John Elihu Hall.

The American journal of science online

. (page 55 of 61)
Online LibraryJohn Elihu HallThe American journal of science → online text (page 55 of 61)
Font size
QR-code for this ebook


tlierefore, as an efficient and independent check upon the trust-
worthiness of the present gas thermometer scale between 0°
and 1100°.

Geophysical Laboratory, Carnegie Institution of Washington,
Washington, D. C, March 20, 1912.

* Sitzungsber. Akad. Wise. Berlin, xliv, 414, 1906.
t Ann. Physik, xxii, 19, 1907. % Loc. oit., p. 8.



Digitized by VjOOQ IC



546 TT. R. Barss — MeasxiremenU of Radio-activity.



Art. XLV. — Note on Measurements of Radio-activity hy
means of Alpha Rays; by W. JR. Babss.

It is a well known fact that in a gas ionized by a-particles
a saturation current is obtained only when a much larger poten-
tial gradient is applied between the plates of the ionization
chamber than is necessary when yS- or X-rays are the ionizing
agents. Bragg and Kleenian* showed that a current through a
gas, ionized by a-particles, was still unsaturated when calculation
showed that the number of ions lost by general recombioation
was small. The effect was ascribed to "Initial Recombina-
tion " ; i.e., to some of the ions being but partially separated
from their parent molecules by the action of the a-particles.
In the absence of an external electric field these ions fall back
on their parent molecules and are thus neutralized. An intense
electric field is supposed to complete the separation of the ions
and to produce saturation. On this hypothesis, lack of satura-
tion would not depend on the size or shape of the ionization
vessel and saturation would be more easily obtained under
diminished pressure.

Kleemanf has shown that lack of saturation with weak ioni-
zation by a-particles is not due to diffusion of the ions, nor
does it depend on the recombination coefficient. He has
shown that " Initial Recombination " is very small in gases
ionized by yS-, 7- and X-rays ; in other words, these ionizing
agents effect a more complete separation of negative ions from
their parent molecules.

Moulin:]: has proposed an explanation of the mechanism of
ionization by a-particles as follows. The ions formed by the
a-particles are not distributed uniformly throughout the gas,
but each a-particle has, associated with it, a column of ions, the
axis of the column being along the path of the particle. Lack
of saturation is explained by recombination of ions of opposite
sign within each column. This recombination between ions of
the same column ought to exceed that which would be obtained
for the same number of ions distributed throughout the volume
of the gas. The amount of the recombination between ions of
the same column should be much greater when the field is
applied in a direction parallel to the direction of the column,
than when it is applied in a direction perpendicular to it ; for
the parallel field would leave the columns intact, while the
perpendicular field would break each column into two parts by

* Phil. Mag., xi, p. 466, 1906.
t Phil. Mag., xii, p. 273, 1906.
% Le Radium, May, 1908, p. 136.



Digitized by VjOOQ IC



W, R. Barss — Measurements of Radio-activity, 547

separating the positive and nes^ative ions. Hence the lack of
saturation should be more apparent in the former case than in
the latter. These facts were experimentally determined by
Moulin. He obtained saturation for the parallel field at 1200
to 1600 volts per centimeter while for the perpendicular field
only about 200 volts per centimeter were necessary.

Moulin concludes that general recombination within the
columns (proportional to the square of the density of ioniza-
tion within the columns) is so much greater than "Initial Re-
combination" that the latter is negligible in comparison.

Ionization by a-particles was further investigated by Whee-
lock ;* among other results, he obtained the following. When
an electric field is applied parallel to the path of the a-parti-
cle and therefore parallel to the axis of the column of ions, the
column would not be broken up and the recombination occur-
ring would be between ions belonging to the same column.
Since each particle makes the same number of ions along its
path, the density of ionization would be the same in any one
column and therefore it would be expected that the ratio of
currents obtained with sources of different intensities would be
constant for different potential gradients applied. When the
field is applied perpendicular to the column and when the
source of ionization is small, very few columns would exist in
the ionization vessel during the time required for the ions to
be carried over to their respective electrodes. Hence there
would be little chance for recombination between the columns,
so that it would be expected that the ratio of currents obtained
with sources of different intensities would be constant as in the
case of the parallel field. When the field is perpendicular
and the source of ionization is stronger, enough columns might
exist between the electrodes at one time to make recombina-
tion possible, not only between ions of the same column but
between those of different columns. In this case the ratio of
currents obtained with different source intensities might not
be constant because of the added recombination of ions of dif-
ferent columns.

