EuzCuo4. An anisotropic Van Vleck paramagnet

Magnetic susceptibility measurements have shown anisotropic Van Vleck paramagnetism in Eu~Cu04 single crystals. This behavior is associated with the singlet ground state ( Fo) of Eu'+ ions, and the measured anisotropy is related to a crystal-field splitting of the excited multiplets ( F&). From the experimental data at low temperatures (T ~ 50 K) a crystal-field parameter A2(r ) = — 93(5) cm ' and a spin-orbit coupling constant /=303(15) cm ' have been estimated. The temperature dependence of the magnetic susceptibility is predicted in terms of the Boltzmann population of the excited multiplets, and a comparison with experimental data up to 350 K is made. The possibility of a magnetic contribution arising from the Cu ions is discussed in connection with some discrepancies observed between the experimental and calculated magnetic susceptibilities.


D. C. Vier and S. Schultz
University of California, San Diego, California 92093 (Received 27 June 1988) Magnetic susceptibility measurements have shown anisotropic Van Vleck paramagnetism in Eu~Cu04 single crystals. This behavior is associated with the singlet ground state ( Fo) of Eu'+ ions, and the measured anisotropy is related to a crystal-field splitting of the excited multiplets ( F&).
The temperature dependence of the magnetic susceptibility is predicted in terms of the Boltzmann population of the excited multiplets, and a comparison with experimental data up to 350 K is made. The possibility of a magnetic contribution arising from the Cu ions is discussed in connection with some discrepancies observed between the experimental and calculated magnetic susceptibilities.
The discovery of high-T, superconductivity in a series of copper oxides, e.g. , (LaSr)zCu04, YBazCu307, Bi2(Ca, Sr)3Cuz09, TlzCazCu30, o, etc. , ' has led to a large amount of experimental and theoretical work in these and other related compounds. All these materials commonly share layered structures with almost square planar Cu-0 arrangement. Some of them are, as mentioned, high-T, superconductors, while other related compounds have been found to order antiferromagnetically, ' such as LazCuO& and YBa2Cu306. Another series of compounds, the (R)2Cu04 family with R=Pr, Nd, Sm, Eu, Gd, form also in a layered tetragonal structure related to the orthorhombic structure of La2Cu04 but with a somewhat larger lattice spacing in the planes. These materials are not superconducting or even metallic, and for this reason it is important to study in detail their magnetic properties in order to obtain more experimental evidence on the interplay of magnetic ordering and superconductivity in these layered copper oxides.
Single crystals of Eu2Cu04 were grown from a PbObased Aux. The crystal structure is tetragonal with lattice constants a = 3.910(1)A and c = 11. 925(3) A.
The magnetic susceptibility was determined from the magnetization measured in fields up to 5 T, and the results are shown in Fig. 1. The temperature dependence of the magnetic susceptibihty corresponds to a Van Vleck paramagnet, as expected for Eu + ions having a singlet ground state (4f, Fo). Our results show a dependence similar to that previously found in polycrystalline material. At the lowest temperatures only the ground state is thermally populated and the susceptibility becomes ternperature independent. The magnetic moment observed is induced by the external magnetic field through the admixture of the excited levels into the ground state. The magnetic susceptibility for such a case is given by S=pii(L+2S) EE+aV2020(L)+. . .
where the ellipsis represents fourthand sixth-order terms, Oz(L) is a Stevens operator, V2 -= Az(r ), and a=2/45 is the Stevens multiplicative factor for the F term. This Hamiltonian splits the F, triplet into a singlet (A, ) and a doublet (E). The Fz multiplet is split into three singlets (A "B"and Bz) and one doublet (E), although two of the singlets (8, and 82) are accidentally degenerate when we limit Eq.
(2) to second-order operators. The different levels are shown in Fig. 2. Although fourthand sixth-order operators are symmetry allowed, we have not included them into our calculation because (i) the crystal-field splitting of the F, multiplet is determined in first-order perturbation theory only by the second-order term of Eq (2); and (ii) even though the neglected higher order terms contribute to the splitting of the F2 levels, we are only interested in their admixture into the ground state and this can be well approximated by using the average energy separation that results from the spin-orbit interaction. In such a case the expressions for the low-temperature limit of the susceptibility reduce to cube of oxygen ions whose contribution to the crystal field is estimated to be A z(r ) =+43 cm ', assuming a shielding factor" (1o z) =0.20. The order of magnitude is correct, but the opposite sign is obtained. Only minor changes are observed if semiempirical parameters" are considered for the Eu +-O interaction. This result is not unexpected, since second-order crystal-field parameters usually have significant contributions from distant ions and also from dipolar or quadrupolar polarization of the ions of the lattice. In fact, we have observed that including neighbors up to a distance of S A reverses the sign of the calculated parameter. At higher temperatures the excited levels become thermally populated and, in the absence of a crystal field, the susceptibility is given by Van Vleck's formula ence of crystal-field effects, result in the following expressions: C", =C,(1+,V', Zk, T) and C, = C, ( 1 --, ' V2 Iktt T) for the first excited multiplet, and similar expressions can be derived for the other excited levels. The Boltzmann factors should also be modified to include the splitting-of the multiplets.
The results of this calculation are shown in Fig. 2 as a continuous line. It should be mentioned that this estimate does not significantly depend on the neglected fourthand sixth-order terms in Eq. (2) The differences between the experimental and calculated values are of the order of 2X10 " emu/mol Eu at room temperature, and are beyond both the experimental resolution and the estimated uncertainties of the calculated curves. The temperature dependence of the magnetic susceptibility is dominated by the energy differences between the ground state and the excited states of the Eu ions. Unfortunately, there are no optical data available to check our estimates for these energies, derived under the assumption that the contribution from Cu planes are as small as in the metallic compound La2Cu04. However, it is interesting to notice that the value estimated for the spin-orbit parameter /=303 cm ' is smaller than the values observed in other compounds, ' i.e. , the measured average low-temperature susceptibility is larger in this material. This fact makes likely the existence of larger magnetic contributions from Cu planes for Eu2CuQ4.
For example, if we arbitrarily assume that their contribution is temperature independent, isotropic, and equivalent to 3 X 10 emu/mol Eu, and subtract it from the measured susceptibilities, then the estimated spin-orbit value would be /=315 cm ' and the crystal-field parameter Vz=--100 cm '. These values imply energies for the first excited levels of 275 cm ' and 335 cm ', respectively. These values are larger than those previously derived and could result in a smaller decrease of the susceptibilities with temperature.
In particular these values would make a better fit of the data.
In conclusion, we have found that Eu2CuO4 is a Van Vleck paramagnet whose magnetic anisotropy can be related to the crystal-field splitting of the excited multiplets. We have also found indications that the Cu-O planes have a non-negligible contribution to the measured magnetic susceptibility. If that is indeed the case, some kind of magnetic order must be present because no significant paramagnetic Curie-like behavior is observed at low temperatures. This observation is consistent with the Mossbauer results' which indicate correlated spin behavior up to temperatures above 400 K, and also with electron spin resonance experiments' that show an antiferromagnetic-like low field resonance line below -200 K. The existence of magnetic ordering at these temperatures is not unexpected since the same order of magnitude coupling is seen here' between Cu spins as in LazCu04 which orders antiferromagnetically at 220 K.