T281

MALLEABILITY AND METALLOGRAPHY

OF NICKEL
By P. D. Merica and R. G. Waltenberg

ABSTRACT

In the manufacture of " malleable " products of nickel and certain nickel
alloys, such as forgings, rolled shapes, and castings, rather unusual metallur-
gical treatments of the molten metal are resorted to in order that the cast
product may be sufficiently malleable and ductile for the subsequent forging
operations. The principal feature of these treatments is the addition to the
molten metal before casting of manganese and magnesium. Ordinary furnace
nickel without these additions will give a casting, in general, which is not
malleable either hot or cold.
It is demonstrated that such " untreated " nickel, or nickel alloys, are brit-
tle because of the presence of small amounts of sulphur—as little as 0.005
per cent. This sulphur is combined with the nickel as Ni 3 S 2 which collects
,

in the form of thin and brittle films around the grains of nickel. This sulphide
also melts (as an eutectic with nickel) at a very low temperature, about
G30° C. Hence, at all temperatures the cohesion between the nickel grains is
either completely destroyed or at least impaired by the presence of these films
of sulphide, and the metal is brittle.
Both manganese and magnesium react with molten NuS- to form sulphides
of manganese or of magnesium. The latter, in particular, has a very high melt-
ing point and. in addition, assumes a form of distribution in the metal mass
which affects the malleability but slightly, viz, that of small equiaxed particles
uniformly distributed throughout the metal grains. The effect of magnesium
addition, therefore, in conferring malleability upon nickel containing sulphur,
isdue principally to the resulting conversion of NI3S2 into MgS.
This investigation was completed under an industrial fellowship at the Bu-
reau of Standards maintained by the International Nickel Co.

CONTENTS
Page
I. Introduction 156
II. Method of investigation 15S
1. Preparation of melts 158
2. Tests for malleability 159
III. Results 160
1. Malleability of pure electrolytic nickel 160
2. Nickel oxide and nickel 161
3. Gases in nickel 163
4. Sulphur in nickel 163
5. Other elements 167
6. Effect of manganes-e and magnesium 171
7. Effect of excess of magnesium 178
IV. Summary and conclusions 179
155
156 Technologic Payers of the Bureau of Standards [Voi.j9

I. INTRODUCTION
For many years after the discovery and isolation of nickel in
1750 by Cronstedt this metal was available only in nonmailable
forms —rondelles, cubes, grains, etc. Indeed, this was true even for
years after the beginning of its commercial use, such as for plating,
alloying, etc. The valuable properties of the pure metal were recog-
nized, but for some time no one was apparently able to cast it in
such a manner as to render it amenable to the usual rolling or draw-
ing processes.
In 1878 the firm of Joseph Wharton, of Philadelphia (which later
became the American Nickel Works, a constituent company of the
International Nickel Co.), exhibited at the Paris Exhibition speci-
mens of rolled pure nickel strip and of drawn nickel wire. Shortly
after this a number of processes were discovered and patented for
1
the production of malleable nickel. Thus, G. A. Boecldicker de-
scribes the use of from iy2 to 3 per cent of manganese as an addition
agent to molten nickel, which renders it malleable. G. Selve and
F. Lotter 2 also describe the use of manganese, and the Bernsdorfer
Metallwarenf abrik 3 claim the use of potassium permanganate solu-
tion to be mixed with nickel oxide and allowed to dry, followed by
reduction and melting with charcoal in the usual manner. J. Ger-
nier 4 mentions the use of manganese for the purpose of rendering
cast nickel malleable, but prefers the use of phosphorus (up to 0.003
per cent) added as a phosphor-nickel alloy.
It is evident that the remedies so proposed were based upon the
thought that the brittleness of the usual cast nickel was due to the
presence of nickel oxide, and, indeed, this thought was usually
and claims of these experimenters.
explicitly stated in the discussion
The most method of producing malleable nickel had
practical
already been discovered and described in 1879 by Th. Fleitmann. 5
Fleitmann discovered and patented the addition to the molten nickel
before casting of metallic magnesium in amounts of from 0.05 to
0.125 per cent, and later amended this practice to use a magnesium-
nickel alloy instead. This operation worked excellently and became
shortly thereafter the recognized method for treating nickel in
order to render it ductile.
Fleitmann did not speak of this operation as designed to remove
nickel oxide, but believed that it removed either carbon monoxide
(by reduction to carbon) or cyanogen compounds. He called at-

1 The Wiggin process for production of malleable nickel; Wagner's Jahresberichte 27,
p. 69; 1881.
- Production of nickel free from oxide; German patent 25798; 1883.
3 German patent 28989 ; 1884.
*On a new process of obtaining malleable nickel; C. R. 01, p. 331; 1880.
s Ber. d. deutsch Gesell., 12, p. 454 ; 1879.
materiberal
Malleability and Metallography of Nickel 157

magnesium secures the production

tention to the fact that this use of
of castings which are both malleable and sound; that is, free from
gas inclusions. He refers also later 6 to the use of manganese to
flux and remove sulphur, a most interesting suggestion, anticipating
in a general way the conclusions reached in this investigation.
After the lapse of about 45 years we find the Fleitmann method,
a bit modified it is true, still used exclusively for the treatment of
nickel intended for malleable castings, as well as for the treatment
of many nickel-bearing alloys, such as Monel metal and other cop-
per-nickel alloys. Even with modern methods of refining and im-
proved furnace practice, nickel-alloy castings are not, in general,
malleable without the magnesium treatment. It is customary to-day
to use for this purpose about 0.1 per cent of metallic magnesium,
usually in conjunction with 0.25 per cent or more of manganese in
one of its usual commercial forms.
The Fleitmann method has, in fact, been so brilliantly successful
that since the day of its discovery no one has apparently found it
necessary to give any consideration to the mechanism by which the
effect is produced. It has worked and does work, why bother about
how it works. The producers of nickel alloys have thankfully
availed themselves of its beneficial effect and, for convenience, re-

ferred to one of cleoxidation and degasification, appar-

its effect as

ently the accepted explanation.

In recent years, however, a number of both major and minor
phenomena have served to focus our attention again upon this ques-
tion, and it has become increasingly evident that there was something-
else in the action of magnesium on nickel and its alloys than deoxi-
dation or degasification.
The investigation here reported was taken up in the summer of
1921 with the specific purpose of ascertaining (1) why ordinary cast
nickel is not malleable, when not treated with magnesium, and (2)
what is the mechanism by which the magnesium treatment produces
malleability in such nickel. It is obvious that such an investigation
has of necessity embraced as well a considerable study of the metal-
lography of nickel and some of its alloys and has brought to light
some other interesting facts only indirectly related to the principal
themes. Of these, perhaps, the most interesting are the results of
the study made of the equilibrium between nickel and nickel oxide.
The metallography of nickel and its alloys has been the subject
of a number of most valuable investigations, of which it is unneces-
sary to mention all in this place. A discussion and bibliography of
these investigations is given in Bureau of Standards Circular No.
100 on Xickel. and we should like to refer also to the very practical

"German patent 28460; 1884.

