High-performance magnetoresistive sensors

High-performance magnetoresistive sensors

Sensors and Actuators 81 Ž2000. 27–31 www.elsevier.nlrlocatersna High-performance magnetoresistive sensors Hans Hauser b a,) , Gunther Stangl b, Jo...

307KB Sizes 2 Downloads 122 Views

Sensors and Actuators 81 Ž2000. 27–31 www.elsevier.nlrlocatersna

High-performance magnetoresistive sensors Hans Hauser b

a,)

, Gunther Stangl b, Johann Hochreiter ¨

c

a Institut fur ¨ Werkstoffe der Elektrotechnik, Technische UniÕersitat ¨ Wien, Gußhausstraße 27–29, A-1040 Vienna, Austria Institut fur ¨ Angewandte Electronik und Quantenelektronik, Technische UniÕersitat ¨ Wien, Gußhausstraße 27–29, A-1040 Vienna, Austria c Dipl.-Ing. Hans Schiebel Elektronische Gerate ¨ Gesellschaft, Margaretenstraße 112, A-1050 Vienna, Austria

Abstract Depending on the angle between magnetization and current density in a thin permalloy film, the anisotropic magnetoresistive effect is utilized for high-performance sensors. Both the parameters of the sputtering process and the sensor layout are discussed. The anisotropic magnetoresistive effect of dc-sputtered Ni 81%–Fe 19% films has been increased up to D rrr s 3.93% at 50 nm thickness and a sensitivity of 0.5 mVrnT has been achieved by an elliptically shaped sensor layout. q 2000 Elsevier Science S.A. All rights reserved. Keywords: Anisotropic magnetoresistive effect; Permalloy film; Cathode sputtering; Demagnetizing field

1. Introduction The anisotropic magnetoresistive ŽAMR. effect is widely utilized in sensor applications, related with the detection of weak magnetic fields w1x. The resistance r in the plane of a thin, ferromagnetic film with uniaxial anisotropy varies with the angle between the current density and the spontaneous magnetization. To achieve a high sensitivity, the coercivity in the hard axis magnetization direction must be very low and the specific resistance variation D rrr must be as high as possible. Depending on both material composition and film thickness, it is Žtheoretically. about 4% in a 50 nm thin Ni 81%–Fe 19% Žmagnetostriction free. permalloy film. The sensitivity S0 can be controlled by the total anisotropy field H0 s 2 Krm 0 Ms : S0 s

DR 1 R 0 H0

.

Ž 1.

The anisotropy constant K, the spontaneous magnetization Ms , the average resistance R 0 , and the field dependent variation D R are determined respectively by the material and geometry of the magnetoresistive element.

) Corresponding author: [email protected]

Fax:

q43-1-504-15-87;

E-m ail:

Furthermore, the layout of the magnetoresistive elements forming a Wheatstone-bridge has to be optimized. Achieving a homogeneous and small demagnetizing field, an elliptical shape of the AMR array is proposed w2x. By applying a flipping field H F , the direction of Ms in the AMR-element can be inverted—this is very useful to overcome offset problems.

2. Sensor layout Three layouts with rectangular permalloy strips of different width and separation distance have been investigated w3x. 1. KMZ1020A: Width 20 mm Žtapering off at both ends., distance 10 m m, R 0 s 1.70 k V , S 0 s 3.31 ŽmVrV.rŽkArm., H F s 400 Arm. 2. KMZ1010B: Width 10 mm, distance 10 mm, R 0 s 5.63 k V, S0 s 1.36 ŽmVrV.rŽkArm., H F s 2000 Arm. 3. H3: Width 10 mm Žtapering off at the ends., distance 10 mm, R 0 s 5.63 k V, S0 s 1.36 ŽmVrV.rŽkArm., H F s 2000 Arm, overall elliptical shape Žsee Fig. 1.. The total sensor area was 1 = 2 mm2 . The AMR film was characterized by D rrr s 1.52 and H0 s 600 Arm Žthese results have been achieved earlier; recent improvements in the AMR film technology are reported below and in Table 1.. Depending both on the demagnetizing factor of the single strips and the total area, the sensitivity of KMZ1020A is about 2.5 times of the KMZ1010B. The

0924-4247r00r$ - see front matter q 2000 Elsevier Science S.A. All rights reserved. PII: S 0 9 2 4 - 4 2 4 7 Ž 9 9 . 0 0 1 6 5 - X

H. Hauser et al.r Sensors and Actuators 81 (2000) 27–31

28

Fig. 1. Layout of H3 Žhalf bridge.; long axis length: 2 mm; the AMR-films are covered with barber-poles Žr.; two bonding pads ŽI. belong to one resistor.

layout H3 was designed as a half bridge and therefore its sensitivity is half of KMZ1020A. The sensor characteristic of both magnetic states is shown in Fig. 2. Rotating the ellipse axes by 908 results in a sensitivity increase by Ž NbrNa . 2 , where Na is the demagnetizing factor in the long axis direction and Nb is the demagnetizing factor in the short axis direction of the ellipse. Furthermore, the signalrnoise ratio is improved by achieving more homogeneous fields as compared to the KMZ layouts w3x.

