WIN!“

1

  

“IWINWIUUUHHWIHWIH

\

mm m

401—;
Ito-I

    

1%? 5307-

 

iiiiim w //f/7,[1777ilif iii/717i mfim
Universitv

 

 

This is to certify that the
thesis entitled

Microstructural Changes in

Boron Doped Ni A1 During

3

High Temperature Low Cycle Fatigue
presented by

Yogesh C. Mishra

has been accepted towards fulfillment
of the requirements for

Master's of Science degree in Materials Science

 

 

)1-7

fl , w
(Dr. ottstein)
_/ \7

Major professor

 

Date JUly 28, 1989

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

PLACE ll RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or More data duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU I: An Affirmative ActiorVEquol Opportunity Institution

MICROSTRUCTURAL CHANGES IN BORON DOPED Ni 3A1
DURING HIGH TEMPERATURE LOW CYCLE FATIGUE

By

Yogesh Chandra Mishra

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Metallurgy, Mechanics, and Materials Science

1989

ABSTRACT

MICROSTRUCTURAL CHANGES IN BORON DOPED Ni 3A1
DURING HIGH TEMPERATURE LOW CYCLE FATIGUE

Yogesh Chandra Mishra

Microstructural changes in boron doped Ni 3A1 during high temperature
strain controlled (0.5% and 1% strain amplitudes) and stress controlled (constant
stress of 85 MPa) low cycle fatigue were studied. From stress-strain plots, it was
found that the true stress (as a function of cumulative strain) showed a max-
imum, followed by a continuous decrease at large cumulative strain. Extensive
grain boundary migration was observed in all but one cyclic tests. Grain boundary
migration could be the major softening mechanism. Grain boundary migration

was inhibited if the material had dispersion in the form of voids that pin down

the grain boundaries.

ACKNOWLEDGEMENTS

I am indebted to my academic adviser, Professor Gunter Gottstein, for his
constant encouragement and guidance, and invaluable suggestions throughout this
study. I would also like to thank, C. S. Kim, C. S. Lee, Moti Tayal, S. R. Chen,
and W. J. Kim for all their valuable help.

Special thanks to the Department of Metallurgy, Mechanics, and Materials
Science for providing the necessary facilities for this study, and the financial sup-.
port in the form of teaching assistantship. Thanks to the Case Center for Com-

puter Aided Design for providing wonderful computing facility.

Finally, the support of U.S. Department of Energy, Office of Basic Science,
under grant no. DE-FGOZ-SSER45205 for the study is gratefully acknowledged.

ii

Table of Contents

LIST OF FIGURES ........................................................................................ v
1. INTRODUCTION .................................................................................... 1
2. LITERATURE SURVEY ......................................................................... 3
2.1 The Material .................................................... 3

2.1.1 Mechanical Properties .................................................... 3
2.1.2 Microalloying ................................................................. 5 ,

2.1.3 Effiact of Boron on Grain Boundary Cohesion ................ 5

2.1.4 Effict of Alloy Stoichiometry ......................................... 6

2.1.5 Efbct of Testing Environment on the Properties ........... 6

2.2 Grain Boundary Migration During Low Cycle Fatigue ................ 11

3. EXPERIMENTAL PROCEDURE ........................................................... 16
3.1 Testing Material .......................................................................... 16

3.2 Specimen Preparation .................................................................. 17

3.2.1 Mechanical Test Specimens ............................................ 17

3.2.2 Optical Microscopy Specimens ....................................... 19

3.2.3 Scanning Electron Microscopy Specimens ...................... 22

3.3 Mechanical Testing ...................................................................... 22

3.3.1 Instron Testing Machine ................................................ 22

iii

6.

>1

iv

3.3.2 MTS Testing Machine ...................................................
EXPERIMENTAL RESULTS ...................................................................
4.1 Mechanical Behavior ....................................................................
4.2 Microstructural Development ......................................................
4.3 Brittle Failure ..............................................................................
DISCUSSION ...........................................................................................
5.1 Mechanical Behavior ....................................................................
5.2 Grain Boundary Migration ..........................................................
5.3 Brittle Failure ..............................................................................
CONCLUSIONS .......................................................................................

BIBLIOGRAPHY .....................................................................................

24

31

31

36

36

47

47

47

48'

49

51

LIST OF FIGURES

Figure Captions

Fig. 1. Yield stress as a function of test temperature for Ni 3Al-base
aluminide alloys, Hastelloy, and type 316 stainless steel.

Fig. 2. Effect of Stoichiometry on fracture mode of B-doped Ni 3A1
at room temperature.

Fig. 3. Plastic strain to fracture (%) of rapidly solidified Ni 3Al-B

as a function of Al and B concentration. The solid lines
indicate constant NiAl ratio (at. %).

Fig. 4. Efbct of Stoichiometry on grain boundary segregation (in
terms of peak-height ratio) and room temperature tensile
ductility of B-doped Ni 3A1 containing 24 to 25.2 at. % Al.

Fig. 5. Plot of tensile elongation of Ni 3A1 alloys as a function
of (AH-Hf) concentration for bare specimens tested at 600°C
in vacuum or for specimens preoxidized at 1100°C/ 2h 4-

850°C/5h and tested in air. Solid symbols for alloys with
0.5% Hf and open symbols for the alloy without Hf.

