The Impact of Bending Stress on the Performance of Giant Magneto-Impedance (GMI) Magnetic Sensors

Transcription

1 sensors Article The Impact of Bending Stress on the Performance of Giant Magneto-Impedance (GMI) Magnetic Sensors Julie Nabias, Aktham Asfour * and Jean-Paul Yonnet Université Grenoble Alpes, CNRS, Grenoble INP (Institute of Engineering Université Grenoble Alpes), G2Elab, F-38 Grenoble, France; Julie.Nabias@g2elab.grenoble-inp.fr (J.N.); Jean-Paul.Yonnet@g2elab.grenoble-inp.fr (J.-P.Y.) * Correspondence: aktham.asfour@g2elab.grenoble-inp.fr; Tel.: Academic Editors: Nicole Jaffrezic-Renault and Gaelle Lissorgues Received: 9 December 216; Accepted: 1 March 217; Published: 2 March 217 Abstract: The flexibility of amorphous Giant Magneto-Impedance (GMI) micro wires makes them easy to use in several magnetic field sensing applications, such as electrical current sensing, where they need to be deformed in order to be aligned with the measured field. The present paper deals with the bending impact, as a parameter of influence of the sensor, on the GMI effect in 1 µm Co-rich amorphous wires. Changes in the values of key parameters associated with the GMI effect have been investigated under bending stress. These parameters included the GMI ratio, the intrinsic sensitivity, and the offset at a given bias field. The experimental results have shown that bending the wire resulted in a reduction of GMI ratio and sensitivity. The bending also induced a net change in the offset for the considered bending curvature and the set of used excitation parameters (1 MHz, 1 ma). Furthermore, the field of the maximum impedance, which is generally related to the anisotropy field of the wire, was increased. The reversibility and the repeatability of the bending effect were also evaluated by applying repetitive bending stresses. The observations have actually shown that the behavior of the wire under the bending stress was roughly reversible and repetitive. Keywords: Giant Magneto-Impedance (GMI); amorphous wire; bending stress; flexible 1. Introduction The use of Giant Magneto-Impedance (GMI) wires as sensing elements for electrical current sensors in real industrial environments raises important issues related to the parameters that influence the sensor. These parameters can dramatically affect the accuracy and reliability of the measurement. They may include but are not limited to the temperature and the surrounding magnetic fields produced, for example, by other conductors in close proximity to the conductor of interest. Mechanical effects can also be encountered in a number of structures of GMI-based current sensors. In fact, in some situations, the GMI wire has to be deformed since it needs to be aligned with the magnetic field produced by the conductor carrying the measured electrical current. Indeed, this is typically the case in some prototypes of current sensors where the sensor has a toroidal structure [1 7], unlike other prototypes that do not involve deformation [8,9]. In the case of toroidal configuration, some studies have already been conducted to evaluate the impact of the deformation of the amorphous wire on the GMI effect [1]. The effects of tensile [1 17] and torsional [18,19] stresses have been intensively investigated in amorphous microwires with low (to vanishing or slightly negative) magnetostriction. Stresses induced during certain fabrication processes, such as cold drawing, can also affect the domain structure of the wires and influence the GMI ratio and field sensitivity [2]. By far, the bending stress effect in GMI amorphous wires is actually less known. Nevertheless, the investigation of this effect is particularly important in applications such as current sensors. In fact, Sensors 217, 17, 64; doi:1.339/s

2 Sensors 217, 17, 64 2 of 1 Sensors 216, 16, x FOR PEER REVIEW 2 of 4 44 By far, bending stress effect in GMI amorphous wires is actually less known. Nevertheless, the it is important to ensure that the potential variations of the intrinsic relevant quantities of the GMI 45 investigation of this effect is particularly important in applications like current sensors. Actually, it is curve due to 46 bending important stress to ensure will that not the affect potential the variations final of response the intrinsic ofrelevant the sensor. quantities Or of at the least, GMI curve the knowledge of the impact47 of the due bending to the bending on stress thesewould relevant not affect quantities the final response has to of the besensor. developed Or at least, sothe asknowledge to allow for adequate 48 of the impact of the bending on these relevant quantities has to be developed so as to allow for solutions to49 minimizing adequate solutions this impact. to minimize this impact. In a sensor 5 application, In a sensor application, the relevant the relevant quantities to be be considered include include the offset resulting the offset from resulting from 51 the field biasing, the intrinsic sensitivity at the bias point as well as the GMI ratio. the field biasing, 52 the These intrinsic quantities sensitivity are illustrated at the by bias Fig. 1 point, which and presents the the GMI typical ratio. nonlinear GMI These53 quantities characteristics are illustrated (Modulus of the in Figure impedance, 1, Z(H), whichas presents a function of typical the magnetic nonlinear field H) GMI for an characteristics 54 amorphous wire. (modulus of the impedance, Z(H), as a function of the magnetic field H) for an amorphous wire. 55 Modulus of Impedance Z, ( ) ( ) (slope) = Bias field 56 Figure 1. A typical GMI curve, Z(H) Figure 1. A typical Giant Magneto-Impedance (GMI) curve, Z(H). 