158 Technologic Papers of the Bureau of Standards [roi.io

remarks on this subject by D. H. Browne and J. F. Thompson. 7

which cover particularly the metallography of nickel with small
amounts of the usual commercial impurities.

II. METHOD OF INVESTIGATION

Inasmuch as the first problem of this investigation was to deter-
mine why and under what conditions cast nickel is not malleable,
recourse was had at once to the preparation, on a small laboratory
scale, of test melts and cast ingots of nickel prepared under a variety
of conditions and with various metallic and nonmetallic additions.
The ingots of these melts were tested both hot and cold for malle-
ability, and many were examined also by thermal and metallographic
methods.
i. PREPARATION OF MELTS
Three furnaces were used in the preparation of these melts in
order to embrace a variety of melting conditions a gas-fired cruci- —
ble furnace, a high-frequency induction furnace, and the Arsem
furnace.
The melts in the induction furnace were made in specially pre-
pared magnesia crucibles or in crucibles of " impervite,'' a commer-
cial refractory porcelain. These crucibles were cylindrical and were
of two sizes,. 22 mm(0.86 inch) in diameter and 75 (3 inches) mm
high, or 35 mm
(1.4 inches) in diameter and 75 mm (3 inches)
high. No ingots could be poured from these crucibles. It was
necessary to melt a charge, make the desired additions, and allow
the melt to solidify in the crucible. Some melts were made in this
furnace by using larger crucibles, additions were made and ingots
poured after each addition. Additions of magnesium were not satis-
factory, as the magnesium would cause violent spitting, caused
possibly by the relatively small cross section of the crucible. The
melts in this furnace were made by direct induction, the current pro-
ducing the heat being developed in the charge.
A
few melts were made in an Arsem furnace to study the effect
of melting under vacuo, hydrogen, and carbon monoxide. These
melts were made in magnesia or " impervite " crucibles 18 (0.7 mm
inch) in diameter and 75 mm
(3 inches) high. The crucible con-
taining the material to be melted was placed in a protecting tube
which was then inserted in the furnace. The furnace was closed,
evacuated, and the charge could then be melted. Hydrogen or
carbon monoxide could be introduced or pumped from the furnace
at any stage of the run. As it was necessary to allow the furnace to
cool before opening, no additions were made to melts made in this
furnace.

7 The physical properties of nickel ; Trans., A. I. M. E., 64, p. 387 ; 1920.

w1m£iberg\
Malleability and Metallography of Nickel 159

The melts in the gas furnace were made in covered No. 10 plum-
bago crucibles, the procedure being as follows: The charge, usually
consisting of 3,000 g (6.8 pounds) of electrolytic nickel with a little

charcoal, would be placed in high

a crucible previously heated to a
temperature. about one and one-quarter hours to heat
It required
the melt to a proper pouring temperature. An ingot weighing about
550 g (1.2 pounds) would be poured, an addition to produce a
desired alloy would be made and a second ingot poured. Further
additions would be made and other ingots poured until the desired
ingots were obtained or until the melt became too cold to pour prop-
erly. This usually occurred after pouring the third ingot. The
melt would then be replaced in the furnace, reheated for about 15
minutes, and the procedure of making additions and pouring con-
tinued. In no case was a melt reheated after an addition of magne-
sium. From three to six ingots were produced from each melt.
These ingots were cast in tapering split-steel molds with a circular
cross section. The diameter of the ingots was 20 mm
(0.8 inch) at
the bottom, 30 mm
(1.2 inches) at the top, with a sink head which
tapered from 30 to 50 mm
(2 inches) in 9 mm
(0.35 inch). The
length of the ingots, exclusive of the sink head, was 110 (4.3mm
inches). The weight of the ingots with sink head was about 550 g
(1.2 pounds).
The sulphur content of our alloys was controlled by adding
weighed amounts of nickel-sulphur alloys of known composition.
These nickel-sulphur alloys were made by adding nickel sulphide
(Ni 3 S 2 ) to molten nickel and contained about 2 per cent sulphur.
Granular metallic manganese and metallic magnesium were used
for the additions of these elements. During the first part of this
work the magnesium were held in nickel tongs and plunged
pieces of
into the molten charge, later the additions were made by wedging
the magnesium in the end of a carbon rod which was then plunged
into the molten charge.
The methods we used for introducing other elements, such as
silicon and arsenic, are described in the section of this report dealing
with the tests made on alloys containing these elements.

2. TESTS FOR MALLEABILITY

The was determined at 1,000 to 1,100° C.

malleability of the alloys
(1,830 to 2,010° F.)by hammering one end of a specimen to a wedge.
If no cracks were visible at the edge and sides of the wedge, the
specimen was termed malleable. No attempt was made to grade the
alloys into more than the two classes, malleable and nonmailable.
The lower end of an ingot, rounded by filing, was usually taken
for this test. A specimen to be forged was placed on the apron of
a gas forge for about 10 minutes, then in the forge, which was held
 :

160 Technologic Papers of the Bureau of Standards

at 1,100° C. (2,010° F.), for about 10 minutes. It was then ham-

mered on an anvil with a 5-pound hammer. The alloys containing
above 0.2 per cent carbon or manganese usually required more than
one heating in order to forge to a satisfactory wedge while hot.
It is believed that this test for malleability as we applied it was
not a particularly severe one as compared with commercial forging
and rolling operations. Some specimens, such as N-l, containing
0.01 per cent sulphur which could be forged to a wedge nicelv but

Appearance of a set of forged nickel specimens illustrating the effect

of additions upon malleability, XI
(a) Specimen 74, nickel containing 0.005 per cent sulphur
(6) Specimen 74A, nickel containing 0.021 per cent sulphur
(c) Specimen 74B, same as 74A after addition of 0.24 per cent manganese
(d) Specimen 74C, same as 74B after addition of 0.1 per cent magnesium

could not be drawn into a bar would crack after repeated heatings
followered b}^ hammering.
The appearance of a set of forged specimens from one melt is
shown in Figure 1.
III. RESULTS
i. MALLEABILITY OF PURE ELECTROLYTIC NICKEL

It Avas determined that commercial electrolytic nickel was malle-

able before and after remelting. The nickel used was obtained from
two sources, the Orford and the Port Colborne works of the Inter-
national Nickel Co. The composition of two shipments of these
nickels, as determined by J. A. Scherrer, chemist of the Bureau of
Standards, was as follows

Ni(by
Cu Co Fe C • differ-
ence)

Per cent Per cent Per cent Per cent Per cent Per cent
0.12 0.30 0.11 0.005 0.002 99.4
.01 .24 .11 0.03 .001

Note.— Si, Me, Pb, llg, As, Sb, and Sn were nut detected in either.