Fig. 6. SEM micrograph showing cyclic grain boundary migration after

testing for 15 cycles in reverse bending at a frequency of

0.17 Hz and a strain amplitude of 10.28%. A preexisting

Page

10

Fig.

Fig.

Fig.
Fig.
Fig.
Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

7.

8.

vi

Figure Captions

twin is at A, and B marks the initial position of the boun-

dary prior to testing.
Two-beam interferometer (monochromic light) showing repetitive

cyclic markings and the presence of sliding after testing for
3 cycles in reverse bending at a frequency of 0.10 Hz and a
strain amplitude of :t0.48%.

Optical micrograph, using Nomarski interference contrast,

showing the presence of a fine structure within the migra-
tion markings after testing for 8 cycles reverse bending at a

frequency of 0.17 Hz and a strain amplitude of :l:0.28%.

9. Fatigue Sample (B-doped Ni 3A1).

10.
11.
12.
l3.

14.

15.

16.

17.

Microstructure of Ni 3A1 annealed at 800°C for 5 hour.
Microstructure of Ni 3A1 annealed at 800°C for 5 hour.
Instron testing machine set up.

Computerized high temperature high vacuum servo-
hydraulic test system.

High temperature grip assembly.

Specimen inserts for the high temperature high vacuum
MTS test system.

Final set up inside the vacuum furnace of the

high temperature high vacuum MTS test system.

Plot of true stress versus cumulative strain for 0.5% strain

Page

12

13

14 A
18
20
21

23

25

26

28

29

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig

18.

19.

21.

23.

vii

Figure Captions

amplitude cyclic deformation in high vacuum at 600°C.
Plot of true stress versus cumulative strain for 1% strain

amplitude cyclic deformation in reducing atmosphere of

(90% N2 + 10% H2) at 600°C.
Plot of true stress versus number of cycles for 1% strain

amplitude cyclic deformation in reducing atmosphere of

(90% N2 + 10% H2) at 600°C.

. Plot of true stress versus cumulative strain for constant

stress controlled test, using constant stress of 85 MPa, in

reducing atmosphere of (90% N2 +10% H2) at 600°C.
Microstructure of Ni 3A1, in the plane perpendicular to

the loading axis, annealed at 800°C for 5 hour, after 100

cycles of 0.5% strain amplitude at 600°C.

. Microstructure of Ni 3A1, in the plane perpendicular to

the loading axis, annealed at 800°C for 5 hour, after 3900
cycles of 1% strain amplitude at 600°C.

Microstructure of Ni 3A1, in the plane perpendicular to

the loading axis, annealed at 800°C for 5 hour.

. Microstructure of Ni 3A1, in the plane perpendicular to

the loading axis, annealed at 800°C for 5 hour, after 790

cycles of constant stress of 85 MPa at 600°C.

. 25. Microstructure of Ni 3A1, in the plane perpendicular to

Page

32

33

34

35

37

38

39

40

viii

Figure Captions

Page

the loading axis, annealed at 1200°c for 2 hour. 41
Fig. 26. Microstructure of Ni 3A1, in the plane perpendicular to

the loading axis, annealed at 1200°c for 2 hour, after 1000

cycles of 0.5% strain amplitude at 600°C. 42
Fig. 27. Microstructure of Ni 3A1, in the plane parallel to

the loading axis, annealed at 800°C for 5 hour, after 100

cycles of 0.5% strain amplitude at 600°C. 43.
Fig. 28. Microstructure of Ni 3A1, in the plane parallel to

the loading axis, annealed at 1200°C for 2 hour, after 1000

cycles of 0.5% strain amplitude at 600°C. 44
Fig. 29. SEM micrograph of the fractured surface of a specimen broken

during cyclic loading in the (90% N2 + 10% H2) atmo-

sphere. 45

Fig. 30. SEM micrograph of the fractured surface of a specimen broken

during cyclic loading in the (90% N 2 + 10% H 2) atmosphere. 46

1. INTRODUCTION

Intermetallic compound Ni 3A1 forms L12 type ordered crystal structure below
the peritectic temperature [1]. Ni3A1 has good high temperature strength, and
oxidation and corrosion resistance [2]. Unlike the most conventional alloys , the
yield strength of Ni 3A1 increases with the increasing temperature [3-5]. Single
crystals of Ni 3A1 are ductile where as polycrystalline Ni 3A1 is brittle at ambient
temperatures [6-10]. The ductility of polycrystalline Ni 3A1 significantly improves '
with the small additions of boron [8] and careful control of aluminum content

[11-13].

The superior high temperature properties and fabricability of boron doped
Ni 3A1 makes it an attractive candidate for the high temperature structural appli-
cations. In the high temperature applications, the material is likely to be subjected
to cyclic loading due to vibrations etc. Generally, the materials subjected to the
high temperature cyclic loading under go microstructural changes leading to the

degradation of properties that may result in the failure of the structure.

It is well known that the high temperature deformation may cause recrystall-
ization and/ or grain growth phenomena in alloys that cause softening of the
materials. Grain boundary migration during high temperature fatigue has been

observed in Pb [14,15], Al [16,17], and Ni [18].

 

The current study is to elucidate the microstructural changes during high
temperature low cycle fatigue. This study would provide a better understanding of
the behavior of the boron doped Ni 3A1, during low cycle fatigue, for the high tem-

perature structural applications.

