Logo: to the web site of Uppsala University

The low-frequency characteristics of planar Hall effect bridge sensors are investigated as function of the sensor bias current and the applied magnetic field. The noise spectra reveal a Johnson-like spectrum at high frequencies, and a 1/f-like excess noise spectrum at lower frequencies, with a knee frequency of around 400 Hz. The 1/f-like excess noise can be described by the phenomenological Hooge equation with a Hooge parameter of g_H=0.016. The detectivity is shown to depend on the total length, width and thickness of the bridge branches. Increasing the total length by a factor of 10 improves the detectivity by a factor of 10^1/2. Moreover, the detectivity is shown to depend on the amplitude of the applied magnetic field, revealing a magnetic origin to part of the 1/f noise.

The applicability of miniaturized magnetic field sensors are being explored in several fields of magnetic field detection, due to their integratability, low mass, and potentially low cost. In this respect, different thin-film technologies, especially those employing magnetoresistance, show great potential, being compatible with micro- and nanotechnology batch processing. For low-frequency magnetic field detection, sensors based on the planar Hall effect, especially planar Hall effect bridge (PHEB) sensors, show promising performance given their inherent low-field linearity, limited hysteresis and moderate noise figure. In this work, the applicability of such PHEB sensors to different areas is investigated. An analytical model was constructed, to estimate the performance of an arbitrary PHEB in terms of e.g. sensitivity and detectivity. The model incorporates a number of approximations and, to validate the results, modelled data is compared to measurements on actual PHEBs. It is concluded that the model slightly underestimated the detectivity, especially at low frequencies and when demagnetizing effects becomes apparent. The model is also sensitive to fabrication process induced variations of the material parameters of the sensors. Nevertheless, accounting for these discrepancies, the modelled data is typically within 10% from the experimental data and the model can be used to estimate the performance of a particular PHEB design. The model is also used to establish a design process for optimizing a PHEB to a particular set of requirements on the bandwidth, detectivity, compliance voltage and amplified signal-to-noise ratio. By applying this design process, the size, sensitivity, resistance, bias current and power consumption of the PHEB can be calculated. The model shows that PHEBs are applicable to several different science areas including archaeological surveying, satellite attitude determination, scientific space missions, and magnetic bead detection in lab-on-a-chip applications.

Submicron sized Magnetic tunnel junctions (MTJs) are most often fabricated by time-consuming and expensive e-beam lithography. From a research and development perspective, a short lead time is one of the major concerns. Here, a rapid process scheme for fabrication of micrometer size MTJs with focused ion beam processes is presented. The magnetic properties of the fabricated junctions is investigated in terms of magnetic domain structure, tunnelling magnetoresistance (TMR) and coercivity, with extra attention to the effect of Ga implantation from the ion beam. In particular, the effect of the implantation on the minimum junction size and the magnetization of the sensing layer are studied. In the latter case, magnetic force microscopy and micromagnetic simulations, with the Object Oriented Micromagnetic Framework (OOMMF), are used to study the magnetization reversal. The fabricated junctions show considerable coercivity both along their hard and easy axes. Interestingly, the sensing layer exhibit two remanent states: one with a single and one with a double domain. The hard axis TMR loop has kinks at about ±20 mT which is attributed to a non-uniform lateral coercivity, where the rim of the junctions, which is subjected to Ga implantation from the flank of the ion beam, is more coercive than the unirradiated centre. The width of the coercive rim is estimated to 160 nm from the hard axis TMR loop. The easy axis TMR loop shows more coercivity than an unirradiated junction and, this too, is found to stem from the coercive rim, as seen from the simulations. It is concluded that the process scheme has three major advantages. Firstly, it has a high lateral and depth resolution – the depth resolution is enhanced by end point detection – and is capable of making junctions of sizes down towards the limit set by the width of the irradiated rim. Secondly, the most delicate process steps are preformed in unbroken vacuum enabling the use of materials prone to forming oxides in the MTJ film stack. Thirdly, the scheme is both uncomplicated and quick and makes it possible to go from design to characterization in the order of hours.

