THE APPARENT EQUILIBRIUM OF THE ALPHA AND GAMMA PHASES OF PURE IRON

The Pennsylvania State College
The Graduate School
Department of Physics
THE APPARENT EQUILIBRIUM-OF THE ALPHA
AND GAMMA PHASES OF PURE IRON
A Dissertation
by
Robert S. Wehner
Submitted in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
A.ugust IS 42
amoved:
4*At fa n
Head of Department
Mai or Professor
17, 17 + 2.______________Date of Approval
A CTCNOVvLEDGEMEN T S
The author is greatly indebted to Professor
Wheeler P. Davey for advice and encouragement during
the exnerimental work and for- help in preparing the
manuscript.
He is also grateful for the assistance
of his wife, Helen, in taking the data.
250341
CONTENTS
Page
I n t r o d u c t i o n .................
Apparatus and Thermometry
Procedure
1
.......................
.........................................
5
Detailed Account of a R u n ............ ..........
8
Method of Correcting Temperature .................
36
C o n c l u s i o n s ......................... ............... 39
R e f e r e n c e s ........................................... 42
Appendix I
X-Ray Tube C i r c u i t ................... 46
Appendix II
The Furnace Spectrometer
..........
48
A upend ix III Specimen H o l d e r ................. - . .
50
Appendix IV
51
A.ppend-ix V
Appendix VI
- Aopendix VII
'Temperature Control Circuits.
. . . .
The G-M Counter and the Counting-Rate
M e t e r ............
Voltage Supply for G-M Counter
Balanced.Filters
........
...
. . . . .
55.
58
60
n?.
LIST OF FIGURES
Page
la.
Part
of the Run
of Aug. 5 - 6 ................. 2.S
lh.
Part
of the nun
of Aug.
1c.
Part
of the Run
of Aug. 5 - 6 ................
Id.
Part
of the Run
of Aug.
le.
5-6 . . -............. ...
5-6. . . . . .
.
18
. . .
gi
Part of the Run of Aug. 5 - 6 ................... £4
If.
Part
of the Run
of Aug.5 - 6 ................. ...
Ig.
Part
of the Run
of A-Ug. 5 - 6 ................. gg
lh.
li.
Part of the Run of Aug. 5 - 6 ..............
Part
of the Run
. 30
of Aug. 5 - 6 ...........
32
2.
Correction Curve for Thermocouple Drift . . .
3.
Suggested Phase Diagram for Pure Iron in
the A 3 R e g i o n ................................ ...
4.
Diagram of.the X-Ray Tube Circuit . . . . . .
5.
Drawing of the Furnace-Speetrometer............47
6.
7.
Specimen Holders..........................
Temperature
Control Circuits
34
46
.
...............
50
52
8 .' G-M
Counter P l a t e a u ...............
9.
G-M
Counter C i r c u i t s .......................... 56
10.
G-M
Counter Voltage Suoply
11.
Balanced Filters ..............................
.
55
................ 59
61
TABLES
Table I.-
Data for Correcting Temperatures . . . .
35
I n an x-ray d i f f r a c t i o n study of the A 3 region in
very pure iron, W a n * s g * r d (.
l) resorted the coexistence of
the alpha and gaaaa p h a s e s
through a temperature range of
several d e g r e e s b e l o w t h e Ac, Point ( 9 1 0 . 5 t 0.6 degrees
C..).
This discovery,
if « ! « ,
importance in any p h y s i c a l
A 3 phase change.
It is
is of considerable
picture of vuiat occurs i n the
thS purpose of the present paper
to present d et ai l ed a d d i t i o n a l evidence apparently
confirming the a l p h a - g a m m e ■equilibrium i n the A, region,
and to point out its e f f e c t on current theory.
■ Apparatus.
The ar)Ps r a 'tus was essentially that used by wangsgard
/
4 r*es
(See
A,p p e„^nGi
CB,J for details.).
'
j_ne ■v— r’^ y tube was a Molybdenum-target Coolidge-type
tube (G.a . v
--^ r a y
Corn.)
operated with Kenotron rectifica-
- --il liamperes, 42 kilovolt peak.
tion at, xh-
The x-ray beam
was c o^l l i sb& ^ e ay
J two 0,040 inch slits Disced 3 and 6
. ,
mcbes
i-r o v
nn th e
specimen.
A sliding filter-bolder vermit-
, , tne a i
insertion of two balanced filters, of
ted
l j-o-rnate
b©- zircoriiuin-o-^-'1'cle and strontium-carbonate, resnectively,
~
' ' into
the x-ray
The
to©3-111*
^i^n.sce-si3ectro3ieter
was thermostated so that the
x
temperature
_
--
the specimen could be controlled to within
0.2 degrees_
-i-n the neighborhood
of 910°C.
—
&
,
of„ the
sDecime-
The temuerature
was measured 'with a Chromel-Alumel thermo-
n u-- ted in place at the end of each run in terms
couule caJLm
of the nielti-
point of U. S. Bureau of Standards silver.
,.
=»r
The e s t
i m a.+
t-,e
^? uncertainties in temperature determinations
'
-i m v s :
were as xox-!--'silver,
n.l
\--J
(l) 0.2 detree for the melting point of
degree in observing the melting point,
(*o')
or•iL C'
Tpg-ree
in observing the Ac3 point of iron, and (4)
\
/ ■
.fc.
0 .2 aegree
<u
to the slight non-linearity of the thermo-
-h-r^ption
curve between 905 and 960°C.
couple c B t 1.!
U ‘
The total
in measurement of absolute temperatures
maximum uric «r'+sintv
e
is t,
h e r eff.o r e
estimated at 0.5 degrees.
Changes in
temperature could be measured to within 0.05 degree.
During every run, a current of oxygen-free, dry, electro­
lytic hydrogen was passed through the furnace so that the
specimen was never allowed to oxidize.
X-rays diffracted by the specimen were detected with
a n-eig er-Mul 1 er quantum counter used with a. conventional
source of high voltage, quenching, pulse-leveling, and
amplifying circuits, and a counting-rate meter involving
a tank circuit consisting of a large condenser, a high
resistance, and a microammeter.
