John Elihu Hall.

The American journal of science online

. (page 56 of 61)
Online LibraryJohn Elihu HallThe American journal of science → online text (page 56 of 61)
Font size
QR-code for this ebook

ature higher than that at which the transformation concerned
would tafce place if equilibrium prevailed. In quenching work
it is the endeavor to make the time sufficiently long that all
transformations which can take place at the temperature of the
furnace may be complete.

In this work the thermoelement junction in the quenching
furnace was distant from the charge 1/2 cm. or more. The
small temperature difference between the charge and the fur-
nace element was determined at frequent intervals over the
temperature range desired by hanging a standard element in
the position of the charge and reading both. The differences
found were applied as corrections to the readings of the fur-
nace element m the course of the work.

The furnace temperature was kept constant by noting it fre-
quently and making the necessary changes in the resistance of
tne circuit. For the more careful work the current from a
storage battery was used.

Several scores of quenchings were made. Of these, only the
more significant will be tabulated (Table II), more especially
those which determine the limit of stability for the various

Digitized by VjOOQ IC

B. L, Bowen — The Binary System.


Table II.
(a) Inversion,



Phases found


at start



after qnenching








6 hrs.

Ne only














6 hrs.

Cg only





Cg only





Cg only




1 hr.

Ne & Cg




1 hr.





4 hrs.

Ne only





1 hr.

Ne & Cg





7 hrs.






4 hrs.

Ne only

Ne & Cg (trace)





1 hr.















* Ne = Nephelite, Cg = Carnegieite. '

(b) Nephelite solidus.
See also (solid solution An in Ne).












6 hrs.
2 hrs.
1 hr.
1 hr.
1 hr.

1 hr.

2 hrs.
2 hrs.

Ne only
Ne only
Ne only
Ne & glass
Ne & glass
Ne only
Ne only
Ne & glass

(c) Ne liquidus.
See also entectio composition.
An & Ne 1320° 1 hr. glass & Ne (bare trace)


(d) Cgliqxiidus

1 hr.
1 hr.
1 hr.


fe) An liquidus.

An & Ne 1330° 1 hr.

An & Ne 1380 1 hr.

An & Ne 1392 1 hr.

glass & Cg (bare trace)

glass only

glass & Cg (rare)

glass only
glass only
glass only

Digitized by VjOOQ IC

558 N. L. Bowen — The Binary System.

(f) Change ofpfuue at * Kniekpunkt,^

33 3 Ne ISol"" 6 hrs. Cg only

34 5 Ne 1357 6 hrs. Cg A glass traoe

35 10 Ne 1357 6 hrs. Cg A glass

36 20 Ne 1357 6 hrs. Cgife glass

37 30 Ne 1357 7 hrs. glass only

38 3 Ne 1346 6 hrs. Cg only

39 5 Ne 1346 6 hrs. Cg & Ne

40 10 Ne 1346 6 hrs. Cg & Ne

41 20 Ne 1346 6 hrs. Ne & glass

42 30 Ne 1346 6 hrs. glass & Ne

(g) Eutectic composition,

43 50 glass 1304"^ 2 hrs. An <fe glass

44 40 glass 1304 2 hrs. Ne A glass

45 47*5 glass 1304 2 hrs. An & glass

46 46 glass 1304 1 hr. An <fe glass

47 45 glass 1304 1 hr. Ne & glass

(b) Solid solution An in Ne,

48 30 glass 1295''ca 2 days Ne only

49 32 glass 1295''ca 2 days Ne only

50 35 glass 1295°ca 2 days Ne & An (mere trace)

(k) Solid solution Ne in An»

51 98-5 glass 1250° 2 days An only

52 98 glass 1250 2 days An & Ne (trace)

53 97 glass 1290 2 hrs. An £ Ne

(1 ) Solid solution An in Cg.

(See also change of phase at * Knickpankt.')

54 3 glass 1350° 6 hrs. Cg only

55 5 glass 1350 6 hrs. Cg & Ne (trace)

Many of these quenching results are worthy of special dis-
cussion, but this may be undertaken more profitably after they
have been combined with the heating curve results and plotted
on a temperature-composition diagram (Diagram II: Equili-
brium Diagram).

Explanations of Signs and Terms.

