UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Supercritical fluid extraction of canola seed Fattori, Michael J. 1985

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

Item Metadata

Download

Media
831-UBC_1986_A1 F47.pdf [ 11.35MB ]
Metadata
JSON: 831-1.0076857.json
JSON-LD: 831-1.0076857-ld.json
RDF/XML (Pretty): 831-1.0076857-rdf.xml
RDF/JSON: 831-1.0076857-rdf.json
Turtle: 831-1.0076857-turtle.txt
N-Triples: 831-1.0076857-rdf-ntriples.txt
Original Record: 831-1.0076857-source.json
Full Text
831-1.0076857-fulltext.txt
Citation
831-1.0076857.ris

Full Text

ci SUPERCRITICAL FLUID EXTRACTION OF CANOLA SEED by MICHAEL J. FATTORI B.Sc, University Of Toronto, 1976, M.Sc, University Qf B r i t i s h Columbia, 1980 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES In t e r d i s c i p l i n a r y Studies, (Chemical Engineering, Bio-Resource Engineering) We accept t h i s thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA November 1985 © Michael J. F a t t o r i , 1985 7> In presenting t h i s thesis in p a r t i a l fulfilment of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t freely available for reference and study. I further agree that permission for extensive copying of t h i s thesis for scholarly purposes may be granted by the Head of my Department or by his or her representatives. It i s understood that copying or publication of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of Bio-Resource Engineering ,Chemical Engineering The University of B r i t i s h Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date: Nov. 1985 i i Abstract The extraction of o i l from fixed beds of Canola seed ( Brassica napus and Brassica campestris ) was studied using carbon dioxide at temperatures and pressures ranging from 25 to 90°C and 10-36 MPa respectively. The highest o i l s o l u b i l i t y in the C0 2 (11 mg/g C0 2) was observed at 36 MPa and 55°C. The equilibrium o i l concentration in the C0 2 phase, was found to be independent of the o i l concentration in the seed phase. The extracts were found to be e s s e n t i a l l y free from phosphorus (<7ppm) and their fatty acid content did not change s i g n i f i c a n t l y as the extraction progressed. The t o t a l amount of o i l recovered from the seeds by C0 2 extraction depended upon the seed pre-treatment. For commercially flaked seed, t h i s amount was comparable to that recoverable by conventional hexane extraction. The C0 2 extraction of simple t r i g l y c e r i d e s at 36 MPa and 55°C was investigated. The s o l u b i l i t i e s of t r i p a l m i t o l e i n , t r i o l e i n , and tr i e i c o s e n o i n were 20 mg/g C0 2, 10 mg/g C0 2, and 4 mg/g C0 2 respectively. The composition of C0 2 extracts of an equimolar mixture of the above t r i g l y c e r i d e s was also studied. It was found that the concentration of each t r i g l y c e r i d e in the extract was equal to the product of i t s mole-fraction in the mixture and i t s s o l u b i l i t y in the C0 2. Equations governing the mass transfer from the Canola seed to the C0 2 solvent were developed. A transient one-dimensional mathematical model based on these equations was used to obtain concentration p r o f i l e s of o i l in both the solvent and seed phases, and to determine the ov e r a l l volumetric mass transfer c o e f f i c i e n t . The calculated concentrations and extraction rates were in good agreement with experimental r e s u l t s . The ov e r a l l volumetric mass transfer c o e f f i c i e n t for the i n i t i a l constant rate period was found to be proportional to the 0.54 power of i n t e r s t i t i a l v e l o c i t y . iv Table of Contents Abstract i i L i s t of Tables v i i L i s t of Figures xiv Acknowledgements xvi I INTRODUCTION 1.1 Introduction 1 1.2 Background 1.2.1 S u p e r c r i t i c a l F l u i d Extraction 6 1.2.2 Oilseed Production 8 1.2.3 Canola Seed Processing 9 1.2 Research Objectives 14 II LITERATURE REVIEW 2.1 Introduction 17 2.2 Review of SFE Literature 17 2.3 S u p e r c r i t i c a l Extraction of Oilseeds 19 2.4 Seed Anatomy 28 III EXTRACTION MODEL 3.1 Introduction 34 3.2 Extraction From a Fixed Bed 34 3.3 Mass Balance Equations 37 3.4 Microscopic Extraction Model 40 3.5 Mass Transfer C o e f f i c i e n t s 45 3.6 Computer Simulation 45 IV EXPERIMENTAL EQUIPMENT AND PROCEDURES 4.1 Introduction 49 4.2 Experimental Equipment 4.2.1 High Pressure Liquid Chromatograph 51 4.2.2 Solvent Flowpath 51 4.2.3 Pressure and Flow Control 53 4.2.4 Extraction Vessels 55 4.2.5 Flow Restricter 62 4.2.6 Extract C o l l e c t i o n System 63 4.2.7 Equipment Ca l i b r a t i o n 69 4.3Materials 4.3.1 Carbon Dioxide 70 4.3.2 Seed Samples 70 4.3.3 O i l Samples 71 4.3.4 Simple t r i g l y c e r i d e s 72 4.4 Seed Treatment Methods Prior to Extraction 4.4.1 Introduction 73 4.4.2 Seed Crushing 73 4.4.3 Seed Chopping 73 4.4.4 Flaking and cooking of seed 74 4.4.5 Pressure Rupturing 74 V 4.4.6 P a r t i a l Extraction of Seed 75 4.4.6.1 P a r t i a l Extraction With Hexane 75 4.4.6.2 P a r t i a l Extraction With C0 2 76 4.5 Experimental Extraction Procedure 4.5.1 Vessel Loading procedure 4.5.1.1 Seed Material 77 4.5.1.2 Liquid O i l Samples 77 4.5.2 Equipment Startup 79 4.5.3 Extract sampling 80 4.5.4 S o l u b i l i t y Determinations 81 4.5.5 Equipment Shutdown 82 4.6 A n a l y t i c a l Procedures 4.6.1 Total Seed O i l Determination 82 4.6.2 Scanning Electron Microscopy 4.6.2.1 Sample Co l l e c t i o n 83 4.6.2.2 Sample Preparation for SEM 84 4.6.3 Seed-bed Sectioning Method 84 4.6.4 Fatty Acid Analysis 84 4.6.4.1 Transesterfication 85 4.6.4.2 Gas Chromatographic Procedure 86 4.6.4.3 Validation Procedure 87 4.7 Phosphorus Analysis 4.7.1 Introduction 88 4.7.2 Digestion Procedure 88 4.7.3 Phosphate Standards 89 4.7.4 Photometric Procedure 89 4.7.5 Phospholipid Calculations 90 4.7.6 Detection Limits of Procedure 91 V RESULTS AND DISCUSSION 5.1 Introduction 92 5.2 O i l s o l u b i l i t y as a Function of Pressure and Temperature 92 5.2.1 P r a c t i c a l Implications of S o l u b i l i t y Data 99 5.3 Equilibrium O i l Concentration in the C02 101 5.4 Eff e c t of Seed Treatment 5.4.1 Seed P a r t i c l e Size 105 5.4.2 Experimental Results 105 5.5 Scanning Electron Microscopy of seeds 111 5.6 Fatty Acid Composition of Extracts 5.6.1 Introduction 115 5.6.2 Fatty Acid Ester Response Factors 115 5.6.3 Fatty Acid Composition of Canola O i l s 116 5.6.4 Fatty Acid Composition of C0 2 Extracts .... 119 5.6.5 Fatty Acid Composition of Hexane Extract .. 126 v i 5.7 C0 2 Extracts of Simple T r i g l y c e r i d e s 5.7.1 Introduction 129 5.7.2 S o l u b i l i t y of Simple Tr i g l y c e r i d e s 130 5.7.3 S o l u b i l i t y of a Tr i g l y c e r i d e Mixture 134 5.7.4 Tr i g l y c e r i d e S o l u b i l i t y Interactions in C0 2 139 5.7.5 Prediction of O i l S o l u b i l i t y in C0 2 140 5.8 Phosphorus Content of O i l s 5.8 1 Introduction 143 5.8.2 Phosphorus in Commercially Produced Canola O i l 144 5.8.3 Phosphorus in C0 2 Extracts of Canola Seed 145 5.8.4 Phosphorus in C0 2 Extracts of Canola O i l 146 5.9 Computer Simulation of Extraction Process 5.9.1 Introduction 150 5.9.2 Results 150 VI CONCLUSIONS 168 VII RECOMMENDATIONS 170 NOTATION 172 LIST OF REFERENCES 174 APPENDIX I Errors Associated With the Fatty Acid Analysis 181 APPENDIX II L i s t i n g of Source Code From Computer Progams 183 APPENDIX III L i s t of Material Suppliers 194 v i i L i s t of Tables TABLE 1.1 Typical values of v i s c o s i t y , density and d i f f u s i v i t y for l i q u i d , gaseous and s u p e r c r i t i c a l carbon dioxide 3 TABLE 1.2 Typical s p e c i f i c a t i o n s of commercially processed crude and refined Canola o i l s (Appelqvist, 1972) .. 14 TABLE 4.1 Dimensions of the extraction vessels used in the experiments 58 TABLE 4.2 Errors associated with the various system parameters as determined experimentally 69 TABLE 4.3 Specifications of commercial siphon grade carbon dioxide 70 TABLE 4.4 Specifications of the t r i g l y c e r i d e samples used during the experiments 72 TABLE 4.5 Gas chromatographic parameters for the fatty acid methyl ester analyses 87 TABLE 4.6 Major phospholipid components of Canola o i l 91 TABLE 5.1 Concentration of o i l in the p a r t i a l l y extracted seed samples , 101 TABLE 5.2 d i s t r i b u t i o n of seed p a r t i c l e sizes, for the d i f f e r e n t methods of seed treatment 105 TABLE 5.3 FID response factors for fatty acid methyl esters r e l a t i v e to methyl palmitate 115 TABLE 5.4 The fatty acid composition of the t r i g l y c e r i d e s in Canola o i l obtained from three v a r i e t i e s of Canola seed (Ackman, 1983) 116 TABLE 5.5 Fatty acid composition of the t r i g l y c e r i d e s in four samples of Canola o i l . For a description of each sample refer to text. The fatty acids C20:1 and C18:3 were not resolved using the chromatographic procedure 118 TABLE 5.6 Fatty acid composition of sequential carbon dioxide extracts of CSP cooked Canola seed. Conditions: vessel #1, pressure 36 MPa, temperature 55°C, flow rate 0.7 g/min 119 v i i i TABLE 5.7 Fatty acid composition of sequential carbon dioxide extracts of CSP flaked and cooked Canola seed. The analysis of o i l extracts obtained from the r e s t r i c t e r valve is also provided. Conditions: vessel #1, pressure 36MPa, temperature 55°C, flowrate 0.7 g/min 122 TABLE 5.8 P r o f i l e by carbon number of the various t r i g l y c e r i d e s present in Canola o i l from two d i f f e r e n t sources (Ackman, 1983) 125 TABLE 5.9 Fatty acid composition of the sequential hexane extract of CSP cooked Canola seed. Conditions: vessel #1, pressure 1.5 MPa, temperature 55°C, flowrate 0.7 g/min 129 TABLE 5.10 Mass fr a c t i o n and mole fraction of components in the t r iglycer ide (TG) mixture 137 TABLE 5.11 Mass fr a c t i o n of C16:1, C18:1, and C20:1 t r i g l y c e r i d e s at each data point as determined using the t r a n s e s t e r i f i c a t i o n procedure. Mass fractions are reported with an error of ± 0.005. Conditions: vessel #1, pressure 36 MPa, temperature 55°C,flow rate 0.7 g/min 138 TABLE 5.12 Comparison of calculated and experimental t r i g l y c e r i d e composition of the C0 2 extract of a t r i g l y c e r i d e mixture. Conditions: vessel #1, pressure 36MPa, temperature 55°C, flow rate 0.7 g/min 140 TABLE 5.13 calculated mole-fraction concentration of t r i g l y c e r i d e in an o i l composed of three fatty acids in the molar r a t i o : C16(P) 0.1, C18(0) 0.8, and C20(E) 0.1 143 TABLE 5.14 Phospholipid content of commercial unrefined Canola o i l . Absorbance readings (Abs) are in absolute values 145 TABLE 5.15 Phospholipid content of refined and bleached commercial Canola o i l . Absorbance reading (Abs) are in absolute units 145 TABLE 5.16 Phospholipid content of C0 2 extracts of flaked and cooked Canola seed. Absorbance readings (Abs) are in absolute units 146 TABLE 5.17 Phosphorus analysis of chloroform washings of the glass wool and glass beads after extraction. Absorbance (Abs) i s shown in absolute units. .... 148 ix L i s t of Figures FIGURE 1.1-Phase diagram for carbon dioxide showing the relationship of the s u p e r c r i t i c a l state to the s o l i d , l i q u i d and vapor states. The t r i p l e point i s designated as TP and the c r i t i c a l point as C 2 FIGURE 1.2 Molefraction s o l u b i l i t y of napthalene in carbon dioxide at 55 °C (Paul and Wise, 1971) 4 FIGURE 1.3 Density of.carbon dioxide as a function of pressure for d i f f e r e n t temperatures 5 FIGURE 1.4 Steps in the processing of Canola seed (Appelqvist, 1972) 6 FIGURE 1.5 Steps in the processing of crude Canola o i l . ..... 13 FIGURE 2.1 S o l u b i l i t y of Sunflower seed o i l in carbon dioxide at 40 °C as a function of pressure (Stahl et a l . , 1980) 21 FIGURE 2.2 S o l u b i l i t y of Soybean and Rapeseed o i l s in l i q u i d carbon dioxide at 20°C and s u p e r c r i t i c a l carbon dioxide at 40°C as a function of pressure ( Stahl et a l . , 1980; Bunzenberger et a l . , 1984) 22 FIGURE 2.3 Density of carbon dioxide at 20 °C and 40 °C as a function of pressure (Newitt et et a l . , 1956) 24 FIGURE 2.4 Photograph of a t y p i c a l Brassica napus seed (magnification 40X) 29 Figure 2.5 Lateral section through a Brassica napus seed showing the seed's basic anatomy (Stanly et a l . , 1 976) 30 FIGURE 2.6 Scanning electron micrograph of a section of a B^ napus seed fragment showing the seeds outer coat (Magnification 200X) 31 FIGURE 2.7 Scanning electron micrograph of a section of B. napus seed showing the seeds r e t i c u l a t e d outer coat (Magnification 400X) 33 FIGURE 3.1 Schematic diagram of a fixed-bed extraction vessel 35 X FIGURE 3.2 Solute concentration p r o f i l e s in the solvent phase (a) and the ' s o l i d ' phase (b) in a fixed bed extractor at di f f e r e n t times; where tO denotes the beginning of the extraction 36 FIGURE 3.3 Concentration of solute in the solvent phase (a) at extractor outlet as a function of time. Cumulative mass of solute (b) extracted as a function of time or cummulative solvent passed through the bed 38 FIGURE 3.4 Cross-sectional diagram of an imaginary cluster of seed p a r t i c l e s , at four stages during the extraction. The black areas of the diagram represent pools of l i q u i d o i l , while the shaded areas represent o i l within the intact seed tissue, a) The p a r t i c l e s are covered with a layer of o i l . b) Bare o i l - f r e e areas of the seed begin to appear. c) The majority of surface of seed p a r t i c l e i s o i l - f r e e , d i f f u s i o n of o i l from the narrow i n t e r s t i c e s becomes s i g n i f i c a n t . d) The entire outside surface of the cluster i s o i l - f r e e . At t h i s stage a l l o i l extraction i s from the narrow i n t e r s t i c e s between the seed p a r t i c l e s 41 FIGURE 3.5 a) O i l surface area (Ap) available for mass transfer on seed cluster as a function of seed o i l concentration; b) Mass transfer c o e f f i c i e n t (K) as a function of seed o i l concentration; c) the product of the mass transfer c o e f f i c i e n t and the o i l surface area as a function of seed o i l concentration 44 FIGURE 4.1 Photograph of the complete experimental s u p e r c r i t i c a l f l u i d extraction system showing: c i r c u l a t i n g cooler (a), modified HP-1081B l i q u i d chromatograph (b) with extract c o l l e c t i o n system (c), volumetric flow-meter (d) and dual tracking chart recorder (e) 50 FIGURE 4.2 Schematic diagram of the experimental extraction system, a) solvent (C0 2) reservoir, b) shut-off valve, c) sintered steel f i l t e r , d) diaphragm pump, e) flow and pressure transducer, f) temperature e q u i l i b r a t i o n c o i l , g) extraction vessel, h) sintered steel f i l t e r , i ) temperature controlled r e s t r i c t e r valve, j) sample c o l l e c t i o n vessel, k) volumetric flow meter. ....... 52 FIGURE 4.3 Photograph of the HPLC pump with the cooling c o l l a r removed . 54 FIGURE 4.4 Photograph of the aluminum c o l l a r which was used to cool the pumphead on the HPLC. The coolant delivery and return l i n e s , with insulation, can be seen at the right 56 xi FIGURE 4.5 Photograph of the three extraction vessels used in the experiments. In the photograph the tops of the vessels are i n s t a l l e d 57 FIGURE 4.6 Cross-sectional view of extractor vessel # 1. A l l measurements shown are in cm, except were indicated 59 FIGURE 4.7 Details of the extraction vessel seal, a) view from the inner surface of the extractor showing the c i r c u l a r sealing ridge, b) cross-sectional view of a portion of the extractor and top, showing the sealing surface and sealing ridge • 60 FIGURE 4.8 Photograph of d e t a i l s of the sealing arrangement on vessel #2 61 FIGURE 4.9 Photograph of the MV-200s r e s t r i c t e r valve shown in the aluminum block which was used to regulate i t s temperature 64 FIGURE 4.10 Photograph of the oven in the extraction system. In the picture the extraction vessels (a,b), the r e s t r i c t e r valve with heating block (c), the stain l e s s steel f i l t e r s (d,e) and the s i l i c a sampling tube (f) are v i s i b l e 65 FIGURE 4.11 Cross-sectional diagram of the sampling head with a top view of the main body of the sampler. A l l measurements shown are in cm unless otherwise s p e c i f i e d . A l l of the components cf the head which are shown were made from 316 stainless s t e e l , except where otherwise indicated. The inset i s a diagram of a 1/16' Valco tube union. The union was used to couple the exit tube from the r e s t r i c t e r to the fused s i l i c a i n l e t tube of the sampling system 67 FIGURE 4.12 Photograph of the extract c o l l e c t i o n system which was used to separate and c o l l e c t the o i l from the C0 2. In the picture the s i l i c a tube (a), the c o l l e c t i o n v i a l (b) and the C0 2 exhaust f i t t i n g (c) are v i s i b l e 68 FIGURE 5.1 S o l u b i l i t y of Canola o i l in C0 2 as a function of pressure at four temperatures. Conditions: 7g flaked seed, vessel #2, C0 2flow rate 0.7 g/min. 94 FIGURE 5.2 S o l u b i l i t y of Canola o i l in C0 2 as a function of temperature at four pressures. Conditions: 7g flaked seed vessel #2, C0 2flow rate 0.7 g/min 95 FIGURE 5.3 S o l u b i l i t y of Canola o i l in C0 2 as a function of C0 2 density at four temperatures. Conditions: 7g flaked seed, vessel #2, C0 2 flow rate 0.7 g/min. .96 x i i FIGURE 5.4 S o l u b i l t y of Canola o i l in C0 2 as a function of temperature at d i f f e r e n t C0 2 densities. Conditions: 7g flaked seed vessel #2, C0 2 flow rate 0.7 g/min 97 FIGURE 5.5 Density of carbon dioxide as a function of pressure at d i f f e r e n t temperatures. The c r i t i c a l point (CP) of the C0 2 i s indicated on the diagram. (Newitt et a l . , 1 956) 98 FIGURE 5.6 O i l concentration in the C0 2 phase at the extractor outlet for d i f f e r e n t seed-bed o i l concentrations. The reduced o i l concentration seed was prepared by p a r t i a l l y extracting samples of flaked seed for d i f f e r e n t lengths of time with hexane. Conditions: pressure 36 MPa, temperature 55°C, C0 2 flow rate as indicated on the Figure 102 FIGURE 5.7 O i l concentration in the C0 2 phase at the extractor outlet for d i f f e r e n t seed-bed o i l concentrations. The reduced o i l concentration seed was prepared by p a r t i a l l y extracting samples of flaked seed for d i f f e r e n t lengths of time using s u p e r c r i t i c a l C0 2 at 36 MPa and 55 ° C. Conditions: pressure 36 MPa, temperature 55°C, C0 2 flow rate as indicated on the Figure 103 FIGURE 5.8 Extraction curves for Canola seed subjected to di f f e r e n t pre-treatments. For comparison purposes the extraction curve for Canola o i l on glass beads i s also shown. Conditions: vessel #1, pressure 36 MPa, temperature 55°C, C0 2 flow rate 0.7g/min 107 FIGURE 5.9 Transformed extraction curves for several Canola seed pre-extraction treatments. The Y axis of the graph represents the concentration of o i l in the C0 2 at the extractor outlet 108 FIGURE 5.10 Scanning electron micrograph of a fragment of flaked Canola seed prior to extraction. (Mag. 620X, 20 Kv, Au-Pd) 112 FIGURE 5.11 Scanning electron micrograph of a fragment of flaked Canola seed after p a r t i a l extraction with C0 2 at 36 MPa and 55°C (Mag. 620X, 20 Kv, Au-Pd). The seed fragment was extracted for approximately 15 minutes at a C0 2 flow rate of 0.5g/min 113 FIGURE 5.12 Scanning electron micrograph of a fragment of flaked Canola seed aft e r being ' f u l l y ' extracted with C0 2 at 36 MPa and 55 °C. (Mag. 780X, 20Kv, Au-Pd). The seed fragment was extracted for approximately 2 h at a C0 2 flow rate of 0.5g/min 114 xi i i FIGURE 5.13 Chromatogram of the fatty acid methyl esters in an e s t e r i f i e d sample of a t y p i c a l C0 2 extract of Canola seed (36 MPa, 55°C, flow 0.7 g/min). Analysis conditions: Column- SP-2330 on 100/120 mesh chromosorb WAW; detector(FID) and injector temp. 250°C; isothermal 200°C 117 FIGURE 5.14 Extraction curve for a 4.2 g sample of commercially cooked Canola seed. The extraction was ca r r i e d out at 36 MPa and 55 °C. The numbered intervals on the curve indicate the regions over which o i l samples were co l l e c t e d for fatty acid analysis 120 FIGURE 5.15 Fatty acid composition of the extracts indicated in F i g . 5.14 121 FIGURE 5.16 Extraction curve for a 4.2 g sample of commercially cooked Canola seed. Conditions: 36 MPa, 55°C, C0 2 flow rate 0.7 g/min. The dotted areas on the curve represent the regions over which samples were col l e c t e d for fatty acid analysis 123 FIGURE 5.17 Fatty acid composition of the Canola seed-C0 2 extracts indicated in F i g . 5.16 124 FIGURE 5.18 Extraction curve for a 3.5 g sample of cooked Canola seed. The extraction was ca r r i e d out using hexane at 1.5 MPa and 55 ° C. The dotted areas on the curve represent the regions over which samples were c o l l e c t e d for fatty acid analysis 127 FIGURE 5.19 Fatty acid composition of the Canola seed hexane extracts indicated in F i g . 5.18 128 FIGURE 5.20 Extraction curves for pure-tripalmitolein (C16:1), t r i o l e i n (C18:1) and tri-11-eicosenoin (C20:1). The extractions were performed from a glass bead matrix using C0 2 at 36 MPa and 55 °C. Conditions: vessel #1, C0 2 flow rate 0.7 g/min 131 FIGURE 5.21 S o l u b i l i t y of the three t r i g l y c e r i d e s indicated in F i g . 5.20 as a function of their molecular weight. The error bar shown i s representative of a l l points 132 FIGURE 5.22 The negative logarithm of the s o l u b i l i t y of the three t r i g l y c e r i d e s indicated in F i g . 5.20 as a function of their molecular weight. A t y p i c a l error bar i s indicated 133 FIGURE 5.23 Extraction curves for an equal weight mixture of t r i p a l m i t o l e i n , t r i o l e i n and tri-11-eicosenoin. The extraction of the mixture was performed from glass beads using C0 2 at 36 MPa, 55 °.C at a flow rate of 0.7 g/min. 135 xiv FIGURE 5.24 Mass fraction of the C0 2 extracts of the t r i g l y c e r i d e mixture at d i f f e r e n t stages of the extraction. The composition of the extracts was determined using the transesterification-GC procedure. Each set of points represents an amount of o i l corresponding to the points on F i g . 5.23. Conditions: vessel #1, pressure 36 MPa, temperature 55°C, C0 2 flow rate 0.7 g/min 137 FIGURE 5.25 Extraction curve for a 4.5 g sample of cooked Canola seed showing the inter v a l s (dotted) over which samples of o i l were co l l e c t e d for phosphorus analysis. Conditions: pressure 36 MPa, temperature 55°C, C0 2 flow rate 0.7 g/min 147 FIGURE 5.26 Extraction curve obtained by passing C0 2 at 55°C and 36 MPa through extraction vessel #4 containing 1.5 g of crushed seed. The C0 2 flow rate was 1.6 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 16.7 cm/min.The computed extraction curves were calculated using three d i f f e r e n t values of ApK 151 FIGURE 5.27 Extraction curve obtained by passing C0 2 at 55°C and 36 MPa, through extraction vessel #1 containing 4.0 g of crushed seed. The C0 2 flow rate was 2.7 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 3.9 cm/min.the computed extraction curve was calculated using an ApK value of 2.0 gC02/cm min 152 FIGURE 5.28 Extraction curve obtained by passing C0 2 at 55 °C and 36 MPa, through extraction vessel #1 containing 4.0 g of crushed seed. The C0 2 flow rate was 2.6 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 3.8 cm/min. The computed curve was calculated using an ApK value of 1.8 gC02/cm min 153 FIGURE 5.29 Extraction curve obtained by passing C0 2 at 55 °C and 36 MPa, through extraction vessel #1 containing 3.8 g of commercially flaked seed. The C0 2 flow rate was 1.7 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 2.5 cm/min. The computed extraction curve was calculated using an ApK value of 1.3 gC02/cm min 154 FIGURE 5.30 Extraction curve obtained by passing C0 2 at 55 °C and 36 MPa, through extraction vessel #1 containing 4.0 g of f i n e l y chopped seed. The C0 2 flow rate was 2.3 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 3.5 cm/min. The computed extraction curve was calculated using an ApK value of 2.0 gC02/cm min 155 X V FIGURE 5.31 Extraction curve obtained by passing C0 2 at 55 °C and 36 MPa, through extraction vessel #1 containing 4.0 g of f i n e l y chopped seed. The C0 2 flow rate was 1.7 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 2.5 cm/min. the computed extraction curve was calculated using an ApK value of 1.5 gC02/cm min 156 FIGURE 5.32 Extraction curve obtained by passing C0 2 at 55 °C and 36 MPa, through extraction vessel #3, containing 12.0 g of crushed seed. The C0 2 flow rate was 1.4 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 0.5 cm/min. The corresponding extraction curve generated by the computer using an ApK value of 0.6 gC02/cm min 157 FIGURE 5.33 Extraction curve obtained by passing C0 2 at 55 °C and 36 MPa, through extraction vessel #1 containing 12.0 g of crushed seed. The C0 2 flow rate was 0.7 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 0.3 cm/min. The computed extraction curve was calculated using an ApK value of 0.4 gC02/cm min 158 FIGURE 5.34 Normalized o i l concentrations in the seeds as a function of normalized distance from the bed entrance aft e r 240 min. The conditions correspond to those shown in Fi g . 5.33. 159 FIGURE 5.35 O i l concentration in the solvent-phase as a function of normalized distance from the bed entrance at four d i f f e r e n t times. The conditions correspond to those shown in Fig 5.26 161 FIGURE 5.36 O i l concentration in the seed-phase as a function of normalized distance from the bed entrance at four d i f f e r e n t times. The conditions correspond to those shown in F i g . 5.26 162 FIGURE 5.37 O i l concentration (y) in the solvent phase as a function of normalized distance from the bed entrance and time. The conditions correspond to those shown in F i g . 5.26 164 FIGURE 5.38 O i l concentration (x) in the seed phase as a function of normalized distance from the bed entrance and time, the conditions correspond to those shown in F i g . 5.26 165 FIGURE 5.39 Volumetric mass transfer c o e f f i c i e n t s (ApK) as a function of i n t e r s t i t i a l v e l o c i t y (v) 166 as a function of i n t e r s t i t i a l v e l o c i t y (v) 166 xvi Acknowledgements I would l i k e to express my sincere thanks to my two supervisors, Dr. N.R. Bulley and Dr. A. Meisen for their f i n a n c i a l support and th e i r advice and encouragement during the course of t h i s project. I would also l i k e to express my deepest appreciation to my friends and family, who helped me throughout the course of th i s endeavour by providing both their moral and f i n a n c i a l support: A. Balabanian, R. Carmichael, 0. Dubek, D. F a t t o r i , P. F a t t o r i , E. Kwong, N. Jackson, J. Kennedy, J. Pehlke, B. S h e l l , V. Semerjian, B.A. Stockwell, D. Whalen. F i n a l l y , I would l i k e to extend a special note of thanks to my friend Milos Horvath whose assistance with chapter 3 and the computer model was invaluable. 1 I. INTRODUCTION 1 .1 Introduction S u p e r c r i t i c a l f l u i d extraction(SFE), also referred to as "dense gas extraction", "dense phase extraction" and " s u p e r c r i t i c a l gas extraction", i s a high pressure solvent extraction process characterized by the use of s u p e r c r i t i c a l f l u i d s in place of conventional l i q u i d solvents. An understanding of the term " s u p e r c r i t i c a l f l u i d " can be obtained by re f e r r i n g to the pressure-temperature phase diagram of a pure substance shown in F i g . 1.1. The lines A-TP, TP-C and TP-B divide the diagram into three regions representing three phases: s o l i d , l i q u i d and vapor. Along each of the l i n e s two phases exist in equilibrium with each other while at point TP, the t r i p l e point, three phases co-exist. The l i n e TP-C, which separates the l i q u i d and vapor regions of the diagram, terminates at the c r i t i c a l point C. Beyond t h i s point, the substance, which w i l l no longer exist as a l i q u i d or vapor i s usually referred to as a s u p e r c r i t i c a l f l u i d . While in the s u p e r c r i t i c a l state, the substance resembles both a gas in terms of i t s v i s c o s i t y and d i f f u s i v i t y and a l i q u i d in terms of i t s density (Table 1.1). Although a variety of substances can be used in the i r s u p e r c r i t i c a l state for extraction, common gases such as carbon dioxide, nitrous oxide and ethane have received the greatest attention. The solvation capacity of a s u p e r c r i t i c a l f l u i d (at constant T) i s strongly dependent on i t s density which in turn 2 oo <D ro Q_ CM LUo EN o CO co UJ °. rr 00 Q_ CO <N 1 1 1 1 gl I l I I l I l I " T " l i I L 1 T "1'": -supercritical ] - liquid ! • . • . • . • . • . * . * . * . • . • . * . * . • . • » : • . • . • : • : • : • : • ' • " • " • ' . • . - . - . • . • . • . * . - . - . * * ~ solid - ^ < > ^ v a p o r J T P -.A.X i i i i i i i i i i • • -100.0 -80.0 -60.0 -40.0 -20.0 0.0 20.0 40.0 60.0 80.0 100.0 120.0 TEMPERATURE (°C) FIGURE 1.1 Phase diagram for carbon dioxide showing the rel a t i o n s h i p of the s u p e r c r i t i c a l state to the s o l i d , l i q u i d and vapor states. The t r i p l e point i s designated as TP and the c r i t i c a l point as C. 3 i s proportional to the external pressure applied to the f l u i d (Brogle, 1982). It follows therefore that the solvent power of a s u p e r c r i t i c a l f l u i d can be controlled by varying the pressure at which the process i s c a r r i e d out. This phenomenon i s i l l u s t r a t e d in F i g . 