195
Sensors and Actuators, AZ-A.23 (1990) 795-798
A Thin-film Magnetoresistive Angle Detector
KEES J M EIJKEL, JOHAN W WIEBERDINK, JAN H J FLUITMAN and THEO J A POPMA
Unzverarry of Twente EL-TDM, P 0 Box 217, 7500 AE Em&& (The NetherIan&)
PETER GROOT and HENK LEEUWIS
Twente Technology Transfer (3T) BV, P 0 Box 545, 7H)o AM Em&de (The Netherlanh)
Abstract
An overview IS gven of the results of our
research on a contactless angle detector based on
the amsotroplc magnetoreslstance effect (AMR
effect) m a permalloy thm lilm The results of
high-temperature anneahng treatment of the
pennalloy !ilm are discussed Such a treatment
suppresses the effects of the umaxlal magnetic
amsotropy that IS present m a permalloy thm fihn
and increases the AMR effect, thus lmprovmg the
detector signal The performance of the detector
throughout a temperature range of 20 to 120 “C
and the results of heat treatment at 125 “C for 1
week have been tested
Introduction
Over the past 5 years, a contactless angle detec-
tor based on the amsotroplc magnetoresistance
effect (AMR effect) m a permalloy thm fihn has
been proposed and developed [ 1,2] This paper
presents an ovemew of the results of our research
on this detector and discloses recent unpubhshed
results
The angle detector consists of a pair of ldentl-
cal pseudo-Hall devices (PHDs), or equivalent
devices such as magnetoresistor bndges, mutually
rotated through 45” and posltloned opposite a
rotatable permanent magnet The magnetic field
at the positlon of the PHDs IS largely m-plane
PHD’s
Fig 1 The angle detector, comstmg of a piur of pseudo-Hall
dewes opposite a rotable permanent magnet
0924-4247/w/33 50
(Fig 1) The magnetiatlon angle 0 m the PHDs
reflects the angular position rp of the rotatable
magnet and IS comprised m the output signals of
the PHDs
V, = I Ap k sm( 20)
V, = I Ap k cos( 28)
I
(1)
in whch I IS the dnvmg current of the PHDs and
k IS a geometrical factor The output angle of the
detector, 0, IS independent of temperature, m
prmclple The magnetoreslstivlty Ap and reslstlv-
ity p of the permalloy are defined as
AP = @II - pJ2
and P = @II + pJ2)
(2)
m which p and pI are the reslstrvlty parallel and
perpenQcu ar
Y
to the magnetiahon m the perm-
alloy The permalloy thm films are sputtered m the
magnetostnction-free composition (81 at % Nl
and 19 at % Fe) This ensures a low m!luencc of
temperature-induced or mechamcally-induced
stresses m the film The values of Ap and p arc
32x 10-9Qmand22x lo-‘0mmafreshGhn
The performance of an angle detector as de-
scribed above depends on Its magnetical and elec-
tncal properties The magnetical propeties of the
detector include the strength and homogenaty of
the magnetic field of the magnet, on the one hand,
and the magnetic properties of the pennalloy thm
film on the other The electrical properties of the
detector, which have been discussed elsewhere [3],
are determined by the magnetoresistive properties
of the permalloy and the geometry of the PHDs
Of course, the stabdlty of all these properties 1s of
major importance for the performance of the de-
tector The electrical and magnetical properties of
the system can be treated separately, because their
mutual interaction (normal Hall effect, magnetic
fields caused by the currents m the thin film etc)
IS neghgble
Magnetic Behavior
A few short remarks concermng the permanent
magnet should be made The magnetic field at the
0 Elsewer Sequola/Pnnted m The Netherlands
196
posltion of the PHDs should be large to ensure a
small Influence of the earth’s magnetic field and
other disturbing a c or d c fields, the field should
be stable (no demagnetization) and it should be
homogeneous to enable a homogeneous magne-
tlzatlon of the PHDs These reqmrements can be
