IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 19, NO. 3, JUNE 2009 2645
Residual Strain in a
Nb
3
Sn
Strand Mounted on a
Barrel for Critical Current Measurements
Ludovic Thilly, Christian Scheuerlein, Uwe Stuhr, Bernardo Bordini, and Bernd Seeber
Abstract—The strain dependence of the critical properties of
Nb
3
Sn
superconducting strands is a major complication for
critical current
( )
measurements. We report neutron diffraction
measurements that have been carried out at room temperature
and at 10 K in order to determine the strain state of a
Nb
3
Sn
powder-in-tube (PIT) strand mounted on a critical current mea-
surement barrel made out of a Ti-6Al-4V alloy.
Index Terms—Diffraction, superconducting wires and filaments.
I. INTRODUCTION
B
ECAUSE of the uncertainty of the 3-D strain state in
superconducting strands that are mounted on
measurement barrels, standardized
measurements can only
provide relative values that allow a comparison between dif-
ferent strand designs and processing schemes. The stress in
the
strands on measurement barrels is influenced for
instance by the mismatch of thermal expansion coefficients of
the different strand constituents and the measurement barrel
material. The bonding techniques used to fix the strand in order
to inhibit strand movement can influence the
results as well.
Neutron diffraction measurements can been used for deter-
mining the residual strain in the different phases of
strands [1], [2]. Here we report neutron diffraction measure-
ments that have been carried out at room temperature (RT) and
at 10 K in order to determine the strain state of the different
phases in a
powder-in-tube (PIT) strand mounted on a
critical current measurement barrel made of Ti-6Al-4V alloy.
II. E
XPERIMENTAL
A. The Sample
The PIT strand B215 has been manufactured by Shape Metal
Innovation (SMI), B. V., the Netherlands (now European Ad-
vanced Superconductors (EAS), Germany). The wire consists
Manuscript received August 22, 2008. First published June 30, 2009; current
version published July 15, 2009.
L. Thilly is with the Université de Poitiers, SP2MI, 86962 Futuroscope,
France (e-mail: ludovic.thilly@univ-poitiers.fr).
C. Scheuerlein and B. Bordini are with the Accelerator Technology
Department, CERN, CH-1211 Geneva 23, Switzerland (e-mail: Chris-
tian.Scheuerlein@cern.ch; bernardo.bordini@cern.ch).
U. Stuhr is with the ETHZ and Paul Scherrer Institut, CH-5232 Villigen,
Switzerland (e-mail: uwe.stuhr@psi.ch).
B. Seeber is with the Institute of Applied Physics, University of Geneva,
CH-1211 Geneva 4, Switzerland (e-mail: Bernd.Seeber@uni-ge.ch).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TASC.2009.2019598
Fig. 1. Transverse cross section of the reacted PIT B215 strand. In the
backscatter electron image (top) the different strand phases
(Nb
0
Ta) Sn
,
Nb-Ta and Cu can be distinguished. The electron backscatter diffraction image
(bottom) shows the grain orientation distribution in the different strand phases
and allows to distinguish between the fine and coarse grain
(Nb
0
Ta) Sn
areas. Courtesy G. Nolze, Federal Laboratory for Materials Research, Berlin.
of 288 Nb-7.5wt.% Ta tubes that are filled with a powder con-
taining NbSn2 and Sn particles. The Nb-7.5wt.% Ta tubes are
embedded in a high purity Cu matrix. The non-reacted strand
has a diameter of
. The Cu to non-Cu
volume ratio of the non-reacted strand determined by the Cu
weight loss measurement is
.
The strand has been studied after a 44 h-695
heat treat-
ment (HT). A transverse cross section of the fully reacted PIT
strand is shown in Fig. 1. The Electron Backscatter Diffraction
(EBSD) image in Fig. 1 shows the grain size and grain orienta-
tion distribution in the different strand phases.
