The Ternary Phase Diagram, Fe-Ni-P - Semantic Scholar
Authors Copy
The Ternary Phase Diagram, Fe-Ni-P A. S. DOAN, Jr. AND J. I. GOLDSTEIN In order to provide the necessary phase equilibria data for understanding the development of the Widmansttten pattern in iron meteorites, we have redetermined the Fe-Ni-P phase dia gram from 0 to 100 pet Ni, 0 to 16.5 wt pct P, in the temperature range 1100° to 550°C. Long term heat treatments and 130 selected alloys were used. The electron microprobe was em ployed to measure the composition of the coexisting phases directly. We found that the fourphase reaction isotherm, where a ÷ liq y + Ph, occurs at 1000° ± 5°C. Above this temperature the ternary fields a + Ph + liq and a + y + liq are stable and below 1000°C, the ternary fields a + y + Ph and y + Ph + liq are stable. Below 875°C a eutectic reaction, liq — ÷ Ph, occurs at the Ni-P edge of the diagram. Altogether nineteen isotherms were determined in this study. The phase boundary compositions of the two- and three-phase fields are listed and are com pared with the three binary diagrams. The a + y + Ph field expands in area in each isotherm as the temperature decreases from 1000°C. Below 800°C the nickel content in all three phases increases with decreasing temperature. The phosphorus solubility in a and y decreases from 2.7 and 1.4 wt pct at 1000°C to 0.25 and 0.08 wt pet at 550°C. The addition of phosphorus to bi nary Fe-Ni greatly affects the a/a + y and y/a + y boundaries below 900°C. It stabilizes the a phase by increasing the solubility of nickel (a/a + y boundary) and above 700°C, it decreases the stability field of the y phase by decreasing the solubility of nickel (y/a ÷ y boundary). However below 700°C, phosphorus reverses its role in y and acts as a y stabilizer, increasing the nickel solubility range. The addition of phosphorus to Fe-Ni caused significant changes in the nucleation and growth processes. Phosphorus contents of 0.1 wt pet or more allow the di rect precipitation of a from the parent y phase by the reaction y a ÷ I’. The growth rate of the a phase is substantially higher than that predicted from the binary diffusion coefficients.
SINCE the early 1800’s metallurgists have studied the transformation which produces the Widmanstätten pattern in iron meteorites. At present the develop ment of the characteristic pattern in iron meteorites is described by the use of the binary iron-nickel phase diagram. However, there is between 0.02 and 3 wt pet P present in the iron meteorites.”2 This element is a ferrite stabilizer whereas nickel is an austenite stabilizer.3 Therefore, we have studied the Fe-Ni-P phase diagram to establish the effect of phosphorus on the precipitation temperature of a ferrite and the Widmanstãtten pattern. The need for phase boundary data in the system Fe-Ni-P was recognized almost forty years ago by Vogel.4 He determined the ternary phase diagram us ing metallographic and thermal analysis. The results of his studies indicated that the two incongruently melt ing binary phosphides (Ph), Fe3P and Ni3P, had com plete mutual solid solubility and that the temperature y + Ph ((Fe of the ternary reaction a + liquid (liq) Ni)3P) is 970°C. His results also indicated that the nickel content of the phosphide phase in the ternary field a + y + Ph increases with decreasing tempera ture. Unfortunately, the solid state reactions that he studied are very sluggish5 below 950°C and the binary diagrams he used to obtain the ternary were incorrect. Because of these factors the phase boundaries deter mined by Vogel are not accurate. Buchwald6 has recently redetermined the iron rich corner of the diagram using sintered powder speci mens. Initial heat treatments at 1100°C for 20 to 40 A. S. DOAN, Jr., is with Planetology Branch, Goddard Space Flight Center, Greenbelt, Md. J. I. GOLDSTEIN, formerly at Planetology Branch, Goddard Space Flight Center, is with Department of Metal lurgy and Materials Science, Lehigh University, Bethlehem, Pa. Manuscript submitted October 8, 1969. METALLURGICAL TRANSACTIONS
hr were followed by anneals of up to 48 hr at 750°C and above, 100 hr at 650°C, and 225 hr at 550°C. Metallographic and X-ray techniques were used to ob tain a series of isothermal sections down to 350°C. The results of his study confirmed the main conclu sions of Vogel and gave the compositions of the ter nary phase field a + y + Ph between 1100° and 35 0°C. Our initial calculations of equilibration times in the Fe-Ni-P system, using measured Fe-Ni diffusion co efficients,7 led us to believe that previous investiga tors had not given their alloys sufficiently long heat treatments. Preliminary work, which used an elec tron microprobe to measure both homogeneity and phase compositions, showed that these calculations were correct and that longer heat treatment times were necessary to approach equilibrium in the ternary system. In order to obtain the phase boundaries of the Fe—Ni-P system, particularly in the solid state range below 750°C, we have redetermined the ter nary phase diagram from 0 to 100 pet Ni and from 0 to 16.5 wt pet P. We prepared over 130 alloys of se lected compositions. Alter long heat treatments, at temperatures between 1100° and 550°C, the coexisting phases were analyzed with the electron microprobe. Direct measurements of homogeneity and phase com positions were obtained. The application of the ter nary ta to the development of the Widmansttten pattern in iron meteorites is described in another paper.8
EXPERIMENTAL PROCEDURE Alloy Preparation The ternary alloys used in this study were prepared by induction melting of weighed quantities of the pure elements. High purity (99.95 wt pet) Fe, nickel with a VOLUME 1,JUNE 1970—1759
Authors Copy Two methods of sample encapsulation were used. For samples run above 900°C, the alloys were placed in alumina crucibles and sealed individually under vacuum (-10-s torr) in quartz tubes. This procedure avoids cross contamination between samples or changes in composition due to phosphorus vaporiza tion. For samples run below 9 00°C, this problem is minimal so that several samples, each individually wrapped in tantalum foil, were sealed off together within the same evacuated quartz tube. Encapsulated samples heat treated at 7 50°C and above were placed directly in the furnaces at the de sired temperature. Encapsulated samples heat treated below 750°C were first heated in the furnace at 900°C for 1 hr and then cooled in a stepwise fashion, 4 hr at 850°C, 2 days at 800°C, 4 days at 750°C, and then fur nace cooled to the heat treatment temperature desired (700°, 650°, 600°, or 550°C). This procedure was adopted in order to prohibit fine particles, produced by large scale nucleation, from forming when the samples were placed directly in the furnace at the desired temperature.
carbon content of less than 20 ppm, and red phos phorus of 99.999 pct purity were used. The ternary alloys were induction melted in an alumina crucible under a protective atmosphere of argon and H2. Small internal voids were noted in most specimens upon metallographic investigation. Altogether some 130 Fe-Ni-P alloys were produced. To prevent phosphorus vaporization each alloy was placed in an alumina crucible, and sealed under vacuum (—10-s torr) in a quartz tube. The alloys were then heat treated at temperatures between 1000° and 1100°C for 5 to 10 days. These treatments served to equi librate the phases present. The alloys were then quenched to room temperature and prepared for ex amination in the probe. Homogeneity of each phase was determined by repeated probe measurements of nickel and phosphorus intensity (5 to 25 points) on each phase. If the data points fell within the range N± 3v’, when N is the average count rate for each phase, the sample was accepted as meeting the homogeneity re quirements. The nominal composition of the bulk sample was obtained within about a ±5 pct relative error by moving the specimen under the electron beam while counting the nickel and phosphorus inten sity. If any phase in the alloy was found to be inhomo geneous, the sample was remelted.
Determination of Phase Compositions Electron probe analysis, which was used in this study, has some important advantages over other con ventional methods (X- rays, quantitative metallography) used for phase diagram analysis. If the alloy phases are at equilibrium at the temperature of interest the electron probe can measure the composition of these phases directly. Tie lines can be obtained directly by measuring the composition of the two coexisting phases at equilibrium. Also in the three phase regions the composition of the three coexisting phases can be meas ured directly and only one alloy is necessary to deter mine the phase field. Even if the various phases are not totally in equilibrium, phase equilibria data can still be obtained by measuring the interface composi tions of coexisting phases.9 This procedure is suitable as long as equilibrium is maintained at the phase in terfaces. After heat treatment, the alloys used in this study were quenched to room temperature. One-, two-, or
Heat Treatments The heat treatments were carried out in tube fur naces controlled to within ±0.5°C or better. In all heat treatments a separate thermocouple was placed next to the samples in the furnace to monitor the actual temperature. A listing of heat treatment temperatures and times are given in Table I.
Table I. Heat Treatments
Temperature, °C
Time, Hr
Estimated Maximum Growth of a, i
1 100 to 995 975 to 925 875 750 700 650 600 550
90 to 120 170 330 200 and 720 720 72Oand 2400 4300 4300
All phases homogenous All phases homogenous All phases homogenous 10 10 6 3 2
875 °C
WT.%Ni 158IP
1000 °C
WT.% Ni
I
70
WT.% P
P
I
I
7.0
WT.%P 16.0
—
—
666NItO
I
• 6.71 N.
-
-
pp
3.0
-
i.o
60
P
1.5
2.0
1.0
5.0 1.0
Tr
50
I 0
I I 30 20 SCANINMICRONS
15,0 2.0
‘54N.
210P
60
15.5
0.70 P
40g.
L°
4.0
0
I I 30 20 10 SCAN IN MICRONS
40.