In Diagram II the circles indicate heating curve breaks ; the
numbered crosses indicate temperatures and compositions at
which quenchings were made, the numbei's corresponding with
those in Table II where the phases present are given. The
data make possible the drawing in of the curves separating the
fields of stability of the v^lous phases. On each field the
phases present at equilibrium are marked. Nephelite is used

Digitized by VjOOQ IC

N, L. Bowen — The Binary Syatem. 559

to indicate the hexagonal form of NajAI^Si^Og and also the
isomorphous hexagonal mix-crystals with 0—35 per cent CaAl,
SijOg. Carnegieite is used to indicate the triclinic form of
Na^Al^Si^Oa and also the isomorphous triclinic mix -crystals
with 0—5 per cent Ca^Al^SijOg. Anorthite indicates triclinic

Diagram U.


CaAlgSijOg and also the isomorphous triclinic mix-crystals with
0-2 per cent Na,Al,Si,0,.

Throughout the text abbreviations are used : Ne to indi-
cate neplielite (NaAlSiO, hex.) ; Cg to indicate carnegieite
NaAlSiO, triclinic) ; An to indicate anorthite (CaAl^SijOg tri-
clinic). Nephelite mix-crystals are indicated, Ne„An„ or Ne^^,-
An„^ the subscript numbers indicating the weight percentafije
of NaAlSiO, and of CaAl^SijO. present in the mix-crystal, in
like manner, carnegieite mix-crystals are referred to, -Cg„An^
and anorthite mix-crystal -An„Ne,. NaAlSiO, and Na^Al,
SijO, are used intercnangeably. The purpose in using the

Digitized by VjOOQ IC

660 jV, Z. Bowen — The Binary System,

latter form ie that of bringing out clearly the analogy between
the two componentd. No knowledge of the molecular weight
ie implied. Diagram II is the only diagram here contained on
which certain points are lettered, and when these points are
referred to by letter in the text, it has not always been con-
sidered necessary to specify Diagram II.

Discussion of Table TI and Equilibrium Diagram,

Inversion Phenomena {Table II (a)). — In the heating curre
work a break was obtained with nephelite (Ne) at 1305°.
The heat effect was small and undoubtedly due to inversion.
In the quenchings it was, however, found that nephelite will
go over to camegieite at a temperature considerably below
1305°. Held at 1245° for 6 hours there was no formation of
camegieite, but at 1252° for only one hour the charge was
largely converted to camegieite. These results indicate an
inversion temperature between 1245° and 1252°. By starting
with camegieite (Cg) confirmatory results were obtained.
Crystallized at a high temperature and then held at successively
lower and lower temperatures, it was found that at 1252° no
change took place in 6 hours, but that at 1245° partial con-
version to nephelite was observed within one hour. ' From
these results the inversion temperature is shown to lie between
1245° and 1252° or at 1248° (approx.). Close to the transition
temperature the inversion is only partial in either direction,
after several hours. At a temperature (about 1300°) more
removed from the transition point nephelite is converted
entirely to camegieite within an hour. The reverse change
does not proceed to completion so readily, but was effected by
heating camegieite over a Bunsen burner for 60 hours.

Glass of composition NaAlSiO^ crystallizes as camegieite
at any temperature above 1248°. At temperatures only a
few degrees below 1248° where nephelite is the stable
phase glass crystallizes as both nephelite and camegieite,
whereas nephelite itself shows no tendency to change to
camegieite at these temperatures. It seems possible that the
unstable glass in assuming the stable form nephelite passes
through the "less unstable" form camegieite which may
persist for a considerable time and is therefore found in the
quenched products. Glass held at temperatures considerably
below 1248° crystallizes only as nephelite. If the *'les8
unstable" form, camegieite, is formed at these temperatures it
inverts quickly to nephelite.

With the nephelite mix-crystal (Ne,,An,„) a small heat effect
suggestive of inversion was obtained at 1335°, somewhat higher
than that in Ke. The results of quenchings indicate again

Digitized by VjOOQ IC

N. L. Bowen — The Bina/ty System. 561

that this temperature is considerably higher than that at which
inversion will take place if the time is sufficiently long.
Ne,^An,o may be held at 1282° (34° above the inversion point
of JNe) for four hours without change. Held at 1294° for one
hour a trace of the carnegieite form and at 1296° for one hour
considerable carnegieite separates.

These results indicate that carnegieite first separates from
Ne,,An,, between 1282-1294° or at about 1290°. This rapid
rise in the temperatures of inversion of nephelite as CaAl^Si^O,
is takea into solid solution was confirmed in Ne^An^ (1266°)
and in Ne^o An^^ which shows only the nephelite form at 1340°.

