LIBRARY UNIVERSITY OF CALIFORNIA. Gl FT OF LELAND STANFORD JUNIOR UNIVERSITY PUBLICATIONS STUDIES IN ELECTRICITY . 1 SOME OBSERVATIONS UPON THE CONDUCTIVITY OF A COPPER WIRE IN VARIOUS DIELECTRICS BY FERNANDO SANFORD, M. S. Professor of Physics in the Leland Stanford Junior University PALO ALTO, CALIFORNIA PUBLISHED BY THE UNIVERSITY 1892 PRICE, FIFTY CENTS LELAND STANFORD JUNIOR UNIVERSITY PUBLICATIONS NO. 1. THE TARIFF CONTROVERSY IN THE UNITED STATES, 1789-1833. With a Summary of the Period before the Adoption of the Constitution. By ORRIN LESLIE ELLIOTT, Ph.D. pp. 272. Price, $1.00. NO. 2. SOME OBSERVATIONS ON THE CON- DUCTIVITY OF A COPPER WIRE IN VARIOUS DIELECTRICS. B y FERNANDO SANFORD, M. S. pp. 44. Price, 50 cents. ADDRESS THE REGISTRAR, Palo Alto, California. LELAND STANFORD JUNIOR UNIVERSITY PUBLICATIONS STUDIES IN ELECTRICITY SOME OBSERVATIONS UPON THE CONDUCTIVITY OF A COPPER WIRE IN VARIOUS DIELECTRICS BY FERNANDO SANFORD, M. S. Professor of Physics in the Leland Stanford Junior University PALO ALTO, CALIFORNIA PUBLISHED BY THE UNIVERSITY SEPTEMBER, 1892 SOME OBSERVATIONS UPON THE CONDUCTIVITY OF A COPPER WIRE IN VARIOUS DIELECTRICS It is probable that no known electrical phenomenon offers greater difficulty of explanation at the present time than the phenomenon of metallic conductivity. Faraday has given reasons for believing that it is essentially re- lated to the phenomenon of static induction, and Maxwell's theory would make both phenomena alike dependentupon an elastic fluid, or ether, which permeates both the con- ductor and the dielectric around it. According to this view, an electric charge consists of a displacement of this elastic fluid, which, on account of its elasticity, tends to return to its original position, while a current consists of a displacement of this same fluid throughout the entire length of a closed conducting circuit, in which case the tendency to return to its original position does not appear. It would, accordingly, appear that this electric elasticity of the ether disappears in conductors, or at the bounding surface between conductors and dielectrics. Since appar- ently the same inducing force may induce different charges in different dielectrics, it seems necessary to assume that the elasticity of the ether varies in different dielectrics. Since, also, the ether displacement for a given force will be inversely as its elasticity, it follows that the 3 4 SOME OBSERVATIONS UPON THE specific inductive capacity of a dielectric will be pro- portional to the reciprocal of the elasticity of the ether in this dielectric when subjected to electric stress. It is now well known that light waves are caused by electric stresses in this same ether, and it seems necessary to assume that they, too, depend upon this same electric elasticity, and, accordingly, that a fixed relation should exist between the velocity of light in a dielectric and the specific inductive capacity of the dielectric. This relation, as determined by Maxwell and Helmholtz, would make the optical index of refraction proportional to the square root of the dielectric constant. While it is true that this relation has not been extensively verified by experiment, it is also true that different determinations of the dielectric constant have given very different results for the same substance. So far as I am aware, it has been assumed that the electric conductivity of a wire or other metallic conductor is uninfluenced by the nature of the dielectric in its field of force,' and depends only upon the nature and temper- ature of the conductor. The close relation which seems to exist between the current in a conductor and the phe- nomenon of induction in a dielectric, and the apparent fact that the ether displacement in a conductor carrying a current is caused by a lateral stress communicated to it by the ether in the dielectric surrounding the conductor, has, however, made it seem to me probable that the amount of this displacement for a given force might also be modi- fied by the nature of the surrounding dielectric, or that the conductivity of a given wire might vary in different dielectrics. More than a year ago, I made some experiments to see if this phenomenon of variable conductivity could be CONDUCTIVITY OF A COPPER WIRE 5 observed, but with only negative results. Wishing to test the matter more fully,! had made a piece of apparatus by means of which