Sébastien Saez

For several years, considerable attention has been devoted to advancing high-sensitivity magnetic sensors to enhance measurement capabilities. A recent trend in this pursuit involves leveraging magnonic devices for sensor implementation. While only a limited number of studies have produced and characterized such devices, initial performance outcomes appear promising. These devices harness the spin wave propagation dependency on external magnetic fields. Typically, yttrium iron garnet (YIG) serves as the magnetic material. The implementation of the sensor system involves several elements. A permanent magnetic field must saturate the material. An ac current passing through an exciting antenna generates spin waves by locally causing a precession of the magnetic moment. A receiving antenna is used to capture the propagated spin wave. A representation of the observed external magnetic field is obtained by demodulating the received signal. Optimizing this system necessitates the development of a comprehensive model. However, existing physical models primarily focus on spin wave propagation and generation. Our proposed approach adopts a holistic, systemic perspective, treating the sensor element as a four-port coupler. This model underscores the intrinsic symmetries of the sensing element and the non-reciprocal nature of spin wave propagation. Notably, our work presents a unified model that accounts for spin wave propagation within microstructured YIG, as well as spin wave reflections along YIG edges, facilitating the assessment of each parameter’s impact on the overall system response. This work presents our latest advancements in the field. Several systems and configurations have been tested, enabling the characterization of various parameters of the couplers. These results were used to explore new ways to optimize the magnonic sensor.

In a ferromagnetic material exposed to a strong magnetic field, the spin wave describes the propagation of the phase shift of precessing spins at the same frequency over the time. In recent years, this subject has been of a great interest in the field of instrumentation and data transmission. Although several studies have shown a strong potential of magnonics for the design of a magnetic field sensor, few studies have been interested in the realization of such sensors and are limited to the measurement of scattering parameters. Here, we describe the development of a magnetic field sensor, using the variations in the propagation of the spin wave induced by the external magnetic field variation. We detailed the full implementation of the system, based on magnonics devices along with its associated electronic. The proposed implementation is mainly based on a quadrature I/Q demodulation to produce an output signal whose variations reflect the variations of the external applied magnetic field. The sensing element is designed from a yttrium iron garnet (YIG) film deposited on the gadolinium gallium garnet (GGG) substrate. It is placed on an optimized microstrip antenna transducer made of two copper lines, respectively, to excite and measure the spin waves. With the help of characterization measurements (S-parameters) carried out with a VNA, field sensitivities, expressed in V/T, could be estimated from S-parameters. Two kinds of magnonic devices have been implemented: one using spin-wave reflections at each YIG edge and other one using spacial filtering by YIG material microstructuring. Their performances regarding the magnetic sensitivity are compared. Finally, this first implementation of such a sensor, based on YIG magnonic material, exhibits a magnetic noise level of about 100 pT/ $\surd $ Hz at 1 kHz and a usable bandpass of 100 kHz. This is quite promising for a new sensor technology, leaving room for further developments.

Since few decades, research in the magnonic area, which studies the phenomenon of spin waves propagation in magnetic materials, has grown simultaneously with the emergence of the so-called spintronic and nano-structured magnetic materials. Recent works exposed the interest of such devices for designing magnetic field sensors, considering the dependence of the transmission parameter S21 regarding the external applied magnetic field, for excitation frequency of several GHz. In this paper, we have investigated the ability of spin wave propagation phenomenon to be used as a magnetometer by characterizing its expected performances, particularly the output sensitivity, expressed in unit of V/T. The sensing element is designed from a YIG film (10 mm × 5 mm × 19.4 μm) deposited on GGG substrate. It is placed on a microstrip antenna transducer made of two copper lines, respectively to excite and measure the spin waves. The excitation is provided by a sine electromagnetic wave at a fixed frequency f0=1.146 GHz. The induced spin waves propagate along the main length direction of the YIG material and induce a voltage signal at the output antenna at the carrier frequency f0. When the external DC magnetic field varies, it yields a modulated signal. The dedicated and reduced electronic conditioning circuitry, based on a I/Q quadrature demodulator, allows to extract the field variations information with an output sensitivity of about 20 V/T. Considering that the dominating noise sources are those due to the electronic conditioning circuitry, a sensor equivalent magnetic noise level of several nT/√Hz in white noise region was obtained.

In this paper, we investigate the effect of temperature variations on the off-diagonal giant magneto-impedance (GMI) response. This paper is based on a theoretical model describing the GMI-based sensor response. The modeling has been extended to include the resistivity dependence versus the temperature. It allows to calculate both temperature sensitivity and intrinsic magnetic field sensitivity, whose ratio leads to the determination of the magnetic field offset drift versus the temperature. Thanks to an original conditioning circuitry based on a double ac current excitation instead of the classical one, we achieved an improvement of the long-term stability and minimized the sensor temperature dependence. Experimentally obtained results have been compared with the expected offset drift for the classical GMI, and with the expected fluctuation reduction while adopting this technique, showing a relatively good agreement with the proposed modeling.

