Surface Acoustic Wave Based Magnetic Sensors

1. Introduction

Since the radar system was invented in 1922, the development of devices communicating by means of reflected power has experienceda continuously growing interest. In 1948, Harry Stockman published a paper [1] in which he laid the basis for the idea of radio frequency identification(RFID), and the first patent had been filed in 1973 by Charles Walton. After decades of research and commercialization, RFID products became a part of everyday life (e.g. logistics, access control, security). With the growing interest in remote and battery-free devices, researchers are pushing the boundaries of RFID technology to find solutions in new fields like sensing applications.

For many sensors such as those operated in remote or harsh environments, the sensitivity is not the only evaluation criteria. The lifetime, especially of the power source, and the complexity added by wiring often demand wireless and passive operation.Batteries have limited lifetime and also add to the size and mass of the sensors. Alternatively, energy harvesting oran RF-based wireless power supply can be employed [2, 3]. The former method depends on environmental conditions such as solar radiation, temperature change, chemical reagents, vibration etc., which are often not constantly or not sufficiently available.RF power sources, on the other hand, transmit power wirelesslyand with full control over amount and timing.

Passive and remote sensors utilizing SAW transpondersaredevices, which are powered by an RF source. These systems require an interrogation device that requests the sensor signal, a SAW transponder plus a sensing element and two antennas. The basic idea is that an RF signal of certain frequencies generated by the interrogator is received by the SAW transponder, which reflects back a signal modified by the sensing element. This signalcontains the environmental information in an amplitude and phase change, which is converted into the physical parameters by the interrogator. In most cases, SAW sensors are coded by having different reflector designs in order to have multiple measurement capabilities from sensors located in the same interrogation area. A great amount of research has been carried out in the past decades in this field, and, as a result, different wireless SAW sensors have been developed to measure a variety of physical and chemical parameters including temperature, stress, torque, pressure, humidity, magnetic field, chemical vapor etc. [4-9]. Several devices are already commercialized [10-12].

SAW-based magnetic sensors have, so far, not been studied in detail. Magnetic sensors are one of the most pervasive kinds of sensors for a large number of applications and are employed in different fields like automotive, biomedical or consumer electronics. Integrating magnetic sensors with SAW transponders enables remote and passive operation, thereby, opens a door for further applications. Early in 1975, a magnetically tuned SAW phase shifter was proposed by Ganguly et al [13]. A thin film of magnetostrictive material was fabricated on the delay line of a SAW device. A phase shift was observed due to the dependence of the wave propagation velocity on the external magnetic field.Recently, a new concept of amagnetic sensor based on a SAW resonator has been published [8]. A magnetostrictive material was used to fabricate the interdigital transducers (IDT) of the SAW device. The resonant frequency of the device changes with an external magnetic field. A different idea was put forth by Hauser and Steindl[14-16] combining a SAW transducer with a giant magnetoimpedance (GMI) microwire sensor. The GMI sensor has a magnetic sensitivity, at least, one order higher than contemporary giant magnetoresistance (GMR) sensors and can be used to measure very low magnetic fields such as those generated by the human heart or muscles. A GMI sensor is operated by an accurrent and the impedance changes upon changes of a magnetic field. This makes it a suitable load for a SAW transponder, which converts this impedance change into a magnitude and phase change of the reflected acoustic waves. In order to reduce the size and improve the level of integration of the senor, a new design of an integrated SAW transponder and thin film GMI sensor has been proposed and developed recently by the authors [17, 18]. The SAW transponder and GMI thin film were integrated on the same chip using standard micro-fabrication technology suitable for mass fabrication.

The ideal SAW-based magnetic sensor is small and highly integrated, inexpensive, passive, remotely controlled and have a high magnetic sensitivity together with a large linear range. With regard to these criterions, an SAW-GMI sensor is avery promising candidate.

SAW-based magnetic sensors have been studied for several years.However, this topic has yet not been comprehensively summarized, and the aim of this chapter is to provide a systematic review of the past research as well as the latest results. The performance of the devices crucially depends on different design parameters in a complex fashion. This will be shown by a detailed description and analysis for a device consisting of a SAW transponder and GMI thin film sensor.

2. SAW based passive sensors

A basic SAW device consists of an input interdigital transducer (IDT) and an output (or reflector) IDT, which are fabricated on a piezoelectric substrate. The area between the input IDT and output IDT is called the delay line. The IDT is made of two metallic, comb-like structuresarranged in an interdigital fashion, whereby the distance between two fingers of a comb defines the periodicity (p) (Fig. 1). Upon application of a voltage, charges accumulate at the fingers of the IDT depending on the capacitance of the structure.The resulting electric field generates stress in the substrate due to the piezoelectric effect.If an ac input voltage is applied, the continuously changing polarity of the charges will excite an SAW (Rayleigh wave) traveling through the substrate. At the operating (resonant) frequency of the SAW device,the value ofp equals the wavelength of the SAW, and the SAW amplitude showsa maximum value due to constructive superposition.

Figure 1.

Schematic of a SAW device.

2.1. Basics of SAW devices

SAW Resonant Frequency:The resonant motion of an acoustic resonator is caused by the coupling between the transducer (IDT) and the acoustic medium. The resonant frequency, or operating frequency, is determined by the periodicity and the acoustic wave velocity(v)

$$f=\frac{v}{p}$$E1

.