0
30 20 10 SCAN IN MICRONS
40
Ni-2 wt pct P heat Fig.1 —Electron microprobe data taken on a ternary alloy of nominal composition 92 wt pct Fe-6 wt pet were measured. X-radiation and PK0 336 hr. for and 750°C hr. 168 for NIKr 93 875CC hr. for l000C at treated 1760—VOLUME l,JUNE 1970
METALLURGICAL TRANSACTIONS
Authors Copy three-phase alloys heat treated above 850°C were homogeneous. However samples heat treated below 850°C had at least one inhomogeneous phase, often the matrix phase. In order to obtain the phase boundary at least six phase interfaces were analyzed with the probe by moving the electron beam in 1 p steps data, across the phase boundaries. Fig. 1 illustrates the type of data obtained with the probe for one alloy with a nominal composition of 92 wt pet Fe, 6 wt pct Ni, 2 wt pct P, annealed at three different temperatures, 1000°, 875°, and 750°C. The a and y phases at 1000°C and the a, y, and Ph phases at 875C are well equilibrated. Compositions for the two phase tie line and three phase fields were therefore quite accurately measured. Data for the 750°C sample shows the presence of gradients in the parent a and y phases which have not had sufficient time to anneal out. Repeated interface measurements do however establish
reproducible trends which allow measurement of the equilibrium compositions to be made. Two A.R.L. (Applied Research Laboratory) probes with takeoff angles of 52.5 deg were used during the course of this study. These intruments were operated at 20 kv with a specimen current between 0.01 and 0.05 pa. Either NiK or NiL radiation was measured on one spectrometer and PK was measured simul taneously on another spectrometer. To determine the nickel content within the a and y phases, a calibra tion curve of relative intensity (j/j0)fL for NiKa and L radiation vs nickel content was determined using nine homogeneous Fe-Ni alloys,7 Fig. 2. Calculations of the absorption correction’°”1 for NiKa radiation in phosphides indicates that the binary calibration curve for Fe-Ni can be used to measure the nickel content of phosphides. However, to measure the nickel content of phosphides using NiL, radiation a calibration curve is necessary. This calibration curve (I/I) vs wt pet Ni, Fig. 2, was obtained by simultaneously measuring the NiK and NiL radiation from several phosphides. The NiK radiation was used to measure the nickel content in the phosphides. Meteoritic schreibersite (FeNi)3P from the iron meteorites Odessa and Carlton, which contains nom inally 15.5 wt pet P, was used as a phosphorus stand ard. To obtain the phosphorus content in a, y, and Ph a linear calibration between phosphorus intensity, alter correction for background, and phosphorus in tensity from the standard was used. Background for phosphorus was measured either by readings off the spectrometer peak or on the Fe-Ni standards.
I .0
/ \, I I I \
°)Ni
L
0.5
0.5
RESULTS The ternary isotherms were constructed using the measured phase boundaries obtained in this study. The currently accepted Fe-Ni equilibrium diagram” was used to establish one binary side of the isotherm. The Fe-P binary data for the a/a + y and y/a + y bound aries were redetermined in this study. Fig. 3 gives the isotherm developed at 875°C and in cludes all the one, two, and three phase data from the
1.0
WT Fraction Ni
Fig. 2—Electron microprobe calibration curve, N1K0 and NiLa vs composition, for the three phases a, y, and phos phide.
875°C (16.2) 16 I Iii 14 I o I = 3— C.!) 12
o = °
I....—
Fig. 3—575°C isotherm.
10
cj u—i
3-
4
I I
I\\ III
I I I I
1.1)
III
11111 III II’ iI
II
+
I+Ph!I
+
I— (
1111 1111 III III
i
8 6
II Ii II III I H
.
z
U
PHOSPHIDE
(5.2, 16.1)
II
2 (1.6)
I y+Ph II
H
Ph IHI IHI (4.3, ill I 1.7)
I’ I I
I I I I I I I I II I I I I
II\ U IH I H
I\(6.7 \I 0.75)
II
I
(0.5) (0.9) 4 100% Fe METALLURGICAL TRANSACTIONS
6
I
I.
.65)
Y
8 10 12 14 16 18 WEIGHT PERCENT NICKEL
20
98 100 100% Ni
VOLUME 1,JUNE 1970—1761
Authors Copy Table II. 875°C Composition twt
Ni 6.0 5.0 5.6 5.8 5.96 6 6.2
92.4 92.5 86.6 92.8 92.1 86 92.7
P 1.6 2.5 7.8 1.4 1.94 8 1.24
Average
BINARY Fe—P
+
Ph Phase Field
Ni
P
Nj
P
0.69 0.70 0.80 0.65 0.75 0.80 0.76
5.6 5.2 5.05 5.4 5.4 4.95 5.2
15.5 16.0 16.4 15.7 15.9 16.5 15.8
5.2 ± 0.4
16.0 ± 0.5
Ni
P
4.3 4.35 4.3 4.3 4.35 4.2 4.3
1.7 1.55 1.75 1.7 1.75 1.8 1.78
6.97 6.8 6.45 6.7 6.65 6.30 6.95
4.3 ± 0.1
1.7 ± 0.2
6.7 ± 0.2
REACTION PATHS TERNARY Fe—Ni—P
Phosphide, (Ph)
7
a
Sample Fe
pct) of a +7
0.75 ± 0.08
•
1100°C BINARY Ni—P (0
LIQUID
16 14 (13.7)
19.9. 12.8)
T
j(l2.5)
—
0
0 0
10
z
U
8
°-
y+Iiq.
y+Iiq.
U
//h/
6
I 0
4 (2.7) 2 (0.4)
I (6.2 2.5)76
1.2)
(12)
V
“‘Ô
8 10 12 14 16 18 WEIGHT PERCENT NICKEL
6
2 4 100% Fe
,
20
98 100 100% Ni
1060°C LIQUID ci)
16 (14.2)
\
d
14 I
a-
(0
Fig. 4—Representation of the ternary equilibria which interlinks the binary and ternary reactions.
12
a+Iiqj II
10
1762—VOLUME 1,JUNE 1970
I I
4 (2.7) 2 (0.4)
I
1’
)U.10
/
/ / // //
II
6
I
i”
2 4 100% Fe
//
/
I
/ / II
II
(I
8F
thirty alloys that were analyzed. To make the ternary isotherm more readable a rectangular coordinant sys tem was adopted. The dashed lines indicate the tie lines developed from analysis of two phase alloys and the solid circles indicate the compositions of single phase alloys. The two phase data was reproducible and the continuous slopes measured for the tie lines show the regularity of the data obtained. The bulk composi tions of the multiple phase alloys were only roughly de termined and are not plotted in Fig. 3. The a + y ÷ Ph phase boundary data, obtained from seven different alloys heat treated at 875°C, is given in Table II. The excellent reproducibility of the data insures not only accurate phase composition values, but also demonstrates that only one alloy is necessary to define the three phase region. Isotherms at 1100°, 1080°, 1070°, 1060°, 1054°, 1040°, 1025°, 10100, 1000°, 995° 975°, 925°, 875°, 750°, 700°, 650°, 600°, and 550°C were determined in this study. All of the phase boundary data and a listing of the al loys that were used are given in a separate publica tion.’3 Figs. 3 and 5 to 8 show representative isother mal sections throughout the temperature range inves tigated. Only selected data points and tie lines are
1
I
/f I
/ a+Iiq/+J
/+/ Iiq
ik / 6
+Iq
10 12 14 16 16 8 WEIGHT PERCENT NICKEL
20
98 100 100% Ni
Fig. 5—1100° and 1060°C isotherms.
included in order not to obscure the essential phase relations. A schematic representation of the ternary equi libria which mterlinks the binary and ternary reac tions is given in Fig. 4. Above the Fe-P eutectic tem perature, 1040° ± 10°C, the a + liq field is separated from the y + liq field by a three-phase region a + ÷ liq, Fig. 5, 1100° and 1060°C. Below 1040°C the a + Ph field forms while the a + liq field is still stable. These binary phase regions must be separated by a three-phase field a + Ph + liq. However this phase re gion was not observed in our samples, Fig. 6, 1010°C. The four-phase reaction plane was determined as 1000° ± 5°C, about 30°C higher than previously re METALLURGICAL TRANSACTIONS
Authors Copy 1010°C
(16.5)L_.
I
16
U)
TN
14 U)
0
\
Ph+Iiq
+
I. Ph+Iiq\ (12.3, 115)___ estimated)
12
= °-
Ph
PHOSPHIDE
—
10
ia
(ested)
H
z
y+Iiq
LU C) LU °-
(11.5)
6,
F-
=
4 LU (2.6) 2 (0.5) 10.31 4 2 U 100% Fe
b
8 10 12 14 16 18 WEIGHT PERCENT NICKEL
20
98 100 100% Ni
1000°C±5°C Ph
412.5)
12 10
P÷h
s+PhJ —
Binary Phase Diagrams
(1.4)
PHOSPHIDE
(16.5)
A summary of the phase compositions for the three phase fields a + y + liq and a + y + Ph are given in Table Ill. The ± values represent an estimate of the accuracy of the measured compositions. The effect of phosphorus on the stability of the twophase field a ÷ y is to shift this region from the Fe-P side of the ternary to the Fe-Ni side with decreasing temperature, Figs. 3 and 5 to 8. The nickel content in a and y is increased by the addition of phosphorus. However below about 725°C, phosphorus decreases the nickel content in y with respect to the y phase in the binary system Fe-Ni.
/
Iiqf
6
I
\
y+liq
/
y+iiq
/
The binary Fe-P diagram was constructed by direct measurement of the Fe-P alloys or by extrapolations of the ternary data to the binary side of the diagram. The binary data is given in Table IV along with an es timate of the errors due to measurement and extrapo lation procedures. This data is plotted on the cur rently accepted Fe-P diagram3’15 in Fig. 9. The solid circles in Fig. 9 indicate the compositions of a and phosphide. The open circles are for a and liquid above the eutectic temperature. Note that the liquid points fall in the region formerly considered to be two phase liq + Ph. We found the eutectic temperature at 1040° ± 10°C which is consistent with the currently accepted value.3 995°C
4 2 ‘0 100% Fe
6
8 10 12 14 16 18 WEIGHT PERCENT NICKEL
20
98 100 100% Ni
Fig. 6—1010° and 1000°C isotherms.
ported,4’5 Fig. 6, 1000°C. The reaction, a + liq y + Ph, is classified as a Class 11-4 phase equiUbrium plane14 where the two fields a + y + liq and a + Ph + Ph and + liq are stable above and the two fields a + y + Ph + liq are stable below the four-phase reaction plane, Fig. 6, 1010°C and Fig. 7, 995°C. The composi tions of the four phases at 1000°C are: a = 7.5 ± 0.1 Ni, 2.7 ± 0.1 P, y 9.8 ± 0.1 Ni, 1.4 ± 0.1 P, Ph = 8.0 ± 0.1 Ni, 16.5 ± 0.5?, and liq = 11.5 ± 1 Ni, 12.5 ± 1 P. At 995°C, Fig. 7, the two three-phase fields are sepa rated by the newly formed y + Ph region. Since no al loys fell in the y ÷ Ph + liq field the boundaries which enclose the three phase region are defined by tie lines + Ph and y + liq fields. Below obtained within the 1000°C, the y + Ph field expands and the y + Ph + liq region moves toward the nickel-rich end of the dia gram. At 87 5°C, Fig. 3, a small amount of liquid re mains, but below about 850°C the Ni-P eutectic reac tion, liq y + Ph, occurs eliminating the y + liq + Ph phase field. Altogether the effect of as little as 0.7 wt pct P can suppress the solidification temperature of the low phosphorus end of the ternary by over 600°C. At lower temperatures the a + y + Ph region ex pands in size, Figs. 7 and 8, 750°, 650°, and 550°C. As the temperature decreases the nickel content of a, y, and Ph increases and the maximum phosphorus con tent in a and y decreases. Phosphorus is more solu ble in a than in y by a ratio of 2:1 at 950°C. This ra tio increases to 2.5:1 at 750°C and to 3.3:1 at 550°C.