1.2. The s o l u b i l i t y of napthalene in C0 2 at 55°C and 10.0 MPa i s about 25 X 10"* moles/mole C0 2. As the pressure of the carbon dioxide i s increased, the napthalene mole fraction r i s e s . At 25.0 MPa the s o l u b i l i t y of the napthalene approaches 500 X 10"* moles/mole C0 2. This represents a twenty-fold increase in s o l u b i l i t y . Compounds extracted by a high pressure s u p e r c r i t i c a l f l u i d can thus be recovered from the f l u i d by a simple pressure reduction. Additionally, e f f e c t i v e separations can also be achieved by a v a r i a t i o n of temperature rather than pressure. TABLE 1.1 Typical values of v i s c o s i t y , density and d i f f u s i v i t y for l i q u i d , gaseous and s u p e r c r i t i c a l carbon dioxide. C0 2 state v i s c o s i t y g/cm s density g/cm3 d i f f u s i o n c o e f f i c i e n t cm2/s l i q u i d gas s u p e r c r i t i c a l f l u i d (a) 1.5 X 10 " 2 (a) 1 .4 X 10 "* (a) 9.1 X 10 "* (b) 0.9 (b) 0.002 (b) 0.9 (c,d) 10 - 5 (c,d) 10 -1 (c,d) 10 " 3 - l 0 "* Source: (a) Newitt et a l . 1956; , 1956; (b) Vukalovich and Altunin, 1968; (c) Randall, 1982; (d) Gere, 1983. The a b i l i t y to vary the density i s an integral feature of the s u p e r c r i t i c a l f l u i d extraction (SFE) process. Consequently, the process i s most often c a r r i e d out near the c r i t i c a l point, where the density v a r i a t i o n i s the greatest. 4 J—I—I I I I I I I I I I I I I ' » ' ' 0.0 2.4 4.8 7.2 9.6 12.0 14.4 16.8 19.2 21.6 24.0 26.4 28.8 PRESSURE (MPa) FIGURE 1.2 Molefraction s o l u b i l i t y of napthalene in carbon dioxide at 55 °C (Paul and Wise, 1971). cn E co <_>«=> • d CN "i—1—1—1—1—1—1—1—1—1—1—1—1—1—1—1—1—1—r J I I I I I I L J I I I I I I L 0.0 4.0 8.0 12.0 16.0 20.0 24.0 28.0 32.0 36.0 40.0 PRESSURE (MPa) FIGURE 1.3 Density of carbon dioxide as a function of pressure at d i f f e r e n t temperatures. The c r i t i c a l point (CP) of the C0 2 i s indicated on the diagram (Newitt et a l . , 1956). 6 Figure 1.3 shows the density of carbon dioxide as a function of temperature. At the c r i t i c a l point the variation of the f l u i d density with pressure is so great that large density variations can even be observed at d i f f e r e n t heights in a f l u i d column due to the gr a v i t a t i o n a l f i e l d ( B a l z a r i n i and Ohrn, 1972). In practice, s u p e r c r i t i c a l f l u i d extraction may employ a mixture of two or more f l u i d s . A l t e r n a t i v e l y , the extraction process may use a f l u i d which has had small amounts of conventional solvents added to i t . The added compounds act to modify the solvation behavior of the s u p e r c r i t i c a l f l u i d and allow i t to be 'tai l o r e d ' for a s p e c i f i c extraction. These compounds are commonly referred to as "entrainers" (Brunner and Peter, 1982). For a more detailed outline of the the o r e t i c a l and p r a c t i c a l aspects of SFE the reader i s referred to the review a r t i c l e by Williams(1981) or the January 1982 issue of Separation Science and Technology (Marcel Dekker, Inc.) 1.2 Background 1.2.1 S u p e r c r i t i c a l F l u i d Extraction Although Hannay and Hogarth reported the unusual solvent power of s u p e r c r i t i c a l f l u i d s over 100 years ago, s u p e r c r i t i c a l f l u i d s have only recently gained prominence-. The long leadtime i s , in part, attributable to the d i f f i c u l t y in understanding the physical nature of s u p e r c r i t i c a l f l u i d s . Moreover, for years the price of petroleum-based solvents which are conventially used in solvent extraction, has been low, thereby discouraging research into a l t e r n a t i v e technologies. 7 During the last decade, however, t h i s situation has changed. The cost of petroleum solvents has in some cases increased twenty-fold. Furthermore, as the o i l c r i s i s of 1973 has shown, the supply of these products i s not secure. During the same period, there has been a growing public concern regarding health issues. Many food additives have become suspect and petroleum solvent residues have become far less acceptable (Hellyar and de F i l i p p i , 1982). S u p e r c r i t i c a l f l u i d extraction (SFE), on the other hand, can u t i l i z e safe, low-t o x i c i t y gases such as carbon dioxide for the extraction of food products (Caragay, 1981; Hubert and Vitzthum, 1978). In l i g h t of these recent trends, the interest in SFE i s growing and some promising results have been obtained. For example, edible o i l s have been co l l e c t e d and p u r i f i e d by SFE. Since the gases used for extraction are very v o l a t i l e , v i r t u a l l y no solvent residue remains in the c o l l e c t e d extract. De-solvent i z i n g , a costly and sometimes lengthy procedure when using conventional solvents i s greatly s i m p l i f i e d (Milligan and Tandy 1974). Although several gases can be used for extraction purposes, carbon dioxide i s most commonly used. The preference for C0 2 i s due in large part to i t s physical and chemical properties. It i s non-flammable and non toxic. Its c r i t i c a l temperature i s low (31°C) and the extraction can be c a r r i e d out at pressures of 10-40 MPa which i s well within the c a p a b i l i t i e s of current technology. Furthermore, numerous compounds are soluble in carbon dioxide (Calame and Steiner 1982; de F i l i p p i , 1982). 8 Other advantages are the low cost and high a v a i l a b i l i t y of C0 2. As the knowledge of SFE increases, i t s applications are becoming more diverse. To date i t has been used experimentally in biotechnology for removing fatty acids from aqueous solutions (Shimshick 1983), in agriculture for producing the i n s e c t i c i d e pyrethrum (Stahl and Shutz, 1980), in perfumery for the extraction of v o l a t i l e aromatics from flowers (Calame and Steiner, 1982)- and in pharmaceutics for the extraction of psycho-active drugs (Stahl and Gerrard, 1982). In the food industry SFE has been used to extract o i l from soybeans (Fri e d r i c h et a l . , 1982), corn germ (Christianson et a l . , 1984) and rapeseed (Bunzenberger et a l . , 1984). Additionally, the process is being used commercially for de-caffeinating coffee (Zosel, 1978; Williams, 1981) and for producing hop extracts used in the making of beer (Vollbrecht, 1982; Hubert and Vitzthum, 1978, Gardner,1982). SFE has been used by the petroleum industry on a large scale for t e r t i a r y o i l recovery (Holm and Josendal, 1974; Stalkup, 1978) and experimentally for improving low grade crudes (Humphrey et a l . , 1984; Gearhart and Garwin, 1976). 1.2.2 Oilseed Production The subject of t h i s thesis concerns the oilseed industry, which may benefit greatly from SFE. In the U.S., over 30 m i l l i o n tons of soybeans are processed each year (Hron et a l . , 1982). Almost a l l of the 6 m i l l i o n tons of o i l from t h i s crop i s recovered using hexane. Since hexane i s a petroleum product i t i s subject to continuously increasing 9 cost and uncertain a v a i l a b i l i t y ( F r i e d r i c h and L i s t , 1982). In addition, hexane does not only extract t r i g l y c e r i d e s , but also unwanted gums. These gums must then be removed from the o i l in a separate process. It has been suggested that s u p e r c r i t i c a l f l u i d extraction of oilseeds could produce an o i l which i s e s s e n t i a l l y gum-free (Friedrich and L i s t , 1982). In Canada, the largest oilseed crop i s Canola (composed of Brassica napus and Brassica campestris ). In 1981 over 2.5 m i l l i o n tonnes of the seed were produced (Pigden, 1983). The majority of thi s seed, l i k e soya, i s extracted using hexane. The extract from this process i s then refined in a series of additional steps, which are outlined below. 1.2.3 Canola Seed Processing The i n i t i a l step in the processing of Canola seed (Fig. 1.4) i s a cleaning procedure in which de t r i t u s harmful to the operation of the mechanical extraction equipment i s removed. The seed, which i s then 99% pure i s preheated to 40°C and introduced into a seed crusher. In the crusher the seed i s forced between two r o l l e r s operating at d i f f e r e n t speeds. The shearing and crushing action of these r o l l e r s ruptures most of the o i l -containing c e l l s . This step i s a necessary pre-requisite for the subsequent processing (Othmer and Agarwal, 1955). The seed material emerging from the crusher consists of thin (> 0.2mm) flakes with a large surface area to volume r a t i o . These flakes are then transferred to large b o i l e r s where they are rapidly heated to 90°C and maintained at t h i s temperature for 20-30 10 SEED STORAGE . 1 OEM I . €R PRE-HEAT CRUSHING FLAKING ROLLERS EJffELLER COOKER SOLVENT EXTRACTOR SOLVENT TOWAL FILTER CRUDE OIL STORAGE SOLVENT RETOVAL M. STORAGE o b FIGURE 1.4 Steps in the processing of Canola seed (Appelovist. 1972). M 11 minutes. This process i s known as "cooking" and serves several functions: 1) i t decreases the v i s c o s i t y of the o i l and allows i t to flow and coalesce; 2) i t adjusts the moisture content of the seed; 3) i t completes the rupturing of the o i l c e l l s ; 4) i t coagulates the proteins in the seed thereby preventing their transfer to the o i l . In addition to the above functions, the cooking of the seed also inactivates the native enzyme myrosinase which catalyzes the degradation of glucosinolates in the seed. The toxic degradation products from t h i s reaction (isothiocyanates and oxazolidinethiones) are soluble in the o i l and interfere with i t s subsequent hydrogenation (Rutkowski et a l . , 1982). Following the cooking procedure, the seed material enters the expeller. This device, which contains a rotating screwshaft running in a c y l i n d r i c a l b a r r e l , squeezes the l i q u i d o i l from the seed. The o i l obtained from t h i s operation i s subsequently c l a r i f i e d by passing i t through a series of f i l t e r s or, a l t e r n a t i v e l y , by centrifugation. The s o l i d leaving the expeller which may contain from 15-I8wt% o i l , i s then broken up and extracted with commercial hexane. After the extraction process, the hexane saturated seed "meal" i s transferred to de-solventizers where the solvent i s flashed from the meal by steam i n j e c t i o n . The meal i s then "finished" in a toasting process and emerges e s s e n t i a l l y free from solvent with a residual o i l content of about 1 to 2wt% (Beach, 1983; Anjou, 1972). The o i l obtained from the hexane extraction process i s then de-1 2 solventized and added to the expeller o i l and the mixture i s subsequently refined. The following i s a summary of additional Canola o i l processing steps (Teasdale and Mag, 1983). In the " r e f i n i n g " step (Fig. 1.5), hot (80°C) crude o i l i s mixed with small amounts of phosphoric acid and water to promote the p r e c i p i t a t i o n of phospholipid gums. This de-gumming step i s necessary since the gums tend to coagulate and p r e c i p i t a t e during storage. After t h i s i n i t i a l de-gumming, the o i l is treated with an a l k a l i (eg. sodium hydroxide) to remove a large portion of the free fatty acids and to further reduce i t s phosphorus content. During a l k a l i r e f i n i n g , the phosphorus le v e l of the o i l i s reduced from 100-200 ppm to approximately 5 ppm (Appelqvist, 1972). The next step, referred to as "bleaching", i s c a r r i e d out in preparation for subsequent hydrogenation and deodorization of the o i l . During t h i s step the o i l i s exposed to surface active clay which adsorbs pigments such as chlorophylls and carotenoids. After bleaching, the o i l i s treated with pressurized hydrogen, in the presence of a c a t a l y s t , to saturate the fatty acids. Saturation increases the oxidative s t a b i l i t y of the o i l and increases i t s melting point. The hydrogenated o i l i s next subjected to steam d i s t i l l a t i o n for deodorizing. In the process, flavor and odor compounds, free fatty acids and various degradation materials are removed from the o i l . The f i n a l product i s a bland tasting, 13 REFINING PHOSPHORIC ACID WATER ALKALI TREATPENT SODIUM HYDROXIDE BLEACHING ACTIVATED CLAY CRUDE i OIL FORMULATION FLAVOR, VITAMINS PROTEIN, WATER HYDROGENATION NICKEL CATALYST SALAD OIL MARGARINE SHORTENING FIGURE 1.5 Steps in the processing of crude Canola o i l . 1 4 l i g h t yellow o i l with excellent storage s t a b i l i t y . The l a s t step in Canola o i l processing i s c a l l e d formulation. As the name implies, i t i s during t h i s process that the o i l i s converted into i t s f i n a l form, i . e . salad o i l , margarine or shortening. Depending on which of these products i s desired, l e c i t h i n , color, vitamins, milk whey, water and s a l t may be added. For a detailed discussion of Canola o i l processing the reader i s referred to Appelqvist (1972) or Teasdale and Mag (1983) In Table 1.2 some t y p i c a l s p e c i f i c a t i o n s for crude and refined Canola o i l are l i s t e d . TABLE 1.2 Typical s p e c i f i c a t i o n s of commercially processed crude and refined Canola o i l s (Appelqvist, 1972). crude refined free fatty acids 1.0% 0.05% water 0.3% 0.05% phosphorus 500 ppm < 5 ppm sulfur 7 ppm erucic acid 2% 2% 1.3 Research Objectives In order to assess the f e a s i b i l i t y and merits of s u p e r c r i t i c a l C0 2 as a solvent in the Canola oilseed industry, a large amount of information i s required. At the beginning of t h i s project i t was decided to focus research on two main areas. F i r s t , the parameters a f f e c t i n g the extraction process had to be established. These parameters include temperature, pressure and flowrate of the carbon dioxide 15 solvent, and the physical state of the seed material. Second, the t r i g l y c e r i d e and phospholipid content of the extracts had to be determined. Both areas are important for determining the optimum extraction conditions (which in turn can be used to compare the SFE process with the standard hexane extraction process) and for gaining insight into the extraction mechanism i t s e l f . The s p e c i f i c objectives of the research were: 1 ) to develop laboratory equipment for investigating the extraction of Canola seed using s u p e r c r i t i c a l carbon dioxide; 2) to determine how the s o l u b i l i t y of the o i l in C0 2 is affected by temperature and pressure; 3) to determine how the extraction process i s affected by the mechanical condition of the Canola seeds; 4) to determine how the composition of the extract varies during the course of the extraction; 5) to determine the mass transfer c o e f f i c i e n t for the extraction process. The research project was car r i e d out in a number of stages. In the i n i t i a l stage, equipment was developed by which small amounts of the seed material could be extracted with a stream of carbon dioxide at d i f f e r e n t temperatures, pressures and flow rates. With the same equipment the extracts could be col l e c t e d for further analysis. In the second stage a method was developed to determine the s o l u b i l i t y of the o i l in the carbon dioxide. Experiments were then conducted to determine the relationship between o i l 1 6 s o l u b i l i t y and carbon dioxide pressure, temperature and seed o i l concentrat ion. In the following stage, the eff e c t of seed treatment on the extraction process was investigated. The seed treatment techniques which were used ranged from commercial flaking to the use of high pressure gas to burst the seeds. In the succeeding section, a n a l y t i c a l techniques were developed to gain information on the fatty, acid composition and phosphorus content of the carbon dioxide extracts of Canola seed. In the f i n a l stage, the equations governing the extraction process were developed and solved numerically. The results predicted by t h i s mathematical model were then compared with experimental data. 1 7 II. LITERATURE REVIEW 2.1 Introduct ion A brief review of the general l i t e r a t u r e pertaining to s u p e r c r i t i c a l f l u i d extraction (SFE) is presented before discussing the l i t e r a t u r e dealing s p e c i f i c a l l y with s u p e r c r i t i c a l extraction of oilseeds. The f i n a l section of this chapter contains a discussion of the anatomy of the Canola seed. 2.2 Review Of SFE Literature The fact that s u p e r c r i t i c a l f l u i d s could act as solvents was f i r s t reported by Hannay and Hogarth in 1879. They demonstrated that a s a l t , potassium iodide, could be dissolved in s u p e r c r i t i c a l ethanol and they observed that the s a l t ' s s o l u b i l i t y was largely dependent on the pressure of the ethanol. These results were subsequently confirmed (Tyrer,l9l0) and i t was shown that the dense gas/salt solution was e l e c t r i c a l l y conductive (Krans, 1922). During t h i s time i t was also suggested that s u p e r c r i t i c a l water in the earth's crust was of importance in geological processes ( N i g g l i , 1912). In 1945 Katz and Whaley (as c i t e d in Randal,1982) showed that natural gas could be used to separate hydrocarbon mixtures. In the late 50's Zhuze demonstrated that s u p e r c r i t i c a l f l u i d s could be used for removing ashphaltenes and resins from petroleum (1958), extracting ozocerite wax from raw ore (1959) and recovering l a n o l i n from wool grease (1958). In the following decade, several a r t i c l e s were published dealing with SFE in the petroleum industry ( E l l i s and V a l t e r i s , 1965; Zhuze, 1960). 18 Extensive surveys of the s o l u b i l i t i e s of compounds in s u p e r c r i t i c a l f l u i d s were reported (Giddings et a l . , 1968; McLaren et a l . , 1968; Giddings et a l . , 1969) and attempts were made to predict s o l u b i l i t i e s in s u p e r c r i t i c a l f l u i d s based on Hildebrand s o l u b i l i t y parameters ( Giddings et a l . , 1969; Czubryt et a l . , 1970). In 1971 Paul and Wise published the f i r s t comprehensive review of the theory and applications of the technique e n t i t l e d "The Pr i n c i p l e s of Gas Extraction" and in 1978 the f i r s t symposium devoted e n t i r e l y to the subject was held in Germany. During t h i s symposium, papers were presented on the t h e o r e t i c a l aspects of SFE (Schneider), empirical methods for determining the s o l u b i l i t i e s of compounds in s u p e r c r i t i c a l f l u i d s (Stahl et a l . ) , the applications of SFE in the food and flavor industry (Zosel, Hubert and Vitzthum), and c r i t e r i a for the design and construction of f u l l - s c a l e SFE plants (Eggers). Zosel also discussed the p r i n c i p l e s of s u p e r c r i t i c a l f l u i d separations and their p r a c t i c a l aspects. Zosel also describes how, by using s u p e r c r i t i c a l ethane with pressure programming, he was able to separate a sample of mixed t r i g l y c e r i d e s into f i f t y f r a c t i o n s . Since t h i s symposium, many additional papers have appeared on both the p r a c t i c a l and the o r e t i c a l aspects of SFE. One notable group of papers on the l a t t e r subject deals with attempts to predict the equilibrium s o l u b i l i t i e s of compounds in s u p e r c r i t i c a l f l u i d s (Vetere, 1979; King et a l . , 1983; King and Bott, 1982). The a b i l i t y to predict s o l u b i l i t y would eliminate the need for d i f f i c u l t and costly experimental procedures. 19 Usually these predictions are based on the equations of state (King and Bott, 1982). The v i r i a l expansion developed by Rowlinson and Richardson (1959) has been used in this regard with some success, but i t i s limited to solids and gases. A more useful method for predicting s o l u b i l i t i e s involves a modification of the Redlich-Kwong equation (King et a l . , 1983). Although the equations can be used with l i q u i d s as well as s o l i d s , their solutions require complex c a l c u l a t i o n s . Mackay and Paulaitus (1979) have reported that multicomponent equilibrium s o l u b i l i t i e s can, in p r i n c i p l e , be calculated using extensions of e x i s t i n g theory. However, the current understanding of the l i q u i d and s u p e r c r i t i c a l states i s i n s u f f i c i e n t to make thi s presently possible (King and Bott, 1982). For a detailed account of the early history of SFE the reader i s referred to Booth and Bidwell (1949), for an extensive survey of the various applications of SFE to Randall (1982) and Williams (1981) and for information on the t h e o r e t i c a l aspects of the process, to Paulaitus et a l . , 1983) and Schneider (1978). 2.3 S u p e r c r i t i c a l F l u i d Extract ion Of Oilseeds One of the f i r s t extensive research projects dealing with oilseed extraction using s u p e r c r i t i c a l carbon dioxide, was undertaken by Stahl et a l . in 1980. The authors investigated the e f f e c t of temperature and pressure on the equilibrium s o l u b i l i t y of the o i l in the carbon dioxide, and the e f f e c t of pre-treatments on the extraction rate and o i l recovery. This equipment consisted e s s e n t i a l l y of a temperature controlled 20 fixed bed extractor which could be f i l l e d with seed material and flushed with a stream of high pressure C0 2. At the outlet of the extractor a two stage pressure reduction sampling system was used to c o l l e c t the o i l . During the extraction, a stream of l i q u i d or s u p e r c r i t i c a l carbon dioxide was passed through the extractor which was f i l l e d with seed. The concentration of o i l in the outlet stream was determined by measuring the mass of o i l extracted and the volume of carbon dioxide used. To determine whether the exit o i l concentration had reached equilibrium, the following method was used: the extraction was begun and the extractor outlet concentrations were measured. When the concentration of o i l at the extractor outlet became constant the values were noted and the extraction was stopped. The bed of seeds was then sectioned into nine segments of equal length and each segment subjected to a hexane extraction in order to determine the remaining o i l content. If the section of seedbed at the extractor outlet retained i t s i n i t i a l concentration, then the stream of carbon dioxide passing through t h i s section must have been saturated with o i l . The concentration of o i l in the exit stream (previously recorded) was therefore assumed to have reached saturation. Figures 2.1 and 2.2 indicate the s o l u b i l i t i e s of seed o i l s as a function of temperature and pressure as determined by the authors. From F i g . 2.1 i t can be seen that the s o l u b i l i t y of sunflower seed o i l in s u p e r c r i t i c a l carbon dioxide depends on the system pressure in an almost linear manner, reaching about 3wt% at 70 MPa. Figure 2.2 shows that the solvation capacity of 21 § i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — r R I i i i i i i i i i i i i i i i i i i i I 0.0 8.0 16.0 24.0 32.0 40.0 48.0 56.0 64.0 72.0 80.0 PRESSURE (MPa) FIGURE 2.1 S o l u b i l i t y of Sunflower seed o i l in carbon dioxide at 40 °C as a function of pressure (Stahl et a l . , 1980). 22 I I I I I I I I I I I I I I I » I I I I I 0.0 4.0 8.0 12.0 16.0 20.0 24.0 28.0 32.0 36.0 40.0 PRESSURE (MPa) FIGURE 2.2 s o l u b i l i t y of Soybean and Rapeseed o i l s in l i q u i d carbon dioxide at 20°C and s u p e r c r i t i c a l carbon dioxide at 40°C as a function of pressure ( B A s t a h l et a l . , 1 980; ••"Bunzenberger et a l . , 1984). 23 l i q u i d carbon dioxide i s also, although to a lesser extent, dependent on pressure. At pressures below 25 MPa, l i q u i d carbon dioxide has a higher solvation capacity, for soybean o i l , than s u p e r c r i t i c a l carbon dioxide whereas at higher pressures the opposite i s true. The " s o l u b i l i t y crossover" e f f e c t has also been reported for other vegetable o i l s (de F i l i p p i , 1982). The ef f e c t i s not unexpected and can be explained by considering the pressure-density function of C0 2 (Fig. 2.3). At 20 MPa and 20 °C the density of l i q u i d carbon dioxide i s 0.94g/cm3. At the same pressure but at 40 °C s u p e r c r i t i c a l carbon dioxide has a density of only 0.84g/cm3. However, at a pressure of 30 MPa, the density of the l i q u i d and s u p e r c r i t i c a l carbon dioxide are very similar (0.99g/cm3 vs. 0.92g/cm3). Since the solvation capacity of carbon dioxide i s greater at higher densities (Brogle, 1982; de F i l i p p i , 1982), i t can dissolve more o i l at 20 MPa in the l i q u i d state than in the s u p e r c r i t i c a l state. However, the solvation capacity i s also dependent on temperature. At the higher pressures the densities of the l i q u i d and s u p e r c r i t i c a l carbon dioxide are more nearly equal and consequently temperature plays a dominant role in the solvation process. Stahl et a l . (1980) also studied the e f f e c t of seed preparation on the extraction process. The seed samples were treated p r i o r to extraction using three d i f f e r e n t methods. When rapeseed was chopped to a very small 24 J 1 1 1 1 1 I I I I I I I I I » I I I 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 PRESSURE (MPa) FIGURE 2.3 Density of carbon dioxide at 20 °C and 40 °C function of pressure (Newitt et a l . , 1956). 25 size (<0.4 mm), up to 98% of the o i l could be removed while less than 1% of the o i l was removed from whole seeds. They reported that not only the size of the p a r t i c l e s in the crushed seed aff e c t the extraction, but the shape of these p a r t i c l e s was also s i g n i f i c a n t . It i s also apparent from the authors' data that the degree of crushing had a substantial effect on the s o l u b i l i t y of soybean o i l in the C0 2. Depending upon the crush, the s o l u b i l i t y varied from 6 to 8 mg o i l / g C0 2. No explanation is given as to why the s o l u b i l i t y of the o i l would be affected by seed condition, and no experimental conditions other than carbon dioxide temperature were reported for these experiments. In a more recent paper by L i s t et a l . (1982) the extraction of soybean flakes using s u p e r c r i t i c a l carbon dioxide was studied both from a physical and a chemical perspective. The extractions were performed in semi-batch mode using a procedure similar to that of Stahl et a l . (1980). During the extractions, the col l e c t e d o i l was p e r i o d i c a l l y sampled and analyzed with respect to free fatty acids, unsaponifiables and phospholipids. The authors found that the average phospholipid content of the carbon dioxide extracted o i l was low in comparison to the o i l produced by hexane extraction (0.17% vs 1.5%) and suggest therefore that carbon dioxide extracted o i l would not require de-gumming. The concentration of phospholipids in the s u p e r c r i t i c a l carbon dioxide extracts were found to increase as the extraction proceeded. Thus, the authors noted that, in a commercial process, there would be l i t t l e merit in extracting the seed beyond a certain point since the o i l removed in the 26 later stages of extraction is of poorer q u a l i t y . In addition to chemical tests, the o i l s , which were extracted using hexane and s u p e r c r i t i c a l carbon dioxide, were also compared by means of sensory evaluation techniques. It was found that the carbon dioxide extracts were not s i g n i f i c a n t l y d i f f e r e n t in taste or odor from the hexane extracts. Bunzenberger et a l . (1984) reported on the s u p e r c r i t i c a l f l u i d extraction of rapeseed. The authors present equilibrium o i l s o l u b i l i t y and extraction rate data at d i f f e r e n t pressures and temperatures using both l i q u i d and s u p e r c r i t i c a l carbon dioxide. The method used to obtain equilibrium values was similar to that used by Stahl et a l . (1980). For comparison purposes the reported o i l s o l u b i l i t y data are plotted in F i g . 2.2. Although the data indicate that the o i l s o l u b i l i t y in the l i q u i d and s u p e r c r i t i c a l carbon dioxide increases with increasing pressure, the effect is not as pronounced as expected from the work of Stahl et al.