Download Determination of the Freezing Point of Mercury using Platinum Resistance Thermometers and more Lecture notes Chemistry in PDF only on Docsity!
FREEZING POINT OF MERCURY
By R. M. Wilhelm, Assistant Physicist.
The temperature at which mercury (^) freezes is of importance in thermometry. It marks (^) the lower (^) limit in (^) the use of mercurial thermometers, and its location (^) at about -39° (^) makes it an im-
portant fixed^ point^ of^ the^ temperature^ scale below 0°^ C.
Among (^) the investigators who (^) made precise determinations of
this constant^ may^ be^ mentioned^ Regnault,*^ who ia 1862 obtained
-38?50, B.^ Stewart,^ 1863, whose^ value -38?85 was the result of a very excellent^ experiment^ in^ which^ a constant volume gas thermom- eter was^ used,^ Vicentini^ and^ Omodei^ in 1888, Chappuis* in (^) 1896, and C. Chree^^ in^ 1898, who obtained the values -38?8o ±o?o2,
— 38?85, and — 38?86, respectively, ustag mercurial thermome-
ters which^ had^ been^ previously^ compared with a gas thermometer. The latest determination^ of this point was made by Henning ®
in 1 914. He^ obtained^ the value —^ 38?89, using platinum^ resist-
ance thermometers^ which^ had^ been compared with a gas ther- mometer. The above (^) values are fairly consistent and would
indicate the freezing point of mercury to be in the neighborhood
of —^ 38?8. However,^ two^ determinations^ of this^ point,^ —^ 39?
and (^) —39^38, attributed to Hutchins^ and Cavendish,^ who inves-
tigated the subject in the period i776toi783, have been published
along with the later values and apparently have, in many cases,
(^1) Mem. d. I'Acad., 26, p. 525; 1862. ^ Phil. (^) Mag., 45, p. 224; 1898. 2 Phil. Trans., (^) 153, p. 425; 1863. 6 Annalen der Physik (^) (4). 43, p. 291; 1914. 3 Atti della R. Ace. di Torino. (^) 23; 1887. 7 phii. Trans. ; (^) 1776. (^4) Compt. rend, de (^) la Conference (^) Generale, p. 291; 1896. 8 phil. Trans.; (^) 1783. 75741°— (^17 10) 65s
656 Scientific Papers^ of the^ Bureau^ of Standards [Voi.
been given equal weight with them, although these early deter-
minations are ob^dously entitled to little weight.
All the evidence at present available indicates that the scale defined by the platinum resistance thermometer, when calibrated at 0°, 100°, and (^) 444^6 (the boiling point of sulphur^), defines temperatures in agreement with those given by the hydrogen thermometer down to —^ 40°^ C. A redetermination of the freezing point of mercurv^ on the scale defined by the resistance ther- mometer seemed, therefore, of value.
SAMPLES OF MERCURY The freezing points of three different samples of mercury were observed. These three samples were furnished by ]\Ir. Mc- Kelw, of this Biu-eau, and were designated as "U. S. P.," "An- ode," and "Hulett Still." The modes of purification were as follows: U. S. P. : Purified to meet the test requirements of the
U. S. Pharmacopoeia. Anode: Electrolyzed in a mercurous ni-
trate solution with the mercury as anode. About 5 per cent was deposited on (^) the cathode (^) to eliminate metals more (^) electroposi-
tive than^ mercur3^^ The^ product^ remaining^ on^ the anode was
then once distilled in vacuum. Hulett Still: An anode sample
was distilled in a Hulett still under reduced pressure with a stream of air bubbling through the mercury. The last named
of the three samples would thus be considered to be the purest.
However, the three samples showed no^ difference in electromotive properties when^ used in a normal^ cadmium^ cell, nor do the re- sults of the freezing-point determinations indicate the existence
of any difference between the three samples.
THERMOMETERS AND MEASURING INSTRUMENTS The three-resistance thermometers used were selected from a number made by H. K. Griffin, of this (^) Bureau, about a year ago.
These thermometers were made with potential terminals^ and^ the coils were wound (^) to give a flexible mounting for the wire in (^) the
manner described by Waidner and Burgess. ^° The (^) highest purity Heraeus wire was used for the coils, and short (2 cm) lengths of the same wire were used for^ the^ connections between (^) the ends of the coil (^) and the gold lead wires. The glass
tubes in^ which^ the thermometer coils were^ inclosed^ were^ closed^ at
the top with plugs, sealed in with Khotinski cement, through
- (^) Henning used 44495 1. but this introduces a difference of only o. "004 between the two scales at the freezing point of mercur-. (^10) Waidner and Btirgess, this Bulletin, (^) 6, p. 155; 191c.
658 Scientific^ Papers^ of the^ Bureau^ of Standards^ [Vol^13
The bridges were calibrated in terms of international ohms, so that they could be used interchangeably. The potentiometer was of the Diesselhorst type made by Otto Wolff. The potentiometer was used to secure a check method but it did not give as high (^) pre-
cision, at least with the low resist-
ance thermometer, as the WHtieat-
stone bridge method.
