IXDUSTRIAL AND ENGINEERING CHEMISTRY
338
ing individual samples in air, or by oxidizing a larger sample and using the method of mixtures, was most satisfactory, When the investigation uses synthetic mixtures or ignited coal ashes, reduction by addition of computed amounts of carbon would be the most direct method. All the methods lack assured precision and chemical analyses of the melts are necessary; determination of ferrous oxide is generally all that is required except for the most exact work. Acknowledgments These investigations were conducted a t the Pittsburgh Experiment Station of the U. 8.Bureau of Mines under the immediate direction of P. Xicholls. The chemical analyses were made by W. H. Frederic under the direction of W. A. Selvig.
VOL. 7, NO. 5
Literature Cited (1) Bowen, N. L., and Schairer, J. F., Am. J . Sei., [ 5 ] 24, 179 (1932). (2) Matsubara, A., Trans. Am. Inst. Mining Met. Engrs., 67, 3-55 (1922). (3) Mellor, J. W., “Comprehensive Treatise on Inorganic and Theoretical Chemistry,” Vol. 5, p. 818, London, Longmans, Green & Co., 1924. (4) Nicholls, P., and Reid, W. T., Trans. Am. SOC.Mech. Engrs., 54, 167-90 (1932). (5) Ibid., 56, 447-65 (1934). (6) Norton, F. H., and Norton, C. L., Jr., J. Am. Ceram. Soc., 13, 161 (1930). (7) Ralston, 0. C., U. S. Bur. Mines, Bull. 296, 16-20 (1929).
R E C E I V ~April D 16, 1936. Presented before the Division of Uas and Fuel1 Chemistry a t the 89th Meeting of the American Chemical Society, New York, N. Y., April 22 t o 2 6 , 1935. Published by permission of the Direotor, U. S. Bureau of Mines. (Not subject to copyright.)
Determining the Specific Gravity of Gases ~
Improvements in the Effusion Method L. C. KEMP, JR., J. F. COLLINS, JR.,A N D W. E. KUHN, The Texas Company, Port Arthur, Texas
T
HE effusion method for determining the specific gravity of gases (9,s) has been widely employed, because of the
simplicity of the apparatus and its manipulation, although
it has been generally recognized that the results obtained can in most cases be regarded only as approximations in view of what were considered to be inherent errors in the method. It is toward the minimization of such errors that the work discussed in this paper was directed. In so far as could be determined, little work had been done in improving the accuracy of the effusion method since its inception in the seventeenth century until about 1917, when Edwards published (1) the results of an extensive investigation of the effusion method of determining gas density. In spite of the soundness of the conclusions and recommendations presented, they apparently received little consideration and certainly had not been put into practice a t the time the work herein under discussion was undertoken.
PREPARATION OF ORIFICE. After some experimentation a method was developed whereby a round and burr-free orifice opening could be effected. A regular steel needle is carefully turned down by means of fine emery so that a very fine, long, and tapering point is obtained. Using this needle, the orifice plate is punched very lightly as nearly in the exact center as possible and the protruding metal is carefully removed by a twist drill (size 80, approximately) run along the side of the plate. The needle drill is again used and, workin from both sides of the plate, the opening is increased to the fesired size. This operation requires much care and patience and frequent examination under the microscope should be resorted t o in order t o insure proper roundness of the opening and freedom from burrs and t o avoid increasing the diameter above the desired amount. An orifice prepared in this way is shown in the upper right-hand photomicrograph, Figure 2.