2. LIT ERAT URE SURVEY

2.1 The Material

2.1.1 Mechanical Properties

Long range ordered crystal structures have attractive properties for the high '
temperature structural applications. Many L12 compounds show increase in flow
stress with the increasing temperature [20-21]. Ni 3A1 falls in this category of
compounds and shows increasing strength, up to 600°C, with the increasing tem-
perature [Fig. 1.] [19]. Ni 3A1 (7' phase) is the strengthening constituent of com-
mercial nickel base super-alloys. It provides high temperature strength and creep

resistance to the super-alloys.

Although single crystals of Ni 3A1 are ductile , the polycrystalline Ni 3A1 is
brittle at both room temperature and at high temperatures [22]. The grain boun-
daries of polycrystalline Ni 3A1 are intrinsically brittle [23,24] and the material

fractures in the intergranular mode.

 

      
  

 

 

 

l
4130
-1 120
— no
.. 100
an "" 90
8 ..
3. so 5
E a:
o {3
'51 — 70 E:
E NCSAIOOZB a;
tn {7:
400 " 6°
3 o
e -. 53
>- 50 i
300* HASTELLOY-X
-. 40
200 »- ._ 3°
sxs‘~~§ _ 20
‘00 316 STAlNLESS erEL/‘\.
-[10
O 41 l L l 1 I L 1 #4 I O

0 2CD 400 600 BC!) 1000
TEMPERATURE (‘Cl

Fig. 1.. Yield stress as a function of test temperature for Ni 3Al-base
aluminide alloys, Hastelloy, and type 316 stainless steel [19].

2.1.2 Microalloying

Small additions (in the ppm range) of dopants can increase the grain boun-
dary cohesion in the polycrystalline Ni 3A1. Boron was found to be most cfbctive
dopant in improving the ductility and fabricability of the polycrystalline Ni 3A1
[11,22,25]. Taub, Huang and Chang [13] reported the beneficial cfbct of boron by
testing Ni 3A1 foil prepared by melt spin technique. With the increasing boron con-
tent, the yield strength sharply increases and the ductility gradually falls [19].
Detrimental Ni 20.41385 phase starts to precipitate as the boron content is

increased beyond 0.3 wt. % [12].

2.1.3 Effect of Boron on Grain Boundary Cohesion

With the small additions of boron, the mode of fracture in the polycrystalline
Ni 3A1 changes from intergranular to a mixture of transgranular and intergranular
fracture [12]. Boron segregates at the grain boundaries [l9] and cases the accom-
modation of slip. It improves the mobility of grain boundary dislocations which
are left behind when lattice distortion which enters the boundary from one side
and leaves it from other [26-27]. Thus the transmission of slip from grain to grain

is possible and the polycrystals, otherwise brittle, become ductile [28].

2.1.4 Effect of Alloy Stoichiometry

The beneficial efbct of boron in the polycrystalline Ni 3A1 is very sensitive to
the Stoichiometry of the alloy [11,22]. By careful control of Aluminum concentra-
tion and thermo-mechanical treatment, Liu et a1. [22] reported tensile elongation
greater than 50% and virtually transgranular fracture in Ni-24 at. % Al + 0.02-

0.10 wt. % B alloy.

Boron is most effictive in improving ductility and promoting transgranular

fracture in Ni 3A1 containing 24 at. % Al [Figs. 2—4] [19].

2.1.5 Effect of Testing Environment on the Properties

In boron doped Ni 3A1 alloys, the tensile ductility strongly depends on the
testing environment, with much lower ductility in air than in vacuum at 6000C
[Fig. 5.] [2]. The loss of ductility is accompanied by a change in fracture mode
from transgranular to intergranular. The embrittlement is due to a dynamic efbct
that simultaneously involves localized stress concentrations, gaseous oxygen, and
high temperature [2]. The severity of environmental efEct on the high tempera-
ture ductility is afbcted by preoxidation in air as well as by the aluminum con-
tent of the alloy. The oxygen embrittlement becomes less severe with a decrease

in aluminum concentration from 24 to 21 at.%.

(o)24°/. Al. e = 49.4%

TRANSGRANULAR
F RACTU RE

  

l
‘0.- - 7 ' n
. 3: rev-.1. ,\ . ”new
erase": .~ 'r‘ .‘varfi‘:
%." -?../-.oo"’l'i.l‘.‘
3. it - '1. 5"...” $.-
t". ' Tug-21.31 ' .359
a -' 2*:"“‘!"\fl - ‘

as???

1’57.
~iriifé. t

.(‘y ‘

50].Lm
I—-—t

 

(b) 24.5 7. Al, e = 37.0%
MIXED FRACTURE
MODE

 

(C) 25.0% Al, 6 = 6.0%

lNTERGRANULAR
FRACTU RE

Fig. 2 Effect of Stoichiometry on fracture mode of B-doped Ni 3A1 at
room temperature [19].

 

 

 

 

Ni=Al RATIO
77:23 76‘24 75=25 74=26
|.5 - "
IO
_ 23 0.0 _
‘g '0 "’ l2 5 '
E _ 3 '23. —
m 27 0.8 0-4
— ls -—
0.5, 23
T ,9 2° 22 0.2 '-
0 13 l 100 0.1 l
22 23 24 25 26
Al (0t pct)

Fig. 3. Plastic strain to fracture (%) of rapidly solidified Ni 3Al-B as a
function of Al and B concentration. The solid lines indicate constant

NiAl ratio (at. %) [19].

