i
THE EFFECT OF SAMPLE PREPARATION PARAMETERS ON
MAGNETORESISTANCE RATIOS (MR%) IN CO/CU
NANOSTRUCTURES
LAU YEE CHEN
A thesis submitted in fulfilment of the
requirements for the awards of the degree of
Master of Science (Physics)
Faculty of Science
Universiti Teknologi Malaysia
MARCH, 2005
BAHAGIAN A – Pengesahan Kerjasama *
Adalah disahkan bahawa projek penyelidikan tesis ini telah dilaksanakan melalui
kerjasama antara _____________________ dengan _________________________
Disahkan oleh:
Tandatangan : ..........................................................
Nama
: ..........................................................
Jawatan
:...........................................................
Tarikh : ..........................
(Cop rasmi)
* Jika penyediaan tesis/projek melibatkan kerjasama.
BAHAGIAN B – Untuk Kegunaan Pejabat Sekolah Pengajian Siswazah
Tesis in telah diperiksa dan diakui oleh:
Nama dan Alamat
Pemeriksa Luar
:
Nama dan Alamat
Pemeriksa Dalam I
:
Pemeriksa Dalam II
:
Nama Penyelia Lain
(jika ada)
:
Disahkan oleh Penolong Pendaftar di SPS:
Tandatangan : ..........................................................
Nama
Tarikh : ..........................
: ........................................................................
iii
To my family. Without their continual support and encouragement this work
would not have been possible.
iv
ACKNOWLEDGEMENT
I would like to express my sincere appreciation to P. M. Dr. Rashdi Shah
Ahmad and Prof. Dr. Samsudi Sakrani, for their excellent guidance, instrumental
assistances and valuable advice which made it possible for me to implement research
leading to this thesis. The valuable advice and opinions from P. M. Dr. Yussof
Wahab, Mr. Mohammad Zaki Hj. Yaacob and some other lectures are also very
much appreciated.
In addition, I am grateful to Dr. Agus Setyo Budi, Mr. Mohd. Nazari
Kamaruddin, Madam Wan Aklim Norsalafiany Wan Ahmad, Mr. Putut Marwoto,
Mr. Md. Sam Ismom, Ms. Carmen Wong and Mr. Hasbullah Antony Hasbi for their
kind assistance in experimental, constructive ideas and valuable suggestion.
The scholarship award and financial support for this research from Universiti
Teknologi Malaysia are really rewarding.
v
ABSTRACT
The research reported in this thesis is primarily aimed at establishing the
fundamental understanding of magnetoresistance (MR) phenomena occurring in
layered magnetic nanostructures of Co/Cu system fabricated using sputtering and
electron beam method. Emphasis is given on the studies of magnetoresistance ratios
(MR%) as functions of Co layer thickness, working pressure, annealing time and
temperature, number of bilayer, direction of magnetic fields, and the application of
buffer layer. The Co/Cu/Co sandwiches in this study were fabricated on corning
glass substrates. The electrical resistance of samples was measured using the four
point Van der Pauw method when magnetic fields of ± 2500 gauss were applied. It
was observed that, the MR% attained almost 10% between 2 - 6 nm of the Co layer
thickness. By varying the working pressure, a maximum MR% of 11.4% was
obtained at a working pressure of 2.6 x 10-3 torr. In the other hand, the MR% also
increases with the increasing of annealing temperature and time. In the bilayers
number, n various MR% was revealed by the existence of up-down fluctuations with
the MR’s peak and valley occurring at n = 5 and n = 8, respectively. It was also
observed that, the magnetic field applied in plane to the samples with and without
chromium buffer layer produced higher value MR% of compared to those applied
perpendicularly. Thus, the results indicate the dependent of MR% on various
preparation parameters.
vi
ABSTRAK
Laporan penyelidikan dalam tesis ini bermatlamat untuk menghasilkan
pemahaman asas mengenai fenomena magnetoresistance (MR) dalam struktur Co/Cu
yang dihasilkan melalui kaedah sputtering dan electron beam. Penekanan diberikan
terhadap faktor-faktor ketebalan lapisan Co, tekanan semasa proses pemendapan,
masa dan suhu pemanasan, bilangan lapisan saput tipis, arah pembekalan medan
magnet, dan lapisan buffer yang mempengaruhi nilai-nilai nisbah magnetoresistance
(MR%). Lapisan Co/Cu/Co dalam pengajian ini dimendapkan ke atas kaca corning.
Rintangan sampel diukur dengan menggunakan kaedah Van der Pauw apabila medan
magnet berjumlah ± 2500 gauss dikenakan. MR% didapati meningkat ke hampir
10% apabila ketebalan lapisan Co berada di antara 2-6 nm. Dengan mengubah
tekanan semasa proses pemendapan, nilai maksimum MR% berjumlah 11.4% telah
dihasilkan pada tekanan 2.6 x 10-3 torr. Selain daripada kesan ketebalan Co dan
tekanan semasa pemendapan, proses pemanasan juga turut meningkatkan nilai MR%.
Kesan bilangan lapisan sampel telah menghasilkan bentuk turun naik dengan puncak
dan lembah MR% masing–masing muncul semasa n = 5 dan n = 8. Disamping itu,
MR% yang lebih tinggi dapat dihasilkan dengan mengenakan medan magnet dalam
arah mendatar kepada sampel, sama ada terdapat lapisan buffer atau tidak. Oleh itu,
hasil penyelidikan ini menunjukkan bahawa MR% adalah dipengaruhi oleh pelbagai
cara penyediaan sample.
vii
TABLE OF CONTENTS
CHAPTER
1
2
TITLE
PAGE
DECLARATION
ii
DEDICATION
iii
AKNOWLEDGEMENT
iv
ABSTRACT
v
ABSTRAK
vii
CONTENTS
ix
LIST OF TABLES
xii
LIST OF FIGURES
xiii
LIST OF SYMBOLS
xvi
LIST OF APPENDIX
xviii
INTRODUCTION
1.1
Literature Review
1
1.2
Research Objective
3
1.3
Research Scope
3
1.4
Thesis Plan
4
THEORY
2.1
Magnetic Material
6
2.1.1
Ferromagnetic Material
6
2.1.2
Antiferromagnetic Material
8
viii
2.2
Band Structure
9
2.3
Thin Film Deposition
10
2.3.1
Radio Frequency (RF) Sputtering
13
2.3.2
Electron Beam (e-beam) Process
16
2.4
2.5
Thickness Measurement
17
2.4.1
17
2.4.2 Dektak3 Surface Profiler
19
The Four-Point Probe
20
2.5.1
2.6
Resistiviy of Arbitrarily Shaped
Samples
20
Giant Magnetoresistance
22
2.6.1
Theoretical Model
23
2.6.1.1 Single-Current Model
23
Benefit Of GMR
26
2.6.2
3
Film Thickness Monitor (FTM)
METHODOLOGY
3.1
Sample Preparation
29
3.1.1 Deposition by RF Sputtering Method
29
3.1.2
3.1.1.1 High Vacuum Coater Setup
30
3.1.1.2 Substrate Pre-Clean
33
3.1.1.3 Pre-sputtering Process
33
3.1.1.4 RF Sputtering Process
33
Deposition by Electron Beam Method
35
3.1.2.1 Edwards Auto 306 Evaporation
Systems
3.2
3.3
36
3.1.2.2 Substrate Pre-Clean
37
3.1.2.3 Electron Beam Evaporation Process
37
Annealing Process
38
3.2.1
40
Temperature Uncertainty Calibration
Measurement
41
3.3.1
Thickness Measurements
41
3.3.1.1 Measurement by Using FTM
42
3.3.1.2 Measurement by Using Dektak3
ix
Surface Profiler
3.3.2
4
MR Measurement
42
45
RESULT AND DISCUSSIONS
4.1
Magnetoresistance for RF Sputtering Film
48
4.1.1
Magnetoresistance (MR) Curve
48
4.1.1.1 Effect of Sample Thickness
50
4.1.1.2 Effect of Working Pressure
54
4.1.1.3 Effect of Bilayers
58
4.1.2 Effect of Annealing Process
61
4.1.2.1 Annealing Time
61
4.1.2.2 Annealing Temperature
64
4.1.2.3 Effect of Annealing Temperature
4.2
4.3
5
for Different Bilayers
68
Magnetoresistance for e-beam Film
71
4.2.1 Effect of Magnetic Field
71
4.2.2
Effect of Buffer Layers
73
4.2.3
Effect of Annealing Time on e-beam Film
76
Comparison Between Sputtering and e-beam Method 78
CONCLUSION AND SUGGESTION
5.1
Conclusion
81
5.2
Suggestion
83
REFERENCES
85
APPENDICIES
Appendix A
94
PRESENTATIONS
96
x
LIST OF TABLE
TABLES NO.
TITLE
PAGE
1.1
Material and their function in system
4
3.1
Label of samples prepared by RF sputtering
35
3.2
Parameters of samples prepared by e-beam method
38
3.3
Parameters of annealing for the sample prepared by
RF sputtering method
3.4
Annealing parameter for sample prepared by e-beam
method
3.5
4.1
40
Thickness detected by using FTM and Dektak3
Surface Profiler
3.6
39
45
Current and resistance values for current source
testing
46
Working pressure and deposition rate
55
xi
LIST OF FIGURE
FIGURE NO.
2.1
TITLE
A typical hysteresis loop of antiferromagnetic
material
2.2
PAGE
7
Variation with the temperature of the susceptibility
for an antiferromagnetic.
8
2.3
Moment spin of an antiferromagnet
9
2.4
Thin film processes
12
2.5
Schematic of the ion-solid interactions and the
sputtering process
13
2.6
The schematic of RF sputtering system
14
2.7
Basic configuration of e-beam
17
2.8
Film Thickness Monitor
18
2.9
Schematic of measurement for Film Thickness Monitor
18
3
2.10
Dektak Surface Profiler
19
2.11
A Collinear Four-point Probe
20
2.12
Four-point Van der Pauw method
21
2.13
The magnetic multilayer type, in which the
magnetizations are forced from natural AF-mode
(θ = 0°) to F-mode (θ = 180°) by H
23
2.13
GMR phenomena showing a) low and b) high resistance
24
2.14
Resistance effectiveness in parallel configuration
25
2.15
Resistance effectiveness in anti-parallel configuration
25
2.16
Basic IBM suspended head design
27
3.1(a)
High Vacuum Coater
31
3.1(b)
Control panel of High Vacuum Coater
32
3.2
Internal part of High Vacuum Coater
32
xii
3.3
Direction of magnetic fields applied to samples
36
3.4
Edwards Auto 306 Evaporation Systems
37
3.5
Set up of annealing process
40
3.6
Graph of quartz temperature versus heater set point
41
3.7
The straight line used for thickness measurement
43
3.8(a)
3
Thickness of (Co/Cu) x 5 measured by Dektak
Surface Profiler
3.8(b)
Thickness of (Co/Cu) x 10 measure by Dektak3
Surface Profiler
3.8(c)
44
44
Thickness of (Co/Cu) x 15 measure by Dektak3
Surface Profiler
45
3.9
Magnetic fields applied in plane to sample
47
4.1
Magnetoresistance curve of Co /Cu /Co (6nm/
2.5nm/ 6nm) sandwich structures
4.2
49
Magnetoresistance curve for Co/Cu for 6 various
thickness of Co layer
51
4.3
Graph of MR% versus thickness
52
4.4
Graph of resistance versus film thickness of Co layer
53
4.5
Graph of resistance change versus film thickness
of Co layer
54
4.6
Effect of working pressure on MR%
56
4.7
Effect of working pressure on resistance of samples
57
4.8
Graph of resistance change versus working pressure
57
4.9
Effect of number of bilayers on MR%
59
4.10
Graph of resistance versus number of bilayers samples
60
4.11
Graph of resistance change versus number of bilayers
samples
60
4.12
Influence of annealing time on MR%
62
4.13
Graph of resistance versus annealing time
63
4.14
Graph of resistance change versus annealing time
63
4.15
Effect of annealing temperature as a function of MR%
in Co/Cu
4.16
65
MR% effect of annealing temperature as a function of
MR% in NiFe/Cu
65
xiii
4.17
Graph of resistance versus annealing temperature
66
4.18
Graph of resistance change versus annealing temperature
67
4.19
Graph of resistance of NiFe/Cu versus annealing
temperature (°C).
4.20
67
Effect of number bilayers of Co/Cu before and after
Annealing at 400°C towards MR%
69
4.21
Graph of resistance versus number bilayers of samples
70
4.22
Magnetoresistance curve of Co/Cu/Co (5.5 nm
/3.5 nm/5.5 nm) when magnetic fields applied along
and perpendicular to the sample
72
4.23
Easy and hard axis in Co hexagonal crystal lattice
72
4.24
Magnetoresistance Curve of Co/Cu/Co (5.5 nm/
3.5 nm/5.5 nm) with and without buffer layers (Cr) layer
4.25
Dependence of MR% on Cr buffer layer thickness in
Co/Cu/Co (5.5 nm/3.5 nm/5.5 nm) sandwich structures
4.26
74
75
Magnetoresistance curve of Cr/Co/Cu/Co (8 nm/5.5 nm
/3.5 nm/5.5 nm) when magnetic fields applied along and
perpendicular to the sample
4.27
7c
Magnetoresistance Curve of Co/Cu/Co (12 nm/2.5 nm
/12 nm) with different annealing time
77
4.28
Effect of annealing on Co/Cu/Co (12 nm/2.5 nm/12 nm)
77
4.29
Resistance of Co/Cu/Co (12 nm/2.5 nm/12 nm) in
different annealing time
4.30
78
Magnetoresistance Curve of Co/Cu/Co (12 nm/2.5 nm
/12 nm) prepared by RF sputtering and e-beam
method at two different working pressures.
80
xiv
LIST OF SYMBOLS
GMR
Gaint Magnetoresistance
MR
Magnetoresistance
MR%
Magnetoresistance ratios
SSF
Surface Spin-Flop
H
Magnetic field
M
Magnetization
µ0
Permeability of free space
β
Bohr magnetron
B
Magnetic induction
χ
Susceptibility
TN
Neel temperature
EF
Fermi Energy
n
Number of multilayers
ρ
Resistivity
Rmin
Resistance in maximum external
magnetic field
Rmax
Resistance in zero field
Rtotal
Total resistance
∆R
Distance between Rmax and Rmin
t
Thickness
tCo
Thickness of Co layers
RF
Radio frequency
e-beam
electron beam
xv
HVC
High Vacuum Coater
Auto 306
Edwards Auto 306 Evaporation Systems
Sccm
Standard cubic centimeter per minute
FTM
Film Thickness Monitor
Z value
Acoustic impedance
3
DEKTAK
Dektak3 Surface Profiler
LVDT
Linear Variable Differential
Transformer
RAM
Random access memory
MRAM
Magnetic RAM
FM-layers
Ferromagnetic layers
NM-layers
Non-magnetic spacer layers
P
Paramagnetic
F-mode
Ferromagnetic mod
AF-mode
Anti-ferromagnetic mode
α
Direction of magnetic field
Co
Cobalt
Cu
Copper
Cr
Chromium
Fe
Iron
Ni
Nickel
NiFe
Nickel Iron
GaAs
Gallium Arsenide
xvi
LIST OF APPENDICIES
APPENDIX NO.