I have been able to observe the suspected pheno- menon in liquid and gaseous dielectrics, and by means of which I hope to accumulate much more valuable data concerning the action of dielectrics upon conducting wires. The apparatus used consists of a cylindrical copper tube 120 cm. long and 2.5 cm. in internal diameter. The ends of the tube are closed air tight by copper plates, which are provided with stopcocks for filling and empty- ing the tube. To the inside of one end plate is fastened a copper wire 1 mm. in diameter, which passes length wise through the center of the tube and out through an insu- lated opening in the other end plate, where it is soldered to a piece of lamp cord, made of many small wires, by which it is connected with the Wheatstone's Bridge. Another similar piece of lamp cord is soldered to the end of the tube through which the wire passes, and is likewise connected with the bridge. Midway between the ends of the tube is another tube 5 cm. long and 1 cm. in diameter entering it from the side. This tube serves to admit the thermometer by which the temperature of the interior of the large tube is measured. The current used for making the measurements passes one way through the tube and returns through the wire. By this arrangement, the entire field of force of the cur- rent to be measured is confined within the tube, and the whole of the dielectric concerned in its transmission can be changed at will. The measurements were made with a Hartmann & Braun combined resistance box and bridge, with bridge arms of 1:1000. The smallest resistance used was .1 ohm, 6 SOME OBSERVATIONS UPON THE which with the above combination, represented a resist- ance in the wire and tube of .0001 ohm. The galva- nometer used is the physiological galvanometer of DuBois Reymoiid, manufactured by the Geneva Society Construc- tion Company. It is provided with a concave mirror and ground glass scale, and the usual method of observing the deflection of a spot of light reflected from the mirror upon this scale was used. This deflection was plainly noticeable for a change of resistance corresponding to .0001 ohm in the wire and tube, and by reversing the current several times, it was possible to obtain a deflection for a change corresponding to one-fifth of the above. The measurements were accordingly estimated with a fair degree of accuracy to .00001 ohm. As the combined resistance of the wire and tube was about .0335 ohm, the average of any set of measurements was certainly not .03 of one per cent wrong. The zero method of measure- ment was adopted throughout. The current used was obtained from a battery of 32 silver chloride cells, usually through 100 ohms resistance between the battery and the bridge, and the strength of current through the wire and tube was always between the extremes of five and eight milli-amperes. The measurements were all made at the temperature of the room, and as this changed very slowly, only a few measurements could be made in a day. CONDUCTIVITY OF A COPPER WIRE 7 COMPARISON OF SOME LIQUID DIELECTRICS WITH AIR. The first set of measurements was made with air and wood alcohol as dielectrics. I thought that, on account of the high specific inductive capacity attributed to this liquid, it would be as likely as any to give results differ- ing from air. In this I was disappointed. After draw- ing a curve representing the resistance of the wire at different temperatures in air, the corresponding curve for wood alcohol was found to practically coincide with it. The wood alcohol was then poured out of the tube, but was not carefully drained off, and the tube was filled with petroleum. An increase in the resistance of the wire was at once noticed. Eleven measurements were made with this dielectric during two days, March 4th and 5th, the liquid being poured off six times and measurements made in the air in the meantime, but the same liquid being poured back each time. The curve representing these eleven measurements was found to indicate a resistance .00008 ohm greater than that represented by the curve for the air readings. As the average resistance of the wire and tube at the temperature at which the measurements were made was about .C3350 ohm, this difference represented about .27 of one per cent of the whole resistance. It was noticed that the liquid had a cloudy appearance, looking neither like wood alcohol nor petroleum, but as "the wood alcohol had caused no varia- tion in the resistance, I supposed this increase of resist- ance to be due to petroleum. 