4 2 ‘‘O 100% Fe
20
98 100 100% Ni
750°C (16.41 16
‘
PHOSPHIDE (6.2, 16.4)
Ph
(15.7)
U)
0 0 ZJZ 0
14 12
0
=
—
METALLURGICAL TRANSACTIONS
8 10 12 14 16 18 WEIGHT PERCENT NICKEL
6
F
z UJ
C) LU 0 I
IC
• a+PhI /I\II 6 I I Ph 1
y+Ph
IY+Ph
4 (7.6, 0.37)
—
(1,0) 2(3.0)4 0 100% Fe
,7
6)6.4)8 10 12 14 16 18 WEIGHT PERCENT NICKEL
,
20
V
98 100 100% Ni
Fig. 7—995° and 750°C isotherms. VOLUME I,JUNE 1970—1763
Authors Copy
a/ca
650°C Ph
PHOSPHIDE T
I
/
1411
0
= a
“
!
/
12
=
i
T
/
(15.5)
(10.5, 15.5)
I
ii
/
,/‘/
I.
z
LU
C.)
7+Ph
LU
a
/1 4jJ+7+Ph a+ 21
2
LU
/
(liD, O.15
(0.61
1’ 0gi2o)
10 12 (148)16 18 8 WEIGHT PERCENT NICKEL
2 (4.9)6 0 100% Fe
20 100% Ni
The a/a + Iiq boundaries are not changed. The ÷ Fe3? boundaries differ near the eutectic temperature and also below 600°C where very long heat treating times (>6 months) were necessary. The a + liq/a boundaries are shifted to higher phosphorus contents. However, there is appreciable error in the composi tions measured because of the difficulty in obtaining reproducible numbers in the liquid phase. Above 650°C the Fe3? phase has about 1 pct more phosphorus than indicated by the stoichiometric formula Fe3P. Below 650°C however the phosphorus content does conform to the stoichiometric formula Fe3P, containing 15.5 wt pet P. Several data points on the Ni-P phase diagram were determined by probe analysis of Ni-P binary alloys. These data points are given in Table V. Although we y + Ph did not define the eutectic temperature liq we found that the present value (880°C)3 is too high. The phosphorus content in liquid is also higher than that of the currently accepted diagram. —
550°C PHOSPHIDE
U)
(15.5’ l4-
/1
,,
= a-
° 0
I
12 I
0
a+Ph
z
LU
/
I
/
I
C) 0 I-
I
I
I
6I.
I
/
(3
=
I
/
A7.4,
2-
/0.26y
in ni
0
I
2
4 (6.3) 8
/
(>19, >13.3)
DISCUSSION
/1
7
/
/)
/ESTIMATED( “+7+Ph I
Electron Probe Method—Precision and Accuracy
I
7+Ph
(19.13 I / +sJ a+)
(2L5)
30N
WEIGHT PERCENT NICKEL
100% Fe
/
10 12 14 16 18 20 22 24 26 28 30
Fig. 8—650° and 550°C isotherms.
In order to assure a precision of ±10 pet or better for measurements of phosphorus contents over 0.2 wt pet, we designed our probe operating conditions in or der to obtain as low a detectability limit for phosphorus as practical. Using a counting time of 40 sec per point, a specimen current of 0.05 (La, and by measuring the background from an Fe-Ni standard of appropriate corn position, we were able to obtain a detectability limit, at the 95 pet confidence limit, of 0.02 wt pet P as calculated by the method of Ziebold.’6
Table Ill. Equilibrium Phase Data Ternary Fe-Ni-P
Wt Pct Ni
Wt Pct P
1100 (1100)6 1080 1070 1060 1040 1025 1010 1000
6.25 ± 0.2 (45) 6.65 ± 0.2 7.2 ± 0.2 6.9 ± 0.3 7.4 ± 0.2 7.3 ±0.3 7.8 ± 0.1 7.5±0.1
2.5 ± 0.1 (2.0) 2.6 ± 0.2 2.55 ± 0.1 2.50 ± 0.1 2.7 ± 0.1 2.7 ± 0.1 2.8 ± 0.1 2.7±0.1
(4.5)
(2.0)
Wt Pet Ni 7.6 ± 0.2 (7.7) 8.7 ± 0.15 9.1 ± 0.1 8.9 ± 0.3 9.4 ± 0.2 9.3 ± 0.1 9.9 ± 0.1 9.8±0.1
Liquid
Ph
7
a Temperature, °C
Wt Pet P
Wt Pet Ni
Wt Pet P
Wt Pet Ni
Wt Pet P
9.9 ± 0.3 (10) 8.4 ± t.8 10.7 ± 2.0 11.3 ± 1.0 10.7 ± 1.0 11.4 ± 1.0 12.3 ± 1.0 11.5± 1.0
12.8 ± 0.5 (9) 12.0 ± 1.0 12.5 ± 0.8 13.5 ± 0.8 11.7 ± 1.0 11.5 ± 1.0 I 1.5 ± 1.0 12.5± 1.0
1.15 ± 0.1 (1.25) 1.2 ± 0.t 1.2 ± 0.1 1.2 ± 0.1 1.3 ± 0.1 1.3 ± 0.1 1.3 ± 0.1 1.4±0.1
8.0±0.1
16.5±0.5
(1.2)
(4.5)
(15.5)
7.0± 0.15 (6.0) 6.2 ± 0.15 5.2 ± 0.4 (10.0) 5.25 ± 0.1 6.2 ± 0.2 (8.0) 8.4±0.5 10.5 ± 0.5 (16) 15 ± 2.5 —19 ± 3 (24)
16.2 ± 0.5 (15.5) 16.3 ± 0.5 16.1 ± 0.5 (15.5) 16.1 ± 0.3 16.4 ± 0.5 (15.5) 15.5±0.5 15.5 ± 0.5 (15.5) 15.5 ± 1.0 >13.3 (15.4)
Fourphase reaction plane (970)6
(9.0)
(12)
(10)
Four-phase reaction plane6 975 (950)6
925 875 (875)6
850 750 (750)6
700 650 (650)6 600 550 (550)6
6.6 ± 0_I (4.5) 5.6 ± 0.1 4.3 ± 0.1 (4.5) 4.2 ± 0.1 4.2 ± 0.2 (4.0) 4.7±0.2 5.3 ± 0.3 (3.8) 5.5 ± 0.5 7.4 ± 0.4 (5.5)
1764—VOLUME 1,JUNE 1970
2.3 ± 0.1 (2.0) 2.05 ± 0.1 1.7 ± 0.15 (2.0) 1.55 ± 0.05 1.03 ± 0.03 (1.3) 0.65±0.1 0.5 ± 0.05 (0.6) 0.32 ± 0.04 0.26 ± 0.02 (0.4)
8.8 ± 0.1 (8.5) 7.7 ± 0.1 6.7 ± 0.2 (7.0) 6.50 ± 0.1 7.6 ± 0.5 (7.0) 9.5±0.5 11.0 ± 0.5 (14) 15 ± 2.0 19 ± 3.0 (24)
1.1 ± 0.05 (1.1) 0.9 ± 0.05 0.75 ± 0.1 (0.8) 0.65 ± 0.03 0.37 ± 0.02 (0.65) 0.27±0.03 0.15 ± 0.04 (0.3) 0.12 ± 0.02 0.08 ± 0.02 (0.1)
METALLURGICAL TRANSACTIONS
Authors Copy Table IV. Fe-P Binary Phase Diagram Data —i
viously reported,4’6 and that the a and Ph phase com positions differ. and Ph phases in the The compositions of the a, three-phase field measured below 700°C differ sig nificantly from previous work. These differences are primarily due to the short heat treatment times used by previous investigators.4’6 The short heat treatment times cause a state of disequilibrium and incomplete phase growth to occur. Therefore, neither the X-ray nor the metallographic techniques used previously are suitable for determining phase compositions at low temperatures (850°C) approaches that given by the precision error alone. For unequilibrated sampies (13.8
0.13 ± 0.03
600
Fe—P
—
Binary
U
w
U
a U
I—
Comparison with Previous Work The phase boundaries of the three phase fields a ÷ y + Ph and a + y ± liq determined in this study are quali tatively in agreement with those of Buchwald,6 Table III, above 700°C. This agreement shows the suitability of either X-ray or microprobe techniques when equi librium is achieved. We also found that the four-phase reaction plane at 1000°C is about 30°C higher than pre METALLURGICAL TRANSACTIONS
0
5
10
15
WI % P
Fig. 9—Fe-P phase diagram. The measured phase boundary values from this study are plotted on the phase diagram. VOLUME 1,JUNE 1970—1765
Authors Copy —(2.7)
(lA)/÷+Ph BOUNDARY (Ni content) Y/a+’i+ Ph (Ni content) (2.05) (0.9) 900 • a+y+Ph (phase field) (0.65) .5) () GIVE THE MAX. P CONTENT INaANDYAT EACH 800 TEMPERATURE 1.0) (0.35) 700 600
\025)
w
500 400 I
0
5
10
I
I
I
I
15
20
25
30
35
I
I
40
45
50
WT% Ni of phosphorus on the sol Fig. 10—Representation of the effect Fe-Ni system. The binary the in ubility limits of a and y nickel concentrations of binary Fe-Ni phase diagram and the the three-phase cor the a/a + y and y/a + ‘ boundaries at on the temperature plotted are field + Ph + a the ners of . diagram composition phase, is formed. This M, a2, a supersaturated bce as the y phase from tion posi com the same has phase to obtain a and y way only which it is formed.’8 The at az at tempera rehe to is tion posi com um of equilibri phase is there a ture T,. The reaction path to form behavior, we this to rast cont fore y —a —a + y. In phosphorus with s alloy i Fe-N for that, rved have obse um phase libri equi the , contents of 0.1 wt pet or more ion react um libri equi The g. a does form on coolin from the y cooled is alloy the as rs occu +y y—a n. Apparently the phase into the two-phase a + y regio tantially influ subs i Fe-N to us phor addition of phos phase. a the of ess proc n eatio nucl the s ence can be calcu te ferri of rate The isothermal growth D, the and d rolle cont on diffusi is lated’9 if growth concen of n interdiffusion coefficient, is not a functio on co diffusi i Fe-N ry bina tration. Assuming that the ns of latio calcu ess, proc growth efficients7 control the t tempera men treat heat the for size cle parti ferrite e I, were made. tures and times used in this study, Tabl te particle ferri yield ns latio calcu these The results of were than ler sizes some 2 orders of magnitude smal the that ed argu be might obtained experimentally. It sys i-P Fe-N the in es phas um libri equi growth of the on diffusi dary boun tem was either controlled by grain on diffusi ry terna bulk or by a large increase in the phorus. coefficients due to the presence of phos
tempera to the binary Fe-Ni system. Plotted on the i phase ry Fe-N bina the are ram diag tion ture composi the a/a + diagram and the nickel concentrations of corners of and )‘/a + y boundaries at the three phase us con phor phos The Ill. e , Tabl field Ph + the a + y several at given s are darie tents of the two phase boun ase incre to is us phor of phos t effec The temperatures. e (a/a + y phas a the of ent cont el nick um maxim the , as tem boundary) over that of the binary. However us phor phos of t amoun the and s ease perature decr phorus soluble in a decreases, the total effect of phos effect of on the nickel content in a is decreased. The inter r rathe is dary boun + the on us y/a phosphor 700°C, esting because of its dual role, Fig. 10. Above g the easin decr by lizer stabi phosphorus acts as an a effect of The e. phas y the in ble solu el of nick t amoun um trans phosphorus is also to increase the equilibri that of the to ct respe with of a re eratu formation temp s contain alloy for , and 700°C below er, Howev ry. bina as an a role its rses reve us ing 9 wi pet Ni, phosphor e in rang bility el solu nick the ases incre stabilizer and the equilibrium transfor ease decr to t is effec Its y. of the mation temperature of a with respect to that for the on ts resul these of ons icati binary. The impl orites mation of the Widmanstätten pattern in mete is discussed in another paper.8
.
