The Raman Spectrum Heat Capacities from 60 to 297°K., Heat of Fusion, Vapor-Pressures and Entropy of Monofluorotrichloromethane Citation Osborne, Darrell Wayne (1938) The Raman Spectrum Heat Capacities from 60 to 297°K., Heat of Fusion, Vapor-Pressures and Entropy of Monofluorotrichloromethane. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/ba3g-ms21. https://resolver.caltech.edu/CaltechTHESIS:11272024-170748854 Abstract No abstract. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Yost, Don M. Thesis Committee: Unknown, Unknown Defense Date: 1 January 1938 Record Number: CaltechTHESIS:11272024-170748854 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:11272024-170748854 DOI: 10.7907/ba3g-ms21 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 16884 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 11 Dec 2024 00:14 Last Modified: 11 Dec 2024 00:16 Thesis Files PDF - Final Version See Usage Policy. 14MB Repository Staff Only: item control page

1 The Raman Spectrum, Heat Capacities from 60 to 297° K., Heat of Fusion, Vapor-Pressures, and Entropy of Monofluorotrichloromethane Thesis Submitted by Darrell w. Osborne in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1938 The investigation of the Raman apectrurr., vapor-pressures, and calorimetric properties of monofluorotrichloromethane (CCl3F) between 60° and 297° K. has been carried out in order to permit c·omparison of the entropy calculated fr-orr. electron diffraction and spectroscopic data with that obtained with the aid of the "third law" of thermodynamics. This compound is also of interest because of its relation to difluorodichloromethane, used extensively as a refrigerant. The sample was kindly furnished by Dr. L. o. Brockway, who obtained it through the courtesy of Kinetic Chemicals, Inc., Delaware, New Jersey. The material was dried over calcium chloride and twice fractionally distilled through a column filled with glass he lie-es. A small Hopkins condenser cooled with iced water was placed at the top of the column to maintain suitable refluxing. The boiling range of the major fraction was about 0.1° c. and the b. p. (corrected) 23.7± 0.1° C. The middle fraction, about 130 ml., was then introduced into a bulb in the vacuum line connected with the calorimeter, the monofluorotrichloromethane frozen by means of liquid air, and the system evacuated w_i th a mercury diffusion pump. prevented eontamination by mercury. A liquid air trap Next the sample was complete- ly dried by distillation through a tube filled with phosphorus pentoxide. The last traces of air were removed by repeating several times the cycle of operations melting, freezing, and pumping. A few ml. were pumped off and then about 80 ml. distilled into the next bulb. Again a little was pumped off, and finally about 70 ml. were condensed into the previously evacuated calorirr.;eter, the temperature of the latter being kept above them. p. of monofluorotrichlororr.ethane so that the filling line would not 2 3 become plugged. The Apparatus. The apparatus was designed for vapor-pressure measurements and for low temperature calorimetry of condensed gases and volatile liquids. It was patterned after the apparatus used by Giauque and co-wor·k ers. 1 The calorimeter- assembly is shown in Fig. 1. The calori- meter, r.epr-esented at 10 in the diagram, is constructed of copper. Its wall thickness is about 0.05 cm. and volume 118 ml. Equally spaced along the inside are ten horizontal, perforated copper discs 0.03 cm. thick, which make a spring fit with the wa.lls and assist in establishing thermal equilibrium. The whole interior, including the disc.s, is gold plated. At 8 a small. Pyrex glass tube is sealed directly to the calorimeter. This tube, 2, serves for filling and also can be connected with a manometer for vapor-pressure measurements. A co.mbined resistance thermometer and heater of No. 40 dou.ble-silk-c·overed copper wire nearly covers the outside of the calorimeter. Its resistance is 167 ohms at o0 c.. It is baked on with Bakelite varnish and covered with gold leaf attached by another coat of lacquer. losses. The gold l.eaf is to minimize radia. ti on No. 36 doub-le-silk-covered copper wires are soldered to the No. 40 wir9s at the calorimeter. The No. 36 and No. 40 wires serve as current and potential leads respectively. Before use the tbermometer-heater was seasoned by cooling in liquid air and warming to room temperature several times. It is cali- brated concurrently with the heat c-apa:c.i ty measur·e ments by means of a sta.ndard copper-constantan thermocouple fastened to the 4 ---------- -----1 ---- ---- - - -- - - --2. -------------- ---1 -- --------"'# F~~f:,;t- -- -- -----5 ---------B ----- - - -'I - - - - - ---iO -- --- - ---11 - --------1.e. ~~~~~~~m-- -- - - --- -13 ½Ac.tuol Siu, 5 bottom of the calorimeter by Wood's metal in the small copper tube shown at 12 in Fig. 1. By means of Wood's metal, 5, the calorimeter is suspended fr-0m the radiation shield, ?, a hollow cylinder of copper weighing a.bout 2 kg. This heavy block is fitted with a constantan heater, 120 ohms resistance, wound on a closely fitting inner sleeve. The exterior is covered with gold foil. A calibrated thermocouple is soldered into the well at 11 with Wood's metal. S-tout silk cords, 4, fron; the top of a vacuum- tight brass can, 9, support the radiation s.hield. The wires from the calorimeter are led through Bakelite plugs, 6, paraffin being used fo,r thermal contact. All the leads are w.ound a:r ·ound the radiation shield a number of times, bound with silk thread, and coated with paraffin. They then go up the G.erman s.ilver tube, 1, and are taken out through a picein joint 80 e:m. above the top of the can. Also, the standard thermocouple is wound a number of times about the calorimeter; altogether about 80 cm. of it ar~ around calori~ meter· and shield. Asbestos paper was wra.p ped around the wires in tbe lower portion of the tube in order to protect them when the can, 9, was soldered to its cover at 3. In addition, low melting bismuth solder was used and wet cloths put abo.ut the ba$e of the tube. The German silver tube eonnects with an oil-diffusion pump. Holes. in the removable bottom of the radiation shield, l3, permit evacuation of the space between the calorimeter and radiation shield as well as that between radiation shield and cal.orimeter. 