The value of vmainly depends on the substrate’s material. A typical SAW velocity for piezoelectric materials is several thousand meters per second. Due to the intrinsic anisotropy of piezoelectric materials, v is dependent on the direction of propagation. Since different acoustic modes have different wave velocities, a device can resonate at different frequencies. The SAWs are Rayleigh waves, which have a longitudinal and a vertical shear component that can couple with any media in contact with the surface. This coupling strongly affects the amplitude and velocity of the wave allowing SAW sensors to directly sense, e.g., mass loads.

Electro-Mechanical Coupling Coefficient: The electro-mechanical coupling coefficient (κ) defines the conversion efficiency of the piezoelectric material between the electrical and mechanical energies, determined by

$${\kappa }^{-2}=\frac{inputenergy}{convertedenergy}$$E2

.

A high coupling coefficient reduces the insertion loss caused by the energy conversion, which results in smaller energy consumption as well as larger effective readout distance of a SAW-based wireless sensor.

SAW Delay Line: The SAW delay line refers to the area between the input IDT and output IDT on the substrate (Fig. 1). It creates a time delay between the input signal and the output signal depending on the SAW velocity and the length of the delay line. Due to this feature, SAW devices are widely used in RF electronics. It is also used in sensing applications, where the measurand causes, e.g., a change in the SAW.

Temperature Coefficient of Delay (TCD): The TCD reflects the temperature dependence of the time delay and is connectedwith the thermal expansion coefficient (α) and the temperature coefficient of the phase velocity (TCV) by

$$TCD=\alpha -TCV$$E3

.

The temperature dependence of the time delay is the basis of SAW temperature sensors, where higher TCD values yield higher sensitivity. However, for other SAW devices, the influence of the TCD on the time delay is undesirableand has to be minimized or eliminated. For this purpose, temperature compensated cuts of the crystalline substrates are employed, where the TCD is minimized over certain temperature ranges [19-20]. Piezoelectric bi-layers are another concept that has been utilized in order to compensate the TCD in sensing applications [21, 22].

2.2. Basic design concepts of passive SAW sensors

Passive SAW sensors typically operate asresonators, delay lines or loaded transponders. In case of resonators, the reflection of the interrogation signal from the SAW device is a function of the SAW device’s resonant frequency, which depends on the measurand. In case of delay lines, the request signal is separated from the response signal by a time difference, whereby this time difference depends on the measurand. Similarly, the request signal and response signal are separated by a time difference in case of a loaded transponder. However, the time difference is constant and the measurand affects the signal amplitude. Intrinsic SAW sensors utilize a change of the substrate’s properties. For example, intrinsic temperature sensors were realized by detecting the resonant frequency or phase change of the SAW in materials with large TCD [23]. Intrinsic stress sensors utilize the length change of the delay line caused by mechanical strain applied to the substrate. The stress can be evaluated by measuring the SAW phase shift [24]. Extrinsic SAW sensors can be realized by integrating a SAW device and an additional sensing element. A common extrinsic sensor conceptutilizes selective thin films on top of the delay line leading to a change in mass by the measurand [13, 25]. This can be, for example,a thin film with high CO₂ solubility and selectivity [26]. As CO₂ dissolves into the film, the additionalmass loadcausesa detectable phase shift in the SAW. Another extrinsic concept utilizes a sensitive IDT. For example, in case of a magnetostrictive IDT, a magnetic field applied to the sensor causes a change of the resonant frequency [8]. A loaded transponder is another extrinsic design, where the output IDT is connected to a sensor, which changes the IDT’s electrical characteristics as a function of the measurand.An example for a load sensor is a pair of conducting rods placed in the earth with a certain distance from each other. As the water level changes, the resistance between the rods changes, which can be detected as a magnitude and phase change of the signal reflected from the load IDT [27]. Another example for a load is a giant magnetoimpedance sensor [16, 17]. Achange in the magnetic field yields a change in the sensor’s impedance. Consequently, changes the reflectivity of the output IDT.

Some SAW sensors, their classification and method of detection are presented in Table 1.

Sensor Type	Commercialization	Year	Intrinsic/; Extrinsic	Design	Detection Method	Access	Paper
Temperature	Yes	1990	Intrinsic	Resonator	Frequency	None	[23]
2003	Intrinsic	Delay line	Phase velocity	None	[28]
Pressure	Yes	2001	Extrinsic	Loaded Transponder	Phase	Capacitive pressure sensor	[25]
2007	Intrinsic	Delay line	Phase	None	[24]
Bio/Chem	No/Yes	2006	Extrinsic	Resonator	Frequency	Thin film	[29]
2011	Extrinsic	Delay line	Phase	Thin film	[26]
2001	Extrinsic	Loaded Transponder	Amplitude/Phase	Conducting rods	[30]
Magnetic	No	1975	Extrinsic	Delay line	Phase	Thin film	[13]
2011	Extrinsic	Resonator	Frequency	Magnetostrictive IDTs	[8]
2006/11	Extrinsic	Loaded Transponder	Amplitude	GMI wire/ thin film	[16,17]
Sound	No	2005	Extrinsic	Loaded Transponder	Phase	Capacitive pressure sensor	[31]
Torque	Yes	1996	Intrinsic	Delay line	Phase	None	[32]

Table 1

SAW-based passive sensors.

