PhysicaC 172 (1990) North-Holland
Flux creep and irreversibility
Pr 1.85Th0.1,Cu04 --y
J.L. Peng and R.L. Greene ’ Centerfor Superconductivity Research, Department ofPhysics and Astronomy, University ofMaryland, College Park, MD 20742, USA Received 10 August 1990 Revised manuscript received
High-quality crystallites of Pr,.ssTh0.15Cu04_, were prepared by a solid state reaction technique and aligned in epoxy by a magnetic field. Magnetic relaxation of the remanent moment was measured with field H parallel to the ab plane. We interpret our data with a thermally activated flux creep model, and show that the effective pinning energy U,, increases with increasing T between 2 and 15 K. We find that cl0 is not strongly dependent on applied field. The crossover from reversible to irreversible magnetization was found to follow the power law H= H( 0) ( 1 - T/T,)‘.‘. The value of ciO for H parallel to the ab plane is found to vary from - 8 to _ 40 meV between 2 and 15 K, respectively. Compared to YBCO, pinning in the electron-doped system is rather weak.
1. Introduction One of the unusual phenomena observed in highT, superconductors is the logarithmic relaxation of the magnetization at temperatures below the so-called irreversibility line. This was first studied by Miiller, Takashige and Bednorz [ 1 ] in LaBaCuO ceramic samples. The key observations made by them have remained the focus of attention to the present day. Significant magnetic relaxation has been found in all high-T, superconductors with Tc> 30 K, in both ceramics and single crystals. Recently, similar phenomena were observed [ 21 in crystals of the low-T, Chevrel phase PbMo&. This material has a short coherence length like the copper oxide superconductors but has a 3D rather than 2D crystal structure. At temperatures and fields below the irreversibility line, the magnetization of superconductor changes with time. While the field-cooled magnetization appears to be in equilibrium within experimental accuracy, the zero-field-cooled magnetization appears to be unstable and to relax slowly downward in magnitude. The remanent magnetization, obtained by turning off the field after field-cooling, also relaxes downward. These relaxations are generally logarith’ Also at IBM Research, 0921-4534/90/$03.50
Heights, NY 10598, USA.
0 1990 - Elsevier Science Publishers
mic in time. Yeshurun et al. [ 31 studied the strong magnetic relaxation found in YBaCuO and BiSrCuO crystals. They attributed this strong magnetic relaxation to weak flux pinning. The combination low pinning energy and high transition temperature results in a “giant” flux creep. Most of the magnetic studies to date have been concentrated on the hole-doped superconductors. The logarithmic magnetic relaxation and the presence of an “irreversibility line” in the H, T plane initially were considered to be unique features of the high-T, superconductors. It is therefore useful to raise the question whether similai properties are found in the electron-doped copper oxide superconductors. The absence of significant broadening of the resistive transition in a magnetic field [4,5] and the larger, in-plane, coherence length have suggested that flux creep may be less important in the electron-doped materials. In this paper, we present the first magnetic relaxation data for an electron-doped superconductor, Pr,.,,Th,.,5CuO~, and evidence for an irreversibility line similar to that found in the hole-doped superconductors. We focus here on relaxation of the remanent magnetization for H parallel to the crystal basal plane (ab). In the entire temperature and field regime the relaxation is linear with the logarithm of
time. We interpret our data within a thermall!, activatcd flux-creep model and present the temperature and field dependence of the resulting pinning potential I’,,.
iI_0.01 2. Experimental
Polycrystalline samples of Pr, HjTh,, ,5CuC~J_,were prepared by a solid state reaction [ 61 from high-puritl Pr,O, ,. [email protected]
, and CuO. Stoichiometric mixtures of the starting materials were ground thoroughly and heated in air using an .Al,OT crucible at 1000-C for I2 h. The resulting powders were reground and pressed into pellets and fired in air at I 1OO’C for I2 h followed by quenching to room temperature. Magnetization measurements indicated that all of the samples quenched in air were nonsuperconducting above 2 K. In order to obtain superconductivity. it was necessary to anneal the sintered pellets in a reducing atmosphere. We placed these pellets in a tube furnace with flowing Ar gas. Next to the samples. we filled an AlzOi boat with a fresh titanium sponge to help getter the oxygen. The samples were then annealed under the argon atmosphere at 830°C for 24 h. Finally the samples were furnace cooled to room temperature. The field-cooled ( FC ) and zero-field-cooled (ZFC) magnetization measurements were performed on a superconducting quantum interference (SQUID) magnetometer (Quantum Design). Figure 1 shows the magnetization data on the unoriented powder sample used for the grain-alignment measurements reported in this work. .After a correction for demagnetization we find a diamagnetic shielding signal close to - 1/47t at 5 K. The superconducting transition temperature 7,. defined at the onset of diamagnetic signal, is 2 1.4 K. while the IO-90% transition width of the field-cooled magnetization is 2 K. Compared to prior work, our samples appear to be of very high quality. Next we ground the powder sample into fine crystallites of particle size - 10 pm. They were mixed in epoxy (epo-tech 30 1) in the ratio of one part powder to two parts epoxy at room temperature and aligned in an applied magnetic field H= 50 kOe for about 10 h. The epoxy was allowed to harden by raising temperature to 80°C for 3 h in the field, producing a cy-
-0.03 L--J 0
1 I 10
T (K) Fig. I. FleWcooled (e ) and xro-field-cooled ( z ) magnct~/alwn as a t’unctlon of temperature for an unaligned powder Pr, n5T’h,, ,rCuO,_, sample I I/= i 0~).
