You are on page 1of 12

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/271224490

Investigation of the Dynamic Strain Aging and Mechanical Properties in


Alloy-625 with Different Microstructures

Article  in  Metallurgical and Materials Transactions A · March 2015


DOI: 10.1007/s11661-014-2717-z

CITATIONS READS

9 414

4 authors, including:

Arnomitra Chatterjee Garima Sharma


Bhabha Atomic Research Centre Department of Atomic Energy
47 PUBLICATIONS   187 CITATIONS    50 PUBLICATIONS   860 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Study of Intermetallic Alloys View project

Study of nanocrystalline materials View project

All content following this page was uploaded by Arnomitra Chatterjee on 21 March 2016.

The user has requested enhancement of the downloaded file.


Investigation of the Dynamic Strain Aging and Mechanical
Properties in Alloy-625 with Different Microstructures
ARNOMITRA CHATTERJEE, GARIMA SHARMA, R. TEWARI,
and J.K. CHAKRAVARTTY

Tensile tests were carried out on service exposed Alloy 625 ammonia cracker tube used at heavy
water production plant to study the effect of microstructure on the serrated yielding and
mechanical properties of the material. Owing to temperature gradient during service exposure,
the microstructure was different in top, middle, and bottom sections of the tube. Variation of
flow stress, ductility, and average work hardening were monitored with temperature. In the
present work, emphasis was given on the study of serrated yielding in the service exposed Alloy
625. Detail investigations were made to study the effect of microstructure on the underlying
mechanism of dynamic strain aging of the material. The study revealed that both the normal
and the inverse Portevin–Le Chatelier effect (PLC) occured in the material at lower and higher
temperature regime, respectively. While the normal PLC dynamics was associated with locking
of dislocations by interstitial carbon atoms, the inverse one was accomplished by the dislocation
pinning by substitutional Mo atoms. Further analyses identified that the basic deformation
mechanism was different in middle and bottom samples as that in the top samples which was
reflected in the difference in their respective activation energy and stress drop magnitude.

DOI: 10.1007/s11661-014-2717-z
 The Minerals, Metals & Materials Society and ASM International 2014

I. INTRODUCTION exposed condition as precipitation of various phases have


been identified during service exposure.[1,2,4–7] Precipita-
THE Alloy-625 is an austenitic alloy which is being used tion-induced hardening accompanied with reduction in
for a variety of components in the aerospace, marine, ductility and toughness during service exposure could be
chemical, and nuclear industries.[1–4] It possesses excellent implicated for the premature failure. Hence, a huge
strength at both ambient and elevated temperatures along amount of work has been done on Alloy-625 to study the
with good oxidation and corrosion resistance.[2–4] More- mechanical properties of the material after service expo-
over, the alloy has a very good resistance to creep sure and to establish the structure property correlation to
deformation and hydrogen attack, excellent formability, predict the life time of the tube from microstructural
and low hydrogen permeability which make this alloy very changes of the material.[1–3] Raman et al.[5] have studied
suitable for use as the cracker tube material in ammonia mechanical behavior of Alloy 625 cracker tubes which had
cracker units of heavy water production plants.[2–4] undergone 3 9 104 hours of the service life. They noticed
Cracker tubes are the primary containment of gaseous the change in mechanical properties along the length of the
mixtures of nitrogen, hydrogen, and deuterium produced tube because of the process-related temperature gradient
by cracking of enriched ammonia at temperature from set up during operation. However, no attempt was made to
923 K to 973 K (650 C to 700 C) at a pressure of around correlate the changes in mechanical properties with micro-
14 MPa. The tubes are approximately 12 m long with a structure. Vani Shankar et al.[3,7] and Sundararaman
circular cross section of diameter 90 mm and wall thickness et al.[6] also studied several features of the microstructure
of 8 mm. They are placed inside a rectangular furnace of the material. Chakravartty et al.[2] carried out a detailed
which heats the tubes from outside. Even though, the Alloy study of the Alloy 625 of cracker tube material after
625 cracker tubes are designed for a lifetime of 105 hours completion of 105 hours of service exposure to evaluate the
under service conditions, several tubes fail before the influence of service temperature on the development of
anticipated lifetime. The microstructure of the solution- microstructure and mechanical properties. Structure prop-
treated virgin material changes considerably in the service erty correlation was developed further depending on the
experimental observations. Apart from the aforemen-
tioned studies, the serrated yielding in Alloy 625 has been
a topic of considerable interest among the researchers.
ARNOMITRA CHATTERJEE, Scientific Officer D, GARIMA
SHARMA, Scientific Officer G, and J.K. CHAKRAVARTTY, Serrations in the flow curves, commonly known as the PLC
Outstanding Scientist, Director Materials Group, are with the effect[8–11] arise from Dynamic Strain Aging (DSA) pri-
Mechanical Metallurgy Division, Bhabha Atomic Research Centre, marily due to repeated pinning-unpinning of mobile
Trombay, Mumbai 400 085, India. Contact e-mail: jayanta@barc. dislocations by diffusing solute atoms. Serrated flow has
gov.in R. TEWARI, Scientific Officer H, is with the Materials Science
Division, Bhabha Atomic Research Centre.
also been reported due to precipitate formation and their
Manuscript submitted August 28, 2013. shearing, formation of shear bands and occurrence of
Article published online December 23, 2014 stacking faults.[12–14] The serrations in stress–strain curves

