International Journal of Solids and Structures 97–98 (2016) 593–612

Contents lists available at ScienceDirect

International Journal of Solids and Structures

journal homepage: www.elsevier.com/locate/ijsolstr

Strain localization during discontinuous plastic flow at extremely low

temperatures
J. Tabin∗, B. Skoczen, J. Bielski
Institute of Applied Mechanics, Faculty of Mechanical Engineering, Cracow University of Technology, Al. Jana Pawła II 37, 31-864 Cracow, Poland

a r t i c l e i n f o a b s t r a c t

Article history: The phenomenon of strain localization in the course of discontinuous plastic flow (DPF) at extremely
Received 8 January 2016 low temperatures is investigated. DPF is observed mainly in fcc metals and alloys strained in cryogenic
Revised 31 March 2016
conditions, practically down to absolute zero. These materials undergo at low temperatures a process
Available online 7 June 2016
similar to dynamic strain ageing, manifested by the so called serrated yielding (DPF). DPF is attributed
Keywords: to the mechanism of local catastrophic failure of lattice barriers (including Lomer–Cottrell locks), under
Multiscale constitutive model the stress fields related to the accumulating edge dislocations. Failure of LC locks leads to massive mo-
Discontinuous plastic flow tion of released dislocations, accompanied by step-wise increase of the strain rate (macroscopic slip) and
Cryogenic temperatures drastic drop of stress. Recent experiments indicate strong strain localization in the form of shear bands
Strain localization propagating along the sample. The plastic power dissipated in the shear band is partially converted to
heat, which results in a local drastic increase of temperature promoted by the so-called thermodynamic
instability (nearly adiabatic process). The Dirac-like temperature function is measured by two thermome-
ters located in the gage length of the sample. Spatio-temporal correlation indicates smooth shear band
propagation, as long as the process of phase transformation remains on hold. A physically based multi-
axial constitutive model presented in the paper describes both DPF and strain localization, accompanied
by temperature distribution represented by Green-like solution of heat diffusion equation. The model ac-
counts for the thermodynamic background, including phonon mechanism of heat transport, accompanied
by specific heat vanishing with the temperature approaching absolute zero. Experimental identification
of parameters of the constitutive model is carried out. A projection of the model to the range where the
phase transformation takes place is discussed.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction
The phenomenon of strain localization in the course of DPF at
extremely low temperatures has not been addressed in the litera-
ture. On the other hand, a significant effort has been focused on
the mechanism of shear bands initiation and propagation during
the so-called PLC (Portevin-Le Chatelier) effect, observed in cer-
tain materials at room temperature. The reason for such a dramatic
lack of information on the behaviour of materials at extremely low
temperatures is related to testing conditions. The experiments are
carried out inside a double-wall cryostat, which is not transpar-
ent and usually tightly equipped with necessary instruments. Thus,
there is practically no possibility to scan the surface of the sample
by means of such instruments like thermographic camera, in or-
der to detect possible effects of strain localization. In order to do
∗
Corresponding author. Fax: +48123743370.
E-mail address: kubatabin@mech.pk.edu.pl (J. Tabin).
so, specific deduction methods have to be used, which will be ex-
plained in the course of the present paper.
1.1. DPF at extremely low temperatures
Fcc metals and alloys are applied in modern cryogenic instal-
lations (including superconducting magnets or cryogenic transfer
and storage systems) within the whole range of temperatures, from
near-0 K to ambient temperature. Such materials like copper and
its alloys, aluminum alloys or different grades of stainless steel
show truly remarkable properties at extremely low temperatures,
including high ductility (cf. Sgobba and Hochoertler, 1999). To give
an example of typical cryogenic application, Fe-Cr-Ni austenitic
stainless steels are used to manufacture components of supercon-
ducting magnets and cryogenic transfer lines, since they preserve
ductility down to absolute zero. Many of these materials undergo
at near-0 K temperatures (for instance in superfluid helium, 1.9 K,
or in liquid helium, 4.2 K) some sort of dynamic strain ageing,
called DPF. DPF manifests itself at the macroscopic level by oscil-
latory mode of plastic flow, consisting in abrupt drops of stress