lindrical specimen with the cylindrical axis parallel to H. No reaction of the crystallites with the cpoxl was detected as determined by either X-ray diffraction or magnetization measurements. To determine the degree of orientation of the specimen. an X-ray diffraction pattern was taken along the direction perpendicular to the specimen cylindrical axis. Figure 2 shows X-ray diffractometer scans for a grain-aligned Pr, ssTh,, , ,CuO,_) ( PrThCuO ) specimen (a) and a powder sample (b). These rcsuits indicate that the U/J planes of the PrThCuO crystallites align along the magnetic field II, as expected since the easy axis of magnetization for the Pr’+ ions is in the ah plane. However, the ah planes are not parallel to each other. Therefore, the (,-axis is randomly distributed perpendicular to the field. Due to this configuration. WC are not able to check the degree of alignment of these ah planes with respect to the applied field by a simple rocking curve diffraction technique and also cannot make measurements with El parallel to the C.-axis. For the relaxation measurement, the time dependence of the remanent moment .V was measured after field cooling the specimen and then turning off the field H at time t=O using a Quantum Design magnetometer. .4fter about 1,,=72 s, required b) the magnetometer to stabilize and become operational. the decay of lzil was recorded over several hours. Possible complications with this experiment were minimized by using short scan lengths (to avoid field gradients). avoiding 7‘or If overshoots and by mcas-
J.L. Peng, R.L. Greene /Flux creep and irreversibility in e-doped Pr, ssTh0 ,,CuO,_,
-3 5 0.24 = 0.22 6 -II
25 30 35 40 45 50 55 60 65 70 75 80
' ' ' ' 10
' s ' 30
Fig. 3. Field-cooled ( A ) and zero-field-cooled (0 ) magnetization M(T) as a function of temperature measured in a field of 5 kOe applied along the ah plane.
70 75 X0
25 30 35 40 45<-65 20 ( degrees
Fig. 2. X-ray powder diffraction patterns for a grain-aligned Pr1.85Th0.1SCu04_-vspecimen (a) andapowder sample (b).
uring only the irreversible part of the magnetization, i.e. the remanent A4 (this automatically eliminates the paramagnetic contribution due to the large Pr3+ moment )
T (K) Fig. 4. Temperature dependence of irreversibility lines for Pr,s5ThOrSCu0,_, along the ab plane. The solid line is a least square fit to a power law (see text).
3. Results The field-cooled (FC) and zero-field-cooled (ZFC) magnetization M(T) were measured for various applied fields H in the basal plane. From the temperature T* at which the FC and ZFC curves merge (within 0.1% difference), we define the irreversibility line N( T*).In fig. 3, the magnetic moment measured in 5 kOe is plotted versus the temperature, where the reversible magnetization is achieved above T*.Their irreversibility line, shown in fig. 4, can be described by the power law: H=H(O)(l-T/T,)“.
Our sample fits eq. (1) with m= 1.69, H(O)=41.9 kOe and T,=21.4 K. The solid line in fig. 4 repre-
sents the best fit to eq. ( 1). This power law for the irreversibility line is similar to the one observed in other high-T, materials. Here H(0) is just a fitting parameter, which is the extrapolation of H on the irreversibility line to T= 0. Figure 5 exhibits the typical time dependence of the remanent magnetic isotherm at 5 K, after cooling in a field H= 5 kOe parallel to the ab plane. The solid straight line represents a best fit to the data. Note that, to achieve the critical state, the applied field should exceed H*, the field where the flux fully penetrates the sample. As will be discussed below, the value of H* depends on the temperature. For the temperature and field regime studied here the relax-
J. L. Peng, R.L. Greene /Flux
creep and irrevrrsihility
in e-doped Pr, 85Th, ,,C‘uO,_
ture range. The relaxation rates for H= 1.jkOe and 5 kOe in fig. 6 increase with decreasing temperature down to 2.5 K. These data are similar to those rcported [ 8.91 for YBCO. In the following discussion, an effective pinning potential is established by onl) using the relaxation rates in the critical state.