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, MARCH 2015—1097


are classified in three generic types: type A, type B and type would be very fascinating to investigate if this is
C.[1,15] Type A serrations are locking serrations character- reflected in the DSA behavior of the material as well.
ized by abrupt rise in stress followed by a drop to or below In general, DSA has been found to influence plastic
the general level of the stress–strain curve.[1] Type B behavior of many engineering materials abnormally and
serrations are oscillations about the general level of the have a detrimental effect on the ductility and toughness
stress–strain curve which occurs due to discontinuous which may lead to the premature failure of the compo-
deformation band propagation arising from DSA.[1] nents. Since Alloy-625 is intended for structural appli-
Finally, for type C serrations, stress drops below the cation, it is very important to investigate to what extent
general level due to the unlocking of dislocations.[1] The DSA influences the mechanical properties of the alloy. It
repeated locking-unlocking of dislocations by solutes is is likely that the partitioning of both substitutional (Nb,
translated into negative strain rate sensitivity (SRS) of the Cr, Mo) and interstitial (such as carbon) elements
system.[16,17] In order to understand the nucleation and between the matrix and precipitates (during service
propagation steps involved with these serration types and exposure), and the size, the distribution and the nature
their relationship with microscopic mechanisms, consider- of precipitates may have considerable influence on the
able experimental and theoretical efforts have been made. dynamics of serrated flow behavior. In the present work,
Zhang et al.[18] investigated the nucleation and propaga- the serrated yielding behavior of the service exposed
tion of PLC bands by a novel digital speckle pattern Alloy 625 has been studied to elucidate the effect of
metrology technique consisting of digital speckle pattern microstructure and to identify the mechanism of DSA.
interferometry (DSPI) and digital speckle correlation The serrated flow behavior has been characterized here
(DSC). P. Zavattieri et al.[19] have reported a detailed by monitoring the variation of yield stress (YS), ultimate
investigation of spatio-temporal characteristics of the PLC tensile strength (UTS) and ductility with temperature
effect in austenitic steel with twinning-induced plasticity and strain rate, evaluating strain rate sensitivity (SRS),
(TWIP). Spatio-temporal pattern of PLC bands has been and estimating activation energy for all the three
investigated by Tong et al.[20] by a technique based on high- microstructural conditions corresponding to top, mid-
speed digital photography and image correlation. In the dle, and bottom sections.
present work, no visual techniques have been adopted to
classify the PLC bands. Instead, band characterization has
been made based on the types of serrations observed in the II. EXPERIMENTS
stress–strain curves.
Kim and Chaturvedi[21] investigated serrated flow in Materials for this study was taken from a used
Alloy 625 in solution-treated condition and found out ammonia cracker tube of Alloy-625 (service exposed for
that the Mo atoms play an important role in serrated 100,000 hours) obtained from a heavy water plant and
yielding behavior of the material. Vani Sankar et al.[1] was the same as used in an earlier work.[2] The nominal
carried out a detailed study on the solution-treated composition of the alloy is given in Table I. The outside
Alloy 625 retrieved from cracker tubes after skin of the top, middle, and bottom locations of the 8 m
60,000 hours service exposure at 873 K (600 C). They long tube were exposed to a temperature of 932 K,
reported serrated yielding behavior in Alloy 625 in a 949 K and 958 K (658 C, 676 C and 685 C), respec-
temperature regime of 523 K to 1023 K (250 C to tively, under operating conditions. Various samples
750 C). They estimated the activation energy for DSA (tensile, impact and fracture toughness) were machined
to be 98 kJ/mol and conjectured that Mo solute is from the central portion of top (T), middle (M) and
responsible for serrated yielding in the solution-treated bottom (B) sections. The details of samples and test
material. During the tensile tests, Chakravartty et al.[2] procedure are given elsewhere.[2] Tensile tests were
could observe serrations in the flow curves in a performed over a temperature regime from room
temperature range of 523 K to 823 K (250 C to temperature (RT) to 973 K (700 C) under a constant
550 C) which was attributed to DSA but no study strain rate condition. Three different strain rates were
was carried out to investigate the DSA phenomenon in chosen for the study: 8.8 9 105, 4.4 9 104, and
the service exposed Alloy 625. It has already been 1.8 9 103 s1. The tests were carried out in a screw
established that the temperature gradient along the driven tensile testing machine using a resistance heating
length of the tube during service exposure renders furnace with temperature control ±1 C. Apart from the
different microstructure development in terms of the conventional tensile tests, a number of strain rate jump
types, size, and distribution of precipitates along the tests were also carried out under different temperatures
length of the tube in the top, middle, and bottom to estimate the SRS of the material. TEM investigations
sections of the cracker tube and thus results in different were done on the tensile-tested samples by slicing
mechanical properties in these sections.[2] Hence, it samples at 45 deg to the fractured region in the guage

Table I. Nominal Chemical Composition of the Alloy 625 Cracker Tube Material

Element Ni Cr Mo Nb + Ta Ti + Al Fe Co Si C P and S
Weight percent bal. 20 to 23 8 to 10 3.15 to 4.15 0.4 each 5 max 1.0 0.5 0.01 0.015 each

