Abraham Clark Freeman John Proffatt.

The American journal of science online

. (page 52 of 59)
Online LibraryAbraham Clark Freeman John ProffattThe American journal of science → online text (page 52 of 59)
Font size
QR-code for this ebook

eight-tenths of an ohm.

The method of measurement employed was as follows:
The strip of alloy to be tested was connected in series with a
comparison standard, the resistance of the two being approx-
imately equal. These, together with a third resistance of about
one hundred and fifty ohms, formed the outer closed circuit of
a single gravity cell. " Potential wires " from the terminals of
the strip of alloy and of the standard, were joined to a switch
of such construction that a mirror galvanometer could be con-
nected in shunt with either. The galvanometer had a resis-
tance of two thousand ohms. A comparison of deflections
when the galvanometer was shunted around the resistance
standard, and around the test piece, afforded data for the cal-
culation of the resistance of the latter.

This method is exceedingly sensitive, and when properly
conducted, it is capable of a high degree of accuracy. In the
experiments under consideration, a check upon errors arising
from fluctuations in the amount of current traversing the test
piece and from changes in the constant of the galvanometer,
was obtained by ever repeated reference to the mdications of
the latter when connected with the terminals of the standard.

In order to relieve the observer of the necessity of main-
taining the reference standard at a constant temperature, or of
applying temperature-corrections to the results obtained, a
compensated carbon standard, of the type recently described
by the writer, was constructed.* The resistance oi this stand-

* On Compensated Resistance Standards ; Transactions of the American In-
stitute of Electrical Engineers, vol. v, No. 10, 1888.

Digitized by


Allays of Ferro-Manganese and Copper,


ard was 0*770 ohms ; and its temperature-coefficient was about
•000010 per degree Centigrade.

The test piece, enclosed in a U-shaped tube of glass, was
placed in an oil bath and alternately heated to 100° C. and
cooled to 20°. The reference standard was kept at the room
temperature. Its changes of resistance were regarded as

The application of the method just described, to a number
of ferro-manganese-copper alloys, brought to light a remarka-
ble and very troublesome property of tnis class of metals. It
was found that they decreased in resistance each time that they
were subjected to a change of temperature, even through the
small range made use of in the attempt to determine their
temperature-coefficients. The character of these changes can
best be illustrated by quoting a series of measurements to
which one of the alloys was subjected. An alloy containing
80*82 per cent of copper and 19*12 per cent of ferro-man-
ganese, had been hard drawn in the process of obtaining a
strip suitable for measurement. Its specific resistance at 20°,
referred to pure copper as unity, was 30-38. It was repeatedly
heated to 100° and cooled to 20° with the following result :

Table I.

Efftci of repeated heating and cooling upon the resistance of Alloy No. 6,

{hard drawn).
Obseryation. Temperature. Specific resistance.

1 20* 30-380

2 100 30-186

3 20 30 163 -99287

4 100 30151 -99255

5 20 30-138 -99202

6 100 30*121 -99180

7 20 30-118 -99134

8 100 30118 -99134

9 20 30-106 -99093

10 100 30.099 -99072

11 20 30092 -99051

12 100 30-104 -99092

13 20 30-079 -99007

14 100 30-104 -99092
16 20 30072 -98986

We have, in the case of this alloy, a substance which in-
creajses in conductivity each time it is heated and cooled
through the small range of 80°, the change in resistance dimin-
ishing in amount with each operation, but still perceptible at
the end of the seventh cycle. At the same time a positive
temperature-coefficient is being developed, which continues to
increase as the heating and cooling process is repeated.

After being heated to 100° seven times, with the result
shown in Table I, the alloy was raised to a red heat. Its tem-

Belative resistance.


Digitized by


474 E. L. Nichols — Electrical JSeHstance of the

peratnre-coefficient was then redetermined with the result
shown in the following table :

Fable II.

Resistance and temperature

coefficient of Alloy No. 8, (after anneaUng at a red hea(y


Specific resistance.


