected into the supply vessel can be calculated. Appendix C gives a sample calculation for calculating the quantity of propane inside the supply vessel. 30 The experiments were conducted at a pressure, Pexp> above the hydrate-propane-water equilibrium. In order to reach that pressure, the crystallizer has to be charged with propane. During the injection of propane into the crystallizer, we made sure that the pressure in the supply vessel was always approximately 0.24 MPa above saturation. This was achieved by pressurizing with nitrogen. The growth stage begins with injection of propane into the crystallizer. In this work, the crystals were grown from either vapour propane or liquid propane. The crystals were allowed to grow at a constant experimental pressure, Pexp t n a t gives a driving force equal to P e X p-P e q. Both the temperature and pressure were monitored. During this period, the crystals were observed for their rate of growth and their morphology. At the end of the crystal growth period, the solution inside the crystallizer is more concentrated than the feed solution thus shifting the hydrate-propane-water equilibrium. The new equilibrium value was also measured by following a similar procedure we used after nucleation. Instead of throttling the gas to point B, the pressure was dropped to the hydrate-propane-water equilibrium, point D. This procedure ensures that only a minimum amount of hydrate crystals is decomposed. This new equilibrium pressure at the experimental temperature was higher than the previous equilibrium pressure that was found after nucleation. At the end of the injection period, by recording the height of the piston, the temperature, and the pressure, all the state variables are known. Thus the quantity of propane inside the supply vessel after propane was injected into the crystallizer is known. The quantity of propane that is dispensed to the crystallizer can be calculated though subtraction of the quantity of propane inside the supply vessel before and after the injection stage. The procedure for crystal growth was different for crystals that were grown in the vapour propane and the liquid propane. The following sections describe in detail these differences. 31 Crystal Growth with Vapour Propane One experiment used vapour propane to grow the hydrate crystals. We wanted to determine if there were any advantages for growing clathrate crystals under gaseous conditions. In this experiment, the crystallizer was supplied continuously with vapour propane at below saturated conditions. Vapour propane was bubbled from the bottom of the crystallizer through the feed solution. The solution was mixed using impeller geometry I. The crystals were allowed to grow until the target quantity of propane was consumed. By measuring the displacement height of the piston in gas supply vessel we could determine the quantity of propane that was dispensed into the crystallizer at all times. Once the target quantity of propane was dispensed into the crystallizer, the valves leading to the supply vessel was turned off. The crystal growth period was considered to be from the beginning of the growth period to the time the valves were shut off. The propane injection period was equal to the crystal growth period. The system was driven to its equilibrium condition. Crystal Growth with Liquid Propane The rest of the experiments used liquid propane to grow the crystals. Liquid propane was injected into the crystalizer in a batch wise manner. At the end of the injection period, a thick layer of liquid propane formed on top of the feed solution. Typically, the injection time required 15 minutes. The height of the liquid propane layer varied. It depended upon the target quantity of propane we chose for that particular.experiment. As the crystals grew inside the crystallizer, the liquid propane was consumed, and the liquid propane layer slowly decreased in size until the liquid layer disappeared. At this point the pressure was still at saturated conditions. To ensure that all the liquid propane was consumed in the crystallizer, the crystallization process was allowed to continue until the pressure dropped below 0.07 MPa from the saturated pressure. The time from the beginning of the crystal growth period to this 32 end point is considered to be the growth period for the hydrate crystals. Next, the system was driven to its equilibrium condition. 3.3.3 Crystal Separation This work conducted three types of separations: a) no washing; b) concentrate displacement with liquid propane; and c) washing with water. These separation processes are illustrated in Figures 3.8, 3.9, and3.10. Each separation process began with draining of the concentrate from the two liquid outlets after the crystallization process was over. While the concentrate was draining, the mixing assembly was turned off. The pressure was maintained at the hydrate-propane-water equilibrium pressure by the pressure controller. The set point of the controller was adjusted to the equilibrium pressure. The draining procedure was set at 15 minutes, with the exception of runs SI and S2. These runs used a draining period of 120 minutes. The draining procedure is conducted in the following manner. The two liquid outlets valves were opened and the concentrated liquid poured out. When propane gas begins to flow out from the two liquid outlet valves, the valves were shut. During the draining period the valves were periodically opened to extract more concentrate until we reached the end of the draining period. Once the crystals were drained of its concentrate the separation procedure began to deviate which depended on the type of experiment. The following sections describe the differences in the separation technique. No Washing Experiments In these experiments, we wanted to determine the quality of the recovered water produced in the clathrate concentration process without any additional separation stage such as washing with water and displacement with liquid propane. It is important to note that in the first two experiments the crystals were drained as usual but the system was allowed to remain idle for two hours. The additional draining period produced a solution that was combined with ee o o o CP CD o o p p CD a CD ^S3 p* 8 CD CD t o CD >-r| CQ CD S 0 B 0 X o a r 1 CD P-0 1 CD CD CD O 0 1 CD f 36 the concentrate. The quantity of concentrate that was produced was weighed, and its characteristics were analyzed by the methods that was discussed in section 3.2 effluent tests. After the draining process, the crystals were decomposed. Gas was vented out of the system to atmospheric conditions. Typically, the crystals required eight hours to decompose. The recovered water was collected and weighted. Again a sample of the recovered water was taken, and it was later analyzed for its characteristics. Experiments with Liquid Propane Displacement In these experiments, we wanted to improve the separation by using liquid propane to displace the concentrate from the bed of hydrate crystals. These crystals are known to agglomerate, and during this process concentrate can be trapped forming occlusions. We want to remove the occluded effluent. We propose to use liquid propane to displace the concentrate from these cavities. Liquid propane and the concentrate are immiscible, and they will readily separate into two distinctive phases. Since liquid propane is less dense than the concentrate, liquid propane will float on top of the concentrate. If we poured liquid propane on top of the bed crystals, the liquid propane can push the concentrate from these cavities. The concentrate will flow downwards, and form a layer of concentrate. This layer of concentrate can be removed by an additional draining process. By removing this additional quantity of concentrate the quality of recovered water should be improved. In preparation for this separation process, a bed of crystals must be deposited at the bottom of the crystallizer. We used two different procedures to form the bed of crystals. Either bed formation procedure A or B was used to prepare the bed of crystals because of different outcome that was experienced during the crystallization process. The first outcome was that most of the crystals formed in suspension. Only a small amount of crystals adhered to the crystallizer's wall. As a result during the draining