6 The assembly is suspended by the German sil.ver tube in a Dewar flask 10 cm. wide and 36 cm. long, inside measur~. A large brass can and a Bakelite cover through which the tube 1 passes enclose the calorimeter assembly and Dewar flask. This system can be made gas-tight and evacuated with a high capacity vacuum pump. Various cooling baths were used in the Dewar flask. The l.owest temperature, 57° K., was pr·o duced by boiling liquid air under reduced pressure. Between 90 and 2.00° K. the bath was liquid air under atmospheric pressure, between 200 and 275° K. carbon dioxide and alcohol, and between 275 and 300° K. ice and water. All electrical measurements are. made w.i th a White Double Potentiometer with 100,000 microvolts range. The manufacturers, Leeds and Northrup, guarantee the instrument to 0.02%, and in addition an interc:ompar.ison of the coils was made. The standard cell was calibrated by Dr. Dunnington of the Physi c.s Department in April, 1.937, against a cell checked at the National Bureau of Standards seven months previously. It was of the unsaturated type, with negligible temperature co.effic.ient, but inasmuch as there is likely to be hysteresis when the temperature is changed, it was kept in an air thermostat. The fixed resista,nces were of manganin ( low temperature c:oef .fici ent and low thermoelectric power against copper) and were c:ompared po-tentiometrically with lab.oratory standards having Leeds, and Northrup certificates abo.ut ten years old. Probably the standards and calibrations were good to ·0.02%. The galvanometer is a Lee'ds and Nor·thrup H S type with an 7 all copper circuit. Its period is 10 sec. It is read with a telescope and a scale a.t 5 meters distance, and with this arrangement the sensitivity is 0.05 micr·ovolt per mm. One can estimate the deflection to about 1/4 mm. The electrical circuits for heating and temperature ment are shown in Fig. 2. measur·e- The thermocouples are not pictured; suitable switches connect them to the P potentiometer. connectors and switches ar-e of copper. All the Most of the connectors are no. 14 or 16 wire, well _insulated with rubber. The system is shielded against stray e. m. f.'s. When all the switches in Fig. 2. are in the up position, the circuit is as shown in Fig. 2a, and temperature measurements can be made with the resistance thermometer--heater R. A very constant thermometer curre·nt is given by 1.6 Willard Low Discharge Batteries, Type DA 2-1, in series. The working cells for the potentiometer also are of the same kind. The resistance of the thermometer is obtained by comparing ER, the potential drop across R, with E1 , the potential drop across a fixed 200 ohm resistance. The ther-mometer current, 0.0004 amp., is small enough for its heating effect to be neglected. The current is allowed to flow continuously througl1 R, except during energy input, and then the batteries are kept in a steady state by discharge through practically the same total resistance (80,000 ohms). E 1 was so c.onsta.nt t.ha t it was measured o,nly near the beginning and end of each heat capacity determination. A copper plug box, r, with manganin resistances, is used to keep the galvanometer critically damped. The sensitivity of the thermometer is about 300 microvolts per degree over most of the range. 8 El(c.rci.ser IT 1.,0 JL 80,000.IL To P Poh,n! iomd:c.r ~ +rr- ~LO,oooJI. 1.ll i<.,o.n. Fig.z He.~ting and T<l.rnp<.roturt. Circ.u'1ts '10,000.Jl. 200.n.. Ex ,o.n. 1.n. EC R son. ER Ev F;g. ?.A T<e.mp<!.r'dt'l.o-e Cire1/1t- Fig 2B ~~ting c;rc.uit 9 To add energy to the calorimeter, switches 2 to 5 are pulled down and at the pr-aper time the rapid throw switch 1 also pulled down. The circuit is then as given in Fig. 2b. The vo1-tage acr-oss the heater terminals is obtained by measuring EV, the potential dr-op across a 50 ohm portion of the 10,050 ohms in para,llel with R. The current through R and the resistance in parallel is obtained from Ee• All the individual r-esisters were accurately calibrated, and as: a check Ee was c-ompared with the po-tential drop ac:ross a standard l ohm resistance temporarily placed in series. The resistance of the potential leads from R to the 10,050 ohm resistance was measured with the calorimeter at .60 and at 290° K., radiation shield at nearly the same temperatures, and cooling baths of liq_uid air and ice and water, respectively, in the Dewar flask. to 0.2%. These leads affect Ev by 0.1 Fro.m Ev and Ee can of course_ be calculated the power, the re:sistance of the thermometer-heater, R, and the fraction of the current passing through the resistance in parallel with R. The source of energy is a battery of 1 to 1-0 lead storage cells. An exer-c.ising rheostat is provided fo-r br-inging the battery to a steady state befor-e supplying ener-gy to the calor-imeter. t' Time measurements are made with a small, syncxxonous motor geared to turn a Veeder-Root counter and to actuate a clicking devic.e, made fr-om a door bell, once per sec. The power c·omes from Boulder Dam, where the frequency is automatically controlled with high acc:uragy. The timer was found to give consistent results by comparison with a stop-watch. A r -easonable estimate of the probable all-over error in timing, including any slight inaccuracy in manual swi tc.hing is>t 0 .05%. 10 As mentioned before, a manometer can be connected to the calorimeter for vapor-pressure measurements. It is constructed of Pyrex tubing, 1.9 cm. i. d., and is enclosed in a wooden case with a plate glass window. A stream of compressed air from the laboratory supply, which changes temperature but slowly, is circulated in the case to eliminate temperature gradients in the mercury columns. • The temperature inside the case is determined from two calibrated mercury thermometers. Gold foil in ·the line protects the calorimeter from mercury vapor. Fur·ther details as to the e:onstruction of the apparatus rr.ay be found in the Doctorate Thesis of Mr. C. s. Garner (1938). Construction and Calibratio~ of the Thermocouples Each thermocouple consists of fiv.e No. 30 doubl.e-silk-cove·red constantan wires about three meters l.ong, connected in parallel to average out small imperfections, and one No. 36 double-silkcovered copper soldered (with soft solder} to each junction, all bound together with silk thread and given a light coating of e1-eetrician' s enameling fl.uid. The ice ends were inserted in small diameter·, paraffin-filled, glass tubes about 45 cm. long. Similar bundles of constantan, taken from the spool just before these, were tested for inhomogeneities by the usual method of meaauring thee. m. f. set up when different portions of the wire are immersed in liquid air, both ends of the loop being joined to copper leads and kept in an ice bath. The maximum e.m.f. found was 1.5 microvolts and the average 0.4 microvolt. The thermocouples were compared with one calibrated in the investigation of Giauque, Johnston, and Kelley, 2 kindly lent to us by Dr. Hugh Huffman. Unfortunately, the copper wire to the 11 variable junction was accidentally broken; it was replaced by a new copper wire, which was found to differ from the old one by 2.4 microvolts at 85° K. This correction was equivalent to only 0.1? 