(At 20")

(At 100')





+ 000062







+ 000052

The effect of repeated annealing upon the resistance and
coeflBcient of these alloys, is still more strikingly exhibited in
the behavior of a specimen containing a larger proportion of
ferro-manganese. The alloy in question consisted of 70*65

Earts of. copper and 29*35 parts of ferro-maganese. After
eing brought into a condition of stability, such that further
heating and cooling through a range of eighty degrees had
but little permanent effect upon its conductivity, it still
showed, when hard drawn, an appreciable negative coefficient
It was then annealed three times at a red heat ; specific resis-
tance and coefficient being determined for the range 20® to
100°, after each annealing. The results are given in Table III.

Table II r.
Effect of repeated annealing upon Alloy Xo, 11.



Specific resistance.—



Rather hard . _ . -

... 4610



— •000024

Once aanealed

.__ 4510

+ •000021

Twice annealed

-_ 44-07



+ •000068

Thrice annealed

.- 42-76



+ •000192

This metal possessed very nearly indeed the composition for
which, in the patent specifications already referred to, the
remarkable property of decreasing resistance with rise of tem-
perature was claimed ; which claim is substantiated, so far as
the hard- drawn alloy is concerned. It will be seen that the
coefficient depends upon the temper of the metal, and that it
would probably be an easy matter to bring the latter into such
a state that the change of resistance, which is, in all conditions
of the alloy, very much smaller than in any other of which
we have definite data, would be too small to be detected.

It appears, moreover, that the conductivity of this alloy
was increased about 2 per cent by each successive annealing
at a red heat, and that even after being thus annealed three
times, it was subject to a further slight but measurable increase
of conductivity, amounting to at least "02 per cent, when sub-
sequently heated to 100® and cooled to 20°.

Digitized by


AUoys of Ferro-Manganese and Copper.


The influence of temper upon the temperatnre^oefficient of
a number of similar alloys waa investigated by the method
just described. The alloys showed, when hard rolled, a co-
efficient very near to zero, sometimes positive, sometimes
negative. After annealing at 300° to 400° a well defined
negative coeflicient, after annealing at a red heat, a still larger
positive coeflicient, was developed. It was found that the
positive coefficient produced by annealing, could be reduced

r*n by rolling^ the alloy. Table IV shows the character of
results obtained. They were verified in every essential
detail by frequent repetitions.

Table IT.

InjkufMA of aitemaU anneaUng and hardening upon the iemperatwe-coeffidmi.

Afloy No. 7. (Copper, 80*40^, Ferro-manganese, \9'60%).
Condition of the alloy. Coefflolent (90°— lOO") .

Hard +000022

Partiallj annealed —000032

Thoroughly annealed +-000066

Re-rolled (hard) +-000021

Again annealed +000046

To determine the relation of the composition of these
alloys to their temperature-coeflBcients, Mr. Blood tested twelve
specimens, in which the amount of copper ranged from 70 per
cent to 99*5 per cent, and also the copper itself from which
the alloys were made. The percentage of copper present in
each test-piece was determined to withm one hundredth of one
per cent, by the method of electro-deposition. The results of
these determinations are incorporated in Table V.

Tahlb V.





Percentage of





AII07 semi-


AII07 bard.


Alloy annealed.



1 100




1 107




' 119




1 11-28





1 11-74




, 1407






— 066621




i - __.





