Physica C 361 (2001) 79±83
www.elsevier.com/locate/physc
Flux jumping and a bulk-to-granular transition in
the magnetization of a compacted and sintered
MgB2 superconductor
S.X. Dou a, X.L. Wang a, J. Horvat a,*, D. Milliken a, A.H. Li a,
K. Konstantinov a, E.W. Collings b, M.D. Sumption b, H.K. Liu a
a
Institute of Superconducting and Electronic Materials (ISEM), University of Wollongong, North®elds Avenue, Wollongong,
NSW 2522, Australia
b
Department of Materials Science and Engineering, LASM, The Ohio State University, College Avenue, Columbus,
OH 43210-1179, USA
Received 5 May 2001; received in revised form 4 June 2001; accepted 4 June 2001
Abstract
In this paper, we report the results of ®eld …H † and temperature …T † dependent magnetization …M† measurements of a
pellet of uniform, large-grain sintered MgB2 . It was found that iron is compatible with MgB2 and can be used as sheath
material for fabrication of Fe-clad MgB2 wires. We show that at low temperatures the size of the pellet and its critical
current density, Jc …H† ± i.e. its M…H † ± ensure low-®eld ¯ux jumping, which of course ceases when M…H † drops below a
critical value. With further increase of H and T the individual grains decouple and the M…H † loops drop to lower lying
branches, unresolved in the usual full M…H † representation. After taking into account the sample size and grain size,
respectively, the bulk sample and the grains were deduced to exhibit the same magnetically determined Jc s (e.g. 105
A/cm2 , 20 K, 0 T) and hence that for each temperature of measurement Jc …H † decreased monotonically with H over the
entire ®eld range, except for a gap within the grain-decoupling zone. Ó 2001 Elsevier Science B.V. All rights reserved.
PACS: 74.60
Keywords: Magnesium boride; Flux jumping; Grain connectivity; Critical current
1. Introduction
The recent discovery of intermediate-temperature superconductivity (ITC) in MgB2 by Akimitsu
et al. [1] and its almost simultaneous explanation
in terms of a hole-carrier-based pairing mechanism
*
Corresponding author. Tel.: +61-2-4221 5722; fax: +61-24221 5731.
E-mail address: jhorvat@uow.edu.au (J. Horvat).
by Hirsch [2,3] has triggered an avalanche of
studies of its structural, magnetic and transport
properties [4±8]. In spite of the signi®cant advancement made on critical current density of HTS
conductors the best transport Jc of Ag/Bi2223
tapes is still limited by the grain connectivity. The
study on the intergranular and intragranular critical current densities revealed that the former is
smaller than the latter by at least one order of
magnitude for the state-the-art Ag/Bi2223 tapes
[9]. This characteristic feature of HTS conductors
0921-4534/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 1 - 4 5 3 4 ( 0 1 ) 0 0 7 8 2 - 1
80
S.X. Dou et al. / Physica C 361 (2001) 79±83
is clearly evident by the rapid drop of Jc in lowmagnetic ®elds which is de®ned as weak link regime. Preliminary results on the newly discovered
40 K superconductor MgB2 reported by Larbalestier et al., have indicated that Jc of this material
is not limited by grain boundary weak links as
HTS [6]. The high-critical current density of 105 A/
cm2 at 11 K and low ®eld and 2 104 A/cm2 at 20
K and 1 T have been reported on the bulk MgB2
samples using magnetisation measurements [5,6].
As a further contribution to the ®eld, we design a
synthesis method to produce bulk MgB2 samples
in which large grains of uniform size were formed.
The grains consist of well connected MgB2 single
crystals. The results of magnetisation measurements performed on these samples allow us to
compare and evaluate the intergranular and intragranular critical currents in MgB2 .