(l980) and de F i l i p p i (1982). In addition, the s o l u b i l i t y of the rapeseed o i l i s s i g n i f i c a n t l y lower than those reported for other vegetable o i l s (de Filippi,1982; L i s t et a l . , 1982; Stahl and Quirin, 1982). No reasons for such low s o l u b i l i t i e s are suggested by the authors. One possible explanation for the low values may l i e in the o i l i t s e l f . Unlike most other vegetable o i l s , certain c u l t i v a r s of rapeseed contain high percentages of erucic acid (Sonntag, 1979). An o i l containing t r i g l y c e r i d e s formed from t h i s high molecular weight fatty acid would be expected to exhibit a lower s o l u b i l i t y in C0 2 (Brunner and Peter, 1982). This explanation 27 cannot be substantiated however, since information on the species of seed or o i l composition was not provided. In a recent paper, Brunner and Peter (1982) reported the equilibrium s o l u b i l i t i e s of pure t r i g l y c e r i d e s and palm o i l in various s u p e r c r i t i c a l f l u i d s . The authors indicate that over the range of temperatures and pressures studied (70,75°C, 10-40 MPa) nitrous oxide and ethane proved to be more e f f e c t i v e solvents for t r i g l y c e r i d e s than either carbon dioxide or t r i - f l u o r o -chloromethane. The reason for these differences i s not readily apparent and can not be explained in terms of density (pC0 2 = 0.645g/cm3, pC 2H 6 = 0.37g/cm3) or p o l a r i t y (C 2H 6 < N 20 < CF 3C1). These results are also contrary to suggestions by Paul and Wise (1971) that compounds exhibit similar s o l u b i l i t i e s in f l u i d s with similar physical properties. A further aspect of Brunner and Peter's research involved the e f f e c t of ethanol entrainer on the s o l u b i l i t i e s of palm o i l in s u p e r c r i t i c a l carbon dioxide. The authors found that, compared with pure carbon dioxide under i d e n t i c a l conditions (70°C, 20 MPa), 10% ethanol in the carbon dioxide increased the s o l u b i l i t y of the o i l by a factor of 20. With the same concentration of ethanol at a higher carbon dioxide pressure (50°C, 30 MPa) the authors reported o i l s o l u b i l i t i e s in excess of 10 wt%. The use of an entrainer i s s i g n i f i c a n t in other respects as well. Due to energy considerations, the recovery of solutes from a s u p e r c r i t i c a l solvent i s more economically car r i e d out by r a i s i n g the temperature of the mixture than by decreasing i t s pressure. If a pure s u p e r c r i t i c a l solvent i s 28 used, solute separation by changing the temperature i s rarely feasible because the required temperature increases are so large that they lead to the degradation of the compounds. By using a suitable entrainer, t h i s problem may be overcome. 2.4 Seed Anatomy Two species of Canola seed, i . e . Brassica napus and Brassica campestris are grown in Canada. The l a t t e r has a s l i g h t l y shorter growing period and a lower o i l content. The seeds of B. Napus (Fig. 2.4) are generally larger (2-3mm), than those of B. Campestris(1.5-2.5mm). The anatomy and location of l i p i d r i c h areas in both species are however s i m i l a r . As seen from F i g . 2.5, the seeds consist of three d i s t i n c t regions: seed coat, cotyledons and r a d i c l e . The dense fibrous seed coat i s composed of st r u c t u r a l carbohydrates, mucilage and l i g n i n ( Yiu et a l . , 1982). This coat (Fig. 2.6) comprises from 12-20wt% of the entire seed weight (Bengtsson et a l . , 1972). The surface of the coat i s observed to be highly r e t i c u l a t e d (Fig. 2.7) and may be somewhat porous (Stanly et a l . , 1976). The endosperm tissue, which i s 1 to 2 c e l l s thick, i s located on the inner surface of the seed coat (Van Caeseele et a l . , 1982). This endosperm tissue i s r i c h in o i l and i s probably the reason why the seed coat contains appreciable amounts of o i l (20wt%)(Anjou, 1972; Yiu et a l . , 1982). The embryo i s composed of the cotyledons which are embryonic leaves and the r a d i c l e from which the root develops ( Bengtsson et a l . , 1972). FIGURE 2.4 Photograph of a t y p i c a l Brassica napus seed, (magnification 40X) Figure 2.5 Lateral section through a Brassica napus seed showing the seed's basic anatomy. (Stanly et a l . , 1976). FIGURE 2.6 Scanning e l e c t r o n m i c r o g r a p h of a s e c t i o n of a B. napus seed fragment showing the seeds o u t e r coat " ( M a g n i f i c a t i o n 2 0 0 X ) . 32 Canola seeds contain approximately 40wt% hexane extractable material by weight (Hofsten 1970; Khan and Hanna, 1983). Generally 95-98wt% of th i s o i l i s in the form of t r i a c y l g l y c e r o l s . The remaining 2-5wt% of the o i l i s composed of mono- and d i - a c y l g l y c e r o l s , phospho- and g a l a c t o l i p i d s , s t e r o l esters, waxes and free fatty acids (Appelqvist, 1972). The o i l is d i s t r i b u t e d evenly throughout the cotyledons and the ra d i c a l (Yiu et a l . , 1982) in the form of small droplets (0.5-1 /urn) bounded by a thin layer of protein. The o i l droplets are located in the cytoplasm of the c e l l s (Hofsten, 1970, Stanly et a l . , 1976). 33 FIGURE 2.7 Scanning e l e c t r o n m i c r o g r a p h of a s e c t i o n of B. napus seed showing the seeds r e t i c u l a t e d o u t e r coat " ( M a g n i f i c a t i o n 400X) . 34 II I . EXTRACTION MODEL 3. 1 Introduction In order to allow for the e f f i c i e n t design of extraction equipment, two basic types of information are necessary: (i) the equilibrium d i s t r i b u t i o n of solute between the feed and the solvent phase and ( i i ) the mass transfer rate of solute from the feed to the solvent. Both of these can, in p r i n c i p l e , be obtained by using a laboratory-scale fixed bed extractor. 3.2 Extraction From A Fixed Bed Consider the fixed bed extractor depicted in Fig 3.1. The solute-free solvent enters at one end of the bed and exits at the opposite end. The bed consists of a mixture of s o l i d inert material and solute. I n i t i a l l y the solute i s d i s t r i b u t e d evenly throughout the bed. The bed can be considered to consist of a series of thin sections having uniform solute concentrations. Provided a dri v i n g force e x i s t s , the solute i s transferred from the ' s o l i d ' phase to the solvent. With time, the concentration of solute in the solvent phase increases, while the solute concentration in the ' s o l i d ' phase decreases. As the solvent passes through succeeding sections, the concentration of solute progressively increases r e s u l t i n g in a solvent phase concentration p r o f i l e as shown in F i g . 3.2a. A corresponding concentration p r o f i l e also develops in the ' s o l i d ' phase (Fig. 3.2b). As the extraction proceeds, the fixed bed w i l l gradually become depleted of solute causing the concentration p r o f i l e in both phases to vary with time (Fig. 3.2a,b). The shape of the 35 Sh I I INLET OUTLET H FIGURE 3.1 Schematic diagram of a fixed-bed extraction vessel. 36 p u o o o BED POSITION SUPPORT PHASE i—i—i—i—r i i O i i i i i i i " i i b 1 1 1 1 T 1 1 1 1 - / t 2 / 7 / t3/ / A A / / \A / / / -t / / t5/ / / • / w / : i u-r i J 7 i i i ,iA i i i i o M I— r i t— Z LLI <_) z o o o BED POSITION FIGURE 3.2 Solute concentration p r o f i l e s in the solvent phase (a) and the ' s o l i d ' phase (b) in a fixed bed extractor at d i f f e r e n t times; where tO denotes the beginning of the extraction. 37 concentration p r o f i l e in each phase i s dependent on the mass transfer area, mass transfer c o e f f i c i e n t , solvent flowrate and the s o l u b i l i t y of the solute in the solvent. From Fig.3.2a i t can be seen that the solute concentration in the solvent phase at the bed outlet also varies with time, giving a curve of the type shown in Fig.3.3a. In the present work, the integrated form of this curve i s obtained experimentally (Fig.3.3b) and i s c a l l e d the 'extraction curve' for the process. It i s understood that the slope at any point on the extraction curve corresponds to the concentration of solute in the solvent phase at the extractor outlet. 3 . 3 Mass Balance Equations In the schematic diagram shown in Fig . 3.1 carbon dioxide enters through one end of the bed and exits via the opposite end. The mass of o i l in the carbon dioxide phase in any element (5h) of the extractor i s , Oil(mass) = (A5hepy) [3.1] where: h - distance along the bed [m] 8h - length of element [m] e - voidage of the bed of seeds [dimensionless3 p - density of the solvent phase [kg/m3] A - cross sectional area of the extractor [m2] y - concentration of o i l in the solvent phase [kg o i l / k g C0 2] 38 SOLVENT PHASE OUTLET CONCENTRATION i i i i i i i i i i i i i i i i i i i i i i i ' a i. i i TIME TOTAL SOLUTE VS TOTAL SOLVENT i i i i i i i i i i i i i i i i i i i i i i i i TIME OR TOTRL SOLVENT FIGURE 3.3 Concentration of solute in the solvent phase (a) at extractor outlet as a function of time. Cumulative mass of solute (b) extracted as a function of time or cummulative solvent passed through the bed. 39 An o i l balance on the solvent phase can be written as: 3(A6hepy) = -pUA9y 5h + ApA5hK(y*-y) dt 9h .[3.2] where: U - s u p e r f i c i a l solvent v e l o c i t y [m/s] Ap - surface area available for mass transfer per unit volume of bed [m2/m3] . K - o v e r a l l mass transfer c o e f f i c i e n t [kgC0 2/m 2s] y* - concentration of o i l in the solvent which i s in equilibrium with seeds having o i l concentration x [kg o i l / k g C0 2] x - concentration of o i l in the seeds [kg o i l / k g o i l - f r e e seed] t - time [s] Four assumptions are i m p l i c i t in Eq. 3.2: ( i ) a uniform o i l concentration in the C0 2 exists across the bed, ( i i ) the amount of a x i a l mixing in the solvent phase is n e g l i g i b l e compared with the convective flow, ( i i i ) t h e o i l concentration in the solvent phase i s small, (iv) an o v e r a l l mass transfer c o e f f i c i e n t can be used to represent the extraction process. If the solvent density and flowrate are constant, Eq. 3.2 reduces to A similar material balance can be written for o i l in the seed phase of the extractor leading to: ep dy = -pU 3y + ApK(y*-y) at 3h [3.3] 3(x( 1-e)A; ) = -ApK(y*-y) at [3.4] 40 where: p -density of o i l - f r e e seeds [kg/m3] s The boundary conditions for Eqs. 3.3 and 3.4 are: t=0 0<h<H x=x0 t>0 h=0 y=0 where H and x 0 denote the t o t a l bed height of seeds and the i n i t i a l o i l content of the seeds, respectively. In order for Eqs. 3.3 and 3.4 to be solved, the equilibrium r e l a t i o n s h i p between the solvent phase o i l concentration and the seed o i l concentration, i . e . y*=f{x}, must be known. This r e l a t i o n s h i p may be determined experimentally by pre-extracting seed samples to obtain a range of x concentrations and then contacting each sample with solvent and measuring the o i l concentration in both phases after equilibrium conditions have been established. 3.4 Microscopic Extraction Model When Canola seed i s crushed the o i l contained within the seed i s li b e r a t e d . The crushed material i s thus composed of fractured pieces of seed and exposed l i q u i d o i l . In the unextracted state i t i s assumed that the o i l associated with the seed i s found in three locations (Fig. 3.5). 1) on the surface of the seed p a r t i c l e s and exposed to the flowing stream of solvent; 2) in the i n t e r s t i c e s between the seed p a r t i c l e s ; 3) within the tissue of the seed p a r t i c l e s . 41 FIGURE 3.4 Cross-sectional diagram of an imaginary cluster of seed p a r t i c l e s , at four stages during the extraction. The black areas of the diagram represent pools of l i q u i d o i l , while the shaded areas represent o i l within intact seed ti s s u e . a) The p a r t i c l e s are covered with a layer of o i l . b) Bare o i l - f r e e areas of the seed begin to appear. c) The majority of surface of seed p a r t i c l e i s o i l - f r e e , d i f f u s i o n of o i l from the narrow i n t e r s t i c e s becomes s i g n i f i c a n t . d) The entire outside surface of the cluster i s o i l - f r e e . At t h i s stage a l l o i l extraction i s from the the narrow i n t e r s t i c e s between the seed p a r t i c l e s . 42 3) within the tissue ( c e l l s ) of the seed p a r t i c l e s . The surface o i l i s removed f i r s t . The o i l in the narrow i n t e r s t i c e s i s primarily removed afte r the surface o i l has been depleted and the channels to the i n t e r s t i t i a l o i l become exposed to the C0 2 solvent. The o i l within the undamaged seed tissue ( c e l l s ) i s not extracted by C0 2. Consider a small cluster of seed fragments in the bed. Let the concentration of o i l in the seed cluster at the beginning of the extraction be represented by x 0. At t h i s time the surface of the seed fragments are covered with a layer of o i l (see F i g . 3.4a). As the extraction proceeds, the surface area of the o i l layer remains constant while i t s thickness decreases (Fig. 3.4b) u n t i l bare or o i l - f r e e regions appear on the outside of the c l u s t e r . Let the corresponding mean concentration of o i l in the seed cluster at t h i s point be denoted by x 1 . At a l a t e r time, the o i l concentration reaches a value of x 2 (where x 2< x, ) and most of the outside surface of the cluster i s o i l - f r e e (Fig 3.4c). The remaining o i l i s contained within the seed tissue and the small spaces between the seed fragments. The concentration value x 3 represents the amount of o i l within the intact seed tissue which can not be extracted by C0 2. Upon further extraction, the outside surface of the c l u s t e r becomes completely bare (Fig. 3.4d). The o i l remaining in the cluster at t h i s stage i s removed from the spaces between the seed p a r t i c l e s by molecular d i f f u s i o n . From Eq. 3.3 i t can be seen that the rate of mass transfer 43 is proportional to the mass transfer c o e f f i c i e n t (K), the drivi n g force (y*-y) and the surface area available for mass transfer (Ap). In the model being suggested here, two di f f e r e n t mass transfer c o e f f i c i e n t s were used. The f i r s t one, K, , governs the transfer of o i l from the surface layer (location 1) to the solvent. The second c o e f f i c i e n t , K 2, governs o i l transfer to the solvent by d i f f u s i o n from the spaces between the p a r t i c l e s (location 2). The value of the f i r s t c o e f f i c i e n t , K, , was expected to be greater than the value of the second c o e f f i c i e n t , K 2. U n t i l the seed-oil concentration reaches a value of x, , the o i l s ' surface area does not change s i g n i f i c a n t l y and, in the model, i t i s assumed to be constant. When the concentration of o i l f a l l s below x, , the area of bare seed surface increases and the area of surface o i l available for mass transfer, where K, i s applicable, correspondingly diminishes. In the model, the decrease of surface area between x=x, and x=0 i s assumed to be a line a r function of the o i l concentration (Fig. 3.5a). For x 2 < x < x, , the surface of the seed i s partly covered with o i l . Because of t h i s , both mass transfer c o e f f i c i e n t s must be considered. However, since K 2 is expected to be very much smaller than Ky , K 2 was set to zero for the concentration range x 2 < x < x, . When the concentration of o i l in the seed p a r t i c l e s f e l l below x 2, only K 2 was considered (Fig. 3.5b). In F i g . 3.5c the relat i o n s h i p between ApK and seed o i l concentration (x) i s shown. I I I I I I I 1 I I I I I I I I I I I I I I I I a • 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 J _ L Seed Oil Concentration l l - i . i i I i i i i i i Seed Oil Concentration a cr I I I l l l l l I I l I I 1 l l l l l I I I l I X 3 X g X<| Seed O i l Concentration ' < i i FIGURE 3.5 a) O i l surface area (Ap) a v a i l a l b l e for mass transfer on seed clu s t e r as a function of seed o i l concentration; b) Mass transfer c o e f f i c i e n t (K) as a function of seed o i l concentration; c) the product of the mass transfer c o e f f i c i e n t and the o i l surface area as a function of seed o i l concentration. 45 3.5 Mass Transfer C o e f f i c i e n t s Equations 3.3 and 3.4 can be solved numerically provided the equilibrium r e l a t i o n s h i p between the solvent and the seeds is known and the parameters U, ps, p, e and Ap are given. The only unknown variable i s the mass transfer c o e f f i c i e n t and i t must be assumed for the purpose of the numerical solution. The correct value of K i s then found by matching the predicted and measured o i l concentrations in the solvent at the extractor outlet as a function of time. It i s known that the mass transfer c o e f f i c i e n t depends on the solvent flowrate, the state of the seeds and the temperature and pressure at which the extraction is carr i e d out (Treybal, 1968). If the .seed material i s crushed or otherwise pre-treated, the surface area available for mass transfer (Ap) i s d i f f i c u l t to measure. In t h i s case i t i s customary to determine the product ApK which i s usually c a l l e d the "volumetric mass transfer c o e f f i c i e n t " . 3 .6 Computer Simulat ion The extraction of Canola seed using C0 2 can be simulated on a computer. The solution of Eqs. 3.3 and 3.4 gives the o i l concentration p r o f i l e s along the bed in both the solvent and seed phase as a function of time. In order to validate the computer model, the predictions made by the model must be compared with experimental data. This can be done using two procedures. In the f i r s t , the calculated bed-oil concentration p r o f i l e s are compared with the 46 experimental p r o f i l e s from the real extractors. This procedure however, i s quite lengthy and involved since i t involves sectioning the experimental seed-bed and performing an o i l -content analysis on each section. In the second procedure the experimental extraction curves are compared with the corresponding extraction curves generated by the model. The extraction curves can be calculated either from the o i l concentration in the seed phase or from the o i l concentration in the solvent phase. Both of these methods are outlined below. The t o t a l mass of o i l extracted at any time can be predicted by using the bed o i l concentration p r o f i l e s , which are calculated from the model. The mass of o i l extracted (Me) equals the t o t a l mass of o i l in the bed before the extraction begins (M) less the o i l which remains in the bed (Mr) : Me = Me - Mr . . . . [3.5] I n i t i a l l y the t o t a l mass of o i l i s d i s t r i b u t e d evenly throughout the bed and i s given by: M o = ( l-e)Ap s Hx o . . . . [ 3 . 6 ] After the extraction begins however, an o i l concentration p r o f i l e develops in the bed. The t o t a l mass of o i l remaining in a small section (8h) of the bed with concentration i s m^ m . = (l-e)Ap x.6h . . . . [ 3 . 7 ] r i s i 47 and the t o t a l o i l content of the bed i s i=n . • • . [3.8] i=l As 5h approaches zero, the t o t a l amount of o i l i s given by H Mr = f (l-OAP s«Jh . . . . [3.g] By rearranging Eq. [3.6] 0 M o = d -e )P s A Hx0 and substituting into Eq. [3.9] the following result is obtained: M H M. r \ - - i x T J X d b . . . . [3.10] 0 Thus from eq. [3.5] H m r ±1 and upon rearranging M = m - " i x d h e ° Hxo'0 [3.11] 1 H M = M ( 1 I xdh) Hx J o 0 [3.12] The o i l concentration along the bed (x=f{h}) at any time can be predicted using the computer model. This o i l concentration p r o f i l e can then be used in conjunction with Eq. [3.12] to calculate the 'extraction curve' (described in section 3.2) for the computer simulation. A l t e r n a t i v e l y the extraction curve can be generated from the solvent phase o i l concentration. Consider the solvent concentration at the extractor o u t l e t . The amount of o i l c a r r i e d 48 out of the extractor with the solvent phase, (6m), during any time period 8t i s : 6m = myo(t) if • • • • [3.13] m mass flowrate of the solvent [kg/s] y (t) concentration of o i l in the 0 solvent phase at the extractor outlet at a given time [kg o i l / k g C0 2] The t o t a l amount of o i l removed from the bed a f t e r time T i s given by . • Mp = m y* y Q(t)dt • [3.14] 0 The o i l concentration at the extractor outlet can be used in conjunction with Eq. 3.14 to generate the extraction curve. By comparing the predicted extraction curve with the! experimentally obtained curve, the v a l i d i t y of the computer, model can be assessed. Once the v a l i d i t y of the model has been! established, the parameters x, , x 2 and ApK can be adjusted to; obtain the best agreement between the computer generated and experimentally obtained extraction curves. 49 IV. EXPERIMENTAL EQUIPMENT AND PROCEDURES 4.1 Introduction In order to study the phenomenon of Canola seed extraction using s u p e r c r i t i c a l C0 2, special equipment and procedures were developed. The following i s a description of the extraction equipment and the various materials used during the extractions. In addition, the s p e c i f i c treatments which were applied to the seed prior to extraction and the techniques used to determine various q u a l i t a t i v e features of both the seed material and o i l s are discussed. Also presented in t h i s chapter are the a n a l y t i c a l chemical techniques which were used to determine the fatty acid composition and phosphorus content of the o i l s . 4.2 Experimental Equipment Since s u p e r c r i t i c a l f l u i d extraction i s a r e l a t i v e l y new procedure, equipment suitable for th i s work was not available and consequently had to be developed. The basic requirements of thi s equipment were to pump a stream of carbon dioxide at di f f e r e n t pressures, temperatures and flow rates through a sample of seed material. It became apparent that the equipment could be developed by adapting an existing high pressure l i q u i d chromatograph(HPLC). Figure 4.1 i s a photograph of the complete experimental extraction system. FIGURE 4.1 Photograph of the complete experimental s u p e r c r i t i c a l f l u i d extraction system showing: Lauda-Brinkman c i r c u l a t i n g cooler (a), modified HP-1081B l i q u i d chromatograph (b), extract c o l l e c t i o n system (c), volumetric flow-meter (d), dual pen chart recorder (e). 51 4.2.1 High Pressure Liquid Chromatograph The Hewlett-Packard 1081B i s a self-contained, microprocessor controlled instrument incorporating a single reciprocating diaphragm pump and a solvent flow system. Flow rates can be selected to f a l l between 0.1 mL/min and 9.9 mL/min against a maximum pressure of 40 MPa. The chromatograph also incorporates a small oven with a temperature range from ambient to 99 °C. 4.2.2 Solvent Flowpath The general operation of the extraction system i s shown in Fig.4.2. Liquid carbon dioxide from a storage cylinder (a) passes through the shutoff valve (b), a Nupro 7 NM sintered metal f i l t e r (c) and then into the cooled diaphragm pump head (d). The cooled l i q u i d C0 2 then flows through a pressure-flow monitoring device (e) and into the HPLC oven. In the oven the C0 2 temperature can be raised or lowered to the desired value by passing i t through 7 m of sta i n l e s s steel tubing ( f ) . The carbon dioxide then enters the extraction vessel (g). From the vessel, the carbon dioxide passes through a 2 Mm f r i t (h) and into a flow r e s t r i c t e r (i) where i t s pressure i s reduced to about 0.1 MPa. The carbon dioxide then flows through 0.6 m of narrow bore (0.25 mm ID), fused s i l i c a tubing into the sampling section (j) where the l i q u i d solute i s c o l l e c t e d . The C0 2 i s subsequently directed through a wet gas-meter (k) and f i n a l l y vented to the atmosphere. A l l of the tubing within the system was 0.1 mm ID stainless s t e e l and the connections were made using 1/16" sta i n l e s s steel 52 FIGURE 4.2 Schematic diagram of the experimental s u p e r c r i t i c a l f l u i d extraction system, a) solvent (C0 2) reservoir, b) shut-off valve, c) sintered s t e e l f i l t e r , d) diaphragm pump, e) flow and pressure transducer, f) temperature e q u i l i b r a t i o n c o i l , g) extraction vessel, h) sintered s t e e l f i l t e r , i ) temperature con t r o l l e d r e s t r i c t e r valve, j) sample c o l l e c t i o n vessel, k) volumetric flow meter. 53 Swagelok f i t t i n g s . 4.2.3 Pressure And Flow Control The design of the 1081B i s such that i t s microprocessor acts as an interface between the user and the solvent flow system. Thus, no direc t manual control of the pumping system i s possible. Instead, 'requests' are made to the processor via the keyboard. Provided the system i s not in the 'error' or 'not ready' condition, the pump w i l l then be activated. The flow system in the 1081B consists of a small, piston-driven diaphragm pump (Fig. 4.3) and a feedback flow control loop. The flow rate through the pump i s a function of stroke length which, in turn, i s controlled by a small stepper motor. The pressure within the system i s monitored continuously by a pressure transducer in the hexane-filled pulse damper (Fig. 4.2, e). The flow i s also measured in thi s damper by monitoring the pressure-decay rate of each pump pulse. The outputs from the pressure transducer, stepper motor and piston stroke length are used by the microprocessor to calculate the flow rate. This value i s then compared with the flowrate setpoint and instructions are sent to the stepper motor to take appropriate action. This feedback flow control i s executed every 0.6 sec. The flow system can be operated in one of two modes. In the f i r s t mode, the pumping system tracks the flow setpoint while the system pressure i s allowed to ' f l o a t ' ; the system responds only i f the pressure goes above the maximum setpoint value. In the second mode, the system tracks the pressure setpoint 5 4 FIGURE 4.3 Photograph of the HPLC pump with the cooling c o l l a r removed. 55 allowing the flow rate to ' f l o a t ' . If the requested flowrate i s s i g n i f i c a n t l y higher than the maximum flow possible for a given pressure, the system w i l l operate in the second mode. Liquid carbon dioxide at room temperature i s d i f f i c u l t to pump because of i t s high compressibility (Vukalovich and Altunin, 1968). For this reason the C0 2 u t i l i z e d in the present experiments was cooled to approximately -5 °C by c i r c u l a t i n g cold ethylene g l y c o l solution (at about 0.25 L/min) through an aluminum c o l l a r f i t t e d to the pump-head (Fig. 4.4). The temperature of the glycol solution was lowered to -20°C in a Lauda-Brinkman c i r c u l a t i n g bath. The pump (with the cooling c o l l a r ) and the coolant flow l i n e s were insulated with 1 cm thick r e f r i g e r a t i o n foam-rubber tubing (Fleck Bros. Ltd., Appendix I I I ) . 4.2.4 Extraction Vessels Four extraction vessels (autoclaves) were used during the course of thi s work (Fig. 4.5). A l l of the vessels were fabricated 'in-house' from 316 sta i n l e s s steel round stock, with the exception of vessel #4, which was a standard unpacked HPLC column (Supelco Corp., Appendix III) each of the vessels can be envisaged as being b a s i c a l l y a thick-walled tube with a flanged top. Figure 4.6 i s a machine drawing of the medium size vessel. 56 1 ^ > H * «MMM> ig 1 t* | """" \'' \ \ \ \z \ \ A, J , AT, 4.4 Photograph of the aluminum c o l l a r which was used t o the pumphead on the HPLC. The c o o l a n t d e l i v e r y and r e t u r n i n e s , w i t h i n s u l a t i o n , can be seen a t the r i g h t . FIGURE 4 . 5 Photograph of the t h r e e e x t r a c t i o n v e s s e l s used i n the e x p e r i m e n t s . In the photograph the t o p s of the v e s s e l s a r e i n s t a l l e d . 58 TABLE 4.1 Dimensions of the extraction vessels used in the experiments. inside inside vessel vessel vessel diameter length volume wall thickness number (cm) (cm) (cm3) (cm) 1 1.27 8.2 10.4 0.6 2 1.27 11.4 14.4 0.6 3 2.54 8.2 41.6 0.8 4 0.48 30.4 5.4 0.1 The tops of the vessels were fabricated with a 'sealing ridge' around their periphery. A seal between the top and the body of the autoclave was made by securing the top, with i t s 'sealing ridge' to the sealing surface of the vessel, using eight s t a i n l e s s steel cap screws (Figs. 4.7 and 4.8). The bolt size for the small vessel was 10/24 (Unified National Coarse (UNC)) and for the large vessels 1/4" UNC. Since stai n l e s s steel tends to 'weld' to i t s e l f , the bolts were l i g h t l y lubricated with a copper-containing anti-seize compound (Jet-Lube SS-30, Fleck Bros. Ltd., Appendix III) prior to each use. Each of the 10/24 bolts was tightened to 50 i n - l b s . The 1/4" bolts were tightened to 100 i n - l b s . The bottom and top of the extraction vessels were f i t t e d with 1/8" National Pipe Thread (NPT) X 1/16" Swagelok tube f i t t i n g s (Columbia Valve and F i t t i n g , Appendix I I I ) . Solvent flow-line connections to and from the vessels were made via these f i t t i n g s . In Table 4.1 the size s p e c i f i c a t i o n s for each of 59 Swagelok tube fitt ing- 1/16" tube to 1/8" NPT 0.16-/. 1/8 National A Pipe t— 1.27 -H ! I Thread (NPT) * 0 . 3 2 " s - 3 2 Unified National Coarse thread •-inch 2J4 - - — 1/8 Swagelok tube fitting as above " National Pipe Thread FIGURE 4.6 Cross-sectional view of extractor vessel #1 (Table 4.1). A l l measurements shown are in cm, except were indicated. 60 a) TOP INNER VIEW 8-32 Unified National Coarse cap screw ; 7/8 inch A r V vessel top sealing ridge sealing surface vessel body b) SIDE VIEW FIGURE 4.7 Details of the extraction vessel seal, a) view from the inner surface of the extractor top showing the c i r c u l a r sealing ridge, b) cross-sectional view of a portion of the extractor and top, showing the sealing surface and sealing ridge. 61 FIGURE 4.8 Photograph of d e t a i l s of the sealing arrangement on vessel #2. 62 the extraction vessels are l i s t e d . 