UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The measurement of the vapour pressure of mercury in the intermediate pressure range Dauphinee, Thomas McCaul 1950

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

Item Metadata


831-UBC_1950_A1 D2 M3.pdf [ 7.89MB ]
JSON: 831-1.0085484.json
JSON-LD: 831-1.0085484-ld.json
RDF/XML (Pretty): 831-1.0085484-rdf.xml
RDF/JSON: 831-1.0085484-rdf.json
Turtle: 831-1.0085484-turtle.txt
N-Triples: 831-1.0085484-rdf-ntriples.txt
Original Record: 831-1.0085484-source.json
Full Text

Full Text

THE MEASUREMENT OF THE VAPOUR PRESSURE Off MERCURY  IN THE INTERMEDIATE PRESSURE. RANGE. by Thomas McCaul Dau.ph.inee oOo A THESIS SUBMITTED IN PARTIAL FULFILMENT QE THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A September, 1950 T H E UNIVERSITY OF BRITISH COLUMBIA F A C U L T Y OF G R A D U A T E STUDIES P R O G R A M M E O F T H E F I N A L O R A L E X A M I N A T I O N F O R T H E D E G R E E O F D O C T O R O F P H I L O S O P H Y T H O M A S McCAUL DAUPHINEE B.A. (Brit. Col.) 1943 M.A. (Brit. Col.) 1945 T H U R S D A Y , O C T O B E R 12th, 1950 A T 2 P.M. IN R O O M 201, PHYSICS BUILDING of COMMITTEE IN CHARGE: Dean H . F. Angus, Chairman Professor G. M. Shrum Professor G. M. Yolkoft Professor G. L. Pickard Professor L . W. Shemilt Professor Barnett Savery Professor R. D. James Professor C. E. Dolman Professor F. A. Kaempfler 0 THESIS T H E M E A S U R E M E N T OF T H E V A P O U R PRESSURE OF M E R C U R Y IN T H E I N T E R M E D I A T E PRESSURE R A N G E An apparatus for the measurement of vapour pressures of metals in the pressure range 5 x 10— : 1 to 102 mm. of Hg. has been built and is in operation. It is believed to be the first generally applicable vapour pressure apparatus to cover this entire range. The apparatus operates on the "streaming" or "trans-piration" principle. A carrier gas (argon) is saturated with metallic vapour at the temperature of the measurement. The vapour is' subsequently condensed out, and the quantity of metal transported is determined by titration. Features that have not been applied before in this pressure range include: (1) bringing the gas to saturation by condensing out an excess of metallic vapour, (instead of depending on evaporation) ; (2) very greatly increased surface over which the vapour and argon come to equilibrium; (3) , increased How of gas which makes measurement at lower pressures possible; (4) use of a completely enclosed system for measurements and for purification of the argon; (5) very precise temperature control. The modifications required to adapt the apparatus to a large number of metals do not affect the basic principles of measurement or the above features. A large part of the system is interchangeable. The vapour pressure of mercury has been measured in the pressure range 5 x 10—3 to 60 mm. (32°C to 240°C) . The results, which are thought to be the most accurate yet obtained below 130°C are significantly above other measure-ments at lower temperatures, but completely confirm the vapour pressure equation recently proposed by Ditchbum and Gilmour. log p = 10.3735 - 3308 - 0.8 log T T Equations for the total vapour pressure of mercury and for the partial pressure of mercury atoms have been derived by Kelly and Dushman, respec-tively from the free energy calculations of the former. (Mercury has a small proportion of Hg 2 molecules at the boiling point). These equations are not consistent in this pressure region and neither is in accord with the present experiment. The vapour pressure table given for mercury in the International Critical Tables is accurate above 160°C, but a graph comparing this table with the equations mentioned above shows a very unnatural change of slope in the 160°C region. The table is not reliable at lower temperatures and it is probably in error by about 15% at the freezing point (—38°C). Between 160°C and the freezing point the Ditchbum and Gilmour equation should be used. The error of this equation is probably not over 4% at the freezing point. G R A D U A T E STUDIES Field of Study: Physics. Electromagnetic Theory—Professor H . D. Smith. Quantum Mechanics—Professor G. M . Volkoff. Nuclear Physics—Professor A. E . Hennings. Spectroscopy—Professor A. M . Crooker X-rays and Crystal Structure—Professor A. E. Hennings Electronics—Professor A. van der Ziel. Theory of Measurements—Professor A. M . Crooker. Electron Optics—Professor J . H . L . Watson. Chemical Physics—Professor A. J. Dekker. Solid State—Professor H . Koppe. Group Theory—Professor H . Koppe. Radiation Theory—Professor F. A. Kaempffer. Other Studies: Physical Chemistry—Professor J . G. Hooley. Theory of Functions of a Complex Variable—Professor W. H . Simons. Advanced Calculus—Professor D. Buchanan. Analytical Geometry—Professor F. S. Nowlan. T H E U N I V E R S I T Y O P BRITISH C O L U M B I A VANCOUVER. CANADA D E P A R T M E N T O F P H Y S I C S October 12, 1950, This w i l l c e r t i f y that the following members on the Committee of Mr. Thomas McCaul Dauphinee have examined his thesis and found that the material f u l f i l s the require-ments of merit and o r i g i n a l i t y prescribed under the regula-tions f o r the Ph.D. degree. Dr. G^M. &hrum Dr. G. M. Volkof Dr. G.L. Pickard Dr. F. A. Kaempffer The research undertaken by Mr. Dauphinee was under the d i r e c t i o n of Dr. G. L. Pickard and Dr. F. A. Kaempffer. ABSTRACT An apparatus for determination of vapour pressures of metals in the pressure range 5 x 10""3 to 10^ mm. Hg. by the streaming or transpiration method is described. The carrier gas i s circulated, and purified, in a closed system. Saturation i s achieved by condensing out an excess of the metallic vapour. Very large condensing surface and increased flow are used, with provision for checking the effects of flow, i n i t i a l quantity of metallic -vapour, total pressure and temperature gradients. The quantity of metal carried over i s determined by t i t r a t i o n . Results of measurements of the vapour pressure of mercury from 5 x 10-3 mm. to 60 mm. (3Q°C. to 240°C.) are given. Low temperature values are significantly above previous determinations, but are in complete agreement with the vapour pressure formula proposed by Ditchbum and Gilmour in 1941. log p. s 10.3735 - 3308 . 0.8 log T T Values given by the International C r i t i c a l Tables are shown to be inaccurate below 140°C. and equations based on free energy calculations are shown to be inconsistent and unsuitable for extrapolation to low pressures. The Ditchburn and Gilmour formula i s recommended for a l l temperatures between 140°C. and -38<>C. (the freezing point). The error at the lower temperature i s probably less than 4$. TABUS OF CONTESTS Introduction 1 Theory 7 (a) The Kirchhoff Equation 7 (bj The Chemical Constant 8 (c) Calculation of the Coe f f i c i e n t of Log T 9 (d) Calculation of Log p from the Free Energy 10 (e) The Work Function 11 Experimental 12 1* The Apparatus 14 2. Description of the Process Cycles 16 3. The Mixer, Saturator and Branch 18 4. The Magnetic Yalvea and Mercury Traps 20 5. The Flowmeter 22 6. Pressure Measurement and Control 23 7. The C i r c u l a t i o n of Argon 24 8. P u r i f i c a t i o n 26 9. Heaters and Control of Temperature 28 10. Thermal Shielding 31 11. Temperature Measurement 31 12. Procedure of Measurement 37 13. Change Oyer 40 14. T i t r a t i o n 41 15. Experimental Error 44 16. Conversion to Other Substances 48 Results 49 1. Calculations 49 2. Readings 51 3. Discussion 54 Conclusion 58 Acknowledgements 60 Bibliography 61 I I . PLATES Plate After Page I* The Apparatus 11 I I . The Control Panel 12 I I I . The Mixer, Saturator and Branch 13 IV. The Magnetic Valves and the Traps 15 V. Plow Diagram 17 VI. The Mixer, Saturator and Branch 20 V I I . The Magnetic Valves and the Traps 21 V I I I . The Temperature Control C i r c u i t 29 IX. The Platinum va. Platinum-10^ Rhodium Thermocouple C i r c u i t 33 X. The Iron-constantan Thermocouple C i r c u i t 34 XI . The Vapour Pressure of Mercury from 30°C. to 240°C. 53 X I I . Comparison between Present Results and Proposed Equations 54 TABLES Table After Page I. Readings 51 I I . Results 52 I I I . IETRODUCTIQlSr The accurate measurement of vapour pressures has been of considerable importance and interest to both the physicist and the chemist for a long time. Using vapour pressure information i t ia possible to calculate the chemical constants (l.a) of various substances and so check the value predicted by s t a t i s t i c a l mechanics (2.a). Latent heats can be calculated for temperatures far from the boiling point, and very high boiling points can be predicted by extrapola-tion of the vapour pressure curve. Vapour pressure informa-tion is also necessary for the interpretation of results of other experiments such as optical absorption measurements and for checking calculated values of the free energy. On the technical side, a knowledge of vapour pressure variation can be used in thermometry and i s essential for high vacuum work and for problems involving contamination or s u i t a b i l i t y for high temperatures. Further examples of the use of these data are provided in this Physics Department in the prepara-tion of thin evaporated films and saturation of crystals. Chemical applications of vapour pressure and vapour density measurements include the calculation of equilibrium constants and determination of the degree of association into mole-cules (3), as well as measurement of molecular weights and dissociation pressures. In spite of this interest in vapour pressures i t was not u n t i l around 1925 that measuring techniques became - 2 -aufficiently refined and the temperature scale sufficiently well-established for agreement between different workers to become quantitative, rather than qualitative, in the lower pressure regions. Since 1925, a f a i r amount of competent research has been done and for a few metals a relatively complete set of results is available, although experimental errors of over 5% are more or less standard in a l l cases. For the great majority of metals there is s t i l l l i t t l e reliable information. The extreme range over which the vapour pressure varies (10 4 to 10"8 mm. of Hg.) makes i t out of the question to get a complete set of measurements for any one element with one apparatus} and this has resulted in a rather spotty distribution of those accurate measurements that have been made. Attempts to extrapolate over wide ranges beyond those covered by the experiments have not been too successful. At the present time there i s great need for a deliberate attempt to f i l l i n the gaps in this experi-mental knowledge and to develop measuring techniques that w i l l enable the experimenter to cover a wide range of pressures with a single apparatus. The types of apparatus that have been used f a l l rather naturally into two main groups: 1. high pressure methods (say from 1 mm. Hg. up) and 2. low pressure methods (from 10""2 mm. Hg. down). It w i l l be noted that the two ranges given do not overlap. - 3 -There do not appear to he any methods which are peculiarly adapted to the intermediate pressure range (4) and i t he comes desirable to develop techniques which w i l l make possible the extension of one of the methods of group 1 or 2, without too great a loss of accuracy, u n t i l i t overlaps with the other group. This has been done in the case of the higher alkalis by a method that applies only to them, and Rodebush et a l (5) (6} (7) have developed and used an absolute manometer in the 1 to 100 micron region which appears to be suitable for use at much higher pressures. Nearly a l l the other low pressure methods break down for fundamental theoretical reasons at about 10 microns, and most of the high pressure methods break down, because of funda-mental limits of sensitivity, between 1 and 10 mm. Moreover, many methods of both groups are restricted i n the temperature range over which they may be used and therefore have very limited application. The method that has been used in the present experi-ment i s a high pressure one, the so-called "streaming method". In this method, a