Untersuchung des Spin-Nernst-Effekts: absolute Momente und Dynamik
Zusammenfassung der Projektergebnisse
The intention of the project was to quantify the spin Nernst effect using micro SQUID technology. This approach was pursued for more than half a decade, however, not successfully. The failure of the approach can be attributed to various reasons. Firstly our SQUID technology never achieved the expected sensitivity which most likely is due to technological problems that still could not be identified. Nevertheless, a geometry problem could also be identified. The concept relies on the detection of the magnetic field that is created by the spin accumulation in a heavy metal. Closer investigation of the combination of the heavy metal stripe and the SQUID shows that it is virtually impossible to create a single edge (for a well-defined spin accumulation) close enough to the SQUID loop. The necessary planarization would make the spacing between the heavy metal and the SQUID too large. The field created by the spin accumulation can be seen as the field of a dipole of a few nm extension that drops equally fast with distance. The only solution to this problem would be to use technologies like chemical mechanical polishing which are not available in most labs. Still the very small effect would make a detection by this technology unlikely. Due to this draw back we pursued two other approaches to identify and measure the Spin Nernst effect. One promising approach is to use scanning SQUID microscopy (SQUID on a tip) as developed by Eli Zeldov. Its extreme sensitivity may allow for the detection. However, the availability of the technique is limited and as first experiments have shown also for these experiments the sample geometry needs to be highly optimized. Unfortunately after some first experiments the collaboration could not be continued. The one approach that got us closest to our goal (though not completely there) is the measurement of the spin Nernst torque using ferromagnetic resonance and time resolved magneto optical Kerr microscopy (TR-MOKE). We applied this technique to a number of samples and we obtained results that may be interpreted as spin Nernst effect. Nevertheless, we have observed several artifacts that surprisingly are not mentioned in literature but that can jeopardize the correct interpretation of the results. We are, however, confident that this method allows either to identify the spin Nernst effect or to set an upper limit to its magnitude that can be compared to literature. In view of the development of the state-of-the-art (only five experimental demonstrations until now, all with very small effects), we can only conclude that the effect is extremely small and with little technological relevance. In most cases even cannot be considered an artifact that needs to be taken into account in spin caloritronic investigations.
