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

Подождите немного. Документ загружается.

188 4 Carbon Problem

400

300

200

100

900 1200

Sintering Temperature (°C)

(A/cm

)

5000 ppm CO

500 ppm CO

50 ppm CO

100% O

1000980960940920

Fig. 4.5. The eﬀect of sintering temperature on J

[958]

We do not know test evidence of secondary phases at grain boundaries in

samples sintered in a 100% O

atmosphere. At the same time, two distinct

types of grain boundaries can be observed in the samples sintered in 0.5%

+99.5%O

[958]. Approximately 10% of the grain boundaries are found

to be wet by a thin layer of a second phase, as shown in Fig. 4.6, and the

remaining boundaries are sharp grain boundaries and free from the presence

of secondary phases. The local change of YBCO stoichiometry leads to the

eutectic reaction, in accordance with the phase diagram, and formation of

BaCuO

and CuO at the grain boundaries. The decomposition of YBCO phase

at the grain boundaries weakens the path for the current, thereby decreasing

the overall J

value. The wetting of the boundaries with non-superconducting

phases decreases the eﬀective contact area between the superconducting grains

and decreases J

. From Fig. 4.6, the width of the secondary phase region along

the grain boundaries is determined to be approximately 50 nm, which is much

larger than the coherence length in this material. Therefore, these secondary

phases block the passage of currents.

There are two potential sources of CO

, when YBCO are processed by

solid-state sintering, namely (i) the CO

, contained in the oxygen gas, used

during sintering and/or annealing and (ii) the CO

, derived from the decom-

position of BaCO

during the calcining step. Each source aﬀects the quality of

the ﬁnal product in diﬀerent ways and can lead to a drop in the critical current

density. An additional potential source is the organic binders that are used

in the YBCO processing. The CO

, formed by the interaction of the organic

binders and the oxygen from the gas atmosphere, could have a deleterious

eﬀect on the performance capabilities of the ﬁnal product. By using thermo-

dynamics, it may by shown [958] that for pressure values, p

< 1atm,a

sintering temperature must increase as p

increases. Therefore, increasing

4.1 YBCO System 189

200 nm

Fig. 4.6. The sample sintered at 970

◦

Cin0.5%CO

+99.5% O

. TEM micrograph

for the grain boundary (GB) which is wet by a layer of BaCuO

second phase [958]

the temperature results in a higher p

. The sintering temperatures can be

lowered if p

can be lowered, and one expedient means of doing this is to

remove the CO

contained in the O

gas.

Carbon also aﬀects negatively the temperature and broadness of super-

conducting transition [1031]. It is known that YBCO samples, sintered to

closed porosity, exhibit broad and suppressed superconducting transitions

(Fig. 4.7). The observations suggested that during sintering of the closed-

porosity material, “impurities” that suppress the superconducting transition

become trapped in the center of the sample, while in the open-porosity sam-

ple and near the surface of the closed-porosity material, volatilization of the

impurity can occur during sintering.

The highest carbon content of ≈1.2 wt% is detected in the grains of the cen-

tral part from the closed-porosity material. Signiﬁcantly lower carbon contents

are detected in specimens, cut from both the exterior of the closed-porosity

material and the central part of the open-porosity material (≈0.6 wt%). Car-

bon removal occurs more readily in the open-porosity material and near

the surface of the denser material, consistent with the suggestion that car-

bon removal is limited by the slow diﬀusion of carbon out of the YBCO

grains.

Under appropriate conditions, carbon has an appreciable solubility in

YBCO bulk. The depression of the transition temperature, detected by mag-

netic susceptibility measurements, correlates well with carbon content. This

correlation suggests that incorporation of carbon in the structure is the dom-

inant cause of the lowering of the transition temperature in the high-dense

superconductors.

190 4 Carbon Problem

Open-Porosity Material

Whole Pellet

Half Pellet

Temperature(K)

−1

Normalized ac-Susceptibility

(a)

1009080706050

Normalized ac-Susceptibility

Closed-Porosity Material

Whole Pellet

Half Pellet

–1

(b)

Re-oxygenated

half (600°C)

Temperature (K)

100

9080706050

Fig. 4.7. Change in ac-susceptibility with temperature for (a) an open-porosity

material (ρ = 87%) and (b) a closed-porosity material (ρ = 92%). In the open-

porosity material both whole and halved samples exhibit sharp transitions, whereas

in the closed-porosity material a stepped transition is seen in the halved sample even

after an additional oxygenation treatment [1031]

4.1 YBCO System 191

In the interior of the closed-porosity material, CO

or CO gas may be

trapped in the pores. As a result, a pressure of CO

or CO gas (in the range

0.01–0.1atm) can develop in the pores that is enough to slow down sintering

processes. Then, several steps may be taken to lower the amount of carbon,

trapped in a dense material, namely:

(1) The external partial pressure of CO

should be as low as possible in order

to favor decomposition of carbonate in the material.