process, the crystals easily fell to the bottom of the crystallizer and it formed a uniform bed of crystals. Nothing was done to aid in the 37 process of forming a bed of crystals. This procedure was known as bed formation procedure A. The second outcome was that some of the crystals formed in suspension, and some of the crystal adhered to the crystallizer's wall forming a hydrate plug. This plug formed above the top impeller, and it adhered firmly to the crystallizer's wall. As a result during the draining process, some of the crystals fell to the bottom of the crystallizer forming a bed of crystals, but the hydrate plug remained adhering to the crystallizer's wall. To form a uniform bed, this plug was broken up. Coolant was evacuated from the heat exchanging surface (cooling jacket). As a result, local melting was induced at the interface where the hydrate plug was in contact with the heat exchanging surface. Heat from the surroundings melted the sides of the hydrate plug enabling it to slide downwards on to the rotating impeller, and it easily broke up this hydrate plug. The crystals fell to the bottom of the crystallizer. Once the new bed had formed, coolant was re-injected back into the heat exchanging surface (cooling jacket). This procedure was referred to as bed formation procedure B. Once the bed of crystal was formed (Bed A or B), liquid propane was poured on top of the bed of crystals via the wash water inlet. Note that the gas supply vessel must be filled with liquid propane and its quantity must be measured prior to injection. Appendix B gives the preliminary set up procedure. Liquid propane was injected at saturated pressure and at the experimental temperature. Typically, the time required to inject the target quantity of propane was thirty minutes. During this process, the pressure inside the supply vessel was kept constant by supplying fresh nitrogen to the vessel. Once the target quantity of propane was injected into the crystallizer, a layer of displaced concentrate was observed to be beneath the liquid propane layer. This displaced concentrate was drained with a procedure that was identical to the procedure used for draining the concentrate. Following this draining process, the crystals were melted in the same fashion as the procedure in the no washing experiments. In each experiment, the concentrate, displaced concentrate, and the recovered water was collected and analyzed. Washing with Water Experiments 38 In these experiments we wanted to examine how well water can displace the concentrate from the crystals. This process is similar to the propane displacement process as shown in Figure 3.10. Again a bed of crystals must be formed, and the procedure for forming the crystal bed was identical to the procedure used for the propane displacement experiments. The procedure began to differ after the formation of the crystal bed. Wash water was injected into the crystallizer from the wash water supply vessel. Appendix B gives the set up procedure for the wash water supply vessel. We used de-ionized water as the wash water, but in one experiment we added a non-ionic dye to the wash water to observe the washing process. By throttling nitrogen manually into the wash water supply vessel, the increased pressure inside the vessel forced the wash water into the crystallizer. When the wash water vessel was empty, the valves connecting to the wash water supply vessel were turned off. During this process the pressure was maintained at the hydrate-propane-water equilibrium at the experimental temperature, T e x p , by manually throttling propane gas out of the system. After water was injected into the crystallizer, the water displaced some of the concentrate from the crystals. This solution was drained from the crystallizer. The procedure for draining this solution was identical to the procedure used to drain the concentrate. This solution was referred to as spent wash water. Again, the crystals were decomposed. The concentrate, the spent wash water, and the recovered water were weighed and analyzed. 39 Chapter 4: Results Because BCTMP effluent is a dark colour liquid that contains a variety of salts and organic compounds, it is difficult to visually observe the clathrate crystallization and separation process. As a result, after experiment E l with effluent we performed experiments with NaCl solutions at 2.5 wt%. In these experiments we can easily observe the clathrate concentration process and the extent of concentration could be measured with only one test, conductivity. 4.1 Salt Solution Experiments Prior to the clathrate hydrate concentration test, the freezing point of the sodium chloride solution was measured. It was found that a solution with a conductivity of 42 000 u,S gave a freezing point depression of 1.40 ° C . The freezing point depression, reported in the 69 t n Ed of The Handbook of Chemistry and Physics, is 1.486 ° C for a 2.5 wt% sodium chloride solution. Nine sets of experiments were conducted with salt solution at 2.5 wt%. All these experiments were conducted with nucleation procedure two, and they all used liquid propane to grow the crystals. Three experiments (SI, S2, and S3) were conducted with no washing. Three experiments (S3, S4, and S9) were performed with water as the washing agent. Three experiments (S5, S7, and S8) used liquid propane to displace the concentrate from the crystals. Table 4.1 describes the operating conditions for all the runs. During these experiments, the experimental temperature fluctuated slightly because of changes in the ambient temperature. Since these experiments used saturated liquid propane, the experimental pressure also fluctuated due to the changes in the ambient temperature. These fluctuations were reported as Tf, and Pf for temperature and pressure, respectively (i.e. T e X p±Tf, and P e X p ± P f ) 4.1.1 Experiments with No Washing 41 We made qualitative observations during nucleation, crystal growth, and separation process. At the end of these experiments, we measured the quantity and conductivity of the recovered water and the concentrate. Nucleation These experiments used nucleation procedure two. Hydrate crystals quickly formed with this procedure. Appendix D gives the results of the hydrate-propane-salt water equilibrium values that were obtained after nucleation was established. These values fell on the same hydrate-propane-water equilibrium line that was produced in Kubota's work on a 2.5 wt% NaCl solution. Crystal Growth As the crystallizer filled up with liquid propane, some of the hydrate crystals that were present accumulated on the interface between the liquid propane and the saline solution. The rest of the crystals were kept in suspension by the mechanical agitation. Typically, the time required to inject the desired quantity of propane into the crystallizer was 15 minutes. Figure 4.1 describes the system at this particular time, the beginning of the growth period. During this process, the crystallizer was well mixed. It was observed that the mixing action created a large vortex. The size of the vortex depended on the type of impeller geometry and the rotational speed used in each experiment. At a rotation speed of 200 rpm, Impeller geometry I created a vortex that occupied 3/4 of the width of the crystallizer with a height of eight centimeters. 