2. Experimental details
Polycrystalline samples of MgB2 were prepared
by conventional solid state reaction. High-purity
Mg and B (amorphous) were mixed and ®nely
grounded well, then pressed into pellets 10 mm in
diameter with 1±2 mm thickness. Extra Mg was
added in order to avoid Mg loss at high temperatures. These pellets were placed on an iron plate
and covered with iron foil then put into a tube
furnace. The samples were sintered at temperatures between 700°C and 1000°C for 1±14 h. A
high-purity Ar gas ¯ow was maintained throughout the sintering process. The sample purity was
found to be very sensitive to oxygen contamination, any leaks in the furnace resulting in the oxidation of Mg to MgO. The purity of the Mg
starting material was also a key factor in¯uencing
the quality of the resulting MgB2 .
The reason for using a wide range of sintering
temperatures and time is to obtain samples with
di€erent grain size and phase compositions as the
sintering time and temperature determine both the
grain size and the phase content. It was found that
the grain size and MgO content increased with
increasing sintering temperature and sintering
time. The wide range of sintering temperatures and
time allows the grain size to change from few mi-
crons to 200 lm and MgO content from 5% to
40%. These samples are desirable for study of
grain decoupling and ¯ux pinning e€ect of MgO
inclusions.
3. Results and discussion
3.1. Compatibility of Fe with MgB2
Phase purity was determined by XRD and grain
sizes by SEM. The XRD pattern for a sample
sintered for 4 h/800°C is shown in Fig. 1, which
shows mostly MgB2 with only a small level of
MgO contamination. The SEM results for the
same sample are illustrated in Fig. 2. The grain size
is of the order of 200 lm. Single crystals forming
the grains were up to 20 lm in size. It is likely that
the large size grains were formed by a partial
melting process. The sample appears to be dense,
hard, and strong. MgO as impurity phase is largely
distributed along the grain boundaries as evidenced by the much lighter contrast at grain
boundaries as MgO is non-conductive compared
with MgB2 .
Chemical analysis was performed using ICPoptical emission spectroscope on the sample by
stripping all the metal elements from MgB2 pellet.
The results show that Fe content is less than
Fig. 1. XRD pattern for the sample showing that all the peaks
can be indexed according to the published crystallography data
of hexagonal MgB2 except for the small peaks of MgO as impurity phase.
S.X. Dou et al. / Physica C 361 (2001) 79±83
Fig. 2. SEM images showing that the sample consists of large
and uniform grains with size ranging from 0.15 to 0.3 mm.
0.01 at.% which is consistent with the published
work on Mg±Fe binary alloy phase diagram that
Mg and Fe have meager mutual solubility, about
0.01% Fe in Mg at its M.P. and 0.2 at.% Mg in Fe
[10]. This and the Tc and Jc results below suggest
that the processing of Mg ‡ B or MgB2 powders in
an Fe container would be accompanied by little
contamination. Fe and Fe alloys may be potential
materials for fabrication of metal-sheathed wires.
3.2. Magnetization and ¯ux jump
The transition temperature Tc was determined
by using AC susceptibility in an AC ®eld of amplitude 1 Oe and frequency 117 Hz. The results,
depicted in Fig. 3, indicate a Tc of 38.2 K with a
Fig. 3. AC susceptibility, measured with frequency and amplitude of AC ®eld of 117 Hz and 1 Oe, respectively.
81
Fig. 4. Magnetic hysteresis loops, measured with the ®eld
sweep rate 50 Oe/s, at di€erent temperatures.
very sharp transition, whose width of only 2 K
again con®rms the high phase purity of the sample.
The magnetization of a 3:5 2:5 1:3 mm3
sample (cut from the initially prepared pellet) was
measured over a temperature range of 5±35 K
using a PPMS induction magnetometer in a timevarying magnetic ®eld of sweep rate 50 Oe/s and
amplitude 9 T. The magnetization results are depicted in Fig. 4. Immediately visible is a band of
¯ux jumping, occurring over a narrow ®eld range
(H < 1 T) at 5 and 10 K, that appears to truncate the maximum attainable magnetization. The
Swartz and Bean sample size limitation on adiabatic ¯ux jump stability [11] can be transformed
into a magnetization limit of the form [12]:
r
0:2
3
Mmax ˆ
109 bT 3 DM0 …T †
1†
3p
p
where bT 3 represents the low-temperature speci®c
heat and DM0 …T † the magnetization temperature
dependence in the form DM=…dDM=dT †. The lack
of speci®c heat data prevents a direct prediction of
Mmax . However a back calculation using the experimental Mmax predicts a speci®c heat for MgB2
at 5 K of 3 10 6 J/cm3 .