1098—VOLUME 46A, MARCH 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A


portion. Samples for TEM investigation were prepared (700 C) at a strain rate of 4.2 9 104 s1, following
by jet polishing 3 mm disks using 20 pct perchloric acid important observations were made
and 80 pct methanol at 20 V and 228 K (45 C). TEM
(i) A decrease in YS was noted in specimens of all
analysis was performed on the Jeol 2000 FX operating
sections with increasing temperature except in the
at 200 kV.
temperature range 523 K to 673 K (250 C to
400 C). While UTS of the B and the M sections
increased with temperature up to 873 K (600 C),
III. RESULTS the T section showed a moderate decreasing trend
of UTS up to about 773 K (500 C) followed by a
Both optical and electron microscopic examinations
rapid drop. At temperatures above 923 K
have been carried out to reveal microstructural features
(650 C) the UTS values of both the B and the M
of the undeformed specimens from the three locations.
sections were found to decrease rapidly. The vari-
The solution-annealed virgin tube exhibited precipitates
ation of the strength parameters with temperature
of different types within the matrix and at grain
appeared to be related to the prevailing micro-
boundaries on service exposure. The details of micro-
structure and chemical composition of the matrix
structural characterization have been reported previ-
phase.
ously[2] and Tables II, III summarizes the
(ii) The ductility (total elongation at break) showed
microstructural features observed in three sections.
an increasing trend with temperature for all the
While the grain boundaries in the T specimens showed
sections
a thin and discontinuous layer of carbides, thicker and
(iii) The T section exhibited higher strength and duc-
continuous precipitates were found in the M (carbides)
tility in the temperature range RT—773 K
and the B (carbides and d) specimens. c¢¢ phase was the
(500 C) compared to that observed for the M
major precipitate within the matrix in all the three
and the B sections.
sections. However, the size of c¢¢ phase was largest in the
(iv) A good correlation between microstructure and
B specimens followed by the M and the T specimens. In
mechanical properties could be established.
addition to c¢¢ phase, Ni2(Cr,Mo) and d were observed
within the matrix of the T and the B specimen, Although, at ambient temperature, stress–strain
respectively. Electron microscopic investigation con- curves were smooth at the strain rates employed in the
firmed the carbides to be of the M23C6 type. It has previous study, serrations appeared in stress–strain
been reported that microstructural development in each curve at higher temperature regime of 523 K to 873 K
section could be matched very well with the temperature (250 C to 600 C). In the following, the experimental
experienced and it was established that the B section results of serrated flow behavior of service exposed
experienced the highest temperature. The complex Alloy 625 are presented.
microstructural development in the three sections Stress–strain curves were obtained for all the three
resulted in distinctly different mechanical properties conditions in the temperature range of RT—973 K
which have also been published.[2] When these mechan- (700 C) at three different strain rates: 8 9 105,
ical properties are compared with the virgin material, it 4.4 9 104, and 1.8 9 103 s1. Serrated yielding was
is seen that there is a significant increase in strength observed between 523 K and 873 K (250 C and
(both YS and UTS) with corresponding decrease in 600 C) in all three sections for all the strain rates
ductility and impact toughness of the service-aged employed. At temperatures up to 473 K (200 C), as
materials.[2] When the service exposed materials were well as at 873 K (600 C) [923 K (650 C) for the T
tensile tested over the temperature range of RT to 973 K samples] and above the flow curves did not exhibit
serrations for all the strain rates employed. Typical true
stress–strain curves for the service exposed Alloy 625 M
Table II. Average Sizes of c¢¢ and Ni2(Cr, Mo) Particles sample deformed with a strain rate of 8.8 9 105 at
Measured at Different Sections of the Cracker Tube 623 K and 873 K (350 C and 600 C) are shown in
Figures 1(a) and (b), respectively. Three types of serra-
Sample c¢¢ Particles (nm) Ni2(Cr,Mo) Particles tions could be distinguished namely Type A, Type B,
and Type C depending on the temperature and strain
Top 6.7 14.8
Middle 39.7 — rate combination. Figure 2 exhibits a typical variation
Bottom 225 — of serrations with test temperature in the M specimen
for the strain rate of 8 9 105 s1. It is quite evident

Table III. Precipitation of Various Phases Formed During Service in Alloy 625 at Different Locations of the Cracker Tube

Formation Site and Type of Precipitate

Sample Within Matrix At Grain Boundary


Top fine c¢¢ and Ni2(Cr, Mo) M23C6 type carbides: very few in volume and discrete distribution
Middle large c¢¢ particles only large chunky M23C6 type carbides: almost continuous distribution
Bottom very large c¢¢ + d particles continuous film of M23C6 carbide + d phase particles

