Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Equilibrium studies on pure compounds : normal propyl alcohol Croil, Thomas Arnold 1959

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

Item Metadata

Download

Media
831-UBC_1959_A7 C7 E7.pdf [ 2.65MB ]
Metadata
JSON: 831-1.0059064.json
JSON-LD: 831-1.0059064-ld.json
RDF/XML (Pretty): 831-1.0059064-rdf.xml
RDF/JSON: 831-1.0059064-rdf.json
Turtle: 831-1.0059064-turtle.txt
N-Triples: 831-1.0059064-rdf-ntriples.txt
Original Record: 831-1.0059064-source.json
Full Text
831-1.0059064-fulltext.txt
Citation
831-1.0059064.ris

Full Text

EQUILIBRIUM STUDIES ON PURE COMPOUNDS: NORMAL PROPYL ALCOHOL by THOMAS ARNOLD CROIL Diploma i n Chemical Engineering, Royal Military College, 1956 B.A.Sc, University of B r i t i s h Columbia, 1957. A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENT FOR THE DEGREE OF MASTER OF APPLIED SCIENCE i n the Department of CHEMICAL ENGINEERING We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA APRIL 1959. ABSTRACT Three grades of normal propyl alcohol have been purified by several methods. The degree of purity has been established by refractometric, ebulliometric, and viscometric measurements, the best product being 99.65 volume % pure as measured by gas chromatographic analysis. A vapor-liquid equilibrium apparatus has been modified and reconstructed i n preparation for vapor pressure measurements of n-propanol up to the c r i t i c a l point. Several semi-empirical equations have been f i t t e d to Young's vapor pressure data for n-propanol with a view to t h e i r u t i l i z a t i o n i n presenting data on an homologous series or on a generalized basis. In a l l cases the per cent difference between calculated and experimental values was less than 2.0$ with a maximum average difference of 0.9$. In presenting t h i s t h e s i s i n p a r t i a l f u l f i l m e n t of the requirements f o r an advanced degree at the U n i v e r s i t y of B r i t i s h Columbia, I agree t h a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r reference and study. I f u r t h e r agree th a t permission f o r extensive copying of t h i s t h e s i s f o r s c h o l a r l y purposes may be granted by the Head of my Department or by h i s r e p r e s e n t a t i v e s . I t i s understood that copying or p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l gain s h a l l not be allowed without my w r i t t e n permission.. Department of Chemical Engineering The U n i v e r s i t y of B r i t i s h Columbia, Vancouver Canada. Date F e b r u a r y ' 1 9 5 9 -i TABLE OF CONTENTS Abstract List of Tables i i List of Figures i i Nomenclature i i INTRODUCTION 1. LITERATURE REVIEW 1. Purification 2. 2. Apparatus 4> 3. Presentation of Vapor Pressure Data 6. PURIFICATION 1. Materials 8. 2. Apparatus 10. 3. Procedure 12. 4. Purity Determinations a) Methods 14. b) Results 17. 5. Discussion a) Apparatus and Methods 24. b) Purification 26. VAPOR PRESSURE APPARATUS 30. VAPOR PRESSURE CORRELATIONS 1. Methods 33. 2. Discussion 36. LITERATURE CITED 38. i i LIST OF TABLES Page 1. Physical Properties of Normal Propanol 19 2. I n i t i a l Investigations 20 3 Physical Data of Normal Propyl Alcohol from the Literature 22 a,b. 4. Constants for Vapor Pressure Correlations 34 5. Correlation Deviations 35 NOMENCLATURE 1. a, b, c,o4, j j constants from the Biot Formula. 2. A, B, C, D = constants for the Riedel, Frost and Antoine Equations T =• temperature, °R. (unless otherwise noted) Tg = reduced temperature (T/T^) P = vapor pressure, p.s.i.a. (unless otherwise noted) P ^ reduced pressure ( P / P Q ) JJ = viscosity, centipoise. Following LIST OF FIGURES I Chromatograms of Normal Propanol from a Flexol Plasticizer Column 20 I I Effect of Suspected Impurities on Chromatogram Peaks 27 I I I Schematic Representation of the Vapor Pressure Apparatus 30 IV Correlation Deviations 35 ACKNOWLEDGEMENT The author wishes to express his appreciation for the financial assistance of the Dr. F. J. Nicholson Scholarship Fund and the National Research Council during the course of this research. 1. INTRODUCTION Chemical engineering design requires an adequate knowledge of the physical properties of materials. While accurate experimental data i s desirable, i t s measurement i s often d i f f i c u l t and time consuming. For this reason chemical engineering literature has recently emphasised methods of correlating and predicting such data, especially i n the f i e l d of thermodynamic properties ( l ) . Because the thermodynamic properties of the normal alcohols, and their solutions, have been the subject of considerable interest i n these laboratories (2,3), this investigation of the vapor pressure of normal propanol was undertaken. Although Speers (4) has examined the thermodynamic network for n-propanol, his calculations were based primarily on the long Standing data of Young (5). While these data are considered most reliable by a l l c r i t i c a l compilers, no measurements above the normal boiling point have been made since that time ' ( f i f t y years ago). In view of the increasing importance of highly accurate c r i t i c a l point measurements, i t was considered worthwhile to re-explore the vapor-liquid relationship of pure normal propanol up to the c r i t i c a l point. This consideration was partly influenced by the availability of a high pressure apparatus designed for multi-component vapor-liquid equilibrium, but equally suitable for such a one component system. The selection of normal propyl alcohol as the single component was also influenced by the fact that benzene-n-propanol had been selected as the f i r s t binary system to be studied i n these laboratories (through a range of pressure). This made the purification, and hence, the physical properties of propanol of considerable interest. The need for such a study became even more apparent i n light of the meagre data available i n the literature on purification techniques. Moreover i t was essential to ascertain to what extent impurities would affect the vapor pressure of normal propanol before confronting the more complex problem of their effect on a binary system. This consideration i s particularly important near the c r i t i c a l point (5,6). Vapor-liquid equilibrium data for one-component systems (vapor pressure data) have traditionally been correlated by theoretical, semi-empirical and 2. empirical equations i n order to enhance their usefulness. While data on individual compounds i s often considered separately, there i s a growing tendency to consider families of compounds (7) and of course to determine a completely generalized approach ( l , 8, 9). Re-evaluation of vapor pressure data on n-propanol can therefore be usefully considered i n a l l of these categories. This thesis, then, presents information on obtaining and measuring the purity of highly purified n-propanol, on the apparatus for making vapor pressure measurements up to the c r i t i c a l point, and on correlations for existing data on n-propanol that can be used for evaluation of new data that may be obtained. LITERATURE REVIEW 1. Purification of Normal Propanol According to Timmermans (10 ) and Speers (4) the only worker who has measured the vapor pressure of propanol above one atmosphere i s Young (5). His f i r s t concern was with the purity of the liquid to be investigated. "The physical properties of a substance, especially at or near i t s c r i t i c a l point, may be seriously affected by the presence of even a very small quantity of impurity; i t i s therefore of the utmost importance that the purification of the substances investigated should be carried out with the greatest possible care," (5) Almost a l l the purification methods for n-propanol involve some form of fractional d i s t i l l a t i o n of a commercial grade. There i s , however, one method, as described by Timmermans and Delcourt ( l l ) , which i s an exception. They stated that there are traces of isomers and homologues which cannot be removed by simple fractional d i s t i l l a t i o n alone. In this case the purification was made by fractional crystallization of a solid propyl ester, l i k e the acid phthalate. The ester was either reduced or hydrolysed back to the alcohol and the usual d i s t i l l a t i o n carried out. However the purity of a Kahlbaum sample (as determined by density and the c r i t i c a l temperature of solution i n petroleum (12)) was found to be the same before and after transformation into the phthalate. Young's n-propyl alcohol (13, 14) was procured from Kahlbaum. It 3. was purified by fractional d i s t i l l a t i o n and then d i s t i l l e d with benzene through a very efficient s t i l l head to remove the last traces of water. The specific gravity of this specimen at 0°/4° was 0.