Effects of MgO impurities and micro-cracks on the critical current density of Ti-sheathed MgB2 wires
ABSTRACT Ti-sheathed monocore MgB2 wires with improved magnetic critical current density (Jc) have been fabricated by in situ powder-in-tube (PIT) method and characterized by magnetization, X-ray diffraction, scanning electron microscopy and electrical resistivity measurements. For the best wire, the magnetic Jc values at 5 K and fields of 2 T, 5 T, and 8 T are 4.1 × 105 A/cm2, 7.8 × 104 A/cm2, and 1.4 × 104 A/cm2, respectively. At 20 K and fields of 0.5 T and 3 T, the Jc values are about 3.6 × 105 A/cm2 and 3.1 × 104 A/cm2, respectively, which are much higher than those of the Fe-sheathed mono-core MgB2 wires fabricated with the same in situ PIT process and under the same fabricating conditions. It appears that the overall Jc for the average Ti-sheathed wires is comparable to that of the Fe-sheathed wires. Our X-ray diffraction and scanning electron microscopy analysis indicates that Jc in the Ti-sheathed MgB2 wires can be strongly suppressed by MgO impurities and micro-cracks.
- [Show abstract] [Hide abstract]
ABSTRACT: We report the phase evolution, microstructure, and critical current density (Jc) of the SiC doped MgB2 superconductors. In our study, all samples were fabricated by hot pressing with a heating rate of 200 °C min−1. The results show that the reaction of 2Mg+SiC = Mg2Si+C can occur at about 500 °C, about 50 °C lower than the formation temperature of MgB2. On the other hand, according to the experimental results and thermodynamic calculations, boron (B) does not react with SiC to form B4C and Si, neither does MgB2 react with SiC to form Mg2Si and B4C. The MgB2 phase was formed via both a solid–solid reaction (during the heating process between about 500 and 650 °C) and a liquid–solid reaction (>650 °C) after the melting of Mg, and the two reactions resulted in differences in the grain size of the MgB2. From scanning electron microscopy, the Mg2Si particles are homogeneously distributed within the MgB2 matrix, with particle sizes ranging from 35 to 230 nm. From the perspective of superconductivity, the C substitution results in strong electron scattering centers in the MgB2 structure. It reduces the electron mean free path and thus may significantly enhance the magnetic Jc. The peak Jc of the 5 wt% SiC doped MgB2 reaches above 106 A cm−2 at 5 K, and decreases slowly with increasing field, remaining high, above 105 A cm−2, at 7 T.Superconductor Science and Technology 06/2009; 22(7):075014. · 2.76 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Some of the interesting features of MgB2 such as low cost, compatibility with existing processing methods and intermediate critical temperatures induced the scientific community to invest in its development with large efforts. In order to promote the use of this compound for military and aerospace activities requiring lightweight devices, we tested titanium as a sheath material. We fabricated Ti-sheathed MgB2 through the powder-in-tube (PIT) process. These wires showed similar results to those sheathed with steel. At 4 K and self field, the critical current density Jc is well above 7x105 A/cm2. Analyses of the interface reaction layer between titanium and MgB2 were carried out using XRD, SEM-EDS and EMPA equipment. In addition to wire development, investigation of the so-called quench events is needed. Quench detection and protection systems depend on how fast the normal zone propagates through the coil, after a disturbance occurs. During this doctorate program two codes were developed. The first is designed to compute the magnetic field of a multilayer solenoid, find the maximum field the superconductor is exposed to and define the critical quench current. Results are used to minimize the wire length required to reach the desired magnetic field. The second code simulates quench events by solving the heat transfer equation for a monodimensional case under adiabatic conditions. Among others, two solenoids were built and tested in order to investigate quench events under different conditions. The first was made with 500 m of commercial MgB2 tape and generates up to 1 T at 16 K. The second was wound with 40 m of the same tape. This thesis was supported by the State of Texas through the Texas Center for Superconductivity and by Ad Astra Rocket company (Houston, TX).01/2007;
Effects of MgO impurities and micro-cracks on the critical
current density of Ti-sheathed MgB2wires
G. Lianga,*, M. Alessandrinib, F. Yenc, M. Hannab, H. Fanga,b, C. Hoyta,
B. Lvd, J. Zengc, K. Salamab
aDepartment of Physics, Sam Houston State University, 1908 Ave. J, Huntsville, TX 77341, USA
bDepartment of Mechanical Engineering, University of Houston, 4800 Calhoun Road, Houston, TX 77204, USA
cTexas Center for Superconductivity, University of Houston, Houston, TX 77204, USA
dDepartment of Chemistry, University of Houston, Houston, TX 77204, USA
Received 27 November 2006; received in revised form 5 February 2007; accepted 21 February 2007
Available online 3 March 2007
Ti-sheathed monocore MgB2wires with improved magnetic critical current density (Jc) have been fabricated by in situ powder-in-tube
(PIT) method and characterized by magnetization, X-ray diffraction, scanning electron microscopy and electrical resistivity measure-
ments. For the best wire, the magnetic Jcvalues at 5 K and fields of 2 T, 5 T, and 8 T are 4.1 · 105A/cm2, 7.8 · 104A/cm2, and
1.4 · 104A/cm2, respectively. At 20 K and fields of 0.5 T and 3 T, the Jcvalues are about 3.6 · 105A/cm2and 3.1 · 104A/cm2, respec-
tively, which are much higher than those of the Fe-sheathed mono-core MgB2wires fabricated with the same in situ PIT process and
under the same fabricating conditions. It appears that the overall Jcfor the average Ti-sheathed wires is comparable to that of the
Fe-sheathed wires. Our X-ray diffraction and scanning electron microscopy analysis indicates that Jcin the Ti-sheathed MgB2wires
can be strongly suppressed by MgO impurities and micro-cracks.