57 As it is well-known, in a classical use of a GMI wire for developing a linear magnetic sensor, the 58 wire has to be biased by a magnetic field,, in the region that exhibits an almost linear behavior. As is 59 well-known, This field biasing in the gives classical rise to an use offset ofvoltage a GMI at the wire final for sensor developing output when athe linear measured sensor, the wire must be biased 6 magnetic by a magnetic field is zero, since field, the Houtput voltage is directly proportional to the value of the impedance b, in the region that exhibits near-linear behavior. This field 61 at the bias field ( ( ) in Fig. 1). Electronic canceling of the offset voltage is usually employed. biasing gives 62 rise However, to anthis offset canceling voltage should at be the effective final only sensor if the value output of (when ) does the not measured change under the magnetic field is zero, since63 the output parameters voltage of influence. is directly proportional to the value of the impedance at the bias field 64 In addition to linearity considerations, the bias point is also chosen to obtain a maximum of ( Z(H b ) in65 Figure 1). Electronic canceling of the offset voltage is usually employed. However, this sensitivity. At the bias point, the intrinsic sensitivity of the sensor is defined by: canceling should 66 be effective only if the value of Z(H b ) does not change under the parameters of influence. ( )= (1) In addition to linearity considerations, the bias point is also chosen to obtain a maximum sensitivity. 67 At the bias point, the intrinsic sensitivity of the sensor is defined by 68 This quantity will dramatically determine the final sensitivity of the sensor in 69 open-loop and, if it is not high enough, it could also impact the closed-loop sensitivity. 7 S(H b ) = Z H. (1) 71 A third quantity to be considered is the GMI ratio, Z/Z, H=Hb defined as: This quantity Z H will dramatically determine the final sensitivity of the sensor in open-loop H=Hb operation and, if it is not high enough, it could also impact the sensitivity in closed-loop operation. A third quantity to be considered is the GMI ratio, Z/Z, defined as Z Z (%) = Z(H) Z(H max) Z(H max ) 1 (2) where a commonly used value for H max is the saturation field of the magnetic material. In practice, H max is generally the maximum field available in a given setup. The GMI ratio obviously contributes to fixing the total dynamic range of the open-loop sensor. In fact, a higher GMI ratio allows more excursion of the measured field around the bias point. Under the parameters of influence, the offset, the intrinsic sensitivity, S, and the GMI ratio are potentially subject to changes that can largely degrade the sensor s performance and reliability.

3 Sensors 217, 17, 64 3 of 1 Sensors 217, 17, 64 3 of 1 According to to the requirements of of each application, one oneor orseveral of of these these quantities quantities have have to to be be optimized. optimized. In In this this paper, an an experimental study is is conducted to to evaluate the impact of of the bending stress, as as an an influence parameter, on on the the offset, the the intrinsic sensitivity, and and the the GMI GMI ratio ratio of of an an amorphous wire. wire. A full full description of of the the measurement setup and and experimental conditions are are given. The first first obtained results results are are illustrated illustrated and and discussed. discussed. Tests Tests of theof repeatability the repeatability and reversibility and reversibility of the bending of the stress bending effect stress are effect also included. are also included. 2. Materials and Methods Figure 2 shows 2 shows the developed the developed experimental experimental setup. The setup. usedthe amorphous used amorphous Co-rich wires Co-rich (Co-Fe-Si-B) wires with (Co-Fe-Si-B) a 1 µmwith diameter a 1 were µm diameter from Unitika were Ltd., from obtained Unitika with Ltd. (Düsseldorf, the in-rotating-water Germany), spinning obtained process. with No the additional in-rotating-water treatments spinning were performed process. No onadditional these wires. treatments were performed on these wires. Current source (DC) Impedance Measurement GMI wire (in a flexible sheath) Bending stress applied with a minimum radius of curvature of 1 cm Figure Figure The The experimental setup. setup. The The GMI GMI wire wire was was 9 9 mm mm in in length length and and excited excited by by an an AC AC current current with with a 1 ma ma amplitude and and a 1 MHz MHz frequency. In In this this figure, figure, the the wire wire is is in in aa bent bent position. position. For all experiments, wire 9 mm long was cut and placed inside flexible envelope (sheath). For all experiments, a wire 9 mm long was cut and placed inside a flexible envelope (sheath). The extremities of the wire were soldered to copper wires for electrical connections and The extremities of the wire were soldered to copper wires for electrical connections and measurement measurement purposes. The external magnetic field,, was applied to the wire using a 27-turn coil purposes. The external magnetic field, H, was applied to the wire using a 27-turn coil placed around placed around the flexible sheath. The maximum available field was 1 ka/m. the flexible sheath. The maximum available field H In all the experiments, the impedance was measured max was 1 ka/m. when the GMI wire was supplied by a In all the experiments, the impedance was measured when the GMI wire was supplied by a high high frequency (HF) current,, with a 1 MHz frequency and a 1 ma amplitude. These values were frequency (HF) current, i only justified by the requirements ac, with a 1 MHz frequency and a 1 ma amplitude. These values were only of our application, where the sensor s energy consumption and justified by the requirements of our application, where the sensor s energy consumption and the the simplicity of implementation of the conditioning electronics are issues. Additionally, with simplicity of implementation of the conditioning electronics are issues. Additionally, with relatively relatively low AC currents, the nonlinear effects of the GMI (nonlinearity between the current and low AC currents, the nonlinear effects of the GMI (nonlinearity between the current and the voltage the voltage across the wire) were avoided [21]. across the wire) were avoided [21]. The bending stress was simultaneously applied to the GMI wire, sheath, and coil. The The bending stress was simultaneously applied to the GMI wire, sheath, and coil. The minimum minimum measured a radius of curvature, which is representative of the applied strain, was about measured a radius of curvature, which is representative of the applied strain, was about 1 mm. 1 mm. This yields a calculated stress of approximately 8 MPa in tension and compression along This yields a calculated stress of approximately 8 MPa in tension and compression along the wire. the wire. 3. Results and Discussion 