A piece of each of these cathode nickels was cold-rolled to about

0.12 inch, then cold-drawn to 0.022 inch, No. 23 B. & S. gauge. No
difficulty was experienced in drawing this wire Avithout any heat
treatment.
ira&Ui rg]
Malleability and Metallography of Nickel 161

Malleable ingots of composition shown in Table 1 were produced

by melting these nickels in vacuo, in an open crucible in an induc-
tion furnace, and in a gas-fired furnace. Nodifference was observed
in the forging properties of these nickels, and both were used in the
preparation of alloys for this investigation.

2. NICKEL OXIDE AMD NICKEL

been considered that the presence in cast nickel
It has generally
of nickel oxide renders it nonmailable. A
number of melts was,
therefore, made to investigate this point.
Xickel melted in the presence of oxygen forms an oxide (NiO)
which forms an eutectic with nickel. This eutectic contains about
0.24 per cent oxygen, corresponding to 1.1 per cent NiO, and melts
at 1,138° C. (2,620° F.), 11°below the melting point of nickel. The
oxygen content of the eutectic was determined by L. Jordan, chemist
of this bureau, by using the vacuum fusion method of analysis, and
was checked by the solution method for oxide determinations used
by the International Nickel Co. The composition of the oxide was
determined as NiO by analyses of the oxide residue obtained in the
latter method of analysis. The microstructures of several specimens
containing the oxide eutectic, and one containing free primary oxide,
are shown in Figure 2.
The eutectic temperature was determined directly by thermal
analysis of specimens placed in vacuo directly upon the hot junction
of a platinum, platinum rhodium thermo-couple. The melting point
of nickel oxide was determined by C. O. Fairchild, physicist of this
bureau, with a micropyrometer. Black nickel oxide from the In-
ternational Nickel Co. melted at 1,652° C. and Kahlbaum nickel
oxide at 1,660° C. Microscopic examination indicated the presence
of only one oxide in nickel. The oxide content in melts made both
in vacuo and under one atmosphere pressure of air, and carrying
from 0.05 to 6.5 per cent oxygen, was shown bj direct analysis to be
NiO. It seems, therefore, that the evidence here presented is suf-
ficiently complete to establish the fact that NiO forms a simple
eutectic series with nickel, stable at least up to 25 per cent nickel
oxide, and probably up to the melting point of nickel oxide itself.
The iiquidus temperatures within this series were not directly de-
termined.
Contrary to expectation, we found that nickel containing nickel
oxide, at least up to the eutectic composition, is malleable. A num-
ber of ingots containing oxygen in varying amounts up to the
eutectic composition were tested, and all were malleable. Further
to demonstrate this point, a wire 0.009 inch in diameter, No. 31 B. & S.
gauge, was drawn from a specimen (No. 00) consisting of the
eutectic of nickel and nickel oxide, of which the microstructure is
24304°— 25 2
 —

162 Technologic Papers of the Bureau of Standards

Fig. 2. Microstructurc of nickel containing various amounts of oxygen, X100

(a) Specimen 1, etched, containing approximately 0.20 per cent of the nickel-nickel
oxide eutectic
(6) Specimen 60, etched, eutectic composition, 0.21 per cent oxygen (1.1 per cent
nickel oxide)
(o) Specimen 59, unetched, contains 11 per cent of the free or primary nickel oxide.
A water solution of nitric and acetic acid (30 parts hy volume concentrated nitric
acid, 30 glacial acetic acid, 40 water) was used for the etching here and elsewhere
writertery]
Malleability and Metallography of Nickel 163

shown in Figure 2b. To produce this wire, the ingot was forged
at 1,050 to 1,100° C. to a round bar about 0.25 inch in diameter. It
was then drawn cold through dies which were in steps of 5 to 10
per cent reduction of area. At 0.097 inch in diameter it was an-
nealed by heating to a bright red over an open gas flame, and the
drawing continued with similar annealing at about every 80 per cent
of area.
A eutectic of the oxide in Monel metal has not been identified.
The oxide apparently separates to the grain boundaries of the
Monel metal as free oxide. It has the appearance of nickel oxide,
and analyses of different specimens gave various proportions of
nickel, copper, and iron oxides. This oxide in small amounts also
has no appreciable effect on the malleability of Monel metal.

3. GASES IN NICKEL

It was considered that the nonmalleability of nickel might be

clue to thepresence of gases, and several melts were, therefore, made
in such a manner as to introduce the maximum amount of certain
gases into the molten metal. Trials were made with electrolytic
nickel (1) by melting and solidifying under hydrogen, (2) by melt-
ing and solidifying under carbon monoxide, (3) by melting in a
gas-fired furnace in an open crucible, and (4) by melting in an
induction furnace in an open crucible. All of the ingots so produced
were malleable. Likewise, a malleable ingot was produced when
nickel oxide was reduced, melted, and solidified under carbon mon-
oxide. A melt of nickel containing oxide was reduced by the addi-
tion of carbon and a malleable ingot produced.
One is forced to conclude from these tests that nickel is not ren-
dered nonmalleable by the presence of CO, C0 2 2 or N 2
, , Cyano- .

gen was not tried, although the open crucible experiments allowed
opportunity for its formation if it were stable under these conditions.

4. SULPHUR IN NICKEL

It will be noted that the experiments up to this stage had yielded

the rather unexpected result that none of the conditions or additions
produced nonmalleability. The addition of even small amounts of
sulphur, however, rendered nickel nonmalleable, both hot and cold-
A number of melts were made under various conditions to demon-
The results are shown in Tables 1 and 2,
strate this satisfactorily.
where the composition of a number of malleable and nonmalleable
specimens of nickel and Monel metal are listed.
 ——

164 Technologic Papers of the Bureau of Standards [Voi.,19

Table 1. Malleable specimens (not treated toith magnesium)

NICKEL

Composition

men No. Furnace used Remarks

C Si Mb

Per cent Per cent Per cent

15 Nil. Trace.
22 do Nil. 0.002
23 do_ Nil. Solidified under hydrogen.
19 do Nil.
do Nil. Contained 0.05 per cent O2.

60 Contained 0.2 per cent O2.

Nil. .002
Bl __ 0.57 .002
Dl _. do .43 .003
63 do _ .06 .003

71 do. .19 .004

65 do .03 .004
64 do .05 .005
74_ do .36 .005
High-frequency induction Nil. .009

69 do .01 .005
Nl Gas .05 .01
33 2 1.0
N2 .05 .01 .50
75 High-frequency induction .17 .005

94 Gas .24 .001

95___ do .37 .001 2 1.75
97C do .17
101
Contained 2 per cent Fe and
0.2 per cent O2.