8 SOME OBSERVATIONS UPON THE On the following day, March 6th, the tube having been carefully drained out and dried was filled with pure petroleum and a neAv set of measurements was begun. These indicated at once a decrease of resistance in the wire. As this seemed to contradict the former measure- ments, I thought that some of the contact resistances about the apparatus were changing. Measurements were accordingly made at irregular periods from March 6th to April 5th, the battery and galvanometer being detached and used for other purposes and the apparatus being allowed to stand from March 14th to 28th. In all, 34 measurements were made in air and 22 in petroleum, the dielectric in the meantime being changed five times, and the temperature changes being from 13.6 C. to 26 C. The whole series was entirely consistent. Not a single reading made with the wire in either dielectric crossed the curve for the other dielectric. The average difference in the resistance of the wire in the two dielectrics was .00006 ohm, which corresponds to about .18 of one per cent of the whole resistance, the conductivity of the wire being that much greater in petroleum than in air. The curves made for these dielectrics are shown in Fig. 1. They were drawn as follows : The points representing the resistance at the different temperatures were platted on cross section paper, none of the points for either dielectric being omitted, and a line was drawn through them with a pencil and ruler as nearly as possible in the true direction of the curve, which did not vary percept- ibly from a straight line. The distance of each point from the line was then measured, and the algebraic sum of these distances divided by the number of the points taken was assumed as representing the true distance of the curve above or below the trial curve from which CONDUCTIVITY OF A COPPER WIRE 9 the distances were measured. The final curve was then carefully drawn parallel to the other curve at this distance above or below it, as the conditions required. The above measurements made it seem probable that the increased resistance with the first petroleum was due to the wood alcohol which was mixed with it. To make this certain, the petroleum was poured off, leaving a little in the tube, and the tube was then filled with wood alcohol. Thirty-one measurements with the wire in this liquid were made on April 5th, 6th, and 7th, the liquid being poured out twice in the meantime and twelve measurements made in the air. The thirty-one successive measurements made for the liquid represented an increase of resistance over the former curve made for air of .00007 ohm, while the twelve air measurements averaged only .000004 ohm from the former curve. This liquid, which consisted principally of wood alcohol and contained only a little petroleum, gave a resistance to the wire .2 of one per cent greater than it had been in air; while the liquid used a month before, consisting principally of petroleum with only a little wood alcohol, gave an increased resistance to the wire of .27 of one per cent. This peculiar effect of using two dielectrics which did not seem to, combine with each other was observed later with wood alcohol and benzine. On May llth and 12th another series of five measurements was made with wood alcohol which seemed to give the wire an increased resistance of .02 of one per cent, a series of seven meas- urements with the wire in benzine gave an increased resistance of .06 of one per cent, and a series of nine measurements when the tube contained nearly equal parts of wood alcohol and benzine gave an increased resistance of .15 of one per cent. 10 SOME OBSERVATIONS UPON THE The other liquids tested were absolute alcohol, ordinary 90 per cent alcohol, carbon bi-sulphide, distilled water, and a mixture of carbon bi-sulphide and oil of turpentine. The absolute alcohol gave the wire an increased resistance over air of .000064 ohm, corresponding to .19 of one per cent of the whole resistance. See Fig. 2. The measurements for 90 per cent alcohol were more unsatisfactory than for most of the others on account of a much wider variation than usual between the individ- ual measurements, both in case of the measurements in alcohol and in air. These measurements were made in two successive days from