3. Results and Discussion Figure 3 shows the change of Z(H) and Z/Z with the applied bending stress. Figure 3 shows the change of ( ) and Z/Z with the applied bending stress. One could note that the Z(H) curves of Figure 3 are slightly asymmetric, in both straight and One could note that the ( ) curves of Figure 3 are slightly asymmetric, in both straight and bent positions. This intrinsic asymmetry, which has been intensively investigated in the literature, bent positions. This intrinsic asymmetry, which has been intensively investigated in the literature, could have several causes [22]. While the study of the asymmetry is well beyond the scope of this could have several causes [22]. While the study of the asymmetry is well beyond the scope of this paper, it seems that its origin is related to the hysteresis of the GMI material. The curves of Figure 3 paper, it seems that its origin is related to the hysteresis of the GMI material. The curves of Figure 3 have actually been plotted in the case of an increasing magnetic field (the asymmetry is reversed for

4 Sensors 216, 16, x FOR PEER REVIEW 4 of 4 12 currents, the nonlinear effects of the GMI (nonlinearity between the current and the voltage across 13 the wire) are avoided [21]. Sensors Sensors 217, 217, 14 17, 17, The bending stress was simultaneously applied to the GMI wire, sheath and coil. The minimum 4 of of measured radius of curvature, which is representative of the applied strain, was about 1 mm. This 16 yields a calculated stress of approximately 8 MPa in tension and compression along the wire. a decreasing field). The asymmetry does not, however, withdraw the validity and the generality of have actually been plotted in the case of an increasing magnetic field (the asymmetry is reversed for a the obtained 17 results. 3. Results and Discussion decreasing field). The asymmetry does not, however, withdraw the validity and the generality of the 18 Figure 3 shows the change of Z(H) and Z/Z with the applied bending stress. obtained results = Bias field = Bias field #$% 111 Figure 3. Change of Z(H) ( ) and Z/Z with the bending stress. Figure 3. Change of Z(H) and Z/Z with the bending stress. 112 One could note that the () curves of Fig.3 are slightly asymmetric, in both straight and GMI GMIRatio 113 and and bent Anisotropy positions. This Field Field intrinsic asymmetry, which has been intensively investigated in literature, could 114 have several causes [22]. While the study of the asymmetry is well beyond the scope of this paper, it For For simplicity 115 seems of of that illustration its origin is and related and explanation, explanation, to the hysteresis we we of are the are GMI interested, interested, material. The without without curves of a loss Fig. a loss 3 of have generality, of generality, in the in the 116 actually been plotted in the case of increasing magnetic field (the asymmetry is reversed for a positive positive fields fields 117 region region decreasing only. only. field). The asymmetry does however not withdraw the validity and the generality of the When When118 the the used obtained used sample results. sample was was not bent not (straight bent (straight position), position), a typicala maximum typical maximum GMI ratio, ( Z/Z) GMI ratio, max, ( / ) of more than, of % more 3.1. GMI was than ratio obtained. 22% was and anisotropy The obtained. field field ofthe the field maximum of the impedance maximum and impedance the GMI and ratio, the HGMI peak, ratio, which is related, which 12 to For the is related simplicity anisotropy to the of illustration field anisotropy and H k, explanation, wasfield nearly we are equal, was interested, to nearly 64and A/m. equal to 64 A/m. without loss of generality, In In the 121 bent bent in position, position, the positive afields a decrease decrease region only. in in the the ratio ratio ( Z/Z) ( / ) max to less to less than than 16% 16% was observed was observed with Hwith peak, 122 When the used sample was not bent (straight position), a typical maximum GMI ratio, increasing, increasing to more to than more 1 than A/m. 1 A/m. 123 ( /), of more than 22 % was obtained. The field of the maximum of impedance and GMI This result 124 ratio, seems #$% to to, be which be in in is agreement related agreement to the with anisotropy with the the field one one that %, was that could nearly could be equal obtained to be 64 obtained A/m. with awith tensile a tensile stress [1 17]. stress [1 17]. 125 In the bent position, a decrease of the ratio ( /) less than 16 % was observed with The usedthe amorphous used amorphous wires have awires nearlyhave zeroa magnetostriction nearly zero magnetostriction [23]. Nevertheless, [23]. thenevertheless, domain structure the 126 #$% increasing to more than 1 A/m. domain of these structure wires 127 is often these considered wires is to often be similar considered to that to of be negative similar to magnetostrictive that of negative wires magnetostrictive [22]. In such wires a case, [22]. thein quenched-in such a case, stresses the quenched-in due to fabrication stresses due by rapid to fabrication solidification by rapid maysolidification cause the surface may cause anisotropy the surface to be circular anisotropy and the to be inner circular anisotropy and the toinner be perpendicular anisotropy to tobe theperpendicular wire axis, thereby to the leading wire axis, to thethereby formation leading of a to specific the formation domain structure, of a specific which domain consists structure, of outer which shellconsists circularof domains outer shell and circular inner core domains roughlyand axial inner domains core [22]. roughly Thisaxial means domains that the [22]. anisotropy This means is circumferential that the anisotropy in the outer is circumferential shell of the wire. in the outer shell of the wire. In Ina first approximation, a bending stress could be beassumed to tobe bea combination of oftensile stress along the outer fiber and compressive stress along the the inner inner fiber fiber of of the the wire, wire, as as illustrated by by Figure Figure There There is is no no deformation deformation of of the the fiber fiber along along the the neutral neutral axis. axis. Compression y Tensile z x Figure 4. Representation of a bending stress. Figure 4. Representation of bending stress.

