PHYSICAL REVIEW B VOLUME 43, NUMBER 13 1 MAY 1991

Initial magnetization and coupling nature of the high-T, grains in Bi-Pb-Sr-Ca-Cu-0

A. Oota and H. NojiDepartment of Electrical and Electronic Engineering, Toyohashi University of Technology, Tempaku cho,-

Toyohashi, Aichi 441, Japan

K. OhbaTechnica/ Research and Deuelopment Laboratories, Topy Industries Ltd. , Akemi-cho, Toyohashi, Aichi 441, Japan

(Received 11 July 1990)

The coupling between grains was studied through low-field ( (100 Oe) initial magnetizationM(H) on bulk samples of Bi-Pb-Sr-Ca-Cu-0 with a single high-T, (110 K) phase prepared underdifferent sintering times t, . The M(H) shows a variation with changing t, due to the weak links be-tween grains and its behavior is characterized by three threshold fields: H &, H2, and H3 ~ The firstflux entry into the sample along the grain boundaries due to flux creep occurs at H, , much lowerthan H„(-7 Oe at 77 K, —15 Oe at 0 K) for Hl c axis determined from the c-axis-oriented tapewith J, =5400 A/cm (77 K, H =0). Between H& and H2, M(H) deviates from a straight line dueto shielding of the sample volume; between H2 and H3 it falls on the Meissner line, an effect as-cribed to electrically decoupled grains. The first flux entry into the grains themselves occurs at H3,above which M(H) is governed by the intragrain magnetic properties of an Abrikosov lattice. Pro-longed sintering improves the weak links between the grains and suppresses both the flux creep atthe grain boundaries and the electrical decoupling between grains, which opens transport currentpaths and leads to the rise in T, (p =0).

I. INTRODUCTION

Since the discovery' of high-T, superconductors, mucheftort has been directed towards achieving high currentcapacity in bulk samples, from both scientific and techno-logical points of view. Superconductor applications athigh temperatures using high-T, superconductors requireone to solve the problems concerning the weak link of thegrains in oxides, the smallness of the coherence length,and the time-dependent relaxation of the critical current,which is negligible in conventional low-T, superconduc™tors. Studies on diamagnetic shielding, Meissner e6'ect,and thermally activated Aux motion (the so-called Auxcreep) using single-crystal and polycrystalline samples arelikely to provide a key to their solution. Of particularimportance in understanding the magnetic behavior ofhigh-T, superconductors is Aux creep. This results in adepinning of Aux bundles due to thermal activation, de-pending upon the ratio of the activation energy Uo to thethermal energy kz T. ' ' Flux creep seems to beenhanced at grain boundaries compared to the interior ofthe grains in polycrystalline samples that have weak linksbetween the grains. '

In a previous paper, we showed the indispensability ofa long sintering period to improve the weak links causingthe lowering of T, (p=O) in polycrystalline samples ofBi-Pb-Sr-Ca-Cu-O, although the high-T, (110 K) phasecould feasibly be singled out from the partially meltedliquid phase in a short time in air by introducing the re-press process. Here we present a systematic study using alow-field (H( 100 Oe) initial magnetization M(H) on thesame samples as used in Ref. 8; the samples are of a single

high-T, phase, but had various values of T, (p=O) de-pending upon their sintering times t, . To establish abasis for discussion, we briefly show the result for a Ag-sheathed tape of Bi-Pb-Sr-Ca-Cu-0 with J, =5400 A/cm(77 K, H=O) from I-V characteristics.

II. EXPERIMENTAL PROCEDURE

The polycrystalline samples of Bi-Pb-Sr-Ca-Cu-0 wereprepared by a solid-state reaction in air. The nominal ra-tio of the samples was

[Bi]:[Pb]:[Sr]:[Ca]:[Cu]= 1:0.3:1:1:1.5 .

