**3. Sensor fabrication and characterization**

#### **3.1. Fabrication procedure of a novel planar Hall sensor**

Nowadays, with the advancement of the accurate sputtering and lithography technologies, the sensor with desired composition in micro-size can be easily fabricated by using a lift off method. The general fabrication procedure of a novel exchange biased planar Hall sensor, for example in typical spin valve geometry, using the lift off method is shown in a simpli‐ fied description in the figure 2 below. However, the same procedure is applied for fabrica‐ tion of other PHR sensors too in different geometries mentioned in this book chapter.

The SiO2 wafer is first cleaned in the acetone and methanol solutions while placing in the ultrasonic bath, then the SiO2 wafer is covered by a commercial photoresist such as Az (5214E, 9260,…) or SU8-(2000, 3000,…) by using a spin coating system with a defined thick‐ ness. The blank cross-junctions are stenciled out on the photoresist coated on SiO2 wafer *i.e*., the sample is aligned and exposed by a mask aligner system. The short wavelength of ultra‐ violet source *i.e.*, 456 or 654 nm is used for the exposure, and then the sample is developed by an appropriate developer followed by cleaning the same in DI water.

The sensor materials, *i.e*., spin-valve structure Ta(5)/NiFe(10)/Cu(1.5)/NiFe(2)/IrMn(10)/Ta(5) (nm), is deposited on the stenciled photoresist layer by using magnetron sputtering system. The base pressure of the system is less than 10-7 Torr and the Ar working pressure is 3 mTorr. During the deposition, a uniform magnetic field of 200 Oe was applied in the thin film plane to induce magnetic anisotropy of the ferromagnetic pinned layer and to define the unidirectional field of the thin films. After the thin film deposition, the sample was lifted off in acetone and methanol solutions in order to remove the photoresist as well as the sen‐ sor material on this photoresist, so that the sensor material exists on the stenciled junctions only.

After fabricating sensor junctions, the electrodes made by Au are connected with sensor junction to establish the external circuitry and to measure the sensors' response. Further, the sensor junctions and the electrodes are passivated with a SiO2 or a Si3N4 layer coated on top of the sensor junctions and electrodes to protect them from the corrosion and fluid environ‐ ment during the experiments. Finally, the sensor is activated by a very thin Au layer for biomolecule immobilization. All these steps are carried out at the same way for all the sensors as the steps for the sensor junction fabrication.

**Figure 2.** The pattern processes for fabricating the sensor junction, electrodes, passivation and Au activation layers of a planar Hall sensor, this sensor is ready for bio-manipulation.

#### **3.2. Sensor characterization**

The comparison results have shown that the PH-AMR or PHE sensor has prominent advan‐

tion (*µ*o*H*min) in the detection of the magnetic field. Furthermore, the voltage profile of a PHE sensor responds linearly to the magnetic field at the small values. This is a prominent ad‐ vantage in detection of small stray field induced from magnetic labels. Therefore, we have chosen and mainly focused on the development of the PHE sensors for bio-applications.

Nowadays, with the advancement of the accurate sputtering and lithography technologies, the sensor with desired composition in micro-size can be easily fabricated by using a lift off method. The general fabrication procedure of a novel exchange biased planar Hall sensor, for example in typical spin valve geometry, using the lift off method is shown in a simpli‐ fied description in the figure 2 below. However, the same procedure is applied for fabrica‐

The SiO2 wafer is first cleaned in the acetone and methanol solutions while placing in the ultrasonic bath, then the SiO2 wafer is covered by a commercial photoresist such as Az (5214E, 9260,…) or SU8-(2000, 3000,…) by using a spin coating system with a defined thick‐ ness. The blank cross-junctions are stenciled out on the photoresist coated on SiO2 wafer *i.e*., the sample is aligned and exposed by a mask aligner system. The short wavelength of ultra‐ violet source *i.e.*, 456 or 654 nm is used for the exposure, and then the sample is developed

The sensor materials, *i.e*., spin-valve structure Ta(5)/NiFe(10)/Cu(1.5)/NiFe(2)/IrMn(10)/Ta(5) (nm), is deposited on the stenciled photoresist layer by using magnetron sputtering system. The base pressure of the system is less than 10-7 Torr and the Ar working pressure is 3 mTorr. During the deposition, a uniform magnetic field of 200 Oe was applied in the thin film plane to induce magnetic anisotropy of the ferromagnetic pinned layer and to define the unidirectional field of the thin films. After the thin film deposition, the sample was lifted off in acetone and methanol solutions in order to remove the photoresist as well as the sen‐ sor material on this photoresist, so that the sensor material exists on the stenciled junctions

After fabricating sensor junctions, the electrodes made by Au are connected with sensor junction to establish the external circuitry and to measure the sensors' response. Further, the sensor junctions and the electrodes are passivated with a SiO2 or a Si3N4 layer coated on top of the sensor junctions and electrodes to protect them from the corrosion and fluid environ‐ ment during the experiments. Finally, the sensor is activated by a very thin Au layer for biomolecule immobilization. All these steps are carried out at the same way for all the sensors

tion of other PHR sensors too in different geometries mentioned in this book chapter.

by an appropriate developer followed by cleaning the same in DI water.

as the steps for the sensor junction fabrication.

only.

) as well as very high resolu‐

tages over others such as very high signal-to-noise ratio (S/Nf

**3. Sensor fabrication and characterization**

202 State of the Art in Biosensors - General Aspects

**3.1. Fabrication procedure of a novel planar Hall sensor**

The quality of the spin-valve structure, Ta(3)/NiFe(10)/Cu(1.5)/NiFe(2)/IrMn(10)/Ta(3) (nm), as observed by a cross sectional transmission electron microscope (TEM) image and also by an energy dispersive X-ray (EDX) spectrum along its thickness, are shown in Fig. 3. The TEM-specimen was prepared by polishing the Si/SiO2 substrate mechanically to a thickness of about 100 *µ*m. After that, the dimpling and hollowing steps were performed at the opti‐ mum conditions to ensure that the sample is undamaged by using a GATAN-691 precision ion polishing system (PIPS), *i.e*., using the Ar-ion beam with an energy of 4.3 keV and under an angle of 6o .

It is evident from Fig. 3(a) that the existence of a multilayer structure is clearly revealed as pointed out by an arrow for each layer. The both seed and top layers of Ta have amorphous behavior and the thickness is about 3 nm. However, there is a difference in the color of the two layers; this can be assumed that the Ta top layer is slightly oxidized [20]. The IrMn layer is well defined, and its thickness is about 10 nm as the nominal thickness when the layer is deposited. In the NiFe/Cu/NiFe region, it is clearly seen that the diffusion of Cu takes place into the adjacent NiFe layers. This kind of diffusion is known to influence the anisotropy significantly [21]. There is also an existence of a rumpling or even a rupture of the layers in some parts. This can be explained when considering the roughness of the NiFe layers; the roughness of NiFe layer is normally about 1.5 – 2 nm [22].

**Figure 3.** A cross sectional TEM image (a) and an EDX spectra (b) of a spin-valve structure Ta(3)/NiFe(10)/Cu(1.5)/ NiFe(2)/IrMn(10)/Ta(3) (nm).

The result of the cross sectional TEM image is supported by the EDX patterns of the same sample shown in Fig 3(b). It can be seen that the peak of Cu is mixing inside the Ni and Fe peaks indicating the diffusion of Cu in NiFe layers. The overlap between the peaks of Ta and Ni (Fe), of Ta and Ir (Mn), and of Ir (Mn) and Ni (Fe) confirms the roughness at the sur‐ face of the NiFe and IrMn layers. Moreover, the shadow in Ni and Fe peaks (black arrows in Fig. 3(b)) indicates the separation of the NiFe pinned and NiFe free layers.