A
DESCRIPTION
Temperature Uncertainty Calibration
PAGE
94
1
CHAPTER 1
INTRODUCTION
1.1
Literature Review
The capacity of magnetic disk systems is growing rapidly year by year with
the advancement in the information-oriented society. From a statistical survey, the
density of magnetic recording is increased by 60% every year (Sato, 1998).
Magnetic thin films have been of great interest recently due to their technological
application in magnetic sensors and magnetic random access memory (MRAM)
modules (Timothy, 2001; and Yamada et al., 2002).
The issue of sensitivity has drawn a lot of attention in giant
magnetoresistance (GMR) materials for applications in sensors, high-density readout heads and other magnetic storage technology in this decade. Thus, the study of
GMR effect was essential to investigate the thin film condition, which can produce
higher GMR values while decreasing the size of the magnetic fields required to
produce the effect.
The phenomenon of magnetoresistance (MR) was first observed by Lord
Kelvin in 1956 (Philip, 2000), where 0.033% rise of electrical resistance was
recorded in a piece of iron subjected to a magnetic fields. However, it only becomes
important when the electrical resistance of permalloy thin film magnetic sensor
changed by up to 2% when its magnetization direction was changed (Mahdi et al.,
2003).
2
Generally, MR sensors are made from ferromagnetic thin films. There are
two major advantages of ferromagnetic thin film over bulk material. These include
the high resistance and the anisotropic characteristic of ferromagnetic thin film,
which can be made uniaxial (Timothy, 2001). According to David (1991), the
anisotropy phenomenon ferromagnetic layers behaves like a single domain. It has
one distinguish direction of magnetization in its plane, called the easy axis.
Barna and Grunberg (1992) reported that, for a given thickness of the nonmagnetic chromium layer in three-layer Fe/Cr/Fe structure, the magnetizations of Fe
layers pointed in the opposite directions. According to Ping (2001), the interlayer
coupling responsible for this anti-alignment is called “antiferromagnetic”.
Antiferromagnetism is a phenomenon in which atomic magnetic moments point in
opposite directions in materials.
In 1988, Baibich et al. observed a similar phenomenon occurring in the
antiferromagnetically coupled Fe/Cr superlattice. However, a considerable drop in
the resistance occurred when a sufficiently high magnetic field of approximately 2T
was applied. This effect is now known as GMR.
GMR sandwich structures were then introduced with capabilities of
producing a higher GMR ratio up to 50% (Mahdi, et al. 2003). Generally, GMR
structures consist of an ultra thin metallic non-magnetic layer of Cu or Ag (≈10 nm)
sandwiched between two ferromagnetic metals, such as cobalt and iron. Several
theoretical studies have been carried out to account for the various mechanisms
occurring in GMR multilayer film (Valet, et al. 1996; Johnson, et al. 1991; and
Barthelemy, et al. 1991)
The novel magnetic alignments were the other interesting aspects in magnetic
thin films. It has been investigated for the academic interest and application in micro
devices (Jiang and Bader, 2002; Matteo et al., 2000; and Crew et al., 2001).
Magnetic multilayers were also proved to be a model system for investigating the
magnetization reversal process, where a transition of surface spin-flop (SSF) was
obtained in an ideal antiferromagnet (Luthi and Hock, 1983; and Rohrer, 1977).
According to Timothy (2001), SSF only occurs when the direction of the top surface
3
magnetization is antiparallel to the bottom surface magnetization. GMR is a new and
developing field. Thus, much more work was needed to explore and make use of the
GMR.
1.2
Research Objectives
It is well known that GMR is very sensitive to the microstructure of the
sample (Ratzke et al., 1999; Herker et al., 2002; and Dinia et al., 2000). Thus, the
main interest of this study is to determine the highest GMR effect that can be
obtained in Co/Cu nanostructures. Although study of GMR effect in Co/Cu
nanostructures have been reported previously, not much work was done in studying
the optimum conditions of thin film, which can produce the highest GMR effect.
This information is important to the computer manufactures to produce read-out
heads with higher density. Thus, the objectives of this research are as follow:
1) To prepare Co/Cu nanostructures.
2) To measure and study the MR% of Co/Cu nanostructures.
3) To determine the highest MR% obtain in the Co/Cu nanostructures.
1.3
Research Scope
The scope of this research involves the preparation of Co/Cu nanostructures
by two different methods, namely the RF sputtering and electron beam (e-beam)
method. Emphasis is given on the studies of magnetoresistance ratios (MR%) as
functions of Co layer thickness, working pressure, annealing time and temperature,
number of bilayer, direction of magnetic fields, and the application of buffer layer in
order to obtain the optimum condition for the highest MR%. The resistance of
samples in this study was measured using the four-point Van der Pauw method.
4
The following materials (Table 1.1) were chosen in order to achieve the aims of
the study.
Table 1.1: Material and their function in the system
Material
Function
1.
Cobalt (Co)
Ferromagnetic material
2.
Copper (Cu)
Non-magnetic but a good conductor
3.
Chromium (Cr)
Buffer layer
Co was chosen as the ferromagnetic material because it is one of the elements
with ferromagnetic characteristic at room temperature (Anderson et al.; 1985 and
David, 1991). Apart from that, the Curie temperature of Co is higher than other
ferromagnetic materials, for example Ni and Fe. Curie temperature is the transition
temperature from ferromagnetic to paramagnetic behavior. Thus, Co was chosen as
ferromagnetic material, since a high temperature is needed to transform into
paramagnetic behavior. Meanwhile Cu, the most common good conducting material
is chosen as a non-magnetic material.
The Cr was used as a buffer layer in the system. This is due to its ability to
enhance the MR% of system (Shen, et al., 1999). Besides, it is easily deposited by
sputtering and electron beam method (e-beam).
1.4
Thesis Plan
This thesis contains 6 chapters. Chapter I is the introductory section on the
development of the research. It reviewed the problem statement and previous work
5
done by other researchers. This chapter also specified the aim of studies, choice of
system, and outline of the thesis plan.
Meanwhile, Chapter II deals with the background of this study, which covers
the theoretical aspects of the magnetic material; i.e. ferromagnetic and
antiferromagnetic material, band structures, thin film deposition and method such as
radio frequency (RF) sputtering and electron beam (e-beam) method, thickness
measurement and GMR.
The details of the sample preparation, design of the experiment, methods of
measurement of the various physical parameters such as thickness and resistivity of
samples, calculation of MR% are described in Chapter III.
Chapter IV discusses all the experimental results obtained from the
investigation on the effects of thin film layers thickness, working pressure, bilayers
of sample, annealing process, buffer layer, direction of external magnetic fields, and
different methods of sample preparation.
Chapter V summarized the findings mentioned and the condition of sample
that can produce highest MR%. Finally, suggestions on future work will also be
mentioned.
6
CHAPTER 2
THEORY
2.1
Magnetic Material
The first known magnetic material or magnetite was discovered in China in
6th century BC. The word magnetism originated from the name of city Magnesia in
Asia Minor, where deposits of magnetite ore can still be found (Timothy, 2001).
Magnetic materials can be divided into ferromagnetic, antiferromagnetic,
paramagnetic, diamagnetic and ferrimagnetic, which display different magnetic
properties and behaviour (Anderson, 1985). In this research, attention will be given
to ferromagnetic and antiferromagnetic materials.
2.1.1
Ferromagnetic Material
Ferromagnets are used because of their high permeability, which enable high
magnetic inductions to be obtained with only modest magnetic fields. Once exposed
to a magnetic field, ferromagnet will retain their magnetization even after the field is
removed. The magnetic properties of ferromagnet can be represent by a plot of
magnetic induction B for various field strength H. Alternatively plots of
magnetization M against H are used, but these contain the same information since
7
B = µ 0 (H + M )
(2.1)
when µ0 is the permeability of free space
The suitability of ferromagnet materials for applications is determined
principally from the characteristics shown by their hysteresis loops. Therefore
materials for transformer applications need to have high permeability and low
hysteresis losses because of the need for efficient conversion of electrical energy.
Materials for electromagnets need to have low remanence and coercivity in order to
ensure that the magnetization can easily be reduced to zero as needed. Permanent
magnet materials need high remanence and coercivity in order to retain the
magnetization as much as possible.
From the Weiss domain theory, large numbers of atomic moments, typically
1012 to 1015, are aligned parallel so that the magnetization within the domain is
almost saturated. However the direction of alignment varies from domain to domain
in a more or less random manner, and it only can be aligned by the external magnetic
fields
Figure 2.1: A typical hysteresis loop of antiferromagnetic material
The magnetic properties in ferromagnetic elements are due to the 3d band
electrons. There is also a 4s electron band but this contains two paired electrons and
so does not affect magnetic properties.
8
The magnetic moments can only arise from unpaired electrons. Thus the
exchange interaction is responsible for creating the imbalance in spin up and spin
down states. In the absence of exchange energy the spin imbalance would be an
excited state but this does not require too much energy in 3d band because of the
high density of states and therefore a positive exchange interaction can be sufficient
to cause the alignment resulting in a spin imbalance and a net magnetic moments per
atom (David, 1991). The high density of states results in a large scattering
probability for electrons, resulting in a high resistivity (Timothy, 2001).
2.1.2 Antiferromagnetic Material
Antiferromagnetic substance has a small positive susceptibility at all
temperatures, but their susceptibilities vary in a peculiar way with temperature. At
first glance, they might therefore be regarded as anomalous paramagnetics.
However, closer study has shown that their underlying magnetic "structure" is so
entirely different that they deserve a separate classification. The theory of
antiferromagnetism was developed chiefly by Neel in a series of papers, beginning in
1932, in which he applied the Weiss molecular field theory to the problem (David,
1991).
χ
P
AF
0
T (°K)
TN
Fig 2.2: Variation of the susceptibility with temperature for an
antiferromagnetic. AF= antiferromagnetic, P= paramagnetic
9
The way in which the susceptibility, χ of an antiferromagnetic varies with
temperature is shown in Fig 2.2. As the temperature decreases, χ increases but
finally goes through a maximum at a critical temperature TN called the Neel
temperature. The substance is paramagnetic above TN and antiferromagnetic below
it. TN commonly lies far below room temperature, so that it is often necessary to
carry susceptibility measurements down to quite low temperatures to discover if a
given substance, paramagnetic at room temperatures, is actually antiferromagnetic at
some lower temperatures.
Below TN, the randomizing effect of thermal energy is so low, so the
tendency toward an antiparallel alignment of moments is strong enough to act even
in the absence of an applied field. The lattice of magnetic ions in the crystal then
breaks up into two sublattices, designated A and B, having moments more or less
opposed. The tendency toward antiparallelism becomes stronger when the
temperature is below TN. At 0 K, the antiparallel arrangement is perfect, as depicted
in Figure 2.3.
A
B
A
B
Figure 2.3: Moment spin of an antiferromagnetic
2.2
Band Structure
Almost all materials properties are influenced by the structural arrangement
of the atoms or ions in the material and their effects on the electron wave function.
Optical properties and electrical and thermal conductivities are good examples of
this. By solving Schrodinger’s equation in a solid, we can know that electrons exist
10
in groups of allowed energies known as energy bands. The energy bands are
separated by regions of forbidden energies know as band gaps (Timothy, 2001).
Insulators have a Fermi energy (EF) half way between the highest occupied
electron energy state and the lowest unoccupied state. The insulator lying in an
energy gap for semiconductors and insulators, while in metals EF lies in the
conduction band. Good conducting noble metals like Cu and Ag have a partially
filled of s-shell and completely filled d-shells. In these metals, the s-electrons are
primarily responsible for conduction (Anderson et al., 1985).
2.3
Thin Film Deposition
Thin film is a layer that consists of solid materials (elements or compound),
which are coated on a substrate surface with the thickness, t smaller than 1000 nm.
When the thickness t ≥1000 nm, it will be considered as a thick film or film only
(Samsudi Sakrani, 1996). A substrate is any solid, e.g. glass, that can support the
deposited thin film without reacting with the surface. The substrate must be smooth
to produce a homogeneous and high quality thin film. The range of thin-film
applications is very broad indeed. Additional functionality in thin film can be
achieved by depositing multiple layers of different materials.
When alternating layers are made using nanometer thickness of semiconducting materials such as GaAs, the result is a “superlattice” that has electrical
properties governed by the constructed periodicity rather than by the atomic
periodicity. Thus, multilayer thin film can behave as completely new, engineered
materials unknown in bulk form. When multiple layering is combined with
lithographic patterning in the plane of the films, microstructures of endless variety
can be constructed. This is the basic technology of the integrated-circuit industry,
and more recently it is being applied to optical wave-guide circuitry and to
micromechanical devices (Vossen et al, 1991).
11
All thin-film processes contain four (or five) sequential steps shown in Figure
2.4. A source of thin film material is provided. It may be a solid, liquid, vapor, or
gas. The deposition takes place when the material is transported to the substrate.
The film is subsequently annealed after the deposition takes place. Finally, the film
is analyzed to evaluate the process. Results of the analysis are then used to adjust the
conditions of other steps for film properties modifications.
12
Solid, liquid,
vapor, gas.
Source
Supply Rate
Vacuum, fluid,
plasma.
Transport
Uniformity
Deposition
Substrate condition,
reactivity of source material,
energy input
Annealing
Structure and composition
Analysis
Process modification
Figure 2.4: Thin-film processes
Structure,
composition,
properties.
13
2.3.1 Radio Frequency (RF) Sputtering
The basic sputtering process is illustrated in Figure 2.5. When a high-energy
incident ion bombards a target, it knocks atoms near the target surface from their
equilibrium positions, causing these atoms to move in the material and undergo
further collisions, and finally causing the ejection of atoms through the target
surface. This ejection process is known as sputtering. Reflected ions, neutral atoms
and secondary electrons may also be produced along with the target atoms. The
ejected target atoms are then made to condense on a substrate to form a thin film.
Primary Ion
+
Sputtered
Particle Ion Or
Neutral Atom
Primary Ion
Penetration
Depth
Implanted Ion
Figure 2.5: Schematic of the ion-solid interactions
and the sputtering process
At low frequency, the average energy of the bombarding ions would be
markedly reduced as a result of positive charge accumulation, so the frequency
should be 10 MHz or more (Siegle, 1972). But if the frequency becomes too high
(for example, >100MHz), the ions cannot respond to the RF field and again the
sputtering rate diminishes. So in practice, the permitted industrial frequency of 13.56
MHz is usually used (Milton, 1985; and George, 1992). It has been reserved for
plasma processing by the Federal Communications Commission.