It is not possible to confirm these inversion temperatures by
starting with the carnegieite form and studying the reverse
change as was done with Ne and Cg because the behavior of
these solid solutions is much less simple than that of the pure

It is well known that according to theory a solid solution
should, in general, melt, not at a definite temperature to form a
liquid of its own composition, but that melting should be dis-
tributed over a temperature interval, the licjuid first formed
differing in composition (concentration) from the solid.
Similarly the inversion of a solid solution takes place over an
interval, the concentration in the first small quantity of the
new crystal form being either greater or less tnan that in the
old. It is mainly in these characteristics that a solid solution
differs from a definite compound.

Both melting interval and inversion interval are exhibited
by the nephelite solid solutions. Any point on the curve E C
represents the temperature of the lower limit of the inversion
interval for the corresponding composition and was located, as
has been described by holding nephelite of that composition at
higher and higher temperatures until evidence of inversion to
carnegieite was obtained. Any point on the curve E D rep-
resents the temperature of the upper limit of the inversion in-
terval for the corresponding composition. Only the end points
E and D are accurately located and the middle portion qualita-
tively indicated (Table II (f) and (1)).

The concentration of CaAl,Si,Oa in the carnegieite solid
solution (E D) is much less than that in the nephelite solid solu-
tion (E C) in equilibrium with it at any temperature. Beckinan*
has developed an equation for the amount of change in the in-
version (or melting) point of a substance by the solution in it

of another substance. It reads A T= — 7- (Cj,— C\) where AT=

change in inversion temperature, T= absolute temperature of
the inversion point in the pure substance, Z=it8 heat of trans-
formation, Cj= concentration of the solution above the inver-

Digitized by VjOOQ IC

662 N. L. Bowen — The Binary System,

sion point and C,= concentration of the solution below. In
the present case, as has been stated, the concentration beloi^
the inversion point (nephelite) is much greater than that above
(carnegieite), therefore C^—G^i^ positive and relatively lar^e
and so also A T. This corresponds well with the sharp rise in
the inversion temperature of nephelite occasioned by the solid
solution of CaAIaSijOg. Qualitatively, the facts till the require-
ments of thermo-dynamic theory.

The Nephelite Solidios and Liquidus {Table 11 (J) and
(c) ). — The melting-interval of the nephelite solid solutions has
already been mentioned. The solidus C F and the liquidus
B 6 have the same significance in this connection as have E C
and E D with respect to the inversion interval.

In mineral work it is not possible to locate solidus and
liquidus by the method of separating solid and liquid phases
from an equilibrium mixture at various temperatures and
determining the composition of each phase. The solidus was
located by finding the temperature at which nephelite mix-
crystals of diflEerent compositions first showed signs of melting
and the liquidus by finding the temperature at which melting
was barely completed. After location of the curves in this
manner the composition of liquid and solid in equilibrium with
each other at various temperatures can be predicted with the
degree of accuracv attained in the location.

The ''KnicJcpunM' {Table II (/) ).— Breaks were obtained
on some of the heating curves at 1852° (approx.). Quench-
ings show that this is the temperature at which the two
solid phases, carnegieite and nephelite, are in equilibrium
with liquid. The carnegieite has the composition Cg„
Anj(D), nephelite Ne„An,«(C) and the liquid has 28*5 per
cent CaAl^SijO^B), the compositions being known from the
positions of these points determined by quenchings and not
from analysis (Table II (a), (b), (c) and (1).

The Eviectic {Table II {g) ). — Breaks were obtained on
many of the heating curves at 1302° (approx.). Quenchings
proved that at this temperature nephelite (Ne^AnjJ and
anorthite (An^gNeJ are in equilibrium with a liquid with 45'5
per cent CaAlgSijOg (eutectic). The exact composition of the
eutectic mixture was obtained by holding the furnace slightly-
above the eutectic point and hanging in it four or five charges
of the same mixture, one above another, on a platinum wire.
The charge at one end of the wire is in the hottest part of the
furnace a few degrees above the eutectic temperature and for
compositions close to the eutectic is completely molten. The
one at the other end is in a cooler part of the furnace a few
degrees below tlie eutectic and completely crystalline. One or
more of the intermediate charges is at a temperature only very

Digitized by VjOOQ IC

N. L. Bowen—The Binary System. 563

slightly above the euteetic point aud will show on quenching
the excess phase. By exanuning all, the excess phase for the
composition nsed will be determined, and by repeating with
other mixtures the euteetic may be placed within narrow
limits. In this manner it was shown that in the mixture with

45 per cent Ca Al^SijO, nephelite was the excess phase, and with

46 per cent, anorthite. The euteetic composition is, therefore,
between the two, approximately at 45 per cent CaA],Si,0^.