4. Discussion I 1000 t (set) Fig. 5. The remanent magnetization at 5 K after cooling in a field of 5 kOe is plotted versus time. The solid line represents a least square fit to the data.
s 2 0.006 -
E F 0.002 5
0.000t 0 T (K)
Fig. 6. Temperature dependence of the relaxation rate dM/dlnl measured with fields of 0.4 kOe, 1.5 kOe and 5 kOe along the ah plane.
ation of the magnetization is linear with the logarithm of time. From the relaxation data of the remanent moment, the relaxation rates dM/dln( t) can be calculated and are shown as a function of temperature in fig. 6. For H=0.4 kOe the rate increases with temperature, peaks at -7 K and drops to zero at -20 K. It should be pointed out that the relaxation rates observed at low temperatures for H= 0.4 kOe are not representative of relaxation in the critical state. At temperatures below 7 K, the incomplete flux penetration in the sample can lead to a more complicated situation, that is, the critical state is not well established. By increasing the magnetic field, the peak of the relaxation rate shifts towards the low tempera-
One object of this work was to compare the pinning properties of the electron-doped superconductors with those of the higher T, hole-doped cuprate superconductors. In YBCO crystals, the estimated coherence lengths range from I S-30 A in the ab plane and from 3 to 7 A in the (‘ direction. Since the coherence length sets the length scale for variation of the superconducting order parameter, the superconductivity will be much more sensitive to small-scale structural or chemical imperfections than in a classic superconductor. In general, the small coherence length gives a low pinning potential, and results in strong flux creep at high temperature. One of the major differences between the electrondoped superconductors (including PrThCuO) and the higher T, oxides is the magnitude of the coherence lengths. Recent experiments [4,5] on single crystals of electron-doped NdCeCuO and SmCeCuO showed that an applied magnetic field causes a parallel shift of the resistivity transition to lower temperatures, as found in conventional superconductors, but unlike the behavior found in the hole-doped high-T, compounds. Ginzburg-Landau coherence lengths estimated from the upper critical magnetic field are 70-80 r\ in the basal plane and 3-I 5 .A along the c-axis. These values are larger than those of the hole-doped high-T, superconductors. Because of this it was thought that flux creep might play a less important role in the electron-doped superconductors. From the Anderson-Kim flux creep model [IO]. the relaxation of the magnetization M can be written M(t)=M,(
In eq. (2), U. is the average energy barrier jumped over by the thermally activated flux lines (pinning potential). It is, in general, field and temperature dependent. The characteristic hopping time 7, is not too well known but is usually estimated to be of the
J.L. Peng, R.L. Greene /Flux creep and irreversibility in e-doped Pr1.B5Th0.1sCuOI-Y
order of lo-” to lo-i0 s. The time t is the effective measuring time. The pinning potential for PrThCuO for H parallel to the ab plane can be extracted from our data by considering the decay rate normalized to the initial moment MO:
R=M--‘M = kT ’
We approximate MO by the first measured moment M( to), where to= 72 s. Because of the slow change of M as a function of temperature, using another M(t) for this approximation of MO does not change the essential nature of present result. Combining the values of the initial moment MO and the flux creep rate dM/dlnt, the pinning potential Uo at various temperatures and fields can be estimated. The results are shown in fig. 7, where the solid line is a guide to the data. The pinning potential U. increases with increasing T, and is not strongly dependent on applied field. This is not consistent with the form of the intrinsic pinning potential (U. - ( 1 - T/Tc)3’2,1B) proposed by Malozemoff [ 71. Similar results were also obtained by Xu et al. [ 81 and Campbell et al. [ 91 in YBCO. They gave a possible explanation for the anomalous temperature dependence of W. by accounting for the time dependence of the critical current density in the creep regime. In their picture the measured pinning potential should increase towards the intrinsic pinning potential as T increases, as found in our experiments. Another possibility suggested by Malozemoff and Fisher [ 111 is that eq. (2) may need
10 T (Kl
Fig. 7. Temperature dependence of the pinning potential L’Oderived from the experimental data at different fields.
