Thermal conductivity of doped YBCO single crystals

Thermal conductivity of doped YBCO single crystals

Physiea C 2 3 5 - 2 4 0 (1994) 1 4 8 7 - 1 4 8 8 PHYSICA North-Holland THERMAL CONDUCTIVITY OF DOPED YBCO SINGLE CRYSTALS A. Inyushkina, A. Taldenk...

235KB Sizes 2 Downloads 48 Views

Physiea C 2 3 5 - 2 4 0 (1994) 1 4 8 7 - 1 4 8 8



THERMAL CONDUCTIVITY OF DOPED YBCO SINGLE CRYSTALS A. Inyushkina, A. Taldenkova, S. Shabanova, and T. Uvarovab

aRussian Research Centre "Kurchatov Institute", 123182 Moscow. Russia bShubnikov Institute of Crystallography 117333 Mosco~L Russia Thermal conducti'~ity K ofthe three systems: ox3'gen deficient YBa2Cu30 x (x = 6.05 -6.95), Yl.zPrzBa2Cu3Ox (z = 0 - 1, x = 6.05 - 6.95), and YBa2(Cul.vZnv)306.95 (y = 0.0 - 0.05) in a-b plane have been systematically studied as fimction of temperature and magfieti6 field. The evolution of the thermal conductivity upon doping depends on the origin of metal-insulator transition. The unusual temperature dependence K(T) of non-metallic crystals suggests that some unconventional mechanism of phonon relaxation dominates in these crystals. Analysis of the experimental data confirms the phononic explanation of the K(T) for Y B a 2 C u 3 0 7.


The heat transport in HTSC is intimately related to the mechanism of superconductivity. The experimental results on K(T), mainly of YBCO. have been interpreted in terms of COlwentional and unconventional models of HTSC. There are two alternative interpretations of the dominant mechanist1 of heat transport: electronic and phononic. Moreov:~r; irrespective of any model, it is unclear what processes dominate the relaxation of heat carriers.

3: °...i



0.01 e~

E o




~ ,,,l 1 ()

Apparently in insulating crystals the phonon transport is responsible for the thermal conductivity. Fig. 1 shows the K(T) of single crystals of nonmetallic cuprates. The simple Cailaway model with a few phonon relaxation processes fits quite well to the K(T) data for the CuO crystal [ 1]. The boundary scattering determines the low-temperature region, point defects limit the phonon mean free path at intermediate temperatures and the phonon-phonon processes cause nearly T -1 dependence of K(T) at high temperatures. These results testify the common behavior of CuO dielectric crystal. In Pr2CuO 4, ,h,, K~T, Is suppressed at T > 3f) K. The suppression is supposed to be due to the phononphonon process and intensive coupling between phonons and Cu spins in CuO 2 planes in broad vicinity of the antiferromagnetic transition (T N ~ 270 K). Our data for YBa2Cu30&o.~ (YBCO6) considerably differ from that of Hagen et al. [2]. We obscn,ed a maximum in K(T) at about !80 K. At high temperatures the dependence K(T) is steeper than T "i and tends to become saturated at T > 400 K. Though there is not local minimum at T N ~ 420 K. we suggest that



i I ~,,l





! 0()

Temperature, K Figure 1. Temperature dependence of thermal conductivi .ty of non-metallic cuprates. the phonon scattering is enhanced by the phonon-spin interaction in Cu(2) position. Below 180 K the K(T) rapidly decreases, displaying unusual for common insulators dependence. The standard sum of relaxation rates in the power-law r,,~--, ,.'"T'" (as in the case of CuO) cannot describe the K(T). Therefore. the data implies that, in YBCO 6, some unusual mcchat ~ i d. ..~.[.I O"l ~ k, U H........ tU3 3 i . it, O..... 3 t ¢ i i l t ;,,;,,. ldli~ nism of phonon l-~' . .O.i l.L i i :" in this range. We suppose that this mechanism must results in a resonance-t)ge phonon scattering or a nonmonotonic temperature factor in the scattering rate. Thermal conductivity ofPrBCO 6 and PrBCO 7 has the same features as that of YBCO 6, however the magnitude of K(T) of the formers is much smaller. The disappearance of the peak in K(T) in Y-diluted (PrY)BCO 7 crystals correlates with dccrcasing of Cu(2) Neel temperature. In this case. the spin-assisted phonon scattering bcco,nes substantial at temperatures

0921-4534/94/S07.00 © 1994 - Elsevier Science B.V. All rights reserved. SSD! 0921-4534(94)01307-1