The magnetic property of the fabricated spin-valve structure used for sensor material is characterized by a vibrating sample magnetometer (VSM) of the make Lakeshore 7407 series with a sensitivity of 10-6 emu. The external magnetic field is swept in the film plane.

In order to achieve the magnetic anisotropy of the free layer in the fabricated spin-valve structure, we measured the magnetization as a function of external magnetic field in the range of ± 80 Oe in both the easy and hard axis, which is presented in Fig. 4. The shift along the external magnetic field axis of the magnetization profile (*M*(*H*)) in the easy axis indicates an effective uniaxial anisotropy field of the spin-valve structure (*H*Keff) by incorporating the free layer shape anisotropy field (*H*demag.) and its uniaxial anisotropy field (*H*K) analyzed from the shift of the *M*(*H*) profile in the hard axis (*H*Keff = *H*<sup>K</sup> + *H*demag. ) [23]. This indicates that the free NiFe layer (active layer) has very good anisotropy characteristic for further study of the PHE sensor. In addition, the inset in Fig. 4 exhibits a two-step hysteresis loop; one is from the interlayer coupling and the other is from the exchange bias coupling. The magnetization of the first hysteresis loop (contributing from 10 nm NiFe free layer) is five times larger than the second one (contributing from 2 nm NiFe pinned layer). The interlayer coupling between the ferromagnetic (F)-free and F-pinned layers separated by a non-mag‐ netic layer (Cu) is determined from the first step of the hysteresis loop. Whereas, the ex‐ change bias field due to the interface between the F-pinned and antiferromagnetic (AF) layers is determined from the second step of the hysteresis loop. The obtained interlayer and interfacial coupling fields are 11 Oe and 550 Oe, respectively. This result elucidates that the NiFe pinned and NiFe free layers are separated by a Cu layer.

**Figure 4.** Hysteresis loops of the spin-valve thin film, Ta(3)/NiFe(10)/Cu(1.5)/NiFe(2)/IrMn(10)/Ta(3) (nm), character‐ ized in the easy and hard axis in the field interval from + 80 to -80 Oe. The inset shows the hysteresis loop character‐ ized in the easy direction in the field range of -800 to 20 Oe.

#### *3.2.1. Microarray of the magnetic sensors*

behavior and the thickness is about 3 nm. However, there is a difference in the color of the two layers; this can be assumed that the Ta top layer is slightly oxidized [20]. The IrMn layer is well defined, and its thickness is about 10 nm as the nominal thickness when the layer is deposited. In the NiFe/Cu/NiFe region, it is clearly seen that the diffusion of Cu takes place into the adjacent NiFe layers. This kind of diffusion is known to influence the anisotropy significantly [21]. There is also an existence of a rumpling or even a rupture of the layers in some parts. This can be explained when considering the roughness of the NiFe layers; the

**Figure 3.** A cross sectional TEM image (a) and an EDX spectra (b) of a spin-valve structure Ta(3)/NiFe(10)/Cu(1.5)/

The result of the cross sectional TEM image is supported by the EDX patterns of the same sample shown in Fig 3(b). It can be seen that the peak of Cu is mixing inside the Ni and Fe peaks indicating the diffusion of Cu in NiFe layers. The overlap between the peaks of Ta and Ni (Fe), of Ta and Ir (Mn), and of Ir (Mn) and Ni (Fe) confirms the roughness at the sur‐ face of the NiFe and IrMn layers. Moreover, the shadow in Ni and Fe peaks (black arrows in

The magnetic property of the fabricated spin-valve structure used for sensor material is characterized by a vibrating sample magnetometer (VSM) of the make Lakeshore 7407 series

In order to achieve the magnetic anisotropy of the free layer in the fabricated spin-valve structure, we measured the magnetization as a function of external magnetic field in the range of ± 80 Oe in both the easy and hard axis, which is presented in Fig. 4. The shift along the external magnetic field axis of the magnetization profile (*M*(*H*)) in the easy axis indicates an effective uniaxial anisotropy field of the spin-valve structure (*H*Keff) by incorporating the free layer shape anisotropy field (*H*demag.) and its uniaxial anisotropy field (*H*K) analyzed

that the free NiFe layer (active layer) has very good anisotropy characteristic for further study of the PHE sensor. In addition, the inset in Fig. 4 exhibits a two-step hysteresis loop;

) [23]. This indicates

with a sensitivity of 10-6 emu. The external magnetic field is swept in the film plane.

Fig. 3(b)) indicates the separation of the NiFe pinned and NiFe free layers.

from the shift of the *M*(*H*) profile in the hard axis (*H*Keff = *H*<sup>K</sup> + *H*demag.

roughness of NiFe layer is normally about 1.5 – 2 nm [22].

204 State of the Art in Biosensors - General Aspects

NiFe(2)/IrMn(10)/Ta(3) (nm).

Fig. 5 shows a complete micro-array of planar Hall resistance (PHR) sensors. In the figure, it was shown that the unidirectional field, *H*ex, and/or the uniaxial field of the thin film were aligned parallel to the terminals *a*–*b*, and a sensing current of 1mA was applied through these terminals. The output voltages were measured from the terminals *c* and *d* at room tem‐ perature under a specific range of external magnetic field applied normal to the direction of the current.

**Figure 5.** Complete micro-array of a 24 element PHR sensor(a), which can even detect a single micro-paramagnetic Dynabead M® -280. Inset of the figure (b) shows a single micro sized cross-junction.

#### **4. Evolution of novel PHR sensors**

Among all the developed magnetoresistive sensors for bioapplications, we mainly focus on the development of PHE sensor because it has prominent advantages compared with others such as signal-to-noise ratio, linearity signal *etc*. Various structures will be used for planar Hall sensor *i.e.*, bilayer, trilayer and spin-valve. Therefore, a short introduction of these mul‐ tilayer structures will be given in this section. Also, the theoretical approach of planar Hall effect in different sensor geometries, such as cross-junction, tilted cross-junction and ring junction, will be discussed. Finally, the description leads to evolution of hybrid AMR-PHR sensor with optimized sensor characteristics for effective use in bioapplications.

#### **4.1. AMR sensor**

The magteoresistive anisotropy in ferromagnetic material depends on the direction of mag‐ netization. The electric field due to the magnetoresistivity is expressed as follows [24];

$$
\vec{E} = \rho\_\perp \vec{j} + (\rho\_\parallel - \rho\_\perp) \vec{m} (\vec{j} \cdot \vec{m}) \tag{3}
$$

where *m* <sup>→</sup> is magnetization vector in single domain, and *j* <sup>→</sup> is current density direction. The *ρ* and *ρ//* are the resistivity when the magnetization vector and current density direction are perpendicular and parallel, respectively. The Δ*ρ*= *ρ// - ρ* is defined by the anisotropic resis‐ tivity, which is the intrinsic resistivity by the spin-orbit scattering in ferromagnetic materi‐ als. In Eq. (3), the electric field can be measured in the current direction as well as perpendicular to current direction due to the anisotropic resistivity, which are called, as mentioned, as the AMR and PHR, respectively.