The mixtures of Bi203, PbO, SrCO3, CaCO&, and CuOwere calcined at 800'C for 12 h. They were pulverized,mixed, and pressed into a disk shape. The pellets weresintered at 840 C for 20 h, and subsequently they werepulverized and re-pressed into a disk shape (the re-press).Finally, they were sintered again at 845'C for variousperiods of t, between 10 and 120 h. The scanning-electron-microscopic (SEM) images of fractured surfacesreveal that the samples composed of platelike microcrys-tals, typically 15—30 pm wide and 0.5 —1 pm thick. Thegrain size and the porosity are not aftected by prolongedsintering; this is consistent with the behavior of the ap-parent density pAP determined by the Archimedes'method. The value of 4.1 g/cm for pAp after the re-pressfirst decreases during sintering with sintering time t, upto 20 h, and then, after 40 h, approaches the saturatedvalue of 3.1 g/cm . Here, we note that no positive evi-dence for the oriented microstructures can be observed inthe SEM images, which are compared with the highly

43 10 455 1991 The American Physical Society

10 456 A. OOTA, H. NOJI, AND K. OHBA 43

oriented textural structures in polycrystalline samplesprepared by the intermediate pressing process.

The Ag-sheathed tape of Bi-Pb-Sr-Ca-Cu-O, 0.1 mmthick and 1.8 mm wide, was prepared by a method com-bining cold working and sintering. The x-ray-diffractionand susceptibility studies, together with SEM images,showed that the tape is of a single high-T, phase withT, = 109 K and has highly oriented microstructures witha tape surface perpendicular to the c axis. The J, valueat 77 K for H=O was determined to be 5400 A/cm fromthe I- V characteristics, and 7600 A/cm from magnetiza-tion measurements. J, at 5 K was found to be 7X10A/cm at 5 K from magnetization measurements. De-tails of the preparation method and the characteristicswill be published elsewhere. '

The initial magnetization M (H) was measured at inter-vals of 1 Oe in low fields of H&100 Oe after zero-fieldcooling by a superconducting quantum interference de-vice (SQUID) magnetometer (Quantum Design MPMS).To avoid the inhuence of the demagnetizing factor, mea-surements were made under a longitudinal configurationon small bars with typical size of 1.2X1.2X7 mm cutfrom sintered pellets. The absolute value of H and the re-sidual field were checked by the measurements usinghigh-purity Pb (T, =7.19 K) as a standard, whenever anew run was performed. All of the data analyses onM(H) were made by a PC 9801 computer.

ni. RESUI,TS ~XD DISCUSSION

As described in Ref. 8, all the samples studied herehave the same difFraction patterns as the ones used for asingle high-T, phase;" they do not reveal any traces ofthe low T, (80 K) ph-ase nor any other impurity phases.In addition, the susceptibility y(T) measured under fieldcooling (FC) at H= 10 Oe falls on nearly the same curve,independent of t, . It has a sharp drop at 108 K andshows no stepwise transition due to the impurity phases.These findings lead us to conclude that all the samples areof a single high-T, phase, independent of t, . Neverthe-less, we also find that T, (p =0), determined from resistivi-ty behavior under measuring current density J-0.2A/cm, is strongly dependent on t, and decreases to farbelow 100 K, especially when the sample is subjected to ashort sintering time of t, 40 h. Considering thatT, (p =0) is regarded as the temperature at which J isequal to the critical current density J„the lowering ofT, (p=0) is likely to originate from the weak links be-tween the grains, causing a disturbance in transportcurrent paths and lowering the value of J, .