3. Magnetoresistive films The magnetoresistive films have been deposited by cathode sputtering Žtriode process.. Fig. 3 shows the arrangement schematically w4x. The target is connected to a negative potential of UT s y800 V and the substrate is biased by US s y60 V. The cathode current is IC s 43 A, the anode current is IA s 3.5 A, and the anode voltage is UA s q50 V against ground potential. The following parameters have been varied: Both target and substrate materials, the temperatures of target ŽTT . and substrate ŽTS ., the distance aT – S between target and substrate, and the film thickness d. This value has been determined by resonance frequency measurements with an accuracy of better than 1%. Depending on the applied magnetic field direction, the specific resistance variation has been measured by a fourwire method. Table 1 shows the parameters and results for some samples, sputtered from a melted and rolled NiFe 81:19

Fig. 2. Sensitivity of H3; the output voltage Ua of the half bridge versus the applied field H y could be about 8 times greater using a full bridge with an elliptical resistor layout rotated by 908 compared to Fig. 1.

target Žpurity 99.3%.. A sintered target of the same alloy Žpurity 99.996%. was used for the sample No. 186 only. The bias magnetic field was H x s 260 Arm. and the substrate material was Si–SiO 2 . The results indicate an increasing AMR effect D rrr with decreasing r Žas a function of the grain size, see Fig.

Table 1 Parameters and results for selected samples sputtered from NiFe targets Sample

TS w8Cx

TT w8Cx

˚ x Sputter rate wArs

d wnmx

aT – S wnmx

D r wmV cmx

r wmV cmx

D rrr w%x

178 182 184 185 186

272 238 275 274 268

560 560 560 560 560

3.0 3.1 2.7 2.9 3.0

50 50 50 20 50

40 40 38 36 36

0.17 0.21 0.21 0.16 0.23

4.42 5.79 5.43 5.03 6.76

3.93 3.54 3.78 3.27 3.43

H. Hauser et al.r Sensors and Actuators 81 (2000) 27–31

29

Fig. 3. Cathode sputtering setup; target and substrate temperatures can be controlled both by bias sputtering and oil heating Žnot shown in the drawing..

˚ . yields also a low 4.. A low sputtering rate Ž0.7–0.8 Ars AMR effect. Both TS and TT have a positive influence on the AMR effect. If TS is about 2708C, D rrr is increased by up to 0.5%. The maximum TT of 5608C increases the AMR effect by 1% up to 3.93% w5x. The target–substrate distance has been varied between 36 mm and 60 mm, yielding a change in the AMR effect by "0.2%. The optimum aT – S is in the range between 38 mm and 42 mm.

The optimum thickness d was about 50 nm. Reducing d to 20 nm, which has often been reported to be the optimum for permalloy, e.g., Ref. w6x, yields a decrease of the AMR effect by 0.5% Žsamples 184 and 185.. The magnetic behaviour also depends strongly on the thickness d. This is demonstrated by the magnetization curves Žmeasured by the magneto-optical Kerr effect w7x. of the two samples above Žsee Fig. 5.. By reducing d from

Fig. 4. D rrr versus r for sputtering processes with different targets. R1: sintered NiFe 81:19, R2: melted and rolled NiFe 81:19, R3: NiFeMo 80:15:5.

30

H. Hauser et al.r Sensors and Actuators 81 (2000) 27–31

Fig. 5. Magnetization M Žphotodetector current, arbitrary units. versus applied field H of sputtered permalloy films, thickness 50 nm Žabove: sample 184. and 20 nm Žbelow: sample 185..

50 nm to 20 nm, both the easy axis coercivity and the hard axis coercivity are increasing. 4. Conclusions For a high sensitivity, the hard axis coercivity should be near zero. This is demonstrated by sample 186 Žsee Fig. 6.,

which shows almost ideal Stoner–Wohlfahrth rotation of the spontaneous magnetization. Providing a high D rrr s 3.43%, this sample is considered to be an optimum AMR sensor material. Using furthermore a Wheatstone bridge arrangement of elliptical shape with barber pole-structured magnetoresistors Žsee layout H3, rotated by 908, which means that the

Fig. 6. Magnetization M Žphotodetector current, arbitrary units. versus applied field H of sample 186.

H. Hauser et al.r Sensors and Actuators 81 (2000) 27–31

permalloy strips are oriented along the short axis. gives a sensitivity of 0.5 mVrnT at a supply voltage of 10 V. The bandwidth depends on the flip-frequency which is about 4 kHz for commercial sensors. With these sensors it is possible to detect the distortion of the earths magnetic field—caused by ferromagnetic objects which have to be located—with a gradiometer arrangement. Also applications for various industrial electronics are considerable because of the low production efforts.

Acknowledgements The authors are grateful to Prof. W. Fallmann and to Prof. G. Fasching for making these investigations possible. Financial support was provided by Dipl.-Ing. Hans Schiebel

31

Elektronische Gerate der ¨ and by the Fonds zur Forderung ¨ Gewerblichen Wirtschaft ŽFFF. under Grant no. 3r9893. References w1x U. Dibbern, Magnetoresistive sensors, in: W. Gopel, J. Hesse, J.N. ¨ Zemel ŽEds.., Sensors, Vol. 5: Magnetic Sensors ŽR. Boll, K.J. Overshott, Vol. Eds.., VCH Verlagsgesellschaft, Weinheim, 1989, pp. 341–380. w2x H. Hauser, Magnetfeldsensor II. Austr. Pat. Appl. No. 1928r96, 1996. w3x H. Hauser et al., Elektrotech. Informationstech. 115 Ž1998. 382–390. w4x G. Aigner, Herstellung dunner Permalloy-Schichten fur ¨ ¨ magnetoresistive Sensoren, Diplomarbeit, Technische Universitat ¨ Wien, Vienna, 1997, pp. 5–49. w5x P. Aigner, G. Stangl, H. Hauser, J. Phys. IV 8 Ž1998. 461–464. w6x Y.-J. Song, S.-K. Joo, IEEE Trans. Magn. 32 Ž1996. 4788–4790. w7x H. Hauser, F. Haberl, J. Hochreiter, M. Gaugitsch, Appl. Phys. Lett. 64 Ž1994. 2448–2450.