MONEL METAL

Gas 0.001
do 0.03 .001
90 .001 Do.
."35"
.002

Contained 0.2 per cent O2.

122 do....

1
Some of these sulphur analyses were made by the evolution and s by the gravimetric method and
re regarded as accurate within ±0.003 per cent.
-'
Amount added to melt.

Table 2. Nonmalleable specimens (not treated with magnesium)

NICKEL

Composition
Furnace used Remarks
men No.
C Si Mn

Per cent Per cent Per cent

Nil. 0.008
High-frequency induction 0.01 .01
.01 .01
67A do .02 .017
.01 .017

.05 .020
do .04 .020
74A do .32 .021
74B do .021 0.24
65B do._ _ !03 .021 .11

re made by the evolution and some by the gravimetric method and

: per cent.
 —

e
¥v ait7nberg]
Malleability and Metallography of Nickel 165

Table 2. Nonmalleable specimens (not treated with magnesium) —Continued

NICKEL —Continued
Composition
Furnace used Remarks
men No.
C S Mn

Per cent Per cent Per cent

0.024
62B do .02 .023 .20
High-frequency induction .01 .023
.02 .023
71B do _ .03 .023 .23

K3 . ..do .07 .024 .37

do .05 .025 .20
.02 .026
.01 .027 .18
D2 . . .do .44 .034

High-frequency induction Nil. .013

Nil. .014
.52 .028
95A do ..... .34 .045 11.75

MONEL METAL

0.02 0.009
6~02~
do .01 .009
90A do _ .015
96A do .35 .004
do .35 .008
96C do ... .35 .010

1
Amount added to melt.
Note.— Malleable ingots were obtained from melts K, 63, 66, 71, 74, and 84 by the addition' of 0.1 per
cent Mg.

To demonstrate that Monel metal low in sulphur is malleable, a

wire 0.009 inch in diameter, No. 31 B. & S. gauge, was drawn from an
ingot (No. 96) containing 0.37 per cent carbon and 0.002 per cent
sulphur. This ingot was cast from a gas furnace melt composed of
69.4 per cent electrolytic nickel, 28.6 per cent electrolytic copper, 2
per cent electrolytic iron, and a little charcoal. This ingot was
poured before any additions were made to the melt. The ingot was
forged at 1,000 to 1,100° C. to a bar about 0.5 inch in diameter. It
was then drawn cold with annealings at 950° C. at about every 80
per cent reduction of area to 0.033 inch in diameter, and then it was
drawn cold to 0.009 inch in diameter.
The reason for this very striking effect of sulphur in nickel is not
far to seek. Bornemann, 8 Hayward, 9 and others have studied the
equilibrium of nickel and sulphur and have shown that sulphur
forms a compound (Ni 3 S 2 ) which forms a eutectic with nickel at
21.5 per cent sulphur that melts at 644° C. This compound when
present in small amounts in cast nickel assumes the form of films
surrounding partially or completely the individual grains of nickel.
8
Metallurgie, 5, p. 13 ; 1908 ; 7, p. 667 ; 1010.
"Trans. A. I. M. E., 48, p. 141 ; 1914.
 —

166 Technologic Papers of the Bureau of Standards lYoi,:

Fig. 3. Physical state in which sulphur may occur in nickel, X500

(a) Specimen S3A, etched, nickel containing 0.23 per cent sulphur. The nickel
sulphide shows as conspicuous films at the grain boundaries
(6) Specimen 83B, etched, same as 83A treated with 3 per cent manganese. The
manganese sulphide appears as a eutectic network at the grain boundaries
(c) Specimen 83C, unetched, same as 83B, treated with 0.5 per cent magnesium.
The magnesium sulphide appears as hexagonal inclusions within the nickel grains
WaitSiterg]
Malleability and Metallography of Nickel 167

It has a characteristic yellowish color and is readily identified under

the microscope. It is shown in Figures 3, 4, and 5. It appears from
the malleability tests that the solubility of this compound in nickel
(solid solution) must be extremely slight, corresponding to less than
0.005 per cent sulphur. The presence of this compound in nickel in
amounts corresponding to as little as 0.02 per cent sulphur (un-
by thermal analysis and microscopically
treated) can be detected both
after annealing. The malleability tests are still more sensitive to its
presence, however, and indicate the lower limit of solubility indi-
cated. Its effect in rendering nickel nonmalleable is readily under-
stood.
The sulphur constituent in Monel metal does not have the same
color as in nickel, and collects in globules at the grain boundaries
rather than as a network. This constituent has a dark blue color
and is attacked by five seconds' immersion in a 10 per cent aqueous
solution of KCN. The oxide in nickel and Monel and the nickel sul-
phide in nickel are not attacked under similar conditions. This con-
stituent apparently is Cu 2 S, and as it is visible in Monel metal con-
taining 0.01 per cent sulphur as cast, and after annealing at 700° C.
its solubility must be very low. The structure of Monel metal con-
taining sulphur, and before manganese or magnesium treatment, is

shown in Ml (fig. 6). The web-like structure of the grains in Ml

isassociated in commercial practice with nonmalleable Monel metal
and is possibly due to a sulphide eutectic. It has not appeared in
any of the small experimental specimens which were treated with
manganese or magnesium. Thermal analyses indicate that this sul-
phide eutectic melts at 570° C. (1,058° F.).

5 . OTHER ELEMENTS
It appeared hardly necessary to give consideration to the possible
effect of certain impurities,such as iron, cobalt, copper, and man-
ganese, since these are known
to form solid solutions with nickel
within wide limits. Likewise, carbon, silicon, and arsenic could
hardly be suspected, but actual trials were made, electrolytic nickel
with additions of these elements being used.
Silicon in amounts usually present in commercial nickel does not
make it nonmalleable. A
nickel containing 0.17 per cent carbon,
0.005 per cent sulphur, and 0.29 per cent silicon was prepared in an
induction furnace and found to be malleable. The silicon was intro-
duced by adding metallic silicon to the molten nickel. The amount
added was 0.35 per cent.
Several ingots containing 0.17 per cent carbon and arsenic in
amounts up to 0.04 per cent were cast from a melt in a gas furnace.
All of these ingots were malleable, so arsenic in amounts occurring
 — .

