6, JUNE 2009Optimization of Spin-Valve Structure NiFe/Cu/NiFe/IrMn for Planar Hall Effect Based Biochips Bui Dinh Tu1, Le Viet Cuong2, Tran Quang Hung3, Do Thi Huong Giang1, Tran Mau Dan
Trang 12378 IEEE TRANSACTIONS ON MAGNETICS, VOL 45, NO 6, JUNE 2009
Optimization of Spin-Valve Structure NiFe/Cu/NiFe/IrMn
for Planar Hall Effect Based Biochips
Bui Dinh Tu1, Le Viet Cuong2, Tran Quang Hung3, Do Thi Huong Giang1, Tran Mau Danh1,
Nguyen Huu Duc1;2, and CheolGi Kim3
Department of Nano Magnetic Materials and Devices, Faculty of Physics Engineering, College of Technology,
Vietnam National University, Hanoi, Vietnam Laboratory for Micro-Nano Technology, College of Technology, Vietnam National University, Hanoi, Vietnam
Department of Materials Science and Engineering, Chungnam National University, Yuseong, Daejeon 305-764, Korea
This paper deals with the planar Hall effect (PHE) of Ta(5)/NiFe( F)/Cu(1.2)/NiFe(P)/IrMn(15)/Ta(5) (nm) spin-valve structures.
Experimental investigations are performed for 50 m 50 m junctions with various thicknesses of free layer (F= 4 8 10 12 16 26
nm) and pinned layer ( P = 1 2 6 8 9 12 nm) The results show that the thicker free layers, the higher PHE signal is observed In
addition, the thicker pinned layers lower PHE signal The highest PHE sensitivity of 196 V/(kA/m) is obtained in the spin-valve configuration with F= 26 nm and P= 1 nm The results are discussed in terms of the spin twist as well as to the coherent rotation of
the magnetization in the individual ferromagnetic layers This optimization is rather promising for the spintronic biochip developments.
Index Terms—Biosensors, Hall effect, magnetization reversal, magnetoresistance, magnetoresistive devices.
I INTRODUCTION
T HE discovery of giant magnetoresistance (GMR) effect
in metallic multilayer has made a strong impact on the
development of computer memory technologies [1], [3]
Recently, this effect has been well developed for biochip
ap-plications due to its large resistance change in small magnetic
field range [4]–[10] The GMR effect is related to the switching
of magnetic domain It has low signal-to-noise ratio (SNR),
leading to a high error in detections of the small stray field
The planar Hall effect (PHE), however, is related to the rotation
process of magnetic domain and is originated as the anisotropic
magnetoresistance This effect exhibits a nano-Tesla sensitivity
and rather high SNR, so it has received great attention for
mag-netic bead detection and biosensor design [2]–[6] The
trans-verse voltage on a planar Hall cross depends on the
orienta-tion of the magnetizaorienta-tion of the ferromagnetic layer with
re-spect to the longitudinal sensing current Thus, a large PHE
is expected to be observed in exchange coupling based
struc-tures because they can ensure a sufficient uniaxial anisotropy
with well-defined single domain state to introduce a
unidirec-tional anisotropy For this purpose, Ejsing et al [6], [7] have
reported a single PHE sensor of NiFe/IrMn/NiFe Furthermore,
a PHE magnetic bead array counter microchip integrated 24 of
single sensors based on a simple NiFe/IrMn bilayer structure has
been successfully prepared [8] Recently, Thanh et al [9] have
found that the sensor signal can be further improved by using
spin-valve structure of NiFe(6)/Cu(3.5)/NiFe(3)/IrMn(10) (nm)
with the size of 3 m 3 m when detecting the 2.8 m
mag-netic beads
The present paper deals with studies of the magnetic field
sen-sitivity as a function of the thickness of the individual free
fer-romagnetic (FFM) and the pinned ferfer-romagnetic (PFM) layers
Manuscript received October 09, 2008 Current version published May 20,
2009 Corresponding author: N H Duc (e-mail: ducnh@vnu.edu.vn).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TMAG.2009.2018580