5 Sensors 217, 17, 64 5 of 1 In fact, the amplitude of the strain, σ, applied on the section of the wire can be calculated using the following formula: σ = E y ρ where E is the Young s modulus of the wire (16, kg/mm 2 in our case), y is the distance of a given fiber to the neutral axis, and ρ is the radius of curvature. According to this formula, σ is positive when the distance y is counted negatively (outer fibers of the wire), and σ is negative when the distance y is counted positively (inner fibers of the wire). Therefore, the strain should theoretically be compensated in the total volume of the wire, as there is the exact same amount of tensile and compressive stress along the wire. However, this simplified model should be considered cautiously, as it appears from our experiments that there is a change in Z(H) with the applied bending stress. This would actually suggest that the effect of the compressive stress is not exactly opposed to the effect of the tensile stress. It seems that the two effects are mixed in some manner, resulting in a dominating tensile stress effect. The effect of the tensile stress has been intensively studied [1 17]. For a negative magnetostrictive wire, a tensile stress applied along the longitudinal axis will inhibit the orientation of the domains along this same axis, giving rise to a favorable domain orientation along the radial and azimuthal axes. This could cause an enhancement of the GMI ratio at low frequencies. However, at high frequencies, the wall mobility is extremely reduced, resulting in a very strong decrease in magnetic permeability [12]. The transverse anisotropy caused by the tensile stress in amorphous wires with vanishing (or slightly negative) magnetostriction reduces the GMI change when the value of the applied stress is large enough. For smaller stresses, the circumferential domain structure of the wires is refined, and the magnitude of the GMI effect can increase [15]. Many works have demonstrated that a commonly observed aspect, resulting from tensile stress, is that the field of the maximum impedance, H peak (which is related to the anisotropy field H k ), is shifted towards higher values [1 17]. This shift is generally accompanied by a decrease in the value of maximum impedance and consequently of the GMI ratio. In our bending stress experiments, all of these observations were verified. While the current study deals only with amorphous wires, it could be of great value to perform a similar investigation for the case of nanocrystalline materials in the future. These materials are usually processed with a controlled annealing of an amorphous precursor. Due to their nanometric grain sizes, these alloys exhibit outstanding soft magnetic proprieties, which make them good candidates for the occurrence of the GMI effect. In Fe-rich alloys, nanocrystallization produced by annealing largely enhances the magnitude of the GMI effect [24]. The same behavior as in amorphous alloys could appear in nanocrystalline alloys with an applied tensile stress, provided that they have the same sign of magnetostriction constant [25]. The tensile stress would enhance the transverse anisotropy of the material. The effect could be larger in amplitude, as the transverse anisotropy induced by tensile stress is larger in nanocrystalline alloys than in amorphous alloys [26]. Nevertheless, nanocrystalline materials are less adapted than their amorphous counterparts for our sensing application, as they exhibit poor mechanical properties. The behavior of the GMI amorphous wires under compressive stress is, to our knowledge, not frequently investigated, despite a very recent experimental study in ribbons [27]. One might, however, argue that, as the magnetostriction constant is negative, a compressive stress applied along the longitudinal axis of the wire will induce a favorable orientation of the domains along this axis, reducing consequently the GMI effect, as the anisotropy is no more circumferential. More quantitative analysis of the wire s behavior under compressive stress is still, however, necessary. The effect of bending or flexion stress on the GMI is even less known than the tensile and compressive stresses. In our results, and for the used length and bending curvature of the wire, it seems that this effect gives similar results to those of tensile stress. This observation should be considered cautiously, since the observed effect could be actually dependent on the bending curvature (3)