4.2.5 Flow Restricter The purpose of the flow r e s t r i c t e r i s to maintain the pressure in the extractor at the desired value. The operation of the extraction system i s affected by the c h a r a c t e r i s t i c s of the flow r e s t r i c t e r . In p r i n c i p l e the device i s quite simple, but a valve suitable for use at the desired high pressures and low flow rates was not available and had to be developed. A variety of flow r e s t r i c t e r s were assessed early in the work with limited success. The simplest, a piece of 1/16" OD crimped stain l e s s steel tubing tended to freeze up p e r i o d i c a l l y and plug. Additionally, i t s flow rate could not be varied e a s i l y . Two metering valves, one made by Whitey and the other by Nupro (Columbia Valve and F i t t i n g , Appendix III) were subsequently used, but both had problems with leaking seals and could not be precisely regulated. Eventually i t was found that a Parker MV-200 metering valve (Surrey F l u i d and Power, Appendix III) could be used to adequately control the C0 2 flow through the system. The Parker valve was designed as a high pressure, l i q u i d metering valve. Its operation in the present extraction system was unconventional because the C0 2 flow was not adjusted by turning the valve stem,, but by f i x i n g i t in one position and varying the temperature of the valve thereby causing the o r i f i c e to expand or contract. This modification allowed the flow resistance of the valve to be varied precisely and, as a result, 63 the system pressure and flow rate could be maintained constant and independent of f l u i d v i s c o s i t y . The valve was located in the HPLC oven and mounted in a heated aluminum block (5cm X 3cm X 2cm) (Fig. 4.9). The valve was f i t t e d with two 1/8" NPT X 1/16" Swagelok adapters and i n s t a l l e d so that the stream of s u p e r c r i t i c a l C0 2 entered via the top of the valve and the mixture of gaseous C0 2 and l i q u i d o i l exited via the bottom. The aluminum block was insulated with about 1 cm of glass fib r e wool and heated with a 100w Chromalux cartridge heater (Chromalux Can., Appendix III) i n s t a l l e d in the block. A heating c i r c u i t based on the Radio Corporation of America (RCA) 3059 integrated c i r c u i t (refer to RCA Application Note ICAN 6182, 1978) was constructed and used to control the temperature of the heating block to within ± 0.5°C. In F i g . 4.10 the r e s t r i c t e r valve i s shown i n s t a l l e d in the HPLC oven. 4.2.6 Extract C o l l e c t i o n Systems Upon passing through the r e s t r i c t e r , the carbon dioxide within the system changes from the s u p e r c r i t i c a l to the gaseous state. Commensurate with th i s change of state i s a large decrease in the solvation capacity of the carbon dioxide. Compounds such as o i l s , soluble in the C0 2 on the upstream side of the r e s t r i c t e r , p r e c i p i t a t e downstream of the r e s t r i c t e r . The sampling systems allowed the separation and c o l l e c t i o n of the resulting o i l droplets. Two extract sampling systems were used during the course of the experiments. The f i r s t system consisted of a 25 cm st a i n l e s s steel tube (0.1mm ID) attached to the r e s t r i c t e r by a 1/8" NPT X 64 FIGURE 4.9 Photograph of the MV-200s r e s t r i c t e r v a l v e shown i n the aluminum b l o c k which was used t o r e g u l a t e i t s t e m p e r a t u r e . FIGURE 4.10 Photograph of the oven i n the e x t r a c t i o n system. In the p i c t u r e the e x t r a c t i o n v e s s e l s ( a , b ) , the r e s t r i c t e r v a l v e w i t h h e a t i n g b l o c k ( c ) , the s t a i n l e s s s t e e l f i l t e r s (d,e) and the s i l i c a s a m p l i n g tube ( f ) are v i s i b l e . 66 1/16" Swagelok tube union. The tube was directed to the exterior of the instrument via the oven wall. A piece of folded 7cm f i l t e r paper was secured to the end of the tube using a paper c l i p such that the flow of C0 2 passed into the envelope of the folded paper. This sampling system, although e f f e c t i v e and used during the i n i t i a l experiments, tended to be somewhat cumbersome, and i f used improperly, led to low values for the o i l c o l l e c t e d . The second sampling system was more complex, but allowed sequential o i l samples to be c o l l e c t e d and the carbon dioxide to be measured by a wet test meter. The system, depicted in F i g . 4.. 1 1 , was as follows: a 5 cm long section of 0.1 mm ID stainl e s s steel tubing was connected to the r e s t r i c t e r valve using a Swagelok adapter. The opposite end of the section was connected to one end of a Valco zero dead-volume 1/16" HPLC union (Chromatographic S p e c i a l t i e s , Appendix I I I ) . A 60 cm, 0.2 mm ID fused s i l i c a tube (Supelco Corp. Appendix III) was inserted and sealed using a 'Micro Graphite Ferrule' (Supelco Corp.). The other end of the tubing was then channelled through the oven wall to the exterior of the instrument and directed into the top of the sampling head, where i t passed concentrically down and through a 5 cm, 0.8 mm OD glass c a p i l l a r y tube into the c o l l e c t i o n v i a l (Fig. 4.12). The two phase gas and entrained o i l mixture, flowed through the s i l i c a tube into the sample v i a l where the o i l was deposited. The o i l - f r e e C0 2 then flowed from the v i a l up a glass c a p i l l a r y and entered the wet test meter. 67 1/16" to 1/8" NPT Swagelok tube fitting top of sampling head 3/4" Unified National Coarse thread (UNC) main body of sampling head 1/16" to 1/16" NPT Swagelok tube fitting TOP VIEW OF MAIN SAMPLING HEAD BODY \ 90° F u M d S * c a Mtam F w n r i n 1/16" OD tube from restricter valve fused silica tube 1.9 CCD-Gh-m 111" UNC thread lower section of sampling head (Teflon) '0' ring 0.15 ID 0.40 OD 8-32 UNC thread Valco 1/16" male tube connector 1/16" OD glass tube Teflon septum top of sampling bottle fused s i l ica capillary FIGURE 4.11 Cross-sectional diagram of the sampling head with a top view of the main body of the sampler. A l l measurements shown are in cm unless otherwise s p e c i f i e d . A l l of the head which are shown were made from 316 except where otherwise indicated. The inset 1/16" Valco tube union. The union was used to tube from the r e s t r i c t e r to the fused s i l i c a sampling system. the components of sta i n l e s s s t e e l , i s a diagram of a couple the exit i n l e t tube of the 0, L^ LL ( IIII  m  iMiiiuiiiilTTTTTnnTTnTTlTnnnm i.l.i.l.i.l.i.lfhlihl.ijUl'mliiilij i l FIGURE 4.12 Photograph of the extract c o l l e c t i o n system which was used to separate and c o l l e c t the o i l from the C0 2. In the picture the s i l i c a tube (a), the c o l l e c t i o n v i a l (b) and the C0 2 exhaust f i t t i n g (c) are v i s i b l e . 69 The sampling head (Fig. 4.11) consisted of a 5 cm long X 3 cm OD stain l e s s steel cylinder with openings for the s i l i c a tube, the glass c a p i l l a r y tube and the gas exit tube. Although the cylinder was constructed in two parts, a 2.5 cm '0' ring placed between the sections ensured that the unit was gas-tight. A 0.4 mm OD '0' ring was also used to ensure a seal between the sampling head and the s i l i c a tube. In order to allow the glass c a p i l l a r y a certain amount of f l e x i b i l i t y , i t was secured in a Teflon plug inserted into the bottom of the sampling head. 4.2.7 Equipment Calibration The extraction equipment was checked p e r i o d i c a l l y to ensure that the solvent flowrate, oven temperature and system pressure were accurately displayed. The c a l i b r a t i o n procedures were performed in accordance with Section 5 of the Hewlett Packard 1081B service manual. The accuracy of the parameters are l i s t e d in Table 4.2. TABLE 4.2 Errors associated with the various system parameters as determined experimentally. parameter error oven temperature <±2% extractor pressure ±2% C0 2 flowrate <±3% t o t a l C0 2 volume <±0.5% 70 4.3 Materials 4.3.1 Carbon Dioxide The carbon dioxide that was used in the experiments was obtained from pressurized steel cylinders supplied by Medigas P a c i f i c (Appendix I I I ) . Each cylinder held 30 kg of carbon dioxide which could be withdrawn in l i q u i d form through a siphon tube within the cylinder. The amount of carbon dioxide remaining in the cylinder at any time was determined by subtracting the tare weight of the cylinder (marked on the wall of the cylinder) from i t s actual weight. Approximately 200-300 h of system run time were obtained with each cylinder. The carbon dioxide s p e c i f i c a t i o n s are l i s t e d in Table 4.3. TABLE 4.3 Specifications of commercial siphon grade carbon dioxide USP siphon grade carbon dioxide >99% pure <10 ppm CO <1 ppm H2S <5 ppm N02 . <0.3 ppm COC12(phosgene) <5 ppm S02 <25 ppm H20 4.3.2 Seed Samples Two v a r i e t i e s of Rapeseed are currently grown in Canada, the "Polish" species ( Brassica campestris ), and the "Argentine" species ( Brassica napus ). Genetic stra i n s , or c u l t i v a r s within these v a r i e t i e s are numerous ( and a l l are 71 named eg. "Midas", "Torch", ,"Regent,"Candle"), due to continuous breeding programs designed to improve both the agronomics of the plant as well as i t s o i l q u a l i t y . Unlike the seed o r i g i n a l l y introduced into Canada in 1943, most of the rapeseed presently grown in Canada, contains only small amounts of erucic acid and glucosinolates. This modern version of the seed i s termed "Canola". One c u l t i v a r from each of the two v a r i e t i e s of Canola was used in t h i s work. The f i r s t , a dark colored seed c a l l e d "Regent", i s a member of the Argentine species and was obtained from a commercial processor (CSP Foods, Appendix I I I ) . The second c u l t i v a r "Candle", i s a member of the Polish variety and was obtained from the Food Science Department of the University of B r i t i s h Columbia. 4.3.3 O i l Samples Four Canola o i l samples were used in the experiments. Three of these samples, "crude", "acid degummed", and "refined" were obtained from CSP Foods Ltd. A l l were extracts of the "Regent" c u l t i v a r . The fourth sample was a commercially refined bleached and deodorized o i l ("Scotchbuy") available in l o c a l supermarkets. The crude o i l sample was a mixture of the o i l obtained from 'pressing' the seed and the hexane extract. The acid degummed o i l was produced by subjecting the crude o i l to an acid treatment procedure, during which phospholipid material i s removed from the o i l (sec. 1.2.3). The refined o i l i s an acid degummed product, e s s e n t i a l l y free of fatty acids, phosphatides 72 and proteinaceous and mucilagenous material. 4.3.4 Simple Tr i g l y c e r i d e s T r i a c y l g l y c e r o l s , more commonly referred to as t r i g l y c e r i d e s , are glycerol esters derived from several d i f f e r e n t carboxylic acids. The number of di f f e r e n t t r i g l y c e r i d e s i s very large and many are commercially available. Three were obtained for thi s research project and used to represent the t r i g l y c e r i d e s found in Canola o i l . A l l of the t r i g l y c e r i d e s were obtained from Nu-Chek Prep Ltd. (Appendix I I I ) . The sp e c i f i c a t i o n s for each are given in Table 4.4. TABLE 4.4 Specifications of the t r i g l y c e r i d e samples used during the experiments name acid t r i p a l m i t o l e i n C16:1 t r i l o l e i n C18:1 molecular molecular formula weight purity t r i - 1 1 -eicosenoin C20: 1 C 5 1 H 9 2 0 6 C 5' 7H 1.0 n06 C.6 3,H ! 1606 801 885 969 >99% >99% >99% When not in use, the t r i g l y c e r i d e s were stored at -20 °C. 73 4.4 Seed Treatment Methods Prior To Extract ion 4.4.1 Introduct ion The Canola seed used during the course of t h i s work was usually treated prior to extraction in one or two ways. In one of the treatments the seed was physically ruptured by. various techniques. In the other treatment the physically ruptured seed was extracted with a solvent to reduce i t s o v e r a l l o i l content. This material in turn, was used to provide data on o i l s o l u b i l i t y in the C0 2 as a function of seed-oil concentration. 4.4.2 Seed Crushing The seed was crushed using a 20 cm mortar and pestle. Approximately 10 g of whole seed were placed into the mortar. Crushing of the seed was done by hand over a f i v e minute period. The crushed seed material was free flowing and contained p a r t i c l e s ranging from approximately 0.1 mm to 0.5 mm. For d e t a i l s see section 5.4.2. 4.4.3 Seed Chopping Finely chopped seed material was produced by placing 50 g of whole seed into a 2L 'Osterizer blender', model "Cyclo-Trol-Ten" for 5 min on the "blend" setting. The seed material produced in this manner was more homogeneous than the crushed material. Individual p a r t i c l e s ranged in size from about 0.05 mm to 0.1 mm (sec. 5.4.2). 74 4.4.4 Flaking And Cooking Of Seed The flaked seed material was procured from a commercial seed processor (CSP). The material was produced by forcing whole seed through a series of r o l l e r m i l l s . During the flaking process the seed i s crushed and flattened (rupturing most of the seeds' c e l l walls) which renders the material more susceptible to solvent extraction. Typical thicknesses of the seed flakes range from 0.2 - 0.5 mm. After f l a k i n g , the seed i s subjected to a short heating process (90 C, 0.5 h). This serves to inactivate certain undesirable enzymes and to enhance the e x t r a c t a b i l i t y of the material (Teasdale and Mag, 1983). Seed treated by t h i s method i s referred to as having been"cooked". For a size d i s t r i b u t i o n of the p a r t i c l e s see Sec. 5.4.2. 4.4.5 Pressure Rupturing Pressure rupturing i s a technique which involves placing a b i o l o g i c a l material in a high pressure gas for an appropriate length of time ( t y p i c a l l y 60 min). During t h i s time the pressurized gas penetrates the material. The pressure i s then quickly released and the contained gas expands, rupturing the c e l l s . The procedure for preparing pressure ruptured seed was as follows. Samples of whole seed ( t y p i c a l l y 12 g) were placed into extraction vessel #3. The extractor was then placed in the HPLC oven for about- 30 min and allowed to equi l i b r a t e at 55°C. The pump was then turned on for about 5 minutes, during which time the vessel pressure rose to 36.0 MPa. Because the r e s t r i c t e r was closed, there was no flow of carbon dioxide through the vessel. 75 At the end of one hour, the top Swagelok f i t t i n g on the extractor was loosened thereby allowing the carbon dioxide to escape rapidly. The ruptured seed material was then removed from the vessel. The pressure-ruptured seeds exhibited a wide variation in p a r t i c l e size ranging from 1 mm (whole seed) to 0.05 mm; the majority of the fragments was greater than 0.5 mm. 4.4.6 P a r t i a l Extraction Of Seed The natural concentration of o i l in Canola seed is about 0.6 g of o i l / g o i l - f r e e seed. Certain experiments however c a l l e d for seed material with an o i l content less than t h i s . Furthermore, i t was required that these samples exhibit a range of concentrations and that the concentration of each be known. Two methods were used to produce these samples. The f i r s t method described i s a standard solvent extraction procedure which u t i l i z e d hexane as the extracting solvent. In the second method the seeds were extracted under high pressure using a stream of s u p e r c r i t i c a l carbon dioxide. 4.4.6.1 P a r t i a l Extraction With Hexane Twelve, ceramic, extraction thimbles (2 cm) were weighed to the nearest 0.001 g. Into each of the thimbles were placed 10 g of crushed seed material. A l l the thimbles were then transferred into a Goldfisch Extraction unit and extracted with hexane. After allowing the extraction to proceed for various lengths of time (depending upon the degree of extraction desired), the thimbles were removed (in pairs) and the seed material within 76 each thimble was dried overnight at 40 °C. The seed material in each thimble was then mixed and subjected to a t o t a l o i l determination (Section 4.6.1) in order to est a b l i s h the degree of extraction. The average value between pairs was used as the working o i l concentration. In t h i s manner samples were prepared ranging in o i l concentrations from 0.05 g o i l / g o i l - f r e e seed to 0.5 g o i l / g o i l - f r e e seed. 4.4.6.2 Pa r t i a1 Extract ion With Carbon Dioxide The apparatus that was used to extract the seed material with C0 2 i s described by Campbell (1983). It consisted of a 1 L temperature controlled, s t a i n l e s s steel vessel, an Aminco double ended diaphragm compressor, a r e s t r i c t e r valve and a C0 2 source. During the extraction, C0 2 flowed from the source tank into the compressor where i t s pressure was raised from 6 MPa to 36 MPa. From here the C0 2 flowed through the heated autoclave (entering via the bottom and exiting through the top) and then through the pressure r e s t r i c t e r valve. The system pressure was a function of the C0 2 flow rate and the r e s t r i c t e r valve o r i f i c e s i z e . A variable speed motor, magnetically coupled to a s t i r r i n g mechanism within the autoclave allowed the vessel contents to be continuously mixed. The extent of the extraction was determined by the length of time the seed material was exposed to the C0 2 gas stream. A t y p i c a l extraction run was as follows: with the autoclave at working temperature (60°C), a sample (100 g) of crushed seed material was placed in the bottom of the vessel. The top of the 77 autoclave (and the s t i r r i n g mechanism) were set in place and secured. The s t i r r i n g rate was t y p i c a l l y 120 rpm. The vessel was then flooded with C0 2 and the compressor activated. By adjusting the r e s t r i c t e r valve, the pressure within the system could be maintained at 36.0 MPa for the required length of time. At the end of the run, the system was depressurized ( over 1 h ), and the seed material removed. To e s t a b l i s h the degree of extraction which had taken place, the seed was subjected to a t o t a l o i l determination (Sec. 4.6.1.). The samples produced in t h i s manner had o i l concentrations similar to those discussed in the previous section. 4.5 Experimental Extraction Procedures 4.5.1 Vessel Loading 4.5.1 . 1 Seed Material At the beginning of each experiment, a weighed amount of seed material (normally 4 g for vessel #1, 7 g for vessel #2, 12 g for vessel #3 and 1.5 g for vessel #4) was placed in the extractor. Fine-spun glass wool was placed at both ends of the vessel to prevent small p a r t i c l e s from entering the tubing. Before closing the top of the autoclave, both the extractor and the sealing surface were wiped, free from d i r t using a chloroform-wetted tissue. In addition, a l l f i b r e s of glass wool protruding from the vessel were cut and removed. It was noted that even a single strand of the glass prevented the completion of the seal and could permanently destroy i t . 4.5.1.2 Liquid O i l Samples Before using the autoclaves with the l i q u i d Canola o i l or 78 the t r i g l y c e r i d e samples, the vessels were f i r s t p a r t i a l l y f i l l e d with fine - (0.3mm) Ottawa sand. The l i q u i d samples were then deposited on the Ottawa sand. This procedure was used for two reasons: 1) the l i q u i d on a matrix of Ottawa sand presented a larger surface area available for mass transfer than would an equal mass of l i q u i d in the empty vessel; 2) the l i q u i d on the Ottawa sand could release dissolved C0 2 during the de-compression stage of the extraction experiments more e f f e c t i v e l y thereby reducing the r i s k of the l i q u i d foaming out of the vessel. The procedure by which the l i q u i d samples were loaded into the vessels was as follows: a known, weight of Ottawa sand--enough to approximately h a l f - f i l l each vessel—was placed in the respective autoclave On top of a small plug of glass wool. The l i q u i d sample, t y p i c a l l y in the range of 1 to 2 g, was then pipetted on top of the beads and allowed to 'sink in' over a period of ten minutes. By performing the procedure in a glass test-tube, i t was established that the method results in an even d i s t r i b u t i o n Of o i l throughout the bed. The mass of l i q u i d transferred was determined to the nearest 0.001 g by weighing the pipette before and aft e r the l i q u i d transfer. Before f i t t i n g the top of the vessel, another plug of glass wool was inserted. During extraction the C0 2 flow through the vessel was from bottom to top. 79 4.5.2 Equipment Startup Approximately 2 h prior to the start of an experiment, the cooler and HPLC were switched on. The cooler was t y p i c a l l y set to -20 °C. During this pre-extraction period the autoclave was washed with chloroform and then acetone. Additionally, the flow system downstream of the vessel, including the f r i t s and the extract c o l l e c t i o n system, was flushed with about 20 mL of a 1/1 chloroform/methanol solution. In order to remove any l a s t trace of solvents, the system was assembled and flushed with s u p e r c r i t i c a l C0 2 for 1 h. Following the system flushing the vessel was placed in the oven and the feed and exit tubes were connected. The shut-off valve on the C0 2 cylinder was then opened and the r e s t r i c t e r valve set to allow a carbon dioxide flow of approximately 50 mL/min at Standard Temperature and Pressure (STP). The temperature of the oven was set to the desired value (usually 55 °C) while the r e s t r i c t e r valve, was usually set to 65°C. The system remained in thi s 'pre-run' state for approximately 30 min. By placing a temperature probe in the bed i t was established that, t h i s period of time was s u f f i c i e n t for temperature e q u i l i b r a t i o n . At the end of thi s period, pressure and flow setpoints were entered. In the majority of experiments the system was required to operate in a 'pressure tracking' mode (Sec. 4.2.4). In order to ensure t h i s , the desired experimental pressure was entered while the flow rate setpoint (usually 9.9 mL/min) was adjusted to greatly exceed the value which corresponded to t h i s pressure. The pump was next switched on by pressing the "pre-run" 80 button on the HPLC. Once the pump was activated, the chart recorder, which provided a permanent record of the C0 2 flow and pressure, was turned on and the coarse control on the r e s t r i c t e r adjusted. This adjustment set the flow of carbon dioxide through the vessel at between 0.5 and 1.5 g /min. 4.5.3 Extract Sampling The sampling procedure normally began immediately aft e r the system flow s t a b i l i z e d , i . e . about 5 to 10 min after a c t i v a t i n g the pump. Two sampling procedures were employed. In the f i r s t of these a piece of folded 10 cm f i l t e r , paper was used to c o l l e c t the o i l . The mass of o i l c o l l e c t e d was determined by weighing the paper before and after using i t . The amount of C0 2 used during each sampling period was determined from the d i g i t a l flow gauge on the instrument. The f i r s t sampling procedure was used in the i n i t i a l stage of the project. This stage included the i n i t i a l experiments in which the s o l u b i l i t y of Canola o i l in C0 2 was determined as a function of pressure and temperature. These experiments were repeated subsequently using the second sampling procedure. In the second sampling procedure the apparatus shown in Fi g . 4.12 was used. The method consisted of f i t t i n g , pre-weighed, 1.8 mL., glass v i a l s on the sampling tube for varying amounts of time. From the 'before' and 'after' weights of each v i a l , the o i l c o l l e c t e d during that period could be determined. The volume of gaseous carbon dioxide that passed through each v i a l during the sampling period was determined using the wet 81 test meter. The corresponding mass of C0 2 was calculated from the C0 2 molar volume corresponding to the temperature of the wet test meter. As each v i a l was removed and replaced with a new one, a mark was made on the chart recorder. From the marks and known chart speed, the length of time that each bottle was exposed to the flow of gas could be determined, as well as the flowrate of the C0 2 during the sampling period. 4.5.4 S o l u b i l i t y Determinations The average o i l concentration in the C0 2 at the extractor outlet, at any int e r v a l during the extraction, can be determined from the mass of o i l c o l l e c t e d during the int e r v a l and the mass of C0 2 passed. If the o i l concentration measurement is made during the i n i t i a l linear portion of the extraction curve (under certain conditions), t h i s measurement w i l l represent the s o l u b i l i t y of the o i l in the C0 2 at that pressure and temperature. Two procedures were used to ensure that t h i s concentration measurement did in fact represent the o i l s o l u b i l i t y . In the f i r s t procedure, two separate extractions were performed on i d e n t i c a l samples of seed under i d e n t i c a l conditions of pressure, temperature and carbon dioxide flowrate. The only difference between the two extractions was that in the second extraction the length of the seed bed was increased. If the slope of the linear part of the second extraction curve was .identical to the f i r s t , i t indicated that the increased bed length did not increase the concentration of the o i l in the C0 2 phase. This in turn suggested that the C0 2 was saturated and 82 that the slope of the curves could be used to calculate the o i l s o l u b i l i t y . In the second procedure the extraction was stopped during the 'linear phase' and the bed of seed divided into sections and each section analyzed for i t s o i l content. If the o i l content of the section of seed nearest the extractor outlet was s t i l l equal to i t s o r i g i n a l concentration, i t indicated that the carbon dioxide was saturated prior to reaching t h i s segment. This in turn i s proof that the slope of the linear portion of the extraction curve represented the o i l s o l u b i l i t y . 4.5.5 Equipment Shutdown At the completion of an experiment, the pump was turned off by setting the system into the "standby" condition and the shut-off valve on the C0 2 cylinder was closed. The system t y p i c a l l y depressurized to ambient over a period of 1 h at which time the equipment was turned off and the extraction vessel was removed. 4.6 A n a l y t i c a l Procedures 4.6.1 Total Seed O i l Determinations The t o t a l amount of o i l within the seed was determined by placing pre-weighed amounts of the seed into c e l l u l o s e extraction thimbles and extracting them with hexane for 8 h at 60°C as outlined in International Union of Pure and Applied Chemistry (IUPAC) Method I.B.2 (Paquot, 1979) The extracted seed samples were then removed and the solvent evaporated at 40°C for an additional 24 h in a small, air-convection oven. By subtracting the weight of the dried seed tissue from i t s weight 83 prior to extraction, the t o t a l amount of hexane extractable o i l within each sample could be determined. At the same time the hexane extracts from the samples were evaporated at room temperature for 24 h and the c o l l e c t e d o i l determined from the before and after weight of the receiving f l a s k s . Each of the extractions was done in t r i p l i c a t e . The f i n a l accepted value of o i l in the seed was the arithmetic mean of the three. 4.6.2 Scanning Electron Microscopy A number of examinations of Canola seed, in a crushed and whole state, were performed using the scanning electron microscope (SEM). The instrument used was an Etec Autoscan with a maximum resolution of 20 nm. The methods by which the seed samples were c o l l e c t e d and prepared prior to th e i r examination are described below. 4.6.2.1 Sample Coll e c t i o n In order to investigate . the seed material at d i f f e r e n t stages of extraction, samples of unextracted, p a r t i a l l y extracted and f u l l y extracted seed were f i r s t obtained. This was done in the following manner. A 4 g sample of flaked seed was placed in the small autoclave and extracted with C0 2 at 36 MPa and 55 °C. Approximately halfway through the linear phase of the extraction, the experiment was stopped and the system allowed to depressurize slowly. The bed of seed was then divided into f i v e equal sections with each section representing a d i f f e r e n t stage of the extraction. 84 4.6.2.2 Sample Preparation For SEM A l l of the seed samples, whether whole or crushed, were f i r s t coated with an a l l o y of gold-palladium in a Technics Hummer (model 4) sputter coater. The maximum plating current which was used during the coating was 4 mi 11iamperes. A l l of the samples were coated for three, 1 min i n t e r v a l s , which gave a coating thickness of approximately 2000 nm. 4.6.3 Seed-bed Sectioning Method At the completion of certain experiments the contents of the extraction vessel were sectioned into several equal volume portions. This was done in the following manner. The l i d of the vessel was removed and the glass wool on top of the seed material was pulled out. A 0.6 mm glass tube attached to a f l e x i b l e Tygon hose, which in turn was connected to a 25 ml vacuum flask, was next inserted into the vessel. By applying a vacuum to the flask, the seed material could be removed from the vessel layer by layer without being mixed. A small plug of glass wool prevented the seed material from being swept into the main vacuum l i n e . 