The room temperature tensile properties of the PIT B215
strand have been determined by tensile tests. The engineering
stress-strain curve of the reacted PIT B215 strand is shown in
Fig. 2. The stress has been partly released every 0.02% in order
to measure the apparent composite E-modulus as a function of
1051-8223/$25.00 © 2009 IEEE
2646 IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 19, NO. 3, JUNE 2009
Fig. 2. Engineering stress-strain curve of the reacted PIT B215 composite
strand measured at RT. The stress has been partly released every 0.02% in order
to measure the apparent composite E-modulus as a function of axial composite
strain. Courtesy B. Rehmer and M. Finn, Federal Laboratory for Materials
Research, Berlin.
axial composite strain. With increasing composite strain the ap-
parent E-modulus determined from the slope of the unloading
curve decreases, possibly because of strain induced damage of
the brittle
phase. The apparent E-modulus measured at
0.2% strain is 116
0.3 GPa.
For
measurements the strand is wound onto a mea-
surement barrel made of a Ti-6Al-4V alloy (see Fig. 3). This
alloy has a similar thermal expansion coefficient as the
phase in the reacted strand. At both ends of the barrel Cu rings
are fixed to which the strand extremities are attached. During
the reaction HT the strand ends are connected to the two Cu
rings by means of two screws and after the HT and before the
critical current measurement the wire ends are soldered to the
Cu rings using a Sn-Ag alloy. The neutron diffraction measure-
ments have been carried out with the strand ends soldered to the
Cu rings.
B. Neutron Powder Diffraction
High resolution neutron powder diffraction measurements
have been performed with the time-of-flight diffractometer
with multiple pulse overlap (POLDI) [3], [4] at the Swiss
spallation neutron source (SINQ) of the Paul Scherrer Institut
(PSI) (Fig. 4).
In order to determine the transverse and axial lattice constants
in the different strand phases, measurements were performed in
two configurations with neutrons scattering at crystallographic
planes either parallel or perpendicular to the strand axis (in the
following referred to as transverse and axial directions, respec-
tively). The instrument was calibrated using Si-powder from the
National Institute of Standards and Technology (NIST standard
Reference Material 640c).
For cooling, the VAMAS sample holder was mounted onto
the cold head of a closed cycle cryo-cooler from CTI Cryo-
genics. The sample holder could be cooled down from room
temperature (RT) to 10 K within 90 minutes.
Fig. 3. Sample holder for critical current measurements of
Nb Sn
supercon-
ducting strands. The PIT strand is wound onto a grooved Ti-6Al-4V cylinder
with an outer diameter of 32 mm. After the reaction HT the wire is soldered
onto Cu cylinders at both ends of the barrel.
Fig. 4. Top view on the POLDI detector, the radial collimator and the sample
area with the attached cryo-cooler.
III. RESULTS
A. Axial and Transverse Lattice Parameters After the First
Cool Down From the Reaction Temperature to RT
The lattice parameters for
, Nb-Ta and Cu
in axial and transverse wire direction measured at RT after
cool down from the processing temperature are summarized in
Table I. The lattice parameters are compared with the nearly
stress free
and Nb-Ta lattice parameters obtained
previously at POLDI using filaments that have been extracted
from the Cu matrix of a similar PIT strand by chemical etching.
The Nb-Ta and
lattice constants measured in
the extracted tubes are 3.3009
and 5.2886 , respectively
THILLY et al.: RESIDUAL STRAIN IN A STRAND MOUNTED ON A BARREL 2647
TABLE I
L
ATTICE
PARAMETERS MEASURED AT
RT FOR THE
DIFFERENT
PHASES IN
THE
Nb Sn
STRAND MOUNTED ON THE
CRITICAL
CURRENT MEASUREMENT
BARREL.THE
NB-TA AND
(Nb
0
Ta) Sn
LATTICE PARAMETERS ARE
COMPARED WITH THE
CORRESPONDING
NEARLY STRESS
FREE LATTICE
PARAMETERS
(3.3009
A
AND 5.2886
A
,RESPECTIVELY)
[2]. It is assumed that these values correspond to the lattice
parameters in a nearly stress free state at RT.
