[1] A new empirical model of the plasmapause location has been developed using density data from the plasma wave receiver onboard the CRRES spacecraft for nearly 1000 orbits. The ''plasmapause'' is identified here as the innermost sharp gradient in density (change of a factor of 5 in <0.5 L). Such a sharp gradient was observed on 73% of the CRRES inbound and outbound orbits that returned data. The plasmapause location is expressed as a linear function of Kp (previous 12 hour maximum) and local time. The model gives the linear best fit location of the plasmapause as well as the standard deviations of the model parameters. We found a slight noon-midnight asymmetry with the plasmapause located on average an L shell farther from the Earth at midnight than in the noon sector. This is in the opposite sense to the noon-midnight asymmetry found previously. Significant variability (with standard deviations up to +/À 1 L shell) in the plasmapause location is seen and suggests that though the mean plasmapause is roughly circular, the instantaneous plasmapause has significant time variable localized structure at all local times but most especially in the duskside sector.

Abstract.A general problem when fitting EXAFS data is determining whether particular parameters are statistically significant. The F-test is an excellent way of determining relevancy in EXAFS because it only relies on the ratio of the fit residual of two possible models, and therefore the data errors approximately cancel. Although this test is widely used in crystallography (there, it is often called a "Hamilton test") and has been properly applied to EXAFS data in the past, it is very rarely applied in EXAFS analysis. We have implemented a variation of the F-test adapted for EXAFS data analysis in the RSXAP analysis package, and demonstrate its applicability with a few examples, including determining whether a particular scattering shell is warranted, and differentiating between two possible species or two possible structures in a given shell.

We present a detailed extended x-ray absorption fine structure ͑EXAFS͒ analysis of the thermoelectric clathrates Eu 8 Ga 16 Ge 30 and Sr 8 Ga 16 Ge 30 , both of which have an unusually low thermal conductivity attributed to a "rattler" motion of the filler atoms Eu and Sr. The EXAFS results show that the Ga/ Ge lattice is quite stiff with a high correlated Debye temperature ϳ400 K. Eu is on-center in the site 1 cage but off-center ͑0.445± 0.020 Å͒ in the large cage called the Eu2 site. The results for Sr are similar, but ϳ75% are off-center 0.40± 0.05 Å and ϳ25% are on-center in the Sr2 site. Both results are in reasonable agreement with diffraction results. The temperature dependence of the nearest neighbor pair distribution widths yield low Einstein temperatures ͑80± 10 and 100± 10 K for Eu1 and Sr1, respectively, and 95± 10 and 125± 10 K for the shortest Eu2-Ga/ Ge and Sr2-Ga/ Ge pairs͒. In contrast, the more distant Eu2 / Sr2-Ga/ Ge pair distributions within the Eu2 / Sr2 cage are strongly disordered even at low T, indicating considerable local disorder. This indicates that the off-center Eu or Sr atom is bonded to the side of the site 2 cage. This has two important implications for the thermal conductivity: it increases the coupling between the "rattler" vibrations and the lattice phonons and it introduces a symmetry-breaking large mass defect.

We present extensive x-ray absorption fine structure measurements on La(1-x)Ca(x)MnO(3) as a function of the B field (to 11 T) and Ca concentration, chi(21%-45%). These results reveal local structure changes (associated with polaron formation) that depend only on the magnetization for a given sample, irrespective of whether the magnetization is achieved through a decrease in temperature or an applied magnetic field. Furthermore, the relationship between local structure and magnetization depends on the hole doping. A model is proposed in which a filamentary magnetization initially develops via the aggregation of pairs of Mn atoms involving a hole and an electron site. These pairs have little distortion and it is likely that they form at temperature T(*) above T(c).

In the filled Ga/ Ge clathrate, Eu and Sr are off center in site 2 but Ba is on center. All three filler atoms ͑Ba, Eu, Sr͒ have low-temperature Einstein modes; yet only for the Eu and Sr systems is there a large dip in the thermal conductivity, attributed to the Einstein modes. No dip is observed for Ba. Here we argue that it is the off-center displacement that is crucial for understanding this unexplained difference in behavior. It enhances the coupling between the "rattler" motion and the lattice phonons for the Eu and Sr systems, and turns on/off another scattering mechanism (for 1 K Ͻ T Ͻ 20 K) produced by the presence/absence of off-center sites. The random occupation of different off-center sites produces a high density of symmetry-breaking defects which scatter phonons. It may also be important for improving our understanding of other glassy systems.

