PUB - Publikationen an der Universität Bielefeld

The emerging field of spin caloritronics, which focuses on the interaction between spin and heat transport in materials, has gained strong interest in recent years. Primarily the prospect that thermal spin transport enables new mechanisms for thermal-to-electric energy conversion, makes the investigation of spin caloritronic effects particularly interesting for energy conversion applications, e.g., waste heat recovery in modern electronics. A promising approach to attain this objective is the tunnel magneto-Seebeck (TMS) effect.
The TMS effect was predicted from ab initio theory by Czerner et al. and experimentally discovered by Walter et al. and Liebing et al. in Co-Fe-B/MgO/Co-Fe-B magnetic tunnel junctions (MTJs) in the year 2011. MTJs are nanopatterned spintronic devices that consist of a nanometer thick insulating tunnel barrier (e.g., MgO) sandwiched between two ferromagnetic electrodes (e.g., Co-Fe-B). The fundamental mechanism behind the TMS effect is the altering of the Seebeck coefficients in an MTJ when the relative alignment of the magnetizations of the electrodes is reversed. If a temperature gradient is applied to the MTJ, this altering of the Seebeck coefficients can be observed as a change of the Seebeck voltage. No additional power source that provides a bias voltage is needed to obtain this signal. Thus, the readout of the MTJs by the TMS effect allows the use of waste heat generated in electronic devices to operate, e.g., memories or sensors, without an additional power source. This makes the TMS effect particularly interesting for reducing the power consumption of future electronic devices.
One important benefit of the TMS effect, compared to other spin caloritronic effects, like the Spin-Seebeck effect, is the fact that it is observed in MTJs, which are already implemented in up-to-date electronic devices, such as hard discs or random access memories. Today, the readout of these de- vices is performed via the tunnel magnetoresistance effect (TMR) under an externally applied bias voltage. The TMR effect describes the change of the resistance of an MTJ depending on the relative magnetization orientation of its ferromagnetic electrodes. It was first discoverd by Julliere using Fe/Ge-O/Co tunnel junctions in 1975. Today, the insulating layer mostly consists of amorphous Al2O3 or crystalline MgO in combination with a wide range of electrode materials, such as ferromagnetic alloys, Heusler compounds, superconductors, and even antiferromagnets. Intense research has led to stable logic devices with effect ratios of several hundred percent. The versatility and robust nature of MTJs has made them the backbone of modern spintronics. Carefully designed MTJs enable effects like spin-transfer torque or spin-orbit torque switching. These discoveries opened a broad field for new devices, such as the magnetoresistive random-access-memory (MRAM). This variety of new effects in combination with the adaptable material parameters of MTJs makes them particularly interesting for the use in spin caloritronic devices.
Although these facts already reveal the large potential of MTJs, the relatively new spin caloritronic counterpart to the TMR effect, the TMS effect, has gained relatively little attention when considering the development of new devices. The reasons for this lack of interest are most likely due to the low Seebeck voltages of only a few microvolt, and the low effect ratios of only a few percent, being much less than for the established TMR effect. However, so far only Co-Fe or Co-Fe-B based MTJs with MgO or Al2O3 barriers have been investigated. These material combinations have proven to be ideal for high TMR effect ratios. Nonetheless, this does not necessarily imply their suitability for the generation of high TMS effects, due to the different transport mechanisms. Accordingly, it is not surprising that recent ab initio calculations by Geisler and Kratzer predict higher TMS effects for MTJs that contain half-metallic Heusler compound electrodes.
The scope of this work is, to obtain a simplified, yet accurate model for the prediction of high TMS effects to enable a fast material screening. Furthermore, this study aims at an experimental evidence that MTJ devices with tailored density of states (DOS), in particular with half-metallic Heusler compound electrodes, can significantly improve the TMS effect.
In order to obtain information on the thermoelectric transport in the MTJs, the Landauer-Büttiker formalism is applied and the influence of different DOSs on the size of the Seebeck coefficients and the TMS effect is evaluated. To prove the feasibility of the proposed material parameters, the TMS effect is investigated using a number of different methods and on various types of MTJs. In this context, firstly, a new method for shifting the chemical potentials through the DOS of the electrodes by using an external bias voltage and simultaneously determining the Seebeck effect in the MTJs is introduced. Secondly, it is investigated whether the nearly half-metallic Heusler compounds Co2FeAl and Co2FeSi provide the desired high TMS ratios.
This thesis is organized as follows: In Chapter 2 the theoretical foundation for the effects that are treated in the experimental sections are presented. This includes a detailed introduction to the TMR and Seebeck effects. With this knowledge the TMS effect is derived, and a model, based on the DOSs and the transmissions of the MTJs in the Landauer-Büttiker formalism, is elaborated. This includes a method to obtain the TMS effect size from the shape of the DOS. Additionally, the nonlinearized Landauer-Büttiker formalism is used to investigate the TMS effect under an applied bias voltage. Furthermore, the nonlinearized formalism can be applied to study the influence of the temperature dependent shifts of the chemical potentials in the electrodes on the Seebeck coefficients.
In Chapter 3 the methods and techniques necessary for the experimental observation of the TMS effect are introduced. This chapter concentrates on the optical and electronic setup, and links the theory to the applied techniques. It also explains the sample design and the methods necessary for the preparation and characterization of the samples. Furthermore, COMSOL simulations are proposed for determining the temperature profile in the MTJs.
Chapter 4 presents the results obtained with the methods in Chapter 3. It begins with basic TMS experiments on Co-Fe-B/MgO/Co-Fe-B MTJs. This is followed by a discussion of the temperature profile in the MTJs for different heating scenarios. After this more general insight into the TMS effect, the analysis focuses more detailed on the connection of the TMS effect to the DOS of the MTJs. First, the behavior of the TMS effect under an applied bias voltage is revealed. This section of Chapter 4 also compares the experimental determined dependence of the Seebeck effect on an external bias voltage to the predicted results from the model in Chapter 2. Second, the TMS effect in Heusler based MTJs is investigated and compared to Co-Fe-B based MTJs. This includes a connection to the model, that has been derived from the Landauer-Büttiker formalism and the DOS in Chapter 2.
Finally, Chapter 5 summarizes the theoretical and experimental findings. It also gives an overview of ongoing experiments, and an outlook on new ideas for future investigations.