? 2007 Elsevier B.V. All rights reserved.
PACS: 74.70.Ad; 84.71.Mn; 74.25.Sv; 75.60.Ej; 61.10.Nz
Keywords: MgB2superconductor; Magnetic Jc; Hysteresis loop; X-ray diffraction; SEM image
For certain type of applications, such as in the growing
sector of electric space propulsion, lightweight supercon-
ducting magnets are preferred. Lightweight magnets
require the use of light metals as the sheath materials in
the powder-in-tube (PIT) process to make MgB2 wires
[1,2]. The most favorable sheath metal should at least meet
the following requirements: low mass density, excellent
chemical compatibility (i.e., non-reactive or nearly non-
reactive with MgB2, B, or Mg during sintering process),
non-magnetic, and appropriate mechanical strength for
cold work. In the last five years, PIT MgB2wires/tapes
have been fabricated with the use of different sheath metals,
such as iron (Fe), nickel (Ni), copper (Cu), silver (Ag), nio-
bium (Nb), and tantalum (Ta) [3–10]. For un-doped MgB2
wires sheathed with some of these metals such as Fe, Jcval-
ues as high as 3 · 105A/cm2have been achieved at 5 K and
in zero field [4,9–11]. However, none of these sheath metals
are suitable for making lightweight MgB2wires because
their mass densities are all greater than 7.8 g/cm3. Titanium
(Ti) seems to be a very promising sheath material for light-
weight and high JcMgB2wires due to its very low mass
density (4.57 g/cm3), excellent chemical stability, non-mag-
netic property [12,13], and high mechanical strength (50%
harder than Fe). Recently, we have successfully fabricated
0921-4534/$ - see front matter ? 2007 Elsevier B.V. All rights reserved.
*Corresponding author. Tel.: +1 936 294 1608; fax: +1 936 294 1585.
E-mail address: email@example.com (G. Liang).
Physica C 457 (2007) 47–54
Ti-sheathed MgB2wires with Jcvalues as high as 4 · 105A/
cm2at 5 K and zero field [14,15]. Currently, much effort has
been made by our groups to improve the Jc. In this paper,
we present the best results on the magnetic Jcfor the newly
developed un-doped Ti-sheathed MgB2 wires, together
with the results from electrical resistivity, X-ray diffraction
(XRD), and scanning electron microscopy (SEM) measure-
ments. This study also provides the first analysis on the
micro-cracks and MgO impurities in Ti-sheathed MgB2
wires to address the observed degradation of Jcin some
of these wires.