This RF sputter deposition method is also widely used for producing films of
metals, alloys, and semiconductors. In the RF sputtering system, the RF power
supply is coupled to the metal electrode where the target, which can either be an
14
insulator or a conductor, is placed. The schematic of RF sputtering system is shown
in Figure 2.6.
Matching
Network
13.56 MHz
Insulation
Cathode
Target
Glow Discharge
Substrates
Anode
Sputtering
Gas
Vacuum
Figure 2.6: The schematic of RF sputtering system
Neutral gas (typically Argon, due to its suitable weight and its chemical
inertness) is introduced into the chamber. When a large RF potential (about 1 to 1.5
kV) is applied across the metal electrodes, glow discharge can be initiated and
electrons oscillating in the alternating field have sufficient energy to cause further
ionizing collisions, and the discharge/plasma (a plasma is a complex gaseous state of
matter comprised of free radicals, electrons, photons, ions, and various neutral
species) can be self-sustaining (Behrisch et al., 1991).
However, the large RF potential required to initiate the discharge is no longer
necessary once it has been attained. Due to their higher mobility as compared to
ions, many more electrons will reach the target surface during the positive half-cycle
than ions during the negative half-cycle, and the target, being mounted capacitively
to the RF source, will become self-biased negatively. The negative dc potential on
the target surface then repels electrons from the vicinity of this surface, creating an
enriched ion sheath in front of the target. These ions will bombard the target, and
sputtering can be achieved.
15
Magnetic field effects are used quite a lot in sputtering systems. In the
conventional sputtering systems, electrons escaping from the inter-electrode space as
a result of random collisions will be lost to the walls and no longer oscillate in the RF
field. Therefore, there will not be sufficient electrons in the plasma to cause
ionization. To minimize this loss, a magnet is placed behind the metal electrodes.
The purpose of using a magnetic field in a sputtering system is to constrain the
electrons, and cause them to produce more ionization.
The RF system requires an impedance matching network to ensure maximum
effective power delivered to the electrodes. Adequate grounding of the substrate
assembly is necessary to avoid undesirable RF voltages developing on the surface.
The working gas pressure typically used ranges from a few mtorr to 0.1 torr,
depending on the factors like target material, RF voltage required etc. If the working
gas pressure in the RF system is too low, the electrons in the plasma do not cause
sufficient ionizations. On the other hand, if the gas pressure is too high, the electrons
are slowed by elastic collisions, resulting in insufficient energy to cause ionizations.
In addition the ions generated may not have enough energy to produce secondary
electrons when they strike the target surface. For both situations, the plasma will not
appear stable.
A considerable amount of energy is dissipated as heat at the target electrode
by the incident ions, and the target gets hot. The maximum temperature attained and
the rate of temperature rise depends on the glow discharge conditions. Although the
sputtering yield for most materials increases with temperature, the target temperature
rise should still be controlled at a tolerable level during sputtering due to some out
gassing problems, which may arise. Hence the target is always cooled with running
water.
Magnetic materials such as Co, Fe and Ni may be deposited by using targets,
which have been thinned by machining. The thinning operation permits sufficient
magnetic field strength to be maintained near the target surface such as that
magnetron operation is achieved (John and Werner, 1978).
16
2.3.2
Electron Beam (e-beam) Evaporation
The history of electron beam evaporation dates only from the days O’ Brian
and Skinner (1933) who developed an apparatus for evaporation of refractory
materials for X-ray targets and for surfaces for excitation potential measurements
(Loretto, 1984).
The basic essential of the e-beam process are an operating vacuum, although
this varies between 10-3 to 10-10 torr; an electron source; electron lenses for forming
an electron probe; detectors to detect the signals, and deflection systems for defining
the probe position (Bakish, 1962).
The electron beam method of evaporation is of interest to both decorative and
functional metallizers. A wide variety of material including refractory metals (such
as tungsten), low vapor pressure metals (such as platinum), and alloys can be
evaporated. Since the electron beam method concentrates large amounts of heat on a
very small area, high rates of deposition are possible, a factor which is of interest to
any production-oriented shop.
The process begins under vacuum of 10-5 torr or less. A tungsten filament
inside the electron beam gun is heated. When the filament becomes hot enough, it
begins to emit electrons. When electron beam strikes the material surface, the kinetic
energy of motion is transformed by the impact into thermal energy (heat). It is
important to remember that the energy given off by a single electron is quite small
and that the heating is accomplished simply by virtue of the vast number of electrons
hitting the evaporant surface. This is the energy, which vaporizes the target material.
The basic configuration of e-beam as shown in Figure 2.7. The energy level
achieved in this manner is quite high-often more than several million watts per
square inch. Due to the intensity of the heat generated by the electron beam, the
evaporant holder must be water cooled to prevent it from melting.
17
Substrate
e-beam Source
Material
Figure 2.7: Basic configuration of e-beam method
2.4 Thickness Measurement
2.4.1
Film Thickness Monitor (FTM)
The FTM (Fig 2.8) is a microprocessor based frequency counter capable of
converting frequency changes into deposition rates and thickness information for a
range of deposition materials. It is used in conjunction with a quartz crystal, which is
placed in the deposition field whose output frequency is controlled by crystal.
The FTM detects frequency changes caused by material being deposited on
the face of the crystal. The change in frequency is used by the FTM to calculate the
rate and thickness of the material being deposited. The frequency data is modified
according to user input data relating to the type of material being deposited and the
geometric relationship between the deposition source, the target substrate and the
crystal sensor. (Edwards, 1992)
18
Figure 2.8: Film Thickness Monitor
FTM Detector
Plate
Disc
Substrate
Magnetron
Vacuum
Chamber
Figure 2.9: Schematic of measurement for Film Thickness Monitor
19
2.4.2
Dektak3 Surface Profiler
In Dektak3, measurements are made electromechanically by moving the
sample beneath a diamond-tipped stylus. The high precision stage moves a sample
beneath the stylus according to a user-programmed scan length and speed. The
stylus is mechanically coupled to the core of an LVDT (Linear Variable Differential
Transformer).
As the stage moves to the sample surface, stylus rides over the sample
surface. Surface variations cause the stylus to be translated vertically. Electrical
signals corresponding to the stylus movement are produced as the core position of
the LVDT changes respectively. An analog signal proportional to the position
change is produced by the LVDT, which in turn is conditioned and converted to a
digital format through a high precision, integrating analog to digital converter (Veeco
Metrology Group, 1998). Figure 2.10 shows the Dektak3 Surface Profiler.
Figure 2.10: Dektak3 Surface Profiler
20
2.5 The Four-Point Probe
The four-point probe method is one of the most common methods for
measuring the resistivity of materials. It was originally proposed by Wenner (1916)
to measure the earth’s resistivity. The four point-probe method measurement
technique is referred as Wenner’s method in Geophysics. It was adopted for
semiconductor wafer resistivity measurements by Valdes in 1954. The probes are
generally arranged in-line with equal probe spacing. But other probe configurations
are possible (Hall, 1967). In the four-point probe method, two probes carry the
current and the other two probes are used for voltage sensing (Schroder, 1990) is
shown in Figure 2.11.
I
I
1 2 3 4
S1 S2 S3
Figure 2.11: A collinear four-point probe
2.5.1
Resistiviy of Arbitrarily Shaped Samples
The collinear probe configuration is the most common four-point probe
arrangement. Arrangement of the points in a square has the advantage of occupying
a smaller area since the spacing between points is only s, whereas in collinear
configuration the spacing between the outer two probes is 3s. However, sometimes
the sample is irregularly shaped, so it is difficult to provide a sample in a square
format (Schroder, 1990).
21
The theoretical foundation of measurements on irregularly shaped samples is
based on conformal mapping developed by van der Pauw (1958). He showed how
the specific resistivity of a flat sample of arbitrary shape can be measured without
knowing the current pattern, if the following conditions are met: the contacts are at
the circumference of the sample, the contacts are sufficiently small and the sample is
uniformly thick.
Consider the flat sample of a conducting material of arbitrary shape, with
contacts 1,2,3, and 4, along the periphery as shown in Figure 2.11. The resistance,
R12,34 is defined as
R12,34 =
V34
I 12
(2.4)
where the current I12 enters the sample through contact 1 and leaves through contact
2 and V34=V3- V4 is the voltage difference between the contacts 3 and 4.
3
V
4
2
Current
Source
1
Magnetic Layer
Non-magnetic layer
Figure 2.12: Four-point Van der Pauw method
22
2.6
Giant Magnetoresistance
The spectacular increase in magnetoresistance (MR) was emphasized by
addition of the adjective “giant” to “MR”, hence the word “giant” is related to its
large resistance changes, i.e. >5%. GMR is a phenomenon that is defined as the ratio
between the change in resistance and initial resistance, which is normally expressed
in percentage (%). Thus, the different between MR and GMR is that the MR film
only shows a slight change in the resistivity when placed in the magnetic fields in
comparison to GMR film, which may produce a significant change.
GMR structures consist of two ferromagnetic layers separated by a nonmagnetic spacer layer. These ultrathin layers, typically a few nm thick, have a
quantum mechanical exchange force that is strong enough to completely overwhelm
dipole (stray magnetic fields) and align the spins within each layer (David, 1991).
Thus, each layer behaves as though it had a single spin vector or magnetic moment,
which varies in orientation but not magnitude according to the strength of an applied
field.
Although GMR was first observed in antiferromagnetically coupled
multilayers, the only requirement are that the magnetizations of successive layers are
antiparallel and can be made parallel by an applied magnetic field. So, GMR is not
restricted to multilayers with antiferromagnetic interlayer coupling. Beside that, in
antiferromagnetically coupled multilayers, the applied fields must overcome a
saturation fields to observe the full GMR effect.
There are two types of GMR. The first type is a magnetic multilayer. This
type of GMR structures is constructed from thin ferromagnetic layers (FM-layers)
separated by non-magnetic spacer layers (NM-layers) in such a way that the
exchange coupling between consecutive FM-layers is antiparallel, i.e in absence of
an external magnetic field the magnetizations of adjacent FM-layers are oppositely
aligned. The function of relatively strong external magnetic fields, H, is to
simultaneously force the magnetizations of all layers from anti-ferromagnetic mode
(AF-mode) to an ultimate ferromagnetic mode (F-mode), see Figure 2.13
23
Spin-valve structures is the other type of GMR, the FM-layers are composed
of two types of material. Each FM-layer is composed of one type of material having
a different coercivity than its adjacent FM-layers. The magnetizations of the FMlayers with the larger coercivity, referred to as hard layers, are pinned to an external
magnetic layer. The remaining soft FM-layers, with the smaller coercivity, may be
rotated on application of a low strength, H (Philip, 2000).
In this study, I will concern only with the first type (magnetic multilayers) of
the GMR. Other than that, magnetoresistance ratios (MR%) were used to denote
both the measurements of resistance changes from MR and GMR effect.
‘3D’ Sandwich
‘2D’ Sandwich
FM NM FM
H
FM
NM FM
H
Figure 2.13: The magnetic multilayer type, in which the magnettizations are forced
from natural AF-mode (θ = 0°) to F-mode (θ = 180°) by H.
2.6.1
Theoretical Model
2.6.1.1 Single-Current Model
The GMR arises from spin-dependent scattering either in the magnetic layer
or at the magnetic/non-magnetic interface (Johnson et al., 1991). At zero fields the
24
magnetizations were antiparallel and the resistivity was at maximum, but in high
fields the magnetizations were parallel to the applied field and resistivity was at
minimum (Smadar et al., 2001). This is because when two magnetizations are
aligned parallel, the electrons transmitted strongly through the first ferromagnetic
layer are easy transmitted through the other, leading to lowered overall resistance.
By changing the relative magnetization of alternate layers from parallel to
antiparallel, a very large room-temperature change in the resistance may be produced
(Anonymous, 1999). Figure 2.14 shows the giant magnetoresistance effect for the
parallel and antiparallel alignment of two magnetizations.
Spin
Down
Spin
Up
ρ‚
ρ
ρ‚
ρ
Spin
Down
Spin
Up
ρ
ρ‚
ρ‚
ρ
Ωmeter
Ωmeter
(a)
Ω
(b)
Figure 2.14:GMR phenomena showing a) low and b) high resistance
The resistivities of spins up and down to the magnetization as:
ρ ↑ = ρ (1 − β )
ρ ↓ = ρ (1 + β )
(2.3)
where β is the number of Bohr magnetron per atom.
In the parallel configuration, one spin polarization experiences ρ↑ through both
layers, the other spin polarization ρ↓ in both layers. This is described in Figure 2.15.
25
ρ (1 − β )
ρ (1 − β )
ρ (1 + β )
ρ (1 + β )
Figure 2.15: Resistance effectiveness in parallel configuration
From Figure 2.15, resistivity of sample in parallel configuration can be
derives as follows
1
ρ parallel
=
=
=
1
ρ↑
+
1
ρ↓
1
1
+
2 ρ (1 − β ) 2 ρ (1 + β )
] [
[
2 ρ (1 + β ) + 2 ρ (1 − β )
]
4 ρ 2 (1 − β )(1 + β )
2ρ
=
2 ρ (1 − β )(1 + β )
1
=
ρ (1 − β )(1 + β )
1
=
ρ 1− β 2
2
(
)
ρ parallel = ρ (1 − β 2 )
(2.4)
In the anti-parallel configuration, each spin channel experience ρ↓ and ρ↑ in series.
The effective resistance is depicted in Figure 2.16
26
ρ (1 − β )
ρ(1+ β )
ρ (1 − β )
ρ (1 + β )
Figure 2.16: Resistance effectiveness in anti-parallel configuration
ρ↑ = ρ↓
ρ ↑ = ρ (1 − β ) + ρ (1 + β )
=ρ+ρ
= 2ρ
1
ρ anti
=
=
=
=
=
1
ρ↑
+
1
ρ↓
1
1
+
2ρ 2ρ
2ρ + 2ρ
4ρ 2
4ρ
4ρ 2
1
ρ
ρ anti = ρ
(2.5)
It can be seen that, from equations 2.4 and 2.5 the effective resistance is greater in
the antiparallel configuration than in the parallel configuration, no matter what the
sign of β is. This model is directly inferred from the two-current model (Fert, 1968),
where the current of both spin orientations are assumed separate. This is a
reasonable model because a scattering event where the electron flips while keeping
its velocity is very rare.
27
2.6.2
Benefit of GMR
Giant magnetoresistance in the ferromagnetic tunnel junction allows more
data to be packed on computer disks. If improvements are made in the interfaces
between magnetic layers in thin-film structures, the number of new applications
could prove irresistible. For example, it would be possible to make computer
operating memories [random access memory (RAM)] that are immune to power
disruptions and ionizing radiation.
GMR also may spur the replacement of RAM in computers with magnetic
RAM (MRAM). By using GMR, it may be possible to make thin-film MRAM that
would be just as fast, dense, and inexpensive. It would have the additional
advantages of being nonvolatile and radiation-resistant. Data would not be lost if the
power failed unexpectedly, and the device would continue to function in the presence
of ionizing radiation, making it useful for space and defense applications.
