IEEE TRANSACTIONS ON MAGNETICS, VOL. 32, NO.

5, SEPTEMBER 1996 343 I

AMR Effect in Spin-Valve Structure

Y.Uehara, K.Yamada, and H.Kanai

Fujitsu Limited, 10-1 Morinosato-Wakamiya, Atsugi 243-01, Japan

Abstract - It i s very important t o estimate the

a n i s o t r o p i c magnetoresistance (AMR) effects i n a
spin-valve structure when designing spin-valve h e a d s . ~ ,/ Four-point Pro&
The AMR e f f e c t i n a free l a y e r can have an e s p e c i a l l y
significant influence on head properties. We measured
the change i n resistance i n s p i n - v a l v e f i l m s and
u-WU
.A$Z-F- Sample
separated the AMR s i g n a l from the total change i n
resistance using an improved means o f measurement. d
With this measurement, we directly obtained the AMR
effect o n the free layer. The sample structure we used
1 ) I
was NiFe / CoFe / Cu / CoFe I FeMn. The AMR effect
NiFe 20nm
was about 1 6 % of the S V effect. This value i s rather
Q
large and can affect the head p r o p e r t i e s .

7s
0.6 - -Measurement
Simulation
Q 0.4
I. INTRODUCTION a 0.2
Spin-valve (SV) elements consist of a free layer, an 0
interlayer anda pinnedlayer[ 1],[2]. The change in resistivity due 0 5 10 15
to the SV effect arises from the relative angle betwecn the pinned
Sample size d/D
and free layers. At the same time, both these layers have an AMR
effect. This AMR effect in the free layw influences SV head
properties, such as output voltage and asymmetry. It is very Fig. 1. Relationship between sample size (d) and AMR signal. The
important to accurately estimate the AMR effect for fabricating horizontal axis is the normalized sample size and the vertical axis is the
normalized AMR signal,. Ap is the intrinsic change in resistivity of the NiFe
SV heads. film and Apd is the measured change in resistivity at sample size (d).
The AMR effect in the SV element was measured by Dicny
[l]. He compared resistances for cumnt applied in parallcl and
perpendicular to the magnetizations of the free and pinned layers. probe (in-line type). When this probe is used to measure the
For both orientations, a suficient field was applied to saturate AMR effect, the signal is dramatically dependent on the sample
the magnetizations of the two layers. The AMR was dcduced by size. Fig.1 shows the relationship between the sample sizeand
using the relative thickness of the two layers. We can evaluate the AMR signal. We used a 20-nm-thick NiFe film. The
the AMR using this measurement, but this isn’t a direct simulation result is also shown. If the sample size is much
measurement and contains some errors from the &duction larger than the probe, the AMR signal cannot be detected.This
process. Furthermore, if the material of the freelayer is different is because the current distribution in the film is modifiedandthe
from that of the pinned layer, we can’t use this measurementany signal, &e to the change in resistance, is canceled When the
more. sample size is ten times larger than the probe size, the AMR
We have successfully measured the AMR effect in the free signal is negligible.
layer and we report its influence on head properties using a On the other hand, the S V signal does not &pend on the
micromagnetic model. distribution of current. The change in resistivity due to the SV
effect can be detected using the four-point probe, even if the
sample size is much larger than the probe size.
11. SPERIMENT These results suggest that the AMR signal can be separated
out from the total change in resistance in S V films. We
Thin film resistivity is usually measured with a four-point conducted two types of measurements, [A] and [B] as shown in

0018-9464/96$05 00 (c) 1996 IEEE

3432

Current source
0.4

0.3
Four-point probe -c
v 0.2
da
0.1

d > 10-D 0
-50 -25 0 25 50
(a) Measurement [AI
Applied fieId (Oe)
(a) SV signal and SV+AMR signal

0.2

0.1

o e Free
h

c
w
da
-0.1

(b) Measurement [Bl -0.2

-50 -25 0 25 50
F+g. 2. Experimental setup for (a) measurement [A] using a four-point probe
and (b) measurement [B] using a pair of extra electrodes Applied field (Oe)
(b) A M R signal

