The Viscosities of Normal and Isopentane as a Function of Pressure Citation Day, Ralph Kohlrausch (1930) The Viscosities of Normal and Isopentane as a Function of Pressure. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/278j-f910. https://resolver.caltech.edu/CaltechTHESIS:08142024-213625177 Abstract No abstract. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Physics) Degree Grantor: California Institute of Technology Division: Physics, Mathematics and Astronomy Major Option: Physics Thesis Availability: Public (worldwide access) Research Advisor(s): Millikan, Robert Andrews Thesis Committee: Unknown, Unknown Defense Date: 1 January 1930 Record Number: CaltechTHESIS:08142024-213625177 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:08142024-213625177 DOI: 10.7907/278j-f910 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 16639 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 15 Aug 2024 21:09 Last Modified: 15 Aug 2024 21:17 Thesis Files PDF - Final Version See Usage Policy. 13MB Repository Staff Only: item control page

THE VISCOSITIES OF NORMAL AND ISOPENTANE AB A FUNCTION OF PRESStroE THESIS BY RALPH KOIIT..RAUSCH DAY In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy, California Institute af Technalogy, Pasadena California, 1930. 1 This thesis contains preliminary results obtained in an investigation on the "Accurate Comparison by Viscosity Methods developed by Milli"kan of the Viscosities and hence of the Mean Free Paths of Ring and Chain Hydrocarbons of the same Molecular Weight and Composition" listed as Project No. 12 of American Petroleum Institute of Research. assistance in this work has been received Financial from a research fund of the American Petroleum Institute donated by John D. Rockfeller. This fund is being administered by the Institute with the cooperation of the Central Petroleum Committee of the National Research Council. The problem here at hand involves the measurement of small differences in viscosities as well as their actual value with the greatest precision possible. The only type of apparatus known to be capable of such precision is the rot~ting cylinder apparatus such as was used by Millikan and his collaborators at the University of Chicago to determine accurately the viscosity of air which was required for the oil drop e·xperiment. The Chicago apparatus was therefore, copied in its essential features with the added feature allowing the apparatus to be heated as a whole to tempera-tures presumably up to 400° centigrade. The isomers of pentane were picked for two reason~; first, because three isomers exist and should be obtainable in high purity. These three compounds are identical in chemical formula and molecular weight, differing only in the 2 way the atoms are eonnected together in the molecule. The organic chemist assigns structural formulas to them as follows:Normal pentane:- CH3-CH2 -CJ½-CH2 -CH3 / Is opentane:- CH3-CH2-CH CH ·3 'CH 3 CH3 I Neopentane: CH3-C-CH3 I CH3 and it is believed that the actual shapes of the molecules are similar to these structural formulae; i.e. the normal pentane is supposed to be a long narrow molecule and the neopentane, a symmetrical one, the isopentane being something intermediate. Unfortunately a sample of pure Neopentane has not yet been obtained so that the measurements here presented are only on the first two of these isomers. The second reason for selecting the pentanes is because their boiling points are near room temperature so that viscosities can be measured at pressures right up to their saturated vapor pressures. Measurements were therefore made on each compound at all pr essures ranging t he up to ~seturated vapor pressure at 25°Centigrade at which temperature all measurements were ma de. Fig. l APPARATUS The apparatus will be described in detail only where important changes were made from the Chicago rotating cylinder apparatus which has already been described in full by Gilchrist1 , Harrington2 and others. Considerable care was taken in the design of the apparatus so that it could pe made to function also at high temperatures. The des- cription of the apparatus will include these features even though the measurements here presented were all made at 25° centigrade and no attempt has as yet been made to heat the apparatus. Fig. l shows a schematic diagram of the apparatus and Fig. 2, the set up as it was. i ' 1 I I I Fig. 2 1. Phys . Rev. l, 124(1913) 2. Phys . Rev. B, 738(1916) • 5 directly in its final state. Before