1 30-38







- -000032













- -000024


Am. Joun,

Soi.— Third Series, Vo


0. 234.— Juke,



Digitized by


476 E. L. Nichols — Electrical Resistance of the

It was found that specimens containing less than eighty {>er
cent of copper could not be worked witnout frequent partial
annealing. Two such specimens were tested, after rolling and
again after subsequent annealing. They showed a negative
coefficient after rolling, which may have been due to the pre-
vious heating, undergone in the process of working them into
the necessary form ; but since they were as nearlv m the con-
dition of the hard-rolled alloys as it was possible to make
them, they have been classified in the table among the hard

The coefficients of the hard-rolled alloys, including those
above mentioned, have been used in the construction of the
accompanying curve, (figure 1). Abscissae are percentages of
f erro-manganese present m the respective specimens ; orcunates
are the. changes of resistance for lOO*' 0. It will be seen that
the coefficient of the unalloyed copper falls considerablv
below Matthiesen's standard, and that the addition of smaU
quantities of ferro-manganese produces a further very rapid







s Ark 8 ^ lo^iJ- ii/^ re-iris" iraa-id-


decrease. With ten per cent of ferro-manganese, the change
of resistance is less than one per cent for one hundred degrees.
Alloys containing from fifteen to twenty per cent of ferro-
manganese possess exceedingly small coefficients, the curve
crossing the base line at the point corresponding to eighteen
per cent. The curve is intended to represent the variations of
alloys which are of the same temper, but it is not possible to
determine in how far it does so. Indeed for the entire rai^
from fifteen per cent to thirty per cent the coefficient may be
given any value between + -00002 and —'00002, by varying

Digitized by


Alloys of Ferro-Manganese and Copper. 477

the temper of the metal. A positive coefficient could be ob-
tained, however, in alloys containing more than twenty per
cent of ferro-manganese, only by thorooghly annealing the
specimen ; whereas in metals containing a larger proportion of
copper, either hardening or complete annetding developed a
positive coefficient.

The marked influence of temper npon the conductivity of
these alloys, renders it difficult to determine the precise law of
the change in specific resistance with the coinposition. It
appears, however, from the results presented in Table V, that
the resistance increases nearly in direct proportion to the per-
centage of ferro-manganese.

Mr. Blood's investigation also included two alloys which
contained nickel as weU as ferro-manganese. The methods of
preparation and measurement were identical with those which
have already been described. The results, which are given
in Table VI, show exceedingly small negative coefficients in
the case of the hardened alloy. Annealing rendered the
coefficient of the alloy containing the smaller percentage of
nickel positive, and reduced the size of the coemcient of the
other specimen.

Table VL
-Compoeltlon of tbe lUoy. . , Temperature coefflcients.-

Copper. Ferro-manganeBo. Nickel. AII07 (bard). AII07 (annealed).

78-28jr U-Otjr V^H -•000011 +000007

52-51jf 31-27jf 16-22jr - OOOOSO -000032

The experiments described in this paper, show that the alloys
of ferro-manganese and copper, so far as their electrical be-
havior is concerned, must be considered as a distinct class.
Up to the time of Mr. Weston's discovery of their properties,
increase of conductivity with rise of temperature was sup-
posed to be confined to electrolytes, and to the single solid con-
auctor, carbon. Recent investigations have added sulphur to
the list, and it is evident that this set of alloys, at least, be-
longs there also. It is not improbable that the further study
of alloys containing metals oi the iron group, will lead to the
discovery of other combinations, possessing tne same interest-
ing and important characteristic.

Digitized by


478 C, Bams — Fluid Volume and its

Art. LVI. — Fluid Volume and its Sdation to Pressvre
and Temperatare;^ by 0. Barus.


1. The present paper contains the introductory part of a
series of experiments on the compressibility of liqaids, now in
progress at this laboratory. The incentive to tne work was

fiven by Mr. Clarence King. In its development I am bene-
ted by his counsel.

The geological purposes in view make it necessary to obtain
a preliminaiT survey of the whole field of inquiry. Experi-
mental details will be filled in later.

2. The literature of the subject has recently been critically
digested by Professor Tait.' Excellent excerpts are to be found
in the Fortschritte der Physik The work of Canton (1762-64),
Perkin (1820-26), (Eersted (1822), CoUadon and Sturm (1827),
R^ault (1847), Grassi (1851), Amaury and Descamps (1869),
is discussed in most text-book& Since that date the contri-
butions have been manifold, and are fast increasing. I shall
therefore principally confine myself to such papers in which
volume changes produced by the simultaneous influence of
both pressure and temperature are considered.