We present detailed local structure measurements ͑using the extended x-ray absorption fine structure tech-nique͒ for the colossal magnetoresistive material La 1−x Ca x MnO 3 ͑0.21Ͻ x Ͻ 0.45͒ as a function of temperature and magnetic field. The local distortions of the Mn-O bonds are parameterized using , the width of the Mn-O pair-distribution function ͑PDF͒. After subtracting thermal phonon contributions, we show that the contributions to 2 from polaron and Jahn-Teller ͑JT͒ distortions, JT/polaron 2 , are a universal function of the magnetization, independent of how the magnetization is achieved via changes in temperature or magnetic field. However this universal behavior is only observed for B fieldsՆ 2 T, likely as a result of domain canting in low B fields. The resulting curve is well described by two straight lines with significantly different slopes. These regimes represent two distinctly differ distortions of the oxygen octahedra about the Mn. For low magnetizations up to ϳ65% of the theoretical maximum magnetization, M T , the slope is low and the distortion removed as the sample becomes magnetized is small-we argue this arises from polarons which have a low distortion around two ͑or possibly three͒ Mn sites. At high magnetizations large distortions per Mn site are removed as these sites become magnetized. The data are also analyzed in terms of a two Mn-O peak distribution using experimental standards for Mn-O. The results agree well with recent neutron PDF results but not with some earlier results. We discuss the limitations of assuming a two peak distribution in view of the two distortions needed to describe the Mn-O distortions as a function of T and B for B Ն 2 T. It is likely that there is a distribution of longer bonds. Finally we show that with increasing B field, the Mn-Mn peak also has a small B-field-induced change-a measure at the unit cell level of magnetostriction but find that there is no observable B-field-induced change in the Mn-La/Ca pair distribution for fields up to ϳ10 T.

A temperature-dependent EXAFS investigation of La 1−x Ca x MnO 3 is presented for the concentration range that spans the ferromagnetic-insulator (FMI) to ferromagnetic-metal (FMM) transition region, x = 0.16, 0.18, 0.20, and 0.22; the titrated hole concentrations are slightly higher y = 0.2, 0.22, 0.24, and 0.25 respectively. For this range of Ca concentrations the samples are insulating for x = 0.16-0.2 and show a metal/insulator transition for x = 0.22. All samples are ferromagnetic although the saturation magnetization for the 16% Ca sample is only ∼ 70% of the expected value at 0.4T. This raises a question as to the nature of the ferromagnetic (FM) coupling mechanism in such insulating samples. We find that the FMI samples have similar correlations between changes in the local Mn-O distortions and the magnetization as observed previously for the colossal magnetoresistance (CMR) samples (0.2 ≤x≤ 0.5) -except that the FMI samples never become fully magnetized. The data show that there are at least two distinct types of distortions.The initial distortions removed as the insulating sample becomes magnetized are small and provides direct evidence that roughly 50% of the Mn sites (associated with the hole charge carriers) have a small distortion/site and are magnetized first. The large Mn-O distortions that remain at low T are attributed to a small fraction (< 30%) of fully Jahn-Teller-distorted Mn sites that are either unmagnetized or antiferromagnetically ordered. Thus the insulating samples are very similar to the behavior of the CMR samples up to the point at which the M/I transition occurs for the CMR materials. The lack of metallic conductivity for x≤ 0.2, when 50% or more of the sample is magnetic, implies that there must be preferred magnetized Mn sites (that involve holes) and that such sites do not percolate at these concentrations.

X-ray Absorption Fine Structure (XAFS) measurements of the colossal magnetoresistance (CMR) sample La 0.79 Ca 0.21 MnO 3 at high fields indicate a decrease in the width parameter of the pair distribution function, σ, as the applied magnetic field is increased for T near Tc. The change in σ2 from the disordered polaron state varies approximately exponentially with magnetization irrespective of whether the sample magnetization was achieved through a change in temperature or the application of an external magnetic field. This suggests a more universal relationship between local structure and the sample magnetization than was previously indicated.