&L923 and the boiling point 97.20°C. at one atmosphere. This requirement of high purity propanol for physical property measurements stimulated further experimentation with purification techniques i n more recent years. de Brouckers and Prigogine (15) purified technical grade propanol by refluxing over lime for five hours and then d i s t i l l i n g through a one meter column. Kretschmer (16) found that a commercial grade of n-propanol contained 1.5$ a l l y l alcohol as i t s main impurity. He carefully purified the alcohol by treating one l i t r e with 15 ml. of bromine. The alcohol was fractionally d i s t i l l e d with a small amount of potassium carbonate through a 75-plate column. The middle fraction of 600 ml. was dried with 1 gram of magnesium ribbon, freshly cleaned with steel wool, i n a storage flask attached to a vacuum system. Before the flask was sealed, 1 gram of 2,4-dinitrophenyl hydrazine was added to react with any propionaldehyde formed by the bromine treatment that had not been removed by d i s t i l l a t i o n . Both Keyes and Winninghoff (17) and Kraus and Bishop (18) dried propyl alcohol with metallic sodium and fractionally d i s t i l l e d . Goldschmidt and Thomas (19) dried 1-propanol with aluminum amalgam and, to remove basic impurities, d i s t i l l e d over s u l f a n i l i c or tartaric acid. Berner (20) boiled n-propanol with lime for six hours and after d i s t i l l i n g warmed the middle fraction with calcium hydride i n a stream of hydrogen. Other workers who purified propanol include Lund and Bjerrum (21) and Brunei, Crenshaw and Tobin (22). The main criterion for purity i s considered to be the constancy of vapour pressure (5, 6) when the l i q u i d i s evaporated, or the vapor condensed, since impurities usually divide themselves unequally between l i q u i d and vapor. Wullner and Grotrian (23) found appreciable differences i n pressure i n the interval between the condensation of the f i r s t drop of l i q u i d from the vapours of several organic liquids and the disappearance of the last bubble 4. of vapor, and these were shown by Tammann (24) to be due to impurities. Tammann found that 0,0001 part of benzene i n water was enough to cause an inconsistency i n pressure during evaporation or liquifaction, the vapor pressure depending on the volume of the vapor phase. Constancy of vapor pressure i s , therefore, an extremely sensitive test of purity, far exceeding boiling point i n delicacy (6). A test of purity i s to evaporate the l i q u i d by pumping off vapour u n t i l only one twentieth the volume of l i q u i d remains, when the vapor pressure should be unchanged (25). Another test i s the constancy of temperature during freezing (26). Young (5) embraced these principles of purity and added others. Amongst these he required close agreement between physical constants of two different specimens of the same liq u i d . . While Weissberger (27) makes no mention of any specific c r i t e r i a , Timmermans (10) states the c r i t e r i a of purity to be the density to the f i f t h place (28) and the c r i t i c a l temperature of solution i n petroleum (12). 2. Apparatus The high pressure apparatus available was designed by Whittle (29) after apparatus described by Sage and Lacey (30). In general i t can be classed as a static or bomb equipment. The work by Young, longstanding and s t i l l highly regarded for i t s accuracy, was carried out by a static method also, and hence should be considered i n more detail for comparison to the method and apparatus intended here. Young*s apparatus consisted basically of a long wrought iron tube having one end f i t t e d with a screw plunger and the other end sealed. This tube, firmly secured i n the horizontal position, had three shorter tubes running i n a vertical direction from i t . Three thickwalled glass tubes, graduated i n millimeters and carefully calibrated, had one end pressure f i t t e d into the iron tubes and the other end sealed. The f i r s t two served as a i r manometers for different pressure ranges, and the third was the experimental tube containing the liquid under investigation. The iron apparatus was f i l l e d with mercury and pressures applied by means of the plunger. The temperature around the experimental tube was controlled by passing vapors from various boiling liquids through a jacket around the tube. 5. When the vapor in the jacketing tube was at the required temperature, readings of vapor pressure were taken with the l i q u i d and vapor i n the experimental tube occupying a series of different volumes. Young corrected his calculated pressure, l ) for the difference i n level of the mercury i n the experimental tube and the manometer, 2) for the pressure of the column of unvaporized l i q u i d , 3) for the expansion of the heated column of mercury, 4) i f necessary, for c a p i l l i a r i t y , 5) for the deviation of air i n the manometer from Boyles Law. He made no correction for the vapor pressure of mercury because he was of the opinion that evaporation through a long column of liq u i d was an exceedingly slow process. The assumptions made by Young i n this statement have been seriously studied i n the past few years. Jepson and Rowlinson (31) have shown that a correction for the v o l a t i l i t y of mercury should be applied to observed pressures of compressed gases where the confining f l u i d i s mercury. The usual correction, when applied, was simply the substraction of the normal vapor pressure of mercury corrected for the hydrostatic pressure (the Poynting effect (32)). They showed that this i s not an adequate treatment of the problem, as the mixture of mercury atoms and compressed gas cannot behave as an ideal mixture. An estimation of the intermolecular forces between mercury atoms and the added gas leads to values of the v i r i a l coefficients from which a revised correction can be computed, assuming that the system i s at equilibrium. This revised correction can be considerably larger than the usual correction and i s often of opposite sign. While i t i s true that i n most vapor pressure measurements li q u i d i s present over the mercury surface, i t s density rapidly decreases as the c r i t i c a l point i s approached. Young's theory may be quite valid for long columns of liquid over mercury but i t seems that the height and density of the liq u i d would be quite important. Jepson and Rowlinson's correction would be particularly applicable near the c r i t i c a l point, i n Young's type of measurement. Kay (33) and later Bahlke and Kay (34) improved Young's apparatus and also carefully considered the corrections required. A similar method has also been employed previously i n these laboratories for measurement of n-butanol (35). The bomb apparatus of Whittle, described below, requires similar 6. corrections to those above. The correction having regard to mercury requires special notice since the mercury surface i s many times larger than i n the case of Young's apparatus. 3. Presentation of Vapor Pressure Data Numerous equations, both empirical and theoretical, have been given relating vapor pressures, P,with absolute temperature, T (6, 36, 37, 38). Young used the Blot Formula (39) to correlate his data on n-propanol. log P = a + b o ^ + eft* (1) where a = 4.479470 log<X,= 0.001641423 log b = T. 3915059 log = T. 99657025 log c - 0.5509601 t = T°C - 20. The constants were calculated from pressures at 20, 80, 140, 200 and 260°C. (13). The agreement between calculated and experimental data was good, but the nature of the equation made i t d i f f i c u l t to use. Other investigators (40) also found the formula inaccurate and i t has subsequently fallen into dis-use. Reid and Sherwood (38) have recently recommended the Eiedel correlation (41) for most accurate work. log P R = A -J3_ + C In T R + DT^ (2) TR (Actually this i s the relation that Riedel used as a basis for his single constant reduced vapor pressure equation). It has i t s disadvantage,; however, i n that the c r i t i c a l temperature and pressure must be known. In many cases the accuracy of these constants cannot be too heavily relied upon. Thomson (36), i n his well known review of 1946, recommends the use of two Antoine equations, 7. log P = A - B (3) T - C where P = vapor pressure, p.s.i.a. T = temperature, °R. C = constant, °R. A,B = constants. one up to T R = 0.8 or 0.85 and the other from T R = 0.8 to T R = 1.0, for most accurate results. The disadvantage here, of course, i s the necessity for two equations when only one i s desired. Among the more recent equations i s that proposed by Frost and Kalkwarf (42), log P = A + B + C log T + DP (4) T "T? i n which they try to explain the reverse curvature of the plot of log P versus 1 on the basis of the non-ideal behavior of the vapor together T with the change i n heat of vaporization with temperature. This equation has been successfully used by Thodos (43) to consider the vapor pressures of a series of the normal paraffin hydro-carbons. Its usefulness here indicates the strong possibility of i t playing a similar role for the n-alcohols. The Antoine equation has also been most successfully used for families of compounds, notably by Dreisbach (7). From the large number of equations available, these three have therefore been selected to be used with the vapor pressure of n-propanol. Young's data are employed, and any later re-evaluation or re-determination could always be compared to these i n the same fashion. The equations selected (Riedel, Antoine, and Frost) a l l have a semi-empirical basis, offer relative simplicity i n calculation, and provide the possibility of interesting comparisons with other members of the normal alcohol series. 