6 Sensors 217, 17, 64 6 of 1 and on the parameters (frequency and amplitude) of the HF excitation current i ac. According to this curvature, some competition could exist between tensile and compressive stress Sensitivity, Offset, and Reversibility The change of the GMI ratio and the anisotropy under a bending stress is important to allow for a fundamental and rigorous understanding of the involved physical phenomena. However, the change in the local sensitivity at a given bias field, the change of the offset, and the reversibility are, from a practical point of view, crucial criteria in a real implementation of a GMI sensor involving bending. The quantification of these parameters is therefore essential. In the absence of a bending stress, the maximum sensitivity for the used sample was achieved at a field of about 32 A/m (remember that we consider, and without loss of generality, the positive fields region only). In a sensor realization, this is generally chosen as a bias field (H b in Figures 3 and 5). The associated sensitivity Sensors 216, is16, noted x FOR PEER S(H REVIEW b ). 7 of 4 = Bias field () Figure 5. Change of the normalized sensitivity, under bending stress. The sensitivity was 4( ) Z(H) 215 normalized with respect to the maximum of S(H sensitivity, b ) H ( ), obtained in a straight position of the 216 wire. 1 Figure 5. Change of the normalized sensitivity, under bending stress. The sensitivity was normalized with respect to the maximum sensitivity, S(H b ), obtained in a straight position of the wire Then, when the wire is bent, the maximum sensitivity (in the region of positive fields) decreases to about 58 % of S(H, ). Moreover, this new maximum is obtained at a higher field of about 45 A/m (Fig. 5). This obviously means that, for the design Hof GMI-based sensors, the value of the initially chosen bias field H, (in the straight position of the wire) will not allow a maximum of sensitivity in bent positions. Actually, at this bias field H,, the sensitivity in our sample was reduced to about 52 % of its initial value. The reduction of the maximum of sensitivity was roughly consistent with the change of the field of maximum impedance, H 5678, and the value of the impedance at this same field Z9H 5678 :. S(H Actually, a rough estimation of the maximum of sensitivity b ) = 1 Z(H) S(H could be given b ) H by Z9H 5678 :/H 5678 [28]. In our experiments, this ratio was decreased in the bent position to about 51 % of its value in a straight position. At least the general trend between the changes of this ratio and the sensitivity is consistent. In the applications of GMI sensor that involve repetitive bendings, the quantification of the loss of sensitivity is then required. When the sensor is operating in closed-loop, it is important to ensure that the lower sensitivity, obtained in the bent position, will still be high so that the final sensor response would be only dependent on the parameters of the feedback (which could be developer-defined) [6]. In our experiments, the sensitivity S(H) = Z(H) was obtained by differentiation of the Z(H) curve of Figure 3. For simplicity reasons, we consider a normalized sensitivity with respect to the maximum sensitivity obtained in a straight position of the wire at H b (i.e., S(H b )). This normalized sensitivity is then defined by S(H) and is plotted in Figure 5 for both straight and bent positions of the wire. Then, when the wire is bent, the maximum sensitivity (in the region of positive fields) decreases to about 58% of S(H b ). Moreover, this new maximum is obtained at a higher field of about 45 A/m (Figure 5). This obviously means that, for the design of GMI-based sensors, the value of the initially chosen bias field H b (in the straight position of the wire) will not allow a maximum of sensitivity in bent positions. In fact, at this bias field H b, the sensitivity in our sample was reduced to about 52% of Another practical issue to be considered is the change of the offset. At the bias field, H,, the its initial value. impedance Z(H, ) was roughly equal to 64 Ω. Under bending, the change of Z(H) at this same The reduction of field the was maximum as high as 38% with sensitivity Z(H, ) = 4 Ω. was The instability roughly of the consistent offset actually with a major the issue in change a practical realization. An efficient offset cancelling device has to be incorporated. Solutions could ( of the ) field of include the use of a difference amplifier [5] or a symmetrical AC bias field [4]. Each of these. maximum impedance, H peak, and the value of the impedance at this same field Z H peak In fact, techniques has some limitations which will not be discussed in this paper. ( ) /Hpeak a rough estimation of the maximum sensitivity could be given by Z H peak [28]. In our experiments, this ratio was decreased in the bent position to about 51% of its value in a straight position. The general trend between the changes of this ratio and the sensitivity is at least consistent. In the applications of the GMI sensor that involve repetitive bendings, the quantification of the loss of sensitivity is then required. When the sensor is operating in a closed loop, it is important to ensure that the lower sensitivity, obtained in the bent position, will still be high so that the