The Ti-sheathed monocore MgB2wires were fabricated
by an in situ powder-in-tube (PIT) method, which has been
described in detail elsewhere [14,15]. In this paper, we only
report results on un-doped MgB2wires sintered at 800 ?C
for 30 min. The milled Mg + 2B powder particles are very
uniform in size and smaller than 2 lm, as checked by SEM
. The cross-sectional areas for the wires and cores are
about 1 mm2(=1 mm · 1 mm) and 0.126 mm2, respec-
tively. The XRD patterns were obtained using a Rigaku
X-ray diffractometer with Cu Karadiation. All of the pat-
terns were calibrated using Si powder as the standard. The
fitting of the XRD peaks was done using software Jade 6.1
provided by Material Data Inc. (MDI). The SEM images
were taken using a JEOL JSM-6330F Field Emission Scan-
ning Electron Microscope. The temperature (T) dependent
resistivity was measured by a standard four-probe dc tech-
nique in a temperature range of 10–100 K. The hysteresis
loops of the magnetization (M) were measured using an
Oxford Instrument’s Maglab 9-Tesla Vibrating Sample
Magnetometer (VSM). The temperature dependent magne-
tization was measured in both zero-field-cooled (ZFC) and
field cooled (FC) modes using a magnetic properties mea-
surement system (MPMS) magnetometer from quantum
3. Results and discussion
Fig. 1 displays the magnetization hysteresis loops,
M(H), measured at 5 K, 20 K, and 30 K for two wire sam-
ples: sample A and B. The difference between the wire sam-
ple A and B is: wire A was made with relatively slower
rolling speed and is micro-crack free; whereas wire B was
made with faster rolling, resulting in producing micro-
cracks (see the SEM result below) in the wire. In the
M(H) hysteresis loop measurements, each piece of wire
sample was prepared with almost identical dimension to
avoid the sample-size effect on Jc. The sample A con-
sisted of six pieces of short wires, each has a dimension
of a · b · c = 0.033 cm · 0.038 cm · 0.73 cm for the MgB2
core. The sample B had a dimension of a · b · c =
0.034 cm · 0.037 cm · 0.75 cm for its MgB2core. The wire
samples were oriented with the c-axis parallel to the direc-
tion of the applied magnetic field. In Fig. 1, the solid curves
are the M(H) data measured with the VSM at a very small
step of 33 Oe. For sample A, it is observed from Fig. 1a
that the M(H) loop measured at 5 K and low fields
(between ?1.7 T and 1.7 T) displays a plateau like shape
with some spikes. This behavior can be attributed to the
flux jumps caused by localized warming due to very high
critical current density and very low specific heat .
For a superconductor, when the ratio of Jcto specific heat
(C) is greater than a critical value, flux jumping occurs .
For MgB2, since Jcincreases and C decreases  with the
decrease of the applied field, flux jumping could occur at
low fields. Such flux jumps can cause the fluctuation of
the magnetization and magnetic critical current density,
as observed previously by some groups [16,19,20] for the
Fe-sheathed MgB2wires. To verify this flux jumping, we
measured the M(H) loop at 5 K and up to ±20 kOe (or
l0H = ±2 T) using the MPMS magnetometer. The data
points are shown in Fig. 1a by the open circles. Indeed,
Fig. 1. The solid curves are the M(H) curves measured for samples A and
B at different temperatures using VSM. The open circles in (a) are the data
points measured at 5 K using MPMS.
G. Liang et al. / Physica C 457 (2007) 47–54
we see that the data points are fluctuating in the plateau
region and match well with the solid M(H) curve in the
regions beyond the plateau. In contrast with the loop mea-
sured at 5 K, the M(H) hysteresis half loops measured at
20 K and 30 K (see Fig. 1b) do not show any flux-jump-
caused fluctuation due to both the substantial decrease in
Jc and increase in specific heat at these temperatures.
Fig. 1c shows that for sample B, the flux-jump-caused mag-
netization fluctuation in the 5 K M(H) curve below 1 T is
much smaller than the fluctuation for the sample A, sug-
gesting a much lower Jcfor sample B.
Shown in Fig. 2 are the field dependent magnetic Jc(H)
curves for these two samples. The magnetic Jcwas calcu-
lated by Jc= 20DM/[a(1 ? a/3b)] [15,16,21] from the Bean
critical state model , where DM (in emu/cm3) is the dif-
ference between the upper and lower branches of the M(H)
curve. There is no contribution to DM from the paramag-
netic Ti-sheath [12,13] because its magnetization is revers-
ible with applied field . Fig. 2 shows that for sample
A, the magnetic Jcat 5 K and fields of 2 T, 5 T, and 8 T
are about 4.1 · 105A/cm2, 7.8 · 104A/cm2, and 1.4 ·
104A/cm2, respectively. At 20 K and fields of 0.5 T and
3.1 · 104A/cm2. In Table 1, we summarize the values of
the magnetic Jcat different temperatures and fields for sam-
ple A and sample B. Fig. 2 shows that at 5 K and 20 K, the
Jcfor sample A at any field is much higher than that for
sample B. At 30 K, Jcof sample A is higher than that of
sample B for l0H < 0.94 T and lower than that of sample
B for l0H > 0.94 T.
For sample A, since the Jcat 5 K and in the low field
regions (for l0H 6 1.7 T) is extremely high, the Bean model
does not apply due to significant flux jumps. For sample B,
Fig. 2 shows that its Jcat 5 K is only about half of Jcof
sample A, the flux jump caused fluctuation in the Jccurve
is very small, below 1 T. Thus, if we use the Jcvalues of the
sample B below 1.7 T and assume that the variation rates
of Jcbelow 1.7 T are the same for both sample A and B,
then the Jcat 5 K and 0 T for sample A is estimated to
be about 1.0 · 106A/cm2. The dashed curve section in
Fig. 2 represents the estimated values of Jcat 5 K and
below 1.7 T for sample A.