http://dx.doi.org/10.1016/j.ijsolstr.2016.06.012
0020-7683/© 2016 Elsevier Ltd. All rights reserved.
594 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612
Fig. 1. Types of plastic flow (serrated versus smooth yielding) as a function of temperature (engineering stress - strain curves).
against strain. Another important feature of DPF consists in the
fact, that it occurs in the temperature range between 0 K and a
material dependent temperature (T0 or T1 ), associated with the
transition from screw to edge dislocations mode. General thermo-
mechanical aspects of DPF are strictly linked to the so-called ther-
modynamic instability, which consists in strong oscillations of tem-
perature due to heat accumulation in near adiabatic conditions,
resulting from vanishing specific heat when the temperature ap-
proaches absolute zero. This effect is accompanied by vanishing
thermal conductivity and vanishing thermal contraction coefficient.
Instabilities occurring at higher temperatures, like the PLC effect,
are usually of different nature and are not addressed in the present
paper.
In order to illustrate the plastic strain induced phenomena that
occur at extremely low temperatures, typical response of selected
alloys (Ni-Ti6Al4V, grade 304 stainless steel), used for structural
applications at cryogenic temperatures, is shown in Fig. 1. In par-
ticular, three domains of response are distinguished: below T1
where the DPF (serrated yielding) takes place, between T1 and Md
where the diffusionless plastic strain induced γ →α ’ phase trans-
formation occurs, and above Md where the smooth plastic flow is
observed.
Generally, DPF represents the so-called oscillatory mode of plas-
tic deformation and is detected in many low (LSFE) and high stack-
ing fault energy (HSFE) materials strained at extremely low tem-
peratures (Obst, Nyilas, 1991). As soon as the serrated yielding oc-
curs, the plastic flow becomes discontinuous in terms of dσ /dɛ.
The DPF takes place below a thermal threshold, characteristic of
given material: T1 for LSFE materials and T0 for HSFE materi-
als (both transition temperatures are material dependent and can
reach some 35 K).
Serrated yielding occurs for a wide range of strain rates (Kom-
nik and Demirski, 1984; Reed and Walsh, 1988; Reed and Simon,
1988; Pustovalov, 2008). When the strain rate exceeds a critical
value, nearly adiabatic accumulation of heat takes place and causes
fast increase of temperature of the sample to the level above T0
or T1 . This leads to restoration of the classical smooth plastic flow
(Fig. 2).
Thus, the DPF is usually bound by the lower and the upper lim-
its of the strain rate. Lower limit does not play a substantial role
and is often close to zero, whereas, the upper limit has an impor-
tant meaning for the way the serrated yielding develops. Typically,
heat transport conditions at the temperature of liquid helium are
close to adiabatic, and enhanced strain rate causes subsequent heat
accumulation and strong temperature rise. For higher strain rates,
the adiabatic overheating leads to gradual increase of temperature
until it reaches the critical level T0 or T1 . As soon as the critical
temperature level is reached, transition from serrated to smooth
plastic flow is observed and DPF temporarily disappears. It is a re-
newed cool down below T0 or T1 that brings the material back to
the oscillatory mode of plastic flow.
1.2. The mechanism of DPF
Recent history of low temperature serrated yielding started
already in the fifties of 20th century. The mechanism of DPF was
analyzed by several authors, among them Basinski, 1957, Schwarz
and Mitchell, 1974, Reed and Simon, 1988, Reed and Walsh, 1988,
Hähner and Zaiser, 1997, Zaiser and Hähner, 1997, Benallal et al.,
2006. Some investigators (cf. Wessel, 1957; Tabachnikova et al.,
1984) postulated that sufficiently high flow stress at extremely low
temperature can induce avalanche-like process of multiplication of
mobile dislocations, which leads to plastic flow instability. Other
authors, among them Estrin and Tangri, 1981, Dolgin and Natsik,
1991, as well as Burns, 1994, developed rather simple models of
DPF. An extensive study of DPF was carried out by Basinski, already
in 1957. Basinski developed a thermodynamic approach, including
the fact that the specific heat and the thermal conductivity tend
to zero with temperature. He based his approach on the adiabatic
heating hypothesis, claiming that any fast dissipative process that
 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612
Fig. 2. Effect of strain rate on activation of serrated yielding for OFE Cu; the strain rate program; transition temperature is equal to 8.5 K (temperature evolution is reflected
by the blue curve). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
occurs at extremely low temperature leads to increase of local
temperature, which results in a drastic decrease of the flow stress.
This early stage hypothesis did not take into account the behaviour
of stress (load) measured directly on the sample. Another expla-
nation of DPF based on creation of lattice barriers for dislocation
motion has been formulated by Obst and Nyilas, 1991, 1998. The
Authors point out that the pile-ups of dislocations on the internal
barriers (Lomer-Cottrell locks, Cottrell atmospheres, Suzuki locks,
groups of interstitial atoms, dislocation forest, non-metallic inclu-
sions etc.) in the lattice may give rise to stress concentrations of
the order of magnitude of theoretical shear strength (cf. Seeger,
1957). In their view, the load drops are due to catastrophic process
consisting in a spontaneous generation of dislocations as soon as
the internal barriers are broken. According to Obst and Nyilas, the
plastic flow instability is of mechanical nature.
Both adiabatic heating hypothesis derived by Basinski, 1957,
and thermo-mechanical approach developed by Obst and Nyilas,
1991, reflect very complex nature of the flow instabilities that oc-
cur at extremely low temperatures.
1.3. Low temperature DPF versus PLC phenomenon
It turns out that the appearance of DPF is quite similar to the
appearance of the so-called Portevin – Le Chatelier (PLC) (Yang,
Tong, 2006; Qi Hu et al., 2012b; Manach et al., 2014; Klusemann et
al., 2015) effect, i.e. the abrupt stress oscillations as a function of
time or strain during a displacement controlled traction test. Zaiser
and Hähner, 1997, attribute discontinuous nature of plastic flow
at cryogenic temperatures to the strain rate softening instabilities,
and point out similarity between the low temperature DPF and the
ambient temperature PLC. Strain rate sensitivity (SRS) consists, in
their view, of positive instantaneous response of the flow stress to
a sudden increase of the strain rate. It is followed by stress relax-
ation to a steady state asymptotic value. Yet, the physical mecha-
nisms of DPF and PLC are rather different, and their origins seem
not to be the same (c.f. Obst, Nyilas, 1991). Judging by the appear-
ances, the interaction of dislocations with the barriers at extremely
low temperatures can be qualified as analogical response to PLC.
Both in the case of PLC and DPF, dislocations meet obstacles and
are temporarily arrested/locked. However, in the course of PLC, so-
lutes diffuse around the dislocations and further strengthening of
595
the obstacles is observed (Bross et al., 2003), whereas, in the case
of DPF the obstacles remain stable because of weakly exited lattice
and lack of thermal energy. Eventually, these dislocations will in
both cases overcome the obstacles reaching sufficient level of lo-
cal resolved shear stress, and will move to the next obstacle where
they are again stopped. The difference between both mechanisms
consists in the fact that in the case of PLC the obstacles are created
by diffusion, whereas, in the case of DPF the obstacles are “diffu-
sionless”.
It is worth pointing out, that the DPF has not been sufficiently
investigated when compared to the PLC effect (Rizzi and Haehner,
20 04; Benallal et al., 20 08; Zavattieri et al., 20 09; and many oth-
ers). In particular, studies of spatio-temporal dynamics of the low-
temperature plastic oscillations are greatly insufficient, both at the
experimental and at the constitutive as well as at the numerical
levels (Zaiser and Hähner, 1997). Therefore, an experimental iden-
tification of the strain localization during DPF within the near-0 K
temperature range has been carried out and is presented (Section
3) in the current paper.
1.4. Multiscale constitutive models of DPF
Multiscale modelling in the domain of plasticity (single and
polycrystals) gained recently a lot of attention. One of the crucial
problems in constitutive modelling of modern materials consists
in the fact that the model should be physically based. An inter-
esting approach, based on physical understanding of evolution of
dislocation structure during strain path changes, was presented by
F. Barlat et al., 2012. In particular, a new concept of “distortional”
hardening has been developed. A combination of multiscale mod-
elling (micro-meso-macro) and physically based approach offers a
chance to build a universal description, valid under different load-
ing conditions.
Plasticity in aluminum single crystals was investigated by Groh
et al., 2009. Multiscale modelling methodology involving two
bridges over three different length scales was developed. Bridging
approach between the nano-, the micro- and the meso-scales was
built. The molecular dynamics simulations at the nanoscale were
performed in order to generate motion of dislocations. A discrete
dislocations numerical tool at the microscale was used and the
work hardening of fcc crystals has been described. Next bridge
596 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612
involved material parameters related to slip system hardening
law, employed in the crystal plasticity models at the mesoscale.
The response of aluminum single crystals subjected to uniaxial
compressive loads was predicted. Multiscale constitutive modelling
of induced plastic anisotropy in bcc metals has been developed
by Hamelin et al., 2011. New framework to predict variation in
local plastic anisotropy due to the crystallographic texture in
bcc polycrystals was developed. The multiscale model has been
developed in order to examine the contribution of microscopic
and mesoscopic effects to the macroscopic response of bcc metals.
A dislocation-based hardening scheme and a Taylor-based crystal
plasticity formulation has been implemented.
Uniaxial model of DPF, including thermo-mechanical nature of
this phenomenon and the relevant thermodynamic background,
has been developed by Skoczen et al., 2010. This model has been
generalized to a multiaxial representation in 2014. In the present
extended and modified constitutive model of discontinuous plastic
flow (DPF) at extremely low temperatures, the principle of physi-
cally based modelling has been incorporated in parallel with cou-
pled thermo-mechanical approach, crucial in view of thermody-
namic instability that occurs in the range of temperatures close to
absolute zero.
1.5. Thermodynamic aspects of strain localization
Materials in which the plastic instability (DPF, PLC or even
Lüders band propagation) occurs exhibit a characteristic oscilla-
tion of stress as a function of strain during kinematically controlled
loading. The plastic deformation takes place locally, within defor-
mation bands, often termed slip bands. These strain localization
effects in a loaded specimen are often associated with heat dis-
sipation and rapid change of temperature (see Fig. 5). Conversion
of plastic power into heat has been investigated in a wide range of
materials by many authors.
For this type of problem, two fundamentally different model-
ing approaches are postulated: models based on thermodynamic
considerations (Chrysochoos, Louche, 20 0 0; Dumoulin et al., 2010;
Klusemann et al., 2015) or models based on the so-called Taylor–
Quinney factor (Taylor and Quinney 1934; Zaera et al. 2013; Bartels
et al., 2015). It is worth mentioning, that the thermo-mechanical
deformation effects are analyzed on the atomic level, in a wide
range of materials. Chowdhury et al. (2015) have been investigat-
ing twinning and dislocation slip associated with plastic flow in
fcc alloy (Co-33%Ni). Moreover, the Authors indicated that the en-
tire surface energy γ is a material parameter, determining whether
the plastic flow is slip or twinning dominated. The model accounts
for the energy balance between the applied work and total energy
expenses as contributed by the atomistic effect related to deforma-
tion. The results of experimental investigation of deformed struc-
ture by means of digital image correlation (DIC) or transmission
electron microscopy, were presented and discussed.
Regarding the strain localization and plastic instability at ultra-
low temperatures, an interesting analysis, leading to better under-
standing of the mechanism of DPF, can be found in Vorobev and
Anpilogova, 2013, 2015. The energy balance and the mechanism of
plastic flow instability at extremely low temperatures in metals are
discussed by the Authors. In particular, the energy of strain jump
has been analyzed. It has been shown that most of the poten-
tial strain energy is converted to work (Vorobev and Anpilogova,
2015). The investigation also shows that the strain jump depends
significantly on the parameters of the loading system, as well as
on the level of material strength (Vorobev and Anpilogova, 2013).
Some recent work on the development of jump-like deformation
at 0.5 K for the polycrystalline samples (Al-Li alloy) with different
microstructure has been carried out by Shumilin et al., 2015.
For what concerns the experimental identification of thermody-
namic aspects of strain localization, there exist experimental meth-
ods to trace the deformation bands associated with heat dissi-
pation. The most popular are imaging techniques: digital image
correlation (DIC) and infrared thermography (IRT). The first tech-
nique produces displacement fields from which strains are derived,
while the second provides temperature fields with which the heat
sources are estimated using a local form of the heat equation
(Chrysochoos et al., 2009). However, such methods can be used
only for narrow range of temperatures, close to room tempera-
ture. Therefore, thermodynamic aspects of strain localization at ex-
tremely low temperatures, near-0 K, have been poorly recognized
until now.
2. RVE based constitutive model of DPF
2.1. Kinetics of serrated yielding (DPF)
Low temperature DPF phenomenon in OFE Copper is illustrated
in Fig. 3. It shows oscillations of stress against strain, measured
inside and outside the cryostat. Also, good correlation of stress and
temperature oscillations has been demonstrated, which proves that
the phenomenon is definitely of thermo-mechanical nature.
From the phenomenological point of view, serrated yielding
consists in frequent abrupt drops of stress as a function of strain
during kinematically controlled loading (Fig. 4). There is enough
experimental evidence to back the hypothesis that the mechanism
of DPF is related to formation of dislocation pile-ups at the in-
ternal lattice barriers, such as the Lomer-Cottrell locks (Obst and
Nyilas, 1991; Obst and Nyilas, 1998; Pustovalov, 2008). The micro-
scopic analysis clearly shows that the back stresses of dislocation
pile-ups substantially hinder the motion of other dislocations. Ac-
cumulation of dislocations in the pile-up leads to gradual increase
of the resolved shear stress at the head of pile-up, until the stress
reaches the level of cohesive strength. As soon as the internal lat-
tice barriers break, liberated glissile dislocations can freely glide
away. Collective and massive character of this phenomenon leads
to sudden drop of stress and further heat production as a result of
plastic power dissipation.
At extremely low temperatures the lattice excitation is very
weak and rather high stresses are observed. The avalanche-like
barrier crossing by dislocation pile-ups is often detected by acous-
tic effects of “dry” sounds, emitted by the specimen.
Typical stress - strain curve for a material that exhibits DPF
(stainless steel 316LN) is illustrated in Fig. 5. Moreover, decompo-
sition of single serration into the strain, stress and temperature re-
sponse as a function of time is shown in the same figure.
In order to deal with production of lattice barriers (locks) and
formation of dislocation pile-ups, a parameter B reflecting the
number of barriers (and pile-ups) per unit volume was introduced.
It is assumed that the production of internal barriers (e.g. the
Lomer-Cottrell locks) is proportional to the accumulated plastic
strain. This means that for monotonic or cyclic loading the number
of created barriers (locks) becomes a function of history of the pro-
cess, represented by the Odqvist parameter. One of possible barri-
ers, the Lomer-Cottrell lock, occurs when two leading Shockley par-
tials combine. They form separate dislocation with the Burgers vec-
tor out of the slip planes. The L-C dislocation (lock) is sessile and
forms a barrier for other dislocations that pile up against it. This
phenomenon is even more pronounced at extremely low tempera-
tures, where the lattice is weakly excited (lack of thermal energy in
the lattice). The number density of barriers is strictly related to the
amount of plastic deformation, which is usually measured by the
plastic strain intensity. To account for the history of the process,
 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612 597

Fig. 3. DPF in OFE-Cu including thermo-mechanical coupling; red and green curves correspond to the load cells located outside and inside the cryostat, respectively; blue
curve denotes the temperature (engineering stress - strain curves). (For interpretation of the references to colour in this figure legend, the reader is referred to the web
version of this article.)