Wheelock found that the ratio of the current produced by a
more intense source of rays to that produced by a weaker
source is constant for the parallel field ; that it is approxi-
mately so for the perpendicular field when the sources are
both weak, and that it increases slightly with the potential
gradient applied when the sources are stronger. Tnis is as
would be expected if the ions formed by a-particles are arranged
in columns.

When the gas is ionized by yS- or X-rays it would not be
expected that tne ratio of currents obtained for different source
*Thi8 Journal, zzz, 238, 1910.



Digitized by VjOOQ IC



548 W. JS. Barss — Measurements of Badio-activity.

intensities would be constant. Here the ions are distributed
throughout the volume of the ^as, and general recombination,
which depends upon the ionization density, i.e., upon the
number of ions per cubic centimeter in the gas, would increase
as the ionization itself is increased, unless a saturating field is
applied.

In a great number of important investigations in the subject
of radio-activity, it has been assumed that the quantity of
radio-active material present was proportional to the ionization
currents produced by the a-rays. In these experiments, elec-
trical fields have been applied which would have been ample
to cause saturation if the ionization had been produced by
)8- or X-rays, but which are now known to be quite inadequate
to produce saturation when a-rays are emploved. Results
which have been obtained in this way are oi fundamental
importance in the theory of radio-active transformation. They
include the determiqation of relative quantities of radio-active
substances by the " Emanation Method " and the method of
thin films, as well as nearly all the measurements of rates of
decay of such substances. It is safe to say that in no case in
which such measurements have been made with an electroscope,
in air at atmospheric pressure, has a saturating potential been
applied, or even very closely approached. The fact that a
fairly consistent body of measurements and constants has been
built up by many investigators, notwithstanding this apparent
flaw in their experimental arrangements, shows that the con-
siderations advanced above must have a considerable degree of
validity. The object of the present experiments is to test this
point specifically in the important case when the a-rays are
produced by an emanation mixed with the ionized gas. In
this case the sources of the rays are scattered through the gas
and on the walls of the vessel, and the paths of the a-particles
and their attending columns of ions extend in all directions ;
so that the geometrical complication is as great as it can
well be.

We might reasonably expect the ratio of currents to be con-
stant in this case, at least for small source intensities. If the
number of a-particles is small, there will be only a few colunms
of ions existing together during the time required for the ions
to be carried to their respective electrodes. It is true that a
portion of the a-particles will cross each other and that the
separated columns of ions will also sometimes cross each other,
thus producing some recombination between ions of different
columns. But even when this happens, the crossing will usu-
ally be at an angle, and the length of each column is so great
compared with the diameter of its cross section that even if
they do intersect, the amount of this recombination will be neg-



Digitized by VjOOQ IC



W. Ji. Bar88 — Measurements of Jiadio-activity. 549



ligible compared to the recombination between ions of the
same column.

As the intensity of the source is increased, the number of col-
umns of ions existing together is also increased. The proba-
bility that different columns will cross each other is greater and
therefore the amount of recombination between ions of dif-
ferent columns will be greater. So that, as in the case of the
perpendicular field, the ratio of the larger current to the smaller
will probably increase as the potential gradient is increased.

In the present experiments, a cylindrical tin chamber was
used IS-S"" high and 10-5''™ in diameter. A central brass

Fio. 1.



§2




8 16



40



80



200
Field in volts



electrode, provided with an earthed guard ring, was connected
to a tilted electroscope of the Wilson type, the leaf of which
was observed by means of a microscope having a graduated
scale in the eyepiece. This central electrode and the leaf of
the electroscope were grounded through a potentiometer by
means of which each deflection due to the ionization current
was calibrated in terms of potential. The capacity of the sys-
tem was kept constant, so that these calibrated readings varied
directly as the actual ionization currents. Different potentials
were applied to the case of the chamber. Kadium emanation
was used as an ionizing agent ; it has a half value period of
about four days, so that it provided a suitable source of varying
intensity.

One series of observed data is given in the following table
V is the potential in volts applied to the case. C, represents
the corresponding ionization current for a given intensity ; C,
the ionization current for a weaker intensity, etc.