3.Passive and remote SAW-based magnetic sensors

Magnetic sensors are one of the most versatile sensors employed not only for the task of measuring magnetic fields but for a large number of different applications, thereby detecting the measurand indirectly, e.g., via a change of material parameters in construction monitoring or a change of distance in position monitoring. A passive and remote operation of magnetic sensors can be advantageous in many cases and considerably increase their applicability.

A SAW-based passive magnetic sensor can be realized either by adding an additional material layer, which is sensitive to magnetic fields, or by loading the output IDT with a magnetic sensor. In the first case, the magnetic layer changes the delay line or the resonant frequency of the SAW device. While in the second case, the sensor changes the reflection signal of the output IDT. Since SAW devices are operated by RF power, the sensor element has to work at the operation frequency of the SAW device. Among the available magnetic sensors, GMI sensors are the most suitable candidates as they have a high magnetic sensitivity as well as a high operating frequency.

3.1. MagnetostrictiveSAW devices

Magnetostriction defines the relationship between the strain and the magnetization states of a material. It is an important property of ferromagnetic materials and was first observed by James Joule in 1842 in nickel samples. For a positive/negative magnetostrictive material, an applied magnetic field causes the material to expand/shrink in the field direction. Inversely, when a stress is applied to the magnetostrictive material, its magnetic anisotropy will change accordingly.

A magnetostrictive-piezoelectric resonator consists of amorphous magnetostrictive material layers as the electrodes sandwiching a piezoelectric core (Fig. 2). An ac signal applied to the electrodes causes the quartz layer to oscillate. The resonant frequency of this oscillation depends on the thickness of the piezoelectric material, the crystal orientation, temperatureand mechanical stress, etc.

Figure 2.

Structure of a compositemagnetostrictive/piezoelectric resonator.The magnetic anisotropy is perpendicular to the external magnetic field H_dc.

When a magnetic fieldH_dc is applied, the length change induced in the magnetostrictive film exerts stress to the piezoelectric material and, consequently, shifts the resonant frequency of the device. Utilizing this concept,a magnetic sensitivity high enough to detect the terrestrialfield has been achieved [33].In a similar work, a magnetostrictive-piezoelectric tri-layer structure has been embedded in a coil. The dc magnetic field sensitivity was as high as 10^-8 T [34].

3.1.1. Passive resonator

A SAW-based, passive resonator for magnetic field detection was developed recently by Kadota et al [8]. Nickel, which is a negative magnetostrictive material, was used to fabricate the sensing IDT on a quartz substrate (Fig. 3). Upon the application of a magnetic field, stress will be induced to the substrate by the IDT change causing a change in the resonant frequency. This sensor showed a frequency change of 200 ppm for a magnetic field of 100 mTapplied perpendicularly to the direction of SAW propagation.

Figure 3.

Schematic of amagnetic sensor device using a magnetostrictive IDT on a SAW substrate.

3.1.2. Passive phase shifter

A magnetically tuned SAW phase shifter is a one-port SAW structure with a magnetic sensing functionality achieved through a delay line sensitive to magnetic fields. This idea was first introduced by Ganguly et al in 1975 [13]. In their device, the acoustic velocity is varied by an external magnetic field. This functionality is facilitated by a magnetostrictive thin film deposited on top of the delay line (Fig. 4). The propagation velocity of the SAW in the film region depends on the magnetic field. Hence, there is a correlation between the time shift of the reflected signal and the magnetic field.

Figure 4.

Schematic of a magnetically tuned SAW wave phase shifter.

Later, research efforts focused on different magnetostrictive materials and measurement methods [9, 35, 36], and a magnetic sensitivity of 10^-4%/kOe was achieved.

3.2. SAW-GMI sensors

A SAW-based, magnetic and passive sensorscomprises a two-port SAW transponder and a magnetic sensor acting as a load at the output IDT. Among the available magnetic field sensors, giant magnetoimpedance (GMI) sensors offer favorable characteristics like high sensitivity to magnetic fields and high operation frequency (compatible with SAW transponders) making them a very suitable load. SAW-GMI sensors have been fabricated by combining SAW transponders with GMI wire sensors as well as thin film GMI sensors. Both of these methods have shown a higher magnetic sensitivity than direct designs.

3.2.1. Basics of GMI sensors

The GMI effect was first observed in Co-based amorphous wires by Panina and Mohri in 1994[37] and has since attracted strong interest due to its sensitivity enabling magnetic field measurement with a nT resolution. The GMI effect isthe impedance change of an ac-powered ferromagnetic conductor upon the change of a magnetic field. The relative impedance change, also called GMI ratio, is expressedas

$$GMIRatio\left(\%\right)=100\text{\%}\times \frac{Z\left(H\right)-Z(H_0)}{Z(H_0)}orGMIRatio\left(\%\right)= 100\text{\%}\times \frac{Z\left(H\right)-Z(H_{max})}{Z(H_{max})},$$E4

whereZ(H₀) is the impedance at zero magnetic field andZ(H_max) is the impedance atsaturation field. Both definitions have particular aspects that should be considered. In case of the first expression,Z(H₀) depends on the remanent state of the magnetic material while,in the second case, Z(H_max) is not always achievable and equipment dependent.