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, MARCH 2015—1099


of recovery mechanisms. These observations are clearly
revealed in Figure 3 where the variations of YS and
UTS with temperature and strain rate are shown for the
three samples with that of the M specimen in the
serrated flow regime given in magnified scale in the inset.
Further it was seen that UTS decreased with strain rate
at the temperature regime 523 K to 873 K (250 C to
600 C). At all strain rates, while the T samples recorded
the highest strength, the B one exhibited the lowest
strength value. The variations of both elongation to
fracture and elongation to UTS with temperature were
also recorded and the latter is depicted for all the three
types of samples deformed at a strain rate of
8.8 9 105 s1 in Figure 4. Unlike the dip in ductility
associated with DSA[22] for example in ferritic steels, the
ductility in service exposed Alloy 625 increased mark-
edly with temperature. Similar observations have been
made by Kim and Chaturvedi[21] in Alloy 625 and
Gopinath et al. in 720Li.[17] The T samples maintained
highest ductility as compared to the B and the M
samples throughout the temperature regime considered
here. Since the ductility is controlled by work hardening
behavior or relative magnitude of the difference (UTS-
YS), the variation of average work hardening with
temperature was monitored for all the samples. A
typical variation of average work hardening with
Fig. 1—A typical true stress–strain curve for the M sample deformed
temperature for the T, M, and B samples are shown in
with a strain rate of 8.8 9 105 s1 at (a) 623 K (350 C) and (b) the inset of Figure 4 to understand the variation of
873 K (600 C). ductility in the serrated flow regime. Two distinct peaks
were observed in the work hardening behavior at the
from the figure that the nature of the serrations changes start and end of the DSA regime. However, it is seen
with change in test temperature. At the lowest temper- that within the DSA regime average work hardening
ature of 523 K (250 C), the serrations are of type A in rate do not vary significantly.
nature which is characterized by an abrupt rise in stress The SRS was also estimated for the material by strain
value over the general trend of stress–strain curve. At a rate jump tests following the equation:[17]
higher temperature of 623 K (350 C), a combination of 
type A and type B serrations are observed. Apart from ln r1=r2
m¼  ;
sudden rise and drop to the general trend of flow curve ln e_ 1=e_ 2
(type A), oscillations about the general trend of stress–
strain curve are also observed, which is identified as type where the stress level changes from r1 to r2 during a
B serrations. On increasing the temperature further, type strain rate jump test from strain rate e_ 1 to e_ 2 for a given
C serrations are found in both 723 K and 773 K (450 C strain and temperature. A typical variation of SRS with
and 500 C). These types of serrations are characterized temperature for the M sample is shown in Figure 5.
as stress drops below the general trend of flow curve
with much higher drop magnitude compared to the type
A and type B serrations. Serrations appeared only after
a finite amount of plastic strain (critical strain). The B. Serrated Flow and Mechanism of DSA
details of serrated flow will be discussed in a section to The investigation of the mechanism of DSA in any alloy
follow. system has drawn huge attention since it enables to identify
the solute responsible for serrated yielding at different test
conditions and further prediction of degradation of
A. Effect of Strain Rate and Temperature on Tensile
mechanical properties can be done by knowing the reaction
Properties
of the solute species with other phases present in the alloy.
The variations of YS and UTS and their strain rate The examination of the serrated flow curves for Alloy 625
dependence on temperature are given in Figure 3. The revealed that the height of serrations (amplitude and
trends of the variations are also shown by the dotted line frequency) increased with increasing temperature and
for all the three conditions. It is seen that in the decreasing strain rate. The types of serrations were also
temperature range of 498 K to 873 K (225 C to observed to be changing with e_ and T combination
600 C), while YS of all the sections exhibit a plateau (Table IV). At the lowest strain rate of 8.8 9 105 s1, in
or small peak, UTS values show a little decreasing trend. the M and the B samples type (A + B) serrations were
At temperatures higher than this range both YS and observed at 573 K (300 C) and below and type C
UTS show a rapid decreasing trend indicating operation serrations were observed above 573 K (300 C). However,

1100—VOLUME 46A, MARCH 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A


Fig. 2—A typical representation of different types of serrations at different test temperatures for the M sample deformed with a strain rate of
8.8 9 105 s1.

in the T samples, the type (A + B) serrations prevailed up and mobile dislocation density (qm) with plastic strain
to a temperature of 623 K (350 C) and above that (Cv  em and qm  eb), K is a constant, Q is the
temperature was observed to be transformed to type C in activation energy, R is the universal gas constant, and
nature. At the intermediate strain rate of 4.4 9 104 s1, T is the absolute temperature. It is evident from the
type (A + B) serrations were observed up to 723 K above equation that (m + b) can be estimated from the
(450 C) for the T samples, where as, for the M and the B slope of the plot ln e_ vs lnec at a constant temperature.
specimens it was up to 623 K (350 C). Above the Finally, Q can be evaluated following the equation from
aforementioned temperature limit, the serration type the slope of a plot of ln ec vs 1/T at constant e_ as
changed to type C. Finally, for the highest strain rate of Q = slope 9 (m + b) 9 R. AE can also be estimated
1.8 9 103 s1, type C serration took place of the type by another method, called the stress drop method.[17,26]
(A + B) serration above the temperature of 823 K It has been found that the stress drop serves as a better
(550 C) for the T samples, whereas the M and the B criterion for studying strain aging compared to the
samples exhibited type C serrations at and above a critical strain as critical strain method seems to be
temperature of 823 K (450 C) for this strain rate. suitable only for very simple alloy systems.[27] Dr is
Previous investigations in the field of DSA suggest defined as the difference of stress level for breakaway of
that the evaluation of critical strain (ec) and its varia- dislocations from their atmosphere and at which they
tions with strain rate and temperature lead to the glide through the lattice. In stress–strain curve, Dr is
estimation of the activation energy (AE) of the DSA and simply the height of individual serrations. In this
hence, carry the signature of the underlying mechanism procedure, Dr at a given strain was calculated for a
of DSA. According to the formulation of McCor- range of e_ and T. A constant value of Dr was then
mick,[10] the much used[23–26] dependence of ec on strain selected and corresponding e_ was determined for each
rate and temperature can be expressed as temperature. Finally, these e_ were plotted as a function
of 1/T and Q was estimated from its slope as
emþb
c ¼ K_e expðQ=RTÞ; Q ¼ ðln e_ RTÞ.
The critical plastic strain (ec) was estimated from the
where m and b are the respective exponents in the
load-elongation curves as the minimum value of the true
relations for the variation of vacancy concentration (Cv)

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, MARCH 2015—1101


Fig. 3—The variation of YS and UTS with temperature at different strain rate for (a) T, (b) M and (c) B samples. Inset shows variation of YS
and UTS for M sample within serration regime.