The AMR properties have been discovered at ferromagnetic material by William Thomson in 1857 [25]. In the AMR response, varying differences between the direction of the magnet‐ izing vector in the ferromagnetic film and the direction of the sensing current passing through the film lead to varying the resistance in the direction of the current. The maximum resistance occurs when the magnetization vector in the film and the current direction are parallel to one another, while the minimum resistance occurs when they are perpendicular to one another. The resistance change by AMR effect in the patterned film with thickness *t*, width *w* and length *l* can be expressed from Eq. (3).

$$V\_{\rm AMR} = I(R\_\perp + \Delta R \cos^2 \theta) \tag{4}$$

where *ΔR* =(*ρll* −*ρ*⊥)*l* / *ωt* is the anisotropic magnetoresitivity and the *θ* is the angle between the magnetization vector and current, *I*. In AMR effect, the MR ratio is expressed as *ΔR* / *R*<sup>⊥</sup> ×100. The AMR effect has an offset resistance of *R*┴. This offset resistance must be reduced to improve the performance by using a compensating voltage or a Wheatstone bridge circuit [26].

#### **4.2. Planar Hall resistance sensor**

**Figure 5.** Complete micro-array of a 24 element PHR sensor(a), which can even detect a single micro-paramagnetic

Among all the developed magnetoresistive sensors for bioapplications, we mainly focus on the development of PHE sensor because it has prominent advantages compared with others such as signal-to-noise ratio, linearity signal *etc*. Various structures will be used for planar Hall sensor *i.e.*, bilayer, trilayer and spin-valve. Therefore, a short introduction of these mul‐ tilayer structures will be given in this section. Also, the theoretical approach of planar Hall effect in different sensor geometries, such as cross-junction, tilted cross-junction and ring junction, will be discussed. Finally, the description leads to evolution of hybrid AMR-PHR

The magteoresistive anisotropy in ferromagnetic material depends on the direction of mag‐ netization. The electric field due to the magnetoresistivity is expressed as follows [24];

and *ρ//* are the resistivity when the magnetization vector and current density direction are perpendicular and parallel, respectively. The Δ*ρ*= *ρ// - ρ* is defined by the anisotropic resis‐ tivity, which is the intrinsic resistivity by the spin-orbit scattering in ferromagnetic materi‐ als. In Eq. (3), the electric field can be measured in the current direction as well as perpendicular to current direction due to the anisotropic resistivity, which are called, as

<sup>r</sup> r r r r (3)

<sup>→</sup> is current density direction. The *ρ*

( )( ) *ll E j mj m*

 rr^ ^ = +- ×

r

<sup>→</sup> is magnetization vector in single domain, and *j*

mentioned, as the AMR and PHR, respectively.


sensor with optimized sensor characteristics for effective use in bioapplications.

**4. Evolution of novel PHR sensors**

206 State of the Art in Biosensors - General Aspects

Dynabead M®

**4.1. AMR sensor**

where *m*

The planar Hall resistance (PHR) in ferromagnetic thin films was considered when the resis‐ tivity depends on the angle between the direction of the current density *j* and the magneti‐ zation *m*. For magnetization reversal of the single domain when *m* makes an angle *θ* with *j*, the electric field is described as follows;

$$E\_{\rm PHR} = j(\rho\_{\parallel \parallel} - \rho\_{\perp})\sin\theta\cos\theta \tag{5}$$

The PHR effect also varies when there is a difference between the direction of the magnetiz‐ ing vector in the ferromagnetic film and the direction of the sensing current passing through the film; however, it leads to varying the resistance in the perpendicular direction of the cur‐ rent only. The longitudinal component of PHR voltage is related to *E*PHR in Eq. (3) and can be revealed when anisotropy of resistivity exists. On the other hand, in this sensor, the meas‐ ured PHR voltage was described as follows:

$$V\_{\rm PHR} = \frac{I(\rho\_{\rm ll} - \rho\_{\perp})}{t} \sin \theta \cos \theta \tag{6}$$

where *t* is the thickness of ferromagnetic film. The PHR in Eq. (6) varies with the angle *θ*. The PHR does not impose the offset resistance. Therefore, it has the advantage of obtaining a large PHR ratio and a linear response characteristic when the angle *θ* having a small value. The PHR effect depends on the intrinsic magneto-resistivity, *Δρ* =*ρll* −*ρ*<sup>⊥</sup> and the sample thickness, *t*. This means that the PHR signal does not depend on the sensor size (width *ω* and length *l*). Therefore, the PHR sensor can be used as the micro- or nano sized sensor for the micro- or nano- bead detection maintaining the large output signal voltage.

In order to analyze the PHR signal with magnetic field, we must know the angle *θ* between the magnetization vector and current direction, which depends on the magnetic field. Fig. 6 shows the general coordinates used to describe the rotational magnetization process under the applied magnetic field in ferromagnetic/antiferromagnetic (F/AF) coupled samples. *H*ex is the exchange coupling field due to the antiferromagnetic layer, and it shows a biasing field effect. *K*u is the effective in-plane anisotropy constant with an angle *γ* from *H*ex.

**Figure 6.** The coordinates for domain rotation process. Here, γ and θ are the angles of the anisotropy constant, and magnetization from the exchange-coupling field, *H*ex, respectively, and *I* is the measuring current.

The applied magnetic field *H* is directed perpendicular to *H*ex, and force the magnetization to rotate by an angle *θ* towards *H*. We introduce the modified Stoner-Wolfforth model with magnetic energy density, *ET* for the F layer in the F/AF sample, which can be written in the following simple form [12, 27]

$$E\_{\rm T} = K\_{\rm u} \sin^2(\theta - \gamma) - HM\_{\rm s} \sin \theta - H\_{\rm ex} M\_{\rm s} \cos \theta \tag{7}$$

where *M*s is the saturation magnetization. The angle *θ* determines the orientation of the magnetization in an equilibrium state with minimum total energy, whose values are calcu‐ lated under the conditions of *∂ET/∂θ*=0.

thickness, *t*. This means that the PHR signal does not depend on the sensor size (width *ω* and length *l*). Therefore, the PHR sensor can be used as the micro- or nano sized sensor for

In order to analyze the PHR signal with magnetic field, we must know the angle *θ* between the magnetization vector and current direction, which depends on the magnetic field. Fig. 6 shows the general coordinates used to describe the rotational magnetization process under the applied magnetic field in ferromagnetic/antiferromagnetic (F/AF) coupled samples. *H*ex is the exchange coupling field due to the antiferromagnetic layer, and it shows a biasing

**Figure 6.** The coordinates for domain rotation process. Here, γ and θ are the angles of the anisotropy constant, and

The applied magnetic field *H* is directed perpendicular to *H*ex, and force the magnetization to rotate by an angle *θ* towards *H*. We introduce the modified Stoner-Wolfforth model with magnetic energy density, *ET* for the F layer in the F/AF sample, which can be written in the

> q

where *M*s is the saturation magnetization. The angle *θ* determines the orientation of the magnetization in an equilibrium state with minimum total energy, whose values are calcu‐

 q*HM H M* (7)

T u <sup>s</sup> ex s *E K*= -- - sin ( ) sin cos

magnetization from the exchange-coupling field, *H*ex, respectively, and *I* is the measuring current.

2

qg

following simple form [12, 27]

208 State of the Art in Biosensors - General Aspects

lated under the conditions of *∂ET/∂θ*=0.

the micro- or nano- bead detection maintaining the large output signal voltage.

field effect. *K*u is the effective in-plane anisotropy constant with an angle *γ* from *H*ex.