GMR recently moved out of the laboratory and into our computers with the
development of “read sensors” for magnetic disk drives (Fig 2.16). this is due to the
capacity of disk drives continues to grow rapidly as they shrink in size.
Figure 2.17: Basic IBM suspended head design.
28
One exciting aspect of GMR devices is their extremely small size. Currently,
computer and electronics manufacturers are struggling to shrink their devices and
keep them working at feature sizes of about 0.5 mm. Operating GMR-based devices
are already 50 times smaller than that, and they tend to work better at smaller sizes.
It has already been shown that GMR can be used to make a transistor.
The application of GMR in motion sensors is also likely to be important in
our homes, automobiles, and factories. It provides a convenient way of sensing the
relative motion and position of objects without physical contact. Just attach a
magnet to one object and a GMR sensor to another. Alternatively, if one of the
objects contains a magnetic material such as iron or steel, the object in motion will
alter any magnetic field that is present. These small changes in the magnetic field
could be detected by a GMR sensor.
Applications of this effect could become widespread in the industrial,
commercial, and military worlds. Here’s a possible list: sensitive detectors for
wheel-shaft speed such as those employed in machine-speed controllers, automotive
antilock brakes, and auto-traction systems; motion and position sensors for electrical
safety devices; current transformers or sensors for measuring direct and alternating
current, power, and phase; metal detectors and other security devices; magnetic
switches in appliance controls, intrusion alarms, and proximity detectors; motor-flux
monitors; level controllers; magnetic-stripe, ink, and tag readers; magnetic
accelerometers and vibration probes; automotive engine control systems; highway
traffic monitors; industrial counters; equipment interlocks; and dozens of other
applications requiring small, low-power, fast sensors of magnetic fields and flux
changes.
29
CHAPTER 3
METHODOLOGY
The experiment is divided into 3 major sections. The first section focused on
sample preparation. Two methods of deposition are employed, namely the
RF sputtering and electron beam method. The second section emphasized on
the annealing process. Finally, the thickness measurement and MR values
will be discussed.
3.1
Samples Preparation
Cobalt (Co) and copper (Cu) with purity higher than 99.9% were used as
starting materials in this work in order to prepare stacked Co/Cu film structures. The
samples were deposited onto corning glass substrate by using two different methods,
namely RF sputtering and e-beam method.
3.1.1
Deposition by RF Sputtering Method
Co/Cu structures were prepared using a High Vacuum Coater (HVC) system
by RF sputtering method. Samples with various thicknesses of Co layers, number of
30
bilayers and working pressure were deposited in order to study the effect of these
parameters on the MR%.
3.1.1.1 High Vacuum Coater Setup
HVC from Penta Vacuum Technology Pte. Ltd as shown in Figure 3.1 is a
coating machine with the combination of RF sputtering, e-beam and ion gun process.
The vacuum pumping system in HVC consists of both Seiko Seiki Turbo Pump, and
Edwards E2M18 Rotary Pump. At the initial stage, the rotary pump will pump down
the pressure of HVC chamber to 1.9 x 10-2 torr. The turbo pump will then operates
in order to pump down the pressure to approximately 9.0 x 10-7 torr. During the
deposition process, Throttle Valve Controller was used to control the pressure in the
HVC chamber. The Gas Matching Control in HVC chamber was used to control the
gas flow in the deposition process. It allows two types of gasses to enter
simultaneously into the chamber when the deposition process is started.
The HVC is equipped with two RF magnetron and one e-beam source, as
well as a shutter in each source. Maximum three substrates can be placed under the
plat disc located at the upper part of the chamber (Fig 3.2). In order to get more
uniform film during the deposition process, the plate disc is rotated by setting the
substrate rotation controller. A cooling water supply that consists of chiller and
water pump is required during the deposition process. The cool water is
continuously recycled in the HVC system with a chiller being set at 20 °C.
Apart from that, the HVC system is also equipped with a loadlock system.
This enables the removal of the plat disc or substrate without opening the operating
chamber. Thus, the operation chamber remains in a vacuum condition. The BOC
Edward EX120 turbo pump in the loadlock system was also backed by the Edwards
E2M18 rotary pump. When the system reaches a high vacuum state, the HVC
chamber and loadlock system were connected by opening the loadlock gate. By
controlling the arm in the loadlock system, the plat disc can either be taken out or put
31
in without venting the HVC chamber. The loadlock gate will close simultaneously
after the arm leave the chamber and followed by flowing in the purified nitrogen.
Pipeline
Lock Lord
System
Vacuum
Chamber
RF Matching
Box
Control Panel
Figure 3.1 (a): High Vacuum Coater
32
FTM
e-beam control
panel
Seiko Seiki
SCU-21D
turbo pump
RF matching
system
GAT gas
matching
control 200A.
BOC Edward
EX120 turbo
pump
Ion gun control
panel
Throttle valve
controller
e-beam power
generator
Substrate
rotation
system
RF power
generator
Main Switch
Figure 3.1 (b): Control panel of High Vacuum Coater.
Loadlock Door
Arm
LoadlockGate
Substrate
Shutters
Loadlock system
Material
RF
Magnetron
Shutter
e-beam
Source
Figure 3.2: Internal part of High Vacuum Coater
33
3.1.1.2 Substrate Pre-Clean
The corning glass was used as a substrate. Firstly, pre-cleaned
the glass ultrasonically in chromic acid solution and distilled water for 40 minutes,
respectively. Next, the corning glass was let to dry. This was then followed by ion
gun pre-clean in the HVC chamber. It was carried out in the presence of argon gas at
8 standard cubic centimeter per minute (sccm) for 10 minutes in each sample.
3.1.1.3 Pre-sputtering Process
Pre-sputtering process was carried out to clean the residue or unwanted
material on the material surface (John et al., 1978). The plasma that formed in this
process will bombard the material and clear off the unwanted material. In this
process, the shutter stays at the close position as shown in Figure 3.2. Thus, the
purity of material can be maintained.
3.1.1.4 RF Sputtering Process
RF sputtering is one of the deposition processes, which can deposit a thin
film of metal. Generally the metal prepared by this method has a high melting point
and cannot be melted easily in the vacuum chamber. The target material with
diameter 7.6 cm and thickness 2 mm was used for RF sputtering deposition process.
In this study, argon gas was used in order to form a plasma during the
sputtering process. The argon gas will flow into HVC chamber when the RF
sputtering deposition process was started. RF power of 200 V was used for the
deposition of both Co and Cu layers. The reflected power value generated was
controlled at a small value. This is because the systems will heat up when a higher
reflected power is produced. Once the system was overheated, it will automatically
cool down to avoid overheating condition in the system. The deposition process will
34
stop when the system was automatically cool down. Thus, the perfect value for the
reflected power was zero.
Plasma will form at pressure higher than 7.5 x 10-3 torr. Thus, the throttle
valve needs to be closed at the early state. The throttle valve was opened after the
formation of plasma and the deposition process started. Both deposition rate and
thicknesses of Co and Cu layers were determined using a Film Thickness Monitor
(FTM) during the deposition process. The Dektak3 Surface Profiler was then used to
measure the thickness of samples in order to confirm the result that was obtained
from FTM.
The sputtered samples were fabricated on corning glass substrates with
various Co layer thicknesses, tCo i.e. 2, 6, 10, 12, 15, 20 nm, while the Cu layer
thickness was fixed at 2.5 nm. This was carried out at a working pressure of 3.0 x 103
torr. In addition, the Co/Cu/Co sandwich structures were prepared under 5 various
deposition pressures, i.e. 2.3 x 10-3, 2.6 x 10-3, 3.0 x 10-3, 4.1 x 10-3, and 4.8 x 10-3
torr. The thickness of both ferromagnetic and non-magnetic layers was 6 and 2.5
nm, respectively. This is to investigate the effect of working pressure towards MR%.
In order to maximize the MR%, various number of bilayers (n) of 6 nm Co and 2.5
nm Cu were prepared under 3.0 x 10-3 torr working pressure. The prepared samples
were labeled as tabulated in Table 3.1.
35
Table 3.1: Label of samples prepared by RF sputtering
Sandwich structure
Parameter
Co/Cu/Co (t nm/2.5 nm/t nm)
*Working pressure = 3.0 x 10-3 torr.
t=2
ST1
t=6
ST2
t = 10
ST3
t = 12
ST4
t = 15
ST5
t = 20
ST6
p = 2.3
SP1
p = 2.6
SP2
p = 3.0
SP3
p = 4.1
SP4
p = 4.8
SP5
n = 3/2
SN1
n=5
SN2
n=7
SN3
n=8
SN4
n=9
SN5
n = 10
SN6
n = 15
SN7
Co/Cu/Co (6 nm/2.5 nm/6 nm)
Co/Cu (6 nm/2.5 nm)
*Working pressure = 3.0 x 10-3 torr.
Label of Samples
Note: t = thickness (nm), p = pressure (10-3), n = number of multilayer
3.1.2
Deposition by Electron Beam Method
Apart from using RF sputtering method, e-beam method was employed in
preparing the thin films. The e-beam method was carried out using Edwards Auto
306 Evaporation Systems (Auto 306). Co/Cu/Co sandwich structures with various
Co and Cu layers thickness were thermally evaporated and deposited onto corning
glass substrates. In order to obtain a higher MR%, buffer layer was also deposited on
the samples. In addition, the MR% was measured with the magnetic field applied
along and perpendicular to the samples as shown in Figure 3.3.
36
Magnetic fields
of ±2500 gauss
was applied
along the
Magnetic fields of ±2500
gauss was applied
perpendicular to sample
Co/ Cu/ Co
Co/Cu/Co
Figure 3.3: Direction of magnetic fields applied to samples
3.1.2.1 Edwards Auto 306 Evaporation Systems
Edwards Auto 306 Evaporation System (Auto 306) shown in Figure 3.4
consists of a filament, shuttle, crucible, voltage and current source, rotary pump,
BOC Edward EX120 turbo pump, FTM and control panel. The rotary pump will
pump down the pressure of HVC chamber to 1.9 x 10-2 torr. The turbo pump will
then operate in order to assist the rotary pump to pump down the pressure to
approximately 9.0 x 10-7 torr.
The filament was located at the upper site of the crucible in the Auto 306
system. Five crucibles can be placed in Auto 306 system for each deposition
process. Thus, Auto 306 system can deposit 5 different materials by rotating the
crucible in the system, without opening the system chamber.
The crucibles used in the deposition process should not contaminate, react, or
become alloy with the evaporant at the evaporation temperature. The most common
sources are cylindrical cups composed of oxides, graphite and refractory metals
(Milton, 1985). In this study, graphite crucible was used for the e-beam process.
37
Vacuum
Chamber
FTM
Control
Panel
Figure 3.4: Edwards Auto 306 evaporation systems
3.1.2.2 Substrate Pre-Clean
The corning glass used as a substrate was pre-cleaned ultrasonically in
chromic acid solution and distilled water for 40 minutes. Then, the corning glass
was let to dry.
3.1.2.3 Deposition By e-beam Evaporation Process
The Auto 306 system was pumped down to a base pressure of 9.0 x 10-7 torr
before the e-beam evaporation process was started. The voltage and current source
was set to 3kV and 0.15A respectively. By supplying the voltage and current source
to the filament, the filament was heated. It begins to emit electrons when the
filament becomes hot enough. When electron beam strikes the material surface, the
kinetic energy of motion was transformed by the impact into thermal energy (heat).
Sample was heated and deposited on the substrate after reaching the melting point.
38
The deposition rate of the Co and Cu layers was 0.1 nm/s. The working pressure for
e-beam method was 1.0 x 10-5 torr.
The deposition rate and thickness of thin film were determined by the FTM.
The shutter will be closed manually when required samples thickness was reached.
Samples that have been prepared by e-beam method were labeled as tabulated in
Table 3.2.
Table 3.2: Parameters of samples prepared by e-beam method
3.2
Sample
Label
Co/Cu/Co (12 nm/2.5nm/12 nm)
EA1
Co/Cu/Co (5.5 nm/3.5 nm/ 5.5 nm)
EB1
Cr/Co/Cu/Co (8 nm/5.5 nm/3.5 nm/5.5 nm)
EB2
Annealing Process
Annealing process may change the magnetic properties of the thin film,
which affects the MR% (Ratzke, et al., 1999; Turilli, et al., 1998). Thus, postdeposition annealing was carried out in order to find out the higher MR% in
Co/Cu/Co sandwich structures.
The samples were placed in a quartz tube and were then heated from room
temperature to the setting temperatures. The uncertainty of the graph was calculated
and discussed in section 3.2.1. Vacuum pump will pump down the quartz tube once
the heating process was started. Thus, annealing process was done under vacuum
condition.
A Co/Cu/Co sandwich structure with respective thickness of (6 nm/2.5 nm/6
nm) was annealed at various annealing temperature and time. Table 3.3 shows the
annealing parameters for samples prepared by RF sputtering method. Samples ST2
39
were annealed at temperature of 300 °C for ½, 1, 1½, 2, and 2½ hours in vacuum
condition as shown in Figure 3.5.
Table 3.3: Parameters of annealing for the samples prepared by RF sputtering
method.
Samples
Parameters
Label of Samples
Co/Cu/Co
y = 200
SY1
(6 nm/2.5 nm/6 nm)
y = 250
SY2
y = 300
SY3
y = 350
SY4
y = 400
SY5
y = 425
SY6
y = 450
SY7
[Co/Cu (6 nm/2.5 nm)] x n
n = 3/2, y = 400
SY8
*Annealing time = 2 hour
n = 5, y = 400
SY9
n = 8, y = 400
SY10
n = 10, y = 400
SY11
n = 15, y = 400
SY12
* Annealing time =2 hour
Co/Cu/Co
(6 nm/2.5 nm/6 nm)
z=½
z=1
SZ1
SZ2
*Annealing temperature =
z = 1½
SZ3
300 °C
z = 2½
SZ4
Note: n = number of multilayer, y = temperature (°C), z = time (hour)
In addition, samples ST2 were then annealed at 200°C, 250°C, 300°C, 350°C,
400°C, 425°C and 450°C at 2 hour. Apart from that, attempts were also made to
measure the MR% for 5 different number of Co/Cu bilayers that were also subjected
to 400°C annealing process for 2 hour. Table 3.4 shows the annealing parameter for
samples EA1 at 300°C for 3 annealing times, i.e. ½, 1, and 1½ hours.
