Electron paramagnetic resonance, electrical resistivity, and magnetization studies in the high Tc superconductors EuBa2Cu3O9−xandEuBa2 (Cu1−yMy)3O9−x (M = Cr, Mn, Fe, Co, Ni, ORZn)

Abstract We have measured electron paramagnetic resonance (EPR), electrical resistivity, and dc magnetic susceptibility from 2 K – 300 K for the high T c oxide superconductor EuBa 2 Cu 3 O 9−x , either undoped or doped with 3d ions (Cr, Mn, Fe, Ni, Co, or Zn), which presumably substitute at the Cu sites. We have observed an EPR line at low temperatures (T ≤ 40 K), which exhibits an increase in intensity and decrease in field for resonance as the temperature is lowered. The EPR linewidth is also temperature dependent and exhibits a minimum at about 15 K. In some of the samples another EPR signal is observed over the entire temperature range studied, with properties that depend on sample preparation conditions. This signal is likely to be due to small amounts of an additional phase. The behavior and origin of these EPR signals are discussed. The variation of T c with 3d ion concentration over the range (1 – 8%) is also presented.

The discovery by Bedaotz and Miiller I of superconductivity above 30 K in La-Ba-Cu-O and the more recent discovery of superco-4_~ctivity above 90K in a series of ox~. 4..SUlmeondoctors RBa2Cu3Og.x, where RfY t-,~ or rare canh ioos, .~ has r~uRed in an unprecedented number of studies on these superconducting oxides, k appears that the R ions have a negligible interaction with the superconducting electrons in these compounds, since replacement by most R's, including those with local moments, does not change the high critical temperature, Tc.4,5 It was suggested that the high temperature superconductivity is associated with a 2-D structure in which two layers of CuO 2 sandwich one CuO 3 chain, with two Ba 2+ per unit cell. 6 Apparently the nlkAliqe earth ions are only weakly coupled to the 2-D network of Cu-O sheets, since the substitution of Ba by Sr and/or Ca also has little effect on Tc.7.8 Because the Cu ions seem to be the essential component of the conduction path in these oxides, it is important to determine whether substituting some of the Ca ions with other 3d ions, particularly those with with a local moment, results in a significant perturbation of their normal and superconducting properties. An additional reason to substitute local moment ions is that they may provide an atomic probe when observed via Electron Paramagnetic Resonance (EPR).
To investigate these ideas we have studied EuBa2(CUl.yMy)3Og. x with M = Cr, Mn, Fe, Co, Ni or Zn and 0.01 < y < 0.15. EPR, ac electrical resistivity, and do magnetic susceptibility measurements were made over the temperature range 2 K -300 K. Extensive studies of RBa2Cu309. x for R = Y, Pr, Nd, Gd, Ho, Er, or Yb, as well as fractional substivation of the R ions with Gd or Er, have also been made, and will be reported elsewhere. 9 Sample preparation techniques have been reported previously. 4,5 X-ray powder diffraction on these compounds confLrmed them as > 97% singie-phaso materials with the desired orthorhombically distorted , oxygen deficient, perovskite-like structure.
Electrical resistivity meas~ts were made at I00 Hz using a four probe configuration. For T > T¢, the resistivity, p, showed a metallic behavior for low dopant concentrations, y < 0.01. At  Fig (2). Although we have also added Mn as a dopant, we have not included these data in Fig.(2) because we Imv¢ found that there is a lack of reproducibility of the suppressi~ ofT c for the samples prepared to date.
The dc magnetic susceptibility, Z, was mcasm~i with a SQUID magnetometer "10 from 6 K to 300 K in magnetic fields from 50 G to 10 kG. In Fig.(3) we present X vs T for EaB~(CUl.y7my) 3 09. x with y = 0.01 and 0.05. Similar data were obtained for the other 3d dopants. We note the large suppression of the diamagnetism below T¢ with increasing Zn dopant concentration.
The EPR measurements were mado on powdered (grain size ~ 10 gm) using an X-band superheterodyne specuvmeter over the mnperamre range 1.5 K -300 IC Below T c the microwave impedance of the sample is m'ongly dependeat on field and temperauae, giving rise to a baseline signal that exhibits significant variations and hysteresis. For most samples we az¢ able to discern a Lorentzian EPR signal superimposed on the field dependent baseline variations. (Some caution mast be exercised in this process because the baseline variations at low field, < 500 G, can often appear to mimic a Lorentzian type of signaL) Below ~ 40 K, an anti-symmetric (A/B ---1.0) Lorentzian EPR line is observed for pure or doped samples, including those whose T c is < 2 K (e.g., sample "E" in Fig.(la)). We will refer to this EPR signal as the low temperature (I.T) line. The properties of the LT line, as the temperature is decreased from 40 K, are: (a) the intensity continuously increases; Co) the EPR linewidth, AI-Ipp, decreases until ~ 15 K and then sharply increases, as shown m Fig.(4a); (c) the resonance field, HR, at 9.2 GHz remains constant down to ~ 15 K and then sharply decreases, as shown in Fig.(4b).