 

 

 

 

 

 

1.0-
E
\
9;
9
.—
:05
5 ' 020
a: 7 ' a
I E
§ 3.
‘i‘ 9
5..
O “0.35...
I
——-o— 0
G
I
X
4
DJ
Q
m: -‘O.1O
.1—
W-z~
zt—v
we
b-O J
O l 1
24 25 - 26

Al CONCENTRATION (08.7.1

Fig. 4. Efibct of Stoichiometry on grain boundary segregation (in terms
of peak-height ratio) and room temperature tensile ductility of B—doped
Ni3Al containing 24 to 25.2 at. % Al [19].

10

 

 

 

 

 

6° 1 l l
‘ _—-—
‘ i
so "' A ‘ 4‘13" A
T \
VACUUM TEST. BARE SPECIMEN
g —
z ‘0 '-
Q
g...
‘<
‘51
c) 30 - .4
.1
W
I.“
.J
«'1':
5 20 ‘" —-4
.—
AIR TEST, PREOXIDIZED SPECIMEN
IO '— .—
¢ 1- “ - '- -.
8%./I/K AIR TEST. BARE SPECIMEN
1
0

 

24 23 22 21
(A! + Hf) CONCENTRATION (It SI

Fig. 5. Plot of tensile elongation of Ni3Al alloys as a function of
(A1+Hf) concentration for bare specimens tested at 600°C in vacuum or
. for specimens preoxidized at 1100°C/2h + 850°C] 5h and tested in air.

Solid symbols for alloys with 0.5% Hf and open symbols for the alloy
without Hf [2].

11

2.2 Grain Boundary Migration During Low Cycle Fatigue

Grain boundary migration during high temperature fatigue is an important
microstructural phenomenon. Snowden [29] first reported the gain boundary
migration and the appearance of slip trace in the bending fatigue test of pure
lead. He found that the gain boundaries tend to migate to positions $450 to
the specimen axis that resulted in the formation of an orthogonal gain structure.
After large number of cycles, the rate of migation decreased possibly due to the

alignment of gain boundaries in the directions of maximum shear stress.

Grain boundary migation and gain boundary sliding have been reported in
Al[16,17,30], Pb[14,15], Pb-Sn solid solution alloy[3l,32], Al-Mg solid solution
alloy[33], and Ni[18]. Room temperature reverse bending and torsion fatigue of
the high purity lead (99.9995%) showed microstructural changes right from the
first cycle. The result showed extensive gain boundary migation and as result a
series of parallel markings at many boundaries could be observed under optical
microscope [Fig. 6.] [14,15]. A well defined one-to—one correspondence between the
number of gain boundary markings and the total number of cycles was observed.
The clarity of each migation marking suggests cyclic gain boundary sliding
which was confirmed using two-beam interferometry [Fig. 7.] [14,15]. Optical
microscopy, using Nomarski interference contrast, revealed a fine structure within
the migation markings [Fig. 8.] [14,15]. In coarse gained specimens, the longer

gain boundaries formed a zig-zag pattern of migation markings.

Abdel-Raouf et al. [34] reported a change in the gain morphology in OFHC

Cu due to the gain boundary migation during cyclic test at 650°C at a low

12

 

Fig. 6. SEM microgaph showing cyclic gain boundary migation after
testing for 15 cycles in reverse bending at a frequency of 0.17 Hz and a
strain amplitude of i0.28%. A preexisting twin is at A, and B marks
the initial position of the boundary prior to testing [14,15].

 

Fig. 7. Two-beam interferometer (monochromic light) showing repetitive
cyclic markings and the presence of sliding after testing for 3 cycles in
reverse bending at a frequency of 0.10 Hz and a strain amplitude of
i0.48% [14,15].

14

... . I
_ occur-- '
. I. I ‘

- ‘ '- I — , _ c .
7 d I m
;:.. :;‘~"’-"4~37-‘-'- at»???

‘ v \.(_
. 'a‘u‘X';

“e
n v."

'.I,f.'.. 20 pm-

.014.

 

—s.-

\

1. . . ~“l ' 54‘; V i .
. . w: ,_; s\_. . -'
- 1' .x' 41v Ail-AC}:

Fig. 8. Optical microgaph, using Nomarski interference contrast, show-
ing the presence of a fine structure within the migation markings after

testing for 8 cycles reverse bending at a frequency of 0.17 Hz and a
strain amplitude of i0.28% [14,15].

15

strain rate ( 4x10’4 3'1 ). The gain shape was found to be strain rate depen-
dent with no change in the gain morphology at high strain rates. Snowden, and
Langdon et al. proposed that the gain boundary migation was stress induced
process. Recently, Chen [35] proposed that the gain boundary migation is

strain induced phenomenon.

Dispersions like impurities, precipitates, or voids slow down or inhibit the
gain boundary migation. That is why, Type 304 stainless steel does not show

any microstructural change at 760°C at the low strain rates.

3. EXPERHVIENTAL PROCEDURE

3.1 Testing Material

A bar shaped casting prepared by arc-melting several times in argon atmo-
sphere to promote homogeneity and then drop-cast in a copper mold was supplied

by Oak Ridge National Lab. The final composition of the casting was as follows: ’

Ni: 76 at. %
Al: 24 at. %
B: 0.24 at. %

One lot of material was received three years back and the another lot was
received one year back. After vacuum annealing, the earlier lot showed a smaller
gain size as compared to the latter lot. This point would be brought up in the

heat treatment section.