40
Table 3.4: Annealing parameter for samples prepared by e-beam method
Sample
Parameter
Label of Samples
Co(12nm)/Cu(2.5nm)/Co(12nm)
z=½
EZ1
*Annealing temperature =
z=1
EZ2
z = 1½
EZ3
200°C
Note: z = time (hour)
Quartz Tube
Vacuum
Pump
Thermocouple
Sample
Furnace
Figure 3.5: Set up of annealing process
3.2.1 Temperature Uncertainty Calibration
The data that was obtained from the temperature uncertainty study was
plotted in a graph quartz temperature versus heater set point (Fig 3.6). The results
show that some of the point were not exactly on the line, thus the uncertainty
calculation of the temperature is needed. The R-squared of graft shown in Figure 3.6
was 0.9983. Thus, the quartz temperatures were almost same with the heater set
point. The uncertainty of temperature was calculated by least square fit method
(John, 1982) as shown in Appendix A.
41
Quart Temperature, °C
500
2
450
R = 0.9983
400
350
300
250
200
150
150
250
350
450
550
Heater Set Point, °C
Figure 3.6: Graph of quartz temperature versus heater set point
From the calculation in Appendix A, the uncertainty of furnace temperature
was ≈ 0.02°C. This indicates that the temperatures of the quartz tube are very close
with the setting temperature.
3.3
Measurement
The thickness and MR% of the prepared samples were measured in the study.
Measurement of samples thickness was carried out using FTM and Dektak3 Surface
Profiler. The MR% measurement was then carry out by the four-point Van der Pauw
method in magnetic fields of ± 2500 gauss.
3.3.1 Thickness Measurements
During the deposition process, the thickness of the prepared samples were
42
determined using the Film Thickness Monitor (FTM). After the deposition process,
the thicknesses of prepared samples were measured again using the Dektak3 Surface
Profiler in order to reconfirm the thickness.
3.3.1.1 Measurement by Using FTM
FTM was used to measure the thickness of the sample by converting the
frequency changes into depositions rates and thickness information as discussed
previously in Chapter 2.
The crystal quartz in FTM was placed behind the plate disc where the
substrate is located. The density and Z value (acoustic impedance) of the deposited
material was entered into the FTM as information of frequency data.
When the deposition process started, the film deposited on the substrate was
also deposited on the crystal quartz behind the plate disc. The frequency detected by
quartz crystal will change according to the amount of film deposited, density and Z
value of material. The frequency data was then converted to the rate and thickness of
film deposited by the computer program.
3.3.1.2 Measurement by Using Dektak3 Surface Profiler
Dektak3 Surface Profiler was used to measure the thickness of the post
deposition sample. Firstly, the sample was soaked in an acetone solution after the
deposition process. The acetone will dissolve the symbol and produces a gap
between the deposited film as shown in Figure 3.7. Thickness of the film can be
determined by measuring the vertical side of the film by using Dektak3 Surface
Profiler. Scan distance and speed for this measurement was set at 2000 µm and 80
µm/s respectively.
43
As the stylus moves vertically from the reference point, the vertical
difference between the sample and the symbol was detected. The thicknesses of the
film were then measured. The reference point was the labeled initial counting value
(vertical side) without the present of thin film. For an example, the thickness of
(Co/Cu) x 15 bilayers that was measured by using Dektak3 Surface Profiler is shown
in Figure 3.8.
Symbol
Thin Film
Measurement
distance
Figure 3.7: The straight line that was used for thickness measurement.
44
Thickness
Measurement Distance
Figure 3.8 (a): Thickness of (Co/Cu) x 5 measure by Dektak3 Surface
Profiler
Thickness
Measurement Distance
Figure 3.8 (b): Thickness of (Co/Cu) x 10 measure by Dektak3 Surface
Profiler
45
Thickness
(c)
Measurement Distance
Figure 3.8 (C): Thickness of (Co/Cu) x 15 measure by Dektak3 Surface
Profiler.
The different between the thickness detected by using FTM and Dektak3
Surface Profiler was shown in Table 3.5.
Table 3.5: Thickness detected by using FTM and Dektak3 Surface Profiler
Samples
FTM (nm)
Dektak3 Surface Profiler (nm)
(Co/Cu) x 5 bilayers
48.5
47.5
(Co/Cu) x 10 bilayers
91.0
92.2
(Co/Cu) x 15 bilayers
133.5
133.4
3.3.2 MR Measurement
Firstly, the current source was tested before the MR measurement was
46
started. It is tested by connecting the current source to resistance of various values.
By changing the value of the resistance, the current going into the resistance
remained the same as shown in Table 3.6. Thus, it confirms that the current used in
the measurement is constant.
Table 3.6: Current and resistance values for current cource testing
No
Current Source (mA)
Resistance (ohm)
1.
15.61
270 x 103
2.
15.61
22 x 103
3.
15.61
15 x 103
4.
15.61
10
5.
15.61
5.6
6.
15.61
1.2
7.
15.61
1.1
The measurement was started by connecting the voltage and current source to
the sample by using the silver paste. The sample was fitted in the EMU-75 magnetic
fields generator that was equipped with a DPS-175 constant current power supply.
The DGM-102 digital gauss meter from Scientific Equipment Roorkee was used to
measure the magnetic field applied. The magnetic field was applied in plane with the
sample prepared by RF sputtering method as shown in Figure 3.9. However, for
samples prepared using e-beam process, the magnetic field was applied in plane and
perpendicular to the sample surface.
The four-point Van der Pauw method (Fig 3.10) was used to measure the
resistance of samples. In this measurement, 15.6 mA constant current (I) was
applied to the sample. The resistance and magnetoresistance ratios (MR%) of the
samples was then calculated using Equation 2.2 and 3.1 respectively where Rmax and
Rmin being the resistivities of the sample without and with the application of magnetic
field, respectively.
MR% =
Rmax − Rmin
x100%
Rmin
(3.1)
47
D i g i t a l G a u ss
M e te r
M agnet
P o le
M agnet
P o le
S a m p le
3cm
Figure 3.9: Magnetic fields applied in plane to sample.
48
CHAPTER 4
RESULT AND DISCUSSIONS
This chapter will discuss the magnetoresistance of samples, which were
prepared by RF sputtering and e-beam method. The effect of various parameters on
the MR% was investigated.
4.1
Magnetoresistance for RF Sputtering Film
Samples prepared by RF sputtering process will be focused in this section.
The effects of sample’s thickness, working pressure, bilayers and annealing process
on Co/Cu GMR were discussed.
4.1.1 Magnetoresistance (MR) Curve
All the deposited sample was measured by four point Van der Pauw method.
The data that was obtained in various magnetic fields was then plotted. Figure 4.1
shows the MR curve that was obtained from sample ST2, which applied in the
magnetic fields of approximately from –2500gauss to 2500 gauss. The MR curve
shows the increases in MR% simultaneously with the decrease in magnetic fields
towards zero. This obtained result is similar to the work that has been reported
49
previously (Ping, 2001). This is due to the magnetizations in two ferromagnetic
layers, which are mutually antiparallel with each other at zero fields (Smadar et al.,
2001). In high magnetic fields, samples show a minimum resistivity. This can be
explaining by the magnetizations of sample alight parallel in high magnetic fields
condition.
12
10
MR%
8
6
4
2
0
-3000
-2000
-1000
0
1000
2000
3000
Magnetic Fields, gauss
Figure 4.1: Magnetoresistance curve of Co /Cu /Co
(6 nm/ 2.5 nm/ 6 nm) sandwich structures
According to Vieux-Rochaz et al. (2000), antiferromagnetic exchange
coupling occurs only for a particular thickness of the spacer. The magnetic moment
in the sample will changes to parallel when external magnetic fields were applied.
and causing the changes in magnetic moment to parallel alignment. Apart from that,
50
both domain wall and domain-wall reflection were eliminated when the
magnetizations change to ferromagnetic alignment (Tagirov et al., 2001). As a
result, the resistivity decreases when a magnetic field was applied either towards
negative or positive values (Philip et al., 2000; and Weng et al., 2001).
The MR% increases as the difference between Rmin and Rmax increases and
will decrease otherwise. Thus, Figure 4.1 also shows the dependency of MR% on
the resistivity of samples.
4.1.1.1 Effect of Film Thickness
From literature review (Vavassori et al., 2003; Takashi et al., 1998; and
Kumar et al., 2001), samples exhibit a superparamagnetic behavior in room
temperature at sufficiently low Co thickness. Apart from that, different Co thickness
will contribute to a different saturation state, which will then affect the MR% (Bass
and Pratt, 2002). In this work, the effect of six various thicknesses of Co layers, tCo
on MR% were studied. Figure 4.2 depicts the maximum MR% obtained from
samples with different thickness of Co layers. Graph of MR% versus thickness
illustrated in Figure 4.3.
It was observed that, the MR% attains almost 10% increase between 2 - 6 nm
of the Co layer thickness (Fig 4.3). This is followed by a rapid drop and further
gradual decrease with the increase in Co layer thickness up to 20 nm. These results
are almost similar with the calculated results of Co/Cu/Co sandwich structural as
reported by Xu et al (2000).
The increase in thickness of ferromagnetic layers will simultaneously
increase the MR% as shows in the early state (tCo ≤ 6 nm). This is explained by the
amount of ferromagnetic spin in the ferromagnetic layers increases due to the
ferromagnetic exchange interaction, as the ferromagnetic Co layer thickness is
increased (Yamada et al., 2002; Vohl et al., 1991). In addition, the average
51
dimension of the Co particles is increased by increasing the tCo, (Vavassori et al.,
2003).
12
10
MR%
8
Co(2nm)
Co(2)
Co(6nm)
Co(6)
Co(10)
Co(10nm)
Co(12nm)
Co(12)
Co(15)
Co(15nm)
Co(20)
Co(20nm)
6
4
2
2220
1628
941
287
-265
-929
-1571
-2220
0
Magnetic Fields, gauss
Figure 4.2: Magnetoresistance curve of Co/Cu for 6 various
thickness of Co layers
52
12
10
MR%
8
6
4
2
0
0
5
10
15
20
Film Thickness of Co Layers (nm)
Figure 4.3: Graph of MR% versus thickness
However, this explanation is only applicable for sample with tCo ≤ 6 nm. This
is due to the magnetic fields efficiency, which is proportional to the particles
dimensions when tCo ≤ 6 nm. Therefore, the MR% decreases with increase in the
number of particles for Co layer (Turilli et al., 1999). Thus, in each applied
magnetic field value, there is an optimum tCo. The optimum tCo that was obtained in
this study is approximately 6 nm. Further increment in tCo will cause a decrease in
MR%. In addition, the reduction of MR% at higher thickness is also due to the
53
decrease in the antiferromagnetic interaction between the ferromagnetic layers near
the non-magnetic layer (Yamada et al., 2002).
From the results, it is also observed that the resistivity of samples decreases
with increase in tCo as shown in Figure 4.4 and Figure 4.5. The resistance value and
difference between Rmax and Rmin (∆R) drop dramatically as the tCo increase from 2
nm to 10 nm. This result in a higher resistance in sample ST1, which is about 4-fold
higher in comparison to sample ST3. On the other hand, resistance of the samples
decreases slightly at a higher thickness, i.e. tCo >10 nm. This results is in agreement
with the findings that was reported by Chapman (1963), who observed that the
resistance change in a thicker sample is smaller in comparison to the thinner samples.
Resistance, ohm (x 10-2)
10.5
9.5
8.5
7.5
6.5
5.5
Rmax
Rmin
4.5
3.5
2.5
1.5
2
6
10
12
15
20
Film Thickness of Co Layers (nm)
Figure 4.4: Graph of resistance versus film thickness of Co layer
Resistance Change, ∆R (ohm, x 10-2)
54
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
2
6
10
12
15
20
Film Thickness of Co Layer (nm)
Figure 4.5: Graph of resistance change versus film thickness of Co layer
4.1.1.2 Effect of Working Pressure
According to Zhang et al. (2000), the MR% depends markedly on the growth
conditions. In fact, various working pressure may produce sample with different
MR% (Higashihara et al., 2004). In this study, Co/Cu (6 nm/2.5 nm) sandwich
structures was prepared under five various pressure, i.e. 4.8 x 10-3, 4.0 x 10-3, 3.0 x
10-3, 2.6 x 10-3 and 2.3 x 10-3 torr. The base pressure of the chamber is around 9 x107
torr. The deposition rates were kept in the range of 0.03 to 0.1 nm/s as shown in
Table 4.1.
55
Table 4.1: Working pressure and deposition rate
No.
Working Pressure (torr)
Deposition Rate (nm/s)
1.
2.3 x 10-3
0.1
2.
2.6 x 10-3
0.09
3.
3.0 x 10-3
0.06
4.
4.0 x 10-3
0.03
5.
-3
0.02
4.8 x 10
The effect of working pressure during sample deposition process on the
MR% was plotted as shown in Figure 4.6. It seems that, a small peak in MR%
occurs between the SP1 to SP3 with a further gradual decreasing trend. The MR%
increases significantly from 9.28% to 11.36% when the working pressure increase
from 2.3 x 10-3 to 2.6 x 10-3 torr. However, it will decrease at working pressure
higher than 2.6 x 10-3 torr. The MR% decrease to 9.24% for samples SP5. Thus, 2.6
x 10-3 torr is the optimum working pressure to produce the maximum MR% in
Co/Cu/Co (6 nm/2.5 nm/6 nm) sandwich structures. This optimum working pressure
may be attributed to the various growth conditions. The growth conditions of
samples resulting in a very different grains size contribution, which will then affect
the MR% (Turilli et al., 1999).
56
12
11.5
11
MR%
10.5
10
9.5
9
8.5
8
2
3
4
5
Working Pressure
Figure 4.6: Effect of working pressure on MR%
The large amount of Co in samples SP1 and SP5 might present as a very
small superparamagnetic grains form. Thus, it gives a neglectable contribution to the
magnetization of the sample. Therefore, it weakly effective for the MR% in room
temperature (Turilli et al., 1999). As a result, the MR% of sample SP1 and SP5 are
smaller in comparison to other samples.
It has to be noted that samples resistivity is influenced by the working
pressure. The effect of working pressure towards the resistance and resistance
change are illustrated in Figure 4.7 and Figure 4.8 respectively. Figure 4.8 clearly
show the decrease in resistance change for all the samples when the working pressure
is increases except for sample SP2, which is slightly higher than SP1. As the
working pressure increase in the range of 3.0 x 10-3 – 4.8 x 10-3 torr, small
differences of resistance is observed. However, the resistance of sample increases at
57
2.6 x 10-3 torr is significantly increased. It is doubled compare with samples grown
at other working pressure.
5.5
Resistance,
ohm (x
10
Resistance
(ohm,
10-2
) -2)
Resistance
5
4.5
4
Rmax
Rmin
3.5
3
2.5
2
2.3
2.6
3
4
4.8
-3
Working
Pressure,
(x10 torr)
Working
Pressure
Working
Pressure
(x 10-3 torr)
Resistance,
Resistance Change, ∆R (ohm, 10-2)
Figure 4.7: Effect of working pressure on resistance of samples
0.5
0.45
0.4
0.35
0.3
0.25
0.2
2
3
4
5
-3
Working Working
PressurePressure
(x 10 torr)
Figure 4.8: Graph of resistance change versus working pressure
58
4.1.1.3 Effect of Bilayers
In order to maximize the MR%, the dependence of MR on the
number of bilayers (n) for samples with structures of (Co/Cu) x n was investigated.