Fig.2. We used a T m m x 50 mm sample and a 3 mm probe. Eg. 3. Separation of A M R signal and SV signal. The field is applied parallel
Measurement [A] was a conventional measurement with the to the exchange anisotropy field, (a) showing SV signal (solid) and SV +
A M R signal (dashed), and (b) showing A M R signal derived from the two
four-point probe, where only the SV signal can be detected. In measurements.
this case, the sample size is much larger than the probe, so the
AMR signal is canceled In measurement [B], we put a pair of
extra electrotks to supply a sense current instead of the two SV structure was 13.0 Q. The sense currents were 5 mA at
electrodes in the probe, as shown in Fig.2(b). In this condition, measurement [A] and 56.2 mA at measurement [B].
the current flow in the film is uniform, and we can dctect both
AMR and SV signals. The difference between the two
measurements represents the AMR effect. 111. RESULTS A N D DISCUSSION
The sample structure we used is Ta (10 nm) / NiFe (6 nm) /
CoFe (1.5 nm) /Cu (3.6 nm)/ CoFe (5.5 nm)/ FeMn (lonm) Figures 3 and 4 show the experimental results. In Fig. 3, an
/ Ta (10 nm) [3]. These films were deposited on AI203TiC extemal fieldis applied parallel to the anisotropy fieldcreated by
substrates using DC or RF magnetron sputtering. We annealed FeMn. This direction matches the orientation of the easy axis of
the sample at 250 T, then sooled it to room temperature in the free layer. The field range is k 40 Oe. The current flows in
applied fields of 2500 Oe. The exch,ange coupling fieldbetween parallel to this direction. The solid line in Fig. 3(a) shows the
the pinned layer and the antifetromagnetic layer was about 400 change in sheet resistance for only the SV effect (measurement
Oe, and the interlayer coupling between the free layer and the [A]) and the dashedline shows the change for both the AMR+SV
pinnedlayer was about 9 Oe. The sample was then shapcd to effects (measurement PI). The difference between the two
form a rectangle. Gold electrocks were placed on two opposite measurementsrepresentsthe AMR effect,as shown in Fig. 3@).
sides of the rectangular pattern, as shown in Fig. 2(b). The total This kind of shape is observed when measuring a thin
sheet resistance was 9.94 Q, and the sheet resistance of just the ferromagnetic film at an applied field parallel to the easy axis.
 /
3433

0.4

0.3
h

2 0.2
da
0.1
-150 -100 -50 0 50 100 150
Applied field (Oe)
0
-50 -2 5 0 25 50 Fig. 5. Simulation of the transfer curve of the SV element without shields.
Applied field (Oe) The element height is 2 pm and the sense current is 5 mA.

(a) SV signal and SV+AMR signal

0.2 direction by exchange coupling with the antiferromagneticlayer.

Measurement [B] - Measurement [A] The element height was 2 pm and the sense current was 5 mA.
0.1 The linear dynamic range in the transfer curve was reduced
Pinned considerablybecause of the AMR effect. Points A andB arethe
Free optimum points that minimize the influenceof the AMR effect,
as shown in Fig. 5. These points correspond to the linearregion
of the AMR signal. The output voltage is also influencedby the
-0.1
AMR effect. The output voltage at point A was 40 % higher
than that at point B. This is because the AMR signal is addcdto
-0.2 the SV signal at point A. The AMR effect in the free layer is one
-50 -2 5 0 25 50 of the most important factors in SV head fabrication.
Applied field (Oe)
(b) AMR signal
IV. CONCLUSIONS
Fig. 4. Separation of the AMR signal and the SV signal with a rotating field.
The horizontal axis represents the component of the field parallel to the
exchange anisotropy field. Wecaniedout anew anddirectmethodof measuring the AMR
effect in SV structures. We demonstrated that the AMR signal
can cleanly be separated based on the change in resistance. The
We canseparate the AMR signal using this measurement,but ratio of the AMR signal to the SV signal is rather large. We
its signal height strongly dependson the orientation of theeasy must consider the effect of the AMR signal in SV headdesign.
axis of the film and the distribution of the anisotropy field In
this case, the distribution of the interlayer coupling also
influences the shape.
Our improved measurement employs a rotating field so that
REFERENCES
the magnetization of the free layer can rotate uniformly. We used
[ l ] B. Dieny, V. S. Speriosu, S. S. P. Parkin, B. A. Gurney, D. R. Wilhois
two pairs of Helmholtz coils to generate the rotating field The and D. Mauri, “Giant Magnetoresistance in Soft Ferromagnetic
strength of appliedfieldwas also40 Oe. Fig. 4 shows the results Multilayers,” Phys. Rev. B vol. 43, pp. 1297-1300, January 1991.
of the improvedmeasurement.The horizontal axis represents the [2] D. E. Heim, “Design and Operation of Spin Valve Sensors,” IEEE Trans.
component of the field parallel to the exchange anisotropy field Mag. vol. 30, pp. 316-321, March 1994.
[3] H. Kanai, K. Yamada, K. Aoshima, Y. Ohtsuka, J. Kane, M. Kanamine,
created by FeMn. The change in sheet resistance due to the SV I. Toda, and Y. Mizoshita. “Spin-Valve Read Heads with
effect was 0.38 R and that h e to the AMR effect was 0.06 Q. NiFeICo90Fe10 Layers for 5 Gbit/m2 Density Recording,” AA-02
The AMR rate was about 16%. INTERMAG’ 96.
To estimate the influence of the AMR effect in the SV head, [4] Y. Uehara, and T. Kutsuzawa, “Play-back Simulation of MR Heads
we calculated the transfer curve of the SV element without using a Micromagnetic Model,‘‘ Digest of The Conference on Magnetics
of Japan, pp. 214, 1995.
shields using a micromagnetic model we developed (Fig. 5 ) [4].
The magnetization of the pinned layer was fixed in the height