0 at 60° K. and was still smaller at higher temperatures. For the comparison the variable junctions of the standard and of five of our thermocouples were imbedded in Wood's metal in small copper receptacles soldered to the lower part of a solid cylinder of copper approximately the size of the calorimeter and in order to eliminate temperature gradients near the junctions, about 80 cm. of each couple wound around the cylinder, bound with silk thread, paraffined, and then wound with adhesive tape. This cylinder was suspended by stout silk casting line in the place later occupied by the calorimeter , and the thermocouple wires and leads to the constantan heater which was wound on the cylinder were led up through the radiation shield and taken out through picein at the top of the apparatus. The ice ends were immersed in a Dewar flask (to a depth of about 25 cm . ) filled with distilled water and finely crushed ice. Comparisons were made at approxi- mately ten degree intervals between 5? and 293° K. , with radiation shield and coppar cylinder at nearly the same temperature and the apparatus evacuated. Referring to the following diagram, Thcz.rmoc.oupl(l. No. S!d. A S 8 7 C D E F ••I C.op p«r - C.011sfanh,n - Co pp ll.r bloc.k a! T" the large potential differences AC and CD were read on the fuite potentiometer and the small. ones such as BD, CE, DF, and similar 12 quantities for the other thermocouples measured by double deflection of the galvanometer. The calibration of the standard thermocouple was in terms of the table given by Giauque, Buffington, and Sc-hulze 3 for their thermocouple 17. Inasmuch as our thermocouples differed from this table by about 270 microvolts at 60° K., it was necessary to construct our own conversion table. To do this, E(l?) - 1.04720 E(7) was plotted against E(17), where E(17) is thee. m. f. of thermocouple 17 of Giauque, Buffington, and Schulze, and E(?) is the corresponding quantity for our thermocouple 7. This graph is shown on a small scale in Fig. 3, from which it can be seen that for many purposes it would be sufficiently accurate to take the ratio of E(l7) to E(7) as a constant. The plot was used in con- junctio.n with the temperature scale of Giauque, Buffington) and Schulze to obtain E(?) at 50°, 60~ 70°, etc., up to 290° K. Linear interpolation between these values is not sufficiently precise, and hence it was decided to interpolate by the procedure of Southard and Andrews. 4 equation The data was fit approximately by the log (6401.80 - E) =- - 0.3233500 + 1.695000 log T, then E - E(7) plotted against E(7) on a large scale, and a smooth curve drawn which missed no point by more than 0.1 microvolt. With the aid of this curve the temperatures corresponding to E(7) were calculated to 0.001° at intervals of 10 microvolts between 50 and 273.lo K. and of 20 microvolts between 273.1 and 300° K. It should be pointed out that the analytical expression alone does not fit the data at all well, deviating as much as 5.5 microvolts, but with the plot of E - E(?) it served very well for interpolation. After the calorimeter was installed, with thermocouple 7 in place on the bottom, the temperature scale was compared with the 13 ':> ~ i: .;:::-. '-" f LlJ 0 (\J t--, ~ 0 "l"'I 1 I ~ ~ "-" uJ 0 -1000 0 ~oo ~o~ £(:11) tn 3""~ JA•V- t.ooo 14 oxygen vapor-pressure thermometer. The oxyg en was prepared by heating c. p. potassium permanganate in a well evacuated, glass system, dried by passage through a tube filled with phosphorus pentoxide, liquefied by means of liquid air, and fractionated, a middle portion being distilled into the calorimeter. Pressure measurements were made with a calibrated Henson comparator and the manometer described previously. observer read E(?). Simultaneously another Data from the International Critical Tables 5 were used to correct the readings to international ems. of mercury. No correction was necessary for capillary depression. The acceler- ation of gravity was taken as 979.57 cm./sec.2, as determined by the Coast Geodetic Survey at this location. results of the comparison. Table I gives the The measurements are given in the order in which they were made. About 2 liters of oxygen gas were pumped off following the thirteenth reading and again after the the seventeenth. In the first column are the temperatures obtained from the thermocouple, and in column two are the pressures calculated from the equation of Cath, 6 logio P (cm.) =-4.19.31/T -t- 7.1173 - 0.00648 T, assuming the tempera tures coming from the thermocouple are correct. The third column lists the observed pressur-es, and column four gives the temperatures calculated from the observed pressures of •oxygen less the temperatures obtained from therm.o couple 7. The satisfac.tory agreement with the oxygen vapor-pressure thermometer gives me confidence that our temperature s,cale, assuming o° C. to equal 273. 10 ° K., is good to ±. 0. 05 °, which is the accuracy claimed for- the thermocouples of Giauque, Buffington, and Schulze. It is true that the last ten degrees or so below 300° K. are extrapolated, but since thee. m. f. is nearly a linear function of temperature 15 in that range, probably little error was introduced. Table I Comparison of Therm.oc-ouple ? with Oxygen Vapor-Pressures p T frem the·rmocouple ? o K. ?O.923 70.941 ?3.954 ?3.968 ?3.98? 74.734 7.4.742 80.578 80.584 84.5?5 84.580 p calculated cm. of Hg. observed cm. of Hg. 5.566 5.584 9.294 9.315 9.344 10.52? 