The GMI effect is explained by classical electromagnetism.The change of the complex impedance mainly originates from the skin effect in conjunction with a change of the complex permeability.Analytically, the complex impedance (Z) of a conductor is defined by

$$Z=\frac{U_{ac}}{I_{ac}}=\frac{{\int }_L^{}\frac{1}{\sigma }{\mathit{J}}_{\mathit{z}}\left(S\right)dz}{{\iint }_q^{}{\mathit{J}}_{\mathit{z}}dq}$$E5

,

whereU_acis the applied ac voltage, I_acis the current, L is the length and σ the conductivity.Sand qrefer to the surface and the cross section of the conductor, respectively.J_zis the current density in the longitudinal direction obtained by solving Maxwell’s equations.In ferromagnetic materials, by neglecting displacement currents ($\dot{\mathit{D}}=0$), Maxwell’s equations can be written as follows:

$$\nabla \times \mathit{H}=\mathit{J}$$E6

,

$$\nabla \times \mathit{J}=-\frac{{\mu }_0}{{\rho }_f}(\dot{\mathit{H}}+\dot{\mathit{M}})$$E7

,

$$\nabla ·\left(\mathit{H}+\mathit{M}\right)=0$$E8

,

J is the current density,H is the applied magnetic field, M is the magnetization of the ferromagnetic material, ρ_fis the free charge density and μ₀is the permeability of vacuum.From Equ.(6) to (8), the expression

$${\nabla }^2\mathit{H}-\frac{{\mu }_0}{{\rho }_f}\dot{\mathit{H}}=\frac{{\mu }_0}{{\rho }_f}\dot{\mathit{M}}-\nabla \left(\nabla ·\mathit{M}\right),\mathrm{ }$$E9

can be derived. Equ. (9) can be solved using the Landau-Lifshitz equation, whichrelates MandH

$$\dot{\mathit{M}}=\mathit{\gamma }\mathit{M}\times {\mathit{H}}_{eff}-\frac{\mathit{\alpha }}{{\mathit{M}}_s}\mathit{M}\times \dot{\mathit{M}}$$E10

,

whereγ is the gyromagnetic ratio,M_sis the saturation magnetization,α is the damping parameterandH_effis the effective magnetic field expressed as [30]

$${\mathit{H}}_{eff}=\mathit{H}+{\mathit{H}}_{\mathbf{a}}+\frac{2A}{{\mu }_0{M}_s}{\nabla }^2\mathit{M}$$E11

whereH is the internal magnetic field that includes the applied field and demagnetizing field, H_ais the anisotropy field andA is the exchange stiffness constant..

By combining Equ.(5) to (11), a theoretical impedance model can be evaluated for GMI sensors with different geometries [37-40].

Although the experimentally obtained GMI effect shows a large sensitivity compared to other effects exploited for magnetic sensors, the theoretically estimatedvalues have not been achieved yet. Therefore, a lot of effort has been putinto improving the magnetic properties of GMI materials [41-44]. At the same time, GMI sensors of different structures have been developed such as glass-coated wires, thin films, multi layer thin films, meander structures, ribbons, etc. [45-47]

3.2.2. Wire GMI sensors

As the first discovered GMI sensor structures, GMI wire sensors have been extensively studied. Based on the classical electromagnetism, the theoretical model of the GMI wire is (Panina et al, 1994) [37]

$$Z=\mathrm{ }\frac{R_{dc}kr{\zeta }_0\left(kr\right)}{2{\zeta }_1\left(kr\right)}$$E12

,

where

$$k=\frac{1+j}{{\delta }_m}$$E13

and

$${\delta }_m=\frac{c}{\sqrt{4{\pi }^{2}f\sigma {\mu }_{\mathrm{\varnothing }}}}.$$E14

R_dc is the dc resistance of the wire, ζ ₀, ζ ₁ are the Bessel functions, r is the radius of the wire, j is the imaginary unit, δ_m is the penetration depth, c is the speed of light, f is the frequency of the ac current, μ_ø is the circumferential magnetic anisotropy. The origin of the GMI effect lies in the dependence of μ_ø on an axial magnetic field resulting in a change of δ_m. In order to obtain a high GMI ratio, the value of δ_mhas to be close to the thickness of the conductor.Hence,the thinner a ferromagnetic conductor and the lower its permeability,the higher the operation frequency required.A well-defined circumferential magnetic anisotropy in combination with a soft magnetic behavior is desirable, since it will provide a large permeability change for small magnetic fields.

Different amorphous and ferromagnetic materials were used to fabricate GMI wires [48], and various fabrication methods were developed such as melt spinning, in-rotating water spinning, glass-coated melt spinning etc. [45, 49, 50].Glass-coated micro-wires (Fig. 5)present outstanding properties in terms of the magnetic anisotropy distribution, which is reinforced by the strong mechanical stress induced by the coating. (CoxFe1-x)72.5Si12.5B15is one of the most typical materials. By adjusting x from 0 to 1, the magnetostriction of the material changes from positive at high Fe content to negative at high Co content. Negative magnetostrictive compositions in combination with the compressive, radial stress induced by quenching and the glass coating provide the best results, since it supports a strong circumferential anisotropy.

Figure 5.

SEM image of a glass-coated amorphous micro-wire (Courtesy of M. Vazquez, Inst. Materials Science of Madrid, CSIC).

Figure 6.

GMI ratio of Co67Fe3.85Ni1.45B11.5Si14.5Mo1.7 glass-coated wires with different geometric ratio ρ(themetallic nucleus diameter to the total microwire diameter)at 10 MHz.