To evaluate the performance of the Ti-sheath on Jc, we
would like to compare the magnetic Jcbetween the Ti-
sheathed and Fe-sheathed MgB2wires. Since magnetic Jc
could be affected by certain factors such as the fabricating
condition, doping content (such as SiC ), and the qual-
ity of the precursor powder, the most appropriate
Fe-sheathed wires for such comparison should be those
fabricated the same way as that used for the Ti-sheathed
wires. Unfortunately, no other groups, except ours, have
fabricated Fe-sheathed, un-doped MgB2 wires with the
same in situ PIT process, fabricating equipment (such as
the groove rolling mill), and fabricating conditions in terms
of Mg and B powders, milling time (2 h), sintering temper-
ature (800 ?C), and size of the wires (1 mm · 1 mm cross
section) . For these Fe-sheathed wires, the magnetic Jc
was measured at 20 K and in a field range from 0 to 3 T
, and the data are presented in Fig. 2 by the open circles
and listed in Table 1. It can be seen from Fig. 2 that for
fields in the range of l0H < 2.5 T, the Jcvalues for the
Fe-sheathed wires are between the values for samples A
and B; for l0H > 2.5 T, the Jcvalues for the Fe-sheathed
wire are lower than values of both sample A and B. This
comparison indicates that even though the best Ti-sheathed
3.6 · 105A/cm2
Fig. 2. The field dependent magnetic Jccurves measured at temperatures
5 K, 20 K, and 30 K for wire sample A (thicker curves) and sample B
(thinner curves). The dashed curve section below 1.7 T for the 5 K Jccurve
of sample A represents the estimated values (see text). The open circles are
the data points of magnetic Jcmeasured at 20 K for a Fe-sheathed MgB2
wires (Ref. ).
The values of the magnetic Jcfor sample A and B at two temperatures, 5 K and 20 K, and at different applied magnetic fields
Jc(105A/cm2) at 5 K
Jc(105A/cm2) at 20 K
0 (T) 2 (T) 5 (T)8 (T)0.5 (T)2 (T)3 (T)4 (T)
Ti-sheathed MgB2wire sample A
Ti-sheathed MgB2wire sample B
For comparison, the Jcvalues for the Fe-sheathed MgB2wires, fabricated previously by us (Ref. ) with the same fabricating conditions, are also listed.
aEstimated (see text).
bFrom Ref. .
G. Liang et al. / Physica C 457 (2007) 47–54
wire (wire A) has a much higher Jcthan the Fe-sheathed
wires, the overall performance of the Ti-sheath on Jcfor
the average Ti-sheathed wires (including wire B) is compa-
rable to the performance of the Fe-sheath.
To explain the observed big difference in Jc between
sample A and sample B, detail analysis on the crystal
phases and micro-structure of the samples is needed.
Fig. 3 shows the powder XRD patterns of the sample A
and B, together with the patterns of three reference com-
pounds: MgB2, MgO, and Ti (all ?325 mesh, from Alfa
Aesar). All of the major peaks in the patterns of sample
A and B can be indexed with the MgB2hexagonal struc-
ture, indicating that the core material of the wires is in
nearly pure MgB2phase. For the pattern of sample A, a
weak impurity peak due to the MgB4phase is observed
located at 2h = 35.51?, which is not seen from the pattern
for sample B. The broad peak located at 2h = 62.33? can
be attributed to the second strongest peak of the cubic
MgO phase, i.e., the (220) peak (also see Fig. 4). The stron-
gest peak of the MgO phase is located at 2h ? 42.9? and
overlaps with the neighboring MgB2(101) peak located
at 2h ? 42.4?, causing some additional broadening effect
on this peak. Such line broadening was confirmed by our
detail analysis on the full width at half-maximum (FWHM)
of the peaks. In the presence of oxygen, MgO impurities
could be formed by the reaction 2Mg + O2! 2MgO dur-
ing the sintering process. The main source of the oxygen
could be from the air trapped in the cores during the
short-time crimping sealing of the end of the Ti tube (with
Mg + 2B mixture packed in) in air. The residual O2con-
tained in commercial argon gas could be the secondary
source. Compared with the pattern of the Ti powder, the
two weak impurity peaks located at 2h = 38.40? and
40.12? in the patterns of sample A and B can be identified
as the (002) and (101) peaks of the a-phase Ti. The inten-
sities of these two peaks are less than 2% of the intensity of
the MgB2(101) peak in each pattern. We believe that the
Ti impurities were introduced into the XRD samples dur-
ing the preparation of the XRD powder slide and thus they
are extrinsic. When the wires were peeled open by a knife,
Fig. 3. Powder XRD patterns for the core materials of the Ti-sheathed
MgB2wire samples A and B. For comparison, the XRD patterns for the
reference compounds MgB2, MgO, and Ti are also shown.