10.540 24.624 24.644 40.870 40.895 5.556 5.559 9.263 9.276 9.330 10.523 10.518 2A. 665 24.672 40.866 40.903 74.004 ?4.036 49.696 49.?38 75.630 75.656 77.538 77.56 ?3.87 89.853 89.857 73.90 49.785 49.801 ?5.66 ?5.68 B6.255 86.258 90.083 90.086 90.327 90.331 77.60 77.63 /:l T o K. -0.010 - o. 026 - 0.021 -0.026 -0.009 - 0.003 ~0.013 -to. 012 +0.008 -0~001 +0.002 +0.017 +o •.01? -O.Ql6 -0.011 -o. 004 -0.003 -0.008 -0.009 Calibration of the Resistance Thermometer The resia,tance thermometer was cali"or.ated during- the heat capacity. runs by comparison with thermocouple 7, on the bottom of the calorimeter. The absolute ac.curacy is of cours.e no better than tha't of the thermocouple, ±0.05°, ·but it is a.ssu.med that the absolute error does not vary appreciably over a small range, so that small temperature intervals can be determined with much higher accura.c;y. Measurements were made at intervals- of approxi- mately 2 to 6 degrees. The data was treated by making two plots, one of deviations of R from a linear function of temperature, and the other of dR/dT. The fir·st curv·e was used to get the average temperature of the heat capacity determination. The sec.end was on a very large scale and was used to convert the 16 r--, I-\ • n 11 I' .u l ':t :I ...... l 11 Ii - j 1 fa 11 . r T l ~ - ..... 4 Cc 0 17 effective rise in resistance of the thermometer-heater into the corresponding temperature rise. Fig. 4 shows this graph much r"educed, with the experimental points omitted since tb.ey are too numer·ous. It was mor-e convenient to have R rather than T as abscissae; temperatur·es corresponding to a few ·points are 13 ohms--60.18° K., 50 ohms--113.80° K., 180 ohms--294.39° K. Heat Capacity Measurements The method is as follows: Hydrogen is admitted into the calorimeter assembly and the calorimeter allowed to fall to the temperature of the cooling bath. Then the container is evacuated. The radiation shiel.d is heated to a suitable temperature, and in c:ase it is much hotter than the c·ooling bath, a small current is adjusted to keep its temperature constant during the run. Below the m. p. of monofluorotrichl.oromethane the radiation shield was between the initial and final temperatures of the calorimeter; above them. p. it was definitely hotter to prevent condensation of the sample in the inlet tube with resulting heat effect which would ruin the heat capacity run. The heating rate of the calorimeter- is then taken by measuring Er once and then Ea at 50 sec. intervals for about 5 minutes. Then a current of 0.05 to 0.1 amp. is passed through the calorimeter heater for a pre-determined time of 5 to 15 minutes, producing a temperature rise of 3 or 4 degrees. For 10 minutes or mor·e prior to the energy input the battery has been discharging through a rheostat set equal to the total resistance in the heater circuit. at suitable intervals. Ey and Ee are measured After· the end of the energy input, the heating rate is again followed, usually for 20 or 25 minutes, depending on how long is required to reach a steady state. The 18 usual time necessary was 1.0 or 20 minutes. While the calorimeter is being heated, both current and voltage change due to the inc-r·e ase in resistance of the thermometerheater. To avoid graphical integration an extension of the method •· l was emp1 oye d . of G 1.~b son an d G1.auque It can b e s h own t'na t l.• f a quantity changes quadratically with time, then the average over a per·iod t equals the average of the values at 0.211 t and 0.789 t, and also equals the aver-age of the values at 0.147 t, 0.500 t, and 0.853 t, r ·egardless of the coefficients in the quadratic·. This method was tested for Ee and Ev at the" lowest temperature, where the percent change in R is greatest for a given temperature rise. In each case it was found that the average Ev or Ee obtained .from frequent readings agreed to 0.01% with that calculated from the values at 0.211 t and 0.789 t, or 0.147 t, 0.500 t, and 0.853 t. This agreement is rather amazing in view of the fact that the change in Ee in the test was aa high as 18%. In c:ale:ulating the e·nergy sup,pl.ied to the heater, a defined calorie equal to 4.185/1.0004 international joules was used. Allowance was made for the resistance of the potential leads (0.1 to 0.2%) and for the fraction of the current passing through the 10,050 ohm resistance in parallel with the heater. could be calculated to a high deg_ree of accuracy. These A ra.ther uncer- tain correction of 0.04% was added for the heat developed in the current leads between calorimeter and radiation shield. It was assumed that half the heat went to the calorimeter and half to the shield. The correction for heat inter-change with the ~urroundings was made on the assum;ption that over a range of 4 or 5 degrees the heat transfer varies linearly with the temperature of the exterior 19 of the c:al.orimeter, the temperature of the radiation shield remaining constant. The maj'or part of the correction was applied by linearly extrap,olating to the middle of the heating period the measurements of ER made be.fore starting energy input and the measurements of ER made after a steady state is reached following energy input. The difference in these extrapolated values of ER, • multiplied by 200/(Er)avg., gives the effective rise in resistance of the thermometer, except for some supplementary c:orrections discussed in the following paragraphs. The correction made as just de scr·ibed would take c.are of heat interchange if the change in ER with time wer·e really constant and if the average temperature of the heater-thermometer during energy input, w,hich can be obtained from the energy measurements, w.ere equal to the average of the initial and final temperatures of the calorimeter and contents. Deviations from the first of these were important only in the case of the empty calorimeter, giving rise to a cor.rection to the heat capacity of the empty calorimeter which increased from about 0.1% at l.30° K. to 2% a.t 295° K. Deviations from the second affected the heat capacity of the em:p,ty calorimeter by 0.2% and the t otal heat capacity by 0.4% ne.ar room temperature. During the measurements of resistance before and after energy input there is a slight thermal head between the resistance thermometer and the average location of heat capac.ity, since the temperature of the calorimeter and contents is changing mostly due to thermal flow through the resistance thermometer'.. The c:orrection for the difference in this thermal head turns out to be e.qual to that due to the incr-eas.e.d temperature of the exterior during energy input. 