Wire-type GMI sensorsprovide the best performance in terms of the GMI ratio with values as high as 615% (Fig. 6)achieved with optimized glass coated microwires (Zhukova et al, 2002) [43].The value of the magnetic field at which the maximum GMI ratio is obtained increasesas the diameter of the magnetic nucleus decreases compared to the diameter of the glass coating. This is attributed to the different anisotropies induced by the stress from the coating. Due to the high sensitivity provided by GMI wiresthey have been commercialized despite the facts that fabrication is not silicon based, does not use standard microfabrication methods and, as a consequence, integration with electronics is complex.

Figure 7.

a) Layout of the commercialized GMI sensor from Aichi Steel Co. (b) Noise output of the GMI sensor.

Fig. 7showsa GMI sensor developed by Aichi Steel Co., which has a very high sensitivity of 1V/μT and a noise level of 1 nT[51].

3.2.3. Ribbon GMI sensors

Magnetic ribbons discussed in this section are planar structuresof rectangular shape with a thickness of a few tens of micrometers and a length and width from several millimeters to centimeters.Similar to the micro-wires, magnetostriction is utilized in order to create certain anisotropies during the ribbon’s fabrication. Magnetic ribbons that exhibit a strong GMI effect have a high permeability as well as a transversal magnetic anisotropy.

For a planar film of infinite width, the impedance is given by

$$Z=\mathrm{ }{R}_{dc}·\frac{jka}{2}coth\left(\frac{jka}{2}\right)$$E15

,

whereR_dc is the dc resistance, a is the thickness of the ribbon, kand δ_m can be obtained fromEqu. (14) with the only difference that μ_ørepresents the transversalpermeability instead of the circumferential one [52].

Again, Fe-and Co-based amorphous alloys are preferably used as the magnetic material. The standard fabrication method for the ribbons is melt spinning, where a rotating copper wheel is used to rapidly solidifythe liquid alloy.This method produces magnetic ribbons with a thickness of about 25μm and a width of several mm. With this thickness, ribbon GMI sensorsoperate at comparably low frequencies of hundred kHz up to a few MHz.A GMI ratio of, e.g., 640% has been obtained with a GMI ribbon made of Fe₇₁Al₂Si₁₄B_8.5Cu₁Nb_3.5at 5 MHz [53].

3.2.4. Thin film GMI sensors

In theory, a single layer magnetic thin film is similar to a magnetic ribbon,and the same analytical expressions are applied for modeling the GMI effect. Practically, the main difference is the fabrication method. Thin film fabrication is a standard micro-fabricationtechnology producing a film thickness of somenanometersup to a few micrometers. Thin film GMI sensorsare of great interest due to the advantages arising from the fabrication in terms of the flexibility in design and integration. They can easily be fabricated on the same substrate as the electronic circuit and other devices. In the context of passive and remote sensors, this is particularly relevant, since the GMI element can be easily integrated with an SAW device. For this reason, GMI thin film sensors will be discussed in more detail and our recent results will be presented.

Compared to wires and ribbons, the results obtained with thin film sensors have not been as good, and the highest GMI ratios reported are around250%[42]. This may be due to the differencesin the magnetic softness as well as the magnetic anisotropy, which is very well established in circumferential and transversal direction in wires and ribbons, respectively, and is difficult to control in thin films. Inthin films transverse anisotropy is mainly realized through magnetic field deposition or field annealing,Fig. 8 (a) and (b) show the magnetization curve and domain structure of aNi₈₀Fe₂₀thin film (100nm thick) fabricated under a magnetic field of 200 Oe during deposition. A magnetic easy axis and domain structures in transverse direction are observed. Due to the small thickness, thin film GMI sensors normally operate at a higher frequency from hundred MHz to several GHz where the penetration depth is in the range of the film thickness.

Figure 8.

a) Magnetization curves obtained by vibrating sample magnetometry of a magneticthin film (100nm of Ni₈₀Fe₂₀) in transversal and longitudinal directions. (b) Domain pattern of the magnetic layer. (c) Schematic of a typical multi layer GMI structure. The arrows in the ferromagnetic material indicate the magnetization of individual domains (simplified). Upon application of an external magnetic field H_ext, the magnetization rotates into the direction of H_ext (dotted arrows).

In general, a GMI sensor with high sensitivity consists of a stack of several material layers. In case of a tri-layer element, one conducting layer is sandwiched between two magnetic layers as shown in Fig. 8 (c). The conducting layer ensures a high conductivity and, in combination with the highly permeable magnetic layers, a large skin effect is obtained [51, 54].An alternating current I_ac mainly flowing through the conductor generates a transversal flux B_tran, which magnetizes the magnetic layers. Upon the application of an external field H_extin longitudinal direction, the magnetization caused by I_acwill be changed. This is equivalent to a change of the transversal permeability of the magnetic layers and is reflected by an impedance change.

The analytical model of the impedance for a magnetic/conducting/magnetic tri-layer structureis given by

$$Z=\mathrm{ }{R}_{dc}(1-2j\mu \frac{d_1{d}_2}{{\delta }_c^2})$$E16

,

whereR_dc is the dc resistance of the inner conductor, 2d₁is the thickness of the conductor,d₂is the thickness of the magnetic layers as shown in Fig. 8 (c) andδ_c is the penetration depth of the conducting layer[39].