The thickness of Co and Cu layers used in this study was 6 nm and 2.5 nm,
respectively.
There were eight various n that had been prepared under pressure of 3.0 x 103
torr, i.e. 1½, 5, 7, 8, 9, 10, and 15 bilayers. Meanwhile, the deposition rate for
samples prepared was 0.06 nm/s. Thus, longer deposition time is needed when n is
increased.
The MR% of samples obtained by four-point Van der Pauw is as shown in
Figure 4.9. It reveals that the MR% increases with n for n < 5. For sample with n >
5, the MR% start to drop and reach 9.97% for (Co/Cu) x 8 bilayers. However, the
MR% increases again when n > 8 and show a maximum value of 12.5% for (Co/Cu)
x 15.
The resistance of the samples increases when the bilayers or total thickness of
sample increased (Fig 4.10 and 4.11). A slight increase is observed when the
bilayers of the sample is smaller than 10. For sample with n = 15, the resistance of
the sample increases significantly, which is 2-fold in comparison to other samples.
The fluctuation between n =5 and n =10 was due to the uncertainty during the
measurement of MR%. Thus, the MR% was increase with the increase of the
bilayers of sample
According to Smadar et al. (2001), when the magnetic fields were applied in
plane to the sample, the total resistance is not equal to sum of the resistances of all
segments (Equation 4.1), but, resistivity of samples is increases with increase in
number of bilayers i.e.
Rtotal ≠ R1 + R2 + R3 + .........
(4.1)
59
13
12.5
12
MR%
11.5
11
10.5
10
9.5
9
0
5
10
15
Bilayers
Figure 4.9: Effect of number of bilayers on MR%
20
60
-2 10-2)
Resistance,
ohm
Resistance Change,
∆R (ohm,
10(x
)
60
50
40
30
Rmin
Rmax
20
10
0
0
5
10
15
20
Bilayers
Figure 4.10: Graph of resistance versus number of bilayers in samples
7
6
5
4
3
2
1
0
0
5
10
15
Bilayers
Figure 4.11: Graph of resistance change versus number of bilayers in
samples
61
4.1.2 Effect of Annealing Process
Thermal treatment is one of the important parameter that modifies the
microstructure of the deposited film (Herker, et al., 2003, and Ratzke, et al., 1999).
Thus the effect of annealing process, which includes annealing time and temperature,
was studied.
4.1.2.1 Annealing Time
Co/Cu/Co (6 nm/2.5 nm/6 nm) sandwich structures which prepared at 3.0 x
10-3 torr (ST2) are annealed at 300°C for ½, 1, 1 ½, 2 and 2½ hour. This is to study
the optimum annealing time, which can generate the highest MR%.
The effect of annealing time on Co/Cu/Co (6 nm/2.5 nm/6 nm) is depicted in
Figure 4.12. Results show that the MR% is enhanced from 8.78% to 18.75%, i.e. >
2-fold, by annealing the sample in vacuum for about 2 hour. However, the MR%
will tend to be constant when the samples were annealed longer than 2 hour.
The fixed magnetic fields that were applied to the sample are unable to
completely saturated the sample at finite temperatures, because of the existence of
very small granules in the size contribution (Wang et al., 1999). Hence, the MR%
increases dramatically after annealing for a short period. The results obtained is in
agreement with the results that was reported by Smith et al. (1998), which show a
drastic increase after annealing at 200°C in short period. Apart from that, Ratzke et
al. (1999) reported that the increase of MR% was due to the decrease in Rmin. This
findings is similar with the obtained Rmin in this study as shown in Figure 4.13.
The resistance of material may be temperature dependent (Boylestad, 2003).
Thus, the resistances of samples in this study decrease slowly as the annealing time
increase, as shown in Figure 4.13. The decrease can be attributed to the grain growth
of Co/Cu/Co structures. This result was found compatible with those reported by
Ratzke et al. (1999) where the ∆R increases for annealing time lower than 3/2 hour.
62
On the other hand, the ∆R is almost constant at above 3/2 hour and thus resulting in a
constant of MR%. In this study, the optimum annealing time for Co/Cu/Co (6
nm/2.5 nm/6 nm) sandwich structure is 2 hour.
21
19
MR%
17
15
13
11
9
7
0
1/2
50 1
3/2100
2
Minute
Annealing
Time (Hour)
Figure 4.12: Influence of annealing time on MR%
5/2
150
63
Resistance,
ohm (x 10-2)
Resistance
6.5
6
5.5
Rmin
Rmax
Rmin
Rmax
5
4.5
4
2
4
6
Annealing Time (Hour)
Resistance
Change,Change
∆R (ohm,10-2)
Resistance
Figure 4.13: Graph of resistance versus annealing time
1
0.9
0.8
0.7
0.6
0.5
0.4
01
2
1/2
13
4
3/2
25
5/2 6
Annealing Time (Hour)
Figure 4.14: Graph of resistance change versus annealing time
64
4.1.2.2 Annealing Temperature
Besides studying the optimum annealing time, the effect of annealing
temperatures, i.e. 200°C, 250°C, 300°C, 350°C, 400°C, 425°C and 450°C on
Co/Cu/Co (6 nm/2.5 nm/6 nm) on sandwich structures were also been studied. The
optimum annealing time (2 hours) that was obtained in the previous section was used
in the investigation of optimum annealing temperature. The MR% obtained are
shown in Figure 4.15.
The MR ratio is found to increase dramatically from 9.82% to 23.69%, as the
annealing temperature under vacuum condition increases from 27 to 400°C.
However, the MR% drops slightly at annealing temperature higher than 400°C and
reaches 20.46% at 450°C.
As proposed by Belozorov et al. (2003) and Coils et al. (2002), a
recrystallization process of Co and Cu occurred during the early state (0-400°C) of
annealing. Accordingly, the Co and Cu species become soluble to each other as they
reach 400°C. Thus, a solubility process is inevitably initiated between them, where
Co atoms gradually precipitate from the Cu matrix and forming the Co clusters.
Their bodies and interfaces with Cu matrix become the conduction electron spindependent scattering centers. A very small Co particles or Co atoms may loose their
magnetic moments due to interaction with the nonmagnetic Cu matrix, which
adversely affects the MR%. On the other hand, when the annealing temperature
exceeded 400°C, Co particles become larger and result in the appearance of a
ferromagnetic interaction between larger Co particles. In other word, an antiparallel
structure that is necessary for MR% does not occur in portions of the samples
(Hecker, et al., 2003; and Yu et al., 1995). The results obtained in this study are
similar with those reported by Liu et al. (2003), where Monte-Carlo simulation
showed a similar optimum MR% at around 400°C. In addition, a lower MR% is
observed at a low annealing temperature. This is explained in term of inactive spindependent tunneling due to the almost perfect alignment of all spins in the lattice.
Thus, no further induced MR effect can be activated.
65
26
24
22
MR%
20
18
16
14
12
10
8
27
27
0
100
200
300
400
500
°C
AAnnealing
nnealingTemperature,
Temperature
Figure 4.15: Effect of annealing temperature as a function of MR% in Co/Cu
14
12
MR%
10
8
6
4
2
0
0
200
400
600
800
Annealing Temperature
Figure 4.16: MR% effect of annealing temperature as a function of MR% in
NiFe/Cu. (From Hecker et al., 2002)
66
∆R decreases monotonically with increasing annealing temperature as shown
in Figure 4.17 and 4.18. The observed behavior of Rmin for Co/Cu is in contrast in
comparison to other systems. For example, NiFe/Cu (Fig 4.19) that was reported by
Hecker et al. (2002) shows the increase of Rmin with the increase in annealing
temperature
Essentially, this difference arises from the different intermixing properties of
the constituent layers. Whereas in the NiFe/Cu system, Ni diffuses preferentially
into the Cu layers, the Co and Cu layers tend to de-mix. Thus, there is no significant
counterbalance in Co/Cu that prevents the decrease of Rmin due to the improved
interface and lattice properties up to annealing at 600°C (Hecker et al., 2003). The
kinetics of the resistance has not yet been analyzed but the changes are expected to
be slower at lower temperatures (Ratzke et al., 1999).
6.5
Resistance, ohm (x 10-2)
6
5.5
5
R
Rmin
max
R
Rmax
min
4.5
4
3.5
3
0
100
200
300
400
500
Temperature(
Degree Celcius)
Annealing Temperature,
°C
Figure 4.17: Graph of resistance versus annealing temperature
Resistance
Change,
10-2 10-2)
Resistance
Change,
Resistance
Change,ohm
∆R x(ohm,
67
1
0.9
0.8
0.7
0.6
0.5
0.4
27
0
100
200
300
400
500
Annealing
Annealing
Temperature
°C
Annealing
Temperature
Figure 4.18: Graph of resistance change versus annealing temperature
1.8
1.7
Resistance, ohm
1.6
1.5
1.4
Rmin
1.3
Rmax
1.2
1.1
1
0.9
0.8
0
200
400
600
800
Annealing Temperature °C
Figure 4.19: Graph of resistance of NiFe/Cu versus annealing temperature
(°C). (From Hecker et al., 2002)
68
4.1.2.3 Effect of Annealing Temperature for Different Bilayers
Experimental results in Section 4.1.2.2 show that 400°C is the optimum
annealing temperature, which can affect the samples condition and produce the
highest MR% within the annealing temperature. In order to study the effect of
annealing temperature for different bilayer, samples with different number of
bilayers were subjected to the annealing process under 400°C for 2 hours. Five
various bilayers were prepared and the results are shown in Figure 4.20.
From Figure 4.20, it is observed an upward shift and improvement in MR%
for Co/Cu bilayers samples that was annealed at 400°C. However, the mode of the
graph remains unchanged if compare with Figure 4.9. For the as-prepared samples
(Fig 4.9), the MR% increases linearly with n except (Co/Cu) x 8. Apparently, MR%
decreases from 26.09% to 17.20% when n is increased between n = 5 to 8. The
MR% starts to rise again in samples with n > 8 and reaches a maximum value of
32.43% for (Co/Cu) x 15.
It was also observed that the resistance of sample increases with the increase
of number bilayers of Co/Cu (Fig 4.21). It shows a slight increase when total
thickness of sample is increased. However, the resistance increases drastically when
the bilayers of Co/Cu change from 10 to 15. Resistances of sample before and after
annealing are almost same for the basic sandwiched structures (Co/Cu/Co). But
when bilayers of samples is increased, the resistance of samples after annealing
process are lower when compare with samples before annealing process.
69
35
30
MR%
25
20
Before
Annealing
After
Annealing
15
10
5
0
5
10
15
20
Bilayers
Figure 4.20: Effect of number bilayers of Co/Cu before
and after annealing at 400°C towards MR%
70
70
Resistance
(miliohm)
Resistance,
ohm
(x 10-2)
60
50
40
Rmin after
anealing
Rmax after
annealing
Rmin before
annealing
Rmin after
Rmax
before
annealing
30
20
10
0
0
5
10
15
20
Bilayers
Figure 4.21: Graph of resistance versus number bilayers of samples
71
4.2
Magnetoresistance for e-beam Film
Samples that were prepared by e-beam method will be discussed in this
section. The effect of direction of magnetic fields, buffer layer and annealing
process on Co/Cu sandwich structures were studied.
4.2.1
Effect of Magnetic Field
Co is one of the samples, which will show anisotropy phenomenon
(Arkadiusz, 1970). Thus, the direction of the applied magnetic field will affect the
MR% in a sample (Timothy, 2001; Errahmani, et al. 2001; and Theeuwen, et al.
2000).
Magnetic fields at approximately ±2500 gauss were applied in plane and
perpendicular to the sample EB1 and EB2. Figure 4.23 show that samples exhibit a
maximum MR%, i.e. 1.57% when the magnetic field is applied along the sample (
=0°). However, it will decrease radically from the maximum value to 0.284% when
the magnetic field is applied perpendicular ( =90°).
This is due to the hexagonal crystal lattice of Co, in which the moments are
aligned along the unique axis [0001] (easy direction). The [1010] axis in the base
plane is the hard axis as shown in Figure 4.22 (David, 1991).
72
1.8
1.6
1.4
MR%
1.2
Current In Plane
1
Current
Perpendicular
0.8
0.6
0.4
0.2
-2
37
0
-1
88
8
-1
03
6
-2
14
60
9
14
08
20
80
23
70
0
Magnetic Fields, Gauss
Figure 4.22: Magnetoresistance curve of Co/ Cu/Co (5.5nm/ 3.5 nm/
5.5 nm) as the magnetic fields is applied along and perpendicular to the
sample.
Figure 4.23: Easy and hard axis in Co hexagonal crystal lattice
73
In the easy axis direction ( =0°), magnetic moments in the samples could
turn over quickly as the applied field changes. Thus, a higher sensitivity was
obtained in this direction. However, the magnetic moments move slowly with the
applied field in the hard axis direction ( =90°), which is the most difficult to be
magnetized, resulting in a smaller sensitivity (Shen et al., 2000). The difference
observed in MR% when the magnetic field is applied along and perpendicular prove
that Co/Cu/Co show an anisotropy phenomena, which consists in the geometrical
dimensions of a ferromagnetic specimen changing under an external magnetic fields
(Arkadiusz, 1970).
A part from that, the spin down electron for sample in which the magnetic
field was applied perpendicular have a smaller scattering probability in comparison
to that of the sample in which the magnetic field was applied along to it (Timothy,
2001). Thus, this explains why sample with magnetic field applied along the surface
will produce higher MR% than those the magnetic fields was applied perpendicular
to the surface.
4.2.2
Effect of Buffer Layers
Previous works showed the influence of a buffer layer on the structure of
GMR multilayers (Takahashi et al., 2002; and Yamamoto et al., 1993). Large MR
ratios were obtained in Co/Cu multilayers with Fe and Ni (Li et al., 2001) as buffer
layer. Accordingly, the buffer layer is found to improve the flatness of the
multilayers and enhances the MR% simultaneously.
Co/Cu/Co sandwich structures with film thicknesses of 5.5 nm/3.5 nm/5.5 nm
were thermally evaporated and deposited onto chromium buffered corning glass
substrates. The thickness of Cr buffer layer prepared is 8 nm. Magnetic fields around
±2500 gauss were applied along the surface of samples. The dependence of MR% in
Co/Cu/Co sandwiches structures on the Cr buffer layer are as shown in Figure 4.24.
It is observed that the MR% of sandwiches is strongly affected by the presence of Cr
74
layer. By adding the Cr as a buffer layer, the MR% increases more than 3-fold, from
1.57% to 5%. The MR% is also found to increase with the increase in thickness of
Cr buffer layer, as shown in Figure 4.25. These results reveal that the optimum
buffer thickness is 8 nm, in which the MR% will decrease with further increase.