slow-moving stream of an inert gas i s saturated at the required temperature as i t flows over the surface of the metal, and the quantity of metal evaporated) together with the volume of the carrier gas, is used to calculate the vapour pressure. This i s actually a vapour density measurement, but for monatomic vapours at pressures - 4 -and temperatures where the perfect gas laws apply no appreciable error arises in the vapour pressure calculation. Part of the reason why a relatively small number of streaming experiments has been performed is undoubtedly the very formidable technical d i f f i c u l t i e s and time-consuming construction that must be surmounted in setting up the i n i t i a l apparatus. Furthermore, measurements are slow to take and continuous readings are impossible. On the other hand, the method has certain distinct advantages which the author feels have been too often overlooked. It i s much less dependent on the surface condition of the sample and on the rate of evaporation. It is selective; that i s , i t can be made to measure the pressure of a particular substance even when other vapours are present. This selectivity makes i t almost the only method that could be used to measure two vapour pressures simultaneously, such as the partial pressures of the constituents of an alloy. It has no theoretical l i m i t of operating temperatures or pressures other than the requirement that the gas laws apply. The only practical temperature limitation i s created by the necessity of finding a container that w i l l be gas-tight and inert to the metal being studied, while the most important pressure limitation i s at the low pressure end where amounts of metallic vapour carried over become too small to measure. The fact that the streaming method is capable of good accuracy over a limited range has already been demon-strated (8). But, to date, no one has developed an apparatus that w i l l give good accuracy over a range of, say, 10 4 of pressure, and, in particular, no measurements have been made in the 1 to 10 micron region where overlapping with the low pressure methods would occur. Mercury was chosen as the f i r s t test substance for several reasons. In the high pressure region (10 mm. and up) the vapour pressure is known quite accurately, but there are only two reliable experimental (9) values below 5 mm. at one end of the intermediate range, and none above 7 microns at the other. K.K. Kelly (3) i n 1935, Ditchbum and Gilmour (4) in 1941 and Dushman (10.a) i n 1949 (on the basis of Kelly's calculations), have proposed equations for the .vapour pressure of mercury which diverge appreciably in this inter-mediate region. The Ditchbum and Gilmour curve runs appreciably above a l l the results of recent determinations from 75°C. down to room temperatures, while the Kelly and Dushman curves are not in agreement with experiments near the freezing point. With mercury, then, i t would he possible to check the operation of the apparatus at higher pressures, perform the f i r s t experimental interpolation across the inter-mediate pressure range and attempt to learn which of the suggested equations is most nearly correct. By means of the apparatus described i n the follow-- 6 -ing pages i t has been possible to measure the vapour pressure of mercury at 10°C. intervals from 5 x 10"*3 mm. to 60 mm. of Hg., thus covering the entire intermediate pressure range. The results differ appreciably from the values given by the Kelly and Dushman equations but agree well within experi-mental error with the equation of Ditchburn and Gilmour. This equation i s probably correct within 2% as far down as ?0°C.f and is the only one that should be used for extra-polating to lower pressures. The streaming method i s established as suitable for interpolation over the inter-mediate range and for checking the results of low pressure methods. THEORY (a) The Kirchhoff Equation The usual way to express the variation of vapour © pressure of metals with temperature i s by an equation of the form -log p » H - B - q log T. (l) (Kirchhoff*s equation) T This equation can be quite simply derived by means of a Oarnot cycle (Fig. I) as follows: The net work in passing around the cycle may be expressed as either w * Zip ( v g - v c) (a) or Al T (b) FIG. I (Q, « heat of evaporation) (vg« volume in the vapour state) (vc» volume i n the condensed state) P Equating (a) and (b), neglecting v G, and replacing v g by H3L on the assumption that perfect gas laws apply, we get an equation having the form -(2) d In p dT RJ2 8 4 may always be expressed by an equation of the form -4 . 4 - - J (G c - Gg) dT (2 a) where 4j. i s heat of evaporation at temperature T.^  A i are any heats of transition of the condensed phase between and T, C c and C g are the specific heats at con-stant pressure of the two phases, and T T^. Where C c - C^  i s approximately constant and no intermediate transitions occur, we get, inserting (2) in (1) and simplifying -a m p s 4 i • ( Q C - cg* T i - (QQ - c g ) ( 3 ) dT RT 2 RT i n p « ^ * ( C * - CS> T l - ^ - ^ l a T i C (4} RT R Since the terms i n brackets are constants, the integration j u s t i f i e s the form of equation (l) or the equivalent -- IL . _- c RI p » &>i T e (b) The Chemical Constant If (C c - Gg) i s not assumed constant, the result-ing formula can be brought to the form -T T d 1 x 1 p 8 4 0 • £ R T / ' C, dT - / G c dT (5) 4T 2 / £ / and T T In p = - i o 4 £ l n T * / £T_ / (C, - Gc) dT • t (6) RT 2 J a i s ) s o o The constant of integration in (6) is called the Chemical Constant. This constant is of interest in chemistry because i t enters into the calculation of the Equilibrium Constanta of chemical reactions. It may be evaluated using equations (4) and (5} from results of experimental measurements. S t a t i s t i c a l mechanics also gives us a value for this constant -(fo-rr 13/2^5/2 \ m * m a s s o f a t o m  (2// m j ^ k ^ * g h.3 / concerned. g = ratio of s t a t i s t i c a l weights of the phases. It w i l l be noted that the properties of the individual substance do not enter into the calculation of M i " in this case* " i " i s a universal constant, and equation (6) should give the same value for a l l substances. As far as can be determined from available data, this i s true. (c) Calculation of the Coefficient of log T. An equation having the form of equation (4) i s not very sensitive to small changes in the power of T used (coefficient of In T). For this reason, the coefficient i s * 10 -extremely d i f f i c u l t to determine experimentally, and many workers prefer to calculate i t from the specific heat data and then determine the other constants from experimental results by the method of least squares. As an example the calculation of the coefficient of In T i s given for mercury. Mercury i s monatomic and therefore Gg la 5. R. The molar specific heat of mercury i s 6.64 Cal./°C. at 20°C. and 6.42 Cal./°C. at about 250°C. The coefficient of In T (also of log T) is thus between 6.64 - 2.5 R 6.42 - 2.5 R R R that i s , between .85 and .72. A good representative value i s 0.8. Ditchburn and Gilmour (4) obtained their equation, log Pmm 3 10.3735 - 3 3 0 8 - 0.8 log T (?) T for mercury by a process of the type described above. They used the very accurately known high pressure values to obtain the other constants. (d) Calculation of log p from the Free Energy If the equation for the variation with temperature of the free energy can be calculated, i t i s possible to calculate the vapour pressure from the equation -i f f 0 « RT In p - 11 -The Kelly (3) and Dushman (10.a) vapour pressure equations were derived in this way. The free energy calcula-tion gives a value for the coefficient of log T of 0.825. (ej The Work Function Equation (2) may he converted to the form -R d l a P . - 4 (8) -d (£) Equation (8) can be used, along with a plot of log p vs. l/T to determine Q,. How 4 T » LXr. • RT (9) where X»j. * work function for the metal at temperature T, L s a conversion factor, and RT i s the work to evaporate 1 mole. We may get a measure of I 0 , the work function at absolute zero, as follows: i . evaporation at temperature T after heating the solid to that temperature, energy required : 3 BI « LXrr. (10) i i . evaporation at T s 0 and subsequent heating of the vapour to temperature T. energy required s L X Q * 2^ from (9), (10} and (11). 4 = L X Q - £ RT X 0 « * BI (12) P L AT E I THE APPARATUS Thermometry Power Panel Panel Thermocouple Switches Zone Block PLATE I. The Apparatus Branch Saturator Mixer Evaporator Yenturi Pump Pressure Control F i l t e r B a l l a s t Volumes Vapour Magnesium Line D i f f u s i o n arc Pump bulb Mercury Bo i l e r s P h i l l i p s Gauge Amplifier - 12 -EXPERIMENTAL The streaming method of measuring vapour pressures does not give the vapour pressure directly. The quantity that i s measured i s the vapour density, and this is used to calculate the vapour pressure using the gas equation -p Y • m R T (13) where p is the vapour pressure, Y i s the volume occupied by the vapour, m la the number of moles evaporated, R i s the gas constant and m/Y i s the density in molar units. The measurement of Y i s accomplished by having the metallic vapour swept along by a carrier gas. The quantity of gas that passes by may be determined elsewhere in the c i r c u i t , and the volumes of the gas and vapour are the same at the point of measurement. The quantity m can be determined by t i t r a -tion after the metal has been condensed out of the gas stream. An equivalent equation which combines the actual measurements made into a direct calculation of the vapour pressure is -P - Pi * (14) M • m where P^ is the total pressure at the point of measurement and M is the quantity in moles of the carrier gas. This is the equation that is usually used. Both of the equations are contingent on the vapour and the carrier gas obeying the perfect gas laws, but under the usual conditions of P L A T E I I THE CONTROL PANELS The Power Panel Manometers Meters Resistors Yariacs PLATE I I . The Control Panels The Thermometry Panel Potentiometer Switches Temperature Control Setting Galvanometer Lamps and Scales Thermocouple Switches White Potentiometer Rubicon. Potentio-meter Tapping Key Thermocouple Switch • - 13 -measurement no appreciable error i s introduced through, t h i s assumption. Extension of the range of the streaming method to lover pressures involves finding some way of increasing the amount of metal that i s carried over so that the t i t r a t i o n process does not become less and less accurate. Increasing the length of time helps, but some increase i n the rate of flow i s desirable i f the runs are not to take several days to complete. On the other hand, with the customary procedure of passing the gas over the surface of the metal a t the temperature of the measurement, the permissible rate of flow tends to decrease as temperatures get lower. This e f f e c t i s accentuated when the metal has to stay i n the apparatus for any considerable time, since the s l i g h t e s t surface contamina-tio n may seriously a f f e c t the rate of evaporation. In t h i s apparatus, increase i n the rate of flow was accomplished, while at the same time ensuring proper satura-t i o n of the gas, by providing a tremendously increased surface i n the form of a condenser column of glass beads. At the same time the usual evaporation procedure was reversed, the gas being f i r s t saturated at a temperature higher than that of the test and the excess subsequently condensed out to equilibrium at the test temperature. The condensing vapour ensured that the large surface area was maintained, and at the same time provided a continuously replaced fresh surface P L ATE I I I THE MIXER, SATURATOR AMD BRANCH P L A T E III Top of glass thermocouple w e l l . Branch. front - manometer sidearm back - flow tube small beads Saturator showing ceramic thermocouple wells. large beads Pre-saturator Mixer and Inlet Tube Evaporator Evaporator Heater - 14 -with a correspondingly lower r e f l e c t i o n c o e f f i c i e n t for any atoms that struck i t . To prevent contamination of the metal "by the large volumes of c a r r i e r gas that swept by, the gas was r e c i r c u l a t e d i n a closed system which included provisions for p u r i f i c a t i o n . In t h i s way about seven l i t r e s of pure gas took the place of several hundred l i t r e s (or f a r more at low temperatures) which would otherwise have had to be supplied. A detailed description of the process, as applied to measure-ment of the vapour pressure of mercury, follows. 