A Co₆₀Fe₂₀B₂₀-based tunneling magnetoresistance multilayer stack with an MgO barrier has been exposed to 30 keV Ga ions at doses corresponding to ion etching and metal deposition in a focused ion beam (FIB) instrument, to study the applicability of these processes to magnetic tunnel junction (MTJ) fabrication. MTJs were fabricated and irradiated to investigate how the exposures affected their coercivity and magnetoresistance. Elemental depth profiles, acquired using electron spectroscopy for chemical analysis, showed that Ga gathered in and around the two Co₆₀Fe₂₀B₂₀ layers. Correlated with the results of the magnetic measurements, this Ga presence was found to cause a reduction of magnetoresistance and an increase in coercivity. Quantitatively, a dose of 10¹⁴ Ga⁺cm^-2 reduced the magnetoresistance by 60%, whereas a dose of 10¹⁵ Ga⁺cm^-2 reduced the magnetoresistance by 67% and also increased the coercivity by 2 mT and changed the dipole coupling between the sensing and the pinning layers by 1.6 mT. The latter was attributed to an imbalance in the synthetic antiferromagnetic structure, where the stack's Ru spacer served as an implantation barrier. The magnetoresistance was lost at a dose of 10¹⁶ Ga⁺cm^-2. Annealing reduced the content of Ga around the magnetic layers but also caused diffusion of Cu from one of the layers in the stack. Apart from the observation and explanation of implantation damages in the multilayer, this work concludes on the applicability of FIB processes for prototyping of MTJs.

Spin-dependent tunnelling devices, e. g. magnetic random access memories and highly sensitive tunnelling magnetoresistance (TMR) sensors, often consist of a large number of magnetic tunnel junctions (MTJs) of uniform quality over the whole device. The uniformity and yield of the fabrication of such a device are therefore very important. A major source of yield loss is the short-circuiting of junctions by redeposition of etch residues. This can be prevented by terminating of the etch in the typically 1 nm thick tunnelling barrier. Here, electron spectroscopy for chemical analysis for monitoring the etching semi-continuously is proposed. The fabrication scheme employs Ar ion milling for etching the MTJs, and photoelectron spectroscopy for analysing the composition of the etched surface in situ. Junctions etched either to or through the barrier were used for this. The quality of the etch stop was investigated using transmission electron microscopy (TEM), and it was confirmed that the etch could be stopped in the MgO barrier. The TEM imaging also showed clear signs of redeposition. Such redeposition was attributed to being partly caused by the reduction of the TMR ratio of the junctions etched through the barrier, which was only 15% as compared with 150% for junctions etched to the barrier. Also, the latter junctions exhibited 2.7 times less noise in the low-frequency regime, resulting in a 27 times improvement of the signal-to-noise ratio with the etch stop. The barrier also proved effective in protecting the bottom contact from oxidation during the capping and contacting of the junctions.

To meet the increasing demand for miniaturized space instruments, efforts have been made to miniaturize traditional magnetometers, e. g. fluxgate and spin-exchange relaxation-free magnetometers. These have, for different reasons, turned out to be difficult. New technologies are needed, and promising in this respect are tunnelling magnetoresistive (TMR) magnetometers, which are based on thin film technology. However, all new space devices first have to be qualified, particularly in terms of radiation resistance. A study on TMR magnetometers' vulnerability to radiation is crucial, considering the fact that they employ a dielectric barrier, which can be susceptible to charge trapping from ionizing radiation. Here, a TMR-based magnetometer, called the spin-dependent tunnelling magnetometer (SDTM), is presented. A magnetometer chip consisting of three Wheatstone bridges, with an angular pitch of 120 degrees, was fabricated using microstructure technology. Each branch of the Wheatstone bridges consists of eight pairs of magnetic tunnel junctions (MTJs) connected in series. Two such chips are used to measure the three-dimensional magnetic field vector. To investigate the SDTM's resistance to radiation, one branch of a Wheatstone bridge was irradiated with gamma rays from a Co-60 source with a dose rate of 10.9 rad min(-1) to a total dose of 100 krad. The TMR of the branch was monitored in situ, and the easy axis TMR loop and low-frequency noise characteristics of a single MTJ were acquired before and after irradiation with the total dose. It was concluded that radiation did not influence the MTJs in any noticeable way in terms of the TMR ratio, coercivity, magnetostatic coupling or low-frequency noise.