plastic strain at which a perceptible load drop of 5 N values estimated from the test data suggested that the
occurred. It was observed that within the strain rate mean values of the (m + b) for the normal PLC
regime of type (A + B) serrations, ec decreased with dynamics were ~1 for all the specimens, whereas those
increasing temperature suggesting the presence of the values in the inverse PLC regime were between 2.0 and
normal PLC effect.[15] At lower strain rate regime where 3.0. AE values were found to be much higher in the
the serrations were of type C in nature, the scenario was inverse PLC regime as compared to the normal PLC
quite the opposite as ec exhibited an increasing trend dynamics. The average AE in the normal PLC regime
with increasing temperature which is attributed to the for the T, M, and B samples were estimated to be 59, 98,
inverse PLC effect.[15,28] In this regime, the serrations are and 72 kJ/mol, respectively. The corresponding AE
generally observed as discontinuous load drops below values at the inverse regime were 90, 181, and 180 kJ/
the general stress–strain curve, in contrast to the locking mol, respectively. The AE derived from the stress drop
type serrations associated with the normal PLC effect at method also yielded similar values and a comparison of
higher strain rates and lower temperatures. However, AE values estimated from both the methods for the M
the origin of the inverse PLC dynamics is not well sample is listed in Table VI. Figure 7 exhibits the typical
understood but it is often attributed to the precipitation plots for the estimation of AE by stress drop method for
before or during tests and depletion of solutes in the the M samples. The AE values derived for the normal
matrix. and the inverse behavior through stress drop method
The values of (m + b) and subsequently the Q values were 87 and 140 kJ/mol, respectively.
were estimated following the method outlined above. A
typical plot for the estimation of (m + b) and corre-
C. Microstructure Evolution During DSA
sponding AE for the M samples are shown in
Figures 6(a) and (b), respectively. In all the three Detailed TEM investigations, as already reported in
samples, the slope changed from positive to negative our previous study,[2] showed presence of c¢¢ and Ni2
with increasing temperature and a transition tempera- (Cr, Mo) precipitates in the T sample with an average
ture could be identified where PLC dynamics switches size of 6.7 and 14.8 nm, respectively (Table II). TEM
from the normal to the inverse one. The average values investigations on the T sample from the PLC regime
of (m + b) and the corresponding Q values at both the showed high density of c¢¢ and Ni2(Cr, Mo) precipitates
normal and the inverse regime of PLC dynamics were and dislocations. As the size of the precipitates is small
represented in Table V. The typical variation of transi- and they maintain coherency with the matrix, they are
tion temperature in different samples deformed at a expected to get sheared by dislocations in the PLC
strain rate of 1.8 9 103 s1 is also listed in Table V. It regime.[13] Figure 8(a) shows the bright field TEM
is evident that the transition temperature did not differ micrograph showing interaction of dislocations with
much between the M and the B samples, where as it was precipitates and tangled dislocations. The dislocation
markedly higher in the T samples. The average (m + b) pairs and shearing of precipitates were also observed as

1102—VOLUME 46A, MARCH 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A


573 to 873 (300 to 600)
623 to 873 (350 to 600)
673 to 873 (400 to 600)
Type C
B
Identification of Strain Rate and Temperature Domain for Different Types of Serrations in T, M, and B Samples

523 to 573 (250 to 300)


523 to 623 (250 to 350)
523 to 673 (250 to 400)
Type (A + B)
Fig. 4—A typical variation of total elongation with temperature for
the samples deformed at a strain rate of 8.8 9 105 s1. The inset

Temperature Regime of Different Serrations [K (C)]


shows the variation of work hardening with temperature for the

573 to 873 (300 to 600)


623 to 873 (350 to 600)
673 to 873 (400 to 600)
three types of sample.

Type C
M

523 to 573 (250 to 300)


523 to 623 (250 to 350)
523 to 673 (250 to 400)
Type (A + B)
Fig. 5—A typical variation of strain rate sensitivity with temperature

723 to 873 (450 to 600)


823 to 873 (550 to 600)
623 to 873(350 to 600)
for the M sample. SRS acquires a negative value over the tempera-
ture regime of DSA.
Type C

shown by arrows in Figure 8(b). TEM investigations of


the T sample from the inverse PLC regime, however,
showed presence of high density of shear bands and
interaction between shear bands (Figure 8(c)). In the
present study, in addition to shear bands, twins–band
interactions were also found to take place as shown in
T

523 to 623 (250 to 350)


523 to 723 (250 to 450)
523 to 823 (250 to 550)

Figure 8(d) with inset showing SAD from twinned


region. It is well known that formation of twin is
Type (A + B)

accompanied by shear strain and this will increase the


Table IV.

plasticity of material if plenty of twins have been formed


during DSA. The twin and shear bands were found to be
almost perpendicular to each other which indicate that
the orientation with respect to loading direction influ-
ence the formation of twin.
TEM investigation of normal PLC regime in the M
and the B sample showed similar microstructure. As
Strain Rate (s )
1

already reported in,[2] there was a dissolution of Ni2(Cr,


Mo) precipitates and growth in c¢¢ precipitate, due to
8.8 9 105
4.4 9 104
1.8 9 103

high exposure temperature in the M and the B sample.


Detailed microstructural investigations in PLC regime
in the M and the B samples showed the presence of shear

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, MARCH 2015—1103


Fig. 6—Typical plot for the (a) estimation of the (m+b) values from the ln e_ vs ln ec plots and (b) evaluation of the activation energies from the
ln ec vs I/T plot for M samples.