in the pattern 50 m 50 m PHE sensors based on NiFe/Cu/ NiFe/IrMn spin-valve structures The highest PHE sensitivity
of 196 V/(kA/m) was obtained in the spin-valve configuration with FFM layer thickness of 26 nm and PFM layer thickness of
1 nm This optimum structure is proposed to apply for magnetic bead detections
II EXPERIMENTALPROCEDURE
The thin films with typical spin-valve structure of Ta(5)/NiFe /Cu(1.2)/NiFe /IrMn(15)/Ta(5) (nm) with
layer thickness , nm are fabricated by dc magnetron sputtering system with the base pressure less than
mTorr The spin-valve structures were sputtered on SiO wafer at room temperature with Argon working pressure
of mTorr During sputtering process, a uniform magnetic field of A/m is applied in plane of
the films, parallel to the Ox direction This magnetic field
induces a magnetic anisotropy in the FFM and PFM layers and then aligns the pinning direction of the antiferromagnetic (AFM) Ir-Mn layer The PHE sensors were structured by using photolithography technique into four-electrode bars with the patterned size of 50 m 50 m [Fig 1(a)] The sensors were passivated by sputtering a 150 nm thick Si N layer to protect against the fluid used during the experimentation The bead array counter (BARC) microchip was fabricated by integrating
10 single sensor patterns as shown in Fig 1(b)
The PHE characteristics of sensors were measured at room temperature by using a nanovoltmeter in the external magnetic fields up to 4 kA/m applied along Oy direction and sensing
currents of 1 mA Longitudinal magneto-resistance was mea-sured by means of a collinear four-point probe method, for sam-ples with the size of 2 mm 10 mm, with applied magnetic field and sensing current are along direction Magnetization was measured by using a Lakeshore 7400 vibrating sample magne-tometer (VSM) on films
0018-9464/$25.00 © 2009 IEEE
Trang 2Fig 1 (a) Top view micrograph of the single 50 m 2 50 m PHR cross The
pinning direction H as well as the direction of the bias field H and sensing
current I are indicated (b) The bead array counter microchip including 10 of
single PHE sensors (with 8 single sensors in the two middle lines and 1 single
sensor in each edge line).
III RESULTS ANDDISCUSSION
A Fixed PFM Layer Thickness Spin-Valve System
Fig 2(a) presents the magnetization data of spin-valve Ta(5)/
NiFe /Cu(1.2)/NiFe(2)/IrMn(15)/Ta(5) (nm) structures with
different free layer thicknesses varying from 4 to 16 nm It
is clearly seen that all the samples exhibit two hysteresis loops
The magnetization accounting from the first loop linearly
in-creases with increasing while that from the second loop is
almost constant These two hysteresis loops are attributed to
the FFM and PFM layers, respectively The FFM layer is
ex-pected to dominate the sensor response at low magnetic fields
The values of the coercivity and exchange coupling
fields determined from the first hysteresis loop [see insert in
Fig 2(a)] are collected and listed in the Table I Note that,
exper-imentally, shows a tendency to increase with while
seems to exhibit a maximum at nm
Illustrated in Fig 2(b) are the GMR data of the samples under
investigation It can be seen from this figure that the reversal of
the magnetization in the FFM layer brings the GMR ratio to its
maximal value of 1.6% for nm This maximum lightly
decreases with the increasing of the FFM layer thickness and
equals to 1.2% for nm Further increasing the
(de-magnetizing) magnetic field, the magnetization rotation in PFM
layers starts rather early in the thin FFM layer thickness
sam-ples and later in the thick FFM layer thickness samsam-ples
How-ever, the final parallel configuration of individual layer
magne-Fig 2 Magnetic hysteresis loops (a) GMR data (b) and low field PHE profiles (c) of spin-valve structures with the fixed PFM layer thickness t = 2 nm and FFM layer thickness (t ) varying from 4 to 16 nm.