8. PURIFICATION 1. Materials I Fisher Certified Grade n-Propanol This material was supplied with the following stated specifications: Acidity (CH^ COOH) 0.002$ Boiling Range 96 - 97.5°C Non-Volatile Matter 0.000$ Substances precipitated by 1^ 0 None This n-propanol i s one of the co-products produced from carbon monoxide and hydrogen i n the high pressure catalytic synthesis of methanol (44). The chief method of separations of these co-products and f i n a l purification of the n-propyl alcohol i s careful fractionation. The higher alcohol mixture produced by this synthesis has been found to contain the following primary alcohols (45) : n-propanol (b.p. 97.19°C.) isobutanol (b.p. 108.39°C.) 2 met hyl-l-but anol (b.p. 128°C.) 2 methyl-l-pentanol (b.p. 148°C.) 2,4 dimethyl-l-pentanol 4 methyl-l-hexanol iso-propanol (b.p, 82.3°C.) 3 methyl 2-butanol (b.p. 114°C) 2,4 dimethyl 3-pentanol (b.p. 140°C.) 2,4 dimethyl 1-hexanol 4 or 5 methyl-l-heptanol 3 pentanol (b.p. 115.6°C.) 2 pentanol (b.p. 119»28°C.) 9. II Canadian Chemical Company Technical Grade  Normal Propyl Alcohol This material i s stated to have the following specifications: Specific gravity at 20/20°C max 0.8074 D i s t i l l a t i o n range °C max 2 Colour APHA max 5 Acidity as Acetic % by wt. max. .003 Alkalinity as NE^ by wt. max. 0.2 Water content % by wt. max. 0.2 Non-volatile materials gms/l00 ml 0.001 Mass spectrometer analysis of the product stream gives (46): High Low Avg. 2-Butanol 4.4# 2.8£ 3.5# 2-Propanol traces traces n-Propanol 97.1 95,5 96 Methoxy-Methylal 0.1 0.0 0.05 for August 18-25, 1958. This propanol i s produced as a by-product of propane oxidation. I I I Eastman n-Propyl Acetate The highest grade Eastman n-propyl acetate obtainable commercially was used. The only manufacturer's specification i s the boiling point: 97 - 102°C. IV Auxiliary Materials The following materials were used as available commercially: Reagent Grade Bromine, Reagent Grade Anhydrous Potassium Carbonate, Magnesium ribbon (freshly cleaned with steel wool), Cylinder nitrogen (purified grade). Reagent Grade Sodium Hydroxide. 10 2. Apparatus a. D i s t i l l a t i o n Apparatus A l l d i s t i l l a t i o n s were done on a Todd Precise Fractionation Assembly, employing a 25 mm. I.D. column packed with single turn case hardened Pyrex brand glass helices 4 mm. i n diameter. The length of the column gave a fractionation efficiency of up to 60 theoretical plates. The apparatus was equipped with jackets and a dual heating unit designed to enable the fractionation column to be operated under adiabatic conditions up to 360°C. An automatic s t i l l head timer controlled the reflux ratio from 2: 1 to 50: 1 i n five integral steps by means of a solenoid operated valve made of teflon and containing a soft iron core. b. Refractometer A l l refractive index readings were made on a Pulfrich refractometer using the light prism with a sodium lamp to provide D - line readings. A constant temperature apparatus maintained the prism at 20°C - 0.1°C. The refractometer , was read to the nearest 0,5 minutes. The normally immersed thermometer well was equipped with a rubber gasket which sealed the top of the cylindrical sample container when the well was lowered into i t . c. Gas Chromatograph A Beckmann GC-2 gas chromatograph was used consisting of the following elements: a chromatographic column, a carrier gas flow control, a heated sample inlet system, a thermal conductivity c e l l , an electronically controlled heater system, and an electronically regulated voltage supply. The columns were f l a t spirals of £ inch copper or stainless steel tubing interchangeable with ones of different packing and different length. Cylinder helium was used as the carrier gas. The conductivity c e l l was an elec t r i c a l l y balanced filament type giving a reproducibility of - 0.1$ of f u l l scale deflection (47). The instrument had a temperature range from 40° to 220°C, maintained by an internal full-proportional heater, 11. electronically controlled. Zero s t a b i l i t y was 2% per hour or better, under normal operating conditions (47). The l i q u i d samples were injected into the chromatograph from a Beckmann 22400 liqu i d sampler. A syringe type instrument, i t was designed for introduction of precisely measured small quantities of liquids (.005 cc to .05 cc). Uniform results were obtained with reproducibility of 0.1$ to 0.5# (47). d. Ebulliometer Boiling points were determined by the comparative method as described by Swietoslawski (48). Two differential ebulliometers constructed according to the standard specifications of Barr and Anhorn (49) were used. Ebulliometer A (50) contained the primary standard, and B (51) the sample to be studied. Both pieces of apparatus consisted basically of a boiler with a thermometer well and drop counter, a condensation temperature element with a thermometer well and drop counter, and a condenser. In addition, B had a rectifying element between the two thermometer wells and was equipped with a s i l i c a gel drying tube above the condenser. Both A and B were well insulated with asbestos. The thermometer wells were bui l t up with cork and insulation so that the thermometer was immersed to the same level as i t was during calibration. These wells were f i l l e d with mercury and covered with a light o i l so that the l i q u i d level would rise to the top of the well when the thermometer was immersed. The boiler sections were wrapped with nichrome heating wire and the heat input controlled with a variable auto transformer. In the case of A, the boiler tube was packed with pyrex glass wool to give undisturbed boiling during operation with the primary standard. The Beckmann thermometer used i n the ebulliometers had 100 divisions per degree. It was calibrated i n a constant temperature o i l bath against a Leeds and Northrup platinum resistance thermometer with a 1955 NBS certificate (52). D i s t i l l e d water having a specific conductivity greater than 800,000 ohms ^ cms ^ was used as the primary standard. 12. e. Viscometer A routine Cannon-Fenske viscometer of the type recommended by the ASTM (54) for testing petroleum products, and as described by Cannon and Fenske (53), was used. 3. Procedure 1 Fisher Certified Grade Normal Propyl Alcohol The n-propanol was charged to the s t i l l pot i n one l i t r e lots with several grams of anhydrous potassium carbonate and a few boiling chips. The apparatus was then purged for several minutes with purified grade cylinder nitrogen to remove any ai r atmosphere. The system was closed and the vent on the d i s t i l l a t e collecting vessel connected to a glass tube which dipped into a flask of propanol. The system was then opened at the vent stopcock allowing N^ i n excess of atmospheric pressure, to bubble out through the propanol. The heater under the s t i l l pot was turned on and the charge brought to boiling. When refluxing was observed from the packing at the bottom of the column the heater was cut back and the two column heaters switched on. When d i s t i l l a t e began condensing i n the top condenser the temperature i n the upper part of the column was adjusted to the column top temperature. Similarly the lower part was adjusted to the s t i l l pot temperature. The column packing was inspected for signs of local heating and then the apparatus was l e f t to come to equilibrium. When this point was reached, the reflux timer was set at 50* 1 and d i s t i l l a t e was collected. The f i r s t 500 mis. were drawn off and then a centre cut of 150 mis. was collected over freshly cleaned magnesium ribbon i n a nitrogen purged flask. The flask was then sealed with a ground glass stopper and stored. Additional runs were made, under exactly the same conditions, with 15 ml. of bromine per l i t r e of propanol i n the s t i l l pot, part of the purification procedure described by Kretschmer (16). I I Canadian Chemical Company Technical Grade Normal Propyl Alcohol 2 l i t r e s of n-propanol were d i s t i l l e d by similar procedure at 13. 