42 Figure 4.1: Initial Condition during the Crystal Growth Period 43 At the same rotational speed, Impeller geometry II generated a vortex that occupied the entire width of the crystallizer with a height of ten centimeters. Impeller geometry JJ produced much better mixing characteristics, which affected the way the crystals were grown. Using impeller geometry I (experiments SI and S2), some of the crystals appeared to accumulate at the interface where liquid propane was in contact with the saline solution. Under stirring, the hydrate crystals initially remained in the free suspension, but as the quantity of crystal grew with time, some of the crystals began to agglomerate and adhere to the crystallizer wall where mixing is not rapid. With time some of the crystals formed a thin concentric wall around the crystallizer about 3 inches in height at the liquid propane and brine interface. As time progressed, the wall became thicker and ultimately it formed a hydrate plug that adhered to the crystallizer wall. Photograph 4.1 displays the hydrate plug in the brine solution at the end of the crystal growth period. Using impeller geometry II (experiment S6), no hydrate plug was formed. This modification increased the torque to the fluid. At the beginning of the growth period, when the impeller speed was monitored at 200 rpm, the torque produced by the impeller assembly was 98 oz-in. The torque value was about three times larger than the torque value produced by the impeller geometry I at the same rotational speed. The increase in torque reflected in a better mixing characteristic, which affected how the crystals were grown. Using impeller geometry II. no hydrate plug was formed but instead a small ring of crystals formed at the end of the growth period. The ring was approximately 3 mm high. It adhered to the crystallizer's wall at the brine and vapour interface. The mixing was sufficient to keep the hydrate crystals from agglomerating and forming the plug. Most of the hydrate crystals were formed in suspension. As hydrate crystals were formed, propane was consumed. Eventually, the pressure dropped below saturated values, and this signaled the end of the crystallization process. Photograph 4.1: Hydrate Plug in Brine Solution 45 As prescribed in chapter three, the system was driven to its equilibrium value. Again, Appendix D gives this hydrate-propane-salt water equilibrium value. These equilibrium values were slightly higher than Kubota's data for a 2.5 wt% NaCl solution because our solution was concentrated by the hydrate concentration process that further shifted the equilibrium to the left. After equilibrium condition was established, the experiment proceeded to the separation stage. Crystal Separation During the experiments with no washing, two different observations were made. When no hydrate plug was present (experiment S6), the crystal separating process began with draining the concentrate from the brine solution. This process required fifteen minutes to drain all of the concentrate. The crystals easily fell to the bottom of the crystallizer forming a bed of crystals. Photograph 4.2 shows the bed of crystals. When a hydrate plug was present after the crystallization stage, we observed that some of the hydrate crystals in suspension quickly floated upwards to the hydrate plug when the agitation had stopped. Photograph 4.3 depicts the floating process. Eventually, some of the free suspending hydrates would agglomerate together, and adhere to the existing plug. When the hydrate crystals finished adhering to the plug, brine was drained from the crystals. From visual inspection, the hydrate crystals had the appearance of wet snow, and the plug contained fine cracks. The concentrate was collected and later analyzed. Again the draining process required fifteen minutes. For experiments S i and S2, the crystals were allowed to sit and drain for an additional 105 minutes. As a result of this extra drainage, 342 and 287 g of solution were removed during runs SI and S2 respectively. This solution was combined with the concentrate. Next, the hydrate crystals were melted, and the recovered water was collected and later analyzed. Table 4.2 gives the results for separation with no washing. Figure 4.2 illustrates these same results. 46 Photograph 4.2: Formation of a Bed of Crystals (Bed Formation Procedure A) 48 49 Figure 4.2 shows that the conductivity values and quantity of feed were similar for all the experiments with no washing. Examining each concentrate, one notices that the concentrate's conductivity is on average 6% higher than the feed solution. The quantity of concentrate depended on the amount of hydrate former used in the experiment. The recovered water decreased in conductivity from the feed and the amount of recovered water depended on the quantity of hydrate formed. The conductivity of the recovered water was still high. Run SI produced recovered water with the lowest conductivity. Table 4.2 Results for Separation with No Washing Run Measurement Feed Concentrate Recovered Water S1 Mass (g) 11000 10075 900 Conductivity. (fjS) 42400 44800 23600 S2 Mass (g) 10550 9330 1200 Conductivity. (JJS) 42000 44000 26400 S6 Mass (g) 9981 7097 2789 Conductivity. (//S) 40800 44200 35500 4.1.2 Experiments with Displacement with Propane Runs S5, S7, and S8 used liquid propane to displace concentrate from the bed of crystals. Table 4.1 gives the specific conditions of each run. These experiments were observed to behave in an identical manner as the previous experiments with no washing up to the draining process. The characteristics of the hydrate crystals were identical to those observed during the "no washing" experiments. Appendix D reports the hydrate-propane-salt water equilibrium 50 values before and after crystal growth period. Again a hydrate plug was formed when impeller geometry I was used. These experiments began to deviate from the no washing experiments at the draining process. For all these experiments, the crystals were drained for 15 minutes. After the draining process, these experiments used either bed formation procedure A or B to form a bed of hydrate crystals in the preparation of propane displacement process. Using procedure A (Runs S7 & S8) to form the bed of crystals, the bed looked like a bed of wet snow that was well distributed at the bottom of the crystallizer. However, using procedure B (Run S5) to form the bed of crystals produced a bed that looked as if wet snow had been dropped into the crystallizer column. The bed appeared to contain clusters of hydrate crystals with large void spaces. After the bed of crystals was formed, liquid propane was injected into the crystallizer. The quantity of propane that was injected into the crystallizer was enough to submerge the crystals entirely. Figure 4.3 illustrates the displacement process. The displaced concentrate and recovered water were collected and analyzed in the manner that was discussed in chapter three. Table 4.3 reports the results of these experiments with propane displacement. Figure 4.4 represents these results on a bar graph. The concentrate's conductivity value is on average 10% higher than the feed solution. Again, the quantity of concentrate varied with the amount of hydrate former used in each experiment. Using propane as a displacement agent, this separation process could extract on average one kilogram of displaced concentrate. The conductivity of the displaced concentrate was on average 7% lower than the concentrate. Again, the recovered water decreased in conductivity from the feed. The amount of recovered water depended on the quantity of hydrate former. Run S7 had the lowest conductivity measurement for the recovered water. 51 Figure 4.3: Brine Displacement with Liquid Propane 52 a . o a o U cd C N 53 Comparing this run with experiment SI (lowest conductivity in the recovered water for the no washing experiments), the recovered water's conductivity was 4% higher than run SI, but it produced 203% more recovered water. This evidence suggests propane displacement improves the separation between crystal and the concentrate. Table 4.3 Results for Propane Displacement Run Measurement Feed Concentrate Propane used in the Displacement Process Displaced Concentrate Recovered Water S5 Quantity(g) 10300 6064 1703 1023 2700 Conductivity OyS) 41700 46900 0 40200 29900 S7 Quantity(g) 9973 6964 665 989 1824 Conductivity OuS) 40500 43900 0 41800 24600 S8 Quantity(g) 9598 6251 650 1075 2258 Conductivity (jt/S) 40850 45800 0 44700 29052 4.1.3 Experiments with Washing with Water Runs S3, S4, and S9 used water to wash the clathrate crystals. Table 4.1 gives the operating conditions of each run. These experiments behaved in an identical manner as the experiments with propane displacement up to the crystal bed forming process. Again the crystals had the same appearance. The hydrate-propane-salt water equilibrium values measured before and after crystal growth stage had the same tendency as the equilibrium results from the "propane displacement experiments". Appendix D gives these results. Again, by using impeller geometry II. the formation of the hydrate plug was eliminated. These experiments deviated from the previous experiments when water was injected into the crystallizer instead of liquid propane. Water flowed downwards on top the bed of crystals. 