To examine the intergrain coupling between the high-T, grains, we measured the M(H) curves on the samesamples as used in Ref. 8, and present the typical resultsfor 20 K in Fig. 1. Although there exists a significantdependence of M (H) on r„the behavior of M (H), whichis characterized by three threshold fields relating to theAux entry into the samples, can be schematically shown(Fig. 2). As H increases, M(H) shows an initial increasewith the slope of 2, , deviates from this straight line atthe field Hi, and then falls on the second straight line

1.2

0.8

0.6

T= 20K

~ 0~ o

o~o~oo a oO~

O~

~ pOQ

M

C.

i

0.4

0.2

ts = 20h

40h

80h

0 120hHigh-Jc Tape

(H J c)I I

20 40

(Oe)

60 80

FIG. 1. Initial magnetization M(H) at 20 K for the bulksamples prepared under sintering time t, . Also shown is the re-sult of M(H) in arbitrary units for Ag-sheathed tape withJ, =5400 A/cm (77 K, H=O} under the Hlc axis.

As

FIG. 2. Comparison for typical behaviors of the magnetiza-tion curve M(H) between the bulk sample and the powder.

with slope 3, in the range between H2 and H3. The ex-trapolation of the second line to H=O Oe is likely tocome near the origin in the M-H plane. At higher fields,H )H3, M(H) deviates from the second line.

The slope 2; is ascribed to the shielding of the samplevolume, due to electrically coupled grains. The Aux ex-pulsion 4'~ A; ~

evaluated from A; at T=20 K using pApranges from —60% to —90% for all the bulk samples.

A, ~seems to increase gradually with t, . This may be

due to the fact that the sample subjected to a smaller t,has larger errors in the value of A; (because of smallerH& ); alternatively, it has a small amount of a secondphase that is not detectable by x-ray diffraction. Howev-er, the latter reason is not as feasible in this case, becausethe y(T) under FC at H=10 Oe falls on nearly the samecurve, independent of t„showing a sharp drop at 108 K

43 INITIAL MAGNETIZATION AND COUPLING NATURE OF THE. . . 10 457

1.2

T Bulk Powder

20K O ~77K

00 ~0 ~eO

0e

0.8—

E0.6—

0~o ~

'

0 ~ Z, Lk

~O

0.4—

0 20 40

H (Oe)

60 80

FIG. 3. Initial magnetization M(H) at 20 and 77 K for thebulk sample subjected to sintering for t, = 120 h and the powdercrushed from it.

and having no stepwise transition. There may be anoth-er explanation —that the London penetration depthcharacterizing the field penetration into the sample sur-face is affected and raised by the weak link of thegrains. ' '

The value of H&

strongly depends on t„rejecting theintergrain coupling between the grains. The shorter theperiod of t„the lower the value of H&. To explainwhether H, is attributable to the lower critical field H, &,

we also measured the M (H) curve for the high-J, tape ofBi-Pb-Sr-Ca-Cu-0 under the Hlc axis, and also includethe typical result in Fig. 1. Using the procedure reportedrecently, ' ' the value of H„is determined from the plotof (H+4vrM)'~ versus H for the data shown in Fig. 1.Here, H, &

corresponds to a minimum value of H, &as a

function of 0, the angle between H and the c axis. Thismethod provides the more precise H, &

compared to thatobtained by a method based on the deviation from thestraight line in Fig. 1. ' The analysis yields H, &

= —15Oe at T-0 K, which is comparable to 9 Oe forBizSrzCaCuzO„. ' However, it is much lower than thevalues of 70 Oe for (La, „Sr„)zCuo~(Ref. 14) and 180Oe for YBazCu30 . ' Evidently, the H& for the bulksample is smaller than the H

&for the high-T, phase.

The lowering of H&

from H, &will be ascribed to the fIux

creep at the grain boundaries, ' which will be discussedlater.