**Figure 7.** (a) Calculated *V*AMR and (b) *V*PHR with applied magnetic field in single ferromagnetic film

Fig. 7 shows the calculated *V*AMR and *V*PHR in ferromagnetic single layer without exchange bias field, *H*ex. The measuring configuration of AMR and PHR voltage is shown in the inset of the figures. The current was applied parallel to the easy axis and the magnetic field was applied parallel to the hard axis of the magnetic thin film. The AMR voltage was measured in the direction of the sensing current passing through the film, while the PHR voltage was measured in the perpendicular direction of the sensing current. In the case of AMR effect, the signal shows the symmetric behavior in the functions of applied magnetic field with off‐ set voltage of *R*⊥. The PHR signal shows the linear behavior in the functions of applied mag‐ netic field with zero offset voltage.

Therefore, the PHR sensor justifies that it can be used as the micro- or nano- sized magnetic field sensor for the detection of the micro- or nano- bead. The PHR signal in ferromagnetic single layer shows large hysteresis behavior. The hysteresis effect is due to the switching of the magnetization in ferromagnetic layer. In order to remove the hysteresis of PHR, the ex‐ change biased F/AF bilayers are considered.

#### **4.3. PHR effect in exchange biased F/AF multilayer structures**

The single ferromagnetic layer with high AMR ratio such as NiFe, CoFe and NiCo alloys has the uniaxial anisotropy. The easy axes for stable magnetization direction are 0 and 180 de‐ grees. If one cycle of magnetic field is applied in the perpendicular direction to the easy axis in ferromagnetic films, the magnetization direction changes from 0 to 90 degrees as the mag‐ netic field increases, and 90 to 180 degrees as the magnetic field decreases. And then the di‐ rection of the magnetization changes from 180 to 270 and then to 360 degrees as the reversed magnetic field increases and decreases. In that case, the AMR and PHR, which are depend‐ ent on the angle between the current and magnetization directions, can show the large hys‐ teresis loop. On the other hand, the exchange biased F/AF bilayers induce the unidirectional anisotropy, which rotates the magnetization direction from 0 to 90 and 90 to 0 degrees as the magnetic field increases and decreases, respectively. It means that the AMR and PHR signal in exchange biased F/AF bilayers show the reversal behavior and the hysteresis can be dis‐ appeared.

**Figure 8.** *V*PHR signal with applied magnetic field in exchange biased F/AF bilayers

Fig. 8 shows the calculated PHR signal for the exchange biased F/AF bilayers. By comparing the PHR signal of the exchange biased F/AF bilayers in Fig. 8 with that of the single ferro‐ magnetic layer in Fig. 7(b), we can clearly confirm that no hysteresis behavior of PHR signal takes place in exchange biased F/AF bilayers. The exchange bias field, *H*ex plays the role of the reversible rotation of the magnetization as the magnetic field changes, which is due to the unidirectional anisotropy compared with the uniaxial anisotropy in single ferromagnetic layer. Also the reversible rotation of the magnetization in exchange biased F/AF bilayers can reduce the Barkhausen noise, which is usually dominated in the irreversible domain motion. Therefore, the signal to noise ratio (S/N ratio) of PHR sensor can be increased by using the exchange biased F/AF bilayers. Further, the PHR effect in exchange biased bilayers shows good linearity and thus it has the advantage for magnetic field sensor application. In the case of GMR or TMR materials, though they have high MR ratios, however, theirs' linearity is not good compared with the PHR signal. Therefore, PHR effect in the exchange biased F/AF bilayers has advantages in use as a bio-sensor for micro or nano bead detection.

#### *4.3.1. Bilayers*

**4.3. PHR effect in exchange biased F/AF multilayer structures**

210 State of the Art in Biosensors - General Aspects

**Figure 8.** *V*PHR signal with applied magnetic field in exchange biased F/AF bilayers

Fig. 8 shows the calculated PHR signal for the exchange biased F/AF bilayers. By comparing the PHR signal of the exchange biased F/AF bilayers in Fig. 8 with that of the single ferro‐ magnetic layer in Fig. 7(b), we can clearly confirm that no hysteresis behavior of PHR signal takes place in exchange biased F/AF bilayers. The exchange bias field, *H*ex plays the role of the reversible rotation of the magnetization as the magnetic field changes, which is due to the unidirectional anisotropy compared with the uniaxial anisotropy in single ferromagnetic layer. Also the reversible rotation of the magnetization in exchange biased F/AF bilayers can reduce the Barkhausen noise, which is usually dominated in the irreversible domain motion. Therefore, the signal to noise ratio (S/N ratio) of PHR sensor can be increased by using the

appeared.

The single ferromagnetic layer with high AMR ratio such as NiFe, CoFe and NiCo alloys has the uniaxial anisotropy. The easy axes for stable magnetization direction are 0 and 180 de‐ grees. If one cycle of magnetic field is applied in the perpendicular direction to the easy axis in ferromagnetic films, the magnetization direction changes from 0 to 90 degrees as the mag‐ netic field increases, and 90 to 180 degrees as the magnetic field decreases. And then the di‐ rection of the magnetization changes from 180 to 270 and then to 360 degrees as the reversed magnetic field increases and decreases. In that case, the AMR and PHR, which are depend‐ ent on the angle between the current and magnetization directions, can show the large hys‐ teresis loop. On the other hand, the exchange biased F/AF bilayers induce the unidirectional anisotropy, which rotates the magnetization direction from 0 to 90 and 90 to 0 degrees as the magnetic field increases and decreases, respectively. It means that the AMR and PHR signal in exchange biased F/AF bilayers show the reversal behavior and the hysteresis can be dis‐

There exists an interfacial coupling in F/AF bilayers. The hysteresis loop of the F layer, in‐ stead of being centered at zero magnetic field, is now displaced from *H* = 0 by an amount noted as the exchange field *H*ex, as if the F layer is under a biased magnetic field. Hence, this phenomenon is also known as exchange bias [28]. In such a structure the anisotropy may behave as unidirectional anisotropy. Technologically, exchange bias is of crucial importance in the field-sensing devices. An example *M*(*H*) loop of Ta(3)/NiFe(10)/IrMn(10)/Ta(3) (nm), which is usually the structure being used for fabricating a sensor, is used for this study. The center of the hysteresis loop of this bilayer, as shown in Fig. 9, is shifted from zero applied magnetic field by an amount *H*ex, the exchange bias field.

**Figure 9.** The shifted hysteresis loop in an exchange biased bilayer thin film

In a bilayer structure, the exchange coupling between the F and AF layers can easily induce the unidirectional magnetic anisotropy of the F layer. In addition, the F layer is improved to be constrained to the magnetization in coherent rotation towards the applied fields, so the sensor can prevent Barkhausen noise associated with the magnetization reversal, and im‐ proves the thermal stability [29]. Because of these advantages, a bilayer structure is a good candidate for developing sensor materials.

Bilayer has been used as PHE sensor materials by M.F. Hansen *et al*, C.G. Kim *et al*, and F.N.V. Dau *et al*. It is revealed from the literature that the sensitivity of a PHE sensor is in‐ creased with the thickness of ferromagnetic layer up to 20 nm [27].