4.6.4 Fatty Ac id Analysi s The fatty acid analysis was performed in two steps. In the f i r s t step the fatty acid moieties were cleaved from the t r i g l y c e r i d e s and simultaneously converted to their methyl esters. In the second step, the methyl esters were i d e n t i f i e d and quantified using gas chromatography. 85 4.6.4.1 Transesterif icat ion The fatty acid composition of t r i g l y c e r i d e s i s one useful c r i t e r i o n for characterizing animal and vegetable o i l s (Ackman, 1983). Normally, t r i g l y c e r i d e s are f i r s t saponified and the resulting fatty acids converted to their methyl esters prior to GC analysis. By contrast, the t r a n s e s t e r i f i c a t i o n procedure converts t r i g l y c e r i d e s d i r e c t l y into their component fatty acid methyl esters (Knapp, 1979). The t r a n s e s t e r i f i c a t i o n procedure used in t h i s work is based on a method developed by Shehata et a l . ( l 9 7 0 ) . The reaction was c a r r i e d out at room temperature using sodium methoxide(NaOCH3) in a single-phase mixture of methanol, petroleum ether and d i e t h y l ether. When the reaction was complete, the mixture was forced to separate into two phases (by the addition of a small amount of water); the fatty acid methyl esters dissolve in the petroleum ether phase. The petroleum ether, and the dissolved methyl esters were then removed from the reaction v i a l by pipette. The s p e c i f i c d e t a i l s of the technique are as follows. A 0.5 N solution of sodium methoxide (Fisher S c i e n t i f i c , Appendix III) in methanol was prepared by adding 0.675 g of the anhydrous powder to 25 mL of absolute methanol. To another 25 mL flask, containing 8.5 mL of anhydrous di e t h y l ether and 5.0 mL of petroleum ether, 12.5 mL of the sodium methoxide solution were added. The e s t e r i f i c a t i o n reagent was prepared on a weekly basis and stored at -10°C in a capped fl a s k . The e s t e r i f i c a t i o n reaction and associated vessels were 1.8 mL b o r o s i l i c a t e - g l a s s 86 screw top v i a l s equipped with Teflon septa. Disposable pipettes were used to transfer l i q u i d s between v i a l s . A l l weights were determined to within ±0.001 g by using a Mettler PT-320 balance. The e s t e r i f i c a t i o n procedure was as follows: between 15 and 50 mg (1 to 3 drops) of the o i l was quantitatively transferred to the reaction v i a l . This was done by weighing the o i l -containing transfer pipette both before and after transfer. The e s t e r i f i c a t i o n reagent (1 mL) was then added to the v i a l . The v i a l was capped, shaken to dissolve the o i l , and allowed to stand for two minutes. After this time, the mixture was forced to separate into two phases by adding 0.5 mL of petroleum ether and one drop of water. The resulting mixture was then shaken for about 30 s to f a c i l i t a t e • t h e transfer of the methyl esters into the petroleum ether phase. The reaction v i a l and i t s contents were then centrifuged for 10 min in order to remove any suspended sodium methoxidei After centrifuging, the top 2 mm of the petroleum ether layer, which contained the methyl esters, were transferred by pipette for further analysis to another 1.8 mL v i a l containing a few grains of anhydrous sodium sulphate (Na 2SO a). 4.6.4.2 Gas Chromatographic Procedure The fatty acid ester solution was analyzed using a Perkin-Elmer (PE) Sigma 2 gas chromatograph (GC), connected to a PE Sigma 10 data system. The esters were separated with a 6 f t . X 1/8 i n . s t a i n l e s s steel column packed with SP-2330 on 100/120 mesh Chromosorb WAW. The column was obtained pre-packed from 87 Supelco Corp. (Appendix I I I ) . A l l GC analyses were performed in accordance with the conditions shown in Table 4.5. TABLE 4.5 Gas chromatograph parameters for the fatty acid methyl ester analyses. Using the above column and conditions, baseline separation of each of the fatty acid esters could be obtained. The integration of the peaks was thus baseline to baseline. 4.6.4.3 Validation Procedure In order to e s t a b l i s h the effectiveness and precision of the e s t e r i f i c a t i o n procedure, a mixture was prepared, consisting of equal weights of t r i p a l m i t o l e i n , t r i o l e i n , tri-11-eicosenoin and t r i e r u c i n . The procedure described in the above section was used to e s t e r i f y 25 mg and 75 mg of the mixture. The resultant esters were then analyzed by gas chromatography using previously determined response factors, and the results compared to calculated r a t i o s . It was determined using t h i s method that the a n a l y t i c a l procedure worked equally well with each of the above t r i g l y c e r i d e s . column SP-2330 on 100/120 mesh Chromosorb WAW column temperature detector temperature injector temperature c a r r i e r gas c a r r i e r gas flow sample size detector 200 °C isothermal 250 °C 250 °C helium 20 cm3/min 1 mL flame ionization 88 4.7 Phosphorus Analysis 4.7.1 Introduction The procedure by which o i l samples were analyzed for their phosholipid content was based on the .work of Duck-Chong (1979). The method consists of two parts. In the f i r s t part the sample i s ashed at high temperature in the presence of magnesium nit r a t e Mg(N0 3) 2 6H20. This step i s required to convert the phospholipids to phosphate and to 'burn o f f non-phospholipid organic, material. In the second step the residual phosphate i s determined using a standard colorimetric procedure and known concentration reference standards. 4.7.2 Digestion Procedure T y p i c a l l y , 7 mg of each o i l sample was ashed. This mass of o i l was s i g n i f i c a n t l y more than that s p e c i f i e d by Duck-Chong (1979). Consequently her method was modified to allow for the increased sample si z e . The modified ashing procedure was then tested using known concentration standards, consisting of a mixture of pure t r i g l y c e r i d e s and pure phosphatidyl choline (Avanti Polar-Lipids, Appendix I I I ) . The modified ashing procedure was as follows: a 7 mg o i l sample was placed into the bottom of an acid-washed 8cm X 1cm Pyrex test tube using a 10/xL pipette. The mass of o i l transferred was determined to the nearest 0.01 mg by weighing the pipette before and after transfer using a Mettler H20T balance. Following t h i s , 30/xL of magnesium n i t r a t e solution, 10%(W/V) Mg(N0 3) 2 6H20 in methanol were then added. The test tube and mixture were then gently 89 heated over a Bunsen burner flame. Digestion of the mixture was achieved by lowering the tube into the hottest region of the flame for several seconds. During th i s time the o i l mixture charred and evaporated, coating the lower portion of the test tube. It was found that i t was necessary to burn this black residue away since i t s presence interfered with the subsequent photometric technique. This was done by further heating the tube for several minutes. After digestion the tubes t y p i c a l l y contained small quantities of white powder. After allowing the tubes to cool for a few minutes, 1 mL of 1 M HC1 was added. Each of the tubes was then covered with a glass marble to minimize evaporation and heated for 15 minutes in a water bath at 90-95 °C. 4.7.3 Phosphate Standards Six standards containing from O(blank) to 0.5/ig of potassium phosphate (KH 2P0 4) were prepared according to the method of Duck-Chong(1979). These solutions were then used to prepare a phosphorus standard curve with which the o i l samples were compared. 4.7.4 Photometric Procedure The amount of phosphorus in the standards and the sample solutions was determined according to the method of Duck-Chong (1979). In t h i s procedure the solutions are reacted with a malachite green-ammonium molybdate reagent. The amount of phosphate-dye complex formed during the reaction, which can be determined by measuring the absorbance of the solution at 650 90 nm, i s an indication of the amount of phosphorus present. In practice, the phosphorus content of each sample was determined by f i r s t measuring i t s absorbance and then comparing the absorbance with that of the standard solutions. A Beckman (model 124) spectrophotometer was used for a l l absorbance measurements. 4.7.5 Phospholipid Calculations The procedure described above detects only the phosphorus present in the samples. In order to be able to calculate the amount of phospholipid present, the r a t i o between the phosphorus and the phospholipid must be known. In p r i n c i p l e , t h i s r a t i o can be calculated from the molecular weight of phosphorus and phospholipid. However, since many types of phospholipids occur in b i o l o g i c a l tissues, the r a t i o in practice i s between the molecular weight of phosphorus and an 'average' molecular weight of phospholipids. The average molecular weight of phospholipids in the Canola o i l was calculated from the concentration and molecular weight of the major phospholipids known to occur in the o i l (Table 4.6), i . e . mol wt(x) = 0.48(781) + 0.2(873) + 0.09(769)/0.77 = 800 The r a t i o of phosphorus to phospholipid i s 31/800 = 0.0386; i. e . 3.86wt% of phospholipid in Canola o i l i s phosphorus. Thus the mass of phospholipid in the o i l can be estimated by multiplying the mass of phosphorus as determined by the 91 aforementioned procedure by 1/0.0386 or 25.9. TABLE 4.6 Major phospholipid components of Canola o i l (Sosulski et a l . , 1981). phospholipid mol form mol wt % occurrence phosphatidyl Ca«H 8 00 8NP 781 48 choline phosphatidyl C»7H 7o0 1 3P 873 20 in n o s i t o l phosphatidyl C a 3H Bo0 8NP' 769 9 ethanolamine 4.7.6 Detect ion Limits Of Procedure The minimum amount of phosphorus which could be determined using the above procedure was found to be 0.05Mg. Since the minimum o i l sample size was 7 mg, the l i m i t s of detection were 0.05yg/7 mg or 7 ppm of phosphorus. This value corresponds to a minimum phospholipid detection l i m i t of approximately 0.02%. 92 V. RESULTS AND DISCUSSION 5.1 Introduction The present section i s divided into three p r i n c i p a l parts. In the f i r s t part, information about select physical aspects of the extraction process i s presented. This information includes such items as the s o l u b i l i t y of Canola o i l in C0 2 as a function of system temperature and pressure, and the effectiveness of seed pre-treatment methods. Part Two deals with chemical aspects of the extraction process, chemical composition of C0 2 extracts of Canola seed and synthetic mixtures of pure t r i g l y c e r i d e s . In the t h i r d part the predictions made by the computer model are compared with experimental data and the v a l i d i t y of the model i s discussed. 5.2 O i l S o l u b i l i t y as a Function of Pressure and Temperature The s o l u b i l i t y of Canola o i l in C0 2 at di f f e r e n t temperatures and pressures was determined from the corresponding extraction curves using the procedure outlined in Sec. 4.5.4. A l l s o l u b i l i t y experiments were repeated at least twice. Add i t i o n a l l y , material balances were ca r r i e d out on the extractions and closed. The data points shown on the figures are the average value obtained from the experiments. The error bars shown on the figures represent the maximum and minimum slopes of the l i n e s which could be drawn through the i n i t i a l portion of the extraction curves. In F i g . 5.1 the s o l u b i l i t y of Canola o i l in C0 2 i s plotted as a function of pressure at d i f f e r e n t temperatures. The figure 93 indicates that as the pressure of the C0 2 increases, the o i l s o l u b i l i t y also increases. In Fig.5.2 the s o l u b i l i t y of the o i l i s plotted as a function of temperature, at d i f f e r e n t pressures. From t h i s figure i t i s evident that the e f f e c t of temperature on the s o l u b i l i t y of the o i l in the C0 2 changes with pressure. At low pressures, the o i l s o l u b i l i t y decreases with increasing temperature, whereas at higher pressures the s o l u b i l i t y curve exhibits a maximum. This temperature e f f e c t i s not unusual and has been reported for napthalene dissolved in s u p e r c r i t i c a l ethylene (Williams, 1981) and in C0 2 (de F i l i p p i , 1982). In Fig,. 5.3 the s o l u b i l i t y of Canola o i l i s plotted as a function of C0 2 density, at d i f f e r e n t temperatures, while in F i g . 5.4 the s o l u b i l i t y is plotted as a function of temperature at d i f f e r e n t C0 2 d e n s i t i e s . These plots show that a simple monotonic relationship exists between o i l s o l u b i l i t y and C0 2 density. The complex re l a t i o n s h i p between o i l s o l u b i l i t y and pressure (at d i f f e r e n t temperatures) can be explained by re f e r r i n g to F i g . 5.5. As indicated in t h i s figure, a r i s e in temperature, at constant pressure, leads to a decrease in C0 2 density. On the other hand, a r i s e in temperature also leads to an exponential increase in the vapor pressure of the o i l (Formo, 1979; Peter and Brunner, 1978). Near the c r i t i c a l point of the C0 2 (Fig. 5.5), the density changes rapidly with temperature. A small temperature change in t h i s region may lead to a large change in C0 2 density and a commensurate change in o i l s o l u b i l i t y . 94 i i i— i—i—i—i—i—i—i—i—i—i—r 10.0 14.0 18.0 22.0 26.0 30.0 34.0 38.0 PRESSURE (MPa) FIGURE 5.1 S o l u b i l i t y of Canola o i l in C0 2 as a function of pressure at four temperatures. Conditions: vessel #2, C0 2 flow rate 0.7 g/min, 7 g flaked seed. 95 l i—i—i—i—i—i—i—i—i—i—i—i—i—r 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 TEMPERATURE (°G) FIGURE 5.2 S o l u b i l i t y of Canola o i l in C0 2 as a function of temperature at four pressures. Conditions: vessel #2, C0 2 flow rate 0.7 g/min, 7 g flaked seed. 96 FIGURE 5.3 S o l u b i l i t y of Canola o i l in C0 2 as a function C0 2 density at four temperatures. Conditions: vessel #2, C0 2 flow rate 0.7g/min, 7 g flaked seed. 97 FIGURE 5.4 S o l u b i l t y of Canola o i l in C0 2 as a function of temperature at d i f f e r e n t C0 2 d e n s i t i e s . Conditions: vessel #2, C0 2 flow rate 0.7 g/min, 7 g flaked seed. FIGURE 5.5 Density of carbon dioxide as a function of pressure at d i f f e r e n t temperatures. The c r i t i c a l point (CP) of the C0 2 i s indicated on the diagram. (Newitt et a l . , 1956) 99 At higher pressures however, the same temperature change has a smaller e f f e c t on the f l u i d density. In t h i s case, the increase in the vapor pressure of the o i l may more than offset the decreased solvent capacity of the f l u i d due to i t s decreased density. The net effect results in an o v e r a l l increase in s o l u b i l i t y . These results support the generally accepted 'rules' that: i) the solvent power of a s u p e r c r i t i c a l f l u i d w i l l increase with density at a given temperature and i i ) the solvent power of a s u p e r c r i t i c a l f l u i d increases with temperature at constant density (Brogle, 1982). 5.2.1 P r a c t i c a l Implications of S o l u b i l i t y Data Figure 5.2 indicates that excellent separation of Canola o i l and C0 2 solvent can be achieved by a simple pressure reduction. The figure also indicates that the pressure reduction need not be to atmospheric since the o i l s s o l u b i l i t y in C0 2 at 10 MPa i s >0.05%. This fact i s s i g n i f i c a n t since i t indicates that, in a r e - c i r c u l a t i n g extraction system, the costs of re-pressurizing the C0 2 could be reduced. Nonetheless, on a large scale, a separation based on pressure reduction may be expensive. In this case a separation based on a temperature change would be more desirable (Peter and Brunner, 1978). The fact that the solvent power of a s u p e r c r i t i c a l f l u i d changes with temperature at constant pressure can be of considerable p r a c t i c a l importance. It Indicates that, in some cases, i t may be possible to e f f e c t a separation of solvent and solute merely by changing temperature. 100 However i t seems unlikely that Canola o i l can be economically separated from C0 2 solely by changing the temperature. As indicated above, the most e f f e c t i v e separations of t h i s kind involve the use of a s u p e r c r i t i c a l solvent near i t s c r i t i c a l point. For C0 2 this would require pressures and temperature's approximately in the region of 7-10 MPa and 30-40°C, respectively. However, under these conditions the solvation capacity of C0 2 for Canola o i l i s too low to make such an extraction worthwhile. A temperature-based separation at higher carbon dioxide pressure would probably also be impractical. For example, i f the extractor operates at 55°C and 36 MPa and the separator operates at 25°C and 36MPa the o i l s o l u b i l i t y i s only changed from 11 to 9 mg/g C0 2 (Fig 5.2). The rec i r c u l a t e d carbon dioxide would thus enter the extraction chamber with a high o i l content thereby lowering the dri v i n g force in the extraction and consequently the rate of extraction. A separation based solely on temperature might, however, be possible provided a suitable entrainer can be found for the C0 2. Brunner and Peter (1982) have demonstrated such a separation using carbon dioxide charged with 10 wt% ethanol. At 50° C and 17.5 MPa, the s o l u b i l i t y of Palm o i l in th i s mixture was reported as approximately 8wt%. At the same pressure but at 90°C, the s o l u b i l i t y was only 2wt%. This indicates that an e f f e c t i v e separation of Palm o i l from the solvent could be achieved with a temperature change of only 40 0 C. Since Palm o i l i s s i m i l a r to Canola o i l , i t would be reasonable to assume 101 that similar separations could be performed with Canola o i l . However, due to the li m i t a t i o n s of the experimental equipment, extraction experiments using entrainers could not be performed. 5.3 Equilibrium Oi1 Concentration In Carbon Dioxide The value of y* has previously been defined as the concentration of o i l in the carbon dioxide solvent phase in equilibrium with seeds having an o i l concentration x. In section 3.2 i t was shown that the relationship between y* and x (y*=f{x}) was required to solve the mass balance equations (Eqs. 3.3, 3.4). Accordingly, two sets of experiments were conducted to provide information r e l a t i n g y* to x. Extractions with C0 2 at 36 Mpa and 55° C were carr i e d out on flaked seed which had been pre-extracted with either C0 2 (36 Mpa, 55° C) or hexane to the levels shown in Table 5.1. TABLE 5.1 Concentration of o i l in the p a r t i a l l y extracted seed samples. g o i l / g o i l - f r e e seed sample C0 2 extracts hexane extracts 1 unextracted seed 0.67 0.48 0.40 0.30 0.22 0.67 0.35 0.23 2 3 4 5 In Figs. 5.6 and 5.7 the value of the o i l concentration in the carbon dioxide at t.he bed outlet i s plotted as a function of the i n i t i a l x value of the the seed for d i f f e r e n t C0 2 flow 102 I 1 I I I I I I I I I I I I I I I I I I I I I I • 0.47 g/min C02 vessel #1 • 0.93 g/min C02 vessel #1 • 1.40 g/min C02 vessel #i A 1.87 g/min C02 vessel #i i -* -• i i i i i i i i i i- i i i i i i i i i t i i i 0.2 0.24 0.28 0.32 0.36 0.4 0.44 0.48 0.52 0.56 0.6 0.64 0.68 BED OIL CONCENTRATION (g oii/g oil-free seed) FIGURE 5.6 O i l concentration in the C0 2 phase at the extractor outlet for d i f f e r e n t seed-bed o i l concentrations. The reduced o i l concentration seed was prepared by p a r t i a l l y extracting samples of flaked seed for d i f f e r e n t lengths of time with hexane. Conditions: vessel #1, pressure 36 Mpa, temperature 55°C, C0 2 flowrate as indicated on the Figure. 1 03 oo O O OS cm £ 8 •— o I s I I I I I I I I I I I I I I I I I I I I I I I • • 0.47 g/min C0 2 vessel # l 0.93 g/min C0 2 vessel # l 1.40 g/min C0 2 vessel # l 1.87 g/min C0 2 vessel # l 0.47 g/min C0 2 vessel # 3 0.2 0.24 0.28 0.32 0.36 0.4 0.44 0.48 0.52 0.56 0.6 0.64 0.68 BED OIL CONCENTRATION (g oil/g oli-free seed) FIGURE 5.7 O i l c o n c e n t r a t i o n i n the C 0 2 phase a t the e x t r a c t o r o u t l e t f o r d i f f e r e n t seed-bed o i l c o n c e n t r a t i o n s . The reduced o i l c o n c e n t r a t i o n seed was p r e p a r e d by p a r t i a l l y e x t r a c t i n g samples of f l a k e d seed f o r d i f f e r e n t l e n g t h s of time u s i n g s u p e r c r i t i c a l C 0 2 a t 36 MPa and 55 ° C. C o n d i t i o n s : p r e s s u r e 36 MPa, te m p e r a t u r e 55°C, C 0 2 f l o w r a t e as i n d i c a t e d on the F i g u r e . 1 04 rates. As can be seen from these figures, the value of y* was constant for the seed samples having o i l contents in excess of about 0.4 g/g o i l - f r e e seed. For the seed samples having lower o i l concentrations, i t i n i t i a l l y appeared as i f y* depended on x. However, i t can also be seen from these figures that, in t h i s region, y* was a function of the carbon dioxide flowrate. As the flowrate decreased, the exit concentration increased thereby indicating that the carbon dioxide had not come to equilibrium with the o i l in the bed. From these experiments, i t can be concluded that, over the range of seed o i l concentrations used, y* is independent of x. This result i s not unexpected. For y* to be dependent on the seed o i l concentration, the o i l and seed tissue would need to exhibit an at t r a c t i o n for each other at the molecular l e v e l , i . e . the o i l would need to be bound to the seed. However, th i s i s not the case. The o i l in the intact seed exists as small o i l droplets (Hofsten, 1970; Yiu et a l . , 1982) and in the crushed seed as a f i l m on the surface of the seed p a r t i c l e s . Within the l i m i t a t i o n s of these experiments i t appears that the seed tissue does not have a 'chemical a f f i n i t y ' for the o i l and acts merely as an inert substrate. The r e l a t i o n s h i p y*=f{x} can thus be represented by y* = constant, where the constant corresponds to the o i l s o l u b i l i t y at a given temperature and pressure. 1 05 5.4 Effect of Seed Treatment 5.4.1 Seed P a r t i c l e Size In Table 5.2 the range of seed p a r t i c l e sizes resulting from the various treatment procedures i s presented. The analysis was performed by passing the seed material through sieves of various mesh sizes and weighing the fraction associated with each tray. TABLE 5.2 D i s t r i b u t i o n of seed p a r t i c l e sizes, for the d i f f e r e n t methods of seed treatment. percent of t o t a l mass size range whole crushed f i n e l y flaked cooked exploded (mm) seed chopped < 0. 149 • - 0.8 8.6 0.2 0.2 0. 1 49-0.589 41.3 70.8 1 9.2 12.6 -0.590-0.850 - . 30.2 15.5 24.8 24. 3 • -0.851-1.00 1.1 12.1 1.5 11 .5 22.8 0.4 1.00 -1.40 25.8 9.7 1.1 19.0 33.3 19.5 >1 .40 73.1 5.9 2.4 25.2 6.7 80.1 5.4.2 Experimental Results I n i t i a l l y , attempts were made to extract whole, unbroken Canola seeds with carbon dioxide at 36 MPa and 55 °C. V i r t u a l l y no o i l was recovered after extraction for 5 h at a carbon dioxide flow-rate of 0.8 g/min. This result agrees with previous reports. Othmer and Agarwal (19551, extracted whole and l a t e r a l l y sectioned soybeans with hexane for 166 h. It was found that less than 0.08% of the o i l o r i g i n a l l y in the whole beans and less than 0.19% of o i l in the half beans were extracted. This indicates that hexane i s unable to penetrate and remove o i l 106 from unbroken c e l l s . Commercial Canola seed extraction is therefore preceded by cooking and flaking prior to extraction (Anjou, 1972; Beach,1983) In F i g . 5.8 t y p i c a l extraction curves are shown, for seeds having undergone various pre-treatments. A l l extractions were ca r r i e d out at 36 MPa, 55° C a t a C0 2 flowrate of 0.7 g/min. In each case, a 4g sample was extracted. For comparative purposes the extraction curve of pure Canola o i l on a bed of 0.3mm Ottawa sand i s also shown. The amount of o i l on the beads (1.6 g) was equivalent to the o i l content of 4g of oilseed. It is evident from t h i s figure that seed pre-treatment greatly a f f e c t s the t o t a l quantity o f o i l removed from the seeds. In F i g . 5.9 the extraction data are plotted in a d i f f e r e n t form. The curves were generated by f i t t i n g a second order polynomial function to the data points in F i g . 5.8 and d i f f e r e n t i a t i n g the function at selected positions. The results correspond to the o i l concentration in the carbon dioxide at the outlet of the extractor (sec 3.1). As indicated in F i g . 5.8, the least e f f e c t i v e pre-treatment procedure was the 'exploding' technique. After extracting t h i s material for 5 h, less than 10% of i t s o i l had been removed. The figure also i l l u s t r a t e s that the carbon dioxide did not reach saturation. This seed pre-treatment technique d i f f e r s from a l l of the others in this respect and indicates that the c e l l u l a r disruption was only s l i g h t . Hence, either the C0 2, which was o r i g i n a l l y used to explode the seeds, did not penetrate to a s i g n i f i c a n t depth and disrupted only the c e l l s on the surface of the seeds or, TOTAL OIL VS.' TOTAL C02 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0 200.0 220.0 240.0 TOTAL C02 tg) FIGURE 5.8 Extraction curves 4 g samples of Canola seed subjected to d i f f e r e n t pre-treatments. For comparison purposes the extraction curve for Canola, o i l on Ottawa sand i s also shown. Conditions: vessel #1, pressure 36 MPa, temperature 55°C, C0 2 flow rate 0.8'g/min. 108 o CM 00 CM CJ <£ Ol \ o O CN I—I *~ I— CE 0 LxJ CJ o ao _J to M O a CN V O CJ a CM n — i — i — i — r » — 0 flaked i — i — i — i — r •+ finely chopped crushed * — * exploded °-—° flaked € cooked - * Ot tawa sand 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 PERCENT OIL EXTRACTED FIGURE 5.9 Transformed extraction curves for several Canola seed pre-extraction treatments. The Y axis of the graph represents the concentration of o i l in the C0 2 at the extractor o u t l e t . 109 a l t e r n a t i v e l y the c e l l walls were s u f f i c i e n t l y robust to withstand the large pressure d i f f e r e n t i a l generated during decompression. The crushing procedure, although considerably more ef f e c t i v e than the pressure-rupturing treatment, evidently l e f t much of the seed intact as well. As indicated in F i g . 5.9, concentrations at the extractor outlet began to decline rapidly after approximately 10% of the o i l had been removed. By the time 60% of the seed o i l was removed, the o i l concentration in the outlet carbon dioxide had decreased to less than 5% of i t s saturation value. For the f i n e l y chopped seed, the outlet o i l concentration began to decrease after about 35% of the o i l had been extracted. The o i l concentration in the carbon dioxide f e l l to 5% of i t s saturation value only after 80% of the seed o i l had been removed. The flaked and cooked seeds had extraction c h a r a c t e r i s t i c s similar to those of the f i n e l y chopped seed in the i n i t i a l stage of the extraction. However, the amount of o i l extracted from these samples exceeded 85% of t o t a l before the o i l concentration of the extractor outlet f e l l to 5% of saturation. Although Beach,(1983) and Clandinin, (1981) have reported that cooked Canola seed releases o i l more readily than uncooked seed, there was no difference between these two samples when extracted with C0 2. As can be seen from F i g . 5.8, the Canola o i l was most e f f e c t i v e l y removed from the glass beads. Over 95% of the o i l in the bead matrix was removed before the outlet o i l concentration 1 10 in the carbon dioxide f e l l to 5% of i t s saturation value. This result i s not unexpected, since the beads are non-porous and a l l of the o i l l i e s on the bead surface. Furthermore, the bead beds are composed of many open channels through which C0 2 can flow. By contrast, the o i l in beds of crushed or flaked seed may be trapped in regions between p a r t i c l e s impervious to the C0 2 or within intact c e l l s . Much of the o i l contained within these regions i s not exposed to the moving stream of carbon dioxide and is only transferred out of these channels by the slow process of d i f f u s i o n . The difference in the extraction c h a r a c t e r i s t i c s of flaked and f i n e l y chopped seed may also be due to the d i f f e r e n t flow patterns in the beds, Unlike the flaked seed, the f i n e l y chopped seed consists predominantly of very small p a r t i c l e s . The passages between these p a r t i c l e s are consequently also small and the resistance to flow through the network i s high. In such a case the carbon dioxide would flow predominantly through a small number of large channels which probably exist in the bed as well. This channeling ef f e c t i s undesirable since only a small fraction of the bed would be exposed to the moving stream of C0 2 . From these experiments i t can be concluded that s u p e r c r i t i c a l C0 2 has the a b i l i t y to extract as much o i l from Canola seed as hexane and that the current commercial pretreatment processes are also suitable for s u p e r c r i t i c a l C0 2 extraction process. 