20 Correction was also made for change in the temperature of the radiation shield during a fe.w of the runs. This can be made negligible by proper adjustment of the current through the shield heater. The heat capacity of the empty calorimeter, about one-third the total heat capacity with the size of sample used, was measured by the method described, about 2 cm. of hydrogen being inside the calorimeter to aid in the distribution of heat. are shown graphically in Fig. 5. The resul ta The smooth curve actually used was on a scale much larger, and the average deviation was about 0.1%. A correction was necessary for vaporization of monofluorotrichloromethane at temperatures above 180° K. The volume of the calorimeter and inlet tube was measured accurately by an expansion method and the volume of the liquid calculated from the mass and density o.f the substance. Ruff and Keim 7 have reported the density to be 2.08 - 0.00185 T gm./tnl. between 180° and 201° K.; and 1.4944 at 17.2° C. Linear interpolation was used to get the density at other temperatures. The vapor-pressure and heat of vaporization were calculated from the vapor-pressure equation given later in this paper. The correction amounted to about 2% near the boiling point, and probably it was good to about 10% of i ta value. It was intended to measure the amount of material in the calorimeter by distilling it into a bulb, connected to the inlet tube by means of a vacuum-tight, ground-glass joint, and weighing it. Unfortunately, the line broke during this operation, and some was lost. Hence the sample was dried and fractionated on the Caf I.¾'· m .:E l=jti; _ ""' ' .. ,:11 ~ , '::i ·,U,} 0 :1 i 3 (.,0 , !:j:j: •.1 ·.f:"'•.. -:x'. +-+t ♦jjj: . ,,, -~~ -- H- 1-1-l .:t fl m: '" "+'rltt·_-ct-L•·•·_ iC }·-. ·L ~-~~-·1·, - -,.. ,!:tr.riJJ ,.-~ ~•,.1--,_ •• .-, .__,._"T"I l _..t;J:i , i fTl rfT?rd 1, F • -~ 100 ~ C['{ffiij ' w - . .. ,.' ,.•. ,,.- ... ~- r:-:: m r 11 l T tt , .,,-~··• ...............~_ '. 0. •_ i- ~/_·; +:=. ' • it · c, f[f'.li'' :c;_(f4'tt. 'r':1~~ .-,-+Jffi. . ·_ •L~•~1~.- ~-~l:iii-'i ,;l J~• • 11W=I 1-i:t:' O ff T~ i- tt tt ;?50 .J- lHl ++-, fE t{~~; .. ' .·: ,::ci, dt ·~·'.·,~-· 300 ·a ~.:;:;:ff;'H-l fflm PJ -1 ~ ~i~ • • L~~t ~-u·fc.t-1TL: :,~ le- l'.l HP -qy_\,ct,LO---JUt-1 E.fE. i;w..u..,, i-,: ·.,Tti Ii-W'Jffil! in ?-1.' 1"i u~L:~rt~r-~. · · ,--:nr '"'' ~ p mmt- -~~ i :,c ~ P.Y\ 91r ~ ___ F[ G.., !7t-, • l+hti -+-i"ti, ''-' • 1 I 1111111111111 ~.f,Ii~ = . "' •' - -- " t+-'= ,-rn, . ~~4-:..; iii•~ '~lliii , -1:. - llllllliil ~ ~ i ~ lr i ~ in l ~ l ~ I ~ i~ l i1 ~ 1-1 ~ lili : _ _ 1 ~ ~ . ;._;. ••• _._-__ • m .j _.. ··~,__ '"-_ ,__-;...-~•: ~:¼ - ' .i:F:•· ·:,... ,'· l ' ,,')~t'· , 'I • ;• - ' .. ~ ... -· . ~ - -- • :__ l · ' • ,oo I '':TC'/ C. -, :-_ •. "'· ~-J,,; . "" "' •-~-i-_• •·! -, . -. ,.cj:j: ' ,,- • , ·- • l s ,:n fll'l •K ~-- J:,_- + ., '. - L __ ·~- "-~~•:.m11f , -., W• £Li•,1·______ t ---t_t•=... ;_,...i :__ , ... ,, ,._ - ,-;:-;:;:: • 'ff_·;•,. rta.,•: :••• '•.:..~+_-_ .. ! ,-, ,- ~ ~ ,m,u., ~' ~ .J!l - "'ct""±l:f:l:ltl -~ '' . ·_1· • s Cr 7' 8 <f 10 11 Jr;, ;f f i+l+m l:l+um •:4~~lliHffl l ,,i.~ ·:-11 G Ii 1-' t\:- 22 vacuum bench once more, weighed in the bul.b, and condeneed in the calorimeter. Several heat capacity measurements were made in a range where the previous results were partic-ularly consistent. The amount pre-sent in the first series was then calculated to be 0.7660 mole, using 137.38 as the molecular weight. uncertainty was about 0.1%. The estimated Weights were corrected to vacuum. The molal heat c_a pac,i ti es of monofluor-otrichloromethane are given in Table II and Fig. 6. The mean deviation from a smooth curve is about 0.2% for the solid and 0.3% for the liquid. The probable error is about O. 3% for the solid and O. 6% for the liquid. Table II Molal Rea t Capac.i ties of Monofluorotr-ichloromethane ~ C cal./degr/mole 62.. 22 65.25 68.08 70.53 75.85 79.85 11.68 1.2.09 •12.3.?J: · 12.56 12.90 13.22 85.37 13.69 89.02 89.29 13.94 13.96 14.28 14.53 14.82 15.12 15 .46 15 .75 16.12 16.46 16.85 17.30 92.86 96.36 99.68 103.0 106.6 ll0.3 114.0 117.9 122.4 126.2 131.2 135.3 137.3 139.1 141.0 141.9 143.2 Run number l 2 3 4 5 6 7 8 59 9 18.13 10 11 12 13 14 15 16 63 17 18 19 18.38 64 18.58 18.83 18.92 20 65 60 • 21 17.77 19.06 continued on next page 23 Table II, continued. 144.7 145.9 14?.0 148.3 152.0 156.4 156.5 157 .1 159 .5 160.0 160.1 162.54 165.9 169.2 171.2 173.0 175.8 · 180.2 183.3 190.9 196.4 201.7 207.0 212.9 222.0 225.0 231.8 241.3 247.9 254.7 260.2 261.0 266.3 272.5 279.3 283.5 285.6 2 90.5 291.8 296.88 Molal Heat Capacities of liiionofluorotrichloromethane. cal ./deg:P/mole C Run number 19.32 19.36 19.49 19.75 19.9? 21.66 21.12 21.92 25.90 26.74 28.20 66 61 22 67 68 23 28 69 24 29 27 melting point 26.10 26.11 26.10 26.18 26.29 26.45 26.37 26.59 26.66 26.63 26.70 26.93 27.05 27.19 27.17 27.42 27.55 2?.70 2?.73 27.87 27.91 28.07 28.21 28.38 28.41 28.60 28.56 30 25 31 26 32 33 70 36 37 38 39 71 73 74 72 75 ?6 77 78 49 79 80 81 83 82 55 84 boiling point In a number of runs there was condensation in the inlet tube, giving high results. These have been rejected and do not appear in Table II. The heat capacities for several degrees below the melting point are high due to premelting. The area between the heat 24 111~ ¥ H, Q ,• ~ - - .L > t I<) N ~ t .- : :i . H -1:t N Cl.i .:? < Ii! :t ,, .... J '! .~~ ,.)"' 'oo 0 ..J 0 :r µ: t\t w ' ~ •• If I 1 I ~ >:I ' .f 1:1: I -~l ,... 0: tl~ 1 ., (tii t-- ..... . ·~ ; Ill) 1":i 0 l4 I ;ij E ll J ,. 1. -- t i"T -, r--, Qt bo ~ c:-. Cl( ~ 0..1 ..., N t N ~ <IL (I.I: 0 <-1 C\l N CL. 2CJ I ~ ~ ~ ~- - ,__ I<) ,...t-- "'--' -~ -:-:: L ... ~ J . •-t-+, ·• ~ ~ ~ ~ ~ 25 capacity curve through the observed points and that obtained by extrapolation from below where pr-emelting is apprec·iable can be used with the heat of fusion and m. p. to calculate the mol-percent of impurity which is soluble in the liquid and insoluble in the solid, assuming that the solution formed is perfect. Thus, at 159.5° K. this heat effect amounts to about 15 cal. per mole; hence one calculates that there is 0.09 mol-percent liquid soluble, solid insoluble impurity in the monofluorotrichloromethane sample. He-at of Fusion The heat of fusion was determined in practically the same way as the heat capacities. The same amount of material, 0.7660 mole, was in the calorimeter. The radiation shield was heated to a fraction of a degree above the melting point and then the current through its heater adjusted to hold the temperature constant. The calorimeter was brought to about 150° K., and after a steady state was reached, the heating rate was observed. Then about O.l amp. was passed through the heater-thermometer until all the solid wa.s melted. The time necessary