1. The Apparatus Pictures of the complete apparatus and of various parts are shown i n Plates I to IT. A block diagram of the apparatus which i s also a flow sheet for the d i f f e r e n t stages of the process i s given i n Plate V. The basic parts of the apparatus were: ( l e t t e r designations refer to Plate V) Mixer (B). (See Plate I I I . ) The chamber i n which the gas (argon) stream was saturated with mercury vapour. Evaporator (C). (See Plate I I I . ) The container i n which the mercury sample was heated. The evaporated mercury passed up a tube to a j e t i n the mixer. Saturator (E). (See Plate I I I . ) A long condenser column f i l l e d with glass beads upon which the excess mercury could condense. At the top of the beads the gas was presumed to be jus t saturated with mercury. 15 Branch. (See Plate I I I . ) The top of the saturator column. The c a p i l l a r y leading to the manometers (P) and (L) was joined on at t h i s point. Plow Tube (G). A heated tube through which the gas passed from branch to magnetic valves (H). Magnetic Valvea (H). (See Plate IT.) A set of b a l l check valves which were used to dir e c t the gas flow into one of a set of traps ( I ) . Main Traps ( I ) . (See Plate IV.) The traps i n which the mercury vapour condensed. The quantity of mercury was determined by t i t r a t i n g their contents. C i r c u l a t i n g and Purifying System. (See Plate I.) This system included a mercury j e t c i r c u l a t i n g pump (0), a f i l t e r system (P) to eliminate i r r e g u l a r i t i e s i n the gas flow, a magnesium arc (Q.) for pu r i f y i n g purposes, two large c y l i n d r i c a l bulbs (N) to act as b a l l a s t f o r the pressure i n the system and a c a p i l l a r y flowmeter (M) to measure the quantity of gas that went through the saturator. The entire system was made of Pyrex glass. B r i e f l y , then, the actual measuring process was as follows: The c a r r i e r gas (pure argon) entered the measuring section at the i n l e t tube (A) to the mixer. In the mixer (B) i t received a large excess of mercury from the evaporator (C). The excess mercury condensed out to P L A T E IV THE MAGNETIC VALVES AND THE TRAPS PLATE IT. The Magnetic Valves and the Traps D i f f u s i o n Outlet Magnetic Outlet Flowmeter Pump Stopcocks Valves Stopcocks Saturator C a p i l l a r y Traps equilibrium before i t arrived at the top of the saturator (E). The temperature of the gas was raised i n the flow tube (G) to prevent condensation and i t passed to the magnetic valves (H). The valves directed the argon-mercury mixture into the traps (I) being used for that measurement. The traps were immersed i n l i q u i d oxygen and i n them the mercury condensed, while the argon proceeded on to the flowmeter (M), through the rest of the c i r c u l a t i n g system, and back to the i n l e t tube. After the measurement, the traps were removed and the amount of mercury determined by t i t r a t i o n . 2. Description of the Process Cycles There were four basic steps to the operation of the apparatus after proper temperature equilibrium (to be d i s -cussed l a t e r ) had been reached. The cycles for each are shown by colored arrows i n Plate V. In Step 1 (a cleaning run) the c a r r i e r gas passed through the outlet stopcocks (S) to the appropriate main trap ( I ) , and from there, through the magnetic valves (H), the flow tube (G) and the branch to the saturator(E)« A l l of these parts were kept at a tempera-ture w e l l above that of the measurement to make sure that a l l mercury i n them would be carried into the saturator (E) where i t could do no harm. Prom the saturator, the gas passed through the mixer (B) and out of the measuring apparatus. After passing through the c i r c u l a t i n g system, - 17 -the gas returned to the traps* In Step 2 (a balance run) the flow of gas through the apparatus was reversed: being passed through the i n l e t trap, the mixer, the saturator, the branch, the flow tube, the valves and a pair of traps that were reserved for balancing-out purposes. These traps were immersed i n l i q u i d oxygen to prevent any mercury from passing through and l a t e r d i f f u s i n g into the traps used for measurement. This condition was maintained u n t i l s u f f i c i e n t time had elapsed to ensure that equilibrium had been reached between the gas stream and the mercury-covered surfaces of the saturator. In Step 3 (the actual measurement) the flow was exactly the same as i n Step 2 except that the argon now passed through the set of traps that had been cleaned out i n Step 1. In Step 4 the flow was turned back to the traps used i n Step 2 and the traps used for the measurement (Step 3) were sealed off at constrictions and removed. I f , f o r any reason, i t was not desirable to st a r t the balance-out run (Step 2) immediately after cleaning out the traps, a cross connection was opened which led from immediately before the i n l e t trap d i r e c t l y to the trap (K) i n f r o n t of the flowmeter (M), thus bypassing the whole main part of the apparatus, but maintaining approximately the same pressure head against the c i r c u l a t i n g system. P L A T E V PLOW DIAGRAM PLATE V. Plow Diagram {_) Stopcocks A. I n l e t trap B. Mixer G. Evaporator D. Pre-saturator E. Saturator P. Manometer G. Plow tube H. Magnetic valves I . Traps J . Seal-off constrictions K. Outlet trap L» Total pressure manometer M» Flowmeter U. B a l l a s t volume 0. Mercury j e t c i r c u l a t i n g pump and b o i l e r P, Pressure control f i l t e r 0. Magnesium arc Key to arrows -* — Step 1 Z. Steps 2, 3 and 4 — C i r c u i t for bypassing the main apparatus C i r c u i t for removing water vapour C i r c u i t for purifying arc - 18 -3. The Mixer, Saturator and Branch. The construction of the mixer, saturator and branch are shown i n d e t a i l i n Plate VI. (See also Plate I I I . ) Pure mercury (G) contained i n the evaporator (B) was heated by r a d i a t i o n from a heater element (D) placed below i t and the vapour passed up a tube to the j e t (E) i n the mixer. A re t u r n tube (P) was provided for the condensed mercury. The i n l e t tube (A), the mixer and the. tube (pre-saturator (G)) leading to the saturator were wound with nichrome heating elements embedded i n asbestos cement (R, S). Two iron-con-stantan (No. 30 B. & S.) thermocouples were mounted as shown next to the glass of the mixer, and a t h i r d was mounted against the i n l e t tube. A platinum vs. platinum-10$ rhodium thermocouple was mounted at the w a l l of the pre-saturator and the leads brought out through the cement at the top end of the heater element. A disc of 0.005 i n . copper f o i l (H) at the top of the pre-saturator eliminated any leakage from the mixer heaters to the saturator couples. The glass beads (I) which f i l l e d the inside of the saturator were supported by the top end (J) of the pre-satur-ator tube which projected into the saturator through a r i n g -s e a l . Por several inches from the bottom these beads were large, and put i n layers of three, while the rest of the beads were 4 mm. i n diameter. A glass thermocouple w e l l (K) passed through a r i n g seal i n the branch and extended down - 19 tiie middle of the saturator through the small beads almost to the larger ones. J?our long ceramic thermocouple wells (3$) (W) were mounted against the w a l l of the saturator and extended past the branch out of the apparatus. The space between the well s was f i l l e d with asbestos and outside them was a three-section furnace (T) which covered the whole of the saturator, and extended about 5 cm. beyond the top of the beads. Each section was about 13 cm. long and had for i t s core a 6 mm. copper sleeve (IT) ( s p l i t f o r mounting purposes) upon which was wound a 50-ohm,non-inductive resistance thermometer (V) of Hytemco wire (Ho. 30 B. & S.). A layer of asbestos tape 5 mm. thi c k separated the resistance thermometer from the heater and the remainder of the i n s u l a -t i o n was asbestos wicking. A platinum-rhodium thermocouple mounted i n one of the thermocouple wells was located near the middle of each of the heaters, the leads coming out of the apparatus at the bottom of the saturator through ceramic tubes ( L ) . The heater for the branch (X) was wound outside the thermocouple wells. I t began immediately above the top saturator heater and extended 5 cm. beyond the branch i t s e l f . A second heater, on the same diameter as the branch, extended from t h i s point for 5 cm. up the thermocouple w e l l s . Heat input was made approximately uniform at the branch point i t s e l f by means of a u x i l i a r y heaters on each side. A thermo-couple w e l l placed against the glass extended from 1 cm. below - 20 the branch, heater past the branch point i t s e l f and out to the top of the apparatus. A thermocouple i n ceramic tubing was buried i n the asbestos beside the branch point and could be moved back and f o r t h to investigate temperature v a r i a -t i o n s . The heater (Y) for the c a p i l l a r y tube that led to the manometer from the branch extended several cm. beyond the end of the c a p i l l a r y . A thermocouple was placed at about the middle of the c a p i l l a r y and the leads were brought out at the end of the heater. The heater was wound on asbestos tape and had asbestos cement i n s u l a t i o n . The flow tube (4) extended about 45 cm. h o r i -zontally from the branch and then turned downward for a further 15 cm. to the magnetic valves. I t s heater (Y) was wound i n two sections which were balanced to give equal heating e f f e c t , and the middle 20 cm. had an extra wrapping of glass wool to prevent cold spots. A platinum-rhodium thermocouple was mounted about 3 cm. from the branch end and two iron-constantan ones were spaced along the horizontal section. The thermocouple leads ran along the glass for about 10 cm. and then came out between turns of the heater. 