#### *4.3.2. Spin-valves*

The spin-valve structure, as shown in Fig. 10(a), which was known as a simple embodiment of the GMR effect, typically consists of two F layers separated by a nonmagnetic conductor whose thickness is smaller than the mean-free path of electrons. The magnetic layers are un‐ coupled or weakly coupled in contrast to the generally strong AF state interaction in Fe-Crlike multilayer; thus the magnetization of F layer with uniaxial anisotropy can be rotated freely by a small applied magnetic field in the film plane, while the magnetization of other magnetic layer had unidirectional anisotropy and was pinned by exchange bias coupling from AF layer. If the relative angle between the magnetization of the two layers changes, a giant magnetoresistance change occurs.

In an illustrative demonstration of the operation of the spin-valve, the applied magnetic field is directed parallel to the exchange biased field and cycled in magnitude. The *M*(*H*) loop and the corresponding magnetoresistance curves are shown schematically in Fig. 10(b) and (c), respectively.

**Figure 10.** (a) Schematic of a typical spin-valve structure, (b) Hysteresis loop, and (c) magnetoresistance of a spin-valve sample of composition, Ta(5)/NiFe(6)/Cu(2.2)/NiFe(4)/FeMn(7)/Ta(5) (nm), at room temperature [27, 30].

The sharp magnetization reversal near zero magnetic field is due to the switching of the free magnetic layer in the presence of its weak coupling to pinned magnetic layer. The more rounded magnetization reversal at higher magnetic field is due to the switching of the pin‐ ned magnetic layer, which overcomes its exchange biased coupling to an AF layer for these fields. Therefore, it was emphasized that a spin-valve here makes use of two different ex‐ change couplings; exchange biased coupling from pinned layer to AF layer and interlayer exchange coupling between two magnetic layers, which in origin, was tentatively assigned to a Ruderman–Kittel–Kasuya–Yosida (RKKY) interaction. The relative orientations of two magnetic layers were indicated by the pairs of arrows in each region of the *M*(*H*) curve where the resistance is larger for antiparallel alignment of the two magnetic layers.

In order to optimize the spin-valve structure for high sensitivity PHE sensor, Kim's group has investigated systematically the effect of the thickness of F-pined and F-free layers (*t*<sup>f</sup> and *t*p) in the spin-valve structure Ta(5)/NiFe(*t*<sup>f</sup> )/Cu(1.2)/NiFe(*t*p)/IrMn(10)/Ta(5) (nm) with *t*<sup>f</sup> = 4, 8, 10, 12, 16, 20 nm, and *t*p = 1, 2, 6, 9, 12 nm. The results show that the sensitivity is increased linearly with *t*<sup>f</sup> and is decreased exponentially with *t*p in the investigated range. As the re‐ sult, the optimized spin-valve structure for highest sensitivity is Ta(5)/NiFe(20)/Cu(1.2)/ NiFe(1)/IrMn(10)/Ta(5) (nm). The details explanation could be found in Ref 13.

#### *4.3.3. Trilayers*

Bilayer has been used as PHE sensor materials by M.F. Hansen *et al*, C.G. Kim *et al*, and F.N.V. Dau *et al*. It is revealed from the literature that the sensitivity of a PHE sensor is in‐

The spin-valve structure, as shown in Fig. 10(a), which was known as a simple embodiment of the GMR effect, typically consists of two F layers separated by a nonmagnetic conductor whose thickness is smaller than the mean-free path of electrons. The magnetic layers are un‐ coupled or weakly coupled in contrast to the generally strong AF state interaction in Fe-Crlike multilayer; thus the magnetization of F layer with uniaxial anisotropy can be rotated freely by a small applied magnetic field in the film plane, while the magnetization of other magnetic layer had unidirectional anisotropy and was pinned by exchange bias coupling from AF layer. If the relative angle between the magnetization of the two layers changes, a

In an illustrative demonstration of the operation of the spin-valve, the applied magnetic field is directed parallel to the exchange biased field and cycled in magnitude. The *M*(*H*) loop and the corresponding magnetoresistance curves are shown schematically in Fig. 10(b)

**Figure 10.** (a) Schematic of a typical spin-valve structure, (b) Hysteresis loop, and (c) magnetoresistance of a spin-valve

The sharp magnetization reversal near zero magnetic field is due to the switching of the free magnetic layer in the presence of its weak coupling to pinned magnetic layer. The more rounded magnetization reversal at higher magnetic field is due to the switching of the pin‐

sample of composition, Ta(5)/NiFe(6)/Cu(2.2)/NiFe(4)/FeMn(7)/Ta(5) (nm), at room temperature [27, 30].

creased with the thickness of ferromagnetic layer up to 20 nm [27].

*4.3.2. Spin-valves*

212 State of the Art in Biosensors - General Aspects

and (c), respectively.

giant magnetoresistance change occurs.

The origin of interlayer coupling in F/spacer/AF trilayer structure is totally different from in‐ terlayer coupling induced in F/spacer/F multilayer thin films. The observation of F/AF ex‐ change coupling across a nonmagnetic layer by Gökemeijer *et al.*, [31] demonstrates that the exchange bias is a long-range interaction extending to several tens of Å. This coupling is not oscillatory but decays exponentially as *J* ~ exp(-*t*/*L*). The range of F/AF exchange coupling is specific to the spacer material, and thus most likely electronic in nature.

In our experiment, we choose Cu as spacer layer in the trilayer structure, Ta(3)/NiFe(10)/ Cu(0.12)/IrMn(10) (nm), because it gives a small exchange coupling with a thin Cu layer. In the sensor application, it can reduce the shunt current resulting in enhanced sensitivity. The exchange coupling of the trilayer structure, determined by the shift of the hysteresis loop in the magnetic field direction and is measured in the order of few tens of Oe, is one order smaller compared with the exchange coupling in a typical bilayer structure (in order of hun‐ dred Oe) as shown in Fig. 9. A comparison of the PHE voltages generated by the bilayer, spin-valve and trilayer structures and their corresponding sensitivities are shown in Fig. 11. Thus, it can be easily seen from the figure that the trilayer structure can improve the field sensitivity of a sensor better than those of the bilayer and spin-valve structures [32].

#### **4.4. Sensor geometry**

The performance of the sensor depends largely on its physical geometry. There were several geometries reported in the literature in the design of planar Hall sensor. Among these geo‐ metries, the cross-junction and circular geometries need special mention as they result better performance of the sensor. Thus, it is intended to present the results of the sensor for better understanding of the sensor performance when the geometries are explored in the form of cross-junction, tilted cross-junction and circular ring junction.

**Figure 11.** Comparison of the PHE performance between the bilayer, spin-valve and trilayer structures.

#### *4.4.1. Cross-junction*

In this part, we discuss the effect of the sensor size on the output voltage of a cross-junction PHE sensor.

**Figure 12.** (left) Illustration of a fabricated PHR sensor, (right) top view micrograph of a single 50 μm × 50 μm PHE sensor junction

Fig. 12 (left) shows the illustration of a fabricated PHR sensor and the Fig.12 (right) shows the SEM image of the passivated single sensor cross-junction of the size 50 *µ*m × 50 *µ*m. The terminals *a*-*b* represents the current line and *c*-*d* represents the voltage line. The unidirec‐ tional anisotropy field, *H*ex, and the uniaxial anisotropy field of the thin film are aligned par‐ allel to the long terminals *a*-*b*. Planar Hall effect (PHE) profiles were measured by the electrodes bar *c*-*d* with a sensing current of 1 mA applied through the terminals *a*-*b* and un‐ der the external magnetic fields ranging from – 50 Oe to 50 Oe applied perpendicular to the direction of the current line and in sensor plane. The induced output voltages of cross-junc‐ tions were measured by means of a Keithley 2182A Nanovoltmeter with a sensitivity of 10 nV. All these sensor characterizations were carried out at room temperature.