In addition to the EPR line just described, we often observed a partially resolved spectrum similar to that found for Cu 2+ ions in an orthorhombic host, with gl -2.27, g2" 2.12, and g3 ~ 2.05. When present, this spectrum is observable over the full temperature range (2 K -300 K). We will refer to this as me high temperature (HT) spectrum. The intensity of the HT spectrum varied greatly from sample to sample, even though the samples were chat'aetefiz~] by X-ray diffraction as having only the EuBa2Cu309. x phase present. In most samples the HT spectrum is either relatively weak, or is not observed, so it is unlikely that it is characteristic of the EuBa2Cu309. x phase. Also, the HT spectrum should not be a consequence of the presence of small amounts of Eu2BACuO 5 (green phase), since we found that this phase by itself exhibits no significant EPR signal. This is in contrast to the case for R = Y, where we have found that Y2BaCuO5 gives such a large HT speclzum intensity that the presence of even a frucdon of I% of this phase in YBa2Cu309. x would give a HI" spectrum comparable in intensity to that foend in our samples with R = Eu. At present, the origin of _this sample dependent HT spectrum is unknown, but it may arise from another phase in amounts small enough (< 2%) to escape detection by X-ray diffi'action. Additional investigations are underway and will be re~ separately. We did not observe any additional EPR signals in samples doped with Cr, Fe, Co, Ni, or Zn. While we have observed an additional EPR line in some Mn doped samples, we find that its presence is sample-dependent. With EPR there is always the concern that an observed signal derives from a spurious phase, and this remains a possibility until the location of the Mn dopant is determined.
None of the EPR signals observed to date have any features which correlate--with the superconducting lxantltion, More explicitly, the HT spectrum, when present, is unaHect~ upon c~ssing T c. The LT line, as previously dtamutsed, is pt~eat in all samples, with a qualitatively similar ~ dependence for H R and AHpp, even for those samples which are non-supercondu~ing (i.e., EuBa2[Cu0.92Zn0.0813Og. x, as well as PrBa2Cn3Og. x and several of the superconducting samples that were subsequently Ar annealed to partially deplete the oxygen). In some of the samples that weredoped with Mn, we also observe an EPR line attributed to the Mn z+ ion and again find no anomaly as we pass through T c.
In general, one may expect to t'md certain features in the temperature dependence of the linewidth and field for resonance of local moments placed in type II superconductors. These features typically arise from : (a) inhomogeneities of the magnetic field  (a) The peak.to-peak linewidth, AHpp, and Co) the field for resonance, H R at a frequency of 9.2 GHz.
when He1 < H R < Hc2, (b) changes in the effects of spin-orbit interactions upon spin relaxation,, or^(c) the diminution of the electronic susceptibility below Tc .L*,~k However, none of these effects would appear to be able to .account for the ~mre.. dependences of the signals we nave ooserven, msteao, it ,s likmy that the large shifts of the field for resonance are due to an into.hal rearrangement of an underlying spin system. It is also most likely that the changes in linewidth are not due to real clumges in spin relaxation, but are rather a reflection of the large shifts in H R, which may be inhomogeneously felt throughout the sample. Meaanrcments of the spin iattic¢ relaxation time, T 1, are underway to clarify this queuion. We .have ~a~ac ~.ts in pro y~es. s to determine whether there is a tmnpetatme.~cence or me susceptibility which can be correlated w,th the temperature dependence of H R and ~ observed for the LT fine. We suggest that an examination of neutron scattering data might also be insighcm.
We conclude with a comment concerning the resistivity data shown in Fig.(2). The data indicate that there is no correlation between the f~e ion magnetic moment of the 34:! impurity ions and their effectiveness in depressing T c. This is in contrast to standard BCS superconductors where magnetic impurities are particularly effective in suppressing T c. Most striking is that the non-magnedc Zn, which one would expect to be Zn 2+, is even more effective in degessing T c than the magnetic 3d ions. One possible explanation is that the diffmamt 3d ions replace the Cu ~. f.erentially on m'tes.of diffmw, nt importance for the supercorglucuvaty.. Anotl~r is. mat substitution of the different 3d tons may change etectron aenmty or other variables to such an extent that these changes become more important in determining T c than the spin-dependent scattering of conduction electrons by these ions. Acknowledgement -The authors wish to acknowledge, with thanks, the technical assistance provided by Mr. Roger Isaacson, University of California, San Diego with the EPR spectrometer.