4. Magnetic Valves and Mercury Traps The magnetic valves and the traps are shown i n Plates VII and IV. The flow tuba (A) (Plate V I I ) , a f t e r turning downward, terminated i n an expanded portion from P L A T E VI MIXER, SATURATOR AND BRANCH PLATE VI Mixer, Saturator and Branch Main Diagram *- Thermocouple A. I n l e t tube B. Evaporator C. Mercury sample D» Heater element E. Mercury j e t P. Return tube G. Pre-saturator H. Copper shielding disc I . Glass beads J . Top of pre-saturator K, Glass thermocouple w e l l L, Ceramic thermocouple insulators M. Heaters and i n s u l a t i o n U. Ceramic thermocouple wells 0. Branch P. Manometer c a p i l l a r y Q,. Plow tuba Cross Sections Asbestos cement Asbestos tape — — Asbestos wicking R. Cross-section of i n l e t tube heater S. Pre-saturator heater T. Cross-section of saturator furnace U. Copper sleeve V. Resistance thermometer W. Thermocouple wells X. Cross-section of branch heater Y, Cross-section of flow tube and manometer heaters - 21 -which, four tubes extended sideways to form a cross and then downward. About 2*5 cm* from the cross, each tube was joined to a 2 cm. piece of 2 mm. c a p i l l a r y tubing (B). The seat (0) so formed was ground to f i t a 1/4* stainless s t e e l b a l l (D) (of the type used for bearings) and three of the ground seats had s t e e l b a l l s i n them. The seal formed by these b a l l s on the ground seat was excellent, a pressure difference of 250 mm. of Hg. being required to drive gas through from the top side at a rate of 2 cc./sec. Since the greatest pressure drop they encountered i n operation was always l e s s than 1 mm. and there was a stopcock closed i n series with them most of the time,' they formed an e f f e c t i v e block for the apparatus end of the traps before the traps were sealed o f f . Moving the b a l l s from one tube to another was quickly accomplished by use of a small electromagnet. The tube that was joined to the bottom of each c a p i l l a r y l e d d i r e c t l y downward to the f i r s t of a pair of traps (F) as shown i n Plate VTI. The enlarged portion of the f i r s t trap was packed loosely with pyrex wool (Qj to ensure that any mercury vapour that might have condensed and frozen i n the argon instead of on the walls would not be entrained with the argon r i g h t through the trapping system. Seal-off constrictions (E) were located as shown i n the diagram and aft e r a run was completed the trapped mercury could be isolat e d from the system very quickly with a hand torch. There were four sets of traps, one of which was required for P L AT E VI THE. MAGNETIC VALVES AND THE TRAPS HATE V I I . The Magnetic Valve a and the Traps A. Flow tube B. C a p i l l a r y C. Seat D. Steel B a l l E. Seal-off c o n s t r i c t i o n s F. Traps G. H e l i c a l heater elements H. Small heater below constrictions I . " Trap tubes J . Aabestoa paper K. Asbestos cement L. Wick packing M. Support H. Convection Screen 0. Ceramic thermocouple tubes P. Spring q,. Glass wool - 22 -balance-out purposes. I t was, therefore, possible to take three measurements before putting on new ones. The heater around the valves was b u i l t i n two half sections for quick removal. Each of the two sections was heated by s i x equally spaced h e l i c a l c o i l s (G) which ran lengthwise and were connected i n series. These c o i l s were near the inner side of a thick layer of asbestos cement (K) and further i n s u l a t i o n was provided by a layer of asbestos wick packing ( L ) • This heater extended 10 cm. past the f i r s t set of seal-off c o n s t r i c t i o n s . Convection currents were prevented by a disc of asbestos paper (N) which f i t t e d the flow tube and rested on the heater. A small C y l i n d r i c a l heater (H) was mounted at the bottom of the furnace i n the space between the four trap tubes. The ceramic in s u l a t i n g tubes (0) f o r the valve thermocouples (Iron-const.) were inserted through holes i n the asbestos convection d i s c . 5. The Flowmeter A c a p i l l a r y flowmeter was used because i t would be nominally independent of pressure at flows where turbulence was not a fa c t o r . In order that i t might be maintained at a constant temperature, i t was b u i l t i n the form shown i n P i g . 2. and was kept i n an ice bath while runs were being taken. A manometer f i l l e d with d i f f u s i o n pump o i l indicated the pressure drop, whichwas usually 8 to 12 cm. of o i l . 23 -A trap immediately before the v flowmeter ensured that no mercury would condense i n i t . tec The flowmeter was calibrated by passing argon through i t at a constant r a t e , pressure and rate being adjusted by needle valves. The quantity of argon was measured by water displacement i n a FIG. 2 large b o t t l e whose volume was accurately known. Flowmeter c a l i b r a t i o n s were reproducible to about 0.5J2. However, on several occasions i t was necessary to flame out the flowmeter to remove mercury that had con-densed there when the main apparatus was bypassed. Each time th i s was done the flowmeter had to be r e c a l i b r a t e d , as the constant changed by one or two percent. Between flamings the constant was stable. 6. Pressure Measurement and Control Measurement of the t o t a l pressure i n the apparatus was made by means of a mercury manometer ( L ) (Plate Y.) which connected to the high pressure side of the flowmeter. This manometer had a 10 mm. tube on one side and a 45 mm. c y l i n d r i c a l bulb on the other. The manometer was c a l i b r a t e d against a standard barometer, and had a c a l i b r a t i o n equation 24 -P = (R p - R 0) 1.0509 where Rp i s the reading at pressure P, and R 0 i s the reading at P • 0. The difference i n pressure between the flowmeter and branch point was measured by a II-tube manometer (P) which was f i l l e d with d i f f u s i o n pump o i l ( S i l i c o n e ) . The o i l manometer for the flowmeter had the same form as the t o t a l pressure manometer. No c a l i b r a t i o n of t h i s manometer was required as the flowmeter was calibrated as a u n i t . I t was o r i g i n a l l y thought that pressure v a r i a t i o n within the system as a r e s u l t of temperature changes would make some form of regulation necessary. Por t h i s purpose, h e l i c a l heater elements were i n s t a l l e d down the centre of each of the 2 - l i t r e b a l l a s t bulbs (N). However, the pressure proved to be much more stable than anticipated and the heaters were never used. The l e v e l of the l i q u i d oxygen about the various traps seemed to be the only factor affecting the pressure and i f t h i s l e v e l were maintained, pressure was quite constant. 7. C i r c u l a t i o n of the Argon The gas was c i r c u l a t e d through the system by a mercury vapour Venturi pump. The general form of the pumping section i s shown i n P i g . 3. throat user - 25 -The nozzle was s l i g h t l y divergent with a minimum diameter of approx-imately 1*5 mm. and the di f f u s e r throat was 3 mm. i n diameter with a taper of about 20% on either side. This pump proved capable of d e l i v e r i n g up to at l e a s t 1 l/min. at a l l pressures used, against a head of 1.5 cm. of Hg., and with a b o i l e r pressure of only 4 to 6 cm. The b o i l e r consisted of two horizontal glass tubes 3.5 cm. i n diameter and 35 cm. long, connected at the vapour outlet and l i q u i d return l i n e . The mercury vapour-l i n e was heated throughout i t s length to prevent droplets of l i q u i d from forming and i n t e r f e r i n g with the smooth operation of the j e t . A constant pressure drop across the apparatus, approximately 1 cm. of Hg. was maintained by a regulator system made up of three b a l l check valves and two c a p i l l a r i e s as shown i n F i g . 4. FIG. 3 V«ntun Pump Apparatus FIG. 4 - 26 The b a l l checks rested on ground seats, and the seats, b a l l sizes and c a p i l l a r i e s were adjusted to give a pressure drop at the low pressure end of approximately 1 cm. of Hg. at a flow of 0.6 1/min. Small adjustments to change the flow were made with electromagnets mounted close to the b a l l s . The c i r c u l a t i n g pump and f i l t e r system gave s a t i s -factory operation but were not quite stable enough to be l e f t without supervision. The change i n temperature required to change the b o i l e r pressure by 0.5 cm. i s only 0.5 degrees, and l i n e voltage changes which occurred from time to time were s u f f i c i e n t to do t h i s . However, attention every 5 to 10 minutes was a l l that was required. The s t e e l b a l l s introduced a s l i g h t r i p p l e on the flowmeter reading, but readings to 0.25$ were quite e a s i l y made. The Venturi pump gave quite stable operation by i t s e l f , and,on occasion, was used at lower flows without the regulator. 8. P u r i f i c a t i o n The f i r s t step i n the p u r i f i c a t i o n process (11) was to pass the mercury sample twice through a pinhole i n a f i l t e r paper. This removed gross contamination. The mercury was than passed through a 150 cm. scrubbing tower containing ether to remove o i l s and greases, and then through a 10% n i t r i c a c i d scrubbing tower. Following t h i s treatment, the base metals in the mercury were oxidized by bubbling f i l t e r e d a i r through i t for 48 hours, the temperature being maintained - 27 -around 10Q°C. After oxidation the slag was removed by f i l t e r i n g twice. The mercury was then passed through a scrubbing tower of sodium hydroxide and two more of n i t r i c a cid. F i n a l l y i t was placed i n a two-stage vacuum s t i l l , d i s t i l l e d twice, and then introduced into the apparatus by two more stages of d i s t i l l a t i o n . The mercury obtained by t h i s process has been used i n several arc lamps i n the Physics Department and shows no detectable impurity l i n e s . Also, when shaken with d i s t i l l e d water, i t forms a foam that l a s t s for 10 to 15 seconds (11). The argon was p u r i f i e d (12) from the commercial grade (99.7%) which was obtained i n large cylinders. I t was f i r s t c i r c u l a t e d for some time through a l i q u i d oxygen trap to remove a l l traces of water vapour and then c i r c u l a t e d f or four to f i v e hours past a magnesium arc operating i n a l ^ l i t r e bulb. A l l pieces of the apparatus were c a r e f u l l y cleaned before being put on, and wherever possible, were heated on hard vacuum to drive off impurities and residual gas. A two-stage mercury d i f f u s i o n pump was used for evacuating and,with suitable trapping, the pressure i n the whole system could be brought down below 2 x 10" 5 mm. (the lower l i m i t of the P h i l l i p s Gauge) quite e a s i l y . Before f i n a l f i l l i n g with argon, the whole system was flushed out several times, evacuated, and then f i l l e d . 28 9. Heaters and Control of Temperature A l l heaters, except those on the saturator, were operated on alternating current. The heaters for each zone (the mixer u n i t , the branch, the flow tube, and the valves) were controlled by a common variac transformer. Individual series r e s i s t o r s made i t possible to vary each heater independently. An A.C. voltmeter was connected to each variac. The saturator heaters were powered from the D.C. mains through a combined voltage regulator and temperature control c i r c u i t . The D.C. l i n e v a r i a t i o n could be s u f f i c i e n t to cause considerable fluctuations i n temperature even with a temperature controlled c i r c u i t , i f the control mechanism were continually being thrown out of balance by the l i n e changes. For t h i s reason, i t was f e l t necessary to incorporate a voltage regulator i n the temperature control c i r c u i t . The c i r c u i t developed for the purpose i s shown i n P l a t e V I I I . The resistance thermometer element (X) previously described (50 ohms, of hytemco wire, wound non-inductively and with manganin leads) was i n series with a 70 ohm, non-inductive, manganin r e s i s t o r (R) to form one side of the A.C. bridge. The other side of the bridge was made up of two 5 ohm r e s i s t o r s ( S i , S2) with a 10 ohm, 10-turn helipot (S) between them. Current for the bridge was supplied from the A.C. mains through a 40-watt l i g h t bulb ( L ) . Adjustment of the bridge setting was made by moving the centre tap of the - 29 -h e l i p o t , while vernier control was provided by a second potentiometer (V) and series r e s i s t o r . The bridge signal was fed into a microphone transformer with the secondary connected to the g r i d of the f i r s t tube. Phase discrimina-ti o n was accomplished by feeding the amplified bridge s i g n a l to the g r i d of a cathode follower whose plate was connected to the A.C. l i n e and was, therefore, i n jphase with the bridge input. The D.C. signal from the cathode follower was super-imposed on the setting of a simple D.C. regulator c i r c u i t having a regulation factor of about 70 and a 2 amp. current capacity. The • 250 Y. and -350 V. for the temperature con-t r o l s came from regulated power supplies. These were of standard design with screen g r i d r i p p l e control and a 6AS7 power tube. They were quite capable of supplying some 15 regulator c i r c u i t s as the c i r c u i t s draw very l i t t l e current. Operation of the regulators was quite simple. The voltage settings were turned to maximum and the thermometer setting to some high value u n t i l the apparatus was close to the desired temperature. The resistance thermometer se t t i n g was now turned down u n t i l the power just shut o f f . When reasonably stable temperatures were obtained the thermometer s e n s i t i v i t y was turned to minimum and the voltage regulator set to about the required voltage. Once th i s was done the s e n s i t i v i t y control