to be modified for the oxides, in which case the meaning of U. is unclear. In YBCO the quoted value of U. at low tempermeV for H parallel to the c-axis ature is -20-50 (perpendicular to the ab plane), at least an order of magnitude smaller than in typical low-T, superconductors. Because the flux creep contribution scales as kT/ Uo, materials with low pinning energies and high Ts can exhibit a “giant” flux creep, several orders of magnitude larger than in conventional type II superconductors. To compare our work with that of YBCO, we need the relaxation data in the c direction. Unfortunately, as stated earlier, we only can align the powder along the ab plane. This complicated our experiment and makes it difficult to make a direct comparison with YBCO. However, we have enough information to draw some general conclusions from our work. We have done a similar analysis of M( t) measured on a randomly oriented powder and find that U. is only about 30% smaller than with H parallel to the ab plane. Since it should be harder to move flux perpendicular to the ab plane, we expect that the pinning potential for H parallel to the ab plane will be larger than for H perpendicular to the ab plane (as found in YBCO crystals). Therefore, in our experiment on aligned powder, the relaxation, and our value for U,, might be affected by a small misalignment of the crystallites with respect to the applied field, i.e., we are measuring the pinning potential averaged over the two directions. At 5 K, the measured pinning potential U. for H parallel to the ab plane is equal to 13 meV (see fig. 7 ). The pinning potential for PrThCuO for H parallel to the c direction should be smaller than 13 meV. In a separate experiment [ 121, we measured the pinning potential for an electron-doped single crystal superconductor, and obtained a value of 6 meV for U. at 5 K for H parallel to the c direction. These results suggest that the pinning in the electron-doped superconductor is rather weak compared to the YBCO. Weak pinning might also be inferred from the large Meissner signal as compared to the shielding signal, shown in fig. 1 for the unaligned powder. Palstra et al. [ 13 ] argued that the different pinning potentials among materials come from the difference in anisotropy. The least anisotropic material has the largest pinning potential. Since YBCO is the least anisotropic among the known
high-l, superconductors, this material should have the largest pinning potential. The electron-doped superconductors are more anisotropic [ 41 than YBCO and thus their smaller pinning potential is consistent with the suggestion of Palstra et al. [ 131. If one examines the resistivity data of the elcctrondoped superconductor more carefully, the resistive broadening is accompanied by a parallel shift in field. Recently. Suzuki and Hikita [ 141 have tried to fit resistivity data from a thin-film electron-doped superconductor in a magnetic field by a thermally activated flux creep model. They use a model in which the pinning potential decreases with increasing temperature. a result in disagreement with our results from magnetization data (fig. 7 ). Although both cxpcriments show that flux creep is important in the electron-doped superconductors, a detailed understanding of the phenomena is clearly lacking and will require further work.
the u/) plane is of order
_ 8 mcV at 2 K. C‘ompared
Scrgent, R. Chevrel and h,l.
and _\.I’. Maloremoff.
[ 51Y. Dalichaouch.
B.W. Lee. (‘.L. Seaman. J.T. Markcrt
M.B. Maple. Phys. Re\. Lctt. 64
We have prepared high-quality PrThCuO superconducting crystallites and successfully aligned them along the basal plane in the magnetic field. The irreversibility line is determined by measuring the field-cooled and zero-field-cooled magnetization. A power law similar to that found in YBCO is found to fit the irreversibility line. We observed logarithmic flux-creep rates in the critical-state regime to increase progressively as Tis lowered down to 2 K. and the effective pinning potential L;, to decrease strong11 with decreasing 7: We also find that C;, is not strong11 dependent on M. The value of CT0for parallel to
( 1990) 23.1.
Physica C‘ I65  Y. Ycshurun
Lett. 58 (1987)
of High (World
M. Sucnaga. .A.R. Moodcnbaugh
Phqs. Rev. 840 (1989)  1.4. Campbell, 64 (1990)
) p. 7 I.
 Y. Xu.
and DC). Welch.
L. FI-uchter and R. C‘abanel. Phys. Rc\. Lctt.
P.W. .Anderson and Y.B. Kim. Rev. Mod. Phqs. 36
M.P.A. Fisher. submitted to Phys. Kc\.
B Rapid. I?] J.L. Pcng. Z.Y. LI and R.L. Greene. unpublished. I3lT.T.M.
and J.V. I%:asfcrak. Phys Kc\.
Suzuki and M. Hiklta.
Phy5. Kc\. B4 I
( I990 I OShh.