54 One experiment, run S9, used a dye solution to wash the crystals. It was observed from this experiment that poor contacting may exist in the washing process. Photograph 4.4 shows a portion of the hydrate bed after wash water was in contact with the hydrate bed. Notice the dye adhering to the surface of agglomerated crystal. The center of the agglomerate doesn't contain much dye. In each experiment, the spent wash water and the recovered water were collected and analyzed. Table 4.4 reports the results from the washing with water experiment. These results are illustrated graphically on Figure 4.5. Table 4.4 Results for Experiments with Water Washing Run Feed Concentrate Wash Water Spent Wash Water Recovered Water S3 Mass(g) 10122 8069.6 3600 3450 2037 Conductivity. OuS) 41700 45500 1 12298 11190 S4 Mass(g) 10300 5839.2 7000 6938 3284 Conductivity. OuS) 42500 47800 1 14870 7460 S9 Mass(g) 9432 5312 3803 3712 3897 Conductivity. (pS) 40850 45900 23.4 15200 23600 Again the initial feed conditions, conductivity and mass, were similar. The concentrate's conductivity value is on average 10% higher than the feed solution. Again, the quantity of concentrate varied with the amount of hydrate former used in the experiment. Large quantities of wash water were used to wash the crystals. In each case the spent wash water was collected, and it was found to be approximately equal to the amount of wash water used to wash the crystals. The conductivity for the spent wash water was 14.1 mS on average. This result demonstrated that dissolved salts can be further removed from the crystal by using wash water. 55 Photograph 4.4: A Bed of Crystals after Washing with a Non Ionic Dye Solution 56 57 However, using large quantities of wash water could not lower the conductivity of the recovered water to desirable levels (5.0 mS). Run S4 produced recovered water with the lowest conductivity measurement, 7.46 mS, but it required 7.0 Kg of wash water. This run produced 3.284 Kg of recovered water. Washing the crystals in this manner does not seem to be an effective method of separating the brine from the crystal. 58 4.2 B C T M P Effluent Five experiments were conducted with BCTMP effluents. The draining time for all these experiments was 15 minutes. Experiment E l through E3 were performed with no washing. In Run E4, the crystals were displaced with liquid propane while in Run E5 water was used to wash the crystals. Table 4.5 summarizes the operating conditions for these experiments. For each experiment, the hydrate-propane-water (effluent) equilibrium data for before and after the crystal growth periods were determined. Appendix D reports these results. 4.2.1 No Washing Experiment E l Unlike all previous experiments, this run used Nucleation Procedure One to nucleate the hydrate crystals. Crystal growth was carried out using vapour propane. Actually, this run was the first experiment we conducted. After this run, we decided to do some experiments with salt solutions for better visual observation. In addition, we decided to work with liquid propane during the growth period. Run E l required 7 days to nucleate a small amount of crystals. During the growth stage, the crystals formed at an experimental pressure of 0.465 MPa, that is well under the saturated pressure, P s a t = 0.490 MPa. The crystals grew slowly inside the crystallizer. It took 14 days to consume 236 g of propane. This translates to an average propane consumption rate of 0.702 grams per hour. The crystals grew on the surface of the crystallizer wall and in the bulk of the solution. With time some of the crystals adhered onto the crystallizer's wall and formed a concentric ring with the crystallizer's wall. As time progressed a hydrate plug had formed above the top impeller blade. Photo graph 4.5 illustrates this phenomenon. The plug was not homogeneous. The bulk of the plug appears to be a jelly-like material. Between this material, a non jelly-like "whitish" layered structure had formed. These hydrates appear to be similar in structure to the deep sea layered hydrates found in the Blake-Bahama Ridge (Sloan, 1989). 5 9 C o C o U 00 .S cd U i <U OL, O in •vr _u X) cd c o a, 00 CO o o o P-c 8 60 a D E u E o o 1 1 2 T3 O •c o « G 58 *a 2 K o bo c •o c « •a a, 0 1 o H o H 11 2 -VO t o CN 1 § vO vn vO r-cs UJ 8P ca O <s o ON o o VO o o vn UJ vO cs U4 c E 8 cs 1 s o o 5 VTl in CO cs o o o o m o o vn n vn VO r-cs vn UJ Photograph 4.5: Hydrate Plug Formed in Experiment E l 61 The standard procedure was used to drain the system, and the concentrate was collected. The crystals were allowed to decompose. During this decomposition period, it was observed that the jelly-like hydrate layers and "whitish" hydrate layers decomposed in a similar fashion as normal hydrates. Bubbling can be observed at the surface of the hydrate crystals due to the release of propane gas. The recovered water was also collected for analysis. Table 4.6 represents the results from this run. Figure 4.6 represents these results on a bar and a pie chart. The crystallizer was fed with 9877 g of effluent. The clathrate concentration process produced 6203 g of concentrate and 3604 g of recovered water. The concentrate's conductivity had increased to 5 % from the feed while the recovered water's conductivity had decreased to 19 % from the feed. In general TS, VS, and FS followed the same trend as conductivity. However, pH, absorbence, and TSS did not follow the same trend as conductivity. Table 4.6 Results for Run E l Feed Concentrate Recovered Water pH 7.78 8.37 8.47 Absorbence 5.52 3.80 5.72 Conductivity. (uS) 24800 26000 20100 TS(g/L) 40.97 43.88 33.11 TSS (g/L) 2.963 2.403 3.731 VS (g/L) 14.11 15.11 12.07 FS(g/L) 26.86 28.77 21.03 Mass (g) 9877 6203 3604 The pH of the feed, concentrate and recovered water did not vary significantly. The concentrate's absorbence decreased on average 31 % from the feed while the absorbence of the recovered water increased on average 3% from the feed. This increase in absorbence found in the recovered water can be attributed to the increase in suspended solids. 62 63 The TSS tests verified that the concentration process produces a recovered water with more suspended solids. The TSS on the concentrate decreased by 19 % from the feed, and the TSS on the recovered water increased by 21%. 4.2.2 No Washing Experiment E2 Run E2 was conducted in a similar manner as Run E l except hydrate crystals were formed with liquid propane, and impeller geometry II was used. The crystal nucleated and grew in a similar fashion as experiment S6 that used impeller geometry II. No hydrate plug was formed. Most of the crystal grew in suspension. This run proceeded at a much faster rate than the previous effluent experiment. It consumed 423 g of propane in 15.5 hours. This translates to an average propane consumption rate of 27.3 grams per hour. The crystals had a uniform granular appearance that was quite different from Run E l . This run used the standard draining procedure, which produced a uniformly distributed bed of crystals. After the draining process was completed, the crystals were decomposed. Table 4.7 describes the result from this run. Figure 4.7 illustrates these results. Table 4.7 Results for E2 Feed Concentrate Recovered Water pH 8.54 7.89 7.89 Absorbence 5.06 4.20 5.32 Conductivity. (p.S) 24900 30400 16200 TS (g/L) 41.27 51.31 26.98 TSS (g/L) 2.902 2.413 3.611 VS (g/L) 14.54 17.48 9.55 FS (g/L) 26.73 33.83 17.42 Mass (g) 9568 5458 4029 65 This run did not form a jelly-liked hydrate, which occludes large quantities of concentrate. As a result, it did a better job at concentrating the feed, and producing cleaner recovered water than run E l . The results for this run show that the conductivity of the recovered water dropped to 35% from the feed, and the concentrate's conductivity increased 18% from the feed. Again TS, VS, and FS seem to increase or decrease in proportion to conductivity. The concentrate's conductivity, TS, VS, and FS values were much higher then the feed solution. The recovered water's conductivity, TS, VS, and FS values yielded much lower then the feed solution. Again, pH, absorbence, and TSS did not correlate with conductivity. 