Setting aside the literature* of critical points, which is too
voluminous for admission here, the work of Cailletet* is first
to be noted, as introducing a long range of pressures (700 atm.).
Amagat's' early work contains a large temperature interval
(0** to 100°), but applies for pressures below 9 atm., only. The
results are discussed with reference to Dupr^'s* eauation. Pass-
ing Buchanan's' and Van der Waals' results on the compressi-
bility of water and solutions, I come to the important step in
the subject made by L6vy*, though he had been considerably
anticipated by Dupr6''. L6vy endeavors to prove that the
internal pressure of a body kept at constant volume is propor-
tional to its temperature. I have already stated" that Levy's
position is antagonized by H. F. Weber,*' Boltzmann," Clausius'*

1 Communicated with the permission of the Director of the Geological Survey.
' Tait : Properties of matter.

> See Landolt aud Boemstein's Phvsikalisch-chemische Tabellen, Berlin, J.
Springer. 1883, p. 62.
* Cailletet: C. R., Ixxv, p. 77, 1872.

5 Amagat: C. R., Ixxxv, pp. 27, 139, 1877 ; Ann. ch. et phys., xi, p. 620, 1877.
« Dupre: C. R., lix. p. 490, 1864 ; ibid., Ixvii, p. 39?, 1868.
' Buchanan: Nature, xvu, p. 439, 1878.
« Van der Waals : Beiblatter, i, p. 5 1 1, 1 877.
» L6vy : C. R.. Ixxxvii, 1878, pp. 449, 488, 676, 664, 649, 826
w Dupr6: Theorie mecan. de la chaleur, Paris, Gauthier-Villars, 1869, p. 51.
" Cf. this Journal, xxxviii, p. 407, 1889.
»« H. P. Weber: C. R.. Ixxxvii, p. 517, 1878.
" Boltzmann: C. R., Ixxxvii, pp. 693, 773, 1878.
" Clausius: C. R., Ixxxvii, pp. 718, 1878.

Digitized by


Relation to Pressure and Temperature. 479

and Masden/ but that important experimental evidence is
given in favor of the probable truth of the principle by Earn-
say and Young. See below. Amagat's' work on the compres-
sion of gases must be mentioned because of its important
bearing on pressure measurement. In two critical researches
Mees* perfects Regnault's piezometer and re-determines the
compressibility of water. Tait* and his pupils, Buchanan',
Marshal, Smith and Omond,' and others carry the inquiry into
compressibility and maximum density of water into much de-
tail, and the data are theoretically discussed by Tait. Pres-
sures as high as 600 atm. are applied. Solutions and alcohol
are also tested for compressibihty. Pagliani,' Palazzo and
Vicentini,* using Regnault's piezometer, publish results for
water and a number of other substances, mostly organic.
They also examined mixtures. Temperature is varied between
O*' and 100°. The results are discussed at length and show
the insufficiency of Dupr6's formula. DeHeen' who has
spent much time in studying the volume changes of liquid,
deduces a formula of his own chiefly in reference to the ther-
mal changes of compressibility. Theory is tested by experi-
ments. The researcines which Amagat,*" published at about
this time are remarkable for the enormous interval of hydro-
static pressure (3000 atm.) applied. Ether and water are oper-
ated on. In later work these researches are extended to include
other liquid substances with the ulterior object of locating the
lower critical temperature. The behavior of water is fully
considered. Grimaldi" critically discusses the earlier work on
the maximum density of water. He also examines the volume
changes produced in a number of organic substances" bv tem-

Ejrature (0"^ to 100*^) and pressure (0 to 25 atm.), and finds
upr^'s, CeHeen's and Konowalow's** formulae insufficient.