The low ranges of the
microammeter were calibrated in counts per minute by means
of a radioactive source and c. Cenco impulse counter.
The specimen holders (See Aupend.ix III) were made
of electrolytic iron rated 38.98 per cent pure.
The electrolytic iron was made by Dr. E, H. V.allace of
the United States Rubber Company.
They held
the specimen with its face in the axis of rotation,
grazed by the incident x-re.y beam.
tried.
just
Various designs were
In one, a hole in the back of the holder extended
through to the back of the specimen, end a sliver of pure
silver was pushed into notches in the sides of the hole
so that it was held close to^but not in contact with the
4.
specimen,
just behind and slightly below the point at
which the thermocouple was attached.
In another successful
design, small holes were drilled in the front of the holder
on each side of the specimen, .one just above the other,
the other just below the irradiated portion of the specimen.
.Silver slivers stuck into these holes permitted determina­
tion of the temperature gradient over the iron under
observation.
The specimens of iron used throughout the experiment
were cut from a sheet of carbonyl iron kindly supplied by
Dr. R. F. Mehi of the Carnegie Institute of Technology.
This iron was extremely pure, the total impurities being
of the order of 0.001 per cent.
It was the same type of
iron as that used by Mehl and his associates in their
dilatometric studies of the A 3 region, and also by Ham
in his work on the diffusion of hydrogen through iron.
Every precaution was taken to prevent contamination, all
cutting work being done with new hack-saw bf aides and new
files.
To increase the chances of obtaining crystals properly
oriented to diffract in the plane of the incident x-ray
beam and the slit of the G-M counter, most of the
specimens used in the later part of the work were given
preferred orientations by rolling.
The iron was passed
between clean 8-3/4 inch rolls until the total reduction
in thickness was about 70 per cent.
The specimens were
then cut to size (0.5 to 1.5 mm thick, 2 to 4 mm wide,
10 - 15 mm long) and placed in the specimen holder
with the rolled fa.ce as the "reflecting” surface and with
the direction of rolling vertical.
This combination of
rolling-direction and mounting tended to give the (1 1 0 )
planes of the BCC chase such an orientation as to diffract
the x-rays in a horizontal plane so that they could pass
through the furnace window and be received by the quantum
counter.
This orientation was automatically favorable
for diffraction by the (ill) planes of FCG iron.(^-)
Procedure
All experimental work was done late at night and
early in the morning, when the A.C. and D.C. line voltages
were constant.
During the day new specimens, thermo­
couples, and frecuently a new specimen holder, were
installed and all preliminary adjustments made.
In the
early evening, after pure dry hydrogen had been passed
through the furnace for half an hour, the heating current
was turned on and the temperature raised slowly to about
900°C.
This recuired several hours, the heating being
6.
particularly carefully done during the last twenty
degrees because - and this must be kept in mind by the
reader - exact temperatures were never known until the
experiment was over.
By the time the specimen had reached
an apparent temperature of 900°C, the x-ray tube, the G-M
counter voltage source, and the amplifier had been turned
on and allowed to reach their steady states.
Once the A 3
region had been neared the rate of heating was reduced
considerably and a search was begun for diffracted lines.
The counter slit was set approximately*the Bragg angle
89 for a given family of planes, and the specimen holder
was rotated slowly until a line was nicked up, as detected
by clicks from the loud-speaker and by deflection of the
microammeter in the co'unting-rate meter.
The positions of
the counter and of the specimen-face could be controlled
very accurately by slow-raotion attachments, and read to
the nearest minute of arc by means of verniers.
search was continued until a BCC line was found.
The
This
was not always successful, by any means, but with rolled
specimens very fair luck was had.
The search was facilitated
by increasing the width of the slit in front of the counter;
i
i
once a line had been ricked up, this slit was narrowed to
•5 to 15 minutes of arc.
The taking of data required the services of two people:
7.
an observer, to manipulate the temperature control,
adjust the x-ray spectrometer, check the constancy of
the tube current, counter voltage, and counter temperature,
and to read off the data; and a recorder.
Once a line had
been found, the settings adjusted, and the temperature
placed under control, the observer merely read off the
readings of the thermocouple ootentiometer and the countingrate meter, and the time at which they were taken.
After
some practice, it was found possible to take data reliably
at the rate of ten sets per minute, while keeping the
specimen temperature controlled.
This rate was necessary
only when the intensities were changing rapidly.
Readings
were taken more leisurely when the counting-rate was
i
constant.
It should be remembered during the following discus­
sion that the intensities of the diffracted lines were
measured in terms of the deflection of a microammeter
which formed, part of a tank circuit including a 100 micro­
farad condenser and a 300,000 ohm resistance.
Pulses from
the G-M counter were amplified and fed into the condenser,
which continually discharged through the resistance and
the meter, the meter reading being simply the time-rate
of discharge prouortional to the counting-rate.
This
circuit had a time-constant of 30 seconds, and consequently
there was always a lag in the meter reading "behind the
actual counting-rate.
Therefore whenever the intensity
of a line changed appreciably, the meter readings did not
show that change for several" seconds, the actual lag
defending upon the magnitude of the change.
Detailed Account of a Hun
The data to be presented in the following pages
consist of an entire run begun on the evening of August
5th and extending into the early morning of August 6 t h , 1942.
This particular run was chosen because of its
length, the variety of phenomena it reveals, and because
on this night the drift of the thermocouple was small
and very nearly linear with time.
The run was made with
a new specimen of rolled Mehl iron, s new specimen holder,
and a new thermocouple.
Throughout the run the x-ray tube
current was maintained at PA milliamperes, the counter
voltage at 870 volts, and the counter temperature at
51°C.
The hydrogen was passed through a fresh" solution
of hydroc.uinone to remove any trace of oxygen and through
drying bottles freshly filled with CaCl 2 and PgOg.
On the graphs which follow, the broad horizontal line
is the time axis, marked off in minutes.
On tne upper
part of the graph the relative intensities of the diffracted
9.
lines are clotted against time in microamperes of
counting-rate meter deflection.
It was necessary to use
two different scales of sensitivity on the meter, one in
which the range 0 -S00 microamperes corresponded to 0-800
counts per minute, the other to O-fQQO counts per minute.
In order to save vertical space on the graphs of Fig.
l(a-i) meter deflections' were plotted instead of the
actual counting rate.
In the lower part of the graph the
corresponding corrected temperatures are plotted (See
’('a.cjes 3fc' 3#
) . Excect for the correction of temperature
for thermocouple drift the data have not been manipulated
in any way.
Since readings could be tahen every 0.1 or
0.2 minute, it follows that, on the scale to which Fig.
1 is drawn,
the points almost touch each other on some
parts of the intensity curves.
Figure la.
Part of the Run of Aug. 5-6.
Figure la shows the preliminary determination of the
-Ac3 point, the temperature of which is to be used in cor­
recting for drift by the met od to be described later.
The
specimen hsd been maintained at slightly less than 300°C
for several hours previous to the start of the run.
with
a very strong BCC (110) line under observation the
temperature was gradually increased.
The data are graphed
10.
beginning at 9:56.0 P.M. at which time the temperatures
of the A 3 region were approached.
The intensity of the
line decreased slightly but steadily as the temperature
was raised, a decrease which may be due to a gradual shift
in the position of the diffracting crystals (the A 3 region
is one of relatively great atomic mobility).
At 9:41.0
at an observed temperature of 906.7°C (lster corrected to
910.5°C') a very sudden drop in intensity was noted and less
than a minute lster the intensity was down to the back­
ground level (5 to 10 on the scale of ordinates).
Note
the abrupt discontinuity in the temperature-curve: the rate
of he airing changed abruptly at the A c 3 point, remaining
almost at zero while the iron in the specimen and in part
of the specimen holder absorbed the energy necessary to
the alpha-gamma transformation (5.6-4.9 cal/gm).
Frequently,
curves made at slower rates of heating show an actual dip
lasting 0.5 to 0.4 ruin.
After entering the gamma phase some two minutes
elapsed (during which the temperature was not allowed to
drop) while a FCC line was found.
(9:45.6 - 9:47.8)
The next four minutes
show an Ar 3 determination.
The gamma-
alpha transition occurred. 4.4 degrees below the Ac 3 at this
re.oid rate of cooling.
Mote the instability,
(hunting),
of the temperature control as the FCC crystals give up their
energy of transformstion.
This is in marked contrast to
11 .
the Suability of the control while crossing the Ac 3 point
at 3:41.0, and is to have been expected.
12 .
Fig. 1 a.
Part of the flun of Aug. 5-6..
COUNTING-RATE
METER
200
READINGS
FCC III
200
100
100
BCC 110
38
44
40
912
912
910
810
908
'
808
Ac
06
90S
TEMPERATURE
FIG.
in
la
i
Figure lb.
Part of the Run of Aug.
5-6.
At 10:50.0 the run was resumed and a strong BCC
(110)
line was nicked up.
Vvhile the temperature was increasing
(10:50.0 to 10:50.8)
the intensity of this line increased
sharply,
reached a maximum,
and started to fall off,
before the A c 3 point was reached.
This change in intensity
of a BCC line as the A c 3 point is approached was
frequently observed.
At the A c 3 point the decrease in
intensity became more rapid,
but before the transition frcrn
BCC to FCC could be completed,
drop,
the
thermostat having
The line did not disappear,
reduced intensity.
the temperature began to
.just previously been reset.
therefore,
At 10:53.0
spectrometer arm was changed
but remained with
the setting of the
to FCC
(ill).
Vvhile making
the adjustments for this line the temperature dropped two
degrees.
ly,
Upon heating again,
the FCC
(ill)
grew irregular­
the humps in-.the intensity curve checking with
discontinuities in the temperature curve if meter lag is
taken into account.
A shift was made at 10:56.0 and a BCC
(1 1 0 ) was nicked up,
at *a different setting of the s p e c i m e n
holder,
and with reduced intensity.
The reduced intensity
was to have been expected in v i e w of the strong FCG line
just observed,
but the fact that the line was found at all
is worthy of note.
14.
There are, apparently, four possible ways of ac­
counting for the existence of a BCC phase at this point.
a) During the interval between 10:55.9 and 10:56.6, for
which there are no temperature measurements, the temperature
dropped momentarily below the Ar 3 point.
b) The FCC phase may have changed suddenly and spontaneously
into the BCC phase at a temperature between Sll and 908°C.
c) The temperature gradient over the length o f .the specimen
may have been so large that a little of the BCC phase may
have rjersisted during the intermediate heating, to act as
a Mseed” for the further growth of the BCC phase as soon
as the.furnace was cooled.
d) The BCC phase did not disappear at 10:5£.4 but a small
amount of it remained, coexistent with the FCC phase from
10:52.4 to 10:56.6.
We shall consider each of these possibilities in turn:
a) This requires that the temperature fall to the Ar 3
point and rise back to 908.2°C. within 0.7 minutes.
This
rate of cooling means that the Ar 3 would be depressed eight
or so degrees below 910.5.
this low value
up
But a rise in temperature from
to 908.£°C would have taxer at least £
minutes, even if the full heating current had been used.
Explanation (a) is therefore untenable.
b) This requires that the Ar 3 point be not lower than 908.£°C.
15
Fig. 1 b.
Part of the Him of AU g. 5-6.
COUNTING-RATE
METER
200
READINGS
FCC
200
FCC,
IOO
100
BGQ
BCC llO
50'
lOh
52
54'
56
58
9!2‘
910
910
908
908
9 06
904
TEMPERATURE
in
To reach, this point from 910.S°C in 0.7 minutes would
reouire a rate of cooling of about 3 degrees per minute.
It has been shown by v/angsgard^^
that even for a rate of
cooling of only 1 degree/minute the A r 3 point lies about
6 °C. below the A c 3 point,
i.e. at 9 0 4 . 5°C.
rate of -5 degrees per minute,
greater.
E x o l a n e t i o n (b)
For a cooling
the depression is even
is therefore quite untenable.
c) This explanation is hard to meet directly,
since it is
obviously impossible to obtain temperature readings all
over the surface of the
specimen.
The way in which the
silver calibrating slivers melted leads one to believe
that the principal temperature gradient in the specimen
holder is horiztonal.
This is consistent with the fact
that the heating coil and the massive heat-equalizer
were unsyrametricsl where they were cut out for the window
The total temperature
1 degree,
gradient over the specimen is about
while the gradient over the effective part of
the specimen is probably less than 0.5 degree.
,t
gradient,
But a
even of one or two degrees C, is clearly
incapable of providing an A.r3 point anywhere on the
specimen during the interval 10:55.9 to 10:56.6.
ICx plana
''"ion (c) is therefore untenable.
d) We are left w i t h the possibility that both BCC and FCC
iron were present from 10:50 to 10:56.6 and were xn
equilibrium with each other.
The simultaneous presence
of two phases of iron over a range of temperatures is
possible from the standpoint of the phase rule, since
x-ray data apply to the state of the pure iron below the
surface, where the vapor phase is absent.
Therefore it
seems impossible to escape the conclusion that the BCC
■phase was present .all the time the FCC phase was under
observation.
From 10:56.6 to 10:58.1 this BCC ( H O )
line suffered
little change as the temperature was raised to 910
degrees.
Then a. shift was made, during which the
temperature probably passed the Ac 3 point, and a very
strong FCC (111) was picked up at 10:58.8.
It increased
in intensity with decrease in temperature, and persisted
with strength until abandoned at 906°C., at time 11:01.0.
18.
Pig. 1 c.
Part of-the -Son of Aug. 5-6.
C O U N TIN G -R A TE
METER
READINGS
200
200
BCC
100
100
Vi io
BCC
FCC 111
912
'
910
910
908
908
Ac-
906
906f
904
—
TEMPERATURE
FIG.
I c.
904
F igure l c .
Part of the R u n of Aug.
5-6.
O b s e r v a t i o n of the pr e c e e d i n g FCC line was stopped at
11:01.0,
and a BCC
(110)
was picked up at 11:02.4.
intensity was only r bout 30-4-0 on the meter scale,
As its
a
search was made for another crystal dif f r a c t i n g this line.
The spectrometer arm (the B r a g g angle
undisturbed and the
about 10 degrees,
(at 11:03.8).
different
setting)
was
spe c i m e n holder was rotated through
u n t i l a much stronger BCC
(110)
was found
(Frequently as many as three crystals with
orie n t a t i o n s were found to diffract the same
lime through the fur n a c e window.)
This n e w BCC line was
followed f r o m 11: 0 3 . 8 to 11:06.0 w i t h little .change in
intensity d u r i n g a change in temperature of over 2 —1/2°C.
On l e a v i n g this- line the thermostat was
temrerature w h i l e
weak FCC
(111)
set to lower the
a. search was made for an FCC line.
was found at 11:09.4.
A
Alth o u g h it is
believed that the t e m p e r a t u r e never rose to wi t h i n a degree
of the A m 3 du r i n g the 3.4 minutes w h i c h elapsed, since the
previous readings,
it is barely possible in this case
that the A c 3 point was exceeded.
is an a rgument against this,
The weakness oi tne line
and it seems more probable
that this FCC line -is d i f f r a c t e d by gamma iron persisting
during the 8 —1/2 m i n u t e s
since the last FCC observation.
20.
This FCC line was followed until the temperature rose to
908.5°C., at 11:11.7.
Then a BCC (110) was picked up (at
11:12.2) at a temperature of 910.20C.
This line immediately
began to drop in intensity, the A c 3 point having been
unintentionally exceeded.
Cooling was begun at once, in
an effort to save the line, and it was left at 11:13.4
still with moderate intensity,
he have, therefore,
between 11:09.4 and 11:13, the interesting and instructive
phenomenon (see Fig. 1c) of a FCC phase existing in pure
iron at a temperature as low as 907°C. and in the same
specimen a BCC phase at a temperature as high as 910°C.
This ’would, be hard to explain except on the theory that
both phases were present the whole time, so that both of
them could have been observed over the entire period had
the aunaratus nermitted.
21.
Fig. 1 d.
Part of the E m
COUNTING-RATE
of Aug. 5-6.
METER
READINGS
20 C
200
! FCC III
BCC 110
FCC 100
100
! FCC
26
20
28
912' Ilh.
-910
910
9 06
906
906
904
904
TEMPERATURE
FIG.
in
Id
22.
F igure Id.
Part
of the Run of.Aug.
5-6.
At 1 1:14.2 o b s e r v a t i o n s were begun on a strong FCC
i/..(Ill) line.
Its
great intensity is doubtless due to the
previous t r a n s i t i o n t h r o u g h the A c 3 point which lasted
for a fevT seconds.
As
908.0 and 9 0 7 . 1°C^
during the time interval 11:14.3 to
11:16.6,
the line
the temnerature varied between
grew very strong and then decreased to
slightly less than its initial intensity.
907.1°C
(time 11:16.6),
temperature,
with the control set for rise of
a shift was made to a BCC
soon as the line was found
lowered.
Then at
(110)
line.
As
the temperature was again
This B C C phase could hardly have been caused
by passing t h r o u g h the A r 3 point.
The rate of cooling
just previous to 11:16.6 was about 1 degree per minute,
-so that an A r 3 t r a n s i t i o n w o u l d had to have been
considerably d e p r e s s e d .
on rate of cooling,
(According to W a n g s g a r d ’s data
at this cooling rate the A r 3 t r a n s i ­
tion should occur at 904.5;
results at 905°C.)
according to the a u t h o r ’s
But the controls were
temperature w h i l e the BCC line was found,
set for higher
and the
temnerature rose from S07.1 to 9 0 8 . 9°C du r i n g that time.
Hence it is d i f f i c u l t to see h o w an A r 3 transition could
possibly have t a k e n place.
The only obvious alternative
23.
is to sdmit that the BCC -ohase persisted during the time
o b s e r v e tions
were being made on the FCC phase (11:14.3 to
11:16.6).
The BCC (110)
11:19.2.
line was followed down to 907°C. at
At 11:19.3 a search was begun for gamma iron.
No lines were found.
'i'he apparent absence of the FCC
phase from 11:19.2 to 11:56.4 is probably due merely to
a slight shift in the orientation of the crystals in the
specimen from the narrow range of positions they can have
and still send a diffracted beam through the furnace
window.
This explanation fits in well with the fact that
at 11:59.4,
after following the BCC phase from 905.0 to
909.7°C. and then-cooling to 908.3°C. a weak FCC (ill) line
was found.
It is hard to believe that both an A_r3 and an
Ac 3 transition occurred during this ranee of temperatures.
The explanation that the failure to find the FCC phase was
due to slight temporary warping of the specimen,
(or at
least of some crystals in the specimen), fits in with the
fact that often in starting a run it was necessary to
shift the temperature up and down through a range of 5 or
.6 degrees below the A c 3 point in order to find a favorable
orientation.
This leaves us once more with the picture
of the coexistence of the BCC and the FCC phases in this
temperature range.
24.
Pig. 1 e.
Part of the liun of Aug. 5-6.
C O U N T I N G - RAT E
METER
FCC
2 0 0 -t
READINGS
111
200
BCC -
IOO
BCC NO 100
110
FCC III
33
37
35
39
912
912'
-+•
910
910
908
906
906
I X
904
904
-
TEMPERATURE
FIG.
le
in ° C.
25.
Figure le.
Fig.
Part of the Run of Aug.
5-6.
le. begins st 11:29.4 with the FCC
(ill)
found at the end of the time-range of Fig. Id.
line
It shows
little change in intensity as the temperature is raised
to 910.0°C and lowered again.
and a BCC
later.
(110)
It was left at 909.7°C
picked up at 909.6°C.
only 0.5 minutes
This BCC line was found at almost exactly the same
temperature at which the FCC
half a minute later.
(111) was left, less than
It is hard to escape the picture
that the BCC phase had been present the whole time since
11:28.2 when it was last observed.
At 11: 28.7 an unintentional A c 3 transition occurred,
and the intensity of the BCC line fell precipitately.
The
drop was so steep that it could only be followed
oualitatively by means of the loud speaker in the counter
circuit.
A shift in setting of the coimter lead to the
discovery of a very
strong FCC
was followed down to 906°C
and
(111) line at 11:34.1.
it is probable that while
the setting was being changed to BCC an A r 3 transition
took place.
At any rate a BCC
904.5 degrees.
It
(110) was picked up at
26.
Fig. I f .
P a r t of the
-film of Aug.
C O U N T I N G - R AT E
5-6.
METER
20C
READINGS
200
100
BCC 110
40
42
BCC 110
44
FCG 100
FCC
46
48
912
llh.
910
910
908
908
9 06
906
9 0 4 ~~|
p
TEMPERATURE
FIG.
If
in
27.
Figure If.
Part of the Run of Aug. 5-6
This figure overlaps the proceeding one by about
i
1—1/2 minutes and shows the BCC (110) line .just mentioned.
A 53 the temperature was lowered to 905.5°C. the line grew
steadily fainter and although it was definitely a line,
a search was made for a more favorable setting.
At
11:42.7", with the temperature still about 903.5, a stronger
BCC (110) was found by rotating the specimen holder
through 5 degrees 12 minutes from the previous setting.
This line was followed to 905.9°C whereupon a shift to
the gamma, phase was attempted.
At "11:45.8 with the
temperature well below 906°C, a strong FCC (ill) was found.
As frequently happens, the intensity and temperature
curves for this line show coinciding discontinuiiiies.
Ift
ft
|
.line was followed to 907.5°C, and a 3CC (110) niched up
1.4 minutes later at 909.4°C.
1
The
Here is another case where
an A.r3 transition could not possibly have occurred; this
BCC (110) is readily explained by assuming coexistence of
I
|
I
the alpha and gamma phases.
On cooling again and shifting
to the FCC setting, an FCC (ill) v,ras found at 903.8°C just above the temperature expected for the Ar3 point at
the rate of cooling' at the time 11:51.
I
ft
28
Fig. 1 g.
Part of the ftm of Aug. 5-6.
CO UNTI N G - R A T E
METER
READI NGS
20C
200
BCC
110
100
100
I FCC llli
FCC III
52
11 h.
54
5S
12 h.
912'
912
910
910
—
908
908
906
1-
906
f
9041
TEMPERATURE
in
904
29.
Figure Ig.
Part of the Eun of' Aug. 5-6
This gra}hi begins at 11:51.9 with the FCC (111)
mentioned.
just
The temperature was raised to 906.4°C; then
the spectrometer arm was
shifted to pick up BCC (110)
907.2°C - well below the
Ac-? point» The specimen was then
cooled to 905.2;
and a shift to FCC was made.
was found a minute later
at
An FCC (ill)
at 11:57.1 at a temperature of
305.3 0C^
This line was followed up to 908.7°C, then
for a BCC
(110) -which was found at 907.2°C. 0.8 minutes
later.
left
T'ais line was followed for almost four minutes with
interesting intensity and temperature discontinuities.
During the whole of the IS minutes covered by this
graph the temperature was never close to1 either the A c 3
nor the A r 3 points, yet lines from both phases were found
repeatedly as rapidly as the counter and specimen settings
could be changed and readjusted.
mht
m
%
m
m
st
1
30.
Fig. 1 h.
Part of the Aun of Aug. 5-6.
COUNTIN G-R A T E
METER
READINGS
200
200
BCC
110
100
100
FCC III
9 I2 (
42
12 h.
J2h._j._ 9 | y
910
-MP of Ag 910
908
Ac
960.5^CZ908
906
906
904
904
TEMPERATURE
FIG.
in
31.
F i g u r e Ih.
Part of R u n of Aug.
Seven minutes
5-6.
elapsed between the end of Fig.
the b e g i n n i n g of Fig.
lh.
lg and
Du r i n g this time interval
the t e m p e r a t u r e was kept close to 907OC while a search
was m a d e f o r FCC lines.
lowed up to 9 0 9 . 0OC.
R o t a t i o n of the specimen resulted
in f i n d i n g a strong BCC
t e m p e r a t u n e had
One was found at last and f o l ­
(110)
gotten out
at 909.£°C.
If the
of control during the shift,
an A c 3 t r a n s i t i o n -would have occurred and no BCC lines
would h ave b e e n found.
coexistence
So again we have an indication of
of the two phases in this range.
Pig. 1 i.
Part of the Ran of Aug. 5-6.
COUNTI NG-R ATE
METER
200
READINGS
200
100
- BCC 211- 100
FCC
912
912
910
908
906
904
TEMPERATURE
FIG.
li
in °C.
DATA AMD CALIBRATION OF THERMOCOUPLES
At this
time it was decided to make s. silver—point
calibration. The BCC
followed up to,
(110)
line niched up
and through,
at IS:14.1
the A c 3 point.
was
Mote the kick
in the temperature curve just before the intensity of the
BCC
(110)
line starts to drop.
After this A c 3 determina­
tion the temperature was raised at a fairly r a p i d .rate
or 5 degrees
(2
per minute) until the melting point of
silver res neared.
From then on, heating proceeded slowly
(less than half a degree per minute)
slivers melted.)
After the silver point had been reached,
the furance was slowly cooled.
Figure li.
until the silver
Fig. li.
L.-nd of Bun of Aug. 5-6
shows the downward
trend of temperature as the A r 3 point was approached.
During this time the counter was set on a fairly strong
FCC (111)
line.
A-fter the A r 3 point eight minutes were
spent in searching for a BCC
found. • However a BCC
(All)
(110), but none could be
line was picked up and used
for the final A c 3 determination.
Silver-point calibrations almost always cause such
disturbance of the positions of the crystals as to ef­
fectively end the run.
Jig.
2.
Correction
Curve
for
CURVE
THERMOCOUPLE
T
0
O
DRIFT
calibration points
points used to check
curve
Drift.
TIM E
for
Thermocouple
CORRECTION
35.
TABLE I
DATA FOR CORRECTING THERMOCOUPLE TEMPERATURE
Fixed
Point
f
me
I
it
w
m
M
fcj
m
8
1
M
Time
Potent. Observed True
Correction
Reading Temp.
Temp. Diff. of obs. temp,
mv
°C
°C
°C
in per cent
Ac 3
9:41.0
37.-630
90S.7
910.5
3.3
0.42
Ac3
12:17.7
37.442
302.0
310.5
3.5
0.94
MP
of
Ag
12:42
59.362
950.3
960.5
10.2
1.07
Ac's
1:18.3
37.392
900.8
910.5
9.7
1.08
uU •
METHOD OF OBTAINING CORRECTED TEMPERATURES
The disadvantage of using unprotected thermo­
couples in an atmosphere of hydrogen is that the metals
in the couple are attacked by hydrogen, particularly at
the -temperatures involved in this experiment.
The
result is a gradual decrease in the thermoelectric e.m.f.
generated at a given temperature.
This fact causes dif­
ficulty in measuring true temperatures during a run
lasting several hours,
since the silver-point calibration
is valid for only a short time before and. after the instant
at which it was made.
During the run just described a
fairly regular downward trend in observed temperatures
was noted,
and in order to correct for this drift of the
thermocouple,
it was necessary to use the Ac3 point of
iron as a fixed point for calibration purposes.
V/angsgard’s (l) value for the A c 3 point, 910.5 plus
or minus'ole degrees
C., is the best in the literature.
The author made numerous determinations of this point, and
all of his values agree with that of Wangsgard within
experimental error.
Furthermore, it is shown botn by
V.angsgard»s data and
the author’s that tnere is apparently
no dependence of the
A c 3 point upon the rate of heating.
Conseouently it seems reasonable to regard tne Ac3 point
as a fixed point- to within the accuracy with which it is
known (0.5 degrees C)- and to use it in addition to the
silver-point as a means of correcting for thermocouple
drift. 'Obviously, Ac3 point determinations may be made
as often as desired during the course of a run, while there
can be only one silver-point calibration.
37.
method of applying this argument may be understood from
the following outline of the procedure followed in cor­
recting the data of the run of August 5-6.
An Ac3 reading
was obtained at the start of the run, another just before
the silver-point determination, and a third at the end of
the run.
These three values, together with the silver-
point value, v/ere treated as shown in Table I.
Table I.
Data for Correcting Thermocouple Temperatures
The oer cent corrections were plotted against the
times as shown in Fig. 2,
Fig. 2.
The Temperature Correction Curves for the
Run of Aug. 5-6.
and a straight line was fitted
to the points.
The correction-curve was then used to
correct each temperature reading taken in the course of the
run.
For the time of each measurement the corresponding
uer cent correction was read from the curve and added to
the observed temperature.
As a check on the curve, it
is interesting to note that values of two unintentional
A.c3 points both occurred at a corrected temperature of
910.5°C.
The closeness of the check is of course ac­
cidental, but it at least indicates that no great error
58.
was made in assuming a linear thermocouple drift with
time.
It is believed that this method permits dispensing
with the silver-point entirely.
CONCLUSIONS
The conclusions drawn from the data presented and
discussed in the proceeding section of this rarer are
summarized in Fig.
Fig. 3.
3.
Suggested Phase Diagram for Pure Iron in
the A 3 Region
In it temperatures are plotted
against the time-rate-of-change of temperature.
heavy horizontal line,
The
the A c 3 isotherm, represents the
A c 3 transition end its apparent independence of the rate
of heating.
The dotted line (drawn from rate-of-cooling
data, obtained by hangsgard(l) , Wells, Ackley, end M e h l ^ ^ ,
and by the author)
represents the A r 3 transition and its
apparent dependence on the rate of cooling.
Between these
two lines is a region, believed to be much larger than
can be attributed to superheating or supercooling, or to
temperature differences over the specimen, in which alpha
and gamma iron are found to exist simultaneously, in ap­
parent eouilibrium.
Close to the upper and lower limits of this region
the vec]uilibriumfI is most unstable,
the transition occur­
ring one wav or the other v/itn great velocity upon
slightly overstepping the temperature boundaries.
But
well within'the region the equilibrium is more stable,
Fig. 3.
Suggested Phase Diagram for Pure Iron
in the A
o
Begion.
FCC
D eg rees
d
(HEATING )
in
909
TEMPERATURE
908
907
906
905
904
BCC
903
0.5
RATE OF CHANGE
in
1.5
1.0
OF
2.0
TEMPERATURE
Degrees C. per
Minute
41.
crystals of the BCC phase growing slov/ly at the expense
of those of the FCC phase, or vice versa at rates that
can be easily controlled by variation of the rate of
heating or cooling of the specimen, but which seem to be
substantially independent of the actual temperature as
long as the temperature is held well within the limits
set by Fig.
3.
42.
REFERENCES
(1) Wangs gard, A. P.
Ph.D. Thesis,"1940
Dept, of Physics, P.S.C.
To be published in
Trans. Am. S o c . for Metals, Sept. or Dec. 1942
(p.) Wells, Ackley and Mehl, Trans. Am. Soc. for Metals,
24, 46, 1936
(3) Post, Lake and Ham, Trans. .Am. Soc. for Metals,
27, 550, 1959
(4) Mehl and Smith, Tech. Pub. No. 521, Iron and Steel
Div., A.I.M.S., 1939
(7) Neher and Harner, Phys. Rev. 49, 940, 1956
(8) Evans and Alder, RSI, 10, 552, 1959
(9) Street and Johnson, JFI, 214, 155, 1932
46~.
Appendix I
gjg.4
Diagram of the X-ray i^ube Circuit
220 v
line
LINE
' 8W
H P PRI
5W
QUl 0
'
-CKD-
transfo rmar
CONTACTOR
MA
2 2 Ov
J
Tmv&cnrv&rBfr-1
HP
SW
X RAY
wa t e r
X RAY
C ! RCU IT
TUBE
CLAMP
a
SLOW MOTION DRIVE
iig .5
3
09
Spec'ti'om etez*
GAS INLET
INSULATION
WINDOW
SPECIMEN
1^^GAS
OUTLET
Furnaoe
water
the
CLAMP a SLOW MOTION
of
COUNTERWEIGHT
ARM SUPPORTING
GM COUNTER
a PRE-AMPLIFIER
Drawing
GRADUATED CIRCL^
WINDOW
Appendix II
The Pumace-Spectrometer
Pig,5 is a full-scale drawing of the furnace.
The upper assembly consisted mainly of parts taken from a
Spencer Student Spectrometer. A central post held the spec­
imen with its face in the axis of rotation,Just grazing the
incident x-ray beam. A counterweighted arm carried the
Geiger-MtUler counter in the horizontal plane of diffraction
Both the post and the counter-arm rotated on double conical
bearings and were fitted with verniers and slow-motion drive
X-rays entered the furnace through a -|--inch hole, covered
with thin aluminum, and were diffracted through a 100 degree
slit, -^-inch wide, in the horizontal plane of the hole. The
slit was covered with a double layer of aluminum and thin
clear mica. A hole in the inner (aluminum) window permitted
visual observation of the melting of silver for calibration
of the thermocouple. The heating element was a 1000-watt
Nichrome coil wound on a refractory core, inside which was
a massive iron heat-equalizer to dampen the ripple in
temperature caused by the thermostat. The Junction of the
control thermocouple was imbedded in the alundum cement
covering the windings at the level of the center of the
specimen. Another chromel-alumel thermocouple,passed down
through the hole in the central post,was spotted welded to
the face of the sample, Just above the x-ray beam. The
insulation was shaped out of firebrick. The furnace was gastight, and fitted with inlet and outlet for circulating any
desired gas through it. Pure dry hydrogen was used.
Appendix III
j,la..6
SPECIMEN
axis
41
50~.
Speolmen H
of
HOLDERS
rotation
thermocouple
s i l v e r __
x- ra y s
speci men
TYPE
A
f r ont vrew
side view
s i.lver
bock
front
TYPE
TYPE
B
side
-§3r.
Appendix IV
The Temperature Control Circuits
Fig, 7 shows the circuits of the thermostat system,
(6 )
It is the photocell type of control used by Sam
. Briefly,
the control thermocouple forms part of a simple potentio­
meter circuit controlled "by a dial resistance "box, vVhen
the temperature of the furnace slightly exceeds a desired
value, a "beam of light reflected from the galvanometer
mirror passes over a photoelectric cell, the current from
which is amplified sufficiently to operate a sensitive
relay, which throws a mercury switch, placing extra resist­
ance in series with the heating coil. As the furnace cools,
the beam moves off the photocell, and the furnace heats
again, Tiy proper adjustment of the rheostat in series with
the heater it is possible to hold the temperature of the
specimen to within 0.2° C. in the neighborhood of 910°0,
Appendix IV
Figure 7
Temperature Control Circuits.
DRY C E L L
DIAL B O X
0,1- 9 9 9 .9 A
SW5
SW
THERMOCOUPLE
TEMPERATURE-CONTROL
C IR C U IT
SW
SW;
6D6
LQAD
T *
DC
AMINCO S.S. R E L A y
PHOTO ELECTRIC
CELL
a
RELAY
sw
SWi
lOvAG
RL.