the final out was taken on the lathe, its surface was equally accurate with the above. The final cut was taken on the insiae of the cyl inder and as follows:- After a good cut was taken on the outside of the cylinder giving it the final uniform machined surface, an iron cylinder was fastened rigidly in the lathe and an inside cut taken on it of exactly the same diameter as the outside diameter of the bronze cylinder. Without removing the iron cylinder from the lathe, both cylinders were cooled with carbon dioxide snow and the bronze one slipped inside the iron one, which was easy when cold because bronze ha s a higher expansion coefficient than iron. The final out was no w taken on the inside of the bronze cylinder while a flame tight fit. was played on the iron cylinder to insure a To remove the bronze cylinder, carbon dioxide snow was used again. In this way there was no chance of changing or hurting the uniformity of the sus,pended cylinder's outer surface and a uniform wall thickness is also assured. The guard cylinders were accurately turned to the same diameter as the suspended cylinder and the gap separating them from the suspended cylinder above and below is about 1/2 mm. The rotating cylinder is accurately centered with respect to the guard cylinders because the top and bottom ball bearings holding the former are centered on the support rods holding the guard cylinders. pivoted on dry ball bearings. All moving par~s are Oil had to be avoided because of the temperature and also because of its solubility in hydrocarbon vapors. 6 To enable the instrument to be adjusted in the cold and then heated up without losing its adjustment, all parts were carefully compensated :tor thermal expansion, metals are matched throughout so that everything will expand together on heating. The Suspension:- The suspension used is the same as Harrington and others found best; namely, an uncoiled watch hair-spring of which the Elgin Watch Co. kindly supplied us a variety of samples. The size selected for this work is 1.8 x 5.8 mils in cross section .and is used at a factor of safety from the breaking weight of not much over two. ftrst strung up there was a rapid drift of zero readjustment of the suspension to its new When due to conditions. This drift had died down to a very small rate in a month's time, but is perceptible even now six months since it was installed • . Due to the small size of the suspension, the natural period of oscillation of the suspended cylinder about its vertical axis is very large - namely 198-1/3 sec. o·n this suspension. We have the elementary formula relating this period {P), the moment of inertia (I) and the sensitivity {S) {angular deflection per unit of viscosity torque) which holds for any kind of a suspension that gives simple harmonic motion:- It is obvious that to get a high sensitivity, a long p eriod must be tolera ted. The moment of inertia has already been cut to a minimum by making the inner cylinder as light as possibl~. The magnetic damping device to be described later makes the taking of readings even with this long oscillation period _ quite easy. The Constant Speed Motor:- Harrington's apparatus was equipped with a chronograph . drive which gave an actual record of the speed ell of the time that it ran. For relative measurements, a . repvoduoible constant speed is desirable while its easy measurement is not so important and all.so for reasons of its convenience and probably greater accuracy a Benioff constant speed m0tor was adopted. This motor was designed by H. Benioff for use in driving the chronograpp drums on the Seismographs in use at the Carnegie Seismograph Station in Pasadena. Many thanks are due him and them for lending their design and helping in its construction. The speed controlling mechanism (Fig. 3) is a steel reed which is attached t0 a heavy base, and vibrates with a small permanent straight bar magnet fastened rigidly to the end of the reed and perpendicular to it. This small magnet acts as a plunger in and out of a coil of wire on each side of it. The alternating current which is set up in the one coil is amplified by means of a vacuum tube and fed back to the other coil to drive the reed. This alternating current . of exactly the frequency of the reed is nON further amplified with vacuum tubes and relays and fed to the motor poles. The reed is in a thermostated box so that its frequency is constant and the motor cannot but be in step with the reed. The error due to inconstancy of the motor speed is smaller than l part in 10,000 as estimated by Mr. Benioff. 