Fig. 4. The mechanism of DPF: massive collapse of barriers at the heads of dislocation pile-ups.

the accumulated plastic strain (Odqvist parameter p) is applied.
 
t˜
2 p
p= ε˙ : ε˙ p dt
t0 3 the model consists in defining the drop of stress in terms of the
Such approach is convenient from the point of view of general-
ization to all possible sources of locks and barriers that depend di-
rectly on the macroscopic strain. Hence, it has been assumed that
the increment of B is strictly related to the increment of the ac-
cumulated plastic strain dp. In view of the fact that increasing in-
tensity of plastic flow generates more barriers for the motion of
dislocations, the following evolution law for the number of lattice
barriers (locks) per unit surface (number density) is applied:
 
+
B˙ = FLC ρ , T , σ p˙ H ( p − pLC )
where B˙ denotes rate of the number of barriers per unit surface,
+
FLC is in general a function of dislocations density ρ , temperature
T and stress σ , whereas, pLC represents the threshold above which
the lattice barriers massively develop and H(..) denotes the Heavi-
side function.
2.2. New features of RVE based constitutive model of serrated
yielding (DPF)
Multiscale constitutive model involves microscopic approach
based on the analysis of evolution of dislocations density (Mecking,
Kocks, 1981; Kocks, Mecking, 2003) and formation of dislocation
pile-ups at the lattice barriers. Moreover, average resolved shear
stress in the lattice and average shear stress at the head of dis-
location pile-up are included. The process of drop of stress dur-
ing single serration is based on a criterion involving number den-
sity of lattice barriers accompanied by dislocation pile-ups and the
(1) shear stress at the head of a pile-up. A completely new element of
evolution of dislocation density rather than using an artificial phe-
nomenological function. The model is defined by means of meso-
scopic representative volume element (RVE), containing the dislo-
cation pile-ups at the lattice barriers (Fig. 6).
It is assumed that the production of lattice barriers (e.g. the
Lomer-Cottrell locks) is proportional to the accumulated plastic
strain. This means that for a monotonic or cyclic loading the num-
ber of created barriers (locks) becomes a function of history of the
process, represented by the increment of the accumulated plastic
(2) strain (Odqvist parameter):

2 p
dp = dε : dεp (3)
3
Thus, for isothermal process and small variation of the flow
stress, the increment of B is strictly related to the increment of
accumulated plastic strain (Odqvist parameter):
+
dB = FLC dp (4)
The average shear stress in the lattice is represented by two
terms: the lattice internal friction represented by the shear stress
τ 0 , and the interaction between the dislocations (Bouquerel, et al,
2006):
√
τ = τ0 + μα b ρ (5)
where μ is the shear modulus, α denotes the coefficient of dislo-
cations interaction, b is the length of the Burgers vector and ρ is
598 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612

Fig. 5. Serrated yielding in fcc metals at 4.2 K: stress - strain response and decomposition of single serration into strain, stress and temperature as a function of time.

increase of the strain rate. Finally, load drops characteristic of ser-

rated yielding are observed and recorded.
The following criterion of avalanche-like failure of barriers (cf.
Skoczen et al., 2010) is applied (Fig. 7a):
B = Bcr f or τ ≤ τmin
τe = τcr f or B ≤ Bmin
B − Bmin τe − τcr
= f or τ > τmin ∧ B > Bmin (8)
Bcr − Bmin τmin − τcr
The core of this criterion is based on the interaction between B
and τ e . A combination of both parameters triggers serration (drop
of stress). Limits Bmin , τ min were introduced since for sufficiently
large value of one of them, the other is insignificant for the phe-
nomenon to occur. The following relations were assumed for sim-
plicity:
Fig. 6. RVE containing the pile-ups of dislocations at the lattice barriers.
Bmin = αB Bcr ; τmin = ατ τcr (9)
where α B , α τ determine relevant fractions of critical values (for
the dislocations density. Based on the average shear stress in the
instance 0.2 and 0.1).
lattice τ , the shear stress at the head of dislocation pile-up is com-
Thus, as soon as combination of number density of lattice bar-
puted (Hull, Bacon, 2011):
riers containing dislocation pile-ups with average shear stress at
π (1 − ν )λ̄τ 2 the head of dislocation pile-up reaches the interaction curve, the
τe = (6) avalanche-like process is triggered.
μb
The process of massive failure of barriers results in an instanta-
where υ is the Poisson ratio. The mean free path of dislocation λ̄ neous decrease of stress (in uniaxial case) σ and abrupt increase
reads: of plastic strain ɛp , that can be interpreted as the amount of slip
1 1 √ √ during the catastrophic failure of dislocation barriers (Fig. 7b).
= + k1 ρ + k2 B (7) In the previous model, drop of stress was determined in an ar-
λ̄ d
tificial way by means of purely phenomenological approach, rep-
where d denotes the average grain size, and k1 , k2 are constants. resented by the function F − LC
. In the present model, entirely new
Here, the relation used by Bouquerel et al., 2006, has been gener- approach based on the evolution of dislocations density is used.
alized so as to include barriers density as a representation of addi- Namely, drop of stress is obtained by means of Eq. (5) in the fol-
tional obstacles for dislocations, along with the grains and disloca- lowing form:
tions themselves. √ √
It is further assumed, that a catastrophic failure of lattice bar-
riers occurs, when the stress concentration within dislocation pile-
up is large enough. Such a local failure of barriers triggers conse-
quently avalanche-like process that becomes massive and usually
spreads over a macroscopic fraction of lattice. Moreover, massive
motion of released dislocations implies generation of new disloca-
tions by Frank-Reed sources. This eventually results in a step-wise
σ = Mμα b( ρ − ρ0 ) (10)
where ρ 0 stands for the after-serration density of dislocations, and
Mstands for the Taylor factor. Such definition makes the approach
more physical, and allows for the evolution of lower envelope of
the stress - strain curve. Assuming that the process is kinematically
controlled, and takes place at a constant value of total strain, the
drop of stress results in the relevant strain redistribution, namely
 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612 599

Fig. 7. (a) Criterion of single serration; (b) the amount of slip (ε serr
p ) during the process of catastrophic failure of lattice barriers; (c) discontinuous plastic flow with four
stages of the process: elastic loading (1), plastic flow (2), drop of stress (3), relaxation (4); (d) experimental result for austenitic stainless steel.

in the decrement of the elastic part identical with the increment The experimental research of the PLC effect was focused mostly on
of the plastic part (E is the Young modulus): recording the temporal and the spatial characteristics. Such well
√ √
Mμα b( ρ − ρ0 ) known techniques like: the shadowgraph technique (Chihab et al.,
ε p =
E
Massive failure of barriers is followed by fast motion of glissile
dislocations in the lattice. Heat released in the course of this pro-
cess depends on the plastic work and the internal friction in the
lattice:
 
dQ = η dW p + dW f
where η denotes the Taylor-Quinney coefficient. In order to per-
form rough estimation of possible temperature fluctuations, the
specific heat has to be expressed as a function of temperature:
dQ
dT =
mCV (T )
It turns out that the temperature fluctuations may even reach
some 40–50 K for 316 LN stainless steel. The process of stress re-
laxation, that usually follows the drop of stress, is temperature
driven and described by the following set of equations:
(11) 1987), acoustic emission (Caceres, Rodriguez, 1987; Lazarev, Vino-
gradov, 2009), use of a multi-zone laser extensometer (Ziegenbein
et al., 2001), infrared thermography (Louche et al., 2005), infrared
thermography (Hu et al., 2012a), electronic/digital speckle pattern
interferometry (ESPI or DSPI)) (Zhang et al., 1999; Petit et al., 2013;
Labergere et al., 2014) and digital speckle correlation (DSC) (Zhang
(12) et al., 2005), or digital image correlation (DIC) (Ait-Amokhtar et
al., 2006; Tong and Zhang, 2007; Nielsen et al., 2010), were used.
Similar methods were developed to observe propagation of the
Lüders bands and to evaluate the strain profile near the Lüders
front. The Lüders bands tracking and identification of local defor-
(13) mation is usually carried out by means of the following meth-
ods: recording surface scratches (Hall, 1951; Iricibar et al.,1977),
using optical ‘‘Topler Schlieren’’ method (Hall, 1951; Sylwestrowicz,
Hall, 1951), applying critical macro-illumination (Boxall and Hundy,
1955), imposing stress-coat (Fisher and Rogers, 1956), applying
clip-extensometer (Moon, 1971), using grid-reflection (Verel and
d ( σ ∞ − σ r ) − σ T Sleeswyk, 1973), imposing surface grid (Lloyd and Morris, 1977),
( σ T ) = ; σ = σ r + σ T (14) applying contrast interferometer (Iricibar et al., 1977), using high
dt tT
speed camera (Miyazaki, Fujita, 1979) and infrared camera (Louche
where σ r , σ ∞ are the stress levels: right after the catastrophic slip
and Chrysochoos, 2001.), applying strain gages (Zhang, Jiang, 2004)
and the asymptotic level (after the transients have died out). Here,
and applying optical interference (Bućko et al., 2014).
tT denotes the characteristic time of the relaxation process. Tem-
For what concerns the cryogenic temperatures, a large variety
perature evolves towards a new steady state value, inducing an ad-
of materials exhibit the plastic flow discontinuities. Some examples
ditional stress increment σ T .
3. Experimental evidence of strain localization during DPF
3.1. Experimental techniques to observe slip band evolution
Different techniques have been developed and implemented to
observe the nature of plastic instabilities under combined loads.
of such materials are presented in Fig. 8.
DPF was experimentally investigated by many authors (Basinski,
1957; Schwarz, Mitchell, 1974; Reed, Simon, 1988; Reed, Walsh,
1988; Hähner, Zaiser, 1997; Zaiser, Hähner, 1997; Benallal et al.,
2006; Skoczen et al., 2014), however, the strain localization ef-
fect was not studied until now. The experimental methods to
observe the propagation of slip bands are largely limited due
600 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612

Fig. 8. Uniaxial tensile test at liquid He for: (a) JK2LB (0.003C-22Mn-13Cr-9Ni-1Mo-0.2 N); (b) Niobium - Ti6Al4V; (c) serrations obtained for Cu-OFE – stress - strain (red)
and temperature-strain (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 9. Experimental set-up for tensile testing at extremely low temperatures.