..I. Invushkin ct al.."Phvsica (' 2.¢5 240 (iqO4l 1487 1488

where the above-menlioned unusual relaxation process have strong influence. The reason for slrong phonon relaxation al T < 180 K in YBCO 6 and PrBCO m a y be the ant~fcrromagnetic correlalions of Cu(I) spins in the CuO chains. There is a clear evidence 131 that they exist in PrBCO 6, where the drop in K(T) is lhe most pronounced. Supposedly these correlations also exist in YBCO 6 and PrBCO 7. In PrBCO 7, we obsen'ed a xvcak suppression of K(T) at T N = 14 K. This indicalcs the substantial phonon coupling with Pr spins.

i Ii

[-- II ¢c L F




i i i IiI ~

1 1! i ! i i i I


"'" vl,co, 1).(114:74 K

0 ~ I


~IIIlmmImmwmiIImwIm I




















Magnetic field, kOe

The estimated lattice thermal conductivity KI(T) of YPCO 7 in Ihc normal slate (To T:. The condensation of frec carriers al T < "1"¢results in rapid increase of the K I. This rcsu!! conlirms the phonon scenario of the thermal conductivity of HTSC 14 I. The evolution of Ihc K(T) upon doping depends slrongly on Ihc origin of the dopant. The compounds near metal-insulator boundat T have the smallest value ~1~

Yo~l'rodK'()7 '01~' . 4.4' .***** (78 K)


./"/ " . I ' . . . . .



..."'" ,**" a ~ .,,,,,."-


Figure 3. Magnetic heat resislance of oxygen deficient and doped YBCO. The substitution of Y by Pr gradually decreases the T c. and consequenll)(he peak in KI,T) disappears al much higher Pr conlcnl. For example, in the (PrY)BCO sample xxith 20% Pr, thc KiT) rise bcloxx T,: is xxcli-pronounccd, ii is inlercsling Ihal in Ihis case Ihc value of thermal conduclivilv is smaller over whole temperature range Ihan Ihal of pure YBCO. This is a result of the phonon scattering by Pr ions which act as poinl defects too.




The nlagnclic heal resislance W=K "1(H) - Kl ( 0 ) iv, t-ITSC superconduclors is associalcd xvilh scalierin L-of (he heal carriers on Ihc core vorllccs II I. This valt,c expresses Ihc difference bcmccn Ihe supcrconducling and Ihc normal SlalC ofCuO 2 planes (Fig. 31. In YBCO:Zn. Ihc W is very small showing the exIremelv high inlensilv of Ihe pair-breaking effect in Ihe CuO 2 planes even for Zn concentration less Ihan 0.0 I. Apparently there is :; high densd.v of slales wilhin SC gap. On the conlrarv, in YBCO6. 6 and (Y-Pr)BCO Ihe magnitude of W is aboul the same :,Is in pure YBCO. Wc inlcrprclcd this as an evidence of wclldeveloped gap in these HTSC. This work is supporlcd by gran| No.93(~88 of Russian Slate Program on HTSC and in parl by Grant No.NI~)G~IHI oflhc IIHcrll:l(ional Science F'oundalion.

~' ().1() .~_


YI~C()"/.u(,, ~J()14)~



1...., 0.01



iJtl t t ~ I ! 00 Temperature, K Figure 2. Thermal conduclivilv of oxygcl~ dcficicn| and doped YBCO. I


of lhermal conduclivitv. The cicclronic propcrlics el" the CuO 2 planes deIcrminc lhc sharpness and relalivc 'ae~m "" • of "h t~ c peak in llic K(T). The reduction of oxygen conccnlralion x from 7 Io 6.6 does tlol change dramalicalh these propcrtics in pure YBCO. as Ihc thermal conductivity dcmonstratcs rapid increase below T:. On the contrary, in Zn doped cr3.stals, the peak is aimos! suppressed al surprisingly small impurily level, indicating the strong reduction of the order paramclcr (Fig. 2).



REFF.RENCES I. V V Florenl'cx et a l . Sxcrkhl)tovodin!ost {KIAE) 3 (199()) 23()2 [Supcrconductivily. 3 (1991) $37S1. -.'~ S.J. Hagen et al.. Phys. Rex'. B 40 (1989) o.,oo~,.,e,:,. 3. N. Rosov era/.. Phvsica C 2{14 (1992) 17 i. 4. L. Tcxvord and T. Wolkhauset~. Solid Slale Commtm. 70 (1989) ,,',o