Analytical solutions for the impedance of thin film GMI sensors can only be found for rather simple structures. In order to calculate the impedance of more complicated geometries, for example,a sandwich structure with isolation layersbetween the conductor and the magnetic layers [41], a meander structure multilayer [46] or to take into account edge effects, the finite element method (FEM) provides a viable solution [55].

Fig. 9 shows the comparison of the GMI ratios simulated for a single magnetic layer, a tri-layerstructure made of a magnetic/conducting/magnetic stack and a five-layer structure with isolation layers between the conducting and magnetic layers using the FEM. The simulated GMI sensors have a width of w = 50 µm and length of l= 200µm. The magnetic layers have a thickness of t_mag = 1 µm andthe conducting layer has a thickness oft_met = 4 µm. The material of the isolation layer is SiO₂ with a thickness of 1µm. The conductivity of the ferromagnetic and conducting layers are 7.69×10⁵S/m ((CoFe)₈₀B₂₀) and 4.56×10⁷S/m (Gold), respectively. All parameters including M_s = 5.6×10⁵A/m, $\gamma$= 2.2×10⁵m/(A s), α = 0.3, H_a= 1890A/m and are taken from literature [56].

Figure 9.

Simulated GMI ratios of single layer, sandwiched multilayer and isolated sandwiched multilayer structures.

The resultsclearly show the performance increase achieved with the multilayer structures. Specifically, the isolated sandwich structure has a superior performance, which is due to preventing the current from flowing in the magnetic layer.

For the fabrication of thin film GMI sensors, Co-based and Fe-based amorphous magnetic alloys were used in earlierstudies. Recently, permalloy, which is a NiFe compound, became popular as it provides very high permeability, zero magnetostriction and simple fabrication. Meander shaped multilayers and different stacks of magnetic and conductive layers using permalloy were developed. Some results are summarized in Table 2.

Year	Material	Frequency	GMI Ratio; (%)	Sensitivity (%/Oe)	Reference
1999	FeNiCrSiB/Cu/FeNiCrSiB	13MHz	77	2.8	[57]
2000	FeSiBCuNb/Cu/FeSiBCuNb	13MHz	80	2.8	[58]
2004	Ni81Fe19/Au/Ni81Fe19	300MHz	150	30	[59]
2004	(Ni81Fe19/Ag)n	1.8GHz	250	9.3	[42]
2005	FeCuNbSiB/SiO2/Cu/SiO2/FeCuNbSiB	5.45MHz	33	1.5	[60]
2011	NiFe/Ag/NiFe	1.8GHz	55	1.2	[61]
2011	NiFe/Cu/NiFe	20MHz	166	8.3	[62]

Table 2

Recent results on thin film GMI sensors.

GMI thin film sensors not only offer the advantages of standard microfabrication and straight-forward integration with SAW devices, but, as can be seen from Table 2, the operation frequency of GMI thin film sensors is also compatible with the one of SAW devices (usually from hundred MHz to several GHz and can be adjusted within a wide range.

3.2.5. Integrated SAW-GMI sensor

In the first studies, SAW transponders and GMI wire sensors werecombined to form remote devices [14-16]. GMI wires were selected for their high sensitivity, and they were bonded totheoutput IDT of the SAW device, which operated as a reflector, in order to act as load impedance. The strong dependence of the impedance on magnetic fields causes a considerable amplitude dependence of the reflected signal on magnetic fields. Even though these studies provided good results for passive and remote magnetic field sensors, the fabrication method for the GMI wires, which is not compatible with standard microfabrication, is a considerable problem with respect to reproducibility and costs, hence, hindering commercial success of such devices. In order to conquer this problem, a fully integrated SAW-GMI design utilizing standard microfabrication processes is required. The most viable option is a thin film GMI sensorsfor the following reasons:

1. Thin film GMI sensors can be produced by the same metallization processes as the SAW transponders and on the same substrate.

2. Standard photolithography technique guarantees an accurate and reproducible alignment of the two devices.

3. Thin film GMI sensors provide a wide range of working frequencies up to GHz, which matches the high frequency requirement of the SAW transponders.

4. Thin film GMI sensor can have a minimized and flexible design as well as large magnetic field sensitivity.

In this section, a detailed description of our recent work on the design, fabrication and testing of an integrated SAW-GMI sensor is presented.

Design

Fig. 10 shows a schematic of a GMI thin film sensor integrated with a SAW transponder. A wireless signal applied to the source IDT (IDT1) is converted to anSAWand propagates towards the other end of the substrate, where it is reflected from the reference IDT (IDT2) and the load IDT (IDT3). The reflected waves containing the reference and load information are received by IDT1 at different time instants and reconverted to a wireless electrical signal sent out via the antenna.

Figure 10.

Schematic of an integrated passive and remote magnetic field sensor consisting of a SAW transponder and thin film GMI sensor.

In order to obtain high magnetic field sensitivity, the GMI sensor is matched to the output port (IDT3) at the working frequency of the SAW device. As the impedance of the GMI sensor changes with an applied magnetic field, the matching deteriorates, which causes the amplitude of the signal reflected from IDT3 to change. Since the piezoelectric material is sensitive to environmental changes, e.g. temperature, a reference IDT is used to provide a signal that enables the suppression of such noise by means of signal processing. Two metallic pads next to the input and output IDTs act as mechanical absorbers and suppress reflections from other structures on the substrate or the edge of the substrate.