to the extreme conditions (cryostat). As for such materials like
austenitic stainless steels the serrated yielding occurs below a spe-
cific threshold T1 (usually less than 30 K), collecting the experi-
mental evidence of localized deformation remains a challenge.
3.2. Cryogenic set-up for tensile tests
Cryogenic tensile tests were performed at the liquid helium
temperature (4.2 K). Since the tests at extremely low temperatures
imply specific conditions and constraints, it is worth explaining
them in detail.
The heart of the system remains the cryostat (vacuum insulated
enclosure around the sample), connected to the liquid helium de-
war by means of helium transfer line (see Fig. 9). Before starting
tensile test, the liquid helium flows from the dewar to the cryostat
until the specimen with the sensors is immersed in the cryogenic
medium. Thermal insulation system of the cryostat consists of in-
sulation vacuum, radiation screens and styrodur blocks, which pro-
vide the thermodynamic stability to the cryogen. Additionally, dur-
ing the test the cryogen can be refilled in the cryostat, in order to
maintain correct level of the cryogen bath. The experimental set-
up is complex and not transparent – it does not allow to observe
the sample directly. Therefore, rather complex instrumentation has
to be used. This, in turn, implies application of high resolution and
high quality data acquisition system.
The cryostat is equipped with an insert containing the sample
holder and the components carrying instrumentation, i.e. the clip-
on extensometers, the thermometers and the piezoelectric force
sensor (Fig. 10).
The instruments are linked to the data acquisition system. Elon-
gation of the specimen, its temperature and the applied loads are
monitored and stored. Selection of the measuring system compo-
nents is crucial for low temperature applications, for instance, se-
lection of appropriate elongation transducer. It is worth pointing
out, that DPF has dynamic character and may generate oscillations
of the extensometer-specimen structure. This, in turn, can lead to
erroneous interpretation of tensile test results. Tabin and Pracik˛
(2015) proposed simple approach to determine dynamic proper-
ties of compound structure that consists of strain gage clip-on
extensometers (dual extensometers) and the specimen. Based on
 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612 601

Fig. 10. Cryostat equipped with suitable instrumentation.

Table 1 The stress - strain and the temperature - strain characteristics

Chemical composition of grade 304 austenitic stainless steel.
indicate the nature of slip band propagation during cryogenic ten-
Element C Si Mn P S Cr Ni sile test performed on 304 stainless steel. Analyzing the results, it
is worth focusing on the temperature distribution within the range
Wt% ≤ 0.07 ≤ 0.80 ≤ 2.00 0.045 0.03 17.0–19.0 9.0–11.0
without visible hardening influence (stage 1, Fig. 11b). The profile
of the temperature - strain curve may suggest the nature of the
slip band evolution during the tensile test at liquid He. The tem-
this approach, appropriate clip-on extensometers were selected to perature distribution in this range has regular form (the comb-like
perform cryogenic tensile tests at extremely low temperatures on profile), which occurs when the slip band during plastic straining
austenitic stainless steels specimen. Finally, precise calibration of propagates from one extremity of the specimen to the other, and
DAQ system for such extreme conditions is necessary. from serration to serration. It is worth pointing out, that during
the period of time from 190 to 285 s (stage 2, Fig. 11b), there is
no recorded temperature spikes in view of the fact that the slip
3.3. Uniaxial tensile test results
band holds the position beyond the specimen gage length (see Fig.
11b and Fig. 15), thereby beyond the sensing range of the temper-
The uniaxial tensile test was carried out by using 304 stain-
ature sensor. Nevertheless, during this period the serrations occur
less steel plate specimen, immersed in liquid helium (4.2 K). The
(see time response of the force transducer). The temperature dis-
uniform gage section of the specimen had a length of 25 mm. The
tribution within the hardening range (stage 3, Fig. 11b) is random,
cross-section of the gage section was of 1.5 mm × 1 mm. The chem-
which reflects irregular and constrained motion of the slip bands,
ical composition of specimen is listed in Table 1.
which is associated with the interaction of collective and massive
During kinematically controlled test (cross-head velocity: Vc =
breaks through the lattice barriers and nucleation of secondary
0.5mm/min), the time responses of the clip-on extensometers,
(martensitic) phase. In other words, initially observed regular trav-
the force transducer and the temperature sensors were recorded
elling slip band is blocked by nucleation of the martensite phase,
(Fig. 11b). Based on these records, the stress - strain and the
whereas, the new slip bands occur randomly within the specimen.
temperature-strain curves were generated (Fig. 11a).
These postulates can be confirmed by means of analysis of time re-
It is worth pointing out, that near-0 K the steel specimens still
sponse of temperature sensors and extensometers, recorded during
preserve suitable physical and mechanical properties, including
the test (Fig. 13 b and c). Based on this analysis, it is possible to
high level ductility. Fig. 12 illustrates fracture surface of 304 stain-
determine the temporal and the spatial characteristics of the slip
less steel specimen tested at 4.2 K, which is characteristic of ductile
band.
decohesion. This kind of rupture is associated with intensive plas-
Looking at the inserted bars in Fig. 13b, it is found that the
tic deformation. From the macroscopic point of view, it is easy to
plastic deformation was first initiated in a local area near the bot-
see that the specimen has been plastically deformed (Fig. 12a).
tom part of the gage section, and has spread over the whole gage
Thermodynamic conditions of DPF are strictly linked to the
section (Fig. 13a). Temperature distribution, corresponding to a slip
so-called thermodynamic instability, related to vanishing specific
band evolving in a regular way with constant velocity, is illustrated
heat when the temperature approaches absolute zero. Macroscopic
in Fig. 14.
character of the DPF is reflected by the stress oscillations (Fig. 11a,
When the slip band approaches the position of the thermome-
red curve), and heat dissipation during each serration generated
ter, the temperature amplitude related to the serrations increases.
during slip band evolution. This is represented by the temperature-
Maximum temperature is registered when the slip band occurs un-
strain curve (Fig. 11a, blue curve). It is worth pointing out, that
der the thermometer. Thereafter, the temperature recorded by the
during tensile test of 304 stainless steel the plastic strain induced
sensor decreases, until the slip band falls outside the measurement
phase transformation (Garion et al. 20 06; Skoczen, 20 04, 20 07; Eg-
range of the thermometer. For what concerns the hardening range,
ner et al. 2015) is also observed. The fcc-bcc phase transformation
the temperature distribution shows random pattern (Fig. 13c). It is
is represented by the hardening effect, visible in the stress - strain
worth pointing out, that the time response of the extensometers
curve (Fig. 11a).
602 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612

Fig. 11. Tensile test of 304 stainless steel at 4.2 K: (a) stress - strain curve (red) and temperature-strain curve (blue); (b) time response of force transducer, clip-on ex-
tensometers and the thermometer. The position of the slip band during the uniaxial tensile test: 1- the slip band propagates from one extremity of the specimen to the
other, 2- the slip band occupies the position beyond the specimen gage length (see Fig. 15), 3- the slip band occurs randomly within the entire length of the specimen. (For
interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 12. SEM analysis of ductile fracture of 304 stainless steel specimen tested at 4.2 K; (a) plastic deformation of specimen cross-section; (b) the fractured surface containing
many microvoids and dimples.

Fig. 13. Cryogenic tensile test of 304 stainless steel specimen; (a) experimental set-up (b) temperature and strain for kinematically controlled test without phase transfor-
mation; (c) temperature and strain during the test accompanied by the phase transformation process.
 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612
Fig. 14. (a) Time response of thermometer during cryogenic tensile test in liquid He (the comb-like profile of temperature); (b) ilustration of the slip band propagation.
Fig. 15. Time response of clip-on extensometers as a function of slip band localization during cryogenic tensile test at liquid He temperature.
during the test allows us to localize the slip band. The behavior
of the extensometer, when the slip band occurs between its arms
and beyond them is different, which is shown in Fig. 15. Drop of
the strain signal is generated by the slip band that occurs outside
the extensometer arms (outside the gage length). During the pe-
riod of time from 190 to 285 s, there is no recorded global increase
of elongation in view of the fact, that the slip band occurs outside
the gage length (stage 2, Fig. 11b, Fig. 15).
During tensile test of 304 stainless steel, the first slip band
occurred near the one of two grips, beyond the extensometer
arms, which is indicated by a drop of strain signal. Then the slip
band propagated across the extensometer arms and the next drop
of strain signal is recorded when the slip band is again beyond
the extensometer arms, near the second grip. On the other hand,
whenever the hardening effect is observed the slip band occurs in
an irregular way (Fig. 13c). Both types of time response of ther-
mometer and the clip-on extensometers are consistent and confirm
the nature of slip band propagation in 304 stainless steel at liquid
He temperature. In order to determine parameters of the slip band
evolution, an individual experiment was carried out.
3.4. Experimental set-up to identify parameters of slip band evolution
In order to determine the nucleation spot and the direction
of slip bands evolution, two thermometers were bonded on 304
stainless steel specimen surface: the thermometer II (red) was
bonded beyond the bottom grip and the thermometer I (green) in
the middle of gage length. Distance between the thermometers is
equal to 16 mm. Kinematically controlled tensile test (cross-head
603
speed: Vc = 0.5mm/min) was carried out at liquid He temperature.
In addition, both clip-on extensometers were used. The experimen-
tal set-up is presented in Fig. 16.
Based on time response of thermometers and clip-on exten-
someters, the parameters of slip bands propagation for two distinct
ranges: without and with hardening can be determined. In partic-
ular, the average velocity of slip bands within the range without
hardening are calculated. Fig. 17a illustrates time response of two
thermometers: the red one, bonded near the bottom grip, and the
green one, bonded in the middle of gage length. Also, time deriva-
tive of extensometers response (strain rate) is included in Fig. 17b.
Time response of extensometers during test becomes an im-
portant source of information (Fig. 13, Fig. 17b). If a slip band oc-
curs outside gage length, drop of extensometers signal is observed.
Therefore, the time derivative of extensometers signal resembles
impulse function with negative value (first peak in blue curve). On
the other hand, when a slip band occurs between extensometer
arms, the time derivative of extensometers signal becomes impulse
function with positive value. Based on the response of thermome-
ters, it can be deduced that during tensile test the slip band oc-
curs near the red thermometer and then propagates towards the
green one (in the middle of specimen). When the thermometer I
(green) is reached by a slip band, the measured amplitudes of tem-
perature oscillations (related to serrations) substantially increase.
The maximum occurs when the slip band reaches the position of
the thermometer. Thereafter, as a consequence of the slip band
propagation, the time response of thermometer I decreases until
the slip band leaves the measurement range of thermometer I. On
the other hand, the thermometer II (red one) was bonded outside
604 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612

Fig. 16. Experimental set-up for tracking slip band evolution during uniaxial tensile test at liquid He temperature.