Fig. 4. XRD patterns in the 2h range of 59? 6 2h 6 65? for sample A,
sample B, and a mixture of MgB2and MgO with 1:1 molar ratio. The
patterns are normalized to the intensity of the MgB2(110) peak in each
pattern. Note that the scale for the intensity of the pattern of the mixture
of MgB2and MgO is reduced by a multiplying factor of 0.25.
Fig. 5. EDS spectrum taken for the entire cross-sectional area of the
MgB2core of sample B.
G. Liang et al. / Physica C 457 (2007) 47–54
some very small Ti particles were stripped away from the
Ti-sheath and fell into the core material. Fig. 5 shows an
energy dispersive spectrometry (EDS) spectrum for the
core of the wire sample B, and it indeed confirms the
absence of the Ti impurities in the cores of the as-sintered
It can be seen from the expanded patterns in Fig. 4 that
the intensity of the MgO (220) peak for sample B is stron-
ger than that for sample A, indicating a relatively larger
content of MgO in sample B. To better understand how
the concentration of the MgO impurities correlates to the
variation of Jcin the MgB2wires, it is very necessary to
have a good estimate about the MgO concentration in
these two samples. For this purpose, we performed a quan-
titative analysis on the MgO (220) and the MgB2(102)
peaks by a similar technique which we recently used to ana-
lyze the concentration of MgCu2impurities in Cu-sheathed
MgB2wires . Fig. 4 shows the XRD patterns of the two
samples in comparison with a pattern measured on a mix-
ture of MgB2and MgO, which has a molar (mol) ratio of
1:1. The pattern for the mixture is used to obtain the cali-
bration line . This pattern for the mixture shows that
the intensity of the MgO (220) peak is much stronger
(about ten times stronger) than that of the MgB2(102)
peak, even though the molar content of these two compo-
nents, MgB2and MgO, are equal. Due to the particle size
effect , the peaks for the sample A and B (with much
finer size) are much wider than that for the powders of
the mixture (particle size ?325 mesh). Note that the pattern
for the Ti powder (in Fig. 3) is not included in Fig. 4,
because for the patterns of sample A and B, the intensity
of the Ti (110) peak located at 2h ? 63? is estimated (using
the intensities of the relevant peaks of the patterns in
Fig. 3) to be only about 3% of the intensity of the MgB2
(102) peak. The percentage weight of MgO in the mixture
of MgO and MgB2 can be expressed as WMgO= KIrel,
where the relative intensity Irelis defined as the ratio of
the intensity of the MgO (220) peak to that of the MgB2
(102) peak, i.e., Irel= I220(MgO)/I102(MgB2); K is a con-
stant which depends only on the Miller indices of the two
selected peaks and can be determined experimentally by
the slope of the WMgOvs. Irelcalibration line . The rela-
tionship between the molar (mol)%, MMgO, and weight
(wt.)%, WMgO, of MgO in a mixture of MgO and MgB2
is given by formula MMgO= WMgO/(r + WMgO? rWMgO),
where r = 0.8776 is the molar mass ratio of MgO to MgB2.
To determine the intensities and peak positions (2h)0of the
overlapping MgB2(110), MgO (220), and MgB2(102)
peaks in the patterns of Fig. 4, these peaks were fitted with
the pseudo-Voigt profile shape functions using the Jade 6.1
XRD analyzing software package. The fitting results are
listed in Table 2 and the values of the Irel, WMgO, and
MMgOare calculated from the fitting result and summa-
rized in Table 3. Table 3 shows that the molar MgO con-
centration for sample B is about 4.8%, which is three
times of the MgO concentration in sample A.