The equivalent magnetic noise spectral densities of off-diagonal giant magnetoimpedance (GMI)-based magnetometers exhibit significant low-frequency excess noise, proportional to 1/f noise. As it represents a serious limitation to the ultimate sensing performances of high sensitivity magnetometers, possible sources of this 1/f noise are under investigation. Low-frequency magnetization fluctuations have been proposed as the noise source in the case of classical GMI-based sensors. Here, we apply this model to off-diagonal GMI-based magnetometers. This requires the inclusion of magnetization fluctuation noise sources, in addition to white noise sources from electronic conditioning in the GMI effect equations. A pessimistic scenario is presented, predicting the upper limit of low-frequency excess noise from material characteristics. The equivalent magnetic noise level is then computed from the sensitivity of each term of the sensing element impedance matrix to the magnetization angle at the static working point (for both axial and circumferential static magnetic field) and to conditioning circuitry. Based on this, it appears that magnetization fluctuations similarly affect all modes of operation of the two-port network sensing element, inducing identical impedance fluctuations. It also appears that this noise depends only upon the static equilibrium condition. This condition is governed by the effective anisotropy of the magnetic wire and by both axial and circumferential static components of the working point.

The lowest measurable magnetic field of a high sensitive magnetometer based on giant magnetoimpedance (GMI) effect will be determined by its noise level. Numerous previous works have showed that the equivalent magnetic noise level of GMI magnetometer presents a 1/f behavior at low frequencies and a white noise. Currently, performances of the magnetometer are still limited by the electronic conditioning as well as the intrinsic equivalent magnetic noise of the GMI sensor. To improve sensor performances, particularly at low frequency, we have investigated two focuses on research: 1) reduce the electronic noise of the conditioning electronic by increasing the voltage sensitivity, in units of V/T and 2) reduce the intrinsic equivalent low frequency sensor noise. As previously reported, the equivalent magnetic noise at 1 Hz can be reduced by increasing the excitation current in addition to a dc bias current. By working on these bias conditions, on the given samples, we almost 1 pT/√Hz in white noise region and 15 pT/√Hz at 1 Hz.

We have investigated the possible impact of low frequency magnetization fluctuations on the equivalent magnetic excess noise of GMI sensors, which we have recently shown to exhibit 1/f noise. This noise component is not associated with the detailed measuring setup nor with the conditioning electronic noise sources, suggesting that it is intrinsic to the sensing element. Various intrinsic GMI noise sources might be able to explain this observation; these include magnetic domain wall motion, hysteresis loop losses, etc. Since GMI elements are excited by a high frequency current, it has been assumed that these low frequency (lf) intrinsic noise sources cannot interact with the carrier and the sensed signal. We recall that the GMI effect is based on an impedance variation, which is governed by magnetization angle variations. These are modulated by the sensed signal and the lf magnetization noise, which then appear as sidebands around the carrier frequency. Applying the fluctuation-dissipation theorem to the GMI model, we have related both the signal and noise to the magnetic susceptibility spectrum and thus quantified the equivalent magnetic noise of GMI sensors at lf. We then present a preliminary comparison to our previous experimental results.

Magnetic sensors based on GMI devices are the subject of intensive research, as they appear promising for magnetometry applications. Performances of GMI magnetometers are often limited by the noise of the electronic setup. Thus, the present challenge is to increase the GMI device sensitivity (expressed in V/T) in order to decrease the equivalent magnetic noise of the system. In our previous work, we showed that the use of a pick-up coil in an off-diagonal configuration improves the magnetic sensor sensitivity and offers a promising approach for developing an inexpensive magnetometer with sub-pT/Hz equivalent magnetic noise levels. Ideally, the use of a coil increases the sensitivity linearly as a function of the number of turns. However, this effect is reduced by the parasitic capacitance of the coil. This affects the device sensitivity, noise level and system performance. The parasitic capacitance can degrade all of these, but also induces a resonance effect, which can help to optimize magnetometer sensitivity, and thus, its noise level. We analyze the effects of the parasitic capacitance on the system (sensitivity and noise) and propose optimization routes. We have obtained sensor sensitivity as high as 700 fT/Hz.

Two versions of an operational giant magnetoimpedance (GMI)-based magnetometer, one of which is capable of working in an unshielded geomagnetic field, has been design, built and benchmarked against a commercial fluxgate magnetometer. The sensing element consists of a detection coil, strongly coupled to the GMI wire, in a two-port configuration. Its optimization is based on the surface impedance tensor of the sensing element, in which we include parasitic capacitance and geometric parameters of the coil. Conditions for optimizing the electronic noise level are deduced from the model recently published in (IEEE Sensors Journal, vol. 11, no. 6, pp. 1317-1324, 2011.). The optimized sensor and its conditioning electronics are operated in a field-locked loop, using an external coil or the detection coil for feedback. Experimental noise performance correlates well with that expected from the sensitivity and noise models and are comparable to those of a commercial Flux-Gate magnetometer, offering a low-cost solution for high sensitivity magnetometry. However, the proposed GMI magnetometers still exhibit significant 1/f noise.