7 Sensors 217, 17, 64 7 of 1 final sensor response will be dependent only on the parameters of the feedback (which could be developer-defined) [6]. Another practical issue to be considered is the change of the offset. At the bias field, H b, the impedance Z(H b ) was roughly equal to 64 Ω. Under bending, the change of Z(H) at this same field was as high as 38% with Z(H b ) = 4 Ω. The instability of the offset is actually a major issue in a practical realization. An efficient offset canceling device has to be incorporated. Solutions could include Sensors the use 217, of17, a64 difference amplifier [5] or a symmetrical AC bias field [4]. Each of these 7 of techniques 1 has limitations that will not be discussed in this paper. sensor response will be dependent only on the parameters of the feedback (which could be Alldeveloper-defined) the experimental [6]. results are summarized in Table 1. Another practical issue to be considered is the change of the offset. At the bias field,, the Table impedance 1. Summary ( ) of was the roughly results equal obtained to 64 Ω. with Under an bending, amorphous the change Co-richof wire, ( ) with at this a same bending stress field applied. was as high as 38% with ( ) = 4 Ω. The instability of the offset is actually a major issue in a practical realization. An efficient offset canceling device has to be incorporated. Solutions could include the use of a difference amplifier [5] or a With symmetrical a BendingAC Stress bias Applied field [4]. Each of these With No Bending Physical techniques Quantities has limitations that will not be discussed (Maximum in this paper. Strain Applied along Relative Change (%) Stress Applied All the experimental results are summarized in Table the 1. Wire: 8 MPa) Maximum GMI ratio, ( Z/Z) max 22% 16% 27% Table 1. Summary of the results obtained with an amorphous Co-rich wire, with a bending stress Peak field, H peak (H peak H k ) 64 A/m 1 A/m +56% applied. Maximum normalized sensitivity, 1 Z(H) With No 1Bending With a Bending Stress.58 Applied (Maximum Relative Physical Quantities 42% S(H b ) H max Stress Applied Strain Applied along the Wire: 8 MPa) Change (%) Maximum GMI ratio, ( / ) Normalized sensitivity at H 22% 16% 27% Peak field, b 1 Z(H) ( (bias field), S(H b ) H (H ) 641A/m 1.52A/m +56% 48% Maximum normalized sensitivity, b) ( ) % Offset, ( Z(H ) b ) 64 Ω 4 Ω 38% Normalized sensitivity at ( ) (bias field), ( % ( ) ) While some Offset, solutions ( ) could be used 64 Ω to reduce the influence 4 Ω of the variations of 38% the offset and to compensate for the loss in sensitivity in the practical utilization of the sensor (utilization that continuously While involves some solutions bending could stress be used of to the reduce sensitive the influence element), of the one variations question of the offset mustand beto carefully compensate for the loss in sensitivity in the practical utilization of the sensor (utilization that addressed. The reversibility and repeatability of the bending effect is actually a major concern. continuously involves bending stress of the sensitive element), one question must be carefully In the case addressed. of poor The reversibility and and repeatability repeatability, of the bending the efficiency effect is actually of solutions a major that concern. cancel In the the offset and compensate case of poor foreversibility the loss ofand sensitivity repeatability, willthe actually efficiency be limited. of solutions This that is cancel why experimental the offset and tests of reversibility compensate and repeatability for the loss of were sensitivity conducted will actually on thebe same limited. sample. This is why experimental tests of Thereversibility wire was and bent repeatability under thewere sameconducted conditions the described same sample. in Section 2 and then relaxed to return to The wire was bent under the same conditions described in Section 2 and then relaxed to return its initial straight position. For example, Figure 6 presents the change in the impedance Z(H). to its initial straight position. For example, Figure 6 presents the change in the impedance ( ) with a with first a applied first applied stress stress followed followed by bya arelaxation step (a step return (a to return the straight to the position) straight and position) then a and then a second applied bending followed followed by by relaxation. relaxation. Figure Figure 7 shows 7 shows also the also related the related curves of curves of normalized normalized sensitivities. (a) (b) Figure 6. Cont. Figure 6. Cont.

8 Sensors 217, 17, 64 8 of 1 Sensors 217, 17, 64 8 of 1 (c) Sensors 217, 17, 64 8 of 1 (d) (c) (d) Figure Figure Reversibility 6. Reversibility and andrepeatability of of of the the impedance curve, ( ), ( ), Z(H), following following two two twoconsecutive bending bending stresses. stresses. (a) (a) Change (a) Change in inthe the modulus of of the impedance, Z(H) ( ), with, with with a 1st a bending a1st 1stbending stress stress applied; applied; (b) Comparison between the after the relaxation of bending (b) (b) Comparison between the themodulus of ofthe theimpedance after the therelaxation of ofthe thebending stress (1st bending) and the initial straight position; (c) Change in the modulus of the impedance, stress (1st bending) ( ), and with a 2nd the bending theinitial straight stress; position; (d) Comparison (c) between (c) Change in the modulus inthe