Effect Y substitution on the microstructure, transport and magnetic proprieties of Bi₂Sr₂Ca₁Cu₂O_8+δ superconducting ceramics

In high Tc superconducting (HTSC) materials the main parameter for practical applications is, beside the critical temperature of transition Tc, the critical current density. This parameter, which limits the transport of current carried by the material, is controlled by the capability of pinning vortices under applied magnetic field. This is why activation energy, or pinning potential, was studied earlier in Bi based high Tc superconducting materials (HTSC) in both single crystals [1-3] and polycrystalline [4, 5] samples. The pinning force balances the Lorentz force. Moving of vortices causes thermal dissipation which may destroy superconductivity. Near Tc, the resistive transition corresponds to the “flux flow” dissipative regime [6, 7] where the Lorentz force dominates [1]. Thermal dissipation is possible also when the pinning force dominates and is caused by “flux creep” [8] which is more evident in magnetic measurements [4, 5]. On the other hand, doping by cation has been intensively studied in order to understand the mechanism of high Tc superconductivity and to improve the properties of the material. Among various atomic elements, the rare earth ones have been used for their capability, when substituting at the Ca site, to reduce the carrier concentration in the CuO₂ planes of the Bi-2212 phase [9-39]. The case of doping by yttrium has been a subject of several studies [10-23, 26, 33, 37, 39]. The pinning properties, expressed in the irreversibility line, have been at this time studied in Y substituted Bi-2212 crystals [33]. To our knowledge, no study has been made before about the effect of Y doping on the activation energy of Bi-2212 ceramic samples as it has been made for undoped Bi-2212 crystal [2], and in more recent works on doped and undoped Bi based polycrystalline samples [41-45]. This effect is presented for the case of Y substitution on the Ca site at the rate varying up to 25 % (x = 0.25), under applied magnetic field having a constant amplitude equal to 0 T, 1 T, 2 T and 3 T.

Bismuth, strontium, calcium and copper nitrates were dissolved in distilled water in a stoichiometric ratio. Y powder dissolved in nitric acid was added to the Bi solution. A solution of triammonium citrate with a concentration of 2 mol/kg, was prepared by reacting citric acid with ammonia solution. The organic gels were made with acrylamide CH₂=CHCONH₂ and N, N-methylene diacrylamide. CH₂=CHCONHCH₂NHCOCH=CH₂ was added in order to complex the Bi, Ca, Sr and Cu cations. The obtained solution was mixed, stirred and heated at 80 °C to 90 °C with a magnetic stirrer on a hot plate. Few drops of a solution of AIBN in acetone were added to accelerate the formation of the gel. The later was transformed to Bi-2212 powder after heating at 400 °C for 2 h and 6 h at 700 °C. The obtained powder was ground in an agate mortar and calcined during 12 h at 860 °C with a rate of 5 °C/min. After calcinations, the samples were ground again, and then pelletized under a pressure of 300 MPa. The pellets were then sintered during 12 h at 860 °C with a rate of 5 °C/min.

Samples have been denoted as Y0, Y0025, Y010 and Y025, where the number following Y indicates the values of the rate x of Y. The formation of the Bi-2212 phase was analyzed by XRD with CuKα radiation in the range 2θ = 10° to 55°. Mi-crostructural and surface morphology of the samples were investigated using scanning electron microscopy (SEM, JEOL 6390-LV). The resistivity was measured for all samples in the temperature range of 10 K to 125 K using a standard DC fourprobe technique with a constant current of 5 mA. The current and voltage contacts were made from a silver paste. The magnetic field was applied parallel to the surface of the pellet and perpendicular to the current flow direction.

The XRD patterns of the samples are shown in Fig. 1. The diffraction results show the presence of Bi-2212 phase as a major component. The structure of Bi-2201 was also detected. The lattice parameter c determined using the software Janna 2006 from the (0 01) peaks of the XRD data is given in Table 1. It is observed that the lattice parameter c decreases significantly with increasing Y content. The Y³⁺ cation has a smaller ionic radius than Ca²⁺. This explains the decrease of c which can be also due to the increase of oxygen content in the system as it has been suggested for Y substitution on Ca site.

Microstructural features were studied using SEM (Fig. 2). The SEM images of all samples were taken at a magnification of 3000. The microstructure of Y0 sample is remarkably different from that of Y0025, Y010, and Y025 samples. One can observe that doping with yttrium decreases the size of the grains and deteriorates the surface morphology.