5 0 : 1.reflux ratio. The f i r s t 5 0 mis. were discarded and last 1 0 0 mis. l e f t i n the boiler. This d i s t i l l a t i o n procedure was repeated twice more and the f i n a l cut stored over Mg ribbon i n a nitrogen atmosphere. I l l Hydrolysis of Normal Propyl Acetate to Normal Propyl Alcohol The production of high purity n-propyl alcohol by hydrolysis of the acetate, as i n the following reactions, was suggested (46): NaOH CH3 C 0 0 C H 2 CH2 CH^ + H 20 = CH^ CH2 CH2 OH + CE^ COOH CH3 COOH + NaOH = CH3 COONa + H 20 1 5 0 0 grams of Eastman white-label n-propyl acetate was fractionated on the Todd s t i l l at 1 5 : 1 reflux ratio, and a heart cut of the 101.6°C. fraction taken. About 1 l i t r e of the rectified propyl acetate was placed i n a 2 - l i t r e flask with 2 0 0 ml. of water and 5 0 grams of sodium hydroxide pellets, and the flask was closed with a rubber stopper. The contents were given vigorous and prolonged shaking by hand, with frequent addition of more sodium.1 hydroxide i n 1 0 to 2 0 gm. portions, u n t i l a sudden generation of heat indicated the commencement of hydrolysis. The stopper was then loosened and agitation was reduced to a gentle swirling, letting the flask cool undisturbed whenever i t became uncomfortably hot to the hand. At this stage the ester odour had been replaced by the alcohol odour. Periods of gentle swirling and cooling were alternated u n t i l further agitation produced no further heat. The mixture was transferred to a 2 - l i t r e s t i l l flask, 2 0 0 mis. of water added, and the n-propanol-water azeotrope was removed by fractionation, collecting the 8 7 » 7 ° C . cut at 1 0 : 1 reflux ratio. The azeotropic cut was treated with an excess of anhydrous potassium carbonate to salt out the alcohol as an upper layer. The upper layer was removed in a separator funnel and dried by shaking with successive portions of anhydrous potassium carbonate. The f i n a l settling period was extended overnight. The dried alcohol was then fi l t e r e d into a 1 - l i t r e flask, 14. and last traces of water removed by fractionating at 15: 1 reflux ratio , collecting a heart cut, when the temperature had steadied near 97° (uncorrected s t i l l head temperature), as pure n-propanol. 4. Purity Determinations a. Methods 1. Refractive Index The refractometer was prepared for use by ensuring that the sample container was clean and dry and that the zero reading was correct. The sample to be analysed was sealed i n a serum bottle with a rubber serum bottle stopper, previouslj'- boiled i n propanol. The sealed bottle was placed on a tray i n the refractometer constant temperature bath and l e f t for about twenty minutes. At the end of this period the bottle was removed from the bath and about 5 mis. of sample withdrawn into a hypodermic syringe. After the air was ejected from the syringe, the sample was run into the refractometer sample container and the thermometer well, (with rubber gasket) lowered into place as quickly as possible. The angle of refraction was read immediately and then at successive time intervals u n t i l the reading was constant within 0.5 minutes. This reading was recorded and then the procedure was repeated with new samples from the serum bottle u n t i l the readings were consistent within 0.5 minutes. 2. Chromatographic Analysis Preliminary investigations were required for determining the optimum values for the variables connected with the chromatograph operation. In the case of flow rate, current and column length, the manufacturer's specified values were used (47). However sample size, column temperature and column composition were determined by a series of investigations. The procedure involved finding a combination of these three variables which would give the best resolution, and hence, the clearest qualitative and quantitative indications of the sample composition. 15. Five recommended columns (46, 47, 54a) of £ inch copper tubing, each six feet long, were packed with 30 - 40 mesh brick dust. The dust was made by grinding up C - 22 Sil-o-cel brick, removing the 30 - 40 mesh cut, washing out the fines with water, and then drying i n an oven. The brick for each column was treated with a different partitioning l i q u i d i n the following way. Six to eight mis. of partitioning l i q u i d were made up to 40 mis. i n a 100 ml. graduate with a dissolving solvent. Complete mixing was ensured by inserting a teflon piston into the graduate and pulling i t back and forth. The piston was l e f t at the bottom of the graduate and 50 mis. of brickdust pored slowly into the li q u i d mixture so that each particle f e l l independently through the f l u i d . After le t t i n g the particles settle for a few minutes, the coated brick dust was removed by extracting the piston. The dust was spread out on a tray to be air dried and subsequently packed into the copper tubing. The five partitioning liquids used were: 1. Tricresol phosphate (reagent grade) 2. Flexol plasticizer - 8N8 (Carbide and Carbon Chemical Company) 3. Vacuum pump o i l (Hyvac-Central Scientific Company) 4. Glycerine (reagent grade) 5. Polyethylene glycol di-2-ethylhexoate (Carbide and Carbon Chemical Company) The optimum sample size and column temperature were determined with the best column of those lis t e d above, following the procedure as outlined below. The warm up period for the chromatograph was normally two hours. In this period the column was purged with helium at the operating pressure and was brought up to the operating temperature. In order to check for zero d r i f t and hence an indication of insufficient warm up period, the attentuator was set at i t s lowest value and the zero adjusted to the 50 m i l l i v o l t position on the chart. I f there was no perceptible shift i n the zero position over a period of ten minutes, the apparatus was considered ready for use. The sample to be injected into the column was drawn into and 16. rejected from the sampling syringe u n t i l i t was obtained i n an a i r free state. At the appropriate moment the chromatogram chart drive was switched on and the sample injected as quickly as possible into the column. During the run the lowest attenuation was maintained to allow for maximum detection of the components. When no more peaks appeared after running at 0 millivolts for several minutes, the chart drive was switched off. However, i n order to detect any possible additional components, the instruments were l e f t running, with the recorder pen on the chart, for ten to twenty minutes after the chart drive had been stopped. Identification of the unknown sample was traced again by a series of investigations. Suspected components were obtained i n a f a i r l y pure state and run separately on the chromatograph. When peaks appeared at the same position on the two chromatograms, i t was generally accepted as positive identification. However, i n cases where more than one suspected component coincided at the same position, known volumes of the suspected components were added to the unknown sample and the effect on the peak i n question observed. The peak height measurement technique was employed to obtain a quantitative analysis of the chromatograms. Only one calibration run (at 100$) was traced for each component, assuming a linear relationship between peak height and composition. The volume per cent of the components was determined as follows: Volume % = component peak height x iOO component calibration peak height In cases where the purest form of the component i n question showed impurity peaks, the calibration peak height was substituted for what was considered a more accurately determined value (see Discussion, b) Results). 17. 