was turned to maximum and a l l further adjustments were made with the coarse and f i n e thermometer P L A T E V I I I THE. TE3S2FERATUEE CONTROL CIRCUIT ELATE V I I I The Temperature Control C i r c u i t X Resistance Thermometer R 70 ohm manganin r e s i s t o r S2 5 ohm f i x e d r e s i s t o r s S 10 ohm, 10-turn h e l i p o t B 40-watt l i g h t bulb Y Yernier control potentiometer L Load (furnace heater) - 30 -settings. For less accurate temperature control the voltage setting was l e f t at maximum. This temperature control c i r c u i t exceeded a l l expectations i n i t s operation. Freed from a l l variations due to l i n e voltage changes, the temperature control part of the c i r c u i t s maintained temperatures i n the saturator to within t 0.02 deg. i n spite of considerable changes i n heater inputs i n nearby c i r c u i t s . After equilibrium had been reached, the variations i n temperature could not be measured with any certainty because they were too close to the funda-mental l i m i t s of s e n s i t i v i t y of the c i r c u i t , but they were cert a i n l y less than i 0.01 deg., and readings corresponding to * 0.0025 deg. were frequently maintained for some hours. Temperature control i n the remainder of the apparatus was a r e l a t i v e l y simple operation once the e q u i l i b -rium temperatures were reached. Temperatures were not at a l l c r i t i c a l , and aside from occasional small changes i n the branch and b o i l e r power input, the heaters needed no attention. The controls, switches and meters for the apparatus were mounted on a panel that extended across the end of the apparatus frame (Plate I I . ) . The regulator c i r c u i t s and power supplies were placed on a table at the other end of the thermometry table to keep them away from the magnetic f i e l d s of the variacs. The operator was only required to 31 -move from h i s position at the controls to turn stopcocks, move the magnetic valves, or add oxygen to the vacuum f l a s k s . 10. Thermal Shielding A curved sheet of polished brass the f u l l height of the saturator shielded i t against r a d i a t i o n from the mercury b o i l e r and the mercury supply l i n e . A second sheet was hung between the valve heater and the saturator. This s h i e l d curved toward the saturator at the top to block r a d i a t i o n from the flow tube as w e l l . 11. Temperature Measurement Thermocouples ware used for a l l temperature measure-ment. C i r c u i t diagrams for the thermocouple c i r c u i t s are given, i n Plates IX and X. In those places where accurate knowledge of the temperature relationships was required, platinum vs. platinum-10$ rhodium thermocouples were used while iron-constantan was used i n the less important measure-ments. Readings of the platinum-rhodium couples (T) (Plate IX) were taken with a White double potentiometer (P) and an L.. & N. Type S galvanometer (G) which had a s e n s i t i v i t y of 0.02 ^v./mm. with the main thermocouple (M). One section of the potentiometer was reserved for t h i s thermocouple which was located i n the glass we l l down the middle of the saturator. A Rubicon student potentiometer (P) (Plate X.) and Type S galvanometer (G) of lower s e n s i t i v i t y were used to read the 32 -iron-constantan thermocouplea (TC). Thermocouple sel e c t i o n was accomplished hy means of three s p e c i a l l y constructed thermocouple switches (S, ST, D) designed by the Author and b u i l t i n the shops of the National Research Council, Ottawa. These switches had copper leads and contacts throughout to minimize intrusive e.m.f'a. and had shielded cases. The main switch i n the iron-constantan c i r c u i t combined thermocouple selection with tapping key action, a feature that greatly increased the speed of operation. The c i r c u i t ends of a l l but the main thermocouple were taken d i r e c t l y from the apparatus to a large brass zone block (2) with a double row of copper connecting plates along the top. Here the s o l i d copper wires (C) of the connecting c i r c u i t s were joined to the thermocouple wires. The platinum-rhodium and iron-cons tan tan cold junctions (CJ) were connected i n the same way. The c i r c u i t ends of the main thermocouple were carr i e d d i r e c t l y to an ic e bath where they were connected to copper leads which went d i r e c t l y to the White potentio-meter. The connection was made i n the ice bath by thrusting the lead wire and the thermocouple wire into a small pool of mercury at the bottom of a closed-end glass tube (B). A l l platinum-rhodium thermocouples were annealed to remove s t r a i n s , following the procedure recommended by the U.S. Bureau of Standards ( l 3 . a ) . The c i r c u i t of the platinum-rhodium thermocouples - 33 (Plate IX.) requires l i t t l e explanation. The leads of the main thermocouple (A) went d i r e c t l y to the nQn c i r c u i t of the potentiometer, while the leads from the other platinum-rhodium couples (TC) and cold junction (CJ) went to the thermocouple switch, which was connected to the "P11 c i r c u i t of the potentiometer. This make of potentiometer has a pro-v i s i o n for switching the battery current into a bypass c i r c u i t i n order that intrusive e.m.f'a. i n the potentiometer, g a l -vanometer leads and galvanometer may be compensated. The compensating c i r c u i t (14) used i s shown at (E). A small current passing through a few turns of f i n e copper wire (R) i n series with the galvanometer, was adjusted u n t i l the potential drop across i t exactly cancelled out the i n t r u s i v e e.m.f' s. of the c i r c u i t . Balance was indicated by zero d e f l e c t i o n of the galvanometer when the tapping key was de-pressed. In practice, t h i s device was hardly necessary as the intruaives i n the c i r c u i t were r a r e l y as large as 0.1 uv. The potentiometer has four d i a l s , and reads to the nearest microvolt, the f i n a l figure (0.1 uv.) being read by deflec-t i o n . The iron-constantan c i r c u i t i s somewhat d i f f e r e n t (Plate X.). The thermocouples (TC) were connected i n two groups, the more important ones being wired d i r e c t l y to the main switch (ST), while each of the others was connected to a thermocouple of the f i r s t group on one side and to the P L A T E THE PLATINUM VS. PLATINUM-10$ RHODIUM THERMOCOUPLE CIRCUIT PLATE IX. The Platinum vs. Platinum-10$ Rhodium Thermocouple C i r c u i t TC Thermocouple P White potentiometer G Galvanometer M Meter Z Zone block C S o l i d copper leads CJ Cold junction B Closed-end glass tube E Compensating c i r c u i t R Resistor of copper wire A Main thermocouple S Thermocouple switch - 34 d i f f e r e n t i a l switch (D) on the other. I t w i l l he noted that t h i s switch was wired i n such a way that when i t was i n i t s f i r s t p o s i t i o n the potentiometer (P) and cold junction (CJ) were connected i n the usual manner through the main switch to the f i r s t group of thermocouples. When the d i f f e r e n t i a l switch was i n any other than the f i r s t p o s i t i o n , the poten-tiometer and cold junction were no longer i n the c i r c u i t and one of tiie second group of thermocouples was substituted i n their stead. When the appropriate tapping key was depressed on the main switch, the galvanometer (G) recorded the difference i n e.m.f. between the two thermocouples. No com-pensating e.m.f • was included i n t h i s c i r c u i t , as the accuracy required did not warrant i t . An external standard c e l l (SC) c i r c u i t was added to eliminate changes i n the setting of the potentiometer. The lo c a t i o n and type of thermocouples connected to the various switches was as follows: 4 potentiometer 0.010 i n . , P t . vs. Pt.- 10$ Rh. The main thermocouple i n the glass w e l l inside the saturator -P potentiometer 0.005 i n . , Pt. vs. P t . - 10$ Rh. Si x thermocouples -(1) on pre-saturator, (2) , (3) and C4) under the windings of the saturator, (5) i n a ceramic w e l l under the branch heater, 1 cm. P L A T E X THE mON-COHSTANTAN THERMOCOUPLE CIRCUIT ELATE X The Iron-constantan Thermocouple C i r c u i t TC Thermocouple ST Tapping key selector switch D D i f f e r e n t i a l switch ZL Zone block C S o l i d copper leads GCT Cold junction E Rubicon potentiometer G Galvanometer SC. Standard c e l l M Meter PLATE X 35 above the top of the top saturator heater, (6) on the flow tube, 8 cm. from the branch. Rubicon - Main switch. #30 (B. & S.} Iron-constantan. S i x thermocouples -(1) i n the ceramic w e l l alongside the top saturator, (2) i n the inner ceramic w e l l alongside the branch, 3 cm. from the top saturator, (3} on the i n l e t tube, (4) on the tube leading from evaporator to mixer, (5) on the flow tube, 5 cm. before the bend, (6) at the top of the valve heater, alongside the magnetic valves. D i f f e r e n t i a l switch. #30 (B. & S.} Iron-constantan. S i x thermocouples and the student potentiometer -(1) Rubicon potentiometer and the cold junction, (2) the pre-saturator, compared with saturator on main switch, (3) on the side of branch under the a u x i l i a r y heater, compared with saturator, (4) on the manometer c a p i l l a r y , compared with the branch, (5) alongside the top part of the glass thermocouple w e l l , .compared with the branch, (6) at the middle of the flow tube, compared with other flow tube thermocouple, (7) inside the valve heater just below seal-off con-36 s t r i c t i o n s , compared with top valve thermocouple. Wherever possible, provision, was made for i n v e s t i -gating the d i s t r i b u t i o n of temperature by s l i d i n g the thermo-couples back and f o r t h . The moving thermocouples were mounted i n 2 mm., 2-hole, ceramic tubing which s l i d e a s i l y inside the thermocouple w e l l s . By this means the temperature d i s t r i b u -t i o n was measured for a l l sides of the saturator and proved to be very uniform. I t seems impossible that any cold spot could have been present i n the apparatus as i t was operated. During e a r l i e r tests of the apparatus, a cold spot, that made the readings uniformly low by about 3$, was detected. Considerable e f f o r t was expended to ensure that the platinum-rhodium c i r c u i t would be free from in t r u s i v e e.m.f's. of a l l kinds. Thermal e.m.f's. i n the potentiometer and galvanometer c i r c u i t s were compensated for by a separate c i r c u i t (already described). In addition, a sh i e l d i n g c i r c u i t (14) (I3.b) was provided which prevented leakage currents from the heating elements or from power leads on the table from entering the thermocouple c i r c u i t . The copper sleeves about the saturator, a copper disc j u s t above the pre-saturator heater, and the whole framework were a l l grounded. The zone block and leads to the switches rested on a grounded sheet of brass and the whole top of the table on which the thermometry instruments and panel rested was covered with a grounded brass plate. The panel on which the switches were mounted was thus completely isolated from the power c i r c u i t s . The - 37 galvanometer lamps and switches were mounted on t h i s panel on grounded plates. Checks of t o t a l intrusives for the whole or parts of the c i r c u i t were conducted from time to time, and were never as much as 0.1 jxv., with the exception of the galvanometer s e n s i t i v i t y switch of the potentiometer, which generated temporary thermals of about 1 uv. i f worked r a p i d l y . These died away quite soon and became ne g l i g i b l e before balance was obtained. The main thermocouple was calibrated very c a r e f u l l y at the ice and steam (15) points, using the procedure recom-mended by the U.S. Bureau of Standards. I t read exactly 0.0 ^iv . at the ice point and deviated from the equation on which the standard tables are based (16) by • 0.6jx*. at 100°C. (a very good agreement). A lin e a r correction equation 6 tables s e meas. ^ "* °»0°0933) was used and temperatures then calculated from the standard equation. With t h i s correc-t i o n , the measured temperatures should be accurate to some-what better than 0.1 deg. up to 150°C, and 0.2 deg. to the highest temperatures used. This c a l i b r a t i o n was checked several times and remained exactly the same throughout the course of the measurements. 