111 5.5 Scanning Electron Microscopy of Seed P a r t i c l e s In section 3.2.3 a mechanism was proposed to account for the extraction c h a r a c t e r i s t i c s of Canola seed (Fig. 3.4). One of the assumptions of the mechanism was that the crushed seed, in i t s unextracted state, was covered with a layer of o i l which gradually became depleted during the extraction process. In order to determine whether th i s assumption was v a l i d , samples of unextracted crushed seed and crushed seed at various stages of carbon dioxide extraction were examined under the scanning electron microscope (SEM). The method by which these samples were obtained was described in Sec. 4.6.2. Fi g . 5.10 i s an SEM photograph of a t y p i c a l seed p a r t i c l e prior to extraction. The presence of an unbroken layer of o i l on the surface of the seed fragment i s suggested by the p a r t i c l e ' s amorphous, globular, l i q u i d - l i k e appearance and the lack of any fine c e l l u l a r d e t a i l . In F i g . 5.11 an SEM photograph of a p a r t i a l l y extracted seed p a r t i c l e i s shown. Regions with fine c e l l u l a r d e t a i l s are v i s i b l e thereby indicating an absence of o i l . In F i g . 5.12 a ' f u l l y extracted' seed p a r t i c l e i s shown. In t h i s figure the c e l l u l a r structure of the seed i s c l e a r l y v i s i b l e and the amorphous globular regions, which were present in the previous two photographs, are completely absent. This photograph suggests that the surface of the f u l l y extracted seed p a r t i c l e i s o i l -free . These photographs generally support the extraction mechanism suggested in Sec. 3.4 and indicate that the seed FIGURE 5.10 Scanning electron micrograph of a fragment flaked Canola seed prior to extraction. (Mag. 620X, 20 Au-Pd) FIGURE 5.11 S c a n n i n g e l e c t r o n m i c r o g r a p h of a fragment of f l a k e d Canola seed a f t e r p a r t i a l e x t r a c t i o n w i t h C 0 : a t 36 MPa and 55°C (Mag. 620X, 20 Kv, Au-Pd). The seed fragment was e x t r a c t e d f o r a p p r o x i m a t e l y 15 minutes a t a C0 2 f l o w r a t e of 0.5g/min. FIGURE 5.12 S c a n n i n g e l e c t r o n m i c r o g r a p h of a fragment of f l a k e d Canola seed a f t e r b e i n g ' f u l l y ' e x t r a c t e d w i t h C0 2 a t 36 MPa and 55 °C. (Mag. 780X, 20Kv, Au-Pd). The seed fragment was e x t r a c t e d f o r a p p r o x i m a t e l y 2 h a t a C0 2 f l o w r a t e of 0.5g/min. 1 1 5 p a r t i c l e s are i n i t i a l l y covered with a layer of o i l which gradually becomes depleted as the extraction proceeds. 5.6 Fatty Ac id Composition of Extracts 5.6.1 Introduction One of the objectives of t h i s study was to determine the composition of the s u p e r c r i t i c a l carbon dioxide extracts as a function of extraction time. Additionally, these extracts were compared with those obtained by conventional hexane extraction 5.6.2 Fatty Ac id Ester Response Factors C o n f l i c t i n g information exists as to whether or not flame ionization detectors (FID) respond with equal magnitude to equal masses of fatty acid esters (Ackman and Sipos, 1964). Since t h i s i s an important consideration, the f i r s t step in t h i s investigation consisted of determining the FID response factors for the relevant fatty acid (FA) esters. TABLE 5.3 FID response factors for fatty acid methyl esters r e l a t i v e to methyl palmitate. fatty < ac id carbon response ester number factor methyl myr i state C1 4:0 1 .000 methyl palmitate C16:0 1 .000 methyl palmitoleate CI 6:1 1 .000 methyl stearate C18:0 1 .000 methyl oleate C18:1 1 .000 methyl 1inoleate C18:2 1 .074 methyl 1inolenate C18:3 1 .000 methyl arachidate C20:0 0.999 methyl eicosenoate C20: 1 1 .000 methyl behenate C22:0 0.991 methyl erucate C22:1 0.999 methyl lignocerate C24:0 1 .000 1 1 6 The a n a l y t i c a l standards used for the above were obtained from Supelco Corp (Appendix I I I ) . In Table 5.3 the response factors determined for several FA esters i s presented. A l l subsequent c a l c u l a t i o n s , which apply to the fatty acid esters, were obtained using the above response factors. 5.6.3 Fatty Ac id Composition of Canola O i l The various t r i g l y c e r i d e s which make up Canola o i l contain both saturated and unsaturated FA moieties ranging in carbon number from 14 to 24. The fatty acid p r o f i l e can thus be used as a means of characterizing the o i l and is often the measure by which o i l from d i f f e r e n t v a r i e t i e s of Canola seed are compared (Table 5.4; Ackman, 1983). In Table 5.5 the fatty acid compositions of Canola o i l from four sources are shown. The C0 2 extract l i s t e d in the table was obtained by extracting flaked Canola seed at 36 MPa and 55°C. The hexane extract was produced from flaked seed using hexane at 55°C. The refined and crude o i l s were obtained from CSP Foods and are described in Sec. 4.3.4. TABLE 5.4 Fatty acid composition of t r i g l y c e r i d e s in Canola o i l obtained from three v a r i e t i e s of Canola seed (Ackman, 1983). Variety 16:0 16:1 % % 18:0 % 18:1 % 18:2 % 18:3 % 20:0 % 20:1 % 22:0 22: 1 % % Andor 3.9 0.2 1 .3 58.2 21.6 12.1 0.5 1 .6 0.4 0.0 Tobin 3.8 0.1 1 .2 58.6 24.0 10.3 0.6 1 .0 0.1 0.3 Jet Neuf 4.9 0.4 1 .4 56.4 24.2 10.5 0.7 1 .2 0.3 0.0 1 17 FIGURE 5.13 Chromatogram of the fatty acid methyl esters in an e s t e r i f i e d sample of a t y p i c a l C0 2 extract of Canola seed (36 MPa, 55°C, flow 0.7 g/min). Analysis conditions: Column- SP-2330 on 100/120 mesh chromosorb WAW; detector(FID) and injector temp. 250°C; isothermal 200°C. 118 A t y p i c a l chromatogram showing the elution sequence of each FA ester and the peak resolution i s presented in F i g . 5.13. TABLE 5.5 Fatty acid composition of the t r i g l y c e r i d e s in four samples of Canola o i l . For a description of each sample refer to text. The fatty acids C20:1 and C18:3 were not resolved using the chromatographic procedure. Canola o i l 14:0 % 16:0 % 18:0 % 18:1 % 18:2 % 20: 1 18:3 % 20:0 % 22:0 % 22: 1 % 24:0 % ref ined o i l 0.2 3.9 1 .8 58.0 22.0 11.0 0.8 0.8 1 .3 0.3 CSP crude 0.2 4.4 1 .8 58.2 21.5 10.9 0.7 0.3 1 .0 0.2 C02 extract 0.1 4.7 2.0 56.9 21.9 11.2 0.8 0.5 0.6 0.3 hexane extract 0.2 4.8 2.1 57.4 22.0 11.0 0.9 0.4 1 .0 0.2 absolute error ± 0.1 0.2 0.2 1 .4 0.3 0.2 0.2 0.2 0.2 0.2 The identit y of each major peak in the chromatogram was obtained i n i t i a l l y by comparing retention times with a n a l y t i c a l standards and was later confirmed using GC/mass spectrometer techniques. The absolute error in each case i s the sum of the t o t a l measurement error involved with the integration of the chromatographic peaks and the error associated with the a n a l y t i c a l standards. Details of the error analysis appear in Appendix I I . It can be seen from Table 5.5 that the hexane and carbon dioxide extracts of the same seed are quantitatively i d e n t i c a l in a l l of the fatty acids with the exception of erucic acid (C22:1). In the hexane extract t h i s fatty acid appeared in 1 19 a s l i g h t l y higher concentration. 5.6.4 Fatty Ac id Composition of Carbon Dioxide Extracts In the f i r s t experiment 4.2 g of cooked CSP seed were extracted at 36.0 MPa, 55 °C, at a C0 2 flowrate of 0.7g/min. Seven consecutive extract fractions were co l l e c t e d (Fig. 5.14). TABLE 5.6 Fatty acid composition of sequential carbon dioxide extracts of CSP cooked Canola seed. Conditions: vessel #1, pressure 36 MPa, temperature 55°C, flowrate 0.7 g/min. fatty extract number acid ester 1 2 3 4 5 6 7 error % C14:0 % 0.1 0.1 0.1 0.1 0.1 0.1 0.2 + 0.1 C1 6:0 % 4.7 5.1 5.2 5.1 ••4.6 4. 1 3.4 + 0.2 C18:0 % 2.0 1 .9 1 .8 1 .9 2.0 2.2 2.5 + 0.2 C18: 1 % 56.9 57.3 57.3 57'.. 9 58.8 59.6 58.5 + 1 .4 C18:2 % 21.9 22.6 22.7 22.4 21.8 20.8 19.6 + 0.3 C20:0 % 0.8 0.6 0.6 0.6 0.7 0.9 1 .3 + 0.2 C20:1 + C1 8 : 3 % 11.2 11.2 11.3 11.1 10.9 10.6 10.7 + 0.2 C22:0 % 0.5 0.3 0.3 0.3 0.3 0.4 1 .0 + 0.2 C22: 1 % 0.6 0.6 0.5 0.5 0.6 0.9 2.5 + 0.2 C24:0 % 0.3 0.2 0.2 0.1 0.1 0.2 0.6 + 0.2 In Table 5.6 the FA ester composition of each extract i s provided while in F i g . 5.15 t h i s information appears in graphical form. As can be seen, only small variations occur in the o i l composition during the extraction. In the f i n a l extract sample, which was obtained after about 80% of the o i l had been removed from the seedbed, the proportion of the heavier FA esters (C20-C24) was higher than in the previous f r a c t i o n s . In an e f f o r t to v e r i f y t h i s and to c o l l e c t extracts after longer times, the experiment was repeated under 1 20 to _ Q o -cr or UJ _ l 00 cr 1— o i <p 1—I I I I I 1 I I I I I I I I I I I I I I I J I I I I I L 0.0 16.0 32.0 48.0 64.0 80.0 96.0 112.0 128.0 144.0 160.0 176.0 192.0 TOTAL C02 PASSED THROUGH BED (g) FIGURE 5.14 Extraction curve for a 4.2 g sample of commercially cooked Canola seed. The extraction was ca r r i e d out at 36 MPa and 55 °C at a C0 2 flowrate of 0.7 g/min. The numbered interv a l s on the curve indicate the regions over which o i l samples were c o l l e c t e d for fatty acid analysis. i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r o—«C16:0 < • C18:0 « »C18:1 »—« C18:2 a—«C18:3 •—* C22:0 *—* C22:1 »—*C24:1 8 -0 0 0 0 o CN t— (_) cr o or-t— < x ilia CO* Li. w O — «=>. w 03 in ID UJ 0_ L£EX 3> i r t — r 4 i i i i l 16.0 32.0 48.0 64.0 80.0 96.0 112.0 128.0 144.0 160.0 176.0 192.0 208.0 TOTAL C02 PASSED THROUGH BED (g) FIGURE 5.15 Fatty acid composition of the extracts indicated F i g . 5.14. 1 22 id e n t i c a l conditions. At the completion of th i s second experiment the r e s t r i c t e r valve was also washed with CHC13 and the o i l analyzed. In F i g . 5.16 the extraction curve is shown along with the positions where the extracts were co l l e c t e d for analysis. In this case the la s t extract was obtained after 90% of the o i l had been removed from the sample. In Table 5.7 the composition of the various extracts i s l i s t e d while in F i g . 5.17 thi s information appears in graphical form. TABLE 5.7 Fatty acid composition of sequential carbon dioxide extracts of CSP flaked and cooked Canola seed. The analysis of o i l extracts obtained from the r e s t r i c t e r valve i s also provided. Conditions: vessel #1, pressure 36 MPa, temperature 55°C, flowrate 0.7g/min. fatty extract number o i l acid from error ester 1 4 7 valve % C14:0 % 0.1 0.1 <0. 1 <0. 1 + 0.1 C1 6:0 % 5.4 5.1 3.3 4.5 + 0.2 C18:0 % 1 .9 1 .9 2.7 2.8 + 0.2 C18:1 % 56.5 56.9 55.2 54.5 + 1 .4 C1 8: 2 % 23.0 21.5 19.9 19.8 + 0.3 C20:0 % 0.5 0.5 1 . 1 1 .3 + 0.2 C20:1 + C18:3 % 11.3 11.4 11.3 10.9 + 0.2 C22:0 % 0.3 0.3 1 .3 1 .4 + 0.2 C22: 1 % 0.6 0.6 4.3 4.3 + 0.2 C24:0 % 0.1 0.1 0.9 0.9 + 0.2 Extracts 1 and 2, which were obtained early and midway through the extraction, were similar in composition to the previous extracts. The proportion of the heavier FA esters was s i g n i f i c a n t l y higher in the la s t extract and in the o i l co l l e c t e d from the r e s t r i c t e r valve. 123 I I I I I I I I I I I I I I I I I I I— X a *~2 co O o 2: « 5 d i i t i » i i i i I I I I L I I I I I I I L 2tU) 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0 200.0 220.0 240.0 260.0 TOTAL C02 PASSED THROUGH BED (g) FIGURE 5.16 Extraction curve for a 4g sample of commercially cooked Canola seed. Conditions: 36 MPa, 55°C, C0 2 flow rate 0.7 g/min. The dotted areas on the curve represent the regions over which samples were c o l l e c t e d for f a t t y acid analysis. 124 I I I I—I I I I—I—I I I I—I—I I I I T 8 oo 0 0 CJ CX o CC V I— » X L U R <o LL. in O en as in •* ro or °-UJp> T—r 1—i—r C16:0 C18:0 C18:1 C18:2 C18:3 C22:0 C22:1 ' C24:0 *20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0 200.0 220.0 240.0 260.0 TOTAL C02 PASSED THROUGH BED (g) FIGURE 5.17 Fatty acid composition of the Canola seed-C0 2 extracts indicated in FIG. 5.16. 125 It i s of interest to note that the concentration of the C24:0 FA in the f i n a l extract f r a c t i o n , exceeds 0.2%, i . e . the maximum sp e c i f i e d for edible Low Erucic Acid Rapeseed o i l (LEAR) by the Codex Alimentarius Commission (Ackman, 1983). The concentration of fatty acids in the f i n a l f r a c t i o n does however, conform to the s p e c i f i c a t i o n s of the Canada A g r i c u l t u r a l Products Standards Act (Boulter, 1983). A possible reason why the composition of the o i l extracts remains constant during most of the extraction, may be obtained by considering the d i s t r i b u t i o n of t r i g l y c e r i d e s which appear in the Canola o i l . In Table 5.8 the t r i g l y c e r i d e carbon number i s l i s t e d for two samples of Canola o i l . The table indicates that a large proportion of the o i l i s composed of t r i g l y c e r i d e s having 55 or 57 carbons. Consequently the MW range of the t r i g l y c e r i d e s in the o i l w i l l be small. TABLE 5.8 P r o f i l e by carbon number of the various t r i g l y c e r i d e s present in Canola o i l from two d i f f e r e n t sources (Ackman, 1983). seed sample t r i g l y c e r i d e carbon number C51 C53 C55 C57 C59 C61 C63 C65 C67 1 2 % % 2 3 6 18 61 7 4 18 70 5 2 1 1 -1 <1 - -Since the various t r i g l y c e r i d e s are also very similar in chemical nature, i t might be expected that they would exhibit similar s o l u b i l i t i e s in the carbon dioxide. Consequently, a large portion of the t r i g l y c e r i d e s w i l l l i k e l y be extracted at 1 26 the same rate and the composition of the extract remains constant for much of the extraction. However, towards the end of the extraction, the small amounts of higher molecular weight t r i g l y c e r i d e s would be expected to constitute a larger f r a c t i o n of the C0 2 extracts since their mole-fractions in the residual o i l are higher. 5.6.5 Fatty Ac id Composition of Hexane Extracts In an e f f o r t to determine whether the hexane extracts obtained with cooked Canola seed were similar to the C0 2 extracts, the above experiments were repeated using hexane at 55°C and 1.5MPa in place of C0 2. At the completion of the extraction, each c o l l e c t i o n v i a l was heated to 55 °C for two hours to f a c i l i t a t e the removal of the hexane from the extracts. Figure 5.18 i s a plot of the t o t a l o i l c o l l e c t e d during the experiment vs the t o t a l amount of hexane used. Indicated on the figure are the regions over which each of the extracts was c o l l e c t e d . In Table 5.9 the fatty acid composition of each extract i s presented while in F i g . 5.19 t h i s information appears in graphical form. It i s evident that the FA composition of the hexane extracts is similar to that of the carbon dioxide extracts(Tables 5.6 and 5.7). It i s worth noting however that the late hexane extracts, unlike the corresponding C0 2 extracts, do not show an increased proportion of the high molecular weight (MW) fatty acids. One explanation for t h i s could be that both high and low molecular weight t r i g l y c e r i d e s are equally soluble in the hexane and therefore neither i s p r e f e r e n t i a l l y removed. 127 cm 0**. UJ *" I— o cr <>i cc -t— X U J e O cq o _J O e I I I I I I I I I I I I I I I I I I I I I I I I **.::<....* • 5 6 7 •-12 -J I I I I I L_l L_J_ J I I I I I I L J I L 0.5 1.5 2.5 3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.5 11.5 12.5 TOTAL HEXANE PASSED THROUGH BED (g) FIGURE 5.18 Extraction curve for a 4g sample of cooked Canola seed. The extraction was carried out using hexane at 1.5 MPa and 55 ° C. The dotted areas on the curve represent the regions over which samples were coll e c t e d for fatty acid analysis. 128 i — i — i — r - 8C16 : 0 -» C18.-0 -»C18:1 C18:2 -oC18:3 •* C22:0 C22:1 -*C24:0 i i i i r 1 — i — r ~ — i — r i — r CO CO CO CO a o I— to X CD U _ i n O in co in ro <->o UJ co CL -—e—. - • B 1 Q • o b— — e — — © — — © — —©—• 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.D 9.0 10.0 11.0 12.0 13.0 TOTAL HEXANE PASSED THROUGH BED (g) FIGURE 5.19 Fatty acid composition of the Canola seed hexane extracts indicated in F i g . 5.18. 1 29 TABLE 5.9 Fatty acid composition of the sequential hexane extract of CSP cooked Canola seed. Conditions: vessel #1, pressure 1.5 MP temperature 55°C, flowrate 0.7g/min. fatty hexane extract number acid ester 1 2 3 4 5 6 * 7 / 9 1 2 error % C1 4:0 % 0.2 0 .2 0. 2 0 .2 0. 5 0. 6 1 .0 0. 5 2.0 + 0.2 C1 6:0 % 4.8 4 .7 4. 8 5 .0 4. 9 5. 3 5 .4 5. 1 5.3 + 0.2 C18:0 % 2.1 2 .0 2. 1 2 . 1 2. 1 2. 1 2 .2 2. 0 2. 1 + 0.2 CI 8:1 % 57.4 57 .6 57. 7 57 . 1 57. 2 57. 0 57 .0 57. 1 56.4 + 1 .4 C1 8:2 % 22.0 22 . 1 22. 1 22 .2 21 . 9 22. 0 21 .7 22. 0 21.4 + 0.2 C20:0 % 0.9 0 .8 0. 9 0 .9 1 . 0 0. 9 0 .9 0. 9 0.9 + 0.2 C20:1 + C18:3 % 11.0 1 0 .9 1 1 . 6 1 1 .0 10. 7 10. 7 10 .5 10. 8 10.2 + 0.2 C22:0 % 0.4 0 .4 0. 4 0 .4 0. 4 0. 4 0 .3 0. 4 0.4 + 0.2 C22: 1 % 1 .0 1 .0 1 . 0 1 .0 1 . 0 0. 9 0 .8 0. 9 0.9 + 0.2 C24:0 % 0.2 0 .2 0. 2 0 .2 0. 2 0. 2 0 . 1 0. 2 0.2 + 0.2 5.7 C02 Extracts of Simple T r i g l y c e r i d e s 5.7.1 Introduction The previous section demonstrated that very l i t t l e f r actionation of the Canola o i l took place during the extraction process and suggested that the reason for th i s was the s i m i l a r i t y of the t r i g l y c e r i d e s which make up the o i l . In order to provide further insight into the extraction process, a simple mixture of pure t r i g l y c e r i d e s was extracted. Since each t r i g l y c e r i d e was composed of a unique fatty . acid, the chromatographic procedures described in section 4.6.4 could be used to 'follow' the t r i g l y c e r i d e - composition of each extract f r a c t i o n . Using t h i s procedure, i t was possible to study how the pure t r i g l y c e r i d e s interact with each other during the 1.30 extraction process. From t h i s information i t was possible to suggest a method for c a l c u l a t i n g the s o l u b i l i t y of other vegetable o i l s in C0 2, based on their fatty acid composition. Two sets of experiments were conducted using the pure t r i g l y c e r i d e s , t r i p a l m i t o l e i n (C16:1), t r i o l e i n (C18:1), and t r i e i c o s e n o i n (C20:1). In the f i r s t set the s o l u b i l i t y of each t r i g l y c e r i d e in C0 2 at 55 °C and at 36.0 MPa was determined. In the second set a mixture of these three compounds was prepared and extracted using carbon dioxide at the above conditions. The extracts from t h i s second set of experiments were co l l e c t e d sequentially and subsequently analyzed for t r i g l y c e r i d e composition. During the experiments vessel #1 was used. The C0 2 flowrate in a l l cases was 0.7 g/min. 5.7.2 S o l u b i l i t y of Single T r i g l y c e r i d e s The s o l u b i l i t y of each t r i g l y c e r i d e in C0 2 was determined by extracting i t from a matrix of Ottawa sand. Figure 5.20 shows the extraction curves for each of the three t r i g l y c e r i d e s . The li n e a r portion of the curves indicate that saturation was achieved during the extraction. The r e s u l t s from these experiments show that the s o l u b i l i t y of the t r i g l y c e r i d e s decreased with increasing molecular weight (Fig. 5.21). When the s o l u b i l i t y i s plotted against molecular weight on a semi-logarithmic scale (Fig. 5.22), a straight 1 irve .. relationship i s observed. Since the three t r i g l y c e r i d e s aire part of a homologous series t h i s behaviour i s not unexpected. A similar s o l u b i l i t y TOTAL C02 PASSED THROUGH BED (g) FIGURE 5.20 Extraction curves for pure-tripalmitolein (C16:1), t r i o l e i n (CV8:1) and tri-11-eicosenoin (C20:1). The extractions were performed from Ottawa sand using C0 2 at 36 MPa and 55 °C. Conditions: vessel .# 1 , C0 2 flow rate 0. 7 g/min. 1 32 CM i—i—i—i—i—i—r i — i — i — i — i — i — i — i — i — i — r CM CM O CM o OI QQ CM ZD ~~ _ l O CO o CD I— 00 CO • i • • • i • I I I I I I I I I I I L 800.0 820.0 840.0 860.0 880.0 900.0 920.0 940.0 960.0 980.0 1000.0 MOLECULAR WEIGHT (amu) FIGURE 5.21 S o l u b i l i t y of the three t r i g l y c e r i d e s indicated in F i g . 5.20 as a function of their molecular weight. The error bar shown i s representative of a l l points. 133 FIGURE 5.22 The negative logarithm of the s o l u b i l i t y of the three t r i g l y c e r i d e s indicated in F i g . 5.20 as a function of t h e i r molecular weight. A t y p i c a l error bar i s indicated. 1 34 effe c t has been observed i n d i r e c t l y when t r i g l y c e r i d e s are dissolved in l i q u i d s . If a homologous series of t r i g l y c e r i d e s i s eluted through an HPLC column i t is found that a linear r e l a t i o n s h i p exists between the number of carbon atoms in the molecule and the logarithm of the volume of solvent required for the passage of the t r i g l y c e r i d e through the column(Hersloff et a l . , 1979; Plattner et a l . , 1977). This elution volume in turn i s proportional to the s o l u b i l i t y of the t r i g l y c e r i d e in the eluting solvent 5.7.3 S o l u b i l i t y of a T r i g l y c e r i d e Mixture Figure 5.23 i s the extraction curve for a mixture of t r i g l y c e r i d e s in the proportions shown in Table 5.10. TABLE 5.10 Mass fr a c t i o n and mole f r a c t i o n of components in the triglyceride(TG) mixture. TG molecular mass mole-weight fraction fract ion C1 6: 1 801 0.330 0.366 C18: 1 885 0.331 0.331 C20:1 970 0.330 0.302 From the graph i t i s evident that the concentration of extract in the carbon dioxide, measured at the extractor outlet, continuously declined during the course of the extraction. The sequentially c o l l e c t e d extracts from t h i s experiment were subsequently analyzed using the transesterification/GC procedure outlined in Sec 4.6.4. By using t h i s analysis, i t was possible to determine the mass fraction of each t r i g l y c e r i d e present. .1 3.5 CN cn • o " " UJ S 3 cr cr UJ O cr O CO 1 1 T 1 1 1 1 1 r v ' i i i i i I I I i i i i — •-_ - --- -- -- -- • -- -- -- • -- -- -- -• -1 1 1 i i 1 1 1 1 I I » « • ! i i i i i i M 16.0 32.0 48Ji 64 JO 80.0 96J3 112.0 128.0 144.0 160,0 176.0 192.0 TOTAL €02 PASSED THROUGH BED (g) FIGURE 5.23 Extraction curves for an equal weight mixture of tripalmitolein., t r i o l e i n and t r i - 1 1-eicosenoin. The extraction of the mixture was performed from Ottawa sand using C0 2 at 36 MPa, 55 °C at a flow rate of 0.7 g/min. 136 from the information i t is evident that the composition of these extracts changed continuously (Table 5.11, Fig . 5.24). The i n i t i a l extracts contained a high percentage of the li g h t e r C16:1 t r i g l y c e r i d e while the f i n a l extracts were composed predominantly of the C20:1 t r i g l y c e r i d e . Hence i t must be assumed that the composition of the t r i g l y c e r i d e phase within the extractor must also have changed continuously. Since the s o l u b i l i t i e s of the t r i g l y c e r i d e s are considerably d i f f e r e n t from one another, i t i s thus possible that saturation can be achieved within the extractor while the ove r a l l rate of extraction continuously changes. For a mixture composed of compounds with widely d i f f e r i n g s o l u b i l i t i e s (eg. the t r i g l y c e r i d e mixture) a constant rate of extraction would only be expected i f the molar r a t i o of each component in the mixture remained constant throughout the course of the extraction. From these experiments i t i s evident that some fractionation of the simple mixture of t r i g l y c e r i d e s occurred during extraction with s u p e r c r i t i c a l carbon dioxide. It i s also evident that the degree of fractionation depends on the s o l u b i l i t y differences that exist among the t r i g l y c e r i d e s in the mixture. It would be expected therefore, that the degree to which Canola o i l could be fractionated would also depend on the s o l u b i l i t i e s of the individual o i l t r i g l y c e r i d e s in the carbon dioxide. Since Canola o i l contains numerous d i f f e r e n t t r i g l y c e r i d e s (Persmark, 1972), many of which would probably exhibit similar s o l u b i l i t i e s in C0 2, i t seems unlikely that simultaneous 137 I I I I I I I I I I I I d CO V— CJ CO CCEJ cr I— CO i—i ° • — (_> ex c«) cr <= CO <N COO* GC I I I I I I I I I » I - C161 v C18=1 • C201 t i \ I » » t i t i | | I I—I—I—L 0.0 8.0 16.0 24.0 32.0 40.0 48.0 56.0 64.0 72.0 80.0 88.0 96.0 PERCENT EXTRACTED (X) FIGURE 5.24 Mass fr a c t i o n of the C0 2 extracts of the t r i g l y c e r i d e mixture at d i f f e r e n t stages of the extraction. The composition of the extracts was determined using the transesterification-GC procedure. Each set of points represents an amount of o i l corresponding to the points on F i g . 5.23. Conditions: vessel #1, pressure 36 MPa, temperature 55°C, C0 2 flow rate 0.7 g/min. 1 38 extraction and fractionation of the o i l could be achieved using only s u p e r c r i t i c a l carbon dioxide. If an entrainer i s used with the C0 2, as suggested by Brunner and Peter (1982), a more e f f i c i e n t extraction and a better separation might r e s u l t . The determination of the solvent mixture conditions under which such a separation might be obtained would, however, be a formidable task. TABLE 5.11 Mass fraction of C16:1, C18:1, and C20:1 t r i g l y c e r i d e s at each data point as determined with the t r a n s e s t e r i f i c a t i o n procedure. Mass fractions are reported with an error of ± 0.005. Conditions: vessel #1, pressure 36 MPa, temperature 55°C, flow rate 0.7 g/min. t o t a l cumulat ive t o t a l C1 6: 1 C1 8: 1 C20: 1 CO 2 (g) mass of extract extract extract extract extract(g) % % % % 8.3 0.020 7.2 50.6 31.0 17.7 16.2 0.019 14,6 48.7 32. 1 18.5 25.4 0.288 21 .,7 45.8 33.6 19.9 34.3 0.378 28.5 42. 1 35.3 22.0 43.3 0.462 34.8 39. 1 36.3 24.2 54.6 0.560 42.2 35.7 36.4 27. 1 65.2 0.652 49. 1 33.6 35.9 29.8 74.3 0.733 55.2 - 33.3 35.4 30.9 83.4 0.811 61.1 32.5 34.8 32.0 92.4 0.886 66.8 30.5 35.5 33.6 101.3 0.949 71 .5 25.3 37.3 36.6 112.4 1 .022 77.0 18.2 37.5 43.8 121.1 1 .072 80. 1 12.0 32.7 54.8 1 30.0 1.121 84.5 8.3 27.4 63.8 139.0 1 . 1 64 87.8 5.8 23.4 70.2 149.2 1 . 176 88.6 4.9 22. 1 72.7 ... - _ '139 5.7.4 T r i g l y c e r i d e S o l u b i l i t y Interactions in Carbon Dioxide When a s u p e r c r i t i c a l f l u i d e xists in equilibrium with a l i q u i d mixture, i t may be possible to write an equation which describes the t o t a l s o l u b i l i t y of the mixture in terms of the s o l u b i l i t i e s of the pure components of the mixture. eg. St = xaSa + xbSb . . . . . [5.1] where St = s o l u b i l i t y of the mixture in the s u p e r c r i t i c a l C0 2 phase Sa,b = s o l u b i l i t y of the pure substance in the C0 2 phase In an e f f o r t to determine whether the s o l u b i l i t y of the t r i g l y c e r i d e mixture could be described by t h i s equation, the calculated and measured concentration values of each component in the C0 2 extract were compared (Table 5.12). The calculated composition of the extract was determined by multiplying the s o l u b i l i t y of each t r i g l y c e r i d e by i t s mole-fraction in the mixture. As the r e s u l t s indicate, both, the c a l c u l a t e d and observed values are within a few percent of each other. The system therefore behaves in an ideal manner. This r e s u l t i s not unexpected for a homologous series since molecular interactions between homologs are comparable to the interactions between molecules of a pure substance. 1 40 TABLE 5.12 Comparison of calculated and experimental composition of the C0 2 extract of a t r i g l y c e r i d e mixture. Conditions; vessel #1, pressure 36MPa, temperature 55°C, flow rate 0.7 g/min. TG TG oil-phase calculated measured s o l u b i l i t y TG mole- composition composit ion g/gCC2 f r a c t i o n g/gC02 g/gC02 C 1 6: 1 0.021 0.366 0.0077 0.0072 C18: 1 0.010 0.331 0.0033 0.0034 C20:1 0.005 0.302 0.0015 0.0016 5.7.5 Prediction of O i l S o l u b i l i t y in Carbon Dioxide Theoretically i t i s possible to calculate the extract composition and t o t a l s o l u b i l i t y of a p a r t i c u l a r o i l in C0 2 provided two types of information are known: i) the identity and proportion of the fatty acids which comprise the t r i g l y c e r i d e s of the o i l , and i i ) the s o l u b i l i t i e s in the carbon dioxide of the various t r i g l y c e r i d e s which aris e from the component fatty acids. A method to predict s o l u b i l i t y would be valuable since gathering the same information experimentally i s costly and time consuming. A c a l c u l a t i o n based on fatty acid composition would be p a r t i c u l a r l y valuable since t h i s information already exists for many seed o i l s . The c a l c u l a t i o n requires, in addition to the above information, two assumptions: 1) the d i s t r i b u t i o n of the fatty acids among the various t r i g l y c e r i d e s i s random 2) the solution of the t r i g l y c e r i d e s in the carbon dioxide i s ideal There i s some evidence that the f i r s t assumption may be v a l i d . V 141 There i s some evidence that the f i r s t assumption may be v a l i d . In a recent study by Merri.tt et a l . (1982), the t r i g l y c e r i d e composition of several o i l s could be predicted accurately from the known proportion of fatty acids present in the o i l . One of the assumptions of this study was that the acids were d i s t r i b u t e d randomly. It i s not known whether the second assumption i s v a l i d . However, some support for i t was provided in the previous section, where i t was shown that the concentration of t r i g l y c e r i d e s in s u p e r c r i t i c a l carbon dioxide could be calculated based on the their mole-fractions in the mixture. An example of how the o i l - s o l u b i l i t y c a l c u l a t i o n would be performed i s .as follows: If a given o i l i s comprised of n fatty acids in proportions p 1,p 2...pn, then from the multinomial d i s t r i b u t i o n i t can be calculated that the pr o b a b i l i t y of the occurrence of a p a r t i c u l a r t r i g l y c e r i d e i s : prob (FFF) = . [5.2] where x represents the number of occurances of a fat t y acid in a p a r t i c u l a r t r i g l y c e r i d e . For example, consider a hypothetical o i l composed of palm i t i t c (C16), o l e i c (C18) and eicosenoic (C20) acids in the proportion 0.1, 0.8 and 0.1 respectively. 