(2) If the densiﬁcation process can be delayed (to allow a longer period, in

which the sample contains interconnected porosity), more decomposition

and gas evolution may occur before the gas becomes trapped.

(3) A successful approach to carbon removal is to heat more slowly to the

sintering temperature.

Carbon in the crystal structure of YBCO interacts with Ba and O, reducing

the Ba–O coordination [1031]. The carbon connected with Ba ion could be

incorporated in the structure either substitutionally or interstitially. In the

ﬁrst case, a direct substitution of oxygen by carbon ions is possible (it is

not supported in tests). An alternative possibility is that the carbon lies in

an interstitial site adjacent to the barium ion. In the absence of a detailed

theoretical understanding of the origin of the superconducting transition in

YBCO, the origin of the depression of T

, produced by the incorporation of

carbon, may be speculated. It is known that there is a strong correlation

between T

and the concentration of holes in the material [966]. The presence

of the carbon causes localization of some of the holes and thus reduces the

mobile holes’ concentration in the sample. This suggestion is consistent with

the location of the carbon in an interstitial site adjacent to the CuO

planes.

Because the carbon concentration can approach one atom per unit cell, the

location of the carbon in such a site could be expected to cause appreciable

local distortions of the CuO

planes and thereby lower T

It may be shown [651] that carbon is released from the matrix at temper-

atures higher than that of the peritectic decomposition. Nevertheless, there

is no doubt that even in a melt-processed sample, depression of T

may still

occur. Consider proper technological procedures, leading to decreasing of car-

bon content in the melt-processed YBCO samples by using as example the

results of [1127]. The melt-texturing was performed by a melt-texture growth

(MTG) process, using a temperature gradient furnace (temperature gradient

about 25 K/cm). As a result, two samples were synthesized, namely, (i) the

sample Sr5a, containing large grains (100 μm) with a very dense, closed poros-

ity structure, and (ii) the sample Sr5b, containing small grains (< 10 μm) with

open porosity. Their superconducting transition curves are shown in Fig. 4.8.

The T

c,onset

of sample Sr5a is depressed to 80 K, with a broad transition.

Sample Sr5b has T

c,onset

at 92 K, which is the optimum value for the Y-123

system. The analysis shows rather a high degree of carbon retention in the

two samples: 2600 ppm for sample Sr5a and 1200 ppm for sample Sr5b. Thus,

192 4 Carbon Problem

20 40 60 100

Temperature (K)

Magnetization,

M (emu/g

Oe)

0.005

0.000

–0.005

–0.010

–0.015

–0.020

−0.025

Sr5a

Sr5b

080

Fig. 4.8. The temperature dependence of magnetization of Y-123 ceramics sintered

at diﬀerent temperatures: Sr5a (sintered at 950

◦

Cinoxygenﬂow)andSr5b(sintered

at 920

◦

C in oxygen ﬂow). Average particle size of precursor (2.5 μm) is the same for

both ceramics [1127]

almost all carbon of the initial powder (3000 ppm) is retained in sample Sr5a,

and its low T

c,onset

is in accordance with the correlation stated above.

In sample Sr5a, three experimental features can explain the carbon content

in this sample as follows:

(1) The obtained lattice parameters (a =3.8288

A,b =3.889

Aandc =

11.667

A) showed that the decrease of the c parameter, which was sensi-

tive to the carbon content, was considerably less than that observed for

samples prepared under a CO

atmosphere (c =11.585

A) [76].

(2) The characteristic features of the YBCO carbonated phases is the exis-

tence of oriented domains, where the c-axes are perpendicular; the for-

mation of these domains is correlated not only to the decrease of the a/b

and a/c ratios, but also to the presence of carbonate groups, which are

preferentially located at the level of the domain boundaries.

(3) In carbon-rich materials, the existence of distributions of local superstruc-

tures, leading to the formation of streaks along a-direction and to short-

range aligned areas, is clearly visible on the images.