To clarify the origin of the second straight line inM(H) for the bulk samples, we measured the M(H)curves for the crushed powder prepared from the bulksample of t, = 120 h, as shown in Fig. 3, together with thecurves for the bulk sample. The typical behavior ofM(H) for the powder is also included and compared withthat for the bulk in Fig. 2. As can be seen from Figs. 2and 3, the M(H) for the powder increases with the slope

of A, , which is nearly the same as 3, for the bulk andhas no second slope. Since the powder consists of electri-cally decoupled fragments, typically 15—30 pm wide and0.5 —1 pm thick, 3; should be attributed to the Meissnercurrent Aowing in the interior of the fragments. In thiscontext we note here that the susceptibility y(T) for thepowder under heating at H=10 Oe after zero-field cool-ing (ZFC) falls on nearly the same curve measured underan FC run. Incidentally, the ZFC and FC data at T-0K turn out to be —6.9 and —6.3 in 10 emu/g, respec-tively, which approach the value of —7.8 in 10 emu/gfor the bulk under FC at H=10 Oe. Taken together, thesecond slope of 3, for the bulk would be attributed tothe electrical decoupling between grains as H increases,which would originate from the same cause as A; for thepowder. The flux expulsion 4'~ A, ~

from the interior ofthe grains, evaluated from both 3, at 20 K and thetheoretical density p,h

=6.3 g/cm, turns out to be80—90% insensitive to t„which approaches perfect Aux

expulsion.The value of H3 =40—45 Oe at 77 K, which is nearly

independent of t„is about six times as large as H&=7 Oe

at 77 K obtained from the high-J, tape. Moreover, theH3 for the bulk sample of t, = 120 h agrees fairly wellwith the deviation field H

&of the powder from the initial

slope in M(H) (Figs. 2 and 3). In a system of randomlyoriented grains, we would expect Aux entry into thegrains perpendicular to H; this begins at H3 as H in-creases. To obtain the H„for the H~~c axis, however, thefield H3 has to be corrected for the demagnetizing factorD. Since H/(1 D) is the f—ield at which the edge of thegrains perpendicular to H is actually encountered, we canapproximate a better H,'i as H3 /(1 D). Accor—dingly,the anisotropy ratio c=H, & /H, &

for the high- T, phasewould be much larger than the ratio of H3 /H„=6at 77K, although it is dificult to determine D. In this contextwe note here that the c. is approximately 20 for the low-

T, phase of BizSrzCaCu~O, " and between 3 and 4 for(La, Sr„)zCu04 (Ref. 14) and YBazCu30„.'

Relating to the first Aux entry into the samples, it isworthwhile mentioning that the sample of t, =120 hshows a logarithmic time relaxation for t (2X10 s inmagnetization under ZFC at H=20 Oe and T=10, 50,and 77 K. Although the logarithmic relaxation observedhere may be explained by a superconducting-glass mod-el, ' we use the conventional Aux-creep model. ' Con-sidering that the measuring field H=20 Oe lies betweenH, and H3, the Aux creep at the grain boundaries(FCGB) (Refs. 4 and 7) would occur in the sample andcause the lowering of H& from H, &. The relaxation rateS =dM/d (lnt) at H=20 Oe is approximately 1, 9, and 2in units of 10 emu/cm at 10, 50, and 77 K, respective-ly. S may show a peak, and decrease with increasing tem-perature. In the present study S at H=20 Oe is muchsmaller than the previous result on magnetic relaxationat 1 kOe for polycrystalline Bi-Pb-Sr-Ca-Cu-O with a sin-gle high-T, phase; this may originate from a large activa-tion energy Uo for FCGB (roughly of the order of 1 eV),or from the difference of H (20 Oe and 1 koe). In thiscontext we note here that the ac-susceptibility study on

10 458 A. OOTA, H. NOJI, AND K. OHBA 43

polycrystalline Y-Ba-Cu-O showed a large Uo for FCGB( —12 eV at H=O and —1.2 eV at 10 Oe) and the de-magnetization study' on single crystals of Y-Ba-Cu-0showed a significant dependence (rise) of S with increas-ing H.