Table V. Average Value of (m + b) and Activation Energy Derived for T, M, and B Samples

Activation Energy
Average (m + b) (range) (kJ/mol)
Transition
Sample Normal Inverse Normal Inverse Temperature (C)
T 1.01 (0.92 to 1.11) 2.58 (1.95 to 2.95) 59 90 550
M 1.07 (0.85 to 1.36) 2.62 (2.36 to 2.81) 98 181 371
B 0.90 (0.81 to 0.96) 2.23 (1.97 to 2.47) 72 180 394
Typical variation of transition temperature for normal to inverse PLC dynamics for the samples deformed at 1.8 9 103 s1 is also shown.

Table VI. Comparison of Activation Energies Derived from Critical Strain Criterion and Stress Drop Method for the ‘Middle’
Sample

Activation Energy (kJ/mol)

Types of Serration From ln ec vs 1:T Plots Using Average m + b Values Stress Drop Method
Type A and B 98 87
Type C 181 140

bands in addition to a high density of dislocations in the IV. DISCUSSION


vicinity of these shear bands (Figure 8(e)). However,
TEM investigations in the inverse PLC regime of the M The distinct differences in several aspects of DSA, like
and the B samples showed high density of deformation transition temperature and activation energy which
twins and twin–twin interactions. Figure 8(f) shows could be readily identified among the T, M, and B
cross twinning in the B samples with inset showing SAD samples appeared to be a reflection of the differences in
from twins. The high density of deformation twins their initial microstructures and alloy chemistry. Cha-
appeared both in the M and the B sample despite the kravartty et al.[2] carried out detail investigation on the
high temperature of inverse PLC regime and this could microstructures of service exposed Alloy 625 along the
be due to the instant higher local stress peaks caused by length of the cracker tube and their findings are
the DSA process. The sample from the T portion in the summarized in Table II and III. The value of (m + b)
inverse PLC regime predominantly showed formation is supposed to carry the signature of the mechanism of
and propagation of shear bands. Whereas the B and the DSA. It is well established in literature that serrated flow
M samples showed formation of deformation twins and resulting from the strain aging is attributed to the
twin–twin interactions as the dominant mechanism in interstitial solutes when the value of (m + b) lies
the inverse PLC regime. These results indicated that between 0.5 and 1.0 and to substitutional solutes when
deformation in the T, M, and B sample is dependent on it lies in the range of 2.0 to 3.0.[26] The observed low
temperature—strain rate combination along with the values of activation energy for type (A + B) serration in
prehistory of these samples. the normal PLC regime along with small value of

1104—VOLUME 46A, MARCH 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A


PLC regime showed presence of high density of shear
bands and interaction between shear bands (see Fig-
ure 8(c)). Interaction of these multidirectional shear
bands is the signature of localized plastic deformation.
At the inverse PLC regime, the dissolution of Ni2(Cr,
Mo) precipitates take place and Cr and Mo atoms go
into the solution. The observed AE (~90 kJ/mol) value
of the inverse PLC dynamics in the T samples is found
to be in close agreement with the AE for Mo atom
migration (~100 kJ/mol) in Ni matrix. Kim et al.[21] and
Vani Shankar et al.[1] have also reported the AE for Mo
migration in Ni matrix to be ~100 kJ/mol which
matched very well with the results obtained for the
inverse PLC dynamics in T samples in the present study.
Thus, the presence of Mo atoms in the solution is
responsible for the serrated behavior in the inverse PLC
regime. Though Ni has a high stacking fault energy
(SFE) ~150 mJ/m2, presence of solutes like Cr, Mo etc.
brings down the SFE of c¢¢ (25 mJ/m2)[32] considerably.
Singh et al.[33] have reported that low SFE promote
planar slip and thereby promoting shear band forma-
tion. The presence of Mo atoms within shear band acts
as pinning agent for dislocations and thereby increases
local stress concentration. When the stress concentra-
tion reaches a critical value, propagation of shear bands
and initiation of formation of shear bands in adjacent
region take place. Thus, deformation in the inverse PLC
regime in the T sample proceeds by the initiation and
propagation of shear bands. It is worth pointing out
that the dislocation density in these shear bands were
Fig. 7—Plots for the determination of the activation energy by stress
drop method: (a) Variation of average stress drop Dr with strain found to be very high. Sundararaman et al.[13] have also
rate e_ , (b) ln e_ vs I/T plot using intercepts from (a). reported the presence of shear bands along with twins in
alloy 718 which is deformed to ~19 pct strains. The
(m + b) (<1.0) suggested that the interstitial carbon deformation strains reported in the present study is,
atom might be playing the crucial role in dislocation however, much higher (~70 pct) and could be the reason
locking here. The average value of activation energy was for higher density of shear bands observed. An atomistic
found to be quite low for normal PLC effect. These analysis by Qu et al.[34] have reported that shear band
values were close to 69 kJ/mol, the activation energy for formation and the dislocation structure within these
diffusion of C in Ni.[29] Keh and Nakada[30] also found bands is rather complex, containing twinning partial
similar value of AE (62 kJ/mol) for normal PLC dislocations, full dislocations, and even stair-rod dislo-
dynamics in Ni-C Alloys which they attributed to the cations. It has also been reported that as the deforma-
core diffusion of carbon atoms. Thus, interstitial carbon tion proceeds, the dislocation density in these shear
atoms could be responsible for the normal PLC behav- bands increases to a critical value and as a result, further
ior in the T sample. TEM investigations for the normal deformation occurs by movement of these bands.
PLC regime showed dislocation–dislocation interaction The microstructural study in the inverse PLC regime
and shearing of precipitates in the T samples. In case of for the M and the B samples showed high density of
the inverse PLC dynamics, where the serration was of deformation twins. In this regard, it is worth mentioning
type C in nature, the corresponding (m + b) values the work reported by Sundararaman et al.[13] where they
(which lies between 2.0 and 3.0) suggested the role of identified a critical size (~10 nm) of the c¢¢ precipitate in
substitutional solutes in locking the dislocations. The Ni-based super alloys below which deformation would
composition of Alloy 625 showed that the major be enabled by the progress of dislocations while with
substitutional elements present were Cr, Mo, Fe, and larger precipitate size it would be governed by shearing
Nb. Since, Cr and Fe are similar in atomic size to Ni; of precipitates through the passage of deformation
they were not expected to cause serrated flow. The twins. This effect can be assumed to be similar to the
activation energy of Cr and Fe or self diffusion of Ni is twin induced plasticity (TWIP) and probably be related
~250 kJ/mol or higher.[26] Hence, the possible solute to the improved effect of ductility. Singh et al.[33] have
forming the Cottrell atmosphere around the dislocations also pointed out the role of deformation twins in service
should be Mo or Nb. Since Nb is known to be a slow exposed Ni-based super alloys. In this work, it has been
diffusing solute in Ni (diffusivity ~1.04 9 106 m2/s, AE observed that the Q values for both the normal and the
~202 kJ/mol),[31] the substitutional element responsible inverse PLC dynamics in the M and the B samples were
for the DSA in this alloy might be Mo.[1] Microstruc- much higher compared to the AE in the T samples. This
tural investigations of the T sample from the inverse difference in deformation mechanism is possibly due to