METHOD The mercury was frozen in (^) a
glass tube 2 cm inside diameter
and about 34 cm long. This tube
was (^) inserted in another (^) glass tube
3 cm^ inside^ diameter.^ The^ ther-
mometer and (^) tubes were held in position by corks. Fig. i is a
sketch of this apparatus showing
one of the thermometers in posi-
tion. About 38 cc of^ mercuT}^ was used which filled the inner
tube to a depth of about 1 2 cm,
which was increased to from (^13)
to 1 6 cm when the^ thermometer
was introduced.
In freezing the mercury' the
tubes containing^ it^ were^ im-
mersed to a depth of about 20
cm in^ a^ stirred^ bath^ which^ could
be cooled to —^ 50°^ by expanding
carbon dioxide^ through^ a^ coil^ im-
mersed in a liquid consisting^ of a
mixture of^ gasoline^ and^ carbon tetrachloride. Under these^ con-
ditions the^ temperature^ of^ the
freezing mercury remained con-
stant sometimes^ for^ as^ long^ as^30
-Freezing point tubes with thermome- (^) j^m^tes. In SOme caseS where ter in position (^) , , r • -* (^) -i. • a complete^ freezing^ and^ meltmg point (^) curve was desired the bath was held at a temperature lower than (^) the freezing temperature until the entire mass of the mercury had frozen and (^) its temperature had begun to fall. If the bath were then allowed to heat (^) up to a temperature higher than the freezing
temperature, the melting point could be obser'ed.
Fig. I.
WUhelm] Freezing Point of Mercury 659
PRECAUTIONS
Care was taken to insure that the thermometer coil was suffi-
ciently immersed in the mercury. Conduction down the leads
might cause a higher^ or lower reading, depending on whether the
surrounding bath, and^ thus the air above the mercury, was at a
higher or lower temperature^ than the mercury. Errors from this
source would be^ more^ likely^ to occur in the use of thermometer
C 22, the^ coil^ of^ which^ was^ longer^ than^ those of the other two.
The effect^ of^ raising^ and^ lowering^ the^ thermometer^ while the
mercury was^ freezing^ was^ observed^ for^ C 22.^ It^ was^ found that
' "
r ^
^n.a'r>
S ^ 5 SI ll^ -^
^
^
t-«-^ rw.:ZJ//^?^ ^^ 0—
c
— at
«39 C " i \ , \ • \
tu son t'oo _^ -J^ \ 50 300 11 _^ _j
Time Fig. 2,
Freezing and point curve
an immersion of 13 cm (measured from end of tubing) was suffi-
cient to eliminate this source of error. The absence of any dis- turbing effects due to conduction is also attested by the fact (^) that
the results obtained were independent of the outer bath tempera-
ture, which varied from below —^ 50°^ up to —^ 30°.
FREEZING AND^ MELTING^ POINT CURVES
Fig. 2 shows a typical freezing and melting point cirrve. A
supercooling of several tenths of a degree precedes the constant
temperattire assumed later when freezing is taking place. No
significant difference (^) was noted between the temperatures of the freezing and melting (^) points.
waheim] Freezing Point of Mercury 66
1
These results show that the precision attainable in the deter-
mination of the freezing^ point of mercury is better than o°.oi. It
seems probable that the^ experimental^ conditions were varied suffi-
ciently in the present^ work^ to preclude the possibility of a sys- tematic error of^ as^ much^ as^ o°.oi^ in the result, expressed on the scale defined^ by^ the^ platinum^ resistance^ thermometer. The differ- ence between this result (^) ( —^ 38°. (^) 87) and (^) that found by Henning
( —^ 38°.89), also^ in^ terms^ of^ the^ scale^ defined^ by the platinum
resistance thermometer, is greater than would be expected from
the accidental errors of either determination. It seems improba-
ble that the discrepancy is due to impurities in the mercury,
although it is known that the method of purification used by Henning (distillation in vacuum) ^^^ is not effective in removing traces of certain metals. SUMMARY The freezing point of mercury was determined, using platiniun
resistance thermometers.
A short historical sketch gives the names^ of previous^ investi- gators, their methods of temperature measurement,^ and^ values obtained. Nineteen determinations were made on three samples^ of^ mer- ciuy purified by different methods. Three resistance thermom-
eters having resistances in melting ice of approximately 2.5, 10,
and (^25) ohms, respectively, were used. Resistance measurements were (^) made both by the Wheatstone bridge method and the poten-
tiometer method.
The (^) result of all the measurements gives —^ 38^.873 for the freezing (^) temperature. The maximum deviation of any determi- nation from (^) the mean is o°.oo5.
Washington, (^) August i, 1916. (^13) Geo. A. Hulett and Howard D. Minchin, Phys. Review, (^) 21, p. 3SS; (^) 1905.