3. Boiling Point Ebulliometer A (50) was f i l l e d with d i s t i l l e d water and the heat input set so that there was rapid boiling. The boiling rate was adjusted to give 5 - 10 drops/minute at the two drop counters. The temperature was measured i n both thermowells to get an additional check on the purity of the primary standard. Ebulliometer B (51) was f i l l e d with a sample of the n-propanol being studied and brought to a steady b o i l giving about 100 drops/minute at both drop counters. The temperature at each thermowell was measured to ascertain the purity (55). A boiling point determination was begun by obtaining a steady-temperature (within ,002°C.) i n the lower thermowell of A. The thermometer was then quickly transferred to the lower thermowell of B. This procedure was repeated u n t i l the temperatures were constant within .002°C. i n both thermowells. The temperature obtained i n A was used in conjunction with the pressure-temperature relationships and the interpolation formulae for water, as recommended by the International Union of Chemistry, to obtain the atmospheric pressure. Assuming a value of dT/dB for n-propanol at 760 mm... as 0.038 C/ mm. (10), the actual boiling point of the sample i n B at 760 mm. was calculated. The pressure range over which this correction had to be applied was a maximum of 10 mm. 4. Viscosity The procedure for measuring the viscosity of n-propanol was described by de Verteuil (56) and his results are presented below. b. Results For the various starting materials, several "grades" of n-propanol were prepared, and determination of boiling point, refractive index, and viscosity, made as detailed above. The results are tabulated i n Table 1 along with estimates of purity and water content. In addition, the results obtained from the i n i t i a l investigations 18. into suitable methods of purification of the available "grades" of n-propanol are presented i n Table 2. Two chromatograms, typical of those used to calculate the percent impurities of Table 1, are illustrated i n Figure I. For comparison purposes, the physical properties of n-propanol from the literature are tabulated i n Tables 3a and 3b. 19. TABLE 1. Physical Properties of Normal Propanol I I I I I bp. °C. a ' 96.0-97.53 96.1-98.13 at 760 mm. b • J L — 7 0 . X I 97.16 97.17* 20 a 1.38546 1.3853 °D b 1.33539 1.38524 1.38524 % impurity 1' 2 a 3.5 3.5 (by volume) b 1.2 0.7 0.35 % Water1 a 0.1 0.1 0.1 (by volume) b 0 0 0 X (cp.) b 1 5 C 2 . 4 8 6 25°C 1 . 9 4 6 30°C 1 . 7 1 8 (See Page 21 for explanatory notes) 20. TABLE 2. I n i t i a l Investigations 1,2 Di s t i l l a t i o n Conditions n ^ (j'j^voiTume) Notes Reflux Ratio before 1.38546 3.5 d i s t i l l a t i o n bromine added 25:1 1.38543 1.7 25:1 1.38543 50:1 1.38539 1.2 before - 3.5 d i s t i l l a t i o n JJ 1st distn. 50:1 - 1.2 2nd distn. 50:1 - 0.7 3rd distn. 50:1 - 0.55 I I I 1st distn. 50:1 - 0.12 5 (See Page 21 for explanatory notes). Following page 20 TIME 2. Purified Technical Grade n-Propanol Figure I CHROMATOGRAMS OF NORMAL PROPANOL FROM A FIEXOL PLASTICIZER COLUMN. 21 Notes for Tables 1 and 2. I Fisher Certified Normal Propyl Alcohol I I Canadian Chemical Company Technical Grade Normal Propyl Alcohol I I I Normal Propyl Alcohol produced by hydrolysis of Normal Propyl Acetate a - before purification b - after purification 1. These volume percentages were calculated from the chromatograms obtained under the following conditions: 2. The percentages here were calculated using 3«5 volume per cent sec-butanol impurity i n n-propanol I l a as the standard. This i s i n accordance with the mass spectrometer analysis from Canadian Chemical Company (see Materials). 3. Manufacturer's specifications. 4» This reading was obtained using a Cenco U.S. Weather Bureau type standard mercurial barometer for the atmospheric pressure determination. The procedure for determining the boiling point was exactly as described above except that the barometer was substituted for ebulliometer A. Pressure readings were made to the nearest 0.1 mtn. Flow rate: 150 cc/min. Sample size: 0.005 cc. Current: 150 ma. Temperature: 70°C. Column Partitioning Liquid: Flexol Plasticizer-8N8 Column Length: 6 feet. 21a. and the brass scale correction applied at room temperature. Reduction of the barometer to latitude 45° was neglected. Since the manufacturer considers this barometer "of the highest type of excellence", i t was assumed accurate to 0.1 mm. Even an error of 0.1 mm., which corresponds to less than 0.005°C, would not affect the boiling point any more than the errors involved i n the comparative method. For this reason the accuracy of the two methods i s considered comparable. 5. This value was obtained using a 5 mm. I.D. column on the Todd Fractionation Assembly, rather than the 25 mm. I.D. column which was used for the f i n a l products, as described on p.19. 22. TABLE 3a. Physical Data for Normal Propyl Alcohol from the Literature. AUTHOR DATE b.p.°C. (760 mm) Young and Fortey 1903 97.19 Dorochewsky 1909 97.20 Dorochew3ky 1911 97.26 Mundel 1913 97.1 Brunei, Crenshaw and Tobin 1921 97.19 Brunei 1923 97.15 Grimm and Patrick 1923 97.19 Trew and Watkins 1933 Timmermanns and Delcourt 1934 97.15 Wojciechowski 1936 97.209 Zepalova-Mikhailova 1937 97.15 Addison 1945 98.0 Vogel 1948 Carley and Bertelsen 1949 97.19 Mumford and P h i l l i p 1950 97.2 Howey 1951 97.2 McKenna, Tartar and Lingafelter 1953 Wetzel, Mil l e r and Day 1953 Purnell and Bowden 1954 97.2 Croil I 1958 I I 97.16 REFRACTIVE INDEX REF. 1.3856 1.38556 1.3862 1.3858 m 1.38539 1.38524 1.38524 (57) (58) (59) (60) 1.3833 (61) 1.3833 (28) (62) 1.38343 (63) (64) (65) (66) (67) (68) (69) (70) 1.3838 (71) 1.3837 (72) 1.3841 (73) 1.3840 (74) This Research 23. TABLE 3b. Physical Data for Normal Propyl Alcohol from the Literature. AUTHOR DATE TEMP. °C. DENSITY VISCOSITY REF. g./ml. cp. Gartenmeister 1890 10 20 30 40 50 .8125 .8052 .7973 .7890 .7802 Thorpe and Rodger 1894 15.06 Thole 1910 25 Baker 1912 25 .8010 Dunstan & Thole 1913 25 English and Turner 1914 25 .7999 Herz 1918 25 Whitman 1930 25 .7957 Timmermanns and Delcourt 1934 15 25 30 .80749 .79957 .79567 Jones 1948 25 .8015 Mumford & P h i l l i p 1950 20 25 .8053 .8016 de Verteuil .8075^ .7998* .7957 and C r o i l 1958 15 n-propanol I I 25 30 2.934 2.273 1.791 1.416 1.148 2.555 1.990 1.971 1,962 1.928 1.915 1.962 (75) (76) (77) (78) (79) (80) (81) (82) 2.522 (64) 1.924(calc) 1.722 2.004 (83) 2.29 (70) 2.015 2.486 1.946 1.718 This research * Assumed Values. 2 4 . 5 . Discussion a. Apparatus and Methods 1. Refractive Index While the refractometer could only be read to the nearest 0 . 5 minutes, successive readings were sufficiently reproducible to give an average deviation from the mean of a group of readings of less than 0 . 2 5 minutes. This corresponds to a maximum error of about 4 i n the f i f t h decimal place of the refractive index value. The rubber gasket, sealing the sample container when the thermo-well was lowered into i t , reduced the exposure of the sample to room ai r to a very small period of time. Thus any moisture pickup by the materials used was insufficient to cause detectable changes i n refractive index. Similar precautions were taken with the l i q u i d before analysis by storing i t i n a serum bottle and retracting samples, as required, with a hypodermic needle. 2 . Chromatographic Analysis The gas chromatographic analysis provided a most useful means of detecting impurities present. The selection of the best column and operating conditions was d i f f i c u l t , however. Of the six columns tested (silicon, t r i c r e s o l phosphate, flexol plasticizer, vacuum pump o i l , glycerine and polyethylene glycol 2-diethyl hexoate) the flex o l plasticizer column alone gave clear distinct peaks for water, n-propanol and an unknown impurity. In the other columns the impurity was either hidden by the propanol peak or the peak was so f l a t i t was d i f f i c u l t to analyze. As a result this column was selected and used for a l l the tests made to provide information on the purity of n-propanol. According to the manufacturer's instructions ( 8 4 ) for determining volume % y a calibration curve should be obtained of composition versus peak height. The errors involved i n taking this relationship to be linear are not l i k e l y to be significant here since the primary interest i s not i n absolute quantitative values but rather i n relative ones. Failure 25. to exactly identify the main impurity makes the former impossible. Other sources of error i n this type of analysis are outlined by the manufacturer (84) but, i n general, have not been considered relevant to these measurements. 3. Boiling Point D i f f i c u l t y was encountered i n controlling the drop rate from ebulliometer A (50), containing the primary standard. The rate at the top of the apparatus always exceeded that at the bottom which i s exactly the reverse of what was expected. It appeared that an error i n construction permitted refluxing liq u i d from the top to return to the boiler section without passing through the bottom drop counter. Nevertheless this seemed to have no particular detrimental effect on the equilibration. The measurement of the temperature i n both thermowells of ebulliometer A gave the same value to within 0.001°G. indicating that the primary standard was of an acceptable purity. Similarly i n the case of ebulliometer B (51), the two temperatures were within .005°C. of each other. According to Swietoslawski (55) this places both liquids i n the f i r s t degree of purity class. Only one thermometer was used to eliminate the errors involved i n using two or more, where the possibility of scale corrections i s presented. On the other hand, i t was almost impossible to measure the temperature i n two ebulliometers with one thermometer, while the room pressure remained constant. Nevertheless, corrections applied, as described previously, provided boiling points at 760 mm. pressure which are considered to have a maximum error of .005°C. This error takes into account possible errors i n calibration, lack of precise equilibrium conditions, and limitations i n reading the thermometer and applying normal corrections to i t . 