4.2.3 No Washing Experiment E3 Run E3 was performed in a similar fashion as Run E2. In addition, the effluent samples were analyzed for TIC and TOC. In this run a new batch of effluent was used. The characteristics of this batch contained fewer impurities than the feed effluent from Runs E l and E2. The conductivity of this batch was five times lower than the previous batch of effluent (Run E l & E2). This run consumed 412 grams of propane in 6.5 hours. This translates to an average propane consumption rate of 63.4 grams per hour. No visual differences were observed between Runs E2 and E3. After draining, the crystals were decomposed. Table 4.8 reports the results of this experiment. Figure 4.8 shows this information on a bar and a pie chart. The conductivity of the recovered water decreased 23% from the feed while the concentrate's conductivity increase 19% from the feed. Again, TS, VS, and FS values increased or decreased in proportion to conductivity. In addition TIC, and TOC also followed the same trend as conductivity. Again, TSS, absorbence, and pH did not correlate with conductivity. 66 67 Table 4.8 Results for E3 Feed Concentrate Recovered Water pH 7.05 7.74 7.03 Absorbence 2.92 1.32 4.56 Conductivity. (uS) 4460 5503 3456 TS(g/L) 6.54 7.52 6.11 TSS (g/L) 2.46 2.51 4.75 VS(g/L) 2.44 2.80 2.15 FS(g/L) 4.08 4.72 3.96 TIC (g/L) 0.083 0.101 0.053 TOC(g/L) 1.810 2.013 L402 Mass (g) 9514 5965 3549 4.2.4 Displacement with Propane Experiment E4 In this experiment, the crystals were grown in a similar manner as experiments E2, and E3. This run consumed 420g of propane in 6 hours. This translates to an average propane consumption rate of 70.0 grams per hour. The crystals grew in suspension. This run used the same batch of effluent as run E3. This run used the standard draining procedure. After drainage, 665 g of liquid propane was used to displace concentrate from the crystals. The displacement process worked in an identical manner as the experiments with salt solution that used impeller geometry II (experiment S7 & S8). Using this quantity of propane all of the crystals were immersed in liquid propane. The displacement process produced 510 g of displaced concentrate. Table 4.9 and Figure 4.9 summarizes the result for this run. 68 O cn p w • r—( o TJ C o O CM o Q S I c o c6 0 o c o o o LL CO C O C O CM O 00 CO CM 69 Table 4.9 Results from run E4 Feed Concentrate Displaced Concentrate Recovered Water pH 7.40 7.51 8.06 7.08 Absorbence 2.91 1.48 1.23 5.21 Conductivity. (uS) 4480 5360 5140 2460 TS(g/L) 6.60 7.39 7.48 5.13 TSS (g/L) 2.13 0.48 0.45 5.08 VS(g/L) 2.85 3.17 3.24 2.78 FS(g/L) 3.75 4.22 4.24 2.35 TIC(g/L) 0.090 0.114 0.093 0.052 TOC(g/L) 1.89 2.117 2.029 1.373 Mass (g) 9652 6338 510 2664 The concentrate's conductivity increased 16% from the feed while the recovered water's conductivity decreased 45% from the feed. Again, TS, VS, FS, TIC, and TOC followed the same trend as conductivity. The amount of increase is proportional to conductivity. Again, pH, TSS, and absorbence did not follow the same pattern as conductivity. The displaced concentrate's characteristics were similar to the concentrate. By extracting this displaced concentrate, run E4 produced a cleaner recovered water than run E3. 4.2.5 Washing with Water Experiment E5 Run E5 was conducted in a similar manner to Run E4 except water was used to displace the concentrate instead of propane. The washing process was observed to be identical to experiment S9 that used impeller geometry II. This run used the same batch of effluent that was used in experiments E3, and E4. Run E5 also consumed 423 grams of propane in 5.5 hr. This translates to an average propane consumption rate of 76.9 grams per hour. This run used 2.000 70 Kg of wash water to displace the concentrate. Table 4.10 and Figure 4.10 gives the results of this run. Table 4.10: Results for Run E5 Feed Concentrate Spent Wash Water Recovered Water pH 6.69 7.81 7.13 7.02 Absorbence 2.94 1.28 2.41 3.66 Conductivity. (pS) 4440 4750 2830 3410 TS(g/L) 6.953 7.789 4.152 5.592 TSS (g/L) 3.104 0.602 1.802 5.063 VS(g/L) 2.318 2.813 1.779 2.172 FS(g/L) 4.635 4.976 2.372 3.42 TIC (g/L) 0.074 0.112 0.012 0.055 TOC(g/L) 1.744 1.996 1.005 1.545 Mass(g) 9679 6100 2312 3245 The recovered water's conductivity decreased 23% from the feed while the concentrate's conductivity increased 7% from the feed. Similar to the other experiments, TS, VS, FS, TIC, and T O C increased or decreased in proportion to conductivity. pH, absorbence, and TSS did not correlate to conductivity. This run produced 2.312 Kg of spent wash water, which contained a large quantity of impurities. One explanation is that a significant amount of concentrate remained with the crystals after the drainage procedure, and the addition of wash water displaced some of the concentrate from the crystals. As a result, both the recovered water and the spent wash water were also high in impurities. 71 .—I . to ~ j _ , eg <u a -o s t/3 — o *o .t-i u> m a ? e «J oo. 72 Chapter 5: Discussion Growth of hydrate crystals from liquid propane instead of gaseous is preferable because it allows for much faster growth as indicated by the average propane consumption rate (mass of hydrate former divided by crystal growth period). In Table 5.1, comparing runs SI and E l , switching from vapor to liquid propane increased the rate by 2.7 times. The impact of impeller geometry and driving force on the growth rate is evident by comparing experiments SI, S2, S4, S6 and E l . Table 5.1 shows the relevant information. Using impeller geometry JJ can increase the average propane consumption rate by 23.6 times (Comparing Runs S2 and S6). The driving force also has an impact on the rate. Increasing the driving force can increase the rate as expected by the crystallization theory. Comparing experiments S2 and S4, Run S4 increased 7.8 times. Table 5.1: Factor Affecting Rate of Propane Consumption. Run Units SI S2 S4 S6 E l AT K 2.5 2.6 4.0 2.6 2.6 AP MPa 0.205 0.215 0.265 0.210 0.195 Growth Condition liquid liquid liquid liquid vapor Rotational Speed 200 300 300 300 210 Impeller Geometry I I I n I Average Propane Consumption Rate g/hr 1.905 2.42 18.9 57.0 0.702 Although the propane consumption rate varied from one experiment to another, it did not appear to be a significant factor in the quality of recovered water. Table 5.2 illustrates this fact. Runs S5, S7, and S8 were compared. All these runs were conducted with liquid propane as hydrate former and at the same rotational speed of 300 rpm. Experiment S7 produced 73 recovered water with the lowest conductivity. This is probably attributed to the fact that the amount of recovered water was lower. Less water was recovered because less hydrate was formed as seen from table 5.2 where the amount of spent propane is shown. Table 5.2: Comparing Results from Propane Displacement Experiments Run Units S7 S8 S5 AT K 2.5 2.1 3.9 AP MPa 0.220 0.180 0.263 Quantity of RW Kg 1824 2258 2701 Impeller Geometry II n I Average Propane Consumption Rate g/hr 57.69 50.13 18.96 Crystal growth period hrs 6 8 24 Quantity of hydrate former 346 401 455 Conductivity of RW mS 24.60 29.05 29.90 The salt experiments demonstrated that a brine solution can be concentrated through clathrate hydrate formation. However, the level of the conductivity of the recovered water indicates a poor separation of the hydrate crystals from the concentrated solution. Table 5.3 indicates the reduction in the conductivity in the experiments without washing ranged from 13 (run S6) to 44.3 % (run SI). This large variation is due to two reasons. First, the quality of the recovered water drops (conductivity increases) when we form more crystals. Second, Runs SI and S2 were drained for an additional 105 minutes that extracted 342, and 287g of concentrate, respectively, which improved the quality of recovered water. Larger reductions were obtained when we used water washing and propane displacement as seen from Table 5.3. Comparing run S7 (lowest conductivity in recovered water with propane displacement) with experiment SI (lowest conductivity in the recovered water for the no washing experiments), the recovered 74 water's conductivity was 4% higher than run SI, but it produced 203% more recovered water. Therefore, propane displacement is preferable. Table 5.3: Summary of Salt Experiments Run Type % Reduction of RW from Feed Quantity ofRW Conductivity ofRW Quantity ofWash Water % g u.S SI No Washing 44.3 900 23600 S2 No Washing 37.1 1200 26400 S3 Water Washing 73.3 2037 11190 3600 S4 Water Washing 82.0 3284 7460 7000 S5 Propane Displacement 28.2 2700 29900 S6 No Washing 13.0 2789 35500 S7 Propane Displacement 39.3 1824 24600 S8 Propane Displacement 28.9 2258 29052 So- Water Washing 42.9 3897 23600 3800 Washing with water could displace some additional impurities from the crystals. However, a large wash water to recovered water ratio (7.0 Kg/ 3.3 Kg) could only reduce the conductivity of the recovered water to 7,500 uS, and it also produced 6.9 Kg of spent wash water at a conductivity of 14,900 uS. The poor separation is probably due to poor contacting between wash water and crystals. One experiment used non ionic dye solution to wash the crystals and it was observed that the wash water was poorly contacted with the bed of crystals. Washing the crystals in this manner does not seem to be an effective method of separating the brine from the crystal. 