1 42 The p r o b a b i l i t y of finding a t r i g l y c e r i d e with the composition C16-C16-C18 would be: The complete d i s t r i b u t i o n of t r i g l y c e r i d e s in the hypothetical o i l calculated using Eq. 5.4 i s l i s t e d in Table 5.13. If the second assumption made e a r l i e r i s also v a l i d , then the t o t a l s o l u b i l i t y of the o i l in the C0 2 becomes the sum of the individual t r i g l y c e r i d e s s o l u b i l i t i e s multiplied by th e i r respective mole-fractions (x) in the mixture, i . e . , If we make an additional assumption that the overiding factor d i c t a t i n g t r i g l y c e r i d e s o l u b i l i t y i s molecular weight then t r i g l y c e r i d e s 2 and 6, 3 and 7, and 4 and 8, in Table 5.13, w i l l have similar s o l u b i l i t i e s . From the s o l u b i l i t y data presented in Fi g . 5.22 the s o l u b i l i t y of the hypothetical o i l can be calculated, using Eq. 5.5, to be 0.011 g o i l / g C0 2 at 36.0 MPa and 55 °C. In order to more f u l l y evaluate the above method for predicting o i l s o l u b i l i t i e s , more research needs to be done. This research would involve making synthetic mixtures of t r i g l y c e r i d e o i l s and comparing the s o l u b i l i t y of the mixtures in the C0 2, as determined experimentally, with the s o l u b i l i t i e s calculated using Eqs. 5.4 and 5.5. prob (P P 0) 0.024 sol(mix) - I X <si> [5.3 ] 143 TABLE 5.13 calculated mole-fraction concentration of t r i g l y c e r i d e in an o i l composed of three fatty acids in the molar r a t i o : C16(P) 0.1, C18(0) 0.8, and C20(E) 0.1 fatty acid molecular mole-number makeup weight fra c t i o n 1 PPP 801 0.001 2 POO,OPO 857 0.192 3 POE,OEP,OPE 885 0.048 4 PEE,EPE 913 0.003 5 PPO,POP 829 0.024 6 PPE,PEP 857 0.003 7 OOO 885 0.512 8 OEO,OOE 913 0. 192 9 OEE,EOE 941 0.024 10 EEE 969 0.001 5.8 Phosphorus Content Of O i l s 5.8.1 Introduction When Canola seed i s crushed, phospholipid components, which appear naturally in the seed's c e l l membranes, are released. These phospholipids are subsequently dissolved by the seeds storage l i p i d s (seed o i l ) . As a re s u l t , the crude o i l obtained by expelling the Canola seed contains 1.5 to 3wt% phospholipid (Teasdale and Mag, 1983). In the conventional hexane extraction process these phospholipid gums are removed along with the seed o i l . Since these gums are undesirable in the finished o i l product, they are removed during the o i l r e f i n i n g process and added back to the Canola meal (Anjou, 1972). There was some indication that C0 2 would not extract phospholipid gums from the crushed seed (F r i e d r i c h and L i s t , 1982). Accordingly, experiments were conducted to investigate t h i s , by using the procedure described in Sec. 4.7. .144 In the f i r s t set of experiments, flaked and cooked Canola seed were extracted with carbon dioxide. Samples of the o i l extract were co l l e c t e d at d i f f e r e n t periods and subsequently analyzed for their phosphorus content. As well, at the termination of the extraction, the system was dismantled and the st a i n l e s s steel f i l t e r s and r e s t r i c t e r valve were analyzed to determine i f any phospholipid had been deposited on them. In the second set of experiments, a quantity of pure crude Canola o i l with a known phospholipid concentration was placed on Ottawa sand and extracted under the same pressure and temperature conditions as above. At the conclusion of the experiment the t o t a l amount of phospholipid remaining in the extraction vessel was determined. By subtracting t h i s value from the i n i t i a l value, the t o t a l amount of phospholipid extracted could be determined. 5.8.2 Phosophorus in Commerci.all.y Produced Canola O i l Prior to determining the phosphorus content of the C0 2 extracts of Canola seed, the phosphorus content of commercially produced crude and refined Canola o i l was studied. Since the phosphorus content of the commercial crude o i l was known to be high, i t was dil u t e d using CHC13 prior to analy s i s . The d i l u t i o n was necessary in order to keep the samples within the range of the a n a l y t i c a l procedure. The d i l u t i o n factor was such that 10 ulL of solution contained 0.498 mg of o i l . The re s u l t s (Table 5.14) indicate that the concentration of phospholipid in the o i l was about 1.87 ±0.09% The commercially refined o i l , considered to be a finished 1 45 product, was known to contain l i t t l e or no phosphorus and consequently was used without d i l u t i o n . Ae indicated in Table 5.15 no phosphorus was detected in the samples. TABLE 5.14 Phospholipid content of commercial unrefined Canola o i l . Absorbance readings (Abs) are in absolute values. sample sample Abs Phospho-number size 1 i p i d mg % 1 0. 498 0.345 1 .86 2 0. 498 0.348 1 .88 3 0. 498 0.346 1 .87 4 0. 249 0. 170 1 .83 5 0. 249 0.171 1 .84 6 0. 249 0. 1 78 1 .92 TABLE 5.15 Phospholipid content of refined and bleached commercial Canola o i l . Absorbance (Abs) readings are in absolute units. sample sample Abs Phospho-number size 1 i p i d mg % 1 7.52 0.002 <0.01 2 7.35 0.000 <0.01 3 7.39 0.010 <0.01 4 7.49 0.002 <0.01 5 7.48 0.008 <0.01 6 7.45 0.002 <0.01 5.8.3 Phosphorus In Carbon Dioxide Extracts Of Canola Seed In t h i s experiment, the phosphorus content of the C0 2 extracts of commercially flaked and cooked Canola seed was investigated. The sample size was 4g and the C0 2 extraction was carri e d out at 36 MPa and 55°C using extraction vessel #1. The C0 2 flow rate was 0.7 g/min. The samples obtained for analysis were c o l l e c t e d over f i v e i n t e r v a l s during the extraction (Fig. 146 5.25). The results of the analysis are given in Table 5.16. TABLE 5.16 Phospholipid content of C02 extracts of cooked Canola seed. Absorbance (Abs) readings are in absolute units. sample sample size Abs Phosholipid number mg % 1 7.26 0.007 <0.01 2 6.87 0.005 < 0.01 3 7.13 0.010 <0.01 4 7.20 0.024 <0.01 5 7.26 0.028 <0.01 At the completion of each experiment the extraction system was taken apart and the f r i t s and r e s t r i c t e r valve were washed with a mixture of chloroform and methanol. The washings were then made up to 2.0 mL using chloroform and four 50mL samples of the washings were analyzed for phosphorus. No phosphorus was detected in any of the washings. 5.8.4 Phosphorus in Carbon Dioxide Extracts Of Canola O i l At the start of the experiment 1.030 g of crude o i l was introduced into extraction vessel #1 which had previously been f i l l e d with 0.3 mm Ottawa sand and packed at both ends with glass wool. From the known.concentration of phospholipid in the o i l (Table 5.14) i t was determined that the t o t a l phospholipid within the vessel was 0.0193 g, which corresponds to a t o t a l phosphorus mass of 749 Mg. After extracting the o i l with C0 2 (flow rate 0.7g/min) at 55 °C and 36.0 MPa for 3.8 h, the glass wool and Ottawa sand were each washed several times 147 FIGURE 5.25 Extraction curve for a 4g sample of cooked Canola seed showing the intervals (dotted) over which samples of o i l were c o l l e c t e d for phosphorus analysis. Conditions: pressure 36 MPa, temperature 55°C, C0 2 flow rate 0.7 g/min. 1 48 with chloroform and the washings made up to 50 mL and 25 mL respectively. Phosphorus determinations were then car r i e d out on each washing. The results from these analyses indicated that at the end of the experiment the glass wool held approximately 497 Mg of phosphorus and the Ottawa sand 246 Mg for a t o t a l of 743 Mg. Within the l i m i t s of error of the experiment, the s u p e r c r i t i c a l C0 2 did not extract any phosphorus from the vessel. The detailed results from th i s experiment are summarized in Table 5.17. TABLE 5.17 Phosphorus analysis of chloroform washings of the glass wool and Ottawa sand after extraction. Absorbance (abs) is shown in absolute units. sample sample Abs phosphorus t o t a l size content phosphorus ML Mg Mg 50 wool 0.450 0.470 10 wool 0. 1 00 0.105 497 20 sand 0.207 0.215 10 sand 0.085 0.089 246 These results indicate that under the conditions tested and within the l i m i t s of the a n a l y t i c a l procedure, s u p e r c r i t i c a l carbon dioxide does not extract phospholipid material from the crushed Canola seed. These findings: are similar to those reported in the l i t e r a t u r e dealing with SFE of corn germ (Christianson et a l . , 1984) and cottonseed ( L i s t et a l . , 1983). In these reports the maximum phosphorus content of the extracted o i l s were 5 ppm and 1 ppm, respectively. In both of these studies the o i l s were extracted using C02 at 50 °C and 54 MPa. In two additional studies, which dealt with the application . 149 of SFE to soybean meal, i t was found that small amounts of phosphorus did appear in the o i l extracts. In the f i r s t of these studies (F r i e d r i c h and L i s t , 1982) i t was found that the extract obtained by using C0 2 • at 34 MPa and 50 °C, had a phosphorus concentration of 60 ppm. In the second study ( L i s t et a l . , 1982), i t was found that the o i l produced by using C0 2 at 54 MPa and 50 ° C, had a phosphorus concentration of 45 ppm. The differences in the phosphorus contents of the s u p e r c r i t i c a l extracts from the various seeds cannot be accounted for on the basis of phospholipid concentration in the d i f f e r e n t meals since a l l have approximately the same value (Sonntag, 1979). However, a possible reason for the differences may be that the phospholipids in each of the seeds are of d i f f e r e n t p o l a r i t i e s . Since i t i s known that a compound's s o l u b i l i t y in a s u p e r c r i t i c a l f l u i d i s strongly dependent on the compound's p o l a r i t y , i t would be expected that d i f f e r e n t p o l a r i t y phospholipids would also exhibit d i f f e r e n t s o l u b i l i t i e s in the C0 2 (Zosel, 1978). The fact that C0 2 does not remove the phospholipids from the Canola meal i s s i g n i f i c a n t . Conventional hexane extracted o i l requires an acid de-gumming step to reduce the phosphorus level s in the o i l to 5-10 ppm (Appelqvist, 1971). Since the o i l produced using s u p e r c r i t i c a l carbon dioxide has a phosphorus content in t h i s region to begin with, i t would not require acid de-gumming. An extraction process based on s u p e r c r i t i c a l extraction would in t h i s respect be simpler and less expensive than the corresponding hexane process. 150 5.9 Computer Simulation Of The Extraction Process 5.9.1 Introduct ion Extraction from a fixed bed i s a semi-batch process and can be mathematically described by d i f f e r e n t i a l Eqs. 3.3 and 3.4 which were introduced e a r l i e r . To solve t h i s system of d i f f e r e n t i a l equations a computer program was developed. The solution to the equations provided by the computer gave both the solvent and seed phase o i l concentration p r o f i l e s along the seed bed as a function of time. An additional program was used to generate the process extraction curves from the obtained concentration p r o f i l e s . Using the mathematical model, which i s described in chapter 3, i t was possible to determine approximate values of the overa l l mass transfer c o e f f i c i e n t s for the extraction process at d i f f e r e n t solvent v e l o c i t i e s . 5.9.2 Results In Figs. 5.26-5.33 the extraction curves generated by the computer are plotted along with the corresponding experimental data for crushed, f i n e l y chopped and flaked seed. The experimental data shown in F i g . 5.26 were obtained from the extractor #4 (p.58) containing 1.5 g of seed. Figs. 5.26-5.31 represent experimental data which was c o l l e c t e d using extraction vessel #1 containing approximately 4 g of seed while Figs. 5.32 and 5.33 represent data c o l l e c t e d with extraction vessel #3 containing 12 g samples of seed. The close agreement between the experimental data and the computer generated data suggests that the model was able to simulate the physical extraction process reasonably well. 151 CO 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 T i i i • • experiment -2 — model tfCTED (g) 0.48 0.56 ApK 5.0 -UL-1— X *« L L l a - -OIL 0.32 _ & i — CO -s C 3 -O CD 1 i i i i i i i i i i i i i i i i i i i i i I I 0.0 8.0 16.0 244) 32 JD 40.0 48.0 56 JD 64.0 72 JD 80.0 88 JJ 96.0 TOTAL C02 PASSED THROUGH EXTRACTOR (g) FIGURE 5.26 Extraction curve obtained by passing C0 2 at 55°C and 36 MPa through extraction vessel #4 containing 1.5 g of crushed seed. The C0 2 flow rate was 1.6 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 16.7 cm/min. The computed extraction curves were calculated using three d i f f e r e n t values of ApK. 152 i—i—i—i—i—i—i—i—i—r • experiment — model i — i — i — r i—i—i—i—r j i i i i i i i i i i i i i ' i » i » 0.0 16.0 32.0 48.0 64.0 80.0 96.0 112.0 128.0 144.0 160.0 176.0 192.0 TOTAL C02 PASSED THROUGH EXTRACTOR (g) FIGURE 5.27 Extraction curve obtained by passing C0 2 at 55°C and 36 MPa, through extraction vessel #1 containing 4.0 g of crushed seed. The C0 2 flow rate was 2.7 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 3.9 cm/min. The computed extraction curve was calculated using an ApK value of 2.0 gC0 2/cm 3min. 1 53 I I I I I I I I I • experiment — model i—i—i—i—i—i—i—i—i—i—i—i—r J i i i i i i i • J i i i i i • « i i 0.0 16.0 32.0 48.0 64.0 80.0 96.0 112.0 128.0 144.0 160.0 176.0 192.0 TOTAL C02 PASSED THROUGH EXTRACTOR (g) FIGURE 5.28 Extraction curve obtained by passing CO? at 55 °C and 36 MPa, through extraction vessel #1 containing 4.0 g of crushed seed. The C0 2 flow rate was 2.6 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 3.8 cm/min. The computed curve was calculated using an ApK value of 1.8 gC0 2/cm 3min. 154 3 i i i i i i i i i i i i i i i i i i i i i i i i i 0.0 40X1 80.0 120.0 160.0 200.0 240.0 280.0 320.0 360.0 400.0 440.0 480.0 TOTAL C02 PASSED THROUGH EXTRACTOR (g) FIGURE 5.29 Extraction curve obtained by passing C0 2 at 55 °C and 36 MPa, through extraction vessel #1 containing 3.8 g of commercially flaked seed. The C0 2 flow rate was 1.7 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 2.5 cm/min. The computed extraction curve was calculated using an ApK value of 1.3 gC0 2/cm 3min. 1 CQ i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r • experiment — model i i i i i ' I t i l l ! 0.0 50.0 100.0 150.0 200.0 250.0 300.0 350.0 400.0 450.0 500.0 550.0 600.0 TOTAL C02 PASSED THROUGH EXTRACTOR (g) FIGURE 5.30 Extraction curve obtained by passing CO? at 55 °C and 36 MPa, through extraction vessel #1 containing 4.0 g of f i n e l y chopped seed. The C0 2 flow rate was 2.3 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 3.5 cm/min. The computed extraction curve was calculated using an ApK value of 2.0 gC02/cm3min. 1 56 c4 3 3 o 3 O f N cr-cr y co O a T — i — i — i — i — i — i — i — i — i — i — r • experiment — model i i i i i i I i i i i J i » » » i i i i i 0.0 16.0 32.0 48.0 64.0 80.0 96.0 112.0 128.0 144.0 160.0 176.0 192.0 TOTAL C02 PASSED THROUGH EXTRACTOR (g) FIGURE 5.31 Extraction curve obtained by passing C0 2 at 55 °C and 36 MPa, through extraction vessel #1 containing 4.0 g of f i n e l y chopped seed. The C0 2 flow rate was 1.7 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 2.5 cm/min. The computed extraction curve was calculated using an ApK value of 1.5 gC0 2/cm 3min. 157 CO 00 cr CN* cr i— UJol 1"^  CD o S l I I I I I i i I • experiment — model i — i — i — i — i — i — i — i — r T — i i — r j i i i i i i i i L J I I I I I I L OJ) 80.0 160.0 240.0 320.0 400.0 480.0 560jD 640.0 720JD 800.0 880.0 9BQJ0 TOTAL C02 PASSED THROUGH EXTRACTOR (g) FIGURE 5.32 Extraction curve obtained by passing C0 2 at 55 °C and 36 MPa, through extraction vessel #3, containing 12.0 g of crushed seed. The C0 2 flow rate was 1.4 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 0.5 cm/min. The corresponding extraction curve generated by the computer an ApK value of 0.6 gC0 2/cm 3min. 158 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0 200.0 220.0 240.0 TOTAL C02 PASSED THROUGH EXTRACTOR (g) FIGURE 5.33 Extraction curve obtained by passing C0 2 at 55 °C and 36 MPa, through extraction vessel #3 containing 12.0 g of crushed seed. The C0 2 flow rate was 0.7 g/min which corresponds to an i n t e r s t i t i a l v e l o c i t y of 0.3 cm/min. The computed extraction curve was calculated using an ApK value of 0.4 gC0 2/cm 3min. 1 59 3 °3 Pr-cr UJ (_) o ' o in I l i o m I I I I I I I I I I I I I I I I I I I I I I I I • experiment — model i i i i i i i i i i i i i i i i 0.0 0.08 0.16 0^4 0.32 0.4 0.48 0.56 0.64 0.72 0.8 0.88 0.96 DISTANCE FROM BED ENTRANCE FIGURE 5.34 O i l concentration in the seeds (g o i l / g o i l -free seed) as a function of normalized distance from the bed .entrance aft e r 240 min. The conditions correspond to those shown in F i g . 5.33 1 60 Experimental data concerning the o i l concentration in the seed phase at d i f f e r e n t points along the bed were obtained by terminating an experiment prematurely, while the extraction curve was s t i l l l i n e a r , sectioning the the bed of seeds and then analyzing each section for i t s o i l content using hexane extraction (Sec. 4.6.1). The extraction curve for t h i s experiment i s shown in F i g . 5.33. In F i g . 5.34 the results from t h i s analysis are presented along with the corresponding values generated by the computer. The large errors which appear in the horizontal d i r e c t i o n of each data point were due to the i n a b i l i t y of precisely determining the position which the seed sections o r i g i n a l l y occupied in the bed. As indicated in the figure, the agreement between the experimental data and the computer data i s good. Computed values for o i l concentrations in the seed- and solvent-phases are shown in Figs. 5.35 and 5.36. In order to generate the data for these figures, the parameters in the model were set so that the model simulated extraction from the smallest extractor, packed with 1.5 g of seed. The nearly constant value of the seed o i l concentration in the extractor at t=45 min represents the si t u a t i o n where the 'readily accessible' o i l on the surface of the seed p a r t i c l e s has been depleted. At t h i s stage of the extraction, nearly a l l o i l removal comes from the i n t e r i o r of the seed tissue and from regions not exposed to the flow of moving solvent , i . e . between the fragments of seed. 161 I I I I I I I I I I I I I 1 I I I I I I I I 45min r i l l \ \ I i ' 0.0 0.08 0.16 0.24 0.32 0.4 0.48 0.56 0.64 0.72 0.8 0.88 0.96 DISTANCE FROM BED ENTRANCE FIGURE 5.35 O i l concentration in the solvent-phase (g o i l / g C0 2) as a function of normalized distance from the bed entrance at four d i f f e r e n t times. The conditions correspond to those shown in Fig 5.26. 1 62 3 1 ' ' I I I I i I I I I I i i i l I l I 1 I I 45min i < I I I I I > i i i i i i i i i i - • • - | 0.0 0.08 0.16 d24 0.32 0.4 0.48 0.56 0.64 0.72 0.8 0.88 0.96 DISTANCE FROM BED ENTRANCE FIGURE 5.36 O i l concentration in the seed-phase (g o i l / g o i l - f r e e seed) as a function of normalized distance from the bed entrance at four d i f f e r e n t times. The conditions correspond to those shown in F i g . 5.26. 163 As indicated by F i g . 5.35, when the extraction has proceeded to th i s stage (t=45 min) the solvent moves through the bed without becoming saturated with o i l . The solvent- and seed-phase o i l concentrations from the above simulation are also plotted as a function of both time and bed position in Figs. 5.37 and 5.38. The values of the o v e r a l l volumetric mass transfer c o e f f i c i e n t (ApK) for the extraction experiments shown in Figs. 5.26-5.33 were also determined. This was done by running the program with d i f f e r e n t values of ApK and matching the computed and experimental curves. The v e r t i c a l bars in F i g . 5.39 represent the range of ApK values which, when used in the model, produced extraction curves that matched the corresponding experimental curve reasonably well..The accepted value of ApK was taken to be the mean of the maximum and minimum of these values. A l l of the computer-generated curves were matched with the experimental data by 'eye'. An example of how the value of ApK a f f e c t s the simulated extraction curve i s shown in F i g . 5.26. In t h i s figure the extraction curves are shown for three d i f f e r e n t values of ApK together with the experimental data. It can be seen that the model predictions are r e l a t i v e l y sensitive to the ApK values; a change in the ApK value by ± 25% s i g n i f i c a n t l y a f f e c t s the shape of the extraction curve. Figure 5.39 is a -log plot of ApK values versus the log of solvent i n t e r s t i t i a l v e l o c i t y . The figure indicates that the value of ApK decreases with decreasing i n t e r s t i t i a l v e l o c i t y as expected. 1 64 FIGURE 5.37 O i l concentration in the solvent phase (g o i l / g C021) as a function of normalized distance from the bed entrance and time. The conditions correspond to those shown in F i g . 5.26. FIGURE 5.38 O i l concentration in the seed phase (g o i l / g o i l - f r e e seed) as a function of normalized distance from the bed entrance and time. The conditions correspond to those in F i g . 5.26. 166 r-LO ^ CO —I CO u . O T "Q_LO c r CO —I—I—I I I 1111 • crushed c flaked o finely chopped i—i—i i i 1111 I I I I I LI » J » • I • Im I J I I I I 111 10 3 5 710° 3 5 710' 3 5 710 2 INTERSTITIAL VELOCITY (cm/min) FIGURE 5.39 Volumetric mass transfer c o e f f i c i e n t s (ApK) as a function of i n t e r s t i t i a l v e l o c i t y (v). 167 No s i g n i f i c a n t difference was observed between the values of ApK for the flaked and finely-chopped seed. The slope of the l i n e in F i g . 5.39 was found to be 0.54 ± 0.2. The value of ApK may thus be related to the i n t e r s t i t i a l v e l o c i t y of the solvent by the following equation: ApK = 0.75v 0' 5" [5.4] where v represents the i n t e r s t i t i a l v e l o c i t y of the solvent in [cm/min]. The units for ApK in t h i s case are [gC0 2/cm 3 min] The value of exponent (0.54) i s reasonable and f a l l s between the value determined for mass transfer in l i q u i d s (0.33) (Wilson and Geankoplis, 1966) and mass transfer in gases (0.6) (Wakao et a l . , 1976) for Reynolds numbers between 0.0016 and 55. The flowrates used in the Canola experiments represented values of Reynolds numbers from 0.24 to 14. The close agreement between the experimental data and the data .generated by computer, indicates that the model simulates the extraction process reasonably well. This in turn suggests that the values of ApK as evaluated using the model are probably r e l i a b l e . These values can be expected to be applicable to a scaled-up version of the extraction process, provided the process- and flow-conditions in the scaled up version are c l o s e l y similar to those used in the experimental extractors. 1 68 VI. CONCLUSIONS This study has demonstrated that s u p e r c r i t i c a l C0 2, under the appropriate conditions, can be an e f f e c t i v e solvent for extracting o i l from Canola seed. However, the maximum s o l u b i l i t y of Canola o i l in C0 2, under the conditions studied, was low compared with o i l s o l u b i l t y in hexane. Nonetheless, the degree to which the seeds could be extracted was similar for both C0 2 and hexane. Additionally, the o i l produced by the C0 2 extraction process contained less impurities than the hexane extracted o i l these results suggest that the C0 2 extraction process could be a viable a l t e r n a t i v e to the conventional hexane extraction process. In order to further evaluate t h i s p o s s i b i l i t y , a p i l o t -plant scale version of the extraction system should be b u i l t and operated. The mathematical model presented herein should be of value in the design and operation of such a system. The major findings of thi s study can be summarized as follows: 1) Over the range of pressures and temperatures studied, the maximum s o l u b i l i t y of Canola o i l in C0 2 (l.1wt%) was observed at 36 MPa and 55°C. 2) The equilibrium concentration of o i l in the s u p e r c r i t i c a l C0 2 was independent of the o i l concentration in the seeds. 169 3) The t o t a l amount of o i l recoverable from the seeds, using C0 2, was dependent upon the method of seed treatment prior to extraction. For commercially flaked seed, t h i s value was comparable to the amount of o i l recoverable by conventional hexane extraction. 4) The • fatty acid composition of the C0 2 extracted o i l was constant for most of the extraction. The o i l extracts obtained very late in the extraction, however, tended to have s l i g h t l y higher concentrations of heavier (C22, C24) fatty acids. 5) The C0 2 extracted o i l was e s s e n t i a l l y free (< 7ppm) from phosphorus and, in t h i s regard, comparable to commercially refined Canola o i l . 6) Experimental and computed o i l concentration p r o f i l e s and extraction rates were in good agreement. 7) The o v e r a l l volumetric mass transfer c o e f f i c i e n t for the extraction process at 36 MPa and 55°C was correlated with solvent i n t e r s t i t i a l v e l o c i t y by the equation: ApK 0.75 v° 5" 170 VII. RECOMMENDATIONS The following recommendations are suggested for future research. 1) A broader comparison between the quality of C0 2- and hexane-extracted o i l should be made. The comparison should include the tests which are routinely used in industry to establish o i l quality eg. Free fatty acid content, color, chromatographic re f i n i n g l o s s . 2) The effectiveness of using an entrainer such as ethanol or acetone in conjunction with the C0 2 should be studied with the aim of enhancing the o i l s s o l u b i l i t y in the C0 2 and providing a means by which the o i l could be separated from the C0 2 solely by a temperature change. 3) The v a l i d i t y of the method for ca l c u l a t i n g t r i g l y c e r i d e s o l u b i l i t i e s in the C0 2 (Sec. 5.7.5) should be established by extracting known composition o i l s and comparing predicted and observed o i l s o l u b i l i t i e s . 4) The v a l i d i t y of the extraction model as discussed in Chapter 3 should be established over a wider range of operating conditions and the values of x, and x 2 (Chap. 3) should be obtained for each d i f f e r e n t type of seed pre-treatment 171 5) A routine should be incorporated into the computer program which would calculate the best f i t value of ApK based on the experimental data. 6) An economic assessment of the C0 2 extraction process should be made; i f the economic analysis seems favorable, a p i l o t - p l a n t scale extraction system should be b u i l t and operated with the aim of gathering the engineering data needed for a f u l l - s c a l e C0 2 extraction plant. 172 NOTATION A -cross sectional area of extractor [m2] Ap -surface area for mass transfer per unit volume [m2/m3] ApK - o v e r a l l volumetric mass transfer c o e f f i c i e n t [kgC0 2/m 3s] H -height of seed bed [m] h -distance along extractor [m] K - o v e r a l l mass transfer c o e f f i c i e n t [kgC0 2/m 3s] m -mass of o i l extracted from seed bed [kg] e m -mass of o i l in seed bed before extraction [kg] o m -mass flowrate of solvent [kg/s] S t - t o t a l s o l u b i l i t y of a mixture in C 0 2 [kg/kg C 0 2 ] sa»b " S o l u b i l i t y of pure components a,b in C 0 2 [kg/kg C 0 2 ] t -time [s] U - s u p e r f i c i a l solvent v e l o c i t y [m/s] v - i n t e r s t i t i a l solvent v e l o c i t y [m/s] x -seed o i l concentration [kg o i l / k g o i l - f r e e seed] x 0 - i n i t i a l seed o i l concentration [kg o i l / k g o i l - f r e e seed] x, -seed o i l concentration at which bare o i l - f r e e surfaces appear [kg o i l / k g o i l - f r e e seed] x 2 - seed o i l concentration at which surface o i l i s depleted [kg o i l / k g o i l - f r e e seed] x 3 - the concentration of o i l in the intact seed tissue which cannot be extracted by C 0 2 [kg o i l / k g o i l - f r e e seed] 173 x a b - m ° l e f r a c t i o n of components a,b in solution [dimensionless] y -concentration of o i l in C0 2 phase [kg o i l / k g C0 2] y 0 -concentration of o i l in C0 2 solvent at extractor outlet [kg o i l / k g C0 2] y* -concentration of o i l in the solvent in equilibrium with seeds having o i l concentration x 8h -height of an element of extractor [m] e -seed bed voids [dimensionless] p -density of C0 2 [kg/m3] p -density of o i l - f r e e seeds [kg/m3] s 174 VII. REFERENCES Ackman, R.G. , Sipos, J. 1964. Application of s p e c i f i c response factors in gas chromatographic analysis of methyl exsters of fatty acids with flame ionization detectors. J. Am. O i l Chem.  