4. Viscosity The apparatus and procedure for making the viscosity measurements 26. tabulated above have been c r i t i c a l l y discussed by de Verteuil (56). b. Purification The purification of n-propanol presents a number of problems. The work described herein has cl a r i f i e d these problems, and proceeded towards their solution. A significant problem i s the vagueness of the literature already available, with i t s omissions, discrepancies and conflicting statements. A second important problem i s the fact that the impurities and normal propanol are so close boiling that not only i s i t very d i f f i c u l t to separate them, but also to detect them by the measurement of physical properties. The purification procedure adopted for n-propanol I was essentially that of Kretchmer (16), although details of the bromine treatment were lacking. Successive t r i a l s at brominating some n-propanol I , to assist i n removing any a l l y l alcohol, showed l i t t l e improvement over the original sample when measured by means of a gas chromatographic analysis. This initiated doubts about l ) the bromination method, 2) differences i n the commercial grades of propanol used, 3) Kretchmer's conviction that the impurity was a l l y l alcohol. Further d i s t i l l a t i o n s at higher reflux ratios began to confirm the doubts about 2) or 3). The impurity peak decreased with d i s t i l l a t i o n at higher reflux ratios even without any bromination. The use of a nitrogen purge was found to be incidental since the product obtained from d i s t i l l a t i o n i n an a i r atmosphere was identical, within the limits of detecting differences i n successive gas chromatograms, to that i n a nitrogen atmosphere. The n-propanol I I , for which a mass spectrometer analysis was available, showed as i t s main impurity sec-butanol. The successive d i s t i l l a t i o n s at high reflux ratios gave a rate of decrease of the impurity peak on a chromatogram corresponding f a i r l y well to that of n-propanol I. In addition the two impurity peaks appeared i n the same positions on the chromatograms. A sample of the n-propanol I I I , as produced from n-propyl acetate, 27. also gave a small impurity peak i n the same position. This n-propanol I I I would have been more easily and efficiently prepared i f potassium hydroxide had been used rather than sodium hydroxide, because of the much greater solubility of the latter. No difference i n purity obtainable would be expected however. On the basis that either a l l y l alcohol or see-butanol was the main impurity indicated as being present, chromatograms were run on f a i r l y pure ( 90 - 95$) samples of each. These were used as standards for determining percentage impurity as described above. In addition a l l y l alcohol-propanol and sec-butanol-propanol mixtures were prepared and chromatograms obtained. In each case small concentrations of the suspected impurity (beginning at 1.0$ by volume) were introduced i n order to detect i t s effect on the impurity peak without completely masking i t . The essential aspects relating to the possible impurity identification as obtained from these chromatograms i s illustrated schematically by Figure I I . Increasing proportions of a l l y l alcohol i n the n-propanol caused distance "a" to decrease as peak A increased, i.e. shifting the impurity peak to the right. Conversely, addition of sec-butanol to the n-propanol caused the impurity peak to shift to the l e f t . This behavior make i t impossible to attribute the impurity effect to either a l l y l alcohol (b.p. 97.08°C. (27)) or sec-butanol (b.p; 99.53°C. (27)), and could not, i n i t s e l f , eliminate the possibility of another compound being present. The decrease i n concentration of the impurity with continued d i s t i l l a t i o n (see Results) would apparently be easier to achieve with sec-butanol having a boiling point difference of more than 2°C. However there i s an azeotrope between a l l y l alcohol and n-propanol boiling at 96.73°C (85) which should separate from propanol quite readily. At the same time, the failure of the bromination procedure to remove the impurity seems to indicate that a l l y l alcohol i s not a major impurity. None of the columns tested could cause separation of these peaks (A, B, C, Fig. I I ) . It was therefore concluded that the impurity was Following page 27 Figure I I EFFECT OF SUSPECTED IMPURITIES ON CHRGMATOGRAM PEAKS 28. l i k e l y a single compound, and probably sec-butanol, although i t s amount could be determined as a percentage of either a l l y l alcohol or sec-butanol. The> percentage impurity, as determined from either of the two chromatograms run as standards, was 5.5 volume per cent. In each case the standard had unknown impurities whose percentage could not be determined exactly, thus making 5*5% an uncertain value. The mass spectrometer analysis of n-propanol I I , with i t s average value of 3»5% sec-butanol, was therefore taken as representing the impurity peak i n n-propanol I and I I . A l l purified sample impurity concentrations were determined on this basis (84). From the chromatographic analysis, Table 1 shows n-propanol I I I to be the purest of the three different samples, with a value of better than 99.65 volume %. The boiling point was determined for n-propanol I I and I I I . (97.16°C and 97.17°C. respectively). Although there i s 0.01°C difference i n the reported values, they are essentially the same within the accuracy obtained. Moreover any error introduced by using two different methods of obtaining the atmospheric pressure is considered negligible. Considering the difference i n impurity percentages for n-propanol I I and I I I , the boiling point should probably not be taken as a primary criterion of purity i n this case. However when compared with Weissberger's value, 97.15°C. (27), these boiling points both give further indication of the presence of higher boiling impurity. The refractive indices decreased with decreasing per cent either of these two compounds. If plots of refractive index versus per cent composition of sec-butanol (in n-propanol) and a l l y l alcohol (in n-propanol) are considered linear, then the fact that n-propanol I I (before purification) l i e s 29. closer to the former plot indicates that sec-butanol i s more l i k e l y the impurity. While water present as impurity lowers the refractive index, of n-propanol, removal was clearly shown by the chromatograms. The best value obtained here i s 1.38524 as compared to 1.38556 selected by Weissberger (27) and based on a determination by Vogel (68), using a simple fractional d i s t i l l a t i o n for purification. From the experience of this investigation one d i s t i l l a t i o n is not sufficient, and Vogel's result should be regarded dubiously. The viscosity values were calculated by de Verteuil (56) assuming the density values as shown i n Table 3b. The densities were carefully selected and agree well with several other investigators. There are large differences among reported viscosity values, however, and while the results reported here are self-consistent and considered to be accurate to 0.5%, they cannot be used as a means of purity comparison and are given only for completeness. 3 0 . VAPOR PRESSURE APPARATUS The apparatus designed by Whittle ( 2 9 ) for vapor-liquid equilibrium measurements can of course be employed for 1-component vapor pressure measurements, as mentioned above, and i n so doing be simplified by eliminating the phase sampling section. Basically the apparatus consists of a glass purification tra i n connected to an equilibrium bomb (in a constant temperature bath) and a mercury storage bomb. The two bombs, of stainless steel, are identical and are connected i n a vertical position by a movable rod used for measuring levels i n the equilibrium bomb. They are also connected by high pressure tubing so that mercury can be transferred from one bomb to the other, thus enabling the rod to be moved without changing the mercury level i n the equilibrium bomb. The temperature and pressure of the sample are varied by means of the constant temperature bath, and nitrogen pressure on the mercury in the mercury storage bomb, respectively. A schematic representation of the bomb assembly, revised from that of Whittle ( 2 9 ) , i s shown i n Figure I I I . The revisions that have been required may be li s t e d as follows: 1 . The synthetic rubber 0-ring between the bottom bomb face and the pressure closure assembly was replaced with a teflon 0-ring. 2 . The level indicator, an N.R.C. design as described by Whittle (29), was replaced by a Pemberthy reflex type level gauge (Model No. V - 9 0 5 ) , pressure tested from 0 to 3 0 0 0 p s i . at 100°F. 3 . The previously silver soldered joint between the measuring head and the connecting rod was welded to eliminate the possibility of mercury attack on the solder. 