75 The results with the effluent experiments were consistent with the results from the salt experiments. Table 5.4 illustrates this fact. Switch from impeller geometry I to II, using liquid propane instead of vapour, and increasing the rotational speed by 90 rpm, Run E2's propane consumption rate increased dramatically, 35.5 times of Run E l . Again increasing the driving force increases the rate as illustrated by Runs E2 through E5. Even though Run E l had the largest driving force, it did not produce the fastest rate because it used impeller geometry I, and vapour propane in the crystallization process. In addition using impeller geometry JJ, the jelly-like structure could be broken. The experiment with the fastest rate was S7 that used liquid propane during the crystallization process with impeller geometry TJ, and the largest driving force. Our objective was to determine the effectiveness of separation and the quality of recovered water. Examining Table 5.4 we observe that Runs E2 through E5 used approximately the same operating conditions (similar driving force, same impeller geometry, same rotational speed, and the similar quantities of liquid hydrate former). The only variables were how the crystals were separated, and the batch of effluent used in each run. Both Runs E2, and E3 were conducted without washing, but these two runs used different batch of effluent. The batch of effluent used in Run E2 contained more impurities than the batch of effluent used in Run E3 (Run E2's conductivity was five times of Run E3). Runs E2 and E3 recovered water's removal of salts were 34.9%, and 22.5%, respectively. This result shows that multiple stages of clathrate hydrate concentration and washing could be used to improve the quality of the recovered water. In multiple stage operations the recovered water is recrystallized, washed, and decomposed followed by any additional stages. The number of stages will be limited by effectiveness of impurity removal from each stage and economic considerations. Our work suggests at minimum one addition stage could be used to improve the quality of recovered water. 76 Table 5.4: Summary of Effluent Experiments Run Units E l a E 2 a E 3 b E4b E 5b Type No Washing No Washing No Washing Propane Displacement Water Washing AT K 2.6 1.5 1.7 1.7 1.7 AP MPa 0.195 0.151 0.174 0.176 0.193 Impeller Geometry I II II n n Quantity of Hydrate Former g 236 423 412 420 423 Phase of Hydrate Former Vapour Liquid Liquid Liquid Liquid Rotational Speed rpm 210 300 300 300 300 Formation of Jelly-Like Structure/Plug Yes No No No No Crystal Growth Period hr 336 17 6.5 6 5.5 Quantity of Hydrate former g 236 423 412 420 423 Average Propane Consumption Rate g/hr 0.702 24.9 63.4 70.0 76.9 Quantity ofRW g 3604 4029 3549 2664 3245 Conductivity of RW uS 20100 16200 3456 2460 3410 RW's % Removal of Salts (from feed) % 19.0 34.9 22.5 45.1 23.1 T O C ofRW 1.402 1.373 1.545 RW's % Removal of TOC (from feed) % 22.5 27.3 11.4 Quantity of Wash Water g 2.314 a Batch from Louisiana Pacific's Pulp Mill from Chetwynd, B.C. Feb./95 b Batch from Louisiana Pacific's Pulp Mill from Chetwynd, B.C. May/95 77 The rest of the effluent experiments, Runs E4 and E5, used the same batch of effluent as Run E3, but these runs used a different separation process. Comparing Runs E3 (no washing), and E4 (propane displacement), experiment E4 produced 22.9% less recovered water than Run E3 but its salt and TOC removal was 2.00 and 1.21 times better than Run E3, respectively. Now comparing Runs E3 and E5 (washing with water), experiment E5 produce 8.5% less recovered water than E3, and its salt and TOC removal were 1.02 and 0.507 times of Run E3 while using 2314g of wash water. Obviously, this run performed poorly, and washing crystals with water in this manner does not seem to be effective. Poor contacting between crystals and wash water may exist. From these comparisons, the best recovered water belongs to the run with propane displacements, E4. This run produced 2664 g of recovered water that had a salt and T O C reduction of 45.1% and 27.3%, respectively. Liquid propane could displace the occluded concentrate from the crystals that improved the quality of the recovered water. In all experiments with effluent, we found that conductivity, TS, VS, FS, TOC, and TIC increased in the concentrate and decreased in the recovered water. However, TSS increased in the concentrate and decreased in the recovered water. In addition the recovered water contained twice as much total suspended solids than its feed solution. Probably, the bed of clathrate crystals behaves as a screen. It filters the large particles such as fibers, which are present in the effluent. Chapter 6: Conclusion and Recommendations We wanted to investigate if a single process unit, crystallizer/wash column/melter, could be used to recover clean water from BCTMP mill effluents using clathrate hydrate crystals. We designed and constructed an apparatus that will do this job. Since BCTMP is a dark color effluent, it is difficult to observe the clathrate concentration process. As a result, we also worked with a 2.5 wt% salt solution. Our objective is to determine the effectiveness of separation with and without washing, and quantify the quality of recovered water. The results 78 from the salt and the BCTMP experiments were consistent with each other. In all the effluent experiments, TOC, TIC, TS, VS, and FS increase in proportion to conductivity in the concentrate, and they also decreased in proportion to conductivity for the recovered water. However, the total suspended solids increased in the recovered water, and decreased in the concentrate. It is likely that the bed of crystals behaves like a screen that filters the fibers from the solution. The crystals were grown from various operating conditions, and they were washed with liquid propane and water. We observed that the rate of crystal growth from vapour propane was slower than using liquid propane. Our experiments demonstrated that the rate of growth can be increased by 2.7 times by using liquid propane. In addition, using an impeller geometry that creates more turbulence (mixing), impeller geometry II. the rate of growth could by increased up to 23.6 times. The As expected the driving force could also increase the rate of growth. All these factors can increase the rate of formation, but evidence with the experiments with salts suggests increasing the rate will not significantly effect the quality of the recovered water. However, the quantity of hydrate former used during crystallization appears to affect the quality of recovered water. Using more hydrate former increased the quantity of recovered water, but the quality decreased. For both experiments with salt and effluent, we found that the recovered water's removal of salts ranged from 13% to 44.3 %. Higher removal of salts could be accomplished with addition separation stage. Displacement with propane could improve the removal of salts to a range of 28.2% to 45.1%. Further reduction was accomplished with washing with water. Using between 2314 to 7000g of wash water, this process had a salt removal in the range of 23.1% to 82 %. Of these experiments, the recovered water from three effluent experiments was analyzed for TOC. The results showed that the TOC values increased or decreased in proportion to salinity. The TOC removal values for no washing, propane displacement and washing with water are 22.5%(E1), 27%(E2), and 11.4%(E3). From these results we can 79 conclude, propane displacement is