CONCLUSIONS
i alloys 1) The solidification temperature of Fe-N with 600°C as much is significantly decreased by as addi In pet). wt 2.8 to us (0.7 phor the addition of phos i-P system Fe-N the in path ion react liquid the tion, by a eutec occurs as follows: liquid is first removed at 1040° Ph, + liq , a Fe-? ry bina tic reaction, in s the nickel content ± 10°C. As temperature decrease phase re of the liquid increases and at 1000°C a fourre de eratu temp As rs. occu +Ph y action a +liq lly enriched in creases below 1000°C, liquid is continua e occurs by nickel and the final loss of the liquid phas ry Ni-P side bina the at Ph + y liq ion react a euteetie of the diagram below 875°C. Fe-P 2) The addition of nickel (0 to 4 wt pet) to the phorus phos the ases incre ry terna the of side binary to lower content of the y loop and pushes the y loop temperatures. expands 3) The a + y ÷ Ph field formed below 1000°C re eratu temp as ons l secti in size within the isotherma in es phas three all in ent cont el nick decreases. The 800°C. The creases with decreasing temperature below from 2.7 s ease decr and y phosphorus solubility in a re pet, wt 0.08 and 0.25 to 1000°C at and 1.4 wt pet . 550°C ied, stud spectively at the lowest temperature tly grea i Fe-N es to us Phas phor 4) The addition of phos Solubility Limits of the a and y from that ges the nucleation and growth process chan de is to iron to wi pet or 0.1 The effect of adding phosphorus of ents of binary Fe-Ni. Phosphorus cont a y loop to the from a crease the stability range of y causing of tion ipita prec t iron is to do just more allow the direc growth The form. The effect of nickel additions to a ion react ÷ the y by phase parent y lity range of that than er high the opposite, that is to increase the stabi rate of the a phase is substantially as might be ex nts. ficie coef on y. In the Fe-Ni-P ternary we found, diffusi ry predicted from the bina of nickel greatly af pected, that the addition of small amounts 5) The addition of phosphorus to Fe-Ni loop y the s expand below 900°C. s darie (0 to 4 wi pet) to the Fe-P binary fects the a/a + y and y/a + y boun of 0.25 wt pet at maximum the ase from a maximum phosphorus content incre to is us phor The effect of phos aining 4 wt pet 0 pet Ni to 0.8 wi pet for an alloy cont + y boundary). e (a/a phas a the of ent cont el nick loop and also minimum Ni. In addition, nickel also pushes the y Above 700°C, phosphorus decreases the res. eratu temp lower to dary) over boun the a/a + y boundary y e (y/a phas + nickel content of the y phosphorus Fig. 10 illustrates the effect of adding ‘.
METALLURGiCAL TRANSACTIONS 1766—VOLUME I,JUNE 1970
Authors Copy that of the binary and increases the equilibrium trans formation temperature of ferrite over that of the bi nary. However below 700C phosphorus reverses its role as an rg stabilizer and increases the nickel sol ubility range in y as well as decreasing the equi librium transformation temperature. 6) The electron probe method can be used very ef fectively to determine phase boundaries in ternary systems. Both equilibrated and nonequilibrated alloys can be analyzed and tie lines in two-phase fields and the composition of ternary fields can be determined directly on the specimens. ACKNOWLEDGMENT
•
The research has been supported in part by the Na tional Science Foundation. The authors are grateful Mrs. for the metallographic and technical assistance of Dr. and ald Buchw Dr. V. with sions . Discus C. Inman L. S. Walter are also gratefully acknowledged. RE FERENCES 1958, i.E. P. Henderson and S. H. Perry: Proc. US. Natural History Museum, vol. 107, p. 339.
Ed., 2. C. B. Moore, C. Lewis, and D. Nava: Meteorite Research, P. M. Miliman, 1969. Press, Reidel 738, p. 3. M. Hansen: Constitution ofBinary Alloys, McGraw.Hhl Book Co., New York, 1958. 4. R. Vogel and H. Baur: Arch EisenhUttenw., 1931, vol.5, p. 269. 5. E. A. Owen and Y. H. Uu: J. Iron Steel Inst., 1949, vol. 163, p. 132. 6. V. F. Buchwald: Acta Potytech. Scand., 1966, p. 1. 7. J. I. Goldstein, R. E. Hanneman, and R. E. Ogilvie: Trans. TMS-AIME, 1965, vol. 233, p. 812. 8. 1. 1. Goldstein and A. S. Doan: Lehigh University, Bethlehem, Pa., unpublished research, 1970. 9. J. I. Goldstein and R. E. Ogilvie: X.ray Optics and Microanalysis, R. Castaing, P. Deschamps, and J. Philibert, Eds., p. 594, Hermann, Paris, 1966. 10. J. Philibert: X-ray Optics and X.ray Microanalysis, H. H. Pastee, V. E. Cosslett, and A. Engstrom, Eda., p. 379, Academic Press, N. Y., 1963. F. 11. P. Duncumb and P. K. Shields: The Electron Microprobe, T. D. McKinley, K. 1966. I. Heintich, and D. B. Wittry, Eds., p. 284, John Wiley & Sons, New York, 12. J. 1. Goldstein and R. E. Ogilvie: Trans. TMS.AIME, 1965, vol. 233, p. 2083. 13. A. S. Doan and J. I. Goldstein: NASA Tech. Note, to be published, 1970. 14. F. N. Rhines: Phase Diagrams in Metallurgy, p. 176, McGraw.Hill Book Co., New York, 1956. 1958, 15. V. N. Svechnikov, V. M. Pan, and A. K. Shurin: Fis. MetalL Metalloved., vol. 6, 662; Phys. Metals Metallog. (USSR), 1958, vol.6, no.4, p. 80, as given in Constitution ofBinary Alloys. First Supplement, R. P. Elliott, Ed., McGrawHill Book Co., New York, 1965. 16. T. 0. Ziebold: Anal. Chem., 1967, vol. 39, p. 858. 17. N. P. Allen and C. C. Earley: J Iron Steel Inst., 1950, vol. 166, p. 281. 18. L. Kaufman and M. Cohen: AIME Trans. 1956, vol. 206, p. 1393. 19. C. Wagner, as given in W. Jost: Diffusion in Solids, Liquids, Gases, p. 68, Academic Press, New York, 1952.
Correction to Trans. TMS-AIME, 1969, vol. 245 and G. M. Lindberg, pp. 1654-55. On the Self-Diffusion of Cotumbium, by Joshua Pelleg Authors Index and Subject Index
appeared in the December issue of Trans. TMS-AIME. It This article was not listed in the Annual Index which vol. 1. will be indexed in 11/Ietatturgicat Transactions, 1970,
Correction to .Tliet. Trans., 1970, vol. 1 als, by J. H. Scaff, pp. 561-73. The Rote of Metatturgy in the Technology of Electronic Materi
Pages 563 and 564 knew the odor” 563 “boron was detected The first five lines which appear in the second column of page “Later” (end of word the follow to 5, 1, line n should be removed from that page and placed on page 564, colum line 4). e of Metatturgical Transactions. The corrected article will appear in the 1970 bound volum .
METALLURGICAL TRANSACTIONS
.