4. For the purposes of vapor pressure measurements the position of the liquid vapor interface i s not essential. For this reason the hot wire anemometer was not installed. 31 5. Instead of using a resistance bridge for measuring the mercury le v e l , an ordinary relay circuit with an indicating light was found to be satisfactory. The circuit was opened or closed by raising or lowering the measuring head out of or into the mercury. The level could be ascertained to within 1.0 mm. This error was partially eliminated by consistently measuring from above the mercury surface. Pressure tests were i n i t i a l l y made on the entire system, excluding the glass purification t r a i n and the level gauge, using nitrogen from a regular storage cylinder up to 1000 p.s.i. and soap solution as a leak detector. Subsequently the system was f i l l e d with S.A.E. 10 o i l and the pressure raised to 4500 p.s.i. at room temperature then reduced to 2000 p.s.i. This pressure was held for three days with no evident sign of leakage. In order to transfer a sample into the equilibrium bomb, the purification t r a i n and equilibrium bomb had to be vacuum tight. A _3 vacuum of better than 10 ^  mm. of merctiry was obtained using a mercury diffusion pump i n conjunction with a Cenco Megavac vacuum pump. The rate of leakage caused a pressure change of approximately 0.001 mm./ minute. To f a c i l i t a t e cleaning of the system, technical grades of benzene, toluene, and n-propanol were circulated through the apparatus* cold. Once the exit solvents became clean, some n-propanol was pumped into the bomb (leaving sufficient room for expansion) and the bomb temperature raised to about 250°C. As the exit solvent from this treatment contained dirt particles and a considerable amount of discoloration, further batches of n-propanol were introduced. As the number increased, the solvent became cleaner and apparently clearer. However, on standing, the discoloration appeared, similar to the f i r s t batch. This was attributed to the air i n contact with the n-propanol causing an aldol condensation reaction which formed a coloured polymer (54a). It was then noted, after the above heating and cooling process, 32. that the equilibrium bomb was no longer vacuum or pressure tight, due to the apparent failure of the teflon V-rings i n the packing gland above gland nut A (Figure I I I ) . Increasing the pressure on the rings, by means of the gland nut and inserted s p l i t steel washers, had no apparent effect on the leak. The failure was ascribed to teflon's lack of geometric st a b i l i t y with respect to the heating and cooling cycle, possibly with some extrusion at the high temperatures. However, the manufacturer's specifications (86) state that teflon i s flexible to 260°C. Above 335°C. i t loses strength and around UOCpC. i t decomposes slowly. Further examination of the packing w i l l have to be made i n order to ascertain whether a major change i n design i s required. Other aspects of the apparatus design appear satisfactory for such measurements as that of the vapor pressure of n-propanol which has a c r i t i c a l temperature of 263°C. and a c r i t i c a l pressure of about 735 p.s.i. 33. VAPOR PRESSURE CORRELATIONS 1. Methods Vapor-pressure-temperature correlations have been made using the data of Young and the selected equations discussed i n the Literature Review. In addition to Riedel's reduced vapor pressure equation, log P R = A - B_ + C In T R + DTR (2) TR i t was considered of interest to check his equation using actual temperatures and pressures; i.e. log P - A - B + C In T + DT6 (5) T The two Antoine equations, of the form log P - A - B_. (3) T-C were evaluated on the basis of the selection of the constant C. Thomson's graphical method for estimating C (36) was used, taking T Q = 97.17°C and P Q = 760 mm. as the point assumed free from error. A plot of log P versus log P - log P q i s linear i f the Antoine T - T o equation holds. When this present data was plotted two straight lines were drawn through the points. The slopes for these two lines, -(To - C), and hence the values of C, -230°C. (below the b.p.) and -176°C. (above the boiling point), were used i n the equations. On conversion to engineering units they became 77.5 and 175°R. respectively. The third equation, proposed by Frost and Kalkwarf, 34. log P = A - B + C log T + DP T T2 (4) was applied directly to the data. In each case the best f i t for the data was obtained by the method of least squares. The regression coefficients were calculated on the U.B.C. electronic d i g i t a l computer, Alwac I I I E, programmed with Routine S-3 (87) for correlation and regression. In a l l cases, at least six significant figures were carried through the computation, reducing the possible error to less than 0.05$. The constants determined for these equations are tabulated in Table 4. TABLE 4. Constants for Vapor Pressure Correlations. Name Equation A S P . 2 Riedel 5 6.5592 6.5984 -3.4819 .0348 Riedel 2 32.694 6323.4 -3.3916 3.0x10 20 Antoine below b.p. 3 6.7282 3276.4 77.5 Antoine above b.p. 3 5.6444 2200.0 175 Frost 32.233 6288.6 -7.6641V 18.31 The per cent deviation of the values of P calculated from these equations i s compared i n Table 5 and Figure IV with those obtained by Young using the Biot Formula. t t log P = a + bpr + qj9 (1) 35 TABLE 5 COBRELATION DEVIATIONS P exp. % Deviation of PCalc from p exp. T°R. p.s.i.a. Biot Riedel (2) Riedel (5) Antoine Frost 491.688 .066519 +1.45 -0.69 -1.10 +1.10 -1.97 509,688 .140386 +1.79 -0.39 -1.03 -0.01 -0.24 527.688 .280385 +1.93 +0.94 +0.79 +0.60 +0.89 545.688 .533701 +1.92 +1.30 +1.24 +0.25 +1.35 563.688 .970717 +1.83 +1.37 +1.14 +0.49 +1.39 581.688 1.68619 +2.06 +1.60 +1.26 +0.67 +1.60 599.688 2.84254 +1.34 +0.73 +1.09 +0.03 +0.30 617.688 4.62154 +0.60 -0.13 +0.09 -0.43 -0.17 635.688 7.27071 -0.18 -1.21 -0.99 -0.69 -1.10 653.688 11.0994 -1.02 -2.00 -2.06 -0.79 -2.00 671.688 16.2915 -0.78 -1.81 -1.80 +0.50 -1.80 689.688 23.3204 -0.65 -1.63 -^1.70 +0.51 -1.60 707.688 32.5442 -0.35 -1.24 -1.30 +0.46 -1.20 725.688 44.3397 +0.12 -0.66 -0.75 +0.58 -0.60 743.688 59.4419 +0.20 -0.38 -0.51 +0.40 -0.32 761.688 78.3535 +0.12 -0.23 -0.64 +0.10 -0.04 779.688 101.790 -0.20 -0.32 -0.49 -0.35 -0.14 797.688 129.461 +0.04 +0.19 -0.22 -0.18 +0.35 815.688 162.102 +0.42 +0.83 +0.58 +0.18 +1.00 832.688 202.381 -0.20 +0.40 +0.15 =0.40 +0.60 851.688 247.533 +0.06 +0.80 +0.55 -0.08 +1.40 869.688 301.174 -0.23 +0.59 +0.36 -0.30 +0.45 887.688 361.196 -0.06 +0.76 +0.78 -0.06 +0.93 905.688 428.392 +0.34 +1.06 +1.04 +0.40 +1.16 923.688 506.513 +0.26 +0.73 +0.63 +0.30 +0.78 941.688 595.290 +0.07 +0.17 +0.12 +0.04 +0.13 959.688 698.124 -0.54 -0.97 -0.90 -0.73 -1.10 966.348 737.126 -0.45 -1.32 -1.13 -0.46 -1.56 Average Deviation 0.69 0.87 0.87 0.40 0.93 Following page 35 J i 1 i | I U — I 510 6 0 0 6 9 0 7 8 0 8 7 0 960°R. Figure IV CORRELATION DEVIATIONS 36. 