the best separation technique, and washing crystals with water in this manner is not an effective way of separating crystal from the concentrate. To examine its suitability of recovered water for reuse, tests must be conducted to use this recovered water in normal pulping operations such as cooking, and bleaching, and examine the quality of the finished pulp. Effluents could be diluted to simulate the recovered water's characteristics and hence be used for testing pulping processes. This work did not examine the suitability of the recovered water because it was not the focus, and it will be continued by another researcher. Our work revealed that the separation of clathrate hydrate crystals from concentrated salt solutions or effluent is not able to achieve the desired purity in one stage. Future work should perhaps investigate additional stages as well as improving the contacting between the washing medium 80 References Barduhn, A.J., "The status of the Crystallization Process for Desalting Saline Waters", Desalination, 5:175-184 (1968) Chowdhurry, J., "CPI Warming up to Freeze Concentration", Chem. Eng. 95(6):24-31 (1988) Coleman, T.C. , Economic Evaluation of Kraft Black Liquor Freeze Concentration for Pulp and Paper Industry Final ReportUS Dept. Energy, 1986. De Graauw, J. and J.J. Tutten, "The Mechanism and the Rate of Hydrate Formation", 3 rd International Symposium on Fresh Water from the Sea, Vol 3, 103-116, 1970 Douglas, J. (1989) "Freeze Concentration: An Energy Efficient Freeze Concentration Process", EPRI Journal, 14(1): 16-21. Englezos, P., "Clathrate Hydrate Concentration of Mechanical Pulp Effluents", Applications for Reasearch and Technology Grant, March 1993. Englezos, P., "Clathrate Hydrates", Ind. Eng. Chem. Res., Vol 32, No 7, 1993. Englezos, P. (1993): "The Freeze Concentration Process and its Applications." Dev. Chem. Eng. Min. Proc. 1993, in press. Fellows, P.J. Food Processing Technology. Ellis Horwood, Ltd., Chichester, West Sussex, England, pp 401-414 (1990). Gaarder, C.(1993) : M.ASc . Thesis, Dept. Chem. Engineering, UBC, Vancouver, B.C., Canada. Glew, D.N., "Solution Treatment", U.S. Pattent 3085832, Oct 16, 1962. Glew, D.N. and M L . Hagget, "Kinetics of Formation of Ethylene Oxide Hydrate, Part I, Experimental Method and Congruent Dissussion. Can. J. Chem. 1968 46 3867-3865 Heist, J.A. (1989): in "Standard Handbook of Hazardous Waste Treatment and Disposal", Freeman, H.M. (ed.), McGraw-Hill, New York, pp 6.133-6.143. 81 Jeffrey, G.A. and W. Saenger, Hydrogen Bonding in Biological Structures. Springer-Verlag Heidelberg New York, ! 991 Kenny R., Gorgal, R.G., Martineau, D., and Prahacs, S. (1992): Pulp and paper Canada, 93 (10): 55-88. Knox, W.G., et al., "The hydrate process, Chemical Engineering Progress", Vol 57, No2, Feb 1961 Kubota, H . , SWrnizu, K., Tanaka, Y. , and T. Makita, "Thermodynamic Properties of R13 (CC1F3), R23 (C1F3), R152a (C2H4F2), and Propane Hydrates for desalination of Sea Water", J. of Chen Eng. in Japan, VI7, no 4, 1984 Natarajan, V. , Bishnoi, P.R., and N. Kalogerakis, "Induction Phenomena in Gas Hydrate Nucleation", Chem. Eng. Science, Vol. 49, No 13, p 2075-2087, 1994. Pavlov, G,D., and I.N. Medvedv, "Gas Hydrating Process of Water Desalination",. Proceedings of the first international Symposium on Water Desalination, Washington, D.C. October 3-9, 1965. Rautenbach, R., Seide, A., "Technical Problems and Economical Problems of the Hydrate Process", Proceedings 6th intern. Sympossium Fresh Water from the Sea, Vol 4, 43-51, 1978. Rousseua, R.W. and E.E. Sharpe, "Freeze Concentration of Black Liquor: characteristics and limitations", Ind. Eng. Chem. Proc. Dev., 19, 201-204, (1980) Rousseau, R.W. (1981): "A view of Separation Processes in Manufacture of Forest Products", AIChE Symposium Series: The Use and Reprocessing of Renewable Resources, 77 (207): 19-24 Shaw, D.J., Introduction to Colloid and Surface Chemistry. 3rd Ed., Butterworth &Co., 1989. Sloan, E.D. (1990): Clathrate Hydrates of Natural Gases Marcel Dekker Inc., New York, NY. Thomas, R.H. P. and R H . Harrison, "Pressure-Volume-Temperature Relations of Propane", J. Chem. Engineering Data 1982, 27, 1-11 82 Tleimat, B.W., "Freezeing methods" In Principles of Desalination, K.S. Sprienger (ed), 2nd ed., Academic Press In. Ch 7: 359-399 1980 Wiley, A.J. Dambruch, L.E., Parker, P.E., and Dugal, H.S. (1978): TAPPI J., 61 (12), 77-88. Willson, R.C., E. Bulot and C.L. Cooney, "Clathrate Hydrate Formation Enhances Near -critical and Supercritical Solvent Extraction Equilibria", Chem. Eng. Comm., 95, 47-55 (1990) Young, J., "Chetwund Pioneers Innovations in Zero-Effluent Pulp Production", Pulp and Paper, 73-75, March 1994. 83 Anendix A: Procedure for Determining TS. VS. FS. and TSS Values TS, VS, andFS 1 Incinerate residual residues from crucibles at 550 ° C for one hour. Cool crucible to room temperature, and then place them in a desiccator for four hours. 2 Weigh the crucible with (Mcg+Eff) and without(McR) effluent sample. 3 Evaporate effluent at 103-105 oC for 48 hours then cool crucible to room temperature and place them inside desicator for four hours. 4 Weigh the crucible (McR+Eff*) 5 Insert crucible into oven at 550°C for one hour, then cool to room temperature and place it desicator for four hours. 6 Weigh the crucible (McR-t£ff*l) 7 Repeat test two more times for each sample. Calculations TS = (McR±Eff_ - M C R ) / ( M c R ± E f f - M C R ) FS = (McR+Eff** - M C R ) / (McR±Eff" MCR) VS = TS-FS 84 TSS 1 Dry a filter paper and a crucible in a furnance at 105 ° C for one hour. 2 Cool filter paper and crucible to room temperature and place them inside a desicator for four hour. 3 Weigh crucible and filter paper M ^ R + F . 4 Clean a Whatman Glass Micro Fibre Filter GF/A (1.6u,m) with deionized water. 5 Seat filter paper to the filter with a small amount of deioned water 6 Measure 50.0 ml of effluent using a graduate cyclinder. 6 Filter the measured sample and rinse the residu from the graduate cylinder with distilled water. 7 Wash the filter paper and sides of the the filter with deionized water. 8 Remove the filter paper and the cake from the the filter, and insert them into the alreaded weighed crucible. Insert the crucible into the furnace at 105 ° C , and dry them overnight. 9 Cool the crucible and filter paper to room temperature, and dry them in a desiccator for four hours. 10 Weigh crucible M C R + F * 11 Repeat test two more time. 85 Calculations TSS-nvtcR+F* - MCR±l)/50 86 Appendix B: Set Procedures Preliminary Set up Procedure for No Washing Experiments The procedure for setting up the supply vessel is as follows. The temperature of bath2 was set at a constant value, T o a t n 2 . This bath cooled the gas supply vessel. Fixing bath2's temperature to 269 K allows us to prepare for the gas supply vessel's filling process. After eight hours, the temperature inside the supply vessel, T s v , reached a steady state value. Before filling the supply vessel, residual air must be purged from the vessel. Purging ensures that the gas in the vessel contains predominantly propane. Propane gas was used to displace the residual air from the supply vessel. The gas was injected into the vessel to a pressure of 0.4 MPa. After which the gas was vented to atmospheric pressure. This purging procedure was conducted at least three times to ensure contamination with air is at a minimum. Next, either propane gas or nitrogen gas was injected into the supply vessel until the piston moved to its target position. At this moment the supply vessel is ready to be filled with propane. Both propane and nitrogen are fed slowly in to their respective chambers in the supply vessel until both chambers reached the propane saturation pressure that corresponded to the temperature inside the supply vessel. At all times the differential pressure between both chambers did not vary more than 0.1 MPa. The piston did not move significantly with this differential pressure restraint. Both the propane and the nitrogen cylinder were supplied by Medigas, and their contents have a purity of 99.9 %, and 99% respectively. Next, propane from the gas cylinder is injected into the gas supply vessel at a pressure above the saturation point that corresponded to the temperature inside the supply vessel. At this point liquid and vapor propane can be observed flowing in a