Parinov I.A. Microstructure and Properties of High-Temperature Superconductors

Подождите немного. Документ загружается.

148 3 Experimental Investigations of HTSC

100 μm

μm

(a)

(b)

Fig. 3.35. SEM micrographs of fracture surfaces of Al

/BPSCCO composite at:

(a) 293 K and (b) 77 K [712]

toughness and fracture modes at room and cryogenic temperatures may be

explained by thermal stresses, forming in the composite during cooling down

to cryogenic temperature. At the same time, strength of the Al

/BPSCCO

composite may decrease in comparison with monolithic superconductor, if the

interfaces between matrix and ﬁbers are weak [679]. In this case, the ﬁbers

act as pores and also as stress concentrations.

Best thermodynamic compatibility of MgO [215] assumes application as

toughening elements of high-strength MgO whiskers that ensure signiﬁcant

increasing stiﬀness and fracture toughness of the monolithic superconductor

3.3 Superconducting Composites, Based on BSCCO 149

10 μm

Fig. 3.36. Surface morphology of a pulled-out ﬁber observed in fracture surface of

/BPSCCO composite at 77 K [712]

[1187]. The toughening mechanisms of this composite are related to whisker

fracture, whisker/matrix interface de-bonding, pushing of whiskers by crack

surfaces and also the fracture modes caused by the crack deﬂection (Fig. 3.37).

An appreciable increasing of plane-strain fracture toughness (K

)isobserved

both for monolithic BPSCCO specimens and (MgO)

/BPSCCO composite

after cooling from room temperature (293 K) to cryogenic (77 K) (Fig. 3.38).

For monolithic ceramic, it can be related to greater surface energy (γ)atlower

temperature by using Griﬃth’s formula [348] as

(T )=



(T )

1 − ν



1/2

, (3.1)

where E and ν are Young’s modulus and Poisson’s ratio, respectively.

A formation of small residual thermal stresses may be expected in the

(MgO)

/BPSCCO composite that introduce minimum number of defects at

interfaces during cooling to cryogenic temperature because of proximity of the

thermal expansion factors for whiskers and superconducting matrix and also

because of their small sizes. Hence, an elevated fracture toughness of compos-

ite is caused in total by the surface energies of BPSCCO matrix and MgO

whiskers [1187]. At the same time, an initial increasing of strength with volume

fraction of whiskers may be replaced by its decrease of growth of the whisker

fraction up to 20% and more. It could be explained by microdefect forma-

tion, including grain boundaries and BPSCCO/MgO interfaces, that creates

signiﬁcant stress concentrations. An increasing of the whisker content leads

to the same eﬀect on change of elastic module. By introducing MgO particles

in BPSCCO matrix, the bending strength increases. In this case, a decreasing

of compliance in comparison with monolithic sample is possible. However, an

application of polymeric addition increases the composite compliance.

150 3 Experimental Investigations of HTSC

10 μm

Crack

(MgO)

(a)

(b)

(c)

(MgO)

Crack

10 μm

Fig. 3.37. SEM micrographs, showing multiple toughening mechanisms in fractured

specimens (three-point notched bending) of (MgO)

/BPSCCO composite, tested at

room temperature: (a) whisker breakage and whisker/matrix interface de-bonding,

(b) whisker pulled out and (c) crack deﬂection [1187]

3.4 Melt-Processed Y(RE )BCO 151

4.0

3.5

3.0

2.5

Volume Fraction of Whiskers (%)

(MPa

1/2

)

293 K

77 K

10 20

Fig. 3.38. Fracture toughness of monolithic BPSCCO and (MgO)

/BPSCCO

composite at room temperature and 77 K [1187]

3.4 Melt-Processed Y(RE )BCO

3.4.1 Microstructure Features

Now, the melt-processing techniques, presented in Sect. 2.5, are considered

the main perspective for preparation of large-grain YBCO samples with high

values of J

. Therefore, the main attention will be directed to investigation

of the melt-processed Y(RE)BCO samples. The normal phases, chemical and

physical heterogeneities at boundaries of a Y-123 (or 123) phase cause weak

links, leading to decreasing of J

in YBCO. In the case of melt-processed multi-

domain samples, a liquid is often observed at boundaries of these regions as

a result of incomplete peritectic reaction (so-called defects caused by grain

growth) and is one of the causes of small J

[196]. Moreover, a non-symmetric

HTSC structure, leading to anisotropy of transport current, is also proper for

small critical current density. So, J

in ab-plane is much more greater than the

corresponding value along c-axis [376]. The value of J

in the melt-processed

YBCO strongly depends on the microstructure of superconducting grains [535,

744, 1134]. Compared to the conventionally sintered samples, many more de-

fects are included in the interior of melt-processed YBCO grains. The defect

density is a function of fabrication conditions including peritectic heat treat-

ment and subsequent oxygenation treatment [535] and also a special concen-

tration of the ab-planar defects (microcracks or lamellae boundaries) around

152 3 Experimental Investigations of HTSC

trapped Y-211 (or 211) particles [196, 1134]. In this case, these defects do not

considerably inﬂuence the transport current parallel to the c-axis [1134]. The

Y-211/Y-123 interface [744] and the structure defects [1134] are proposed as

ﬂux-pinning centers of a Y-123 phase. Other advantages, associated with the

Y-211 additives, are the following: the improved fracture toughness [298], ho-

mogeneous microstructure [536] and suppression of the microcrack formation

[744].