was about 4500 sec. Readings of Ec and Ey, f'rom which the current and voltage could be calculated, were taken at intervals. Then after the heating period the drift was again followed until it became constant. surr,n;ariz.ed in Table III. The results are The first column gives the temperature interval through which the calorimeter and contents rose. Column two gives the energy added electrically, and column three lists the heat received from the surroundings. The fourth column gives the heat required to raise the temperature of the calorimeter and the 0.?660 mole of the solid to them. p. (corrected for premelting) plus the heat to raise the calorirr.eter and the liquid to the final 26 temperature. The la.st column gives the molal heat of fusion. The pr·obable err·or is about 0.2%. Table III Heat of Fusion of Monofluorotrichloromethane dh. ~H cal.,r'mole 4.0 399.3 1641.0 5.2 332.0 1642.? Mean 1641.8 t3.2 Temp. interval Energy input cal. Interchange cal. 150.73-166.35 1652.3 150.59-163.90 1585.l K. J C dT Melting Point Determination The melting point was determined by observing the equilibrium temperatures with different amounts melted, calculated from the amount of energy added and the heat of fusion. presented in Table IV. The results are The change of melting po.int with percent mel.ted indicates about 0.01 mol-percent of impurity, in contrast with the estimate of 0.09 mol-percent obtained from the premelting effect. PeJZhaps there was lack of equilibriurr. in the premelting range. The melting point of the pure compound from our measure- ments is 162.54 +- 0.05° K. Table IV Ruff and Keim 7 give -111° C. Melting Point· of Monofluorotrichlaromethane % '.I' ( aK.) T (DK.) from res. thermometer melted from thermocouple 7 20 162.493 32 l .6 2.523 162.531 45 162.538 162.539 58 162.542 162.540 71 162.543 162.540 84 162.541 162.542 27 Vapor-Pressure.a The vap-or-pressures of monofl.uorotrichloromethane were measured in the manner described f o.r the comparison of thern.oc ouple 7 with • the oxygen vapor-pressure thermometer. The observations are shown in Table V along with the pressures calculated from the equation log1 0 P (cm. )~ in coluILn three. 8.21671 - (1610.04)/T - 0.0030743 T, Temperatures, colurr.n one, were obtained from thermocouple 7 and are given to 0.001° because of the high relative accuracy. The observed pressur·es are in c olurm two. All the pressures are expressed in international ems. of mercury. Tab.le V Vapor-Pressures of Monofluorotrichlororr.ethane T oK. 236.640 2.3.6 .680 2.3 9. 742 2.39.812 251.489 251.526 251. 571. 265.183 265.229 274.799 2,74.809 280.918 280.938 280.945 282.229 282.268 285.980 285.987 291.103 291.109 291.211 293.844 293.893 295.330 295.381 p ems. of Hg p calculated ems. of Hg 4.854 4.851 5.781 5.781 11.. 019 11.040 11.052 21.398 21.410 32.509 32. 510 41.869 41.884 41.894 44.120 44.192 51.027 51.035 , 61.802 61.840 62.046 68.205 68.376 71. 93 7 72.125 4.848 4.858 5.807 5.830 11.003 11.024 11.050 21.381 21.426 32.578 32.592 41.853 41.887 41 . 901 44.088 44.156 51.006 51. 0 19 61.793 61.807 6 2 .040 68.253 68.371 71.962 72.093 observed Pobs. - Peale. ems. of Hg. + 0.006 - 0.007 - 0. 026 - 0.049 + 0 .016 + 0 .016 - 0.002 + 0.017 -0 . 0 16 - 0 .069 - 0.082 -t- 0.016 -0 .003 -0.00? + 0.032 +o.036 + 0. 021 -t- 0.014 + 0.009 + 0 .033 + 0.006 - 0.048 + 0.005 - o.025 + 0.032 28 The boiling point calculated from the vapor-pressure equation is 296.88 + 0.05° K. (23.'78° c.), which is in agreement with the value obtained by distillation of the sample, 23.? + 0.1° C. (corrected to 76 cm. of mercury by means of the vapor-pressure equation.) The equation of Ruff and Keim, 7 log 10 P (cm.) = 7.540 - 1385/T, gives 24.1° C. as the boiling point, but since the highest pressure from which this equation was obtained was less than 60 cm., the fit should not be expected to be very close near the boiling point. The only other value of the boiling point recorded in the literature is 2,4.gJO C. S The heat o.f vaporization of monofluorotrichlorometbane at its boiling point was calculated from the empirical vapor-pressure equation and the Clapeyron equation, using 1.48 gm./ml. as the density of the liquid and assuming that the vapor follows a zr;odified Berthelot equation of state pv = RT [ l + 9pTc (1 - 6T~/T2)/128pcT j. The critical constants Tc and Pc have not been determined for this substance, but an estimate may be obtained from the fact that the ratio of critical . temperature to boiling temperature for similar compounds is fairly constant. The average of the ratio 9 for carbon tetrachloride, 9 1.590, and for chloroform, 1.603, was. used to get 4'74.1 ° K. as Tc for monofluorotrichloromethane. A linear extrapolation to 4'74.1° K. on a plot of log P against 1/T for the four highest vapor-pressures measured gives pc=- 50.0 a tm. This is nearly equal to Pc for· carbon tetrachloride, 45.0 atmospheres. 9 The heat of vaporization calculated using these constants is 5905 cal./mole at the boiling point. correc.tion amounts to -220 cal./mole. The Berthelot 29 Entr-opy fr-om the Calorimetric Data The calculation is summarized in Table VI. uncertainty lies in the extrapolation too° K. The greates t Several methods were considered, but no satisfactory one was found. Presumably one could fit the observed curve 'oy a combination of Einstein and Debye functions; however, this procedure is extremely difficult. Then-formula of Lewis and Gibson 10 and the method of Kelley, Parks, and Huffman 11 were tested on the data for carbon tetrachloride; 12 the values of the entropy obtained were both about 15% low. The method actua l l y used wa s to plot against log T the molal heat capacities of monofluorotrichloromethane (above 60° K.) 12 and those of car-b an tetrachloride given by Lord and Blanchard, and then to shift the heat capacity curve for carbon tetrachlo~ide along the axis of log T until it joined on to the other curve at 60° K. Graphical integration of the curve obtained in this manner gave the value 10.12 cal./deg./mole listed in Table VI. The change in entropy between 60° and them. p. and between them. p. and the b. p. was calculated by graphical integration of a smooth curve through the experimental heat capacities (corrected for premelting in the interval just below them. p.) plotted on a large scale against log T. A small correc t ion for gas imperfection was derived from the equation of state assumed for the calculation of the heat of vaporization together with one of Maxwell's relations, (dS/dP)T: - (dV/dT)~. In order to get the change in entropy of the gas between the b. p. and 298.10° K., at 1 atm. pressure, the molal heat capacity of the gas at constant pressure and these temperatures was calculated from the spectroscopic and electron diffraction data. 30 Table VI Entropy of Monofluorotrichloromethane from the Calorimetric Data s cal ./deg ./mole O - 60.ooo K., extrapolation 10 .12 ± 1.5 60.00 - 162.540 K., graphical 15. 