Matching the sensor load to the optimal working point of IDT3 is a crucial aspect in the device design.Therefore, the influence of the load on the signal reflected from IDT3 issimulated. The interaction of a SAW with an IDT can be described by the P-matrix model introduced by Tobolka [63]. As shown in Fig. 11, P₁₁ is the acoustic wave reflection at the output IDT [24]. Specifically, the dependence of P₁₁ on the load impedance Z = Z(H_ext)+ Z_m, where Z(H_ext) is the impedance of the GMI element and Z_m is the matching impedance, is expressed as

$$P_{11}\left(Z\right)=\mathrm{ }{P}_{11,sc}+\mathrm{ }\frac{2\bullet P_{13}^2}{P_{33}+\frac{1}{Z}}$$E17

,

whereP_11,sc is the short circuit reflection coefficient, P₁₃ is the electro-acoustic transfer coefficient andP₃₃ is the input admittance of the transducer. In order to have a large change of P₁₁, which is equivalent to the sensitivity of the SAW device loaded by an impedance sensor, the influence of Z in equation (17) needs to be large. Therefore, a SAW transducer with a small P_11,scand largeP₁₃ will provide a large sensitivity. P_11,sc can be minimized by using a double electrode IDT design as shown in Fig. 10, which provides cancelation of the internal mechanical reflections of the IDT.

Figure 11.

Electric and acoustic ports of the SAW sensor

The electro-acoustic transfer coefficient P₁₃ and input admittance P₃₃ can be obtained by,

$$P_{13}=\frac{1}{2\sqrt{2Z_a}}{r}_m(1-e^{-j{\phi }_m})$$E18

$$P_{33}=j\omega C_{IDT}+\frac{r_m^2}{Z_a}(1-e^{-j{\phi }_m})$$E19

Where r_m is the ratio of the electrical to acoustical transformer, C_IDT is the capacitance of the IDT,Za is the acoustic impedance and ${\phi }_m$ is the transit angle [63].

Since the GMI sensor is an inductive element, matching is accomplished by a series capacitance resulting a load impedance

Z = 1/jωC_m + R +jωL(H_ext),(20)

whereC_m is the matching capacitance, R is the average resistance (over the considered magnetic field range) of the GMI sensor and L(H_ext) the inductance of the GMI sensor.

Fig. 12 (a) shows the simulation result of the IDT’s reflectivity as a function of the load. The slope of this plot corresponds to the magnetic field sensitivity. Therefore, the optimum matching capacitance can be determined. Fig. 12 (b) presents the rate of change of P₁₁for 1nH inductance changes (corresponding to a field change of approximately 50A/m). The result shows that with the optimum matching capacitance a maximum reflectivity change rate of 0.3dB/nH can be achieved. As the fabricated GMI sensor has an inductance change from 5nH to 15nH, a reflectivity change of 3dB can be expected.

Figure 12.

a) Magnitude P₁₁ as a function of the matching capacitance and sensor inductance. (b) Rate of change (absolute value) ofP₁₁ for 1nH load inductance change.

The piezoelectric substrate chosen for this application is LiNbO₃ as it provides a strong electromechanical coupling corresponding to a high value ofP₁₃. The detailed design parameters of the SAW transponder are shown in Tab. 3. The working frequency of the device is 80MHz, resulting in a periodicityp of 50μm (Equ. (1). The value ofp determines the electrode width and gap. The distances between the IDTs yielda1.25μs delay between IDT1 and IDT2 and a 0.625μs delay between IDT2 and IDT3.

Design parameter				Design parameter
Substrate material	LiNbO3 (128 deg. Y-X cut)			Electrode material	Gold
Center frequency	80MHz			Aperture	30λ
Periodicity		50μm		Electrode/gap width		6.25μm
Electrodes per segment		4		IDT segment number		30

Table 3

Design parameters for the SAW device.

The GMI sensor consists of a tri-layer structure with two ferromagnetic layers of 100nm in thickness made of Ni₈₀Fe₂₀ and a conducting copper layer with a thickness of 200nm. The sensor has a rectangular geometry of 100 μm× 4000 μm. The conducting layer is connected to the IDT3 [18].

Fabrication

The fabrication of the combined device is accomplished in several steps as shown in Fig. 13. On a LiNbO₃ wafer, a 40 nm Ti adhesion layer and 200 nm gold layerare sputter deposited and patterned by ion milling into individual SAW devices. The leads and SMD footprints are designed together with the SAW device to facilitate an on-chip impedance matching circuit, which was accomplished by a 150pF capacitor connected in series with the GMI element.The GMI element comprises a tri-layer structure (Ni₈₀Fe₂₀(100nm)/Cu(200nm)/Ni₈₀Fe₂₀(100nm)) deposited at room temperature with a uniaxial magnetic field of 200 Oe applied in the transversal direction.

Figure 13.

Fabrication flow chart of the integrated SAW-GMI device.