The above two sintered samples were further processed by MTG. The re-

sults presented in Fig. 4.9 show that this procedure allowed to further decrease

the carbon content. However, an excessive amount of carbon can still be re-

tained, if the carbon content of the initial precursor is too high, as shown for

sample Sr5amt. Therefore, the absolute necessity of controlling the quality of

the precursor material before melt texturing is obvious.

Then, in order to examine the inﬂuence of the 211 ﬁne-disperse additives

and Pt doping on carbon retention, two sets of samples were prepared, one

without Pt addition (5avmt) and another with the addition of 0.5 wt% PtO

(5avpmt). Both sets had nominal composition Y

1.4

2.2

3.2

. Addition of

4.1 YBCO System 193

Temperature(K)

0.25

–0.31

–0.88

–1.44

–2.00

Sr5amt

Sr5bmt

Susceptibility(a.u.)

10080604020

Fig. 4.9. The dependence of magnetic susceptibility on temperature of melt-

processed samples. Sr5amt (or Sr5bmt) is an MTG sample, using Sr5a (or Sr5b)

as a pre-sintered precursor [1127]

Y-211 does not signiﬁcantly decrease carbon retention in the Y-123. In con-

trast, as it follows from Fig. 4.10, for sample 5avpmt with added Pt, T

c,onset

is enhanced up to the optimum value of 92 K, and a remarkably sharp transi-

tion (ΔT

= 2 K) is obtained. Analysis shows that the carbon content in this

sample is as low as 200 ppm. At the same time, the sample 5avmt contains

100

Temperature (K)

0.3

−0.3

−0.9

−1.4

−2.0

5avmt

5avpmt

Susceptibility (a. u.)

0 20406080

Fig. 4.10. The temperature dependence of the susceptibility curve for the samples

processed by an MTG method with starting composition Y

1.4

2.2

3.2

,with

(5avpmt) and without (5avmt) addition of Pt [1127]

194 4 Carbon Problem

about 870 ppm of carbon. Although the exact mechanism by which the plat-

inum acts to decrease the carbon content is unknown, it may be suggested

that Pt acts as a catalyst for de-carbonation.

4.2 BSCCO Systems

The behavior of BSCCO samples due to carbon eﬀect somewhat diﬀers from

the behavior of YBCO superconductors. The dependencies of sintering tem-

perature and density of Bi-2223 precursor samples on carbon content are

shown in Fig. 4.11. Obviously, the higher sintering temperature and increas-

ing density of the Bi-2223 samples cause the lower carbon content.

740 750 760 770 780 790 800

400

300

200

100

Sintering Temperature (°C)

Carbon Content (ppm)

(a)

Densit

(

/cm

)

Carbon Content (ppm)

400

300

200

100

234

(b)

Fig. 4.11. The dependencies of (a) sintering temperature and (b) density of Bi-2223

precursor samples on carbon content [1164]

4.2 BSCCO Systems 195

Oxide superconducting powder can absorb atmospheric moisture and then

react with CO

, worsening superconducting properties. Therefore, a prepara-

tion process, proposed for Bi-2223 superconductor [928], in the ﬁrst place

restricts absorption of moisture in order to suppress the deterioration of the

calcined powder. The preparation process includes the processes of dissolving,

coprecipitation, ﬁltering, primary calcining (T = 800

◦

C), wet pulverizing to

obtain a ﬁne powder, drying and secondary calcining (T = 800–815

◦

C). Then,

in order to fabricate bulk sample the calcined powder is molded by using a

coaxial pressing technique. The sample is sintered at 850

◦

C sometimes by us-

ing intermediate cold isostatic pressing (CIP). The results obtained for three

various samples are presented in Table 4.1. The intermediate pressing leads

to increasing of the sample density (Fig. 4.12a) and the corresponding rise of

critical current density (Fig. 4.12b). The saturation of the results occurs after

third intermediate pressing. The results obtained conﬁrm that superconduct-

ing properties of the bulk superconductor are found by content of moisture

and carbon in the calcined powder. Carbon and admixture phases segregate

at the grain boundaries, forming weak links and decreasing J

The cause of the porosity formation and broken uniformity of Bi-2212/Ag

tapes may be as a result of the gas release, such as CO

, when Bi-2212 phase

reforms from the liquid phase [1196]. Carbon exists in the sample in the car-

bonate form and can transfer into CO

gas in proper condition during the

heat treatment, such as at high temperature by the reaction:

SrCO

→ SrO+CO

. (4.7)

gas may be released and causes severe bubble formation and porosity,

depending on the carbon content. It may also dissolve into the liquid phase

and deﬁne its behavior. The carbon content eﬀects on the melting behavior

of the Bi-2212 phase [1195] state that increasing the carbon content of the

powder can greatly decrease the melting and reformation temperatures of the

Bi-2212 phase in the OPIT tapes.