8 Fig. 3. The motor has therefore one definite normal speed as determined by the frequency of the reed. speed This normal is about 2 revolutions per second and is, reversible. By m:,ans of gears, the motor can be made to transmit speeds which are exact multiples of the normal speed by the gear r~tio. The Magnetic Coupling and Worm Gear Drive:- Harrington's method of connecting his motor drive outside the gas chamber with the rotating cylinder inside the chamber was by using a train of gears and shafts going right through a barometric column • .Mercury has been considered intolerable in my apparatus because of high temperatures giving high mercury vapor pressures and because a liquid air trap cannot be used when condensable vapors are to be measured. To solve the 9 same difficulty in my apparatus, therefore, a magnetic coupling through the walls of the gas chamber was used and proved quite satisfactory. The device is shown in Fig, 4. The inside bar or soft iron which follows the electromagnets around is blocked out in brass partly to give inertia to the inside system and also to stop any possible hunting between the systems. Hunting would produce eddy currents in the brass which would tend to damp it. In starting the motor some care was necessary to keep from giving the magnetic coupling too great or too sudden an acceleration at which times the inside rotor would be lost and it was necessary to start over again. Once going, the coupling was never lost. Fig. 4. 10 Since the motor runs normally at about 2 revolutions per second and it is desirable to have the outer cylinder rotate at about l rev. pe~. minute, a worm gear drive of ratio 120:l was used and was put inside the vacuum chamber. The worm gear and rotating cylinder are the only parts going at slow speed so that any jerkiness due to the motor ,or the magnetic coupling, is ironed out by the high angular mementum of those systems and the outer cylinder ·g ets a very uniform The worm gear drive as everything else inside the rotation. chamber, must run witho~t oil, so the worm gear was made of cast iron and the worm itself was ground out from a piece of Agate. A little graphite sprinkled oYer the worm is its only lubrication. This lower section of the apparatus containing the magnet ic coupling, worm gear drive and thre·e ball bearings is separated from the main apparatus by a water jacket and kept cool. The upper main central ball bearing is the only bearing which must run hot. The lower cold section is in a soldered brass housing involving two set-in rubber gaskets and a flat paper gasket. Duca was applied afterwards to make a sure seal. The Magneti ~ Damping Device:- Much effort was spent in trying to damp the oscillations of the suspended cylinder by means of a strong stat ionary magnetic field through the walls of the suspended cylinder. This was abandoned because the hi gh fields necessary are cumbersome and involve many other difficulties . The bronze suspended cylinder itself was found to be quite magnetic, for one thing. 11 Finally,a rotating three phase field has been adopted as an eddy current twisting device and has proved itself to be a perfect and very convenient method of damping am. is also free from any theoretical objections which may be raised with respect to a stationary field. There is no chance of permanently magnetizing any iron or steel parts of the instrument with aJl.ternating current and when turned off there is: no question of any residual magnetism. device is shown in Fig. 5. Fig. 5. The 12 Three coils of 500 turns each of #29 insulated copper wire are placed symmetrically 120° apart around the outside apparatus housing, connected in a closed delta and through three resistances of 300 ohms each to the 110 volt-3-phase line. This :rermits a current of about 0.2 amps. to flow. A reversing switch r:e·:rmits reversing or cutting off the field at any time. The device is really, as before mentioned, a twisting device and it can be used to bring the cylinder to rest in any position desired. There is no kick of any kind when the field is thrown on or off. There is no measureable heating due to the ene1·gy dissipated in the apparatus walls. This improvement over Harrington's apparatus makes the taking of readings very much simpler and easier so that successive readings caa be taken at ten minute intervals. This means that the ti me during which the gas is in the apparatus, previously