12. Procedure of Measurement The procedure i n taking readings was as follows: The heaters, including the c i r c u l a t i n g pump b o i l e r s , but not the evaporator, were turned on and the apparatus brought up 38 -to the desired temperature as quickly as possible. While this was being done, the valves, flow tube, branch and manometer c a p i l l a r y were kept hotter than the saturator so that mercury would not diffuse into them and condense. The regulator controls were then set i n such a way that they gave approximately the r i g h t power with zero signal on the bridge, and then the bridge settings were adjusted to give a s l i g h t negative gradient from bottom to top of the satura-tor. This gradient did not appear to be c r i t i c a l but was kept less than 0.05 deg./cm. for the top two saturator sections. While t h i s f i n e adjustment was being made, the valves, branch, etc., were maintained at a f a i r l y high temperature. As soon as the saturator temperatures were set, the gas was started through the apparatus from traps to saturator, and the traps required for the measurement were flamed to clear them of mercury. When s u f f i c i e n t time had elapsed to ensure that a l l tubes back to the saturator were c l e a r , the flow was stopped. The temperatures of the flow tube and branch were then dropped to that required for the measurement and the heater for the evaporator was turned on. Temperatures of manometer and flow tube were not at a l l c r i t i c a l and were kept at a point which would prevent condensation without disturbing the r a d i a l symmetry of the saturator. Prom 5 to 20 deg. above the temperature of the saturator was quite 39 sat i s f a c t o r y although, the flow tube was r a r e l y run as cold as 5 deg. for fear of condensation further along. The branch temperature was adjusted u n t i l the main thermocouple, and the thermocouples i n the ceramic wells just outside the satura-t o r , indicated a long f l a t minimum of temperature near the top of the beads. In practice i t was found that t h i s condi-t i o n would be obtained i f a point i n any one of the outer wells about 1 cm. above the top of the top saturator heater were kept 0.3 to 2.0 deg. hotter than the saturator. A posit i v e gradient was maintained from there to the branch point i t s e l f , the excess temperature at the branch being between 10 and 25 deg. Once this setting had been made i t ra r e l y needed readjustment. The pre-saturator and mixer were kept hotter than the saturator but were given no p a r t i c u l a r care and the excess temperature ranged from 3 to 50 deg. without apparent e f f e c t . The evaporator heater was set any-where between 55 and 75 volts over which range the mercury flow varied from an occasional bubble to a f a i r l y v i o l e n t b o i l which maintained an excess pressure i n the evaporator of several cm. of Hg* As soon as the above conditions were obtained, the traps to be used were immersed i n l i q u i d oxygen and the balance run was begun. While the balance run was on, a f i n a l check of temperatures and rate of flow was made. This run lasted for approximately the same time as the actual measure-- 40 -ment. The*actual run was started by opening the valve heater, quickly moving the b a l l check from the tube that was to be used for the measurement to the balance trap and closing the heater again. Frequent readings of flow, pressure and temperature were made during the run which was terminated i n exactly the same way that i t was started. A calibrated stop-watch was used to measure the time between seatings of the valve. As soon as the run was over, the heater was opened again and a gas-oxygen torch used to seal off the f i r s t con-s t r i c t i o n . The valve heater was then replaced and the remaining constrictions sealed, allowing the temperature i n each trap to r i s e to about 0°C. before sealing i t . The entire procedure for the run was now repeated at new temperatures, using the two remaining pairs of traps. When the apparatus was to be cooled down, the evaporator, mixer and saturator heaters were turned off f i r s t and the other heaters were l e f t running f or some time to ensure that condensation did not occur i n the branch and flow tube. After the apparatus had cooled, the traps were removed for t i t r a t i o n . 13. Change Over Change over to new traps was quite simply made. Argon was admitted to the saturator, flow tubes and valves (through the vacuum li n e ) u n t i l the pressure was several cm. - 41 -above atmospheric. With t h i s pressure maintained, the glass tubes between the valve c a p i l l a r y and the f i r s t c o n s t r i c t i o n were cut and the f i r s t traps (with cross tubes to next trap attached} were blown on. The second traps were then attached using the needle valve i n l e t (Plate Y.} for blowing purposes. The saturator was then is o l a t e d from the vacuum system and the traps evacuated and tested for leaks with a Tesla c o i l . The b a l l valves were quite s u f f i c i e n t to hold back the gas i n the saturator while t h i s test was being made. Once gas tightness was assured the saturator was evacuated and closed off again while the traps were flamed vigorously to drive off adsorbed water. Flaming was continued u n t i l a pressure of 0.1 to 0.2 microns could be maintained with the traps hot. After the whole system had been pumped hard a l l night, the pure gas was l e t i n and a new run begun. 14. T i t r a t i o n After the traps were removed from the apparatus they were opened and the mercury dissolved out with hot 50$ n i t r i c a c i d . The f i r s t trap was opened at the top end (nearest the valves}, and the acid was poured i n u n t i l the tube was almost f u l l . (The acid did not enter the other half of the tube because of the a i r locked in.} After s u f f i c i e n t time for the mercury to be dissolved, the other end of the trap was removed and the a c i d allowed to r i s e into the glass wool. The wool was thoroughly wetted and used to scrub - 42 the sides of the trap before being pulled out into a small beaker. The n i t r i c acid-mercury s o l u t i o n , and two r i n s i n g s of b% n i t r i c a c i d , were also poured into the beaker. The second trap was opened at the large end and rinsed with a small amount of hot 5Q/£ n i t r i c acid followed by two rinses with 5$ n i t r i c acid. Some pairs of traps were t i t r a t e d separately to check the e f f i c i e n c y of the f i r s t trap, but i n most cases the whole of the trap contents was t i t r a t e d (1?) together. The solution was prepared for t i t r a t i o n by b o i l i n g off the oxides of nitrogen, adding water to reduce the n i t r i c a c i d concentration, and cooling to about 15°C. After cooling, 1 ml. of saturated f e r r i c alum s o l u t i o n per 100 ml. of solu-t i o n was added. The s o l u t i o n was t i t r a t e d slowly and with good a g i t a t i o n , with 0,1 I , or 0.025 H. potassium t h i o -cyanate, to a permanent reddish-brown color. The thiocyanate s o l u t i o n was standardized against a standard 0.1 N. mercury s o l u t i o n prepared by d i s s o l v i n g a weighed (10.030 g. approx.) quantity of pure mercury and d i l u t i n g to exactly one l i t r e i n a volumetric f l a s k . Standardization was quite exact and could be reproduced to 0*1%, including the making up of fresh standard mercury solu-t i o n . T i t r a t i o n s of large amounts of mercury presented no problems, provided that the t i t r a t i o n was done below 15°C. Above 15°C. there i s a tendency for a premature end point owing to the d i s s o c i a t i o n of mercuric thiocyanate. Color - 43 -from t h i s cause disappears when the solution i s cooled and the t i t r a t i o n ean then he completed. Recognition of the end point i s very sharp and requires no correction. With smaller amounts of mercury and the 0.025 H. solution, a tendency to a premature end point, at low temperatures, was discovered. The end point appeared hut faded very slowly, and appreciable quantities of thiocyanate sometimes had to be added before the permanent and point appeared. (In. the case of s i l v e r , adsorption of s i l v e r ion by the precipitated Ag CHS causes a simil a r a f f e c t (17.b), but i t i s not known i f the same type of reaction applies here.} The concentration of n i t r i c a c i d did not appear to aff e c t the r e s u l t as long as temperatures were kept low but at higher temperatures there was a very gradual fading which i s a t t r i b u t e d to a side reaction i n which the a c i d takes part. High concentrations of n i t r i c a c i d had a very d e f i n i t e retarding e f f e c t on the rate at which the thiocyanate and mercury combined, since the increased UO3 concentration presumably reduced the mercuric i o n concentration. P r e c i p i t a t i o n of mercuric thiocyanate usually occurred near the end point but t h i s d i d not i n t e r f e r e appreciably with the t i t r a t i o n . In those cases where pre-c i p i t a t i o n did not occur, the premature end point was much more i n evidence and more care was required. The premature color disappeared in" a few seconds when p r e c i p i t a t i o n started. 44 -The equations for various reactions involved i n the t i t r a t i o n are given below: (1) Solution of Hg. i n H¥0 3 -Hg 4 4HM>3 —f Hg(&0 3) 2 • 2H20 4 2N0 2 (15) (2) T i t r a t i o n , before the end point -Hg(E0 3} 2 4 2KCHS » 2KH03 4 Hg(CMS)2 ( l 6 l (3) The i o n i c reaction giving the color change -Ve*** 4 6CNS~ FelCNSjJ (17) (The FeCCHSjg ion gives the color indication.) (4) The following ionic r e action along with (17) gives the premature end point -Hg«mS) 2 Hg + + 4 2 cus"" (IS) 15. Experimental Error The thermocouple c a l i b r a t i o n w i l l not be i n error by more than 1 0.1°C. below 15Q°C. or more- than t 0.2°C. at 240°C. This corresponds to an error in-the measurement of less than 0.5$. The d i s t r i b u t i o n of temperature was inves-tigated using the same thermocouple for a l l measurements and the temperatures i n the glass w e l l corresponded to those i n the outer wells to better than 0.05 deg. Gradients were such that the maximum error from t h i s cause could not be greater than that amount. Errors due to i n s t a b i l i t y of temperature were completely n e g l i g i b l e . Search was made for hot or cold spots and no i n d i c a t i o n of either was found. - 45 -The t o t a l error from thermometry should be less than. 0.8$. I t i s d i f f i c u l t to estimate the magnitude of errors a r i s i n g from the degree of saturation, but testa were per-formed that indicate that they must be very small. The gradient i n the saturator was varied by several hundred per cent, and a l l temperatures i n the lower part of the apparatus were varied over wide ranges (the evaporator even being l e f t completely unheated}, without any detectable difference i n r e s u l t s . A 1QQ$ change i n the rate of flow made no d i f -ference, but the smaller flow of the test (0.5 1/min.) was used as a maximum for other reasons. In view of the tremen-dous surface exposed and the very small temperature gradient i n the saturator, i t seems c e r t a i n that saturation errors were l e s s than 0.3$ and were probably random. At these pressures, droplet size (19.a} (20) and t o t a l pressure (19.b) of the gas would not have had an appreciable e f f e c t on the r e s u l t . Possible losses of mercury down the manometer c a p i l l a r y have been calculated (lO.b) and were n e g l i g i b l e at the pressures used. Estimates baaed on rates of leakage under pressure indicate that i f a l l the outlet stopcocks were l e f t open, the losses by leakage past the valves would be about 0.1$, and i n practice they were c e r t a i n l y n e g l i g i b l e . Losses of mercury by passing through the traps were checked by t i t r a t i