In Fig. 4, we present schematic diagrams for the fluxentry into the sample. In the region H &H&, whereB

M(H) has the initial slope A;, shielding currents fiow onthe sample surface because of electrical coupling of thegrains, so that no magnetic Aux enters into the sample[Fig. 4(a)]. As H increases further, exceeding H I, fiux en-try due to Aux creep at the grain boundaries occurs, sothat M(H) shows a deviation from the straight line ofslope 2; [Fig. 4(b)]. In the region H2 (H (H3, whereM(H) has the second slope of A, , the Aux enters solelyalong the grain boundaries, while no Aux entry into thegrains occurs because of the Meissner current Aowingthrough the interior of the grains [Fig. 4(c)]. Consideringthat the A, for the bulk agrees well with the A; for thepowder, the FCGB presumably becomes the Aux Aowthat is characterized by the predominance of the Lorentzforce over the pinning force. Further increase of Hbeyond H3 allows the fiux to enter into the grains [Fig.4(d)], where the behavior of M(H) would be governed bythe magnetic properties of the intragrain Abrikosov lat-tice. We note here that threshold fields in M(H) charac-terizing the flux entry, which are difFerent from H, &, havebeen previously suggested for polycrystalline samples ofLa-Sr-Cu-O. 12

The values of A, and A, , together with y(T) mea-sured under FC at H= 10 Oe, are shown as functions oftemperature for the bulk sample of t, =120 h in Fig. 5.For comparison, the data of A; for the powder are also

3.0

2.5—

A'

0 A,'~ A',

X(T)

2.0—

o 1 5

( FC H = 100e)

20 40 60

T (K)

80 100 120

included. As mentioned before, the A, agrees well withthe A; over the whole temperature range of T & T„andnearly approaches the y(T) value, except that y(T) isslightly lowered by the Aux pinning in the interior of thegrains. At low temperatures, where grains are electrical-ly coupled with each other, A; shows a deviation from

FIG. 5. The initial and second slopes A; and A, obtainedfrom the magnetization curve M (H), together with the suscepti-bility data g(T) under field cooling (FC) at H=10 Oe, for thebulk samples subjected to sintering for t, = 120 h. Also shown isthe initial slope A; of the magnetization curve M(H) for thepowder.

120 30

100— —25

Tc80— —20

H', &H&H,'{b)

60— O15

40— —10

I I I I 020 40 60 80 100 120 140

ts (h)

H', &H&H;{c)

H&H,'{d)

FIG. 4. Schematic diagrams for the Aux entry into the bulksample with increasing H.

FIG. 6. The threshold field H& for the magnetization curveM (H) and the decoupling temperature TD of the grains, togeth-er with the T, (p=0) determined from resistivity measured un-der J-0.2 A/cm, for the bulk sample plotted vs sintering timet, .

43 INITIAL MAGNETIZATION AND COUPLING NATURE OF THE. . . 10 459

100-

80—

60—

40—

20—

100

80

s = 120h

20 40 60 80 100 120

creases from —100 to —50 K. Such a decrease of TD isascribed to the enhancement of the weak links betweenthe grains. It is evident from Fig. 6 that both TD and H,obtained here are strongly related to the value ofT, (p =0) from resistivity behavior measured underJ-0.2 A/cm . Prolonged sintering improves the weaklinks and enhances the value of TD and H &, which leadsto an increase of T, (p =0).

The magnetic diagrams for the bulk samples of t, = 120and 20 h from the behaviors of M(H) are shown in Fig.7; they clearly illustrate the intergrain coupling betweenthe grains. It is evident that there are four distinctive re-gions; these are ascribed to the shielding current, FCGB,Meissner current (presumably fiux fiow), and Abrikosovlattice, respectively. They are separated by three charac-teristic fields H&, H2, and H3, where H3 is in goodagreement with H

&for the powder. The weak link of the

grains lowers both H, and H2, while causing no notablechange in H& and/or H&. As a consequence, it narrowsthe shielding region, leading to the region vanishing in anextreme case, while expanding the Meissner region.

IV. SUMMARY

60

40

20

20 40 60

(K)

80 100 120

FIG. 7. The magnetic diagram for the bulk samples subjectedto sintering for t, = 120 h (a) and 20 h (b).