evacuated is a minimum and therefore the impurities introduced due to adsorbed gases in the apparatus and small leaks is alro a minimum. A new apparatus housing of brass was constructed to repiace the welded steel one for this first work at room temperature which permits the rotating field to pe.netra te through the casing to the inner cylinder. I firmly believe that working at high temperatures will not eliminate the possibility or using this rotating field device although this has not yet been worked out. 13 DEFLECTION MEASUREMENTS The defle,ctions were measured by the usual telescope and scale method. The scale is a meter stick of maple wood with millimeter scale divisions. estimated. Tenths of a millimeter were The mirror to scale distance is about 214 cm. The scale is ~erpendicular to the line from the mirror to the exact center of the scale. Sum.mary of Apparatus Dimensions and Constants Length of rotating cylinder 18¼ in. Inside diameter of rotating cylinder - 4.690 in. Length ot· ee.ch guard cylinder - Length of suspended cylinder - 10-1/16 in. Outside diameter of suspended cylinder Weight of suspended cylinder Length of suspension 3-7/8 in. 4.218 in. - 478.5 gms. 9-3/8 in. Period of oscillation of inner cylinder - 198-1/3 sec. TEMPERATURE CONTROL This work was done in a sub-basement room which was thermostated at 25° centigrade. The variation of temperature was never greater than 0.2 degrees and during these neasurements, temperature control was mostly much better thmthis. vaJi.ues of Sutherland's constant (C) for air and isopentane as given in the International Critical Tables are:air:- C = 120 isopentane :- C ~ 500 The 14 It is presumed that the value of "C" for normal :[Bntane is similar to that of isopentane. Sutherland's equation may be written as follows to compare small differences ip temperature:- 1'+ C ( T+ C )¾ \ 1 I Putting T = zga°K, we find that 0.2° centigrade changes the viscosity of air 1 part in 1800 and of pentane 1 part in 1300. The temperature control therefore may account for small deviations from the correct value but the accuracy in its wm st case is about equal to the accuracy of reading. PRESSURE. MEASUREil/.!ENTS As before mentioned, mercury could not be used in connection with the apparatus. A pressure gauge without mercury was therefore necessary and the pressure measurements should be independant of the nature of the gas or vapor, A Bourdon spiral gauge fulfills these requirements. Such a spiral as is shown in Fig. 6 was blown from PYREX and set up to twist a small galvanometer mirror with a scale about a meter away. 1.3 millimeters deflection on this scale represents a pressure difference of one mm. Fig. 6 15 or mercury between the inside and outside of the spiral. The inside of the spiral is connected to the apparatus system and the spiral itself is enclosed in a chamber eonnected to a separate system in which the air pressure is regulated at will and measured by means of a closed tube mercury manometer. The s~iral is thus used as a zero instrument capable of reading small deviations and pressure readings are taken directly on the mercury manometer. The final measurements are accurate to about 1 mm of mercury. THE ORGANIC COMPOUNDS The two isomers of pentane which occur in natural gasoline were obtained in a very pure state. Normal pentane was obtained directly from the Eastman Research Laboratory. Isopentane was found unavailable in a sufficiently pure state, so that our Chemistry Department undertook to synthesize it for this work. To test the purity of these compounds, the freezing points and the range of temperature ower which freezing takes place were measured. This physical property was chosen, first because it is the property which organic chemistry uses in general as a criterion of purity and further because the difference in the freezing points of the two isomers is 2s 0 c, whereas the difference in boiling point, for instance, is only s 0 c. Therefore, a reasonable check of freezing points and a small range of freezing should indicate high purity. A single junction copper-constantan thermocouple was used and calibrated roughly with liquid oxygen and solid CO 2 . Since 16 both freezing points are equally far in error from the measured values , it • is suspected that the calibration may be that far The chief interest was in the range of freezing so wrong. that no further effor ts were taken to check the calibration. The hydrocarbon was put into a smal.l Dewar flask . evacuated to