168 Technologic Papers of the Bureau of Standards [Voi.j$

f ,'

'
•
X . ^-—^«v.
Sj-rtf^'.
'''
^
^t
H^J- .

i#

CL

/ \ ?*
*V"' -p.

V .._

-i
b
m
'

P?F I '•¥•'-"

13
0; \
^
1

;
" v
•T •
,

:;
' &. /

•53-

# ,- *
u
c
d> :'

^ . .i -..
.
&'
Fig. 4. Effect of treatment with manganese and with magnesium upon the
condition in which sulphur occurs in nickel, X500
(a) Specimen 74A, etched, nickel containing 0.21 per cent sulphur. Note the inter-
crystalline sulphide film
(5) Specimen 74B, etched, same as 74A after treatment with 0.24 per cent manganese.
Note the manganese sulphite globules at grain boundaries
(c) Specimen 74C, etched, same as 74B after addition of 0.1 per cent magnesium.
The pits resulting from the decomposition of the magnesium sulphide by the etching
reagent used, indicate the distribution of this constituent
 —

wm?nberg\ Malleability and Metallography of Nickel 169

Fig. 5. Occurrence of sulphur in a commercial melt of nickel as affected by

metallurgical treatment, X500
(a) Specimen XI, etched. This was taken during the tapping- of the furnace and
contains 0.025 per cent sulphur. Note the intercrystalline films also the globules within
the grains
(h) Specimen X2, etched. Same as XI after the addition of 0.5 per cent mangauese.
Note the globules at the grain boundaries
(e) Specimen X3, etched, same as X2 after the addition of 0.1 per cent magnesium.
Note the globules of magnesium sulphide within the grains
 —

170 Technologic Papers of the Bureau of Standards [Vol.

Fig. G. Occurrence of sulphur in a commercial melt of Monet etal as affected

by metallurgical treatment, X500
(a) Specimen Ml, etched. Sample of Monel metal taken during tapping of the furnace,
0.029 per cent nickel. A sulphur constituent appears as globules at the grain boundaries
(6) Specimen M2, etched, same as Ml after addition of 1.65 per cent manganese.
The manganese sulphide occurs as globules at the grain boundaries
(e) Specimen M3, etched, same as M2 after addition of 0.1 per cent magnesium. Note
the sulphide within the grains
 c
w ain"nbcrr] Malleability and Metallography of Nickel 171

in commercial nickel does not make it nonmailable. The arsenic

was added molten nickel as metallic arsenic inclosed in nickel
to the
capsules. The total amount added was 0.045 per cent, and a chemi-
cal analysis on the final ingot gave the arsenic content as 0.04 per
cent.
Commercial practice of adding carbon and manganese to nickel
amounts do not make nickel
indicates that these elements in ordinary
nonmalleable. To determine their effect on electrolytic nickel sev-
eral melts were made containing various amounts of carbon and
manganese. The ingots from these melts were malleable, and the
compositions of some of them, containing carbon or manganese or
both, are listed in Table 1.

6. EFFECT OF MANGANESE AND MAGNESIUM

Having thus demonstrated, as we believe, that it is sulphur and sul-

phur alone of the common impurities which renders ordinary cast
nickel nonmalleable without treatment, it remained to study the na-
ture and mechanism of that treatment.
The behavior of these additions to nickel and Monel was demon-
strated in the following manner Several
: with varying
series of melts
carbon content (0.05 to 0.5 per cent) were obtained in the gas fur-
nace, each melt furnishing several ingots after successive treatments
or additions to the melt. The first ingot from a melt of electrolytic
nickel invariably would prove to be malleable, then about 0.015 per
cent sulphur was added to the melt in the form of a high sulphur al-
loy containing about 2 per cent sulphur, and the second ingot poured,
which on forging proved to be brittle. Then 1 per cent manganese
was added to the remaining molten nickel and the third ingot poured,
which could be forged a little, but would develop cracks. A final
addition to the " melt " of 0.1 per cent magnesium was made and the
last ingot poured, which was always malleable. This same pro-
cedure was followed and gave similar results with Monel metal made
by melting a charge consisting of 69.4 per cent electrolytic nickel,
28.6 per cent electrolytic copper, and 2 per cent electrolytic iron.
The appearance of a set of such specimens of nickel after forging is
shown as Figure 1.
Table 3 lists the composition of a number of nonmalleable ingots,
and those melts from which malleable ingots were produced after the
addition of magnesium are designated.
An addition of manganese
is not required for producing malleable

metal from nickel or Monel metal containing sulphur. The follow-

ing table gives the results of tests on specimens containing sulphur,
but no manganese, which were made malleable by the additions of
magnesium.
 —

172 Technologic Papers of the Bureau of Standards [vol

Table 3. Effect of addition of magnesium

Composition
Specimen number Behavior on forging
C S Mg

Per cent Per cent Per cent

0.52 None.
.52 !028 0.03
.52 .1

.35 .042 .02

.35 .042 .03 Do.
.35 .042 .25

.21 .046 4 None.

.21 .046 .039 '

.21 .046 .080 Do.

.40 .104 None.

.40 .104 .15
.40 .104 .24 Do.

.02 Do.

It appearsfrom the results tabulated in Table 3 that an addition

of magnesium equivalent to the sulphur content is all that is required
to make nickel malleable.
A
melt of nickel (No. 95) containing 1.75 per cent manganese,
0.045 per cent sulphur, and 0.34 per cent carbon was made mallea-
ble by the addition of 0.038 per cent magnesium. As a similar melt
containing no manganese (No. 94) was made malleable by the addi-
tion of 0.039 per cent magnesium, it seems that the manganese con-
tent has no effect on the amount of magnesium required to make
nickel malleable.
The most plausible explanation for the effect of manganese and of
magnesium is that each will take the sulphur from the nickel and by
forming manganese or magnesium sulphide eliminate the dangerous
nickel sulphide. This was readily demonstrated in several ways.
It is known that manganese has a somewhat greater affinity, and
that magnesium has a very much greater affinity, for sulphur than
has nickel.
It was demonstrated that both manganese and magnesium react
with nickel sulphide (Ni 3 S 2 ) at the temperature of molten nickel.
A fusion of metallic manganese and nickel sulphide (Ni 3 S 2 ) in an
alundum crucible produced a metallic button and a slag. This slag
decomposed in cold 0.25 per cent (by volume) hydrochloric acid,
with evolution of hydrogen sulphide, but was not decomposed by
cold or boiling distilled water. A similar fusion of nickel sulphide
and magnesium produced a metallic button and a slag. This slag
decomposed in cold distilled water, hydrogen sulphide being given
off. As nickel sulphide (Ni 3 S 2 ) was not attacked under similar
conditions, it was evident that both the manganese and the mag-
nesium had reacted with the nickel sulphide.
 A