16

17

3.2 Specimen Preparation

3.2.1 Mechanical Test Specimens

(1) Ni 3A1 cyclic test samples were machined from the blocks cut from the sup-
plied cast bars. The blocks were annealed in a vacuum furnace at 800°C for 3

hour so as to ease the machining of the material. Specimens were machined

to the shape and dimensions as shown in Fig. 9. [35].

(2) Heat Treatment

All the specimens were annealed in a vacuum furnace at a pressure of approx-
imately 10’3 Pa at 800°C for 5 hour except for one sample which was annealed at

1200°C for 2 hour. Annealed specimens were chemically polished at 60°C to

remove any surface irregularity, using the following solution:

Nitric Acid:
Phosphoric Acid:
Distilled Water:
Acetic Acid:

Hydrofloric Acid:

Hydrochloric Acid:

20 ml

20 ml

20 ml

10 ml

10 ml

10 ml

18

 

Fig. 9. Fatigue Sample (B-doped Ni3Al) [35].

19

The two batches of Ni 3A1 were received two years apart. The gain size of
the heat treated specimens from earlier batch was small and gains were equi-axial
[Fig. 10.] where as the gain size of latter batch was larger and the microstructure
was dendritic [Fig. 11.]. This difference can be explained as some variation in

fabrication procedure of the material.

3.2.2 Optical Microscopy Specimens

After the completion of cyclic tests, some microgaphs were taken by polish-
ing the plane parallel to the loading direction. Specimens were cut perpendicular A
to the loading direction with a slow diamond wheel cutting machine (Buehler
Isomet). Subsequently, they were mounted in cold setting resin and polished on
abrasive git paper and then on a cloth by alumina powder to get a mirror like

surface. The polished samples were etched with the following solution:

CuSO 4: 5 g
HCl: 20 ml
H 20: 20 ml

All the Optical microgaphs were taken with a Neophot 21 Microscope.

20

 

 

Fig. 10. Microstructure of Ni 3A1 annealed at 800°C for 5 hour.

21

 

 

Fig. 11 Microstructure of Ni 3A1 annealed at 800°C for 5 hour.

22

3.2.3 Scanning Electron Microscopy Specimens

The fractured surface was ultrasonically cleaned with acetone. A Hitachi

SEM was used to study the fractured surface. The applied voltage was 25kV.

3.3 Mechanical Testing

Two different testing machines and set-up were used to perform the high

temperature low cycle fatigue tests.

3.3.1 Instron Testing Machine

High temperature low cycle fatigue tests were conducted on a floor model
electro—mechanical Instron testing machine with a 500 Kg tension-compression
reversible load cell. The pull rods and button grips were made of 310 heat resis-
- tant stainless steel. To avoid the oxidation of the specimen, during high tempera-
ture testing, a cylindrical protective chamber was designed as shown in Fig. 12.
[35]. A stainless steel ring with a clearance of 0.75 mm to pull rod was welded at
the top of the tube, while the lower ring which fitted to the lower pull rod was
made of Invar and the dimensions were chosen such that at temperature higher
than 200°C, a tight fit with the lower pull rod was obtained due to the difference

in the thermal expansion between the Invar ring and lower pull rod.

23

 

 

 

 

 

 

 

 

 

 

 

I
UPPER
PULL R00

0 0
0 a SIAINLESS STEEL
., ml W
0
“’ W
0 Q

HEATER Ma ¢
o L J a
G G
q; /spccnucu a
3 TC O

C1 fl

9 o
0 0
. r .
0 o
O 0
3 ~ 3
:- I 0 INVAR
0 o

 

 

 

 

Fig. 12. Instron testing machine set up [35].

24

During the test, a reducing gaseous mixture (90% N 2 + 10% H 9 was flown
at the flow rate of 12 liter/ hour at 5 psi to minimize the oxidation. Higher flow
rates were avoided to reduce burning of the gas at the top of the chamber and the

unwanted back pressure.

Computer codes were developed to computer control the Instron by digital
output fi'om IBM XT personal computer and for data acquisition of load cell sig-
nals using an A/ D converter. Load and displacement data from test could be
stored at chosen interval. The resolution of data acquisition was 4 um per data

set with a accuracy better than 0.1%.

3.3.2 MTS Testing Machine

In the later part of the project a new computerized high temperature high
vacuum servohydraulic test system was used. The essential parts of the system
are schematically shown in Fig. 13. [35]. The system includes a closed loop frame
(MTS 810) with 10-100 kN (MTS 661.20A-03) or 0.5-10kN (MTS 661.19) load cell,
a 25.4 mm gauge length water cooled high temperature extensometer with max-
imum strain range of 30% and minimum strain range of -10%, and a Centorr S-
60 high vacuum high temperature furnace. The maximum resolution of the exten-
someter is 0.1 pm, however, due to limitation of 12 bit A/D converter the resolu-

tion is 0.488 um.

Specially designed gips [Fig. 14.] [35] were used to avoid the crossover

phenomenon during change in the loading direction in tension-compression fatigue

Ana 82$? 53
ozsfieangbm Esau? $2 3383an :wE canteen—9:00 .2 .wi

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

25

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

   

0; wk!