The resistivities recorded under different applied magnetic fields are reported for all the samples in Fig. 3. These records are normalized to the value ρ_n of the resistivity at 100 K where the samples are supposed to be in their normal state and near the beginning of the superconducting transition. The curves in Fig. 3 show that the applied magnetic field increases the broadening of the transition whose shape depends on the rate of yttrium. Two temperatures are indicated by arrows in the figure to characterize the broadening: the critical temperature T_c0 indicates the end of the zero resistance state and, in the transition, a threshold temperature T_t where the different curves are no longer separated. The values of these temperatures (T_c0max and T_c0min have been measured without and with applied magnetic field) together with the critical transition onset temperature T_{c onset}.

These results show that doping with Y increases the width of the transition and lowers the T_c0. Applying magnetic field increases the width of the transition in the doped samples more than in the undoped ones. The enhancement in the undoped sample for 3 Tesla is about 7.3 K and goes from 16 K to 7.6 K as the content of Y increases. This enhancement is roughly equal to the decrease of T_c0 except for the undoped sample where the decrease of T_c0 is higher. Moreover, the variation of T_c0 with magnetic field is higher in the undoped samples than in the doped ones. The shape of the transition varies also from the undoped samples to the doped ones and with the content of Y. In the doped samples, the transition presents a kind of knee much more evident in samples Y0025 and Y025 at about 30 K. The knee is also present for 0 T at about 50 K for Y0025 and 45 K for Y025. In samples Y0 and Y010, the knee is not present at 0 T, but appears with magnetic field at about 60 K in Y010.

In sample Y025, under magnetic field, the resistivity varies slowly between T_t and 30 K (position of the knee). This kind of shape (the knee) has been seen in measurements of resistivity of Bi-2212 crystals with magnetic field applied perpendicular to ab planes when the configuration of the applied current and field has not resulted in any Lorentz force [46-48]. In the last case, the resistivity shows an important enhancement before the sharp decrease of the superconducting transition. Such behavior, as shown by several authors [47-52], indicates a transition from a gas of 2D pancake vortices to a liquid of 3D vortices, and is above the irreversibility line in the (H, T) plane.

In the ceramic samples studied here, the grains have plate-like shape where the largest dimension is along ab planes. When a magnetic field is applied perpendicular to these plates, their individual demagnetizing field coefficient Ng along the same direction increases the field by a factor 1/(1 - N_g) (Ng ≤ 1) [53]. Curves in Fig. 3 suggest that in samples Y0025 and Y025 more grains have their ab planes oriented perpendicular to the applied field. Moreover, N_g gives a much higher effective field which destroys the conduction of the Josephson junctions, forming the weak links between the grains, but this behavior is supposed to occur at low field.

Fig. 4 shows the Arrhenius plot of the natural logarithm of the normalized resistivity of the samples (ln(ρ/ρ₁₀₀) versus 1/T). This figure gives a better view of how the yttrium doping changes the effect of magnetic field at low temperatures. The horizontal scale is the same in Fig. 4a and Fig. 4c but different in Fig. 4b and Fig. 4d, where there is a multiplication of about five. Without yttrium (sample Y0 in Fig. 4a) the curves are well separated and the absolute value of their slope at low temperatures decreases with the strength of the magnetic field. With yttrium, the magnetic field moves the curves toward much lower temperatures (higher values of 1/T) but their separation is smaller and they are almost identical as for sample Y0025 (Fig. 4b).

These curves allow the calculation of the activation energy as suggested by the model of “flux creep” [7]. In this case, the resistivity obeys the law: ρ(H,T)=ρ0exp(−U0(H)/kBT)$$\rho \left( {H, T} \right) = {\rho _0}\exp \left( { - {U_0}\left( H \right)/{k_B}T} \right)$$(1)

Fig. 1

Fig. 2

Fig. 3

where U₀ is the activation energy, k_B the Boltzman constant and ρ₀ is a pre-factor to be determined. The slope of ln(ρ/ρ₀) in the Arrhenius plot gives the value of U₀/k_B.

The problem is to limit the range of temperature where this slope may be calculated by a fit to a linear regression using the least square method. The values of ρ/ρ₀ used by other authors [2] were much lower than 0.1 reaching about 10^-6. The value of 0.1 has currently been used to estimate the irreversibility field H_irr. For the estimation of U₀, the curves were taken for values of ρ/ρ₀ between 0.1 and a value corresponding to the deviation from linearity. The obtained values verify the criterion of U₀ >k_BT which limits the ‘‘flux creep” regime and where the “flux flow” regime begins.

The variation of U0 versus the applied magnetic field H is revealed for each sample in Fig. 5. This figure shows that U0 decreases when yttrium is introduced. For H ≠ 0, this decrease goes through a maximum with the rate x of doping. There is also a decrease of U₀ with H except for sample Y010 where it remains quasi constant.