Malleability and Metallography of Nickel 173

wZitmberff]

To determine if similar reactionsoccurred in nickel, three ingots

were obtained from a melt of nickel containing 0.23 per cent sulphur.
The first ingot (83A) was obtained from a melt of electrolytic nickel
to which 0.23 per cent sulphur was added as Ni 3 S 2 the second ingot
;

(83B) was poured from the same melt after the addition of 3 per
cent manganese, and the third ingot (83C) was poured after the ad-
dition of 0.5 per cent magnesium. When specimens of these ingots
were drilled under 5 per cent (by volume) hydrochloric acid, 83
gave off no hydrogen sulphide and 83B and 83C did. When these
specimens were drilled under water, 83A and 83B gave no hydrogen
sulphide and 83C did. These tests were not so satisfactory as might
appear, because the speed and sharpness of the drill by affecting
the temperature of the drillings had considerable effect on the amount
of hydrogen sulphide liberated. However, they confirmed the pres-
ence of MnS in the B sample and of MgS in the C sample.
A more satisfactory indication of differences in the sulphur con-
stituents was obtained by sulphur printing specimens from the three
ingots. Two per cent sulphuric acid (by volume) was used, and
the specimens were in contact with the photographic printing paper
for two minutes. When the specimens were polished in air, a light
print was obtained on 83A and 83C, while a dark print was ob-
tained on 83B. When the specimens were prepared by polishing
under absolute alcohol, a light print was obtained on 83A and darker
prints on 83B and 83C. These prints are shown as Figure 7, a.
A similar set of sulphur prints were obtained on specimens from a
commercial melt of Monel metal. The melt was No. 4,952, and the
chemical analysis of this melt was C, 0.16 per cent; Mn, 1.65 per
cent; Fe, 1.82 per cent; Si, 0.07 per cent; and S, 0.029 per cent.
Specimen No. 1 was taken from the furnace before any addition,
specimen No. 2 was taken from the ladle after the addition of man-
ganese, and specimen No. 3 was taken after the addition of 0.1 per
cent magnesium. These specimens were sulphur printed with 2 per
cent sulphuric acid, and the time of contact was 10 minutes. These
prints are shown as Figure 7, b.
It is believed that these sulphur prints indicate that sulphur ma}^
be present in nickel or Monel metal in any of three forms, depending
upon the treatment the metal has received.
The mechanism of the effect of treating sulphur-bearing nickel
with manganese and magnesium is also Avell demonstrated by ther-
mal analysis. Inverse rate heating and cooling curves are shown in
Figure 8. Specimen 83A contained 0.23 per cent sulphur but no man-
ganese or magnesium. Specimen 100 contained 2.33 per cent sulphur
and 7.96 per cent manganese, and specimen 83C was cast from the
same melt as 83A after the addition of 3 per cent manganese and (15
per cent magnesium. It will be observed that specimen 83A, which
 —

174 Technologic Papers of the Bureau of Standards

Fig. 7. Sulphur prints of nickel and Monel metal melts indicating differences
in the sulphur constituents
(a) Nickel specimen S3A, 0.23 per cent sulphur, 83B, and 83C (fig. 3) polished in air
and moistened with water, (1) as cast, (2) same treated with 3 per cent manganese,
(3) same after further treatment, 0.5 per cent magnesium
(6) Same specimens) repolished under alcohol
(c) Monel metal specimen 4953, 0.29 per cent sulphur, polished in air, and kept wet
with water, (1) as cast, (2) same treated with 1.65 per cent manganese, (3) same alter
further treatment with 0.1 per cent magnesium
(<Z) Same specimen as c repolished under alcohol
 Malleability and Metallography of Nickel 175

llAO'C

logo'
j
i %

5 o
a §
I

900' \

1
5 o
!

800' :

700'
c
j

{
S 3 h
i

y>
i 1

?62t;~e*3' ;

600'
\ \ j

o \ .. \
K 'rises' i
a
+1 1
1 \

-*S2f: {

500'
t \ J^—S<L \
)

£
\

\ j

±go°
1
1 il l

\
S CO

^ 6
1 I

300'

200° \ \

V.. B3A /V. 100 H..83C

0.Z37.S £ .337. S 0.237.S
7 .967.f1n 3.007-Hn
0.507.n$ i a o a

02 cc

Time in seconds to neat or cool 2'C

 — :

176 Technologic Papers of the Bureau of Standards [Yol.19

was untreated, exhibits the nickel sulphide eutectic arrest at 630°

C. (1,166° F.), as well as an arrest at a lower temperature, 530°
C. (986° F.). This latter arrest is presumed to be due to a trans-
formation in Ni 3 S 2 as thermal analyses of fused Ni 3 S 2 gave very-
,

sharp and long arrests at that point. 10

Treatment with manganese, as well as that with manganese fol-
lowed by magnesium, eliminates the nickel sulphide arrests com-
pletely. The MnS eutectic arrest is above the range of these ob-
servations and was demonstrated to occur at 1,325° C. (2,417° F.)
by thermal analyses of nickel melts carrying sulphur and treated
with manganese.
The microstructure of this series so analyzed is typical and is
shown in Figure 3. Ni 3 S 2 is present in the untreated sample (fig.
3, a), MnS in the second specimen treated only with manganese (fig.

3, b), and MgS in the specimen treated with magnesium (fig. 3, c).
In nickel containing sulphur in ordinary amounts about 0.02 per —
cent —
to which manganese has been added the sulphur constituent
appears as globules in the grain boundaries and has a bluish gray
color. When etched with the nitric-acetic acid solution these inclu-
sions are attacked and stains are formed around them as shown in
74B (fig. 4, b) and X2 (fig. 5, b). The inclusions in Monel which
has been treated with manganese have the same appearance, but the
adjacent Monel metal is not stained on etching. This manganese-
sulphur constituent is not attacked when the specimen is immersed
for five seconds in 10 per cent KCN, but the copper-sulphur con-
stituent is These inclusions in Monel are shown in M2
attacked.
(fig. 6, b). Observations on specimens of nickel containing large
amounts of sulphur and manganese have shown that the manganese
sulphide forms a eutectic with nickel containing about 3.1 per cent
sulphur. This eutectic melts at 1,325° C. (2,417° F.), and its appear-
ance is shown in specimen 100 (fig. 9) and in specimen 83B (fig. 3, b).
The reaction between manganese and nickel sulphide in molten
nickel does not appear to be quantitative unless one regard the sul-
phide as having the composition Mn^S,. This is shown by examina-
tion of some specimens having larger amounts of sulphur and man-
ganese, the results of which are given in Table 4

Table 4. Equilibrium condition s for nickel, sulphur, ai d manganese

Free Free
Specimen No. Mn B
Mn" Mn '
Remarks

Per cent Per cent Per cent Per cent

4.97 2.60 0.5 -1.7 Large amount N13S2.
3.90 1.44 1.4 .3
7.96 2.33 4.0 2.0
7.56 2.6 .1

» Assuming sulphide of manganese to be MnS.