.
VOLUME I,JUNE 1970—1767
The Ternary Phase Diagram, Fe-Ni-P A. S. DOAN, Jr. AND J. I. GOLDSTEIN In order to provide the necessary phase equilibria data for understanding the development of the Widmansttten pattern in iron meteorites, we have redetermined the Fe-Ni-P phase dia gram from 0 to 100 pet Ni, 0 to 16.5 wt pct P, in the temperature range 1100° to 550°C. Long term heat treatments and 130 selected alloys were used. The electron microprobe was em ployed to measure the composition of the coexisting phases directly. We found that the fourphase reaction isotherm, where a ÷ liq y + Ph, occurs at 1000° ± 5°C. Above this temperature the ternary fields a + Ph + liq and a + y + liq are stable and below 1000°C, the ternary fields a + y + Ph and y + Ph + liq are stable. Below 875°C a eutectic reaction, liq — ÷ Ph, occurs at the Ni-P edge of the diagram. Altogether nineteen isotherms were determined in this study. The phase boundary compositions of the two- and three-phase fields are listed and are com pared with the three binary diagrams. The a + y + Ph field expands in area in each isotherm as the temperature decreases from 1000°C. Below 800°C the nickel content in all three phases increases with decreasing temperature. The phosphorus solubility in a and y decreases from 2.7 and 1.4 wt pct at 1000°C to 0.25 and 0.08 wt pet at 550°C. The addition of phosphorus to bi nary Fe-Ni greatly affects the a/a + y and y/a + y boundaries below 900°C. It stabilizes the a phase by increasing the solubility of nickel (a/a + y boundary) and above 700°C, it decreases the stability field of the y phase by decreasing the solubility of nickel (y/a ÷ y boundary). However below 700°C, phosphorus reverses its role in y and acts as a y stabilizer, increasing the nickel solubility range. The addition of phosphorus to Fe-Ni caused significant changes in the nucleation and growth processes. Phosphorus contents of 0.1 wt pet or more allow the di rect precipitation of a from the parent y phase by the reaction y a ÷ I’. The growth rate of the a phase is substantially higher than that predicted from the binary diffusion coefficients.
SINCE the early 1800’s metallurgists have studied the transformation which produces the Widmanstätten pattern in iron meteorites. At present the develop ment of the characteristic pattern in iron meteorites is described by the use of the binary iron-nickel phase diagram. However, there is between 0.02 and 3 wt pet P present in the iron meteorites.”2 This element is a ferrite stabilizer whereas nickel is an austenite stabilizer.3 Therefore, we have studied the Fe-Ni-P phase diagram to establish the effect of phosphorus on the precipitation temperature of a ferrite and the Widmanstãtten pattern. The need for phase boundary data in the system Fe-Ni-P was recognized almost forty years ago by Vogel.4 He determined the ternary phase diagram us ing metallographic and thermal analysis. The results of his studies indicated that the two incongruently melt ing binary phosphides (Ph), Fe3P and Ni3P, had com plete mutual solid solubility and that the temperature y + Ph ((Fe of the ternary reaction a + liquid (liq) Ni)3P) is 970°C. His results also indicated that the nickel content of the phosphide phase in the ternary field a + y + Ph increases with decreasing tempera ture. Unfortunately, the solid state reactions that he studied are very sluggish5 below 950°C and the binary diagrams he used to obtain the ternary were incorrect. Because of these factors the phase boundaries deter mined by Vogel are not accurate. Buchwald6 has recently redetermined the iron rich corner of the diagram using sintered powder speci mens. Initial heat treatments at 1100°C for 20 to 40 A. S. DOAN, Jr., is with Planetology Branch, Goddard Space Flight Center, Greenbelt, Md. J. I. GOLDSTEIN, formerly at Planetology Branch, Goddard Space Flight Center, is with Department of Metal lurgy and Materials Science, Lehigh University, Bethlehem, Pa. Manuscript submitted October 8, 1969. METALLURGICAL TRANSACTIONS
hr were followed by anneals of up to 48 hr at 750°C and above, 100 hr at 650°C, and 225 hr at 550°C. Metallographic and X-ray techniques were used to ob tain a series of isothermal sections down to 350°C. The results of his study confirmed the main conclu sions of Vogel and gave the compositions of the ter nary phase field a + y + Ph between 1100° and 35 0°C. Our initial calculations of equilibration times in the Fe-Ni-P system, using measured Fe-Ni diffusion co efficients,7 led us to believe that previous investiga tors had not given their alloys sufficiently long heat treatments. Preliminary work, which used an elec tron microprobe to measure both homogeneity and phase compositions, showed that these calculations were correct and that longer heat treatment times were necessary to approach equilibrium in the ternary system. In order to obtain the phase boundaries of the Fe—Ni-P system, particularly in the solid state range below 750°C, we have redetermined the ter nary phase diagram from 0 to 100 pet Ni and from 0 to 16.5 wt pet P. We prepared over 130 alloys of se lected compositions. Alter long heat treatments, at temperatures between 1100° and 550°C, the coexisting phases were analyzed with the electron microprobe. Direct measurements of homogeneity and phase com positions were obtained. The application of the ter nary ta to the development of the Widmansttten pattern in iron meteorites is described in another paper.8
EXPERIMENTAL PROCEDURE Alloy Preparation The ternary alloys used in this study were prepared by induction melting of weighed quantities of the pure elements. High purity (99.95 wt pet) Fe, nickel with a VOLUME 1,JUNE 1970—1759
Authors Copy Two methods of sample encapsulation were used. For samples run above 900°C, the alloys were placed in alumina crucibles and sealed individually under vacuum (-10-s torr) in quartz tubes. This procedure avoids cross contamination between samples or changes in composition due to phosphorus vaporiza tion. For samples run below 9 00°C, this problem is minimal so that several samples, each individually wrapped in tantalum foil, were sealed off together within the same evacuated quartz tube. Encapsulated samples heat treated at 7 50°C and above were placed directly in the furnaces at the de sired temperature. Encapsulated samples heat treated below 750°C were first heated in the furnace at 900°C for 1 hr and then cooled in a stepwise fashion, 4 hr at 850°C, 2 days at 800°C, 4 days at 750°C, and then fur nace cooled to the heat treatment temperature desired (700°, 650°, 600°, or 550°C). This procedure was adopted in order to prohibit fine particles, produced by large scale nucleation, from forming when the samples were placed directly in the furnace at the desired temperature.
carbon content of less than 20 ppm, and red phos phorus of 99.999 pct purity were used. The ternary alloys were induction melted in an alumina crucible under a protective atmosphere of argon and H2. Small internal voids were noted in most specimens upon metallographic investigation. Altogether some 130 Fe-Ni-P alloys were produced. To prevent phosphorus vaporization each alloy was placed in an alumina crucible, and sealed under vacuum (—10-s torr) in a quartz tube. The alloys were then heat treated at temperatures between 1000° and 1100°C for 5 to 10 days. These treatments served to equi librate the phases present. The alloys were then quenched to room temperature and prepared for ex amination in the probe. Homogeneity of each phase was determined by repeated probe measurements of nickel and phosphorus intensity (5 to 25 points) on each phase. If the data points fell within the range N± 3v’, when N is the average count rate for each phase, the sample was accepted as meeting the homogeneity re quirements. The nominal composition of the bulk sample was obtained within about a ±5 pct relative error by moving the specimen under the electron beam while counting the nickel and phosphorus inten sity. If any phase in the alloy was found to be inhomo geneous, the sample was remelted.
Determination of Phase Compositions Electron probe analysis, which was used in this study, has some important advantages over other con ventional methods (X- rays, quantitative metallography) used for phase diagram analysis. If the alloy phases are at equilibrium at the temperature of interest the electron probe can measure the composition of these phases directly. Tie lines can be obtained directly by measuring the composition of the two coexisting phases at equilibrium. Also in the three phase regions the composition of the three coexisting phases can be meas ured directly and only one alloy is necessary to deter mine the phase field. Even if the various phases are not totally in equilibrium, phase equilibria data can still be obtained by measuring the interface composi tions of coexisting phases.9 This procedure is suitable as long as equilibrium is maintained at the phase in terfaces. After heat treatment, the alloys used in this study were quenched to room temperature. One-, two-, or
Heat Treatments The heat treatments were carried out in tube fur naces controlled to within ±0.5°C or better. In all heat treatments a separate thermocouple was placed next to the samples in the furnace to monitor the actual temperature. A listing of heat treatment temperatures and times are given in Table I.
Table I. Heat Treatments
Temperature, °C
Time, Hr
Estimated Maximum Growth of a, i
1 100 to 995 975 to 925 875 750 700 650 600 550
90 to 120 170 330 200 and 720 720 72Oand 2400 4300 4300
All phases homogenous All phases homogenous All phases homogenous 10 10 6 3 2
875 °C
WT.%Ni 158IP
1000 °C
WT.% Ni
I
70
WT.% P
P
I
I
7.0
WT.%P 16.0
—
—
666NItO
I
• 6.71 N.
-
-
pp
3.0
-
i.o
60
P
1.5
2.0
1.0
5.0 1.0
Tr
50
I 0
I I 30 20 SCANINMICRONS
15,0 2.0
‘54N.
210P
60
15.5
0.70 P
40g.
L°
4.0
0
I I 30 20 10 SCAN IN MICRONS
40.