tizations seems to be completed at the same magnetic field of
kA/m for all samples This finding is comparable with the magnetization data mentioned in Fig 2(a) These GMR results are consistent with those reported in [11]
Shown in Fig 2(c) are the PHE voltage profiles, , as
a function of the applied field First, the PHE voltage initially develops rather fast at low fields reaching a maximal value at
A/m and finally decreases with further increasing
in the magnetic fields It is interesting to note that the magnetic
Trang 32380 IEEE TRANSACTIONS ON MAGNETICS, VOL 45, NO 6, JUNE 2009
TABLE I
V ALUES S ENSOR S ENSITIVITY (S), C OERCIVE (H ), A NISOTROPY (H ),
E XCHANGE C OUPLING (H ) F IELDS , AND M AXIMAL F IELD (H ) AT GMR
P EAK FOR S PIN -V ALVE S YSTEM W ITH D IFFERENT F REE L AYER T HICKNESSES
field, at which GMR reaches the maximum , is
systemat-ically close to the sum of (see Table I) For this fixed
PFM layer spin-valve system, the maximal value of the PHE
voltage increases with increasing FFM layer thickness It
in-creases from the value of 15 V for the sample with nm
to the value of 48 V for nm Consequently, the sensor
sensitivity ( , see below) is enhanced from the value
of 21.4 V/(kA/m) to 95.5 V/(kA/m), respectively (Table I)
It is well known that when the magnetization vector makes
an angle with an easy axis along the direction (and/or with
), the transverse induced PHE voltage (or ) parallel
with direction is given as follows:
(1)
resis-tivity measured with the current parallel and perpendicular to the
magnetization, respectively; is the free ferromagnetic layer
thickness
Typically, these curves are fitted well by using the
single domain model with the magnetic energy per unit of the
magnetic layer In this case, the Stoner-Wohlfarth energy can be
expressed as [12]
(2) Here, the and are the angles between the magnetization
of the free and pinned layers and the easy axis direction,
respec-tively; is the effective anisotropy constant,
is the saturation magnetization of the free layer, and is the
interlayer coupling constant that can be extracted from the
re-lation with the exchange coupling field between two FM layers
If the exchange bias field between PFM and AFM layers is
strong enough, the angle between magnetization and the easy
axis direction of the PFM layer will be fixed at low applied
mag-netic fields, i.e., equals to zero This can be applied for the
present case, where the magnetization reversal of the free and
pinned layers occurred separately [see in Fig 2(a)]
For small angles, , the PHE voltage exhibits a linear
characteristics as well as high sensitivity at low fields (
A/m) In this case, the sensitivity of sensor is given as
(3)
Applying this theoretical approach to experimental data, we
can determine the values for and fields as well as the
Fig 3 Hysteresis loops (a) GMR data (b) and low field PHE profiles (c) mea-sured in spin-valve structures with the fixed free layer thickness x = 10 nm and different pinned layer thickness (t) from 2 to 12 nm.