6 Assuming sulphide of manganese to be MnjSi.
10
See Bornemann, footnote 8 and Hayward, footnote
; 9, p. 165.
 — ,

Merica 1
Malleability and Metallography of Nickel
Waltenbcrg] 177

As was found when the sulphur prints were made, the sulphur
constituent in nickel andMonel metal which has been treated with
magnesium decomposes in the presence of water. When polished in
the ordinary manner, by using water with an abrasive, the specimens
contained pits distributed through the grains of the metal. These
pits were the result of the decomposition of the magnesium-sulphur
constituent and are shown in 74C (fig. 4, c). The magnesium-sul-
phur constituent was identified by using alcohol in place of water
with the abrasive while polishing specimens. It was then found
that the magnesium-sulphur constituent is distributed through the
grains of nickel or Monel as hexagonal inclusions with a bright blue
color. These inclusions are shown in nickel in 83C (fig. 3, c), and

'"•'

; i\^ ^xxxxoV ^! 1

^-i 1

"VV: Kv^
>V.xH
,. • <* .\\, \> _
,-.

*• * •:••.-•'•'
.•..".•H.«v-;V-

./'•
1

\\[ ':':':':-. '-''."^

\ '»v \> 'VvV.vv '-.

Fig. 9. Microstructure of nickel containing manganese sulphide eutectic, X500

Specimen 100; etched, 2.53 per cent sulphur and 7.96 per cent manganese (fig.8)

in X3 (fig. 5, c). Their appearance in Monel is shown in M3 (fig.

6, c).These inclusions decompose when the specimens are im-

mersed in water, so it is concluded that they are magnesium sul-
phide. Their distribution and appearance indicate that the mag-
nesium sulphide is insoluble in nickel and Monel metal, and that it
solidifies at a higher temperature than either nickel or Monel metal.
The presence of manganese in either nickel or Monel metal ap-
parently has no effect on the formation or appearance of magnesium
sulphide.
An interesting fact is that nickel or Monel metal containing sul-
phur and made malleable by the addition of manganese and mag-
nesium becomes nonmailable when melted in an oxidizing atmos-
phere. Apparently under such conditions the magnesium-sulphur
 :

178 Technologic Papers of the Bureau of Standards [voi.,19

constituent breaks down, and the objectionable nickel sulphide con-

stituent is formed. Therefore, it is desirable that additions of
magnesium be made shortly before pouring a melt.
Although the role of magnesium with respect to sulphur has
here been emphasized as constituting the essential mechanism of the
restoration of malleability, the authors realize that it is not neces-
sarily the only role magnesium plays in the treatment of furnace
nickel. Magnesium undoubtedly does so react where used commer-
cially as a finishing agent for nickel or Monel metal. That it does
serve to reduce the gas content of these metals is proved by analyses
recently made at the Bureau of Standards by L. Jordan, chemist, of
samples submitted by The International Nickel Co. Three samples
of a regular Monel metal heat were taken (1) before any treatment,
(2) after treatment with the usual amount of ferromanganese, and
(3) after additional treatment with the usual amount of magnesium.
The total oxygen and hydrogen contents of these samples are given
below

Sample Per cent by weight

Oxygen Hydrogen

1 0.0214
.0156
.0029

It is quite evident that the magnesium removed oxygen which was

not removed by the manganese addition.
Inasmuch as the magnesium added to molten nickel or Monel metal
reacts with both oxides and with sulphides, in practice the amount
added must divide itself between these two reactions and probably
with greater affinity for the oxide reactions. Hence, if the metal
before treatment with magnesium is not already sufficiently deoxi-
dized by other means, all of the magnesium may be used up in com-
pleting the deoxidation proper. In this case none or insufficient
magnesium will be left for the sulphide conversion, and the metal
produced will not be malleable. Cases of irregularity in the results
produced by adding magnesium in commercial practice may prob-
ably be traced to this cause.

7. EFFECT OF EXCESS OF MAGNESIUM

As magnesium is usually added in excess of that required for the

sulphur present, a few melts were made with the object of determining
what effect an excess of magnesium has on the malleability of nickel
and Monel metal. As much as 0.5 per cent was added to nickel with-
 e
w du7nber ]
Malleability and Metallography of Nickel 179

out appreciably affecting its malleability. The maximum amount of

magnesium found by chemical analysis was 0.33 per cent. An excess
of magnesium in Monel metal is more harmful, specimens containing
more than 0.2 per cent magnesium are not malleable, and when the
magnesium content is in excess of 0.4 per cent Monel metal is brittle
at forging temperatures (1,050 to 1,100° C).
The magnesium-nickel constituent appears in nickel as irregularly
shaped inclusions at the grain boundaries. Before etching these in-
clusions are faintly evident, but a light etching with the nitric-acetic
acid solution revealsthem as a gray inclusion which may be eutectic.
The appearance of the magnesium sulphide and magnesium-nickel
constituents are shown in Figure 10', a and b.
Guertler, in his Metallographie, 1912, page 418, states that a com-
pound (Xi 2 Mg) is formed in nickel rich alloys of magnesium. This
compound forms a eutectic with nickel at 11 per cent magnesium and
the melting point of this eutectic is given as 1,082° C. (1,980° F.).
He magnesium is soluble at the eutectic
indicates that about 2 per cent
temperature. Thermal analyses of some of the specimens indicate
that the eutectic temperature is 1,095° C. (2,003° F.). As the mag-
nesium-nickel constituent is evident in specimens containing 0.1 per
cent magnesium and after heating to 1,100° C. (2,012° F.),
as cast
its solubility in solid nickel must be less than 0.1 per cent.