The Limits of Solid Solution {Table II (A) {k) <& (I).— A
solution may be defined simply as a homogeneous mixture, and
with this in mind it is readily seen that the limit of solid solu-
tion might be determined by finding the first composition
(concentration) which shows any trace of inhomogeneity. This
method is, as a rule, the best in mineral work. The most
important solid solutions encountered in this system are the
nephelite mix-crystals. At a temperature slightly below the
euteetic point the first mixture which showed anorthite in
excess was that with 35 per cent CaAl,Si,Og ; in this the
anorthite was a mere trace. The extreme solid solution is,
therefore, approximately Ne.aAn„.

The maximum concentration of CaAl,Si,0^ in carnegieite is
obtained at the temperature of the 'knickpunkt' 1352^, where
the composition of the carnegieite mix-crystals is Cg„An.

The Liquidus of Anorthite {Table II {e) ). — The upper por-
tion of the anorthite liquidus was established by heating-curve
breaks. With lesser concentration of anorthite the magnitude
of the euteetic break obscured the upper point, and in these
mixtures the liquidus was located by quenching from various
temperatures and finding that temperature at which the last
trace of anorthite melted.

In a similar manner part of the carnegieite liquidus was
located (Table II (d)).

Optical Study,

In the microscopic examination of the product of each
experiment the method employed for identifying constituents
was that of determining refractive indices in the crushed

f rains by immersion in liquids of known refractive index,
his method is especially useful for detecting small quantities
of any component. If tlie preparation be placed in a liquid
which matches the mean index of the excess constituent, even
very minute quantities of another constituent show up markedly,
except in the rare case of different substances with nearly
identical refractive indices. Moreover, it is possible to deter-
mine in the crushed grains most of the properties by which
minerals are identified in thin section, such as order of bire-
fringence, optical character, etc.

Digitized by VjOOQ IC

564 iV^. Z. Bowen — The Binary System.


The optical properties of anortliite (An) had previously
been determined in detail at the Geophysical laboratory by H.
E. Merwin, and the figures given are the results of his work.

a = 1-576, y = 1-589, y— a= '013. 2V— 80°±3°
/3 = 1-583, G = 2-766.


Carnegieite (Cg) always shows polysynthetic twinning simi-
lar to that in albite, with high extinction, up to 36°, against
the twinning lines. Other Jess persistent twinning lamellae
appear crossing these at various angles, often giving a raicro-
cline-like structure. On account of the universal intricate
twinning, many of the optical properties are difficult of deter-
mination. The indices, obtained by matching with a liquid in
sodium light and determining the index of the liquid on the
Abb6 refractometer, gave

y =1-514, a = 1-509, y-o = -005

The mean index is therefore low, comparable with that of leu-
cite. The birefringence, measured with the compensating
quartz wedge,* gave -0052 and -0048 in different plates. Opti-
cal character negative, 2 V small estimated at 12°-15°, proba-
bly triclinic.

All efforts to obtain euhedral crystals failed. In all charges
showing carnegieite and glass (area A D B, Diagram II) the
carnegieite is always in rounded crystalline globules. An
attempt to grow crystals in a flux gave nothing but similar

The density of crystalline carnegieite (Cg) is 2-513 at 21**,
determined in Thoulet's solution on carefully selected material,
free from air bubbles.


In all charges consisting of nephelite and glass (area C B G
F, Diagram II), the nephelite appears under the microscope in
perfect hexagonal basal sections, dark with crossed nicols, or
rectangular prismatic sections with parallel extinction, indicat-
ing the hexagonal nature of these mix-crystals.

Of the pure sodium compound, nephelite, euhedral crystals
were prepared in sodium tungstate over a Bunsen burner.
These were in all cases bounded by the prism and terminated
only by the base. One of these was measured on tl^e two-
circle reflecting goniometer and the hexagonal nature con-
firmed. The maximum deviation from 60° of the measured
prism angle was 16', and the average deviation 8', with fair to

Digitized by VjOOQ IC

N, L, Bowen—The Binary System. 565

good signals. The indices, determined by the immersion
method in sodium light, gave a> = 1*537, € = 1-533. Bire-
fringence -004, slightly, but noticeably less than that in

About half a gram of these crystals was tested for tungstic
acid by boiling with hydrochloric acid. Only pure white
silica was left with no yellow tinge. The amount of tungstic
acid could not have been more than a trace.