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, MARCH 2015—1105


might have induced the dissolution of Ni2(Cr, Mo) at a
lower temperature. Hence, for the T samples, the
depletion of Mo atoms was delayed further in compar-
ison to the M and the B specimens. As a consequence,
the transition from the normal to the inverse PLC
dynamics in the T samples shifted to a relatively higher
temperature than those of the M and B ones. Apart
from DSA, the difference in mechanical properties in the
T, M, and B samples also were related to their different
50 nm 50 nm microstructure. The finely dispersed c¢¢ and Ni2(Cr, Mo)
particles rendered the high strength to the T samples, as
(a) (b)
compared to the M and B ones where only c¢¢ particles
100 nm 200 nm of larger size were present. The usual dip in ductility
during DSA was not observed here. Kim and Chaturv-
edi[21] and Gopinath et al.[17] also reported the absence
of dip in ductility during serrated yielding in nickel-
based super alloys. Kim et al.[21] associated the phe-
nomenon with the longer traveling distance of defor-
mation fronts. In the present work, the plateau observed
in work hardening rate during DSA suggested that the
rate of recovery was quite strong even in the DSA
regime. The lowest ductility and UTS observed in the B
samples can be explained in terms of the presence of
(c) (d) needle-shaped d particles, as reported in our previous
200 nm 200 nm study.[2] These precipitates acted as internal notches at
the grain boundaries which along with the continuous
carbide layer prohibited the slip propagation through
grain boundaries. These conditions led to cracking
locally and thus hindered the attainment of ultimate
strength, resulting in lower UTS and ductility.

V. CONCLUSIONS
The investigation of the Dynamic Strain Aging in
(e) (f) service exposed Alloy 625 revealed that serrated yielding
in the material was confined in a temperature domain of
Fig. 8—BF TEM image showing (a) dislocation precipitate interac- 523 K to 873 K (250 C to 600 C) for the strain rates
tion in normal PLC regime in the T sample; (b) dislocation pairs and considered here. For the T, M, and B samples, the PLC
precipitate shearing in normal PLC regime in the T sample; (c) shear
band interactions in inverse PLC regime in the T sample; (d) shear dynamics exhibited both normal and inverse behavior at
band twin interaction in inverse PLC regime in the T sample with low and high temperature regime, respectively. The
inset showing SADP for twin; (e) dislocation-shear band interaction normal behavior was associated with type (A + B)
in normal PLC regime in the B sample; (f) twin–twin interaction in serrations and interstitial carbon atoms could be
inverse PLC regime in the B sample.
responsible for the normal PLC behavior in the T
sample. TEM investigations for the normal PLC regime
variation in the microstructure of the samples. The high showed dislocation–dislocation interaction and shearing
AE in the B and the M samples could be due to of precipitates in the T samples. On the contrary, the
requirement of high local stress concentration for the serrations were of type C in nature in the inverse PLC
formation and propagation of deformation twins at high regime and were attributed to the locking of dislocations
temperature. Cui et al.[35] related the inverse PLC effect by substitutional Mo atoms and nucleation and prop-
in Ni-Co super alloy with the occurrence of continuous agation of twins and shear bands. The differences in AE
stacking faults within the c¢ precipitates. However, such in the M and the B samples with that of the T samples
stacking faults could not be observed in the present revealed basic difference in the deformation mechanism
study. owing to different microstructure. The transition tem-
The study of the microstructures of Alloy 625 after perature from normal to the inverse PLC dynamics also
service exposure suggested that the T samples contained varied appreciably in the T samples than that of the M
Ni2(Cr, Mo) phases in the matrix which might have and the B ones which could be explained in terms of the
dissolved at high temperature releasing Mo atoms in the delayed depletion of Mo solutes in solution for the T
solution. Vani Shankar et al.[3] reported dissolution of samples. The tensile properties were also investigated in
the Ni2(Cr, Mo) phase at 923 K (650 C). However, in the DSA regime. The SRS took a negative value in the
the present case, the T samples were deformed with DSA regime. Even though YS and UTS showed
simultaneous exposure to elevated temperature which marginal peak in this regime, the ductility did not show