We showed that the initial magnetization M(H) pro-vides available information on the coupling nature of thegrains in sintered bulk samples, that is not obtained byother measurements such as resistivity and susceptibility.The behavior of M(H) shows a significant differenceamong the bulk samples of Bi-Pb-Sr-Ca-Cu-0 preparedunder different periods of t„although all of them are of asingle high-1; phase, according to x-ray-diffraction andsusceptibility studies. This difference can be attributed tothe difference in the intergrain coupling between thegrains. The results also show that the four distinctive re-gions due to shielding current, Aux creep at the grainboundaries, Meissner current, and the Abrikosov latticeappear in the M(H) curve as H increases.

ACKNOWI. EDGMENTS

An increase in temperature causes the electricaldecoupling of the grains, due to a decrease in transportcritical current, and at high temperatures 2; agrees with

and/or A, . As the sintering time t, decreases from120 to 20 h, the decoupling temperature of the grains, TD(defined as the temperature at which 2; falls on A, ), de-

We would like to thank Mr. S. Fukutomi for technicalassistance, and Mrs. A. Yata and K. Hayashi for samplepreparation of the Ag-sheathed tape. This work was par-tially supported by the Ishida Foundation, and also bythe Foundation for the Promotion of Material Scienceand Technology of Japan.

~J. G. Bednorz and K. A. Muller, Z. Phys. B 64, 189 (1986).Y. Yeshurun and A. P. Malozemo6; Phys. Rev. Lett. 60, 2202

(1988).C. W. Hagen and R. Griessen, Phys. Rev. Lett. 62, 2857 (1989).

4M. Nocolo and R. B. Goldfarb, Phys. Rev. B 39, 6615 (1989).5M. A-K. Mohamed, J. Jung, and J. P. Franck, Phys. Rev. B 39,

9614 (1989)~

H. S. Lessure, S. Simizu, and S. G. Sankar, Phys. Rev. B 40,5165 (1989).

7S. L. Sinde, J. Morrill, D. Goland, D. A. Chance, and T.McGuire, Phys. Rev. B 41, 8838 (1990).

A. Oota, K. Ohba, A. Ishida, H. Noji, A. Kirihigashi, K.Iwaski, and H. Kuwajima, Jpn. J. Appl. Phys. 29, L262(1990).

T. Asano, Y. Tanaka, M. Fukutomi, K. Jikihara, andH.Maeda, Jpn. J. Appl. Phys. 28, L595 (1989).

' A. Oota, A. Yata, K. Hayashi, and K. Ohba, J. Appl. Phys.69, 3182 (1991).

U. Endo, S. Koyama, and T. Kawai, Jpn. J. Appl. Phys. 27,L1476 (1988).S. Senoussi, M. Oussena, M. Ribault, and G. Collin, Phys.Rev. B 36, 4003 (1987).

10 460 A. OOTA, H. NOJI, AND K. OHBA 43

T. L. Hylton and M. R. Beasley, Phys. Rev. B 39, 9042 (1989).M. Naito, A. Matsuda, K. Kitazawa, S. Kambe, I. Tanaka,and H. Kojima, Phys. Rev. B 4I, 4823 (1990).

I~B. D. Biggs, M. N. Kunchur, J. J. Lin, S. J. Poon, T. R.Askew, R. B. Flippen, M. A. Subramanian, J. Gopalakrish-nan, and A. W. Sleight, Phys. Rev. B 39, 7309 (1989).

L. Krushin-Elbaum, A. P. Malozemoff, Y. Yeshurun, D. C.Cronemeyer, and F. Holtzberg, Phys. Rev. B 39, 2936 (1989).

' C. Rossel, Y. Maeno, and T. Morgenstern, Phys. Rev. Lett.62, 681 (1989).Y. Yeshurun, A. P. Malozemoff, F. Holtzberg, and T. R.Dinger, Phys. Rev. B 38, 11 828 (1988).