the extent that complete freezing took place in • . about 20 minutes when the flask wit~ ·the liquid hydrocarbon, thermocouple and stirring rod inside, was immersed in liquid air. A time temperature curve was then plotted. follow:Freezing Meas. Point Int. Cri t. Tables N. Pentane - 132.8°c - 131.5°c Isopentane - 161.1°c - 159.7°c Difference The results Range of Freezing The introduction of the hydrocarbon into the apparatus involved sone. difficulties at first because of its solvent action on stopcock grease and waxes. Red sealing wax which is apparently insoluble, still underwent some decomposition when it was used a s a stopcock lubricant, especially when it was heated to turn at which time a yellow vapor was given off discoloring the P2o5 .tube. Phosphoric acid as a stopcock lubricant worked fine for a while, but would not last out the number of turns required for a set of mea surements. The phos phoric acid does not have the rubber-like quality necessary for standing pressure without slowly being pushed away until a break develops in the film around the cock and it leaks. 1'1 B • To For, -Voc. j To For. -Voc. Fig . ? • The arrangement shows in Fig . ? was finally adopted. An all metal-stopcock (B) using a single Sylphon- bellow and closing with a large angle tapered needle on a silver seat was finally introduced and found very satisfactory . This cook is used between the flask with the hydrocarbon and the apparatus system, and is connected to the glassware on each side with a copper- PYREX seal . To isolate the apparatus system , a large glass cook (A) is used with the usual grease , but this need only be turned at the beginning of the run after which it simply presents a grease surface to the system. The hydrocarbon will be ab sorbed easil y and probably in considerable quantity in this grease yet since the grease has no appre c iable vapor pressure 18 by itself, it is not conoeivable how the organic vapor oan be in anyway contaminated by being absorbed and evaporated any number of times by the stopcock grease. Before going from one compound to another, this cock, connecting the pumps and another smaller one connecting the :MoLeod gauge to the apparatus were carefully cleaned and lubricated with fresh grease. In addition to these two grease stopcocks, other parts which could absorb hydrocarbons by solution were several rubber gaskets and a little soft wax used to plug up leaks. The same argument as to contamination holds here and it is held that, the compounre introduced were not in anyway contaminated by the apparatus itself. Contamination of the vapors is still possible in a variety of ways. Air leaks in the apparatus have been the greatest source of trouble. The apparatus, all told, has a tremendous surface area inside, so that adsorbtion on these walls of any kind of gas is easily possible. Stacy* speaks of a "Hydrogen Effeet" in his measurements on air at low pressures, this being presumably due to adsorbed Hydrogen in the apparatus. Troubles of thts nature will be described later in discussing the measurements. EXPERIMENTAL PROCEDURE The equation used which gives the viscosity co-efficient in terms of instrument constants and the deflection angle of the inner cylinder assumes a motion of the gas between *Stacy - Phys. Rev. (1923} 21, p 245. 1<3 the walls of the cylinders due only to the motion of the rotating cylinder. At times when this condition of a "static" gas was not satisfied such as when there was a large leak or if some condensed hydrocarbon was evaporating inside due to a recent fast reduction in pressure from a saturated vapor, the deflected position of the inner cylinder was irregular and a reading quite impossj_ble. procedure was therefore as follows: The experimental The pe!ntane in its flask is frozen by immersing this in liquid air . The flask with the frozen pentane is easily sealed on as shown in the diagram and the whole system evacuated with metal stopcock {B) open and the liquid air still on the pen tane. When a good vacuum is obtained, stopcock (B) is closed and the liquid air ~emoved allowing the pmtane to warm up to room temperature. The apparatus is now isolated from the pumps by closing stopcock (A) and a small portion of pentane let in by opening stopcock (B), and again closing . it. The pressure :is. read and the motor driving the outer cylinder is started. The swing of the inner cylinder is arrested within a few millimeters of its deflected position using the rotating field damping device. Maxima and Minima are observed until exactly ten minutes after the motor was started, at which time, its direction is reversed and readings taken in the other direction. The deflected reading taken is the mean of the la.st three readings corrected by interpolation such that the middle one comes ?