met using common magnetic matenals like an-
isotropic ferroxdure In practical mtuatlons, errors
caused by disturbing magnetic fields (e g the
earth’s magnetic field) or by the mhomogenelty of
the field can be kept below 0 1 o [l] If very large
field strengths and high stablhty of the magnet are
required, new magnetic matenals like SmCo, and
NdFeB can be applied
Permalloy films show magnetic amsotropy,
which causes an onentafion difference between the
magnetization and the magnetic field, resulting m
an angle detector measurement error The mag-
netic amsotropy can be onented If the film 1s
deposited m the presence of a magnetic field [4]
Under normal deposltlon condltlons, the an-
isotropy field strength Hk m a sputtered or evapo-
rated film 1s about 350 A/m, leadmg to a
maximum angle detector measurement error of
about 0 5” If a magnetic field of 20 kA/m 1s used
In many cases this error IS not acceptable A
number of techmques have been developed to
reduce this error [l] The use of a permalloy
bllayer wth perpendicular m-plane onentatlon of
the amsotropy m both sublayers proved to be very
successful The results have been pubhshed else-
where [S]
Alternatively, one can use a permalloy film
conslstmg of many small areas mth randomly
onented magnetic amsotropy If such a fihn expe-
nences a sticlently large m-plane magnetic field
(larger than the local amsotropy field Hk) wth
arbitrary onentatlon, the mean magnetitlon m
the fdm will have the onentatlon of the field [l]
To manufacture such a magnetically lsotroplc
film, depontlon of the permalloy m a rotatmg
magnetic field was considered [6] Tlus techmque
was not successful We discovered m several ex-
penments that a permalloy layer with a well-
defined amsotropy dctates the amsotropy
onentatron of a second layer sputtered on top of
the first, even m a moderate (2 kA/m) m-plane
magnetic field of arbitrary but constant onenta-
tlon This indicates that the onentatlon of the
amsotropy 1s transferred to the second layer by
means of some nonmagnetic mteractlon (e g by
the surface structure of the first film) or that the
magnetic amsotropy IS confined to and transferred
by certain (small) areas m the film whch are not
influenced at the field strength use’d The fact that
Goto et al [q managed to reonent the anisotropy
m the second layer using a stronger deposition
field indicates that the latter 1s vahd or that a
balance exists between field-mduced and struc-
ture-induced amsotropy
Another method to produce a magnetically
lsotrop~ permalloy fihn 1s the use of a high-
temperature anneahng step to Qsonent the an-
isotropy locally mslde the permalloy layer
Expenments performed by Metzdorf [8] show that
changes m amsotropy otrentation, mduced by an-
neahng m a magnetic field at temperatures up to
400 “C, are partly reversible when annealing at
lower temperatures Stable films are obtamed after
annealing at temperatures above 400 “C [4] In
our case, a desonentatlon of the magnetlzatlon m
the permalloy 1s necessary durmg annealing ms
can be achieved by exceeding the Curre tempera-
ture (for permalloy at a 8 l/19 composition, Just
above 500 “C) We annealed a number of films at
dfferent temperatures between 200 “C to 600 “C
for 1 h and measured the coerclvlty H, of the film
(using an mductlve hysteresis loop tracer), cr, (the
angle which includes 90% of all amsotropy onen-
tations m the fihn, measured using the Crowther
method [9]), and the maxunum angle between the
magnetic field and the mean magnetization m the
(cp - &nax,
at a field strength of 2 kA/m usmg
1000
x
1
500
d
I
I I I
I
0 100 200 so0 400 500 600
(a)
T [‘Cl
(W
ago t&9=-1
Fig 2 (a) I#, and a, as a function of anneahng temperature
(annealmg tune 1 h) The error bars mdxate the spread of the