> Massiea: C. R., Ixxxvii, p. 731, 1878.

^ Amagat: C. R., Ixxxix, p. 43T, 1879; ibid, xc, pp. 863, 995, 1880; ibid., xci,
p. 428, 1880, BDd elsewhere.

» Mees: Beiblatter, iv, p. 612, 1880; viii, p. 435, 1884.

^ Tait: Nature, xxiil, p. 595, 1881 ; Challeoger Reports, ii, 1882, app., pp. 1 to
40; Proc. Roy. Soc. Ed., xi, p. 813. 1882-83 ; ibid., xii, 1883-84, pp. 45, 223, 226,
757 ; ibid., xiii, p. 2, 1884-86.

* Buchanan: Proa Roy. Soc. Ed., x, 1880, p. 697.

« Marshal, Smith and Omond: Proc. Roy. Soc. Ed., xi, pp. 626, 809, 1882-83.
^ Pagliani and Palazzo: Beibl., viii, p. 795, 1884; ix. pp. 149, 626, 809, 1886.

* Pagliani and VioeotiDi: Beibl , viii, p. 794, 1884; Joum. de phys., (2) xxx, p.
461, 1883.

* DeHeen: Bull. Soc. Roy. Belg., (3), ix, 1885, p. 550; Journal de phys., (2),
viii, p. 197, 1889.

^^ Amagat: C.R., ciii, p. 429, 1886.

" Amagat: C. R., civ, p. 1159, 1837; ibid., cv, pp. 165, 1120, 1887.
" Giimaldi: Beiblatter, x, p. 338, 1836.

^Grimaldi: Beiblatter, xi, pp. 136, 137, 138, 1887; Ztschr. fQr phys. chem.,
ii, p. 374, 1888.

' Konowalow: Ztschr. fur phys. chem., ii, p. 1, 1888.

Digitized by


480 G. Baru9 — Fluid Volume a/nd its

At this stage of progress, the point of view gained in the
extensive researches of Eamsay and Young* throws new light
on the subject. They prove experimentally that if pressure^
and temperature fl vary linearly (p^b0^a\ the substance
operated on does not change in volume. Ether, methyl and
ethyl alcohol, acetic and carbonic dioxide are tested. Excep-
tional values for the case of acetic acid and nitrogen tetroxiae
are referred to dissociation. Utilizing James Thomson's' dia-
gram, they point out that the locus of the isothermal minima
and maxima intersect the locus of maximum vapor tensions
at the critical point Data are given for ether. Fitzgerald'
reasoning from Ramsay and Youngs results arrived at the
theoretic results virtually given by X6vy (see above). Tait*
who is still actively at work on high pressures has recently
made further publication on the effect of dissolved substances
on internal pressure. An endeavor to formulate Andrew's
classical results is due to Dickson.* Fitzgerald* recently applied
Clausius's' equation to a discussion of Ramsay and Young's data,
and Sarrau* has similarly discussed Amagat's data.

3. A few references to thermal expansion of liquids, which
enters incidentally into the present paper, must be added.
Many formulae have recently been devised and tested by
A venarius,' DeHeen," Mendeleeff," Thorpe and Rucker," Jouk"
and others," not to mention older observers. None of these
forms are satisfactory when long ranges of temperature are in-
troduced, as was shown by Bartoli and Stracciati," testing
Mendeleeff's, and Thorpe and Riicker's formulae, and by the
discussion between Mendeleeff and Avenarius.

Special mention must be made of the celebrated papers of J.
Willard Gibbs," by whom the scope of graphic methods for
exhibiting the thermo- dynamics of fluids was surprisingly

' Ramsay and Young: Phil. Mag., (6) xxiii, p. 435, 1887 ; ibid., xxiv, p. 196,
1887 ; Proc. Roy. Soc. Lond., xlii, p. 3, 1887.

« James Thomson : Pbil. Mag., (5), xliii, p. 227, 1872 ; Nature, ix, p. 392, 1873.