_._

 

 

 

 

 

 

 

 

26

      
 
 
 
 
  
    
    
  
   
 
 
  

__ 1' Jlllflul

l

 

J .; ;
\\ =' " rat-
I

,§

III/III III)

III]

,1 C
[III]

TL—
{i
\\

Iran“

.3

 

1)
r

 

L T
ll
1.
\\:\7

   

VACUUM SEAL
COOLING IN

COOLING ou1\_h v. -1.“ PRESSURE IN

PRESSURE RETURN

Fig. 14. High temperature grip assembly [35].

27

test. The two ends of the specimen were preloaded by applying hydraulic pressure
on the pistons residing inside the pull rods. The specimen insert included two pair
of collet 3, four tungsten pins and two extension pistons shown in Fig. 15 [35]. All

the hot zone parts are made of high temperature molybdenum alloys (TZM).

Mechanical test controlling system comprises of one Microconsole (MTS
458.20), one MicroProfiler .(MTS 458.91), one AC controller (MTS 458.13), and
two DC controller (MT S 458.11). The MicroProfiler has 52 KB memory and can
store ninety nine progams. Remote progammability allows the progams to be
sent from computer via RS 232. Signals from the load cell, extensometer, or

LVDT can be used to perform the load, strain, or stroke tests.

The Centorr Model S-60 high vacuum furnace is mounted on the load frame
of MTS. Two bellows for the 50.8 mm diameter rods accommodates the move-
ments of the actuator up to 150 mm. The hot zone size is 76.2 mm diameter by
203.2 mm high with two halves of tungsten mesh heating elements. The heat radi-
ations are blocked by six layers of heat shields made of tungsten and molybde-
num. The extensometer is fixed in the furnace chamber outside the heat shields.
Fig. 16. [35] shows the final set up for the test. The maximum operating tempera-
ture of the furnace is 2000°C in vacuum or controlled inert gas (argon, nitrogen,
helium) atmosphere. Two "C" W5%Re/W26%Re type thermocouples are used
along with a 20 segment progammable Honeywell Universal Digital Controller.
Three mode controller provides temperature control stability within i1°C accu-
racy. Over temperature controller is installed to protect the system from over

heating in the case of damaged or broken main thermocouple.

28

 

III
C

 

 

 

 

EXTENSION PISTON

 

 

 

 

SIDE VIEW

 

BACK VIEW
W
/
tour I

 

 

 

 

 

 

 

FRONT VIEW

 

Fig. 15. Specimen inserts for the high temperature high vacuum MTS
test system [35].

29

 

Fig. 16. Final set up inside the vacuum furnace of the high temperature
high vacuum MTS test system [35].

30

Machine control and data acquisition is performed by a Zenith AT personal
computer with 1MB main memory. One high performance AID Input Output
board ( Data Translation DT 2818) is connected to the one of the expansion slots
of the personal computer. The DT 2818 can be progammed from the computer
compiled BASIC language to perform A/ D conversions, D/A conversions, and
digital input and digital output transfers. It has a 12 bit A/ D converter with a
maximum data sampling rate of 13.7 kHz. Using Direct Memory Access (DMA),
the data can be moved directly into or out of BASIC variable arrays at high
speeds. PCLAB, a real-time application software package containing a number of
machine language routines designed to be called from computer compiled BASIC
language geatly reduced the progamming required to use DT 2818. High level
routines for data acquisition and control were written by Chen [35] in MicroSoft

QuickBASIC language.

4. EXPERIMENTAL RESULTS

4.1 Mechanical Behavior

In order to characterize the high temperature deformation at 600°C, two
strain amplitudes (elastic + plastic) of 0.5% and 1% were used. A plot of true
stress versus cumulative strain, obtained from a test conducted in high vacuum -
using 0.5 % strain amplitude, is shown in Fig. 17. In the first few cycles, the true
stress rapidly increased before reaching saturation, then decreased in the latter
cycles. The initial steep increase in the true stress was due to large strain harden-
ing rate. The subsequent decrease in the true stress at large cumulative strain
indicates the occurrence of a strong softening mechanism. Fig. 18 exhibits results
corresponding to 1% strain amplitude cyclic test, using a reducing (90% N2 +
10% H2) atmosphere. This test also shows a maximum followed by a continuous
decrease in the true stress. Fig. 19 is a plot of the true stress versus the number
of cycles for 1% strain amplitude cyclic test, using a reducing (90% N 2 + 10%
H2) atmosphere. It corroborates the true stress behavior represented by Fig. 18.
Fig. 20 shows the results of stress controlled test ( constant stress of 85 MPa) in a

reducing (90% N 2 + 10% H2) atmosphere.

31

32

0.68 3 8:3? .32 E :otaecouou one? ”63:95
52:. .wemd .28 59:» 033388 “35> v.85 on: me SE 2 .mi

 

  

a“ 3 55:... rats as a
. _ _ S as...
a
a
. . _.
______:__,____ _:__E:___________________:__:____: , m
____________________ __________:____:____________:__________:___,_ a
, , a
m
+ __
._
. H
_ _ _ Ga.—

 

 

 

33

coca

.968 a a: s2
+ a 2 «8% .«o absence; €6.62 5 eotaficaoe 0:98 avian—Eu
59:“ am; he 52:. 033388 name? 38; on: Co 85 M: .mE

 

 

  

E =55 3:5 83

 

   

 

:3

8

Illa-C:

r—-n==:=2 wr—ccfiflflmm

34

.988 a 3e 82 +
u 2 fioev mo Pangaea; 98.62 5 eomuafieouoe one? 88298
$83 8“ com 83.0 mo beau: 383 «8:» 25 .«o 83 3 .mmm

 

 

8.26 3 .3832
8mm . 8.3 . 8.2 . . 8....