As reported by Li et al. (2001), Cr buffer layer will reduce the roughness of
the interfaces and thus induces good-quality structure of the lower Co layer in
Co/Cu/Co sandwich. As a result, the MR% increases in the present of buffer with an
optimum thickness. It also observed that, the maximum MR% decrease to 3.06%
when magnetic fields were applied perpendicular ( =90°) to the length of the sample
as shown in Figure 4.26. The reasons for the increment in MR% are similar to those
been discussed previously in Section 4.2.1.
6
5
4
MR%
With Buffer Layer
3
Without
Buffer
Witout
Buffer
Layer
Layer
2
1
2370
2080
1408
609
-214
-1036
-1888
-2370
0
Magnetic Fields, Gauss
Figure 4.24: Magnetoresistance curve of Co/Cu/Co (5.5 nm/3.5 nm/
5.5 nm) with and without buffer layers (Cr) layer
75
6
5
MR%
4
3
2
1
0
0
5
10
15
Thickne ss of Cr La ye r (nm )
Figure 4.25: Dependence of MR% on Cr buffer layer thickness in
Co/Cu/Co (5.5 nm/2.5 nm/5.5 nm) sandwich structures
6
5
MR%
4
Magnetic Fields
along to Sample
3
Magnetic Fields
Perpendicular to
sample
2
1
-2
37
-1 0
88
-1 8
03
6
-2
14
60
9
14
08
20
80
23
70
0
Magnetic Fields, Gauss
Figure 4.26: Magnetoresistance curve of Cr/Co/Cu/Co (8 nm/5.5nm/2.5 nm/
5.5nm) when magnetic fields applied along and perpendicular to the sample.
76
4.2.3 Effect of Annealing Time on e-beam Films
Annealing process is also one of the essential parameters in modifying the
microstructure of the samples. Samples EA1 was annealed at 200°C for ½, 1 and 1 ½
hours in order to study the effect of annealing time.
Figure 4.27 and 4.28 show the magnetoresistance curve and the effect of
annealing on Co/Cu/Co (12 nm/2.5 nm/12 nm) sandwich structures, respectively. It
is found that, the MR ratio is enhanced from 0.80% to 2.25% by annealing the
sample in vacuum for about 1 hour. At the initial state (annealing time < 1 hour),
MR% increases proportionally with the annealing time. As the annealing times are
longer than 1 hour, the MR values will tend to be constant. This result is in
agreement with the results that obtained from samples prepared by RF sputtering
process, as discussed in Section 4.1.2.1. It shows a drastic increase after been
annealed at 200°C in a short period. Simply, annealing time will increase the
granular size. Thus, it indirectly increases the MR%(Wang, 1999).
A part from that, the resistivity of samples reduces with the increase of the
annealing time (Fig 4.29). The resistance of sample is found to decrease
proportionally from approximately 0.055 to 0.045 ohms when the samples were
annealed at 200°C for 2 hours.
77
2.5
2
As-deposited
1/2 hour
1 hour
1 1/2 hour
MR%
1.5
1
0.5
-2
51
0
-2
06
0
-1
67
4
-9
72
-1
74
57
7
13
42
19
07
21
70
0
Magnetic Fields, gauss
Figure 4.27: Magnetoresistance curve of Co/Cu/Co (12 nm/2.5nm/12 nm)
with different annealing time
2.5
MR%
2.3
2.1
1.9
1.7
1.5
1.3
1.1
0.9
0.7
10
2
1/2
31
43/2
Annealing Time (Hours)
Figure 4.28: Effect of annealing on Co/Cu/Co (12 nm/2.5 nm/12 nm)
78
5.6
Resistance,
ohm
(x 10-2)
Resistance
(miliohm)
5.4
5.2
5
Rmin
Rmax
4.8
4.6
4.4
4.2
10
1/2
2
31
3/2
4
Annealing Time (Hours)
Figure 4.29: Resistance of Co/Cu/Co (12 nm/2.5 nm/12 nm) in
different annealing time
4.3 Comparison Between Sputtering and e-beam Method.
At the heart of any deposition process is the creation of a flux of condensable
species. In the case of sputtering and evaporation, these species originate from
spatially constrained sources like a target or an intermediate liquid contained by a
crucible or filament (Jansen, 1997). Structure of film is dependent upon method of
deposition. Naturally this structure will affect the electrical properties of the film
(Coombe, 1967). Thus, indirectly the samples deposited by these two different
methods may have different MR%.
Figure 4.30 illustrates the different of MR% between films deposited by
sputtering and e-beam process. The sample deposited by sputtering and e-beam were
labeled as ST4 and EA1, respectively. The maximum MR% for samples ST4 are
higher than sample EA1, which is 4-fold of sample EA1.
79
This may due to the different conditions of film preparation. Samples ST$
and EA1 were prepared under working pressure of 3.0 x 10-3 and 9.0 x x 10-5 torr
respectively. This is due to the different deposition requirement that needed for
sputtering and e-beam method. Other than that, the RF sputtering and e-beam
process were carried out using HVC and Auto 306 system, respectively. Samples
prepared in HVC are more dirt free compare with samples prepared in Auto 306
system. This is because samples prepared in HVC have substrates pre-cleaned
ultrasonically in chromic acid solution and by ion gun. On the other hand, substrate
for samples deposited in Auto 306 systems were only pre-cleaned ultrasonically in
chromic acid solution.
Beside that, pre-sputtering process was carried out for sample prepared in
HVC. Pre-sputtering process is done to clean and equilibrate target surfaces prior to
film deposition (John and Werner, 1978).
Apart from that, the interaction of the deposition material with the crucible
giving rise to impurity atoms in the film deposited by e-beam method (Coombe,
1967). RF sputtering is one of the deposition methods that can avoid this kind of
problem because crucible is not required in RF sputtering. The deposition rate will
also affect the uniformity of sample. Thus, this indirectly contributes to the
difference of MR%, in which sample ST4 is more uniform compared to sample EA1.
This is due to the lower deposition rate for sample ST4, which may produce a sample
with higher uniformity (Milton, 1985). The deposition rate of sample ST4 and EA1
were 0.06 and 0.1 nm/s respectively.
RF sputtering method may produce a film with good adhesion to the substrate
compared with films deposited by e-beam method, which is easily loose. Thus, film
defects are easily formed in sample prepared by e-beam method. This may affect the
MR% that depends on the condition of film.
80
5
4.5
4
3.5
MR%
3
Deposited by
Sputtering method
2.5
Deposited by e-beam
method
2
1.5
1
0.5
-2
51
-2 0
02
-1 0
33
8
-3
80
57
7
15
08
20
60
25
30
0
Magnetic Fields, Gauss
Figure 4.30: Magnetoresistance curve of Co/Cu/Co (12 nm /2.5
nm/12 nm) prepared by RF sputtering and e-beam methods at two
different working pressures.
81
CHAPTER 5
CONCLUSION AND SUGGESTION
5.1
Conclusion
The objectives of this project were successfully achieved. The effect of
various parameters on GMR Co/Cu/Co sandwich structures have been carried out
and studied. The optimum conditions, which can produce the highest MR% were
determined.
The resistance of samples in this study to increase with the decrease of
magnetic fields towards zero. These results were clearly connected to the relative
arrangment of the magnetization in the magnetic layers. The magnetic moments in
two ferromagnetic layers show an antiparallel alignment between each other and
produced a maximum resistance at zero fields. However, the external magnetic field
applied will change the direction of magnetizations to parallel alignment, thus
generated lower resistance values when external magnetic fields were applied.
The Co/Cu/Co (6 nm/2.5 nm/6 nm) sandwich structure produces a maximum
MR% of 9.82% within the various thicknesses of Co layers. The MR% increases as
the thickness of Co layers increase from 2 nm to 6 nm. It is associated with the
increases of ferromagnetic spin in the ferromagnetic layers and the average
dimension of the Co particles. However, the increase of MR% is followed by a rapid
drop and a further gradual decrease with the increase in Co layer thickness up to 20
nm.
82
Apart from that, a small peak in MR% occurs between the working pressures
of (2.3 – 2.6) x 10-3 torr with a further gradual decreasing trend. MR% of sample
deposited at working pressure of 2.6 x 10-3 torr shows highest MR% compared with
other samples.
The study of MR% for various number of Co/Cu bilayers (n) were also
carried out to measure. It reveals that the MR% increases with n when n < 5. For
sample with n > 5, the MR% starts to drop and reaches 9.97% for (Co/Cu) x 8
bilayers. However, the MR% increases again when n > 8 and shows a maximum
value of 12.5% for (Co/Cu) x 15.
Co is one of the sample that shows anisotropy phenomenon. Thus, Co/Cu/Co
shows higher MR% when the magnetic field is applied along ( =0°) compare with
magnetic field applied perpendicular ( =90°) to the sample. This can be explained
by the magnetic moments in samples could turn over quickly as the applied field
changes in the easy axis direction ( =0°). On the other hand, it is difficult to
magnetize the sample when the magnetic field is applied in the hard axis direction (
=90°).
By depositing 8 nm of Cr as buffer layer to the Co/Cu/Co (5.5 nm/3.5 nm/5.5
nm) sandwich structures, the MR% increases from 1.57% to 5%. The Cr buffer layer
improves the flatness of the interfaces and induced good-quality structure of the
lower Co layer in Co/Cu/Co sandwich. It is also observed that, the MR% increased
to maximum value with the increasing of Cr layer thickness to 8 nm. When the Cr
layer is thicker than 8 nm, the MR% started to drop. The Cr/Co/Cu/Co also shows
anisotropy phenomenon. This is evidenced by the hexagonal crystal lattice in Co
structure.
The post-deposition annealing processes will modify the microstructure of the
deposited film. Thus, it will significantly affect the MR% of samples. MR% for
samples prepared by both RF sputtering and e-beam method increased after been
annealed in vacuum condition for a short period. The MR% tend to be constant for
the longer annealing time.
83
In addition, the MR% increases by a higher ammealing temperature. The
MR% reaches 24% at maximum temperature 400°C. This is due to the
recrystallization process of Co and Cu during the early stage (0-400°C) of annealing
However, it degrade to 20% and even lower when the annealing temperature was
higher than 400°C. Co and Cu species is weakly soluble to each other at higher
annealing temperature, i.e. >400°C, resulting the immediate drop of the MR% after
400°C.
Bilayers samples annealed at 400°C show an improvement in MR% compare
with sample before undergo the annealing process. The MR% of samples reached
32.43% when the (Co/Cu) x 15 was annealed at 400 °C. However, the mode of the
graph remains unchanged compared with graph bilayers of sample before annealing
process.
The sample prepared by RF sputtering method produces higher MR%
compared to the samples prepared by e-beam method. This is due to the different
deposition requirement that were needed in sputtering and e-beam method. Other
than that, it can produce higher purity, uniformity and good adhesion to the substrate.
As a final conclusion, the results were indicative of MR% dependency on
selected experimental parameters. A part from that, the highest MR% in this study is
obtained in the [Co (6 nm)/Cu (2.5 nm)] x 15 bilayers. This sample was deposited
on the corning glass at working pressure of 2.6 x 10-3 torr. Other than that, it also go
through the annealing process at 400 °C for 2 hour.
5.2
Suggestions
Generally, this research focused on the MR% of the sample. Thus, structural
characterization of the samples by X-ray diffraction (XRD) and scanning electron
microscopy (SEM) can be carried out as a further work.
84
In this study, the MR% was based on the magnetic multilayer. Thus, a study
of spin-valve structures, in which FM-layers composing of two types of material can
be done. In addition, other combination of sample such as Ni/Cu, Fe/ Cu and
NiFe/Cu may be carried to obtain different information in GMR study.
The highest annealing temperature in this study was 450°C for 2½ hour.
Therefore, it is suggested that a study of annealing effect at higher temperature, i.e.
>500°C and longer time, i.e. >2½ hour can be investigated further.
From the experimental results, the MR% increases with the increase of the
number bilayers of sample. For further works, samples with more bilayers can be
studied.
.
85
References
Anderson, J. C., Leaver, K. D., Rawlings, R. D., and Alexander, J. M. (1985).
Materials Science: 3th ed. England: Van Nostrand Reinhold (UK) Co. Ltd.
Anonynous, H. (1999). Irreversible magnetic field dependence of the magnetic
structure in GMR Co/Cu multilayers. Journal of Research of the National
Institute of Standards and Technology. 104:207-214.
Arkadiusz, G. (1970). Nonparametric Magnetic Amplifiers. Warszawa: Pitman
Publishing
Baibich, M. N., Broto, J. M.,Fert, A.,Nguyen F. V. D., and Petroff, F.(1988)
“Giant Magnetoresistance of (001)Fe/(001)Cr Magnetic Superlattices” Phys.
Rev. Lett. 61, 2472–2475.
Bakish, R. (1962). Introduction to Electron Beam Technology. London: John
Wiley & Sons, Inc.
Barthélémy A. and Fert A. (1991). Theory of the Magnetoresistance in Magnetic
Multilayers: Analytic Expressions from a Semiclassical Approach," Phys. Rev.
B43(13):124 –131.
Barna, J and Grunberg, P.(1992). Spin waves in exchange-coupled epitaxial
double-layers. Journal of Magnetism and Magnetic Materials. 82(2-3): 186189.
Bass, J. and Pratt, W. P. (2002). Version 7/21/01 current-perpendicular to plane
(CPP) magnetoresistance. Physica B: Condensed Matter. 321(1-4): 1-8
Behrisch, R. and Wittmaack, K. (1991). Sputtering by Particle Bombardment III.
Germany: Spring-Verlag.
86
Belozorov, D. P., Derkach, V. N., Nedukh, S. V., Ravlik, A.G, Roschenko, S. T.,
Shipkova, I.G., Tarapov, S. I., Yildiz, F., and Aktas, B. (2003). Magnetization
and impedance measurements of multiplayer Co/Cu structures in millimeter
waveband. Journal of Magnetism and Magnetic Materials. 263: 315-323.
Boylestad, R. L. (2003). Introductory Circuit Analysis. Tenth Edition. United
State of America: Prince Hall.
Butler W. H., Zhang X.-G., Nicholson D. M. C., and Maclaren J. M. (1995). First
Principles Calculation of Electrical Conductivity and Giant Magnetoresistance
of Co-Cu Multilayers. Phys. Rev. B 52(13): 399-410
Camley R. E. and Barna, J. (1989). Theory of Giant Magnetoresistance Effects in
Magnetic Layered Structures with Antiferromagnetic Coupling. Phys. Rev.
Lett. 63:664-667
Canali, L. (2000). Novel Scanning Probes Applied to the Study of Nanostructures.
Technische Universiteit Delft: Ph D. Thesis.
Chapman, R. M. (1963). The effect of fabrication variables on chromium thin film
resistors. Vacuum .13(6): 213-221.