1106—VOLUME 46A, MARCH 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A


the usual dip in serrated yielding in this material, rather 17. K. Gopinath, A.K. Gogia, S.V. Kamat, and U. Ramamurty: Acta
it followed an increasing trend with temperature. Mater., 2009, vol. 57, pp. 1243–53.
18. Q. Zhang, Z. Jiang, H. Jiang, Z. Chen, and X. Wu: Int. J. Plas-
ticity, 2005, vol. 21, pp. 2150–55.
19. P.D. Zavattieri, V. Savic, L.G. Hector, Jr., J.R. Fekete,
W. Tong, and Y. Xuan: Int. J. Plasticity, 2009, vol. 25,
pp. 2298–2330.
20. W. Tong, H. Tao, N. Zhang, and L.G. Hector, Jr.: Scripta Mater.,
REFERENCES 2005, vol. 53, pp. 87–91.
21. S. Kim and M.C. Chaturvedi: Trans. Jpn. Inst. Met., 1987, vol. 28,
1. V. Shankar, M. Valsan, K.B.S. Rao, and S.L. Mannan: Metall. pp. 205–12.
Mater. Trans. A, 2004, vol. 35A, pp. 3129–39. 22. P. Rodriguez: Bull. Mater. Sci., 1984, vol. 6, pp. 653–63.
2. J.K. Chakravartty, J.B. Singh, and M. Sundararaman: Mater. Sci. 23. K.W. Qian and R.E. Reed-Hill: Acta Metall., 1983, vol. 31,
Technol., 2012, vol. 28, pp. 702–10. pp. 87–94.
3. V. Shankar, K.B.S. Shankar, and S.L. Mannan: J. Nucl. Mater., 24. S. Venkadesan, C. Phaniraj, P.V. Sivaprasad, and P. Rodriguez:
2001, vol. 288, pp. 222–32. Acta Metall. Mater., 1992, vol. 40, pp. 569–80.
4. Q. Guo, D. Li, S. Guo, H. Peng, and H. Jie: J. Nucl. Mater., 2011, 25. I.S. Kim and E.O. Hall: Metals Forum, 1978, vol. 1, p. 95.
vol. 414, pp. 440–50. 26. C.L. Hale, W.S. Rolling, and M.L. Weaver: Mater. Sci. Eng. A,
5. V.V. Raman, G.E. Prasad, and P.D. Gupta: Trans. IIM, 1987, 2001, vol. 300, pp. 153–64.
vol. 40, pp. 113–20. 27. R.W. Hayes and W.C. Hayes: Acta Metall., 1982, vol. 30,
6. M. Sundararaman, L. Kumar, G.E. Prasad, P. Mukhopadyay, and pp. 1295–1301.
S. Banerjee: Metall. Mater. Trans. A, 1999, vol. 30A, pp. 41–52. 28. J.M. Robinson and M.P. Shaw: Int. Mater. Rev., 1994, vol. 39,
7. V. Shankar, M. Valsan, K.B.S. Rao, and S.L. Mannan: Scripta pp. 113–22.
Mater., 2001, vol. 44, pp. 2703–11. 29. W.F. Gale and T.C. Totemeier, eds.: Smithells Metals Reference
8. F. Le Chatelier: Rev. Metall., 1909, vol. 6, pp. 914–17. Book, Butterworth-Heinemann, London, 2004, pp. 3–25.
9. A. Portevin and F. Le Chatelier: Trans. Asst., 1924, vol. 5, p. 457. 30. Y. Nakada and A.S. Keh: Acta Metall, 1970, vol. 18, pp. 437–43.
10. P.G. McCormigk: Acta Metall., 1972, vol. 20, pp. 351–54. 31. R.V. Patil and G.B. Kale: J. Nucl. Mater., 1996, vol. 230, pp. 57–
11. E. Pink: Scripta Metall. Mater., 1994, vol. 30, pp. 767–68. 60.
12. S. Banerjee and U. Naik: Acta Mater., 1996, vol. 44, pp. 3667–77. 32. A. Francois, G. Hug, and P. Veyssiere: Philos. Mag. A, 1992,
13. M. Sundararaman, P. Mukhopadhyay, and S. Banerjee: Acta vol. 66, pp. 269–88.
Metall., 1988, vol. 36, pp. 847–64. 33. J.B. Singh, J.K. Chakravartty, and M. Sundararaman: Mater. Sci.
14. B.H. Tian, Y.G. Zhang, and C.C. Chen: Mater. Sci. Eng. A, 1998, Eng. A, 2013, vol. 576, pp. 239–42.
vol. 254, pp. 227–33. 34. S. Qu, H. Zhou, and Z. Huang: Scripta Mater., 2011, vol. 65,
15. K. Chihab and C. Fressengeas: Mater. Sci. Eng. A, 2003, vol. 356, pp. 715–18.
pp. 102–07. 35. C.Y. Cui, Y.F. Gu, Y. Yuan, and H. Harada: Scripta Mater.,
16. R.C. Picu: Acta Mater., 2004, vol. 52, pp. 3447–58. 2011, vol. 64, pp. 502–05.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 46A, MARCH 2015—1107

View publication stats

You might also like