With an orifice prepared in the manner described, a correlation such as that shown by Figure 3 was obtained by deApparatus and Operation termining the specific gravities of a wide range of refinery gases on both the effusion bottle and the Edwards balance. Following unsuccessful attempts typified bv Figure 1 to The average deviation of these results develop a satisfactory correlation between from the correlating line is negligible and, the results obtained with a standard make except for one sample showing a negaof effusion bottle and with an Edwards tive deviation of 6 per cent, the maximum specific gravity balance (previously calideviations are less than 5 per cent. I n brated by means of pure gases), a careful this instance the effusion bottle results study of the operating technic was made. are approximately 5 per cent lower than This, however, failed to reveal any those obtained with the Edwards balance. source of appreciable error and attention ENLARGEMENT OF GAS CHAMBER.In was accordingly directed to the equiporder to effect further improvement in ment. Bearing in mind that Edwards the accuracy of the method, consideration had stressed the necessity of having a was given to means of obtaining an incarefully prepared orifice, the plate was creased time of flow. It was finally removed from the apparatus and subdecided to accomplish this through injected to microscopic examination. Even c r e a s i n g t h e size of t h e c a l i b r a t e d a casual inspection (upper left, Figure 2) chamber in preference to reducing the is sufficient to establish its inferiority. size of the orifice, since the latter proCloser inspection indicates that the irregucedure is o b j e c t i o n a b l e both from a larities are due to the employment of a FIGURE 1. CORRELATION OF REs t a n d p o i n t of i n c r e a s e d tendency to’ punch in effecting the opening, and inSULTS WITH STANDARD EFFUSION deed such a procedure is generally recomBOTTLE AND EDWARDSSPECIFIC e r r o r t h r o u g h plugging of the orifice and of difficulty in preparing a smaller GRAVITY BALANCE mended in the literature. Y
”
-
SEPTEMBER 15, 1935
ANALYTICAL EDITION
orifice which is satisfactorily round and burr free. Figure 4 shows the revised apparatus in some detail; the chamber volume has been increased from 100 to 200 cc. Figure 5 presents the results obtained w i t h t h e revised equipment correlated with results obtained with an Edwards balance on gases of approximately the same range of specific gravity as previously employed. In this instance the average deviation from the correlating line is less than 0.1 per cent, with maximum positive and negative deviations of 4.5 and 2.9 per cent, respectively. The effusion results are, on the average, 2.75 per cent lower than those obtained with the Edwards balance. STAINLESS S T E E LO R I F I C E . Figure 6 presents the results obtained with an effusion apparatus employing the enlarged chamber and an orifice (lower left, Figure 2) having an opening 0.240 mm. in diameter in a sheet of stainless steel 0.052 mm. thick. These results, in which the average deviation is less than 0.3 per cent and the maximum positive and negative deviations are2.61 and 1.64 per cent,respectively, s i m p l y i n d i c a t e that stainless steel maybe substituted for Dlatinum where service conditiins will permit. It is entirely probable that other metals which can be obtained in satisfactorily thin sheets can be employed as particular conditions encountered may warrant. :EFFECTOF ORIFICEDIAMETER AND PLATE THICKNESS ON ACCURACY. Figures 7 and 8 present the results obtained with two platinum orifices, one (lower right, Figure 2 ) having an opening 0.300 mm. in diameter in a plate 0.092 mm. thick
FIGURE2. PHOTOMICROGRAPHS
and the other having an opening 0.260 mm. in diameter in a plate 0.130 mm. thick. The results of these determinations, as well as the two previously discussed, are summarized in Table I. While it is realized that above data are inadequate to establish definite limits as to the size of the orifice opening or the thickness of the plate, it is believed justifiable to set up an
FIGURE3. CORRELATION OF SPECIFIC GRAVITYDETERMINATIONS Using orifice prepared by author’s method
339
FIGURE4. REVISED APPARATUS
FIGURE 5. CORRELATION OF RESULTS WITH REVISED EQUIPMENT AND EDWARDS BALANCE
INDUSTRIAL AND ENGINEERING CHEMISTRY
340
I 0
l
l
I
0 E L//o
1
I Pt /
I
I
c
l
1
1
//r
FIGURE6. RESULTSWITH ENLARGED AND STAIXLESS STEEL CHAMBER ORIFICE
FIGURE7. RESULTSWITH PLATINUM ORIFICE
TABLE I. RESULTS OBTAINED 11 ITH VARYINGORIFICE DIAMETERS AND PLATE THICKNESSES
Mm.
Mm.
0.052 0.130
0.240 0.260
0.110
0.185
0.092
0.300
--Percentage DeviationF~~~ Maximum from curve Edwards Positive Negative -3.64 -2.15 -2.75 -4.85
2.61 7.02 4.54 5.72
FIGURE 8. RESULTSWITH PLATINUM ORIFICE
Lower right, Figure 2
opening of approximately 0.240 nun. diameter in a plate of less than 0.10 mm. thickness as a tentative standard. Experience has indicated that difficulty is encountered in obtaining round, smooth openings of a diameter less than 0.200 mm. and this should be the lower limit. The results indicate that, a t 0.300 mm. diameter, the operation is unsatisfactory so that the diameter should propably not exceed 0.260 mm. It is believed desirable to employ as thin a sheet as possible with due regard to providing sufficient strength to resist normal wear. Anything above 0.05 mm. should be satisfactory in this respect. Considerably more importance should be attached to the diameter of the opening than t o the thickness of the plate.