 

 

 

 

.om I

is
N
°dw ssens em;

.om

 

35

0°88 3 at $2 + N2 83 8 22885;
wEusooH 5 £52 mm Mo 838 5528 9:2. .82 3:828 838
E3250 5“ 588 0322383 2.8? 828 02.: mo SE om .wE

mm 5

 

as E as; gas

 

 

a

a

H

 

94¢:

9‘9:th cat—accr-I‘lmw

36

4.2 Microstructural Development

Substantial differences were observed in the microstructure of the specimens
in the initial and final states. A very few large grains are observed in Fig. 21
(0.5% strain amplitude cyclic loading in vacuum) as compared to Fig. 11 (initial
structure). Extensive grain boundary migration can 'be observed by comparing
Fig. 10 (initial microstructure) to Fig. 22 (after 1% strain amplitude cyclic test).
Similar microstructural changes can be observed by comparing Fig. 23 (initial
microstructure) to Fig. 24 (after stress controlled test). However, one high tem-
perature test done in vacuum did not show any significant microstructural change

even after 1000 cycles of loading [Figs. 25 and 26].

For a specimen annealed at 800°C for 5 hour and fatigued at 0.5% strain
amplitude in high vacuum, the microstructure in the plane parallel to the loading
direction showed a tendency of grain boundary alignment at :l:450 to the stress
axis [Fig. 27]. Another specimen annealed at 1200°c for 2 hour and fatigued at

0.5% strain amplitude in high vacuum, showed no such tendency [Fig. 28].

4.3 Brittle Failure

The fractured surface of a specimen that broke in 0.5% strain amplitude
cyclic test in a reducing (90% N 2 + 10% H 2) atmosphere showed extensive inter-

granular cleavage with a very little transgranular fracture ( Figs. 29 and 30).

37

 

Fig. 21 Microstructure of Ni 3A1, in the plane perpendicular to the
loading axis, annealed at 800°C for 5 hour, after 100 cycles of 0.5%
strain amplitude at 600°C.

38

 

Fig. 22 Microstructure of Ni 3A1, in the plane perpendicular to the
loading axis, annealed at 800°C for 5 hour, after 3900 cycles of 1%
strain amplitude at 600°C.

39

 

Fig. 23 Microstructure of Ni 3A1, in the plane perpendicular to the
loading axis, annealed at 800°C for 5 hour.

40

 

Fig. 24 Microstructure of Ni 03A1, in the plane perpendicular to the
loading axis, annealed at 800 C for 5 hour, after 790 cycles of con-
stant stress of 85 MPa at 600°C.

41

 

Fig. 25 Microstructure of Ni 8A1, in the plane perpendicular to the
loading axis, annealed at 1200 C for 2 hour.

42

 

Fig. 26 Microstructure of Ni 3A1, in the plane perpendicular to the
loading axis, annealed at 1200°c for 2 hour, after 1000 cycles of
0.5% strain amplitude at 600°C.

43

 

Fig. 27 Microstructure of Ni 3A1, in the plane parallel to the loading
axis, annealed at 800°C for 5 hour, after 100 cycles of 0.5% strain
amplitude at 600°C.

44

 

Fig. 28 Microstructure of Ni 3A1, in the plane parallel to the loading
axis, annealed at 1200°c for 2 hour, after 1000 cycles of 0.5% strain
amplitude at 600°C.

45

 

Fig. 29 SEM micrograph of the fractured surface of a specimen bro-
ken during cyclic loading in the (90% N 2 + 10% H 2) atmosphere.

46

,
ease. zsxseifiuflx-

 

Fig. 30 SEM micrograph of the fractured surface of a specimen bro-
ken during cyclic loading in the (90% N 2 + 10% H 2) atmosphere.

5. DISCUSSION

5.1 Mechanical Behavior

All the cyclic hardening curves [Figs. 17-20] show a maximum followed by a
continuous decrease in the true stress. The nature of curve remains same for
different strain amplitudes and stress controlled test. The continuous decrease in
the true stress after the maximum indicates the occurrence of a strong mechanism
at large cumulative strain. The softening could be due to the grain boundary

migration.

5.2 Grain Boundary Migration

Extensive grain boundary migration was found in all but one cyclic tests.
Grain boundary migration during LCF has been reported in Al[16,17,30],
Pb[14,15], Pb-Sn solid solution alloy[31,32], Al-Mg solid solution alloy[33], and
Ni[18].

No evidence of grain boundary migration in one of the test [Figs. 25,28] could

be attributed to the voids present in that particular specimen [Fig. 28].

47

48

Dispersions like precipitates and voids pin down the grain boundaries thus inhibit

the grain boundary migration.

5.3 Brittle Failure

Liu and White [2] have reported oxygen embrittlement of B-doped polycry-
stalline Ni 3A1 during the high temperature deformation. The sample observed
under SEM had clean grain boundary cleavage with a very little ductile fracture.
Although a reducing mixture of (90% N 2 + 10% Hz) was used in the Instron
machine, the presence of the traces of air could not be ruled out. The brittle
failure could be attributed to the presence of elemental oxygen at the grain boun-

daries and/ or formation of glassy B 203 phase.