themodulus of the of impedance ofthe theimpedance, after the ( ), Z(H) with, relaxation witha a2nd 2ndbending of the stress; bending stress; (d) (2nd (d) Comparison bending) between and the the initial straight themodulus of position. ofthe theimpedance after the relaxation of ofthe bending stress (2nd bending) and the initial straight position (a) (a) (b) (b) Straight position 1 st bending -5 5 Magnetic Straight field position H, (A/m) Magnetic field H, (A/m) (c) 1 1 st bending Straight position End of 1 st bending Magnetic field Straight H, position (A/m) End of 1 st bending Magnetic field (d) H, (A/m) (c) (d) Straight position 2 nd bending Magnetic field H, (A/m) Magnetic field H, (A/m) Figure 7. Reversibility and repeatability of the sensitivity following two consecutive bending stresses. (a) Change.5 Straight position in the normalized sensitivity with.5 Straight position a 1st bending stress; (b) Comparison between 2 nd bending End of 2 nd bending the normalized sensitivity after the relaxation of the bending stress (1st bending) and the initial straight position; (c) Change -5 in the normalized 5 sensitivity -5 with a 2nd bending stress; 5 (d) Comparison between the normalized Magnetic sensitivity field after H, the (A/m) relaxation of Magnetic the bending field stress H, (A/m) (2nd bending) and the initial straight position. Figure 7. Reversibility and and repeatability of the of sensitivity the sensitivity following following two consecutive two consecutive bending bending stresses. stresses. (a) Change (a) Change in the normalized in the normalized sensitivity sensitivity with awith 1st bending a 1st bending stress; stress; (b) Comparison (b) between between the the normalized sensitivity after after the relaxation the relaxation of theof bending the bending stress stress (1st bending) (1st bending) and theand initial the straight initial straight position; position; (c) Change (c) Change in the normalized in the normalized sensitivity sensitivity with a 2nd with bending a 2nd bending stress; (d) stress; Comparison (d) Comparison between between the normalized the normalized sensitivity sensitivity after theafter relaxation the relaxation of the bending of the bending stress (2nd stress bending) (2nd bending) and the and initial the initial straight straight position. position..5 Straight position End of 2 nd bending

9 Sensors 217, 17, 64 9 of 1 We have shown, for illustration purposes only, the effect of two consecutive bendings. About 1 consecutive bendings were conducted and showed that the change in the Z(H) curve did not exceed a few percent of the absolute value of Z(H). This value can be included in the error margin of two consecutive measurements (within our current experimental setup, two consecutive applied stresses can slightly differ). Nevertheless, we can conclude that the first observations showed that the effect of the bending stress was roughly reversible and repetitive (obviously within the limit of elasticity of the wire). This result is actually of great value for sensor applications. While this study has addressed an initial overall behavior of amorphous GMI Co-rich wires under bending stress, it is still obviously required that these experimental results be confirmed by theoretical analysis of model and completed by further experimental investigations. First of all, and unlike the effect of tensile stress, which has been largely studied in the literature, the effect of compressive stress should be carefully addressed from theoretical and experimental points of view. The general approximation made in considering that the bending effect is a combination of tensile and compressive effects must be refined. One of the two effects could dominate according to the diameter of the curvature of the bent wire. The influence of the frequency and of the amplitude of high frequency excitation current, i ac, should also be considered. All of these issues comprise the subject of our ongoing and future works. 4. Conclusions We addressed an experimental study to evaluate the impact of the bending stress on the GMI effect in amorphous Co-rich wires. This bending stress was considered an influential parameter in our application of electrical current GMI sensors. Its effect on the relevant quantities for GMI sensor implementation was investigated. The results show that the GMI ratio, the intrinsic sensitivity, and the offset in the studied sample were affected by the bending. For the considered experimental conditions of bending and the used parameters of excitation current of the wire, both GMI ratio and maximum sensitivity were reduced (by roughly 5%). A net change in the offset was also observed. It has also been shown that the effect of bending seemed to be reversible and repetitive. However, more intensive work is still necessary to allow for a better understanding of the physical phenomena related to bending and compression. This understanding must also be confirmed by more experimental investigations. Author Contributions: Julie Nabias contributed in the ideas of this