For studying the size effect in planar Hall sensor, cross-junctions with various sizes of *x* × 50 *µ*m2 and 50 × *x µ*m2 , (*x* = 30, 50, 100) using spin-valve structure Ta(3)/NiFe(10)/Cu(1.5)/ NiFe(2)/IrMn(10)/Ta(3) (nm) were fabricated. For estimating the free layer magnetic aniso‐ tropy of the fabricated spin-valve structure, we measured the magnetization as a function of the external magnetic field in the range of ± 80 Oe in both the easy and hard axis (refer to Fig. 4). As mentioned, the shift along the field axis of the magnetization profile in the easy axis indicates that the free NiFe layer (active layer) has very good anisotropy characteristic for further studying the PHE voltage profiles of the sensor.

The PHE voltage profiles of the fabricated sensors with various junction sizes are given in Fig. 13. Analogous to the other PHE results, the PHE voltage in all the sensor junctions ini‐ tially changes very fast and appears linear at low fields, reaches a maximal value at *H* ~ 11 Oe and finally decreases with further increase in the magnetic fields.

*4.4.1. Cross-junction*

214 State of the Art in Biosensors - General Aspects

PHE sensor.

sensor junction

In this part, we discuss the effect of the sensor size on the output voltage of a cross-junction

**Figure 11.** Comparison of the PHE performance between the bilayer, spin-valve and trilayer structures.

**Figure 12.** (left) Illustration of a fabricated PHR sensor, (right) top view micrograph of a single 50 μm × 50 μm PHE

**Figure 13.** The PHE voltage profiles of the various size sensor junctions based on the spin-valve thin film Ta(3)/ NiFe(10)/Cu(1.5)/NiFe(2)/IrMn(10)/Ta(3) (nm)

It is noteworthy that the maximum value of the PHE voltage profile is obtained at the field close to the effective uniaxial anisotropy field, *H*Keff, of the free layer. This finding was stud‐ ied systematically in a spin-valve structure and has been reported, previously [31-34]. More‐ over, it is observed in the linear response region (at the field range from -11 Oe to 11 Oe) only despite having variation in the junction size, and the slope of the PHE voltage profile remains constant. That means there is no change in the field-sensitivity when the sensor junction is varied either in length or width.

The theoretical voltage profile of the fabricated PHE sensor was also calculated with a set of following parameters: *K*u = 2×103 erg/cm3 , *M*<sup>s</sup> = 800 emu/cm3 for the NiFe, *J* = 1.8×10-3 erg/cm2

(*J* = *tM*s*H*int), *H*K = 2*K*u/*M*s, *I* = 1 mA and *V*<sup>o</sup> <sup>=</sup> *<sup>I</sup>*(*ρ*// <sup>−</sup>*ρ*⊥) *<sup>t</sup>* = 62 *µ*V and the calculated result is represented as solid line in Fig. 13. The excellent agreement between the theoretical and ex‐ perimental results confirms the point that the field-sensitivity of the PHE sensor is inde‐ pendent of the size of the cross-junction.

This result is important for the bio-applications because the sensitive detection of low bimo‐ lecular concentration is proportional to the junction size.

#### *4.4.2. Tilted cross-junction*

The idea behind the study of the tilted cross-junction is to combine some of the magnetore‐ sistive effects, such as GMR, AMR and PHE and to explore how beneficial the sensor could be in its performance [35]. Therefore, the spin-valve structure which has GMR effect causing by spin scattering of electron between two F layers through a spacer layer, AMR and PHE effects causing by the spin-orbit coupling in the F layer are the best candidates for a sensor material.

To study the tilted cross-junction bars, 100 *µ*m × 50 *µ*m, with various tilt angles of *ζ* = 0o , 4o , 8o , 10o , 30o , 45o using Ta(5)/NiFe(6)/Cu(3)/NiFe(3)/IrMn(15)/Ta(5) (nm) spin-valve structure are fabricated. The tilted cross-junction bar with a tilt angle *ζ* is shown in Fig. 14, in which the angle between the electrodes *a–b* and *c–d* is deliberately altered from 90° to 45°. The unidirectional field, *H*ex, and the uniaxial field of the thin film were aligned parallel to the long terminals *a-b*, and sensing current of 1 mA was applied through these terminals. Output voltages were meas‐ ured from the short terminals *c* and *d* at room temperature under the external magnetic fields ranging from - 45 Oe to 45 Oe applied normal to the direction of the current bar.

In general, the GMR and AMR effects could be obtained from the parallel direction to the current bar or longitudinal part while the PHE can be obtained from the transverse part of the sensor junction. Therefore, in a novel design of the sensor based on the tilted cross-junc‐ tion the longitudinal and transverse contributions could be combined together in one sen‐ sor. In this tilted junction, we observed that there is an enhancement of PHE sensitivity and better linearlity of MR longitudinal component.

In Fig. 15 we demonstrated the output voltage profiles of the sensor junctions with different tilted angles. It clearly shows an increase in amplitude of the output voltage profile with in‐ creasing tilted angle *ζ* and the upward shift of the drift voltage. In particular, a significant en‐

It is noteworthy that the maximum value of the PHE voltage profile is obtained at the field close to the effective uniaxial anisotropy field, *H*Keff, of the free layer. This finding was stud‐ ied systematically in a spin-valve structure and has been reported, previously [31-34]. More‐ over, it is observed in the linear response region (at the field range from -11 Oe to 11 Oe) only despite having variation in the junction size, and the slope of the PHE voltage profile remains constant. That means there is no change in the field-sensitivity when the sensor

The theoretical voltage profile of the fabricated PHE sensor was also calculated with a set of

represented as solid line in Fig. 13. The excellent agreement between the theoretical and ex‐ perimental results confirms the point that the field-sensitivity of the PHE sensor is inde‐

This result is important for the bio-applications because the sensitive detection of low bimo‐

The idea behind the study of the tilted cross-junction is to combine some of the magnetore‐ sistive effects, such as GMR, AMR and PHE and to explore how beneficial the sensor could be in its performance [35]. Therefore, the spin-valve structure which has GMR effect causing by spin scattering of electron between two F layers through a spacer layer, AMR and PHE effects causing by the spin-orbit coupling in the F layer are the best candidates for a sensor

To study the tilted cross-junction bars, 100 *µ*m × 50 *µ*m, with various tilt angles of *ζ* = 0o

ranging from - 45 Oe to 45 Oe applied normal to the direction of the current bar.

better linearlity of MR longitudinal component.

, 45o using Ta(5)/NiFe(6)/Cu(3)/NiFe(3)/IrMn(15)/Ta(5) (nm) spin-valve structure are

fabricated. The tilted cross-junction bar with a tilt angle *ζ* is shown in Fig. 14, in which the angle between the electrodes *a–b* and *c–d* is deliberately altered from 90° to 45°. The unidirectional field, *H*ex, and the uniaxial field of the thin film were aligned parallel to the long terminals *a-b*, and sensing current of 1 mA was applied through these terminals. Output voltages were meas‐ ured from the short terminals *c* and *d* at room temperature under the external magnetic fields

In general, the GMR and AMR effects could be obtained from the parallel direction to the current bar or longitudinal part while the PHE can be obtained from the transverse part of the sensor junction. Therefore, in a novel design of the sensor based on the tilted cross-junc‐ tion the longitudinal and transverse contributions could be combined together in one sen‐ sor. In this tilted junction, we observed that there is an enhancement of PHE sensitivity and

In Fig. 15 we demonstrated the output voltage profiles of the sensor junctions with different tilted angles. It clearly shows an increase in amplitude of the output voltage profile with in‐ creasing tilted angle *ζ* and the upward shift of the drift voltage. In particular, a significant en‐

, *M*<sup>s</sup> = 800 emu/cm3 for the NiFe, *J* = 1.8×10-3 erg/cm2

*<sup>t</sup>* = 62 *µ*V and the calculated result is

, 4o , 8o ,

erg/cm3

junction is varied either in length or width.

pendent of the size of the cross-junction.