results (b) Measured and cakulated values of (9 - 6), as a
function of cr, in a field of 2 IA/m
the measurement system described m ref 10 The
fihns were annealed 111 a mtrogen atmosphere and
were protected by a 75-mn 40 ti The results
are shown m Fig 2(a) and (b) It IS clear that a,
Increases mth mcreasmg annealing temperature,
as expected The iihns annealed at 600 “C are
magnetically lsotroplc In these films, a, cannot
be determmed wth the Crowther method
To gam more ins&t into the propeties of the
annealed fihns, we performed computer slmula-
tions of films consistmg of a large number of small
areas Hrlth no mutual magnetic m&action, each
havmg a well-defined magnetic amsotropy (Hk
equal to that of a fresh fihn, 350 A/m), with an
onentatlon datnbuuon as measured m our an-
nealed fihns The calculated values of @I - @,,,
at a field strength of 2 kA/m are shown m Fig
2(b) Two possible mechamsms can cause the
observed differences
-The amsotropy field strength of small areas
wth well-defined amsotropy decreases durmg an-
nealing
-A large fraction of the total amount of
these areas has random amsotropy one&&on
This fraction 1s not vlslble m measurements of
(rp - &ll~X
or m the Crowther measurements, but
It reduces the mfiuence of the remamng part of the
film on the onentatlon of the mean magnetition
The areas mth well-defined amsotropy onenta-
tlon m high-temperature ( b 500 “C) annealed
films are small In Kerr rotation measurements
using a small hght spot of a few tens of rmcrome-
ters, comparable magnetic behavior m Merent
areas 1s found
Detenoratlon of the permalloy m hlgh-temper-
ature annealed films could be responsible for the
observed increase m coerclvlty Capacitor struc-
tures of 100 x 100 pm formed by a permalloy
film, a 75-nm $0 layer and an alummum layer
showed low resistance m some cases, but good
lsolatlon m some other cases, all on the same
wafer This indicates that the BO layer IS not
completely closed (e g due to pm-holes) No VISI-
ble damage to the film surface could be found
wth nucroscoplc mspectron and SEM
It 1s concluded that a film annealed at 500 “C
for 1 h 1s appropnate for use m an angle detector
((rp -@,ax<O 1”
for H=20kA/m) figh-
temperature annealing has another advantage It
lowers the reslstlvlty of the film [ 111, thus mcreas-
mg the magnetoreslstlvlty ratio Ap/p from 1 5%
m fresh films to 2%
TechOOlOgy
The manufacture of our PHDs and magnetore-
slstor bndges 1s straightforward A pennalloy film
IS deposited by RF sputtermg on an oxldlzed slhcon
wafer The fihn geometry IS produced usmg pho-
tohthographlc techmques and wet chenucal etchmg
Flhn and substrate are covered with a 75-nm BO
Ghn The wafer 1s annealed for 2 h at 500 “C and
a stable, magnetically lsotroplc fihn IS obtamed
Via-holes are etched m the SlO layer by reactive-ion
etchmg m a CF,/O, atmosphere The permalloy 1s
backsputtered to provide a clean surface for the
contacts Stable contacts with low mterdfluslon at
high temperatures can be prod& using MO and
Cr as a contact matetral [ 121 We use a 25-nm
chrommm film followed by a 0 5-pm alummum
layer The ahmunum and chrommm are wet-etched
using the same photoreslst mark The wafer 1s shced
and the devices are mounted on a substrate
StdllQ
The stablhty of the electrical and magnetic
properties of the devices can be determmed m two
ways In order to determme the behavior of the
devices at merent temperatures from 20 “C to
120 “C, the devices were mounted on a heating
plate m an angle detector set-up [lo] Altema-
tlvely, the devices were heated to 125 “C and their
properties measured after 1, 2 and 7 days All
measurements were performed on PHDs of 3 x 3
and 1 x 1 mm
Concermng the magnetic amsotropy m the