Orifice plate orifice Thiokness Diameter
VOL. 7, NO. 5
1.64 6.00 2.92 6.46
"$$$~ No. 6 8 5 7
ELIMINATION OF CONDENSATION ON ORIFICE. In order to eliminate, if possible, the tendency towards accumulation of small drops of water around the orifice, attempts were made to employ acidified solutions of zinc sulfate and ferrous sulfate as confining media in the effusion bottle. Considerable difficulty was encountered, however, with precipitation of zinc sulfide and iron sulfide in the bottles and no improvement in the accuracy was obtained. Predrying of the samples just before entering the apparatus likerise had no apparent effect on the accuracy of the method. DECREASED ACCURACY AT HIGHER SPECIFIC GRAVITIEB.An inspection of the graphical presentations shows that the greatest deviations occurred for the most part in the specific gravity range above 1.30. This may be attributed to the tendency towards condensation of some of the heavier components of the gas around the orifice. It was attempted to eliminate this effect by placing the entire apparatus in a thermostatically controlled box so that determinations might be made a t a higher temperature. This, however, was not found satisfactory, because of increased vaporization of the water used as the confining medium. Since, a t the present time, there appears to be no way of improving the accuracy in the higher gravity range, it is recommended that the results so obtained be accepted only with reservation.
Sources of Error Experience with the effusion method has indicated that the following avoidable factors may contribute to inaccurate results: (1) loss of confining medium between determinations, (2) leakage of the apparatus, (3) accumulation of dirt in the orifice. The air rate should be frequently checked to insure that no radical change has occurred which might indicate the influence of one of the above factors. It has also been found advisable to clean the orifice periodically by washing with alcohol and ether. A clean hair may be used to remove any accumulated material from the opening. If possible the orifice should be examined occasionally under a microscope to insure its being in a satisfactory condition. CALIBRATION OF EACH ORIFICE. Included in Table I is a column showing the per cent deviation from Edwards which varies from 2.15 to 4.85 per cent, the effusion results being too low in every case. It is for this reason recommended that, wherever possible, each orifice be calibrated against an Edwards balance or by means of gases of known specific gravity. I n the event no method of calibration is available it is recommended that the specific gravity as determined directly from the effusion apparatus reading be arbitrarily increased by 2.5 per cent.
Conclusions The results obtained in this work are such as to warrant the acceptance for routine work of gas specific gravities below 1.30 obtained by the effusion method employing the enlarged gas chamber and corrected by means of a correlation of the apparatus being used with a specific gravity balance. The probable accuracy so obtained is within 2.5 per cent. Values above 1.30 may be used, but the probable accuracy is considerably lessened above that figure. The data, although not entirely conclusive, indicate that the orifice plate should have a thickness of less than 0.10 mm. and probably above 0.08 mm. The diameter should be approximately 0.240 mm. and should not exceed 0.260. Below 0.20 mm. diameter difficulty is encountered in obtaining a satisfactory opening. Stainless steel is recommended as an acceptable substitute for platinum in making the orifice plate for service where stainless steel will withstand the corrosive agents present.
Acknowledgment Considerable thanks are due to G. C. Fairhairn of The Texas Company for his assistance in developing a suitable drill
SEPTEMBER 15, 1935
ANALYTICAL EDITION
for the production of satisfactory orifice openings and in other phases of the work.
Literature Cited Edwards, J. D., Bur. of Standards, ~
~Paper 94 ~ (
J~20,~
1917).
341
Lunge, George, ‘‘Technical Gas bnalysis,” p. 181, New York, D. Van Nostrand Co., 1914. (3) Stone, C. H., “Practical Testing of Gas and Gas Meters,” pp. 260-3, New York, John Wiley &Sons,1909.