Am. Jour. Sci.— Fourth Series, Vol. XXXIII, No. 198.— June,' 1912.
86



Digitized by VjOOQ IC



650 W. H. Barss — MeasuremenU of Radio-activUy.



V





c.


c.


C4


C.


2


•72


•65


•40


•23


•16


4


1-60


1-16


•85


•60


•38


8


2-50


1-90


1-40


•80


•55


16


2-90


2-20


1-50


•96


•66



40 3-35 2^40 1^75 1^05 -75

80 3-53 2-58 1-86 MO "78

200 3-65 2-80 1-95 1-16 •SO

400 3-80 2^90 2*05 -83

600 3-90 3-00 2-08

A series of curves plotted from these data is given in fig. 1:
abscissae represent the potential V applied to the case. Curve
1 has for its ordinates tne values given in C, above, curve 2 the
values in C„ etc.

Ratios of ionization currents are given in the following table:

V 'Ci/C. C,/C. C/C, C/Cs



2


1-31 •


1-80


3-13


4-50


4


1-39


1-88


3-20


4-21


8


1-31


1-80


3-12


4-54


16


1-32


1-93


3-02


4^45


40


1-39


1-91


3-19


4-46


80


1-36


1-89


3-21


4-52


200


1-30


1-87


314


4-66


400


r3i


1-85




4-57



600 1-30 1-87

It is evident tliat the current ratios are constant within the
limits of experimental error.

In the above data the potential applied to the case was nega-
tive. A series of readings was made with the potential posi-
tive giving similar results.

The radium emanation used was drawn from carnotite,
the amount of emanation being equivalent to the amount in
equilibrium with about ID"* gm. of radium. It remains to
be tried to what degree the intensity may be increased before
there is a change in the current ratios.

Summary.

When the a-particles are moving in all directions with
respect to the electric field, and when the source of ionization
is not too intense, the ratio of the currents obtained from two
sources of different intensities is constant for different poten-
tials applied to the ionization chamber.

No great errors are involved even when currents are used
less than one-fifth of the saturation value.

In conclusion, I want to thank Professor Bumstead for his
many suggestions throughout the experiment.

Sloane Physical Laboratory, Yale University, New Haven, Conn.



Digitized by VjOOQ IC



N. L. Bowen — The J3ina7y System, 551



Abt. XLVL— rA(? Binary System : Na^Al^8\0jNepMUe,
Camegieite) — CaAlJSifi^ (Anorthite) ; by N. L. Bowen.*

Contents :

Introduotion.
Preparation of Materials.
Preliminary Study.
Heating Curves.

Method.

Results— Table I.
Quenehings.

Method.

Results— Table II.
Equilibrium Diagram.
Discussion of Equilibrium Diagram.
Optical Study.

Properties of Anorthite.
Camegieite.
Nephelite.
Solid Solutions.
Work of Earlier Investigators.

Camegieite.

Nephelite.

Anorthite.
Application to Natural Minerals.

Anorthite.

Nephelite

Nephelite Mix-Crystals.

Other Solid Solutions.

Camegieite.
Thermometry.
General.

Introduction.

The study of the system Na, Al,Si,0,— Ca AI,Si,0, was
UDdertaken because of the very considerable importance of
these compounds as rock-forming constituents.

The compound CaAl,Si,0, occurs in nature as the mineral
anorthite. The preparation of anorthite in the laboratory has
been accomplished by a number of workere.

The compound Na, Al,Si,Og (NaAlSiOJ approximates in com-
position the natural mineral nephelite and has been prepared in
a form resembling that of nephelite. It is mentioned in text-
books of mineralogy as artificial soda-nephelite and is given a
place in order to bring out the chemical relationship within the
hexagonal group of which nephelite is the best known member.

Preparation of Materials,

In the preparation of mixtures for exact thermal work, it is
important that only the purest material should be used. The
calcium carbonate was of tested purity, the alumina was freed

* Presented in partial fulfillment of the requirements for the degree of
Doctor of Philosophy in the Department of Geology at the Massachusetts
Institute of Technology ; May, 1912.