» Fitzgerald: Proc. Roy. Soc, xlii, p. 50, 1887.

*Tait: Challenger Reports, Phys. and Chem., ii, (4), 1888; Proc. Roy. Soc
Ed., XV, p. 426, 1888.

• Dickson: Phil. Mag., x, p. 40, 1880.

• Fitzgerald: Proc Roy. Soc Lond., xlii, p. 216, 1887.
■^ Clausius: Wied. Ann., p. 337, 18 .

• Sarrau: C. R., xciv, pp. 639, 718, 845, 1882; ibid., ci, p. 941, 1885.
» Avenarius: Beibl., ii. p. 211, 1878; ibid., viit p. 806, 1884.

>o DeHeen: Bull. Ac. Roy. Belg., (3), iv, p. 526, 1882 ; Journ. Chem. Soc, xlv,
p. 408, 1884.

" MendeleeflE: Chem. Ber., xvii, p. 139, 1884; Beibl, viii, p. 806, 1884.

" Thorpe and Rucker: Joum. Chem. Soc. xlv, p. 135, 1884.

" Jouk: Beibl.. viii, p. 808, 1884.

*^ Rosenberg: Journ. d' Almeida, vii, p. 350, 1878.

i» Bartoli and Stracciati : Beibl., ix, p. 510, 1885.

'• J. Willard Gibbs: Trans. Conn. Acad., ii, (2), pp. 809, 382, 1873.

Digitized by


JSdcUion to Pressure and Temperature, 481

enlarged. The terms isometric^ isopiestiCj isothermalj isentrop-
icSj etc., are need in the present paper in the sense defined by
Gibbs (1. c, p. 311).

4. Surveying the above researches as a whole, it appears at
the outset that more work has been spent on the compressi-
bility of water than the exceptional behavior of this substance
justifies at the present stage of the inquiry. In other words
the volume-pressnre-temperature changes of the great majority
of liquids probably conform closely with a general thermo-
dynamic law, due respectively to l)upr6 and L6vy, and to
Kamsay, Young and Fitzgerald. It is therefore first desirable
to find the full importance of this law experimentally, and then
to interpret the exceptional cases with reference to it. Again
it appears that researches in which long pressure ranges are
applied simiiltaneouslv with long temperature ranges, are
urgently called for. It is from such work that further eluci-
dation of this important subject is to be obtained.

Among pressure experiments the late researches of Amagat
stand out by their originality and importance. One can not
but admire the experimental prowess wnich has enabled him to
penetrate legitimately into a region of pressures incomparably
nigh, without lowering his standard of accuracy.


5. Force vuin/p and appurtenances, — In making the experi-
ments detailed in the following pages I used a large Cailietet
force pump, of the form constructed both by Ducretet of Paris,
and by the Societe G6nevoise. Its eflBciency is 1000 atm. It is
made to be fed with water, but I found by using thin sperm
oil, it was possible to facilitate the operations, because there is
less danger of rusting the fine polished steel parts of the ma-
chine. The pump consists of two parts : the pump proper and
the strong cylindrical trough into which the compression tubes
are to be screwed. The trough is cannon-shaped, its axis ver-
tical, and the open end uppermost. Pump and trough are
connected bv strong phosphor-bronze tubing, and similar tubes
lead to the large Bourdon gauge by which the pressures are

Digitized by


6. Com/pression tubes and aj
to be testei

Fig. 1. — Capillary tube
with appurtenances for
measuring compressibil-
ity of fluid& Scale \.

f. — The substance
is enclosed in capillary tubes
of glass of very fine bore (-03^ and less)
and about -6*^ thick, such as are used for
thermometers. The length of these tubes
ahc^ figure 1, was about 50"° or 60°". To
insert them into the trough, the solid glass
of the tube was swelled or bulged at h

Online LibraryAbraham Clark Freeman John ProffattThe American journal of science → online text (page 52 of 59)