fihns, It could be concluded that small amsotropy
could be induced by anneahng at 125 “C m a
magnetic field The resultmg values of (rp - 6),,
m different fihns m a field of 20 kA/m were deter-
mmed to be well below 0 1” for all fihns
Devices Hrlth a large overlap of the contact
matenal over the SlOcovered permalloy showed
large, irreversible changes m signal amphtude and
offset durmg heat treatment, presumably due to
the fact that the S10 layer was not completely
closed The followmg results concern devices urlth
a small overlap which showed good stablhty
Amphtude imbalance between two as-manufac-
tured devices m an angle detector 1s usually of the
order of l-5% Temperature-induced imbalance
m the temperature range of 20 to 120 “C shows a
maxlmum of 0 5%, yleldmg a measurement error
of less than 0 1’
Offset stablhty m the deuces IS very cntical,
because of the sensltlvlty of the offset voltage to
geometrical vanahons The offset was tnmmed
to zero at 20 “C using the correction clrcmt shown
m Ag 3 In subsequent measurements, the
tmnmed devices showed an offset of less than 1%
m the temperature range 20 “C to 120 “C The
offset changes are reversible and are atulbuted to
positiondependent changes m the properties of
Rg 3 Offset correction clrcult
the contacts The changes m offset after anneahng
at 125 “C for longer times are generally lower than
0 75% We can conclude that offset attnbutes less
than 1 75% to the signal under normal condltlons
m the temperature range 20 “C to 120 “C, leadmg
to angle detector measurement errors of 0 5” max-
lmum
It IS concluded that the offset stablhty 1s the
hmltmg property with regard to the accuracy of
the angle detector Further research ~11 focus on
this problem
c0ne1usioos
With the techniques descr&d above, angle
detectors wth an accuracy of 0 5” over a tempera-
ture range of 20 to 120 “C can be produced There
1s no mdlcation that the temperature range cannot
be extended to lower temperatures The accuracy
1s mamly hmited by temperature-induced changes
m the offset of the devices If the technology can
be adapted to provide more stable contacts, an
accuracy of 0 2 to 0 1” should be attainable
Ackoowkdgemeot
This work was supported by the Netherlands
Technology Foundation (STW)
1 K J M &kel, A thm-tilm magnetomuve angle detec-
tor, P/ID Thesrr, Umverslty of Twente, Enschede, The
Netherlands, 1988
2 C J M Btikel and J H J Fhutman. Contacbs angle
transducer, US Patent A&c 06 913 760 (1986)
3 K J M EiJkel and J H j- Flmtman, Opt&&m of the
response of magnetoreslstive elements, IEEE Trans hfagn ,
(1989) m press
4 R F Soohoo, Magnetu: Thm Fdms, Harper and Row,
New York, 1965
5 J W W&erbk and K J M Eqkel, Permalloy multiy-
ers to reduce the effects of umaxtal amsotropy, IEEE
Truns Magn , (1989) m press
6 M Takayasu, S Uchquna, K Takaha&u and T FUJU,
Isothermal anneahng behawor of zero-magnetostmtive
pennalloy lib, IEEE Trans Magn MAG-ZO (1974) 552-
555
7 M Goto, H Tange, T Kamtmon and H Kubo, Dead
amsotropy layer of IO-permalloy films evaporated m mag-
netlc field, 12th Znt Colloq Magnerrc Fdms and Surfaces
(ICMFS), Frunce, 1988, ThP-15
8 W Metzdorf, I)le hnflussbarkat der umaxlalen mag-
netlschen Amsotropte aufgedampfter Permalloyschxhten
durch Tempern nn Magnetfeld, Z Angew Phys , 18 (1964)
534-540
9 T S Crowther, Angular and magmtude dqnxslon of the
amsotropy in magnettc films, J Appl Phyx , 34 (1%3)
580-586
10 K J Eqkel, Measurement of the amsotropy m permalloy,
IEEE Trans Magn , MAG-24 (1988) 1957- 1959
11 A D Butherus and S Nakahara, Anneahng kmettcs of
thm permalloy films, IEEE Trans A&n, MAG-21 (1985)
1301-1305 -
-
12 M I(ltada and H Yamamoto, Reactlon between pennal-
loy and several thm metal 6hns, Thm Solui Fdmt, 12:
(1984) 173-182