Evidence to be presented below indicates a sample, initially a bulk superconductor, that breaks
down into a granular assembly beyond a certain
critical value of ®eld and temperature. Such a
breakdown will also occur during a ¯ux jump. The
82
S.X. Dou et al. / Physica C 361 (2001) 79±83
¯ux jumps depicted are partial, rather than complete. This suggests that a jump is terminated when
the magnetization of the outer shell grains, and
that of the residual core, drop below the critical
value. Within the ¯ux jump zone of Fig. 4 (for
T < 10 K and H < 1 T) we note the magnetization
dropping to about one-®fth of its initial (virtual)
value during the partial jump.
3.3. Magnetic critical current density
A magnetic critical current density can be derived from the height of the magnetization loop
DM using a suitable variant of the ``semi-bean''
relationship DM ˆ kJc d, where k is a constant and
d the thickness that the sample presents to the
applied ®eld. DM itself vs. H is plotted semi-logarithmically in Fig. 5. Here we note what are
essentially a high-®eld and low-®eld regions of
curves, the transition between which is complete at
a fairly constant value of DM. Since the low-®eld
family covers three orders of magnitude in DM, the
high-®eld sets are not visible in the usual linear
type of M…H † loop (Fig. 4).
The transition from low-®eld step to high-®eld
step corresponds to a drop in DM by more than an
order of magnitude. It is interesting to note that
from the image of SEM of the sample the average
grain size is less than one tenth of the sample size.
It is evident that the drop in DM is attributable to
Fig. 5. Width of the hysteresis loops, DM, as a function of ®eld,
with the sweep rate of the ®eld 50 Oe/s.
the breakdown of grain matrix as a result of ¯ux
penetration to the grain boundaries which may
contain impurities. In the low-®eld region the
current circulates mainly over the entire sample
size, which is the intergranular current. The induced current circulates only in the individual
grains in the high-®eld region, which is the intragranular current [9]. The transition from low-®eld
step to high-®eld step represents a transition from
intergranular current regime to intragranular current regime.
Based on the full sample size at low ®elds and
the grain size at high ®elds, the magnetic Jc is
calculated using the relationship for a plate in
perpendicular ®eld: Jc ˆ 20DM=…a a2 =3b†. For
low ®elds, the lateral dimensions of the sample are
used: a ˆ 0:25 cm, b ˆ 0:35 cm, whereas for the
high ®elds the average dimension of the grains is
used: a ˆ 170 lm, b ˆ 260 lm. No calculation is
applied to the intermediate-®eld transitional zone
wherein the grain decoupling is taking place. The
results are depicted in Fig. 6 wherein we note a
continuous ®eld dependence of Jc throughout the
entire ®eld range. The same results were obtained
on two more samples, prepared under slightly
di€erent conditions, i.e. containing a small degree
of impurities.
Granular structure seems to be a common
property of MgB2 pellets and powder [6,13,14]. In
Fig. 6. Critical current density vs. ®eld, with corrections for
di€erent screening length of the currents in small and high
®elds.
S.X. Dou et al. / Physica C 361 (2001) 79±83
most cases, the size of the grains is smaller than for
our samples. This was probably the reason why
other groups did not report the step in DM vs. H
for their samples. The connections between the
crystals in the grains are more resistive to the ®eld
penetration than the connections between the
grains, as evidenced by Figs. 5 and 6. Further insights into the mechanisms of grain formation and
physical properties of the grains could help improve the ®eld dependence of Jc for bulk MgB2 .
Our results are consistent with magneto-optical
imaging of bulk MgB2 [6], where regions of strong
superconductivity of 150 lm in size were identi®ed.
However, our samples exhibit an order of magnitude higher intergrain, as well as intragrain Jc .
Further, it was noted that there is no texturing of
crystals in the grains [6,14]. In contrast to this, our
preliminary results with torque magnetometer indicate an anisotropy of Jc in the plate-like grains of
about 2.5.