work. She designed the setup and realized the measurements. She also performed the data processing, contributed to the interpretation of the results, and wrote the paper. Aktham Asfour and Jean-Paul Yonnet contributed ideas and participated in the setup design, results interpretation, and the paper writing. Conflicts of Interest: The authors declare no conflict of interest. References and Note 1. Rheem, Y.W.; Kim, C.G.; Kim, C.O.; Yoon, S.S. Current sensor application of asymmetric giant magnetoimpedance in amorphous materials. Sens. Actuators A Phys. 23, 16, [CrossRef] 2. Mapps, D.J.; Panina, L.V. Magnetic Field Detector and a Current Monitoring Device Including Such a Detector. U.S. Patent No. 7,564,239, 21 July Ripka, P. Current sensors using magnetic materials. J. Optoelectron. Adv. Mater. 24, 6, Malátek, M.; Ripka, P.; Kraus, L. Double-core GMI current sensor. IEEE Trans. Magn. 25, 41, [CrossRef] 5. Zhan, Z.; Yaoming, L.; Jin, C.; Yunfeng, X. Current sensor utilizing giant magneto-impedance effect in amorphous ribbon toroidal core and CMOS inverter multivibrator. Sens. Actuators A Phys. 27, 137, [CrossRef] 6. Asfour, A.; Yonnet, J.-P.; Zidi, M. A High Dynamic Range GMI Current Sensor. J. Sens. Technol. 212, 2, [CrossRef]

10 Sensors 217, 17, 64 1 of 1 7. Fisher, B.; Panina, L.V.; Fry, N.; Mapps, D.J. High Performance Current Sensor Utilizing Pulse Magneto-Impedance in Co-Based Amorphous Wires. IEEE Trans. Magn. 213, 49, [CrossRef] 8. Han, B.; Zhang, T.; Zhang, K.; Yao, B.; Yue, X.; Huang, D.; Huan, R.; Tang, X. Giant magnetoimpedance current sensor with array-structure double probes. IEEE Trans. Magn. 28, 44, Kudo, T.; Tsuji, N.; Asada, T.; Sugiyama, S.; Wakui, S. Development of a small and wide-range three-phase current sensor using an MI element. IEEE Trans. Magn. 26, 42, [CrossRef] 1. García, C.; Zhukov, A.; Blanco, J.M.; Zhukova, V.; Ipatov, M.; Gonzalez, J. Effect of Tensile Stresses on GMI of Co-Rich Amorphous Microwires. In Proceedings of the INTERMAG Asia 25, Digests of the IEEE International Magnetics Conference, Nagoya, Japan, 4 8 April 25; pp Ciureanu, P.; Khalil, I.; Melo, L.G.C.; Rudkowski, P.; Yelon, A. Stress-induced asymmetric magneto-impedance in melt-extracted Co-rich amorphous wires. J. Magn. Magn. Mater. 22, 249, [CrossRef] 12. Knobel, M.; Sanchez, M.L.; Velazquez, J.; Vázquez, M. Stress dependence of the giant magneto-impedance effect in amorphous wires. J. Phys. Condens. Matter 1995, 7, L115. [CrossRef] 13. Knobel, M.; Pirota, K.R. Giant magnetoimpedance: Concepts and recent progress. J. Magn. Magn. Mater. 22, 242, [CrossRef] 14. Knobel, M.; Vázquez, M.; Kraus, L. Giant magnetoimpedance. Handb. Magn. Mater. 23, 15, Atkinson, D.; Squire, P.T. Experimental and phenomenological investigation of the effect of stress on magneto-impedance in amorphous alloys. IEEE Trans. Magn. 1997, 33, [CrossRef] 16. Knobel, M.; Vazquez, M.; Hernando, A.; Sánchez, M.L. Effect of tensile stress on the field response of impedance in low magnetostriction amorphous wires. J. Magn. Magn. Mater. 1997, 169, Mandal, K.; Puerta, S.; Vazquez, M.; Hernando, A. The frequency and stress dependence of giant magnetoimpedance in amorphous microwires. IEEE Trans. Magn. 2, 36, [CrossRef] 18. Blanco, J.M.; Zhukov, A.; Gonzalez, J. Effect of tensile and torsion on GMI in amorphous wire. J. Magn. Magn. Mater. 1999, 196, [CrossRef] 19. Kim, C.G.; Yoon, S.S.; Vazquez, M. Evaluation of helical magnetoelastic anisotropy in Fe-based amorphous wire from the decomposed susceptibility spectra. J. Magn. Magn. Mater. 21, 223, [CrossRef] 2. Zhao, Y.; Hao, H.; Zhang, Y. Preparation and giant magneto-impedance behavior of Co-based amorphous wires. Intermetallics 213, 42, [CrossRef] 21. Seddaoui, D.; Ménard, D.; Movaghar, B.; Yelon, A. Nonlinear electromagnetic response of ferromagnetic metals: Magnetoimpedance in microwires. J. Appl. Phys. 29, 15, [CrossRef] 22. Phan, M.; Peng, H. Giant magnetoimpedance materials: Fundamentals and applications. Prog. Mater. Sci. 28, 53, [CrossRef] 23. Co-Based SENCY TM, Unitika Ltd. Datasheet 24. Hernando, B.; Sanchez, M.L.; Prida, V.M.; Tejedor, M.; Vázquez, M. Magnetoimpedance effect in amorphous and nanocrystalline ribbons. J. Appl. Phys. 21, 9, [CrossRef] 25. Li, Y.F.; Vazquez, M.; Chen, D.X. Giant magnetoimpedance effect and magnetoelastic properties in stress-annealed FeCuNbSiB nanocrystalline wire. IEEE Trans. Magn. 22, 38, [CrossRef] 26. Yang, X.L.; Yang, J.X.; Chen, G.; Shen, G.T.; Hu, B.Y.; Jiang, K.Y. Magneto-impedance effect in field-and stress-annealed Fe-based nanocrystalline alloys. J. Magn. Magn. Mater. 1997, 175, [CrossRef] 27. Beato, J.J.; Algueta-Miguel, J.M.; de la Cruz Blas, C.; Santesteban, L.G.; Perez-Landazabal, J.I.; Gomez-Polo, C. GMI Magnetoelastic Sensor for Measuring Trunk Diameter Variations in Plants. IEEE Trans. Magn [CrossRef] 28. Ménard, D.; Seddaoui, D.; Melo, L.G.C.; Yelon, A.; Dufay, B.; Saez, S.; Dolabdjian, C. Perspectives in Giant Magnetoimpedance Magnetometry. Sens. Lett. 29, 7, [CrossRef] 217 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (