THE INFLUENCE OF HEAT TREATMENT,
NEUTRON IRRADIATION AND
DEFORMATION ON THE
IN METASTABLE 3 -BRASS
JOHN WAYNE KOGER
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The author wishes to express his sincere appreciation to his
committee chairman, Dr. R. E. Hummel, who contributed much valuable
aid and time in fruitful discussions during the course of the research.
Appreciation is also extended to Dr. J. J. Hren for his help in
the interpretation of the electron micrographs and to Drs. F. N. Rhines,
R. G. Blake, and J. B. Conklin, Jr. for serving as members of the au-
Thanks are given to Mr. E. J. Jenkins for the preparation of elec-
tron microscopy samples and to Mr. V. Pashupathi for help in perform-
The author would like to thank Dr. J. Stanley of the Oak Ridge
National Laboratory for his assistance in the irradiation procedures.
The author is grateful to the Graduate School of the University
of Florida and the National Aeronautics and Space Administration for
providing his financial support.
The author is indebted to the United States Atomic Energy Com-
mission which supported this investigation under Contract No. AT-(40-1)-
Last but not least, the author would like to express his everlast-
ing gratitude to his wife, Rosemary, whose love, understanding and sac-
rifices made the completion of this work pos,
TABLE OF CONTENTS
ACKNOWLEDGMENTS. . . . ... .. . ii
LIST OF TABLES . . . . . . vi
LIST OF FIGURES. . . . . . . vii
ABSTRACT . . ... . . . . xii
I. INTRODUCTION . . . . . 1
II. SURVEY OF PREVIOUS WORK. . . . . 4
2.1 Historical Information. . . ... .. .. 4
2.2 General Features of Martensite Transformations. ... 5
2.3 The Phases Involved in the Martensitic Trans-
formation of 31-Brass . .... .. .. 6
2.4 The Order-Disorder Transformation . . 8
2.5 The Massive Transformation. . ... .. .. 8
2.6 Martensite Formed by Cooling a1-Brass . .. 10
2.7 The Characteristic Temperatures of the Martensitic
Transformation. . . ... . . 10
2.8 The Kinetics of the Transformation and Growth of
the Martensite Phase. . . . . 11
2.9 The Crystallography of the Martensite Phase Formed
by Cooling. . . . . ... 12
2.10 Deformation . . .... . . 13
2.11 Theoretical Considerations. . . ... 15
2.12 Electron Microscopy . . . ... 16
2.13 Heat Treatment of Quenched 81-Brass . ... 16
2.14 Radiation Damage . . . . .
III. EXPERIMENTAL METHODS . . . . .
3.1 Sample Preparation . . . .
3.2 Quantitative Analysis . . . .
3.3 Cryostat and Temperature Control . . .
3.4 Resistance Measurements . . . .
3.5 Metallography . . . . .
3.6 Irradiation and Safety Precautions . .
3.7 X-Ray Diffraction . . . . .
3.8 Electron Microscopy . .. . . .
IV. RESULTS . . . . . .
4.1 The Basic Transformation . . . .
4.2 Thermal Cycling . . . . .
4.3 Effect of a-Phase on the Martensitic Transformation
4.4 Plastic Deformation . . . .
4.5 Electron Microscopy Observations . . .
4.6 Transformations at Elevated Temperatures . .
4.7 Neutron Irradiation . . . .
V. DISCUSSION . . . . . .
5.1 The Temperatures of the Martensitic Transformation.
5.2 Retained Martensite . . . .
5.3 Untransformable Phases . . . .
.5.4 Plastic Deformation . . . .
5.5 Electron Microscopy . . . .
5.6 a,-Brass Martensite Structures. .. . .
5.7 Neutron Irradiation . . . .
VI. CONCLUSIONS. . . . . ... ..... 133
I. THEORIES OF MARTENSITIC TRANSFORMATIONS. . ... 135
A.1 Crystallography . . . . ... .135
A.2 Thermodynamics.. .. . . . ... 137
A.3 Kinetics and Growth . . . .. .138
II. APPLICATIONS OF RESISTANCE MEASUREMENTS FOR THE
STUDY OF TRANSFORMATIONS IN METALS . . ... 142
REFERENCES . .. . . . ..... 146
BIOGRAPHICAL SKETCH. . . . . ... . 149
LIST OF TABLES
1. Summary of neutron irradiation effects on martensitic
transformations .... . . . ... . 20
2. Resistivity of the various phases. . . ... 35
3. Visual observations on B1-brass after various degrees
of rolling at room temperature . . ... 67
4. Percent martensite present at various temperatures (1) 82
5. Percent martensite present at various temperatures (2) 82
6. Influence of fast neutron bombardment on M -temperature. 113
LIST OF FIGURES
1. Cu-Zn binary system. . . . . 7
2. Shape of the sample for resistance measurements. ... 22
3. The cryogenic unit . . . .... ... 25
4. Circuit for resistance measurements. . . ... 27
5. Cold unit for the Phillips x-ray diffractometer. ... 30
6. Dimpling unit for preparation of transmission electron
microscopy samples. . . . . .. 31
7. Relative resistivity versus temperature of 01-brass
(38.8 wt. percent zinc). . . . ... 34
8. Transformation curves. . . . . ... 37
9. Relative resistivity versus temperature of 01-brass
(38.8 wt. percent zinc) containing retained martensite 40
10. Optical micrograph of 01-brass (38.8 wt. percent zinc)
after cycling to low temperatures (room temperature,
magnification 250x). . . . . ... 41
11. Transmission electron micrographs of 01-brass martensite
(38.8 wt. percent zinc). . . . ... 43
12. Retained martensite in percent at room temperature
versus M -temperature. . . . . 44
13. Relative resistivity versus temperature of 01-brass
(38.8 wt. percent zinc). . . . ... 45
14. Relative resistivity versus temperature of B1-brass
(38.8 wt. percent zinc). . . . ... 48
15. Percent transformation versus temperature. . ... 49
16. Parts of transformation curves with various amounts of
A- and B-phases (calculated) . . ... 51
17. Parts of transformation curves with various amounts of
A- and B-phases (calculated) . . . .. 53
18. H versus percent P1-phase. . . . ... 54
19. Relative resistivity versus temperature for 81-brass
(38.8 wt. percent zinc) at five different degrees of
deformation. . . . . ... ..... 58
20. Shift of Ms(AMs ) versus deformation. . . ... 59
21. Shift of M(AM) versus deformation. . . ... 59
22. 6 and H (percent a1-phase present) versus percent
deformation. . . . . ... ..... 60
23. Optical micrographs of deformed 81-brass (38.8 wt.
percent zinc). . . . . .. ..... 62
24. Optical micrographs of deformed a1-brass (38.8 wt.
percent zinc). . . . . .. ..... 63
25. Optical micrographs of deformed 81-brass (38.8 wt.
percent zinc). . . . . .. ..... 64
26. Optical micrographs of deformed 81-brass (38.8 wt.
percent zinc). . . . . .. ..... 65
27. Optical micrographs of deformed 81-brass (38.8 wt.
percent zinc). . . . . .. ..... 66
28. Transmission electron micrographs of 81-brass (38.8
wt. percent zinc) deformed in tension. . . ... 69
29. Specific resistivity (in p2CM) versus temperature for
a1-brass (38.8 wt. percent zinc) at two different de-
grees of deformation . . . .... .70
30. Percent transformation versus temperature. . ... 71
31. Area of hysteresis loop of percent transformation
curve versus deformation . . . .. .. 72
32. Transmission electron micrographs of 81-brass (38.8 wt.
percent zinc) where martensite formed during thinning. 74
33. Transmission electron micrographs of 81-brass (38.8 wt.
percent zinc) showing striated region which makes up the
martensite formed during thinning. . . ... 75
34. Transmission electron micrograph of $8-brass (38.8 wt.
percent zinc) in cold stage showing growth of martensite
at various temperatures below M -temperature .. 77
35. Transmission electron micrographs of B1-brass (38.8 wt.
percent zinc) in cold stage showing growth of martensite
at various temperatures below M -temperature ...... 78
36. Volume percent martensite versus AT from the
M -temperature . . . .... ..... 79
37. Transmission electron micrographs of i1-brass (38.8
wt. percent zinc) in cold stage showing growth of
four needles of martensite . . . .... .80
38. Transmission electron micrographs of 81-brass (38.8 wt.
percent zinc) in cold stage showing growth of four
needles of martensite. . . . . ... 81
39. Transmission electron micrographs of a1-brass (38.8 wt.
percent zinc) in cold stage with large martensite
needle still growing . . . .... .83
40. Transmission electron micrographs of B1-brass (38.8 wt.
percent zinc) in cold stage. . . . ... 84
41. Transmission electron micrographs of Bi-brass (38.8 wt.
percent zinc) in cold stage showing the martensite
needle disappearing. . . . . .. 85
42. Transmission electron micrographs of Bl-brass (38.8 wt.
percent zinc) in cold stage showing the martensite
needle almost gone . . . .... 86
43. Transmission electron micrograph of B1-brass (38.8 wt.
percent zinc) in cold stage with martensite needle gone. 87
44. Transmission electron micrograph of 81-brass martensite
(38.8 wt. percent zinc) in cold stage. . . ... 88
45. Transmission electron micrograph of 81-brass martensite
(38.8 wt. percent zinc) in cold stage showing striations
at the center of a plate . . . .... .89
46. Transmission electron micrograph of a1-brass martensite
(38.8 wt. percent zinc) in cold stage with the boundary
between two plates . . . .... 90
47. Transmission electron micrograph of 81-brass martensite
(38.8 wt. percent zinc) in cold stage with plate left of
center much wider at the top than at the bottom. .. .... 91
48. Transmission electron micrograph of 81-brass martensite
(38.8 wt. percent zinc) in cold stage showing inner struc-
ture . . . . . . 92
49. Transmission electron micrograph of Bl-brass martensite
(38.8 wt. percent zinc) in cold stage with a boundary
containing dislocation . . . .... .93
50. Transmission electron micrograph of B1-brass martensite
(38.8 wt. percent zinc) in cold stage. . . ... 94
51. Transmission electron micrograph of a1-brass martensite
plates (38.8 wt. percent zinc) in cold stage ...... 95
52. Transmission electron micrograph of Bl-brass martensite
(38.8 wt. percent zinc) in cold stage. . . ... 96
53. Transmission electron micrograph of B1-brass martensite
(38.8 wt. percent zinc) in cold stage with dislocations
at boundaries. . . . ... ..... 97
54. Transmission electron micrograph of Bl-brass and martensite
(38.8 wt. percent zinc) in cold stage. . ... 98
55. Electron micrograph of replica of B1-brass martensite
(38.8 wt. percent zinc) showing the surface distortion .. 100
56. Transmission electron micrographs of B1-brass (38.8 wt.
percent zinc) showing slip traces with dislocations and
stacking faults. . . . . ... .. 101
57. Transmission electron micrograph of Bl-brass (38.8 wt.
percent zinc) showing tangles of dislocations. . ... 102
58. Transmission electron micrograph of B1-brass (38.8 wt.
percent zinc) showing dislocation line-up. . ... 103
59. Transmission electron micrograph of B1-brass (38.8 wt.
percent zinc) showing pilu-ups of dislocations . .. .104
60. Relative resistivity of B1-brass (38.8 wt. percent
zinc) versus five minute annealing temperatures. ... 105
61. Percent resistance change of Bl-brass (38.8 wt. percent
zinc) versus time in minutes at various temperatures 106
62. Relative resistivity of B1-brass (38.8 wt. percent zinc)
versus temperature before and after heating. . .. .108
63. Relative resistivity of Bl-brass (39.4 wt. percent zinc)
versus temperature before and after irradiation. ... .109
64. AM versus integrated neutron flux . . .. 111
65. AM versus integrated neutron flux (logarithmic scale) 112
66. Transformation curves. .. .. 115
66. Transformation curves. .............. ............ 115
67. Superposition of a percent transformation curve with
three different temperature coefficients of resistivity
(a) . . . . . . 118
68. Retained martensite at room temperature in percent
versus shift of M (AM ). . .. .. ..... .... .120
69. Optical micrograph of B1-brass (38.3 wt. percent zinc)
showing BI- (light), martensite, and a- (dark) phase 124
70. Optical micrograph of B1-brass (38.3 wt. percent zinc)
showing 8i- (light), martensite, and a- (dark) phase 125
71. Optical micrograph of B1-brass (38.3 wt. percent zinc)
showing B1- (light), martensite, and a- (dark) phase 126
Abstract of Dissertation Presented to the Graduate Council
in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy
THE INFLUENCE OF HEAT TREATMENT, NEUTRON IRRADIATION,
AND DEFORMATION ON THE MARTENSITIC TRANSFORMATIONS
IN METASTABLE 81-BRASS
John Wayne Koger
Chairman: Dr. R. E. Hummel
Major Department: Metallurgical and Materials Engineering
A study was made on the martensitic transformations in meta-
stable 81-Cu-Zn (with 38.3 to 39.4 wt. percent zinc) to determine
the effect of repeated thermal cycling, deformation, annealing, bom-
bardment with fast neutrons, and the presence of untransformable
phases on these transformations. The main tool used to monitor the
transformations was the measurement of electrical resistance which
was complemented by optical and electron microscopy and x-ray dif-
The principal characteristics of the transformation on cooling,
determined from resistance measurements, were found to be the M -tem-
perature, the temperature coefficient of electrical resistivity of
the phases, the area of the hysteresis loop of the (percent) trans-
formation curve, and the rate of transformation.
The M -temperature was found to change according to the treat-
ment given the sample. Elastic deformation and the presence of re-
tained martensite would raise the M The M -temperature was lowered
after plastic deformation and neutron irradiation but was not affected
by heat treatments up to 1350C.
Retained martensite formed during the first thermal cycle was
found to be present at room temperature in samples which had an M -
temperature above -330C. The amount of retained martensite increased
with increasing M -temperature. The shift of M between two cycles
is a linear function of retained martensite. The retained martensite
was observed to grow above the previously found M -temperature.
Untransformable phases, such as a and deformation-martensite,
acted as obstacles to the formation of low-temperature martensite.
As the amount of these phases increased, the area of the hysteresis
loop of the transformation curve increased; the maximum rate of trans-
formation decreased; and the temperature range of transformation was
spread over a wider range, thus increasing the amount of energy needed
to complete the transformation. In samples which were irradiated with
fast neutrons no such increase in the area of the hysteresis loop was
detected, showing that the irradiation induced defects are only small
Using cold-stage electron microscopy it was observed that mar-
tensite formed immediately after a heavy bending movement of the foils.
Differences in the growth mechanism during cooling and heating were
The fact that many alloys and metals undergo a martensitic trans-
formation has been of great interest for a long time. This interest
has stemmed mainly from the shear-like nature of the transformation
and the sometimes enhanced properties of the transformed phase, such
as the increased hardness of the steel martensites.
A relatively new and effective means of influencing a state of
matter that could be used for studying the nature of phase transfor-
mations is neutron irradiation. It is usually of interest as a method
of introducing various kinds of defects and producing certain changes
in the properties of metals and alloys. However, the irradiation may
alter the state of an initial phase and cause some change in its sta-
bility and thus affect the kinetics of subsequent phase transforma-
tions. The strictly regular directed shift of atoms in a martensitic
transformation is extremely sensitive to any disturbance of the lattice
structure in the original parent phase.
Investigations of the effect of irradiation on the martensitic
transformation have been done on alloyed steels which were complex
in both behavior and composition, and the results were not adequate
to give a complete picture of the processes which cause changes in
the transformation. Therefore, it was felt that further irradiation
work was needed on another martensite system which is nonferrous and
would have less variables. The Cu-Zn system was chosen.
Quenched Bl-phase Cu-Zn alloys with compositions from 37 to 43 wt.
percent zinc undergo a martensitic transformation when the samples are
cooled to sub-zero temperatures. The occurrence of the transformation
at relatively low temperatures would insure that imperfections introduced
by room-temperature irradiation would not anneal out. Cu-Zn alloys have
a relatively low residual radioactivity after neutron bombardment. (The
only isotope with a long half life is the Zn65 with T/2 = 245 days.)
The previous studies of the martensitic transformation of B1-brass
have concerned the composition dependence of the martensitic-start tem-
perature, the crystal structure of the martensite phase, the observation
of the transformation by optical microscopy, and some effects of deforma-
tion on the transformation.
The purpose of this research was to determine the influence of
heat treatment, neutron irradiation, and deformation on the martensitic
transformations of B1-brass and to investigate whether neutron irradia-
tion causes the same effects as cold-working 81-brass, or produces ef-
fects similar to those found for steels and iron alloys. The trans-
formation was monitored by high precision electrical resistance measure-
ments. Optical and electron microscopy were also used to supplement
the resistivity data.
In the course of the investigation, it was found that very little
was known about the characteristic properties of the transformation and
the parameters which were landmarks of the transformation. Studies there-
fore were made on the effect of repeated thermal cycling. It was neces-
sary to know how the characteristics of the transformation would be af-
fected if the samples were transformed several times so that any effect
due to irradiation would be separated from the cycling effect. The effect
of an added non-transformable phase on the transformation was studied.
Due to the mode of quenching, a-phase could always be present along
with the B1-transformable phase. This study was also helpful in the
interpretation of results gained by deformation. Since it had been
noticed in experiments concerning radiation damage that deformation
caused the same effects as neutron irradiation, the influence of de-
formation on the martensitic transformation was investigated. These
preliminary studies provided a better basis for the interpretation
of property changes which occurred due to neutron bombardment of al-
loys with martensitic transformations.
The results of this investigation will provide an understanding
of the effect of many controlled variables on the martensitic trans-
formation in $1-brass, will allow an understanding of the basic trans-
formation, and will provide knowledge of the effect of neutron ir-
radiation on martensitic transformations.
SURVEY OF PREVIOUS WORK
This chapter will provide a summary of the types of studies made
on 81-brass and its martensitic transformations. The previous investi-
gations concerning the effect of irradiation on martensitic transfor-
mations are also given. Appendix I gives more details about martensitic
transformations and discusses ferrous alloy martensitic transformations
where similar studies were not made on 81-brass.
2.1 Historical Information
Martensite, originally considered just a hard constituent in steel,
was named in honor of A. Martens, a German metallurgist, as proposed by
Osmond in 1895.
In 1946 Troiano and Greninger (1) suggested that the diffusion-
less reactions found in the non-ferrous alloys are similar to the mar-
tensite reaction in steel and that these transformations should be
referred to as martensitic.
Pure metals and alloys found to have a martensitic transformation
Na Cu-Si Fe Re-Mg Li-Mg U-Cr
Li Cu-Zn-Ga Fe-Ni Ti Cr-Mn NH NO3
Zr Ag-Cd Fe-Mn In-Ti Ar BaTiO3
Cu-Zn Ag-Sn Fe-C In-Ta Ar-N2 KNbO3
Cu-Sn Ag-Sb Fe-Ni-Mn In-Cd Ir-Cd V3Si
Cu-Al Au-Zn Fe-Ni-C Ti-Mh Co Au-Cu-Zn
Cu-Mn Au-Cd Fe-Cr-Ni Ti-Ni Co-Ni Cu-Ni-Al
Cu-Ga Au-Mn Fe-Ti TI-Mo Al-bronze
and other ternaries involving the above systems.
The previous list is possibly not complete but does contain most of
the systems where information has been published. Almost all of the
references for the studies of these systems can be found in the sur-
veys given by Christian (2) and Barrett and Massalski (3).
2.2 General Features of Martensite Transformations
Phase transformations in metallic systems are usually considered
to be either nucleation-and-growth transformations or martensitic
transformations. The nucleation-and-growth processes are considered
to be isothermal transformations in that the transformation can take
place at constant temperature with thermal activation and diffusion
playing the important parts.
In general, the characteristics of the martensitic transformations
have been that it is diffusionless (interchange of atoms over a dis-
tance less than the interatomic spacing); it is athermal (transforma-
tion occurring with changing temperature and transformation stops when
temperature change is halted); the transforming region undergoes a
change in shape; there is a definite orientation relationship between
the product and parent phase; and the transformation is aided by cold-
work. Because of exceptions to many of the above characteristics,
"martensitic" now usually refers to diffusionless transformations
having macroscopic shear or a shape distortion.
The following items are also observed in martensitic transforma-
tions. The interchange of atoms takes place through the propagation
of a semi-coherent interface. It is assumed that this interface must
be composed of dislocations whose movement permits the martensitic
product to form rapidly at any temperature. The martensite plates
in many alloys contain a fine structure of slip bands or twins as a
direct result of this transformation mechanism. In alloys which are
ordered, the martensite phase has been found to be ordered. The a-
mount of transformation is characteristic of temperature, provided
other variables such as grain size are held constant. There is no
dependence on cooling rate for the transformation. Martensitic re-
actions are reversible in the sense that an initial atomic configura-
tion can be repeatedly obtained. Thermodynamically, the transforma-
tion is not reversible as seen by the hysteresis loop and the energy
dissipation. It is assumed that the plates which form on cooling
have the same size and shape and appear in the same regions of the
2.3 The Phases Involved in the Martensitic Transformation of 61-Brass
Figure 1 shows a portion of the Cu-Zn phase diagram. The symbols
used to designate the various phases concerned with the martensitic
transformation in 1B-brass have been given by several authors.
B Disordered body centered cubic B-brass
BI Ordered body centered cubic B-brass
8" Martensite formed by cooling (4,5)
a Face centered cubic a-brass
ia Martensite formed on deformation (6)
a Massively transformed brass (3)
The above were those phases involved in this investigation. The fol-
lowing phases are those mentioned in other studies:
a' Quenched martensite-face centered tetragonal (4,5)
a" Quenched martensite with unknown structure (4,5)
a2 Transition phase between B and al (6)
B' Bainite structure (4,5)
WEIGHT PER CENT ZIC0
30 40 80
ATOMIC PER CENT ZINO
Figure 1. Cu-Zn binary system.
Bl Transition lattice between 8' and a'(7)
B"'- Theoretical martensite phase beyond 0"(8)
2.4 The Order-Disorder Transformation
Above 4540C to 4680C, the 8-phase in Cu-Zn is disordered body
centered cubic. Below these temperatures it is ordered body centered
cubic and has the cesium chloride structure. As is well known, the
ordering is found to occur so rapidly that even rapid quenching cannot
suppress the ordering.
The order-disorder transformation is indicated in phase diagrams
by a single line running from 4540C to 4680C in the 8-phase region.
Work has indicated that perhaps 0 and 81 fields are separated by a
normal two-phase field (8+a1) (9).
2.5 The Massive Transformation
In order to obtain the transformable (i.e., metastable) 81-phase,
the alloy must be quenched from the 8-region. The composition of the
alloy before quenching is very important since not all 01-alloys trans-
form to cooling martensite and because of the massive transformation.
Phillips (10) first reported that Cu-Zn, containing between 37.0
and 39.0 wt. percent zinc, when quenched from the 8 region would give
a uniform a structure. As there has been a question on the location
of the boundary between the a and a+a region, Phillips proposed that
the single phase a region lies below 39.0 wt. percent zinc and that
the above mentioned boundary is a straight line at the 39.0 wt. per-
cent zinc composition. This transformation from B to a was seen to
occur with great speed during cooling.
The structures obtained when low zinc content B-brass was quenched
in different manners has been discussed by Schimmel (11). The alloy
was 37.8 wt. percent zinc, rest copper, which is B-phase from 850C to
8950C and a below 550C. On quenching from 8500C, there was a consid-
erable quantity of finely precipitated a disseminated through the 8
crystals with a little massive a along the grain boundaries. On quench-
ing from 890C there were large 8 crystals with a feather-like a along
the grain boundaries. With a more drastic quench, there was a more
general precipitate of a particles with actually more a present than
before. With an extremely rapid quench, the structure became all a.
The explanation of these structures was that with cooling at a moderate
rate, the transformation from 8 to a must take place by diffusion, and
the time will not be long enough for this. If the 8-phase is cooled
rapidly below 550C, no diffusion is required for the transformation,
and it required scarcely any time with only the structure change of
The above-mentioned transformation in 81-brass was recognized as
a massive transformation by Hull and Garwood (12).
A study of alloys that exhibit the massive transformation by
Massalski (13) reaffirmed that alloys in the low zinc range of 81-
brass transformed into a face-centered cubic phase on quenching.
This structure obtained by the massive transformation was regarded
as a supersaturated extension of the equilibrium a phase. For a
quenched alloy of 37.95 wt. percent zinc, a was found to be 3.700 A.
Recent work on massive transformations (3) showed that the grow-
ing crystals of the massive phase cross the prior grain boundaries of
the parent phase. While the final product of the massive transfor-
mation is usually the equilibrium phase, corresponding to a lower
temperature, the actual reaction may well occur at much higher
temperatures from which slow cooling would produce a two-phase mixture.
The massive transformations require certain critical conditions of
supersaturation and quenching rates. Probably the kinetics of the mas-
sive transformation would be intermediate between that of an equilibrium
reaction and that of a possible martensitic transformation.
2.6 Martensite Formed by Cooling B1-Brass
Martensite is formed when alloys of Bl-brass with 37 to 43 wt. per-
cent zinc are cooled to sub-zero temperatures.
The first mention of this type of transformation in B-brass was
made by Kaminski and Kurdjumov (14) in 1936 but was not recognized as
a martensitic transformation until later. In 1938 Greninger and Moora-
dian (15) also noted this transformation on cooling but did not observe
a critical temperature for the transformation. Bassi and Str6m (4),
Massalski and Barrett (16), and Pops and Massalski (17) found that al-
loys of Bl-brass with more than 43 wt. percent zinc do not transform
to martensite on cooling.
2.7 The Characteristic Temperatures of the Martensitic Transformation
The temperature at which martensitic nuclei start growing (18) has
been designated as Ms. In some iron-nickel alloys a large amount of
martensite will form at one time in a burst. When a burst-type mar-
tensite occurs after the Ms-temperature, a temperature MB is designated
as the martensite burst temperature (19). MF is the temperature at
which the transformation is complete. As is the temperature at which
the parent phase starts re-forming on heating, and A is the tempera-
ture at which the parent phase has completely re-formed.
Titchener and Bever (20) studied the temperature range of the
martensitic transformation as a function of composition in 81-brass.
The compositions of the samples used were from 38.54 to 40.04 wt. per-
cent zinc. It was shown that the M -temperature decreased with decreas-
ing copper content. M -temperature of the samples ranged from about
-200C to -1300C. Electrical resistivity measurements were used,and M
was defined to be that temperature at which the resistivity-temperature
curve on cooling deviated from a straight line. Because of a lack of
accuracy, the temperature of minimum resistivity was taken for M The
plot of M -temperature versus zinc content had a somewhat different
slope than the slope of a similar curve given by Kurdjumov (5).
Pops and Massalski (17) proposed that there were two stages in the
transformation. A thermo-elastic phase (appeared and disappeared as
cooled and reheated) formed initially at the M -temperature. On further
cooling, an additional martensitic phase formed in rapid bursts at the
lower temperature MB. The M -temperature was considered to be the tem-
perature at which the first evidence of martensite was optically seen.
This temperature could not be observed from their resistivity data.
The M -temperature, as obtained from the optical microscopy, was said
to be equivalent to the minimum temperature of the resistivity-tem-
perature curve. Pops and Massalski also concluded that the M -temper-
atures found by Titchener and Bever were actually MB-temperatures.
2.8 The Kinetics of the Transformation and Growth of the Martensite
The martensitic transformation of $1-brass is athermal; that is,
the transformation only proceeds with a lowering of the temperature.
The time at a certain temperature or during cooling has no effect on
the fraction transformed.
Hull and Garwood (12) and Pops and Massalski (17) observed the
growth of the martensite in detail using cold-stage optical microscopy.
First seen were thin "needles" which grew in the lengthwise direction
and then thickened slowly. The plates did not reach their final size
at once but did definitely grow when the temperature was lowered.
2.9 The Crystallography of the Martensite Phase Formed by Cooling
Kaminski and Kurdjumov (14) and Greninger and Mooradian (15) de-
termined in early work that the structure of the martensite was near
that of a face centered tetragonal lattice.
Using B1-brass containing approximately 39.0 wt. percent zinc,
Garwood and Hull (21) employed a Laue x-ray method to determine the
lattice orientation relationship between the martensite and the parent
phase. The values obtained were seen to be the same as exist in the
martensite transformation in the Cu-Al system. When the measured
shear was applied to a theoretical unit cell of the 8 lattice, a body-
centered triclinic lattice was produced, which did not agree with ex-
perimental results. Garwood and Hull assumed that the lattice of the
martensite phase would be close-packed because of the identification
of a prominent basal plane in each martensite plate. They felt that
in view of the diffusionless nature of the transformation at sub-zero
temperatures that the ordered structure which is not suppressed by
quenching would be retained in the martensite product. In other alloy
systems there are faults in the stacking sequence of the martensite
phase after the transformation (22). These faults may also be present
in this system. Distortions of these types are mainly responsible for
the complexity of the x-ray spectra reported for the low-temperature
Jolley and Hull (23) determined from x-ray and electron diffrac-
tion data that the martensite obtained on cooling a 81-brass sample
containing approximately 39.0 wt. percent zinc had an orthorhombic
crystal structure. Using these and other data, they applied the
Wechsler, Lieberman, and Read theory. The calculated habit plane
compared favorably to the experimentally determined habit plane.
In a very detailed crystallographic study with x-ray measurements
at low temperatures, Kunze (7) showed a monoclinic (also called body-
centered pseudorhombic) transition lattice (8') between parent and mar-
tensitic phase (W"). The superlattice cell of the martensite phase B"
was said to be triclinic face-centered on one face (pseudomonoclinic).
About 25 percent of this transition lattice was always present with
the martensite at liquid nitrogen temperature.
Greninger and Mooradian (15) noticed "markings" in deformed B1-
brass containing 39.22 wt. percent zinc and determined the crystal-
lographic directions to which these markings were parallel. X-ray
work on hammered powder samples showed some lines corresponding to
a face-centered tetragonal structure.
Reynolds and Bever (24) elastically compressed a B1-brass sample
containing approximately 39.98 wt. percent zinc and therefore induced
plates of deformation martensite. After the stress was released, the
plates disappeared. When samples were stressed and unstressed in suc-
cessive cycles, the plates appeared in the same locations. They also
cooled a specimen which had residual strain-induced martensite plates.
As thh specimen was cooled below room temperature, some of the plates
increased in size, and others were formed. The thermally induced and
strain-induced plates had a similar metallographic appearance. It was
concluded that strain-induced martensite was mechanically reversible
with considerably hysteresis and that elastic stresses were operative
in this reversal. Suoninen, Genevray, and Bever (25) investigated the
effect of elastic stress on the transformation. It was shown that the
M -temperature increased with increasing stress. Also, an alloy con-
training approximately 39.5 wt. percent zinc was partially transformed
by cooling it under stress. The stress was released, and the marten-
site disappeared. After further cooling, the martensite started form-
ing as usual. This was considered to be consistent with the concept
of "thermo-elastic" martensite first considered by Kurdjumov (5) and
confirmed by Kurdjumov and Khandras (26).
Massalski and Barrett (16) studied the effect of cold work on the
transformation in B1-brass alloys with compositions from 39.73 to 51.80
wt. percent zinc. Alloys which would not transform on cooling; i.e.,
above 43 wt. percent zinc, were seen to transform on cold work. All
the alloys did transform on cold work even though in some cases low
temperatures were required. A temperature, MD, has been defined by
McReynolds (27), as was previously done in an Fe-Ni alloy, as being
the temperature above which no martensite will be formed on deforma-
tion. The M- temperatures were determined for some of the high zinc
content alloys, and a plot was made of MD-temperature versus zinc con-
tent. The structure of the deformation martensite was found to be
face-centered cubic with stacking faults for the lower percent zinc
alloys and hexagonal close packed for the higher percent zinc alloys.
The face-centered cubic structure had parameters that could be extra-
polated from the a-phase region assuming line shifts from the stacking
faults. For alloys with compositions above 45.59 wt. percent zinc,
the transformation product reverted to the parent phase after a few
weeks. Thus, B1-brass with 37.0 to 42.0 wt. percent zinc is metastable
with regard to cooling and deformation and B1-brass with above 42.0
wt. percent zinc is metastable with regard to deformation. Massalski
and Barrett assumed that in the latter case only the disordered 8-
phase was metastable. Therefore, the cold work disordered the alloy
and the low temperatures caused the transformation.
Hornbogen, SegmUller, and Wassermann (6) studied the transforma-
tion during deformation using elastic and plastic tensile stresses.
It was seen with the use of x-ray measurements that after 10-15 per-
cent deformation the Bl-brass (composition between 39.5 and 39.8 wt.
percent zinc) would transform to a tetragonal phase a,. This phase
was described as an ordered a-brass of the Cu-Au type. A transition
phase a2 was seen between the a1- and al-phases. At higher degrees of
deformation the al-phase would transform into a disordered supersat-
Investigations from the Southern Research Institute (28) also
found martensite in Muntz metal (59.98 wt. percent copper and 38.71
wt. percent zinc) which had been rolled to 53 percent reduction at
240C and to 28 percent reduction at -1960C. The studies were made
using optical photomicrographs and were performed to improve the
2.11 Theoretical Considerations
Kunze (8) considered the theory of elasticity and the contribu-
tion of the Fermi energy to the transformation. The Fermi energy,
which'controls the stability of the 81-lattice at room temperature,
was also said to carry the principal weight in controlling the steps
of the low temperature transformations. This kas shown from the
calculations of the Brillouin zones of the low temperature phases.
It was predicted from these electron theoretical considerations that
the martensite may possibly undergo a further transformation at low
temperatures. This transformation from 6" ~ "'would be a close
packing of (110) atomic planes. He assumed that the first step
(shearing), 81 + 1', of the low temperature transformation corresponded
to the first step in the martensitic transformation by plastic deforma-
tion (1 -* ac) and quenching (,1 a') and that the unknown structure
a2 between Bl and al is identical to that of the transition lattice
8'. The martensitic transformation by deformation proceeds by other
mechanisms than the low temperature transformation after the first
2.12 Electron Microscopy
Hull (29) observed in thin foils of S1-brass, containing approxi-
mately 39.0 wt. percent zinc, martensite which apparently occurred
during the preparation of the foils in thin regions at the edges of
the specimens. Orientation relationships were determined, and it was
seen that the interface between the martensitic phase and parent phase
did not follow any particular habit plane. This has been seen in other
alloy systems which undergo a martensitic transformation.
2.13 Heat Treatment of Quenched B1-Brass
In 1924 Homerberg and Shaw (30) determined strength characteristics
of 81-brass at various temperatures during reheating after quenching.
Later, Hansen (31) determined that heating the quenched 81-brass above
1500C caused a rise in resistivity of about 10 percent. Changes in
resistivity and hardness were correlated with composition, temperature,
and time changes.
Garwood (32) studied the isothermal decomposition of 81-brass
samples containing 41.3 wt. percent zinc at temperatures from 170C
to 4700C. A bainitic transformation was assumed to occur at temper-
atures between 1700C and 2250C. The structure of this phase was not
determined. Above 225C thea-and 8-phases (evidenced by x-ray dif-
fraction) occurred as expected.
Bassi and Strom (4) heat-treated a 81-brass sample of 40.5 wt.
percent zinc for 73 hours at 150C and also detected a new phase but
could not determine its structure.
Hornbogen (33) aged quenched 81-brass containing about 40 wt.
percent zinc between 2000C ad 3000C. Precipitation occurred which was
connected with an increase in hardness. The final transformation to
a-brass was accompanied by a decrease in hardness. Two tetragonal
intermediate phases were found before the transformation was complete.
Hornbogen and Warlimont (34), using the electron microscope,
studied the mechanism of the isothermal transformation of quenched
8-brass and proposed a generalized definition of the bainite trans-
formation. The bainite transformation is between the martensitic
and nucleation-and-growth transformations. Lattice defects in great
density are created which allow a segregation of the atoms, and they
act as nucleation sites. The transformation, itself, occurs like
martensite by a shear process.
2.14 Radiation Damage
Little work has been done on the effect of radiation on the mat-
tensitic transformation. Zakharov and Maksimova (35) studied the ef-
fect of neutron irradiation on the martensitic transformation of
hi-carbon alloy steels and Fe-Ni-Mn alloys. The samples were irra-
diated at ambient reactor temperature (4 = 5 x 1016 and 1017n/cm2)
and the course of the transformation was followed with magnetic meas-
urements. In the steels which contained carbon, the irradiation
raised the transformation start temperature 150C. This was considered
an activating effect. (Porter and Dienes assumed that this was due to
precipitation of austenite.) The "activating effect" was decreased
after larger irradiation times and was removed after aging the speci-
mens at temperatures below 1000C. Also, more martensite was formed
at a given temperature than before. In the Fe-Ni-Mn alloys, which
contained hardly any carbon, the reverse effect was obtained. The
stability of the austenite was increased, the martensitic start tem-
perature was lowered, and the amount of transformation was decreased.
It was concluded from these results that during irradiation there is
a simultaneous development of structural changes which affect the
austenite stability in opposite directions. Therefore, the obser-
vation of opposing effects depends on the total neutron flux and the
properties of the alloy. In the early stages of damage, transforma-
tion is favored. Later, there are structural changes which have the
Porter and Dienes (36) investigated the effect of neutron irra-
diation on the martensitic transformation in Fe-Ni alloys. They
found that the M was lowered for doses above 2.5 x 1017 nvt at 1000C.
They observed no change in the kinetics. This effect was thought to
be associated with an increase in the critical shear stress of the
matrix material. For samples which were partially transformed to
martensite, neutron irradiation lowered the M point and partially
recovered the plastic deformation produced by the prior transformation.
This strain recovery was thought to be due to the annealing of shallowly
trapped damage by enhanced diffusion and by the mutual annihilation of
Weiss-Hollerwager (37) studied the effect of neutron irradiation
on the martensitic transformation of a chrome steel which contained
1.26 percent carbon. For doses of 1017 and 5 x 1017n/cm2, 5 percent
more martensite than before was found at the end of the transformation.
During irradiation with doses of 1018n/cm2, 25 percent martensite was
found. Cooling created only about 10 percent more martensite.
A summary of the past work that concerns irradiation effects on
the martensitic transformation is given in Table 1.
TABLE 1. Summary of neutron irradiation effects on martensitic transformations.
Ms IRRADIATION EFFECTS
AUTHOR and YEAR MATERIAL
UNIRRADIATED (INT. FLUX IN NEUTRJCM2)
AILZAKHAROV 8 0.48% C 0=5.1016: Ms+50C ------ -60
O.P MAKSIMOVA 7.7 % Mn IO1C 10% more end-mortensite s.o16 '
(1957) 2.2 % Cu u -20t
Balance Fe 0
-22q (TOO -100 0
B.WEISS- I. 26% C 0=10. and 5.10: ..
HOLLERWOGER 0.36% Mn 5-- 5C Ms=+10*C and ca.+ 35 C unirr.
(1960) 0.42%Si 5% more end- mortensite iO8
5.40 %Cr = 101:22% martensite at "
Balance Fe room temp. and 20% less .10
end-martensite T*C -150 -too -50 0
A.I. ZAKHAROV 8 0.02 % C ~1017: Ms=-60*C 20
O.P.MAKSIMOVA 22.4%Ni -39C 3 % less end-martensite o 7-1 10
(1957) 3.48 %Mn
Balance Fe (T C)-100 00
L.F. PORTER 8 26.37% Ni +2.60C ,* 1.6.1018 Ms=-12"C 7t 10
G.J. DIENES 0.018 %C (Variation between T
'*e*o'-7 IR e
(1959) 0.197% Mn +11.2 and +2.6*C) I
Balance Fe TC -32-20 -10 0
3.1 Sample Preparation
81-brass samples containing between 38.3 and 39.4 wt percent zinc
were prepared. These compositions were selected in order that the
temperatures of the transformation would be in a range practical for
the experimental apparatus.
The purity of the samples was very important. Unstable impurities
which might precipitate or dissolve during the various temperature
changes could affect the characteristics of the martensitic transforma-
tion, and some impurities when irradiated could give unwanted products
with long radiation half-lives.
The alloys were made from 99.999 percent copper rods obtained from
A. D. Mackay, Inc. and 99.9999 percent zinc pellets supplied by United
Mineral and Chemical Corporation. The copper and zinc were melted in
an evacuated vycor tube using an induction furnace. The melted alloy
was shaken manually and quenched in water to eliminate segregation.
The outside layer was removed on a lathe after which the alloys were
worked, re-incapsulated,and homogenized for 300 hours at 800C. Then
the alloys were rolled into long strips of about O.lmm thickness. Al-
loys'containing the higher percentages of zinc had to have an inter-
mediate heat-treat before being rolled to the final size. The samples
were cut into the shape shown in Figure 2 and heat-treated in a quartz
_______________________________________: --- -i
Shape of the sample for resistance measurements.
tube. The tube was first repeatedly evacuated and flushed with argon
with the samples at the cold end of the tube. With the argon atmos-
phere present, the samples were moved into the hot end of the tube,
heated at 8700C for five minutes, and quenched into previously boiled
water of 20C. The quench yielded the metastable ordered Bl-phase.
Scraps of brass around the samples were used to prevent de-zincing.
About 15 samples in a pack were heat treated together. The com-
position of some samples was such that the "massive transformation"
to the supersaturated a-phase occurred. Because the samples had to
fall about two feet before being quenched, some were cooled slightly.
This lowered the quenching temperature of these samples into the a+B
region where some a precipitated out. Because of these procedures,
various amounts of a-phase were present along with the 81-phase.
3.2 Quantitative Analysis
The amount of a and B1-phase present in the samples was obtained
after the final quench using x-ray diffraction, resistivity measure-
ments, and optical microscopy. In the x-ray method, the amounts were
determined by comparing the integrated intensities of two of the main
peaks of each of the phases. This method was non-destructive and very
convenient for this work. Since the electrical resistivity of a two-
phase material is a linear function of the volume fractions of the
two phases present, the amount of each phase could also be determined
by knowing the resistivity of the sample and the individual resistivi-
ties of the two phases. The third method of determining the amounts
of the phases present was by quantitative metallography. All three
methods yielded, within the error limit, the same ratio of a and
B1-phase. Chemical analysis was provided by V. Horrigan of the Ana-
conda American Brass Company. The zinc content of the samples varied
less than 0.06 percent.
3.3 Cryostat and Temperature Control
Two units could be used for the temperature bath. In the tem-
perature range between +10C and -160C, which covers almost every
M -temperature in these experiments, isopentane was employed as the
cooling medium. Isopentane could not be used much above +10C be-
cause of its high vapor pressure and below -1600C because of high
viscosity. The isopentane, which was held in a one-gallon metal Dewar
flask, was cooled by liquid nitrogen pumped under pressure (provided
by an argon tank and regulator) through a 15-foot copper tubing coil
immersed in the flask. The flow was regulated by a Linde temperature
controller using two thermistors in the cooling bath. The tempera-
ture was found to be constant within 20C. The bath was continuously
circulated by means of a magnetic stirrer. An electric immersion heat-
ing coil was built into the Dewar flask for use when heating was needed.
For alloys with an M -temperature near 00C, the samples were ini-
tially placed in a water bath whose temperature could be maintained
between +20C and +980C. As cooling was desired, the water was allowed
to circulate through an ice bath. The amount of cooling was regulated
by a heating device which worked at the temperature specified by a
The temperature of the samples was taken during each resistivity
measurement with a copper-constantan thermocouple and a Leeds and
Northrup potentiometer with temperature calibration.
A schematic of the cryogenic unit is given in Figure 3.
Controller --- -
Figure 3. The cryogenic unit.
3.4 Resistance Measurements
The resistance measurements were obtained using a Leeds and
Northrup K-3 potentiometer to measure the voltage drop across the
potential leads of the sample. A one-amp direct current was made
to flow through the sample. The current was regulated by a vari-
able resistor of 12 ohms and monitored across the potential leads
of a 0.001 ohm standard resistor by an auxilliary input of the K-3
potentiometer. Three 12-volt batteries in parallel were used as a
current source, thus affording a more nearly constant current than
only one battery. The resistance measurements were found to be re-
producible at a given temperature within 0.03 percent. The circuit
diagram is shown in Figure 4.
Three sample holders of micarta were used so up to three samples
could be measured at one time. Tests were made on the reproducibility
of the resistance of samples removed from the sample holders, turned
over, and replaced in the sample holders. The change in resistance
was again within 0.03 percent.
A Starrett micrometer was used to measure the dimensions of
the samples in order to obtain resistivity data from resistance meas-
For metallography work, the specimens were polished by standard
techniques through Linde B and diamond paste. The specimens were
etched for a few seconds in a solution of 25 mL of ammonium hydroxide,
35 ml.of water, and 1 ml. of hydrogen peroxide. a-phase, B1-phase,
and martensite could be differentiated as they all gave different
Figure 4. Circuit for resistance measurements.
appearances after etching. A Bausch and Lomb metallograph was used
for the optical microscopy.
For a quick examination of samples to determine if the Bl-phase
was present, a macroetch of equal parts HNO3 and water was used. The
81-phase appeared as small crystallites which were very distinguish-
able from the larger unetched grains of a-phase.
3.6 Irradiation and Safety Precautions
The samples were irradiated in the Oak Ridge National Laboratory's
Bulk Shielding Reactor with integrated fluxes between 5 x 1016 and
5 x 1018n/cm2, E > 1 mev. The samples were under helium atmosphere
with a pressure of one atmosphere and the temperature was 510C.
The irradiated samples, when not being used, were kept in an
enclosed lead container with a wall thickness of 1 inch. With all the
irradiated samples in this lead container, a Geiger counter at a dis-
tance of 12 inches showed a radiation level not beyond the cosmic
radiation level. A wall of 2-inch lead brick and a bottom plate of
1-inch steelwere placed around the cryostat during measurements. A
Geiger counter showed no increased radiation level outside this wall.
3.7 X-Ray Diffraction
A Norelco diffractometer with a high-intensity copper tube was
used for the x-ray work. Modifications were made to the radiation
shield so the samples for these experiments could be handled without
bending and so cold-stage work could be done.
Dry nitrogen from a tank forced liquid nitrogen into the elongated
radiation shield, whose window was closed from the atmosphere by mylar.
After cooling started, any areas which could let air in became frosted
and were sealed. Thus, no icing took place on the sample as a nitro-
gen atmosphere was formed. Figure 5 shows this unit.
3.8 Electron Microscopy
The same material which was used in the resistance studies served
as stock material for the transmission electron microscopy.
The first sample preparation step was the "dimpling" of small
disks. The dimpling technique consisted of allowing a thin stream
(jet) of liquid electrolyte (50 percent orthophosphoric acid) to im-
pinge on the disk and produce a concave surface. Initially, a disk
would be thinned until a hole appeared. Then a smaller amount of elec-
trolyte would be used to produce the dimple in the other samples. The
jet producer and container for the electrolyte were made of glass and
contained a platinum wire which constituted the cathode. The speci-
men was the anode (Figure 6).
After dimpling, the sample was electro-polished until the very
first small hole appeared. For this, the sample was held with plat-
inum-tipped tweezers between two point electrodes. There was a po-
tential difference between the tweezers and the electrodes. The elec-
trolyte was again 50 percent orthophosphoric acid which was cooled
below room temperature with acetone and dry ice to slow down the re-
action. During this final polish the sample was periodically removed,
washed with alcohol, and examined for a small hole with the aid of a
light placed behind the specimen. The area around the first hole was
usually thin enough for transmission work. If the sample was polished
very long after the first hole appeared, all the thin area was de-
stroyed. The yield of useable samples was very small. If a satis-
factory sample was obtained, care had to be taken to wash the sample
well with alcohol to prevent contamination.
Figure 5. Cold unit for the Phillips x-ray diffractometer.
Dimpling unit for preparation of
transmission electron microscopy samples.
A cold stage built for the Phillips 200 Electron Microscope by
Ladd Research was used in this work. The unit consisted of a sample
holder, dewar, and thermocouple. After insertion of the sample in the
microscope, liquid nitrogen was added to the dewar and the cooling
would begin. The temperature was monitored by a built-in copper-con-
stantan thermocouple. The speed of cooling could be adjusted by the
amount of liquid notrogen added and by a built-in heating unit.
4.1 The Basic Transformation
The main features of the martensite transformation curve of BI-
brass, as monitored by electrical resistance measurements at changing
temperatures, are discussed in order to determine the characteristics
of the transformation.
Figure 7 shows relative resistance or relative resistivity (set
equal to 1 at 0C) vs. temperature. (This will be referred to as the
resistivity-temperature curve in the future.) The relative resistance
decreases with constant slope as the temperature is lowered. At a
certain temperature, the curve deviates from a straight line. This is
the starting point of the transformation (Ms). Since the slope changes
very gradually at this inflection point, the M -temperature can only
be determined to within an accuracy of 1fC. At a lower temperature,
the resistivity-temperature curve shows a resistivity minimum which
is designated as M. At the temperature MF, the transformation is
finished, after which the resistivity-temperature curve again becomes
a straight line. On heating the sample, the curve again deviates from
a straight line. This is the A -temperature, lower than the M -tem-
perature, where the parent phase starts forming. The transformation
is completed at a temperature AF after which the resistivity-temper-
ature curve is a straight line.
S -10 -20 -30 -40 -50 -60 -70 -80 -90 T (0)
I I I I I i i I i I I I
Figure 7. Relative resistivity versus temperature of B1-brass (38.8 wt.
To find the amount of martensite formed at a given temperature
and the degree of hysteresis after cooling and heating, a plot of
percent transformation vs. temperature was made from the resistance
data (Figure 8a). This curve was constructed by taking the ratio
of the difference in resistivity between the straight line or extra-
polated straight line portion of the curves for the parent phase and
the martensite phase at a given temperature to the difference in
resistivity between the curve and the straight line or extrapolated
straight line of the curve for the parent phase at the same tempera-
percent martensite = (100) (see Appendix II).
The rate of transformation-temperature curve, which is equivalent to
the amount of martensite formed per degree, was constructed from the
derivative of the data of the percent transformation-temperature curve.
The martensite phase formed by cooling had a higher specific
resistivity (p) and a greater temperature coefficient (6) than the a1-
phase. Table 2 gives the resistivities and temperature coefficients
of all the phases studied in this work.
TABLE 2. Resistivity of the various phases.
Phase p at 0OC (p-ohm-cm) e(1/deg)
81 5.3 .0128
8" (cooling martensite) 6.9 .0179
a 6.8 .0116
al (deformation martensite) 7.4 .0133
Figure 8. Transformation curves.
a. Percent transformation versus temperature.
b. Percent transformation per degree versus
temperature. (Calculated from Figure 7.)
Li 60 -
-20 -30 -40 -50 -60
-70 -80 --90
The reason for using the resistance measurements to monitor the
martensitic transformation of Bl-brass was that the resistivity of
the martensite in 81-brass is much higher than the resistivity of
the parent phase and resistance can accurately and easily be measured.
Thus, very small amounts of transformation could be detected. (More
details are given in Appendix II.)
All factors such as long-range order, short-range order, anti-
phase boundaries, and lattice defects which cause changes in resis-
tivity were not changed during cooling.
Local stresses set up by the formation of martensite plates might
be expected to increase the resistance of both the martensite and the
surrounding parent phase and thus influence the resistivity-temperature
curve. The amount of the increase in resistance should not be greater
than would be caused by extensive cold work. This increase is usually
not more than one percent for most alloys. Experiments were made that
showed that the samples of 81-brass had to be bent enough to create
a permanent crease in the metal to cause a 0.5 percent change in re-
sistance. It is assumed that the transformation strains would never
be this great. Therefore, the stresses accompanying the transforma-
tion introduce no serious error.
4.2 Thermal Cycling
The M -temperature and other characteristics of the martensitic
transformation were found not to be reproducible when samples of a
certain composition were repeatedly cooled to the temperature of com-
plete martensite formation and heated to room temperature.
These effects were attributed to retained'martensite which was
substantiated from the following observations:
1. After one full cycle, the resistivity curve obtained from
heating the sample (later referred to as heating curve) does
not meet the resistivity curve obtained from cooling the
sample (later referred to as cooling curve) in the linear
region around 0OC; i.e., the specific resistivity (p) of the
sample is higher after cycling (Figure 9). Since the martensite
phase has a higher specific resistivity, the increase of p at
0C is thought to mean that some martensite is retained at
2. The slope of the linear portion of the resistivity vs. tem-
perature curve (the temperature coefficient 8) at temperatures
below MF is larger than e at temperatures above Ms. This is
indicated in Figure 9, where a line parallel to the linear
portion at temperatures below MF is drawn on the cooling curve
near OOC. This line is labeled martensitee." Figure 9 also
shows that 8 from the heating curve at temperatures above AF
is larger than 0 from the cooling curve in the same temperature
range. Since a larger 6 implies the presence of martensite, it
is deduced that the observed larger 6 at temperatures above A ,
obtained after cooling and heating, implies retained martensite.
3. Metallographic evidence of martensite plates was found at
room temperature after a full cycle and was interpreted to
mean that some martensite was retained (Figure 10). The needles
could not be detected in all sections examined and this was
attributed to the fact that the amount of retained martensite
is small and to the possibility that the martensite plates
which were retained have a size below that of the resolution
of the optical microscope.
-20 -30 -40 -50 -60
I I 1 I 1 I I I I I
-80 -90 -100
I I I I I
Figure 9. Relative resistivity versus temperature of B1-brass (38.8
wt. percent zinc) containing retained martensite.
Figure 10. Optical micrograph of B$-brass (38.8 wt. percent zinc)
after cycling to low temperatures (room temperature,
magnification 250 x) .
4. Retained martensite was also observed in transmission-electron
micrographs made from a sample which had been cooled below M"
and heated back to room temperature in a cold stage (Figure 11).
No attempt was made to compare the amounts of retained mar-
tensite observed by optical and electron microscopy because of
the difference in behavior of thin films and bulk samples.
5. A third cycle reproduced the second cycle almost perfectly.
This means that after the second cycle no additional retained
martensite is left at temperatures above AF.
The amount of retained martensite at 0OC was computed by taking
the ratio of the difference between the resistivity curves before and
after cooling to the difference between the parent phase resistivity
curve and the extrapolated straight line portion of the martensite
phase resistivity curve. The amount of retained martensite calculated
by this method was found to be dependent upon the Ms-temperature. At
M -temperatures lower than about -330C no martensite is retained at O0C.
With increasing M -temperature, the amount of retained martensite in-
creases. Figure 12 shows that up to 5 percent of the total amount of
martensite can be retained at 0OC.
When the same sample is cooled a second time, Ms is shifted to a
higher temperature (Figure 13). No change is noted in Ms on the third
In summary, some martensite, formed when B1-brass is cooled to
sub-zero temperatures, is retained at room temperature. This marten-
site is stable and only occurs in samples with high Ms-temperatures.
After the first cycle the samples which contained martensite had in-
creased M -temperatures.
Transmission electron micrographs of 81-brass martensite
(38.8 wt. percent zinc). Magnification 14000x.
a. During cooling.
-o1 -20 -30 -40 M8 [C]
Figure 12. Retained martensite in percent at room temperature
versus M -temperature.
Relative resistivity versus temperature of 81-brass
(38.8 wt. percent zinc).
4.3 Effect of a-Phase on the Martensitic Transformation
In the course of this investigation it was found that the resis-
tivity-temperature curves of a group of samples which had the same
composition and heat treatment were not the same (Figure 14). Optical
microscopy, x-ray diffraction, and resistivity measurements showed the
presence of varying quantities of a-phase, which was first observed
by Phillips (10).
From the quantitative analysis it was seen that the sample of
Figure 14a was about 100 percent B1-phase, and the sample of Figure 14b
contained about 54 percent 01-phase. Five main differences of these
curves were observed.
1. The difference between M and M was greater in Figure 14b than
in Figure 14a.
2. There is a considerable variation in M between the curves in
Figure 14a and Figure 14b.
3. At OC the resistivity difference between the parent phase and
the martensite phase (pM-p ) is about twice as large in Figure
14a as in Figure 14b (pM at O0C was obtained by extrapolation).
pM-P 1 is called H and may also be taken as the amount of 81-
4. The degree of hysteresis (defined as the temperature difference
between the means of maximum and minimum resistance of cooling
and heating (15)) is about 50 percent larger in the curve of
Figure 14b than in the curve of Figure 14a.
5. The slope of the straight-line portion of the curve above the
M -temperature is larger in Figure 14a than in Figure 14b.
Percent transformation-temperature curves (Figure 15) and rate of
transformation-temperature curves were constructed as described earlier.
Figure 14. Relative resistivity versus temperature of 81-brass
(38.8 wt. percent zinc).
a. 100 percent Bl-phase in sample.
b. 54 percent 81-phase in sample.
I I I I I I I
-10 -20 -30 -40 -50 -60 -70 -80 -90
--- TEMPERATURE (C)
100% 0, PHASE
0 -20 -40 -60 -80-100-120-140
C-- TEMPERATURE (C)
54% 0, PHASE
Percent transformation versus temperature.
a. 100 percent B1-phase.
b. 54 percent Sl-phase.
Percent transformation per degree versus t
c. 100 percent a1-phase.
d. 54 percent 81-phase.
(Calculated from Figure 14)
The area of the hysteresis loop increased, the maximum rate of trans-
formation per degree decreased, and the temperature range of transforma-
tion (Ms-M_) increased as the amount of a-phase in the samples increased.
Theoretical resistivity-temperature curves were made to determine
the effect of the presence of the a-phase.
The resistivity-temperature curve of a-brass was found to be a
straight line with a slope less than that of the linear part of the
01-brass transformation curve at temperatures above M These measure-
ments were made to assure that no transformation of any type occurred
in the a-brass at temperatures in the range used in these experiments.
Using the actual resistivity-temperature curves of a sample of
100 percent a-phase and a sample of 100 percent $1-phase, curves with
different amounts of 81 and a were constructed. The proportion of the
resistivity of each of the two phases was taken at a given temperature
(using the assumed amounts of each phase), added together, and plotted
The following features were noted from the theoretical curves:
1. The M -temperature was not influenced by the mathematical
addition of a-phase to 81-phase.
2. Additions of a-phase to BS-phase increased the temperature
difference between M and M.
3. With increased additions of a-phase the resistivity difference
H at O0C decreased.
4. At a critical amount of a-phase (around 15 percent) no minimum
resistivity was observed.
5. The slope of the straight line portion of the curves above the
M -temperature approached the slope of the curve of the a-phase
as the amount of a-phase in the samples was increased.
as the amount of a-phase in the samples was increased.
25 % B
S 10 -00 -30 -40 -50 -60 -70
I I I l I I I I I I I I T ( C)
Parts of transformation curves with various amounts of A- and
B-phases (calculated). 6 of A smaller than 0 of B.
No changes are noted in the percent transformation-temperature
curves or rate of transformation-temperature curves since all the
theoretical curves were mathematically constructed from the 100 per-
cent B1-phase transformation curve.
To describe, in general, the behavior of an added non-transform-
able phase, another curve was constructed assuming that the non-trans-
formable phase had a greater slope than the linear portion of the 100
percent B1-brass transformation curve above the M -temperature (Figure
17). The same features as listed above were noted.
The resistivity difference (H), taken from the theoretical curves,
plotted versus percent B1-phase present (Figure 18) was used to de-
termine the amount of B1-phase present from experimental resistivity-
Any changes seen in both the theoretical and experimental resis-
tivity curves are geometrical effects due to the added a-phase. The
increased difference in M and M with increasing amounts of a-phase
and the very similar shapes of the curves including the changes in H
and the slope are examples.
The M -temperature of samples originally the same composition which
contain equal amounts of a-phase after heat treatment may not be the
same because of one or all of the ways the a-phase can be formed.
1. If a sample has a composition in the range where the massive
transformation occurs, the quenched samples will contain dif-
ferent amounts of a-phase due to the quenching speed. The
M -temperature of these samples will probably be a function
of the amount of a-phase present.
2. When several samples are quenched together, some may lose more
zinc than others. The loss of zinc will change the composition
Parts of transformation curves with various amounts of
A- and -B-phases (calculated). 6 of A larger than 6
L B 6o
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Figure 18. H versus percent 81-phase.
and thus change the M -temperature.
3. The a-phase may also be formed if the sample is cooled into
the a+B region and quenched. Samples treated in this manner
will contain different compositions of the 81-phase and dif-
ferent amounts of the a-phase, both of which will change the
Thus, with one or a combination of the three occurrences, the true
effect of the a-phase on the M s-temperature is "masked."
Therefore, the changes due to the actual physical presence of
the a-phase are seen in the experimental percent transformation-tem-
perature curves of samples containing different amounts of a-phase.
The area of the hysteresis loop increased, the maximum rate of trans-
formation per degree decreased,and the temperature range of transforma-
tioned widened as the amount of a-phase in the samples increased.
4.4 Plastic Deformation
Bl-brass samples were deformed to learn more about transformation
caused by deformation and how plastic deformation influences the char-
acteristic points when the sample undergoes the martensitic transforma-
tion on cooling. These experiments were also done in view of the fact
that very often neutron irradiation and plastic deformation cause simi-
lar effects to metals.
After a pre-deformation run, i.e., a measurement of resistance from
room temperature to a temperature of complete transformation and back to
room temperature, the samples were deformed a small amount and run again.
The samples would then be deformed many times, with runs made in between
each deformation. The deformation was by rolling and was measured by
the percent reduction in thickness. From the resistance-temperature
curves the changes in the characteristics of the curves could be noted.
Figure 19 presents five different resistivity-temperature curves
corresponding to the transformation from the 81 to B" phase after dif-
ferent degrees of reduction in thickness on the same sample. The fol-
lowing observations were made from these curves.
1. With increasing degree of deformation the M -temperatures are
shifted to lower temperatures. This shift amounts to some 350,
compared to the undeformed state. This is also indicated in
Figure 20 where AM (the shift of M compared to the undeformed
state) is plotted versus the degree of deformation. AM in-
creases rapidly to about 16 percent reduction. For higher
deformations the further increase of AM was about negligible.
2. With increasing degree of deformation the temperature of mini-
mum resistivity (M) is also shifted to lower temperatures.
This shift is about 60*C, compared to the undeformed state.
In Figure 21, AM is plotted as a function of deformation. It
is seen that AM increases rapidly until about 16 percent de-
formation then it levels off.
3. The temperature coefficient of resistivity, e, i.e., the slope
of the straight-line portions of the resistivity-temperature
curves at temperatures above Ms,is seen to be a function of
deformation as it decreases with increasing amount of deforma-
tion. Three different slopes, corresponding to three degrees
of deformation are drawn in Figure 19 on the curve of 15.3 per-
cent deformation. In Figure 22 e is plotted as a function of
deformation. A rapid change in e is observed up to deformations
Figure 19. Relative resistivity versus temperature for $1-brass
(38.8 wt. percent zinc) at five different degrees of
0 -20 -40 ,-60 -80 -100 -120
-- TEMPERATURE (C0)
Shift of M (AM ) versus deformation.
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
Figure 21. Shift of M(AM) versus deformation.
2 4 6 8 10 12 14 16 18 20 22 24 26 28
I 1 a I I a a a
0 2 4 6 8
10 12 14 16
18 20 22 24 26
Figure 22. 6 and H (percent B1-phase present) versus percent deforma-
. I I I I a a I a
of 12 or 13 percent. With higher degrees of deformation 6 is
4. A further characteristic property of the transformation curve
is the resistivity difference between the cooling curve and
the extrapolated linear portion of the heating curve at 00C.
This quantity was denoted previously as H and was taken as a
measure of the amount of the 01-phase present (Figure 18). H
(the percent B1-phase present after deformation) is seen to
decrease with increasing amounts of deformation (Figure 22).
At small deformations, H decreases rapidly up to 12 to 16
percent deformation after which it decreases only gradually.
5. With increasing degree of deformation, A is shifted to lower
temperatures. AA plotted versus deformation would be similar
to the AM vs. deformation curve. The A -temperature, after
deformation, is shifted 330C, compared to the undeformed state.
The five above mentioned properties all point to a common behavior.
At the beginning of deformation they change rapidly, but after a criti-
cal degree of deformation (12 to 16 percent) they almost become constant.
It had been mentioned by other investigators (24) that when de-
formed, B1-brass would transform to martensite. In this research, this
martensite is called deformation-martensite. The proof of the presence
of deformation-martensite in this work is given by optical micrographs,
electron microscopy, resistivity measurements, and x-ray diffraction.
A series of a1-brass samples with 38.8 weight percent zinc which
were plastically deformed by cold-rolling and investigated with optical
microscopy are shown in Figures 23-27 and the results are listed in
Table 3. At deformations of 3 percent and larger, the martensite needles can
Figure 23. Optical micrographs of deformed 81-brass (Q8.8 wt. percent
zinc). Magnification 560x.
a. 1 percent deformation.
b. 2 percent deformation (needles are deformation-martensite
Figure 24. Optical micrographs of deformed 81-brass (8.8 wt. percent
zinc). Magnification 560x.
a. 3 percent deformation.
b. 6.5 percent deformation.
Optical micrographs of deformed S1-brass (38.8 wt. percent
zinc). Magnification 560x.
a. 13.4 percent deformation.
b. 17.0 percent deformation.
Optical micrographs of deformed B1-brass (38.8 wt. percent
zinc). Magnification 560x.
a. 23.2 percent deformation.
b. 35.5 percent deformation.
Optical micrographs of deformed 81-brass (48.8 wt. percent
zinc). Magnification 560x. 43.8 percent reformation.
TABLE 3. Visual observations on 81-brass after various
degrees of rolling at room temperature.
Degree of deformation in percent Observations
(reduction in thickness)
1 Only 81-phase and a little a-
2 Only 81-phase and a little a-
phase. Needles shown are those
of deformation-martensite formed
when the sample was cut.
3 Isolated needles of martensite.
6.5 Martensite needles seen in almost
all areas, some untransformed
left. Grain boundaries are still
13.4 Thick needles in all areas.
17 Deformed thick needles.
35.5 } No individual needles are seen
be seen. For deformations above 13.4 percent, the martensite needles
begin to be deformed.
Figure 28 shows electron micrographs of B -brass that had been
pulled in tension and then prepared for the electron microscope. The
plates seen are those of deformation-martensite.
In Figure 29 the absolute resistivity-temperature curves of the
transformation of a sample, undeformed and 13 percent deformed, are
given. The resistivity change after deformation is shown. This sample
contained 21 percent a-phase (calculated as mentioned earlier) and 79
percent transformable 1B-phase. The 13 percent deformation transformed
a portion of the 81-phase to deformation-martensite so that after de-
formation, the sample contained 41 percent deformation-martensite, 38
percent $1-phase, and 21 percent a-phase. It can be seen also from
this plot that the temperature coefficient of resistivity of the de-
formation martensite is lower than that of both the cooling martensite
and the parent $1-phase.
The percent transformation versus temperature curves were constructed
for the deformed samples as described in section 4.1 (Figure 30). These
curves, drawn for cooling and heating, form a hysteresis loop. The first
derivative of the above data for the cooling curve was taken to obtain
the rate of transformation per degree and was plotted versus the temper-
The area of the hysteresis loop was seen to increase as the amount
of deformation increased (Figure 31). Also, the maximum rate of trans-
formation per degree on cooling decreased and the temp rature range of
transformation was spread as the amount of deformation increased.
The investigations described in this section showed that five char-
acteristic properties of the transformation curve changed rapidly up to
Figure 28. Transmission electron micrographs of 81-brass (38.8 wt.
percent zinc) deformed in tension. Magnification 8200<.
Specific resistivity (in pQCM) versus temperature for s1-brass (38.8
wt. percent zinc) at two different degrees of deformation.
20 0 -20 -40 -60 -80 -100 -120
20 0 -20 -40 -60 -86 -100 -120 -140
Percent transformation versus temperature.
0 percent deformation.
15.3 percent deformation.
Percent transformation per degree versus temperature.
0 percent deformation.
15.3 percent deformation.
(Calculated from Figure 19)
> 0 -20 -40 -60 -80 -100 -120 -140
_j __ i
< 10 -
< 5 -
O 2 4 6 8 10 12 14
/o Plastic Deformation
Figure 31. Area of hysteresis loop of percent transformation curve versus
about 15 percent deformation and remained constant at higher deforma-
tions. The hysteresis loop was widened, the maximum speed of transforma-
tion decreased, and the temperature range of transformation was spread
over a greater range when the samples were deformed. These property
changes were due to the formation of deformation-martensite which was
seen in optical and electron micrographs.
4.5 Electron Microscopy Observations
Although some of the observations of the martensitic transforma-
tion which were made by using the electron microscope have been mentioned
earlier, it is believed that a separate section would be beneficial.
It was seen in these investigations and in studies by Hull (29)
that after preparation of some B1-brass samples for transmission elec-
tron microscopy, martensite needles appeared near the edge of the
foils at the very thin regions (Figures 32 and 33). This martensite
was interpreted to be deformation-martensite as the structure was face-
centered cubic (as determined from electron diffraction) which is the
same as that of highly deformed Bl-brass. This phenomenon did not
occur in each sample but in enough to make their appearance signifi-
cant. This martensite (deformation-martensite) did not grow when the
foil was cooled and transformed in the electron microscope. The most
distinctive features of the martensite are the striations within the
In evaluating the absolute temperatures obtained with a thermo-
couple, it must be noted that the electron beam will heat up a portion
of the sample and it is not really known if the temperature recorded
Transmission electron micrographs of $1-brass (38.8 wt.
percent zinc) where martensite formed during thinning.
a. Magnification 12600x.
b. Magnification 17000x.
Figure 33. Transmission electron micrographs of B1-brass (38.8 wt..
percent zinc) showing striated region which makes, up the
martensite formed during thinning.
a. Magnification 25200x.
b. Magnification 58900x.
is that of the part of the foil being examined. However, the tempera-
ture differences should be accurate.
As the prepared foils were cooled in the electron microscope and
just before any of the cooling martensite was seen, there was a great
movement of the contours of the foil, somewhat of an undulating motion
believed to be due to a bending of the foil. Immediately after this
the first martensite was seen. If the cooling was fast, the martensite
plates "exploded" into view and grew until stopped by a barrier. With
slow controlled cooling, the plates were seen as thin lines growing
first in length then in thickness when further cooled.
Figure 34a shows the first evidence of a martensite plate at sub-
zero temperatures. A thin line was first seen. (This foil was very
dirty because of problems with the vacuum system of the electron micro-
scope.) The black dot is an impurity used as a landmark. As the tem-
perature was lowered, the amount of martensite increased. Figures 34
and 35 show the plates at various stages of the transformation. A
second needle and the thickening and joining of the two needles is
easily seen. The percent of martensite seen in the same section of
each picture is measured by the area of the martensite compared to the
area of the section and is givenin Table 4 and Figure 36.
Figures 37 and 38 show the growth of martensite in another area.
Table 5 gives the temperature difference and amount of martensite seen
in a common area of each of the pictures. These pictures were made at
the temperatures of the greatest martensite growth. The inner striated
structure can be seen in these needles. The inner structure, in most
cases, was seen to form after the needle was formed.
Transmission electron micrographs of 81-brass
percent zinc) in cold stage showing growth of
at various temperatures below M -temperature.
a. M -temperature AT = 0.
b. AT = 50C
c. AT = 70C
d. AT = 80C
2.9 vol. percent martensite.
9.6 vol. percent martensite.
16.7 vol. percent martensite.
Transmission electron micrographs of 81-brass
percent zinc) in cold stage showing growth of
at various temperatures below M -temperature.
tion 8200x. s
a. AT = 10C
b. AT = 120C
c. AT = 150C
d. AT = 180C
39.5 vol. percent martensJte.
64.0 vol. percent martensite.
Volume percent martensite versus AT from
the M -temperature.
Figure 37. Transmission electron micrographs of B1-brass (38.8 wt.
percent zinc) in cold stage showing growth of four needles
of martensite. Magnification 8200x.
a. AT = 0 37.5 vol. percent martensite.
b. AT = 3"C 59.4 vol. percent martensite.
Figure 38. Transmission electron micrographs of $1-brass
percent zinc) in cold stage showing growthlof
of martensite. Magnification 8200x.
a. AT = 5C
b. AT = 60C
77.1 vol. percent martensite
85.0 vol. percent martensite.
TABLE 4. Percent martensite present at various temperatures (1).
Figure AT(OC) Vol. percent martensite
34 5 2.9
35 10 39.5
TABLE 5. Percent martensite present at various temperatures (2).
Figure AT(OC) Vol. percent martensite
37 0 37.5
38 5 77.1
When the foils were heated from temperatures of complete marten-
site formation, the martensite disappeared. The striated bands across
the needles first disappeared, starting at one end of the needle and
going to the other. There was no change in the width (Figures 39-43).
Thus, there were differences in the way the needles grew and disap-
Figures 44-54 show foils at various temperatures during the mar-
tensite formation. These pictures and the ones above of 81-brass mar-
tensite formed in the electron microscope using a cold stage were
probably the first of their kind. Many different areas of the foils
are shown. The inner structure of the needles and their boundaries
are particularly interesting.
Figure 39. Transmission electron micrographs of 81-brass (38.8 wt.
percent zinc) in cold stage with large martensite needle
still growing. Magnification 8200x.
b. T 60C.
percent zinc) in cold
micrographs of B1-br ss (38.8 wt.
stage. Heating the ample. Mag-
a. T + 460C.
b. T + 580C.
Figure 41. Transmission
electron micrographs of 81-b ass (38.8 wt.
in cold stage showing the ma tensite needle
a. T + 47C.
b. T + 810C.
Transmission electron micrographs of a1-bi
percent zinc) in cold stage showing the mu
almost gone. Magnification 8200x.
a. T + 890C.
b. T + 1020C.
ass (38.8 wt.
Transmission electron micrograph of 81-brass (38.8 wt.
percent zinc) in cold stage with martensite needle gone.
Magnification 8200x. T + 108C.