FURNACE
FURNACE
FLOW
C IR C U IT
SW
5&*
Appendix V
The Geiger-Mtlller Counter and the Counting-Aate Meter
The counter used during most of the work was
made at the Bartol -^search Laboratories. It contained a
mixture of 94% Argon, 6% oxygen at a pressure of 9 cms. I’he
cylindrical cathode had a long narrow (3 mm wide) slit
parallel to the anode w i r e , and was mounted behind a
depressed-bubble window of thin ^onex glass, The counter
was mounted in a metal can supported by the counterweighted
arm so that it could be rotated around the furnace window.
The ffont of the can was shielded with lead, except for an
adjustable slit over the counter window. xhe can also
i.
contained the pre-amplifier circuit.
,j-,he counter was found to operate best with
a '57-tube quenching circuit of the Meher-Harper type (7).
With this circuit its plateau was flat over a range of
about 80 volts (see I'ig. 8, page 55), and it counted
reliably at rates as high as 2500 counts per minute as
long as its temperature was controlled. l‘he counter was
cooled by an air jet. x'he pre-amplifier, or quenching
circuit, had entirely separate heater and bias supplies;
its only jgonnection with the main amplifier was through
the coupling condenser.
The main amplifier included a counting­
rate meter almost identical with that developed by Evans
and Alder^ ® ^ , and the additional stages of amplification
-54t
needed to operate a midget loud-speaker and a Cenco
impulse counter, '-^he microammeter in the counting-rate
meter circuit had four ranges, one of which was calibrated
at 0-800 counts per minute, another at 0-2000 counts per
minute. I’he calibration ohanged slowly and irregularly
with change in circuit characteristics, making absolute
intensity measurements rather difficult.
700
Appendix
V
MINUTE
600
Slg.e
OPERATING
POTENTIAL
PER
4 00
s-M
i
FOR
BARTOL
WITH
100
G-M
770
COUNTER
NEHER-HARPER
QUENCHING
counter
CURVE
Plateau
COUNTS
GHARACTERI STIC
Counter
300
CIRCUIT
temp#30°C!
790
810
830
VOLTAGE
850
ACROSS
870
-COUNTER
890
910
56~.
Appendix V
Fig, 9
G—M
Counter.
Circuits
GM
R,
EZ 4.5 v
lOv AC
P R E - A M PLIFIER
_c=.v/
© r
3:
COUNTING
Z~ =L~
RATE
METER
6-3v
A U X ILIA R Y
MAGIC E Y E
a
Tftru
i IQ v AC
//
AMPLIFIER
IIOv AC
o~&a
C«
O
r
63 v
$
POWER
SUPPLY
2
SV 3
Parts List for the
Counter Circuit
-5Y.
Pre-Amplifier
10 meg.
4 meg.
%
• • • •
1 watt.
• • • •
2
22 mmfd.
"
5000 volts
Counting-iiate Meter
ft
gl.fis
XL*
i
5.%.
£1.0
'
*L2
‘
3
<
0.5 meg.
0.25 "
0.5 "
0.1 "
0.2 11
10.000 ohms w.w.
50.000 ohms w.w.
3,500 ohms w.w.
0.3 meg.
I watt
2 n
1 ir
1 w
1 it
10 n
ii
10
10 it
2 it
50,000 ohms
Centralab 72-104,with switeh.
P 2*P 3
4
C
C 2»°S
100 mfd
500 mmfd
300 mmfd
0.1 mfd
0
Cornell-Dubilier 1047
2000 volts
2000 "
400 ”
Auxiliary Amplifier
5° p"*
2 meg.
0.2 "
0.1 "
1500 ohms w.w.
10 meg.
2000 ohms w.w.
60,000 ohms w.w.
1 meg*
•<
1 watt.
2
1
10
1
10
10
1
If
It
If
If
II
If
II
Centralab 72-105
. 50,000 ohms.
;e.°io
'9
'13
*'
500 mmfd
250 mmfd
500 mmfd
... . 2000 volts
..... 2000 "
.... 2000 «
Power Supply
... 8 mfd
....
400 volts
Condensers
...
30
henries
..
th.
64C05
Choke Coils
..
700
volt
C.t.
th.
67JR93
transformers
Appendix
VI
Voltage Supply for G-M. Counter
fhe stabilized high-voltage source for
(9)
the counter was of the Street-Johnson type
; its
circuit is shown in J’ig. 10„ page 59.
FIL SW
Ov AC
'Zfmr-
K2538
57 U
866
PLATE
STABILIZED
f or
HIGH
GM
VOLTAGE
COUNTER
SUPPLY
s i
-66-.
Appendix VII
Balanced
-Filters
llhe filters used were made by the
Patterson Screen Company.
The zirconium oxide filter
(0.072 gras, of Zr per sq.cm.) was matched against
a
strontium carbonate filter (0.092 gras. Sr per sq.cm.).
The filters were mounted in a special sliding holder
which permitted either filter being inserted into the
x-ray beam, between the collimating slits and the
furnace. The effectiveness of the
Sr filter in stopping
the Mo-Kalpha radiation is shown in Pig. 11, page 61. *he
solid-line curve is the BCC (110) band passed by the Zr
filter, while the dashed-line curve shows the same band
when the Sr filter is in the beam. The data used to obtain
these curves were taken with a new specimen of Mehl iron
at a temperature of 900°C. Using a slit subtending about
15 minutes of arc at the center of the sample in front of
the counter, the counter was set at a given setting and the
total counts received in a five-minute interval recorded by
a Cenco counter. Then the filters were changed and another
run was made. This procedure was repeated twice before
moving to another setting.
If at any time during a run there was
doubt as to whether an apparent line was really K-alpha
diffraction or merely intense white radiation, it was
necessary only to replace the Zr filter by the Sr filter.
In the former case the line disappeared,in the latter it
remained with almost undiminished intensity.
-ear.
Appendix Vli
BCC
MO
1500
IRON
Balanoed Filters
L I NE
DIFFRACTED
SPECIMEN
Pig. n
BY
OF MEHL
AT 9 0 0 °C
£K
UJ
Q.
UDOO
o
with Zr
— with
Sr
Filter
Filter
500
40
20'
SETTING
OF
GM
COUNTER