26 ± 0. 05 Fusion, 1641.8/162.54 10.10 ± 0.02 162.54 - 296.88° K., graphical 16.34 ± 0.10 Vaporization, 5905/296.88 19.89±,0.20 Entropy of actual gas at b . p. 71. 71 + l. 5 296.88 - 298.10° K. 0 . 0? Correction for gas imperfection Entropy of ideal gas at 298 .10° K., 1 atm. 71.92 ± 1. 5 Raman Spectrum of Monofluorotrichloromethane The Raman spectrum of liquid monofluorotrichloromethane was photographed, using a water-cooled mercury arc in Pyrex and a 0 short-focus glass sp-ectrograph with a dispersion of about 60 A . 0 per mm. near- 4400 A. The compound was distilled in vacuo into the Raman tube, which w.as of Pyrex and about 25 ml. capacity. An iron arc provided the comparison spectrum. An exposure of about one-half hour was sufficient. In Table VII are listed the observed Raman shifts, together with the number of times each shift appeared on the plate (including both Stokes and anti-Stokes lines) and also the relative intensities. 2 e:m . -1 The accuracy for the frequencies was about except for the weakest line, which could not be measured with the comparator. 31 Table VII Raman Spectrum of Monofluorotrichloromethane ,.., -1 Relative No. of /::.11 in cm. appearances intensity 246 3 6 351 3 6 398 4 6 536 5 10 838 2 4, broad 1072 2 2 1540 1 0 It should be remarked that the frequency 838 c.m. -1 s h ows some indication of being a doublet. By the methods of group theory 13 it has been shown that a molecule such as FCC1 3 has six normal modes of vibration. Of these, three are single and three double, and all are active both in the Raman effect and in the infrared absorption spectrum. In the non-degenerate vibrations there is a change in the electric moment along the symrr.etry axis and in the doubly-degenerate ones a change perpendicular to the symmetry axis. The lines arising from the doubly-degenerate vibrations have the special depolarization ratio 6/7 for unpolarized and 3/4 for plane-polarized incident light. In order to make the assignment, the non-degenerate frequencies of FCC1 3 were computed by the following procedure. Rosenthal and Voge have given a normal coordinate treatment for molecules of the type XYZ 3 and have derived a set of potential constants for the -CC1 3 group from the fr·equencies for CH 4 , ClCH 3 , and cc1 4 • This potential function and a C-H bond constant obtained in the derivation are sufficient to calculate the three parallel fre- 32 quencies of ncc1 3 and DCC1 3 fairly accurately. To treat FCC1 3 I used Rosenthal and Voge's formulae and potential function for the -CH 3 group plus the frequencies of FOU 3 to calculate a C-F bond constant. Then this was used with the potential function for the -Cc1 3 group to get the following values for the parallel frequencies of FCC1 3 : 341, 499, and 1380 cm.- 1 Another value 15 for the C-F bond constant, given by Sutherland and Dennison, led to the values 339, 493, and 12?6 cm.- 1 Now if one assumes ·that the intense Raman lines observed are fundamentals and that the calculation just sketched is reasonably accurate, then one can assign 351, 536 and 1072 ( or 1540 ) cm. -1 to the non-degenerate vibrations and 246, 398, and 838 cm.- 1 to the doubly degenerate vibrations. The fundamental frequencies of HCC1 3 and of DCCl3 are also useful in making the assignment. For HCC1 they are 16 263 (2), 3 366 (1), 668 (1), 768 (Z), 1222 (2), and 3025 (1), where the numbers in parentheses are the degeneracies, obtained from polarization measur·ements. For DCC1 3 they are 17 262 (2), 366 (1), 651 (1), 736 (2) , 908 (2), and 2256 (]} cm.- 1 Four of the fre- quencies are shifted but little in going from HCC1 3 to DCC1 3 , and hence it seems reasonable that approximately these same frequencies should appear for FCC1 3 . 536 (1), and 838 (2) cm.- 1 These are 246 (2), 351 (1), The two highest frequencies for the chloroforms apparently involve mainly the C-H bond, since they depend so markedly upon the ~ass of the hydf ogen . The frequencies for Fcc1 which involve mainly motions of the F against the CC1 3 3 pyramid are pro.bably 398 and 1072 ( or 1540) cm. -l Of these, the firat consists of bending and the second of stretching the C-F bond, since the former is degenerate and the latter non-degenerate. 33 1072 cm.-1 may possibly be an overtone of 536 cm.- 1 , but the exact agreement with 2 x 536 is suspicious. The most intense Raman line, 536 cm.; 1 should be assigned to symmetrical stretching of the C-Cl bonds, and it may be noted that the frequency in the HCc1 3 spectrum with which we correlate it (668 cm.- 1 ) is also assigned to this mode of vibration. 16 The only other non-degener- • 1 a t e f requency 1• s 3hl v cm. -l ; 1·t mus t c orrespon d t o th . e symme t r1ca bending of the C-Cl bonds. As for the remaining frequencies, . 246 and 838 cm.,-1 they may be assigned to degenerate bending and stretching of the C-Cl bond, respectively. It must be emphasized that this characterization of the modes of vibration is very approximate. Entropy from Raman Spectrum and Elec.tron Diffraction The translational entropy was calculated frorr; the SackurTetrode equation. The computation of the rotational entropy was made with the electron diffraction data of Brockway 18 for monofluorotrichloromethane, assuming the molecule acts as a rigid, symmetric top. The effect of the isotopes of chlorine was taken into account merely by using the average atomic mass for chlorine; such a procedure in the case of carbon tetrachloride yields practically the same result as the more refined n,ethod used by Lord and Blanchard . 12 The product of the three principal rr.oments of inertia is (6.14±0.23)x 10- 113 gm • . cm. 2 A symmetry nurr.ber of 3 was used, in accordance with the discussion of Lord and Blanchard 12 regarding the effect of isotopes on the symmetry number. The equation for the translational and rotational entropy, neglecting nuclear spin, is then Strans. + rot.