The microstructure of melt-processed YBCO, as a rule, consists of Y-123

lamellae (pseudo-grains), oriented almost parallel to the ab-planes and pos-

sessing common c-axes (Fig. 3.39). However, the formation process of this

structure is yet to be understood. For example, it is argued in [639] that

lamellae formation occurs during peritectic solidiﬁcation and is associated

with the irregular growth morphology of a Y-123 growth front in the presence

of a build-up of Y-211 particles. On the other hand, it is suggested that the

lamellae structure forms during the tetragonal to orthorhombic phase transi-

tion (TOPT), which occurs at 400–500

◦

C, in subsequent sample oxidization

rather than at high temperature during peritectic solidiﬁcation [1134]. The

latter point of view can be supported by the absence of lamellae in a tetrag-

onal, unoxygenated Y-123 grain, at simultaneous existence of a number of

lamellae in the oxygenated Y-123 grain [530]. Therefore, it may be suggested

that the driving force for lamellae formation is the stress, induced during the

TOPT, and the subsequent instability of the Y-123 phase.

Besides twins and microcracks, another defect frequently observed in the

melt-textured microstructure is a CuO stacking fault. The stacking fault

presents itself as an additional CuO layer between two BaO layers [1134].

These additional CuO layers exist together with partial dislocations both in a

Y-123 matrix and around trapped Y-211 particles (Fig. 3.40). In [10], it was

proposed that CuO stacking faults formed during high-temperature peritectic

Y–211

10 μm

Fig. 3.39. Scanning electron micrograph, showing BaCuO

lamellae along the

ab-plane of a Y-123 domain [530]

3.4 Melt-Processed Y(RE )BCO 153

Y-211

010

100

0.5 μm

Fig. 3.40. Stacking faults and partial dislocations, developed around a trapped

Y-211 particle [530]

reaction. However, in [541], it was found that the stacking faults were formed

during oxygen annealing in the same manner as that of the lamellae, because

Y-123 was not stable at oxygen annealing and decomposed into other stable

secondary phases, following the reaction [1145]:

4YBa

7−x

+(1/2−3/2δ+2x)O

→ 2Y

BaCuO

+3Ba

6−δ

+CuO .

(3.2)

The Y-123 decomposition seems not to be associated with the formation of

a Y-211 phase, because yttrium-enriched phase regions are observed around

the trapped Y-211 particles [1134]. Similar to the BaCuO

lamellae, a driv-

ing force for the formation of CuO stacking faults may be stress induced

by the TOPT and also stress caused by the diﬀerence in thermal expansion

between Y-211 and Y-123 phases. The stresses caused by the TOPT in the

melt-processed YBCO are many more than corresponding values of sintering

ceramic due to comparatively greater Y-123 grain size. It also can provoke

decomposition of Y-123 phase around trapped Y-211 particles. In the case of

application of the melting processing (PMP), stacking faults form in the sam-

ple because of incomplete diﬀusion during peritectic reaction between Y-211

particles and the liquid (BaCuO

and CuO). The amount of defects could be

decreased using an additional annealing of the sample under elevated temper-

ature. The considerable decreasing of J

, observed after this annealing, proves

that stacking faults are proper magnetic ﬂux-pinning centers [1194].

There are two main sources of void formation in Y(RE)BCO. (i) Inert gas

in technological atmosphere remains in the form of voids in sample after its

fabrication. In this case, the void number increases with the partial pressure

154 3 Experimental Investigations of HTSC

of inert gas, which cannot diﬀuse from the sample [824]. (ii) A great amount

of oxygen realises as a result of decomposition of Y-123 phase, forming voids

[909]. In this case, there are two ways for oxygen evaporation: usual diﬀusion

and O

bubbles, moving through liquid. In low-temperature region, oxygen

diﬀusion is dominated, but at high temperatures, bubble formation occurs,

which may be accompanied by sample deformation [197, 909]. A melting in

pure oxygen renders eﬀective the decreasing of Sm-123 porosity; however, in

this case, a decreasing of superconducting properties is possible [465].

3.4.2 Growth Processes in Seeded Sample

The growth of Y-123 crystal occurs in the form of a parallelepiped with (100),

(010) and (001) habit planes [475]. Basically, three modes of crystal growth

from the seed are observed (Fig. 3.41), namely:

(1) Epitaxial growth of single Y-123 crystal from the seed with its c-axis par-

allel to the sample axis (the case of using Nd-123 plate-shaped seed and

direct seeding at the beginning of solidiﬁcation) [197]. In this case, there

are ﬁve growth sectors (GSs), namely: four a-GSs with habits perpendic-

ular to the [100], [

100], [010] and [0

10] directions and a c-GSwithhabit

perpendicular to the [00

1] direction. The top angle of the c-GS depends

on the ratio of the growth rates in the c-anda-directions.