Results

A network analyzer (Agilent E8363C) is used to applyan RF signal to IDT1 and measure the electric reflection coefficient (S₁₁) of the input IDT, which is related to the admittance matrix of the whole device and P₁₁.The time domain signal of S₁₁is converted from the frequency domain using fast Fourier transform. As shown in Fig. 14 (a),two reflection peaks at 2.45μs and 3.55μsare observed indicating the reflections from the reference IDT and the load IDT accordingly. The magnetic response of the integrated device is determined by applying a variable magnetic field in longitudinal direction to the device. A 2.4dB amplitude change of the reflection signal can be observed. A comparison of the simulated and experimental results together with the measured GMI ratio curve is shown in Fig. 14 (b).

Figure 14.

a) Time domain measurement of the SAW-GMI device. Inset: Frequency domain measurement. (b) Comparison of the simulated and experimental device response together with the measured GMI ratio curve.

4. Potential applications

Magnetic field sensors are one of the most widely used sensors and employed for many different applications. Current commercial magnetic sensors are wire connected to a circuit providing power and readout. These wire connections prevent the sensors from being used for certain applications. In addition, as the complexity and the number of devices, increases in modern systems such as automobiles, wire connections are becoming an increasing problem due to limited space. For those and other reasons, wireless solutions are being much sought after.

As pointed out in the previous sections, SAW-based sensors have been developed for different applications, and this technology also provides a platform for realizing wireless and passive magnetic sensors. They can provide solutions for various applications, for example, where the sensors have towithstand harsh environmental conditions or are required to have a long lifetime without maintenance.

Out of the countless applications for SAW-based passive and remote magnetic field sensors, a few will be highlighted in the following.

Nanotechnology and miniaturized systems are becoming increasingly popular in the biomedical field. Technologies based on magnetic effects are of particular interest since they can be controlled remotely via magnetic fields. For example, NVE Corporation recentlydeveloped a battery operated magnetic sensor to be used as a magnetic switch for implantable devices. When a magnetic field is applied, the sensor turns on triggering a specified action. It turns off when the field is removed. The sensor works at a stable operating point of 15 Oe [64]. Magnetic beads have been extensively used in many biomedical applications. These magnetic beads are made of ferromagnetic material ranging in size between 5 nm to 500 um. A new application of such particles promises benefits in cancer therapy by employing the particles either as drug carriers or heat sources (hyperthermia) [65]. In order to have better control of the treatment, magnetic sensors are considered to measure and detect the concentration of these magnetic particles.

The automotive industry extensively uses magnetic sensors for different purposes, for example, to measure current in electric vehicles [66] or the rotation speed of gears [67]. Another application employs magnetic sensors to detect passing vehicles using lane markers [68]. Such a system could also be used to detect vehicle speed by measuring the time between two markers of a fixed distance. In yet another application, developed by Stendl et al [69], the wear of a vehicle’s tire is detected by measuring the field of magnetic beads embedded in the rubber of tire treads. As the tread size decreases, the magnetic field alsodecreases. A wireless magnetic sensor is placed just below the threads.

Construction monitoring is an upcoming application for wireless sensor. Long-term monitoring of metallic reinforcements in, e.g., bridges or buildings requires passive and remote sensors, which are capable of detecting changes of the metal. Similarly, the detection of internal defects or corrosion of pipelines is of great interest. Gloria et al [70] developed an Internal Corrosion Sensor (ICS) consisting of a magnet and a Hall sensor. A disturbance in the magnetic field caused by changes of the metal changes the sensor readout. This information is used for both to detect and size the defects.

5. Conclusion

In this chapter we discussed different types of SAW-based, magnetic sensors including resonators, phase shifters and loaded transponders. Sensitivity to magnetic fields can be achieved byeither changing the properties of the IDT or delay line utilizing magnetostrictive materials or loading the output IDT with a magnetic field sensor. GMI sensors feature a very high sensitivity and wide range of operating frequencies and, therefore, constitute an especially suitable load. The principle of GMI sensors is described in detail and different GMI structures are discussed. While the highest sensitivity has been obtained with GMI microwires, thin film GMI sensor are advantageous because they can be produced using standard microfabricationmethods, and they can be easily integrated with a SAW transponder on the same substrate. These features are crucial with respect to production complexity and costs.

A SAW transponder combined with a GMI element connected to the output IDT is a passive and remote magnetic field sensor, which responds to an interrogation signal with a delayed response signal. The design of such a device needs to take into account different aspects like operation frequency, dimensions of IDTs and delay line or matching the load with the output IDT. In order to obtain a high sensitivity, an impedance change of the GMI element caused by a magnetic field, has to yield a large change in the SAW reflected from the output IDT. A model is presented to simulate the electro-acoustic interaction of the output IDT with the GMI sensor’s impedance and the impedance matching capacitance. The simulation results provide information regarding the matching parameters and are invaluable for obtaining anoptimized performance. A detailed description of the fabrication of an integrated SAW-GMI sensor is providedusingstandard microfabricationtechnologies.The GMI ratio of the fabricated sensor is 45 % The SAW-GMI sensor provides a sensitivity of 3 dB/mT, and its output corresponds well with the simulation results.

Magnetic field sensors have countless applications and are widely used in many different fields. The trend towards wireless operation, which is generally observed nowadays, drives the development of passive and remote magnetic field sensors. Several concepts of such sensors employing SAW devices have been presented in this chapter. The most promising one is a SAW-GMI sensor, which has been discussed in detail and which features wireless and battery less operation as well as durability and the ability to withstand harsh environments. This kind of sensor is considerable not only for providing existing applications with a wireless mode; it also largely expands the potential applications of magnetic field sensors.