(*J* = *tM*s*H*int), *H*K = 2*K*u/*M*s, *I* = 1 mA and *V*<sup>o</sup> <sup>=</sup> *<sup>I</sup>*(*ρ*// <sup>−</sup>*ρ*⊥)

lecular concentration is proportional to the junction size.

following parameters: *K*u = 2×103

216 State of the Art in Biosensors - General Aspects

*4.4.2. Tilted cross-junction*

material.

10o , 30o

**Figure 14.** The geometry of a tilted cross-junction. The width of current and voltage bars are 100 μm and 50 μm, re‐ spectively. The inset shows the micrograph of the cross-junction with tilt angle ζ = 10o.

hancement of sensor sensitivity by about 30% is observed when the cross-junction is tilted with an angle of 45o , and in this case, the sensitivity about 9.5 *µ*V/Oe is reached. It is also notewor‐ thy to observe a gradual change in the shape of the output voltage profile from asymmetric to symmetric which implies a corresponding increase of longitudinal MR voltage due to the in‐ crement of tilted angle in the cross-junction, *i.e*., for the first case when *ζ* = 0o , the voltage pro‐ file corresponds to the PHE only. In the other tilted cross-junctions (*ζ* ≠ 0<sup>o</sup> ), the output voltage profiles consist of the PHE, AMR and GMR components.

In order to understand the voltage contribution from each effect in a titled cross-junction quantitatively, we have performed systematic investigations on the role of the MR and PHE in the tilted cross-junction. In such case, it was noticed that the active PHE region and active MR region are from the transverse part and longitudinal part of the sensor, respectively. When the tilt angle of cross-junction increases, the length of the transverse part (*x*<sup>t</sup> in Fig. 14) decreases and the length of longitudinal part (*x*<sup>l</sup> in Fig. 14) increases accordingly.

It is observed that the PHE voltage is independent of the junction size irrespective of its change in the length or the width in previous part. Therefore, the PHE voltage component in the tilted cross-junction is always a constant. Then the transverse PHE component (corre‐ sponding to *ζ* = 0o ) is decomposed from experimental data for different tilted cross-junc‐ tions. The decomposed results are illustrated in Fig. 16 for the sensor junction with *ζ* = 10o . Clearly, a strong contribution of the longitudinal MR component is evidenced. However, the PHE dominates good linearity and high sensitivity at low magnetic fields.

**Figure 15.** The experimental and theoretical voltage profiles of cross-junctions with different tilt angles of 0o, 4o, 8o, 10o, 30o, 45o.

Applying the above mentioned decomposition procedure for all investigated sensor junc‐ tions, one can derive the values of the drift (minimal) voltage (*V*MRmin), the MR voltage (or the MR voltage change in external magnetic fields) (*ΔV*MR = (*V*MRmax − *V*MRmin)) and the per‐ centage of voltage change of the longitudinal MR voltage profile (Δ*V*MR/*V*MRmin ).

The results are listed in Table 2. Note that, *V*MRmin and Δ*V*MR increases as the tilted angle in‐ creases and thus the Δ*V*MR enhances the total output voltage profiles.


**Table 2.** The sensor sensitivity (*S*) and values of the minimal voltage (*V*MRmin), MR voltage change in the applied fields (Δ*V*MR), relative voltage change of the longitudinal MR voltage profile (Δ*V*MR/*V*MRmin) of different tilted cross-junctions

**Figure 16.** PHE and MR voltage components are decomposed from the experimental voltage profile of the sensor junction with the tilt angle ζ = 10o (a) at the field range of ±45 Oe and (b) at the field ranging from 0 to 8 Oe to illus‐ trate the linearity of the sensor. In this figure, the origin of the PHE voltage component is adjusted to the minimum voltage of the MR components.

Applying the above mentioned decomposition procedure for all investigated sensor junc‐ tions, one can derive the values of the drift (minimal) voltage (*V*MRmin), the MR voltage (or the MR voltage change in external magnetic fields) (*ΔV*MR = (*V*MRmax − *V*MRmin)) and the per‐

**Figure 15.** The experimental and theoretical voltage profiles of cross-junctions with different tilt angles of 0o, 4o, 8o,

The results are listed in Table 2. Note that, *V*MRmin and Δ*V*MR increases as the tilted angle in‐

**ζ (o)** *S* **(μV/Oe)** *V***MRmin (μV) Δ***V***MR(μV) Δ***V***MR/***V***MRmin×100 (%)**

**Table 2.** The sensor sensitivity (*S*) and values of the minimal voltage (*V*MRmin), MR voltage change in the applied fields (Δ*V*MR), relative voltage change of the longitudinal MR voltage profile (Δ*V*MR/*V*MRmin) of different tilted cross-junctions

0 7.4 - - -

 7.5 1799 11.0 0.61 7.6 3877 24.0 0.62 7.7 5021 30.5 0.61 9.1 15752 94.5 0.60 9.5 22385 136.0 0.60

centage of voltage change of the longitudinal MR voltage profile (Δ*V*MR/*V*MRmin ).

creases and thus the Δ*V*MR enhances the total output voltage profiles.

10o, 30o, 45o.

218 State of the Art in Biosensors - General Aspects

Generally, the longitudinal MR component was contributed from AMR and GMR effects [25,36]. The total output voltage induced from these effects satisfies the following equation [33]:

$$V\_{\rm MR} = I \times R\_{\rm s} \times \sin \zeta \times (1 + 0.5 \times \text{GMR} \times (1 - \cos(\theta - \theta\_{\rm p})) + \text{AMR} \times \cos^2 \theta) \tag{8}$$

In this equation, *θ*p is the angle between the magnetization direction of the F-pinned layer and the easy axis of F-free layer, and the drift voltage term (*I* ×*R*s×sin*ζ*) was modified from Ref. [33] in accordance with the investigated sensor junctions, because it depends on the length of the sensor junction. The increased length of the active region of the MR compo‐ nents depends on the sinusoidal function of tilt angle *ζ*. From Eq. (8), if the sensor junction has no tilt angle, *V*MR is zero, in which case the sensor has only the PHE contribution. When the junction starts to tilt, the MR components contribute to the total sensor output voltage. The drift voltage and then the MR voltage depend on sinusoidal function of the tilt angle (~ *I* ×*R*s×sin*ζ*) [37].

The decomposed MR voltage profiles can be described with values of the sheet resistance *R*<sup>s</sup> = 28.5 Ω, GMR = 1.8 % and AMR = 0.4 %. Other parameters are kept the same as for the PHE voltage profile calculations. The trend of the calculated results of representative sensor junc‐ tion with *ζ* = 10o is presented by the red solid line in Fig. 16.

Finally, the total output voltage profiles of the tilted junctions are calculated by combining both the PHE and MR components represented in Eq. (8). The results are shown by solid lines in Fig. 16, where the calculated drift voltages are adjusted to the experimental drift vol‐ tages. It is clearly evident that a rather good consistence between the experimental and the calculated data is obtained. Thus, the tilted cross-junction exhibited not only a better sensi‐ tivity in comparison with individual PHR sensor but also a better linearity compared with individual MR sensor.