6. CONCLUSIONS

The present study on B-doped Ni 3A1 yields the following results:

1. Under high temperature low cycle fatigue, the true stress amplitude (as a
function of the cumulative strain) shows a maximum, followed by a continu-

ous decrease at large cumulative strains.

2. There is an extensive grain boundary migration during the high temperature

low cycle fatigue.

3. Dispersions in the form of voids, pin down the grain boundaries and inhibit

grain boundary migration.

4. The ductility of B-doped Ni 3A1 strongly depends on the test environment.
Oxygen, if present even in traces, adversely affects the ductility of the

material.

49

7. BIBLIOGRAPHY

[1] M. Hansen, Constitution of Binary Alloys, McGraw Hill Book Co., New
York, p. 119, (1958).

[2] C. T. Liu and C. L. White, Acta Metall., vol. 35, No. 3, pp. 643-649,
(1987) .

[3] P. H. Thorton, R. G. Davis and T. L. Johnson, Metall. Trans. 1 p. 207,
(1977).

[4] P. A. Flinn, Trans. TMS. AMIE 218, p. 145, (1960).
[5] R. G. Davis and N. S. Stoloff, Trans. TMS. AMIE 233, p. 714, (1965).

[6] . M. Grala, Mechanical Properties of Intermetallic Compounds, ed. J.

. Westbrook, John Wiley and Sons, Inc., New York, p. 358, (1960).

mm

[7] R. Moskovic, J. Mater. Sci. 13, p. 1901, (1978).

[8] K. Aoki and O. Izumi, Trans JIM 19, p. 203, (1978).

[9] A. V. Seybolt and J. H. Westbrook, Acta Metall. 12, p. 449, (1964).
[10] K. Aoki and O. Izumi, Nippon Kinzoku Gakkaishi 41, p. 170, (1977).

[11] C. T. Liu and C. C. Koch, Technical Aspects of Critical Materials Used
by the Steel Industry, vol. IIB, NBSIR-83-26—A-Z National Bureau of
Standards, p. 42-1, (1983).

[12] C. T. Liu, C. L. White and J. A. Horton, Acta Metall., vol. 33, p. 213,
(1985).

[13] A. I. Taub, S. C. Huang and K. M. Chang, private communication
(1984).

50

51

[14] T. G. langdon and R. C. Gifkins, Acta Metall., vol. 31, pp. 927-938,
(1983).

[15] T. G. langdon, D. Simpson and R. C. Gifkins, Acta Metall., vol. 31, pp.
939-946. (1983).

[16] P. Yavari and T. G. langdon, Acta Metall., vol. 31, pp. 1595-1603,
(1983).

[17] P. Yavari and T. G. langdon, Acta Metall., vol. 31, pp. 1605-1610,
(1983).

[18] S. R. Chen and G. Gottstein, Acta Metall., vol. 36, pp. 3093-3101,
(1988).

[19] C. T. Liu and C. L. White, Design of ductile polycrystalline Ni3Al
alloys, (in the press).

[20] E. M. Schulson and J. A. Roy, Acta Metall., vol. 26, p. 29, (1978).
[21] S. Takeuchi and E. Kuramuto, Acta Metall., vol. 26, p. 415, (1973).

[22] C. T. Liu, C. L. White, C. C. Koch and E. H. Lee, High Temperature
Materials Chemistry -II, Munir & Cubicciotti, eds, vol 83(7), pp. 32-41,
(1983). '

[23] T. Takasugi, E. P. George, D. P. Pope and O. Izumi, Scripta Metallur-
gica, vol 19, pp. 551-556, (1985).

[24] T. Ogura, S. Hanada, T. Masumoto and O Izumi, Metallurgical Transac-
tions A, vol 16A, pp. 441-443, (March 1985).

[25] K. Aoki and O. Izumi, Nippon Kinzoku Gakkaishi 43, p. 1190, (1979).

[26] E. M. Schulson, T. P. Weihs, I. Baker, H. J. Frost and J. A. Horton,
Acta Metall., vol. 34, p. 1395, (1986).

[27] E. M. Schulson, T. P. Weihs, 1. Baker, H. J. Frost and J. A. Horton,
Scripta Metall., vol. 19, pp. 1497-1498, (1985).

[28] E. M. Schulson, 1. Baker, H. J. Frost, Metallurgical Transactions A, vol
18A, p. 1995, (Nov. 1985).

52

[29] K. U. Snowden, Phil. Mag., vol. 6, p. 321, (1961).

[30] V. Raman and T. G. langdon, J. Mat. Sci. Lett., vol 2, p. 180, (1983).
[31] H. S. Betrabet and V. Raman, Metall. Trans., vol. 19A, p. 1437, (1988).
[32] V. Raman and T. C. Reiley, Scripta Meta., vol. 20, p. 1343, (1986).
[33] V. Raman and T. G. langdon, Acta Metall., vol. 37, p. 725, (1989).

[34] H. Abdel-Raouf, A. Plumtree and T. H. Topper, Metall. Trans., vol. 5,
p. 267, (1974).

[35] S. R. Chen, Ph. D. thesis, Michigan State University, (1989).

"‘11111111111111.1111ES