-1/2 minutes after the motor was started. If the right and left readings agree by a rough calculation and apparatus conditions 20 are all perfect, more pentane is let in and the procedure repeated. In most cases, one deflection in each direction was sufficient and the mean of these was taken to represent the viscosity at this pressure and this value plotted as a point on the curve. A run consists of a series o1r' measure- ments at increasing pressures after which liquid air was put on the pentane flask and measurements taken at decreasing pressures until it was all frozen out again. By going back and forth on the pressure in each run, it is possible to check up on any differences between tr.early" and '-"later" readings taken at the same pressure which would be caused by leaks in the apparatus or other sources of contamination which would vary with the time. In general, it will be noticed that "later" viscosities tend to be lower than earlier ones. The procedure regardless of the pressure range or the number of readings taken at a pressure was al ways to deflect in alternate directions for ten minutes each. This procedure was also followed on calibration runs on air and on hydrogen. The reason for this as well as computing the deflection mean for the 7-1/2 minutes after the motor was started, is to cancel. out the effect of the "nachwirkung" of the suspension. After the inner cylinder is deflected and damped, there is a small drift of the deflected mean position to increasing deflections with time due to slow adjustment of the suspension to its new conditions in the deflected position. By following the above procedure on readings and calibration both, the error due to this drift is a minimum. 21 COMPUTATION OF RESULTS The following diagram shows the method of measurements: T ~ lf 'JOO C M eo :X.-1 l C is the center of too scale and the foot of the perpendicular from the mir ror. measurements. It is the origin for all xo , ~l ' and ,:.r are the distances from C of the zero position, the deflection to the left and the deflection to the right respectively. G..{)8/L and e~ are the angles between MC and the beams of light to the three positions above. The d i stance MC is called ttd" in the computations. The results are calculated according to the followi ng formulae:- ~+ eo f}.TI.. -D=Clo 80= .Cl UA + () )L.== t o nor, -X:..,.e ..L t a n q-/"'tlO -I -I XA _ -, /::) 22 The standard formula for the rotating cylinder a pparatus(l) is:- ,r eI (bl- aj 1~ a1-.b 2 P 2S2-€ In all of these measurements a comparison is made between the deflection for the gas in question and dry air so that the relative viscosity equation reduces itself to:- Harrington's vallue for 1 air is used in the calcula tions. '? air {at 23°c.) : 1822.6 xio- 7 The formula for the variation of the viscosity of air with Temperature near 23°C is given by Prof. Millikan( 2 ): - ~T='?23o 0 + .000000493 ( T -23) This gives ~ 25 = 1832.5 xio- 7 RESULTS In the following table "p" is the pressure in mm. of mercury, "t" is the time in minutes between the reading and the time when the apparatus was shut off from the pumps. The calibration deflection on air at 25°c is taken as the mean of a calibration run immediately before and after the readings taken. 1: Harrington: - loc. cit. 2: Millikan:- Ann. der Physik (1913) 41 p.759 23 Normal Pentane: Isopentane: ✓e ==- 4 a ir °. 5101 ·0 : : 40. 5081 • : eir NORMAL PENTANE No. Run #1 Run #2 Run #3 7 p. t. xl xr 7 9 8 17 25 50 102 50 25 17 9 15 35 55 75 95 115' 135 155 175 l?.101 17 . .108 17.099 l?.090 17.068 17.088 17.086 17.079 17.085 14.964 14.9?0 14.982 14.976 14.965 14.967 14.979 14.979 14.971 676.98 677 .. 22 6-77.30 677.02 676.29 676.77 676.98 676.81 676.77 17.016 17.100 17.118 1'7.124 17.mW 17.069 17.048 17.058 17.078 17.083 17.08'9 14.908 14.994 15.014 15.021 14.976 14.954 14.935 14.940 14.948 14.958 14.956 673.97 677.51 678.40 678.64 679.02 676.08 675.23 675.56 676.16 676.49 676.57 17.080 17.058 17.046 17.041 17.038 17.030 17.062 I 17.080 14.912 14.882 14.881 14.865 14.852 14.855 14.856 14.890 675.43 674.34 674.05 6?3.40 673.28 673.12 673.85 674.97 9 10 11 12 13 14 15 16 17 18 19 0.38 1.3 20 23 24 25 26 51 146 250 200 100 25 8 14 34 54 74 94 114 134 154 174 199 224 27 28 29 30 31 32 33 34 200 300 402 479 484 455 350 250 28 63 88 210 230 300 345 365 21 22 4 9 /'Yl_ xl0 24 ISOPENT.ANE Run #1 Run #2 Run #3 1 xl0 7 No. p. t. xl Xr 1 2 3 4 5 6 7 8 2.9 11.5 25 50 101 50 11.5 30 55 75 95 115 135 155 175 17.860 17.882 l?.868 l?.858 17.850 l?.855 17.85'7 l?.855 15.124 15.142 15.141 15.118 15.112 15.111 15.108 15.110 69 6. 