2. Discussion In general, a l l the plots of Figure 17 show the same tendencies. Positive peaks occur at 581.7, 815»7, 851.7, 905.7 °R. and negative ones at 653.7, 833.7, 869.7 °R. i n nearly a l l cases. The large negative peak near the boiling point i s of particular interest. This occurs for Frost, Riedel, and Biot equations but i s eliminated by the use of two Antoine equations. Of the four equations compared, the Antoine correlation appears to give the best f i t . However, as was mentioned previously, the disadvantage of using two equations must be taken into consideration. Although Thomson (36) recommends T R = 0.8 as the intersection point of the equations, T R = 0.45 was used here for the best results. The extension of the equations beyond the intersection point results i n increased deviations, as i s shown by the dashed lines i n Figure IV. This tends to support the choice of T R = 0.45 as the intersection point. The next best f i t for the data i s given by the Biot formula. Although the deviations are quite large i n the low temperature regions they become considerably less significant above the boiling point. Blot's formula does have the advantage of covering the f u l l range of data, but i t i s questionable whether i t i s any better i n the low temperature region than the extension of the high temperature range Antoine equation. However this consideration becomes insignificant when the evaluation of the Biot constants i s taken into account. The solution of this five constant equation, following Prony's Method of Interpolation by Expotentials (88), requires much computation and almost excludes the possibility of using the method of least squares. The f i t of the two Riedel equations appear to be almost identical indicating that the c r i t i c a l data have been carefully measured. Both equations f i t this data well enough to consider applying them to other members of the n-alcohol series. However 37. there i s a greater possi b i l i t y of chain length relationships between the constants of a reduced form of the equation used for a homologous series. On the other hand, the equation using actual temperatures and pressures i s not dependent on the c r i t i c a l data for i t s correlation. This would be an advantage where c r i t i c a l constants have not been accurately determined or where available data i s fragmentary. The deviations of the Frost equation are very similar to those of the Riedel equations. These are, of course, identical except for the last term and the graphs indicate that this difference i s not very significant. I t would appear that the Frost equation does not account for the reverse 11 s " shaped curve i n the plot of log P versus VT. as Thodos (43) found i t did for the normal hydrocarbons. In the f i n a l analysis, the Antoine equations are the simplest to use. Although they f i t the data very well, there i s the dis-advantage of two equations. The Biot equation should not be considered because of the d i f f i c u l t y involved i n calculating the constants. There i s l i t t l e to choose between the Frost and two Riedel equations. Possibly the reduced form of the Riedel equation i s most useful because of i t s applicability to the theorem of corresponding states. 38. LITERATURE CITED 1. Lydersen, A. L., Greenkorn, R. A, and Hougen, 0. A., University of Wisconsin, Eng. Experimental Station Report No. 4 (1955). 2. Shemilt, L. W., The Thermodynamic Properties of the Normal Alcohols, Joint Conference on Thermodynamic and Transport Properties of Fluids, The Institute, London (1957). 3. Esplen, R. W., M.A.Sc, Thesis i n Chemical Engineering, University of B r i t i s h Columbia 1950. 4. Speers, E. A., M.A.Sc. Thesis i n Chemical Engineering, University of Br i t i s h Columbia 1958. 5. Young, S., Sci. Proc. Roy., Dublin Soc. 12, 374, (1910). 6. Partington, J. R., An Advanced Treatise on Physical Chemistry Vol. I, para. 3 VII B, Longmans, Green, London, 1951. 7. Dreisbach, A. M., Pressure-Volume-Temperature Relationships of Organic Compounds, Handbook Publishing, Ohio, 1952. 8. Riedel, L., Chem. Eng. Techn., 24, 353 (1952). 9. Pitzer, K. S., J. Am. Chem. Soc., 77, 3427 (1955). 10. Timmermans, M. J., Physics-Chemical Constants of Pure Organic  Compounds. Elsevier. N. Y., 1950. 11. Timmermanns, M. J., and Delcourt, M. Y., Journal de Chimie Physique, 31, 101 (1934). 12. Crismer, L., Bu l l . Soc. Chim. Belg., 18, 18 (1904) and 20, 382 (1906). 13. Ramsey,W., and Young, S., Trans. Roy. Soc, 180, 137 (1889). 39. 14. Fortey, and Young, S., Trans. Chem. Soc, 81, 717 (1902). 15. de Brouckers and Prigogine, B u l l . Soc Chlm. Belg., 47, 399 (1938). 16. Kretchmer, C. B., J. Phys. and Colloid Chem., 55, 1351 (1951). 17. Keyes, F. G., and Winninghoff, W. J., J. Am. Chem. Soc, 38, 1178 (1916). 18. Kraus, C. A., and Bishop, J. E., J. Am. Chem. Soc, 43, 1568 (1921). 19. Goldschmidt, H., and Thomas, L., Z. physik Chem., 126, 24 (1927). 20. Berner, E., Z. physik Chem., 141A, 91 (1929). 21. Lund, H., and Bjerrum, J., Ber. Deut. Chem. Gesell, 64, 210 (1931). 22. Brunei, R. F., Crenshaw, J, B., and Tobin, E., J. Am. Chem. Soc, 43, 561 (1921). 23. Wullner and Grotrian, Ann. Phys., 11, 545 (1880). 24. Tammann, ., Ann. Phys., 32, 683 (1887). 25. Shepherd, ., Bur. Stand. J. Res., 12, 185 (1934). 26. Taylor, ., J. Am. Chem. Soc, 52, 3576 (1930). 27. Weissberger, A., Organic Solvents Interscience, N.Y., 1955. 28. Brunei, R. F., J. Am. Chem. Soc, 45, 1334 (1923). 29. Whittle, D., M.A.Sc. Thesis i n Chemical Engineering, University of Br i t i s h Columbia, 1958. AO. 30. Sage, B. H., and Lacey, W. N., Trans. A.I.M.E., 174, 102 (1948). 31. Jepson, W. B., and Rowlinson, J. S., J. Chem. Phys., 23, 1599 (1955). 32. Beattie, J . A., Proc. Am. Acad. Arts Sc i . , 69, 389 (1934). 33. Kay, W. B., Ind. Eng. Chem., 28, 1015 (1936). 34. Bahlke, W. H. and Kay, W. B., Ind. Eng. Chem., 24, 291 (1932). 35. Esplen, R. W., M.A.Sc. Thesis, University of B r i t i s h Columbia, 1950. 36. Thomson, G. W., Chem. Revs., 38, 1 (1946). 37. Weissberger, A., Physical Methods of Organic Chemistry Interscience, N.I., 1949 second edition Vol. I, Part I , Ch. V. 38. Reid, R. C, and Sherwood, T. K., The Properties of Gases and Liquids McGraw-Hill. N.Y., 1958. 39. Biot, ., Compte Rende., 12, 150 (1841). 40. Ham, Churchill and Ryder, J. Franklin Inst., 15, 186 (1918). 41. Plank, R., and Riedel, L., Ing. Arch., 16, 255 (1948). 42. Frost, A. A., and Kalkwarf, D. R., J. Chem. Phys., 21, 264 (1953). 43. Perry, R. E., and Thodos, G., Ind. Eng. Chem., 44, 1649 (1952). 44. Private correspondence with Fisher Scientific Company Limited. 41. 45. Groggins, P. H., Unit Processes In Organic Synthesis. McGraw-H i l l , N.Y., p.624, 1958. 46. Private correspondence with Canadian Chemical Company Limited. 47. Beckmann Instruments Inc. catalogues: 545-C, 561-A. 48. Swietoslawski, W., Ebulliometric Measurements Reinhold, New York (1945). 49. Barr, W. E., and Anhorn, V. J., Scientific and Industrial Glass Blowing Laboratory Techniques. Pittsburgh, Instruments Publishing, 1949. 5 0 . Ibid, p.310, Fig. 179. 51. Ibid, p.311, Fig. 180. 52. Whittle, D., M.A.Sc. Thesis i n Chemical Engineering, University of B r i t i s h Columbia, 1958. 53. Cannon, M. R. and Fenske, M. R., Ind. Eng. Chem. (Anal. Ed.) 10, 297 (1938). 54. A.S.T.M. Standards on Petroleum Products and Lubricants, p.192, 1954. 54a. Consultation with the Department of Chemistry, University of B r i t i s h Columbia, 1958. 55. Swietoslawski, W., op. c i t . , p.79. 56. de Verteuil, G. F., M.A.Sc. Thesis i n Chemical Engineering, 1958. 57. Young, S., and Fortey, E. C, J. Chem. Soc. (London) 83, 45 (1903). 58. Dorochewsky, A. G., J. Russ. Phys. Chem. Soc, 41, 972 (1909). 42. 59. Dorochewsky, A. G., J. Russ. Phys. Chem. Soc, 43, 66 (1911). 60. Mundel, C. P., Z. Physik Chem., 85, 435 (1913). 61. Brunei, N. F., Crenshaw, J. L., and Tobin, £., J. Am. Chem. Soc, 43, 501 (1921). 62. Grimm, F. W. and Patrick, W. A., J. Am. Chem. Soc, 45 , 2794 (1923) 63. Trew, V. C. G. and Watkins, G. M. C, Trans. Faraday Soc, 29, 1310 (1933). 64. Timmermanns, J. and Delcourt, Y., J. Chim. Phys., 31, 85 (1934). 65. Wojciechowski, M., J. Research Nat. Bur. Standards, 17, 721 (1936). 66. Zepalova, ., Mikhailova, L. A., Trans. Inst. Pure Chem. Reagents (Moscow) No. 15, 1 (1937). 67. Addison, C. C, J. Chem. Soc, 98 (1945). 68. Vogel, A. T., J. Chem. Soc, (1948) 1814. 69. Carley, J. F. and Bertelsen, L. W., Ind. Eng. Chem., 41, 2806 (1949). 70. Mumford, S. A. and P h i l l i p , J. W., J. Chem. Soc, 75, (1950). 71. Howey, G. R., M.A.Sc. Thesis i n Chemical Engineering, University of Bri t i s h Columbia, 1951. 72. McKenna, F. E., Tartar, H. V., and Lingafelter, E. C i , J. Am. Chem. Soc, 75, 604 (1953). 73« Wetzel, F. H., M i l l e r , J. C. and Day, A. R., J. Am. Chem. Soc, 75, 1150 (1953). 43. 74. Purnell, J. H. and Bowden, S. T., J. Chem. Soc, 539 (1954). 75. Gartenmeister, R., Z. physik Chem., 6, 524 (1890). 76. Thorpe, T. E. and Rodger, J. W., P h i l . Trans., 185A, 464 (1894). 77. Thole, F. B., Z. physik Chem., 74, 686 (1910), J. Chem. Soc, 105, 2011 (1914). 78. Baker, F., J. Chem. Soc, 1409 (1912). 79. Dunstan, A. E., and Thole, F. B., J. Chem. Soc, 12? (1913). 80. English, S., and Turner, V. E. S., J. Chem. Soc 105, 1658 (1914). 81. Hertz, W., Z. Anorg. Chem., 47, 104 (1918). 82. Whitman, J. L. and Hurt, D. M., J. Am. Chem. Soc, 52, 4762 (1930). 83. Jones, W. J., S. T. Bowden, W. W. Yarnold and W. H. Jones, J. Phys. and Colloid Chem., 52, 753 (1948). 8 4 . Beckmann Instruments Inc. Catalogue No. GC-86-M1, Peak Height and Peak Area Measurement Techniques. 1958. 85. Azeotropic Data. Advances i n Chemistry Series No. 6, American Chemical Society 1951. 86. Anchor Packing Company Limited, Montreal, Catalogue No. 2. 87. Computing Centre, University of B r i t i s h Columbia, 1959. 88. Whittaker, E. T., and Robinson, G., The Calculus of Observations. Blackie, London, 1932, p.369. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0059064/manifest

Comment

Related Items