Fig. 17. Visualization of slip band evolution in 304 stainless steel: (a) regular slip band propagation (no hardening); (b) irregular slip band propagation (phase transformation
induced hardening).

the gage length, therefore a slip band recorded by thermometer II
causes simultaneously drop of extensometer signal. Next, the same
slip band occurs within the gage length and propagates across the
gage length in the proximity of thermometer I. When the slip band
falls outside the gage length, the negative peak is again observed in
the rate of elongation chart. The slip band propagation in the range
without hardening has regular form, and the slip band during the
test propagates from grip to grip. Hardening caused by the fcc-
bcc phase transformation affects considerably the slip band prop-
agation. In particular, slip bands occur at random places and very
soon are blocked by obstacles (secondary phase). Therefore, time
response of thermometers does not have regular pattern. Also, the
strain rate suggests that the slip bands occur in random places,
within or outside the gage length.
Experimental set-up presented in Fig. 16 allows to determine
the basic parameters of slip band propagation under test condi-
tions initially specified (cross-head velocity Vc = 0.5mm/min). Dis-
tance between two thermometers is equal to lT = 16mm. Based on
time response of two thermometers, it is possible to determine
time increment when a slip band occurs under the thermome-
ter II and, later on, under the thermometer I (the test conditions
are fixed). Therefore, the average slip band velocity is equal to
Vsb = 48.5mm/min, when no phase transformation occurs. Similar
effect was observed by Louche and Chrysochoos (2001) during uni-
axial tensile test of S355MC steel at room temperature. A single
band, propagating at constant velocity was observed.
4. Thermodynamic model of temperature oscillations
4.1. General aspect
In view of Section 2 it is worth keeping in mind, that in the
proximity of absolute zero several thermo-mechanical parameters
like thermal conductivity and thermal contraction coefficient, or
 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612 605

state functions like specific heat and entropy tend to zero. It has x ∈ (−∞, ∞ ); t ∈ 0, ∞ ) (17)
fundamental consequences for the behavior of fcc materials at
with the following initial condition:
ultra-low temperatures, especially due to the dissipative phenom-
ena related with discontinuous plastic flow. The general aspect of θ (x, 0 ) = f (x )
discontinuous yielding mechanism is presented in Sections 2.2 and x ∈ (−∞, ∞ ) (18)
2.3, including explanation of dissipation character of DPF in fcc
The general solution of the heat Eq. (17) is given by (Tichonow
materials. Thus, during DPF, when the abrupt drop of stress begins,
and Samarski, 1963):
the temperature locally increases. The heat increment is related  ∞
to the temperature increment by Eq. (13). The heat quantity pro-
θ (x, t ) = C (ω ) e−iωx e−kω t dω
2
(19)
duced during the massive rupture of lattice barriers is a function −∞
of plastic work and internal friction in the lattice (Eq. (12)). There- The initial condition (18) can be rewritten as:
fore, the DPF combines the mechanical and the thermodynamic ef-  ∞
fects. In order to perform quantitative analysis of DPF Skoczen et θ (x, 0 ) = f (x ) = C (ω ) e−iωx dω (20)
al. (2014) proposed a numerical model. −∞
For now, only general aspects of the model will be presented. In This equation implies that the function f(x) is the inverse
particular, the thermo-mechanical model assumes, that for single Fourier transform of the function C(ω) which, in turn, implies that
serration there is no change of temperature as long as the elastic C(ω) is the Fourier transform of f(x):
deformation lasts. Subsequent smooth plastic flow generates plas-  ∞
1
tic power that is converted to heat. The sample warms up and the C (ω ) = F (ω ) = f (x ) eiωx dx (21)
2π −∞
heat is partially transported by convection from the sample to the
surrounding helium. Temperature increment is based on the heat Substituting C(ω) to Eq. (19), the general solution can be writ-
balance equation, and reads: ten as:
 ∞
  ∞

ησ ε p θ (x, t ) =
1
f (ξ ) eiωξ dξ e−iωx e−kω t dω
2
T = (15)
2π
(22)
ρmCV + 4λενε−1 g−2 −∞ −∞

where T denotes temperature increment, η is the efficiency fac- whereξ is used as a dummy integration variable. Now, the order of
tor, ρ m stands for mass density, CV means specific heat, λ denotes integration can be changed over:
thermal conductivity, g is the thickness of the sample, and  ε    
vε is ∞
1 ∞
−kω2 t −iω (x−ξ )
the estimated time increment. θ (x, t ) = f (ξ ) e e dω dξ (23)
−∞ 2π −∞
After the stage of continuous plastic flow, abrupt drop of stress
occurs within extremely short time counted in micro-seconds. For Expression in the bracket forms heat kernel and has the follow-
this reason, the heat exchange is neglected (nearly adiabatic condi- ing form (Tichonow and Samarski, 1963):
tions) and total plastic power is converted to heat. The increment  ∞
1 1 −(x−ξ )2
G (x − ξ , t ) = e−kω t e−iω (x−ξ ) dω = √
2
of temperature can, therefore, be derived as: e 4kt (24)
2π −∞ 4π k t
ησ ε p
T = (16) Therefore, Eq. (23) can be rewritten as:
ρmCV  ∞

During relaxation stage, there is no significant heat source and θ (x, t ) = f (ξ ) G (x − ξ , t ) d ξ (25)
−∞
the sample is cooled down to the liquid helium temperature.
It is worth pointing out, that the proposed by Skoczen et al. Eq. (25) represents a spatial convolution of the functions f(ξ )
(2014) numerical model assumes that the RVE travels together and G(x − ξ , t ).
with the slip band. In other words, the thermometer moves during Let us consider special case when the initial temperature distri-
the kinematically controlled test along with the slip band. There- bution f(ξ ) is concentrated at point x0 , namely:
fore, the peak value of temperature produced by the model is al- f ( ξ ) = Aδ ( ξ − x 0 ) (26)
most constant (see Fig. 23). Nevertheless, the experimental results
where δ denotes the Dirac function. This implies that the intensity
indicate that the temperature distribution, measured by spatially
of heat source at time t = 0 in an infinite rod reads:
fixed thermometer, has different shape (comb-like profile), as il-  ∞  ∞
lustrated in Fig. 18. Q = cV ρ f (ξ ) dξ = cV ρ A δ (ξ − x0 ) dx = cV ρA (27)
The difference between the numerical and the experimental −∞ −∞
temperature distributions is related to the assumption that the For t > 0, the temperature function θ (x, t) is given by:
RVE travels together with the slip band. However, as regards the  ∞
experiment, the thermometer is spatially oriented (mounted on θ (x, t ) = f (ξ ) G (x − ξ , t ) d ξ
−∞
the specimen surface) when the slip band propagates through the  ∞
specimen. = A δ ( ξ − x0 ) G ( x − ξ , t ) d ξ = A G ( x − x0 , t ) (28)
In order to correlate the experimental results with the numeri- −∞
cal simulations, a heat conduction analysis is required. where the fact that the only contribution to the integral comes
from the point at which the argument of the Dirac delta function
4.2. Solution of heat diffusion equation is zero (x0 ), has been used. Thus, one obtains:
Q −(x−x0 )2
The heat diffusion equation is a starting point for the analysis θ (x, t ) = A G(x − x0 , t ) = √ e 4k t (29)
of temperature distribution generated by slip band evolution cV ρ 4π k t
during DPF. For a flat specimen the heat diffusion equation can The total area under this distribution is equal to A for any time
be markedly simplified to one-dimensional partial differential t.
equation: The convolution in Eq. (25) represents a weighted average of
the function f(ξ ) about the point x0 (Eq. (26)). As the time in-
∂θ ∂ θ 2
=k 2 creases, this weighted average includes more points away from x0 ,
∂t ∂x
606 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612

Fig. 18. Tensile test results: the stress – strain curve (red) and temperature –strain curve (blue) for 304 stainless steel at liquid He temperature. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 19. Temperature distribution around the position x0 = 0.5 at defined moments of time ti = [0.15 s, 0.20 s, 1.20 s, 1.40 s, 2.60 s, 6.00 s].

Table 2 Fig. 19 shows a graph of temperature distribution, that exhibits

Properties of grade 304 stainless steel at liquid He temperature.
a concentration around the position x0 = 0.5. At time tn > 0, por-
Thermal conductivity k= 0.23
W tion of heat has been transported from the localization zone to re-
mK gions situated far away from x0 = 0.5.
J
Specific heat CV = 2
kg K In view of Fig. 18, periodic dissipation of energy during each
kg serration occurs. It means that for the kinematically controlled ten-
Mass density ρm = 7998
m3 sile test and spatially oriented thermometer, the traveling slip band
m2 generates within each serration some amount of heat, which is
Thermal diffusivity α = 0.0038
s
recorded by the thermometer in the form of temperature peaks.
W
Intensity of internal heat source Q = 900 3 Similar effect can be obtained with respect to the solution of heat
m
diffusion equation. Each curve in Fig. 20 results from a solution of
Eq. (29) for experimentally determined parameters (such as: serra-
tion frequency fs , slip band velocity Vsb , etc.). The results of numer-
ical simulation of time response of the thermometer during kine-
which accounts for the smoothing of the temperature distribution matically controlled uniaxial tensile test of 304 stainless steel spec-
θ (x, t). Fig. 19 illustrates the qualitative aspects of such smoothing. imen at liquid He temperature are presented in Fig. 20. The basic
In order to perform a simulation of temperature relaxation process, parameters of the test are presented in Table 3.
grade 304 stainless steel has been selected. Its physical proper- It is worth pointing out, that the presented heat conduction
ties at liquid He temperature are presented in Table 2 (Marquardt, analysis is based on a number of simplifications, such as the
Radebaugh, 20 0 0).
 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612 607

Fig. 20. Time response of “numerical thermometer”, based on the solution of heat diffusion equation.