60 697.45 69?.12 696.42 696.13 696.22 696.20 696.20 9 10 11 12 13 14 15 16 17 18 19 20 2.1 ll.5 25 50 101 150 eo1 251½ 350 250 150 50 15 35 55 75 95 115 135 155 175 195 215 235 l?.939 17.954 17.946 17.942 l?.924 17.916 17.906 17.901 17.894 17.905 17.919 17.934 15.031 16.035 15.030 15.023 15.012 15.008 14.997 14.990 14.983 14.9?6 14.991 14.999 696.30 696.70 696.42 69 6.20 695.59 695.32 694.89 694.63 694.34 694.42 695.05 695.53 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 25 50 100 202 300 350 400 450 497 550 650 658 600 553 400 300 149 15 35 55 75 95 115 135 165 195 235 448 525 570 595 625 645 665 17.966 17.960 17.951 17.92g 17.910 17.908 17.903 17.901 17.871 17.868 17.8?4 17.867 17.855 17.875 17.892 17.89'7 17.931 15.003 14.995 14.986 14.975 14.956 14.946 14.945 14.932 14.924 14.919 14.911 14.903 14.902 14.906 14.915 14.934 14.957 696.28 695.gs 695.60 6g4.9l 6g4.12 693.87 693.?4 693.43 692.62 692.46 692.41 692.10 691.84 692.33 24 692.89 693.40 694.58 25 DISCUSSION OF THE ACCURACY OF THE RESULTS Since the measurements were made with the apparatus in a condition such that small leaks were constantly developing in a cracked soldered joint, some consideration is necessary to justify the experimental results. As mentioned before, the "runs" were all taken by going first up the pressure scale and then down without evacuating between readings. It is a safe presumption that any contamination due to a leak would be a continuous process with time and therefore it should not be possible to go up and down the pressure scale and check the results. In most ·cases viscosities on the up-scale half of the run are higher than those on the later down scale half although Run #3 on normal pentane for instance shows no regularity in this re~pect. The explanation of these lower viscosities after the pentane nas been in the apparatus for a longer time, is entirely unknown to me although in most cases the results are all within the limit of error so that no anxiety need be felt about it. It will be noticed that the largest deviations are at the very low and very high pressures. difficult pressures at which to work. These are the The apparatus, as mentioned before, has stopcock grease, rubber gaskets and other materials inside which tend to absorb the pentane. These absorptions produce the most trouble in the very low and very high pressure ranges. At low pressures, small quantities of evaporating or condensing vapors are large percentages of the total vapor present and at pressures near saturation, very 26 large amounts of pentane, even puddles, may accumulate. The evidence ~or this is in the large changes in pressure on standing which were noticed at the high pressures. During these changes of pressure, the deflections were always uncertain which is probably due to currents of vapor blowing on the inner cylinder. The reasons that the initial low pressure readings are often higp may be because of the displacement of air or other gases then adsorbed on the walls of the apparatus. Attempts at measuring the effect on the viscosity caused by small air impurities i a isopentane gave the result that up to 0.2% air made no detectable difference and larger percentages caused a rise in Viscosity al though the percentage rise is still much smaller than the percentage of air impurity. Air impurities therefore do not seem to affect the viscosity of the p~ntanes very much. Still the fact that air leaks increase instead of decrease the viscosity, cannot account for the previously mentioned fact that the viscosities on the return part of a run going to lower pressures tend to be lower than the early measurements. The errors due to fluctuations in speed of the motor o~ any imperfections in the apparatus, are believed to be within the accuracy of reading because of the consistency of readings taken on air and hydrogen. The air-hydrogen ratio was also checked to within the accuracy of reading. The accuracy of· reading of the deflections is estimated at 0.1 mm on the scale. Since deflections on the pentanes were about 150 mm., this 27 is 1 part in 1500 or 0.07%. The maximum error due to