Effect of Filament Sausaging on E-J Characteristics in Bi-2223 Tapes Yuri Ueno, Edmund Soji Otabe, Masaru Kiuchi, Teruo Matsushita Abstract_ Sausaging of superconducting filamentsTstrong-he lengthwise cross-se* *ctional view of the tape was ob- ly influences the E-J characteristics in superconductingserved with an optical* * microscope and the thickness of a su- tapes. In this research the effect of filament sausaging was investigated. It was found that the thicknesspofefilamentsrconducting filament * *was measured every 15.6 ~m along was distributed within a factor of about 3.5talonghtheelength.length and the th* *ickness distribution was obtained. This promises that the uniformity of the criticalTcurrenthden-e current-voltage* * curves were measured by the four sity can be drastically improved when the filament thickness can be made uniform. probe method under the magnetic fie* *ld parallel to the c- Keywords_Bi-2223 tape, E-J characteristics,asausagingxofis. The pulse transpor* *t current with the period of 1 s filaments was applied to the specimen to redu* *ce the joule heat at current leads and the voltage was m* *easured across the volt- I.Introduction agedterminalseseparatednbys1.0icm.t* *yThewcriticalacurrents determined by the electric field criterion of T is known that the E-J curve in superconductingEBi= 1:0 x 10-3 V/m. I-2223 tapes is not so sharp as in metallic superconduc- tors. This feature in the Bi -2223 tapes is characterizedIII.Results and Discus* *sion by their low n values especially at high fieldsFandiatghigh. 1 shows the obtain* *ed histogram of the supercon- temperatures. The low n value is caused by adveryuwidecting filament thickness,* * d. It is seen that the histogram distribution of the critical current density.cTheawidendis-be approximately exp* *ressed by the Gaussian distribu- tribution is attributed to the weak links betweentsupercon-ion of = 2:56 x * *10-5 m and oe2 = 3:36 x 10-11m2. If ducting grains as well as an essential randomwdistributione define the maximum * *and minimum thicknesses by 5% of the pinning strength. However, it is alsooattributedftothe peak in the distr* *ibution, the variation factor given the sausaging of superconducting filaments. Thatbis,ysincethe ratio of maximum * *value to minimum value is 3.49. the ceramic oxide is much harder than the silver matrix, the oxide layer thickness varies largely alongWtheelength-definebthe effective * *virtual critical current density, wise direction of the tape when rolled. In thisJcasecthe0, in the creep-free c* *ase and assume that the temperature local critical current density also varies alongathenlengthdinmagnetic field de* *pendencies are expressed as proportion to the cross-sectional area. In order to improve " ` '2#m * * ` 'ffi the n value in Bi-2223 tapes, therefore, it is necessarybtoJc0= bA1 - T_ Bf* *l-11(-1B__); evaluate quantitatively the effect of each factor on the n Tc * * Bc2 value. In this paper, the effect of sausaging is especially fo- cused and the distribution of thickness of superconductingwhere bA, m, fl, ffi * *are pinning parameters. In practical su- filaments is investigated. perconductors the flux-pinning stre* *ngth is inhomogeneous and distributed, and the filament t* *hickness is also dis- II.Experimental tributed. For simplicity, we assume* * that such distributions Specimen was a superconducting multifilamentaryaBi-re approximated by a simple* * distribution only of bAof the 2223 silver sheathed tape prepared by the powder-in-tubeform: method. The cross-section of the specimen was 3 mm x " * * # 0.22 mm, and the volume fraction of superconductingffila-(Ab) = bKexp-(logbA-_l* *ogbAm)2_;(2) ments was approximately 24%. The c-axis of the specimen 2boe2 was oriented normal to the flat surface of the tape. The critical temperature, Tc, was 109.5 K and Jcinwthehself-fieldere bAmis the most* * probable value, boerepresents the de- at 77.3 K was 45 A. gree of deviation and bKis a consta* *nt determined by the Y. Ueno is with Kyushu Institute of Technology,c680-4,oKawazu,ndition of norma* *lization. In the flux creep-flow model [1] Iizuka 820-8502 Japan (Telephone: +81-948-29-7683,the-mail:e pinning potential* *, U0, is obtained using Eqs. (1) and ueno@aquarius10.cse.kyutech.ac.jp) (2) and the E-J curve can be calcul* *ated. Above param- E. S. Otabe is with Kyushu Institute of Technology,e680-4,ters are adjusted so* * as to obtain an agreement between Kawazu, Iizuka 820-8502 Japan (Telephone: +81-948-29-7683, e- mail: otabe@cse.kyutech.ac.jp) experimental E-J curves and theoret* *ical ones. The solid M. Kiuchi is with kyushu University, 6-8-1,lHakozaki,iHigashi-ne in Fig. 2 sho* *ws the estimated distribution of bAwith ku,Fukuoka 812-8581,Japan boe2= 0:01 and bAm= 1:28 x 109. T. Matsushita is with Kyushu Institute of Technology, 680-4, Kawazu, Iizuka 820-8502 Japan (Telephone: +81-948-29-7663,Fe-rom the two distri* *butions in Figs. 1 and 2, the distri- mail: matusita@cse.kyutech.ac.jp) bution only of the pinning strength* *, can be estimated. For Fig. 1. Distribution of the superconductingFfilamentithickness.Theg. 2. Distri* *bution of experimental and theoretical values of bA. The solid line shows the Gaussian distribution ofssuperconductingofil-lid and bro* *ken lines show the theoretical distribution and the ament thickness. experimentally determined one. simplicity, we assume that the virtual critical current den- sity, Jc0, has the same dependencies as in Eq. (1) and that the parameter representing the pinning strength, A, has the same distribution as in Eq. (2). Since the critical current is proportional to d unless d is smaller than the correlation length of the order of 1~m, the distribution of bAcan be simply regarded as that of (d/)A. Thus, the distribu- tion of A is adjusted so that the obtained distribution of bA accords with the experimentally determined distribution. Fig. 3 shows the distribution of A with oe2 = 0:0011 and Am = 1:3 x 109 and the solid line in Fig. 2 represents the distribution resulted from the two distributions. A slight disagreement comes from the assumption that the distri- bution function of A is of the same form as that of bA. Thus, the real distribution of pinning strength is much sharper than the distribution of the effective pinning Fig. 3. Distribution o* *f A. strength influenced by the sausaging. That is, the influ- ence of sausaging on the E-J characteristics is very large. In other words, there is a large room for improvement3of.theIt is expected the * *E-J characteristics can be greatly E-J characteristics by reducing the sausaging ofifilaments.mproved by reducing * *the sausaging of filaments. Acknowledgements IV.Summary The authors would like to acknowle* *dge Dr. J. Fujikami The effect of sausaging on the E-J characteristicsaisnfo-d Dr. K. Ohmatsu of S* *umitomo Electric Industries for cused and the distribution of thickness of superconduct-preparing Bi-2223 speci* *men for the measurement. ing filaments is investigated. The following results are ob- tained. References 1.The thickness of filaments is distributed[over1factor]M. Kiuchi, K. Noguchi,* * T. Matsushita, T. Kato, T. Hikata and 3.5. K.BSato,i"Scaling-of2current-volt* *age2curves2in3superconductingsilver-sheathed tape wires", Physica C 278 (1997* *)@ 2.The estimated real distribution of pinning strength is much sharper than the directly observed distribution of the effective pinning strength.