During cooling, when Bi-2212 phase re-forms from liquid, [C] will be re-

leased in accordance with the reaction:

[C] + O

→ CO

, (4.8)

where [C] is the carbon, which is dissolved in the liquid phase.

Table 4.1. Moisture content and carbon content of calcined powders [928]

Sample Process Moisture content

(wt %)

Carbon content

(wt %)

A After drying in air 1.01 0.610

2nd-calcined powder at 800

◦

C 0.13 0.041

B After drying in vacuum 0.65 0.486

2nd-calcined powder at 800

◦

C 0.11 0.028

C After drying in vacuum 0.64 0.482

2nd-calcined powder at 815

◦

C 0.12 0.025

196 4 Carbon Problem

(b)

(kA/cm

)

01234

CIP Number

5.5

4.5

01234

3.5

CIP Number

Density (g/cm

)

(a)

Fig. 4.12. Increase of (a) sample density and (b) J

(77 K, 0 T) due to intermediate

pressing (CIP) and sintering [928]

4.3 Carbon Embrittlement and Fracture of YBCO Superconductor 197

The volume change, caused by the formation of CO

gas from the solid-

state carbon, can be tremendous. A rough calculation indicates that 200 ppm

of carbon can cause about 36% porosity in the core, if all carbon form CO

at high temperature [1196].

The eﬀect of the introduction of carbon into YBCO and Bi-2212 bulks,

obtained by using a partial melting (both samples were heated some above

the peritectic temperature), was studied in [962]. The presence of carbon has

strong eﬀect on the YBCO melting, but is not essential for Bi-2212. In this

case, the exposed carbon excess can cause an increasing of the magnetic ﬂux

pinning. At slow cooling of the YBCO and Bi-2212 from the temperature of

partial melting, it is necessary to strictly control temperature and gas atmo-

sphere in order to obtain high superconducting properties. Most important

factors are the oxygen gas release [143, 408, 742, 961] and the presence of

carbon and CO

in the sample volume [1101, 1195]. Thus, due to the re-

action of YBCO and BSCCO with CO

atmosphere, the carbon trapped in

the sample can render unfavorable eﬀect on the sample density [1196], in-

tergranular boundaries [307, 958], critical currents [969], and critical current

density [1101, 1158]. Besides atmospheric sources, carbon may be introduced

due to carbonized precursors and organic binders during sample processing

[212] and also by solvents [304]. Finally, the carbon can be precipitated into

superconductor due to carbonized gases and liquids. Existing techniques can

form carbon particles with size of some nanometers, which can be the centers

of magnetic ﬂux pinning into superconducting grains. Then, on one hand,

carbon can render useful eﬀect, in particular, using Pb and some rare-earth

elements [742], and, on the other hand, it can deteriorate a structure of the

intergranular boundaries and structure-sensitive properties of HTSC.

4.3 Carbon Embrittlement

and Fracture of YBCO Superconductor

Carbonates, forming in result of the chemical reactions (4.3)–(4.6), are the

brittle phase, leading to embrittlement of superconductor. It may lead to

delayed carbonate cracking, a subcritical crack growth mechanism with the

next formation and fracture carbonate. The carbon-induced embrittlement is

a complicated mechanism, which results from the simultaneous operation of

several coupled processes, namely (i) carbon diﬀusion, (ii) carbonate precipi-

tation, (iii) non-mechanical energy ﬂow and (iv) material deformation. In this

case, carbon diﬀusion occurs due to gradients of chemical potential and tem-

perature [189, 974]. Similar to hydride formation in metals [69, 620, 839, 872],

it may be proposed that the carbon chemical potential depends on stress,

and therefore the carbon diﬀusion is coupled with material deformationand

non-mechanical energy ﬂow. The ﬁnal carbon dissolution into material de-

pends on the thermal stresses due to carbonate expansion at precipitation

deﬁning coupling between carbonate precipitation, material deformation and