temperature fluctuations is of the same order. The measurements here presented may be said to be accurate to An impur ity might very easily get into the hydrocarbon due to docidental causes or through the medium of the apparatus. There is no reason to believe that there has been impurity introduced and every pre caution was taken I mention this only because it would be so to prevent it. easy to absorb an organic impurity and extremely difficult to detect it. For this reason the relative values on each compound are probably more accurate than the absolute values. PREVIOUS MEASUREMENTS The only previous work which was found on either of these compounds, is the measurements by Rappenecker (ref.1 on isopentane at 100°c and 212.5°c using the capillary tube m,ethod. His values are:- 1 - 885.1 xio- 7 at 100°c 1 = 1164 xio-7 at 212.50c Sutherland's constant (C) = 500 Extrapolating his values with the help of his Sutherland constant and Sutherland's equation gives i = 691 xio- 7 at 25°c. He used "Kahlbaum" isopentane and worked at a few centimeters of mercury above atmospheric pressure . This extrapolated value agrees with my high pressure value of 691.5 xio-7 at ref. Rappenecker - Zeit. Phys. Chem. 72, 695 (1910) 28 650 mm, whieh :ia a nice check though not much weight can be put on this as a check when one considers the different conditions under which the two experiments were performed. DISCUSSION OF RESULTS After all the measurements were made, they were plotted and it is quite apparent that a straight line fits best the points of each curve. It will be further noticed that the slopes of the lines on the two isomers are of the same order of magnitude (ratio:85/?6). The difference between the absolute values at any pressure is about 2.8% the isopentane having the greater viscosity. These values are as would be expected. The iso- pentane being a more compact molecule would be expected to have a smaller collision aroo.and consequently a longer mean free path and larger viscosity. The decrease in viscosity with increase of pressure on each compound seems to indicate an increase in the effective size of the molecule with increase of pressure. This fact and the apparent linear relationship need careful theoretical consideration. Kinetic Theory interpretation of this decrease of viscosity with increase of pressure and of the apparent linear relationship may lead to some interesting and new knowledge of the behavi or of gaseous vapors from the molecular viewpoint. MEAN FREE PATH The Landolt-Bornstein Tables* use the following * Vol. I. p.119 29 formulae to calculate viscosities into mean free path data:- L (' l1 Cm.) -= 1T I 8 • 0.I49 • -,J_Q_l_3....,J 2.,,.....5_0_ 1 i, where SL ::: _a__ . 7... 73 . B_ iT L - mean f'ree path; m . ~o = viscosity at o0 c; &1 :: molecular velocity; R -- gas constant; m = molecular weight. Since the measurements here presented are at 25°c, and Sutherland's constant has not been measured in this range, L will be calculated at 25°c, using '7. 25~ and J1 calculated at 25°c. di,5° = ; Li.so : • l98 • Ji = 8. 7634x 10 8 Z.34/ 4x /01..x '72.s-o which gives:normal pentane pressure L 25°c 10 mm 1.5853 x10-6 500 mm l.5?58 x10- 6 isopentane 10 mm 1.6306 xio- 6 650 mm 1.6191 xio- 6 30 SUMMARY ~he vapor viscosities of both normal and isopentane have been accurately measured by the rotating cylinder method at different pressures ranging from a few millimeters pFessure to the saturated vapor pressures and at 25° centigrade. Each isomer shows a linear decrease of viscosity with increasing pressure, the total decrease in each case being slightly less than one percent. The difference between the absolute values of the viscosities of the isomers at any pressure is about 2.8%, the isopentane having the higher viscosity, which is consistent with what would be expected from the organic chemists structural picture of these isomers. The measured values are:Normal Pentane pressure viscosity (c.g.s.units) 10 mm 677.1 xlo- 7 500 mm 673.0 x10-7 Isope.ntane 10 mm 696.4 xio-7 650 mm 691.5 x10- 7 These results may also be expressed as follows:normal pentane: rrc 25ex10 7 ; i so, pen tane: '1_ 25ex107 - 696.5-.00??p 677 .2-.0084p where p. pressure of the vapor in millimeters of mercury. 51 The Research was carried out under the personal direction of Prof. R. A. Millikan. Credit is due Mr . Pearson and his shop for help in design and accurate construction of the apparatus and thanks are expressed to Mr. Wm. Bleakney for a ssistance in the work.