Magnetic Properties of Pd–Fe Alloys

Abstract

Results of high‐field susceptibility measurements and nuclear spin‐lattice relaxation times on Pd–Fe alloys are presented. The susceptibility measurements encompassing the concentration range 3 ppm to 100% Fe were performed at 4.2°K in magnetic fields up to 150 kG. This range of impurity concentration, c, significantly extends previous data particularly in the low‐concentration region.^1,2 The interpretation of magnetic susceptibility for dilute magnetic alloys is difficult since one cannot simply separate the measured susceptibility into conduction‐electron and localized‐moment contributions. However, by measuring the high‐field susceptibility χ_HF, when the impurity spins are fully polarized, one obtains the susceptibility of the conduction electrons alone. (We do not consider the contributions of the orbital susceptibility or smaller terms.) From these measurements one can obtain the concentration dependence of impurity enhancement on the susceptibility.^3,4 However, these effects should be small under the present experimental conditions. In the ferromagnetic state ${\text{(c≳0.1\%,\hspace{0.17em}T≃4}}^{\circ }\mathrm{K})$ , where the up and down spin‐bands are split, the concentration dependence of the susceptibility can give information on the density of states. The gross features of the high‐field susceptibility data are as follows: In the concentration range up to about 0.5% Fe in Pd, the susceptibility is large and almost constant. For concentrations higher than 0.5%, χ_HF falls sharply. One might expect a peak in χ_HF at a concentration for which the splitting of the bands places the Fermi level near the conjectured peak in the density of states.^5,6 The splitting of the bands is given by E_s≃2c J S/[1 − N (0) V]. If we use values for J, S, and 1/[1 − N (0) V] of 0.15 eV, 3/2, and 10, respectively,^3,4,7 the concentration necessary to obtain a splitting of 80°K is approximately 0.2%. The rapid decrease in χ_HF above 0.5% we attribute to further splitting of the bands away from the supposed peak to regions in which the density of states is small compared with that of pure Pd. It should be noted that for no impurity concentration did χ_HF at 4.2°K exceed the value of χ for pure Pd at the 80°K maximum. Further discussion of the pure Pd problem will be discussed elsewhere.⁸ Preliminary results for measurements of the spin‐lattice relaxation time for ⁵⁷Fe in the magnetically ordered state of samples containing 0.5, 1.5, and 5.0 at. % iron show a large increase in T₁ as the Fe concentration is increased with T₁ values at 1.5°K (in zero magnetic field), ranging from about 150 μsec for the 0.5% sample up to about 20 msec for the 5 at. % sample. The echo decay is not a simple exponential and the range of values quoted correspond to large separations of the exciting pulses. The value of T₁ at each concentration increases markedly with applied field and decreasing temperature, e.g., T₁ is larger by a factor of two in the 1.5% sample in an applied field of 3 kOe. The very rapid increase with c of T₁ may be related to the rapid decrease in χ_HF observed for Fe concentrations greater than 0.5%. A similar behavior of T₁ for Pd is also observed with increasing iron concentration. Details of these investigations will be published elsewhere.