The obtained values of U₀ lie between 0.003 eV and 0.060 eV corresponding to about 30 to 750 K. These values are the same as those obtained in Bi-2212 crystal by Shi et al. [2]. The last authors used a modified law based on a field dependent activation energy U₀ = A(H)(1-T/Tc)^m suggested by Yeshurun et al. [54]. Shi et al. [2] explained in this way the curvature of the Arrhenius plot in their results. This curvature, which occurs in most of the cited works, is evident in the results presented in Fig. 4. Nearly the same values of U₀ were also obtained by Sharma et al. [45] in the measurements up to 14 T on Bi-2212 polycrystalline samples, prepared by sol-gel method and annealed between 760 °C and 840 °C, where they showed an influence of annealing temperature on U₀. With lead containing (Bi, Pb)2212 samples doped by a rare earth on Sr site the obtained values are much higher, 0.075 eV to 0.308 eV with an applied field of 0.28 T for Gd doping [40] and 0.077 eV to 0.47 eV with an applied field of 0.32 T and 0.63 T for Tb doping [41].

Fig. 4

Fig. 5

The effect of yttrium is also evident in the plot of the natural logarithm of normalized resistivity (ln(ρ/ρ₁₀₀) versus the normalized temperature (Uo/T) shown in Fig. 6. These curves have been plotted for the same values of ρ/ρ₁₀₀ as those used for the estimation of U₀.

Fig. 6 shows that for H ≠ 0, the different curves become less separated as the content of yttrium increases. In the undoped sample Y0 (Fig. 6a), the curves are well separated. Moreover, the curve obtained without applied magnetic field (H = 0) is obviously separated from the curves with applied field in samples Y0 and Y0025.

Fig. 6

The results shown in Fig. 7 give information on the pre-factor corresponding to the intercept of curves in the limit 1/T → 0. This intercept has been calculated for the curves shown in Fig. 6 for the Y doped Bi-2212 samples. The behavior observed in Fig. 6 is confirmed by the intercept, which has a quasi-constant value when H ≠ 0 in the samples doped with Y (Y0025, Y010 and Y025). Moreover, the intercept is very different when H = 0 in the samples Y0 and Y0025, which confirms the behavior observed in Fig. 6a and Fig. 6b.

Karppinen et al. [55] showed that doping with yttrium at Ca site gives, for the concentrations used in our samples (remark: x corresponds to 1 - x in that work), an overdoped Bi-2212 phase which is nearly optimally doped for Y010 and Y025.

Fig. 7

Increasing x decreases the c axis parameter of the unit cell but the number of holes increases in both CuO₂ plane and BiO layer [55]. The variation of the c axis parameter changes the coupling between the coupling of CuO2 planes by changing their distance and thus their Josephson coupling. The electromagnetic coupling of the pancake vortices is also influenced by these variations. The variation of the density of holes influences the Tc. Thus, 3D to 2D transition of the vortices is influenced by these variations. When the c axis parameter changes, the transition field, where the 3D vortices lattice becomes a 2D vortices gas, changes also. A similar effect is obtained when Tc changes, but for the temperature of the 3D to 2D transition.

On the other hand, the texture and the granular quality of the sample may influence also the 3D to 2D transition. The angle between H and the ab planes is not constant in granular samples and its distribution is more random when the texture is not good. There is also a distribution of the shape and size of the grains and consequently of their demagnetizing factor Ng. These characteristics combined with a possible lack of homogeneity in yttrium distribution may give grains, in which the vortices form a 3D liquid lattice and other a 2D vortices gas. Fig. 3b and Fig. 3d suggest a higher number of grains with 2D vortices gas in samples Y0025 and Y025 than in the other samples. This may explain the lower activation energy found in these samples (Fig. 5). Fig. 6 shows that when there is yttrium and the field is applied, the curves overlap. Increasing the concentration of yttrium the overlap is more and more obvious. This effect is a consequence of several changes induced by doping with yttrium, such as density of holes in CuO₂ planes, length of c axis parameter and texture of the ceramic sample. More investigations on how these changes can influence the 3D to 2D transition of the vortex lattice are needed.

Yttrium lowers the activation energy of the Bi-2212 phase when it substitutes at the Ca site, for concentrations less or equal to 25 %, at the applied field up to 3 T. The shape of various resistive transitions indicates that this effect is mostly a consequence of the granular nature of prepared samples rather than the variations of the density of holes in the CuO₂ planes or the c axis parameter. When the texture is poor, an important part of the grains may be oriented in the magnetic field so as to promote a 3D to 2D transition of the vortex lattice. The effect is more evident in the higher part (higher values of temperature near T_{c, onset}) of the superconductive transition than in its lower part (lower values of temperature near T_c0) where the resistivity decrease obeys the usual exponential law corresponding to a flux creep of the vortex lattice.