Fig. 6 shows the SEM images for these two samples. The
images clearly show that there exist a large number of
cracks in sample B and no cracks in sample A. A close
inspection of the outer surface of the Ti-sheath of the sam-
ples also revealed that some cracks were developed in cer-
tain portion of the Ti-sheath of wire sample B but not in
the wire sample A. These micro-cracks could be produced
due to the relatively faster feeding of the wire sample B
during the wire rolling process. The SEM result supports
the following explanation for the big difference in the
MgO concentration between sample B and sample A: with
the existence of substantial micro-cracks in both the Ti-
sheath and the core region of sample B, the oxygen in
the air could percolate or diffuse along the cracks into
the core region from outside of the wire and then react with
Mg to form extra MgO during the sintering process. Such
extra MgO could not be produced in the crack-free sample
A. Correlating to the Jcresults shown in Fig. 2, it appears
that the decrease of Jcfrom sample A to sample B could be
due to the substantial increase of MgO content from 1.6%
in sample A to 4.8% in sample B. Similar correlation
between Jcand MgO content was also seen from the recent
results of Chen et al. [25,26] that Jcis decreased substan-
tially (by a factor of 3 or more) with slight increase of
the MgO content in their MgB2samples. It was shown
The peak fitting result for the MgB2(110), MgO (220), and MgB2(102) peaks in the XRD patterns shown in Fig. 4
SampleMgB2(110) peakMgO (220) peak MgB2(102) peak
MgO + MgB2
Fitting parameters: (2h)0is the centroid position of the peak, FWHM is the full width at half maximum, and I is the peak intensity measured by the area
under the fitting curve for each peak. The values of intensity are normalized to the intensity of the MgB2(110) peak for each XRD pattern.
The values of the relative intensity Irel, WMgO, and MMgO, which are
defined respectively as I220(MgO)/I102(MgB2), MgO wt.%, and MgO
mol%, for the following three samples: the mixture of MgO and MgB2
powders with 1:1 molar ratio, sample A, and sample B
MgO + MgB2
G. Liang et al. / Physica C 457 (2007) 47–54
previously that the MgO grains can be formed along the
boundaries of or even inside the MgB2grains [27,28]. Such
MgO impurities at the grain boundaries could make the
MgB2inter-grain connectivity worsen and thus cause sub-
stantial degradation in Jc.
In addition to MgO, MgB4 impurities could also be
formed during the sintering process by either the reaction
2MgB2! MgB4+ Mg or Mg + 4B ! MgB4[29–31]. One
interesting question related to the production of MgO
and MgB4is whether the formation of MgB4could affect
the concentration of the MgO and vice versa. Previously,
Li et al.  observed that MgB4impurities were formed
in their MgB2samples sintered at 800 ?C but not in samples
sintered at 660 ?C. This observation indicates that MgB4
may be formed only at high temperatures, at least above
660 ?C. On the other hand, since the autoignition tempera-
ture of Mg is 473 ?C, Mg can react easily with O2to form
MgO at temperatures between 473 ?C and 660 ?C. Thus,
we propose the following picture: In the sintering process
of our Ti-sheathed wires, it is very possible that prior to
the formation of MgB4, some starting Mg particles had
already reacted with all (or most) of the O2trapped in
the cores of the wires to form MgO. After the temperature
reached 800 ?C, small amount of MgB4could be produced
under certain conditions by the decomposition of MgB2.
At present, the detail conditions for the formation of
MgB4in the wires are not clear and further study on this
issue is needed. Since all of the O2trapped in the core of
the wire had been consumed earlier in forming MgO, no
oxygen would be available for the newly produced Mg
(via 2MgB2! MgB4+ Mg) to be oxidized to form new
MgO, instead, the Mg could either remain in the core
(may not be detected by XRD due to low concentration)
or escape from the open ends of the wires. In such pro-
posed picture, the concentration of the MgO in the samples
should be predominantly (or solely) determined by the ini-
tial content of the oxygen trapped in the core before the
sintering process. This means that the MgO concentration
should be either independent of or nearly unaffected by the
formation or concentration of MgB4impurities. In another
word, less production of MgO does not necessarily mean
less production of MgB4and vice versa, as observed in
the XRD patterns (see Fig. 3) of the samples. This explana-
tion is also consistent with the observation that for many
MgB2-based materials, MgO impurities usually present
without the accompanying formation of MgB4[9,25,31].
The SEM images in Fig. 6 show that there exists a large
amount of spherical holes or voids in the cores of sample A
and B. Most of the holes/voids are about 1–2 lm in diam-
eter, which is close to the size of the Mg powder particles in
the milled Mg + 2B powder precursor. These voids could
be produced by the volume reduction in the Mg + 2B !
MgB2reaction, it also could be partially attributed to the
evaporation of the Mg particles during the sintering of
the wires. It can be seen from Fig. 6 that the density of
the voids for the sample B is higher than that for sample
A. This is understandable because the micro-cracks (which
do not exist in sample A) in the core and certain portion of
the Ti-sheath of sample B provided more ‘‘doors’’ for the
evaporated Mg particles to escape during the sintering,
leaving behind, in the core, more voids/holes. Previously,
the existence of voids/holes was also observed in the cores
of the Fe-sheathed MgB2wires [6,32]. Thus, besides more
content of MgO impurities, the existence of micro-cracks
and more voids in sample B in contrast with those in sam-
ple A could be another contributing factor for the depres-
sion of Jcdue to the worsening connectivity between larger
Fig. 7 shows the 4pv(T) curves for the wire samples,
where v (= M/H) is the dc magnetic susceptibility. The crit-
ical transition temperature, Tc, defined as the onset of the
diamagnetism, is about 36 K for both the two samples.