once filling the tube with alco- hol. Four of the air measurements were made before the alcohol measurements and nine were made afterwards. The average of the eight measurements in alcohol was greater than that of the thirteen in air by .000024 ohm, but the irregularity in both series was so great that but little value should be attached to these results. The cause of this irregularity is unknown to me. The curve for carbon bi-sulphide differed very little from that made for air. In this case, the air curve was made before filling the tube with the carbon bi-sulphide. Eleven measurements were made with this liquid on May 9th and 10th. The seven measurements made on the 9th all indicated a decreased resistance in the wire, though of a very small amount, while the four measurements on the 10th indicated an increased resistance of a somewhat greater amount. It seems probable that some contact resistance in the apparatus had increased by a small amount in the meantime. The mixture of carbon bi-sulphide and oil of turpen- tine gave a decreased resistance to the wire of about .00003 ohm, or nearly .09 of one per cent. CONDUCTIVITY OF A COPPER WIRE 11 With distilled water the irregularity was also very great. When the water was first poured in it seemed to give an increased resistance to the wire, but this increase seemed to disappear slowly as the water continued to stand in the tube. This seemed to me to be due to an increased conductivity of the water caused by the solu- tion of salts of copper from the wire and tube. The following tables give the figures relating to the different measurements as preserved in my note book. It will be observed that these measurements were made at all times of the day, so that any difference of tempera- ture between the wire and the resistances used for com- parison would not always affect the result in the same direction. The same resistance box and thermometer were used throughout. DATE HOUR DIELECTRIC TEMP. RESIST- ANCE IN OHMS 1892 Feb. 22 Air 23.0 .03385 n 12 M. it 22.8 . 03380 a 2 P. M. l( 25.0 .03410 ft 5 " u 25.2 .03410 Feb. 23 9 A. M. a 19.1 03340 u 4 P. M. it 21.9 .03370 11 6 " a 23.0 . 03385 Feb. 24 8.30 A. M. \ it 17.7 .03320 12 SOME OBSERVATIONS UPON THE DATE HOUR DIELECTRIC TEMP. RESIST- ANCE IN OHMS Feb. 24 10 A. M. Air 18.3 . 03330 It 11.30 A.M. it 20.0 .03350 tt 3.15 P.M. a 22.0 .03370 tt 5 it 22.8 .03383 Feb. 25 10 A.M. tt . 18.0 .03327 Feb. 26 9.30 " it 19.5 .03347 a 3.20 P. M. tt 21.6 .03368 tt 3.45 " Wood Alcohol 21.1 .03366 a 5.20 " a 21.7 . 03367 n 9.00 tt 22.6 .03377 Feb. 27 4.00 " tt 21.0 .03365 Feb. 28 9.30 A.M. tt 18.0 .03328 Feb. 29 10 tt 19.6 .03343 u 2.30 P. M. tt 21.7 .03370 a 6.00 " tt 21.0 .03360 March 1 4.00 " tt 22.0 .03374 March 3 10.30 A.M. it 19.9 .03350 CONDUCTIVITY OF A COPPER WIRE 13 Poured out wood alcohol and filled with petroleum. Some wood alcohol left in the tube. DATE HOUR DIELECTRIC TEMP. KESIST- ANCE IN OHMS 1892 March 4 ii Petroleum and Wood Alcohol a 21.0 22.0 .03370 .03385 tt 4.00 P. M. it 22.6 .03388 tt 5.00 " it 23.0 .03392 March 5 9.00 A. M. n 17.7 .03330 u 9.30 " Air 19.0 .03340 a n 9.45 " 10.30 " Petroleum and Wood Alcohol Air 19.1 21.0 .03347 .03365 u 10.45 " Petroleum and Wood Alcohol 20.0 .03360 Disconnected apparatus and cleaned tube by draining out the liquid and forcing air through it with a bellows. DATE HOUR DIELECTRIC TEMP. 24.1 RESIST- ANCE IN OHMS March 6 7.45 p. M. Air .03418 " 7 8.15A.M. tt 21.0 .03390 a tt 9.30 " n 22.0 .03400 14 SOME OBSERVATIONS UPON THE DATE HOUR DIELECTRIC TEMP. RESIST- ANCE IN OHMS March 7 10.45 A. M. Air 23.1 .03418 li it 11.10 " a 23.4 . 03420 ( i a 3.00 p. M. a 24.7 .03431 n u 3.50 " a 25.0 .03434 a a 4.50 " t i 26.0 .03442 " 8 9.15A.M. i i 19.9 .03374 u u 9.50 i . 20.3 .03383 ii 10.30 " a 20.8 .03384 a 11.45 " Petroleum* 22.2 .03390 a 12.15 p. M. i i 22.0 .03390 a u 1.30 " u 22.7 .03400 u a 2.10 " u 23.0 .03406 a u 4.00 " ti, 24.8 .03426 " 9 9.00 A. M. a 20.0 .03370 u a 10.00 " a 21 2 .03386 a a 11.15 " t i 21.7 .03392 u 12.00 M. u 22.0 .03396 u a 3.00 p. M. a 23.1 . 03408 * Whittier, Fuller & Go's. 150 Fire Test " Star " Kerosene. CONDUCTIVITY OF A COPPER WIRE 15 DATE HOUR DIELECTRIC TEMP. EESIST- ANCE IN OHMS March 9 5.00 P. M. Petroleum 24.0 . 03415 " 10 10.00 A. M. Air 21.6 .03392 u a 11.00 " ti 21.7 .03394 tl It 1.15 P. M. u 21.8 .03395 It l( 4.00 n 20.1 .03374 11 10.00 A. M. n 20.6 .03382 " 12 2.10 P. M. 11 19.1 .03366 " 13 10.00 A. M. it 17.8 . 