4. Summary
In summary, we presented evidence for decoupling of the grains in sintered MgB2 , both through
partial ¯ux jumping and a step in ®eld dependence
of Jc . Similar step was also observed in the hightemperature superconductors. However a signi®cant di€erence here is that the step occurs at much
higher ®elds and it corresponds to a much smaller
drop of Jc (Fig. 5). Therefore, grain connections in
MgB2 can sustain rather large Jc , in contrast to
high-temperature superconductors. The ¯ux jump
and step in Jc …H † was observed thanks to large
value of Jc and large size of the grains. Our samples gave Jc of 105 A/cm2 at 20 K and zero ®eld, as
obtained from magnetic hysteresis loops with
sweep rate of ®eld of 50 Oe/s. The value of Jc
would be much larger at lower temperature.
However it was not possible to obtain it there from
the hysteresis loop because of the ¯ux jump. The
connections between the single crystals in the
grains, which are a common property of bulk and
MgB2 powder, are much more resistant to the ®eld
penetration than the connections between the
83
grains (Figs. 5 and 6). This shows a way for improving the ®eld dependence of Jc for bulk MgB2 .
In addition, the present results show that iron is
compatible with MgB2 showing non-poisoning to
MgB2 at reaction temperatures and can be used as
sheath material for fabrication of Fe-clad MgB2
wires.
Acknowledgement
The authors would like to thank Australian
Research Council for ®nancial support.
References
[1] J. Nagamatsu, N. Nakagawa, T. Muranaka, Y. Zenitani,
J. Akimitsu, Nature (London) 410 (2001) 63.
[2] J.E. Hirsch, Presented at the Third International Conference on New Theories, Discoveries and Applications of
Superconductors and Related Materials (New3SC-3),
Hawaii, January 2001, Physica C 364±365 (2001), in press.
[3] J.E. Hirsch, Phys. Lett. A 282 (2001) 392.
[4] C.U. Jung, M.-S. Park, W.N. Kang, M.-S. Kim, S.Y. Lee,
S.-I. Lee, Physica C 353 (2001) 162.
[5] Y. Takano, H. Takeya, H. Fujii, H. Kumakura, T.
Hatano, K. Togano, preprint.
[6] D.C. Larbalestier, M.O. Rikel, L.D. Cooley, A.A. Polyanskii, J.Y. Jiang, S. Patnaik, X.Y. Cai, D.M. Feldmann,
A. Gurevich, A.A. Squitieri, M.T. Naus, C.B. Eom, E.E.
Hellstrom, R.J. Cava, K.A. Regan, N. Rogado, M.A.
Hayward, T. He, J.S. Slusky, P. Khalifah, K. Inumaru,
M. Haas, Nature 410 (2001) 186.
[7] D.K. Finnemore, J.E. Ostenson, S.L. Bud'ko, G. Lapertot,
P.C. Can®eld, Phys. Rev. Lett. 86 (2001) 2420.
[8] S.L. Bud'ko, G. Lapertot, C. Petrovic, C.E. Cunningham,
N. Anderson, P.C. Can®eld, Boron isotope e€ect in
superconducting MgB2 , Phys. Rev. Lett. 86 (2001) 1877.
[9] K. Muller, C. Andrikidis, H.K. Liu, S.X. Dou, Phys. Rev.
B 50 (1994) 10218±10224.
[10] M. Hansen, Constitution of Binary Alloys, McGraw-Hill,
NY, 1958, pp. 662±663.
[11] P.S. Swartz, C.P. Bean, J. Appl. Phys. 39 (1968) 4991±
4998.
[12] E.W. Collings, M.D. Sumption, E. Lee, IEEE Appl.
Supercond. 11 (2001) 2567.
[13] F. Simon, A. Janossy, T. Feher, F. Muranyi, S. Garaj, L.
Forro, C. Petrovic, S.L. Bud'ko, G. Lapertot, V.G. Kogan,
P.C. Can®eld, cond-mat/0104557.
[14] Y. Boguslavsky, G.K. Perkins, X. Qi, L.F. Cohen, A.D.
Caplin, Nature 410 (2001) 563.