stratified manner through the 87 translucent tubing into the supply vessel. Liquid propane continued to flow into the supply vessel for approximately fifteen minutes When liquid propane had stopped flowing to the supply vessel, nitrogen was fed into the upper chamber of the supply vessel until the pressure is approximately 0.24 MPa greater than the saturated pressure corresponding to the temperature inside the supply vessel. The piston inside the supply vessel moved slightly downwards, and it compressed the liquid propane to a the same pressure as the nitrogen chamber. This procedure ensures all of the propane is in the liquid state. The supply vessel is ready to be used for an experiment. While the supply vessel was being set up, the crystallizer was simultaneously prepared for its preliminary set up. First, the crystallizer's temperature bath was set at a constant value, Tbathl- This bath cooled the crystallizer. Setting crystallizer's bath temperature keeps the crystallizer experimental temperature at a constant. Next, feed solution who's characteristic have been previously determined was injected into the crystallizer. Section 3.2, Effluent Tests, describes how the characteristic of the feed solution was evaluated. Typically, the feed solution ranged from 9.8 to 11.0 Kg. Injecting feed solution in to the crystallizer allows the feed solution to be precooled to the experimental temperature prior to crystallization process. At this point stirring was commenced. After eight hours of cooling, the temperature inside the crystallizer reached a steady state value. To purge the crystallizer of residual air, propane gas is injected from the supply vessel, manually, into the crystallizer to a pressure of 0.276 MPa followed by venting of the gas out of the crystallizer. This procedure is repeated at lease three times to ensure contamination with air is at a minimum. Upon completion of the purging procedure, the nucleation procedure can begin. 88 Preliminary Set Up for Propane Displacement Experiments The crystallizer and the supply vessel had to be set up in an identical fashion as the no washing experiments. The gas supply vessel must be must be purged of residual air in the system, its temperature bath must be set at the target temperature, and the quantity of propane that was injected into the crystallizer for crystal growth must be measured. Similarly, the crystallizer must be purged of residual air, its temperature bath must be set at the target temperature, and its feed solution must be injected into the crystallizer. An additional set up procedure was necessary for these experiments with propane displacement. During the crystal growth period, the supply vessel was filled up in preparation for the propane displacement process. The filling procedure was identical to the procedure that was used previously. Preliminary Set up for Wash Water Experiments Again the crystallizer and the supply vessel were set up in an identical manner as the previous set of experiments. Like the experiments with propane displacement, an additional set up procedure was necessary for the preparation of the washing procedure. The wash water supply vessel was filled up with a known quantity of wash water, and this vessel was cooled to the experimental temperature. In one experiment non ionic dye was added to the wash water to observe the washing process. This run used 655.45 mg of dye per Kilogram of water. Once these preparations were conducted, experiments with wash water can begin. 89 Appendix C: Calculation for the Quantity of Propane that was injected in the Crystallizer Measured values P s v Pressure inside the supply vessel T s v Temperature inside the supply vessel A H S V The change in height of the piston for before and after injection. All the propane is in the liquid state (well above saturated conditions). The pressure and temperature values are known. Thus the molar volume of propane, Vp, can be looked up in published vapour-liquid propane equilibrium data. This worked used the data from Thomas and Harrison 1982. The volume of propane inside the supply vessel can be calculated from the following equation. V = I T J R 2 A H s v where R= radius of the gas supply vessel The following equation gives the mass of propane that is injected into the inside the crystallizer. where M ^ = Molecular weight of propane Vp = Data from Thomas and Harrison, 1982 at Psv and Tsv M p ^ V M ^ / v p All units were in SI. 90 Appendix D: Equilibrium Results Equilibrium data for the salt and effluent experiments are presented in Tables D l , and D2, respectively. Figure D l plots these points on a partial phase diagram for hydrate-propane-water system. Examining the figure, one notices that after crystal growth period, the equilibrium pressure is elevated from the previous value. Table Run Temperature Equilibrium Pressure Before Crystallization After Crystallization K MPa MPa S1 273.7 0.250 0.291 S2 273.8 0.257 0.290 S3 272.8 0.207 0.230 S4 272.0 0.180 0.226 S5 272.1 0.205 S6 273.6 0.263 S7 273.5 0.265 S8 274.4 0.326 S9 274.3 0.318 Table Run Temperature Equilibrium Pressure Before Crystallization After Crystallization K MPa MPa E1 274.9 0.250 0.270 E2 276.7 0.373 0.394 E3 276.4 0.343 0.355 E4 276.5 0.353 0.366 E5 276.2 0.330 0.342 CD CD -h O -1 CD 0 -2 CO r+ 9L F3' 01 5" 3 O O CD CO cn W 0) CO CD -S o CD 0 CO l-f 01 N 01 O 3 o o CD CO CO m CD CD O CO N 01 o 3 O o CD CO CO s o CO N 01 o 3 o 0 CD CO CO V> 01 to In 01 O c CT O —. -n T5 3 p C CD O -t TO CD 01 I_J w - i ~« s. ^ " £. CD 01 O < » r+ 01 ^ O 7 \ c c =1 ° " — 2. CD c c 3 1 6 Appendix E : Propane Consumption Test for Vapour Pronane in B C T M P Effluent 92 Procedure We wanted to determine how the rates changed with increasing concentration. We decided to measure the initial rate of vapour propane consumption and the finial rates of vapour propane consumption during the crystal growth period with BCTMP effluent. In this experiment, impeller geometry I was used to mix the contents inside the crystalizer. During the crystallization process, the propane feed was turn off at the beginning the crystal growth period. The pressure was allowed to drop to approximately 0.450 MPa. At this moment the pressure was recorded, and the timer was turned on. This precise moment was defined as time zero. The crystallization process was allowed to take place. During the crystallization process the pressure Vs time relationship was recorded. The crystallization process was allowed to continue until the pressured dropped to 0.418 MPa. At this point, the feed was turn back on, and the crystallization process continued until the target quantity of propane was consumed. At this point, the feed was shut off, and the pressure was allowed to drop to 0.450 MPa. As before, the pressure Vs time relationship was monitored. This time the crystallization process was allowed to consume vapour propane until the pressure reached 0.433 MPa. Once this pressure was reached the tset was finished. Vapour Propane Consumption Experiments Results In these run propane consumption rates were measured for beginning of the crystallization period, KE1, and at the end of the crystallization period, KE2. The temperature in the liquid and vapour phase at beginning, Run KE1, were 274.25 K, and 276.65 K respectively. Whereas, the temperature in the liquid and vapour phase for the end, Run KE2, were 274.85 K, and 275.35 K, respectively. At the beginning, Run KE1, when the conductivity in the solution was at 24800 u,S, the rate was slow. It took 29 minutes to drop from pressure 93 of 0.449 to 0.434 MPa. Table E . l and Figure E . l present the consumption data at the beginning. At the end when the conductivity in the solution was 26600 u,S, the rate was extremely slow. The time required to drop a pressure from 0.449 to 0.434 MPa was 265 minutes. Table E.2 and Figure E.2 display the propane consumption data for this run. Table E . 1: Propane Consumption Data at the Beginning of the Experiment (Run KE1) Time Pressure Min. MPa 0 0.4485 5.5 0.4451 11.25 0.4417 19 0.4383 27.75 0.4349 36.75 0.4316 47 0.4282 57.75 0.4248 69.75 0.4214 85 0.4180 Table E.2: Propane Consumption Data at the End of the Experiment (Run KE2) Time Pressure Min. MPa 0 0.4502 6 0.4468 24.5 0.4417 85.5 0.4366 265.5 0.4332 95 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058540/manifest

Comment

Related Items