0
30 20 10 SCAN IN MICRONS
40
Ni-2 wt pct P heat Fig.1 —Electron microprobe data taken on a ternary alloy of nominal composition 92 wt pct Fe-6 wt pet were measured. X-radiation and PK0 336 hr. for and 750°C hr. 168 for NIKr 93 875CC hr. for l000C at treated 1760—VOLUME l,JUNE 1970
METALLURGICAL TRANSACTIONS
Authors Copy three-phase alloys heat treated above 850°C were homogeneous. However samples heat treated below 850°C had at least one inhomogeneous phase, often the matrix phase. In order to obtain the phase boundary at least six phase interfaces were analyzed with the probe by moving the electron beam in 1 p steps data, across the phase boundaries. Fig. 1 illustrates the type of data obtained with the probe for one alloy with a nominal composition of 92 wt pet Fe, 6 wt pct Ni, 2 wt pct P, annealed at three different temperatures, 1000°, 875°, and 750°C. The a and y phases at 1000°C and the a, y, and Ph phases at 875C are well equilibrated. Compositions for the two phase tie line and three phase fields were therefore quite accurately measured. Data for the 750°C sample shows the presence of gradients in the parent a and y phases which have not had sufficient time to anneal out. Repeated interface measurements do however establish
reproducible trends which allow measurement of the equilibrium compositions to be made. Two A.R.L. (Applied Research Laboratory) probes with takeoff angles of 52.5 deg were used during the course of this study. These intruments were operated at 20 kv with a specimen current between 0.01 and 0.05 pa. Either NiK or NiL radiation was measured on one spectrometer and PK was measured simul taneously on another spectrometer. To determine the nickel content within the a and y phases, a calibra tion curve of relative intensity (j/j0)fL for NiKa and L radiation vs nickel content was determined using nine homogeneous Fe-Ni alloys,7 Fig. 2. Calculations of the absorption correction’°”1 for NiKa radiation in phosphides indicates that the binary calibration curve for Fe-Ni can be used to measure the nickel content of phosphides. However, to measure the nickel content of phosphides using NiL, radiation a calibration curve is necessary. This calibration curve (I/I) vs wt pet Ni, Fig. 2, was obtained by simultaneously measuring the NiK and NiL radiation from several phosphides. The NiK radiation was used to measure the nickel content in the phosphides. Meteoritic schreibersite (FeNi)3P from the iron meteorites Odessa and Carlton, which contains nom inally 15.5 wt pet P, was used as a phosphorus stand ard. To obtain the phosphorus content in a, y, and Ph a linear calibration between phosphorus intensity, alter correction for background, and phosphorus in tensity from the standard was used. Background for phosphorus was measured either by readings off the spectrometer peak or on the Fe-Ni standards.
I .0
/ \, I I I \
°)Ni
L
0.5
0.5
RESULTS The ternary isotherms were constructed using the measured phase boundaries obtained in this study. The currently accepted Fe-Ni equilibrium diagram” was used to establish one binary side of the isotherm. The Fe-P binary data for the a/a + y and y/a + y bound aries were redetermined in this study. Fig. 3 gives the isotherm developed at 875°C and in cludes all the one, two, and three phase data from the
1.0
WT Fraction Ni
Fig. 2—Electron microprobe calibration curve, N1K0 and NiLa vs composition, for the three phases a, y, and phos phide.
875°C (16.2) 16 I Iii 14 I o I = 3— C.!) 12
o = °
I....—
Fig. 3—575°C isotherm.
10
cj u—i
3-
4
I I
I\\ III
I I I I
1.1)
III
11111 III II’ iI
II
+
I+Ph!I
+
I— (
1111 1111 III III
i
8 6
II Ii II III I H
.
z
U
PHOSPHIDE
(5.2, 16.1)
II
2 (1.6)
I y+Ph II
H
Ph IHI IHI (4.3, ill I 1.7)
I’ I I
I I I I I I I I II I I I I
II\ U IH I H
I\(6.7 \I 0.75)
II
I
(0.5) (0.9) 4 100% Fe METALLURGICAL TRANSACTIONS
6
I
I.
.65)
Y
8 10 12 14 16 18 WEIGHT PERCENT NICKEL
20
98 100 100% Ni
VOLUME 1,JUNE 1970—1761
Authors Copy Table II. 875°C Composition twt
Ni 6.0 5.0 5.6 5.8 5.96 6 6.2
92.4 92.5 86.6 92.8 92.1 86 92.7
P 1.6 2.5 7.8 1.4 1.94 8 1.24
Average
BINARY Fe—P
+
Ph Phase Field
Ni
P
Nj
P
0.69 0.70 0.80 0.65 0.75 0.80 0.76
5.6 5.2 5.05 5.4 5.4 4.95 5.2
15.5 16.0 16.4 15.7 15.9 16.5 15.8
5.2 ± 0.4
16.0 ± 0.5
Ni
P
4.3 4.35 4.3 4.3 4.35 4.2 4.3
1.7 1.55 1.75 1.7 1.75 1.8 1.78
6.97 6.8 6.45 6.7 6.65 6.30 6.95
4.3 ± 0.1
1.7 ± 0.2
6.7 ± 0.2
REACTION PATHS TERNARY Fe—Ni—P
Phosphide, (Ph)
7
a
Sample Fe
pct) of a +7
0.75 ± 0.08
•
1100°C BINARY Ni—P (0
LIQUID
16 14 (13.7)
19.9. 12.8)
T
j(l2.5)
—
0
0 0
10
z
U
8
°-
y+Iiq.
y+Iiq.
U
//h/
6
I 0
4 (2.7) 2 (0.4)
I (6.2 2.5)76
1.2)
(12)
V
“‘Ô
8 10 12 14 16 18 WEIGHT PERCENT NICKEL
6
2 4 100% Fe
,
20
98 100 100% Ni
1060°C LIQUID ci)
16 (14.2)
\
d
14 I
a-
(0
Fig. 4—Representation of the ternary equilibria which interlinks the binary and ternary reactions.
12
a+Iiqj II
10
1762—VOLUME 1,JUNE 1970
I I
4 (2.7) 2 (0.4)
I
1’
)U.10
/
/ / // //
II
6
I
i”
2 4 100% Fe
//
/
I
/ / II
II
(I
8F
thirty alloys that were analyzed. To make the ternary isotherm more readable a rectangular coordinant sys tem was adopted. The dashed lines indicate the tie lines developed from analysis of two phase alloys and the solid circles indicate the compositions of single phase alloys. The two phase data was reproducible and the continuous slopes measured for the tie lines show the regularity of the data obtained. The bulk composi tions of the multiple phase alloys were only roughly de termined and are not plotted in Fig. 3. The a + y ÷ Ph phase boundary data, obtained from seven different alloys heat treated at 875°C, is given in Table II. The excellent reproducibility of the data insures not only accurate phase composition values, but also demonstrates that only one alloy is necessary to define the three phase region. Isotherms at 1100°, 1080°, 1070°, 1060°, 1054°, 1040°, 1025°, 10100, 1000°, 995° 975°, 925°, 875°, 750°, 700°, 650°, 600°, and 550°C were determined in this study. All of the phase boundary data and a listing of the al loys that were used are given in a separate publica tion.’3 Figs. 3 and 5 to 8 show representative isother mal sections throughout the temperature range inves tigated. Only selected data points and tie lines are
1
I
/f I
/ a+Iiq/+J
/+/ Iiq
ik / 6
+Iq
10 12 14 16 16 8 WEIGHT PERCENT NICKEL
20
98 100 100% Ni
Fig. 5—1100° and 1060°C isotherms.
included in order not to obscure the essential phase relations. A schematic representation of the ternary equi libria which mterlinks the binary and ternary reac tions is given in Fig. 4. Above the Fe-P eutectic tem perature, 1040° ± 10°C, the a + liq field is separated from the y + liq field by a three-phase region a + ÷ liq, Fig. 5, 1100° and 1060°C. Below 1040°C the a + Ph field forms while the a + liq field is still stable. These binary phase regions must be separated by a three-phase field a + Ph + liq. However this phase re gion was not observed in our samples, Fig. 6, 1010°C. The four-phase reaction plane was determined as 1000° ± 5°C, about 30°C higher than previously re METALLURGICAL TRANSACTIONS
Authors Copy 1010°C
(16.5)L_.
I
16
U)
TN
14 U)
0
\
Ph+Iiq
+
I. Ph+Iiq\ (12.3, 115)___ estimated)
12
= °-
Ph
PHOSPHIDE
—
10
ia
(ested)
H
z
y+Iiq
LU C) LU °-
(11.5)
6,
F-
=
4 LU (2.6) 2 (0.5) 10.31 4 2 U 100% Fe
b
8 10 12 14 16 18 WEIGHT PERCENT NICKEL
20
98 100 100% Ni
1000°C±5°C Ph
412.5)
12 10
P÷h
s+PhJ —
Binary Phase Diagrams
(1.4)
PHOSPHIDE
(16.5)
A summary of the phase compositions for the three phase fields a + y + liq and a + y + Ph are given in Table Ill. The ± values represent an estimate of the accuracy of the measured compositions. The effect of phosphorus on the stability of the twophase field a ÷ y is to shift this region from the Fe-P side of the ternary to the Fe-Ni side with decreasing temperature, Figs. 3 and 5 to 8. The nickel content in a and y is increased by the addition of phosphorus. However below about 725°C, phosphorus decreases the nickel content in y with respect to the y phase in the binary system Fe-Ni.
/
Iiqf
6
I
\
y+liq
/
y+iiq
/
The binary Fe-P diagram was constructed by direct measurement of the Fe-P alloys or by extrapolations of the ternary data to the binary side of the diagram. The binary data is given in Table IV along with an es timate of the errors due to measurement and extrapo lation procedures. This data is plotted on the cur rently accepted Fe-P diagram3’15 in Fig. 9. The solid circles in Fig. 9 indicate the compositions of a and phosphide. The open circles are for a and liquid above the eutectic temperature. Note that the liquid points fall in the region formerly considered to be two phase liq + Ph. We found the eutectic temperature at 1040° ± 10°C which is consistent with the currently accepted value.3 995°C
4 2 ‘0 100% Fe
6
8 10 12 14 16 18 WEIGHT PERCENT NICKEL
20
98 100 100% Ni
Fig. 6—1010° and 1000°C isotherms.
ported,4’5 Fig. 6, 1000°C. The reaction, a + liq y + Ph, is classified as a Class 11-4 phase equiUbrium plane14 where the two fields a + y + liq and a + Ph + Ph and + liq are stable above and the two fields a + y + Ph + liq are stable below the four-phase reaction plane, Fig. 6, 1010°C and Fig. 7, 995°C. The composi tions of the four phases at 1000°C are: a = 7.5 ± 0.1 Ni, 2.7 ± 0.1 P, y 9.8 ± 0.1 Ni, 1.4 ± 0.1 P, Ph = 8.0 ± 0.1 Ni, 16.5 ± 0.5?, and liq = 11.5 ± 1 Ni, 12.5 ± 1 P. At 995°C, Fig. 7, the two three-phase fields are sepa rated by the newly formed y + Ph region. Since no al loys fell in the y ÷ Ph + liq field the boundaries which enclose the three phase region are defined by tie lines + Ph and y + liq fields. Below obtained within the 1000°C, the y + Ph field expands and the y + Ph + liq region moves toward the nickel-rich end of the dia gram. At 87 5°C, Fig. 3, a small amount of liquid re mains, but below about 850°C the Ni-P eutectic reac tion, liq y + Ph, occurs eliminating the y + liq + Ph phase field. Altogether the effect of as little as 0.7 wt pct P can suppress the solidification temperature of the low phosphorus end of the ternary by over 600°C. At lower temperatures the a + y + Ph region ex pands in size, Figs. 7 and 8, 750°, 650°, and 550°C. As the temperature decreases the nickel content of a, y, and Ph increases and the maximum phosphorus con tent in a and y decreases. Phosphorus is more solu ble in a than in y by a ratio of 2:1 at 950°C. This ra tio increases to 2.5:1 at 750°C and to 3.3:1 at 550°C.
