We studied the magnetoresistance of normal metal (NM)/ferromagnet (FM) bilayers in the linear and nonlinear (current-dependent) regimes and compared it with the amplitude of the spin-orbit torques and thermally induced electric fields. Our experiments reveal that the magnetoresistance of the heavy NM/Co bilayers (NM = Ta, W, Pt) is phenomenologically similar to the spin Hall magnetoresistance (SMR) of YIG/Pt, but has a much larger anisotropy, of the order of 0.5%, which increases with the atomic number of the NM. This SMR-like behavior is absent in light NM/Co bilayers (NM = Ti, Cu), which present the standard AMR expected of polycrystalline FM layers. In the Ta, W, Pt/Co bilayers we find an additional magnetoresistance, directly proportional to the current and to the transverse component of the magnetization. This so-called unidirectional SMR, of the order of 0.005%, is largest in W and correlates with the amplitude of the antidamping spin-orbit torque. The unidirectional SMR is below the accuracy of our measurements in YIG/Pt.Comment: 14 pages, 5 figure

The reliability of microelectronic devices is largely dependent on electronic packaging, which includes heat removal. The appropriate packaging design therefore necessitates precise knowledge of the relevant material properties, including thermal resistance and thermal conductivity. Thin materials and high conductivity layers make their thermal characterization challenging. A steady state measurement technique is presented and evaluated with the purpose to characterize samples with a thermal resistance below 100 mm(2) K/W. It is based on the heat flow meter bar approach made up by two copper blocks and relies exclusively on temperature measurements from thermocouples. The importance of thermocouple calibration is emphasized in order to obtain accurate temperature readings. An in depth error analysis, based on Gaussian error propagation, is carried out. An error sensitivity analysis highlights the importance of the precise knowledge of the thermal interface materials required for the measurements. Reference measurements on Mo samples reveal a measurement uncertainty in the range of 5% and most accurate measurements are obtained at high heat fluxes. Measurement techniques for homogeneous bulk samples, layered materials, and protruding cavity samples are discussed. Ultimately, a comprehensive overview of a steady state thermal characterization technique is provided, evaluating the accuracy of sample measurements with thermal resistances well below state of the art setups. Accurate characterization of materials used in heat removal applications, such as electronic packaging, will enable more efficient designs and ultimately contribute to energy savings.