In axial direction the
lattice parameter is about 0.14%
smaller than the nearly stress free parameter while in transverse
direction it is about 0.08% larger. Thus, in the PIT strand
mounted on the VAMAS sample holder at RT
is under about 0.14% compressive axial pre-strain. Similar
differences are found for the respective Nb lattice parameters
(
0.12% and 0.08%).
At RT (before the 1st cool down to 10 K) the Cu lattice pa-
rameter is in axial direction about 0.03% larger than in trans-
verse direction, indicating that the Cu matrix is under axial ten-
sile stress.
B. Influence of First Cool Down From RT to 10 K on the Axial
Pre-Strain in the Different Strand Phases
The lattice parameters of the different strand phases measured
at 10 K after the 1st cool down are presented in Table II. The
lattice parameter variation between the values measured at 10
K and RT is presented as well.
The transverse lattice parameter variations during cool down
from RT to 10 K measured with POLDI are in good agree-
ment with the lattice parameter variations that are predicted
from published thermal contraction factors (
0.18%, 0.15%
and
0.3% for , Nb and Cu, respectively [5]).
During the 1st cool down from RT to 10 K the difference
between the axial and transverse
lattice parameters re-
mains nearly constant, indicating that the residual compressive
strain in axial direction is only slightly changed during
the 1st cool down from RT to 10 K.
TABLE II
L
ATTICE PARAMETERS OF THE DIFFERENT PHASES IN THE
Nb Sn
STRAND
MOUNTED ON THE CRITICAL CURRENT MEASUREMENT BARREL MEASURED
AFTER THE 1ST COOL DOWN OF THE SAMPLE HOLDER TO 10 K. THE
LATTICE PARAMETERS ARE COMPARED WITH THE CORRESPONDING VALUES
MEASURED AT RT, PRIOR TO THE 1ST COOL DOWN
C. Influence of Thermal Cycling on the Axial Pre-Compression
in the Different Strand Phases
In order to study a possible influence of thermal cycling on
the
lattice parameters at 10 K, lattice parameters have
been re-measured after a second cool down from RT to 10 K
and after subsequent heating of the sample holder to RT. The
differences of the lattice parameters measured after the 1st and
2nd cool down to 10 K are relatively small and are approaching
the resolution of the experiment. Nevertheless, the diffraction
results indicate a significant reduction of the
axial com-
pressive pre-strain by thermal cycling between RT and 10 K.
D. Strand Length and Cross Section Changes During the
Reaction HT of the Free Standing PIT Strand
After the strand reaction HT the PIT B215 strand length is
reduced by about 0.15% with respect to the non-heat treated
strand. After a 6 h-200
annealing HT a similar contraction of
0.17% is measured (the estimated accuracy of the contraction
measurement is
0.04%). Thus, it can be concluded that the
strand contraction is caused by the Nb-Ta tube relaxation during
the Cu annealing heat treatment, while the
formation
has only a minor influence. A similar contraction is observed
for cold drawn Nb-Ti/Cu and Nb/Cu binary composite wires.
The axial tensile stress of the Nb-Ta tubes develops during the
cold working of the strand as a result of the plastic mismatch of
Nb-Ta tubes and the Cu matrix [6]. During the reaction HT the
strand diameter increases by 1.6% to
,
which corresponds with a cross section increase of 3.4%.
The non-Cu cross section increases by about 7% (mainly due
to the presence of porosity in the reacted non-Cu part). In con-
trast, the Cu cross section increases only slightly by 0.15% be-
2648 IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 19, NO. 3, JUNE 2009
cause of the strand contraction. Thus, the true Cu to non-Cu
volume ratio in the reacted strand is somewhat smaller than the
ratio determined for the non-reacted strand (
vs. , respectively).
IV. D
ISCUSSION AND
CONCLUSION
The phase in the PIT B215 strand mounted on
the VAMAS sample holder is at RT under about 0.14% axial
compressive pre-strain.