Table 3 Plastic flow observed at a macroscopic level is accompanied by

The parameters adopted for numerical test.
evolution of microscopic variables. In particular, the evolution of
Type of test Kinematically controlled dislocations density reads (Ma et al., 2006; Bouquerel et al., 2006;
mm
Madec et al., 2002):
Cross-head velocity Vc =0.5  √ 
min
1 k 1 ρk
Slip band velocity Vsb =48.5
mm ρk+1 = ρk + εkp+1 M + − k a ρk (30)
min db b
Serration frequency f s = 1 Hz
where d is the average grain size, b is the length of the Burgers
vector and ka stands for the dislocation annihilation constant. As
soon as the threshold ɛLC is reached, the lattice barriers evolve
boundary conditions imposed at infinity or 1D geometry of the and their number density is calculated according to Eq. (4), with
test specimen. Therefore, the numerical distribution of tempera- Bcr corresponding to such a number density beyond which the ser-
ture does not entirely converge with the experimental one (e.g. ration occurs immediately:
the long-lasting tail, Fig. 20). Nevertheless, the peak values of tem-  
perature spikes (or the envelope of temperature distribution) have if εkp+1 > εLC then Bk+1 = min Bk + FLC
+
εkp+1 ; Bcr (31)
a significant meaning in view of the DPF model, which presently Modified average shear stress in the lattice reads (cf. Kubin,
takes into account the travelling slip band (carrying the RVE), and 2003; Bouquerel et al., 2006):
a fixed thermometer located in the middle of the sample. √
Based on Eq. (29), initial conditions (Eq. (26)), and experimen- τk+1 = τ0 + μα b ρk+1 (32)
tally identified test parameters (Table 3) the numerical time re- with the mean free path of dislocation calculated as:
sponse of the thermometer during uniaxial tensile test of several
1 1 √ 
materials at liquid He temperature has been obtained (Fig. 21b for = + k1 ρk+1 + k2 Bk+1 (33)
grade 304 stainless steel). λ̄k+1 d
Finally, the average shear stress at the head of dislocation
pile-ups is computed according to Eq. (6), and the criterion of
5. Numerical simulations of DPF versus experimental results avalanche-like failure of lattice barriers is checked.
The drop of stress (stage 3) is defined on the basis of the micro-
Explicit incremental scheme is applied for stress integration, scale variables, with a newly proposed approach in which pre-
controlled with a constant value of the total strain increment. The viously used heuristic function F − is eliminated. The relation be-
LC
return mapping algorithm is used as soon as the hardening line tween the stress in the lattice and the density of dislocations (Eq.
is reached at the end of the elastic range. Following the experi- (5)) is defined as:
mental results, it is assumed that each serration occurs at a con- √
stant value of total strain. Hence, redistribution of the elastic and σk+1 = M (τ0 + μα b ρk+1 ) (34)
the plastic parts of strain takes place. Moreover, both density and where M is the Taylor factor, and τ 0 stands for the shear stress of
amplitude of serrations slightly increase in the course of loading internal friction. Hence, the difference in stress values before and
process. Also, the plastic strain threshold for initiation of each ser- after serration are expressed as:
ration gradually increases. Drop of stress is followed by a temper- √

ature rise that indicates a process of mechanical origin. For 304 σ serr = σkbefore
+1 − σ after
k+1 = M μα b ρ k+1 − ρ 0
k+1
(35)
stainless steel linear plastic hardening has been assumed and pe-
riodically repeated serrations do not affect the hardening curve Recalling the observation that the serration starts at a given
stress value σk+1 and takes place at a constant value of total
be f ore
(Skoczeń et al., 2010).
608 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612

Fig. 21. Time response of thermometer during tensile test of 304 stainless steel at liquid He temperature (a) experimental results, (b) numerical results.

strainɛserr , and that the increment of plastic strain corresponds to On the other hand, during stage 4 (relaxation) the sample is
the equivalent decrement of elastic strain, one obtains the plastic gradually cooled down to the temperature of helium bath. Ther-
strain increment (see Fig. 7c) in the form: modynamic balance takes the form:
√ 
Mμα b ρk+1 − ρk0+1 ρmCV Tk+1 = Q
=
4λε (Tk+1 −T0 )
(40)
ε serr
p = (36) V g2 νε
E
Temperature decrement is computed as long as the term in the
Furthermore, it is assumed that the density of dislocations af-
parenthesis is positive:
ter serration ρk0+1 is not retained at its initial levelρ00 , but evolves
during the loading process and so does the after-serration number 4λε (Tk+1 − T0 )
Tk+1 = (41)
density of barriers (e.g. L-C locks) B0 . The threshold for the barriers ρmCV g2 νε
evolution is also updated:
The thermodynamic functions like the specific heat and the
Bk+1 = B0k+1 , ρk+1 = ρk0+1 , εLC = (εLC )k+1 + εkp+1 (37) thermal conductivity are functions of temperature. Their values are
It is proposed to approximate the after-serration dislocation illustrated in Fig. 22 for grade 304 stainless steel (Marquardt et al.
density as: 20 0 0):
⎛ ⎞
log Cv
⎜ 0.0 0 031 ⎟
ρk0+1 = ⎝ 
3 + 0.2⎠ρk+1 (38) = 22.0061 − 127.5528 log T + 303.6470(log T ) +
2
2
εkp+1 3
− 381.0098(log T ) + 274.0328(log T ) − 112.79212(log T ) +
4 5

Note, that the serration increment of the plastic strain εpserr 6

+ 24.7593(log T ) − 2.239153(log T )
7
(42)
does not influence the yield surface and no additional hardening
is assumed. This is because of the fact that the increase of plastic log λ = −1.4087 − 1.3982 log T + 0.2543(log T ) +
2
strain due to serration is caused by a different mechanism (fast 3 4 5
slip) than in regular continuous yielding (motion of dislocations). − 0.6260(log T ) + 0.2334(log T ) + 0.4256(log T ) +
6 7 8
Eventually, at the end of stage 3, the stress and the unchanged − 0.4658(log T ) + 0.1650(log T ) − 0.0199(log T )
strain are obtained:
(43)
εk+1 = εk , σk+1 = σk − σ serr
After serration, stage 4 is in general expected, when the stress
relaxation takes place. The integration of constitutive relations dur-
ing this phase can be found in Skoczeń et. al, 2010.
The algorithm for calculation of temperature change was dis-
cussed in Skoczeń et al, 2014, for bi-axial stress state. The approach
developed in the present paper includes formation and motion of
slip bands, accompanied by dissipation of plastic power converted
to heat. It is assumed that no change of temperature is observed
during elastic deformation (stage 1), since there is no dissipation of
energy. During stage 2 (classical plastic flow), the dissipated plas-
tic power is converted to heat. Part of heat is transported to the
surrounding liquid helium by convection. Temperature evolution is
computed from the heat balance equation according to Eq. (15) for
stage (2), and Eq. (16) for serration stage (3).
(39)
The numerical results are illustrated in Figs. 23 and 24. The fol-
lowing set of parameters was adopted for the numerical analysis
(Table 4):
Good agreement when compared to the experimental data is
obtained. Especially, serration frequency and amplitude fall within
the range obtained from experiments. It is worth pointing out,
that the temperature evolution predicted by the model shows sud-
den peaks, but the effect of “distant thermometer” is not included.
Hence, the result resembles a fictitious situation of a thermometer
travelling with the slip band. This issue is addressed in Section 4.
Fig. 24 illustrates the evolution of dislocation density as a function
of total strain. The peak points correspond to maximum values ob-
tained during plastic deformation (Eq. (30)), whereas, the lowest
values represent the after-serration density, when the barriers (L-C
locks) are broken and the spontaneous motion is triggered.
 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612 609

Fig. 22. Thermal conductivity and specific heat for 304 stainless steel as a function of temperature.

Fig. 23. Evolution of stress and temperature as a function of strain computed during DPF.