(2) A cubic Nd-123 seed leads to ﬁve domain samples with c-GSc dominating

in each grain [1066]. In this case, four narrow a-GSs develop along both

sides of the 90

◦

boundaries in each grain. These 90

◦

high-angle grain

boundaries between grains are strongly coupled and ab-microcracks do

not disturb current in ab-planes.

(3) Seed from MgO single crystal does not dissolve in the partially melted

sample and it can be placed on a cooled green sample before melt pro-

cessing. Due to the misﬁt of the MgO and 123 lattices, it often happens

that samples with the ab-plane parallel to the sample axis are produced.

They consist of two c-GSc and three a-GSs [429].

In Y-123 crystals, grown from partially melted bulks, subgrains (SGs) can

form, which present crystal regions, divided by low-angle grain boundaries

[197]. In general, these boundaries are not to act as weak links; however, some

low-angle grain boundaries can act as weak links in high-ﬁeld regions [200].

Therefore, for high-ﬁeld applications, it is necessary to control the low-angle

grain boundaries, that is, it is important to study the formation mechanism

of the subgrains. Five diﬀerent types of subgrain can be classiﬁed, accord-

ing to growth directions active in their formation [202, 203, 204], namely:

(i) a-subgrains (a-SGs) have subgrain boundaries (SGBs) parallel to the

a-axis; (ii) a-a subgrains (a-a-SGs) have SGBs tilted from the a-direction

(Fig. 3.42a); (iii) a-c subgrains (a-c-SGs) have SGBs parallel to the c-axis and

develop at the stepped planar a growth front; (iv) c-subgrains (c-SGs) have

SGBs parallel to the c-axis; and (v) c-a subgrains (c-a-SGs) have SGBs par-

allel to the a-axis and develop at the stepped planar c growth front. The ﬁrst

3.4 Melt-Processed Y(RE )BCO 155

(a)

(b)

(c)

GSBs

90°-GBs

[001]

GSB

[110]

max

[100] or [010]

Calcined Rod

Liquid Phase

(d)

= R

/ 2 = (R

max

/ 2) cos θ

= R

max

sin θ

= 2 tg θ

c-GS

a-GS

√

Fig. 3.41. Schematic illustrations of GSs, divided by GS boundaries (GSBs), formed

in the melt-grown 123 bulk: TSMG-processed single-grain bulks with the c-axis

parallel (a) or perpendicular (b) to the sample axes and in the ﬁve-grain bulk

(c). Schematic illustration of longitudinal interface for typical sample, prepared by

directional solidiﬁcation (d) [197]

three SGs form in the a-GS and the last two SGs form in the c-GS. It is inter-

esting that the c-SGs are always larger than the a-SGs in cross-section. The

arrangement of subgrains in the single-grain melt-grown bulk is demonstrated

in Fig. 3.43.

In particular, subgrains are manifested by the changes in the direction of

microcracks parallel to the ab-plane (Fig. 3.44) [202]. In this case, the SGBs

are clean and not cracked.. The growth-related subgrains, observed in melt-

grown Y-123, is supposed to be formed by the dislocation arrangement into

156 3 Experimental Investigations of HTSC

–

growth

a – a SGB

(a)

(b)

–

growth

300 μm

–

c SGB

–

c SGB

–

growth

100 μm

Solidified Liquid

GSB

Fig. 3.42. Polarized optical micrographs of TSMG-processed Y-123/Y-211: (a)a

higher macrostep at the growth front along a-axis, forming a-a-SGB; (b) the cross-

section both parallel to the c-axis and the growth direction, c-a-SGBs, which are

connected with the inner corners of the steps, form on the stepped growth front [197]

dislocation walls during the crystal growth [478]. The subgrain formation is

assisted by the edge dislocations with Burgers vectors parallel to the growth

front. In the case of melt-grown 123 grains, the dislocation formation due to

Y-211 particle seems to be the most plausible mechanism because Y-123/Y-211

interface is incoherent [203]. The density of dislocations should be propor-

tional to the density of 211 particles, which is supported by the fact that the

subgrain size increases with an increase in the average spacing between 211

particles [199]. Moreover, it is observed that the subgrain size is reduced with

an increased cooling rate [1022]. The fact that subgrains are not observed for

3.4 Melt-Processed Y(RE )BCO 157

a-Subgrains

Growth Front

c-Subgrains

a-Growth

c-Growth

Fig. 3.43. The subgrain arrangement in the single-grain melt-grown bulk [197]

10 MκM

Fig. 3.44. The tilting of ab-microcracks at SGBs in a section parallel to the (100)

plane [197]