Digitized by VjOOQ IC



552 N, Z. Bowen — The Binary System.

from all but the last trace of alkalis by boiling with a solution
of ammonium chloride and igniting. A very pure silica was
obtained by grinding selected quartz and treating with hydro-
chloric acid. Anhydrous sodium carbonate was made from
the hydrated compound by heating for eight hours at 300°C.
It was proved that the material so obtained corresponded with
the formula Na,C(), by converting to chloride and weighing-

To make anorthite (/aCO„ A1,0„ and SiO, were mixed in
the proper proportions and fused in a Fletcher gas furnace at
about 1600°. Three fusions, the product being finely ground
before each, gave a glass which crystallized to a perfectly
homogeneous mass of anorthite.

In the preparation of the soda compound Na,CO„ A1,0„ and
SiO, were mixed and the same procedure followed, but not
with equal success. When the resultant glass was crystallized,
instead of the homogeneous product to be expected, the micro-
scope showed, scattered throughout, plates and needles of A1,0,
(corundum) uncombined. Kepeated fusion did not remedy
the difficulty. Finally, it was determined by analysis of the
glass that some Na^O had been volatilized at the high tempera-
ture of the gas furnace. The materials had been mixed in the
proportion to give : —

Na,0 Al.O, SiO, Sum

21-77 35-89 42-34 = lOO'OO

Analysis showed 20-50 36*62 42*85 = 99-97

By heating the intimately mixed oxides at a low temperature
(about 800°C) in an electric furnace, grinding and reheating
four times, it was found possible to cause the oxides to com-
bine without loss of soda. The whole could then be melted to
a glass which gave on slow cooling a perfectly homogeneous
crystalline mass with no excess of alumma.*

Even after combination some alumina develops in the per-
fectly homogeneous mass if held near its melting point for
several days, from which it appears that some soda may still be
lost after long heating, l^o very definite statements can be
made concerning this phenomenon inasmuch as every effort
was concentrated on its avoidance.

Preliminary Study.

A stock of the pure end members being ready, intermediate
mixtures corresponding to each 10 per cent interval were
made up.

* The loss of soda gave, of course, an excess in both silica and alumina.
The silica undoubtedly went to form a small quantity of a more highly sili-
cated compound, probably albite, which, however, did not appear as such
and, therefore, must have disappeared in solid solution. Apparently no

compound in which the ratio, x^A-^, is greater than unity could form under

the existing conditions and the alumina was left free.



Digitized by VjOOQ IC



iV^. Z. Bowen — The Binary System, 558

A preliminary study was carried out by putting some of each
mixture in a separate platinum crystallizing disn and holding
all at an approximately constant temperature in an electric fur-
nace for a couple of days. The products were then examined
under the microscope. This procedure was repeated at mod-
erate temperature intervals over some range. The detail of
this work need not be given here because it is not in itself of
great importance, but is a very useful preliminary to the more
precise determinations. It was noted that all' compositions
crystallize readily ; that, in the mixtures, no reactions pro-
ducing new components took place, leaving the system truly of
two components. The existence of a eutectic point in the
neighborhood of 1300° C. was indicated. Two different forms
of the soda compound were observed, the one appearing at low
and the other at higher temperatures. The low temperature
form is analogous to nephelite and will henceforth, for con-
venience, be referred to simply as nephelite. The high tem-
perature form had formerly been prepared at this laboratory,
and given the name carnegieite. At low temperatures mixtures
containing up to 30 per cent anorthite are perfectly homoge-
neous, showing the ability of the low temperature form to hold
over 30 per cent of the anorthite molecule in solid solution.
The lime compound was observed in only one form, anorthite.

Heating Curves,

With this preliminary information it was possible to proceed
with the exact determination more expeditiously.

Small quantities of each mixture were crystallized at about
1200°C. (below the eutectic point indicated above), and heating
curves run on each. The method employed was that found at
the Geophysical laboratory to be the best in mineral work.'*
The charge is of about 2 gms. ; the bare thermocouple of plati-
num-platinum-rhodium dips into the charge and is connected
with a potentiometer system which measures the E.M.F. set
up at the thermocouple contact. A curve has been prepared
giving the E.M.F. corresponding to temperatures between 0°
and 1550° C. for standard elements calibrated against the gas
thermometer. Such standard elements are used to calibrate
the elements employed in the course of the work.

The furnace in which the charge is heated is an electrical
resistance furnace in which a coil of platinum wire is the con-
ducting material.