sensor sensitivity The obtained results of is also sum-marized in Table I Note that the values of and obtained from the fits of the PHE data are in excellent agreement with those derived from experimental data The calculated values
of , however, are systematically larger than determined from the magnetization measurements The increasing of the sensitivity in these sensor junctions is usually explained simply
by the shunting current in the spin valve thin films The more FFM layer is thick, the more shunting current from other layers
is small Other explanations will be extended below
Trang 4TABLE II
V ALUES S ENSOR S ENSITIVITY (S), C OERCIVE (H ), A NISOTROPY (H ),
E XCHANGE C OUPLING (H ) F IELDS AND M AXIMAL F IELD (H ) AT GMR
P EAK FOR S PIN -V ALVE S YSTEM W ITH D IFFERENT F REE L AYER T HICKNESSES
B Fixed FFM Layer Thickness Spin-Valve System
Fig 3(a) presents the magnetization data of Ta(5)/NiFe(10)/
Cu(1.2)/NiFe /IrMn(15)/Ta(5) (nm) spin-valve structures
with different PFM layer thickness varying from 1 to
12 nm Here, all samples exhibit the two hysteresis loops too
However, contrary to the fixed PFM layer thickness system, the
magnetization accounting from the first loop is almost constant
while that from the second loop increases with increasing
The values of the coercive and exchange coupling
fields determined from the first hysteresis loop are collected
and listed in the Table II Note that lightly varies around
the value as small as 80 kA/m, whereas strongly increases
with increasing
Typical magnetoresistive characteristics of spin-valve
struc-tures are presented in Fig 3(b) The magnetic field interval for
the existence of the antiparallel configuration between FFM and
PFM layer magnetizations decreases with increasing the PFM
thickness The final parallel configuration is completed at the
, and nm, respectively This behavior is consistent
with the magnetization data reported in Fig 3(a) The maximal
GMR ratio, however, increases from 0.85% to 2.84% when
increases from 1 to 12 nm
Shown in Fig 3(c) are the PHE voltage profiles as a
func-tion of the applied fields For this fixed free layer spin-valve
system, it is clear that with increasing , the maximal value
of the PHE voltage decreases In addition, this peak (at )
shifts to higher magnetic fields and once again the relation
be-tween and the sum of is found (see Table II)
Consequently, the sensor sensitivity is strongly reduced from
the value of 110.6 to 42.7 V/(kA/m) when increases from
1 to 12 nm The values of anisotropy , exchange coupling
fields and sensor sensitivity derived from the
theoret-ical fits show an excellent consistence with experimental results
While a rather large difference between and (calculation)
is observed for nm
C Optimal Spin-Valve Structure for PHE Sensor Sensitivity
It was provided from above mentioned investigations that the
large PHE sensor sensitivity can be reached in spin-valve
tures with thin PFM and thick FFM layers In spin-valve
struc-tures, the PHE is strongly contributed from the FFM layer
By increasing the thickness of this layer and optimizing the
thicknesses of other layers, the shunting current can be reduced
through remain layers, leading to the observed higher sensitivity
of our PHE sensors On the other hand, the high PHE sensitivity
may also be related to the spin twist as well as to the coherent
Fig 4 Hysteresis loops (a), GMR data (b) and low field PHE profiles (c) mea-sured in Ta(5)/NiFe(26)/Cu(1.2)/NiFe(1)/IrMn(15)/Ta(5) (nm) spin-valve struc-ture, i.e., with t = 26 nm and t = 1 nm.
rotation of the magnetization in the individual ferromagnetic layers This can be understood as follows In the PFM layer, the well-aligned spin part is usually formed near PFM/AFM in-terface Further increasing the pinned layer thickness will lead
to an enlarging of the twist structure where the magnetization is pinned in different directions from the easy axis (i.e.,