4 2 ‘‘O 100% Fe
20
98 100 100% Ni
750°C (16.41 16
‘
PHOSPHIDE (6.2, 16.4)
Ph
(15.7)
U)
0 0 ZJZ 0
14 12
0
=
—
METALLURGICAL TRANSACTIONS
8 10 12 14 16 18 WEIGHT PERCENT NICKEL
6
F
z UJ
C) LU 0 I
IC
• a+PhI /I\II 6 I I Ph 1
y+Ph
IY+Ph
4 (7.6, 0.37)
—
(1,0) 2(3.0)4 0 100% Fe
,7
6)6.4)8 10 12 14 16 18 WEIGHT PERCENT NICKEL
,
20
V
98 100 100% Ni
Fig. 7—995° and 750°C isotherms. VOLUME I,JUNE 1970—1763
Authors Copy
a/ca
650°C Ph
PHOSPHIDE T
I
/
1411
0
= a
“
!
/
12
=
i
T
/
(15.5)
(10.5, 15.5)
I
ii
/
,/‘/
I.
z
LU
C.)
7+Ph
LU
a
/1 4jJ+7+Ph a+ 21
2
LU
/
(liD, O.15
(0.61
1’ 0gi2o)
10 12 (148)16 18 8 WEIGHT PERCENT NICKEL
2 (4.9)6 0 100% Fe
20 100% Ni
The a/a + Iiq boundaries are not changed. The ÷ Fe3? boundaries differ near the eutectic temperature and also below 600°C where very long heat treating times (>6 months) were necessary. The a + liq/a boundaries are shifted to higher phosphorus contents. However, there is appreciable error in the composi tions measured because of the difficulty in obtaining reproducible numbers in the liquid phase. Above 650°C the Fe3? phase has about 1 pct more phosphorus than indicated by the stoichiometric formula Fe3P. Below 650°C however the phosphorus content does conform to the stoichiometric formula Fe3P, containing 15.5 wt pet P. Several data points on the Ni-P phase diagram were determined by probe analysis of Ni-P binary alloys. These data points are given in Table V. Although we y + Ph did not define the eutectic temperature liq we found that the present value (880°C)3 is too high. The phosphorus content in liquid is also higher than that of the currently accepted diagram. —
550°C PHOSPHIDE
U)
(15.5’ l4-
/1
,,
= a-
° 0
I
12 I
0
a+Ph
z
LU
/
I
/
I
C) 0 I-
I
I
I
6I.
I
/
(3
=
I
/
A7.4,
2-
/0.26y
in ni
0
I
2
4 (6.3) 8
/
(>19, >13.3)
DISCUSSION
/1
7
/
/)
/ESTIMATED( “+7+Ph I
Electron Probe Method—Precision and Accuracy
I
7+Ph
(19.13 I / +sJ a+)
(2L5)
30N
WEIGHT PERCENT NICKEL
100% Fe
/
10 12 14 16 18 20 22 24 26 28 30
Fig. 8—650° and 550°C isotherms.
In order to assure a precision of ±10 pet or better for measurements of phosphorus contents over 0.2 wt pet, we designed our probe operating conditions in or der to obtain as low a detectability limit for phosphorus as practical. Using a counting time of 40 sec per point, a specimen current of 0.05 (La, and by measuring the background from an Fe-Ni standard of appropriate corn position, we were able to obtain a detectability limit, at the 95 pet confidence limit, of 0.02 wt pet P as calculated by the method of Ziebold.’6
Table Ill. Equilibrium Phase Data Ternary Fe-Ni-P
Wt Pct Ni
Wt Pct P
1100 (1100)6 1080 1070 1060 1040 1025 1010 1000
6.25 ± 0.2 (45) 6.65 ± 0.2 7.2 ± 0.2 6.9 ± 0.3 7.4 ± 0.2 7.3 ±0.3 7.8 ± 0.1 7.5±0.1
2.5 ± 0.1 (2.0) 2.6 ± 0.2 2.55 ± 0.1 2.50 ± 0.1 2.7 ± 0.1 2.7 ± 0.1 2.8 ± 0.1 2.7±0.1
(4.5)
(2.0)
Wt Pet Ni 7.6 ± 0.2 (7.7) 8.7 ± 0.15 9.1 ± 0.1 8.9 ± 0.3 9.4 ± 0.2 9.3 ± 0.1 9.9 ± 0.1 9.8±0.1
Liquid
Ph
7
a Temperature, °C
Wt Pet P
Wt Pet Ni
Wt Pet P
Wt Pet Ni
Wt Pet P
9.9 ± 0.3 (10) 8.4 ± t.8 10.7 ± 2.0 11.3 ± 1.0 10.7 ± 1.0 11.4 ± 1.0 12.3 ± 1.0 11.5± 1.0
12.8 ± 0.5 (9) 12.0 ± 1.0 12.5 ± 0.8 13.5 ± 0.8 11.7 ± 1.0 11.5 ± 1.0 I 1.5 ± 1.0 12.5± 1.0
1.15 ± 0.1 (1.25) 1.2 ± 0.t 1.2 ± 0.1 1.2 ± 0.1 1.3 ± 0.1 1.3 ± 0.1 1.3 ± 0.1 1.4±0.1
8.0±0.1
16.5±0.5
(1.2)
(4.5)
(15.5)
7.0± 0.15 (6.0) 6.2 ± 0.15 5.2 ± 0.4 (10.0) 5.25 ± 0.1 6.2 ± 0.2 (8.0) 8.4±0.5 10.5 ± 0.5 (16) 15 ± 2.5 —19 ± 3 (24)
16.2 ± 0.5 (15.5) 16.3 ± 0.5 16.1 ± 0.5 (15.5) 16.1 ± 0.3 16.4 ± 0.5 (15.5) 15.5±0.5 15.5 ± 0.5 (15.5) 15.5 ± 1.0 >13.3 (15.4)
Fourphase reaction plane (970)6
(9.0)
(12)
(10)
Four-phase reaction plane6 975 (950)6
925 875 (875)6
850 750 (750)6
700 650 (650)6 600 550 (550)6
6.6 ± 0_I (4.5) 5.6 ± 0.1 4.3 ± 0.1 (4.5) 4.2 ± 0.1 4.2 ± 0.2 (4.0) 4.7±0.2 5.3 ± 0.3 (3.8) 5.5 ± 0.5 7.4 ± 0.4 (5.5)
1764—VOLUME 1,JUNE 1970
2.3 ± 0.1 (2.0) 2.05 ± 0.1 1.7 ± 0.15 (2.0) 1.55 ± 0.05 1.03 ± 0.03 (1.3) 0.65±0.1 0.5 ± 0.05 (0.6) 0.32 ± 0.04 0.26 ± 0.02 (0.4)
8.8 ± 0.1 (8.5) 7.7 ± 0.1 6.7 ± 0.2 (7.0) 6.50 ± 0.1 7.6 ± 0.5 (7.0) 9.5±0.5 11.0 ± 0.5 (14) 15 ± 2.0 19 ± 3.0 (24)
1.1 ± 0.05 (1.1) 0.9 ± 0.05 0.75 ± 0.1 (0.8) 0.65 ± 0.03 0.37 ± 0.02 (0.65) 0.27±0.03 0.15 ± 0.04 (0.3) 0.12 ± 0.02 0.08 ± 0.02 (0.1)
METALLURGICAL TRANSACTIONS
Authors Copy Table IV. Fe-P Binary Phase Diagram Data —i
viously reported,4’6 and that the a and Ph phase com positions differ. and Ph phases in the The compositions of the a, three-phase field measured below 700°C differ sig nificantly from previous work. These differences are primarily due to the short heat treatment times used by previous investigators.4’6 The short heat treatment times cause a state of disequilibrium and incomplete phase growth to occur. Therefore, neither the X-ray nor the metallographic techniques used previously are suitable for determining phase compositions at low temperatures (850°C) approaches that given by the precision error alone. For unequilibrated sampies (13.8
0.13 ± 0.03
600
Fe—P
—
Binary
U
w
U
a U
I—
Comparison with Previous Work The phase boundaries of the three phase fields a ÷ y + Ph and a + y ± liq determined in this study are quali tatively in agreement with those of Buchwald,6 Table III, above 700°C. This agreement shows the suitability of either X-ray or microprobe techniques when equi librium is achieved. We also found that the four-phase reaction plane at 1000°C is about 30°C higher than pre METALLURGICAL TRANSACTIONS
0
5
10
15
WI % P
Fig. 9—Fe-P phase diagram. The measured phase boundary values from this study are plotted on the phase diagram. VOLUME 1,JUNE 1970—1765
Authors Copy —(2.7)
(lA)/÷+Ph BOUNDARY (Ni content) Y/a+’i+ Ph (Ni content) (2.05) (0.9) 900 • a+y+Ph (phase field) (0.65) .5) () GIVE THE MAX. P CONTENT INaANDYAT EACH 800 TEMPERATURE 1.0) (0.35) 700 600
\025)
w
500 400 I
0
5
10
I
I
I
I
15
20
25
30
35
I
I
40
45
50
WT% Ni of phosphorus on the sol Fig. 10—Representation of the effect Fe-Ni system. The binary the in ubility limits of a and y nickel concentrations of binary Fe-Ni phase diagram and the the three-phase cor the a/a + y and y/a + ‘ boundaries at on the temperature plotted are field + Ph + a the ners of . diagram composition phase, is formed. This M, a2, a supersaturated bce as the y phase from tion posi com the same has phase to obtain a and y way only which it is formed.’8 The at az at tempera rehe to is tion posi com um of equilibri phase is there a ture T,. The reaction path to form behavior, we this to rast cont fore y —a —a + y. In phosphorus with s alloy i Fe-N for that, rved have obse um phase libri equi the , contents of 0.1 wt pet or more ion react um libri equi The g. a does form on coolin from the y cooled is alloy the as rs occu +y y—a n. Apparently the phase into the two-phase a + y regio tantially influ subs i Fe-N to us phor addition of phos phase. a the of ess proc n eatio nucl the s ence can be calcu te ferri of rate The isothermal growth D, the and d rolle cont on diffusi is lated’9 if growth concen of n interdiffusion coefficient, is not a functio on co diffusi i Fe-N ry bina tration. Assuming that the ns of latio calcu ess, proc growth efficients7 control the t tempera men treat heat the for size cle parti ferrite e I, were made. tures and times used in this study, Tabl te particle ferri yield ns latio calcu these The results of were than ler sizes some 2 orders of magnitude smal the that ed argu be might obtained experimentally. It sys i-P Fe-N the in es phas um libri equi growth of the on diffusi dary boun tem was either controlled by grain on diffusi ry terna bulk or by a large increase in the phorus. coefficients due to the presence of phos
tempera to the binary Fe-Ni system. Plotted on the i phase ry Fe-N bina the are ram diag tion ture composi the a/a + diagram and the nickel concentrations of corners of and )‘/a + y boundaries at the three phase us con phor phos The Ill. e , Tabl field Ph + the a + y several at given s are darie tents of the two phase boun ase incre to is us phor of phos t effec The temperatures. e (a/a + y phas a the of ent cont el nick um maxim the , as tem boundary) over that of the binary. However us phor phos of t amoun the and s ease perature decr phorus soluble in a decreases, the total effect of phos effect of on the nickel content in a is decreased. The inter r rathe is dary boun + the on us y/a phosphor 700°C, esting because of its dual role, Fig. 10. Above g the easin decr by lizer stabi phosphorus acts as an a effect of The e. phas y the in ble solu el of nick t amoun um trans phosphorus is also to increase the equilibri that of the to ct respe with of a re eratu formation temp s contain alloy for , and 700°C below er, Howev ry. bina as an a role its rses reve us ing 9 wi pet Ni, phosphor e in rang bility el solu nick the ases incre stabilizer and the equilibrium transfor ease decr to t is effec Its y. of the mation temperature of a with respect to that for the on ts resul these of ons icati binary. The impl orites mation of the Widmanstätten pattern in mete is discussed in another paper.8
.