The density of these crystals is 2*619 at 21°. The degree of
correspondence of the artificial compound with the natural
mineral is shown in the following table :

Cryst'n. Char. « e G habit

NaAlSiO, (artif.) hex neg 1*637 1*533 2*619 prism and

Nephelite (nat.) hex neg 1*541 1-537 256-265 prism and


It may be fortuitous that there is close correspondence with
Gladstone's law in the relation between the mean indices and
densities of the two forms of the soda compound, but the fact
is, perhaps, worthy of note. Gladstone's law states that the

ifi \

specific refractive energy —3— is constant, whatever the state of

molecular aggregation.
We have then

1-512— 1
for carnegieite ~~o~^^ ~ -2037

for nephelite L^-^Zll = -2042

^ 2-619

Nephelite Mix- Crystals,

It was found possible to determine definitely the variation,
with composition, of the optical properties of the hexagonal
mix-crystals or solid solutions from Ne to Ne„An„. With
increasing proportion of the lime molecule, the birefringence
of -004 (negative) grows less, passes through zero and finally
becomes *002 (positive).

The indices were determined in sodium light at room tem-

?erature by the immersion method. The results are shown in
'able III and in the plot Diagram III, in which the size of
the circles is a rough measure of the accuracy of the determi-

Am. Jour. Sci. -Fourth Series, Vol. XXXin, No. 198.— June, 1912.

Digitized by VjOOQ IC

666 N, L. Bowen — The Binary System.

Table III.

Indices of Refraction of Nephelite Mix-crystals.

Optical charactei*















very weak







1-537 +

very weak







It will be noted that the refractive index of the ordinary i-a v
remains sensibly constant, while that of the extraordinary ray
mounts from 1-533 to 1-539. At the composition Ne*,An„

Diagram III.

10 20

(approx.) we have the unique case of hexagonal crystals iso-
tropic with respect to sodium light.

Two different kinds of material were available for these
determinations : homogeneous crystalline aggregates prepared
by crystallizing the various compositions at about 1200° (area
E C F H, Diagram II) and euhedral crystals Ne,,An„-Ne,,
AUj^ embedded in glass obtained from quenchings within the
area C B G F, Diagram II. With the former material the
optical character of nephelite (Ne) was determined as negative
and that of nephelite (Ne^An,,) as positive by the ordinary
interference figure method. Witli the latter material the
optical character could be determined where the birefringence
was extremely small because the relation of indices to crystal-
lographic directions was known. Important aid was thus given

Digitized by VjOOQ IC

N, Z. Bowen — The Binary System. 567

in determining the refractive indices; for example, both
indices of crystals varying in composition from iNe^^^n^ to
Ne„An„ appear to match the 1*537 liquid, but Ne^,An„ is
weakly birefracting and determinable in the prismatic section
of a crystal as definitely negative and therefore o>>e. In a
similar manner it was shown that i^e„An„ is positive and
therefore e>a). The positive character of crystals NeeftAn,,
indicated by the interference figure was confirmed in these
crystals of known orientation.

A curious fact which gave rise to some difficulty in some of
the quenchings may be noted here. The nephelite mix-crystals
with 23 per cent CaA],Si,Og (approx.) are in equilibrium at
1335° with a melt containing 36 per cent CaAl^SijO^ (approx).
(Area C B F G, Diagram II.) When quenched the crystals are
sensibly isotropic and have an index 1'537 and the glass, it so
happens, has the same index.* Neither by a difference of
index nor by the presence of birefringence do the crystals
become distinguishable from the glass. Thus, in several
charges quenclied from 1335^ the presence of nephelite in the
glass could not be detected, although the necessity of its pres-
ence could be proven from charges quenched at a slightly
higher or lower temperature, in which charges the extreme
similarity of nephelite with the glass was approached, but not
quite attained. Vov this reason the attempt to accurately
locate the intersection of the liquidus B Q on the ordinate of
35 per cent CaAl^SiaOg failed. The curve may, however, be
regarded as sufliciently established by the other known points.

The Work of Earlier Inveatir/ators,

Several different workers have succeeded in preparing the
nephelite form of NaAlSiO,, but the carnegieite form has
seldom been encountered. This is rather a curious fact, for
carnegieite crystallizes readily from the pure melt, whereas
nephelite is obtained only with very slow coolingf or with the

Online LibraryJohn Elihu HallThe American journal of science → online text (page 56 of 61)