The width of the transition (10–90% of the full drop in v)
is about 1.5 K. This Tc value is comparable to the Tc
(35–37 K) reported for some Fe-, Nb-, and Cu-sheathed
MgB2wires [7,8,33,34]. Shown in the inset of Fig. 7 are
the electrical resistivity q(T) curves for the samples, which
give the same onset Tcand transition width as the values
determined from the 4pv(T) curves. Above Tc, the q(T)
curve displays a metallic behavior similar to the q(T) curves
for the MgB2pellet samples . Fig. 7 shows that the val-
Fig. 6. SEM images of the cores of the Ti-sheathed MgB2wires: (a) for
sample A, and (b) for sample B. The surfaces of the cores were polished
before taking the images. These SEM images show that some micro-cracks
exist in sample B but not in sample A.
G. Liang et al. / Physica C 457 (2007) 47–54
ues of ?4pv at 5 K in the ZFC condition are about 0.75
and 0.63 for sample A and B, respectively, corresponding
to 25% non-superconducting volume fraction in the core
of sample A and 37% in sample B. Since the non-supercon-
ducting volume fraction represents the sum of the volume
of all open holes, voids, cracks, and non-superconducting
materials in the cores, this result is consistent with the
SEM result that more voids (including cracks) were formed
in sample B when compared those formed in sample A. For
both sample A and B, the ?4pv values in the FC mode is
very small (only about 3%). The large difference between
the ?4pv vales in the ZFC mode and FC mode indicates
that there exists a strong flux pining force in both sample
A and B, which holds most of the magnetic flux in the cores
under the FC condition.
In summary, we have successfully fabricated Ti-
sheathed, undoped monocore MgB2wires by in situ PIT
method with Jc improved substantially from previously
fabricated wires. Particularly, for our best wire (sample
A), the Jcis much higher than that of the Fe-sheathed
mono-core MgB2wires fabricated with the same in situ
PIT process and under the same fabricating conditions.
However, our Jcresults also suggest that the overall Jc
for the average Ti-sheathed wires including wire sample B
is comparable to the Jc of the Fe-sheathed wires. Our
XRD and SEM analysis reveals that MgO impurities and
micro-cracks could be the possible factors for the degrada-
tion of critical current density in these Ti-sheathed MgB2
wires. The results from the XRD, magnetic Jc, SEM,
EDS, M(T), and q(T) measurements are mutually consis-
tent and well correlated.
The authors thank Dr. Z. Tang, Dr. J. Meen, A. Scotti,
M. Crush, and Dr. J. Horvat for either their assistance in
measurements or helpful discussion. This work was sup-
ported by Sam Houston State University’s Faculty Re-
search Grant and 2006 EGR grant, an award from
Research Corporation, and by the State of Texas through
 J. Nagamatsu, N. Nakagawa, T. Muranaka, Y. Zenitani, J. Akimitsu,
Nature 410 (2001) 63.
 D.C. Larbalestier, L.D. Cooley, M.O. Rikel, A.A. Polyanskii, J.
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. Hass, Nature 410 (2001) 186.
 R. Flu ¨kiger, H.L. Suo, N. Musolino, C. Beneduce, P. Toulemonde, P.
Lezza, Physica C 385 (2003) 286.
 H. Fang, Y.Y. Xue, Y.X. Zhou, A. Bailalov, K. Salama, Supercond.
Sci. Technol. 17 (2004) L27.
 G. Liang, H. Fang, D. Katz, Z. Tang, K. Salama, Physica C 442
 S. Soltanian, X.L. Wang, J. Horvat, S.X. Dou, M.D. Sumption, M.
Bhatia, E.W. Collins, P. Munroe, M. Tomsic, Supercond. Sci.
Technol. 18 (2005) 658.
 P. Kovac, I. Husek, T. Melisek, M. Kulich, V. Strbik, Supercond. Sci.
Technol. 19 (2006) 600.
 X.L. Wang, Q.W. Yao, J. Horvat, M.J. Qin, S.X. Dou, Supercond.
Sci. Technol. 17 (2004) L21.
 W.K. Yeoh, J.H. Kim, J. Horvat, X. Xu, M.J. Qin, S.X. Dou, C.H.
Jiang, T. Nakane, H. Kumakura, P. Munroe, Supercond. Sci.
Technol. 19 (2006) 596.