[13] In this context, the twisted part can be assumed to be elim-inated in the structure with thin pinned layer nm Prac-tically, the maximal PHE voltage and the highest sensitivity of
Trang 52382 IEEE TRANSACTIONS ON MAGNETICS, VOL 45, NO 6, JUNE 2009
sensor were observed in this configuration For the FFM layers,
the magnetic influence and then the twist part can be
estab-lished near NM/FFM interface only The thick free layers thus
dominate the collinear ferromagnetic part and enhance the PHE
voltage
Combining these two optimal tendencies, we prepared the
Ta(5)/NiFe(26)/Cu(1.2)/NiFe(1)/IrMn(15)/Ta(5) (nm)
spin-valve structure, i.e., with nm and nm Its
magnetization and PHE data are presented in Fig 4 Although
the magnetization reversal is mainly contributed to the first
magnetic hysteresis loop [Fig 4(a)], the rotation of the
magneti-zation in the PFM layer to re-establish the parallel configuration
is well evidenced in the magnetoresistance [Fig 4(b)] Here,
the most interesting result is that the PHE voltage reaches
its maximal value of about 62 V at A/m and this
spin-valve configuration shows a sensor sensitivity as large as
196 V/(kA/m) Additionally, the values of and fields
are as small as 160 and 330 A/m, respectively
IV CONCLUSION
The influence of the individual free and pinned layer
thick-ness on the sensitivity of PHE sensor based on the spin-valve
structure of NiFe /Cu(1.2)/NiFe /IrMn(15) (nm) with
size of 50 m 50 m has been studied The results show that
the thicker free ferromagnetic layers enhance the PHE signal,
whereas the thicker pinned ferromagnetic layers lower the PHE
one For a good combination, the highest PHE sensitivity of
196 V/(kA/m) is obtained in the spin-valve configuration
with nm and nm The results are discussed
in terms of the spin twist as well as to the coherent rotation
of the magnetization in the individual ferromagnetic layers
This optimization is rather promising for the spintronic biochip
developments
ACKNOWLEDGMENT
This work was supported by Vietnam National University, Hanoi under Grant QG.TD 07.10
REFERENCES
[1] M Johnson, Magnetoelectronics. Amsterdam, The Netherlands: El-sevier, 2004.
[2] S Maekawa, Concepts in Spin Electronics. Oxford, U.K.: Oxford Science Publications, 2006.
[3] C Chappert, A Fert, and F N Van Dau, “The emergence of spin
elec-tronics in data storage,” Nature Mater., vol 6, pp 813–823, 2007.
[4] A Schuhl, F N Van Dau, and J R Childress, “Low-field magnetic.
sensors based on the planar Hall effect,” Appl Phys Lett., vol 66, pp.
2751–2753, 1995.
[5] N V Dau, A Schuhl, J R Childress, and M Sussiau, “Magnetic
sen-sors for nanotesla detection using planar Hall effect,” Sens Actuators
A, vol 53, pp 256–260, 1996.
[6] L Ejsing, M F Hansen, A K Menon, H A Ferreira, D L Graham, and P P Freitas, “Planar Hall effect sensor for magnetic micro-and
nanobead detection,” Appl Phys Lett., vol 84, pp 4729–4731, 2004.
[7] L Ejsing, M F Hansen, A K Menon, H A Ferreira, D L Graham, and P P Freitas, “Magnetic microbead detection using the planar Hall
effect,” J Magn Magn Mater., vol 293, pp 677–684, 2005.
[8] B D Tu, T Q Hung, N T Thanh, T M Danh, N H Duc, and C.
G Kim, “Planar Hall bead array counter microchip with NiFe/IrMn
bilayers,” J Appl Phys, vol 104, p 074701, 2008.
[9] N T Thanh, B P Rao, N H Duc, and C G Kim, “Planar Hall
resis-tance sensor for biochip application,” Phys Stat Sol., A, vol 204, pp.
4053–4057, 2007.
[10] T Q Hung, P H Quang, N T Thanh, S J Oh, B Bharat, and C.
G Kim, “The contribution of the exchange biased field direction in
multilayer thin films to planar Hall resistance,” Phys Stat Sol., B, vol.
244, pp 4431–4434, 2007.
[11] B Dieny, V Speriosu, S S P Parkin, B A Gurney, P Bumgart, and
D R Wilhoit, “Magnetotransport properties of magnetically soft
spin-valve structures,” J Appl Phys., vol 69, pp 4774–4779, 1991 [12] R C O’Handley, Modern Magnetic Materials. New York: Wiley, 2000.
[13] S Wang, Y Xu, and K Xia, “First-principles study of spin-transfer
torques in layered systems with noncollinear magnetization,” Phys Rev B, vol 77, p 184430, 2008.