n g the traps separately. The second trap never contained as much as 1$ of the t o t a l mercury and losses from - 46 -i t must have been extremely small. Losses a r i s i n g from f a i l u r e to get a l l mercury out of the traps were checked by pouring fresh acid i n the traps and t i t r a t i n g the washings. Hone of these second t i t r a t i o n s ever amounted to more than 0.1$ of the o r i g i n a l value. T i t r a t i o n tests and blanks were run with known quantities of mercury i n solution and d i f f e r e n t amounts of glass wool, n i t r i c a c id, i n d i c a t o r , etc. These indicated that the t i t r a t i o n error for a l l measurements above 70°C. was not more than 0.5$. Large accidental errors i n preparation for t i t r a t i o n would almost c e r t a i n l y be low. At lower pressures, there was greater uncertainty both i n removal of mercury from the traps and i n the t i t r a t i o n . However, aside from very lowest temperatures, the r e s u l t should not be more than 1$ high or 2$ low. The stopwatches used i n the time measurement were checked against WW and were correct to better than 0.03$. The uncertainty of s t a r t i n g and stopping was approximately 1 s e c , with the two operations cancelling one another to some extent. This caused, at most, a 0.1$ error. The flowmeter c a l i b r a t i o n was correct, to t 0.5$ and,when referred to the mid-point of the flowmeter c a p i l l a r y , was quite accurately independent of pressure and flow. Readings of the flowmeter during a run could quite e a s i l y be made to 0.25$. Total pressure readings of high accuracy were only required for the 47 c a l i b r a t i o n of the flowmeter, because of the form of the equations, and these measurements were ea s i l y made to 0.05$ under c a l i b r a t i n g conditions. The difference pressure manometer (Oil) required reading to the nearest millimeter to give 0.03$ accuracy to the pressure correction and t h i s was e a s i l y done. Pressure gradients due to d i f f u s i o n of mercury down the manometer c a p i l l a r y and to thermo-molecular pressure were not a factor at the t o t a l pressures of t h i s experiment. Deviation of. argon from the perfect gas laws has been calculated (l.b) and i s l e s s than 0.1$. Calculations based on the Yan der Waal's Constants (IS.a) for mercury indicate that about 0.1$ deviation should be expected at 5 cm., the highest pressure measured. A more important error was introduced at high pressures where mercury i s s l i g h t l y associated into Hgg (3). This would cause readings to be high. A further discussion of t h i s point i s given l a t e r . I t i s impossible to estimate the effect of con-tamination on the r e s u l t s . Any contamination would almost c e r t a i n l y have made the r e s u l t s low (21) (lO.c), p a r t i c u l a r l y at low pressures, but such effects were minimized by the condensation technique being employed. Great care was taken i n p u r i f i c a t i o n , cleaning and assembly, and i t seems u n l i k e l y that any metallic contamination (the chief danger) would get i n . Contamination that might get i n from the arc would be - 48 -i n oxide form and would only af f e c t the rate of evaporation. Because the t i t r a t i o n i s selective (except for s i l v e r (17.a)), other small vapour pressures have no e f f e c t . For temperatures above 70°C. the t o t a l of system-a t i c errors i s 0.8$ while the t o t a l of random errors i s 1.8$. For the region, below 70°C. the systematic error may be around 2$, random errors t o t a l about 3$, and the greatest tendency for both i s to be low. In the region above 70°C. the yST2" » 0.8$. random 16. Conversion to Other Substances The apparatus, as described, could be used to make measurements on any l i q u i d which has a vapour pressure within the range of the apparatus at temperatures below the soften-ing point of pyrex glass. I f necessary, iron-loaded glass valves could be used. Few metals can be handled i n pyrex, but conversion to quartz would permit measurements on a large number of interesting metals, at temperatures towards 1000°C. For metals other than mercury and for so l i d s generally, a d i f f e r e n t mixer section i s required. A suitable form would be composed of a heated U-tube coated on the i n -side with the metal to be measured. The metal on the surface of the tube could then be replenished between runs by evapora-t i o n i n vacuo from the bottom of the U. Changes required i n the saturator and traps are of very minor nature. - 49 RESULTS 1. Calculations (a) Calculation of Vapour Pressure The "basic formula on which the vapour pressure ca l c u l a t i o n i s based i s -P - P 8 — U9) M • m p s vapour pressure P a s t o t a l pressure i n the saturator m a moles of mercury carr i e d over M s moles of argon The formula i s based on the assumption that the gas laws apply and that the mercury and argon are monatomic. In t h i s experiment, P s was obtained from the formula -P a . P m * P d (20.a) P m st t o t a l pressure at high pressure side of flowmeter P a s difference manometer pressure reading m was determined from the t i t r a t i o n procedure by the equation - "*t x P t m = (20.b) 200.6 Y^ s volume of t i t r a t i n g s o l u t i o n P t = mercury equivalent of 1 ml. of solution 200.6 • Atomic weight of mercury while M was obtained from the c a l i b r a t i o n of the flowmeter 50 and the readings of flow, pressure and time, using the equation -Cf x Rf x B p x t Iff - -= = = (20.c) 760 x 22.415 C f - flowmeter constant Rf a flowmeter manometer reading Ef • pressure at mid-point of the capillary; t s time of the measurement c f x R f * rate of flow of gas (l/min.J, approx. 0.5 l/min. at 0°C. The constant of the flowmeter, C^, was obtained from the c a l i b r a t i o n data by using the equation -Gf » ( Y * £ a ) *LLl x ± . * J L (*>.d) \ T J J Pf' t + R f tank flowmeter x Y s volume of c a l i b r a t i n g tank P a s p a r t i a l pressure of argon i n the tank T * absolute temperature (b) Calculation of Temperature The correction to be applied to the reading of the thermocouple at the steam point was determined as follows; The b o i l i n g point was calculated from the barometric pressure and the b o i l i n g point tables (I8.b). The standard thermo-couple e.m.f. was then calculated from the equation -e = 5.4501S T • 10.9517 x 10~ 3 T2-11.535 x 10~ 6 T 3 T -°G. U l ) upon which the platinum-10$ rhodium tables are based. The correction required to bring the thermocouple to t h i s reading - 51 -was then found and an equation set up of the form -ex T . , = e m ( 1 4 c ) (22) tables T.G, ~S ' V ' T.C. In the case of the main thermocouple, t h i s equation was -_ s e m C 1 - 0.000933) (22.a) tables T.C. The temperature corresponding to any corrected reading i s determined by using equation (21) to calculate the e.m.f. for some temperature close to the measured one. The slope of the curve i s then used to calculate the temperature of the read-ing. For temperatures above 232°C., the following equation i s used e r - 35.01 4 6.10222 T 4 7.28131 x 10~ 5 T 2 - 5.0249 x 10~ 6 T 3 (23) The temperatures calculated from the thermocouple e.m.f !s. were converted to the absolute scale and the various values (1000/T, log T, etc.) required for the Kirchhoff vapour pressure equation -l o g p s A - B - c log T (1) T were tabulated. 2. Readings Before the f i n a l set of readings were taken, a 52 large number of preliminary runs were made to determine the range of operating conditions of the apparatus. The f i r s t set of readings was taken with, the branch and flow tube very hot and a large temperature gradient immediately above the beads. Another set of readings was taken with a small gradient and with no power i n the branch heater, a l l the heat for the top end being supplied by the a u x i l i a r y heaters about the branch, point i t s e l f and by conduction from the manometer and flow tube heaters. In the f i r s t case, the r e s u l t s were on the average 3$ high of the f i n a l readings and more widely d i s -t r i b u t e d . In the second case, the r e s u l t s were about 5$ low. On a f u l l i n v e s t i g a t i o n of the temperature d i s t r i b u t i o n , i t was discovered that the actual minimum, of temperature was located on the wal l of the tube leading from saturator to branch, about 4 cm. above the top of the beads. This condition was corrected by adding power to the branch heater u n t i l a thermocouple s l i d i n g up and down i n the ceramic wells outside the saturator showed a very f l a t minimum just below the top of the beads, and a slowly-increasing gradient of temperature from there upward (See procedure of measurement). Under such conditions, the temperature r i s e over the f i r s t 4 cm. above the beads never exceeded 0.4 deg. i n the outside wells, or 0.2 deg. i n the glass w e l l down the centre. No other changes i n conditions appeared to a f f e c t the r e s u l t s . TABLE I READINGS Reading No. t e .m.f. T(°C.) * I M M H O 2 1 13.66 1284.3 181.74 49.14 .23463 .4530 2 16.48 1199.2 171.42 48.27 .26683 .3629 3 60.48 619.3 96.66 47.08 .97378 .04821 4 59.57 715.8 109.79 37.19 .76933 .09471 5 54.16 817.6 123.29 38.28 .70579 .16613 6# 39.82 885.0 132.05 37.00 .50133 .17131 7 22.61 953.1 140.78 29.25 .22554 .15016 8 27.18 1016.9 148.85 28.93 .26824 .25732 9 21.25 1108.1 160.23 28.58 .20080 .31115 10# 201.83 302.7 50.60 54.13 3.6583 .009028 11 152.28 445.2 71.97 53.75 2.7048 .028115 12 122.45 503.1 80.34 39.73 1.6279 •038886 13 100.58 558.4 88.20 39.29 1.3141 .05027 14 35.73 885.9 132.17 39.43 .44090 •15052 15# 61.48 646.8 100.43 49.98 1.0119 .07303 16 19.50 1523.2 210.12 49.54 .30967 1.6051 17 19.21 1784.4 240.37 50.49 .30198 3.9475 18 184.47 362.3 59.71 37.95 2.3191 .01587 19 301.40 240.7 40.90 37.80 3.7538 •006866 20 590.33 184.7 31.88 37.74 7.3943 .006129 Run 6. The f i r s t trap was sealed off at too low a temperature and the f i n a l pressure was above atmospheric. When the trap was opened, a surge of escaping gas blew mercury up into the end being opened and some was l o s t . This reading represents a lower l i m i t rather than a true value. See Chart I I . TABLE I I RESULTS Ditchbum Reading No. T(°K) 1000 T ' m^m Dauphinee iOg Pmm & Gilmour log ?mm log P 1 454.88 2.1984 9.3062 .9638 .9749 -.0060 a 444.56 2.2494 6.477 .8114 .8138 -.0024 3 369.80 2.7042 2.3295 1.3673 1.3736 -.0063 4 382.93 2.6114 0.4573 1.6602 1.6685 -.0033 5 396.43 2.522S 0.8989 1.9537 1.9506 •.0031 6# 405.19 2.4630 1.2600 .1004 .1233 -.0229 7 413.92 2.4159 1.9347 .2866 .2882 -.0016 8 421.99 2.3697 2.749 .4392 •4343 •.0049 9 433.37 2.3075 4.361 .6396 • 6309 •.0087 io# 323.74 3.0888 0.01336 2.1257 2,1474 -.0217 11 345.11 2.8975 0.05586 2.7471 2.7583 -.0111 12 353.48 2.8289 0.09488 2.9772 2.9769 •.0002 13 361.34 2.7674 0.15025 1.1768 1.1726 •.0042 14 405.31 2.4672 1.3415 .1276 .1255 •.0021 15# 373.57 2.6768 .3607 1.5571 1.4608 •.0963 16 483.26 2.0694 24.412 1.3876 1.3806 •.0070 17 513.51 1.9475 58.370 1.7662 1.7623 •.0033 18 332.85 3.0044 .02597 2.4144 2.4172 -.0028 19 314.04 3.1843 .006914 3.8397 3.8422 -.0025 20 305.02 3.2785 .003128 3.4953 3.5407 -.0455 Run 10. This was one of the f i r s t t i t r a t i o n s with 0.025 N. thiocyanate. The problems mentioned i n the section on t i t r a t i o n f i r s t arose here and the t i t r a t i o n error i s somewhat larger on that account. Run 15. No reason for t h i s large deviation i s known, but the value i s so f a r off i n comparison with other read-ings that i t i s presumed to be a reading error. 53 The main data fox the f i n a l readings are shown i n Table I. The r e s u l t s as calculated from the data are given i n Table I I , These represent a complete set, with none omitted, and those readings which are for some reason ques-tionable, have been starred and a footnote added i n explana-t i o n . The value of log p has been included i n Table I I together with the corresponding value of log p as calculated from the proposed equation of Ditchburn and Gilmour -The r e s u l t s of the measurement are shown i n graphical form i n Plates XI. and X I I . Plate X I . i s a normal plot of log p vs. 100Q/T. The s o l i d l i n e i s a p l o t of the Ditchburn and Gilmour equation. Since the differences are very d i f f i c u l t to see on such a small scale, the scale has been expanded i n Plate X I I . by p l o t t i n g differences i n log p values, using Equation (?) as reference, against 1000/T. This procedure makes possible a much more open seala which i l l u s t r a t e s the d i s t r i -bution of the experimental points. The K e l l y (3) and Duahman (10.a) equations -log p. 10.3735 - 3308 . 