In contrast, the
lattice parameters measured
with POLDI in a straight free standing PIT strand show that
in this case
is not under axial pre-compression,
but may be even under slight axial tension [2]. A slight axial
tension of
in the free standing PIT strand might be ex-
plained by the stronger
thermal contraction with respect
to the contraction of the unreacetd Nb barrier, assuming that the
fully annealed Cu strand part is so soft that it cannot transmit
stresses onto the filaments during the cool down from the pro-
cessing temperature to RT.
The difference between the results obtained for the free
standing PIT strand and the PIT strand reacted on the VAMAS
barrel may be related to the fact that in the latter case the strand
ends are fixed to the measurement barrel, which inhibits the
strand contraction of about 0.15%, which takes place during
the reaction HT of the free standing strand [7], [8].
Thermal cycling between RT and 10 K can affect the axial
pre-strain in the A15 phase. Changes of the axial
pre-strain have been reported to occur also under cyclic tensile
loading [9] and bending [10] of bronze route strands at RT.
A reduction of the axial pre-strain in the A15 phase should
cause a measurable increase of
after the 2nd cool down, with
respect to
measured after the 1st cool down.
A
CKNOWLEDGMENT
The authors thank B. Rehmer and M. Finn from the Federal
Laboratory for Materials Research (BAM), Berlin, for the ten-
sile tests and G. Nolze from the BAM for the EBSD measure-
ments. We are grateful to A. Godeke for stimulating discussions
and suggestions.
The authors acknowledge the PSI for beam time at the POLDI
experiment.
R
EFERENCES
[1] R. Flükiger, W. Schauer, W. Specking, L. Oddi, L. Pintschovius, W.
Mullner, and B. Lachal
, Adv. Cryo. Eng., vol. 28, pp. 361–370, 1982.
[2] C. Scheuerlein, U. Stuhr, and L. Thilly, “In-situ neutron diffraction
under tensile loading of powder-in-tube
Cu
=
Nb Sn
composite wires:
Effect of reaction heat treatment on texture, internal stress state and
load transfer,” Appl. Phys. Lett., vol. 91, no. 4, p. 042503, 2007.
[3] U. Stuhr, Nuclear Instruments and Methods in Physics Research A, vol.
545, pp. 319–329, 2005.
[4] U. Stuhr, H. Spitzer, J. Egger, A. Hofer, P. Rasmussen, D. Graf, A.
Bollhalder, M. Schild, G. Bauer, and W. Wagner, Nuclear Instruments
and Methods in Physics Research A, vol. 545, pp. 330–338, 2005.
[5] N. Mitchell, “Finite element simulations of elasto-plastic processes in
Nb
Sn
strands,” Cryogenics, vol. 45, no. 7, pp. 501–515, 2005.
[6] L. Thilly, P. O. Renault, S. Van Petegem, S. Brandstetter, B. Schmitt, H.
Van Swygenhoven, V. Vidal, and F. Lecouturier, “Evidence of internal
bauschinger test in nanocomposite wires during in situ macroscopic
tensile cycling under synchrotron beam,” Applied Physics Letters, vol.
90, p. 241907, 2007.
[7] M. Pojer and L. Rossi, Development and Characterisation of ITD
Multifilamentary
Nb
Sn
Superconductors for 10–15 T Field Magnets
INFN report, INFN/TC-99/25, 1999.
[8] D. R. Dietderich, J. R. Litty, and R. M. Scanlan, “Dimensional changes
of
Nb
Sn
,
Nb Al
, and
Bi Sr CaCu O
conductors during heat
treatment and their implication for coil design,” Adv. Cryo. Eng., vol.
44b, pp. 1013–1020.
[9] S. Ochiai and K. Osamura, “Influence of cyclic loading at room temper-
ature on critical current of
Nb
Sn
superconducting composite wire,”
Cryogenics, vol. 32, pp. 584–590, 1992.
[10] S. Awaji, H. Oguro, G. Nishijama, P. Badica, K. Watanabe, S. Harjo, T.
Kamiyama, and K. Katagiri, “Neutron diffraction study of prebending
effects for bronze route
Nb
Sn
wires without reinforcement,” IEEE
Trans. Appl. Supercond., vol. 16, no. 2, pp. 1228–1231, 2006.