The magnesium-nickel or copper constituent in Monel metal has

not been identified. When
no sulphur is present with increasing-
magnesium content, from 0.1 to 0.6 per cent, there are increasing
numbers of small brownish inclusions throughout the grains of
Monel. From their distribution one might expect their effect on the
malleability to be less than the effect of Ni 2 Mg in nickel, but, as was
stated above, such is not the case.

IV. SUMMARY AND CONCLUSIONS

The may be summarized as follows:
results of this investigation
First, in order to determine whether pure nickel is naturally non-
malleable, or whether it is rendered so by the presence of impurities,
a series of small laboratory melts was made of pure electronic
nickel, to which were added in some cases known amounts of different
impurities. The ingots of these melts were then hammered hot and
cold and otherwise tested for malleability and ductility. From these
tests certain definite conclusions were reached.
1. Of those elements or compounds naturally occurring in com-

mercial cast nickel, viz, carbon, silicon, iron, copper, arsenic cobalt,
manganese, oxygen, and sulphur, the latter alone is responsible for
itsnonmalleability when untreated. The presence of gases in nickel,
such as carbon monoxide, carbon dioxide, hydrogen, and nitrogen.
 —

180 Technologic Papers of the Bureau of Standards [Voi.id

has likewise no effect on its malleability, except in so far as the

presence of blowholes impairs it.

Fig. 10. Microstructure of nickel treated ilh ,1)11

(a) Specimen 83C, unetched, 0.23 per cent sulphur treated with 3 per cent manganese
and 0.5 per cent magnesium (fig. 8), unetched. Note the magnesium sulphide inclusions
and the Mg-Ni compound
(&) Same alter etching with the nitric acid-acetic acid reagent. The magnesium
sulphide has heen pitted and the magnesium-nickel compound is evident at the grain
boundaries

2. Commercial electrolytic nickel containing less than 0.005 per

cent sulphur is and cold, before and
essentially malleable, both hot
after remelting and without any manganese or magnesium treat-
ment. It can absorb the usual amounts of such impurities as carbon,
w&terff] Malleability and Metallography of Nickel 181

silicon, manganese, iron, copper, cobalt, oxygen, and arsenic and

remain quite malleable. 11
3. As little as 0.01 per cent of sulphur introduced into remelted
electrolytic nickel renders it almost completely nonmailable; in
order that cast nickel may be malleable without treatment the sul-
phur content must be (probably substantially) below 0.005 per cent.
4. This profound effect of such a small amount of sulphur is due to

the formation of a film of nickel sulphide (Ni 3 S 2 ) which surrounds

each metallic grain of nickel and lowers effectively the intercrystal-
line cohesion of the mass of grains even at low temperature. This
sulphide (eutectic) melts at 630° C, and at higher temperatures
there is in consequence practically no cohesion between the nickel
grains.
The addition of small amounts of manganese (0.50 to 1.0 per
5.

cent) to nickel containing 0.01 per cent sulphurand over slightly

improves its malleability. The addition of still smaller amounts
of magnesium (0.05 to 0.10 per cent) completely restores its
malleability.
6. The facts noted (1 to 5) apply equally well to Monel metal and

probably to other alloys of high nickel content.

Having thus determined the origin of the nonmalleability of nickel
it remained to study the mechanism of the curative action of

manganese and magnesium, chiefly by investigating the chemical

reactions resulting from the addition of manganese and megnesium
to nickel containing sulphur. The conclusions of these studies were:
7. Manganese reacts with nickel sulphide (Ni 3 S 2 ) to form man-
ganese sulphide (probably MnS) and nickel; this reaction is
probably not quantitative, but reaches an equilibrium determined by
the concentration of the excess manganese in solution in the nickel.
The manganese sulphide forms an eutectic with nickel melting at
1,325° C, and its distribution in the solid nickel, although along the
grain boundaries, takes the form of small globules, and not films as
does nickel sulphide. Its deleterious effect on the grain cohesion is
not so great as that of the latter. In consequence, the addition of
manganese improves the malleability of nickel but does not restore
it completely.

8- Magnesium has a higher " affinity " for sulphur than either

nickel or manganese and reacts apparently almost quantitatively

with either manganese or nickel sulphide to form magnesium sul-
phide. The latter is insoluble in molten nickel and, indeed, is solid
at these temperatures (about 1,460° C), forming small crystalline

11 After the completion

of this work B. Bogiteh (Compt. rend. 178, p. S55 1924) states
;

that 0.005 per cent sulphur makes nickel cubes or rondelles nonmalleable, whereas the.\
are malleable when the sulphur content is lower. These statements confirm very well
the conclusions of this investigation.
182 Technologic Papers of the Bureau of Standards [V01..19

particles which are entrapped during the solidification of the nickel

and dispersed throughout the grains. They are not concentrated
at the grain boundaries and exert practically no effect in con-
sequence of their uniform mode of distribution on the cohesion or
malleability of the nickel. Magnesium additions, therefore, restore
the ductility of cast nickel containing sulphur by changing the form
of the sulphur from Ni 3 S 2 to MgS and its distribution from that of
intercrystalline films of Ni 3 S 2 (of low melting point) to that of
uniformly dispersed equiaxed particles of MgS (of high melting
point). In connection with this investigation a study was made of
the equilibrium between nickel and nickel oxide.
9. Although the principal function of magnesium, and of man-

ganese to a less extent, in rendering cast nickel malleable is thus

demonstrated to be its effect in changing the form of the sulphur
present, it should not be thought that this is the only role it plays.
Magnesium undoubtedly reacts, as Fleitmann originally claimed,
with oxides, and particularly with carbon monoxide present, and
this also serves to " clegasify " the molten metal as well. Irregular-
ities in the effects produced by added magnesium may, indeed, be
traced very probably to a varying distribution of the amount be-
tween these two functions.
10. Nickel oxide (NiO) and nickel appear to form a simple
eutectic series, the eutectic containing 0.24 per cent oxygen or 1.1
per cent NiO melting at 1,438° C. The solubility of nickel oxide
in solid solution in nickel appears to be slight.
This work was carried on by R. G. Waltenberg as research as-
sociate of The International Nickel Co. at the Bureau of Standards,
and the authors desire to express their deep appreciation of the
privilege of this association with the bureau, and of the many
courtesies and aid of its director, Dr. George K. Burgess. They are
indebted as well to H. A. Bright and J. A. Scherrer, of the bureau,
for many chemical analyses, and to E. J. Huh, research assistant, for
general assistance also to Miss I. J. Wymore, assistant chemist, for
;

most of the thermal analyses included in this paper. To A. J. Wad-

hams, manager of development and research department of the In-
ternational Nickel Co., a special measure of thanks is due for the
generous and sympathetic support he has accorded this work at all
times.

Washington, September 19, 1924.