0.8 log T (7) T K e l l y log P, mm = 10.474 - 3327 - 0.826 log T (24) T (Hg * Hg 2) Dushman (Hg) log p; mm « 10.377 - 3285 _ 0.8254 log T (25) T P L A T E XI THE VAPOUR PRESSURE OP MERCURY PROM 30°C. TO 240°C. HATE XI. The vapour Pressure of Mercury from 30°G. to 240°C The present experiment The Ditchburn and Gilmour equation P L A T E X I I COMPARISON BETWEEN PRESENT RESULTS AND PROPOSED EQUATIONS 4 PLATE X I I . Comparison. Between Present Results and Proposed Equations The present experiment Unreliable readings Ditchburn and Gilmour equation Dushman equation K e l l y equation International C r i t i c a l Tables 2% deviation from Ditchburn and Gilmour equation p = log p - log p reading Ditchburn and Gilmour or equation - 54 -and the values given i n the International C r i t i c a l Tables (22) are plotted i n the same way to i l l u s t r a t e the divergence between the curves. 3Por reference purposes, l i n e s corres-ponding to Z% deviation from the Ditchburn and Gilmour curve have been added. 3. Discussion I t w i l l be seen from the tables and graphs that the apparatus gave very consistent r e s u l t s down to about 6Q°C, but that below t h i s temperature, the scatter was some-what more. I f the readings of doubtful r e l i a b i l i t y are omitted (see Table I. for explanation), the picture i s better. The low r e s u l t at 30°C. may be caused by the f a c t that a l l heater power had to be turned to an extremely low value to get to t h i s temperature at a l l , and a cold spot could have developed. However, i t i s thought that a new trap design might make an improvement at these very low pressures. There was no i n d i c a t i o n of f a i l u r e at the upper end. The t o t a l useful range of the apparatus was, therefore, from 5 x 10~ 3 mm. to more than 60 mm., a factor of at l e a s t 10 4 i n pressure. This range compares very favourably with that of any other apparatus that has been used. I t w i l l also be seen from the graphs that the agreement between the present work and the equation of Ditchburn and Gilmour i s extremely close. The mean deviation of the points shown on the graph (excluding the reading at - 55 3 microns and one value which i s presumed to be a reading error) from the Ditchburn and Gilmour curve i s 0.35$. I f a l l unreliable points are excluded the mean deviation i s 0.05$. The l a t t e r group of values have a standard deviation from the mean of 1.3$. I f one bears i n mind that at high temperatures the values obtained by t h i s method may be s l i g h t l y high because of the s l i g h t molecularity of the mercury, and that the accuracy of the measurements i s les s (with the greater p r o b a b i l i t y of th e i r heing low) a t low temperatures, i t i s evident that there i s no j u s t i f i c a t i o n for suggesting another equation. On the other hand, every r e l i a b l e value obtained i n t h i s experiment i s higher than the corresponding, value given by the K e l l y (Hg • Hg^) equation (24), i n most cases by more than the t o t a l error of the experiment. The Dushman (Hg) equation (25) i s seen to be completely at variance with these r e s u l t s and with the Ditchburn and Gilmour equation (7), although by coincidence, i t i s i n agreement at 100°C. This equation i s also inconsistent with the (Hg • Hgg) equation for i f both equations were correct they would con-verge at low temperatures instead of diverging. The diver-gence at high temperatures i s to be expected because of the presence of Hgg molecules, but i n view of the r e s u l t s of t h i s experiment and the fact that the two graph l i n e s cross at a f a i r angle instead of converging, i t , seems obvious that 56 they give no more than a q u a l i t a t i v e estimate of the degree of association. Assuming, however, that the expressions do give some estimate of the degree of association at higher temperatures, then the two highest temperature points should be s h i f t e d downward (say, to the positions marked -f- on the graph), and a l l suggestion of a difference of slope between the present measurements and Equation (7) i s removed. In view of t h i s i t seems certain that the Ditchburn and Gilmour equation i s the most accurate of the three at these tempera-tures. I t i s probably correct to 1 2% down to 6Q°C instead of to 13Q°C. as was o r i g i n a l l y proposed* Up to t h i s point no mention has been made of the vapour pressure table given i n the International C r i t i c a l Tables (22) of. 1928. (The same table i s given i n the 1949 Handbook, of Chemistry and Physics (18.c).) This Table agrees very w e l l with the Ditchburn and Gilmour equation at higher temperatures, but begins to diverge r a p i d l y i n the i n t e r -mediate region. The shape of the curve i n the 180° to 140°C. region does not inspire confidence and the Table probably cannot be r e l i e d on below t h i s point. The experimental determinations that have been made within the lower end of the range of t h i s apparatus are those of Neumann and Volker (24) and Volker and Kirchhoff (25). The result s of Neumann and Volker are uniformly 10$ below the present r e s u l t s and Equation (7) from 7Q° to 20°C. 57 -and are considerably below a l l the suggested equations. The Volmer and Kirchhoff values are for a small range below 10~ 2 mm. and are about 5$ below Equation (7). On the other hand, the r e s u l t s of th i s experiment are i n excellent agreement with the very consistent r e s u l t s of Menzles (26), Smith and Menzles (27) and Hodebush and Dixon (28) i n the high pressure region, since i t i s upon these r e s u l t s that Equation (7) i s based. I t seems probable that the present experiment gave r e s u l t s that are more r e l i a b l e than either of the low pressure measurements mentioned, i f only because a good check on i t s operation has been provided over an appreciable part of i t s range by other measurements that have been made. For t h i s reason, and for the further reason that Equation (7) i s i n good agreement with other workers at lower temperatures (29) (30), i t i s f e l t that the re s u l t s of Volmer and Kirchhoff and of Neumann said Volker should be disregarded. 58 CONCLUSION The re s u l t s of this experiment have shown that i t is. f e a s i b l e to make vapour pressure measurements on metals by the streaming method down to pressures at least as low as 5 microns. The condensation technique appears to give as r e l i a b l e r e s u l t s as the previously-used evaporation process, while i t allows larger gas flows and probably w i l l operate to lower pressures. The range over which a single streaming apparatus can operate covers a factor of at least 10 4 of pressure and extends over the whole of the intermediate pressure range. The apparatus used i n t h i s experiment was apparently the f i r s t of any generally applicable type to make such an i n t e r p o l a t i o n . ^ The vapour pressure of mercury has been measured at approximately 10 deg. intervals over the temperature range 30°C. to 24Q°C. Results agree w e l l within experimental error with, the vapour pressure equation suggested by Ditchburn and Gilmour -log p = 10.3735 - 3308 - 0.8 log T (7) T The formulae for the t o t a l vapour pressure of mercury and for the p a r t i a l pressure of mercury atoms which have been derived from K e l l y ' s free energy c a l c u l a t i o n s , The only other type of apparatus to do so i s one which applies only to the a l k a l i s 123). - 59 -are shown to he inconsistent and neither i s suitable for extrapolation into this pressure range. This i s taken to mean that these equations require r e v i s i o n before being used i n the higher temperature region and that c a l c u l a t i o n s of the degree of association based on them are not r e l i a b l e . The International C r i t i c a l Tables data on the vapour pressure of mercury are r e l i a b l e down to 14Q°C. but are inaccurate below t h i s point. For lower temperatures, the equation of Ditchburn and Gilmour should be used. The r e s u l t s of the present experiment indicate that values of the vapour pressure calculated from t h i s equation are correct to at least t 2% at 60°C. and t 4$ at the freezing point. - 60 -ACKNOWLEDGEMENTS I would l i k e to express my gratitude to the National Research Council of Canada for the generous f i n a n c i a l assistance that has been given to t h i s project, for permission to use various pieces of equipment, and for a studentship, a fellowship, and two summer scholar-ships that have been awarded to me. I also wish to thank Professors G. L. Pickard and P. A. Kaempffer for their advice and assistance while supervising the project. The following people have contributed materially to the success of the experiment and I acknowledge their assistance with thanks: Professor G* M. Shrum, Head, Department of Physics Professor A. M. Crooker, Department of Physics Professor R. E. Delavault, Department of Geology and Geography Professor M. Ki r s c h , Department of Chemistry My Wife Mr. W. Peria . Mr. John Lees Miss Ruth I r i s h Various s t a f f members and senior students of the Physics and Chemistry Departments who have contributed technical advice and assistance. 61 BIBLIOGRAPHY 1. Roberta, J.K., Heat and Thermodynamics, 3rd e d i t i o n , B l a c k i e , 1940. (a) p. 181 (b) p. 84 2. Gurney, R.W., Introduction to S t a t i s t i c a l Mechanics, McGrara H i l l , 1949. (a} p. 229 3. K e l l y , K.K., U.S. Bureau of Mines B u l l e t i n 383 (1935). 4. Ditchburn, R.W., and Gilmour, J.G., Rev. Mod. Phys. 13, 310, (1941) 5. Rodebush, W.H., and Coons, C.E., J . Am. Chem. Soc. 49, 1953, (1927) 6. Rodebush, W.H., and Henry, W.P., J . Am. Chem. Soc. 52, 3159, (1930) 7. Deita, V., J . Chem. Phys. 4, 575, (1936) 8. Theile, E., Ann. d. PhysiJc 14, 937, (1932) 9. Menaies, A.W.C., J . Am. Chem. Soc. 41, 1783, (1919) 10. Dushman, S., S c i e n t i f i c Foundations of Vacuum Technique, Wiley, 1949. (a) p. 781 (b) p. 74 11. S c i e n t i f i c Applications and Methods, E..H. Sargent and Co., F a l l , 1948. 12. Tolanaki, S., High Resolution Spectroscopy, Methuen, p. 38, 1947. 13. Temperature, I t s Measurement and Control i n Science and Industry, Reinhold, 1941. (a) p. 289 - (To) p. 279 14. Armstrong, L.D., and Dauphinee, T.M., Can. J . Res. A 25, 357, (1947) 15. Meuller, E.F., and S l i g h , T.S., R.S.I. 6, 958, (1922) - 62 -16. Roesser, W.F., and Wensel, H.T., J . Res. N. B. S. 10, 275 (1933) 17. K a l t h o f f , I.M., and Stenger, V.A., Volumetric Analysis, V o l . I I . , Interscience (1947) (a) p. 336 (b) p. 295 18. Handbook of Chemistry and Physics, 31st e d i t i o n , Chemical Rubber Pub. Co., 1949 (a) p. 1836 (b) p. 1854 (c) p. 1858 19. Glasstone, S., A Textbook of Physical Chemistry, Van. Mo strand, 1940 (a.) p. 495 (b) p. 445 20. Adam, U.K., The Physics and Chemistry of Surfaces, Oxford, 1941 21. Halban, H.H., Helv. Phys. Acta 8, 65 (1935) 22. International C r i t i c a l Tables, V o l . I I I . , p. 206, (1928) 23. Lewis, L.C., Z e i t s . f . Physik 70, 737 (1931) 24. Neumann, K., and Volker, E., Z e i t s . f . physik. Chemie 161, 33 (1932) 25. Volmer, M.,, and Kirchhoff, P., Z e i t s . f . physik. Chemie 115, 233 (1925) 26. Menzies, A.W.C., Z e i t s . f . physik. Chemie 130, 90 (1927) 27. Smith, A., and Menzles, A.W.C., J . Am. Chem. Soc. 32, 1412 (1910) 28. Rodebush, W.H., and Dixon, A.L., Phys. Rev. 26, 851 (1925) 29. Halban, H.H., Helv. Phys. Acta 7, 856 (1935) 30. Volmer, M., and Esterman, I . , Zeits f . Physik 7, 1 (1921) 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items