CONCLUSIONS
i alloys 1) The solidification temperature of Fe-N with 600°C as much is significantly decreased by as addi In pet). wt 2.8 to us (0.7 phor the addition of phos i-P system Fe-N the in path ion react liquid the tion, by a eutec occurs as follows: liquid is first removed at 1040° Ph, + liq , a Fe-? ry bina tic reaction, in s the nickel content ± 10°C. As temperature decrease phase re of the liquid increases and at 1000°C a fourre de eratu temp As rs. occu +Ph y action a +liq lly enriched in creases below 1000°C, liquid is continua e occurs by nickel and the final loss of the liquid phas ry Ni-P side bina the at Ph + y liq ion react a euteetie of the diagram below 875°C. Fe-P 2) The addition of nickel (0 to 4 wt pet) to the phorus phos the ases incre ry terna the of side binary to lower content of the y loop and pushes the y loop temperatures. expands 3) The a + y ÷ Ph field formed below 1000°C re eratu temp as ons l secti in size within the isotherma in es phas three all in ent cont el nick decreases. The 800°C. The creases with decreasing temperature below from 2.7 s ease decr and y phosphorus solubility in a re pet, wt 0.08 and 0.25 to 1000°C at and 1.4 wt pet . 550°C ied, stud spectively at the lowest temperature tly grea i Fe-N es to us Phas phor 4) The addition of phos Solubility Limits of the a and y from that ges the nucleation and growth process chan de is to iron to wi pet or 0.1 The effect of adding phosphorus of ents of binary Fe-Ni. Phosphorus cont a y loop to the from a crease the stability range of y causing of tion ipita prec t iron is to do just more allow the direc growth The form. The effect of nickel additions to a ion react ÷ the y by phase parent y lity range of that than er high the opposite, that is to increase the stabi rate of the a phase is substantially as might be ex nts. ficie coef on y. In the Fe-Ni-P ternary we found, diffusi ry predicted from the bina of nickel greatly af pected, that the addition of small amounts 5) The addition of phosphorus to Fe-Ni loop y the s expand below 900°C. s darie (0 to 4 wi pet) to the Fe-P binary fects the a/a + y and y/a + y boun of 0.25 wt pet at maximum the ase from a maximum phosphorus content incre to is us phor The effect of phos aining 4 wt pet 0 pet Ni to 0.8 wi pet for an alloy cont + y boundary). e (a/a phas a the of ent cont el nick loop and also minimum Ni. In addition, nickel also pushes the y Above 700°C, phosphorus decreases the res. eratu temp lower to dary) over boun the a/a + y boundary y e (y/a phas + nickel content of the y phosphorus Fig. 10 illustrates the effect of adding ‘.
METALLURGiCAL TRANSACTIONS 1766—VOLUME I,JUNE 1970
Authors Copy that of the binary and increases the equilibrium trans formation temperature of ferrite over that of the bi nary. However below 700C phosphorus reverses its role as an rg stabilizer and increases the nickel sol ubility range in y as well as decreasing the equi librium transformation temperature. 6) The electron probe method can be used very ef fectively to determine phase boundaries in ternary systems. Both equilibrated and nonequilibrated alloys can be analyzed and tie lines in two-phase fields and the composition of ternary fields can be determined directly on the specimens. ACKNOWLEDGMENT
•
The research has been supported in part by the Na tional Science Foundation. The authors are grateful Mrs. for the metallographic and technical assistance of Dr. and ald Buchw Dr. V. with sions . Discus C. Inman L. S. Walter are also gratefully acknowledged. RE FERENCES 1958, i.E. P. Henderson and S. H. Perry: Proc. US. Natural History Museum, vol. 107, p. 339.
Ed., 2. C. B. Moore, C. Lewis, and D. Nava: Meteorite Research, P. M. Miliman, 1969. Press, Reidel 738, p. 3. M. Hansen: Constitution ofBinary Alloys, McGraw.Hhl Book Co., New York, 1958. 4. R. Vogel and H. Baur: Arch EisenhUttenw., 1931, vol.5, p. 269. 5. E. A. Owen and Y. H. Uu: J. Iron Steel Inst., 1949, vol. 163, p. 132. 6. V. F. Buchwald: Acta Potytech. Scand., 1966, p. 1. 7. J. I. Goldstein, R. E. Hanneman, and R. E. Ogilvie: Trans. TMS-AIME, 1965, vol. 233, p. 812. 8. 1. 1. Goldstein and A. S. Doan: Lehigh University, Bethlehem, Pa., unpublished research, 1970. 9. J. I. Goldstein and R. E. Ogilvie: X.ray Optics and Microanalysis, R. Castaing, P. Deschamps, and J. Philibert, Eds., p. 594, Hermann, Paris, 1966. 10. J. Philibert: X-ray Optics and X.ray Microanalysis, H. H. Pastee, V. E. Cosslett, and A. Engstrom, Eda., p. 379, Academic Press, N. Y., 1963. F. 11. P. Duncumb and P. K. Shields: The Electron Microprobe, T. D. McKinley, K. 1966. I. Heintich, and D. B. Wittry, Eds., p. 284, John Wiley & Sons, New York, 12. J. 1. Goldstein and R. E. Ogilvie: Trans. TMS.AIME, 1965, vol. 233, p. 2083. 13. A. S. Doan and J. I. Goldstein: NASA Tech. Note, to be published, 1970. 14. F. N. Rhines: Phase Diagrams in Metallurgy, p. 176, McGraw.Hill Book Co., New York, 1956. 1958, 15. V. N. Svechnikov, V. M. Pan, and A. K. Shurin: Fis. MetalL Metalloved., vol. 6, 662; Phys. Metals Metallog. (USSR), 1958, vol.6, no.4, p. 80, as given in Constitution ofBinary Alloys. First Supplement, R. P. Elliott, Ed., McGrawHill Book Co., New York, 1965. 16. T. 0. Ziebold: Anal. Chem., 1967, vol. 39, p. 858. 17. N. P. Allen and C. C. Earley: J Iron Steel Inst., 1950, vol. 166, p. 281. 18. L. Kaufman and M. Cohen: AIME Trans. 1956, vol. 206, p. 1393. 19. C. Wagner, as given in W. Jost: Diffusion in Solids, Liquids, Gases, p. 68, Academic Press, New York, 1952.
Correction to Trans. TMS-AIME, 1969, vol. 245 and G. M. Lindberg, pp. 1654-55. On the Self-Diffusion of Cotumbium, by Joshua Pelleg Authors Index and Subject Index
appeared in the December issue of Trans. TMS-AIME. It This article was not listed in the Annual Index which vol. 1. will be indexed in 11/Ietatturgicat Transactions, 1970,
Correction to .Tliet. Trans., 1970, vol. 1 als, by J. H. Scaff, pp. 561-73. The Rote of Metatturgy in the Technology of Electronic Materi
Pages 563 and 564 knew the odor” 563 “boron was detected The first five lines which appear in the second column of page “Later” (end of word the follow to 5, 1, line n should be removed from that page and placed on page 564, colum line 4). e of Metatturgical Transactions. The corrected article will appear in the 1970 bound volum .
METALLURGICAL TRANSACTIONS
.
.
VOLUME I,JUNE 1970—1767
