ISSN:
1089-7550
Source:
AIP Digital Archive
Topics:
Physics
Notes:
We have performed inelastic neutron scattering study of the cerium γ↔α valence transition (T0(approximately-equal-to)150 K) in polycrystalline Ce0.74Th0.26 at the Intense Pulsed Neutron Source (IPNS) of Argonne National Laboratory. An incident neutron energy of 300 meV was used to measure the excitation energy spectra of Ce0.74Th0.26 at 100, 140, 155, and 200 K by chopper spectrometers. By comparing the neutron total scattering of Ce0.74Th0.26 with that of La0.74Th0.26, an isostructural nonmagnetic alloy measured under idential experimental conditions, the magnetic scattering function of Ce0.74Th0.26 averaged over all q in the Brillouin zone, Savemag (Q,E), is derived. We find that the magnetic response of Ce0.74Th0.26 at all temperatures is predominately due to moments of 4f character. The obtained magnetic scattering function in this temperature region consists of a broad quasielastic peak, which is well fitted by a spin relaxational spectral function, namely, a Lorentzian centered at 0 energy. The peak shifts to higher energies and broadens as the temperature is lowered. As the temperature decreases across the transition temperature, the magnetic intensity drops sharply, accompanied by an abrupt broadening of the linewidth corresponding to a spin fluctuation energy much higher than the thermal energy. Γ, the Lorentzian HWHM's, were found to be 16, 25, 63, and 110 meV at T=200, 155, 140, and 100 K, respectively. Within experimental precision, we find no evidence of additional inelastic peaks due to crystal-field excitations. These results are in good agreement with those from an earlier neutron experiment1 using thermal-energy neutrons in which the measured spectra were limited to about 70 meV. The static single-site susceptibility obtained by a Kramers–Kronig analysis agrees well with the bulk susceptibility.1 The 4f occupation per Ce atom, deduced by summing theneutron data from −100 to 230 meV, is about 0.6 and 0.4 for the γ and α phase, respectively. These values are considerably smaller than those obtained from photoemission2 and other measurements.1 This indicates that the neutron experiment did not extend to high enough energies to account for all the intensity. In the high-temperature γ phase the relaxational model represents a reasonable approximation to the spin dynamics. By integrating the Lorentzian scattering function obtained from the fits of the neutron data over an energy interval of −0.1 to 2 eV, we obtained an f occupancy close to unity in the γ phase. For the α phase there is currently no analytical expression of the magnetic scattering function from first-principle calculations. Model calculations3,4 of the ground state (T=0) for a single f impurity in the metal, on the other hand, predicts a magnetic response function having a thresholdlike rise at a finite energy, followed by a long tail extending to high energies. In order to be able to compare with the theory, we have also undertaken the measurements of the magnetic scattering function of Ce0.74Th0.26 at 10 K with an incident neutron energy of 1.2 eV. We find that Savemag shows an inelastic peak at about 138 meV and a tail at higher energies. The line shape of the measured spectrum agrees qualitatively with the single-impurity theory3,5,6 and the result of a recent polarized neutron study7 of α-Ce. By summing the measured magnetic intenstiy up to 500 meV, we estimated a 4f occupancy of 0.76 per Ce atom. The static susceptibility at 10 K and the electronic specific heat coefficient obtained from the neutron data and the theory3,8 agree well with the values obtained from bulk measurements.1,9,10 The above results are also in fair agreement with the electronic spectroscopy data.2 Since the neutron measurements were made using polycrystalline samples, we are unable to detect, if any, coherence effects6,11 due to interaction of the f electrons in the lattice. A detailed report on the inelastic neutron scattering investigations is presented elsewhere.12
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1063/1.338923
Permalink