=- 21.15 + 18.300 log 10 T - 4.5750 log 10 Patm. 34 The vibrational entropy was calculated from the Raman frequencies, using the equation S =. R£ P..:{x/(eXl- 1) - ln(l - e-x,.,)J, vib. "" where x"-:::: hcti'✓kT and P.t is the degeneracy of the i th fundamental frequency. This relation is derived on the assu~ption that the normal modes of vibration are harmonic oscillators weakly coupled. Tables of the function under the summation sign may be found in several places. 19 No allowance was made for the different symmetry types due to the presence of isotopes of chlorine, for the isotope shift should be small. The results are presented in Ta~le VIII. The values of all natural constants used ar·e those given in the In t ernational Critical Tables. 5 Table VIII Entropy of Monofluor otrichloromethane at 298.10° K. and 1 Atmosphere from Spectroscopic and Electron Diffraction Data. S, cal./deg./mole Translational and rotational 66.43 Vibrational 7.67 Total (neglecting nuclear spin) 74.10 Between the 11 calorimetric" and "spectroscopic" values for the entropy there is a discrepancy of 2.18 cal./deg./mole. Probably the chief reason for this divergence is the poor extrapolation below 60° K. No definite conclusion can be drawn as to whether or not there is agreement until the heat capacity measurements are extended to lower temperatures. There is, of course, the possibility that the Raman spectrum was incorrectly analyzed. However, it is only necessary to know the fundamental frequencies and their degeneracies. The highest non-degenerate frequency 35 can be either 1072 or 1540 cm. -1 without much change in the entropy. Acknowled gements I wish to thank Professor Don M. Yost and M:r. C. s. Garner for their cooperation in this research, Dr. Hugh Huffman for the loan of a standard thermocouple and for many helpful discussions, Dr. Kenneth Pitzer for information about the calculations, Ivir. Newton for asistance in constructing apparatus, and my wife, Marjorie Osborne, for help in drawing diagrams and graphs. Summary An apparatus for vapor-pressure measurement and low temperature calorimetry has been assembled. The heat capacities of monofluorotrichloromethane have been measured between 60° K. and the b. p. The heat of fusion at the m. p., 162.54:t: 0.05° K. (-110.56° c.), is 1641.8± 3.2 cal./mole. The vapor-pressures between 5 and 72 cm. are well represented by the equation log 10 P (cm.)::; 8.21671 - (1610.04)/T - 0.0030743 T. The b. p. from the vapor-pressure equation is 296.88± 0.05° K. (23.78° C.) and the heat of vaporization, including a Berthelot correction of -220 cal./mole, is 5905+ 60 cal./mole. The entropy from the cal6rimetric data, making an extrapolation below 60° K. amounting to 10.12 cal./deg./mole, is for the ideal gas 71.92 ±1.5 cal./deg./mole at 298.10° Kand 1 atmosphere. The Raman spectrum has been photographed and the following shifts observed, with the relative intensities indicated: 246 (6), 351 (6), 398 (6), 536 (10), 838 (4, broad), 10'72 (2), -I and 1540 (o)c.m. An assignment of the frequencies has been made 36 and with the aid of electron diffraction data the entropy of monofluorotrichloromethane at 298.10° K. and 1 atmosphere, neglecting nuclear spins, calculated to be 74.10 cal./deg./mole. References 1 Gibson and Giauque, J. Am. Chem. Soc. 45, 93 (1923); Giauque and Wiebe, ibid. 50, 101 (1928;. 2 Giauque, .J ohnston, and Kelley, ibid. 49, 236? (1927). 3 Giauque, Buffington, and Schulze, ibid. 49, 2343 (192'7). 4 Southard and Andrews, J. Franklin Inst., 2QJ., 323 (1929). 5 International Critical Tables, McGraw-Hill. Book Co., New York, ( 1926), vol. I. 6 Cath, Comm. Phys. Lab. Univ. Leiden, 152d, 1918. 7 Ruff and Keim, 8 Swarts, Bull. Acad. Roy. Bel.g . (3) 24, 309 (1892). 9 International Critical Tables, vol. III, p. 248. 10 Lewis and Gibson, J. Arr.. Chem. Soc. 39, 2565 (191?). 11 Kelley, Parks, and Huffman, J. Phys. Chem.~~' 1802 (1929). 12 Lo~d and Blanchard, J. Chem. Phys. 13 Rosenthal and Murphy, Rev. Mod. Phys.~, 317 (1936). Wilson, J. Chem. Phys., g, 432 (1934) 14 Rosenthal and Voge~ ibid.!, 134 (1936). z. anorg. allgem. Chem. gg_!_, 245 (1931). Voge and Rosenthal, ibid. i, 707 (1936). 1, 137 (1936). 15 Sutherland and Dennison, Proc. Roy. S oc. 148, 250 (1935). 16 Kohlrausch, "D.er Smekal-Raman Effekt, 11 Springer, Berlin (1931) p. 116. 17 Wood and Rank, Phys. Rev. !§, 63 ( 1935). Redlich and Pordes, Naturwiss. gg, 808 (1934)~ 18 Brockway, J. Phys. Chem. 41, _1 85 ( 1937). 19 Gordon and Barne s ,, .I. C.herr... Phys . l, 297 (1933) . 37 Propositions Submitted for Oral Examination 1. In calibrating a set of fractional weights against a brass standard to an accuracy of 0.01 mg. it is unnecessary to make buoyancy corrections, provided that when the weights are used the buoyancy corrections are applied as though the weights vvere brass. 2. Precise measurements of specific heats of substances with low vapor-pressures can better be made by the adiabatic method than by the method in which the temperature of the environment is kept constant. 3. For precise determination of temperatures and temperature intervals, strain-free resistance thermometers are superior to therrr. ocouples. 4. The method for auto-calibrating a White double potentiometer described by the manufac t urers of the instrument (Leeds and Northrup) cannot be used without modification. 5. Heat capacity data are of little value for determining symmetry numbers of molecules. 6. The discrepancy of 7.2 calories per degree per mole between the spectro!copic and "third law" values for the entropy of tetramethylmethane at its normal boiling p oint may be due to (a) "False equilibrium" in the crystal at the lowest temperatures reac:hed in the experiments. ( b ) Lack of free rotation of the methyl groups. (c) Inadequacy of the symmetry number approximation. 7. The Nernst Heat Theorem should be regarded as a useful generalization w~th some exceptions rather than as "always ir~elevant and useless." 8. Graduate students. working on structural, physical, or inorganic chemistry should be encouraged to take a course in qualitative organic analysis. A rea_Eonable assignment_ for the Raman,.., shifts of liquid CC1 3 F is ~, : 351 cm. ; J 3 " 536 j ~~=10 72 ( or perhaps 1540) -v V,_ - 246 ~ ; Vi( ' 838 ; -Ve.,= 398 ~ ,_, where 11, , -i)'..3 , and V0 are non-degenerate and -i7;).., , V'f , and 1/" doubly-degenerate vibrations involving mainly bending of the C-C l bonds, stretching of the C-Cl bonds, and stretching of the C-F bond respectively. 9. (1) Aston and Messerly, J. Arr::. Chem. S oc. 58, 2354 (1936). (2) Fowler and Sterne, Rev. iviod. Physics 1, 707 (1932).