(2)
R E C ~ I V EMay D 18, 1935. Presented before the Division of Petroleum Chemistry, Joint Symposium on the Chemistry of Gaseous Hydrocarbons, at
~~t h e 89th ~ ~ Meeting l of the, American Chemical Sooiety, New York, N. Y., April 22 to 26,1935.
A Line-Operated Vacuum-Tube Voltmeter R. L. GARMAN
AND
M. E. DROZ, Washington Square College, New York University, New York, N. Y.
S
INCE the introduction by Goode (4) of the vacuum-
tube device for measurements in electrochemistry, many circuits employing this device have been described (3). I n general they have been battery-operated and in many cases of single-tube assemblies their sensitivity has proved inadequate. The recent circuits of Dubridge (1, 2) and others (7, 9) have been designed for high sensitivities. The same principles may be used in the construction of a line-operated device. However, it is impossible to attain the same degree of stability without an accompanying loss of sensitivity because of the high line-voltage fluctuations as compared to battery fluctuations. The balanced single-tube circuits of Dubridge and Turner were designed primarily for special electrometer tubes in which the filament formed part of the balancing network. The direct adaptation of these circuits t o ordinary commercial radio tubes, where alternating current line operation is desired, proves difficult in practice because of the necessity of rectifying the relatively high filam‘ent currents (0.3 t o 2.5 amperes). The bridge circuits of Wynn-’Williams (IO),Nottingham ( B ) , and others (6) avoid this difficulty, but depend upon the selection of matched tubes for successful operation. A stable and yet sensitive circuit which will operate from both alternating and direct current lines can be designed by combining the mu-balance (8) and conventional bridge circuits. A circuit design based on these considerations which was found practical is shown in Figure 1.
-
FIGURE1. CIRCUIT DESIGN c1, cz. 8-mfd. electrolytic
RI, Ra. 1 megohm, 1 watt Rs, R4. 400-ohm control Rs. 200 ohms 15 watta Re. 8000 ohm;, 30 watts R7, RB. 6000-ohm control Re1 Ria. 20,000 ohms, 2 watts Rii. 140 ohms, 100 watts R12. 300 ohms. 16 watts R13. 20 ohms,i6 watts R14. 10,000-ohm tapered control L1. 30 H. choke, 200 ohns Bi, Bz. 3v. flashhght batteries
denser S. P. 5. T. Yaxleyjack s1. D.P. D. T. Y a x l e y button switch SS. D . P . D . T . Y a x l e y switch S A . S. P. S. T. Yaxley jack SI.
conswitch pushjack
L4duo-triode (6A6) tube is used in which the plate impedances switch . ~ ~ _ V I . AAa of the tube form two arms of the bridge and the resistors Re-Rlc Vz. 84 and R7-Rs the other two arms. Balance is det,ermined by a mioroamMA. Weston 0-200 microammeter connected across the bridge R7-Rg. The difficulty meter of matching has been minimized by the use of the 6A6, since its two triode sections were manufactured at the same time and operate in the same environment. an 84-type rectifier and appropriate filter system to permit While a certain degree of stability is attained by the use of the operation on alternating and direct current. matched triodes in the bridge circuit, the stability of the device has been further increased by placing the two sections of the tube in individually balanced mu-balance circuits (2, 8). The Operation ratio of the resistance Re to the effective resistance of the network Rb-R3-R12-R4 is chosen to be slightly higher than the ampliThe proper operating condition may be determined by fication factor (manufacturer’s sDecifications) of the tube, RO that individual adjustments to cdmperforming the following operapensate for slight differences in the tions : two sections can be made by adjusting the resistors RI and R4. Open 8,;throw AS’,to the right and s Batteries Bl and B, are introduced Xs to the left; S4 should be closed. to increase the sensitivity by opposDisconnect the microammeter from ing the high negative grid potentials the circuit by removing the “jumper” which would normally result from across binding posts PI and P2 and the mu-balance network (8). These connect a 0-2 milliammeter between batteries are small, since there is no RQand Rlo. Adjust resistor Rq until current drained from them. the change in plate current with The apparently complex switching change in line voltage becomes a arrangement adapts the instrument to minimum. Changes in line voltage :?NE \‘CLTAOE a variety of purposes. can be most easily secured by placing The power supply is designed with a variable resistance in series with the FIGURE2. STA~ILITY CURVE