03350 n a 10.40 " it 18.9 .03360 " 14 10.00 < 11 20.3 .03380 . " 28 10.45 " it 16.3 .03330 a a 11.20 " it 16.3 .03330 a ti 4.00 P. M. ti 17.0 . 03340 " 29 10.30 A. M. ti 17.7 . 03348 " 30 9.30 " it 15.6 .03325 April 2 2.00 P. M. it 17.9 . 03349 tt tt 2.10 " Petroleum 18.9 .03355 it u 4.40 " ti 17.7 . 03343 16 SOME OBSERVATIONS UPON THE DATE HOUR DIELECTRIC TEMP. RESIST- ANCE IN OHMS April 2 5.30P.M. Petroleum 17.6 . 03340 " 3 9.50 A. M. n 13.6 .03293 n u 10.10 " a 14.0 .03298 it l( 10.50 " 1 1 15.0 .03310 it 11 12.10 P. M. t t 16.6 .03330 (i it 5.40 " i t 16.6 .03330 " 4 9.50 A. M. Air 15.1 .03320 n u 11.40 " 15.1 .03315 a u 4.00 P. M. a 18.9 .03364 a tt 4-10 " a 18.9 .03365 u u 4.45 " t ( 19.2 .03367 u u 4.50 " i i 19.2 .03367 u 5.30 " u 19.3 .03368 u u 7.20 " Petroleum 19.2 .03364 " 5 8.30A.M. u 13.6 .03298 a a 10.10 " a 14.0 . 03304 a a 11.00 " 1 1 14.6 .03310 CONDUCTIVITY OF A COPPER WIRE It The curves representing the above measurements for air and petroleum are given in Fig. 1. The petroleum was mostly poured off and the tube filled with Wood Alcohol at 11.35 A. M. DATE HOUR DIELECTRIC TEMP. RESIST- ANCE IN OHMS April 5 12.00 M. Wood Alcohol and Petroleum 14.8 .03308 it 12.30 P. M. tt 15.3 .03320 < . t< 1.45 " 16.4 .03334 i '. a 3.00 " it 17.6 .03352 a a 4.50 " it 18.4 .03363 a a 5.30 " u 18.9 .03368 a it 6.00 " 11 19.1 .03368 ' ft :- ' 4< 7.35 " it 19.0 .03364 " 6 8.30 A. M. u 16.1 .03333 < > it 9.45 " i , 17.1 .03345 a a 10.45 " it 18.3 .03363 a a 11.45 " tt 18.7 .03370 U (t 12.35 P. M. U 19.0 .03372 (l tl 1.45 (t 19.5 .03376 t( It 3.30 " tt 20.7 .03395 18 SOME OBSERVATIONS UPON THE DATE HOUR DIELECTRIC TEMP. RESIST- ANCE IN OHMS April 6 4.40 P. M. Wood Alcohol and Petroleum 21.5 .03405 n n 5.30 " Air 21.9 .03395 tl it 5.515 " tt 21.6 .03392 it It 7.40 " a 21.0 .03386 u a 9.20 " 15.3 .03318 u u 10.00 ' a 18.6 .03355 7 7.25 A. M. a 15.2 .03316 u u 9.15 " tt 18.4 .03355 tt U 10.05 tt 19.0 .03362 tt tt 10.20 " Wood Alcohol and Petroleum 18.0 .03357 tt 10.50 " it 18.9 .03366 tt u 11.30 " it 19.1 . 03368 tt a 12.30 P. M. i i 19.6 .03373 u u 1.45 a 20.0 . 03381 it, U 4.00 " Air 23.0 .03414 n u 5.30 " a 22.1 . 03405 tt tt 5.45 " a 24.5 . 03434 it a 8.00 " u 21.1 .03390 CONDUCTIVITY OF A COPPER WIRE 19 It will be noticed that the first measurement after pour- ing the wood alcohol into the tube which had just con- tained the petroleum gave the same resistance that the wire had shown in petroleum, and that not until the fourth measurement, made over three hours after the wood alcohol was poured in, did the true resistance in the mixed liquid seem to be reached. This phenomenon was frequently observed by pouring off a liquid dielectric and making a measurement at once, and in nearly every case the resistance seemed to be either that given by the liquid dielectric at that temperature, or one between that of the liquid and the air resistance at the same tempera- ture. From this, I was led to think that the thin layer of this dielectric in contact with the wire had a marked influence upon the wire's conductivity.* For this reason when the liquid was poured out the tube was usually carefully drained out and dried by blowing air through it with a bellows or drawing it through with a filter pump. After the last measurement in air given above, the tube was filled with ordinary 90 per cent alcohol, and the following measurements were made. * Since the above was in type, my attention has been called to the re- marks made by Prof, von Helmholtz before the recent meeting of the British Association at Edinburgh, as reported in "The Electrician " of Aug. 12, 1892. The substance of the report, so far as it applies to the question in hand, is that a mercury column in a glass tube may have a greater re- sistance due to the film of air adhering to the inside of the tube, even when the thickness of this film does not exceed .0005 of a wave length of light, and that the lowest resistance of the column may be obtained by allowing a drop of petroleum to spread itself over the inside of the tube. It has seemed to me possible that this phenomenon may be re- lated to the one described above. 20 SOME OBSERVATIONS UPON THE DATE HOUR DIELECTRIC TEMP. RESIST- ANCE IN OHMS April 7 9.20P.M. Alcohol 19.0 .03365 " 8 7.30 A. M. it 14.0 . 03304 (I It 8.40 " 1 1 15.1 .03325 11 S 9.40 " i i 16.6 .03341 it 10.50 " 1 i 17.9 .03355 tl it 12.30 P. M. 19.0 .03370