Colis, S., Guth, M., Arabski, J., Dinia, A., and Muller, D. (2002). Thermal
stability of spin valve sensors using artificial Co/Ir based ferrimagnets.
Journal of Magnetism and Magnetic Materials. 240:186-188.
Coombe, R. A. (1967). The electrical properties and applications of thin film.
London: Sir Isaac Pitman and Sons Ltd.
Crew, C. D., Kim, J., Lewis, L. H., and Barmak, K. (2001). Interdiffusion in
bilayer CoPt/Co films: potential for tailoring the magnetic exchange spring.
Journal of Magnetism and Magnetic Materials. 233(3): 257-273
87
David, J. (1991). Introduction to Magnetism and Magnetic Materials. London:
Chapman and Hall.
Dinia A. Persat, N., Coils, S., Ulhaq-Bouillet C., and van den Berg, H. A. M.
(2000). Effect of the structural quality of the buffer on the magnetoresistance
and the exchange coupling in sputtered Co/ Cu sandwiches. Eur. Phys. J. B.
18: 413-419.
Donald, L. S. (1995) Thin Film Deposition. London: McGraw-Hill, Inc.
Edwards (1992). Instruction Manual (FTM5 Film Thickness Monitor). UK:
Edwards High Vacuum International.
Errahmani, H., Berrada, A., Schmerber, G., and Dinia, A. (2001) Comparative
study between the effect of annealing and substrate temperature on the
magnetic and transport properties of Co20Cu80 granular alloys. Materials
Letters. 51:48-55
Fert, A., Campbell, I. A. (1968). Two-current conduction in Nickel. Phys. Rev
Lett. 21:1190.Evolution from multilayer to granular behavior via Cobalt layers
fragmentation in Co/Cu multilayers. Journal of Magnetism and Magnetic
Materials. 262(1): 7-11.
George, J. (1992). Preparation of Thin Films. New York: Marcel Dekker, Inc
Higashihara, S., Oomi, G., Suenaga, K., Ono, T., and Shinjo, T. (2004). Effect
of pressure on the giant magnetoresistance in Fe/Cr multilayers on
SrTiO3(100) substrate. Physica B: Condensed Matter. 57:4-16.
Hecker, M., Thomas, J., Tietjen, D., Baunack, S., Schneider, C. M., Qiu A.,
Cramer, N., Camley R. E. and Celinski, Z. (2003). Thermally induced
modification of GMR in Co/Cu multilayers: correlation among structural,
transport, and magnetic properties. J. Phys. D: Apply. Phys. 36: 564-572.
88
Jansen, F. (1997). Handbook of Thin Film Process Technology. London: IOP
Publishing Ltd
Jiang, J. S. and Bader, S. D. (2002). Magnetic reversal in thin film exchangespring magnets. Scripta Materialia. 47(8): 563-568
John, L. V., and Werner, K. ed. (1978). Thin Film Processes. London: Academic
Press.
John, R T. (1981). An Introduction to Error Analysis: The Study of Uncertainities
in Physical Measurements. United State of America: Oxford University Press.
Johnson, B. L., and Camley R. E., (1991). Theory of Giant Magnetoresistance in
Fe/Cr Multilayers: Spin-Dependent Scattering from Impurities. Phys. Rev. B
44: 9997-10002.
.
Kakashi, S, and Osamu, N. (1998). Structure and giant magnetoresistance of
Co/Ag granular alloy film fabricated by a multilayering method. Journal of
Thin Solid Films. 334(1-2): 206-208
Kiwi, M. (2001). Exchange Bias Theory J. Magnetism and Magnetic Materials
234(3): 584-595.
Kumar, D., Narayan, J., Kvit, A. V., Sharma, A.K, and Sankar, J. (2001)
High coercivity and superparamagnetic behaviour of nanocrystalline iron
particles in alumina matrix. Journal of Magnetism and Magnetic Materials.
233(3):161-167
Liu, J. M., Yang, Y., Zhou, X. H., Chen, X. Y., and Liu, Z. G., (2003). Magnetic
polaran mechanism of electron transport and magnetoresistance in spin
systems: a Monte-Carlo simulation. Materials Science and Engineering B.
99:558-562.
Loretto, M. H. (1984). Electron Beam Analysis Of Materials. London: Chapman
89
and Hall.
Luthi, B. and Hock, R. (1983). Dipolar surface spin waves in antiferromagnets.
Journal of Magnetism and Magnetic Materials. 38(3): 264-2.68
Mahdi A. E., Panina, L. and Mapps, D. (2003) Some new horizons in magnetic
sensing: high- Tc SQUIDs, GMR and GMI materials. Sensors and Actuators
A. 105:271-285.
Maissel, L. I. and Glang, R. (1970). Handbook of Thin Film Technology. New
York: McGraw Hill
Masashige, S., Hideyuki, K, and Kazuo, K. (1998). Giant Magnetoresistance in
Ni-Fe/Co/Al-AlOx/Co/Ni-Fe/Fe-Mn Ferromagnetic Tunnel Junctions. Journal
of FUJITSU Sci. Tech. 34(2):204-211
Matteo, A. Angelo, R., and Maria, G. P. (2000). Exchange-spring behavior of
hard/soft magnetic multilayers: optimization study of the nanostructure.
Physica B:Condensed Matter.275(1-3): 120-123.
Milton, O. (1985) The materials science of thin films. United States of America:
Academic Press, Inc
Philip, J. (2000). Quantum transport in disordered magnetoresistive systems.
Rijks Universiti Groningen: Ph.D. Thesis.
Ping, H. (2001). Studying and Modelling of the Giant Magnetoresistance Sensors.
University of Manchester: Ph.D. Thesis
Ratzke, K., Hall, M J, Jardine, D. B., Shih, WC, Somekh, R E and Greer, A. L.
(1999) Evolution of microstructure and magnetoresistance in Co/Cu
multilayers during annealing. Journal of Magnetism and Magnetic Materials.
204(1-2):61-67
90
Rohrer, H., Derighetti, B. and Gerber, Ch. (2002). A spin-flop bicritical line in
GdAlO3. Physica B+C.86-88 (2): 597-598.
Samsudi Sakrani.(1996). Pencirian Saput Tipis. Kursus Teknologi Bahan.
Universiti Teknologi Malaysia. 29-30 Oktorber 1996.
Sato, M., Kikuchi, H. and Kobayashi, K. (1998)." Giant Magnetoresistance in NiFe/Co/Al-AlOx/Co/Ni-Fe/Fe-Mn Ferromagnetic tunnel Junctions". J. Sci.
Tech.34(2): 204-210.
Scroder, D. K., (1990). Semiconductor material and device characterization. New
York: A. Wiley-Interscience Publication, John Wiley & Sons Inc.
Siegle, G. (1972). MRV Metallpraxis/Oberflachentechnik. London: Academic
Press, Inc. 247
Shen, H L, Li, G. X., Shen, Q. W., Li, T., and Zou, S. C. (2000) Giant
magnetoresistance and structural properties in Co/Cu/Co sandwiches with Si
and Cr buffer layers. Thin Solid Films. 375(1-2):55-58
Smadar, S. and Nathan, W. (2001). Magnetoresistance of magnetic multilayers:
understanding Ohm’s law. Journal of Physica A: Structural Mechanics and its
Applications. 302(1-4): 382-390.
Smith, N., Zelster, A.M., Parker, M.R. (1996). GMR multilayers and head design
for ultrahigh density magnetic recording. IEEE Trans. Magn. 32: 135.
Tagirov, L. R., Vodopyanov, B. P, and Efetov, K. B. (2001). Ballistic versus
diffusive magnetoresistance of magnetic point contact. Phys. Rev. B. 63:
10442.
Takahasgi, D., Miura, S., Tsunoda, M., and Takahashi, M. (2002). Enhanced
91
lateral grain growth and enlarged giant magnetoresistance in Co/Cu
multiplayer by Fe-Si buffer layer. Journal of Magnetism and Magnetic
Materials. 239:282-284.
Takashi, S., and Osamu, N. (1998). Structure and giant magnetoresistance of
Co/Ag granular alloy film fabricated by a multilayering method. Thin Solid
Films. 334(1-2): 206-208
Timothy, R. C. (2001). Growth, Characterization, and Properties of Co/ Re
Superlattices. West Virgina University: Ph.D. Thesis.
Turilli, G. Pareti, L. and Castaldi. (1999). Effects of layering and working
pressure on magnetic and magnetotransport properties of as sputtered
multilayered granular Co/Cu films. Journal of Superlattices and
Microstructures. 25(4): 591-600
Valet, T. and Fert, A.(1993). Theory of the Perpendicular Magnetoresistance in
Magnetic Multilayers. Phys. Rev. B. 48:7099-7113.
Valdes, L. B. (1954).Resistivity measurements on germanium for transistors.
Proc. IRE. 42: 420-427.
Vavassori, P. Spizzo, F., Abgeli, E., Bisero, D., and Ronconi, F. (2003). Evalution
from multiplayer to granular behavior via cobalt layers fragmentation in
Co/Cu multilayers. Journal of Magnetism and Magnetic Material. 262(1):
120-123.
Veeco Metrology Group (1998). Installation, operation and Maintenance Manual
of DEKTAK 3 surface Profile Measuring System. California: Veeco.
Vieux-Rochaz, L., Cuchet, R., and Vaudaine, M. H. (2000) A new GMR sensor
based on NiFe/Ag multilayers. Sensors and Actuators. 81:53-56.
Vohl, V., Wolf, J. A., and Grunberg, P. (1991). Exchange coupling of
92
ferromagnetic layers across nonmagnetic interlayers. Journal of Magnetism
and Magnetic Materials. 93: 403-406
Vossen, J. L., and Kern, W. (1991) Thin Film Process II. London: Academic
Press, Inc.
Wang, W. D., Zhu, F., Lai, W. et al. (1999). Microstructure, magnetic properties
and giant magnetoresistance of granular Cu-Co alloy. J. Phys. D: Apply Phys.
15:1990-1996.
Wenner, F. (1915) A method of measuring earth resistivity. Bulletin of the Bureau
of Standards. 12: 469-478.
Wieder, H. H. (1979). Four terminal nondestructive electrical and
galvanomagnetic measurements. New York: Plenum Press
Xu, M., Fan, Y., Luo, G., and Mai, Z. (2000). Dependence of giant
magnetoresistance on the thickness of magnetic and non-magnetic layers in
spin-valve sandwiches. Physics Letters A, 272:282-288.
Yamada, A., Houga, T., and Ueda, Y. (2002). Magnetism and magnetoresistance
of Co/Cu multiplayer films produced by pulse control electrodeposition
method. Journal of Magnetism and Magnetic Materials. 239:272-275.
Yamamoto, H., Motomura, Y., Anno, T., Shinjo, T. (1993).Magnetoresistance of
non-coupled NiFe/Cu/Co/Cu multilayers. Journal of Magnetism and
Magnetic. Materials. 126: 437.
Yang, W. W., Sun, H, Zhu, X.Y., and Song, W. J. (2001) A phenomenological
theory of giant magnetoresistance in magnetic granular alloys. Physica B.
299:77-82.
Yu, R. H., Zhang, X. X., Tejada, J., Knobel M., Tiberto P. and Allia, P. (1995).
93
Magnetic properties and giant magnetoresistance in magnetic granular
CoxCu100-x alloys. Phys. D: Appl. Phys. 28:1770-1777.
Zhang, Z., Zheng, W., Zhang, G., Wang, A., X, L., Chen, J., and Ge, S. (1999)
Effect of preferred orientation on GMR and saturation fields in Co/Cu
multilayers. Journal of Magnetism and Magnetic Materials. 198-199: 49-51.
94
Appendix A
Temperature Uncertainty Calibration
x2
xy
(y-c-mx)2
205
40000
41000
12.6025
250
246
62500
61500
30.14
300
303
90000
90900
2.1609
350
354
122500
123900
5.9049
400
396
160000
158400
31.4721
425
428
180625
181900
1.8769
450
454
202500
204300
5.5225
2375
2386
858125
861900
89.6798
x (Heater Set
y (Quartz
Point)
Temperature)
200
∑
∆ = N ∑ xi − (∑ xi )
2
2
= 7(858125) − (2375)
2
= 6006875 − 5640625
= 366250
m=
1⎡
⎤
n∑ xi y i − ∑ xi ∑ y i ⎥
⎢
∆⎣ i
i
⎦
1
[7(861900) − (2375)(2386)]
366250
1
=
(6033300 − 5666750)
366250
= 1.0008
=
c=
1
∆
[∑ x ∑ y − ∑ x ∑ x y ]
2
i
i
i
i
i
1
[(858125)(2386) − (2375)(861900]
366250
1
=
2047.48625 × 10 6 − 2047.0125 × 10 6
366250
= 1.29
=
[(
) (
)]
95
1
( yi − c − mxi )2
∑
N −2
1
(89.6798)
=
7−2
= 17.93596
σ2 =
σm = N
σ2
∆
= 7
(17.93596)
= 0.0185 ≈ 0.02
366250
96
Publications
1.
Lau Yee Chen, Samsudi Sakrani, Rashdi Shah Ahmad and Yussof Wahab.
(2003). Modeling of giant magnetoresistance in ultra thin metallic films.
Proceedings: IIS Symposium on Fundamental Science Research, 20-21
May 2003, Johor Bharu.
2.
Lau Yee Chen, Rashdi Shah Ahmad, Samsudi Sakrani, and Yussof Wahab.
(2003). Giant magnetoresistance in Co/Cu/Co nanostructures.
Proceedings: International Conference on Advancement Science and
Technology (iCAST), 5-7 August 2003, Kuala Lumpur.
3.
Lau Yee Chen, Rashdi Shah Ahmad, and Samsudi Sakrani. (2003).
Annealing effects on Co/Cu/Co nanostructures. Proceeding: National
Conference on Physics 2003 (Perfik 2003), 15-17 August 2003,Fraser Hill.
4.
Lau Yee Chen, Rashdi Shah Ahmad, Samsudi Sakrani, and Yussof Wahab.
(2003). Thickness dependent of giant magnetoresistance in Co/Cu/Co
nanostructures. Proceeding: Conference on Public Institutions Of Higher
Learning ( IPTA ), 2-5 October 2003, UPM Kuala Lumpur.
5.
Lau Yee Chen, Rashdi Shah Ahmad, Samsudi Sakrani, and Yussof Wahab.
(2003). Temperature dependence of GMR in Co/Cu nanostructures.
Proceeding: The XX Regional Confrence On Solid State Science And
Technology, 12-14 December 2003, UPSI Perak.
6.
Samsudi Sakrani, Lau Yee Chen, Yussof Wahab, and Rashdi Shah Ahmad.
(2004). GMR Dependence of Film Thickness, Annealing Temperature,
Multilayer and Gas Rate in Co/ Cu/ Co Nanostructures. Acceptance
Paper: 2004 E-MRS FALL MEETING, September 6-10, Warsaw
University of Technology, Poland.