Soc. 41: 377. Ackman, R.G. 1983. Chemical composition of Rapeseed o i l . High  and Low Erucic Acid Rapeseed O i l s . J.G. Kramer, D.S. Frank, Eds. Academic Press. Toronto, Can. pp. 85-124. Anjou, K. 1972. Manufacture of Rapeseed O i l and Meal. Rapeseed:  C u l t i v a t i o n , Composition, Processing and U t i l i z a t i o n . L.A. Appelqvist, Ed. Elsevier Pub. Co. New York. pp. 198-217. Appelqvist, L.A. 1972. Chemical constituents of Rapeseed. Rapeseed: C u l t i v a t i o n , Composition, Processing and U t i l i z a t i o n . L.A. Appelqvist, Ed. Elsevier Pub. Co. New York. pp. 123-173 B a l z a r i n i , D.; Ohrn, K. 1972. Coexistence curve of sulfur hexafluoride. Phys. Lett. , 29(13): 840-842. Beach, D.H.C. 1983. Rapeseed crushing and extraction. High and  Low Erucic Acid Rapeseed O i l s . J.G. Kramer, D.S. Frank, Eds. Academic Press. Toronto, Can. pp. 181-184. Bengtsson, L.; von Hofsten, A.;. Loof, B. 1972. Botany of Rapeseed. Rapeseed: C u l t i v a t i o n , Composition, Processing and U t i l i z a t i o n . L.A. Appelqvist, Ed. E l s v i e r Pub. Co. New York. pp. 36-48. Booth,H.S.; Bidwell, R.M. 1949. S o l u b i l i t y measurements in the c r i t i c a l region. Chem. Revs. 44: 477-513. Boulter, G.S., 1983. The history and marketing of Rapeseed O i l in Canada. High and Low Erucic Acid Rapeseed O i l s . J.G. Kramer, D.S. Frank, Eds. Academic Press. Toronto, Can. pp. 62-82. Brogle, H. 1982. C0 2 as a solvent: i t s properties arid applications. Chemistry and Industry June: 385-390. Brunner, G.; Peter, S. 1982. On the s o l u b i l i t y of glycerides and fatty acids in compressed gases in the presence of an entrainer. Sep. S c i . Tech. , 17(1): 199-214. Bunzenberger, G.; Lack, E.; Marr, R. 1984. C0 2-extraction: comparison of super-and s u b c r i t i c a l extraction conditions. Ger.  Chem. Eng. , 7: 25-31. Calame, J.P.; Steiner, R. 1982. C0 2 extraction in flavor and perfumery industries. Chemistry and Industry June 19, pp. 399-402. Campbell, H. 1983. MASc. Thesis. Department of Chemical Engineering, University of B r i t i s h Columbia, Vancouver, B.C. 175 Caragay, A.B. 1981. S u p e r c r i t i c a l f l u i d s for extraction of flavors and fragrances from natural products. Perfumer and  Fl a v o r i s t 6(4):46-46,48,51-55. Christianson, D.D.; F r i e d r i c h , J.P.; L i s t , G.R.; Warren, K.; Bagley, E.B.; Stringfellow, A.C.; Inglett, G.E. 1984. Su p e r c r i t i c a l f l u i d extraction of dry-milled corn germ with carbon dioxide. J. Food S c i . , 49(1): 229-232. Clandinin, D.R. 1981. Canola Meal for Livestock and Poultry Pub. No. 59. D.R. Clandinin Ed. Canola Council of Canada. Czubryt, J.J.; Myers, M.N.; Gidding, J.C. 1970. S o l u b i l i t y phenomena in dense carbon-dioxide gas in the range 270-1900 atmospheres. J. Phys. Chem. 74(24): 4260. de F i l i p p i , R.P. 1982. C0 2 as a s o l v e n t — a p p l i c a t i o n to fats, o i l s and other materials. Chem. Ind. (London) 2(12): 390-394. Duck-Chong, C.G. 1979. A rapid and sensitive method for determining phospholipid phosphorous involving digestion with magnesium n i t r a t e . Lipids 14(5): 492-497. E l l i s , S.R.M.; V a l t e r i s , R. 1965. Recent developments in extraction processes. Chem. Ind. (London) : 2027 Formo, M.W. 1979. Fats in the d i e t . Bailey's Industrial O i l and  Fat Products. Vol. 1, Fourth Ed. D. Swern Ed. Wiley-Interscience Pub. New York. F r i e d r i c h , J.P.; L i s t , G.R. 1982. Characterization of Soybean o i l extracted by s u p e r c r i t i c a l carbon dioxide and hexane. J.  Agric. Food Chem. , 39: 192-193. F r i e d r i c h , J.P.; L i s t , G.R.; Heakin, A.J. 1982. Petroleum-free extraction of o i l from Soybeans with s u p e r c r i t i i c a l carbon dioxide. J. Agric. Food Chem. , 28(6): 1153-1157. Gardner, D.S. 1982. Industrial scale hop extraction with l i q u i d C0 2. Chemistry and Industry , June 19, :402 404. Gearhart, J.A.; Garwin, L. 1976. Resid-extraction offers f l e x i b i l i t y . O i l Gas J. June 14, :63. Gere, D.R., 1983. S u p e r c r i t i c a l f l u i d chromatography Sc ience 222: 253-259. Giddings, J . C ; Myers, M.N.; King, J.W. 1969. Dense gas chromatography at pressures to 2000 atmospheres. J. Chrom. S c i . 7: 276. Giddings, J . C ; Myers, M.N.; McLaren, L.; K e l l e r , R.A. 1968. High pressure gas chromatography of nonvolatile species. Sc ience 162: 67. 176 Hannay, J.B.; Hogarth, J. 1879. On the s o l u b i l i t y of solids in gases. Proc. Roy. Soc. (London). 29: 324-326. Hellyar, K.G.; de F i l i p p i , R.P. 1982. Extraction processes using solvents near their thermodynamic c r i t i c a l point. The Chemical  Engineer. A p r i l : 136-138. Hofsten, A.V. 1970. C e l l u l a r structure of Rapeseed. Proc. Int.  Conference on the Science, Technology and Marketing of Rapeseed  and Rapeseed Products. Quebec, Can. Sept. 20-23: 70-85. Holm, L.W.; Josendal, V.A. 1974. Mechanism of o i l displacement by C0 2. J . Petr. Tec. 22: 1058. Hron, R.J.; Koltun, S.P.; Graci, A.V. 1982. Bio-renewable solvents for vegetable o i l extraction. J. Am. O i l Chem. Soc. , 59(9): 674A-684A. Hubert, P.; Vitzthum, 0. 1978. F l u i d extraction of hops, spices and tobacco with s u p e r c r i t i c a l gases. Angew. Chem. Int. Ed.  Engl. 17(10): 710-715. Humphrey, J . ; Rocha, J.; F a i r , J.R. 1984. The essentials of extraction. Chem. Eng. Sept. 17. pp. 76-95. Khan, L.M.; Hanna, M.A. 1983. Expression of o i l from oilseeds-a review. J. Agric. Engng. Res. 28: 493-503. King, M.B.; Alderson, D.A.; Fallah, F.H.; Kassin, D.M.; Kassim, K.M.; Sheldon, J.R.; Mahmud, R.S. 1983. Some vapour/liquid and vapor/solid equilibrium measurements of relevance for s u p e r c r i t i c a l extraction operations and their c o r r e l a t i o n . Chemical Engineering at S u p e r c r i t i c a l Conditions. M.E. Paulaitus, J.M.L. Penninger, R.D. Gray, P. Davidson, Eds. Ann Arbor Science Pub. pp. 31-80. King, M.B.; Bott, T.R. 1982. Problems associated with the development of gas extraction and similar processes. Sep. S c i .  Tech. , 17(1): 119-150. Knapp, D.R. 1979. Handbook of Ana l y t i c a l P e r i v a t i z a t i o n  Reactions. John Wiley and Sons, New York. pp. 164-165. Krans, C.A. 1922. The properties of e l e c t r i c a l l y conductive systems, A.C.S. Monograph No. 7, The Chemical Catalog Co. Inc. New York. Mackay, M.E.; Paulaitus, M.E. 1979. So l i d s o l u b i l i t i e s of heavy hydrocarbons in s u p e r c r i t i c a l solvents. Ind. Eng. Ch. Fundls. 18: 149-153. Maron, S.H.; Prutton, G.F. 1965. Pri n c i p l e s of Physical  Chemistry. 4th Ed. MacMillan. Co. New York, N.Y. 1 77 McLaren, L.; Myers, M.N.; Giddings, J.C. 1968. Dense-gas chromatography of nonvolatile substances of high molecular weight. Sc ience 159: 197. Merritt, C ; Vajdi, M.; Kayser, S.G.; Halliday, J.W.; Bazinet, M.L. 1982. Validation of computational methods for t r i g l y c e r i d e composition of fats and o i l s by l i q u i d chromatography and mass spectrometry. J . Am. O i l Chem. Soc. , 59(10): 422-432. M i l l i g a n , E.D.: Tandy, D.C. 1974. D i s t i l l a t i o n and solvent recovery. J. Am. O i l Chem. Soc. , 51(8): 347-350. Newitt, D.E.; Pai, M.V.; Kuloor, N.R.; Huggill, A.W. 1956.• Thermodynamic Functions of Gases , Vol. 1. Din, F.; Ed. Butterworths S c i e n t i f i c Publications; London, England. N i g g l i , P. 1912. The Gases in magmas. Z. Anorg. Chem. 75: 161-188. Othmer, D.F.; Agarwal, J.C. 1955. Extraction of Soybeans: theory and mechanism. Chem. Eng. Progr. , 51(8): 372-378. Paquot, C. 1979. International Union of Pure and Applied Chemistry (IUPAC), Method I.B.2. Standard Methods for the Analysis of O i l s , Fats and Derivatives. 6th Ed. Part 1. Ed. C. Paquot. Pergamon Press, N.Y. Paul, P.F.M.; Wise, W.S. 1971. The Pr i n c i p l e s of Gas Extraction. M and B Monographs. M i l l s and Boon, Publishers, London, U.K. Paulaitus, M.E.; Penninger, J.M.L.; Gray, R.D.; Davidson, P. Eds. 1983. Chemical Engineering at S u p e r c r i t i c a l Conditions. Ann Arbor Science Pub. pp. 183-373. Persmark, U. 1972. Analysis of Rapeseed o i l . Rapeseed:  Cu l t i v a t i o n , Composition, Processing and U t i l i z a t i o n . L.A. Appelqvist, Ed. Elsev i e r Pub. Co. New York. pp. 174-197. Peter, S.; Brunner, G. 1978. The separation of non-volatile substances Angew. Chem. Int. Ed. Engl. 17(10): 746-750. Pigden, W.J. 1983. World production and trade of Rapeseed and Rapeseed products, high and Low Erucic Acid Rapeseed O i l s . J.G. Kramer, D.S. Frank, Eds. Acedemic press, toronto. Can. pp. 21-58. Randall, L. 1982. The present status of dense(supercritical) gas extraction and dense gas chromatography: impetus for DGC/MS development. Sep. S c i . Tech. , 17(1): 1 -118. 1 78 Rowlinson, J.S.; Richardson, M.J. 1959. S o l u b i l i t y of solids in compressed gases. Advances in Chem. Phys. I. Prigogine, Ed. Interscience Pub. Inc. 2: 85-118. Rutkowski, R.; Gwiazdas, S.; Krygien, K. 1982. Sulfur compounds af f e c t i n g the processing of Rapeseed. J. Am. O i l Chem. Soc. , 59(1): 864-866. Schneider, G.M. 1978. Physiochemical p r i n c i p l e s of extraction with s u p e r c r i t i c a l gases. Angew. Chem. Int. Ed. Engl. 17(10): 716-727. Shehata, A.Y.; de Man, J.M.; Alexander, J.C. 1970. A simple and rapid method for the preparation of methyl esters of fats in milligram amounts for gas chromatography. Can. Inst. Food Technol. J. 3(3): 85-89. Shimshick, E.J. 1983. Extraction with s u p e r c r i t i c a l C0 2. Chemtech, June: 136-138. Sonntag, 1979. Composition and c h a r a c t e r i s t i c s of individual fats and o i l s . Bailey's Industrial O i l and Fat Products. Vol. 1, Fourth Ed.; D. Swern Ed. Wiley Interscience Pub. New York, N.Y. Sosulski, F.; Zadernowski, R.; Babuchowski, K. 1981. Composition of polar l i p i d s in Rapeseed. J. Am. O i l Chem. Soc. , 58(4): 561-564. Stahl, E.; Gerrard, D. 1982. Detoxification of Absinthe herb using C0 2. Planta Med. 45(3): 147. Stahl, E.; Quirin, K.W. 1982. Extraction et fractionnement de l i p i d e s et d'autres produits natural a l'aide de gaz supercritiques et l i q u e f i e s . Revue Francaise Pes Corps Gras. 29(6-7): 259-268. Stahl, E.; S c h i l z , W.; Schutz, E.; W i l l i n g , E. 1978. A quick method for the microanalytical evaluation of the dissolving power of s u p e r c r i t i c a l gases. Angew. Chem. Int. Ed. Engl. 17(10): 731-738. Stahl, E.; Schulz, E.; Mangold, H.K. 1980. Extraction of seed o i l s with l i q u i d and s u p e r c r i t i c a l carbon dioxide. J . Agr. Food 28(6): 1153-1157. 1 79 Stahl, E.; Schutz, E. 1980. Extraction of natural compounds with s u p e r c r i t i c a l gases. 3. Pyrethrum extracts with l i q u i f i e d and s u p e r c r i t i c a l carbon-dioxide. Planta Med. 40(1): 12-21 Stalkup, F.I., 1978. Carbon dioxide miscible f l o o d i n g — p a s t , present and outlook for future. J. Petro. Tec. 30(aug): 1102-1112. Stanly, D.W.; G i l l , T.A.; de Man, J.M.; Tung, M.A. 1976. Microstructure of Rapeseed. Can. Inst. Food S c i . Technol. J . 9(2): 54-60. Teasdale, B.F.; Mag, T.K. 1983. The commercial processing of low and high erucic acid Rapeseed o i l s . High and Low Erucic Acid  Rapeseed O i l s . J.G. Kramer, D.S. Frank, Eds. Acedemic Press. Toronto, Can. pp. 198-228. Treybal, R.E. 1968. Mass Transfer Operations. Second Ed. McGraw-H i l l Book Co. New York. Tyrer, D. 1910. S o l u b i l i t i e s below and above the c r i t i c a l temperature. J . Chem. Soc. 97: 621-632. Van Caeseele, L.; M i l l s , J.T.; Sumner, M.; G i l l i s p i e , R. 1982. Cytological study of palisade development in the seed coat of Candle Canola. Can. J. Bot., 60: 2469-2475. Vetere, A. 1979. A predictive method for ca l c u l a t i n g the s o l u b i l i t y of sol i d s in s u p e r c r i t i c a l gases: application to apolar mixtures. Chem. Eng. S c i . , 34: 1393-1400. Vollbrecht, R. 1982. Extraction of hops with s u p e r c r i t i c a l C0 2 . Chemistry and Industry. June 19, pp. 397-399. Vukalovich, M.P.; Altunin, V.V. 1968. Thermophysical Properties  of C0 2. London, Wellingborough, C o l l e t s . Williams, D.F. 1981. Ext.raction with s u p e r c r i t i c a l gases. Chem.  Eng. S c i . , 36(11): 1769-1788. Yiu, S.H.; Poon, H.; Fulcher, R.G.; Altosaar, I. 1982. The microscopic structure and chemistry of Rapeseed and i t s products. Food Microstructure , 1: 135-143. Zosel, K. 1978. Separation with s u p e r c r i t i c a l gases: p r a c t i c a l applications. Angew. Chem. Int. Ed. Engl. 17(10): 702. Zhuze, T.P. 1958. S o l u b i l i t y of substances in compressed gases. Itaqi Nauki, Khim. Nauki. Akad. Nauk S.S.S.R. 2: 450-477. Zhuze, T.P. 1960. Use of compressed hydrocarbon gases as solvents. Petroleum (London) 23: 298-300. 180 Zhuze, T.P. Yushkevich, G.N. Gekker, I.E. 1958. Extraction of l a n o l i n from wool fat with the aid of compressed gas. Masloboino-Zhirovaya Prom. 24(6): 34-37. Zhuze, T.P.; Yushkevich G.N.; Safronova, T.P. 1959. Extraction of raw ozocerite from ozocerite ores with compressed gases. Trudy Inst. N e f t i , Akad. Nauk S.S.S.R. 13: 275-279. 181 APPENDIX I_ Errors Associated With The Fatty Ac id Analysis It i s assumed that the t r a n s e s t e r i f i c a t i o n procedure i s quantitative and that the reagent reacts no n - s p e c i f i c a l l y with a l l the fatty acid moieties present in the o i l , i rrespective of their carbon number or their position within the t r i g l y c e r i d e (Knapp, 1979). Additionally, since the fatty acid composition of the o i l samples was determined only in a r e l a t i v e manner, there are no errors associated with sample preparation. Two sources of error are however s i g n i f i c a n t : i) measurement error associated with the chromatograph, i i ) accuracy of o r i g i n a l a n a l y t i c a l standards upon which response factors are based. Measurement error In order to assess the f i r s t source of error, a standard mixture (Supelco Rapeseed O i l Mixture-CT) of fatty acid esters was analyzed f i v e times over a period of days. The results of these analyses and the measurement errors associated with the technique are l i s t e d in Table 1. TABLE 1. Composition of Supelco CT fatty acid ester mixture as determined by chromatograph. Values are in weight %. ester sample sample sample sample sample standard standar 1 2 3 4 5 deviat ion error C1 4:0 1.21 1 .03 1 .05 1 .02 1 .07 0.08 0.1 C1 6:0 4.46 4.12 4.23 4.13 4.14 0.14 0.1 C18:0 3.10 2.99 3.02 3.02 3.02 0.01 0.1 C18:1 60.23 59.95 60. 1 1 59.70 59.76 0.22 0.2 C18:2 1 1 .55 11.41 11.61 1 1 .50 11.53 . 0.07 0.1 C20:0 3.00 3.12 3.07 3.13 3.13 0.06 0.1 C20:1+C18: 3 6.03 6.02 6.09 6.05 6.05 0.03 0.1 C22:0 2.91 3.12 2.97 3.13 3.09 0.10 0.1 C22: 1 4.87 5.24 4.97 5.24 5.18 0.17 0.2 C24:0 2.76 2.99 2.86 3.05 3.00 0.12 0.1 182 Quantitative Standards In order to determine the GC response factors for the fatty acid esters, Supelco GLC standard mixtures (GLC-10, GLC-40, GLC-50 and GLC-60) were used. The standard error associated with these mixtures, as spe c i f i e d by Supelco, was ±0.5%. For a mixture containing 4 compounds t h i s translated to 2% r e l a t i v e error for each component. Each measurement reported by the GC contains both of these errors. For the high concentration fatty acid esters (C18:1, C18:2, C18:3), the error associated with the accuracy of the response factors predominates, while for the remaining esters, the measurement error associated with the instrumental techniques i s be the most important source of error. In Table 2 a summary of the errors i s l i s t e d . TABLE 2. Errors associated with measurement technique and response factors . A l l values are in Weight %. fatty measurement response t o t a l ac id standard factor absolute ester error error error C14:0 0.1 0.02 0.1 C1 6:0 0.1 0.08 0.2 C18:0 0.1 0.06 0.2 C1 8:1 0.2 1 .20 1 .4 C18:2 0.1 0.23 0.3 C20:0 0.1 0.06 0.2 C20:1 + C18:3 0.1 0.12 0.2 C22:0 0.1 0.06 0.2 C22:1 0.1 0.10 0.2 C24:0 0.1 0.06 0.2 183 APPENDIX II_ The program which was used to generate the concentration p r o f i l e s can be roughly divided into three sections. In the f i r s t section, values are assigned to a number of variables associated with the problem. These variables were of two types: those which could be measured d i r e c t l y , and those which could not be measured d i r e c t l y and for which values had to be assumed. Among the f i r s t type of variable were extractor volume, o i l s o l u b i l i t y in the solvent, seed packing density, seed weight, solvent flow rate, solvent density and the i n i t i a l seed o i l concentration. The second type of variable included the seed o i l concentration at which bare o i l - f r e e seed surfaces f i r s t appear (described in Chap. 3), the density of o i l - f r e e seed tissue and the volumetric mass transfer c o e f f i c i e n t s for the extraction process. In the second section of the program the various parameters used in the model are scaled in order to obtain results which can be compared with those obtained from the experimental extractors. This was necessary because the main program solves the mass balance equations for a unit volume extractor. The t h i r d section of the program i s e s s e n t i a l l y a series of subroutines which numerically solve the set of d i f f e r e n t i a l equations (Eqs. 3.4, 3.5). This package of subroutines was developed at the Argonne National Laboratory (Hyman, 1976), and was made available through the UBC computer centre under the l i b r a r y name "M0L1D". In addition to the main program a second program was used to produce the process extraction curves. The program generated 184 the curves from the o i l concentration p r o f i l e s provided by the main program. Also included are the programs which were used to generate the three-dimensional o i l concentration p r o f i l e s of the solvent and seed phases respectively. 185 C MODEL OF EXTRACTION FROM A FIXED BED C IMPLICIT REAL*8(A - H,0 - Z) REAL*8 MFLOW DIMENSION UZ(2,101), XM(lOl), M0RD(2,3), TOUT(401) COMMON /MOLPLT/ XAL, NPLT, MG, UAL, XAXIS(2), UAXIS(2,2) COMMON BEPS, RHOG, RHOS, XC1, XC2, XC3, Y2, F, C, D, E, V COMMON YEQ, CON1, CON2, RHOGV, VBEPS, DX, NPTS c c BEPS BED VOIDAGE [-] c RHOG - DENSITY OF C02 PHASE [g/cm3] c RHOS - DENSITY OF OIL-FREE SEED PHASE [g/cm3] c XC2 - OIL CONCENTRATION CONSTANT [g OIL/g OIL-FREE SEED] c XC1 - OIL CONCENTRATION CONSTANT [g OIL/g OIL-FREE SEED] c XC3 - OIL CONCENTRATION CONSTANT [g OIL/g OIL-FREE SEED] c Y2 - MAXIMUM VALUE OF ApK [g C02/cm3min] c F - RATIO OF ApK(at XC2) to ApK(at XC1) c VOLU - VOLUME OF EXPERIMENTAL EXTRACTION BED c AREA - CROSS-SECTIONAL AREA OF EXPERIMENTAL EXTRACTOR [cm2] c YEQ - SOLUBILITY OF THE OIL IN THE C02 [g o i l / g C02] c CON1 - SPECIFIC VOLUME OF C02 [cm3/g] c CON 2 - SPECIFIC VOLUME OF SEED [cm3/g] c APK - VOLUMETRIC MASS TRANSFER COEFFICIENT [g C02/cm3min] c BPACK - OIL-FREE SEED BED PACKING DENSITY [g/cm3] c TRES - RESIDENCE TIME IN EXTRACTOR [min] c SEED - MASS OF SEED IN EXPERIMENTAL EXTRACTOR [g] c OILC - CONCENTRATION OF OIL IN THE WHOLE SEED [g o i l / g seed] c EOIL EFFECTIVE OIL CONCENTRATION ie CONCENTRATION OF OIL IN c SEED, EXTRACTABLE BY C02 [g o i l / g seed] c RFLOW - EXPERIMENTAL C02 MASS FLOWRATE [g/min] c SUPER - EXPERIMENTAL SUPERFICIAL C02 VELOCITY [CM/MIN] c VELOC - EXPERIMENTAL C02 INTERSTITIAL VELOCITY [CM/MIN] c MFLOW C02 MASS FLOWRATE -UNIT VOLUME EXTRACTOR [G/MIN] c V - SUPERFICIAL C02 VELOCITY -UNIT VOL EXTRACTOR [CM/MIN] c c VBEPS INTERSTICIAL C02 VELOCITY- UNIT VOL EXTRACTOR [CM/MIN] c c c SET UP EXTRACTION PARAMETERS c SEED = 1.46D0 BEPS = 0.61D0 RFLOW = 1.600D0 RHOS = 0 .8D0 RHOG = 0 .88D0 YEQ = 0. 01 1D0 Y2 = 2 .00D0 F = 30. 0D0 c REASONABLE VALUES OF XC2 RANGE FROM 0.25 TO 0.67 XC2 = 0. 35D0 c REASONABLE VALUES OF XC1 RANGE FROM 0.35 TO 0.0 XC1 = 0. 20D0 C CANOLA SEED VARIES IN OILC FROM 0.35 - 0.45 OILC = 0.40D0 C REASONABLE VALUES OF XC3 RANGE FROM 0.0 TO 0.05 XC3 = 0.02D0 EOIL = (OILC - (OILC*XC3)) BPACK = RHOS*(1-BEPS) C CROSS SECTIONAL AREA OF EXTRACTOR #3 C XAREA = 5.067D0 C CROSS SECTIONAL AREA OF EXTRACTOR #1 AND #2 C XAREA = 1.267D0 C CROSS SECTIONAL AREA OF EXTRACTOR #4 XAREA = 0.178D0 C C CALCULATE FLOWRATES, VELOCITIES, VOLUMES AND OTHER CONSTANTS C C = (Y2/(XC2 - XC1)) * (1.D0-1.DO/F) D = Y2 - C * XC2 E = Y2 / (XC1*F) MFLOW = (BPACK/(SEED*(1 - EOIL))) * RFLOW VOLU = SEED * (1 - EOIL) / BPACK AREA = 1.0000D0 V = MFLOW / (RHOG*AREA) CON1 = 1.D0 / (BEPS*RHOG) CON2 = 1.D0 / ((1.D0-BEPS)*RHOS) RHOGV = RHOG * V VBEPS = V / BEPS TRES = BEPS / (MFLOW/RHOG) SUPER = (RFLOW/RHOG) / XAREA VELOC = SUPER / BEPS C WRITE (7,10) C, D, E 10 FORMAT CC = ', F6.3, 2X, ' D = ' , F7.3, 2X, ' E = ' , F5.3) WRITE (7,20) BEPS 20 FORMAT ('EXPERIMENT BED VOIDAGE = ', F5.3) WRITE (7,30) V 30 FORMAT ('MODEL SUPERFICIAL VEL (CM/MIN) = *, F5.3) WRITE (7,40) TRES 40 FORMAT ('EXTRACTOR RESIDENCE TIME (MIN) = ', F7.3) WRITE (7,50) VBEPS 50 FORMAT ('MODEL INTERSTITIAL VEL (CM/MIN) =', F5.3) WRITE (7,60) MFLOW 60 FORMAT ('MODEL MASS FLOW (G/MIN) = *, F5.3, 4X) WRITE (7,70) VOLU 70 FORMAT ('EXPERIMENT VOLUME OF BED (CM3) = ', F7.3) WRITE (7,60) VELOC 80 FORMAT ('EXPERIMENT INTERSTITIAL VELOC (CM/MIN) = F7.3) WRITE (7,90) SUPER 90 FORMAT ('EXPERIMENT SUPERFICIAL VELOC (CM/MIN) = ', F7.3) C C START OF MOL1D PROGRAM C NPDE = 2 C 'NPTS' SPECIFIES THE NUMBER OF SPACE SECTIONS THAT THE BED C IS DIVIDED INTO NPTS = 31 KEQN = 4 KBC = 2 DO 100 I = 1, NPDE MORD(1,1) = 0 MORD(I,2) = 0 100 MORD(1,3) = 0 MORD(2,l) = 0 METH = 20 EPS = 1.D-4 TINT = 0.D0 C 'TLAST' IS THE ENDTIME (IN MIN) OF THE EXTRACTION TLAST = 45.DO MOUT = 1 187 KMOL = 0 C THE NUMBER OF ITERATIONS OF THE DO-LOOP MUST EQUAL 'TLAST' DO 110 I = 1, 45 110 TOUT(I) = 1 * 1 . 0 D 0 C C DEFINE INITIAL CONDITIONS C DX = 1.D0 / DFLOAT(NPTS - l ) DO 120 I = 1, NPTS XM(I) = DFLOAT(I - 1) * DX UZ(1,1) = 0.D0 UZ(2,I) = EOIL / (1 - EOIL) 120 CONTINUE C CALL MOL1D(NPDE, NPTS, KEQN, KBC, METH, EPS, MORD, TINT, TLAST, 1 MOUT, TOUT, UZ, XM, KMOL) STOP END SUBROUTINE FUNC(F, U, UX, UXX, T, XM, IX, NPDE) IMPLICIT REAL*8(A - H,0 - Z) DIMENSION F(NPDE), U(NPDE), UX(NPDE), UXX(NPDE) C C CALCULATE FLUX FUNCTION, IF IT EXISTS C RETURN END SUBROUTINE BNDRY(T, UL, AL, BL, CL, UR, AR, BR, CR, NPDE) IMPLICIT REAL*8(A - H,0 - Z) DIMENSION UL(NPDE), AL(NPDE), BL(NPDE), CL(NPDE), UR(NPDE), 1 AR(NPDE), BR(NPDE), CR(NPDE) C C SET BOUNDARY CONDITIONS C AL(1) = 1.D0 RETURN END SUBROUTINE PDE(UT, U, UX, UXX, FX, T, XM, IX, NPDE) IMPLICIT REAL*8(A - H,0 - Z) DIMENSION U(NPDE,1), UT(NPDE,1), UX(NPDE,1), UXX(NPDE,1) DIMENSION FX(NPDE,1), XM(1) LOGICAL F1, F2 COMMON BEPS, RHOG, RHOS, XC1, XC2, XC3, Y2, F, C, D, E, V, YEQ, CON1 , 1 CON2, RHOGV, VBEPS, DX, NPTS C C CALCULATE TIME DERIVATIVES C F1 = .FALSE. F2 = .FALSE. DERV = 0.D0 DO 40 I = 1, NPTS IF (U(2,I) .LE. XC2) GO TO 10 APK = C * XC2 + D GO TO 30 10 IF (U(2,I) .LE. XC1) GO TO 20 APK = C * U ( 2 , I ) + D F2 = .TRUE. GO TO 30 20 APK = E * U(2,I) F1 = .TRUE. 30 CONTINUE 188 IF (T .EQ. O.DO) APK = O.DO IF (XM(I) .GT. VBEPS*T) APK = O.DO IF (I .NE. 0) DERV = (U(1,I) - U(1,I - 1)) / DX OT(1,I) • CON1 * (-RHOGV*DERV + APK*(YEQ - U(1,I))) UT(2,I) •= -CON2 * APK * (YEQ - U(1,D) 40 CONTINUE IX = NPTS RETURN END 189 L i s t i n g of SUM at 14:04:27 on DEC 18, 1985 for CCid=DUST Page 1 1 REAL*8 TIME,ULAST,U,SUM,SEED,BPACK,MFLOW,RFLOW,CMASS 2 SEED = 1.46D0 3 RFLOW = 1.600D0 4 BPACK = 0.3150D0 5 SUM = 0.018D0 6 ULAST = 0.00D0 7 MFLOW •= (BPACK/(SEED*0.63) )*RFLOW 8 1 READ(4,200,END = 2)TIME FORMAT(22X,F5.1) 9 200 10 READ(4,300)U 11 300 FORMAT(36(/),15X,E11.4) 12 UA = (U + ULAST)/2 13 SUM = SUM + (UA*MFLOW*((SEED*0.63)/BPACK)) 14 CMASS = TIME*RFLOW 15 WRITE(5,400)CMASS,SUM 16 400 FORMAT( 'TOTAL C02(G)',2X,F5.1,6X,'TOTAL OIL G',F10.5) 17 WRITE(6,500)CMASS,SUM 18 500 FORMAT(2X,F5.1,6X,F10.5) 19 ULAST = U 20 GOTO 1 21 2 STOP 22 END C THIS PROGRAM CREATES A 3D PLOT FROM SCATTERED DATA C POINTS. THE INPUT DATA MUST BE IN THE FROMAT OF X,Y,2. C THE INPUT DATA MUST BE IN THE FOLLOWING RANGE: X (0-1.0) C Y (0-45), Z (0-0.015) REAL X(1395),Y(1395),Z(1395),ZMAT(48,29) DIMENSION WSPC(8000),IWORK(4900) C C SET PLOTTING DEVICE C CALL DSPDEV('PLOT') CALL BGNPL(O) C C SET AXIS PARAMETERS AND ALPHABETS C CALL ZAXANG(90.0) CALL COMPLX CALL BASALF('L/CSTD') CALL MIXALF('STANDARD') CALL HEIGHT(0.30) C CALL HEIGHT SETS LETTERSIZE. THE AXES LABELS C WILL BE SET TO THE SIZE (IN INCHES) C THAT HAS BEEN SPECIFIED IN BRACKETS. C THE AXES NUMBERS ARE SET TO 5/7 OF C THIS VALUE. THE TITLE IS SET TO 1.5 TIMES C THIS VALUE. C C DEFINE 3D WORK AREA AND AXES C CALL PAGE(14.,11.) C C CALL PAGE ALLOWS THE USER TO SPECIFY THE OUTPUT SIZE (INCH) C IF "$" FOLLOWS THE TITLES THE PROGRAM AUTOMATICALLY C CALCULATES THE STRING LENGTH C C CALL TITL3D('(SOLVENT PHASE OIL)$',100,8., C *8. ) CALL TITL3D(*()$',100,8.,8) C C THE LAST VALUES HERE SPECIFY THE RATIO OF THE XAXIS C YAXIS AND ZAXIS RESPECTIVELY C CALL AXES3D('(DISTANCE)$',100,'(TIME) MIN$',100, 1'(OIL CONCENTRATION) Y$',100,1.0,1.0,1.0) C C DEFINE VIEWPOINT AND PROVIDE AXES VALUES C THE DISTANCE THAT THE VIEW IS TAKEN FROM IS IN USER C SUPPLIED UNITS eg IF THE AXIS IS 10 UNITS AND THE C REQUESTED VEIW IS 30, IT WILL BE THREE TIMES AXIS LENGTH C CALL VIEW(-29.0,1000.0,0.120) C C GRAF3D ALLOWS USER TO GIVE COORDINATES TO GRAPH C IN X,Y,Z, FIRST NUMB=ORIGIN, SECOND=LABEL STEPS-C AND THIRD=MAX VALUE CALL GRAF3D(0.0,0.2,1.00,0.0,9.0,45.0,0.0,0.003,0.015) C C READ DATA C DO 10 1=1,1395 READ(4,1) X ( I ) , Y ( I ) , Z(I) 1 FORMAT(2X,F5.3,5X,F5.2,4X,F7.5) 10 CONTINUE NP=1500 CALL BGNMAT(48,29) CALL GETMAT(X,Y,Z,1395,1WORK) CALL ENDMAT(ZMAT,IWORK) CALL SURMAT(ZMAT,2,48,1,29,WSPC) C C STOP C CALL ENDPL(O) CALL DONEPL STOP END C THIS PROGRAM CREATES A 3D PLOT FROM SCATTERED DATA C POINTS (X,Y,Z) C THE INPUT DATA MUST BE IN THE C FOLLOWING RANGE: X (0-1.0), Y (0-45), Z (0-0.7) C X IN THIS CASE REPRESENTS THE NORMALIZED DISTANCE FROM C THE BED ENTRANCE; Y THE TOTAL TIME OF THE EXTRACTION; AND C Z THE CONCENTRATION OF OIL IN THE BED REAL X(1395),Y(1395),Z(1395),ZMAT(48,29) DIMENSION WSPC(5000),IWORK(4000) C C SET PLOTTING DEVICE C CALL DSPDEV('PLOT') CALL BGNPL(O) C C SET AXIS PARAMETERS AND ALPHABETS C CALL ZAXANGO0.0) CALL COMPLX CALL BASALF('L/CSTD') CALL MIXALF('STANDARD') CALL HEIGHT(0.3) C C DEFINE 3D WORK AREA AND AXES C CALL PAGE(14.,11 . ) C C CALL PAGE ALLOWS THE USER TO SPECIFY THE OUTPUT SIZE (INCH) C IF "$" FOLLOWS THE TITLES THE PROGRAM AUTOMATICALLY C CALCULATES THE STRING LENGTH C C CALL TITL3D('(BED OIL PROFILE)$',100,8., C *8.) CALL TITL3D('( )$',100,8.,8) C C THE LAST VALUES HERE SPECIFY THE RATIO OF THE XAXIS C YAXIS AND ZAXIS RESPECTIVELY C CALL AXES3D('(DISTANCE)$',100,'(TIME) MIN$',100, 1'(OIL CONCENTRATION) X$',100,1.0,1.0,1.0) C C DEFINE VIEWPOINT AND PROVIDE AXES VALUES C THE DISTANCE THAT THE VIEW IS TAKEN FROM IS IN USER C SUPPLIED UNITS eg IF THE AXIS IS 10 UNITS AND THE C REQUESTED VEIW IS 30, IT WILL BE THREE TIMES AXIS LENGTH C CALL VIEW(-29.0,1000.0, 7.0) C C GRAF3D ALLOWS USER TO GIVE COORDINATES TO GRAPH C IN X,Y,Z, FIRST NUMB=ORIGIN, SECOND=LABEL STEPS C AND THIRD=MAX VALUE CALL GRAF3D(0.0,0.2,1.0,0.0, 9.0,45.0,0.0,0.1,0.7) C C READ DATA C DO 10 1=1,1395 READ(4,1) X ( I ) , Y ( I ) , Z(I) 1 FORMAT( 2X,F5.3,4X,F6.2,4X,F6.4) 10 CONTINUE NP=1450 CALL BGNMAT(48,29) CALL GETMAT(X,Y,Z,1395.IWORK) CALL ENDMAT(ZMAT,IWORK) CALL SURMAT(ZMAT,2,48,1,29,WSPC) C C STOP C CALL ENDPL(O) CALL DONEPL STOP END APPENDIX III  L i s t of Material Suppliers Avanti Polar-Lipids Inc. 2421 High Bluff Road, Birmingham, AL. 33216. (205) 822-9162 Chromalux Canada Ltd. 2455 Beta St., Burnaby, B.C. (604) 294-1801 Chromatographic S p e c i a l t i e s Ltd. 300 Laurier Blvd, B r o c k v i l l e , Ont. (613) 342-4678 Columbia Valve and F i t t i n g Ltd. 380 E. Esplanade, Vancouver, B.C. (604) 986-5251 CSP Foods Ltd. Box 580, Nipawin, Sask. (306) 862-4686 Fisher S c i e n t i f i c Inc. 196 W. 3rd Ave., Vancouver, B.C. (604) 291-8866 Fleck Bros. Ltd. 110 Alexander St., Vancouver, B.C. (604) 684-8131 Hewlett-Packard Inc. 10691 Shellbridge Way, Richmond, B.C. (604) 270-2277 Hyseco F l u i d Systems Ltd. 145 W. 2nd Ave., Vancouver, B.C. (604) 879-8851 Intek Elect r o n i c s Ltd. 10-8385 St. George St., Vancouver, B.C. (604) 324-6831 Intertechnology Ltd. Richmond, B.C. (604) Medigas P a c i f i c Ltd. 6841 Palm Ave., Burnaby, B.C. (604) 438-5276 1 94 Items phospholipid standards cartridge heaters 1/16" stainless s t e e l tubing valves, f i t t i n g s Canola seed and o i l samples E s t e r i f i c a t i o n chemicals r e f r i g e r a t i o n insulation tubing HPLC, supportive hardware Parker MV-200s valve RCA, Teledyne electronic components pressure transducer carbon dioxide 195 Nu-Chek Prep Ltd. P.O. Box 295, Elysian, MN. (507) 267-4689 56028 t r i g l y c e r i d e samples RAE Ind u s t r i a l Electronics Ltd. 3455 Gardner Crt., Burnaby, B.C. (604) 291-8866 Supelco, Inc. Supelco Park, Bellefonte, PA, (814) 359-3446 16823-0048 temperature transducers GC a n a l y t i c a l standards, GC columns Surrey F l u i d and Power Ltd. 203-13395-76th ave., Surrey, B.C. (604) 594-3461 Parker MV-200S regulating Valve 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0076857/manifest

Comment

Related Items