6. Discussion

The above presented constitutive model contains a number of

new features. The first new element reflects the way of computing
drop of stress within single serration. It has been assumed that the
drop of stress is entirely determined by dislocations density be-
fore and after the serration (Eq. (10)). Also, the amount of plastic
slip follows similar description (Eq. (11)), which makes its physi-
cal interpretation more obvious. Moreover, dislocations density be-
fore and after serration becomes function of time-like parameter,
which leads to correct evolution of the minimum and the maxi-
mum stress during each serration. This effect has been experimen-
tally confirmed.
Fig. 24. Evolution of dislocation density as function of strain computed during DPF.
Another crucial and new feature of the model is related to
the observation of slip band propagation along the sample gage
length during kinematically controlled test. In the course of pre-
Table 4
vious studies of this phenomenon (starting from early sixties), it
The parameters adopted for numerical analysis of stress, temperature and dis- was not clear whether the strain localization effect takes place at
location density (Figs. 23 and 24). extremely low temperatures. Recent experiments, reported in the
Parameter equation value present paper, clearly indicate strong strain localization in the form
of slip bands propagating along the sample. Plastic power dissi-
σ0 550 pated in the slip band is partially converted to heat, which re-
τ0 Eq. (5) = MPa
M 3
sults in a drastic increase of temperature promoted by the ther-
E (1 + ν ) 2.1 · 105 (1 + 0.3)
μ Eq. (5) = MPa modynamic instability (nearly adiabatic process). The localization
2 2
α Eq. (5) 0.4 effect is reflected by temperature oscillations measured by spa-
tially fixed thermometer(s). Such experimental evidence of strain
b Eq. (5) 2.58 · 10−10 m
localization process must be taken into account in the constitutive
d Eq. (7) 5 · 10−6 m
model by locating the RVE inside the slip band. Thus, the plas-
k1 Eq. (7) 0.25 tically active process, including the plastic slip accompanied by
k2 Eq. (7) 0.01 abrupt drop of stress, takes place inside the slip band (RVE) and
tT Eq. (14) 100 s the plastic work is partially converted to heat. Due to nearly adi-
kg abatic conditions (see the specific heat and the thermal conduc-
ρm Eq. (15) 7860
m3 tivity in the temperature range 0–5 K), the temperature increase
g Eq. (15) 0.003m measured by one or two thermometers is significant. Moreover,
νɛ Eq. (15) 1.8 · 10−3 s−1 as the slip band moves along within the gage length of the sam-
ka Eq. (30) 0.01 ple, the thermometer located in the middle of the sample mea-
sures increasing temperature spikes. When the slip band crosses
610 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612
Fig. 25. Comparison of the experimental and the numerical stress - strain curves corresponding to DPF in 304 stainless steel.
the gage length of the sample (from one extensometer grip to an-
other) the measured temperature spikes follow comb-like profile
with the peak temperature corresponding to the alignment of the
slip band with the thermometer. This shows very ordered nature of
the slip band propagation process. However, as soon as the phase
transformation process takes place, the motion of slip band is hin-
dered by formation of the secondary phase. For this reason, the
temperature spikes occurring during the phase transformation are
randomly redistributed. Most probably, the free path of a slip band
is limited to the average distance between the secondary phase is-
lands and the initiation of slip band has random character. This
part of the process, seen from the point of view of temperature
measurements, has disordered character. Nevertheless, the present
paper is focused on the first part of the process, without phase
transformation.
One of the important features of the slip band propagation in
the absence of phase transformation consists in very regular tem-
perature pattern, measured by the thermometer located in the
middle of the gage length. When the slip band starts at one ex-
tremity of the sample the temperature spike measured by the ther-
mometer is small. As soon as the slip band approaches the ther-
mometer, the spike gradually reaches its maximum. Finally, when
the slip band moves to the opposite extremity of the sample,
the measured temperature spike decreases again. This very reg-
ular pattern of temperature fluctuations reflects regular and un-
constrained motion of the slip band, that can be described as a
heat source generating heat pulses as soon as the abrupt slip takes
place. For what concerns the thermodynamic aspect of the slip
band propagation in the first part of the process, the heat genera-
tion can be estimated by means of simple energy balance. It leads
to evaluation of temperature increase resulting from the thermo-
dynamic instability. However, this process can be described in a
much more precise way by means of heat diffusion analysis, in-
cluding temperature response at a distance from the heat source.
Temperature profile obtained as a result of heat diffusion analysis
(involving thermodynamic parameters and variables depending on
temperature) turns out very similar to what is measured during
the experiment (Fig. 21).
7. Conclusions
The available studies of dynamics of plastic flow oscillations at
extremely low temperatures were greatly insufficient from the ex-
perimental and from the numerical point of view. For this rea-
son, the above experimental identification of the strain localiza-
tion phenomenon during discontinuous plastic flow in the near-
zero K regime has been carried out. It reveals an interesting fea-
ture of the plastic flow discontinuities consisting in strong local-
ization in the form of slip band travelling across the gage length
of the sample. Another interesting feature consists in a coupling
of plastic flow discontinuities with the phase transformation pro-
cess, which makes the nucleation of slip bands and their free glide
randomly redistributed over the gage length of the sample. The
present paper deals with the regular way of nucleation and free
unconstrained glide of the slip bands in the sample, in the absence
of phase transformation and obstacles in the form of secondary
phase. The measured profile of temperature spikes has been ex-
plained by building a model of travelling slip band, carrying the
RVE, and fixed thermometer located in the middle of the sample. A
thermodynamic problem of heat diffusion in the sample has been
solved, taking into account travelling source of a given heat gen-
eration rate. Numerically obtained profile of temperature spikes is
similar to the profile measured during the experiment. Moreover,
the model of stress oscillations has been improved by expressing
their lower bound in terms of dislocations density.
Thus, it has been assumed that the drop of stress is entirely
determined by dislocations density before and after the serration
and the amount of plastic slip follows similar description, which
makes its physical interpretation more obvious. A comparison of
the experimental stress - strain curve with the numerical results
is shown in Fig. 25, and indicates correct behaviour of the consti-
tutive model. The numerical stress - strain curve does not entirely
match the initial behaviour, because the DPF model is based on lin-
ear elastic deformation combined with linear hardening (after the
stress gains the initial yield point). Such assumption is sufficient
for the constitutive description of DPF. There is also a good per-
spective to describe the interaction of DPF with the fcc-bcc phase
transformation process, however, it will be an objective of a future
paper as soon as the experimental evidence is sufficient.
Acknowledgements
This work has been supported by the National Science Cen-
tre through the Grant No. UMO-2013/11/B/ST8/00332. The authors
wish to thank Stefano Sgobba (PhD) and Dawid Marcinek (MSc) for
their suggestions and help.
References
Ait-Amokhtar, H., Vacher, P., Boudrahem, S., 2006. Kinematics fields and spatial ac-
tivity of Portevin–Le Chatelier bands using the digital image correlation method.
Acta Mater 54, 4366.
Barlat, F., Haa, J., Grácio, J.J., Leea, M-G, Rauch, E.F., Vincze, G., 2012. Extension of
homogeneous anisotropic hardening model to cross-loading with latent effects.
Int. J. Plast. 46, 130–142 2013.
 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612
Bartels, A., Bartel, T., Canadija, M., Mosler, J., 2015. On the thermomechanical cou-
pling in dissipative materials: a variational approach for generalized standard
materils. J. Mechanics Phys. Solids 82, 18–234.
Basinski, Z.S., 1957. The instability of plastic flow of metals at very low tempera-
tures. Proc. R. Soc.London, Ser. A 240, 229–242.
Benallal, A., Berstad, T., Borvik, T., Clausen, A.H., Hopperstad, O.S., 2006. Dynamic
strain aging and related instabilities: experimental, theoretical and numerical
aspects. European J. Mech. A/Solids 25, 397–424.
Benallal, T., Berstad, Borvik, T., Hopperstad, O.S., Koutiri, I., Nogueira de Codes, R.,
2008. An experimental and numerical investigation of the behaviour of AA5083
aluminium alloy in presence of the Portevin–Le Chatelier effect. Int. J. Plast. 24,
1916–1945.
Bouquerel, J., Verbeken, K., De Cooman, B.C., 2006. Microstructure-based model
for the static mechanical behaviour of multiphase steels. Acta Materialia 54,
1443–1456.
Boxall, T.D., Hundy, B.B., 1955. Photographing stretcher-strain markings with the
Vickers projection. Metallurgia 51, 52–54.
Bross, S., Hähner, P., Steck, E.A., 2003. Mesoscopic simulations of dislocation motion
in dynamic strain ageing alloys. Comput. Mater. Sci. 26, 46–55.
Bućko, S., Jodłowski, H., Trzebicki, M., 2014. An experimental-numerical analysis of
plastic zones initiation around holes in shields made of steel C45 under tension.
Arch. Mechanical Eng. 61 (1), 197–198.
Burns, T.J., 1994. A simple criterion for the onset of discontinuous plastic defor-
mation in metals at very low temperatures. J. Mech. Phys. Solids 42 (5), 797–
811.
Caceres, C.H., Rodriguez, A.H., 1987. Acoustic emission and deformation bands in
Al-2.5% Mg and Cu-30% Zn. Acta Metall 35, 2851.
Chihab, K., Estrin, Y., Kubin, L.P., Vergnol, J, 1987. The kinetics of the Portev-
in–LeChatelier bands in an Al-5at%Mg alloy. Scripta Metall 21, 203.
Chowdhury, P., Sehitoglu, H., Abuzaid, W., Maier, H.J., 2015. Mechanical Response of