In running a heating curve the temperature of the furnace
is caused to rise gradually and regularly. The temperature of
the thermocouple in the charge is read at regular time inter-

* For references see list at end of article.



Digitized by Google



654 N. L. Bowen — ITie Binary System.

vals, and temperature is plotted against time. If there is no
energy change within the charge, its temperature also rises reg-
ularly, but at the appearance of a new phase the rate of heating:
of the charge (the gradient of the heating curve) experiences a
change depending upon the energy involved in the change of
phase.

During the change of phase, on account of heat absorption,
the temperature of tlie charge either remains constant or rises
more slowly than the temperature of the furnace, but upon
completion of the change of phase the temperature rises rap-
idly to that of the furnace. It is. this sudden increase in the
rate of heating of the charge which is commonly termed a
'break' on the heating curve. The temperature at which a
break occurs is, then, the temperature at which the disappear-
ance of a phase of the system is completed. In Table I the
temperatures at which breaks were obtained on heating curves
are given opposite the corresponding composition.





Table I.






Wt. j«


Upper point


Eutectic


Inversion 1


Cnlckpni]


CaAUSi.Oe













1527





1304


....





1626


- - - -


1305


....


10


1473


. - - -


1335


1352


10


. - - - -


....


....


1354


20


,-__- -__-


- - . -


....


1360


20


.-_-. - -


_ - -


....


1352


30


._-_- - -.








1349


40


,_ - - . - -


1310''




....


50


,.. - ....


1305


- - . -


....


50


.-_-_ - -


1304


- . _ -


....


60


.-.-_ - -


1302


- . - -


- _ - -


70


1437


1307


_ _ . _


....


80


1484


1306


- - . .


....


90


1522







. . - .


100


1650


- - . -




- - . .


100


1549


- - - -


....


....



These results are plotted on the temperature-composition
digram (Diagram I).

1^ the diagram each circle shows a temperature at which a
break occurred on the heating curve for the corresponding
composition. The general form of the equilibrium diagram is
suggested. There is a eutectic point at about 1302°. Another
triple point, possibly a ' knickpunkt,' occurs at about 1352**.
The general form of the liquidus curves is indicated. Small
heat effects obtained in the pure sodium compound and in the



Digitized by VjOOQ IC



N. Z. Bowen — The Binary System,



555



Diagram I.



1 KJL^




















p

/


L


















f


/




\

\
\
















/
/
/






\
\
\














)


/

(




I480


(


\
\












/
/
/










\
\
\












/
/






I440
1 %AC




\
\










/

/


1








1


i
%








/

/

/












\

\






/


/

r














\






/

/
/

/










Id40


i,


1 — -(




>




/

/
/










1^20
1300


<






\


/


/










»






J


> — ^


► ^


r-^


\ — <


k ,^




i^do























10 ao 30 ^O ,TO CO 70



OO ©O lOO



mixture with 10 per cent CaAljSi,Og are shown by small
circles and are undoubtedly due to inversion.

Quenchinga,

In Table I there are many blanks. Doubtless, some of these
could have been filled in by especially careful work with heat-
ing curves, but a diflFerent metnod was adopted for obtaining
the same information.

The use of the quenching furnace' combined with the micro-
scopic examination of the charges offers a trustworthy method



Digitized by VjOOQ IC



556 N. L. Bowen — The Binary System,

of obtaining information as to temperatures at which phases
appearand disappear, and at the same time a knowledge of the
nature of the phases themselves.

A small charge of any mixture, wrapped in platinum foil, is
held at a definite temperature for a period of time which is
deemed sufficient to insure equilibrium for that temperature
and composition, and is then quenched by allowing it to fall
into a dish of mercury at room temperature.

The chilling is so abrupt that any phase present at the fur-
nace temperature is ' fixed ' and ready to be studied under the
microscope. By running a series of quenchings progress is
made towards a knowledge of the phases present at equilibrium
for each composition, at all temperatures — in short, towards
the data necessary for an equilibrium diagram.

By the quenching method many changes of phase may be
detected and studied which either take place too slowly or
involve too little energy change to give an appreciable break
on a heating curve. With many transformations, moreover,
especially in the more viscous substances, superheating is likely
to occur when the heating is rapid, as in running a heating
curve, so that the actual break observed may come at a temper-



Online LibraryJohn Elihu HallThe American journal of science → online text (page 55 of 61)