#### *4.4.3. Ring junction*

The idea to develop the sensor based on a ring is to combine both the PHE and AMR com‐ ponents in one ring junction [38]; thus, the output voltage of the sensor can be enhanced. In the following, the role of the output signal as well as the optimization results will be dis‐ cussed.

Firstly, for studying the role of the signal in the ring junction, we design the ring with differ‐ ent configurations. These rings have the same diameter of 300 *µ*m and the same width of 20 *µ*m. The illustration schemes and the tested results corresponding to each configuration us‐ ing exchange biased structure Ta(3)/NiFe(50)/IrMn(10)/Ta(3) (nm) are given in Fig. 17

**Figure 17.** Designed rings with different Au electrode configurations and their corresponding output voltage profiles for the case of AMR arms (a), PHR elements (b) and a full ring (c) in the exchange biased structures shown in the inset.

It is evident from Fig. 17 that the signal change in the case of a full ring (350 *µ*V) is close to the sum of the signals in the cases of a AMR arms (300 *µ*V) and PHR elements (50 *µ*V). Based on these obtained results we assume that, in the full ring junction, there exist two components AMR (Fig. 17(a)) and PHR (Fig. 17(b)).

#### **4.5. Hybrid AMR and PHR ring sensor – Optimized performance**

nents depends on the sinusoidal function of tilt angle *ζ*. From Eq. (8), if the sensor junction has no tilt angle, *V*MR is zero, in which case the sensor has only the PHE contribution. When the junction starts to tilt, the MR components contribute to the total sensor output voltage. The drift voltage and then the MR voltage depend on sinusoidal function of the tilt angle (~

The decomposed MR voltage profiles can be described with values of the sheet resistance *R*<sup>s</sup> = 28.5 Ω, GMR = 1.8 % and AMR = 0.4 %. Other parameters are kept the same as for the PHE voltage profile calculations. The trend of the calculated results of representative sensor junc‐

Finally, the total output voltage profiles of the tilted junctions are calculated by combining both the PHE and MR components represented in Eq. (8). The results are shown by solid lines in Fig. 16, where the calculated drift voltages are adjusted to the experimental drift vol‐ tages. It is clearly evident that a rather good consistence between the experimental and the calculated data is obtained. Thus, the tilted cross-junction exhibited not only a better sensi‐ tivity in comparison with individual PHR sensor but also a better linearity compared with

The idea to develop the sensor based on a ring is to combine both the PHE and AMR com‐ ponents in one ring junction [38]; thus, the output voltage of the sensor can be enhanced. In the following, the role of the output signal as well as the optimization results will be dis‐

Firstly, for studying the role of the signal in the ring junction, we design the ring with differ‐ ent configurations. These rings have the same diameter of 300 *µ*m and the same width of 20 *µ*m. The illustration schemes and the tested results corresponding to each configuration us‐

**Figure 17.** Designed rings with different Au electrode configurations and their corresponding output voltage profiles for the case of AMR arms (a), PHR elements (b) and a full ring (c) in the exchange biased structures shown in the inset.

ing exchange biased structure Ta(3)/NiFe(50)/IrMn(10)/Ta(3) (nm) are given in Fig. 17

is presented by the red solid line in Fig. 16.

*I* ×*R*s×sin*ζ*) [37].

220 State of the Art in Biosensors - General Aspects

tion with *ζ* = 10o

individual MR sensor.

*4.4.3. Ring junction*

cussed.

In order to optimize the performance of the sensor using a ring junction, efforts were made to design an hybrid AMR and PHR ring sensor. It is known that the maximum voltage of the AMR and PHR voltages in the ring can be calculated using:

$$\begin{aligned} V\_{\text{AMRo}} &= \frac{r}{o\nu} \frac{I \Delta \rho}{t} \\ V\_{\text{PHRo}} &= \frac{I \Delta \rho}{t} \end{aligned} \tag{9}$$

where *r* and *ω* are the radius and the width of the ring junction, *I* is the applied current, *t* is the thickness of the sensor material.

It is clear from the above that the PHR component is always constant while the AMR com‐ ponent increases linearity with the increase in *r*/*ω* ratio. It is noteworthy that when *r*/*ω* = 1 the AMR voltage is equal to the PHR voltage, in which case the ring becomes the full disk. By fixing *I*Δ*ρ*/*t* = 1, the output signal of the sensor is calculated, and the result is shown in Fig. 18.

The results in Fig. 18 ensure that the higher the *r*/*ω* ratio the larger the output voltage of the ring. To increase the *r*/*ω* ratio, basically, we can increase the radius, *r*, or reduce the width, *ω*, of the ring. However, for integrating with the other devices using present silicon technol‐ ogy, the ring size must be restrained to a certain limit. We assume that the ring size should be limited to about 300 *µ*m, corresponding to the radius of *r* = 150 *µ*m. The second problem that must be considered for optimizing the sensor performance is the width of the ring; the thinner the width, the higher the resistance, therefore, the higher output voltage can be ach‐ ieved. But the width can not be made so thin, because the heat generated during the work‐ ing time will burn the sensor junction. By considering these parameters, the optimized ring will have the radius of 150 *µ*m and the width of 5 *µ*m (*r*/*ω* = 30).

The results of the sensitivity versus *r*/*ω* of the ring sensor using bilayer and trilayer struc‐ tures (Ta(5)/Ru(1)/NiCo(10)/IrMn(10)Ru(1)/Ta(5) and Ta(3)/NiFe(10)/Cu(0.12)/IrMn(10)/ Ta(3) (nm) are illustrated in Fig. 19. It is abundantly clear from the figure that the ring sensor using trilayer structure has higher sensitivity compared to that of bilayer structure. So the best performance of the ring is obtained using the trilayer structure, in which case the sensi‐ tivity is about 340 *µ*V/Oe, and this is a much improved sensitivity compared to the sensitivi‐ ty of an AMR or a PHR sensor (normally, the sensitivity of PHR sensor < 15 *µ*V/Oe).

**Figure 18.** The calculation and experimental results of PHR and AMR output voltage components versus r/ω ratio of the ring. The insets show schematics of a ring junction with defined r and ω, and a representative PHE voltage profile of ring sensor for *r* = 150 μm, ω = 20 μm.

**Figure 19.** Experimental results of the sensitivity versus *r*/ω ratio of the rings using a Ta(5)/Ru(1)/NiCo(10)/ IrMn(10)Ru(1)/Ta(5) (nm) bilayer thin film and trilayer thin film Ta(3)/NiFe(10)/Cu(0.12)/IrMn(10) /Ta(3) (nm).

It can be summarized from the above that the systematic investigations on the ring junc‐ tions revealed that there exist both PHR and AMR voltages contribute to the output volt‐ age profile. The PHR voltage component is always kept constant when varying the size of the ring, while the AMR voltage component linearity increase due to the increasing the *r*/*ω* ratio of the ring. For practical and application aspects, the ring must be opti‐ mized both in terms of its size and performance. The optimized radius and the width of the junction are 150 *µ*m and 5 *µ*m, respectively. By using the trilayer structure, the best performance of the sensor is obtained. In such case, the highest sensitivity sensor is about 340 *µ*V/Oe. This hybrid sensor is very much improved in the sensitivity compared to an AMR or a PHR sensor.