Low Stacking Fault Energy Co-Ni Alloys - Continuum, Mesoscopic and Atomic
Level Treatments. Int. J. Plast. 71, 32–61.
Chrysochoos, A., Louche, H., 20 0 0. An infrared image processing to analyse the
calorific effects accompanying strain localisation. Int. J. Eng. Sci. 38, 1759–
1788.
Chrysochoos, A., Wattrisse, B., Muracciole, J-M., El Kaïm, Y., 2009. Fields Of Stored
Energy Associated With Localized Necking Of Steel. J. Mech. Mater. Struct. 4 (2),
245–262.
Dolgin, A.M., Natsik, V.D., 1991. Criteria of instability and kinetics of jumps under
unstable low temperature plastic flow. Acta Universitatis Carolinae – Mathemat-
ica et Physica 32 (1), 77–88.
Dumoulin, S., Louche, H., Hopperstad, O.S., Børvik, T., 2010. Heat sources, energy
storage and dissipation in high-strength steels: Experiments and modelling. Eu-
ropean J. Mech. A/Solids 29, 461–474.
Egner, H., Skoczen, B., Ryś, M., 2015. Constitutive and numerical modeling of cou-
pled dissipative phenomena in 316 L stainless steel at cryogenic temperatures.
Int. J. Plast. 64, 113–133 (2015).
Estrin, Y., Tangri, K., 1981. Thermal mechanism of the anomalous temperature de-
pendence of the flow stress. Scripta Metallurgica 15, 1323–1328.
Fisher, J.C., Rogers, H.C., 1956. Propagation of Lüders bands in steel wires. Acta Met-
all 4, 180–185.
Garion, C., Skoczen, B., Sgobba, S., 2006. Constitutive modelling and identification
of parameters of the plastic strain-induced martensitic transformation in 316 L
stainless steel at cryogenic temperatures. Int. J. Plast. 22, 1234–1264.
Groh, S., Marin, E.B., Horstemeyer, M.F., Zbib, H.M., 2009. Multiscale modeling of the
plasticity in an aluminum single crystal. Int. J. Plast. 25 (8), 1456–1473.
Hähner, P., Zaiser, M., 1997. From mesoscopic heterogeneity of slip to macroscopic
fluctuations of stress and strain. Acta Mater. 45 (3), 1067–1075.
Hall, E.O., 1951. The deformation and aging of mild steel: II. Characteristics of the
Lüders deformation. III. Discussion of results. Proc. Phys. Soc. B 64, 742–753.
Hamelin, C.J., Diak, B.J., Pilkey, A.K., 2011. Multiscale modelling of the induced plastic
anisotropy in bcc metals. Int. J. Plast. 27 (8), 1185–1202.
Hu, Q., Zhang, Q., Cao, P., Fu, S., 2012. Thermal analyses and simulations of the type
A and type B Portevin–Le Chatelier effects in an Al–Mg alloy. Acta Materialia 60
(4), 1647–1657.
Hu, Qi, Zhang, Qingchuan, Cao, Pengtao, Fu, Shihua, 2012. Thermal analyses and
simulations of the type A and type B Portevin–Le Chatelier effects in an Al–Mg
alloy. Acta Materialia 60 (4), 1647–1657.
Hull, D., Bacon, D.J., 2011. Introduction to dislocations, Fifth Edition Elsevire Ltd.
ISBN: 978-0-08-096672-4.
Iricibar, R., Mazza, J., Cabo, A., 1977. On the Lüders band in mild steel-I. Acta Metall
25, 1163–1168.
Klusemann, B., Gottfried, F., Böhlke, T., Svendsen, B., 2015. Thermomechanical char-
acterization of Portevin–Le Châtelier bands in AlMg3 (AA5754) and model-
ing based on a modified Estrin–McCormick approach. Int. J. Plast. 67, 192–
216.
Kocks, U.F., Mecking, H., 2003. The physics and phenomenology of strain hardening.
Prog. Mater. Sci 48 (3), 171–273.
Kubin, L., 2003. Discrete dislocation dynamics & multiscale analyses of plasticity.
Theoretical and Computational Materials Science, June 2-6, 2003, 1st Summer
School in Stuttgart, International Max Planck Research School for Advanced Ma-
terials.
Labergere, C., Guelorget, B., Francois, M., 2014. Strain rate distribution and lo-
calization band width evolution during tensile test. Int. J. Solids Struct. 51,
3944–3961.
611
Lazarev, A., Vinogradov, A., 2009. About plastic instabilities in iron and power spec-
trum of acoustic emission. J. Acoustic Emission 27.
Lloyd, D.J., Morris, L.R., 1977. Lüders band deformation in a fine grained aluminium
alloy. Acta Metall 25, 857–861.
Louche, H., Chrysochoos, A., 2001. Thermal and dissipative effects accompanying
Lűders band propagation. Mater. Sci. Eng. A 307, 15–22.
Louche, H., Vacher, P., Arrieux, R., 2005. Thermal observations associated with the
Portevin-le Chatelier effect in an Al-Mg alloy. Mater Sci Eng, A 404, 188–196.
Ma, A., Roters, F., Raabe, D., 2006. A dislocation density based constitutive model
for crystal plasticity FEM including geometrically necessary dislocations. Acta
Materialia 54, 2169–2179.
Madec, R., Devincre, B., Kubin, L.P., 2002. From dislocation junctions to forest hard-
ening. Phys. Rev. Lett. 89 25 255508-1 – 255508-4.
Manach, P.Y., Thuillier, S., Yoon, J.W., Coër, J., Laurent, H., 2014. Kinematics of Portev-
in–Le Chatelier bands in simple shear. Int. J. Plast. 58, 66–83.
Marquardt, E.D., Le, J.P., Radebaugh, R., 20 0 0. Cryogenic material properties
database. presented at the 11th International Cryocooler Conference, June
20-22, 20 0 0, Keystone, Co.
Mecking, H, Kocks, U.F., 1981. Kinetics of flow and strain-hardening. Acta Metall 29,
1865–1877.
Miyazaki, S., Fujita, H., 1979. Dynamic observation of the process of Lűders band
formation in polycrystalline iron. Trans. Japan Inst. Metals 20, 603–608.
Moon, D.W., 1971. Strain distribution through a propagating Lűders band front.
Scripta Metall. 5, 213–216.
Nielsen, K.L., Pardoen, T., Tvergaard, V., de Meester, B., Simar, A., 2010. Modelling of
plastic flow localisation and damage development in friction stir welded 6005A
aluminium alloy using physics based strain hardening law. Int. J. Solids Struct.
47, 2359–2370.
Obst, B., Nyilas, A., 1991. Experimental evidence on the dislocation mechanism of
serrated yielding in f.c.c. metals and alloys at low temperatures. Mater. Sci. Eng.
A137, 141–150.
Obst, B., Nyilas, A., 1998. Time-resolved flow stress behavior of structural materi-
als at low temperatures. In: Advances in cryogenic engineering (Materials) 44.
Plenum Press, New York, pp. 331–338.
Petit, J., Montay, G., Francois, M., 2013. Strain rate measurements by speckle inter-
ferometry for necking investigation in stainless steel. Int. J. Solids Struct. 51, 550.
Pustovalov, V.V., 2008. Serrated deformation of metals and alloys at low tempera-
tures (review). Low Temp. Phys. 34 (9), 683–723.
Reed, R.P., Simon, N.J., 1988. Discontinuous yielding in austenitic steels at low tem-
peratures. Int. Cryogenic Mat. Conference, Boulder 2, 851–863.
Reed, R.P., Walsh, R.P., 1988. Tensile strain rate effect in liquid helium. Adv. Cryo-
genic Eng. Mater. 34, 199–208.
Rizzi, P., Hähner, P., 2004. On the Portevin–Le Chatelier effect: theoretical modeling
and numerical results. Int. J. Plast. 20 (1), 121–165.
Schwarz, R.B., Mitchell, J.W., 1974. Dynamic dislocation phenomena in single crystals
of Cu-10.5-at.%-Al alloys at 4.2 K. Phys. Rev., B 9, 3292–3299.
Seeger, A., 1957. Dislocations and mechanical properties of crystals. Wiley, New
York.
Sgobba, S., Hochoertler, G., 1999. A New Non-Magnetic Stainless Steel for Very
Low Temperature Applications. In: Proceedings of the International Congress
on Stainless Steel: Science and Market, 2. Chia Laguna (I), 6-9 June 1999,
pp. 391–401.
Shumilin, S.E., Isaev, N.V., Zabrodin, P.A., Fomenko, V.S., Grigorova, T.V., Gei-
darov, V.G., 2015. Peculiarities of Plastic Deformation of SPD Al_Li alloy at 0.5 K.
Acta Physica Polonica A 128 (4), 536–539.
Skoczen, B., 2004. Compensation systems for low temperature applications.
Springer-Verlag, Berlin, Heidelberg, New York.
Skoczen, B., 2007. Functionally graded structural members obtained via the low
temperature strain induced phase transformation. Int. J. Solids Struct. 44 (16),
5182–5207.
Skoczen, B., Bielski, J., Marcinek, D., Sgobba, S., 2010. Constitutive model of discon-
tinuous plastic flow at cryogenic temperatures. Int. J. Plast. 55, 198–218.
Skoczen, B., Bielski, J., Tabin, J., 2014. Multiaxial constitutive model of discontinuous
plastic flow at cryogenic temperatures. Int. J. Plast. 26 (12), 1659–1679.
Sylwestrowicz, W., Hall, E.O., 1951. The deformation and aging of mild steel. Proc.
Phys. Soc. B 64, 495–502.
Tabachnikova, E.D., Dolgin, A.M., Bengus, V.Z., Yefimov, Y.V., 1984. Fiz. Metallov Met-
allovedenie 58, 573.
Tabin, J., Pracik,
˛ M., 2015. Methods for identifying dynamic parameters of clip-on
extensometer–specimen structure in tensile tests. Measurement 63, 176–186.
Taylor, G.I., Quinney, H., 1934. The latent energy remaining in a metal after cold
working. Proc. Royal Soc. London 143, 307–326.
Tichonow, A.N., Samarski, A.A., 1963. In: Równania fizyki matematycznej,
państwowe wydawnictwo Naukowe- PWN, Warszawa (in English: equations of
mathematical physics, XVI. Macmillan, New York, p. 765.
Tong, W., Zhang, N., 2007. On serrated plastic flow in an AA5052-H32 sheet. ASME
J. Eng. Mater. Technol. 129, 332–341.
Verel, D.J., Sleeswyk, A.W., 1973. Lüders band propagation at low velocities. Acta
Metall 21, 1087–1098.
Vorobev, E.V., Anpilogova, T.V., 2015. Kinetics of low-temperature discontinuous de-
formation of metals. Cryogenics 68, 59–66.
Vorobev, Е.V., Anpilogova, T.V., 2013. Numerical analysis of the deformation insta-
bility of metals in liquid helium. Comput. Mater. Sci. 68, 66–72.
Yang, S-Y., Tong, W., 2006. A perturbation analysis of the unstable plastic flow pat-
tern evolution in an aluminum alloy. Int. J. Solids Struct. 43, 5931–5952.
612 J. Tabin et al. / International Journal of Solids and Structures 97–98 (2016) 593–612
Zaera, R., Rodríguez-Martínez, J.A., Rittel, D., 2013. On the Taylor–Quinney coefficient
in dynamically phase transforming materials. Application to 304 stainless steel.
Int. J. Plast. 40, 185–201.
Zaiser, M., Hähner, P., 1997. Oscillatory modes of plastic deformation: theoretical
concepts. Phys. Stat. Sol. (b) 199, 267–330.
Zavattieri, P.D., Savic, V., Hector Jr., L.G., Fekete, J.R., Tong, W., Xuan, Y., 2009. Spa-
tio-temporal characteristics of the Portevin–Le Châtelier effect in austenitic steel
with twinning induced plasticity. Int. J. Plast. 25, 2298–2330.
Zhang, J., Jiang, Y., 2004. A study of inhomogeneous plastic deformation of 1045
steel. ASME J. Eng. Mater. Tech. 126, 164–171.
Zhang, Q., Toyooka, S., Meng, Z., Safrapedi, 1999. Study of propagation and pulsation
of clip band using dynamic digital speckle interferometry. SPIE 3385, 389–398.
Zhang, Q.C., Jiang, Z.Y., Jiang, H.F., Chen, Z.J., Wu, X.P., 2005. On the propagation and
pulsation of Portevin–Le Chatelier deformation bands: An experimental study
with digital speckle pattern metrology. Int. J. Plast. 21 (11), 2150–2173.
Ziegenbein, A., Hähner, P., Neuhäuser, H., 2001. Propagating Portevin–LeChatelier de-
formation bands in Cu-15 at.% Al polycrystals: experiments and theoretical de-
scription. Mater. Sci. Eng. A309–A310, 336–339.