 X. Xun, M.J. Qin, K. Konstantinov, D.I. Dos Santos, W.K. Yeoh,
J.H. Kim, S.X. Dou, Supercond. Sci. Technol. 19 (2006) 466.
 A. Serquis, X.Z. Liao, Y.T. Zhu, J.Y. Coulter, J.Y. Huang, J.O.
Willis, D.E. Peterson, F.M. Mueller, N.O. Moreno, J.D. Thompson,
V.F. Nesterenko, S.S. Indrakanti, J. Appl. Phys. 92 (2002) 351.
 E.W. Collins, P.C. Gehlen, J. Phys. F: Met. Phys. 1 (1971) 908.
 R.J. Wasilewski, J. Appl. Phys. 40 (1969) 2677.
 M. Alessandrini, H. Fang, M. Hanna, P. Putman, Y.X. Zhou, K.
Salama, Supercond. Sci. Technol. 19 (2006) 129.
 G. Liang, H. Fang, M. Hanna, F. Yen, B. Lv, M. Alessandrini, S.
Keith, C. Hoyt, Z. Tang, K. Salama, Supercond. Sci. Technol. 19
 S. Soltanian, M.J. Qin, S. Keshavarzi, X.L. Wang, S.X. Dou, Phys.
Rev. B 68 (2003) 134509.
 M. Tikham, Introduction to Superconductivity, McGrow-Hill, New
York, 1985, p. 184.
 H.D. Yang, H.L. Liu, J.-Y. Lin, M.X. Kuo, P.L. Ho, J.M. Chen,
C.U. Jung, Min-Seok Park, Sung-Ik Lee, Phys. Rev. Lett. 87 (2001)
 S. Jin, H. Mavoori, C. Bower, B. Van Dover, Nature 441 (2001) 563.
 S.X. Dou, S. Soltanian, J. Horvat, X.L. Wang, S.H. Zhou, M.
Ionescu, H.K. Liu, P. Munroe, M. Tomsic, Appl. Phys. Lett. 81
 E.M. Gyorgy, R.B. van Dover, K.A. Jackson, L.F. Schneemeyer, J.V.
Waszczak, Appl. Phys. Lett. 55 (1989) 283.
 C.P. Bean, Phys. Rev. 8 (1962) 250.
 G. Liang, C. Hoyt, H. Fang, S. Guchhait, J. Markert, unpublished.
 L.V. Azaroff, M.J. Buerger, The Powder Method, McGrow-Hill
Book Company, New York, 1958, Chapter 13.
 S.K. Chen, K.A. Yates, M.G. Blamire, J.L. MacManus-Driscoll,
Supercond. Sci. Technol. 18 (2005) 1473.
Fig. 7. Temperature dependent dc magnetization, measured in ZFC and
FC modes in a field of 20 Oe and temperatures between 5 K and 50 K, for
sample A (filled circles) and sample B (open circles). The inset shows the
temperature dependent electrical resistivity curves for these two samples
between 20 and 80 K.
G. Liang et al. / Physica C 457 (2007) 47–54
 S.K. Chen, M. Wei, J.L. MacManus-Driscoll, Appl. Phys. Lett. 88
 R.F. Klie, J.C. Idrobo, N.D. Browning, K.A. Regan, N.S. Rogado,
R.J. Cava, cond-mat/0107324.
 Y. Zhu, L. Wu, V. Volkov, Q. Li, G. Gu, A.R. Moodenbaugh, M.
Malac, M. Suenaga, J. Tranquada, cond-mat/0105311.
 P. Kovac, I. Husek, T. Melisek, J.C. Grivel, W. Pachla, V. Strbik, R.
Diduszko, J. Homeyer, N.H. Andersen, Supercond. Sci. Technol. 17
 P. Lezza, R. Gladyshevskii, H.L. Suo, R. Flukiger, Supercond. Sci.
Technol. 18 (2005) 753.
 A.H. Li, X.L. Wang, M. Ionescu, S. Soltonian, J. Horvat, Y. Silver,
H.K. Liu, S.X. Dou, Physica C 361 (2001) 73.
 S.H. Zhou, A.V. Pan, M.J. Qin, H.K. Liu, S.X. Dou, Physica C 387
 Y. Ma, X. Zhang, A. Xu, X. Li, L. Xiao, G. Nishijima, S. Awaji, K.
Watanabe, Y. Jiao, L. Xiao, X. Bai, K. Wu, H. Wen, Supercond. Sci.
Technol. 19 (2006) 133.
 X. Zhang, Y. Ma, Z. Gao, Z. Yu, G. Nishijima, K. Watanabe,
Supercond. Sci. Technol. 19 (2006) 479.
G. Liang et al. / Physica C 457 (2007) 47–54